International Handbook on Industrial Policy

INTERNATIONAL HANDBOOK ON INDUSTRIAL POLICY

International Handbook on Industrial Policy

Edited by

Patrizio Bianchi Professor of Applied Industrial Economics and Policy and Rector, University of Ferrara, Italy

Sandrine Labory Lecturer of Applied Industrial Economics and Policy, University of Ferrara, Italy

Edward Elgar Cheltenham, UK • Northampton, MA, USA

© Patrizio Bianchi and Sandrine Labory 2006 All rights reserved. No part of this publication may be reproduced, stored in a retrieval system or transmitted in any form or by any means, electronic, mechanical or photocopying, recording, or otherwise without the prior permission of the publisher. Published by Edward Elgar Publishing Limited Glensanda House Montpellier Parade Cheltenham Glos GL50 1UA UK Edward Elgar Publishing, Inc. 136 West Street Suite 202 Northampton Massachusetts 01060 USA

A catalogue record for this book is available from the British Library

Library of Congress Cataloguing in Publication Data Bianchi, Patrizio, 1952– International handbook on industrial policy / Patrizio Bianchi, Sandrine Labory. p. cm. Includes bibliographical references and index. Contents: Contents: Industrial policy in a global perspective — Competition and regulatory issues — Research, technology and innovation — Territory, industrial development and small firms — The use of game theory in industrial policy analysis – Particular views. 1. Industrial policy. 2. Globalization. I. Labory, Sandrine. II. Title. HD3611.B533 338.09—dc22

2006 2005058952

ISBN-13: 978 1 84376 836 4 (cased) ISBN-10: 1 84376 836 4 (cased) Printed and Bound in Great Britain by MPG Books Ltd, Bodmin, Cornwall

Contents vii viii

Acknowledgments List of contributors Introduction Patrizio Bianchi and Sandrine Labory PART I

xv

INDUSTRIAL POLICY IN A GLOBAL PERSPECTIVE

1. From ‘old’ industrial policy to ‘new’ industrial development policies Patrizio Bianchi and Sandrine Labory

3

2. Emerging issues in the new economy and globalization Ashish Arora and Alfonso Gambardella

28

3. European industrial policy Jacques Pelkmans

45

4. Industrial policy in developing countries: what can we learn from East Asia? Sanjaya Lall

79

5. Do informational service activities translate into new industrial policy requirements? Jacques De Bandt

98

PART II

COMPETITION AND REGULATORY ISSUES

6. Competition policy and innovation Jochen Lorentzen and Peter Møllgaard

115

7. State aid to business Stephen Martin and Paola Valbonesi

134

8. Mergers and concentration policy Hans Schenk

153

9. Liberalization and regulation of public utility sectors: theories and practice Tom Björkroth, Sonja Grönblom and Johan Willner 10. State-owned enterprises, privatization and industrial policy Andrea Goldstein

v

180 198

vi

International handbook on industrial policy

PART III

RESEARCH, TECHNOLOGY AND INNOVATION

11. Intellectual property rights and market dynamics Fabrizio Cesaroni and Paola Giuri 12. New approaches to intellectual property: from open software to knowledge-based industrial activities Nicolas Jullien and Jean-Benoît Zimmermann

217

243

13. Science-based industries and spin-offs Marco Giarratana and Salvatore Torrisi

265

14. Science parks and high-tech clustering Jan Annerstedt

279

15. Technology policy and social policy: how industrial policy applies to health Daniele Paci and Stuart O. Schweitzer

298

PART IV TERRITORY, INDUSTRIAL DEVELOPMENT AND SMALL FIRMS 16. Decentralizing industrial policies: threat or opportunity in developing countries? Leandro Sepulveda and Ash Amin

321

17. The local dimensions of industrial policy Marco Bellandi and Marco Di Tommaso

342

18. Business support policies Nicola Bellini

362

19. Small and medium-sized enterprise policies in Europe, Latin America and Asia Patrizio Bianchi, Sandrine Labory, Daniele Paci and Mario Davide Parrilli

380

PART V THE USE OF GAME THEORY IN INDUSTRIAL POLICY ANALYSIS 20. The genesis of game theory and its role in industrial economics Luca Lambertini

405

21. Differential game-theoretical analysis and industrial policy Roberto Cellini and Luca Lambertini

415

PART VI

PARTICULAR VIEWS

22. Industrial policy: perspectives, experience, issues Christos N. Pitelis

435

Index

451

Acknowledgments We are grateful to all the contributors to the Handbook but a particular expression of our gratitude goes to Marco Bellandi, Jacques De Bandt, Alfonso Gambardella, Paola Giuri, Andrea Goldstein, Sanjaya Lall, Jacques Pelkmans, Hans Schenk and Johan Willner, who provided comments and suggestions at various stages of the Handbook and/or read and commented on chapters from other contributors. David Ulph also provided useful suggestions for the Handbook at its earliest stages.

We dedicate this book to Sanjaya Lall, who unfortunately died in June 2005.

vii

Contributors Ash Amin is a professor and former Head of Geography at Durham University. His latest books include Cities: Reimagining the Urban, Polity Press, 2002 (with Nigel Thrift); Architectures of Knowledge, Oxford University Press, 2004 (with Patrick Cohendet); Placing the Social Economy, Routledge, London, 2002 (with Angus Cameron and Ray Hudson); and the Blackwell Cultural Economy Reader, Blackwell, 2004 (with Nigel Thrift). Jan Annerstedt is the co-founder of Interlace-Invent, an international consultancy focused on the design and management of innovation environments such as science parks, corporate R&D units and academic centres. Annerstedt holds the UNESCO Chair at the Copenhagen Business School, where he is a specialist in international economics and business strategy. He has been an advisor to various multinational companies and to regional and national governments and a consultant to international organizations such as the OECD, the European Commission, UNIDO, UNDP, UNCTAD and UNESCO. Ashish Arora has a PhD in Economics from Stanford University (1992). He is a Professor of Economics and Public Policy at Carnegie Mellon University, Pittsburgh, and co-director of the Software Industry Center at Carnegie Mellon University. Arora’s research focuses on the economics of technological change, intellectual property rights, technology licensing and international technology transfer, and he has published extensively on the growth and development of biotechnology and the chemical industry. More details are available at HYPERLINK (http://www.softwarecenter.cmu.edu/Arora.htm). Marco Bellandi is Doctor of Research in Political Economy (1987) and Full Professor (straordinario) of Applied Economics (2005) at the University of Florence (Faculty of Economics), where he teaches Political Economy, and Industrial Economics and Policy. His research builds on Marshall’s contributions, and on contemporary models of analysis and empirical investigation of industrial organization, with particular regard to external and internal economies, industrial districts, processes of innovation and change in local production systems. Nicola Bellini is Professor of Business Economics at the Sant’Anna School of Advanced Studies in Pisa. He has been an advisor for economic policy and planning at the Office of the President of the Emilia Romagna region in Italy and Research Fellow at the Nomisma Economic Research Institute in Bologna. He is the author of several books and articles on industrial and innovation policy issues, with special reference to local and regional development and innovation policies, the management of regional development agencies, the management and marketing of business support services and area marketing. Patrizio Bianchi is Professor of Applied Industrial Economics and Policy at the Faculty of Economics of the University of Ferrara. He has wide practical experience in policy decision making and implementation in the fields of industrial policy, development viii

Contributors

ix

policy and privatization. He has advised various regional and national governments and also international organizations such as UNIDO and the European Commission. He has been Dean of the Faculty of Economics and is now Rector of the University of Ferrara. Tom Björkroth obtained his doctoral degree from Åbo Akademi University in 2003 and is now Senior Research Fellow at the Institute for Competition Policy Studies, Turku School of Economics and Business Administration. His research interests include empirical industrial organization, infrastructure industries and, in particular, telecommunications and competition analysis and policy. Roberto Cellini (PhD in Economics at the University of Bologna) is Professor of Economics at the University of Catania (Italy). The main fields of his research are the theoretical and econometric problems of economic growth and game theory, where he focuses on differential games with reference to industrial organization issues. His publications include articles in leading economics journals, such as the Journal of Economic Theory and the Journal of Economic Dynamics and Control. Fabrizio Cesaroni is an assistant professor at the University of Lecce and research associate at Sant’Anna School of Advanced Studies (Pisa, Italy), where he obtained his PhD with a thesis on ‘Technology strategies in the knowledge economy’. His main research interests are in the economics and management of technology licensing and technology transfer, both by for-profit firms and by non-profit research organizations, such as universities and public research laboratories. Jacques De Bandt, with a Doctorate in Law and a PhD in Economics, is First Class Director of Research (Emeritus) at the French National Center for Scientific Research, and Professor at the University of Sophia Antipolis. His field of specialization is industrial economics, with special emphasis on business services. He is a founder and editor of the Revue d’Economie Industrielle. He is Chairman of AMARES and editor of FAIRE SAVOIRS (Inter-disciplinary scientific Association and Journal in Human and Social sciences). Marco Di Tommaso (PhD University of Naples and MSc University of Manchester) is Professor of Industrial and Development Policy at the University of Ferrara, Italy. In the field of industrial policy he has worked with international institutions (UNIDO, UNDP, CEPAL), local and national governments (Italy, Syria, China) and universities (UCLA, Birmingham, South China University of Technology). His research ranges from hightech industry to SMEs, clusters and developing countries’ industrialization policies. His latest book has been published by Edward Elgar (2005): Health Policy and High-Tech Industrial Development (with S.O. Schweitzer). Alfonso Gambardella (PhD, Stanford, 1991) is Professor of Economics and Management at the Università Commerciale ‘L. Bocconi’, Milan, Italy. His research interests are in the economics and management of technology and innovation. He has also taught courses on the economics of technology at the Department of Economics of Stanford University. His website is www.alfonsogambardella.it.

x

International handbook on industrial policy

Marco Giarratana is Assistant Professor of Management and Strategy at the Department of Business Administration of the Carlos III University in Madrid. He holds a PhD in Economics and Management from the Sant’Anna School of Advanced Studies in Pisa and a Laurea degree in Economics from the Bocconi University of Milan (Italy). His research interests centre on industrial dynamics, entrepreneurship, firm strategy and innovation. His works have been published both in international journals and in edited book chapters. Paola Giuri is Assistant Professor in Industrial Economics and Management at Sant’Anna School of Advanced Studies in Pisa. She obtained her PhD at Sant’Anna School, and spent two years as a visiting study fellow at SPRU, University of Sussex within the EC Marie Curie Programme. Her research interests are in the economics of innovation and patenting, entrepreneurship and innovation in different regimes of IPRs (including open source software) and large and small firms’ technology strategies. Andrea Goldstein is a senior economist at the OECD Development Centre. He was previously at the World Bank Group (Foreign Investment Advisory Services), the OECD Economics Department and Consob, and the Italian Securities and Exchange Commission, and has also been a consultant for the Inter-American Development Bank and the UK Department for International Development. His current research interests include multinationals from emerging, transition and developing countries, regulatory reform in network industries, the political economy of the global aerospace industry and high-skill migration from Latin America. Sonja Grönblom is a research associate at Åbo Akademi University, specializing in industrial organization, in particular applications to public sector reform and non-profit organizations. Nicolas Jullien defended his PhD on the economy of free software in 2001. He is today in charge of coordinating a research group on the uses of IT applications in Brittany (France), called Marsouin (http://www.marsouin.org). He also manages the European research project Calibre on free software for GET, one of the participants in the project (further details available at HYPERLINK http://www.calibre.ie). Sandrine Labory holds a PhD from the European University Institute of Florence (Italy) and is a lecturer of Applied Industrial Economics and Policy at the University of Ferrara. She has practical experience of European economic policies, having worked in various projects for European institutions. Her research interests include European policies and regulation, the organization of industrial production and industrial policy. The late Sanjaya Lall was Professor of Development Economics at Oxford University’s Department of International Development. He was chief editor of Oxford Development Studies, a former staff member of the World Bank and a consultant to several international organizations on technology, foreign investment and industrialization policies.

Contributors

xi

Luca Lambertini (DPhil, Oxon) is Professor of Economics at the University of Bologna, and fellow of ENCORE, Faculty of Economics & Econometrics, University of Amsterdam. He has been a research fellow at the Institute of Economics and the Centre for Industrial Economics, University of Copenhagen. His research ranges from oligopoly theory to international trade and differential games. He has published in leading scientific journals. Jochen Lorentzen is Chief Research Specialist in the Human Resources Development Programme of the Human Sciences Research Council (HSRC), South Africa’s statutory research council for the applied social sciences. He has taught at universities and business schools throughout Europe and in the USA and South Africa. His principal area of research is technological learning and innovation in latecomer countries. Stephen Martin is Professor of Economics at Purdue University. His PhD is from the Massachusetts Institute of Technology. He has worked at Michigan State University, the European University Institute, the University of Copenhagen and the University of Amsterdam. He has been managing editor of the International Journal of Industrial Organization and President of the Industrial Organization Society. Peter Møllgaard is Professor of Law & Economics and Head of the Department of Economics at the Copenhagen Business School. He is associated with the Centre for Industrial Economics and with the Centre for Competition Policy at the University of East Anglia. His principal area of research is the application of industrial organization to competition policy. He has done work on horizontal and vertical coordination as well as technology transfer. Daniele Paci is a researcher at the University of Ferrara. He collaborates with scholars of the University of California Los Angeles and is a lecturer for L’Institute Graduate School in Industrial Development Policy and for the Master in Economics and Management of Health Services at the University of Ferrara. His research interests include high-tech clusters, intangible assets, patents and innovation in the biotech and pharmaceutical industry and industrial policy. Mario Davide Parrilli holds a PhD from Birmingham University and an MPhil of the IDS of Sussex University. He is a lecturer at the University of Birmingham. He is also co-ordinator of the Central American Graduate School in Industrial Development and SME Policies, Managua, Nicaragua and San José, Costa Rica. He has been a lecturer at the University of Ferrara and Scientific Co-ordinator of a Master’s in ‘International co-operation and social economy’ at the Centre for Didactic Innovation and Distance Education (CARID) at the same university. Jacques Pelkmans holds the Jan Tinbergen Chair of European Economic Integration at the College of Europe, Bruges, where he is Director of the Economic Studies department. He also is Council member of the WRR, a think-tank on public policy analysis attached to the Dutch Prime Minister’s Office in The Hague. Between 1990 and 2001, he was Senior Research Fellow at CEPS in Brussels and Professor of Economics at Maastricht

xii

International handbook on industrial policy

University. His current research interests include liberalization and regulation of network industries in the single market, ‘better’ (that is, more efficient and effective) EU regulation, and the development of a functional EU subsidiarity test. Christos N. Pitelis is Director of the Centre for International Business and Management (CIBAM) and Reader in International Business and Competitiveness at the Judge Institute of Management, University of Cambridge. He has realized projects for governments, the European Commission, the United Nations, the Commonwealth Secretariat and the private sector. He is an honorary research fellow at the Birmingham Business School, fellow of L’Institute, Associate Professor at the University of Athens, fellow of the Athens Laboratory of Economic Policy Studies, and a non-executive director of the Medical Group of Companies, Athens. Hans Schenk is founding professor of organizational economics and academic chair of the Tjalling C. Koopmans Research Institute at Utrecht University’s School of Economics. He graduated from the University of Oregon, obtained his MSc and MBA at the University of Louvain and his doctorate in economics at the Université de Nice–Sophia Antipolis. He was previously a professor at Tilburg University, Erasmus University Rotterdam and the University of Groningen and held visiting positions in China, France and the UK. He has acted as consultant to Fortune 500 firms, the United Nations, the European Commission and various governments. Stuart O. Schweitzer is Professor of Health Services at the UCLA School of Public Health. Additionally, he is Visiting Professor of Economics at the University of Ferrara. His areas of teaching and research are health and pharmaceutical economics and industrial policy towards health. The second edition of his book, Pharmaceutical Economics and Policy, has been published by Oxford University Press. Leandro Sepulveda is a research fellow at Middlesex University (School of Health and Social Sciences) in the UK. His current research interests are industrial policy and SMEs, local and regional development, informal economic activity, and entrepreneurship amongst ethnic minorities. Prior to this he worked for the University of General Sarmiento (1997–2004) and ECLAC UN in Argentina. His work has largely been commissioned by public and private agencies at local and national levels. He holds a PhD from Durham University, UK. Salvatore Torrisi obtained his PhD from the Science Policy Research Unit (SPRU), University of Sussex, Brighton, UK and his Laurea degree in Economics from Bocconi University Milan, Italy. He is Associate Professor of Economics and Management at the University of Bologna. He teaches in the European Doctoral Training Programme, ‘Industrial Dynamics and Firm Behaviour – Empirical Aspects’ (ETIC) at MERIT, University of Maastricht. His main research interests are in the economics of innovation, the impact of new technologies on firms’ productivity and the dynamics of high-tech industries. Paola Valbonesi (PhD in Economics, 1995, European University Institute, Florence) has been Associate Professor of Economics at the Department of Economics, Padua

Contributors

xiii

University, since 2002. She has held visiting positions at the Hungarian Academy of Science (1991), the University of York (1997) and the Centre for Industrial Economics at the University of Copenhagen (1998). Her research interests include various topics in industrial organization, focusing on competition policy and regulation of public utility. Johan Willner obtained his doctoral degree at Åbo Akademi University, where he has been Professor of Economics since 1988. His research interests include theoretical and applied industrial organization, and to some extent public economics and corporate governance. He is one of the founding members of the European Network on Industrial Policy (EUNIP). Jean-Benoît Zimmermann is CNRS Research Director at GREQAM (Groupement de Recherche en Economie Quantitative d’Aix-Marseille) and IDEP (Institut d’Economie Publique) in Marseilles. His main topics of interest are economics of interactions and networks, proximity dynamics, innovation and intellectual property economics (Open Source Software). More details are available at HYPERLINK (http://www.vcharite.univ-mrs.fr/ greqam/index_fr.php).

Introduction Patrizio Bianchi and Sandrine Labory

Putting together a Handbook on Industrial Policy appears challenging at first sight. Both the reflections within industrial economics and the practical experiences over time and across the world point to a richness of contexts, approaches, factors and determinants that make it difficult to define the problems and delineate the fields of application. Hence we chose to start from empirical observation to arrive at general properties and systematic characteristics that theoretical perspectives should take into account and explain (an approach that classical economists took, and that the pioneers of industrial economics such as Bain also adopted). The Handbook has been put together with the aim of providing an overview of all the approaches and relevant themes of the ample field of industrial policy. Our objective was not to promote a particular approach or school of thought. We think it is a task of the reader to derive the most relevant questions and the most relevant approaches from the richness of perspectives proposed in the Handbook. All chapters are written by specialists with deep knowledge of particular sectors, regions or policy fields that they have observed empirically. This make it possible to point to relevant current policy issues and derive implications for theory and methodology. The meaning of the term ‘industrial policy’ has changed a lot over time. Until the 1980s, the term meant the direct intervention of the state in the economy, the direct control by the government of large parts of the production apparatus and a set of public actions aimed at limiting the extent of the market and at conditioning productive organization. Nowadays, the term ‘industrial policy’ indicates instead a variety of policies which are implemented by various institutional subjects in order to stimulate firm creation, to favour their agglomeration and promote innovation and competitive development in the context of an open economy. The new industrial policies are therefore mainly industrial development policies, where industry is implicitly considered as the organization, and the strategic management of human competencies and technical capacities. Industry in this sense is the main driver of what we may call the Wealth of Nations, meaning the capacity to produce and distribute wealth within a community. Some important changes have indeed occurred in the economy and have generated the need for firms to adapt and restructure. These changes are manifold, ranging from purely economic aspects (increase in world trade, improvements in financial markets), to technological ones (diffusion of the information and communication technologies) and political ones (end of the division of the world into two political blocks). All these changes are often confused or only partly considered and we think they all constitute what everyone tends to call ‘globalization’. Globalization, understood in this sense, has led to the necessity for firms to undertake dramatic structural changes. Very often firms have called on governments for help: to better understand the changes going on or to be directly supported. As a result, the new industrial development policies have been defined within the constraints imposed by xv

xvi

International handbook on industrial policy

international agreements such as the WTO rules or the European Union legislative framework for European states. Industrial development policies are all policies aimed at favouring and accompanying structural change. They are dynamic in nature and primarily consist of programmes that evolve over time according to the evolutions of the economy and its context and according to the realizations achieved. It is possible to define two broad sets of policies. First, there are policies aimed at defining the rules of the competitive or competitiveness game. These are anti-trust, property right legislation, international rules on world trade, and so on. Second, there are policies aimed at providing or improving firms’ and individuals’ capabilities to take part in the competitive game. These are technology and research policy, training and education, policies favouring linkages between firms, universities and local or regional authorities, and so on. We have therefore articulated this Handbook around these themes, defining four axes that also fit the information that emerges out of not only the economic literature but also the documents published by the European communities and by other international institutions. These four axes constitute the four main Parts of this Handbook. First, there is an important part that we have to call ‘market opening’ that is centred on the actions aimed at integrating capital and product markets so as to increase the extent of the market (Chapter 3). It also includes actions aimed at helping agents to take part in the global competition game (Chapter 4) and actions aimed at helping to exploit the opportunities of the global game (Chapters 2 and 5). Hence these actions require an analysis of the changes in worldwide production processes, that is performed by the authors of Part I of the volume. Thus Arora, Gambardella, Lall and De Bandt show, among other things, the emergence of global-scale production systems where firms from developing countries may develop capabilities that are complementary to the capabilities of firms in developed countries. The second axis which we think is emerging is that of the ‘rules’, that is, market and competition guarantee, that includes the redefinition of the role of anti-trust authorities (Chapters 6, 7 and 8), the actions of privatization of public services and the regulation of the risks of creation of new monopolies (Chapters 9 and 10). This axis is therefore one concerned with ensuring that markets function efficiently and that competition is as free as possible (Chapters 6, 7 and 8). In today’s globalized world the problem of the national state no longer being the ultimate reference of economic development and the absence of a world authority able to guarantee the rules of the game is strongly felt, as are the problems of taking account of market dynamics in competition cases (Chapters 6 and 8). The third axis concerns both rules and capabilities. It is about the role of research, of innovation and of the so-called ‘new science – new industry’ relationship in making markets and competition more dynamic. Innovation requires the two extreme levels of industrial policy: national or supranational regarding the definition of rules such as property rights; and local regarding the creation of knowledge (spin-offs and science parks examined in Chapters 13 and 14), that is, capability building. Regarding rules, the changes outlined above have led to the definition of new property rights regimes such as the open source model (Chapters 11 and 12). Innovation policies also interact with other policies, such as social policies (Chapter 15). The role of local realities constitutes the fourth axis of the new industrial development policies, which is the core of their capability-making aspect. Local realities and therefore

Introductiom xvii territories have been rediscovered as important determinants of productive development, in particular through the role of SMEs and their clusters (of which industrial districts represent a type), especially when they are locally rooted and when the territory is taken as a resource that can be exploited (Chapters 16, 17 and 19). For this purpose, provision of local public goods such as business support services is essential (Chapter 18). Decentralized industrial policies have proved to be useful but their effectiveness depends on their being integrated to a national (centralized) strategy (Chapters 16 and 19). The analysis conducted within the four axes points to a number of issues regarding, not only the practical implementation of industrial policies, but also their analysis in the field of economics. Hence Part V of the Handbook focuses on methodological and conceptual issues. A specific consideration of the important tool that has greatly influenced industrial economics, namely game theory, is first undertaken, showing that game-theoretical analyses have increasingly taken dynamics into account (Chapters 20 and 21). The Handbook is concluded by a particular view of concepts and approaches (Chapter 22). Overall, we think that the Handbook shows that, while the concept of industrial policy has been put to one side for a long time, numerous scholars have continued working on the subject and have developed a very rich field: studies are rich in methodological and conceptual approaches, in theoretical and empirical considerations, hence the very fertile and influential field. The new industrial development policies turn out to be programmes or strategies (objectives and means) of industrial development defined by a country or a group of countries (as in the EU). They primarily aim at providing the conditions for business performance, but they also include, as experience throughout the world shows, measures aimed at sustaining and promoting industrial restructuring, that is, at orientating comparative advantages (specializations). We provide below a brief summary of each chapter in order to guide the reader through the Handbook. Part I Industrial policy in a global perspective Patrizio Bianchi and Sandrine Labory provide an analysis of the new industrial development policies as they emerge out of the contributions to the Handbook, together with a historical perspective and theoretical interpretation of the move from ‘old’ industrial policies to the ‘new’ industrial development policies. The main characteristics of the new policies and the implications in terms of research agenda are pointed out. The chapter by Ashish Arora and Alfonso Gambardella discusses a number of questions raised by globalization on the basis of an analysis of a particular case, that of the significant growth of software production in some emerging economies. The chapter discusses the effects of the rise of the software industry in five countries – Brazil, China, India, Ireland and Israel – on the development of the sector in the USA, Europe and Japan. It is shown that, while the USA is likely to remain an important R&D centre for software, specializing in high value-creating phases of the production process, Europe and Japan incur more risks, essentially because they are not as able as the USA to attract a skilled workforce from the emerging countries and because the ability to relate to the firms in the emerging countries which have complementary activities is higher in the USA. The analysis of the chapter has a number of implications, including countries’ specializations, outsourcing and human capital policies.

xviii

International handbook on industrial policy

Chapter 3, by Jacques Pelkmans, examines the evolution of industrial policy in the European Union. It first provides a rigorous analysis of the decision-making framework for industrial policy within the European Union, given by the rules defined by the Treaties. It explains the two levels of application of industrial policy, namely national and European, and shows the complementarity of the two levels. The chapter thus shows the overwhelming importance of European integration in determining structure and performance of industry in Europe, as a result of the deepening, widening and enlargement of the EU. Sanjaya Lall determines how proper industrial development policies should be elaborated in order for countries’ competitiveness to improve, on the basis of an analysis of the East Asian experience. He reviews country performance over the period 1980–2000 in the developing world in terms of market shares for manufacturing value added and manufactured exports. He shows that the countries that performed best were East Asian ones. Within the East Asian region, the best performing were those, such as Korea and Taiwan, which implemented comprehensive industrial policies, that is, policies aimed at both the rules (effective institutions, bureaucracies, regulation) and capabilities (sustaining domestic firms). Thus most of them used infant industry protection, export subsidies and targets, credit allocation, local content rules and so on to build their industrial capabilities base. Jacques De Bandt provides some considerations in Chapter 5 on the changes that economies are experiencing worldwide and the consequent need for industrial policies. He stresses that what is going on is very complex to understand and is inducing a change in production systems, whereby knowledge management and creation is fundamental. Hence the economy is turning to a ‘knowledge-based’ economy in the sense that methods of production involve intense knowledge creation and management. Industrial policy should therefore change objectives and priorities. Policies should help to understand the complexity of production realities and systems and, on this basis, provide appropriate training to the population, promoting multidisciplinary approaches, system analysis and modelling, information and knowledge engineering, and so on. Thus policy should help develop competencies, the capabilities for producing and transforming knowledge into operational solutions or applications. The organizational set-up that implements policies should also be carefully designed, with an emphasis on the ‘bottom-up’ approach, but ‘with the understanding that the emphasis is not on individual actors, but on specific decentralized production systems’. Part II Competition and regulatory issues Chapter 6, by Jo Lorentzen and Peter Møllgaard, examines the implications of the intensification of innovation for competition policy. It has indeed been argued that globalization has meant increasing importance of innovation for firms. To face the challenge of more frequent innovation, one strategy that firms have been adopting is to form technological alliances. Policies have favoured such collaboration in order to increase innovation and its diffusion. However, these policies raise questions for anti-trust policy, since collaboration may continue at the final stages of production. The chapter examines this paradox. First, a review of the rationale behind technological alliances is presented. Second, the organization of industries in a dynamic context is analysed and consequences for competition policy are derived. The authors conclude that, despite the complexity of analysis, competition authorities should monitor the behaviour of high-tech firms during and also after judging cases. The authors call for more empirical research on where

Introductiom xix technological trajectories, especially in the new sectors, are heading. For this purpose, multidisciplinary research is needed. The chapter by Stephen Martin and Paola Valbonesi provides an objective and complete account of state aid in both the European Union and the USA. The authors show that state aid is present in both regions, despite legal impediments and resulting economic inefficiencies. They first review the main theoretical literature on the subject, namely strategic trade policy, tax competition and rent seeking, and conclude that none of the literature makes a case for state aid as an efficient or welfare-improving tool. The authors subsequently review the legal framework for state aid in the USA and the EU. The main competition-distorting question raised by state aid concerns inter-state competition in the USA and inter-member state competition in the EU. The empirical evidence is that, in both regions, state aid is still implemented in significant amounts. In the EU, enlargement to Central and Eastern European countries also raises new issues for state aid, since these countries employ these instruments to help their convergence to the rest of the EU. Hans Schenk concentrates on the anti-trust case of mergers in Chapter 8. He shows that, although mergers and acquisitions operated by firms in all industries represent enormous value, since firms spend amounts comparable to countries’ GDP to merge with or to acquire other firms, empirical studies do not provide evidence of positive efficiency effects of these operations. Evidence, however, is mainly on static efficiency effects, not dynamic efficiency effects. The author goes on to review empirical practice in the treatment of merger cases by competition authorities in the EU, the USA, Canada and Australia. Here again, dynamic efficiencies have been considered only in a few cases. The author shows that each region tends to have its own system without international coherence. Chapter 9, by Tom Björkroth, Sonja Grönblom and Johan Willner, deals with theoretical and empirical aspects of liberalization and deregulation of public utility sectors. The theoretical part of the chapter analyses the impact of competition and ownership on allocative efficiency and the significance of the cost structure for the possible trade-off between the imperfections associated with different regimes. Theory seems to suggest that the threshold for liberalization to be beneficial may be higher than usually believed. The empirical part focuses on telecommunications and electricity provision, with short discussions of other industries such as water and public transport. Deregulation in telecommunications has in most cases been followed by lower prices, but some incumbents were also technically progressive, and in those cases there are some question-marks related to the counterfactual. Similar concerns have been raised in the energy sector, but there have in addition been doubts about reliability, lost advantages of vertical integration or excessive consolidation. The chapter by Andrea Goldstein concludes Part II on competition and regulation, concentrating on nationalization versus privatization issues in both developed and developing countries. The chapter reviews alternative explanations of state ownership as a tool of industrial policy. It provides evidence on the role that state-owned enterprises (SOEs) played in the economic growth process of some developed and non-developed countries, on the motives for privatization, on the conditions in which SOEs have been sold and on the effects of such a policy choice. The case of France is also discussed in more depth. Goldstein concludes that the institutional framework within which privatization is undertaken is very important in determining its success or failure; governments should be as objective as possible, in the sense of resisting ideology, in order to raise the probability of success.

xx

International handbook on industrial policy

Part III Research, technology and innovation This part contains five chapters, each of which deals with specific key issues regarding technology and research policy. Fabrizio Cesaroni and Paola Giuri analyse the crucial issue of innovation protection. More specifically, the authors examine two models of intellectual property protection, namely the strong intellectual property rights (IPR) model and the open source/open science model, and their implications for innovative activities and market dynamics. The IPR and the open science models present advantages and disadvantages in terms of incentives to innovate, adoption of technological innovations and diffusion of technologies. The chapter focuses on some key and emerging issues on intellectual property protection and examines the effects of (stronger or weaker) patent protection on the innovative activity of large and small firms and on entry and growth of new technology-based firms. The main policy implications are, first, that the pros and cons of each model are influenced by the differences in the characteristics of the various industrial sectors; second, that the presence of complementary assets also influences the preferred model. Nicolas Jullien and Jean-Benoît Zimmermann also analyse the issue of the IPR versus the open source model of intellectual property protection, arguing that intellectual property protection is in practice always the result of a compromise between the various actors involved: producers, consumers and public authorities. Historically, protection has experienced phases where stronger and weaker protection have alternated. The authors argue that, in the software industry, the trend tends to be towards weaker protection, as shown by the diffusion of the open source model. More generally, in the knowledge-based economy the trend should be, contrary to what tends to happen nowadays, going towards weaker protection since there are enormous gains to be made from the sharing of knowledge and innovation as realized under the open source model. The definition of intellectual property rights is crucial in determining the incentives to innovate and to commercialize innovations. Such commercialization generally arises out of spin-offs from existing companies or from universities. Marco Giarratana and Salvatore Torrisi analyse spin-offs in detail, addressing a number of questions. Are they an important form of science-based entrepreneurship? Where do they come from (academic institutions, established firms)? What about their post-entry performance relative to other new science-based or technology-based firms? Why do some countries and institutions differ in the rate of new science-based firms formation? The authors answer these questions first by illustrating the importance of science-based industries and spin-offs; second, they explain the empirical evidence using different theories; and third, they analyse different policies aimed at science-based spin-offs. The creation of science parks and business incubators particularly favours spin-offs. In Chapter 14, Jan Annerstedt reviews the theoretical and empirical evidence on a particular form of innovative networks that has been a focus of technology policy over recent years, namely science parks. Science parks are geographically limited areas where business, private and public research centres and local governments are brought together in order to create ‘innovation environments’, that is, environments where the close and dense network relationships between the various actors of innovation favour innovation and firm creation. The chapter reviews the historical evidence on science parks and defines three generations of such parks, with an increasing integration of the parks into the local urban areas and into global flows of knowledge. The chapter subsequently goes beyond the science park

Introductiom xxi evolution to address more general policy-relevant issues of locally based clustering of innovative firms and their supporting institutions. Science parks are areas that favour firm creation, be they start-ups or spin-offs from public institutions or from companies. Daniele Paci and Stuart Schweitzer present an interesting and often neglected issue related to technology and research policy, namely the interaction between technology policy and other policies. They concentrate on a high-tech sector, the pharmaceutical sector, and make two points. First, the way a sector is defined in the analysis influences the policy implications one reaches; second, taking a broader view allows us to take important interactions between policies into account. Thus defining the health industry as comprising all the sectors that realize products aimed at providing health services (medical machinery, drugs and so on) with a market composed of not only patients but also health ‘providers’ (doctors, chemists and so on) leads the authors to outline crucial trade-offs, such as the interaction between technology policy that aims at favouring innovation in the sector allowing for monopoly rents over innovations (patents) and social policy that aims at providing health to all at low cost, hence favouring competition between firms. Part IV Territory, industrial development and small firms Part III showed that industrial policy in practice has had a tendency to decentralize, in line with the new ‘bottom-up’ approaches. Thus high-tech clusters have been a focus of technology policy in many countries over recent years. Part IV is therefore specifically dedicated to local industrial development and policies. Leandro Sepulveda and Ash Amin start this part by providing an analysis of the new decentralized industrial policy approach. They argue that this new approach has many merits, but at the beginning of its implementation there has been a tendency to decentralize too much and forget the importance of centralized policies. If industrial development is first and foremost localized on specific territories where it embeds, some degree of centralization is also required, in order to define coherent rules throughout the country and to allow successful experiences to spill over into other regions of the country. These arguments are illustrated with the case of Argentina. Marco Bellandi and Marco Di Tommaso then go deeper into the analysis of the rationale and forms of local industrial policies, that is, the local dimension of industrial policy. They also stress that local industrial policies have to be coherent with strategic objectives defined at a more central level. They carry out a deep and rigorous analysis of local industrial policies and show their various dimensions. They also show that the fundamental reason for the local level of industrial policy is that industry develops only if it embeds locally. This ‘embeddedness process’ determines the success of industrial development and can be supported by the public provision of public goods of various types. In Chapter 18 Nicola Bellini examines a particular kind of local industrial policy, namely business support policies. These policies are specifically aimed at assisting enterprises or entrepreneurs to develop their businesses and to respond effectively to the challenges of their business, social and physical environment. The chapter discusses three main approaches that are based on different research paradigms. It ends by pointing to the lack of evaluation of these policies. According to Bellini, evaluation is rarely carried out owing to the many difficulties of this exercise, such as technical difficulties which are explained in the chapter. However, evaluation has started to be performed in various ways. One example

xxii

International handbook on industrial policy

is analyses testing the additionality of business support policies. Other examples are examined, showing that evaluation should be a priority on the research agenda. Local industrial development primarily involves the creation and development of small and medium-sized enterprises (SMEs). The last chapter of Part IV, by Patrizio Bianchi, Sandrine Labory, Daniele Paci and Mario Davide Parrilli, focuses on SME policy. It provides an overview of the policy tendencies in both developed and developing countries, focusing on three continents, namely Europe, Latin America and East Asia. It shows that globalization has led to the focus on networking as a tool to consolidate SME development. The comparison of SME policies in developed and developing countries leads the authors to conclude that developing countries often have SME policies but which lack coherence such as is provided, for example, in Europe by the European decision-making framework. In addition, attention is often focused on the larger firms among SMEs and not on the myriad of very small firms that represent an important share of employment and contribute significantly to socioeconomic stability. This therefore risks creating a fracture within the national production systems. Part V The use of game theory in industrial policy analysis The various chapters of the Handbook so far have shown that globalization poses new challenges not only to the practice of industrial policy, that is, policy decision making and practical implementation, but also to the analysis of industrial policy by economists (and possibly other scholars). Part V addresses these challenges in more detail, focusing on an important tool of industrial economics, game theory. Chapter 20, by Luca Lambertini, shows how game theory has come to play an important role in industrial economics, providing a rigorous objective tool for the theoretical perspective. However, only the static game-theoretical framework was first used extensively. This approach has a number of drawbacks, including restrictive and often unrealistic assumptions and the sensitivity and variability of results to the particular assumptions made. Hence this perspective has been increasingly questioned, not least because of the evolution of the competitive conditions and of the industrial policy problems faced by decision makers to which static game theory was increasingly unable to provide satisfactory answers. Dynamic game theory has developed recently and rapidly. The first chapters of the Handbook show that looking at the reality of industrial development and policies leads to evidence that a proper approach is necessarily dynamic. In fact, Roberto Cellini and Luca Lambertini in Chapter 21 show how results and policy prescriptions change when dynamic rather than static games are considered. Part VI Particular views Christos Pitelis reviews the various methodological approaches to industrial policy analysis, focusing on the example of the EU and providing a bird’s eye view of extant alternative perspectives on industrial (and competition) policies. The chapter also derives implications in terms of industrial development policies in developing countries.

PART I INDUSTRIAL POLICY IN A GLOBAL PERSPECTIVE

1

From ‘old’ industrial policy to ‘new’ industrial development policies Patrizio Bianchi and Sandrine Labory

Introduction The last 20 years have been characterized by significant changes in productive structures and international competition. The entry of new competitors, such as Japan in the 1980s and China and India more recently; technological changes such as the diffusion of information and telecommunication technologies (ICTs) and scientific breakthroughs such as genetic engineering and the development of biotechnologies; institutional changes such as the deepening of European integration and enlargement of the European Union (EU); social and demographic changes such as the ageing of the population in developed countries; and so on have led to the necessity of structural adjustments so significant as to redesign the economy and the society beyond the boundaries of national states. All these changes imply the definition of new industrial policies, at both national and international levels. In fact, new approaches to industrial policy have been tried out throughout the world, in particular the formulation and implementation of policies ‘from the bottom’ (local or regional), in a context in which the central authorities seem no longer to have the coercion and command powers they used to have. In contrast, old-type industrial policies were direct interventions by the central state in markets to pick winners and support the development of particular sectors. This policy functioned in a context where the national state was indeed the reference. We define industrial policies as a variety of public actions aimed at guiding and controlling the structural transformation process of an economy. The implicit assumption is that the industrialization process is essential for the transformation of the economy as a whole and it is possible to act on this process in order to guide the entire structural change mechanism. Public actions are multiple and varied, ranging from actions defining the rules of the competitive game to actions aimed at favouring the participation and performance of particular individuals and firms in the competitive game (the framework conditions). Industry has the capacity to organize production, mobilizing both tangible and intangible assets, in the sense of using existing assets and creating new necessary assets. Industrial policies therefore concern all productive activities. Industrial policies in this sense originate from the affirmation of the modern state and the emergence of capitalism. It is indeed as capitalism develops that the capacity to organize production, rather than the natural availability of scarce resources, becomes the main determinant of the wealth of nations. However, the rationale for and content of industrial policies has changed greatly since the beginning of capitalism and especially over the last 20 years. The main elements of industrial policies remain the same, namely industrial strategy definition, competition policy and regulation, technological policy, regional and territorial policies, methodological approaches and conceptual analysis. These six main elements constitute the six Parts of the Handbook. The practical experience analysed in the Handbook shows two things. First, there has 3

4

International handbook on industrial policy

been a shift of emphasis of governments towards providing the conditions for business to prosper (especially with the liberal push). Second, these framework conditions do not exclude actions aimed at promoting structural change. In fact, most governments throughout the world (in the EU, in the USA, also in developing countries) adopt measures to define and guarantee the rules of the competitive game (anti-trust, trade policy and WTO rules, and so on), but they also take measures to promote structural change (favouring the creation of resources such as human capital and knowledge and their exploitation). In this way a country’s comparative advantages can be influenced: developing countries have thus built specializations in electronics (Korea, Taiwan) and software (India) after building specializations in heavy and chemical industries (Korea) or consolidating some low technology sectors (Taiwan’s exports were based on labour intensive textile products, they are now concentrated in the man-made fibres’ sector). When the country’s industrial development reaches a certain level (as in some Asian NIEs such as Korea, Taiwan and Singapore or in industrialized countries at their present stage of development), measures aimed at promoting structural change are more concentrated towards the ‘strategic’ industries, that is, high tech industries which require high R&D investments. In this introduction we provide a brief historical overview of industrial policies across the world in order to put the current industrial policies into context, with special emphasis on American and European history. We also identify what emerges out of the Handbook in order to define the new industrial policies. The practical experience examined in the Handbook regards all countries in the world, both developed and developing. We call the new policies, ‘industrial development policies’, not to give yet another new name to old things but in order to stress their dynamic character: they aim at the dynamic adjustment and development of firms and sectors and therefore are not once-and-for-all punctual measures. Rather, they are programmes that evolve over time as conditions change. All chapters in the Handbook stress the importance of dynamics. We aim to raise the most important issues, not to be exhaustive in listing the issues, which would be an impossible task, given the richness of the various contributions to the volume. The present chapter is therefore structured as follows. In the first section we provide a historical analysis of industrial policy in order to outline the evolution of the practice of industrial policies by governments across the world. In the second section we outline the main rationale and contents of the new industrial policies, which we call ‘industrial development policies’ for precise reasons which we explain in more detail. We conclude with some questions and some suggestions for future research. 1 Historical perspective Industrial revolution and the modern state The control of the means of wealth production and political power are always closely linked to each other. However, it is with the birth of capitalism and with the consolidation of the modern state that the terms of the issue takes a new meaning. The English industrial revolution stemmed from an extraordinary convergence of social forces that determined an unprecedented structural break in world history. First of all, there was the construction of a unitary nation with a legislative framework and a management structure that emerged out of centuries of conflict between the growing parliamentary power and the monarchy which increasingly had to recognize human and local communities’ rights. The industrial revolution was itself the expression of a ‘bourgeois’ revolution that

‘Old’ industrial policy to ‘new’ industrial development policies

5

replaced the king and his divine mandate with a nation and a state. In that nation individuals no longer asserted themselves according to rights derived from ancestors but thanks to the capacity to accumulate wealth sufficient for them to be co-opted in the command system of the society in which they lived. The modern state was asserting itself, as an organization in which political power incorporates and expresses, thanks to explicit assets such as a set of rules, precisely defined roles and resources used to serve a unitary and specific sets of interests and aims (Poggi, 1992, p.33). The upheaval produced by the industrial revolution originated from the simultaneous consolidation of a framework of social governance and of a productive organization that was redesigning the overall organization of society while boosting its development, thereby expanding both resources and rights. We use here Dahrendorf’s analysis of the conflicts of modernity and define development as both the growth of resources and the increase in the rights to participate in the creation and use of resources (Dahrendorf [1987] 1990). The stimulus to trade, to innovation and to the organization of productive forces stems from the possibility to transform this economic tension into effective rights and legitimate political power. In the same manner the wealth of nations rapidly becomes the power of nations and therefore the government stimulates, sustains and promotes the new industry because it is the source of national power (Supple, in Cipolla, 1980, p.285). In this perspective, the political content of the three aspects always underlined in the classical literature on backward development has to be considered: the creation of the large national market, the availability of raw materials of limited costs and the access to capital to be invested in the new productive activities. The creation of a large internal market is necessary for the development of productive organization and for the full use of static and dynamic economies of scale so as to reduce prices and initiate competition against local producers. The creation of a large internal market also requires an organized central state that is able to unify fiscal systems, financial and monetary rules, weight and measure systems and, above all, to stabilize a system of property rights recognition, not only for physical goods, but also for intangible goods. When this internal market is no longer sufficient the expansion of productive forces beyond national boundaries has to be supported; thus, English economic growth starts exerting political pressure on other countries. The availability of raw material stems from the possibility of specifying property rights on lands and licences for the exploitation of the ground, hence on a public good. The availability of investment resources requires both the possibility of transferring resources across sectors, for example from agriculture to the infant industry, and of a cohesion and integration between ruling classes and a capacity to regulate the new financial markets. The expansion of the productive system again requires its international expansion, that necessarily is transformed into a conflict between nations. In this sense the industrial revolution and the consolidation of the modern state are unified in that ‘bourgeois’ revolution that marks the contemporary era. The development of capitalism not only consolidates the modern state, but also opens it beyond its boundaries in a race that assumes in the 19th century the form of imperialism. Imperialism is no longer and only the search for gold and silver in new lands to conquer, as in the Spanish model, or the need for colonization in order to resolve the problem of scarce land, as in

6

International handbook on industrial policy

the case of North America or Australia, but rather a means to guarantee a protected market in which to buy cotton and other raw materials and sell manufactured goods. This trade increases and increasingly becomes unequal between contractors. The extraordinary growth in Britain and subsequently in Germany and France led to the design of a new geography of the power of nations. As argued by Landes (1969), English progress represented an inevitable challenge for continental states. The governments of European states had been considering the development of trade and manufacture as the necessary element to guarantee the wealth of the country through a good trade balance, state wealth through proper fiscal receipts and the wealth of the population through stable and growing employment. All countries encouraged the development of trade and local production, with particular support for arms production. In most cases, however, these activities were realized in craft form and often managed by the state’s structures. As pointed out by Landes (1969), the English revolution introduced new modes of production, productive volumes and product typologies that changed the equilibrium of economic forces, so much so that industrialization was from the start a political imperative for all governments which had to follow the new English industrialists (Landes, 1978, p.183). Accepting English imports would mean being dependent on another country and experiencing a reduction in state income and a rise in unemployment. Facing the English challenge meant forcing industrialization but without the technical knowledge, the financial capital and the commercial networks that the United Kingdom had. Prussia was the most active in this forced industrialization and the King of Prussia implemented, until the unification of the Reich, an active policy of promotion of the new industry, of schools and universities for the diffusion of new technologies and research, such as the Prussian Gewerbeschulen or the Gewerbe Institut of Berlin and the massive obligatory schooling that resulted in a schooling rate of 97 per cent (Landes, 1978, p.447). The King of Prussia urged, or even forced, the local nobility, the powerful Hanseatic communities and the Jewish bourgeoisie to invest in manufacture while the state secured a strong presence through commercial protection, school creation and large orders of mainly warfare and infrastructural material such as railways. In the meantime the king promoted a customs union between Prussian states, the Zollverein of 1831, which significantly contributed to redefining the commercial and political forces in Europe. Economic growth received the same push in France, since the government favoured both the creation of banks and shareholders’ associations (the so-called ‘Caisses’) and a programme of investment in railways and infrastructure. The creation of the Crédit mobilier of the Péreire brothers (1852) represents a significant innovation in investment financing. These banks were rapidly diffused in France, Germany and then over all of Europe and were simultaneously deposit, discount and business banks, with large capacity in booming phases, when deposits were used in industrial investments, thereby producing the necessary liquidity, but an extraordinary fragilility in slump phases when the financial lever inverted and the whole system had no guarantees at its disposal (Gille in Cipola, 1980, p.223). Landes (1978, p.254) argues that ‘the period between 1850 and 1873 was the period in which European industry became adult’. Although English supremacy was maintained by a global orientation, the French, German and, to a lesser extent, Belgian industrial and financial systems were consolidating and needed to expand to the rest of the European continent and to the colonies. In this context, the entry of the United States into the world economy is quite original.

‘Old’ industrial policy to ‘new’ industrial development policies

7

As explained by David Riccardo at the beginning of the 19th century, the English and European economies were based on the exploitation of lands belonging to landowners who delegated its use to entrepreneurs. The latter used labour compensated at a subsistence level to produce for the market. Therefore, once landlords’ rents and labour wages were paid, profits depended on both the choice of the most efficient techniques of production organization and the extent of the market in which they sold. In such a context the possibility of increasing production depended on the possibility of increasing the extent of the market, hence of exporting outside the country. Therefore the competition between firms became a competition between countries in an international trade model in which each country tries to maintain the competitiveness of its national champions in order not to lose economic and hence political sovereignty. In the new American nation, conditions were different since neither lands nor labour supply were constrained. Therefore the possibility of production was technically without limits since demand expansion was ensured by luring a growing number of consumers or by conquering new territories. In this case dangers are presented by the lack of the risks were rather overcapacity of production or the creation of monopolies. Thus two approaches to industrial development emerge. In the United States, the federal state intervenes to guarantee the rules of the game and at the same time takes actions to sustain demand such as large structural investments or massive military spending. In continental Europe, the national state intervenes to sustain a system with limited dimensions in a competitive game that turns into a competition between nations. The state intervenes in Europe to create agents able to compete internationally or to close the market when local firms are not able to compete against foreign firms; it also tends to nationalize firms in strategic sectors (infrastructure or goods with military applications) in order to protect them from competitive forces. Anti-trust policies and natural monopoly regulation As a consequence of the above considerations, anti-trust policies differ widely between Europe and the USA. While in Europe the economic process was an optimization of limited resources, in the USA the constraint of limited resources did not exist. Economic growth in the USA relied on the initiative of entrepreneurs, the free interaction of whom determined the organizational forms that were adequate to the country’s development, through a Darwinian process. The consolidation of local powers, however, tended to limit this evolutionary process, since the control of all the strategic resources of an area or a sector could lead them to block the entire growth process of the nation. Thus, in the period following the Civil War and the Indian wars, when the geographic expansion of the country ran from the Atlantic to the Pacific oceans, no resistance existed anymore. Therefore strong concentration processes occurred in some crucial sectors such as oil, thereby creating difficulties both for the economy owing to the risk of monopolization by strong industrial trusts and for the state whose legitimacy was based on the free initiative of every citizen. The publication of the Sherman Act in 1890 reflects the importance of the political consolidation of the state against attempts to monopolize the economy. Any attempt to monopolize directly or through agreements between producers is considered as a criminal infringement that directly affects the autonomy of the state and hence one of the cultural pillars of the young American nation.

8

International handbook on industrial policy

Adam Smith had already warned in The Wealth of Nations that monopoly had to be avoided not only because it distorts the allocation of resources but also because the resulting distribution of wealth does not optimize the overall use of existing resources and creates social injustice. Monopoly also generates a feudal constraint between the monopolist, who can discriminate the quantity and prices of goods and the population, which has to consent and cannot give voice and hence has no freedom. The English tradition of Common Law had many cases of contracts that were declared null because they had been realized under a bargaining asymmetry between the contracting parties. In American courts too some cases had resulted in the cancellation of clauses of obligatory buying by exclusive suppliers. The Sherman Anti-trust Act attributed to these norms the value of fundamental law to regulate the entire American economy in its expansion phase. The subsequent Clayton Anti-trust Act of 1914 recognizes that in some cases, however, mergers can strengthen the smaller competitors to the point of making them effective contenders of consolidated incumbents. The prohibition of firms’ agreements had in fact induced the merger of local companies into large national companies that dominated the national market. The subsequent norms of the 1930s and 1950s further influenced the development of these firms, inducing them, first, to form conglomerates covering various fields of activities, so as not to take too dominating a position in specific sectors, and then to internationalize, in order to avoid their strength on the national market being seen as monopoly powers. Anti-trust norms have therefore not only avoided monopolies putting the legitimacy of the central state at risk but also oriented the development of American large firms. The AT&T case is an extreme case of this power of orientation. AT&T was divided into many local companies in order for the regulatory reform of the telecommunications sector to be effectively put in place, and this led to a reorganization of the whole American industrial system. The opening of sectors characterized by a natural monopoly to competition (namely the regulatory reform or liberalization of utilities) also included the air transport sector and had significant effects on the productive organization of the sector, not only in the USA but also worldwide. In the American tradition anti-trust action has been accompanied by actions to sustain demand through public procurement for civil or military uses, especially in sectors characterized by strong technological innovation. Another important element in the USA is the effective role of the financial market in the selection, stimulus and sanction of entrepreneurial initiatives. The crucial element, however, is the realization of a large market where competitive dynamics between subjects who recognize the common regulation imposed by the federal state are ensured. The European industrial history is very different because Europe is characterized by internal markets that are not large enough to guarantee efficient competitive dynamics. Therefore competition is between firms that represent national interests in a conflict between nations. In this context anti-trust policies have been different, if not opposed to American anti-trust. As stressed by Marshall (1919), the emulating race to reach the UK’s level of industrialization led to the promotion of industrial agglomeration processes which in turn resulted in the creation of trusts that were backed by the states. In Germany, the ‘Konzerns’became the instruments of forced industrialization and led to a strong industrial concentration and thus the creation of national champions able to compete against English

‘Old’ industrial policy to ‘new’ industrial development policies

9

leaders. In Japan, too, the ‘zaibatsus’were supported in order to create national champions. These national champions, however, contributed to the growing rigidity of the economy. This rigidity was of course reinforced by the consolidation of the authoritarian political regimes which progressively transformed the German and Japanese economies into war economies where all economic activities and civil life are subject to the national state. At the end of World War II, competition was introduced into the German economy by the allied forces which controlled the country. The Konzerns such as IG Farben in the chemical sector, or the big groups in the steel and metal sector that had played an important role in the armament of Germany, were broken up. The creation of the Federal Republic that brought together the territories occupied by the American, French and English troops, initiated a phase of economic reconstruction based on the concept of a social market economy. The social market economy was characterized by competition as a reference but with an attention to economic solidarity where even town halls, local states, foundations and trade unions were part of the ownership of firms which progressively reconstructed the big groups that were eliminated at the end of the war. German capitalism took shape within a dense network of public relations and was characterized by strong competitiveness in final productions and strong regulation at source, that is, governed by cross shareholdings through a limited set of large banks. Antitrust norms were implemented only in the 1970s, when the oil crisis created difficulties for the new economic giant and required the temporary admission of crisis cartels in order to face problems in some crucial sectors of the country. Japanese history compares to that of Germany. After the Meiji revolution which took place in the same year that the German Reich was created, 1870, an extraordinary technological and organizational catching up with regard to Western countries started. The feudal economy was rapidly reorganized into large banking and industrial conglomerates, the zaibatsus, that were also broken up by the temporary allied government at the end of World War II and progressively rebuilt in the shape of new industrial export-oriented groups thereafter. The role of the Ministry of International Trade and Industry (MITI) in the orientation and promotion of the strategic convergence between these groups has been widely discussed in the literature. However, the most important aspect is the possibility of reconstructing the groups which had to adopt aggressive commercial policies in order to grow without monopolizing the internal market. In this respect the Italian case is also interesting. Italy was also defeated in World War II and experienced extraordinary growth during the 1950s and 1960s. The industrial development of Italy took place much later than the English one, that is, at the beginning of the 20th century and 40 years after the country’s unification. Prior to the unification, Italy was divided into various states that were too small to allow the development of an efficient industry. Only the north-eastern region of the country had a developed textile industry thanks to its links with the Austro-Hungarian empire. Unification implied the need to develop a heavy industry that could sustain the willingness of the new state to increase power. Economic growth was sustained by public orders of armaments and infrastructure while foreign capital was attracted, first from France and then from Germany, in order to invest in business banks and to develop large firms in the steel and chemical sectors. This system, based on public demand for military goods and on inflation without limits, led to a high growth of firms that, however, remained deeply fragile. After World War I and during the crisis of the 1930s, the system

10

International handbook on industrial policy

collapsed, and in 1933 the state became the owner of a large part of the industry after the rescue of the main large banks and the creation of the Institute for Industrial Reconstruction (Istituto per la Ricostruzione Industriale, IRI). After World War II, a debate took place on the privatization of IRI, but Italy did not dispose of sufficient capital to acquire the state-owned enterprises. In addition the strategy of export orientation was considered as the best for the country’s development even by a large proportion of private entrepreneurs and required transforming raw material and semi-finished goods, the capital for which only the state could guarantee. A system of symbiosis developed between state-owned enterprises – that produced steel, energy, transport, that is, investment goods – and public companies that produced consumption goods such as cars, fridges, clothing and food. Such a system allowed a long period of growth for the country where capital and raw materials were scarce. The rules characterizing the system remained in place until the 1990s. In the new world and European context a process of massive privatization took place in the 1990s that affected both the banking sector and IRI, which was dissolved in the year 2000. A vast policy of competition guarantee and monopoly regulation was implemented through the creation of a network of independent authorities that regulate the various markets. This policy, however, has left ample margins for concentration in particular in the strategic information sector, thereby generating an anomalous situation of strong concentration of economic and political interests. If the three defeated countries experienced in the immediate post-war years an intense phase of reconstruction followed by an economic boom, the evolution of the victorious countries was more complex, where the economic and political costs of decolonization were added to the cost of reconstruction. The UK, France, the Netherlands and Belgium had to experience in the 1950s a difficult exit from colonialism that had allowed the internationalization of these relatively small countries. In the UK, the first merger and monopoly regulation was implemented in 1948, in a difficult phase of reorganization of the economy. British industry had been formed on the basis of an Empire of worldwide dimensions; it was left with an internal market of reducing dimensions, despite the attempt to create a customs union between the countries of the Commonwealth. The risks of monopolization were therefore high. British industry experienced a long period of decline in which nationalization and privatization alternated and in which the anti-trust norms accompanied a difficult restructuring. The UK’s massive privatization programme of the 1980s was accompanied by a stricter application of anti-trust policy, in particular in the sectors of telecommunications and information. The Broadcasting Act was adopted in 1980 and strictly regulated the propriety of journals and televisions in order to prevent too high a concentration of these sectors affecting not only the competitive but also the political framework. This can be considered as a relevant preoccupation, given the evolution of the Italian situation where the lack of application of anti-trust norms has allowed the major owner of journal and television channels to become the prime minister, thereby intertwining the political and economic powers. In France, following the tradition of control over industrial concentration, the government promoted merger between national firms in order to create national champions able to face the ‘American challenge’.1 After the phase of decolonization and during the 1970s, the French government claimed that the relaunch of the French economy could be realized

‘Old’ industrial policy to ‘new’ industrial development policies

11

only through support of large firms to the point where they could become dominant and therefore comparable to the American multinationals already present in Europe. This policy has been implemented by governments of different political stripes, sometimes through the nationalization of industry, sometimes using privatization. In this context, anti-trust policy initiated in the 1980s has had the effect of internal monitoring, with the idea that French competitiveness had to be defended primarily on international markets. The new orientation of the European industrial policy The European case is therefore different from the American tradition where market monopolization is a criminal infringement per se. In the European experience, articles 85 and 86 of the Treaty of Rome (now articles 81 and 82 of the EC Treaty) constituted the basis for the action in favour of competition in the European Community. However, these articles essentially provided safeguards of the proper functioning of markets which were being integrated into the new Common Market. Simultaneously, some norms regulating both state aids to industry and dumping were implemented, while state ownership was not regulated at the European level, member states being only obliged to make sure that stateowned enterprises were not favoured on the common market at the expense of other firms. Derogations, however, were permitted and firms’ agreements and state aids were allowed in cases when this would increase the competitiveness of the whole European industry. This common action significantly weakened during the 1970s, when the oil crisis led many large European firms to suffer difficulties and thereby create wider economic and social problems, up to and including political terrorism phenomena in some European countries. Hence public interventions aimed at sustaining industrial reorganization were implemented, much like the policies for industrial reconstruction that were implemented after the crisis of the 1930s and after World War II. These years of the 1970s were also characterized by the progressive decline of state-owned enterprises, in the sense that many of these enterprises became lame ducks, while the interventions used as social buffers during the most difficult years were generating market distortions that could not be sustainable over time. After the most acute phase of the crisis, when the strategies of ‘downsizing’ of many large firms started to bear fruit and when new entrepreneurs emerged, the process of European integration was given new impetus with the idea that it would help the European economy to recover competitiveness. Thus the Single European Act of 1986 and subsequently the Maastricht Treaty implemented in 1993 initiated a new phase of European integration, where the elimination of residual tariff and non-tariff barriers was claimed to be a necessary condition for European firms to be competitive, since the increase in the extent of the market would allow firms to exploit higher economies of scale, to increase in size and to offer European consumers lower prices and higher variety (Cecchini Report, 1985). Simultaneously, the integration of financial markets that had been postponed since the creation of the EEC was progressively started. The relaunch of the integration of European markets implied a wave of mergers and concentrations across European firms. This phenomenon was at the origin of both the more rigorous application of articles 85 and 86 of the Rome Treaty and the adoption of the Merger and Concentration Regulation in 1990. The regulation in fact gave the European Commission some powers to influence the industrial restructuring process that

12

International handbook on industrial policy

was taking place. In addition to this, both article 90 of the Rome Treaty (on state-owned firms) and article 92 on state aids were implemented more accurately, with important implications such as the decision by the Italian government to realize relatively rapidly a massive privatization of state-owned holding IRI that was agreed by the Italian minister Andreatta and the Commissioner of competition policy Van Myert, who had contested to the Italian government the excessive debt of the IRI and some of its main firms, such as Finsider in the steel sector and Alitalia in the air transport sector. This turning point of European policy led to the reduced focus on the policies that directly supported crisis sectors and that were implemented by national governments in agreement with the European institutions in the 1970s to avoid the decline of key sectors like steel, shipbuilding or sugar and textiles. European Commissioner Davignon had argued in 1970 that it was necessary to preserve the existence of strategic industrial sectors by helping them to rationalize and to renew equipment and buildings. This action, however, could no longer be justified in the second half of the 1980s and in the 1990s, when a new economic and political phase had started. In particular, the crisis of the former Soviet Union was generating a new scenario for Europe. The European Community became the European Union, having changed considerably since its beginning; to the six founding countries a number of new members had been added, from the Northern European states that were members of EFTA to the Southern European states that joined as soon as they turned into democratic regimes. The relaunching of the construction of the European market therefore took place in a new context, that of a Union between countries of varied structural histories and dynamics, instead of the relative homogeneity of the six founders. In this context it was clear that a policy focused on ensuring competition only would have risked not allowing the weakest firms in the Union, especially in the least developed regions and the periphery, to develop, thereby creating a widening gap between the competitive centre of the EU and the increasingly backward periphery. Hence a European structural policy was defined, aimed at favouring the development of marginal areas and their full participation in the competitive dynamics of the EU. In this process, important innovations in terms of methods of industrial development policy were introduced (progressively from the mid-1980s). Structural policies are first characterized by the necessary involvement of all public and private institutions of the concerned area in their formulation. Second, their elaboration requires the definition of development programmes that aim at enabling the territory and its firms to play an active part in the new European market. The definition of such a programme requires a deep analysis of the economic and social reality of the area to identify both the potentialities and the positive externalities (for example, transport infrastructure, schools and so on) that have to be created in order to sustain the development process. Cooperation between local firms also has to be favoured in order to produce the certifications that promote growth, such as quality trademarks for food products or technical standards for mechanical products. Within these programmes some investment aids can be provided after verification that they do not distort the proper functioning of the Single Market. This methodology of programming based on development plans has been diffused throughout Europe, even in the smallest towns. European structural policies have certainly not always been successful, but the successes in Ireland, northern Spain, Portugal,

‘Old’ industrial policy to ‘new’ industrial development policies

13

and also some areas in southern Italy, show that the large internal market can grow, not only strengthening the most competitive incumbents, but also allowing the entry of new competitors. In addition to the redefinition of territorial policies, the European relaunch has led to a new definition of research and innovation policies. The European Community had based this policy on the heritage of Euratom, which was the union of the six founding countries for the common development of atomic energy. The Euratom Treaty, adopted jointly with the Treaty of Rome, was centred on the institutional obligation to develop research on energy, using a common research centre. Euratom was subsequently merged into the EC treaties, while specific initiatives had been implemented to promote industrial research and innovation in crisis sectors. As a result, a set of varied and often incoherent measures were taken in the field and started to be coordinated within Framework Programmes (starting in 1984). However, the bulk of research and innovation policies were implemented at national level. The Maastricht Treaty subsequently put emphasis on the role of research and innovation in determining competitiveness. This emphasis was amplified in the declaration of the Lisbon strategy at the European Summit of 2000 when the EU heads of state took a strong position in favour of research and innovation to ensure higher competitiveness of European industry and higher levels of employment in Europe. Thus, to a certain extent, a European industrial policy has been defined, but one that is very different from the traditional meaning of the term. The shift of emphasis from declining industries to strategic industries (strategic in the sense of using a generic technology, that is, an advanced technology that has many potential applications to various sectors and that requires large R&D investment) is strong at the beginning of the 1990s, when Commissioner Bangemann publishes a report on industrial policy (the Bangemann Report: European Commission, 1990). In this document, he argues that governments should play the role of pathfinders and catalysts of industrial development. Without going into details (see Chapter 3) the new approach is then formalized in article 130 of the Maastricht Treaty so that for the first time industrial policy is explicitly mentioned in EC Treaties. The industrial policy that has emerged over the years is defined in the Communication, ‘Fostering Structural Change: an Industrial Policy for an Enlarged Europe’.2 This Communication follows the 2002 Communication on ‘Industrial Policy in an Enlarged Europe’3 and the 2003 Communication ‘Some Key Issues in Europe’s Competitiveness – Toward an Integrated Approach’.4 With these Communications the European Union has provided the concept of industrial policy with a new legitimacy, but precisely defining its new contents. In the new approach, industrial policy means a variety of policies aimed at creating an environment that is favourable to industry development. The European Commission recognizes the need for restructuring that most European industry is facing and ‘intends to mobilise industrial policy to accompany the process of industrial change’ (p.2). This industrial policy has three major aspects. The first is to improve the regulatory framework. A number of programmes and measures have been undertaken since the 1990s regarding this aspect, at both European and national level, such as the SLIM programme to simplify legislation and administrative procedures that business has to go through or the BEST programme aimed at benchmarking member states’ actions for SMEs (see Pelkmans et al., 2000; Pelkmans in Chapter 3 of this book). A second aspect refers to

14

International handbook on industrial policy

the ‘synergies between different Community policies having an impact on industry’s competitiveness’ (p.3) that have to be better exploited. Five particular synergies are identified: knowledge (innovation, research, training), functioning of markets (taxation, competition policy), cohesion (regional and employment), sustainable development (in particular eco-sustainable production) and finally an international dimension of industrial policy (to favour the access of Community companies to third countries). A third aspect regards sectoral policies. A return to sectoral policies is not proposed, but a need for adjustment of measures to the specificities of the sectors is suggested. The European Commission thus proposes to ‘ensure firstly a monitoring of the competitive situation of the main industrial sectors and the evolution of the process of structural change’ (p.38). Such monitoring makes it possible to identify the specific problems of each sector and, accordingly, to adapt the industrial policy measures taken at European level (which are horizontal in nature). Hence European industrial policy as defined in the Communications of the Commission is not a return to old-type industrial policy, but is a definition of the new industrial development policies in that it is defined as measures that favour the dynamic adjustment of industry to competitive challenges. This means both ensuring the conditions for business to develop and taking some measures to promote restructuring (development of and access to new technologies, training, access to financing, and so on). As a conclusion, the term ‘industrial policy’ is again legitimate in Europe, but in the precise and new sense of creating an environment favourable to industrial development, given that industry is still largely the main element of development and of European integration. The ultimate aim of industrial policy is economic growth and social stability (mainly through employment), in that it is part of a growth strategy. At European level, these industrial development policies are grouped under the term ‘competitiveness policies’ (see Chapter 3 by Pelkmans), thereby underlining the fact that these policies do not just aim at helping the manufacturing sector but aim at industrial development for economic growth and social stability. This change in perspective is linked to the evolution of the EU, but it has also reflected the dramatic changes that have occurred in the competitive context and that are often summarized as ‘globalization’. It has also been supported by a long and heterogeneous theoretical thinking in economics that has developed since about the mid-1980s. These reflections led to a change in the whole theoretical framework, towards a strong emphasis both on the role of competition and market opening and on the role of regions, cities and more generally the civil society, as well as the role of innovation as an engine of productive transformation. The following section therefore explains the changes in both the competitive context and the theoretical thinking that have led to the definition and implementation of the new industrial development policies. These new policies will be examined below. 2 New industrial development policies Globalization and new industrial policies Globalization is an all-embracing term that scholars and policy makers use to mean different things, from the intensification of worldwide exchanges to the diffusion of information and communication technologies (ICTs) that result in the advent of the ‘knowledge-based economy’.

‘Old’ industrial policy to ‘new’ industrial development policies

15

According to the present authors it is a term used to mean numerous changes that have occurred in the real economy over the last 20 years and that together have resulted in necessary structural adjustments for firms. The increase in worldwide exchange (be it trade or foreign direct investment, FDI) is not new and has been going on throughout the 20th century; however, it has accelerated in the last two decades (for the evidence, see Audretsch, 2003; Lall, 2004), owing to various factors, including in particular the diffusion of ICTs, the continued pressure for trade liberalization within the GATT and then the WTO and the improvements in financial markets. Lall, in Chapter 4 of this volume, summarizes the changes as ‘the intense pressures created by rapid and inexorable technical change, failing transport and communication costs (shrinking economic distance), widespread liberalisation and the spread of global production networks’. De Bandt argues in Chapter 5 that changes include ‘increased communication at lower costs, increased international flows of production factors (. . .), world-wide dissemination of production facilities and capabilities, continuous and rapid growth of international trade’ and, in developed countries, ‘the ageing of the population’. A last, but not least, factor is the change in the international political landscape, with the end of the division of the world into two blocks that had been taking place since the end of World War II. As a result, new countries have emerged in the international economy, for instance Eastern European countries and China. All these factors imply a progressive change in competitive conditions that we can call ‘globalization’. It means a change in the extent of the market (demand and supply conditions) that has induced the need for firms to adjust. Non-price strategies have to be adopted in most sectors, meaning increasing the differentiation of products by rapid product renewal and production of many varieties. Products incorporate consumer tastes and new technologies so that their knowledge content increases. The most value-creating phases of the production process are no longer those in manufacturing, as in the mass production system, but the phases located just before and after manufacturing, respectively research and product development (pre-manufacturing) and marketing and commercialization (post). As argued by Bianchi and Labory (2004), these phases are also the phases which are most intensive in intangible assets such as knowledge (innovation as a creation of knowledge), human capital, organizational capital (the way the teams of researchers and marketers are organized determines their effectiveness in creating appropriate products) and the social capital (the conventions, norms and values of a society that are expressed in its formal and informal institutions and that make relationships easier, more trusting and more efficient). Hence we agree with De Bandt (Chapter 5 of the present volume) that the structural changes firms face include the design of new production systems. Firms are encountering difficulties in adapting to the changes, for different reasons that De Bandt outlines in his chapter. First, it is difficult to understand ‘what has been really changing, on a permanent basis, and what the implications are’; second, economic actors in general are often reluctant to change; third, the changes that have to be made appear to be very complex and deep, implying a long learning process. This necessity of deep structural changes has led many firms to ask for protection and assistance, which governments have taken seriously, hence the debate on the possible return to old-type industrial policies. The contrast between the ‘old’ and ‘new’ industrial policies should not only be seen in a historical perspective, as above, but also as a debate

16

International handbook on industrial policy

that emerges again each time the relative position of countries weakens with respect to new entrants. As illustrated by Martin and Valbonesi in Chapter 7, the policies of direct subsidies to industry through state aid remain very diffused actions although they are widely recognized as inefficient. The reason is that the social groups affected by the crisis ask governments for direct help. The emergence in particular of China, just as with that of Japan in the 1980s, has implied the re-emergence of a request for protection even in very liberal countries. This request has been seen by some scholars and other observers as a return to old-type industrial policy. However, a return to old-type industrial policies is impossible. Not only, very often, does the competitive framework shaped by the laws and regulation (anti-trust, product and labour regulation and so on) not allow the implementation of such policies, as for instance the EC treaties prohibit such a return in the European Union (see Chapter 3) but also international rules such as those of the WTO do not allow (in general) a government to subsidize its domestic firms at the expense of foreign ones. Hence the debate over whether we are experiencing a return to old-type industrial policies is largely irrelevant. What is relevant is that the challenges brought by globalization as we intend it here have led to the progressive definition of new industrial development policies. Before turning to an explanation of these new policies it is necessary to analyse the change in economic thinking that has been at least partly induced by globalization and that has also been influential in framing the new policies.5 The years of crisis and the renaissance of economic thought The neoliberal revolution which involved, in the 1980s, all the international organizations and all national governments, put an end to the old approach to industrial policy; the new ideology even led to the elimination of the term ‘industrial policy’ from the vocabulary of economists. However, over the same period, a new approach to policies for industrial development was elaborated, which increasingly influenced the actions of both international institutions and national and regional governments. The liberal line of policy was defined as a general reorganization of the economy, centred on the opening of economies, the privatization of state-owned companies and the deregulation of economic activities. The new line was subsequently stylized as the ‘Washington Consensus’ (Stiglitz, 2002). Experience shows that this approach did not function per se; that is, it did not lead obviously to the expected fruits: growth and development. Various chapters of this Handbook show this: Lall in the case of East Asia, Bellandi and di Tommaso regarding local industrial policies in developing countries, Bianchi et al. in the case of SME policies. Evidence is also provided in the Handbook that the liberalization of utilities has not always provided the expected fruit either. Thus Björkroth et al. show that there is neither theoretical nor empirical uncontroversial evidence that the opening of network industries (industries characterized by an essential infrastructure or natural monopoly such as electricity, gas, water supply and so on) has led to lower prices and higher quality, hence higher welfare. Telecommunications, though, is a case apart since the natural monopoly no longer exists thanks to technological innovations in the sector. In this latter case, the effects of the opening to competition have been shown to be largely positive.

‘Old’ industrial policy to ‘new’ industrial development policies

17

Together with the new liberal prescription, some countries considered both the capacity to stimulate development from below and market forces as the dynamic engine of the economy. This stimulus was realized through two ‘discoveries’, the emphasis of two intellectual and operative fields of thought that had remained at the margin of economic thinking and most of all of political action, namely (a) the role of innovation and research, and (b) the role of the territory and of cooperative clusters of firms. Both approaches have very ancient roots but they really emerged in the 1980s to stimulate more complex economic visions than the ones offered by the existing orthodoxy. Eminent scholars refocused their work on industrial organization dynamics and several policy practitioners pioneered new approaches to local development in an open economy. Growth theories had long been focused on macroeconomic views and the accumulation of physical capital and labour (in the sense of quantity of labour, not quality, that is, skills), such as the famous model of Solow (1956). The oil crisis and the long years of stagnation and inflation determined a strong need to search for the hidden dynamics of growth. Hence growth theory was enriched by the consideration of endogenous factors of growth, mainly technological capital (Romer, 1986,6 1990) and human capital (that is, skills, as in Lucas, 1988). Numerous theoretical and empirical studies have been performed to show the importance of endogenous factors which are essentially countryspecific (for instance, Barro, 1991; Barro and Lee, 1993, show the importance of human capital and public goods in determining growth, while Caballero and Jaffe, 1993, assess the effect of the stock of knowledge and its evolution on growth). The work of Schumpeter was reconsidered together with the development of an evolutionary vision of the economy where the competitive dynamics determine development patterns that are not predictable ex ante but are path-dependent (Nelson and Winter, 1982; Dosi, 1991; Stoneman, 1995; Fagerberg and Verspagen, 2004). Besides this evolutionary approach, a rediscovery of institutions occurred. Institutions include firms and states and are viewed as agents that can act as intermediaries in the relationships between individuals and are able to provide rules that can internalize these conflicts (North, 1990). Firms were viewed again as complex organizations within which contracts have to be organized (Williamson, 1975, 1985; Williamson and Winter, 1993; Alchian and Demsetz, 1972; Jensen and Meckling, 1976) or in which behavioural routines are generated (Nelson and Winter, 1982) and where phenomena of learning and knowledge accumulation have a key role in the growth of firms (Teece et al., 1994; Dosi and Marengo, 1994). These routines ensure the coherence of productive systems within firms and greatly condition the capacity of managers to react to external events and to reply in innovative ways. Within sectors one could observe strategies of differentiation according to the capacity to innovate and therefore to attack competitors in competitive wars that game theory tools allowed to stylize in theoretical analyses (Tirole, 1988; Lambertini, Chapter 20 in this volume), just as the consolidating strategic management corpus was increasingly used in firm management (Porter, 1990). Innovation, first intended as technological change and then extended to also mean organizational transformations and the introduction of new products that change tastes and consumption patterns, is recognized as the engine of structural change and therefore increasingly risks generating monopolistic abuses that tend by nature to block change, with very negative effects on democracy (for instance, Cowling et al., 1999).

18

International handbook on industrial policy

Overall these new theories all point to the complex characteristics of social and economic dynamics that allow a community to improve its capacity to generate surplus, to accumulate wealth and to favour the birth of new firms. This is rather close to the ideas of Hirschman (1963), who argued that development arises as a consequence of disequilibria (p.77) and that industrial development is the engine of economic development. The theory of balanced development (Nurske, 1953; Lewis, 1954) was consolidated in the 1950s and stylized the growth process as the capacity to develop harmoniously the various sectors of the economy, so as to avoid generating disequilibria between demand and supply that could slow the growth process down. Development appears to arise out of a big push on demand operated mainly by the state, together with a careful regulation of imports in order to avoid the new demand being met by foreign, already strong firms, and being met instead by the new local firms. This is close to the theory of infant industry protection first developed by Stuart Mill in the UK and List in Germany. The structuralist approach represented by Myrdal (1957) reached the same conclusion and recommended protection because of the rigidities of the markets in developing countries. These ideas translated into the policy of import substitution implemented in Latin America in the 1950s. This policy consisted in protecting the local firms that produce in the sectors where the country tends to import a high volume of products. This policy led to the development of a number of industrial sectors in Latin America which, however, never managed to gain significant autonomy. As shown by Rodrik (2004) or Chapters 16 and 19 in the case of SME policy, the problem was not that no industrial policy was implemented, since a number of measures such as public ownership, public R&D or subsidized credit were used to support domestic industries; the problem is that industrial policy has not been concerted and coherent. The studies on innovation and technological change, and on science as a generator of new economic development possibilities in a competitive context, have had a significant impact on the political debate both in North America and in Europe. In the USA, the rapid growth of new sectors related to information and communication technology, together with the diffusion of new financial means, determines the emergence of the so-called ‘New Economy’, that leads to the redefinition of the industrial structure of the country, with the emergence of new leaders and the consolidation of new success factors for the business community. Stress is put on both the necessity for government to invest massively in fundamental research (for instance, the genome) or in new communication infrastructure (for instance, the Gore project on information highways) and the local aspects of high-tech and high science sectors’ development (Silicon Valley, Massachussets, Texas), that is, areas of anciently rooted development where the division of labour between university, large firms, high-tech SMEs and the financial system is redefined. In Europe, a new impetus is given to the integration of the European markets and a new emphasis on innovation policy actions that are grouped within Framework Programmes in order to provide some coherence. This approach also emphasizes the necessity of clustering of universities, firms and public institutions across Europe, in order to create transnational networks of innovation and also to consolidate the technical and scientific capacity necessary to spin off new productive activities at local level (see Chapters 13 and 14).

‘Old’ industrial policy to ‘new’ industrial development policies

19

The need for a public regulation therefore emerges again. In particular, guaranteeing access rights and property rights without limiting competitive dynamics appears to be very important. The new challenges to anti-trust are discussed both in the USA and in Europe. In the USA, the long debate on contestability is back to the fore with spectacular cases such as Eastman Kodak, Microsoft and Intel. Anti-trust policy revives and intensively uses game theory (especially in the case of Eastman Kodak against Image Technical Service in 1992). The emphasis on innovation increases the interest in and attention to the institutions that generate and accumulate innovation, such as universities, the training system and the public and private research centres, and therefore in the rules aimed at promoting, guaranteeing and transferring innovation. The systemic character of these institutions is stressed and innovation systems are defined as sets of public and private, national and local, institutions that design both cooperative and competitive innovative relations (national innovation systems, see Freeman, 1987; Lundvall, 1992; Metcalfe, 1995). In the same manner the rediscovery of the multiplicity of organizational forms and therefore of the cooperative and competitive relationships that are built between firms induce a necessity to rediscover the territory as a social and cultural support for clusters that, in order to be stable, require constraints of trust that cannot be summarized in price negotiations only. During the uncertainty and the problems of the 1980s, indeed, the crisis of large firms led to the emergence of new organizational models. Models of quasi-markets were rediscovered, where groups of firms bounded by trust constraints acted on highly innovative and flexible dynamics (Piore and Sabel, 1984). The analysis of Italian industrial districts points to cultural and social parameters as determining the strength and the advantages of the linkages existing between the firms in highly innovative production systems (Becattini, 1979, 1990, 2002; Brusco, 1982, 1989); the same analysis is extended to very dynamic contexts such as the film industry (Storper, 1997) and is used in taxonomies of clusters which exist in very diverse environments (Porter 1998a, 1998b). Both discoveries give importance to the historical context in which economic relations are built. A new evolutionary political economy approach pays more and more attention to the social dynamics which support economic change (Amin and Thrift, 1994). A new economic geography redesigns the relationships between countries, raising issues that seemed forgotten, such as the difference between a nation and a region (Krugman, 1995). The debate on regional development merges with the debate on local development and a new vision emerges whereby not only firms but also territories are in competition in the global economy. The importance of productive clusters in the competitiveness of territories is pointed out and discussion therefore focuses on the ways to create agglomerations where the historical conditions have not endogenously generated these productive organizational forms. The importance of the coherence between the decentralized industrial policies and the policy framework provided by the state is also stressed (Bianchi, 1994; Bianchi et al., 1998). The two discoveries mentioned above also give importance, once again, to the organization of production as the unit of analysis of industrial economics, hence the attention to industrial organization (Tirole, 1988; Milgrom and Roberts, 1992). The 1980s and the 1990s therefore seemed to be years in which the consolidation of the neomarginalist orthodoxy was reducing economic analysis to abstract and stylized

20

International handbook on industrial policy

forms, whereas in fact this new orthodoxy was providing new vigour and strength to a new thinking on society and industrial organization. In fact, here again, the evolution of the real economy was driving changes in intellectual approaches. The new industrial development policies and the role of the state The experience of industrial policy practice around the world reviewed in the Handbook, show that the new approach to industrial policy that has emerged in the EU can be generalized. The new industrial development policies consist in a variety of measures that aim at favouring firms’ structural adjustment and helping the development of new sectors. For this purpose, the conditions (framework) for business has to be built and function well, but the government can also play a role of ‘pathfinder’, ‘catalyst’ (European Commission, 1990) or of ‘athlete’, ‘coach’ and ‘referee’ (Chu et al., 2001), according to the stage of development of the country. Chapter 4 shows that East Asian governments were very active in orientating and promoting development at early stages (building institutions, rules and markets, and favouring the growth of specific sectors by concentrating resources there). As development proceeds, however, governments become ‘coach’, that is, ensure rules and adapt more horizontal (functional) measures, that apply to the whole industry and are not specific to some sectors. Both developed and developing countries implement measures to favour the development of high-tech sectors which are considered as ‘strategic’ in terms of growth potential and spillovers to other industries (R&D programmes with subsidies, creation of research institutions, science parks, and so on; training of scientists and engineers; public orders; and so on). Both developed and developing countries also support weaker actors, namely SMEs (see Chapter 19) or specific territories that may be backward (Chapters 16 and 17) and also crisis sectors. State aids, in the form essentially of subsidies, although inefficient from an economic point of view and illegal in most countries, are still widely used (see Chapter 7 on the cases of the EU and the USA). The various measures can be grouped into two broad types, rules and capabilities. Rules represent the interventions that design the institutional set-up of the system, the rules of the competitive game. They include anti-trust legislation, regulation of sectors (network industries or utilities in particular), protection of intellectual property rights, regulation of products and regulation of labour markets. Regarding anti-trust, a number of issues emerge among which the consideration of dynamics in cases. Both Chapters 6 and 8 highlight this point. As shown by these chapters and by Chapters 20 and 21, game theory has been essential in providing a rigorous tool on which to base judgments. The development of the dynamic game-theoretical framework appears to be promising in allowing us to take market dynamics into account, although more research is needed, as stressed by Cellini and Lambertini. Alternatives to traditional intellectual property protection such as patents are being developed, in particular the open source model (Chapters 11 and 12). This model is based on a particular juridical tool, namely the ‘General Public Licence’ or GPL, whose principle is to persuade its adopters to disclose the source code of the concerned programmes and of any further improvement if they circulate them, as well as the free circulation of the code under the sole condition of maintaining its ‘open’ character. New knowledge is immediately available in the public domain, but any agent that uses it has also to make his

‘Old’ industrial policy to ‘new’ industrial development policies

21

inventions available publicly. As stressed by Jullien and Zimmermann, the GPL model favours the constitution of collective knowledge creation, which is essential in the globalized or intangible world as we intend it. Capabilities are actions aimed at providing economic agents with the capacity to take part and perform well in the competitive game. They represent the way in which collective mechanisms for the definition of growth opportunities are activated. They are actions aimed at favouring the availability of tangible (finance, infrastructure, and so on) and, especially, intangible assets (knowledge, human capital): policies to develop human capital, to create knowledge; that is, to innovate. They are implemented at various levels: local because linkages appear to be better created at the local level (see Chapter 17), but also at a higher level because they are spillovers and economies of scale to be realized (R&D, human capital mobility and so on). Coherence between the various measures and between the various levels is provided in part by creating linkages between various agents aimed at generating the conditions for growth of an entire production system. If therefore the definition of ‘industry’ is again the organization of production, the industry nowadays increasingly involves integration between manufacture and services, because the creation of value is increasingly determined by intangible assets related to knowledge, as we stressed above and as De Bandt develops in detail in Chapter 5. Therefore service activities and in particular information services are essential for development because they make a bridge between diverse activities thanks to the transfer of knowledge between sectors (Chapter 5). Business support services are important in contexts in which new productive firms are clustering, although evidence on their effective implementation and on their effect is hard to find (Chapter 18). The increasing importance of intangible assets is even greater in high-tech sectors, where innovation is linked to basic research, such as the life science sectors (Chapter 15) or ICTs (Chapters 11, 12 and 13). These sectors have represented a focus of the new industrial development policies because they are a major source of growth and competitiveness, as Giarratana and Torrisi stress. In these sectors, the capacity to generate intangible assets is primarily determined by territorial clusters and linkages (Chapters 13 and 14). Knowledge has been shown to be more effectively communicated and created in local networks, especially when knowledge is of the tacit (non-codified) kind. The deep reason for this effectiveness appears to be the embeddedness of relationships into local communities, that share norms and values (social capital) that ease communication. The importance of the embeddedness of industrial activities in territories is also stressed in Chapter 17. The importance of networking for innovation and knowledge diffusion recognized by policy makers and realized in implemented policies has also taken the form of promoting relationships between all actors of innovation systems, including universities and other research centres and institutions of various types. The role of universities has changed profoundly over the last 20 years. From being essentially providers of knowledge to students, universities have taken an active role in knowledge creation and diffusion, building relationships with firms in order to understand their competence requirements and also commercializing innovation through spin-offs. University spin-offs have indeed been an important and increasing phenomenon (Chapter 13). The end of the political division of the world into two blocks and the entry of new large countries in the worldwide competitive arena and integration processes of large areas such

22

International handbook on industrial policy

as Europe have led not only to the redefinition of the extent of the market on which firms compete, but also to the redesigning of the frontiers of the state and therefore the extent of the rules within which firms could define their strategies. Globalization is therefore not only the intensification of exchange within an increasingly integrated world economy, but also a complete change in the reference values that structured the world economy from the end of World War II onwards. The declining possibility of the state to act as the ultimate reference of sovereignty implies the emergence of new subjects able to formulate and implement industrial development policies: the cities, the regional governments and the international agencies. Hence a multiple consensus has to be found in order to formulate and implement policies that concern firms that are localized in territories with historical specificities but open to international competition. Policy making is no longer an action decided by a central government to limit or subsidize individual behaviour but rather a programme that involves all the institutions of a territory and that is aimed at consolidating an environment able to increase the collective competitiveness and therefore to stimulate the latent innovation capacity. It addresses the dynamics of the economy by aiming at favouring the development of industry and its structural adjustment when needed. Rodrik (2004) in a similar way argues that ‘the right model of industrial policy is not that of an autonomous government applying Pigovian taxes or subsidies, but of a strategic collaboration between the private sector and government with the aim of uncovering where the most significant obstacles to restructuring lie and what types of intervention are most likely to remove them’ (p.3). The role of local and regional administrations therefore changes. They were subordinates to the national and central level; they are now active in the decision making of policies aimed at attracting investment and at developing the local endowment of intangible assets (human capital and knowledge). Thus globalization is characterized not only by the intensification of competition between firms, but also by the competition between territories and between production systems. Hence the rich local dimensions of industrial policy, as stressed by di Tommaso and Bellandi in their chapter. It appears that, initially, in the late 1980s and 1990s, decentralization of industrial development policies was advocated (essentially by scholars, but also by some international organizations such as UNIDO) as being the best solution to foster development or help developed countries resolve their competitiveness problems (Europe in particular). Hence ‘bottom-up’ approaches were advocated, where action was essentially at local level while the national level (or above) would coordinate actions and favour the exchange of information and experiences across territories (Bianchi et al., 1998). The chapters of the Handbook show that in practice actions at local level have been widely implemented but without sufficient centralization: Sepulveda and Amin (Chapter 16) show this for the case of Argentina; Lall (Chapter 4) shows that the most successful East Asian countries have been those that have implemented national strategies or programmes aimed at the development of the country and including some local actions too (for instance, orchid production in Taiwan, discussed by Rodrik, 2004). Chapter 14 also raises the question of the impact of locally limited experiences such as science parks on the rest of the economy and argues that the diffusion of global production and value creation processes implies the need to favour the integration of local experiences into such global chains. The old-type policies are not entirely abandoned either. As we stressed earlier on, it appears that protectionist policies are effective at some stages of development. State

‘Old’ industrial policy to ‘new’ industrial development policies

23

ownership has also been extensively used in both developed and developing countries at industrialization stages (or reconstruction for the latter states), as shown in Chapter 10 and as shown by the historical review of industrial policy in the first part of this chapter. A country’s specializations are therefore not determined by a given resource endowment only; they are also determined by learning processes and by competencies that are developed mainly at local levels (capabilities to develop own intangible assets and to attract intangible assets such as knowledge and technology, and human capital), and that can be favoured by government’s policy, as shown in Chapter 2, by Arora and Gambardella. Arora and Gambardella show in Chapter 2 that the policies implemented in India in the area of Bangalore (but also in four other countries, Brazil, China, Ireland and Israel) in the software sector have brought positive effects not only locally but also in the USA, because they have stimulated the competition between firms and the cooperation between institutions inducing human capital mobility that has accelerated the diffusion of knowledge and therefore of innovation. In this example the industrial development policies implemented by developing countries aimed at high-skill and high-tech sectors have changed the international competitive advantages, at the expense of Europe, which does not seem to be fully aware of this dynamics. As a consequence, globalization (as defined here) implies the irrelevance of the old debate on industrial policy and the definition of industry. The new industrial policies are actions that governments implement in order both to define the rules and to help develop capabilities. Thus, for example, in the EU, the new approach to policy outlined earlier (pp.12–14) aims at providing the appropriate framework for business to be competitive, and this means both providing the rules (anti-trust, property rights) and also implementing structural policies that help backward regions and weaker actors (such as SMEs) to develop appropriate capabilities. Structural policies also include the important actions that help the creation and diffusion of knowledge (technology and research policies). The new industrial policies are therefore effective in generating the growth of developing countries. Thus Lall in Chapter 4, shows that what we learn from the East Asian experience is that the development process accelerates only if the industrial core of the country can be made dynamic. This occurs with national and concerted programmes. Many countries in Latin America have had difficulties. Chapter 19 shows that, at least concerning SMEs, the problem with Latin America is not that no industrial policies have been implemented, but that they have been implemented incoherently, as they were not part of a national strategy of industrial development. Chapter 16 also shows in the case of Argentina the lack of action at the national level that undermines the effect of local industrial development policies. As pointed out by Rodrik (2004), ‘the challenge in developing countries is not to rediscover industrial policy, but to redeploy it in an effective manner’ (p.5). There has been a debate on whether industrial policy has determined the rapid development of East Asian economies such as Taiwan, Korea and Singapore. The debate has focused on the role of selective industrial policies (specific to sectors), as shown by Chang (2001). Our definition of industrial development policies implies a new view of the debate: what is relevant is not whether selective or functional policies are useful but what is the appropriate mix of measures to determine industrial development. Industrial development

24

International handbook on industrial policy

policies are programmes or strategies consisting of a variety of measures (both rules and capabilities), the mix of which is specific to a country’s conditions and historical experience. Conclusions Industrial policy is a field of economic research which is and has long been very rich in approaches, methods and empirical fields of observation. Both the practical experience in countries in all parts of the world and the literature reviewed in the Handbook show the richness of the field and the fertility of research. What emerges from the present volume and from our historical analysis can be summarized in two major points. First, ‘globalization’ has led to the definition and implementation of new industrial development policies, where globalization is considered as the numerous economic, technological and political changes that have occurred over the last 20 years or so. These changes have induced a change in the extent of the market of many firms, that is, of competitive conditions, that have implied their necessary restructuring to adapt to changes. Firms have often called for industrial policies in order to help these adaptations. Second, the industrial policies defined nowadays do not represent a return to the old-type industrial policies that aimed at picking winners and directly intervening in markets to correct once and for all market failures. The ‘new’ industrial policies consist of large sets of measures that aim at providing the appropriate framework for industrial development to take place, via both the creation of firms and sectors (creation of comparative advantages) and the structural change of existing firms (structural change). This does not mean, however, that old-type industrial policies have completely disappeared. The amounts of state aid spent by governments throughout the world show that direct support is still there. In addition, it appears that old-type industrial policies are particularly useful at earlier stages of development. The difficulty for governments nowadays is to define the set of measures in coherent ways, so that their effects converge towards the same point. Coherence involves two aspects. First, there is coherence between the various measures, that is, measures defining rules and measures promoting the development of capabilities, so that the whole is a synthesis of the various actions rather than a sum where interaction effects are not taken into account. Second, coherence must also be ensured between the different levels of intervention, namely local, regional, national and supranational. The trade-offs between the costs and benefits of centralization have been identified for a long time (see Neven and Röller, 2000), but without considering both the complexity of the new production systems and the creation of global value creation processes (as stressed by Annerstedt in his chapter). We have shown in the present chapter that a clear tendency of industrial development policies has been to ‘go local’, but the importance of centralization to guide structural changes and ensure spillovers between local experiences has been too often neglected (both in the case of policies in developing countries and in the case of innovation and knowledge diffusion policies). Consistency over time is also essential. Another difficulty is that these sets of measures that constitute coherent programmes or strategies have to be constantly adapted because industrial development is inherently a dynamic process. The need to take account of dynamics is thus stressed in many chapters of the present volume, from those on anti-trust (Chapters 6 and 8) to those on territories (Chapters 17 and 19) and science, technology and innovation policies (Chapter 14).

‘Old’ industrial policy to ‘new’ industrial development policies

25

Overall, we suggest below a (non-exhaustive) number of issues that emerge out of the Handbook. First, developed and developing countries face different problems but with globalization the solutions to problems (the industrial policy measures) are increasingly related (global production systems; exchange of human capital and other complementary intangible assests). Second, there does not appear to exist a general recipe for industrial development; the mix of measures must be adapted to a country’s conditions (availability of resources, stage of development) and historical experience. Third, measures are implemented at various levels of government and more research is needed on how consistency is ensured. Fourth, evaluation of policies is a field of research that has to be developed. Often politicians do not like it since it may criticize some of the policies they have adopted and financed (hence they do not fund evaluation studies); evaluation is in any case a difficult exercise. Fifth, there are many approaches to industrial economics: neoclassical, evolutionary, . . .; all seems to converge in putting the production system at the centre of the analysis and in claiming that a dynamic view is needed. The need to take dynamics into account is (implicitly or explicitly) outlined for example in Chapters 4, 5, 6, 8, 11, 12, 14, 19, 20 and 21. The new industrial development policies are therefore essentially pragmatic. They consist in developing a national programme of industrial development within which local actions are defined, involving all the actors concerned, such as business, universities and other institutions. They are thus no longer specific actions aimed at specific market failures, because market failures are often too difficult to identify and to resolve (information required is complex and incomplete or too many factors are at play to determine with any degree of precision the right level of tax or subsidies). They do not aim at ‘picking the winners’ either, because winners often come out of random evolutionary processes that are impossible to predict. They aim at favouring the development of industrial sectors and the dynamic adjustment of firms to competitive conditions by providing both rules and capabilities. Industrial development policies are generally horizontal (applying to all sectors) but are also vertical in some instance: sectors have specificities and policies have to be adapted to these specificities. These policies therefore comprise a variety of measures, all aiming at providing the appropriate framework for industrial development, and the main problem of policy makers is to ensure coherence. The experience reviewed in the various chapters of the Handbook shows that the best performing countries in terms of development and growth are those where the governments elaborate and implement coherent development strategies or programmes, such as Korea and Taiwan in South East Asia. We think that future research on industrial policy should focus on the above issues and particularly on the issue of how to ensure the coherence and consistency of industrial development policies. Notes 1. This expression is the title of a book published by a French journalist and that became very famous, Le Défi Américain (The American Challenge), in 1969. 2. COM(2004) 274 final, 20.04.2004, Brussels. 3. COM(2002) 714 final, 11.12.2002, Brussels. 4. COM(2003) 704 final, 21.11.2003, Brussels.

26

International handbook on industrial policy

5. Such influence is also demonstrated by Pitelis in Chapter 22 of this volume. 6. This is a seminal paper of endogenous growth theory.

References Alchian, A.A. and H. Demsetz (1972), ‘Production, information costs and economic organisation’, American Economic Review, 62. Amin, A. and N. Thrift (eds) (1994), Globalization, Institutions, and Regional Development in Europe, Oxford University Press. Audretsch, D.B. (2003), ‘Introduction’, in D.B. Audretsch (ed.), SMEs in the Age of Globalization, Cheltenham, UK and Northampton, MA, USA: Edward Elgar. Barro, R.J. (1991), ‘Economic growth in a cross-section of countries’, Quarterly Journal of Economics, 106, 407–43. Barro, R.J. and J.W. Lee (1993), ‘Losers and winners in economic growth’, NBER working paper, no. 4341. Becattini, G. (1979), ‘Dal settore industriale al distretto industriale: Alla ricerca dell’unita d’analisi dell’economia industriale’, Rivista di Economia e Politica Industriale, 1, 7–21. Becattini, G. (1990), ‘The Marshallian industrial district as a socio-economic notion’, in F. Pyke, G. Becattini and W. Sengerberger (eds), Industrial Districts and Inter-firm Cooperation in Italy, Geneva: International Institute for Labour Studies. Becattini, G. (2002), ‘Industrial sectors and industrial districts: tools for industrial analysis’, European Planning Studies, 10(4), 483–93. Bianchi, P. (1994), ‘Industrial strategy in an open economy’, in P. Bianchi, K. Cowling and R. Sugden (eds), Europe’s Economic Challenge. Analyses of industrial strategy and agenda for the 1990s, London and New York: Routledge. Bianchi, P. and S. Labory (eds) (2004), The Economic Importance of Intangible Assets, London: Ashgate. Bianchi, P., K. Cowling and R. Sudgen (1998), Industrial policies and economic integration: learning from European experiences, London and New York: Routledge. Brusco, S. (1982), ‘The Emilian model: productive decentralisation and social integration’, Cambridge Journal of Economics, 6. Brusco, S. (1989), ‘A policy for industrial districts’, in E. Goodman and J. Bamford (eds), Small firms and industrial districts in Italy, London: Routledge. Caballero, R. and A. Jaffe (1993), ‘How high are the giant’s shoulders: an empirical assessment of knowledge spillovers and creative destruction in a model of economic growth’, NBER working paper no. 4370. Cecchini Report (1985), Brussels: European Commission. Chang, H-J. (2001), ‘Rethinking East Asian industrial policy – past records and future prospects’, in P-K. Wong and C-Y. Ng (a cura di), Industrial Policy, Innovation and Economic Growth: the Experience of Japan and the Asian NIEs, Singapore: Singapore University Press, pp. 55–84. Chu, Y-P., T-J. Chen and B-L. Chen (2001), ‘Rethinking the development paradigm: lessons from Taiwan – the optimal degree of state intervention’, in P-K. Wong and C-Y. Ng (a cura di), Industrial Policy, Innovation and Economic Growth: the Experience of Japan and the Asian NIEs, Singapore: Singapore University Press, pp. 197–244. Cipolla, C.M. (1980), Storia Economica d’Europa. La rivoluzione industriale, Turin: UTET. Cowling, K., C. Oughton and R. Sugden (1999), ‘A reorientation of industrial policy? Horizontal policies and targeting’, in K. Cowling (ed.), Industrial Policies in Europe. Theoretical Perspectives and Practical Proposals, London and New York: Routledge. Dahrendorf, R. (1990), Il conflitto sociale nella modernità, Rome: Laterza (original edition, 1987). Dosi, G. (1991), ‘Perspectives on evolutionary theory’, Science and Public Policy, 18, 353–61. Dosi, G. and L. Marengo (1994), ‘Towards a theory of organisational competences’, in R.W. England (ed.), Evolutionary Concepts in Contemporary Economics, Ann Arbor: Michigan University Press. European Commission (1990), ‘Industrial policy in an open and competitive environment’, (Bangemann Memorandum), Working paper, 14 December, Brussels. Fagerberg, J. and B. Verspagen (2004), ‘Innovation, growth and economic development: why some countries succeed and others don’t’, Centre for Technology, Innovation and Culture, working paper no. 02/04, Oslo, Norway. Freeman, C. (1987), Technology Policy and Economic Performance: Lessons from Japan, London: Pinter. Hirschman, A.O. (1963), The Strategy of Economic Development, New Haven: Yale University Press. Jensen, M.C. and W.H. Meckling (1976), ‘Theory of the firm: managerial behaviour, agency costs and ownership structure’, Journal of Financial Economics, 3(4). Krugman, P. (1995), Development, Geography and Economic Theory, Cambridge, MA: MIT Press. Lall, S. (2004), ‘The employment impact of globalisation in developing countries’, in E. Lee and M. Vivarelli (eds), Understanding Globalization, Employment and Poverty Reduction, Basingstoke: Palgrave Macmillan.

‘Old’ industrial policy to ‘new’ industrial development policies

27

Landes, D.S. (1969), The Unbound Prometheus, Cambridge: Cambridge University Press. Landes, D.S. (1978), Prometeo liberato. Trasformazioni tecnologiche e sviluppo industriale nell’Europa occidentale dal 1750 ai giorni nostri, Milan: Einaudi (Italian translation of Landes, 1969). Lewis, W.A. (1954), ‘Economic development with unlimited supplies of labour’, Manchester School, XXII, 139–91. Lucas, R.E. (1988), ‘On the mechanics of economic development’, Journal of Monetary Economics, 22, 3–42. Lundvall, B.A. (ed.) (1992), National System of Innovation. Towards a Theory of Innovation and Interactive Learning, London: Pinter. Marshall, A. (1919), Industry and Trade, London: Macmillan. Metcalfe, S. (1995), ‘The economic foundations of technological policy. Equilibrium and evolutionary perspectives’, in P. Stoneman (ed.), Handbook of the Economics of Innovation and Technological Change, Oxford: Blackwell. Milgrom, P. and J. Roberts (1992), Economics, Organisation and Management, Englewood Cliffs, NJ: PrenticeHall. Myrdal, G. (1957), Asian Drama: an Inquiry into the Poverty of Nations and the Challenge of World Poverty, Allen Lane: The Penguin Press (1968). Nelson, R.R. and S.G. Winter (1982), An Evolutionary Theory of Economic Change, Cambridge: Belknap Press of Harvard University Press. Neven, D. and L-H. Röller (2000), The Political Economy of Industrial Policy in Europe and the Member States, Berlin: Edition Sigma. North, D. (1990), Institutions, Institutional Change and Economic Performance, Cambridge and New York: Cambridge University Press. Nurske, R. (1953), Problems of Capital Formation in Developing Countries, Oxford: Blackwell. Pelkmans, J., S. Labory and G. Majone (2000), ‘Better EU regulatory quality: assessing current initiatives and new proposals’, in G. Galli and J. Pelkmans (eds), Regulatory Reform and Competitiveness in Europe, vol. I, Cheltenham, UK and Northampton, MA, USA: Edward Elgar. Piore, M. and C. Sabel (1984), The second industrial divide, New York: Basic Books. Poggi, G. (1992), Lo Stato, Bologna: Il Mulino. Porter, M. (1990), The Competitive Advantage of Nations, London: Macmillan. Porter, M. (1998a), ‘Location, clusters and the “new” microeconomics of competition’, Business Economics, 33(1), 7–17. Porter, M. (1998b), ‘Clusters and new economics of competitiveness’, Harvard Business Review, December, 77–90. Rodrik, D. (2004), ‘Industrial policy for the 21st century’, paper prepared for UNIDO. Romer, P. (1986), ‘Increasing returns and long-run growth’, Journal of Political Economy, 94, 1002–37. Romer, P. (1990), ‘Endogenous technical change’, Journal of Political Economy, 98(5), S71–S102. Solow, R.M. (1956), ‘A contribution to economic growth theory’, Quarterly Journal of Economics, 70(1), 65–94. Stiglitz, J.E. (2002), Globalization and Its Discontents, New York and London: W.W. Norton & Company. Stoneman, P. (ed.) (1995), Handbook of the Economics of Innovation and Technological Change, Oxford: Blackwell. Storper, M. (1997), The Regional World: Territorial Development in a Global Economy, New York: Guilford Press. Teece, D.J., R. Rumelt, G. Dosi and S. Winter (1994), ‘Understanding corporate coherence: theory and evidence’, Journal of Economic Behaviour and Organisation, 23, 1–30. Tirole, J. (1988), The Theory of Industrial Organisation, Cambridge, MA: MIT Press. Williamson, O. (1975), Markets and Hierarchies: Analysis and Antitrust Implications, New York: The Free Press. Williamson, O. (1985), The Economic Institutions of Capitalism, New York: The Free Press. Williamson, O. and S.G. Winter (eds) (1993), The Nature of the Firm. Origins, Evolution and Development, Oxford: Oxford University Press.

2

Emerging issues in the new economy and globalization Ashish Arora and Alfonso Gambardella

1 Introduction The 1990s have raised new issues for industrial development, and hence for industrial policy, following the opportunities opened up by information technology (IT) and the socalled ‘new economy’. In this chapter we look at one phenomenon observed at the global level which we employ as a case for discussing some broader questions. This is significant growth of software production in some emerging economies. Its peculiarity is that such a spectacular development has occurred in regions where one would have hardly predicted the rise of human capital-intensive activities. We will discuss software growth in five countries – Brazil, China, India, Ireland and Israel – following a two-year research that we coordinated on this topic (Arora and Gambardella, 2005a). As we shall see, the phenomenon is certainly striking. For example, a country like India had practically no software production in the late 1980s and it is now a leading world exporter in this industry. Most importantly, the story of software growth in these regions exemplifies some relevant new questions for industrial development and policy. As a basis of our analysis we rely on some recent innovative contributions on industrial policy and related subjects. Gomory and Baumol (2000) have argued that the theory of comparative advantages, as a basis for understanding trade and specialization across countries, is at best indeterminate. We will discuss their contention that, even if we restrict the potential specializations of a country or region to fields in which they have comparative advantages, we are still left with so many options that it is difficult to predict where the country or region will actually specialize. Rodrik (2004) and Hausmann and Rodrik (2003) make a similar point. They too argue that there are many potential outcomes of the search for the right specialization. Therefore a country or a region has to single out its own specializations or comparative advantage through a process of ‘self-discovery’. By the latter they mean that the country or region has to find out its profitable industries or opportunities by experimenting with different potential fields through entrepreneurs and entrepreneurial search. Finally, we adopt Klepper’s (2004) perspective that a good deal of the competencies to form new industries or areas of activities arise from previous capabilities in similar realms. That is, the process of industrial development is not made of jumps and discontinuities. New successful industries or firms arise from previous successful industries or firms through related diversification. In this respect Klepper’s framework also suggests that the competencies created within firms, and particularly within successful firms, play a notable role in fostering industrial growth. Against this background our software industry stories enable us to discuss three issues in the context of the new economy and globalization. First, the growth of the software industry in our five countries hits head-on the widely debated topic of outsourcing. The fear of job and competitiveness loss in the advanced countries, along with other consid28

Emerging issues in the new economy and globalization 29 erations, means that the subject is ripe for a better understanding of the industrial policy implications for both the advanced and the developing economies. Second, our stories illustrate some effects of the growth in the supply of human capital in some developing economies and the implications for the developed world. We will discuss industrial policy on this subject for both the advanced and the developing countries. Finally, we compare the potential responses of the USA and Europe. We will argue that not only does the USA appear to be better fortified to cope with the growth of human capital-intensive activities in developing economies, but it is also better equipped to take advantage of fruitful complementarities with them. Our conclusions will provide additional remarks about the industrial policy consequences that arise from these three issues, by taking the Hausmann–Rodrik, Gomory–Baumol and Klepper frameworks and using them to elaborate on potential policy actions in this context. The next section briefly presents the Gomory–Baumol, Hausmann–Rodrik and Klepper frameworks. Section 3 illustrates the spectacular development of software in our five countries. Section 4 discusses whether these patterns are beneficial for the advanced economies and particularly for the USA and Europe. Section 5 examines whether they are beneficial for the developing countries. Section 6 concludes by discussing emerging industrial policy implications and some related actions. 2 Industrial policy for the 21st century Gomory and Baumol (2000) argue that the industrial specializations of countries in the 21st century are no longer based on natural endowments, but are determined by processes of learning, education and so on. This produces a fundamental novelty. When the domestic specializations depended on the endowment of immobile resources, there was a relatively narrow range of activities that could be effectively developed: those that were gauged by the resource endowments. By contrast, if the specializations depend on learning, the acquisition of skills and other factors that are not tied to what is available locally, many more activities can be developed. For example, a country rich in some mineral resource can become an international specialist in its extraction and possibly in industrial productions that make use of it. But when specializations are acquired via learning, education or skills, there is no reason to confine the country’s trajectories in one direction or another. In principle, anything can be learned. This influences the benefits of the trade relations between countries. The question that Gomory and Baumol ask is how a country will be affected if a trading partner becomes more productive in a given industry. They argue that, if a country produces a vast range of goods compared to its trade partner, both the country and the partner will gain. This is because the countries can have very heterogeneous productivity levels across industries. Hence, if a country produces many goods, it will allocate resources to industries in which it is very productive but also to industries in which it is not productive. This is inefficient because the country has to ‘waste’ some of its resources in industries in which it is not productive. If the partner becomes more productive in some domain, there is a chance that it will be in one of the activities in which the former country is not productive. The first country will then buy the good from the partner, will redeploy the saved resources onto domestic activities in which it is more productive and may gain additional efficiency because of its greater specialization. The trade partner will of course gain because it has become more productive in some of its industries.

30

International handbook on industrial policy

By contrast, if the trading partners are already quite specialized, already focus on the industries in which they are most productive, sell these goods both domestically and to the other country and buy the other goods from the latter, then an increase in the productivity of the partner in an industry in which the original country is operating can hurt it. This is because, if the first country now buys from the more productive partner, or the more productive partner produces for its domestic market and no longer resorts to the other country, the latter loses domestic industrial activities in which it employed resources efficiently. As resources are moved to the remaining domestic industries, the advantages from the increase in specialization are minor because at the aggregate level the gains from specialization were already high. Moreover, because the deployed resources were allocated productively, the gains from the additional specialization have to be high in order to compensate for the lost productivity of the very same resources. Thus Gomory and Baumol argue that, when two countries are unbalanced in terms of levels of growth and international trade volumes, they can both gain if the less advanced partner becomes more productive in some activities. By contrast, if two countries are at a comparable level of development, then, as a country becomes more productive, it will gain at the expense of the other. Put simply, an advanced nation gains from a more productive developing country partner, but there will be conflicts between two advanced countries if one of them becomes more productive in some activity. A key aspect of Gomory and Baumol’s argument is that the particular industries in which the countries specialize are hard to predict ex ante. Like Gomory and Baumol, Hausmann and Rodrik (2003) note that what makes the new industries unpredictable is that, unlike resource endowments, assets like learning or competencies (technological, organizational, managerial) can be applied to many realms. Comparative advantages do not disappear. For example, we shall see that the success of the Indian software industry rests in part on its wide supply of university graduates. But comparative advantages define at best a (large) set of potential specializations. The specializations that are actually pursued cannot be easily singled out. In principle, there were many potential areas where a human capital-intensive production, based on a large supply of academic graduates, could be fruitfully employed in India. Likewise, Hausmann and Rodrik ask why Bangladesh exports garments such as hats but not bed sheets, while Pakistan exports bed sheets but not hats. More generally, they point out that in developing countries the bulk of exports consist of a narrow range of products. Typically, a small proportion of products account for 60–80 per cent of their total exports. The authors then note that, as the Bangladesh–Pakistan example illustrates, similar countries end up specializing in different products within similar industrial activities and the specific products in which they specialize cannot be predicted by comparative advantages. According to Hausmann and Rodrik (2003) and Rodrik (2004), the way in which a country can find out its specializations is by a process of self-discovery or experimentation. This means encouraging entrepreneurship. In a very Schumpeterian way, they think of entrepreneurs as individuals who try new avenues, experiment with new industrial activities, produce innovations and so on. Moreover, by encouraging a large number of trials (via many entrepreneurs) society as a whole can experiment with many trajectories and find out which trajectories are more fruitful than others. The problem is that the private returns of the entrepreneurs are lower than those of the society. This is because, even if the entrepreneurs fail, society can gain. For example, failure provides information

Emerging issues in the new economy and globalization 31 that an industrial activity is unsuccessful or very costly and therefore advises others not to pursue it. Also entrepreneurial failures may nonetheless entail production of valid learning, information and the like, which may benefit second-comers. Thus, while entrepreneurship is the way to discover one’s specializations, it may be undersupplied, a point also made by Levinthal and March (1993). Hausmann and Rodrik (2003) and particularly Rodrik (2004) thus maintain that a key element of an ‘industrial policy for the twenty-first century’ is to support entrepreneurship in the search for successful specializations. Once the specializations are discovered, society will invest larger scale resources in them. This no longer requires industrial policy interventions because, if a direction proves to be successful, the private incentives to invest will be high enough and most likely not lower than those of the society as a whole. The key support, as Rodrik (2004) notes, is on discovery/experimentation, not on exploitation once the discovery is made.1 One potentially misleading feature of Hausmann and Rodrik’s framework is that the successful set of activities for a country is exogenously specified, awaiting discovery by a lucky entrepreneur. Klepper’s (2004) argument suggests that the successful specializations typically arise from related activities in which a country or a region has been successful in the past. To be sure there is a clarification to be made here. Hausmann and Rodrik think primarily of developing countries and particularly of countries in which there has been little or no previous industrialization. Their story is, then, one in which the countries have to find out their specialization, starting almost from scratch. By contrast, Klepper studies the rise of new industries in the developed world, where a good deal of industrial activities has been present for many years. To reconcile these views, and to clarify our perspective, the more a region is poor and industrially underdeveloped the more one expects the successful avenues to be exogenously determined because there are no previous comptencies on which to build. But if there are pre-existing competencies, one expects them to affect the rise of new opportunities. In this respect the pre-existence of competencies does not mean that the realm of activities is more restricted, as with given resource endowment. For one reason, as we shall also see from Klepper’s industry examples below, the reliance on previous competencies is not the same as similarity in products. Competencies are often general with respect to specific products, industries or activities. For example, competencies in organizing and managing firms or basic technological competencies can be applied to many fields. Moreover, competencies are a source of innovation, not just technological but also entrepreneurial. They are the basis for discovering new industries and product opportunities and for opening up new fields or activities. This is exemplified by Klepper’s stories on the genesis of four major US industries: TV receivers, automobiles, tyres and lasers.2 Klepper collected systematic data on the origins of the early firms in these industries and the evolution of firms and industries for some decades after the industries birth. The successful TV producers arose as a process of diversification from the leading US radio producers. The early automobile and tyre producers also emerged as diversifiers from related industries (for example, carriage and wagons or bicycles for cars and rubber for tyres). But in these two cases the successful producers that emerged soon after the birth of the industry were spin-offs from early successful car and tyre makers. This is not a contradiction of our more general point. Whether the new industries come from diversifiers or from successful early entrants, our point is that the pre-existence of competencies is an important factor in determining the directions of future specializations.

32

International handbook on industrial policy

There are some related remarks. First, Klepper emphasizes that the successful motor or tyre spin-offs or the successful TV producers came from successful motor or tyre firms or successful radio producers. That is, it is the successful firms, which hoard successful competencies, that spawn successful new competencies and therefore successful new firms or industries. In other words, it is not the presence of an industry or firms per se but of successful industries or firms that has to be looked for as a source of new industrial specializations. Second, Klepper shows that these patterns explain the location of industries. Because new firms arise from existing competencies, it is most natural to expect that they will locate where the parent competencies were located. Finally, the laser industry exhibits a slightly different pattern in that it remained fragmented over time and did not experience a shake-out leading to concentration in a few firms after some time. Klepper explains this by the fact that laser as a product is highly differentiated, suggesting that it relies on distinct submarkets. This entailed several sources of differentiated competencies with no drive for concentrating them in few firms. Relatedly, this did not lead to geographic concentration because the industry was scattered geographically when it started and it remained pretty much so because there was no force pushing for its convergence. Nonetheless, many successful laser firms did emerge as spin-offs of successful parents. 3 The software industry in Brazil, China, India, Ireland and Israel Basic facts During the 1990s, India, Ireland and Israel emerged as significant software exporters. In the same period, Brazil and China also developed an extensive software sector relying largely on their domestic market and they are now attempting to move to exports. Table 2.1 reports some estimates of the growth rates of the software industry in these countries and of their export shares. The table documents the spectacular growth of the industry in these countries and confirms that, while the 3Is (India, Ireland and Israel) rely on exports, Brazil and China have exploited their large domestic markets. Table 2.2 shows some additional figures for these countries and, by way of comparison, for the three leading software producers in the advanced world: the USA, Germany and Japan. From Table 2.2, the industry in our five countries is sizable, even though still smaller than in the three advanced nations. The success of the industry in the 3Is, Brazil and China is implied by the fact that they were non-existent in the late 1980s. Thus an industry of the size observed today was formed over a decade or so. This is the consequence not only of the high Table 2.1 Countries Brazil China India Ireland Israel

Brazil, China and the 3Is: software industry growth and export shares Average growth in the 1990s (%)

Exports as % of sales (2002 or latest available year)

20 30–35 40 20 20

1–2 11 80 85 70

Source: Arora and Gambardella (2005a).

Emerging issues in the new economy and globalization 33 Table 2.2 The software industry in Brazil, China and the 3Is and, by comparison, in the USA, Japan and Germany (2002 or latest available figures) Countries Brazil * China India Ireland (MNE) Ireland (domestic) Israel * USA Japan ** Germany *

Sales ($ bn)

Empl. (000s)

Sales/Empl. (000s)

Software sales/ GDP (%)

Software development index

7.7 13.3 12.5 12.3 1.6 4.1

160 ** 190 ** 250 15.3 12.6 15

45.5 ** 37.6 ** 50.0 803.9 127.0 273.3

1.5 1.1 2.5 10.1 1.3 3.7

0.22 0.23 0.96 0.34 0.04 0.17

2.0 2.0 2.2

0.05 0.08 0.09

200 85 39.8

1024 534 300

195.3 159.2 132.7

Notes: * = 2001; ** = 2000. MNE is the Irish figures for the multinational firms located in Ireland, while domestic is the figures for the domestic Irish firms. The sales figure for the MNE in Ireland might be inflated because many MNE account their European sales in Ireland for tax purposes. The software development index is the ratio between software sales over GDP (in per cent) and the GDP per capita of the country (in US$ 000s) (See also Botelho et al., 2005). Source: Arora and Gambardella (2005a).

average growth rates of the 1990s, reported in Table 2.1, but also of the fact that two-digit growth rates are still common in these countries (particularly in India) in very recent years. Moreover, the notable growth of software in these countries has made this industry an important new specialization in their overall economy. India and Israel have ratios of software sales to GDP even higher than those of the three leading nations (2.5 per cent and 3.7 per cent against 2.0 per cent and 2.2 per cent for the leading countries). The share is lower for the Irish domestic firms, Brazil and China (between 1.1 per cent and 1.5 per cent).3 However, these shares are higher than the share of software production in other advanced economies, such as Italy, the UK or France, at around or below 1 per cent. Moreover, in China, the growth rate of software has accelerated since the end of the 1990s and it has been systematically higher than the growth of GDP. Therefore the Chinese software to GDP ratio is expected to rise further. Finally, the software specialization of these countries is even more apparent if we look at the software development index in Table 2.2, which is the ratio between the software to GDP share and the GDP per capita of the country. This index is orders of magnitude higher for our five countries compared to the three leading nations. They are far more specialized in software than expected, given their overall level of development. Table 2.2 also shows that our countries differ in their sales per employee figures, reflecting qualitative differences in their software productions. Israel exhibits the highest ratio of sales per employee. The figure is even higher than that of the USA, which is the highest among the three advanced economies. As a matter of fact, software in Israel is largely product- and R&D-oriented and it focuses on these higher-end tasks probably even more than the USA, whose large industry comprises many activities including services not at the very top of the value chain. The ratio for Ireland is lower than the advanced countries’

34

International handbook on industrial policy

but higher than that of Brazil, China and India. This is also consistent with the stylized facts. The software industry in Brazil, China and India has lower value added and in India in particular it is heavily service-oriented. In Ireland, there is a group of product-oriented firms and several small consultancy and niche firms. Finally, the growth of the software industry in these countries has relied on a significant degree of entrepreneurship and of industrial as well as firm-level experimentation. In all five cases the development of the industry has been accompanied by a notable rise in the number of firms. In India, the membership of NASSCOM (the Indian software industry association) increased from around 100 in 1990 to 797 in 2000 (Athreye, 2005). Similarly, the number of new Irish software firms increased from less than 300 in 1991 to 760 in 2000 (Sands, 2005). Botelho, Stefanuto and Veloso (2005) report that, of a sample of 685 Brazilian software firms in existence in 2001, a little less than a third (210) were founded between 1996 and 2000, and a slightly larger fraction, 221, were founded between 1991 and 1995. There are no specific figures for the creation of new software firms in China or Israel. However, the existing scattered accounts support the idea of high rates of entry in software in both countries. Human capital and the supply of skills One regularity among our five countries, and particularly in the 3Is, is that they have exhibited an ‘excess’ supply of human capital and specifically an excess supply of engineering and technology graduates relative to their demand from manufacturing and related services. Simply put, these countries have produced more engineers than their hitherto lacklustre industrial sector could absorb.4 This process started before the rise of the software industry. From the mid-1970s to the late 1980s, these countries grew only modestly, while continuing to show a good rise in the number of college graduates, especially in science and engineering, relative to their level of development. Thus, when the IT boom started, they already exhibited a reservoir of underemployed college graduates. The underemployment of the college graduates is crucial here. In the more advanced countries as well as in the rapidly growing Asian countries such as South Korea, Taiwan and Singapore, the science and engineering graduates, whose numbers were also growing, especially in the Asian Tigers, faced a high opportunity cost of working in the software sector. Plentiful job opportunities in industry and in well established firms with good opportunities meant that there would be fewer entrepreneurs setting up software firms and, more generally, the nascent software firms found it difficult to attract talented engineers. Why the 3Is were relatively abundant in technically skilled workers is not well understood, but having such an abundance turned out to be crucial for software success. To be precise, there are two issues here. The first has to do with the level of supply of the relevant human capital. The second is about the elasticity of the supply of graduates in these countries. In this respect the education institutions in all five countries have responded with sizable increases in the number of graduates as the demand for their services rose over the 1990s. For example, the accredited stock of engineers in India increased from around 60 000 in 1987–8 to around 340 000 in 2003, while IT professionals increased from around 25 000 to nearly 250 000. NASSCOM figures also indicate that in India the number of IT graduates increased from 42 800 in 1997 to 71 000 in 2001. By comparison, the number of IT

Emerging issues in the new economy and globalization 35 graduates in the USA increased from 37 000 in 1998 to 52 900 in 2000. During this period the IT workforce (which does not directly correspond to IT degree holders) in the USA was probably eight to ten times larger than the IT workforce in India. Other countries show a similar pattern. Tschang and Xue (2005) report that, in China, the number of IT graduates increased from 29 000 to 41 000 in 1999–2001. Botelho, Stefanuto and Veloso (2005) note that the 18 000 graduates in IT in Brazil in 2000 is a greater per capita number than that of China and India. OECD figures show that Ireland, with 34 000 graduates per year, has the third highest share of adult population with tertiary degrees (30 per cent of which are in science and technology), after Canada and the USA (OECD, 2003). Such increases were accomplished through a mix of private and public efforts, with the mix varying across countries. In Ireland, Israel and China, the bulk of the efforts were probably in the public sector. By contrast, in India, a substantial fraction of additional engineering capacity created during the 1990s was created in the private sector. Figure 2.1 reports data from the All India Council on Technical Education showing that the almost 80 per cent of the accredited intake capacity for engineering students at the undergraduate level in 2003 (about 340 000) was in privately financed colleges (those that did not receive grants from the government). The role of privately financed colleges, the vast majority of which were created in the 1990s, is even more marked if one looks only at ITrelated engineering programmes. Such a rapid expansion of engineering training capacity has raised valid concerns about the quality of the education and a variety of other social costs. Nonetheless, this supply response is a manifestation of a flexibility that is rare in more advanced countries and also speaks for the perceived returns on human capital.

Per cent private not granted

The ‘diaspora’ One peculiar but crucial element of the picture has been the role of human capital flows. The 3Is, but China and Brazil as well, have directly supplied human capital particularly to the USA. Some estimates provided by Carrington and Detragiache (1998) indicate that the stock of highly-skilled (more than 13 years of schooling) immigrants in the USA from China, India and Brazil were about 400 000, 300 000 and 60 000, respectively.5 Kapur and

Figure 2.1

100 90 80 70 60 50 40 30 20 10 0

84

82 76

92

86 80 80

76 66 1987–88 1995–96

42 26

25

2002–03

12 13 5 East

Central

North Region

West

South

Share of privately financed colleges in Indian IT capacity, by region and year

36

International handbook on industrial policy

McHale (2005) show that China is the second country after India ranked by the number of approved H-1B visa petitions by the US Immigration and Naturalization Service (INS). Indians alone account for more than 42 per cent of all the H-1B visas approved in this period, over half of which went to computer-related occupations. Independent estimates of the US IT workforce indicates that more than 15 per cent of this workforce is from Asia, of which slightly less than one-third from India, implying that Indian born account for about 5 per cent of the IT (mostly software) workforce of about 3.2 million (IT Workforce Update, 2003). The diaspora can provide links, act as ‘reputational intermediaries’ and, upon returning, can bring back valuable skills and expertise. Despite much talk of IT professionals returning to India and China, only in Ireland has the return of skilled emigrants clearly increased the domestic supply of skilled people (Kapur and McHale, 2005). Similarly, the Sands (2005) survey of 58 Irish software entrepreneurs indicates that 66 per cent of the founders of the Irish software firms in her sample had worked abroad, 55 per cent had worked for multinational companies and 74 per cent of her companies had one founder who worked abroad, largely consistent with earlier findings (O’Gorman, O’Malley and Mooney, 1997). The Israeli case is similar to that of India. There is little evidence of systematic returns of immigrants to the USA. At the same time, there is no evidence of a major skill shortage in Israel. The number of US PhDs awarded to Israeli nationals has remained stable since the early 1990s, while the number of PhD students who plan to stay in the USA has increased slightly, from 51 per cent to 57 per cent, albeit well below the corresponding figure for Indian and Chinese PhD recipients in the USA, which is around 5–10 per cent. Despite the many potential benefits of a diaspora, the outflow of skilled engineers, scientists and doctors does represent a net loss of talent and of the considerable investment in training the emigrants. The broad question about the net effects of the international mobility of skilled people on the home country is a complex one and well beyond the scope of this chapter. However, Kapur and McHale (2005) conclude that the benefits from the diaspora outweigh the costs for the 3Is in the particular case of the software industry that we are discussing here. In other words, setting aside the question whether the brain drain represented by the diaspora was a good thing or not, the software industries of India, Ireland and Israel certainly benefited from having one. Given the domestic market orientation of the software industry in China and Brazil, the trade-off has likely been less favourable. 4 Is the growth of the Indian software industry good for the USA and Europe? In this section we turn to the question posed by Gomory and Baumol in the specific context of the software industry. Since outsourcing of software activities to India, from both the USA and Europe, has been the most significant in terms of employment, we focus on the Indian case to exemplify our argument. We start by discussing the relationships between the USA and India in software and then turn by comparison to the other advanced countries and particularly Europe. Since the 1990s, when the productivity of Indian software has increased, the USA has outsourced a good deal of software activities and services to this country. The bulk of the outsourcing, however, has involved software activities that are (a) complementary to valueadded activities done in particular in the USA and (b) at a ‘lower level’ of the value chain. Moreover, in spite of its notable growth, the Indian software industry is still small com-

Emerging issues in the new economy and globalization 37 pared to the USA. As shown by Table 2.2, the Indian software industry employs about 250 000 people while the US software industry employs over a million people. The difference in the total software workforce is even larger because many software professionals are employed by the user industries and the latter are far more developed in the USA than in India. As noted earlier, the total software workforce in the USA is estimated to be about 3.2 million as against 400 000–500 000 in India (Arora and Gambardella, 2005b). We will not enter the current debate on whether the current job losses due to growth in software production overseas will be quickly made up by the creation of new opportunities at home. This debate echoes earlier debates that appear to pit free trade against job losses in existing industries. The conclusion of most mainstream economists, with which we broadly agree, is that, although free trade creates winners and losers, the gains are greater than the losses and thus, if the winners were able to compensate the losers, free trade ought to be supported. In this instance, we think that policies such as retraining, or even some redistributive policies, which lessen the impact of such job losses, are a good idea.6 Our focus is on two other questions. The first is whether, following outsourcing, the physical separation of lower end (typically software coding and maintenance) from higher end (typically product design) software activities will affect US productivity in the higher end. This again echoes an earlier debate about whether product innovation requires proximity to manufacturing. In fact not only are there many industries in which the locus of production and the locus of invention are physically separated, but this is particularly true when the body of knowledge underlying the invention has a strong technical basis. For example, building on earlier work by Lamoreaux and Sokoloff (1996, 1997), Sutthiphisal (2004) studied the location of production and invention in three industries during the second industrial revolution: textile, footwear and the electrical industry. He finds that the locus of invention did not shift with the locus of production as the latter moved to other locations. Moreover, he finds that the link between location of production and invention is weaker in the more ‘science-based’ electrical industry. Using data from a century later than the second industrial revolution, Mariani (2001) studies the location of R&D and production facilities by the Japanese MNCs in Europe. She finds that, in low and medium R&D industries, R&D labs are more likely to be located close to production facilities than in more R&D-intensive industries. Our own work on the history of the chemical industry provides additional evidence. In particular, as discussed in Arora and Gambardella (1998), the US chemical industry has developed since the 1950s an independent sector of specialized engineering firms (SEFs) that supplied new process technologies via licences and chemical process engineering services to the rest of the world. Through both SEFs and chemical manufacturers the USA is still the world leader in the provision of chemical process technology and engineering services in spite of the fact that many chemical productions have been decentralized to other countries and particularly to the developing countries. The second question is about the evolution of these processes. As the Indian software industry becomes more sophisticated and productive, it may increasingly erode higher end activities. However, there are reasons to believe that the USA will remain the centre of software innovation for the foreseeable future. Two resources are crucial to remaining the centre of innovation in software: (a) access to talented designers, software engineers and programmers, and (b) proximity to a number of large and technically sophisticated users. The USA dominates on both counts.

38

International handbook on industrial policy

As noted, emerging software-producing regions are leveraging their abundance of relatively underused human capital, especially of engineers and IT professionals. The USA also produces abundant human capital. But that is not all. The very process of globalization, which has led to the growth of software overseas, has also worked to attract the best and the brightest to the USA, as also noted in our discussion about the diaspora. Even more significant is the relationship with the users. New software applications depend on knowledge about demand and the use of the applications. This is apparent, for example, in telecommunications and semiconductors, where the software needed to design the chip is part of the product itself. More generally a substantial fraction of software is used for running businesses and business processes. Therefore proximity to business activities is crucial for innovation in this industry, which consists of solutions to emerging business problems, often specific to particular industries such as banking, finance, retail, telecommunications or manufacturing. The development of new commercial applications or ‘solutions’ is a very special comparative advantage of the USA. India’s economy is growing and sophisticated domestic demand for software may eventually rise. At present, however, the USA lead is overwhelming. Moreover, as was true of human capital, globalization may reinforce this lead because innovative companies from Israel, Ireland and even India are increasingly moving their operations to the USA to be closer to the users or to sources of equity financing, legal services and thick and well-functioning labour markets. There are of course countertendencies as well. Insofar as Indians have a preference for staying in their home country, the cost of scientists and engineers will be lower in India, so that the cost of R&D activities that are human capital-intensive and relatively less intensive in physical infrastructure will be lower. There is anecdotal evidence that India is being viewed as an attractive location for certain types of R&D activities, although anecdotal evidence also suggests that US firms are not locating mission critical activities in India, nor are they moving activities at the technological frontier. These remarks suggest that the increased Indian competitiveness in software can more seriously affect other advanced nations, and in particular Europe and Japan. First and foremost, Europe and Japan are far less capable of attracting talents than is the USA. This is the most important problem. Relatively speaking, the problem of the dearth of IT users is not as serious. Europe and Japan have well developed IT user industries. Overall, the European and Japanese user industries may be less stimulating than the USA for software innovations. But, on the one hand, Europe and Japan have leading large companies in manufacturing, telecommunications or services that have provided several stimuli to software innovation. On the other hand, the differences from the USA are, in this respect, less significant than the gaps in the ability to attract talents. Apart from IT graduates and professionals from India or other developing countries with large supplies of skills, Europe in particular is unable to keep many of its own talents who are flowing steadily to the USA. While the inability to retain its own skilled workers is an obvious problem, the difficulties in attracting graduates and professionals from India, China or similar countries should be recognized as a problem as well. There are at least two reasons. First, these regions are gradually becoming world suppliers of skills. As Europe and Japan fail to tap such a supply, or if suppliers look overwhelmingly to the USA, they can hardly fill the skill shortage that they are facing, a problem that is exacerbated, especially in Europe, by the demographic trends. In a period in which growth is largely affected by the ability to expand the employment of skilled labour, an inadequate supply of skills can be a serious constraint. It could

Emerging issues in the new economy and globalization 39 explain in good part the low growth rates of these countries, and particularly of Europe, compared to that of the USA. The second reason is that, because the principal competition for the US IT and software firms is still from countries like Germany, France or Japan, the US ability to lever outsourcing and to link to the skilled labour markets in India or similar countries may produce strategic benefits as well. The US firms can gain competitive advantages over their European and Japanese competitors in terms of lower costs, greater flexibility and shorter product development cycles, as well as in terms of attraction of talents from these countries, which reinforces the productivity of the US firms. The problem is less severe for the European or Japanese multinationals which can locate in the USA and partly take advantage of the US trends. It will be more serious for the European and Japanese domestic industries. US firms may move to Europe or Japan, given that they have large commercial markets, and possibly contribute to the growth of the European or Japanese IT or software sectors. But the risk that Europe and Japan are running is that a good deal of the invention process in IT and software will increasingly be elsewhere and this may impinge on the overall competitiveness and growth of these industries. 5 Is the globalization of software good for India? The other twist of the question is whether the growth of software, and particularly its globalization, has net benefits for India.7 In addition to the standard benefits from specialization according to comparative advantage, it was a crucial strategy for India to specialize in activities that were complementary to the international industry. If India had moved up quickly onto higher value added products, as many had suggested, perhaps because of the common belief that moving up the value chain is the only way to sustain an industry, it would have most likely conflicted with the USA. Ireland and Israel faced this problem, especially during the downturn of the international IT demand after 2000. Because software product development was more important in these countries than in India, it was natural that the reduced international demand for IT had a bigger impact on them. Indeed, as shown in Arora and Gambardella (2005b), the performance of the Irish and Israeli software industries faltered in 2001–2, with reduced profitability and employment for the first time since the early 1990s. By contrast, the growth of the Indian software industry slowed but only to about 20 per cent per year in this period. The international migration of human capital also requires some thoughts. As Table 2.3 shows, the number of Indian-born residents in the USA increased nearly threefold between 1990 and 2000, with over 80 per cent of the Indian-born residents having tertiary-level education. Brazilian and Chinese-born residents have lower levels of tertiary education than those for Indians, but still above 50 per cent, and have also increased rapidly in the last decade. To put these figures in perspective, around 50 per cent of the native-born adult population in the USA has received tertiary education, second only to Canada. Such large-scale migration of human capital has conflicting welfare implications. To the extent that the increase in the US labour supply curbs US wages, this lessens the welfare of the native-born workers. By contrast, the Indian workers gain because the salary of the emigrants increases, as does the salary of the workers who stay in India (or at least it does not fall as much as it might have, given the large increases in the supply of Indian IT workers). The standard argument, however, also implies that, because the migration follows more productive opportunities in the USA, the reallocation of Indian workers to the USA implies a higher combined output of the two countries than if migration did not

40

International handbook on industrial policy

Table 2.3

India Brazil China

Selected foreign born populations in the United States aged 25 and over 1990

2000

Change (%)

% of 2000 population entering post1990

304 54 405

837 154 847

175 186 109

55 49 66

Educational Attainment (2000) Primary Secondary Tertiary (%) (%) (%) 5 9 20

15 36 26

80 55 54

Source: Kapur and McHale (2005); original data from the March 2001 Current Population Survey of the United States.

take place. Hence proper redistribution policies, or the extent to which Indian workers ship some of their benefits back to India (for example, links to Indian software companies, returnees, consumption of other Indian commodities), may in the end enhance the overall value for India as well. One problem might be if the people who migrate from India to the USA are the more productive individuals who also produce great externalities on others; for example, because they are more entrepreneurial, have better links to the USA or have greater technological or managerial capabilities. In this case, while their migration increases productivity because it helps restrain the growth in the supply of skills, it reduces the productivity of the remaining workers. We cannot exclude the possibility that the migrating skilled workers are the most productive ones. However, the available evidence is that the Indian software salaries have increased faster than those in the USA. Athreye (2005) showed that the US salaries for a variety of software occupations such as programmer, project leader, quality assurance specialist and systems designer increased by about 21 per cent between 1995 and 1999, while Indian salaries increased by nearly 45 per cent. It might be that without migration the salaries of the Indian software professionals might have increased even more. But migration has definitely not impoverished them. This opens the question whether India’s ability to sustain a continuous flow of human capital may turn this into a major opportunity. We have seen that India, but also Ireland, Israel, China and Brazil, have created several new academic institutions in the 1990s which expanded considerably their supply of graduates without – as shown above – reducing their salaries. As people move, and are replaced by new workers, the flow of workers to the USA feeds its software industry with talented programmers and motivated entrepreneurs. It may well be that the main Indian (or Chinese) resource could be its academic system supplying human capital to the world. 6 Implications for industrial policy The discussion so far suggests several implications for industrial policy. We focus on four issues. First, our software stories are suggestive of the unpredictability of the industries that may become the successful specializations of a country in the 21st century. As noted, it was hard to predict that any of our five countries could become a software powerhouse. It was even more unpredictable that they could exhibit comparative advantages in human

Emerging issues in the new economy and globalization 41 capital-intensive industries. Even within human capital-intensive industries it is not easy to explain why Ireland has become a specialist in software and not Scotland, for example, which has instead been more active in semiconductors, or why software developed in Brazil or China and not in Argentina or in Russia (which has a good supply of skilled mathematicians). As with ‘self-discovery’ all these countries found out gradually that they could become software producers. This stemmed from entrepreneurial activities that experimented with the industry early on. As Rodrik (2004) or Levinthal and March (1993) pointed out, self-discovery creates externalities that make the process more valuable socially than for the individuals that pursue it. The support of entrepreneurial activities for self-discovery is then a natural target for industrial policy. By contrast, the individual incentives are well in place once these industries develop, which suggests that later on there is little need for interfering with their growth. A second important area for industrial policy is outsourcing. Our discussion in this chapter suggests that, as long as a country has the ability to outsource complementary activities that are carried out relatively more efficiently abroad, the net effect of outsourcing is positive. In the particular case of the software industry, we noted that proximity to the users and the ability to form and attract talents are critical for supporting innovation capabilities. The USA excels in both. This makes it well positioned to benefit from outsourcing and to preserve innovative capabilities in higher end segments. Europe and Japan also have potentially stimulating users, but they are not that good at attracting human capital or even retaining their own. Outsourcing can then be a more serious problem than in the USA. We hesitate to generalize from one industry example; however, in Europe and Japan, industrial policy should design ways to promote the role of the users as a basis for stimulating innovation in upstream firms and sectors. This means paying special attention to the formation of links between industries and their suppliers, vertical relationships that sustain cooperative innovation efforts and so on. Even more importantly in Europe and Japan, industrial policy should concentrate on the formation and attraction of human capital. The problem is particularly severe in Europe because of its demographic trends. Europe should devise ways to tap into the new world supply of skills coming from countries like India or China. In this respect it is not unfair to say that the US ability to attract skills from these countries is one of its major competitive advantages today. The USA, however, faces another problem associated with outsourcing. It seems to be moving from being a country with comparative advantages in skill-intensive activities to a country with comparative advantages in very skill-intensive activities.8 The implied dualism could be serious. Moderately skilled workers (such as college graduates and white-collar workers, rather than just high-school graduates or lower) could face greater problems than in the past because of the rise of new competitors in some human capitalintensive industries. As noted earlier in this chapter, we agree with the general view that free trade based on comparative advantages is beneficial. However, industrial policy should work on the potentially severe distributional problems that could be associated with the transition. The specific form of these interventions is beyond the scope of this chapter. They could involve retraining of individuals or anything that could ease the asperity of change. But the US problem of the reduced comparative advantage of people who are ‘just’ skilled (and not very skilled) should not be neglected. Third, the role of the users can be cast in the more general framework of the sources of learning and competencies that can give rise to new comparative advantages. We discussed

42

International handbook on industrial policy

Klepper’s argument that successful new firms or specializations come from successful firms or industries through related diversification and more generally from competencies formed in pre-existing successful sectors. Our software industry stories are consistent with this view, as discussed in Arora, Gambardella and Klepper (2005). Here we can just mention the case of the Israeli software industry which has taken off from its strong and successful IT hardware industry and successful IT user sectors, as well as from the competencies that resided in the military. Similarly, the multinationals provided important sources of competencies in Ireland and even in Brazil. As a result, an effective industrial policy may not address the specific industries or firms that may constitute the new specializations, but the existing sources of competencies (industries, firms or others). This feeds self-discovery and entrepreneurial experimentation by creating the right incentives to spin off successful new firms and industries. The idea of encouraging an industry or a firm to spin off resources, competencies (particularly valuable ones) or even potential competitors may be counterintuitive. However, many resources or competencies of a firm are of greater value outside than inside it. For example, individual employees may have an incentive to pursue projects that are of little interest to the company because the latter has opportunity costs. By letting these employees go, the company can hire new employees. This rejuvenates its workforce and spins off employees who might be frustrated inside the company, and hence are less loyal or productive, because they cannot advance their projects or aspirations. Note that a successful company, precisely because it is successful, creates a lot of new opportunities, typically more than it can pursue. Tensions with its employees, who want to pursue certain projects or who want to leave, are then inevitable. From his long-standing experience at Intel and as one of the father founders of Silicon Valley, Gordon Moore noted this point (Moore and Davis, 2004). He also mentioned that one way to avoid competition from the employees of the company is to encourage them into activities that use the competencies formed in the successful company to perform complementary rather than competitive activities. For example, at Intel, once they figured out that they could do little to stop their employees’ spin-offs, they tried to encourage them onto the production of semiconductor equipment, which Intel could buy, rather than the semiconductor chip, which was the core business of Intel. Industrial policy could help drive the spin-offs of successful firms or industries into supplier rather than competitor sectors. This would also have the advantage of policy being consistent rather than in conflict with the incentives of the successful parent industries or firms. Finally, in this chapter we added our voice to the chorus that calls for a greater attention to human capital. We also noted the gap between the USA, on the one hand, and Europe and Japan, on the other. In this respect education policy, to enhance the quantity and quality of human capital, has become an important section of industrial policy. It is intriguing that in some European countries, such as Spain, there has been a notable increase in the share of graduates in the population in the past few years, while in others, such as Italy, this has not occurred (OECD, 2003). Moreover, in countries such as Italy there is resistance to moving academia into a research-oriented, international and meritocratic system. In the 21st century, this will have serious negative implications that go beyond education (or labour) policy per se and will impinge upon industry, industrial policy and therefore on the competitiveness of a country. At the same time, apart from the domestic supply of human capital, Europe and Japan should tap into the world supply of skills. The USA

Emerging issues in the new economy and globalization 43 has benefited significantly from its ability to mingle domestic and international human capital and to transmit its entrepreneurial culture to many graduates from other countries. Industrial policy in the 21st century will increasingly be tied to immigration policy. Apart from expanding the supply of human capital, which can in particular help Europe, given its demographic trends, openness to human capital from developing countries like China or India will have the additional benefit of expanding the skilled workforce with individuals whose ambitions, expectations and desire to grow are much higher than those of a domestic population that is in large part economically satiated. Notes 1. Hausmann and Rodrik cite a World Bank study by Rhee and Belot (1990) who provide several examples of the discovery of successful specializations in developing countries through entrepreneurial experimentation or sparked by specific events which demonstrate the unpredictability of the specializations and their heterogeneity across countries. 2. For a summary of these stories and the references to the original Klepper papers, see Arora, Gambardella and Klepper (2005). 3. As noted in a footnote of Table 2.2 the high share of the software multinationals’ sales in Ireland is largely a consequence of inflated invoicing in this country, for tax purposes. 4. See the country chapters in Arora and Gambardella (2005a) for details on the reasons for the excess supply of human capital in the five countries. Regressions of the stock of scientists and engineers on per capital GDP and other factors (not reported here) showed that India, Ireland and Israel had actual stocks that were higher than the predicted stocks of scientists and engineers. 5. Brazilian emigration of skilled workers, though lower in absolute volume, when expressed as a percentage of the stock of skilled population, is comparable to that of India and China. Unfortunately, the small absolute size of the stock of Brazilian immigrants in the USA has meant that little is known about their role in the growth of the Brazilian software industry. 6. On the debate about the outcomes for the USA of the outsourcing game in software and IT more generally, see for instance Mann (2003) and the McKinsey Global Industry report (MGI, 2003). 7. Once again, we focus on India for brevity and emphasis. 8. This remark was made to us by Dick Nelson. Of course he is not responsible for any improper way in which we are using it.

References Arora, A. and A. Gambardella (1998), ‘The evolution of industry structure in the chemical industry’, in A. Arora, R. Landau and N. Rosenberg (eds), Chemicals and Long Term Economic Growth, New York: John Wiley and Sons. Arora, A. and A. Gambardella (eds) (2005a), From Underdogs to Tigers: The Rise and Growth of the Software Industry in Brazil, China, India, Ireland, and Israel, Oxford: Oxford University Press. Arora, A. and A. Gambardella (2005b), ‘The globalization of the software industry: perspectives and opportunities for developed and developing countries’, in A. Jaffe, J. Lerner and S. Stern (eds), Innovation Policy and the Economy, vol. 5, Cambridge, MA: The MIT Press. Arora, A., A. Gambardella and S. Klepper (2005), ‘Organizational capabilities and the rise of the software industry in the emerging economies: lessons from the history of some emerging industries’, in A. Arora and A. Gambardella (eds), From Underdogs to Tigers: The Rise and Growth of the Software Industry in Brazil, China, India, Ireland, and Israel, Oxford: Oxford University Press. Athreye, S. (2005), ‘The Indian software industry’, in A. Arora and A. Gambardella (eds), From Underdogs to Tigers: The Rise and Growth of the Software Industry in Brazil, China, India, Ireland, and Israel, Oxford: Oxford University Press. Botelho, A.J., G. Stefanuto and F. Veloso (2005), ‘The Brazilian software industry’, in A. Arora and A. Gambardella (eds), From Underdogs to Tigers: The Rise and Growth of the Software Industry in Brazil, China, India, Ireland, and Israel, Oxford: Oxford University Press. Carrington, W. and E. Detragiache (1998), ‘How big is the brain drain?’, IMF working paper no. 98, IMF Washington, DC. Gomory, R.E. and W.J. Baumol (2000), Global Trade and Conflicting National Interests, Cambridge, MA: The MIT Press. Hausmann, R. and D. Rodrik (2003), ‘Economic development as self-discovery’, Journal of Development Economics, 72, 603–33.

44

International handbook on industrial policy

IT Workforce Update (2003), ‘The outlook in 2003 for information technology workers in the USA’, Commission on Professionals in Science and Technology (available from http:// www.cpst.org/ ITWF_Report.cfm; last viewed, 3 May 2004). Kapur, D. and J. McHale (2005), ‘Sojourns and software: internationally mobile human capital and the software industry in India, Ireland and Israel’, in A. Arora and A. Gambardella (eds), From Underdogs to Tigers: The Rise and Growth of the Software Industry in Brazil, China, India, Ireland, and Israel, Oxford: Oxford University Press. Klepper, S. (2004), ‘The geography of organizational knowledge’, mimeo, Carnegie Mellon University. Lamoreaux, N. and K. Sokoloff (1996), ‘Long term change in the organization of inventive activity’, Proceedings of the National Academy of Science, 93, 12686–92. Lamoreaux, N. and K. Sokoloff (1997), ‘Location and technological change in the American glass industry during the late nineteenth and early twentieth centuries’, NBER WP #5938, Cambridge, MA. Levinthal, D. and J. March (1993), ‘The myopia of learning’, Strategic Management Journal, 14, 95–112. Mann, C. (2003), ‘Globalization of IT services and white collar jobs: the next wave of productivity growth’, International Economics Policy Brief, Number PB03-11, December. Mariani, M. (2001), ‘Next to production or to technological clusters? The economics and management of R&D location’, Journal of Management and Governance, 6 (2), 131–52. McKinsey Global Institute (MGI) (2003), ‘Offshoring: is it a win–win game?’, San Francisco, CA: McKinsey & Co. Inc. Moore, G. and K. Davis (2004), ‘Learning the Silicon Valley way’, in T. Bresnahan and A. Gambardella (eds), Building High-Tech Clusters: Silicon Valley and Beyond, Cambridge, UK: Cambridge University Press. OECD (2003), Education at a Glance, Paris: OECD. O’Gorman, C., E. O’Malley and J. Mooney (1997), ‘Clusters in Ireland. The Irish indigenous software industry: an application of porter’s cluster analysis’, NESC, National Economic and Social Council, Research Series, Research Paper no. 3, November. Rhee, Y.W. and T. Belot (1990), ‘Export catalysts in low-income countries’, World Bank Discussion Paper, World Bank, Washington, DC. Rodrik, D. (2004), ‘Industrial policy for the XXI century’, mimeo, Harvard University. Sands, A. (2005), ‘Eye of the tiger: evolution of the Irish software industry’, in A. Arora and A. Gambardella (eds), From Underdogs to Tigers: The Rise and Growth of the Software Industry in Brazil, China, India, Ireland, and Israel, Oxford: Oxford University Press. Sutthiphisal, D. (2004), ‘The geography of invention in high- and low-technology industries: evidence from the second industrial revolution’, working paper, Department of Economics, UCLA. Tschang, T. and L. Xue (2005), ‘The Chinese software industry: the implications of a changing domestic market for software enterprises’, in A. Arora and A. Gambardella (eds), From Underdogs to Tigers: The Rise and Growth of the Software Industry in Brazil, China, India, Ireland, and Israel, Oxford: Oxford University Press.

3

European industrial policy Jacques Pelkmans

Introduction The present chapter discusses industrial policy of the European Union. The reader will encounter complexity, two levels of policy making, drastic changes over time and the overall difficulty of where to draw the boundaries of EU industrial policy. In the European Union industrial policy is pursued at (at least) two levels of government: the EU and the member states’ levels. Furthermore, owing to shifting convictions in economic analysis as well as the logic of pro-competitive market integration (which has greatly deepened over time), the nature and intensity of European industrial policy has drastically changed over the five decades since the treaty of Paris.1 These two ‘European’ features add to ‘general’ complications in analysing industrial policy, such as the contrast between normative economic (‘welfare’) and positive analytical approaches of industrial policy, the variations in classification of what industrial policy is, and the potentially vast range of tools, quasi-tools and soft forms of persuasion in this area. The aim of the present chapter is to set out and explain the great significance of European integration in determining (changes in) structure and performance of industry in Europe. This is due to the deepening of economic integration since the 1970s, the widening of its scope and the enlargement of the club. Member states, drawing on the treaties but also on changing insights about the role and dynamics of markets, have agreed to bind themselves ever more by the pro-competitive logic of deepening market integration and to abide by the ever stricter constraints that this implies. In addition, they have not (or better, only marginally and selectively) shifted to the EU level the tools for interventionist policy making member states themselves used to employ in the past. The structure of the chapter is as follows. Section 1 discusses at some length the division of powers between, and the complementarity of, the member states’ and EU level of government when it comes to industrial policy. A comprehensive survey is provided, based on a wide and fairly detailed classification of industrial policy instruments. Section 2 will address the place and meaning of three cross-cutting issues, namely, the indiscriminate use of the ‘competitiveness’ label in the EU circuit of business and of policy makers, the relation between services and EU industrial policy and, finally, that of European infrastructure. Sections 3, 4 and 5 will deal, respectively, with the three building blocks of the wide concept of industrial policy distinguished in section 1, that is, the EU framework of market integration, EU horizontal industrial policy and EU sectoral or specific industrial policy. The chapter ends with the main conclusions. The reader, it is hoped, realizes that the subject area is much too vast to be surveyed comprehensively. Choices have to be made. One such choice is that I shall focus more on the framework and the horizontal aspects, as the incentive structure for industry and industrial markets is dominated by these two. A corollary is that section 5, on sectoral and specific policy, is limited to an overall perspective, with (brief) illustrations only about ICT and technology, while refraining from details on sectors. Another choice is to remain complementary to other authors in this 45

46

International handbook on industrial policy

volume dealing with a range of specific economic analyses which often are relevant to EU policy. 1 Industrial policy: what the EU and member states do and do not do There is a great deal of confusion about what industrial policy is, only surpassed by the confusion about what European industrial policy might be. The latter is even more complicated because the EU level has ample constraining powers vis-à-vis member states’ instruments and their use, while at the same time the EU level itself is also restricted in its actions and does not dispose of anywhere near the same range of tools as member states might employ, in a constrained fashion or not. 1.1 Scope and constraints of European industrial policy Figure 3.1 should help to reduce such confusion to an appreciable extent. A complete resolution is beyond the capacity of the author, for the simple reason that no taxonomy can fully respect the range of views on industrial policy which can be found in the literature. The figure combines three features: a comprehensive classification of industrial policy and of other policies somehow influencing industry, a sixfold breakdown of how EU and national powers are related (see ‘Notes’) and an application of the latter to the many distinct elements of industrial and other policies specified in the classification. The classification begins by distinguishing two sets of policy which influence industry, yet are not part of industrial policy, namely, ‘policies not for industry which affect industry’ for obvious reasons, such as macroeconomic stability with fiscal and monetary instruments; redistributional tools; agricultural and services policies; tax policy; energy; land-use and so on, and ‘policies which directly help or constrain industry but are not meant (only) for industry’, such as price controls, buy-national campaigns, tied development aid or environmental policies addressing specific hazards such as poisonous chemicals or smog. The remainder is defined as a ‘wide concept of industrial policy’ and consists of three building blocks: framework aspects, horizontal industrial policy and sectoral and specific industrial policy. For all the elements specified in Figure 3.1, a proxy indication is provided about the division or combination of powers at EU and member states’ levels. For a full understanding of the nature, scope and potential of EU industrial policy, all these elements would have to be explained separately, though in a coherent fashion. Given the space constraint, it should be sufficient to concentrate on the three building blocks of ‘wide’ industrial policy, with occasional references to other elements in the figure where appropriate. Some special attention will be paid to three such references in section 2. Broadly speaking, the economic and institutional importance of EU industrial policy is greatest for the EU framework aspects. The respective weights of horizontal and sectoral and specific policies cannot be established in general terms. Both are constrained by the EU framework aspects as well as by other subsidiarity considerations which often – though not invariably – mean that the EU level of government has few ‘hard’ instruments to act. It is crucial to keep these limitations in mind when studying the flurry of EU policy documents on industrial policy, ‘competitiveness’ strategies or ‘entreprise’ policies. Most such documents are long on analytical aspects and on recommended directions on how to mend one’s ways in European industrial markets, but short on relatively ‘soft’ or concrete Actions or the employment of significant funding and other instruments. One may regard this as the inevitable outcome of a much more market-driven approach to industrial

European industrial policy 47 Policies (somehow) affecting industry

Policies NOT for industry

Policies for industry

but affecting industry such as • macroeconomic (EU/nat.) • redistributional (nat.) • wage policies/industrial rel.(nat.) • agricultural/fisheries policies (EU/nat.) • services policies (EU/nat.) • tax policies (nat. (
EU)) • land-use policies (nat.) • infrastructure policies (nat. (
EU)) • energy (security) policies (nat. (
EU))

Non-industrial policy measures directly affecting industry • buy national campaigns (EU) • regional planning/ development (nat./EU) • price controls (EU) • tied development aid (nat. (
EU)) • overall export promotion (EU) • specific environmental policies (nat./EU)

(Wide concept of ) industrial policy

framework aspects

establishment IM (EU) proper functioning IM (EU) • (lack of justified) harm.n. • lack of standards & C.A. • removal distortions competition policy (EU/nat.) • state aids • network industries regional/cohesion policies (general) (EU/nat.) better regulation (EU/nat.) state ownership (nat.)

horizontal industrial policy

sectoral/specific industrial policy

research strategies (nat./EU) innovation stimulus (nat./EU) entrepreneurship & risk capital (nat./EU) skills & human capital (nat.) restructuring funds (nat./EU) competitiveness tests for other policies (e.g. ENV; etc.) (EU) public procurement (nat./EU)

sectoral interventions (EU) sector policies (EU; EU/nat.) clustering & filières policies (nat.) trade policy (EU) specific aspects of regional/cohesion policies (EU/nat.) technology policies (EU/nat.) defence procurement (nat.)

Notes: EU → EU powers; EU/nat. → shared powers; nat. → member states’ powers; nat./EU → shared powers, mainly national; nat. (
EU) → national & EU constraints or marg. inputs; EU → little nat’l. leeway, strict EU constraints.

Figure 3.1

EU and national powers of industrial policy instruments

48

International handbook on industrial policy

change, which is no doubt correct, but it might just as well be seen as an explicit choice not to endow the EU level with forceful instruments and powers to influence or ‘engineer’ the future path of Europe’s industries. Writing in 2005, the two perspectives boil down to very much the same result in terms of industrial policy potential. The framework aspects follow directly from what the EU is. The European Union is, in economic terms, essentially an Economic and Monetary Union. The Economic Union consists of the internal market (an extremely comprehensive concept with a strong legal basis and a powerful institutional backing) and a very modest set of ‘cohesion’ policies. The cohesion policies largely consist of regional instruments for relatively poor regions in the Union, and to some extent for providing funds for relatively poor countries in the EU. ‘Cohesion’ would seem to refer to a kind of redistribution from richer to poorer areas in the EU. Nevertheless, cohesion amounts to an efficiency-based long-run strategy of ‘catch-up growth’. The interventions are meant to accelerate catch-up growth and, once this is achieved, cohesion policies are terminated. Cohesion policies do influence industry in a number of ways but especially by improving the determinants of location and investment. However, it ought to be noted that cohesion policies do not and cannot address specific industries (or services, for that matter). Such policies are more horizontal, that is, are mainly oriented to hard (networks and roads/bridges/tunnels) and soft (human capital, technical schools, retraining facilities, administrative capacity) infrastructure. The Monetary Union in strict treaty terms comprises not only the ‘eurozone’ of (now) 12 countries2 but also the non-eurozone countries insofar as they have obligations under what the treaty calls the second stage of EMU.3 1.2 Internal market rationale The hard core of the Economic Union is the internal market. A good understanding of the internal market, the logic of its design and of its proper functioning is indispensable to grasp the nature, scope and potential of EU industrial policy at the two levels of government. An economically meaningful definition of the internal market reads as follows: ‘the free movement of goods, services and factors of production as well as the right of establishment across intra-EU frontiers, accompanied by all necessary common regulation and/or policies for this internal market to function properly’.4 What the treaty calls the ‘establishment’ of the internal market boils down to the accomplishment of free movement and (company) establishment, but, in actual practice, this accomplishment can often only be had if degrees of common regulation or mutual recognition are agreed first. Thus, the establishment and proper functioning are frequently pursued simultaneously. However, proper functioning implies the absence of market failures and this refers to competition policy as well. Therefore the framework box, bottom left in Figure 3.1, mainly concerns the internal market: when its establishment is incomplete, it leaves a degree of freedom to member states (the few examples left nowadays include the absence of an internal labour market and the lack of a true EU patent), and when the proper functioning is not (yet) adequately dealt with, it might refer to a lack of justified harmonization (due to one or a few recalcitrant member states), a lack of common standards or the slow liberalization of network industries (indeed, without a common EU regulator, unlike the case of national markets). For the most part, however, the internal market is well established and its functioning improves with deepening and widening of scope. This will be discussed in section 3. The

European industrial policy 49 common regulation in ‘approximated’ (harmonized) or centralized form has recently become much more rigorously scrutinized under the label ‘better regulation’ (see section 4). The new approach of subjecting all substantive Commission proposals to regulatory impact assessment, based on an analytical methodology, is bound to influence whatever industrial policies are going to be proposed. Finally, the framework box in Figure 3.1 mentions state ownership. It is good to underline that the EU treaty is neutral with respect to ownership (art. 295, EC) and that privatization is strictly a matter of the member states. The casual substitution between terms such as deregulation, liberalization and privatization when it comes to the liberalization of and EU-wide competition in network industries is incorrect, even though it is routine in many circles. For the EU, privatization is strictly the change of ownership from state (or provincial or municipal) to private shareholders. The Union is interested in the combination of competition, also over intra-EU frontiers, hence free movement and (company) establishment, with justified regulation in network industries, not in state versus private ownership. Equally important, although state ownership is a matter of the member states, the treaty does not allow any concrete advantage for such firms in the internal market. It is hard to think of any economic reason why a company should remain in state hands, other than a protection against unwanted takeovers. Even fresh capital invested by the state in the company is seen as illegal (because distorting) state aid if an independent investor would not have been willing to do the same. 1.3 Horizontal industrial policy Horizontal industrial policy is of relatively recent origin. Indeed, of the seven categories specified in the middle box at the bottom of Figure 3.1, none of them existed at EU level in, say, 1980 in a meaningful way, or at all. The EU had no programmes or in-depth analyses in areas such as innovation stimuli, entrepreneurship and risk capital, skills and human capital and competitiveness tests for EU policies. Around 1980, almost all research funding was spent on nuclear and other energy research in the framework of Euratom. The first ESPRIT (on computers and IT) programme dates from 1981. An EU restructuring fund did not exist except for steel and was paid by the steel mills themselves in the framework of the ECSC. And EU public procurement is too tiny to be utilized for strategic purposes, and no such objective had been made explicit anyway. However, the issues in the horizontal industrial policy box were influenced by the EU level when imposing disciplines on national industrial policies under some of these rubrics. Thus research subsidies of a ‘fundamental’ nature are not caught by the state aid regime, but if research is ‘applied’ subsidies are disciplined depending on how close the applications are to reaching the market. National restructuring programmes did exist and were subjected to complex state aid disciplines, more often than not politically influenced at the highest level of EU politics via ad hoc decisions on rescue operations. National public procurement was formally disciplined by non-discrimination provisions in the treaty and a few weak directives on goods and on public works which were hardly enforced in those days. The later shift to horizontal EU industrial policy, beginning with the Bangemann report (1990), therefore required a wholesale rethinking of the way government intervention could usefully facilitate the static and dynamic efficiency of markets in Europe and what market institutions might be of help (for example standards, certification and quality incentives, or innovation structures, or skill enhancement). Next to no such thinking had been present at EU level and no policies had been initiated around 1980.

50

International handbook on industrial policy

At the EU level, the 1980s and early 1990s consisted of enormous deepening and widening (of scope) of the framework box in terms of the internal market in goods, services and capital. The corollary of the Single Market Programme consisted in the widening of competition policy, and complemented by the EC merger regulation. All this was bound to have many implications for what member states have become accustomed to do. Friction merged on the interface between the EU level and the member states’ level, in regular struggles between the Commission and the Council and, to some extent, between member states in the Council. The upshot was to discipline and improve the sectoral and specific industrial policies both at EU and member states’ levels. Thus the middle box of horizontal issues was of marginal importance at best. Writing in 2005, approaches have radically changed. As section 4 will elaborate, today’s emphasis in European industrial policy at both levels of government consists of a combination of the framework and the horizontal boxes, with relatively little attention paid to the sectoral and specific box except for technology policies. 1.4 Sectoral and specific industrial policy Sectoral and specific industrial policy lies at the origin of the Community. European economic integration began in 1952 with a ‘deep’ free trade area in coal and steel, called the European Coal and Steel Community. Its ultimate aim clearly being the cementing of peace between former adversaries, the legal commitments consisted of a combination of intra-ECSC free trade in coal and steel with extensive actual and potential interventions in the sectors, including investment plans and (under the admittedly extreme scenario of a ‘manifest crisis’, art. 58, ECSC) even the administrative organization of intra-ECSC trade. Looking back at the ECSC five decades later, the contrast with today’s economic thinking about the functioning of markets and specific interventions is rather sharp. It is true that the ECSC initiated a kind of European competition policy and that state aids were completely forbidden,5 but otherwise the ECSC hardly managed to create a functioning internal steel market and never accomplished an internal market for coal. In fact, the steel market enjoyed its share of the ‘golden growth era’ of the EU between the mid-1950s and 1973, but an undistorted internal market in steel only came about just before the ECSC treaty was integrated in the EC treaty (in 2002). In fact, the early and unstoppable decline in coal and the subsequent decline of steel have caused tremendous adjustment problems, often concentrated in certain regions as well, prompting all kinds of regional jobs programmes and ad hoc responses. In the same period of concluding the ECSC treaty, attempts to agree on a similar sectoral treaty for agriculture (the so-called ‘Green Pool’) only just failed. However, they formed the basis of what later was to become the Common Agricultural Policy in the EEC. The Rome treaty of 1957 did not have a sectoral slant, at least not for industry, except for a minor clause in shipbuilding (which has caused subsidy problems until today). The only sectoral bias one finds is that in (the common) transport policy, which was first interpreted in a highly interventionist fashion, resulting in a delay of two decades before the ECJ in 1985 ordered the Council to speed ahead with the liberalization of intra-EU transport in all six modes. The long EU struggle to come to grips with the enduring steel crisis has ultimately helped the Commission to rationalize the EU state aids regime. This is the subject of Chapter 7. Suffice it here to say that, first, the consistency of EU state aid discipline over different sectors had to be accomplished;6 second, the residual political control of the

European industrial policy 51 Council over the Commission’s regime under art. 89, EC had to be pre-empted;7 third, the regime had to be organized in ways similar to the rules of antitrust with well-justified block exemptions rather than allowing scope for ‘negotiating rules’ for member states time and again;8 and, finally, the underlying economics of justifying state aids could be addressed, as was done after an invited paper by Besley and Seabright (1999). The result of the long haul of rationalizing state aid in the EU has been a drastic reduction of national state aid in relative terms and, to some degree, even in absolute terms, as well as a more rational rearrangement of the types of aid given. A quick glance at the sectoral and specific industrial policy box, bottom right in Figure 3.1, makes it clear that the member states’ discretion for such policies today is limited at best. Sectoral policies other than state aids, of course, are not allowed to be anticompetitive, whilst discrimination in, for example, filières or clustering goes against the treaty, too. Other than defence procurement (and, indeed, sales worldwide), exempted under the treaty,9 national leeway is modest. EU policies can have sectoral or specific effects, notably under trade policy which is, by definition, as detailed as individual tariff lines at the six-digit level. However, after many GATT rounds, a host of EU bilateral free trade area agreements with eastern and southern neighbours as well as with South Africa and Mexico, and the EU version of the Generalized System of Preferences, industrial MFN tariff protection is low overall, cases of MFN tariffs higher than the highest clothing tariffs of 13 per cent are exceedingly rare and most countries (other than non-European OECD countries) do not even pay these MFN tariffs but lower ones, or none at all. Quotas vis-à-vis WTO members are outlawed and all that remain are a few quotas for aluminium and steel against exporters from several ex-Soviet Union countries not yet in the WTO. Clothing quotas under the Multi-Fibre Arrangement have expired with the arrangement itself (late 2004). EU policies under ‘cohesion’ cannot be sector-specific but compensation measures (retraining, quality upgrading and so on) are possible. Furthermore, infrastructural investments and other actions improving the investment climate as well as otherwise burdensome environmental clean-ups can be funded under regional and cohesion policies. What remains boils down to technology policies where both EU and member states have been active. One can discern a trend with the emphasis shifting more to the EU level of analysis and bringing together the experts at research and sectoral level. Indeed, the line between the EU research strategy in the horizontal box and the EU technology policies in the specific policies box is hard to draw. As sections 4 and 5 will show, both tend to shy away from ‘picking winners’. For all these reasons, section 5 will only briefly touch upon two ‘evergreens’ in the specific policy box, namely, ICT strategy and technology policy. 2 The economic relevance of cross-cutting issues Even a wide concept of industrial policy will not encompass all influences on industrial structure and performance. Conversely, when trying to encompass all such influences, the term ‘industrial policy’ eventually becomes meaningless. The bottom half of Figure 3.1 already stretches the concept very far. Before going into the details of the three boxes in the following three sections, however, it is likely to serve the understanding of European industrial policy if some reflections are offered on a few cross-cutting issues of relevance to the performance of industry in Europe. Many issues in the two top boxes in Figure 3.1 would be eligible for such reflections. A few examples can illustrate how wide the discussion would stretch. Thus, although (industrial) business is capable of surviving under

52

International handbook on industrial policy

almost any macroeconomic policy regime, or system of industrial relations, or strategy for energy security, there can be little doubt that the quality of such regimes matters a lot for the efficiency, profitability and long-run dynamism of industry. Again, these three are ruled by quite distinct combinations of EU and national powers. Thus macroeconomic stability has become a firm obligation in the treaty for all member states – a major and commendable achievement – even if the division between monetary and fiscal policy differs between eurozone (not having a monetary policy at national level anymore) and non-euro countries (having national monetary policies, but under considerable constraints). In contrast, the system of industrial relations is truly national, with all the differentiations this implies. Similarly, the EU has only limited powers over national policies concerning energy security, indeed, virtually none over the very choice of what energy source to give priority to in the first place. A warning is also in place before linking corporate taxation to European industrial policy. First, the EU has no corporate taxation of its own and there is no legal basis to create an EU-wide corporate tax floor. Second, the harmonization of corporate tax in the EU has thus far proved to be extremely difficult, in part because corporate tax competition might be better, in part because treasuries dislike the loss of what they regard as fiscal sovereignty, in part because the short-run losses due to harmonization of the tax base are feared by some. Third, corporate taxation relates to business in general and most businesses are found in the services, not in industry. Of course, one can (rightly) argue that tax issues may make cross-border mergers unduly difficult, but again that does not apply solely to industry. A recent fear amongst some political leaders is that corporate tax competition, especially from the new member states, is inviting a ‘race to the bottom’. If this race is not going to be resisted, companies will have to migrate to Central Europe and accentuate a process of ‘deindustrialization’ already going on in Western Europe. Apart from the fact that, once again, this applies just as much to services, it is far from clear whether or not low effective tax rates (combining the statutory rate with the tax basis; see, for example, Devereux, Griffith and Klemm, 2002) compensate for relatively unfavourable location or other handicaps in Central Europe and whether (many) other determinants of direct investments are decisive. The three cross-cutting issues selected to be elaborated here are the following: does EU industrial policy amount to ‘competitiveness’ policies, do services matter for industry and, if so, does the EU approach to services matter for industrial policy, and what can Europe do about infrastructure and how critical is that for industrial policy? 2.1 Competitiveness policies for industry? The European Union has a Competitiveness Council, combining the previous Councils for the internal market, research and industry. Art. 157, EC,10 under the title ‘Industry’, specifies that the Union and the member states ‘shall ensure that the conditions for the competitiveness of the Union’s industry exist. For that purpose, in accordance with a system of open and competitive markets, their action shall be aimed at (a) speeding up the adjustment of industry to structural changes; (b) encouraging an environment favourable to initiative and to the development of undertakings throughout the Union, particularly small and medium-sized undertakings; (c) encouraging an environment favourable to cooperation between undertakings;11 (d) fostering better exploitation of the industrial potential of policies of innovation, research and technological development’.

European industrial policy 53 This is to be done in two ways, as the third part of this article indicates: first, the Union ‘shall contribute . . . through the policies and activities it pursues under other provisions of the treaty’ which is what Figure 3.1 essentially depicts; second, ‘The Council, acting unanimously . . . may decide on specific measures in support of action taken in the Member States.’ Interestingly, the draft constitution has added to the last phrase on specific measures: ‘excluding any harmonisation of the laws and regulations of the Member States’. This competitiveness clause for industry was inserted during Maastricht. One can view it as a confirmation of the essential message hidden in Figure 3.1, following the debate on and the backing of the 1990 Bangemann report. In any event, art. 157, EC, is inconsistent with the remnants of interventionist industrial policy still around during the conclusion of the Maastricht treaty. Since the Maastricht treaty came into force (November 1993), Commission and Council have displayed a new style of activism in policies related to EU industry and its competitiveness. The main tools are, first of all, analysis, a flurry of horizontal and sectoral policy papers as well as all kinds of networks for horizontal (innovation, new forms of standardization, SMEs and so on) and sectoral (technology, structural adjustment of certain industries and so on) stimulus; and, second, reliance on EU policies and regulation (including the huge area of the internal market itself) other than selective intervention for industry, which is what art. 157, EC, requires. From an economic point of view, therefore, it would seem to make sense to integrate the three old Councils of the internal market, industry and research into a single, more strategic competitiveness Council.12 There are pros and cons of the Union’s shift to competitiveness policies as restrictively defined by the treaty and elaborated since 1993. First, the pros. The strategic advantages are crucial and may well dominate. They include the (virtual) termination of short-sighted and distortive interventions to rescue failing firms or sectors in decline owing to comparative disadvantage as well as the very high barriers against a policy reversal in this respect.13 The focus is on being competitive in Europe and worldwide and this mind-set, even more than treaty texts, has pervasively changed the policy landscape in the Union. The economic advantages, especially in the long run, are obvious. Compared to the more interventionist landscape, one should expect structural adjustment both to be better anticipated by companies and sectors – knowing that the political market is largely taboo – and to be faster if severe problems nevertheless arise. In turn, this will raise productivity overall, hence growth, while presumably reducing the costs of intersectoral labour flows. Other advantages include the more strategic economic perspective of industry as a whole rather than the crisis-prone ad hocery of interventions and a politicized state aids regime. This perspective is (a) profoundly market-oriented,14 (b) far more explicitly focused on how to improve the economic performance of all sectors, not least the ones already doing well, (c) emphasizing future growth sectors in a horizontal fashion (via innovation policies, framework programmes for R&D, attention to ‘entrepreneurship’ and risk capital and so on) and vertically (via strategic High Level Groups and stimulus of standardization, networking, benchmarking, special studies and so on), without ‘picking winners’. Finally, a co-benefit of the competitiveness orientation of the Union is undoubtedly that numerous regulatory and other measures are better scrutinized as to their impact on competitiveness. This might imply additional support for liberalization (for example, in the case of network industries or services, more generally) or an extra form of discipline on regulation before imposing costs on European industry.

54

International handbook on industrial policy

There are significant drawbacks, too. The first seems academic, yet is likely to have profound implications for the policy-making process and to some extent its substance. Whereas competitiveness is a straightforward notion for a single company, it is more elusive for sectors and problematic, if not controversial for countries and the EU. It is not unfair to speak of a gulf between academic economists here, and national and international policy makers.15 For reasons of space, we shall merely refer to some of the implications. Competitiveness is now so widely defined that it serves as a ‘container’ in which almost any idea can be dumped. Not only are ill-defined policies rarely good policies, the nebulous approach acts as an open invitation for (industrial) lobbies and national ministers (not least, once they occupy the rotating EU presidency) to argue attention to almost anything, resulting in waves of ‘fashionable’ topics. This seems to capture what has happened at EU level.16 In fairness, the Commission understood early on that detached analysis and an apolitical organ suggesting agenda items with authority could counterbalance these risks. Thus it initiated a Competitiveness Advisory Group bringing out a series of special reports with the prominent help of Alexis Jacquemin. In 1997, it started annual Competitiveness Reports, largely written by independent economists, and gradually providing a far more reliable analytical basis than hitherto for EU policy making and for enterprises in Europe in their strategic thinking. Another manifestation of the ‘container fallacy’ is the unproductive ‘repackaging’ and relabelling of instruments and policies. At the symbolic level, the term ‘industrial policy’ went out of fashion in the mid-1990s and ‘competitiveness policies’ was the label employed, even though the two Bangemann reports, as the origin of the U-turn to competitiveness concerns in a pro-market approach, utilized ‘industrial policy’. In 1999, however, ‘entreprise policy’ suddenly appears as a new term,17 and seems to be employed, first, as a subset of competitiveness policy dealing with issues of special (but not exclusive) importance to SMEs,18 and, later, as a new label for that part of competitiveness strategy dealing with industry.19 The former idea was developed as policies for ‘entrepreneurship’ or ‘better environment for business’ in a range of activities20 which have stimulated member states to reduce the barriers to start-ups and selectively to cut red tape. The latter notion, that ‘enterprise policy’ would become the new label for competitiveness policies, was swiftly dropped when Chancellor Schroeder started calling for greater EU concern about manufacturing industry in a new industrial policy. The term ‘industrial policy’ was already back in late 2002 and more prominently in 2004.21 Repackaging also took place after the Lisbon goal of making the EU by 2010 ‘the most competitive and dynamic knowledge-based economy in the world, capable of sustainable growth with more and better jobs and greater social cohesion’ had been set in 2000. The formulation of the goal, though vague, strongly suggests a similarity to the OECD/EU definition of competitiveness, mentioned above. But such a ‘holistic’ approach is likely to conflict with the call by Europe’s political leaders to have more specific regard to manufacturing industry as restated in the European Council in October 2003 and in earlier public letters by Schroeder, Chirac and Blair to Commission president Prodi. Their concern was caused by a fear of ‘deindustrialization’ in the light of globalization and of too costly regulatory initiatives such as REACH (the draft chemicals laws proposed in 2003). Courageously, the Commission preferred to dismiss the deindustrialization thesis on analytical grounds,22 while offering major concessions in unburdening the chemical industry in the REACH process. In doing so, it has stuck to what it calls an ‘integrated approach’ to industrial

European industrial policy 55 policy. This fits the holism of Lisbon but remains very hard to come to grips with, let alone to assess rigorously in terms of effectiveness. The treaty language of art. 157, EC and Lisbon concurs in that the direct policy influence of the EU level is modest and selective, at best, except insofar as the (open) internal market and competition policy is concerned. Lisbon is largely a matter of the member states (see Pelkmans and Casey, 2004, for elaboration). In other words, the container of industrial policy is not only filled with a great many disparate policies but most of these policies are effectuated at the national level. Starting at pure EU initiatives, without realizing this context, would be flawed. 2.2 Does EU services market integration matter for EU industrial policy? Services generate about 70 per cent of the value-added of the EU. Clearly, from the demand side the development of the services sector is important for industry as it is bound to be the largest aggregate customer. The question we address here is on the supply side. Do services matter for the dynamic economic performance of EU industry and, if so, does EU services market integration matter for EU industrial policy? Although all services may have some significance for efficiency, productivity trends and innovation in industry, it is useful first to distinguish government from market services and, in turn, business-related services from other market services. We focus on businessrelated services as they interact most intensively with industrial manufacturing.23 In the EU-25 of 2003, business-related services provided 37 per cent of all market employment, compared to 24 per cent for manufacturing and construction together. Although businessrelated services are active suppliers to each other and to other services sectors (and, to some extent, even government services), some 30 per cent of the intermediate output from business-related services is consumed by manufacturing services.24 In the EU-15 businessrelated services employment has grown from 32 per cent in 1980 to 41 per cent in 2001, whereas the share of manufacturing fell from 27 per cent to 17 per cent. Business-related services are also a growth sector in cross-border trade, especially in the internal market, despite the persistence of a range of barriers: measuring intra-EU trade from the import side, the average annual growth of business-related services trade over 1996–2002 was 14.8 per cent (compared to 9.5 per cent for goods). Their significance for the internal services market is overwhelming given that, in 2002 for EU-15, these services represented no less than 61 per cent of intra-EU services trade, with tourism (24 per cent) being second. The rapid growth of business-related services is usually attributed to outsourcing, implying that a good deal of the fall of manufacturing jobs represents in fact a migration of jobs to services. This is undoubtedly the case, resulting from a search for cost cutting and a focus on ‘core’ business, but there are several other, more deep-seated reasons for this growth as well. Certain relatively new services have become vital for industrial performance, IT services and software above all, but also logistics (with scale-driven production, logistics are critical to performing in the entire internal market both for deliveries and for just-in-time processes). Furthermore, with production ‘europeanizing’ in Europe-wide firms, a range of professional services and selected other ones no longer have to be sourced locally because these providers may ‘follow’ the europeanization of the MNCs either by local establishment or by cross-border services on request (but based on a relationship of trust, indispensable for services supply). It is therefore important for the competitiveness of European industry that the internal markets for services function properly. The recent liberalization of network industries in

56

International handbook on industrial policy

the Union has thus far selectively helped business (industry and services) to cut input costs such as in lower telecoms, postal, air transport and energy bills for a given set of services. The liberalization of financial markets has been favourable to wholesale customers (sizable business clients) but far less, unfortunately, for SMEs and private clients. And there is a hesitant beginning of greater market pressures for professional services and more liberal regulation for example on entry. In marketing and advertising, serious restrictions in some EU countries have remained (despite vigorous efforts, led by the Commission, to overcome them) on the grounds of consumer protection.25 It is probably inappropriate to generalize about the development of service quality and of the emergence of innovative services but it is possible to identify submarkets where EU liberalization has improved the quality of services26 and significantly widened the range and innovativeness of services.27 So the internal market for services matters to industry and a deepening, widening and better functioning of it is crucial for the competitiveness of manufacturing in Europe. The deepening and better functioning of the internal services market is currently hindered by a great many barriers and restrictive regulations. To a large extent they are tackled or at least consistently brought into an EU framework (say, for harmonization where justifiable) in the so-called ‘services draft directive’ in COM (2004) 2 of 13 Jan. 2004, also known as the Bolkestein directive. In particular, for SMEs this directive is crucial. More generally, after a long adjustment period, the scope for arbitrary protection and new restrictions would become minimal once this framework directive was in force. Remaining regulation would have to be justified or liberalized under free movement or establishment. It would improve economic performance, too. Both CPB (Kox, Lejour and Montizaan, 2004) and Copenhagen Economics (2005) expect greater cross-border services trade and establishment (by up to 35 per cent more) in the Union, counting only the direct effects, not the long-run impact of competitive exposure. However, European industry should not focus solely on the internal market for services. A very dynamic subset of services, namely ICT services, turn out to be distinctly less dynamic than their counterparts in the USA,28 and this is not so much due to defects in the internal market as to other restraints of competition and attitudes being more riskaverse, besides less deep ICT investment and a skill gap (see Denis et al., 2005). In this sense, horizontal policies which would probably benefit both services and industry are relevant here, but most of these issues (such as educational; skill gaps; subtle restraints of competition in services via excessive reliance of consumer protection; hindering rapid change because of employment protection laws) have to be addressed at the member states level, given the current assignment of powers to the EU level. 2.3 Is European infrastructure crucial for European industry? In a trivial sense, of course, infrastructure matters greatly for any modern economy. The questions for this contribution are, first, whether it matters in a special way for (European) industry, and, second, whether the EU does or should have a role in (European) infrastructure in a way especially of interest to industry. It is good to realize that the second question was long taboo in European integration. Before the Maastricht treaty, no provision about or reference to infrastructure could be found in the treaties. Consider the sharp contrast with the internal ‘economic integration’ of Canada and of the United States. Both countries only began to integrate after major infrastructural investments (especially, East–West, both railroads and highways) had been initiated. In Europe, post-war

European industrial policy 57 cross-border infrastructure was weakly coordinated, on a voluntary basis, in the framework of the UN–ECE in Geneva and/or bilaterally. Both approaches often failed, remained suboptimal or suffered from long delays. Infrastructure was regarded as a national, not in any way as a European, issue for reason of domestic sensitivity, because of budgetary burdens and, not least, because of a perceived or actual discrepancy between payment for and use of the facilities (especially in case tolls were rejected as a user charge). The taboo was broken in 1984 by the European Round Table of Industrialists in its report on ‘Missing links’ (ERT, 1984). Member states, trapped in their inward-looking inhibitions, failed to appreciate that broader free trade, investment and, later, services trade as well, let alone the deepening of the internal market, necessitated a European perspective on infrastructure, besides national and local networks. Transiting countries was often unnecessarily costly because links were missing or inefficient routing was imposed by a purely national outlook. This applied to roads and freight rail but also to network infrastructure in (then not yet liberalized) public utilities. A European market with business quickly ‘europeanizing’ required European infrastructure without having to rely solely on national investments for local needs. Writing more than two decades later, the view has meanwhile been widely accepted, but its consequences seem too painful for member states’ governments for a wholehearted embrace of the idea. The first question hinges on the role of the state and on the specificity of industry as a user of infrastructure. Because of profound structural changes in the economy and increasing congestion problems, European industry often complains nowadays that (public) infrastructure suffers from neglect and underinvestment, in turn worsening the business climate in Europe and acting as a threat to long-run productivity and economic growth. This would apply in particular to Western Europe. In addition, the special problems of cohesion in countries such as Portugal and Greece (and, until recently, Spain) as well as the Central European member states and the candidates Romania and Bulgaria, necessitate additional infrastructure investments as a condition for catch-up growth, since the lack of it would reduce the attractiveness of FDI and make exports as a handmaiden of their growth more difficult. Of course, this is linked to the second question by arguing for pan-European investment schemes facilitating their exploitation of the opportunities provided by market integration. Insofar as Western Europe is concerned, the problem of neglect requires the answer to three difficult empirical questions. First of all, has there been an actual downward trend in real EU-15 infrastructure investment by governments? The careful survey by Valila, Kozluk and Mehrotra (2005) brings out that indeed there has been a downward trend in this public investment for decades but public investment was still growing (albeit slower than GDP) such that the stock of public capital has grown significantly. The authors empirically reject a number of popular assertions about why the down trend happened, including the strictness of EMU rules, the decreasing need for infrastructure, or privatization. What remains as a statistically significant explanation are long-drawn-out periods of fiscal consolidation long before EMU and disconnected from it (the UK being the most severe case). For Central Europe, they find that, during transition, infrastructure investment did not collapse and that the absence of reliable estimates of public capital stock render it impossible to state whether it is too low for their development at this stage. Another implied question to be answered is the relation between public capital and economic growth. This complicated issue has generated a huge literature. The survey by

58

International handbook on industrial policy

Romp and de Haan (2005) finds that the recent, more sophisticated, analyses do find a positive relationship but far more modest than the influential one in Aschauer (1989). The heterogeneity amongst countries is also great, for reasons of the size and quality of the existing capital stock, non-linearities caused by network effects and institutional and political factors. The authors warn that additional investment, though perhaps politically more rewarding, is no good measure if it pulls away money required for a proper maintenance of the physical stock in place, a problem not rarely encountered in Europe. A third question to be answered is whether the EU-15 suffers from a lack of public capital, that is, whether the stock of public capital is ‘suboptimal’. Again, this cannot be a pure industry or even a pure business issue. Both modelling (see, for example, Romp and de Haan, 2005) and empirical analytical problems (see Kamp, 2005) explain why it is so difficult to come up with authoritative answers. Kamp holds that, in 2000, all EU-14 countries but two had ratios of public to private capital that exceeded the growth-maximizing value; the same is true for the EU-14 average. Taking uncertainty margins into account, Kamp suggests that ‘in most EU countries there is neither a shortage nor an excess of public capital . . . relative to private capital’ (2005, p.85). If long-run trends are extrapolated, the author finds that Austria, Belgium and the UK currently seem to grossly underinvest in public capital. Recent public–private partnerships in the UK might compensate but this is not at all guaranteed. Thus, if European industry insists on infrastructure spending being increased in EU-15 (for Central Europe, this is not controversial) as a precondition for improved competitiveness in general, it would seem that economic analysis does not support such an argument. It is not excluded, however, that specific elements or sectoral shortcomings might have to be addressed because the ‘public capital’ literature deals with larger aggregates. The question for horizontal EU industrial policy is whether the EU level should be involved to ensure or at least foster a European perspective and the development of ‘European’ infrastructure in the sense of missing links, greater overall (EU) network efficiency and cross-border (seamless) interconnections for network industries sharpening competition in such former ‘national’ utilities. It is here that subsidiarity strongly suggests a justified EU role (see Pelkmans, 2005b) given cross-border externalities and, sometimes, scale. This was hesitantly recognized in the Maastricht treaty as TransEuropean Networks (TENs), but without much financial funding (mainly feasibility studies and minor subsidies) and no other competences than voluntary agreements in TEN projects. The EC White Paper on Growth and Competitiveness of December 1993 (see EC, 1993) contained the first economic underpinning of the TEN approach, followed by the Christopherson report of 1994 and the Essen European Council decision about a range of TEN projects. As a generalization, it is not unfair to say that TENs have generated more disillusions than concrete improvements in European infrastructure.29 Nevertheless, the awareness of a number of serious shortcomings in the ‘Europeanness’ of networks or the absence of them has undoubtedly increased. Of course, the regulatory, environmental, spatial planning and budgetary constraints lead to enormous delays even for projects which tend to be supported in the wider EU interest. In network industries such as gas and electricity, markets themselves normally pay for infrastructure. The question here is mainly whether interconnectors across borders are built (and by whom), knowing that this would invite more competition given third-party access. In freight rail the relevance for industry is probably most obvious. Rail freight in

European industrial policy 59 Europe is expensive, slow and of mixed quality as far as business customers’ wants are concerned. Liberalization (see di Pietrantonio and Pelkmans, 2004) should help improvements on all three accounts. One critical condition for that to happen is access to crossborder, long-haul freight rail corridors. Thus freight rail liberalization began with freeing the TEN RF corridors for competition and free entry. Still, owing to the dual use of rail track for freight and passengers in Europe, which is very costly for freight (some 40 per cent cost increase only on that account), capacity problems are expected once a modal shift to rail at lower cost and better quality takes place. A somewhat different issue played in air transport, where delays and expensive rerouting of planes were due to the absence of a common air traffic infrastructure (with tremendous inefficiencies as a result) and a routing system where national interests (usually the military) were allowed to assume a very large part of the air corridors without any weighing of competing European civil aviation interests. In telecoms and the Internet, in contrast, the liberalization of the former has generated major expenditure on B2B high-speed networks (indeed, with excess of so-called ‘dark fibre’) and Europe is rapidly responding to the broadband challenge for B2C and C2C connections (see COM (2004)759 of 2 Dec. 2004 on telecoms), especially in the EU-15. A special case of electronic networking, where the EU has assumed a true European perspective, is the powerful GEANT network for universities (with terabit speeds by 2010) and some 10 per cent of direct EU subsidies. The GALILEO satellite GPS system (with a major EU contribution) should allow surface and maritime navigation and facilitate logistics and freight tracking systems. Finally, for road networks, the revised Eurovignette system, adopted in 2005, allows mark-ups for expensive bottlenecks (more often than not, cross-border ones like the Brenner Pass), thereby reducing a significant obstacle to upgrading precisely of cross-border infrastructure. More generally, the ‘missing links’ idea would seem to be better pursued after October 2003, when the European Council of government leaders agreed that the cross-border segments of infrastructure could receive much higher percentages of EU support.30 Except perhaps on a case study basis, the author is not aware of analytical work singling out the (special?) relevance of industry for TENs. Should one call the TEN approach a kind of horizontal industrial policy, with co-benefits for consumers, and more indirectly for growth? This seems to go too far and, moreover, fails to take into consideration that TENs might crowd out other investment that, perhaps, might have a higher social marginal productivity. The latter problem has sometimes played a role in EU cohesion policies which emphasize infrastructure investments with a view to improving the capacity of poor regions to participate in the internal market. Depending on how this is done (see, for example, Martin, 1999; de la Fuente and Vives, 1995) it may or may not aggravate agglomeration effects and might boost economic growth in poor regions at the cost of growth in a country as a whole. In Central Europe, if multinational investors look for a production base with main outlets in Western Europe, clearly, the quantity and more often the quality of infrastructure will have to be improved over a period of decades. EU cohesion and structural funds can provide so much support that the risks of crowding out could be minimized (although co-financing will always imply some risk in this respect). 3 EU framework conditions for industrial policy Both national and EU industrial policies are severely constrained by the EU framework of the internal market, complemented by its common policies and cohesion. This is not

60

International handbook on industrial policy

surprising as it reflects the fundamental choice of member states to achieve economic progress, formulated in the EU’s economic objectives, via a deep and well-functioning internal market. The internal market is governed by far-reaching instruments such as free movement and free establishment which purposefully reduce national policy autonomy in areas which might affect, actually or potentially, cross-border economic intercourse. Article 4, EC, speaks about the internal market being governed by the principle of an ‘open market economy with free competition’. Even without entering into the details of internal market governance, including competition policy and different degrees of common regulation where justified, it should be clear that there is not much scope for a direct impact on industrial activity. Indeed, (1) EU industrial policy cannot itself be inconsistent with this ambitious regime, even if the additional constraint of art. 157, EC (see section 2.1) did not exist; (2) the notion, implied in art. 157, EC, that national industrial policy has more degrees of freedom and greater choice than EU industrial policy must be understood against these framework conditions resulting from the internal market regime. Of course, since member states have explicitly agreed with this set-up, one could interpret the framework as a quasi-constitutional denial to fall back on interventionist industrial policy in the future. In Figure 3.2 the EU framework governing industrial policy is spelled out in some detail. Both so-called ‘negative integration’ (the establishment of the internal market via prohibitions for member states to intervene with cross-border flows – see top of the figure) and ‘positive integration’ (that is, where member states cooperate or regulate together or centrally – see middle and bottom of the figure) constrain the industrial policy of member states and of the Union in numerous ways. Most of these aspects are under a ‘hard’ legal regime, ultimately enforced by the Commission as the ‘guardian of the treaty’ and, where necessary, by the European Court of Justice. All the main areas have been specified. Attention should also be called to the column of guiding principles. Several of those principles are of a constitutional nature and can suffice, depending on the case, to nullify specific measures of national industrial policy. Two are more policy-oriented. The ‘better regulation’ principle has been embraced by the Union as a fundamental approach for all EU regulation, be it ‘new’ or when reviewing packages of the existing regulatory ‘acquis communautaire’. It is briefly discussed in section 4. Moreover, member states are already bound by the legal duty to respect ‘proportionality’, which says that measures should go no further than what is needed to accomplish the objective. The notion of ‘cooperative federalism’, coined by political science, is meant to express the idea that the two levels of government are not effectively separated (as is often the case in the USA) but, in contrast, often work together to pursue specific objectives or implement rules or codes. This is important for horizontal industrial policy where joint initiatives with a view to complementarity are routine. For reasons of space, further study of Figure 3.2 is left to the reader.31 It is important, however, to appreciate the systemic obstacles this framework generates against a return to old, interventionist and politicized industrial policy. For a while the probability of a policy reversal seemed slim. But, since 2004, if not before, leading politicians in countries such as France, Germany and Italy, and sometimes industrialists, have occasionally called for a return to industrial policy. These calls tend to be vague because it is well known that national autonomy is greatly circumscribed and, in addition, the European mind-set is much more pro-competition or pro-market than several decades

European industrial policy 61 Key Aspects Establishment of internal market

Proper functioning of internal market

Regional/ structural & cohesion policies

Substantive Elements

Principles

Free movement • goods • services • capital • codified technology • workers (host c.c.) Free establishment • entrepreneurship • FDI

– non-discrimination – proportionality – no frontiers

Overcoming market failures (a) via rules & codes • harmonization/approximation – old & new approaches – global approach (C.A., risk-based) • common regulation – uniform rules (and selected, autonomous agencies) • mutual recognition (no EU rules) • competitive public procurement • standards and voluntary C.A. Overcoming market failures (b) via competition policy • complementarity EU & member states • EU → anti-trust → network industries (see (a)) → state aids prohibition/control Common EU policies • transport (link with infrastructure) • trade (see ‘specific industrial policy’) • agriculture (link with cohesion/ environment) • environment (IM priority, cost/benefit & competitiveness tests)

– subsidiarity – ‘better regulation’ – ‘cooperative federalism’

– measures strengthening hard & soft infrastructure (locational attractiveness) – measures for retraining/skill enhancement – support for adjustment after industrial decline/rural stagnation – cross-border regional facilitation – removing environmental legacies

– long-run catch-up growth – co-financing & additivity

– ownership (EU neutral)

– ‘open & competitive’

Notes: C.A. = conformity assessment (testing, certification, accreditation); host c.c. = host country control (protectionist).

Figure 3.2

EU framework governing industrial policy

62

International handbook on industrial policy

ago. Looking at Figure 3.2, national industrial policies cannot affect free movement and free establishment. The case law in this area goes very far and severely limits policy action that is in some sense discriminatory (for companies from other EU countries) or distorts free movement/establishment. Thus ‘golden shares’ of national governments in privatized former utilities have been condemned by the ECJ as going against the free movement of capital, here, the equal opportunity of all potential shareholders in the Union. Similarly, in overcoming market failures of the internal market (see (a) in the second row, second column), all five items greatly limit national discretion, even if full uniformity of EU rules is not applicable. Thus mutual recognition even dispenses with common rules, but it imposes equivalence and otherwise falls back on the strict requirements of free movement (see, for example, Pelkmans, 2002). Overcoming market failures via competition policy exercises equally constraining effects (see (b)). In industrial policy one tends to think that this applies with particular force to state aid control (which is correct) but one should not forget that the Commission has taken a restrictive view on crisis cartels, that network industries (where interventionist industrial policy was routine only 20 years ago) are being or have already been liberalized,32 and that merger control (begun at EU level in 1990) has further reduced the options for national policy makers. Last but not least, the internal market being ‘open’ to the world economy (at least, for industrial products) further curtails the toolkit of national industrial policy, if ever one should wish to return to interventionism. Quotas have been abolished in GATT and cannot come back. Industrial tariffs are low or zero and are ‘bound’ in Geneva, which means that any increase beyond the bound level will cause an upheaval and, in any event, will have to be compensated. In rare instances, a trade measure of ‘damage control’ (think of the 2005 restraints on China for selected clothing items) is possible for a short period, but one cannot build a true policy on that. Just before this book went to the press, France announced that it had selected ten sectors which must be prevented from falling into ‘foreign’ hands. The reader should note that the way this was suggested as being compatible with European law was that national security and military goods were involved, so that art. 296, EC would apply as well as an escape clause in the EC Takeover Bid directive 2004/25. This escape route is so extreme that its use is already a powerful signal of how strong the constraining effect of the EU framework is.33 4 Horizontal EU industrial policy The main emphasis nowadays in the Union is on horizontal industrial policy. A sharp definition of what ‘horizontal’ means in actual practice is not easy. The term probably originates from the desire not to intervene sectorally (vertically) and to defend policy makers against ad hoc pressures for rescue operations and direct interventions in single enterprises. Such ‘specific’ industrial policy, another term often used, is not desirable in case declining industries or badly performing companies ask to be helped by the state, nor is the state any better than market players at ‘picking winners’ for the future. At the same time, the boundaries between the framework conditions (see section 3) and horizontal policy are not always clear-cut and the distinction between the two is, at times, fluid. Another problem in setting out a proper picture of EU horizontal industrial policy is that the overall scope of this approach is vast, too vast for a chapter like the present one. To get around this difficulty, this section will employ two summary representations of crucial EU policy papers which give a rich impression of the EU horizontal approach

European industrial policy 63 • EU industry has improved its competitiveness • but faces mounting international competition, emergence of the information society, continued needs in less developed regions, inadequate European networks, and demand for intangible investment

Drive

→ European Information Society action plan → four horizontal priorities (see below)

Priorities

Promote intangible investment – general support for intangibles in investment (via tax) – research, with regard to market needs & spin-offs – promoting alliances – promoting quality – promoting clean technologies, with economic incentives – incorporate vocational training in other policies

Develop industrial cooperation

Ensure fair competition

Modernize role of public authorities

– removal of legal/ fiscal obstacles to industrial cooperation – facilitate (BC Net) – round tables – various cooperation schemes for cooperation with companies in Latin America, MED countries and Asia – energy & standards cooperation with Central Europe

– a post-Uruguay Round (EU) agenda on international competition rules – rendering contingency protection in EU trade policy more efficient – stricter & more coherent EU state aids regime – pursuing the internal market for network industries

– regulatory reform – greater simplification EU laws (later Molitor report and SLIM) – improve administrative cooperation between member states & Commission (that is ‘cooperative federalism’) – streamlining EU industrial policy decision making in various councils

Note: Horizontal aspects only. Source:

COM(1994) 319 of 14 September 1994.

Figure 3.3

EU horizontal industrial policy proposed in 1994

and which can be used for annotations in the text. Elaboration of the economic analysis underlying the initatives (or the lack of it) is impossible for reasons of space. Doing justice to the many interesting initiatives and their rate of success is equally unfeasible. In Figures 3.3 and 3.4, the Commission papers on industrial policy of 1994 and of 2004 are summarized. After the treaty revision from Maastricht came into force the EU did not wait a moment before taking multiple action on the industrial policy front. Late 1993 and 1994 yielded no fewer than three major initiatives (not counting the Christopherson group on TENs, see section 2.3). First, the White Paper on Growth, Competitiveness and Employment of December 1993 links industrial policy with macro variables such as growth and employment. It stresses how crucial it is for the EU to be present, if not strong, in the markets ‘of the future’ (for example, telecoms, biotech, environmental protection, new materials and energy). Commissioner Bangemann personally led a group on the Information Society,

64

International handbook on industrial policy

working out an action plan in the autumn of 1994. At about the same time, the Commission published its first post-Maastricht industrial policy paper which is largely of a horizontal character. Extracts from the latter are shown in Figure 3.3. The idea is to articulate further the current ‘unprecedented effort to restructure and innovate’ in the EU, induced by EC-1992 and greater trade openness. The European Information Society can be seen as ‘sectoral’ although it is of course an extremely broad and elusive concept. Otherwise, four priorities are identified, all clearly ‘horizontal’. The first, labelled ‘intangible investment’, turns out to be a collection of initiatives which would seem to have in common that they are formulated in a pro-market fashion (for example, research with more regard to market needs and aiming at spin-offs; economic incentives for clean technologies; promoting quality). As could be expected given art. 157, EC, the EU level has little more than persuasion and some, rather limited, funding available, whereas selected results will have to come from the member states, such as tax treatment of intangibles and economic incentive measures promoting clean technologies (with possible exemptions at EU level). Some initiatives, such as quality (which should be driven by market incentives), vocational training and business alliances, are mainly at the enterprise level. Indeed, a quality policy at the EU level was resisted later as inappropriate even if voluntary certification and testing are (weakly) encouraged via EOTC. On alliances, the EU level walks a tightrope between strong powers disciplining individual (anti-competitive) alliances and soft, non-interventionist encouragement of high-tech alliances in general. This is also evident from the second priority, where the EU level can mainly try to remove legal and fiscal barriers to ‘industrial cooperation’ (which requires directives) and otherwise little more than encouragement, a ‘marriage bureau’ (like BCnet) and active liaison functions in other continents under trade and cooperation agreements. The third priority (fair competition) is remarkable because it is mainly about a post-Uruguay agenda for worldwide competition rules, a wish meanwhile removed from the Doha agenda under the pressure from developing countries. Other items, such as a stricter and more coherent state aid regime,34 and the pursuit of the internal market for network industries, have been taken up with vigour and much has been achieved in the meantime. The desire to render contingency protection (such as anti-dumping) more efficient seems benign but one can ask questions about the ‘fair competition’ nature of EU anti-dumping. It has often been suggested that, at least in part, EU anti-dumping serves as the last instance of protectionist and selective industrial policy (see Evenett and Vermulst, 2005, for a survey). The fourth priority (modernizing public authorities) is mainly about regulatory simplification and reform, which has grown into a major plank of EU (and increasingly of national) policy making. The 1994 Commission paper anticipates the Molitor report (1995) on regulatory simplification for competitiveness as well as the SLIM programme (on simplification, see Pelkmans, Labory and Majone, 2000, for a critical review), initiatives which have since been completely overtaken by a far more intrusive and broader approach of RIAs. The prudent idea of streamlining industrial policy decision making has served as a prelude for the eventual merger of the internal market, industry and research councils with the Competitiveness Council, a decade later. The overall impression of the 1994 Commission paper is that the EU was still searching for an effective common approach between the member states, the EU level and EU industry as the principal actor. The horizontal approach of 2004, summarized in Figure 3.4, is definitely wider in scope and almost certainly better equipped with ideas,

European industrial policy 65

Drive

• worrying combination of (too) slow structural change and selective (but not general) ʻdelocalizationʼ • disappointing EU performance in productivity growth, research and innovation • too weak risk taking & entrepreneurship • EU still has assets and many opportunities (for example intra-branch) • Eastern enlargement → reorganize value-chains for competitiveness and growth

Priorities

→ ʻbetter law-makingʼ approach, combining a ʻdeeperʼ internal market, with a lower regulatory ʻburdenʼ → ʻintegrated approachʼ to competitiveness, in five areas (knowledge, the IM, cohesion, sustainable development, worldwide dimension)

Integrated (multi-policy) approach competitiveness

Better EU law-making – IM and one-stop-shop regulations benefit industry → pursue – RIAs – alternative methods to regulate – a fresh look at competitiveness impact of existing acquis – verify cumulative effects of regulation

– European Research Area (IM for R&D) (more funding, reversing brain-drain, better framework conditions for business R&D) – innovation policy (business-driven, diffusion, funds, innovation systems) – human capital & skills – competition policy for innovation & tech. transfer and new merger control (with ʻefficiency testʼ) – deepening/widening internal market (services, standards, financial markets, company law, fiscal) – cohesion/industrial clusters, TENs – incentives for ʻgreen businessʼ, clean energy/tech. – facilitating access-to-markets outside the EU (worldwide; neighbourhood strategy)

Notes: Horizontal aspects only; IM = internal market. Source:

COM(2004) 274 of 20 April 2004.

Figure 3.4

EU horizontal industrial policy proposed in 2004

funds and (sometimes) instruments. This contrast is not to be interpreted as that, therefore, it is more ‘effective’ in accomplishing goals or somehow ‘better’. Such an assessment is very difficult to make, given the vast areas of activities, the constraints in terms of powers specified in Figure 3.1 and the potential or actual gaps between the desired results in policies and the genuine impact on European business. It is good to see the main

66

International handbook on industrial policy

reasons for the greater ambition in horizontal EU industrial policy after the turn of the century. The instrumental reasons include (a) larger funding over time for research and for technology in the respective framework programmes; (b) a more useful composition of these funds, with greater concentration on fewer areas, better exploitation of crossborder networking and the links between researchers/engineers and business strategists more systematically pursued; (c) a focus on ‘hard’ instruments where possible (that is, legal or financial rather than talkshops); (d) an increase in funds for cohesion, which can be used for horizontal industrial policy (and only since the enlargement of 1 May 2004 has there been a threat of decline per country/region). But there were also reasons emerging from the very top of European politics, such as the fear of delocalization or deindustrialization (in particular, China and India have not failed to leave a deep impression on political leaders). An explicit public appeal was made by three leading prime ministers (an initiative of Chancellor Schroeder) to have much more concern about the impact on European industry when regulating or intervening at EU level. Probably the greatest drive for a more aggressive industrial policy arises from the frustration at prolonged, very low growth in Western Europe since late 2001 and the hopeless failure of the Lisbon process, as far as economic performance is concerned (see Kok et al., 2004). One could, however, argue just as well that, given so many powerful drives and the predicament of Europe, the industrial policy proposed is still too little, too late. Figure 3.4 presents the drives as set out in the Commission proposal. One can appreciate that the Commission convincingly argues against a doom scenario of delocalization and underlines that reallocation of productive factors and structural change more generally are inevitable processes (for productivity trends to go up) which should not be resisted. However, its analysis shows that a too slow structural change, combined with selective delocalization, is a menace to Europe’s recovery and growth trend. The Commission is right in emphasizing the problem of too weak risk taking and entrepreneurship but, surprisingly, nowhere in the paper is its initiative on ‘Entrepreneurship in Europe’ even mentioned.35 Of course, it is exceptionally difficult to influence deep-seated inhibitions against risk taking in Europe, but both the limited impact from taking barriers away for SMEs and de novo companies and the insistence that risk taking be rewarded and not discouraged is undoubtedly worth including in its horizontal industrial policy. Fortunately, the drives in Figure 3.4 also include opportunities, not least the Eastern enlargement. Horizontal industrial policy is nowadays based on what is called an ‘integrated’ (across a range of policies) approach to competitiveness.36 Thus the focus on only two priorities is a little misleading, although there is no denying that many of these policies are related. The first priority, ‘better lawmaking’, goes beyond industry alone, of course. However, since much of the acquis can be found in the area of goods regulation, occupational health and safety, and environment, affecting industry directly, European industry regards a lower regulatory burden as a handmaiden to competitiveness. The question is really whether the EU can deal with market failures in a least-cost fashion without compromising the objectives. Since these organs are the supreme political organs in the EU, it seems hard to imagine that they will accept such self-discipline all the time. Moreover, a good deal of potential excess burdens might well arise from the detailed implementation and this can depend on what member states do in the transposition process or what ‘comitology’ (the EU/member states technical committees dealing with implementation or revisions, an instance of ‘cooperative federalism’) or the Commission decides in the

European industrial policy 67 follow-up of political decisions. After more than a decade of wavering, the EU has finally introduced RIAs for all proposals or amendments of a non-trivial character.37 Detailed manuals38 have meanwhile been developed and are compulsory for internal RIA preparation of the Commission. Amazingly, an interinstitutional agreement between the EP, the Council and the Commission extends the RIA process to all major amendments of Commission proposals in the process of adopting EU legislation. This chapter is not the place to assess the RIA approach and its effectiveness for the EU public interest (but see Radaelli, 2003 and Renda, 2006, for instance). It is also too early to assess whether the ‘regulatory burden’ for European industry is likely to go down as a result of these initiatives. One should not forget that there may well be justifiable reasons for those burdens if market failures have to be overcome. The empirical issue is, rather, whether European industry is unduly burdened because interventions are not based on market failures or not on well-defined objectives, or entail higher costs than benefits for society (hence a regulatory failure replaces a market failure), or because costs of rules are disproportionate even if the net social benefit is positive. The biggest case so far (on REACH, the new chemicals regulatory regime) shows how complex such assessments are likely to be (Pelkmans, 2005b). It is nevertheless progress that these issues are now squarely faced in the Union and thus can at least pre-empt true excesses while forcing political decision makers to be clear and accountable for the choices they make. The status of the other two items (impact of the existing acquis and the problem of ‘cumulative impact’) is much less clear. They prompt large question marks because, by definition, these are far more sensitive, and inevitably will take a decade or more to tackle satisfactorily. The multi-policy approach to competitiveness is rich and better endowed with resources and powers than was possible ten years ago. Every single item of the list under this heading in Figure 3.4 is worth a chapter in the present volume. The following can be no more than a few brief annotations for the guidance of the reader. We shall limit ourselves to three areas: research, innovation and human capital. The European Research Area has emerged from the Lisbon process’s requirement that public and private R&D in the EU should be pushed up from 1.9 per cent (in 2001) to 3 per cent of GDP. Even more than public R&D, it is business R&D in Europe which is lagging behind the USA and Japan.39 The subsequent action plan40 envisaged three lines of policy: improving the effectiveness of public support for research, redirecting public budgets towards R&D, and improving the framework conditions for research. For the first line, the so-called ‘open process of coordination’ in the Lisbon framework was utilized, which is of doubtful utility, to put it mildly (for example, Pelkmans and Casey, 2004). The Sapir report (Sapir et al., 2003) rightly attacked the avoidance rather than the search for excellence and recommended a European Research Council based on the competitive model of the USA. But, since the bulk of public money for R&D is national in the EU (some 94 per cent), it would only help if a good deal of this national money was also subjected to competition and peer review on a European basis. Redirecting public money towards R&D is always tough, but particularly so when budget constraints are tight in a period of very low growth. At EU level economists have suggested a radical switch away from agricultural support to R&D,41 but this is fiercely resisted. The framework for research has everything to do with the overall prospects for top quality business and academic research in the Union and with the motives behind the brain drain of many thousands of researchers towards the USA. These issues, too long disregarded in the

68

International handbook on industrial policy

Union, are now frantically studied and the debate is at last beginning to be led by leading researchers themselves.42 However, the inhibitions are deep-seated in academic Europe and career prospects, sustained funding of basic research and a range of other problems will have to be addressed in a convincing fashion. The Union has long been concerned about Europe’s incapacity to transform inventions into innovations, but little more has been done than ‘policy by speech’. This has changed since the beginning of the 21st century.43 After the Barcelona European Council of March 2002, mostly devoted to an acceleration of the Lisbon process, a series of initiatives were undertaken, such as the European Innovation Scoreboard44 and a new EU approach.45 The ‘open method of coordination’ approach among the member states which amounts to benchmarking and learning from one another’s innovation systems became more structured.46 For the medium term, the innovation activities are going to be intensified and closely linked with the policies to foster ‘entrepreneurship’ in Europe. As part of the medium-run EU budget proposals for 2007–2013 (the so-called ‘EU Financial Perspectives’, setting expenditure ceilings for a period of seven years, with the purpose of pre-empting annual budget battles between the member states or between the Council and the EP), the Commission proposed a Competitiveness and Innovation programme, one subprogramme being that on Entrepreneurship and Innovation.47 A major plank of this subprogramme is support for young, innovative SMEs in Europe, in particular improving access to start-up finance. It is held that European capital markets are too risk-averse and companies without an established reputation and with high-risk investment needs tend to be locked out. Already, since 2000, the EU has run a Risk Capital programme to promote and complement venture capital activities. In these new proposals four financial instruments would be established: venture capital support for innovative SMEs in their early stage, as well as in their expansion stage, guarantees and counterguarantees for SME loans, and capacity building. Daringly, the impact assessment suggests the effect on job creation might be as high as nearly 400 000 jobs, although the analytical basis for this statement is not provided. It is good to see that this new subprogramme also undertakes a range of horizontal activities which have to do with ‘innovation governance and culture’. The emphasis on human capital and skills is of course another very broad (and not solely industrial) area of attention. Its importance for innovation and growth is well known.48 The EU level has only very limited competences here as this is a national (and, at times, regional) power par excellence. The broad scope as well as deep concerns about weakening educational performance and lifelong learning in Europe has prompted the Lisbon process to give it priority. Since the Lisbon process is member states-driven, there is no conflict with subsidiarity. But it also means that the results are uncertain and disparate over the Union. Human capital and skills at EU level are more a derivative of employment policy and of cross-migration issues than of industrial policy, even if the concerns largely overlap. In the Kok (2003) report on European employment policies,49 a separate chapter is devoted to more investment in human capital and lifelong learning. Similarly, in the European Employment Strategy,50 it plays a role, with specific quantitative indicators as interim targets. Where it comes to cross-border migration, the EU level is more appropriate and recent activity has been commendable.51 The Lisbon process has also addressed more generally the question of knowledge promotion on a European scale and of quality assessment in higher education. This began in earnest with the ‘Education and training 2010’

European industrial policy 69 programme.52 At university level the initial reservations would seem to have reduced and several initiatives by the Commission and the universities themselves in 2005 may make a more European approach, belated as it is, more effective.53 The irony of Europe taking no less than 50 years after the Messina declaration of May 1955, which began the ‘relaunch’ of European integration culminating in the EEC treaty in 1957, to come to terms with the idea of a Europe-wide search for excellence and competition at intercontinental level is particularly great. In Messina, the foreign ministers called for a European top university to pre-empt a brain drain of the finest European minds to the USA. Subsequently, negotiations took 17 years, aiming essentially at preventing a European ‘Harvard’ coming into being, solely because the elite club of the oldest universities did not accept the emergence of a possibly superior competitor.54 For a reversal of the current brain drain to the USA, however, much will have to be fundamentally reformed in Europe. The other four areas mentioned in the right/bottom column of Figure 3.4 are in fact other policy areas with an interface with industrial policy: aspects of EU competition policy, further deepening of the internal market, cohesion and questions of access to third markets. This testifies to the incredible scope and complexity of horizontal industrial policy. At times, it is more insightful to call this approach the constant attention to opportunities and risks for industry in Europe when it comes to other EU policies and strategies, rather than a policy of its own. 5 Sectoral and specific European industrial policy Sectoral and specific industrial policy of the old days has largely disappeared. In this sense, the sectoral and specific industrial policy of the EU has modernized and become far more modest. Sectoral interventions are limited to a few remnants of the past (such as lingering but dwindling subsidies in coal; perhaps the R&D support for Airbus under the ‘Community interest’ clause (art. 87.3.b., EC) and some remaining subsidies in shipbuilding; a handful of more explicit forms of industrial protection via trade policy such as a few quotas in steel and aluminium against some former USSR republics). All of these are national except the quotas, with the Commission exercising supervision, and on the way out, some faster than others. Modern EU-specific industrial policy can have a sectoral slant but only in a noninterventionist way. Since 1991, sectoral Commission papers typically scan all impacts of EU policies and rules (or indeed, free movement), consider what adverse impact can be reversed and what beneficial effects might be amplified, conduct joint studies on competitive positioning and (may) search for R&D elements which the EU could usefully support. This fairly consistent approach is the counterpart of the focus on horizontal policy and the reliance on the overall EU framework. It is undoubtedly a significant improvement on protectionist, ad hoc, politicized or anti-adjustment interventions or plain nationalism. Two and a half decades ago, an entire range of sectors was subject to sectoral European industrial policy of the old variety, in one way or another, via the relaxing of state aid supervision, explicit (often ‘grey’) protectionism, special regulations, anti-dumping and so on. Around 1980, the range came to include cars, aircraft, shipbuilding, coal, steel, textiles and clothing, railway rolling stock, telecoms equipment, consumer electronics and so on. Picking winners took place as well in often ingenious ways. Nowadays, while such bad policy making is avoided, it is not so easy to find out whether good policy has taken its place. Good industrial policy might be verified as correcting or

70

International handbook on industrial policy

improving upon market functioning, in particular dynamic efficiency, and as being leastcost. The present chapter cannot go into the numerous sectoral policy papers the Commission has published since 1991; it would be too encyclopaedic and it would be impossible to do justice to specialized and useful analyses. Many sectors have been subject to such broadly encouraging but largely non-interventionist ‘policies’, based on specialized study groups and high level panels and with wide consultation of the sectors involved. Examples include textiles and clothing, cars, trains (both for new technology and for interoperability purposes), aircraft, new materials like ‘new ceramics’ and superhard fibres, chemicals (culminating in new proposals to transform regulations in such a way that innovation would not be discouraged, as is now the case; this has become the REACH process), biotechnology, shipbuilding, telecoms equipment (especially the attempt to keep the advantages EU companies had acquired with GSM, in 3G equipment), environmental technologies and so on. Such approaches have not been accompanied by trade protection any more, or by large subsidies. When this chapter went to the publishers, the Commission published a paper on what it calls a ‘new, integrated industrial policy’ (see COM (2005) 474 of 5 Oct. 2005 and SEC (2005) 1215, for further details) which attempts to combine systematically its given horizontal perspective with a detailed screening of 27 sections, with respect to initiatives which can be taken to address weaknesses or new challenges at the sectoral level, always coherent with the horizontal framework underlying it. It is nonetheless worthwhile to digress briefly on two aspects because they are critical for the reader to appreciate some remaining specifics of EU industrial policies of today, namely the EU’s preoccupation with ICT, and its changing approach to technology policy. The Union can look back on 25 years of what is now called ICT policies, beginning with ESPRIT in 1981 and currently concentrating on what is called the i2010 programme starting in 2006. Much has been written on the EU efforts but it is hard to escape the painful conclusion that the results of much collaborative research and co-production and (altogether) very considerable sums of EU money (usually 50 per cent of the costs) have not led to an improvement of the competitive position of the European ICT industries at any point in time during this quarter of a century, be it in hardware or software. It is probably fair to summarize by noting that the Union has held on far too long to a protective (and sometimes protectionist) approach, only shifting to a more pro-competitive, demandoriented perspective in the course of the 1990s. The EU’s ICT industry value added is about 80 per cent in services and only 20 per cent in manufacturing. The manufacturing ICT sector finds itself squeezed between the proven innovative capacity of US industry and the cost-effective industrial processes in China and East Asia more generally. The software industry, though bigger, is probably even weaker in relative terms (Dang Nguyen and Genthon, 2005) when assuming a global view. What matters for the latter, far more than for the former, is the widespread presence and fostering of knowledge, of users’ sophistication and of a pro-competitive and innovative climate. In turn, this requires low entry barriers, risk taking and access to venture capital (a private approach to ‘backing winners’), possibly at the margin supported by extra capital venture for SMEs from EU funds but on a market basis. The eEurope programme up to 2005 and the i2010 programme are, in these respects, incomparable to the old approach to ICT. eEurope is promoting a knowledge-based economy, about getting Europe ‘on line’, including e-government and education and eHealth. It also aims to stimulate new services based on ICT (broadband, in particular).55 The question can be asked whether

European industrial policy 71 eEurope can accelerate or prompt developments which, in the final analysis, are marketdriven for the most part, especially because the EU instruments amounted to little more than persuasion projects, studies and modest funding. It is also crucial to understand the complementarity between eEurope, R&D programmes on ICT and regulation where appropriate. Nonetheless, the attention on user’s demands and market processes in the ‘Information Society’ (but not as a blueprint any more) stands in marked contrast to the old approaches to ICT. This can be further improved upon by focusing on those aspects that the EU ought to cover from a subsidiarity perspective, such as interoperability, open standards and security issues in the Internet, because they are by definition cross-border and in the European public interest. They happen to be critical from a user perspective. Indeed, the uncertainty or incompleteness holds back deeper investments in many ways while suppressing new e-practices such as eMoney. From a still broader perspective, that of Lisbon and economic growth, the swifter adoption and wider deployment of ICT hardware and software is crucial to obtaining a higher trend growth in Europe.56 In future, the demands aspects are to be widened to multi-media and contents orientation, among others. The idea is that ICT investment, much lower in Europe than in the USA and in Japan, relative to GDP, can only be stimulated in the long run if the dynamics of the (internal) market provides the opportunities. It remains to be seen whether i201057 is genuinely going to be more effective, but it would seem to represent a policy approach no longer built on grand illusions and no longer prioritized from the existing (EU) supply side. EU technology policy began in earnest with the Single Act of 1985, when it was given a legal base. Framework programmes began a few years earlier but were heavily biased towards energy under Euratom. Nevertheless, even the term ‘technology policies’ is not clear at the EU level because R&D is often called ‘research’ whereas large parts of framework programmes also dealt and deal with applications in potential or close-to-market technologies. Besides, there was the EUREKA programme, largely market-driven and not accompanied by guaranteed subsidies. Two decades of such programmes in EU-wide networks of companies and technology centres have radically altered the mind-set of European firms, from ‘local’ to European, if not global. As the MERIT/CATI database developed by John Hagedoorn of Maastricht University has shown,58 how EU firms have been driven to extensive networking in Europe in this way, while engaging in numerous strategic networks and alliances with (basically, non-competing) partners from other continents for market penetration purposes. With the internal market gradually becoming more integrated and cross-border mergers easier over time, the Europeanization of industry has also changed the technology landscape drastically. It is a moot question whether EU technology programmes were effective in the first decade or so. They tended to be splintered over far too many ‘priorities’ and compromised excellence by replacing technological leadership with equity criteria (such as all member states ‘getting reasonable shares’). They suffered from initial asymmetry of information between the Commission and business about where ‘needs’ could be identified, risking picking winners or protecting losers. Gradually, in scaling up to the sixth and seventh framework programmes with ever-rising sums of EU money (substituting for national efforts, presumably), priorities have been reduced to a few and a much more rational assessment of the state of the art preceded choices, with a shift to a more market-driven approach such as the new ‘technology platforms’. Technology

72

International handbook on industrial policy

platforms bring together for a number of years the entire array of companies and centres having received R&D subsidies for a specific subject matter in order to induce synergies and mutual inspiration, and possibly a larger range of marketable products. Such platforms may be the origin of new standards, new ideas, new combinations and so on. They attract market participants looking for certain applications as well. The platforms do not receive money other than the original R&D subsidies, while the Commission expects progress reports to be done there, with all the expert and commercial exposure available. Where the EU has only slowly improved its technology policy is in the corporate environment. First, the common EU patent was voted down in 2004 on trivial and provincial language excuses of only two countries, a very costly failure; patenting in the EU costs on average five times as much as in the USA. Second, the competitive environment has become tougher in the internal market over time and, for companies beyond a minimum size, this might be expected to spur the search for new technologies and products. Unfortunately, the overall R&D carried out by EU business is much less, relatively, than business R&D in Japan or the USA. Moreover, some relocation of high-tech research (for example in pharmaceuticals) to the USA and even to China (in electronics) has prompted questions about the lack of stimulus for quality research in technology in Europe. Finally, Europe has only moderate R&D in defence, in sharp contrast with the USA. To the extent that defence R&D generates spin-offs sooner or later, or creates a common cost base for civil and military products, the Union may be at a disadvantage. Repeated attempts to combine at least defence research in the EU have not been very effective. The probable sidelining of the draft Constitution is a setback in this respect because it would provide a basis for a special agency in this area, based on voluntary collaboration. All in all, the sectoral and specific EU industrial policies are by no means dead, but have altered radically in nature while having become on the whole non-interventionist. Of course, this chapter cannot hope to assess their factual effectiveness, nor does it pretend to do so. The only major weakness so far not seriously addressed is the failure to optimize the two-level problem, so strongly argued by Eaton et al. (1998). One cannot be surprised about the problem of ‘free riding’ at the member states level in a two-level technology policy, without strict two-level coordination. The EU level has no say on the national policies, their ‘waste’ or excessive duplication horizontally or with the EU, or indeed on their deplorable quality where relevant. However, the national spending is easily more than 90 per cent of total European expenditure. That is why a future European Agency for Research, if ever adopted by the Council, should be given a mandate to deal as well, and on a competitive basis, with a part of national subsidized R&D. The issue here is not whether a member state would thereby lose its power to subsidize, but, rather, whether ‘peer review’ can ensure a high minimum level of quality and originality. This is similar to technology platforms where a kind of perhaps informal peer review, but with applications in mind, disciplines and stimulates high-tech output at high minimum levels. 6 Conclusions European industrial policy cannot be understood by focusing solely on technology, ICT and a few other specific stimulating policies. It is indispensable to assume a broad perspective with a good appreciation of the EU framework determining largely both the constraints at member states level and the considerable limits of action at the EU level,

European industrial policy 73 besides an understanding of what initiatives are feasible, and how, with respect to ‘horizontal’ and ‘specific’ industrial policies. The upshot has been that industrial dynamics in the Union have become ever more firmly determined by market pressures in a very big Economic Union (with no fewer than 455 million consumers, up to 485 million by 2007) and by exposure to globalization, given very low industrial trade protection and the absence of barriers to FDI (as well as national treatment). Inevitably, EU industrial policy has therefore drifted into endeavours to strengthen fundamental determinants of competitiveness, to some degree corresponding to dynamic comparative advantage. Hence the ‘horizontal’ emphasis on research strategies, skills and human capital and a general promotion of innovation throughout industry. Moreover, this accords well with the deepening of the internal market, still going on, and the consistency of EU competition policy, in reducing distorting state aid and severely limiting the scope for discriminatory or distortive promotion of ‘national champions’. In addition, endeavours have been made to facilitate SMEs and ‘entrepreneurship’ by cutting red tape and making access to venture capital easier. Over the last few years, political leaders, calling for a new industrial policy against deindustrialization, acknowledged this horizontal perspective when they merely insisted on reducing the burden on industry arising from new EU regulation, while refraining from a systemic policy reversal. Sectoral and specific industrial policy, so familiar from the 1950s, 1960s and 1970s has almost entirely disappeared at the member states level. It is no longer tolerated at the EU level, given that treaty amendments and accepted supervision policy have become routine. At the EU level, such interventionist policy (other than once tolerating it for member states) could only be supported via trade policy. Also the latter has lost a good deal of its tools, with the possible exception of anti-dumping action (under WTO constraints and tightened EU law). Nowadays, the EU institutions are very active in stimulating a better sectoral environment in many sectors but this is no longer done by means of protection, distortive subsidies (funds go to adjustment, retraining or research) or permissive competition policy. Policy reversals are not totally excluded (witness the temporary constraints in Chinese textile exports) if pressures are extreme but there is no doubt that competitiveness and hence adjustment retain overriding importance. The EU has greatly increased technology funding (quite often as a corollary of research programmes) in areas such as new materials, space, hydrogen, health and environmental technologies. Also here, the EU level seeks to avoid a ‘picking winners’ approach by letting technology forums develop where many ideas and applications can blossom. This chapter has not discussed the (in)effectiveness of EU industrial policies. Even if partial, this would be a difficult exercise given the multidimensional nature of policies and the importance of constraints and/or the absence of powers for the EU level. For research and technology, however, the most visible element presumably of policy today, there is undoubtedly a serious lack of coordination in a two-level game. It is essential that the research climate in Europe become more demanding, challenging and promising (hence stopping the brain drain), that quality be stimulated by a tough peer review approach, and that excessive duplication at the two levels be avoided. However, even these measures may not be sufficient. After all, European industry (and services even less) invests much less in R&D than Japanese and US business and, unless their output and subsequent innovation to markets are systematically far better than their rivals in Japan and the USA – an assumption that would go against all evidence – this underperformance is bound to erode

74

International handbook on industrial policy

competitiveness amongst developed economies in the longer run. Apparently, it is not so much the size of the internal market but its proper (and dynamic) functioning which might lead EU companies to hold back on R&D. Notes 1. 2.

3. 4. 5. 6. 7.

8.

9. 10. 11. 12.

13.

14.

The treaty of Paris of 1951 laid the basis for the European Coal and Steel Community, the archetype of former interventionist industrial policy. Several new member states may join Euroland in 2006 or 2007, after the compulsory two years in the ERMII without too much turbulence. These might include Estonia, Slovenia and Lithuania. Other new member states will follow in due course, but this may be stretched out over a longer period. Whether the UK, Sweden and Denmark (already in ERM-II but a referendum in 2003 came out with a NO vote) will join any time soon is unclear. Unlike the new member states, they are not under the obligation to join. Before joining the euro, all countries must have an independent Central Bank; all EU countries are bound by price stability, fiscal prudency and completely unrestricted capital movements. For a detailed exposition and economic analysis of the EU internal market, see Pelkmans (2006, chs 3–10). See also COM(2003) 238 of 7 May 2003 on the Internal Market Strategy and its annual implementation reports. In art. 4, ECSC, the huge national subsidies making a mockery of internal free trade and specialization were typically defended on the basis of an escape clause in art. 95, ECSC. Thus rescue operations in some sectors were ‘politically’ assumed to be more acceptable than in other ones, such as shipbuilding, airlines (with their repetitive ‘last aid and never again’ subsidies), steel in East Germany, and so on. Formerly art. 94, EC, the article can be interpreted as giving the Council the option of imposing exceptions and more political criteria on the Commission’s regime which, otherwise, should be least distortive for the internal market. In 1990, under the Italian presidency, minister Battaglia led an attempt by a group of member states to do just that, but it failed. It should not be forgotten that state aid differs from antitrust in that the parties called to alter conduct are not companies but member states, and this may prompt political interventions at the top, sometimes in turn prompted by demonstrations in national capitals. Generally accepted rules by the Council, for example, in block exemptions turn the Council into a ‘co-owner’ of the regime and make it easier for the Commission to act. But it begs the question of what ‘military goods’ are because many of them are or can be so-called ‘dual purpose’ goods, and the latter have come under increasingly strict EU supervision. The European Constitution, not yet ratified, has largely adopted this clause as art. III – 279, substituting Union for Community. But ‘not . . . any measure which could lead to distortion of competition . . . ’ as the last sentence of the article says. In actual practice, the transaction and information costs of this heavy Council appear to be rather high, however. Especially, the numerous legal and technical subjects of the internal market do not accord well with the interest and expertise of specialists in industrial analysis and even less with those dealing with the research community in the Union. Also this Council, already cumbersome with no fewer than 25 ministers, often sees different ministers or their deputies appear, given the diverse national portfolios for them. In other words, there seems to be a trade-off between the opportunities to discuss in strategic terms at ministerial level how to foster competitiveness and the far more technical agenda that the Council also has to deal with. Consider the dire predictions of sceptical economists that the Italian car quotas against Japan (gradually relaxed and removed in the 1990s) would not actually disappear; similarly, the textile and clothing quota under the Multi-Fibre Arrangement (ended late 2004) did go; even huge import jumps in textiles and clothing from China in 2005 – though prompting a kind of voluntary export restraint by China – led only to selected and temporary import ‘quotas’, with some growth and at much higher levels than earlier imports while no state aids will be permitted; remember the way the Commission dealt with restrictive practices and state aids in steel in 1994; the much stricter state aids regime for airlines was first fiercely resisted but did prevail at last; the shipbuilding sector, with decades of struggle over hidden subsidies, has adjusted dramatically and state aids are now a trickle compared to the 1980s; the very narrow scope of action for member states when they announce rescue operations, as with Alsthom in France in 2004 or with EKO Stahl in (East) Germany in the late 1990s; the tightly controlled clearance under EU competition policy for crisis cartels, and so on. The Maastricht treaty introduced novel and firm language about the ‘economic order’ of the Union in this respect. Both arts 4, EC and 5, EC (new numbering) impose the condition that member states and the EU act, in all policies and measures, ‘in accordance with the principle of an open market economy with free competition’.

European industrial policy 75 15.

16.

17. 18. 19. 20.

21. 22. 23.

24. 25.

26. 27. 28. 29.

30.

Although the debate is not entirely comparable with that on the (multiple) definitions of industrial policy, it does echo some of its traits and is equally conceptual and perpetual. For discussion, see the attack by Krugman (1994), and partial reconciliations by Jacquemin and Pench (1997) and Galli and Pelkmans (2000); see also Lawton (1999). The Commission embraced the OECD definition of ‘competitiveness’ in the mid-1990s as ‘the ability of companies, industries and regions, nations or supranational regions to generate, while being and remaining exposed to international competition, relatively high factor income and factor employment levels on a sustainable basis’. The ultimate object here is standards of living (for example of the EU) and the suggestion is that nations ‘compete’ as if they are firms. The Krugman critique applies here, of course. The World Economic Forum, influenced by Porter-type of notions that countries somehow do compete (because their natural endowments matter less and less) by means of alternative economic institutions and strategies, clearly appeals more to policy makers, even if the analytical and statistical basis of the WEF reports leaves much to be desired. Among the many questions which can be raised when using such notions of competitiveness, the most important one is what this would (or should?) imply for industry. A few examples: the December 1993 Commission White Paper on Growth, Employment and Competitiveness, Bulletin EC, supplement 93/6, emphasizes SMEs, infrastructure and new technologies; the second Bangemann Memorandum, COM(1994) 319, focused on industrial cooperation and on intangibles; a benchmarking exercise in COM(1996) 463 stresses quality management and quality promotion; in COM(1999) 465 of 5 October 1999, weaknesses in structural change and adjustment are viewed as the ‘main challenge for policy makers’. In the Commission, DG Industry was renamed DG Enterprise as well. See, for example, COM(1999) 569 of 9 November 1999, Report on concerted action with the member states in the field of enterprise policy. DG Magnus Lemmel, From industrial policy to enterprise policy, 25 Nov. 1999 in Brussels at the conference on ‘A Sharper Cutting Edge’. A Green Paper on Entrepreneurship, COM(2003) 27 of 21 Jan. 2003; COM(2002) 610 of 7 Nov. 2002 on ‘Better environment for enterprises’; the BEST procedure focused on facilitating SMEs; see, for example, COM Staff Working Paper SEC (2002) 1212; and various benchmarking exercises in this area, as in SEC(2002) 1213 and SEC(2002) 1214. See COM(2002) 714 of 11 Dec. 2002 on industrial policy in an enlarged Europe; COM(2004) 274 of 20 April 2004, ‘Fostering structural change, an industrial policy for an enlarged Europe’, a more elaborate and updated version of the former. See COM(2004) 274, pp.6–16 The Commission defines business-related services as business services, distributive trades, network services and financial services. Business services include, on the one hand, professional services (IT and management consulting, R&D services, engineering consultancy, advertising, accountancy/auditing, legal and so on) and, on the other hand, operational services (such as industrial cleaning, security, secretarial, specific call-centres and so on). See DG Enterprise, working paper, ‘Business-related services – a key driver of European competitiveness’, an enhanced economic analysis, Dec. 2004; and COM(2003) 747 of 4 Dec. 2003. It should be noted that EU services statistics are both less disaggragated and less harmonized than those for goods, or, labour, for that matter. This can be a problem for measurement and empirical economic analysis. See also Woelfl (2003). All data from DG Enterprise, 2004, as in the previous footnote. For some background and data supporting these contentions, see for example SEC(2004) 866 of 23 June 2004, ‘Evaluation of the performance of network industries and so on’, European Commission; regular reports on gas/electricity (latest COM(2004) 863 of 5 Jan. 2005, and on telecoms (latest COM(2004) 759 of 2 Dec. 2004); restrictions are analysed at length in COM(2002) 441 of 30 July 2002, ‘The state of the internal market for services’; COM(2004) 83 of 9 Feb. 2004, report on competition in professional services; and so on. For example, in postal, see WIK (2004), ‘Main developments in the European postal market’ (http:// europa.eu.int/post/doc/studies/2004-wik-final_en). For example, in financial markets, this is attempted to be measured in the Financial Integration Monitor; for 2004, see SEC(2004) 559, June 2004, Commission Staff Working document. See Bart van Ark (2004), CPB colloquium; for a less alarming analysis of the ICT ‘gap’, see Denis, Morrow, Roeger and Veugelers (2005, pp. 19–36), even if the authors agree on the crucial role of ICT services. In a survey chapter like this, it is hard to substantiate this statement in detail. After the relative optimism of the first years, see, for example, the annual reports in COM(1998) 356 of 3 June 1998 and COM(1999) 410 of 15 September 1999, and the hope that public–private partnerships could help fill the huge funding gaps (see High Level Group, 1997), progress became disappointing and highly selective. See also the implied criticism in Boxes 1 and 2 of the European Initiative for Growth, COM(2003) 690 of 21 Nov. 2003, an attempt to revitalize the TEN policy. See COM(2003) 690 of 21 Nov. 2003, ‘A European initiative for growth’, pp.10ff.

76

International handbook on industrial policy

31. 32.

Greater detail can be found in the literature, for example Pelkmans (2006) or other surveys. Which implies quite an intrusive mix of competition policy and pro-competitive regulation; see (Pelkmans, 2001). In principle, art. 296, EC goes further than mere state ownership because the latter can prevent any (including hostile) takeover but cannot change the competitive environment of the company. Its application might be possible for military goods and for goods and services vital for national security, but this would only work if the goods and services were of a purely military nature. As soon as so-called ‘dual-purpose goods’ are produced (and that is likely to be the case for most of these sectors), much of the discretion disappears since free movement and undistorted competition apply. Coherence especially with structural policy and with criteria for EU (as opposed to national) subsidies. COM(2003) 27 of 15 Jan. 2003, Green paper, ‘Entrepreneurship in Europe’; COM(2004) 70 of 11 Feb. 2004, ‘Action plan: the European agenda for entrepreneurship’. See also COM(2003) 704 of 21 Nov. 2003, ‘Some key issues in Europe’s competitiveness – towards an integrated approach’. See COM(2002) 276 of 5 June 2002 on RIA policy and methodology, and COM(2005) 97 of 16 March 2005, ‘Better regulation for growth and jobs in the EU’. The latest version is SEC(2005) 791 of 15 June 2005, ‘Impact Assessment Guidelines’. See, for example, COM(2002) 499 of 11 Sept. 2002, ‘More research for Europe: towards 3% of GDP’. See COM(2003) 226 of 4 June 2003, ‘Investing in research: an action plan for Europe’. Sapir et al. (2003); Gros and Micossi (2005). See, for example, High Level Expert group, 2005, ‘Frontier Research, the European Challenge’, Brussels, European Commission DG Research; EIROforum, 2005, ‘Towards a Europe of Knowledge and Innovation’ (http://www.eiroforum.org/efarchives/efsciencepolicy05.pdf). For at least a decade, the EU has been actively interested in stimulating innovation. It is strongly emphasized in the Lisbon strategy. The text will only briefly point to recent initiatives and attempts to measure improvements as concretely as possible. For instance, see SEC(2003) 1255 of 10 Nov. 2003, 2003 European Innovation Scoreboard. COM(2003) 112 of 11 March 2003, ‘Innovation policy: updating the Union’s approach in the context of the Lisbon strategy’. See DG Enterprise, 2004, Innovation policy in Europe 2004, TrendChart, via www.cordis.lu/trendchart comprising an elaborate comparison of the national innovation systems and the lessons one might learn from it. COM(2005) 121 of 6 April 2005, ‘Proposal for a decision [. . .] establishing a Competitiveness and Innovation Framework programme’ (2007–2013); also SEC(2005) 433, plus an Annexed Commission Staff Working Document providing an Impact Assessment. An interesting illustration is given by a box in the Sapir report, based on unpublished work from Philippe Aghion, showing the growing importance of human capital for innovation at the productivity frontier. For both total factor productivity and labour productivity, the regression of the interaction between distance to the technological frontier and skilled labour is positive and significant. See Sapir et al. (2003), p. 33. Kok et al. (2003), Jobs, jobs, jobs, creating more employment in Europe, Brussels, November. European Commission (2004), ‘More and better jobs for all’, Brussels, January; and the Council Decision 2003/578 of 22 July 2003 on ‘Guidelines for the employment policies of the Member States’, item 4, which underlies it. See the Action Plan on Skills and Mobility, COM(2002) 72; the reform of the EURES system on job vacancies (see Commission decision 2003/8 of 23 December 2002, in Official Journal EC, L 5 / 16 of 10 January 2003; the many calls for student mobility and the various programmes the Commission has undertaken since 1987 to promote cross-border student exchange such as ‘Erasmus’, etc.; its attempts to stimulate researchers’ mobility over frontiers; see COM(2001) 331 of 26 June 2001 on ‘Mobility within the European research area’; its cautious support of the so-called ‘Bologna process’, which is intergovernmental and involves non-EU countries as well but can be helpful in the longer run to facilitate transparency about national academic degrees, and should – ceteris paribus – lower considerably the transaction and information costs for cross-border mobility of students and graduates. In addition, it has the advantage of making it easier for Asian and American students to come to Europe and for multinationals to comprehend European degrees. The details of the Bologna process do matter, however, and misgivings abound in Europe about the risk of lowering quality or diluting certain master’s degrees. For mutual recognition to work in actual practice, however, some measure of comparability and simplicity over a continent of more than 40 countries is badly needed). See the Interim report in COM(2003) 685 of 11 Nov. 2003. See COM(2005) 58 of 5 Feb. 2005, ‘The role of the universities in the Europe of knowledge’; COM(2005) 152 of 20 April 2005, ‘Mobilising brain power of Europe, enabling universities to make their full contributions to the Lisbon strategy’; European University Association, 2005, Glasgow Declaration, ‘Strong universities for a strong Europe’, Brussels, April.

33.

34. 35. 36. 37. 38. 39. 40. 41. 42. 43. 44. 45. 46. 47. 48.

49. 50. 51.

52. 53.

European industrial policy 77 54. 55. 56. 57. 58.

The European University Institute was the result and started formally in 1972 (de facto in 1975) in Florence. See COM(2002) 263 comprising the eEuropeAction Plan, and COM(2004) 108 of 18 Feb. 2004, ‘eEurope 2005 Mid-term review’. See in particular Denis, McMorrow and Roeger (2004), and Denis, McMorrow, Roeger and Veugelers (2005), as well as van Ark (2004); see also chapter 2 of the Commission’s European Competitiveness Report (2003) on the intra-firm organizational implications of ICT and the productivity gains. See COM(2005) 229 of 1 June 2005, ‘i2010 – A European information society for growth and employment’; Reding (2005). See, for example, Hagedoorn, Link and Vonortas (2000) for a survey.

References Ark, B. van (2004), ‘What do we know about services productivity in Europe?’, CPB workshop on Productivity in services, 10 June (www.cpb.nl), The Hague. Aschauer, D. (1989), ‘Is public expenditure productive?’, Journal of Monetary Economics, 23, 177–200. Bangemann Memorandum (1990), ‘Industrial policy in an open and competitive environment’, Commission Working Paper, 14, September. Besley, T. and P. Seabright (1999), ‘The effects and policy implications of state aids to industry: an economic analysis’, Economic Policy, 28. Christopherson, H. et al. (1994), ‘TransEuropean Networks’, Final (Christopherson Task Force) report, Brussels. Copenhagen Economics (2005), ‘Economic assessment of the barriers to the internal market for services’, Brussels/Copenhagen, January (DG Enterprise, European Commission). Dang Nguyen, G. and C. Genthon (2005), ‘Has the European ICT sector a chance to be competitive?’, unpublished paper, ENST Bretagne and College of Europe, Bruges. Denis, C., K. McMorrow and W. Roeger (2004), ‘An analysis of EU and US productivity developments’, Economic Papers, 208(July), DG EcFin, European Commission, Brussels. Denis, C., K. McMorrow, W. Roeger and R. Veugelers (2005), ‘The Lisbon strategy and the EU’s structural productivity problem’, Economic Papers, 221(February), DG EcFin, European Commission, Brussels. Devereux, M., R. Griffith and A. Klemm (2002), ‘Corporate income tax reforms and international tax competition’, Economic Policy (October). Eaton J., E. Gutierrez and S. Kortum (1998), ‘European technology policy’, Economic Policy, 27(October). ERT (1984), Missing links, Report of the European Round Table of Industrialists, Brussels/Paris. European Commission (1993), ‘White paper on growth, competitiveness and employment’, EC Bulletin, Supplement 93/6 (December), Brussels. Fuente, A. de la and X. Vives (1995), ‘Infrastructure and education as instruments of regional policy – evidence from Spain’, Economic Policy, 13–51. Galli, G. and J. Pelkmans (2000), ‘Introduction: aims, structure and overview’, in G. Galli and J. Pelkmans (eds), Regulatory Reform and Competitiveness in Europe, Cheltenham, UK and Northampton, MA, USA: Edward Elgar. Gros, D. and S. Micossi (2005), ‘A better budget for the European Union. Move value for money, more money for value’, CEPS Policy Brief, n.66, Brussels. Hagedoorn, J., A. Link and N. Vonortas (2000), ‘Research partnerships’, Research Policy, 29, 567–86. High Level Group (1997), ‘On public–private partnership financing of TransEuropean transport network projects’, Final report, Brussels, May. Jacquemin, A. and L. Pench (1997), ‘What competitiveness for Europe? An introduction’, in A. Jacquemin and L. Pench (eds), Europe competing in the global economy, Cheltenham, UK and Northampton, MA, USA: Edward Elgar. Kamp, C. (2005), ‘Is there a lack of public capital in the EU?’, EIB Papers, 10(1). Kok,W. et al. (2003), Jobs, jobs, jobs, creating more employment in Europe, Report of the Employment Task Force, November, Brussels. Kok,W. et al. (2004), Facing the challenge, the Lisbon strategy for growth and employment, Report from the High Level Group, November, Brussels. Kox, H., A. Lejour and R. Montizaan (2004), ‘The free movement of services within the EU’, CPB document, 69, October, The Hague. Krugman, P. (1994), ‘Competitiveness: a dangerous obsession’, Foreign Affairs, 73, 28–44. Lawton, T. (ed.) (1999), European Industrial Policy and Competitiveness, London: Macmillan. Martin, P. (1999), ‘Are European regional polices delivering?’, EIB Papers, 4(2). Molitor, B. et al. (1995), Report of the group of independent experts on legislative and administrative simplification, Brussels: European Commission.

78

International handbook on industrial policy

Pelkmans, J. (2001), ‘Making EU network markets competitive’, Oxford Review of Economic Policy, 17(3). Pelkmans, J. (2002), ‘Mutual recognition in goods and services, an economic perspective’, BEEP Briefings, 2, College of Europe, Bruges (see www.coleurop.be/ECO). Pelkmans, J. (2005a), ‘Testing for subsidiarity’, in T. Bruha and C. Nowak (eds), Die Europaeische Union: Innere verfasstheit und globale Handlungsfaehigkeit (The EU, internal constitutionalism and global effectiveness), Baden-Baden: Nomos. Pelkmans, J. (2005b), ‘REACH, getting the chemistry right in Europe?’, in D. Hamilton and J. Quinlan (eds), Deep Integration, How Transatlantic Markets are Leading Globalisation, Brussels (CEPS) and Washington, DC: CTR, Johns Hopkins University, SAIS. Pelkmans, J. (2006), European Integration, methods and economic analysis, 3rd rev. edn, Harlow: Pearson Education. Pelkmans, J. and J.P. Casey (2004), ‘Can Europe deliver growth?’, BEEP Briefing, 6, College of Europe, Bruges (see www.coleurop.be/ECO). Pelkmans, J., S. Labory and G. Majone (2000), ‘Better EU regulatory quality: assessing current initiatives and new proposals’, in G. Galli and J. Pelkmans (eds), Regulatory Reform and Competitiveness in Europe, Vol. 1, Cheltenham, UK and Northampton, MA, USA: Edward Elgar. Pietrantonio, L. di and J. Pelkmans (2004), ‘The economics of EU railway reform’, Journal of Network Industries, 5(3–4). Radaelli, C. (2003), ‘Impact assessment in the EU: innovations, quality and good regulatory governance’, background report for the Conference on EU Impact assessment, 3 December 2003, European Commission, Brussels. Reding, V. (2005), ‘i2010, the European Commission’s new programme to boost competitiveness in the ICT sector’, Speech 05/61, Prague, 31 January. Renda, A. (2005), ‘Impact assessment in the EU’, draft paper, CEPS, August. Romp, W. and J. de Haan (2005), ‘Public capital and economic growth: a critical survey’, EIB Papers, 10(1). Sapir, A., P. Aghion, G. Bertola, M. Hellwig, J. Pisany-Ferri, D. Rosati, J. Vinals and H. Wallace (2003), ‘An agenda for a growing Europe: making the EU system deliver’, Report of the independent High level Group, July, Brussels (see http://europa.eu.int/comm/dgs/policy/advisers/experts_groups/ps 2/docs/agenda_en.pdf). Valila, T., T. Kozluk and A. Mehrotra (2005), ‘Roads on a downhill? Trends in EU infrastructure investment’, EIB Papers, 10(1). Woelfl, A. (2003), ‘Productivity growth in service industries’, STI Working paper, 2003–07, OECD, Paris.

4

Industrial policy in developing countries: what can we learn from East Asia? Sanjaya Lall

1 Introduction There is growing concern with industrial competitiveness in all economies.1 The intense pressures created by rapid and inexorable technical change, falling transport and communication costs (‘shrinking economic distance’), widespread liberalization and the spread of global production networks (with large shifts in the location of production: see Dicken, 1999) are posing difficult challenges everywhere. While it is the advanced industrial countries that are the most active in mounting competitiveness strategies, the real concern – and the intensity of the underlying threat – is greatest in the developing world. Many poor countries, particularly the least developed ones, are failing to compete internationally and, as a result, are suffering low or negative growth rates of manufacturing as they liberalize (Lall, 2001; UNIDO, 2002). However, this is not universal. Some developing countries are succeeding, often quite dramatically: it is therefore not the case that globalization, technical change and liberalization per se are harmful. The issue is how countries cope with these forces. The fault, if that is the right term, lies more in the countries than in the external environment, but the external environment also exerts immense pressures that constrain the ability of countries to become more competitive. This is the topic of this chapter. What determines whether and how countries become internationally competitive? The view that dominates mainstream policy thinking is that they become competitive by resorting to free markets. The best strategy for developing countries is to remove government interventions in markets, provide a stable macroeconomic setting and clear rules of the game, and invest in infrastructure and generic human capital. This will optimize static comparative advantage (based on existing factor endowments) and also dynamic competitiveness (resource allocation over time). This approach, the ‘neoliberal’ or ‘Washington Consensus’ approach, has many merits. It has helped remove some of the more inefficient, rent-creating interventions that marked early industrialization in many countries. It appeals to many governments that tried, often unsuccessfully and sometimes at huge cost, to build competitive industry behind high protective barriers, with restrictions on foreign investment and a lead role for the public sector. The approach is strongly backed by the industrialized countries and the leading development institutions; it is enshrined in the new ‘rules of the game’ implemented by the WTO. The new rules are based on a strong premise: markets are ‘efficient’ (in the economic sense), the institutions needed to make markets work exist or will arise soon, and no deviations from optimality (market failures) can be remedied effectively by policy. The final element of the premise, essentially that governments are necessarily less efficient than markets and even with market failures a neoliberal policy is necessarily better, is a political rather than an economic judgment. The arguments are based on a particular view 79

80

International handbook on industrial policy

of the technological basis of industrial competitiveness (Lall, 1996, 2001) and backed by reference to the experience of successful industrializing economies. An alternative view puts less faith in free markets as the driver of dynamic competitiveness and more in the ability of governments to intervene efficiently. It questions the theoretical and empirical arguments that free market forces alone account for industrial success. Accepting that past industrialization strategies were often wrongly designed and poorly implemented, it argues that greater reliance on market forces actually requires a strong proactive role for the government. Markets are often deficient and the institutions needed to make them work efficiently are often weak or absent. With the challenges noted above, healthy and sustained industrial development cannot proceed without interventions to improve markets and build institutions. This chapter illustrates these arguments with examples from East Asia. Section 2 examines indicators of the competitiveness of developing countries from the early 1980s to 2000, and shows that East Asian tigers have performed better than other developing countries, although they have not directly implemented the policies of the Washington Consensus. Section 3 shows the industrial policy implemented by these mature East Asian Tigers (Hong Kong, Singapore, Korea and Taiwan). Section 4 draws some conclusions for policy. 2 Recent industrial performance This section describes industrial performance in the developing world from the early 1980s to 2000,2 using two measures of performance: market shares for manufacturing value added (MVA) and for manufactured exports. Let us start with MVA performance. The developing world performed well in these two decades: its share of global MVA rose by 10 percentage points, from 14 to 24 per cent, and its annual rate of growth (in constant dollars) was 5.4 per cent, over twice the 2.3 per cent recorded by the industrialized world.3 However, industrial success was highly concentrated (Figure 4.1). East Asia (shown with or without China) dominated MVA values and growth, raising its total world market share from around 4 per cent to nearly 14 per cent – exactly the 20 percentage point rise for the developing world as a whole. It came from well behind LAC in 1980 to account for over two and a half times the latter’s share by 2000.4 Figure 4.2 shows per capita MVA in developing regions and again shows how EA performed far better than other regions, and 8.0 1980 1990 2000

7.0 6.0 5.0 4.0 3.0 2.0 1.0 0.0 EA 2

Figure 4.1

China

S. Asia

LAC

MENA

World shares of MVA by developing regions (%)

SSA 2

S. Africa

Industrial policy in developing countries

81

800 1980 1990 2000

700 600 500 400 300 200 100 0 LAC

Figure 4.2

E. Asia

EA 2

South Asia

MENA

SSA

Per capita MVA in developing regions (constant 1990 US$)

how LAC and Sub-Saharan Africa ended the period with lower MVA per capita than they started with. LAC, the region that liberalized the most (and the most rapidly), was the worst performer. East Asia, the best performer, was the most export-oriented and its export growth matched its production growth (the precise causation between growth and exports is still debated, but in my view export orientation clearly played a very important role in stimulating the development of competitive capabilities). However, EA was not ‘liberal’ in the Washington Consensus sense: most economies there used infant industry protection, export subsidies and targets, credit allocation, local content rules and so on to build their base of industrial capabilities (Amsden, 1989; Stiglitz; 1996; Wade, 1990; Westphal, 2002; World Bank, 1993). There were differences within this general approach. The leading economies such as Singapore, Korea and Taiwan mounted very strong interventionist strategies. They invested massively in human capital (particularly technical skills), fostered R&D and built strong support institutions (Lall, 1996, 2001). They tapped FDI in different ways: Singapore by plugging into global production systems and the other two by drawing on its technologies via arm’s-length means like licensing, copying and original equipment manufacture (OEM) contracts. Singapore industrialized under free trade conditions, aiming directly at world markets, while Korea and Taiwan combined infant industry protection with a strong export orientation, seeking to build national capabilities rather than plug into global value chains under the aegis of MNCs. The second wave of Tiger economies such as Malaysia, Thailand, Indonesia and Philippines relied more heavily on FDI in export-processing enclaves and less on building indigenous capabilities; their export success was thus largely driven by global value chains, particularly in electronics.5 China has a blend of different strategies, some similar to its neighbours and others, like public enterprise restructuring, uniquely its own. The fall in LAC shares is offset by rises in South Asia (largely India) and MENA (partly Turkish growth and partly oil refining in some other countries). The leading industrial economies in East Asia (China, the Republic of Korea and Taiwan Province) have liberalized cautiously, retaining a significant role for the government in industrial development (Rodrik, 2001). As Stiglitz says in a ‘special contribution’ to the 2003 Human Development Report,

82

International handbook on industrial policy China and other East Asian economies have not followed the Washington consensus. They were slow to remove tariff barriers, and China still has not fully liberalised its capital account. Though the countries of East Asia ‘globalized’, they used industrial and trade policies to promote exports and global technology transfers, against the advice of the international economic institutions. (UNDP, 2003, p.80)

It is worth exploring the differences between LAC and East Asia, the leading industrial powers in the developing world, slightly more. The two regions used very different approaches to industrialization, initially in building up the sector and, then, in opening it up to world markets. In the first phase, LAC, in common with most other developing regions, relied heavily on protected import substitution, sheltering enterprises from international competition but failing to offset this with incentives or pressures to export. It did little to attract exportoriented FDI (in EPZs – Export Processing Zones) and so missed the surge in global production systems in electronics. It did not deepen local technological activity (by encouraging R&D) or develop the new skills needed for emerging technologies. In concert with widespread macroeconomic (and in some cases political) turbulence, this meant that LAC failed to develop a broad base of industrial capabilities that would drive competitiveness as it liberalized. As a comparatively high wage region, LAC needed competitive advantages in complex activities to offset labour cost disadvantage vis-à-vis Asia. Despite its tradition of entrepreneurship and good initial base of skills, its industrial strategy failed to foster the necessary capabilities. There were exceptions, such as the automotive industry in the larger economies and resource-based activities more generally. But many such activities were not growing rapidly in world trade and, as shown below, LAC failed to increase its export market shares rapidly, the outstanding exception being Mexico, but owing more to NAFTA privileges than to strategy. In the second (liberalization) phase, policy reform in LAC was rapid and sweeping, with no strategy to foster competitive capabilities and concentrate on promising activities. Again, there were exceptions, including the motor industry (restructured with the help of complementation programmes, banned under new WTO rules), agro-based exports in Chile or national export ‘champions’ like Embraer in Brazil, but the general lack of strategy on industrial competitiveness meant that the region failed to catalyse export dynamism. Its main growth was in resource-based sectors where it was largely exploiting static comparative advantages. Some other developing regions that also used import substitution strategies liberalized more slowly and carefully – India is a good example – and did better in terms of MVA growth (but almost as poorly in terms of export competitiveness). The resulting differences in competitive performance are worth noting, and the next two figures deal with MVA in these two regions, separating China in East Asia and Mexico in Latin America. Both these economies are outliers in their regions, the former because of its size, industrial growth and international competitiveness, the latter because of its export performance driven by its location and privileged access to the US market under NAFTA. While both have done very well in manufactured exports (see below), the link between export and MVA growth has been far stronger in China than in Mexico. This may be due to several reasons: Chinese industry is less exposed to import competition than is Mexico’s (China has just joined the WTO), it is more competitive and it has greater local content in its export activity; a mixture of the three is probably the correct explanation.

Industrial policy in developing countries

83

Figure 4.3 shows MVA market shares within the developing world (not, as earlier, in the world as a whole) for EA without China, China, LAC without Mexico, and Mexico. Figure 4.4 shows changes in these market shares over 1980–90 and 1990–2000. In 1980, LAC as a whole accounted for 47 per cent of developing world MVA and East Asia as a whole for 29 per cent; two decades later, the shares were 22 per cent and 58 per cent, respectively. The main surge in MVA growth in EA 2 (excluding China) was in the 1980s, with a significant slowing down in the 1990s (reflecting the impact of the financial crisis and the general global recession). In China the trends are reversed, with the more rapid growth occurring in the 1990s, taking its share of developing world MVA to a higher level than the rest of East Asia put together. In LAC, the region excluding Mexico loses its MVA shares more rapidly than does Mexico, with the 1980s (the ‘lost decade’ after the debt crisis) being much worse than the 1990s. It is the 1990s that are more interesting: despite relative stability and a reasonable period after trade liberalization, the region continues to perform poorly. LAC2, for instance, sees MVA growth in this decade of 1.9 per cent p.a., much lower than developing countries as a whole (6.4 per cent), East Asia

40 35

1980 1990 2000

30 25 20 15 10 5 0 EA exc. China

Figure 4.3

China

LAC exc. Mexico

Mexico

East Asia and LAC, shares of developing world MVA (%)

10 1980–90

1990–2000

5 0 EA 2

China

LAC 2

Mexico

5 10 15

Figure 4.4

East Asia and LAC, changes in shares of developing world MVA (%)

84

International handbook on industrial policy

(9.5 per cent) or even South Asia and MENA (5.9 per cent and 4.3 per cent). Mexico has more robust growth, at 4.4 per cent, up from the 2.1 per cent in the 1980s, but this still fails to match EA 2 (6.7 per cent) and China (13.1 per cent). We now consider export performance. Figure 4.5 shows world market shares for manufactured exports by developing regions for 1981–2000 and the value of such exports in 2000, separating China from East Asia 2 and Mexico from LAC 2.6 East Asia as a whole accounted for 18 per cent of world manufactured exports in 2000, up from 7 per cent in 1981; within the region, EA2 raised its share from 6 per cent to 11 per cent and China from 1 per cent to 7 per cent. Note that China accounts for a much higher share of regional MVA than exports: its industry is, not surprisingly in view of the size of the economy, far less export-oriented than its neighbours’. LAC as a whole lost world market share between 1981 and 1990, falling from 3.2 per cent to 2.4 per cent, then rose to 5.2 per cent by 2000 (Figure 4.6). The initial fall was due only to LAC 2 (from 2.7 per cent to 1.9 per cent), with Mexico maintaining its 0.5 per cent world market share. In the 1990–2000 period, LAC 2 raised its share slightly while, with the advent of NAFTA, Mexico shot ahead to 3.0 per cent, a sixfold increase. Figure 4.6 shows these changes. It also separates South Africa from the rest of SSA, showing that in the 1990s the former raised its share while the rest of the region stagnated. What can we conclude from these figures? First, as noted, MVA performance is correlated with manufactured export performance. However, the fit is far from perfect. EA 2 and Mexico, for instance, fare far better in export than in MVA shares in the 1990s, while the opposite is true of South Asia and MENA. Second, neither is strongly related to liberalization (in the Washington Consensus sense). Third, industrial success is highly concentrated in the developing world and there is little sign that liberalization and globalization are reversing this divide. 25% East Asia excl. China ($878 bn)

WMS in 2000

20%

15%

10%

China ($318 bn)

5% Mexico ($144 bn) 0%

MENA ($79 bn)

0% S. Asia 1% SSA 2% ($30 bn) ($54 bn)

LAC excl. Mexico ($106 bn)

3%

4%

5%

6%

7%

8%

WMS in 1981

Figure 4.5 World market shares for manufactured products in 1981 and 2000, and values of manufactured exports in 2000 ($ bn)

Industrial policy in developing countries

85

8% 7% 6% 5% 4%

1981–90

1990–2000

SSA excl. S. Africa

S. Africa

3% 2% 1% 0% 1%

Figure 4.6

East Asia excl. China

China

South Asia

MENA

LAC excl. Mexico

Mexico

Changes in world market shares for manufactures (% points)

The neoliberal premise, that free market forces are necessary and sufficient to promote healthy industrial growth, is not supported by the evidence. This chapter argues that some liberalization is probably desirable, but full liberalization is not necessary – on the contrary, it may be harmful if done too rapidly and without adequate strategies to build capabilities – and it is certainly not sufficient. This is illustrated by the experience of the mature Asian Tigers. 3 Industrialization strategies in the mature Asian Tigers While East Asian industrialization has been studied intensively and there is little need to discuss it at any length, it is useful to spell out some aspects of policy relevant to this chapter. To start with, note that there is no general ‘East Asian model’. Each country had a different model within a common context of export orientation, stable macro management and a good initial base of skills. Each model reflected different objectives and used different interventions (though some, like support for exporters and SMEs, were very similar). As a result, each had a different pattern of industrial and export growth, reliance on FDI, technological capability and enterprise structure. However, for none, even the least interventionist, was simply ‘getting prices right’ a sufficient explanation for industrial success. The different objectives of the NIEs are shown in a simplified form in Table 4.1. Let us start by looking at some broad performance indicators for the past two decades. Figure 4.7 shows rates of MVA growth for these four countries, China (as the regional giant), and industrialized and developing countries for 1980 to 2000. Hong Kong stands out for its weak industrial performance. Korea is the best performer among the mature Tigers, but China outshines the four (and the rest of the region). Figure 4.8 shows manufactured export growth from 1981 to 2000, with very similar patterns except that Singapore marginally outperforms Korea in the 1990s. Finally, we look at the growth rate of HT (high technology) exports (Figure 4.9). Note the high rates of growth for the developing world as a whole which, however, reflects the

86

International handbook on industrial policy

Table 4.1

Industrial policy objectives of NIEs Deepening industrial structure

Raising local content

FDI strategy

Raising technological effort

Promotion of large local enterprises

Hong Kong

None

None

Passive open door

None

Singapore

Very strong push into specialized high-skill/ tech industry, without protection

None, but subcontracting promotion now started for SMEs

Taiwan

Strong push into capital, skill and technology intensive industry

Strong pressures for raising local content and subcontracting

Aggressive targeting & screening of TNCs, direction into high value-added activities Screening FDI, entry discouraged where local firms strong. Local technology diffusion pushed

None except technology support for SMEs None for local firms, but TNCs target to increase R&D

Korea

Strong push into capital, skill and technologyintensive industry, especially heavy intermediates and capital goods

Stringent local content rules, creating support industries, protection of local suppliers, subcontracting promotion

FDI kept out unless necessary for technology access or exports, joint ventures and licensing encouraged

Strong technology support for local R&D & upgrading by SMEs. Governmentorchestrated high-tech development Ambitious local R&D in advanced industry, heavy investment in technology infrastructure. Strategic technologies targets

None, but some public sector enterprises enter targeted areas

Sporadic: to enter heavy industry, mainly by public sector

Sustained drive to create giant private conglomerates to internalize markets, lead heavy industry, create export brands

Note: SMEs refers to small and medium-sized enterprises, FDI to foreign direct investment, TNCs to multinational corporations, R&D to research and development.

rapid growth of only two countries, China and Mexico, and the relatively small initial base. HT exports by other developing regions grew fairly slowly, and the amounts involved were tiny. Let us consider briefly the industrialization strategies of each economy.7 Hong Kong was nearest to the neoliberal ideal, combining free trade and no selective targets with an open door policy to FDI. Does its success provide an object lesson in the virtues of free markets to other countries? Not necessarily: Hong Kong had unique initial conditions

Industrial policy in developing countries 14 12

1980–90

1990–2000

10 8 6 4 2 0 Singapore

Korea

Taiwan

China

Industrialized Developing

2 Hong Kong

Figure 4.7

Growth rates of MVA (% p.a.)

1981–90

1990–2000

25 20 15 10 5 0 5

Figure 4.8

Hong Kong

Singapore

Korea

Taiwan

China

Industrialized Developing

Growth rates of manufactured exports (% p.a)

1980–90

1990–2000

35 30 25 20 15 10 5 0 Hong Kong

Figure 4.9

Singapore

Korea

Taiwan

Growth rates of HT products (% p.a.)

China

Industrialized Developing

87

88

International handbook on industrial policy

and its industrial performance is rather ambiguous. Its initial conditions included a long entrepôt tradition (with global trading links), established infrastructure of trade and finance, presence of large British companies (the ‘Hongs’) with immense spillovers in skills and information, and influx of entrepreneurs, engineers and technicians (with considerable past learning) from the mainland. This unique background allowed it to launch into export-oriented light manufacturing: entrepôt strategies in other developing countries have not led to similar competitive success. In addition, the government did intervene to help industry. It allocated scarce land to manufacturers, and set up strong and wellfunded support institutions like the Hong Kong Productivity Council, an export promotion agency, a textile design centre, a technical university and, recently, a technology park with co-financing for high-tech start-ups. The absence of a strong industrial policy, however, constrained the deepening and growth of manufacturing as inherited capabilities were ‘used up’. Hong Kong started and stayed with light labour-intensive manufacturing where learning costs were relatively low and predictable. There was some progress in terms of product quality and diversification within light industry, but little industrial or technological deepening over time, in striking contrast to Singapore, a smaller entrepôt economy that pursued a strong industrial policy. As a result, Hong Kong deindustrialized as costs rose; today manufacturing accounts for less than 5 per cent of GDP compared to over 25 per cent at the peak. Its manufacturers shifted to other countries, mainly China, and its own exports went into decline in the 1990s. The economy has been growing more slowly than the other Tigers, and its main competitive advantage (providing financial and other services to China) is under severe threat as China builds its own service capabilities. In any case, as far as industrial development goes, its experience does not convince one of the unalloyed benefits of free trade. Singapore used highly interventionist policies to promote and deepen industry, but in a free trade setting, showing clearly how industrial policy can take many other forms apart from import protection. With half the population of Hong Kong, even higher wages and a thriving service sector, Singapore did not suffer a similar ‘hollowing out’ of manufacturing. Its industrial structure, with strong policy support, deepened steadily over time, allowing it to sustain rapid industrial growth. It relied heavily on TNCs but, unlike Hong Kong, the government concentrated on activities for promotion and aggressively sought and used FDI as the tool to achieve its objectives (Wong, 2003). Singapore started with a base of capabilities in entrepôt trading, ship servicing and petroleum refining. After a spell of import substitution, it moved into export-oriented industrialization, based overwhelmingly on FDI. There was little influx of new technical and entrepreneurial know-how from China, and a weak tradition of local entrepreneurship. After a decade or so of light industrial activity, the government acted firmly to upgrade the industrial structure. It guided TNCs to higher value-added activities, narrowly specialized and integrated into their global operations. It intervened extensively to create the specific skills needed (Ashton et al., 1999)8 and set up public enterprises to undertake activities considered in the country’s strategic interest, where foreign investment was unfeasible or undesirable. Such specialization, along with the heavy reliance on foreign investments, greatly reduced the initial need for local technological effort. Over time, however, the government mounted strong efforts to induce TNCs to establish R&D facilities and to foster innovation in local enterprises (Wong, 2003). This strategy has worked fairly well, and

Industrial policy in developing countries

89

Singapore now has the third highest ratio in the developing world of enterprise financed R&D in GDP, after Korea and Taiwan (UNIDO, 2002). The two larger Tigers, Korea and Taiwan, adopted the most interventionist strategies, spanning product markets (trade and domestic competition) as well as all factor markets (skills, finance, FDI, technology transfer, infrastructure and support institutions). They had a strong preference for promoting indigenous enterprises and for deepening local technological capabilities, and assigned FDI a secondary role to technology importation in other forms. Their export drive was led by local firms, backed by a host of policies that allowed them to develop impressive technological capabilities.9 The domestic market was not exposed to free trade; a range of quantitative and tariff measures were used over time to give infant industries ‘space’ to develop their capabilities. The deleterious effects of protection were offset by strong incentives (in the case of Korea, almost irresistible pressures) to export. Korea went much further in building heavy industry than Taiwan. To compress its entry into complex, scale and technology-intensive activities, its interventions had to be far more detailed and pervasive. Korea relied primarily on capital goods imports, technology licensing and OEM agreements to acquire technology. It used ‘reverse engineering’ (taking apart and reproducing imported products), adaptation and its own product development to build upon these arm’s-length technology imports and develop its own capabilities (Amsden, 1989). Its R&D expenditures are now the highest in the developing world, and ahead of all but a handful of leading OECD countries. Korea alone accounts for some 53 per cent of the developing world’s total enterprise-financed R&D (UNIDO, 2002). Korean industrial targeting and promotion was pragmatic and flexible, and developed in concert with private industry. Moreover, only a relatively small number of activities were supported at a given time, and the effects of protection were offset by strong export orientation (see below). These features strongly differentiate its interventions from those in typical import-substituting countries, where infant industry protection was sweeping and open-ended, non-selective, inflexible and designed without consultation with industry. One of the pillars of Korean strategy, and one that marks it off from the other Tigers (but mirrors Japan), was the deliberate creation of large private conglomerates, the chaebol. The chaebol were hand-picked from successful exporters and were given various subsidies and privileges, including the restriction of TNC entry, in return for furthering a strategy of setting up capital and technology-intensive activities geared to export markets. The rationale for fostering size was obvious: in view of deficient markets for capital, skills, technology and even infrastructure, large and diversified firms could internalize many of their functions. They could undertake the cost and risk of absorbing very complex technologies (without a heavy reliance on FDI), further develop it with their own R&D, set up world-scale facilities and create their own brand names and distribution networks. This was a costly and high-risk strategy. The risks were contained by the strict discipline imposed by the government: export performance, vigorous domestic competition and deliberate interventions to rationalize the industrial structure. The government also undertook various measures to encourage the diffusion of technology, putting pressures on the chaebol to establish supplier networks. Apart from the direct interventions to support local enterprises, the government provided selective and functional support by building a massive technology infrastructure and creating general and technical skills.

90

International handbook on industrial policy

Korea today has the highest rate of university enrolment in the world, and produces more engineers each year than the whole of India. Its enrolments in technical subjects at the tertiary level come to over twice the ratio in the OECD. Even more striking than its creation of high-level skills was its promotion of industrial R&D. Enterprise-financed R&D in Korea as a percentage of GDP is the second highest in the world, after Sweden, and exceeds such technological giants as the USA, Japan and Germany. Such R&D has grown dramatically in the past two and a half decades as a result of the promotion of the chaebol, export orientation, incentives, skill availability and government collaboration. All this was an integral part of its selective industrial policy. Taiwan’s industrial policy encompassed import protection, directed credit, selectivity on FDI, support for indigenous skill and technology development and strong export promotion (Wade, 2003). While this resembles Korean strategy in many ways, there were important differences. Taiwan did not promote giant private conglomerates, nor did it attempt a similar drive into heavy industry. Taiwanese industry remained largely composed of SMEs and, given the disadvantages to technological activity inherent in small size, it supported industry by a variety of R&D collaboration, innovation inducements and extension assistance. Taiwan has probably the developing world’s most advanced system of technology support for SMEs, and one of the best anywhere. In the early years of industrialization, the Taiwanese government attracted FDI into activities in which domestic industry was weak, and used several policies to ensure that TNCs transferred their technology to local suppliers. It directed FDI into areas where local firms lacked world-class capabilities. The government played an active role in helping SMEs to locate, purchase, diffuse and adapt new foreign technologies. Where necessary, the government entered joint ventures, for instance to get into technologically very difficult areas such as semiconductors and aerospace (Mathews and Cho, 1999). In Taiwan, early trade policies had ‘extensive quantitative restrictions and high tariff rates [that] shielded domestic consumer goods from foreign competition. To take advantage of abundant labour, the government subsidised light industries, particularly textiles’ (World Bank, 1993, pp.131–3). As import substitution started to run out of steam, by 1960 ‘a multiple exchange rate system was replaced by a unitary rate, and appreciation was avoided. Tariffs and import controls were gradually reduced, especially for inputs to export. In addition, the Bank of Taiwan offered low-interest loans to exporters. The government also hired the Stanford Research Institute to identify promising industries for export promotion and development. On the basis of Taiwan’s comparative advantage in low-cost labour and existing technical capabilities, the institute chose plastics, synthetic fibres and electronic components. Other industries subsequently promoted included apparel, consumer electronics, home appliances, watches and clocks (ibid.). In the 1970s, the Taiwanese government again drew upon foreign advice, now from consultants Arthur D. Little, to upgrade the industrial structure and enter secondary import substitution. These interventions included the setting up of ‘capital-intensive, heavy and petrochemical industries to increase production of raw materials and intermediates for the use of export industries’. In the 1980s, as its light exports lost competitiveness, Taiwan’s government again moved to restructure the economy.10 After extensive consultation with domestic and foreign advisors, the government decided to focus on high-technology industries: information, biotechnology, electro-optics, machinery and precision instruments, and environmental technology industries.

Industrial policy in developing countries

91

‘The shift to a high-technology economy necessitated the close coordination of industrial, financial, science and technology, and human resource policies.’ Individual tariff rates still varied widely, with widespread quantitative restrictions in use: the use of these protective instruments was made conditional on prices moving towards international levels in two to five years. The average legal tariff rate in 1984 was as high as 31 per cent, higher if additional charges are added; this is higher than the 34 per cent prevalent in the developing world (Wade, 1990, p.127). Mathews (2001) describes one of the most successful and distinctive recent tools of industrial policy used in Taiwan, R&D consortia. It is worth quoting him at some length. Unlike the case of many of the collaborative arrangements between established firms in the US, Europe or Japan, where mutual risk reduction is frequently the driving influence, in the case of Taiwan it is technological learning, upgrading and catch-up industry creation that is the object of the collaborative exercises. Taiwan’s R&D consortia were formed hesitantly in the 1980s, but flourished in the 1990s as institutional forms were found which encourage firms to cooperate in raising their technological levels to the point where they can compete successfully in advanced technology industries. Many of these alliances or consortia are in the information technology sectors, covering personal computers, work stations, multiprocessors and multimedia, as well as a range of consumer products and telecommunications and data switching systems and products. But they have also emerged in other sectors such as automotive engines, motor cycles, electric vehicles, and now in the services and financial sector as well. Several such alliances could be counted in Taiwan in the late 1990s, bringing together firms, and public sector research institutes, with the added organizational input of trade associations, and catalytic financial assistance from government. The alliances form an essential component of Taiwan’s national system of innovation. Taiwan’s high technology industrial success rests on a capacity to leverage resources and pursue a strategy of rapid catch-up. Its firms tap into advanced markets through various forms of contract manufacturing, and are able to leverage new levels of technological capability from these arrangements. This is an advanced form of ‘technological learning’, in which the most significant players have not been giant firms (as in Japan or Korea), but small and medium-sized enterprises whose entrepreneurial flexibility and adaptability have been the key to their success. Underpinning this success are the efforts of public sector research and development institutes, such as Taiwan’s Industrial Technology Research Institute (ITRI). Since its founding in 1973 ITRI and its laboratories have acted as a prime vehicle for the leveraging of advanced technologies from abroad, and for their rapid diffusion or dissemination to Taiwan’s firms . . . This cooperation between public and private sectors, to overcome the scale disadvantages of Taiwan’s small firms, is a characteristic feature of the country’s technological upgrading strategies, and the creation of new high technology sectors such as semiconductors. It is Taiwan’s distinctive R&D consortia that demonstrate most clearly the power of this public–private cooperation, in one successful industry intervention after another. Taiwan’s current dominance of mobile (laptop) PCs for example, rests at least in part on a public–private sector led consortium that rushed a product to world markets in 1991. Taiwan’s strong performance in communications products such as data switches, which are used in PC networks, similarly rests on a consortium which worked with Taiwan’s public sector industry research organization, ITRI, to produce a switch to match the Ethernet standard, in 1992/93. When IBM introduced a new PC based on its PowerPC microprocessor, in June 1995, Taiwan firms exhibited a range of computing products based on the same processor just one day later. Again this achievement rested on a carefully nurtured R&D consortium involving both IBM and Motorola, joint developers of the PowerPC microprocessor, as external parties. Taiwan is emerging as a player in the automotive industry, particularly in the expanding China market, driven by its development of a 1.2 litre 4-valve engine. Again, this is the product of a public–private collaborative research endeavour involving three companies, which have now jointly created the Taiwan Engine Company to produce the product. Thus, the R&D consortium is an inter-firm

92

International handbook on industrial policy organizational form that Taiwan has adapted to its own purposes as a vehicle for catch-up industry creation and technological upgrading. The micro-dynamics of the operation of these alliances or consortia, is therefore a matter of some substantial interest.

This outline of industrial policy in the mature Tigers leads to the following conclusions: ●

●

●

●

●

●

●

●

Selective as well as functional interventions played vital roles in the industrial and technological development of the most dynamic economies in the developing world (Hong Kong is the odd one out since its story is largely one of truncated industrial development). Each mixed selective and functional policies in each area of intervention. There is thus no reason to partition policy into these categories: any effective policy has elements of both. In practice the distinction is impossible to draw and its importance is largely ideological. The extent of technological deepening in the three industrially dynamic Tigers is directly related to their selective interventions in industry. Those who argue that intervention was irrelevant to their industrial success show a (perhaps deliberate) lack of understanding of the real capability-building processes underlying industrialization. It is only the capability perspective that allows a full appreciation of the role of policy in industrial development. Governments in the Tigers could devise and implement complex interventions effectively. In Korea and Taiwan, that used trade interventions, export orientation imposed a strict discipline on both industry and governments. In Singapore, trade openness and the need to attract, retain and upgrade FDI did the same. In all three, government capabilities were improved over time, with growing levels of skill, remuneration and insulation allowing bureaucrats to operate efficiently and autonomously. 11 The nature and impact of interventions differed as the governments focused on different market and institutional failures. The failures were addressed by different policies, reflecting location, size, history, culture and political economy. FDI played different roles in technology development. The countries that wanted to promote indigenous technological deepening restricted foreign entry and guided MNC activities. The one that chose to rely on TNCs had to intervene to concentrate on investors, guide their allocation and induce them to set up more complex functions than they would otherwise have done. The options and compulsions applicable to the larger economies, with greater scope for internal specialization and local content as well as better established indigenous enterprises, were different from those open to small states with weak indigenous entrepreneurship and a tiny internal market. Given the need to spread technological development more widely, the former had to take more direct steps to assist local firms.

Finally, the contrast between the success of industrial policy in the Tigers and its failures elsewhere suggests that there is no justification for the general Washington Consensus case against selective interventions. It shows instead that the outcome depends not on whether governments intervene but how they do so. On ‘how to intervene’, the differences

Industrial policy in developing countries

93

between typical import-substituting strategies and those used in the Tigers lay in such things as the following: ● ● ● ● ● ● ● ● ●

Selectivity (picking a few activities at a time) rather than promoting all industrial activities indiscriminately and in an open-ended way. Picking activities and functions that offered significant technological benefits and linkages. Forcing early entry into world markets, using exports so as to discipline and monitor both bureaucrats and enterprises. Giving the lead role in productive activity to private enterprises but using public enterprises as needed to fill gaps and enter exceptionally risky areas. Investing massively in skill creation, infrastructure and support institutions, all carefully coordinated with interventions in product markets. Using selectivity in FDI help build local capabilities (by restricting FDI or imposing conditions on it) or to tap into dynamic, high-technology value chains. Centralizing strategic decision making in competent authorities who could take an economy-wide view and enforce policies on different ministries. Improving the quality of bureaucracy and governance, collecting huge amounts of relevant information and learning lessons from technological leaders. Ensuring policy flexibility and learning, so that mistakes could be corrected en route, and involving private sector in strategy formulation and implementation (Lall and Teubal, 1998).

The list could be extended but suffices to show that there are many ways to design and implement industrial policy; the analysis of the successes and failures offers important lessons on what to do now. There are also many levels of selectivity, and adopting ‘industrial policy’ does not mean that the country has to copy the comprehensive and detailed interventions used in Korea or Singapore. In fact, the new setting may provide a case for lower degrees of selectivity in some areas. At the same time, the rigours imposed by globalization and technical change may well strengthen the case for more intervention in others. The mistakes of some industrial policy attempts should not be allowed to overshadow the success of others. The evidence on the benefits of their effective use is overwhelming (and stretches so far back in history, well beyond the post-war period covered here) and that on the effects of the alternative (passive and rapid liberalization) is very disappointing for countries with weak capabilities. To insist on the difference between selective and functional interventions and to condemn the former outright seems to fly in the face of theory and evidence – it carries the hallmarks of ideology. 4 Conclusions There is a valid theoretical and empirical case to promote industrialization.12 The interventions used have to be selective as well as functional: given the differences in learning and externalities between activities and technologies it makes no sense to use only functional policies or to protect them equally. The evidence suggests that selective interventions have not only worked well but have been critical to rapid and sustained industrial development. And the evidence comes not only from the newly industrializing economies

94

International handbook on industrial policy

of East Asia but also from the early industrial history of the currently rich countries (Chang, 2002). Selective interventions have to go well beyond protection; much of the debate on industrial policy is confused by equating it with trade barriers. Protection, if indiscriminate, high and open-ended, cannot lead to healthy capability development. Exposure to international competition in some form is vital to provide the incentive and information enterprises need to invest in this difficult, costly and uncertain process. The East Asian Tigers (and other industrializing countries before them) used export orientation to provide the discipline and motivation that firms needed to invest in capabilities. The success of export orientation in Asia did not lie, as neoliberals argue, in ‘getting prices right’ (to realize static comparative advantage). It lay in providing a setting in which selective interventions could work efficiently to create new forms of dynamic comparative advantage that would otherwise have been realized much later, or not realized at all. Industrial policy involves more than giving ‘breathing space’ and incentives to enterprises to learn. Learning is rife with externalities and information failures. It requires coordination across firms and industries linked to each other in the production chain. Such coordination is almost impossible with free markets, particularly in less developed economies where information and institutions are weak, investment uncertain, inter-firm linkages poor, and learning processes unfamiliar and unpredictable. Enterprise coordination would not work unless factor markets improved in line with technological needs. Ensuring this requires coordination at a higher level, between the productive sector and factor markets. Education, training, technology, credit and other institutions lack the information, resources and risk-taking ability to respond promptly to the needs of new technologies. Moreover, when the technological structure of industry is influenced by government, only policy can ensure that the coordination takes place. The coordination involves more than freeing up labour or credit markets, privatizing education or infrastructure, and so on; it needs the selective creation of new skills and support institutions (see Ashton et al., 1999, on skill formation in East Asia). Typical marketfriendly solutions are unlikely to suffice. Competitive success is not a once-for-all achievement. It has to be sustained with consistent policy support, not just to improve capabilities within activities but also across them. This is because different activities offer different opportunities for technology upgrading, productivity increase, spillover benefits and learning. They also have different prospects for export growth. In general, complex and technology-intensive activities offer a combination of learning, spillover, productivity and export benefits. The essence of industrial growth lies in shifting from low-learning, low-technology and slow-growing activities to those that offer greater benefits. This shift may not, given the structural impediments arising from weak markets, occur rapidly without policy intervention. Competitive success depends on the ability of the government to mount policies to build local capabilities, strengthen learning systems and tap the potential of global production systems. In East Asia this was done in different ways by different countries.13 Some relied mainly on building domestic capabilities, some mainly on tapping foreign ones (for example, plugging into labour-intensive activities in global value chains) and some by combining the two. Passive reliance on FDI to drive upgrading did work for a while, but it is now running out of steam with rising wages, changing technologies and the emergence of competitors. In the long term, therefore,

Industrial policy in developing countries

95

there is only one way to succeed in industry: develop and diversify local capabilities. FDI can certainly help in this, but only to the extent that it is privately profitable for TNCs to strengthen local production skills; only the government can provide the public goods of capability development and promote difficult learning processes. Building local capabilities does not follow automatically from opening up to trade and investment. On the contrary, national capabilities determine how well each country responds to the stimulus of liberalization. Without capabilities the response can be feeble or non-existent (beyond realizing inherited endowments like natural resources), the story of much of recent industrial experience in the developing world. The capability approach explains much better than the Washington Consensus why recent liberalization has evoked such a poor competitive response in most countries, even after macroeconomic stabilization. The neoliberal explanations rely on insufficient liberalization, poor investment climate or weak governance to explain industrial failures; they rarely include technological learning and the failures involved. Yet, without addressing such structural factors, the explanation does disservice to the development cause. Finally, there are larger policy implications arising from the cumulativeness of the process. Not only is capability development cumulative and path-dependent, initial success provides more resources to invest in better factors and institutions, leading to further divergence. There is little inherent in this system to reverse this: further liberalization in a setting of rapid technical change and globalization can simply lead to the widening of the gap between the industrial haves and have-nots. Only strong policy intervention can remedy this unhealthy and unsustainable trend. Notes 1. 2.

3. 4. 5. 6. 7. 8.

This chapter draws upon a paper prepared by the author for the Group of 24, the UNIDO Industrial Development Report (2004) and earlier publications. The advice and encouragement of the editors, Sandrine Labory and Patrizio Bianchi, is gratefully acknowledged. This section draws on work done by the author for UNIDO (2004). We define the regions as follows: East Asia includes China and all the countries in the South East and East Asian region apart from Japan (and EA 2 excludes China). LAC (Latin America and the Caribbean) includes Mexico, except where specified (LAC 2 excludes Mexico). MENA (Middle East and North Africa) includes Turkey but not Israel, which is counted as an industrialized country. SSA (Sub-Saharan Africa) includes South Africa except where specified (as SSA 2). The MVA data are taken from UNIDO files. China has a major role to play in this rise; see Lall and Albaladejo (2004). Their future competitiveness faces the problem of whether they can, in the face of intense competition from lower-wage China, upgrade domestic capabilities rapidly enough to enter higher-value, more advanced segments where their wage costs must drive them. See Lall and Albaladejo (2004). All the data are taken from the UN Comtrade database. This draws upon Lall (1996). The Singapore government invested heavily in creating high-level skills to drive the required upgrading of the industrial structure. The university system was expanded and directed towards the needs of its industrial policy, its specialization changed from social studies to technology and science. In the process, the government exercised tight control of curriculum content and quality, and ensured its relevance for the activities being promoted. Apart from formal education, the government also directed considerable effort to developing the industrial training system, now considered one of the best in the world for high-technology production. Singapore is a regional leader in employee training programmes held outside the firm. It set up the Skill Development Fund in 1979, along with a Skill Development Fund Levy, which collected a levy of 1 per cent of payroll from employers to subsidize the training of low-paid workers. This marked the identification of a technology-intensive and knowledge-intensive industrial structure and high value-added orientation as national objectives with policy thinking focused on the importance of ensuring suitable human resources. The SDF levy is disbursed to firms that send their low-paid employees to approved training courses.

96 9. 10.

11.

12. 13.

International handbook on industrial policy On the semiconductor industry, see Mathews and Cho (1999). On the use of OEM as a tool of technology transfer, see Cyhn (2001). A good example of the Taiwanese government’s support for industrial restructuring is in a mature industry, textiles. Textile exports, Taiwan’s second largest foreign exchange earner in the early 1990s ($12 billion in 1993), consist mainly of synthetic fibres, since labour-intensive garments have been largely relocated to lower-wage countries. Faced with rising labour costs and intensifying competition from cheaper countries, the government embarked in the late-1980s on a major programme of restructuring and upgrading the industry. The Industrial Development Bureau of the Ministry of Economic Affairs developed a $95.4 million programme, of which 95 per cent was to be in grants to private firms to speed up technological renovation, encourage R&D, improve design capabilities and train technical and managerial personnel. Over 250 textile plants received financial and technical assistance under this programme. A number of other public and private agencies were involved. The Taiwan Textile Federation and the CETRA Industrial Design Centre provided information, design training and design shows. The China Productivity Centre sent out technical teams to visit plants and advise on automation. Banks provided low-interest loans to SMEs to move their facilities overseas and had a special credit line for them (up to $60 000 each) to import new equipment. These efforts induced textile firms more into the latest open-end rotor spinning and waterjet and air-jet weaving technologies and improve their technical skills. Indigenous designers were beginning to establish a reputation in export markets and the simpler facilities were relocated to China and South East Asia. There was no ‘super-bureaucracy’ in East Asia, and the process of building administrative competence was slow and halting. It often focused on the critical operational parts of the government rather than covering the whole apparatus. Thus there are important transferable lessons on improving government capabilities from the Tigers: it is difficult to argue that their ability to mount industrial policy was unique and unrepeatable. See Evans (1998) and Cheng et al. (1998). We take as given the desirability of appropriate framework conditions: good macroeconomic management, realistic exchange rates, political and social stability and a supportive environment for the private sector. For an analysis of the strategies adopted by the new Tigers (Malaysia, Thailand, Philippines and Indonesia), not discussed in this chapter, see Lall (2001).

References Amsden, A. (1989), Asia’s Next Giant, Oxford: Oxford University Press. Ashton, D., F. Green, D. James and J. Sung (1999), Education and Training for Development in East Asia: The Political Economy of Skill Formation in the East Asian Newly Industrialised Economies, London: Routledge. Chang, Ha-Joon (2002), Kicking Away the Ladder: Development Strategy in Historical Perspective, London: Anthem Press. Cheng, T-J., S. Haggard and D. Kang (1998), ‘Institutions and growth in Korea and Taiwan: the bureaucracy’, Journal of Development Studies, 34(6), 87–111. Cyhn, J. (2001), Technology Transfer and International Production: The development of the electronics industry in Korea, Cheltenham, UK and Northampton, MA, USA: Edward Elgar. Dicken, P. (1999), Global Shift: transforming the world economy, 3rd edn, London: Paul Chapman Publishing Company. Evans, P. (1998), ‘Transferable lessons? Re-examining the institutional prerequisites of East Asian economic policies’, Journal of Development Studies, 34(6), 66–86. Lall, S. (1996), Learning from the Asian Tigers, London: Macmillan. Lall, S. (2001), Competitiveness, Technology and Skills, Cheltenham, UK and Northampton, MA, USA: Edward Elgar. Lall, S. (2004), ‘The employment impact of globalization in developing countries’, in M. Vivarelli and E. Lee (eds), Understanding Globalisation, Employment and Poverty Reduction, Geneva: International Labour Office. Lall, S. and M. Albaladejo (2004), ‘China’s competitive performance: a threat to East Asian manufactured exports?’, World Development, 32(9), 1441–66. Lall, S. and M. Teubal (1998), ‘ “Market stimulating” technology policies in developing countries: a framework with examples from East Asia’, World Development, 26(8), 1369–85. Mathews, J.A. (2001), ‘The origins and dynamics of Taiwan’s R&D consortia’, Research Policy, 30. Mathews, J.A. and D.S. Cho (1999), Tiger Technology: The Creation of a Semiconductor Industry in East Asia, Cambridge: Cambridge University Press. Rodrik, D. (2001), ‘The global governance of trade as if development really mattered’, April, background paper for UNDP (http://ksghome.harvard.edu/~.drodrik.academic.ksg/papers.html). Stiglitz, J.E. (1996), ‘Some lessons from the East Asian miracle’, The World Bank Research Observer, 11(2), pp. 151–77.

Industrial policy in developing countries

97

UNDP (2003), Human Development Report 2003, New York: United Nations. UNIDO (2002), Industrial Development Report 2002/2003, Vienna: United Nations Industrial Development Organization. UNIDO (2004), Industrial Development Report 2004, Vienna: United Nations Industrial Development Organization. Wade, R.H. (1990), Governing the Market, Princeton: Princeton University Press. Wade, R.H. (2003), ‘Creating capitalisms’, introduction to new edition of Governing the Market, Princeton: Princeton University Press. Westphal, L. (2002), ‘Technology strategies for economic development in a fast changing global economy’, Economics of Innovation and New Technology, 11, 275–320. Wong, P-K. (2003), ‘From using to creating technology: the evolution of Singapore’s national innovation system and the changing role of public policy’, in S. Lall and S. Urata (eds), Competitiveness, FDI and Technological Activity in East Asia, Cheltenham, UK and Northampton, MA, USA: Edward Elgar, 191–238. World Bank (1993), The East Asian Miracle, Oxford: Oxford University Press.

5

Do informational service activities translate into new industrial policy requirements? Jacques De Bandt

1 Introduction Industrial policies were in their heyday in the 1970s. While transfers from the state to enterprises were increasing at an accelerating pace, in most European countries they launched violent debates within OECD spheres. For one thing, the interpretation (De Bandt, 1983, p.36) being that the countries had to face and to manage the recent crisis (already a crisis of deindustrialization, because of the NICs), these policies were, very pragmatically, recognized as a kind of inescapable response, aiming at preventing extreme industrial damage. Even while the real nature of the crisis was not clear, for some time, the crisis was there and governments had to assume their responsibilities. But, on the other hand, such massive state interventions (in deeply affected sectors) were difficult to accept from the standpoint of the orthodox market economy. A compromise of principle was found, by introducing (rather hypocritically) a distinction between positive and negative industrial policies (Michalski, 1983, p.126), according to whether they were in line or not with what market mechanisms would have worked out in any case. The concept of ‘positive adjustment policies’ was proclaimed in a 1982 OECD Declaration of Ministers. But by then the industrial policy tide was over. The move towards deregulation, privatization and so on, towards more orthodox market principles, initiated in the Anglo-Saxon world, was rapidly gaining momentum. Even while the role played by industrial policies – in the form of industrial development strategies, referring to some kind of plan rationality (Johnson, 1982) – in the success stories of Japan and South Korea was progressively to be recognized in the 1990s, this kind of recognition came too late, as the trend away from industrial policies had gained institutional bases. With huge concessions for R&D or innovation policies, industrial policies, in the form of transfers to enterprises, came to be completely forbidden. Today, governments (both national and local governments) and enterprises (both big and small enterprises) are again asking for industrial policies, that is, for massive state interventions aiming at securing competitiveness and further growth. Of course such interventions should be labelled as being somehow R&D, technological or innovation policies, that is, as pre-competitive (even if enterprises are advocating mainly competitive requirements), or in any case should be ‘positive’. The main argument refers again to deindustrialization, but this time essentially as a result of delocalizations (to the Far East). The question is whether or not this is just a repetition of what happened in the 1970s: industrial policies aiming again at facing severe deindustrialization trends. While, in the industrialized countries, industrial employment reached a peak some time in the 1970s and has been substantially reduced since (like the share of industrial value added in GDP), employment reductions, and this time also in many tertiary activities, are again 98

Informational service activities 99 seen as significant now and for years to come. If this is the main reason, then industrial policies may indeed appear as constituting a kind of repetition of what happened in the 1970s: to restructure declining industries and help develop new industries. But at the same time, since the early 1990s, there has been much talk, not only about the ambiguous ‘new economy’, but also, more deeply, about the emerging information, knowledge or learning economy, or about the new emerging technical (or sociotechnical) system. If this is so – and who can have any doubts about this? – then the problem may appear at once to be substantially different. If the technical and economic systems are changing, this then means that the basic data or conditions are different and that industrial policies may have to be more or less completely different: having different objectives and targets, subjects (or variables), actors and agents, procedures and so on. The question raised here is whether the new informational and knowledge activities, which have already become quantitatively dominant in the production sphere, are indeed changing things more or less fundamentally. In order to answer such a question, we have first to try to get some more precise idea of what has changed in the production sphere and is imposing new ways of doing. Once this is understood, it may become possible to explain why industrial policies are indeed necessary, for a combination of reasons, translating into a huge gap between (knowledge production) requirements and capacities. On that basis an attempt will be made to define the new so-called ‘industrial policies’, the new ways and means of state interventions as required within the framework of the new technical (or sociotechnical) system.1 2 New conditions, new modes of production The first step thus consists in trying to understand how and to what extent production realities have changed (and are changing further). One has first to have a look at the world of production and more particularly at the new production conditions, that is, at the new conditions for producing not only products (goods and services), but values and incomes. Of course everybody knows that our economic systems have been changing dramatically in recent decades, let us say since 1970. The analysis (De Bandt, 2002a, 2002b) of the impressive series of transformations which have been going on, over a period of 25 or 30 years, shows a combination of different types of transformations. Some of these transformations are rather well known: they consist of contextual macroeconomic phenomena (inflation, unemployment, globalization, US deficits and the like). Global transformations during the period include some much debated phenomena, such as the much heralded ‘small is beautiful’, or the ‘productivity paradox’, but which appear to have been more or less transitional. Major transformations of the period also refer to a series of major technological breakthroughs, opening up a wide range of new opportunities, all over the place. But then also (and this is what is of interest here) a whole series of ‘new ways of doing’: new behaviours (such as much increased R&D, cooperation, networks, externalization, mergers and acquisitions, the ‘control revolution’), new rules (such as fluctuating exchange rates, environmental norms) and new policies (such as innovation policies, privatization, deregulation). As concerns this last category of transformations, corresponding to new ways of doing, at the level of both private and public decision makers, the interpretation would necessarily seem to be that those actors are facing new realities and constraints, and are having difficulties in doing so. The new technologies have been of enormous help in controlling and rationalizing the direct (material, industrial)

100

International handbook on industrial policy

production activities. But while this allows for substantial cost reductions, this is not enough for creating new value, on the basis of new productive solutions. They thus try to adapt their behaviour and to design and implement, more or less on a trial and error basis, new ways of doing in their production activities. And they are more or less, but on the whole not very, successful. The main reasons seem to be of three different, but rather strongly interrelated, kinds. First, it appears very difficult, for most decision makers, to understand what has been changing really, on a permanent basis, and what the implications are. All actors were accustomed to the workings, organizational modes, rules and procedures of the industrial system, as they had been maturing over the long run, and it has been and still is difficult, or even impossible in many instances, to imagine that the system is changing and that the ways and means for producing products and value have really changed. It is quite difficult to imagine where the system is heading. A second reason is of course the reluctance to accept change: there is of course a strong tendency to stick both to existing facilities and to traditional ways of doing, be they forms of organization, rules or procedures. But a third reason seems to be the importance and length of the learning processes they have to go through and of the implied (non-material) investments they have to make. And these reasons are reinforcing each other. The reluctance for change is the stronger, the stronger the existing uncertainties as concerns the new production conditions and the higher the required efforts and investments. And the importance of both learning processes and investments is related to existing uncertainties. How are the new production realities and conditions to be characterized? Is it at all possible to give a clear picture or representation of the actual world of production? One has of course to take account, first, of the general context, which can be described essentially in terms of both globalization and competition. These include increased communications at lower costs, increased international flows of production factors (including transfers of technology and competencies), dissemination of production facilities and capabilities worldwide, continuous rapid growth of international trade, much increased competition all over the place. One central preoccupation of industrialized countries today does concern the pace – both the present pace and more particularly the likely future pace – of delocalization moves of production facilities and jobs, including skilled jobs. To this must be added, in order to characterize the context, one process which is to become an absolutely central feature of our economies in all their dimensions, namely the ageing population. Within this context, production realities are characterized by two rather fundamental phenomena. The first of these phenomena does concern the purely physical reproduction of products. In the last 15 years or so, the costs and relative prices of purely reproducible goods, that is products which are reproduced identically, with the same technical specifications, have been undergoing (since some time in the 1980s) strong downward tendencies. For some basic products, for example in the clothing industry (such as tee-shirts or jeans), prices have been enormously reduced and can even be said to tend asymptotically towards zero. This is due to two quite different mechanisms. On the one hand, there is the enormous pressure of low wages. The question of low wages, of competition from low-wage countries, has been discussed for many years now. Of course in labour-intensive industries (textiles, clothing, footwear and so on), it has been known since the 1970s that huge wage differences constituted a thread in the industrialized countries, which have responded by

Informational service activities 101 introducing new more capital-intensive techniques, thus increasing labour productivity. As was expected, from the theoretical point of view, varying production factor combinations were being used throughout the world, depending on the relative prices of (immobile) production factors. Many of the disadvantages of high wages were thus compensated for by the use of more capital-intensive techniques, but for some cases (such as clothing) in which substitution elasticities were too low. With the diffusion of production techniques and capabilities, all over the place, similar production factor combinations tend to be used and the wage differences are again a decisive factor in all those activities where products are being reproduced (more or less) identically.2 On the other hand, the ‘control revolution’ was launched some time in the 1980s (Beniger, 1986). On the basis of progressively more and more sophisticated business software, it has become possible, over a period of 15 to 20 years, to record and control automatically all operations and flows within production processes, on a continuous basis, and thus to push the rationalization of such operations to its limits. While, until the 1970s, much importance was attached, from the organizational point of view, to the distinction between continuous (‘process’) and discontinuous production processes, most industrial production have somehow become process industries. The brutality of the competitive process has thus suddenly become quite severe. While the emerging countries are accelerating their contribution to and the growth of their share of world production, on the basis of both the delocalization of basic production facilities from the industrialized countries and their entry into many of the new production activities, the industrialized countries have to respond by accelerating their move towards the new information economy. Having lost the competitive battle for purely reproducible products, industrialized countries have no choice but to try to find their way along the innovation path. A steady-state innovation regime is now required. This of course is not new. For many years now (progressively, since the 1980s), innovation policies – or various combinations of science, technology and innovation policies – have been designed and implemented, in order to promote innovations systematically. One way or another, more actors and processes have come to be mobilized in order to accelerate knowledge transfers towards (profitable) applications. National and local innovation systems have become central. Depending on the criteria, innovation policies are considered as more or less successful or unsuccessful. On the basis of the comparative analysis of the experiences so far (on a large scale), they can certainly be improved and more resources can still be allocated. But the pure extension of what actually exist would not seem to be up to requirements, as will become clear below. The second of the fundamental phenomena to be noted is the much increased (systemic) complexity and uncertainty of production realities. Of course the above phenomenon is already explaining part of the story: the number of players and the variety of conditions under which they are playing as production actors are already increasing substantially the number of elements (both variables and unknowns) in the system and thus of possible combinations and interactions within the system. And with those numbers of such combinations and interactions, the numbers of possible uncertain combinations and interactions are likely to increase more than proportionately. But this is indeed only part of the story. Beyond pure numbers of actors, economic and production systems have become multidimensional. And the diversity of dimensions

102

International handbook on industrial policy

(aspects of production realities) which have to be taken into account is steadily increasing. To mention just a few: variety and evolution of technologies and techniques, of capital goods, of specific competencies, of consumers, of uses, of markets, of materials, of environment (different aspects of environment), risks, hazards, various kinds of uncertainties, locations, transport. All these aspects are more or less interrelated, but with many interrelations being unknown or uncertain. Increased complexity may be said to have two main consequences. A first consequence is the importance of decreasing returns within our developed production systems: the complexity of production realities is translating into the necessity to allocate more resources than before in order to obtain the same or a similar production result. Environmental constraints or norms, or new specifications required from the user, or new risks, they all translate into more resources being allocated and into decreasing returns. This is more directly visible in such complex systems as the health or education systems, in which organizational costs are strongly increasing, but this is the case all over the place, in most production spheres. Of course, there are also increasing returns: thanks mainly to the new ICT, rather strong increasing returns are at work (in the USA) or in any case possible. Increasing and decreasing returns are at work at the same time, and are somehow compensating each other to a certain extent, but it is difficult to disentangle within observed facts the relative importance of those opposite forces. But this should not lead to an underestimation of decreasing returns (nor, for that reason, of increasing returns). Major facts are as follows: (a) the share of the purely ‘industrial’ component (raw material and direct labour) of goods is now well below 20 per cent in most industrialized countries; (b) the considerable increase in non-material business investments and in business-related service activities, part of which is a reflection of decreasing returns, activities aiming at compensating for the complexity of the systems and the problems being encountered; (c) the contradiction (during the ‘productivity paradox’ period) between important productivity increases at the plant or micro levels, and the low productivity increases at more global levels; systemic complexities translate into decreasing returns which are (more or less) compensating for the micro-level productivity increases.3 If this is so, part of the productivity gains is used in order to pay for the complexities and decreasing returns affecting the production system. And it is necessary to find solutions for reducing or solving these decreasing returns, before we can take the full benefit of the ICT-driven productivity increases in order to feed growth. The second consequence of the increased complexity of production realities is the necessity for all the actors involved to face this complexity. They are confronted, on a more and more continuous basis, with complex or multidimensional problems – we say usually, with increased uncertainties (because of the number of unknown variables and interactions) – and have to find new, better solutions. For that purpose they have to enter, more or less systematically, into the production of (more or less) complex knowledge. Production is becoming much more knowledge-intensive (knowledge in general) but, taking account of what has been said of the complexity of production realities, this concerns particularly complex knowledge.4 While simple knowledge can easily be automated, and while part of ‘problem’ knowledge (in cases in which the problem is mainly related to numbers of variables and possible combinations) is actually in the process of being automated, no such automation is possible in the case of complex knowledge, whose production require series of interactions within very demanding organizational learning processes.

Informational service activities 103 Non-material investments, which are progressively being substituted for material investments, correspond to the new type of roundabout methods of production, through the production of knowledge. Direct production activities being marginalized, more importance and more resources are being allocated to indirect production activities aiming at designing and implementing new solutions. To summarize, conditions have changed dramatically, let us say since the 1970s. In a context of globalization and generalized competition, the competitive battle for purely reproducible goods is lost. The ‘industrial’ economy, meaning the production of goods and value through the rationalization of the energetic transformation of raw materials into identically reproduced goods, is somehow exhausted. Simultaneously, the systemic complexities of production realities have grown enormously, and uncertainties have increased dramatically. Of course, manufacturing matters (Cohen and Zysman, 1987), but only to the extent that new ways of doing manufacturing, on the basis of informational activities, can restore the capacity to produce value. For both these reasons – necessity to face complexities and necessity to find new innovative solutions – new modes of production of products, value and profits have to be developed, very systematically. This is what may be called the new information or learning economy: roundabout methods of production based on the production of (complex) knowledge, in order to come up, continuously, with better adapted, specific solutions. Better adapted means that, on the basis of existing opportunities, that is, the spectrum of scientific and technological possible solutions, new solutions are designed which are on the one hand feasible (taking account of the existing constraints of all kinds) and on the other hand relevant, with reference to the specific needs of the users or consumers. 3 Industrial policies are seriously needed Why is it that these phenomena and the implied new modes of production do require ‘industrial’ policies, and not only that, but both increased state or collective interventions, and policies of another kind? There are four types of strongly interrelated reasons. The first type of reasons has to do with the importance of the unsatisfied needs for such new modes of production. The second type has to do with the enormous difficulties involved in entering the new learning economy, that is, in trying to meet those needs and to develop the required modes of production. Those two first types of reasons together translate currently into a big gap between requirements and capacities, and these are both increasing. The third type of reasons has to do with the degree of inadequacy of market mechanisms for organizing and securing the necessary dynamics. The fourth, correlative, type of reasons then has to do with the importance of the collective, cooperative dimensions of the required organizational learning processes. Let us give some rather short indications as concerns these different types of reasons. The production actors, both private and public decision makers, are strongly in need of new modes of production. For some time, in the 1980s and part of the 1990s, they all became progressively aware of the fact that the traditional or usual ways of doing (in the field of production activities), were less and less appropriate and that they were more and more confronted by new types of more complex problems and new organizational requirements. Many examples were observed of the disfunctioning of firms which had been considered until then as being well-organized and well-managed. Traditional management models and methods became visibly inadequate. While firms were resorting more

104

International handbook on industrial policy

systematically to non-material investments, they found out progressively that, because of the complex problems they had to face and the implied competence problems, they were becoming less autonomous, and had to find ways to complement their own resource and competence bases. Many alternative solutions were tried: mergers and acquisitions (in order to enlarge their competence basis); also (for the same purpose) more R&D; more training expenditures; outsourcing (in order to concentrate their activity on their ‘comparative advantages’); and the use of external service providers; inter-enterprise cooperation agreements (in order to unite efforts); decentralization (in order to give more freedom of initiative and action at decentralized levels); entering into networking schemes, including those at the local level; but also, finally, trying more and more to promote new training activities (trying to influence existing structures or to develop new structures, on cooperative bases).5 One should add here ‘economic intelligence’, seen as non-material investments (approaches and tools) for better understanding production realities and systems in their complexity. While many of these solutions have proved to be only poorly successful (the proportion of mergers and of cooperative agreements which have failed is considerable, and the use of business service providers has proved to be often rather deceptive or disappointing), and while economic intelligence is still in its infancy, it is true that some of the experiences with the organization of competence poles, on a local basis, seem more promising. What this all means is that, while firms have realized that they are confronted, on a continuous basis, by complex questions and problems, and increasingly so, they do not seem to have learned so far how to produce all the complex knowledge which is required. This is typically what Drucker (1999) is repeatedly saying: as concerns information jobs, we have not learned yet and are still in the Pre-Taylor era. This is (and this constitutes the second type of reasons) because of the inherent difficulties of knowledge production activities and processes, the more so when what is at stake is the production of rather complex knowledge. There are three main requirements, the definition of which immediately shows the difficulty involved and the time which may be needed. Firstly, the production of knowledge, and more especially of complex knowledge, implies that the necessary competencies are available, and have thus been developed. Of course people have competencies which have been accumulated in the past and, confronted by the new more complex problems, are trying to solve those problems as best they can and gain new experiences. But in order to do better with more adapted knowledge and methods, they need to have followed new training programmes that are adapted to the new paradigm upstream and to have accumulated, on those new bases, more fruitful experiences. This is obviously both a difficult and a long process. Secondly, in order to produce the required complex knowledge, as concerns a specific question or problem in a particular production domain, one must have the comprehensive knowledge (the understanding) of the concerned production system in its complexity.6 As systems analysis is seldom included in teaching programmes as yet, only a very few people have the required competencies for applying (even approximately) systems modelling techniques to the particular production systems. Thirdly, because of the multidimensionality of the production systems, the production of complex knowledge requires the cooperation and interactions of different complementary competencies. Within the organizational learning processes, different types of interactions have to be organized.7 Needless to say, bringing together different people (with

Informational service activities 105 different competencies, of different origin, training, status, interests, personality) to work together of course raises a series of difficult problems, from cognitive and language problems, to various organizational problems, and to a series of property and income problems. This indicates that the difficulties are huge, and that, even if progress is obviously being made along various lines (expert systems, knowledge engineering, economic intelligence, systems modelling), such progress tends to remain rather slow, as a matter of fact too slow. As indicated above, those two types of reasons translate currently into a wide gap between knowledge needs and knowledge production capacities. This is revealed particularly by what has been happening (let us say since the mid-1990s) as concerns knowledgeintensive business services. The rapid growth of these informational service activities clearly shows that the needs are rapidly increasing. However, and this has been showing up more recently, such informational service provision appears to be quite disappointing, and the clients have become very critical. It appears that the needs but also the knowledge production capacities are increasing, but the danger exists that the gap might be widening. One possible interpretation of the slowness of the progress being made in the required learning processes relates to the inadequacy of the market mechanisms for inducing such learning processes. As suggested, when confronted by complex problems, enterprises are trying to meet their knowledge needs, by using, more or less on a trial and error basis, what appears to be the possible available solutions. They have neither the bases nor the resources for deciding, with reference to market signals, the much huger non-material investments which would be required for producing and accumulating experiences in more fundamental solutions. Because of their ‘industrial’ representation of production activities, they probably underestimate the possibilities of making profitable such non-material investments in their knowledge production capacities. They visibly remain, with exceptions of course, quite reluctant to go much further. When they decide to go further and, for that reason, to enter into relations with other actors, they are confronted by enormous information asymmetries, and the corresponding adverse selection and moral hazard phenomena. The fourth type of reasons, which can be seen as the exact corollary of the previous one, has to do with the importance of the collective and cooperative dimensions of complex knowledge production processes. There are two major aspects of this, which have already been indicated above. Knowledge production appears, under the form of non-material investments, as typical roundabout methods of production. The organizational learning processes have to contribute to building and accumulating non-material (knowledge) assets, part of which, like infrastructures, can only be developed and are in any case better seen as collective assets. While enterprises are clearly less and less self-reliant in terms of required competencies, they try to compensate by relying on the competencies of other actors, and enter into a diversity of cooperation schemes, networks and groupings. Of course, scientific and technological developments have always been social products, based on many dispersed cumulative contributions, but, compared to such rather anonymous processes, the collective dimensions have come to be much more explicit and visible. What this means is that, in order to be more efficient and, to the extent possible, successful, organizational learning processes, at the level of particular actors, have to be integrated in or to rely on larger, collective, cumulative, more comprehensive knowledge production processes, within more or less well-organized production systems. Of course, successful examples of such (often locally) organized production systems are known, but they remain

106

International handbook on industrial policy

exceptions and to a certain degree experimental, and not enough is yet known about the ways and means of their success. Again, progress along such lines is real but remains rather low. Is it not so that we stick to the traditional representation of production systems consisting of individualistic, autonomous firms linked mainly through the market place and that the a priori against any form of collective intervention remains very strong? Putting together these four types of reasons means that not only does there exist a significant gap between significant and increasing complex knowledge requirements and complex knowledge production capacities within enterprises, but also the necessary developments towards the new modes of production are only rather slow, because of the degree of inadequacy of market mechanisms and the insufficiency of compensating collective actions or policies. 4 New requirements and new forms of ‘industrial’ policies One approach to the problem would consist in renewing, if necessary, and in any case, in reinforcing and developing actual policies belonging to the domain of industrial policies. Pure transfers of financial resources to enterprise (for restructuring or investments purposes) are somehow eliminated, even if such transfers continue to take place. But, instead, R&D, technology and/or innovation policies are being developed all over the place, the emphasis being put on incentives and the allocation of resources to knowledge production and transfer processes. And competition policies, which systematically try to eliminate dominant position and abuses of market power, are explicitly accepting interenterprise agreements and cooperation at the so-called ‘pre-competitive’ level. If people consider that innovation policies are helpful and efficient, it is of course possible and probably necessary to continue to increase such policies, with still more incentives and more resources. Two remarks must be made here. On the positive side, specialists consider that it should indeed be possible, by addressing new populations of enterprises which have not yet entered into the innovation game, to increase further the scope of such policies aiming at the quantitative increase of innovations at the firm level.8 SMEs are a central target. On the negative side, one possible interpretation may be that innovation policies (the way they are designed) have exhausted much of their potential and cannot help the promotion of further increases in innovation. Innovation policies are mainly based on the model of the individual innovator, pursuing his process as autonomously as possible, but, with the necessity to take hold, in his environment, of appropriate and as much useful knowledge as possible. Rival entrepreneurs may be looking at the same time for the same or similar innovation. But the domains and possibilities for such individualistic innovation processes, which are in any case limited, are at any point of time rapidly exhausted. It is true that, in the last 10 years or so, new developments have taken place on the innovation scene which constitute clear (at least attempts of) departures from the interventions and policies designed with reference to the individualistic model. For one thing enterprises have entered into a diversity of cooperative schemes. Even if these are not necessarily a departure from the individualistic process, they introduce (irreversibly) the cooperative dimension. Different types of groupings based on local proximity enlarge the scope for cooperation to series of other actors. This will have to be integrated within the realm of the new policies. The main idea here is that innovation policies as usually understood are not up to actual requirements. Even if some of the more recent developments are seemingly going in the

Informational service activities 107 right direction, these are being made more or less on a trial and error basis, more than with reference to solid basic principles, referring to the new production realities. Much more is apparently needed in order to enter the new learning economy. If we take for granted the fact that production processes are becoming decisively knowledge-intensive, meaning that the main driving force within the production of products, value and profit is the production of (more or less) complex knowledge, and accept the observation that the individual actors cannot meet, by themselves or on the basis of market mechanisms, the corresponding requirements, then the new industrial (or industrial knowledge) policies have mainly to foster the relevant knowledge productions. What then should the required (industrial) knowledge or information policies consist of ? Industrial policies or strategies should consist of various complementary components: objectives and targets, of course, but also have the necessary resources and finally the organizational set-up (the way the organizational requirements of the collective actions are supposed to be met). A first component thus is made of objectives, goals and targets. We are not speaking here of general objectives, in terms of economic performance (growth, profits, employment, competitiveness and exports and the like), but of objectives in terms of production capacities and developments in specific areas. Such objectives must, for practical purposes, be realized in terms of particular targets (scope, time horizon, products). What is at stake is the development of particular, more or less comprehensive, production systems.9 There has always been much debate as concerns the difference between so-called ‘horizontal’ industrial policies – more or less automatic and undifferentiated interventions with reference to eligibility criteria – and ‘vertical’ policies, implying the concentration of efforts on the development of selected domains or activities. Already in the past, the industrial policies which have proved to be efficient were essentially those which, referring to some type of plan rationality, were pursuing specific targets, whether ‘grand programmes’ (in the case of France) or sectoral restructuring policies (in many countries) or industrial development strategies (in the case of Japan). With increased complexities and uncertainties, the justifications for such a selective approach and for the concentration of efforts and resources on specific domains, projects or subsystems are again much stronger. While the systematic development of human resources (training, competencies) is obviously needed on a large scale (see below), such development is in any case difficult to achieve, unless this is done within the framework of specific target domains. What is thus needed is to create the conditions for knowledge accumulation processes in particular domains. A second component includes the allocation of the required resources, taking account of the fact that some of these (financial) resources will be used in order to produce and accumulate other (mainly human) resources over time. In order to develop the above production systems, what is required here, very fundamentally, is the development of series of complementary competencies and non-material investments and, on that basis, of innovative supply and production capacities. The central objective being to enhance complex knowledge production capacities within the production system and to support and facilitate complex knowledge production processes, it is possible to define a series of well-defined goals, aimed at increasing the basic resources which are made available for knowledge production activities. While enterprises are not doing and cannot do the necessary investments purely on their own,

108

International handbook on industrial policy

appropriate policies (supporting systems, collective actions, incentives) of different kinds have to be developed. 1.

2.

3.

4.

Understanding the complexity of production realities and systems: decision makers (both public and private) being confronted by much more complex production realities, there is a great need for the development of ‘economic intelligence’, that is, for appropriate information and knowledge systems, both basic information methodologies and systems for the systematic (and systemic) analysis of those production realities. Statistical systems of the past were designed essentially for accounting purposes (from population and employment to national products) on the basis of the measurement of flows and stocks with reference to technical units and their aggregation on the basis of prices. This needs to be revised, very fundamentally, not only because the accounting purposes prove to be more and more deceptive, but because the basic production realities are now non-material (not to be counted with reference to technical units) and have to be understood in their multidimensional or systemic complexity, in a dynamic and future-oriented perspective. This is an enormous challenge. Meeting the new training requirements: the contents of all training programmes, at the different complementary levels, should be systematically adapted to the required increasing knowledge intensity of all production processes. This is far from being the case, most of the programmes having been changed only marginally over the years. Of course, changing the contents of training is a difficult and time-consuming process. As a matter of fact, even if here also some partial and dispersed progress has been made, not much has been done on the whole in order to give to the new generations of students the necessary bases for systematically increasing their capabilities for producing complex knowledge and for acquiring the necessary competencies: multidisciplinary approaches, systems analysis and modelling, information and knowledge engineering, communication, cognitive interactions and so on. Building up competencies: on the basis of such new training, what is really at stake is the systematic development of competencies, that is, the capabilities for producing and transforming knowledge into operational solutions or applications. To the extent that such competencies include tacit knowledge which is developing and accumulating on the basis of experience, and to the extent that such experience can only be acquired on the spot, there is a clear necessity for promoting and facilitating such experience and the accompanying learning processes and to induce enterprises to participate actively in the systematic organization of such experiences and learning processes. What this really means is that, while firms are of course pursuing their own interest in building up their own competence bases, they have also an interest in more collective competence-building processes and have the capability to contribute to these. As many examples have shown, positive sum games are clearly possible. Incentives, but also eventually support systems, are needed in order for them to do so, on a sufficient scale. Contributing to the development of collective non-material assets: to the extent that, within the dynamics of innovation processes, the relative roles and weights are shifting from individual actions towards more collective cooperative actions, that is, from the individual actor (enterprise) to the production system as such, it is easy to

Informational service activities 109 understand that such more collective processes will be fostered by making collectively, at the level of the production system, the corresponding non-material investment. This is akin to the ‘open-source’ approach in the field of software: many different actors are developing collectively, on a cooperative basis, the basic software, on the basis of which each one can develop its own specialized and proprietary applications. A third component of such policies is the organizational set-up, the design of the organization for implementing concretely the policies as defined. In its most elementary form, in the case of resource transfers policies, the organizational set-up consists in defining the administrative procedure for attributing the subsidies with reference to eligibility and allocation criteria, as defined. But in most cases the implementation of the policy (who is doing what and how) needs series of more direct interventions and has thus to meet series of organizational requirements. Industrial policies, in the past, have been mostly designed and decided ‘top-down’, that is, at the political and administrative levels. If we agree that more (and different) industrial policies are needed, we have to be conscious that ‘government failures’ may be a real danger. In the past, government failures, in the domain of industrial (or production) policies, but including the domain of innovation (or technology) policies, were due to several possible phenomena: political bias in the decision processes (or even bribery), lack of information (or delays in getting the necessary information), lack of competencies, lack of incentives, lack of (human) resources for implementing and controlling the application. The experience of the 1970s (particularly) has also shown that, while information and competence problems were the most decisive (at least in developed countries), such problems could be and had been successfully resolved in other countries (such as Japan). While many of these reasons can be said to remain potentially valid, but with possible solutions, account must be taken here of the much increased complexity of production realities. This complexity makes it much harder to imagine possible ways and means for designing and implementing such policies successfully, top-down and from the outside, as it were. This obviously requires a deep understanding of the complexity of the production realities which are involved and such understanding can only be realized on the basis of the systematic analysis of the system and of much experience. Such analysis and understanding can only very exceptionally be found in the administration.10 For such reasons, the approach has to become mainly ‘bottom-up’ in the sense that, within the institutional and resources framework as defined by the policies (at some global level), most of the decisions and actions as regards the specific strategic targets, the resources and their development and the organization (communication, coordination) have to take place at the level of the concerned actors, who have to assume responsibility for their decisions and actions. This obviously means that, even while some basic top-down decisions have to be taken, in order to allocate the necessary resources, to develop the necessary structures, infrastructures and institutions and to propose the necessary incentives, much more emphasis must be put on alternative, intermediate (‘meso’) approaches. As opposed to traditional ‘top-down’ approaches, this can be called ‘bottom up’, but with the understanding that the emphasis is not on individual actors, but on specific decentralized production systems. Initiatives are taken, cooperative projects are organized and dynamics are developed at

110

International handbook on industrial policy

that level, and this is where collective action has to be stimulated. This is also the level where most of the organizational requirements have to be met. This brings as to a final remark, which concerns the local dimensions of the new policies. If we consider the various types of solutions (mergers and acquisitions, interenterprise agreements, subcontracting, decentralization, networks, local systems) which enterprises are using in order to meet the organizational requirements of the learning economy (which can be synthesized by the following question: how to bring very different people to produce knowledge and production solutions together?), the solutions which have so far shown the biggest potentials are local production systems or competence poles. Not all such local systems are successful. Far from it. But several of them are successful and show some kind of rather successful combination of ‘ways and means’ for meeting the organizational requirements of the learning economy: resource bases (training and research facilities, both fundamental and applied research), entrepreneurial support systems, incubators, risk capital, specialized domains, complementary activities, selforganized systems and dynamics, bottom-up strategic approaches, networks and cooperative structures and procedures, intermediate bodies, multidimensional interactions, collective semi-public assets, information systems, and so on. Some additional components, which are still less developed, are in some places being developed, such as economic intelligence, knowledge engineering, knowledge systems, new training systems and virtual enterprises. This tends to show that such systems, in order to succeed and to be dynamic, are quite demanding. This is the very reason why supporting policies are obviously needed. Notes 1.

2.

3. 4.

Of course the term ‘industrial policy’ may be and is misleading, as it (at least) suggests a continuity and strong similarity to the industrial policies of the earlier period (following the strong after the World War II industrial growth period). The term is certainly not appropriate any more, if one refers to the share of industrial activities proper to the production of value (value added), which is now of secondary importance, compared to the share of information or knowledge jobs and activities. But at first sight this question of terminology may not appear to be that important. And why not maintain this term, to which people are really accustomed? The question is whether and to what extent there are more advantages than disadvantages in keeping this term than in trying to define more appropriate concepts and terms. Is it not so that, at this stage (in 2005), the disadvantages of keeping this term are now significantly greater than the advantages? The whole concept of industrial policy typically referred to the workings and organizational requirements of the industrial system (devoted to the energetic transformation of materials into products) as it existed before, since it had become the centre and the main driving force of the whole productive system. However, it no longer exists that way. In addition, keeping this term of ‘industrial policy’ means also, possibly or at least partially, keeping the same industrial representation of the productive system, which may be terribly misleading. Production realities are now really of another kind. But then one has to recognize that it is not easy to find some more appropriate term. And competition among the low-wage countries, under the pressure exerted by the importing developed countries, implies that, for the time being at least, these wages tend to remain low or even to decrease. Owing to the powerful drive of China, many of the countries which were recently exporting textile and clothing to developed countries are now being out-competed and some of them try to react by reducing their wages. One possible interpretation of the ‘productivity paradox’ would be that the productivity effects of ICT (‘computers everywhere’) have been strongly underestimated and this was because we have failed to recognize the importance of simultaneous decreasing returns, due to systemic complexities. Following Dibiaggio (1998), a useful distinction is made, within question–response processes, between simple, ‘problem’ and complex knowledge. In the case of simple knowledge, question and answer are well defined beforehand. In the case of ‘problem’ knowledge, question and response are new, but knowing, on the basis of similar questions and responses (in the past), both that a relevant solution is likely to be found and more or less how. In the case of complex knowledge, question and response are equally new, but it is

Informational service activities 111

5.

6. 7.

8.

9.

10.

possible to know beforehand neither whether a relevant solution can be found, nor how such solution can possibly be found. It is even impossible, for that reason, to know beforehand, whether the question which is being raised is the relevant one. In the case of France, many new institutions, called university, observatory, school or institute, have been created, in the recent past, on a cooperative basis (for a particular activity, or a particular function), in order to develop cooperative research and the corresponding teaching on the production system they belong to. As the production systems gain in complexity, the traditional hypothesis of (quasi-)decomposability of these systems is losing its validity, thus eliminating the justification of partial analyses. In an innovation process, schematically speaking, at least five types of complex interactions must be organized: (i) at the level of the problem or question (in order to come up with the appropriate or relevant question); (ii) at the level of scientific and technological possibilities (in order to come up with the various possible answers); (iii) at the level of the engineering and production activities (in order to come up with a feasible solution); (iv) with the users, clients or consumers (in order to come up with the relevant answer); and (v) between the above four different levels of interaction. It is of course very difficult to find out which innovations are to be counted as being the results of innovation policies. The question is in principle very simple. By increasing incentives, a series of innovations which were not profitable before become profitable and are implemented. But things are much less clear. In a first period, let us say in the 1970s and part of the 1980s, innovation policies appeared to be successful, but the growing number of innovations included innovations which would have been profitable anyway, without the incentives, and would have been introduced in any case. Growth was to a large extent due to better information. But then, in a second period, in order for the number of innovations to grow further, growing incentives and efforts were needed in order to bring less profitable innovations into being, but decreasing returns were rapidly setting in. The famous ‘European paradox’ (in the 1990s) was somehow a reflection of the relative inefficiency of innovation policies. But this is not the end of the story. On the basis of experience, industrial policies are being improved. For one thing, some kind of innovation culture has been diffused everywhere. But also, feasibility conditions are better known and support systems have been developed everywhere. The term ‘production system’ is used in a general sense, referring to any ‘system’ consisting of a series of interrelated elements and actors, developing their production activities in the same or complementary domains. They can be local or sectoral, be devoted to more or less narrow specializations, be more or less comprehensive, and so on. This does not mean either that the necessary knowledge cannot be acquired by an outsider (not belonging to the concerned system) or that the necessary knowledge and understanding cannot be developed within the administration. As concerns the first point, is it not the case that, in industrial economics (as in other domains), some specialists or experts do acquire, on the basis of systematic comparative analyses of the production realities, a better understanding of the system than the insiders, who have more partial views? Of course many so-called specialists, without any direct knowledge of those realities, have only secondhand expertise and knowledge. As concerns the second point, the only question is whether or not the administration decides to allocate the necessary resources for acquiring such deep understanding.

References Beniger, J.R. (1986), The Control Revolution: Technological and Economic Origins of the Information Society, Cambridge, MA: Harvard University Press. Cohen, S. and J. Zysman (1987), Manufacturing Matters, New York: Basic Books. De Bandt, J. (1983), ‘La politique industrielle: réponse de l’Etat-Nation à la crise?’ in Les Politiques industrielles, Special Issue, Revue d’Economie Industrielle, 23, 36–56. De Bandt, J. (2002a), ‘L’émergence du nouveau système technique ou socio-technique’, Revue d’Economie Industrielle, 100, 9–38. De Bandt, J. (2002b), ‘La “nouvelle économie”: économie, système ou société?’, Faire Savoirs, 2, 91–99. Dibiaggio, L. (1998), Information, connaissance et organisation, doctoral thesis, Université de Nice Sophia Antipolis. Drucker, P. (1999), Management Challenges for the 21st Century, New York: Harper Business. Johnson, C.A. (1982), MITI and the Japanese Miracle: the growth of industrial policy, 1925–75, Stanford: Stanford University Press. Michalski, W. (1983), ‘Les politiques d’ajustement positives. Un concept stratégique pour les années 80’, in Les Politiques industrielles, Special Issue, Revue d’Economie Industrielle, 23, 126–37.

PART II COMPETITION AND REGULATORY ISSUES

6

Competition policy and innovation Jochen Lorentzen and Peter Møllgaard

1 Introduction Technological progress harbours the prospect of a future in which there would be cures for and vaccines against cruel diseases such as HIV/AIDS, cancer or malaria. People the world over would enjoy the benefits of mobility in safe, zero-emission vehicles and make use of effective, affordable communication devices for both work and play. And industry would meet the growing demand for goods and services by an increasing global population while drastically reducing its negative impact on environmental sustainability. Making this future a reality requires innovations. To bring them about, firms as the major agents of technological change must invest in new products and processes. In so doing, they face costs associated with R&D that they often cannot bear individually. They confront technical risks and uncertainties related to market acceptance and, because of the increasing complexity of new applications, they come up against limits of their internal capabilities to solve a constantly evolving set of problems single-handedly. Firms have responded to this challenge through mergers and by cooperating with each other. For example, they have formed international technology alliances, including those with universities and science institutes. These partnerships aim to share costs, pool risks and complement or enhance individual capabilities through joint R&D or technology development. At the latest count there were almost 6000 of such alliances, concentrated in the Triad economies of the USA, Europe and Japan (NSF, 2004, Table 4.16). The tally is conservative in that it undercounts collaborative relationships among small firms and in certain technology fields. They are concentrated in sectors where pushing the technology frontier depends increasingly on a judicious combination of scientific (including basic) research with command over deep financial, organizational and technological resources. Examples include information technology, biotechnology, advanced materials, aerospace and defence, automotive technology and chemicals. Fewer than two out of ten registered partnerships are equity-based. Especially in areas where technological trajectories are hard to predict, these alliances tend to take the form of flexible, non-equity agreements. To the extent they are successful – and therein of course lies their promise – technology alliances may help solve the problems referred to above. From the 1980s, governments have acknowledged the role of collaborative relationships among firms for innovation, and accommodated them in a competition policy framework increasingly more tolerant of practices that earlier would have been suspected of collusion and would thus have been illegal. The situation in the USA is a telling example of this relaxation. For example, the 1980 Baye–Dole Act made it legal for government contractors to exploit the fruits of research undertaken on behalf of the US government. The 1984 National Cooperative Research Act allowed firms to collaborate on generic, pre-competitive research. Since 1993, firms may additionally engage in collaborative production. Hence a critic’s worries about ‘the shadow that neoclassical thinking [by which he meant the 115

116

International handbook on industrial policy

suspicion of horizontal agreements] casts over antitrust policy’ seemed largely a thing of the past (Teece, 1992, p.23). A similar liberalization has informed competition policy in the European Union, which aims to exploit private and public investments in the knowledge economy in its endeavour to increase its competitiveness (European Commission, 2003). In sum, the strategic responses by firms to the increased complexity and speed of innovation are reflected at least in part by evolving competition rules and industrial policies that aim to facilitate technological change. It is this connection – between innovation, competition policy and industrial policy – that this chapter addresses. What we have said so far may suggest that all is well: firms pursue novelties in cooperative arrangements, governments marshal international competitiveness, pundits wax lyrical about the knowledge economy, and the challenge of innovation is what holds it all together. But it is not quite as easy as that. The emphasis on innovation creates problems for theoretical analysis and policy practice in anti-trust. Theoretically, the uncertainty surrounding technological change in highly dynamic sectors implies that it is difficult or impossible to evaluate the likely effect of interfirm collaborations that are in principle open to anti-competitive practices. Also, even if overall R&D expenditure is lower in a cooperative set-up, the welfare effects, post innovation, may still be positive provided the former R&D partners now compete in the product market and this market does not resemble what a long defunct pop group once lamented as The Winner Takes It Aaaaall . . . For policy practice, this means that anti-trust officials do not have tried and tested general guidelines (or at least much less so than in the past) that would allow them to pronounce on the welfare implications of an alliance without facing the criticism of the firms involved, parts of the legal and economics profession, or ultimately the courts entrusted with oversight and litigation of the relevant cases. Section 2 briefly reviews the rationale behind technological alliances and provides a snapshot of their role in global competition, especially insofar as it is based on intellectual capital. This chapter is not about technological alliances per se, but, apart from their obvious link to innovation, they nicely illustrate the increased importance of horizontal agreements and thus establish the relevance of the topic. Section 3 discusses the organization of industries in a dynamic context and draws out consequences for competition policy. Section 4 concludes with an outlook on the underlying tensions between technology alliances, competition policy and industrial policy. In other words, the chapter starts with a perspective on firms, moves on to policy, and finally settles with a number of open questions that concern firms, policy makers and researchers of innovation alike. 2 Technology alliances: rationale and importance Technology alliances involve at least two partners from the public and/or private sector that collaborate around R&D or technology development. The terminology is somewhat loose; research partnerships and strategic technical alliances also frequently bounce around the literature and are in essence interchangeable. An alliance always involves legally distinct entities and is thus different from a merged company. Technology alliances are a form of strategic alliance in that the involved parties have a common goal and agree on how to reach this goal by means of pooling resources and coordinating attendant activities (Teece, 1992; see also Dunning, 1997). Alliances are formed for many reasons besides issues surrounding technology and innovation, but evidence suggests that these

Competition policy and innovation

117

are indeed among the most important. Research joint ventures (RJVs) are a subset of technology alliances; they may be entered into by firms and institutions from higher education or the science sector and, of course, imply control via equity (for an overview, see Caloghirou, Ioannides and Vonortas, 2003). Technological knowledge is subject to externalities, uncertainty and the danger of opportunistic behaviour. This results in underinvestment in knowledge creation in arm’slength markets unless firms can counter prohibitively high transaction costs resulting from incomplete contracts. Insofar as they limit the risk exposure of each participating firm while creating new or diffusing existing knowledge, technology alliances are a mechanism to achieve just that. This is especially so in high-tech sectors and in activities where technological change tends to be rapid and radical rather than incremental and gradual and where prospects for future market dominance are weak. Globalization is also a reason behind the growth in alliances. This is because of a growing mismatch between the kinds of knowledge firms require to be internationally competitive, and the technologies that are locally available. The distinct technological trajectories that countries travel over time evolve slowly, while markets converge globally at a higher speed. So, for firms to keep up with what is going on internationally, they may opt to internalize the relevant knowledge through an alliance, especially when other forms of technology acquisition are unavailable (Narula, 2003). Strategic considerations complement the picture. Thus firms may enter into a technology alliance in the pursuit of a first-mover advantage. The increased scope of activities from an alliance reduces the response time to new developments for its members who may consequently be faster onto the market. Alliances offer advantages in very complex technologies insofar as they allow firms to leverage their own capabilities with needed external competences they can acquire with a view to matching the innovative challenge of whatever it is they are after, with the requisite portfolio of human, technical and managerial capital. In the process, firms can learn from each other, thus improving their capabilities along the way. In this view, working together is not so much aimed at reducing transaction costs as at ensuring access to and internalization of external knowledge that would otherwise remain elusive (Hemphill and Vonortas, 2003; Caloghirou, Ioannides and Vonortas, 2003). Over time alliances can develop a reputation for the mutually beneficial character of relationships and thus bestow trust upon the firms involved. They can therefore act as a mechanism that allows firms to hedge their bets on future technologies with highly uncertain prospects without staying out of the game entirely. In other words, evolving multi-market and multi-project alliances provide incentives for engaging with innovation ‘on the cheap’. Empirical research largely confirms these propositions concerning why technology alliances are formed in the first place and what difference they make (for an overview, see Hagedoorn, Link and Vonortas, 2000). Over the last quarter of the 21st century, alliances have become a fixture of contemporary capitalism. A total of 185 alliances were formed in 1980; by the 1990s, the annual increase amounted to two to three times this figure (see Figure 6.1). Non-equity alliances dominated in all but two years in this period, and the ratio of non-equity to equity alliances increased significantly, especially from about the middle of the 1990s. Initially, the dominant sector was IT, but in later years biotechnology attracted the highest number of new collaborative relationships. Most alliances involve US firms, followed by firms in Europe and, finally, Japan.

118

International handbook on industrial policy 400 300 200 100 00

98

20

96

19

94

19

92

19

90

19

88

19

86

19

19

84

82

19

19

19

80

0

IT

New materials

Automotive

Biotech

Aerosp./defence

Chemistry

Source: National Science Foundation (2004, Appendix table 4.42)

Figure 6.1

International technology alliances: sectors

600 400 200

US

Japan

Europe

Non-Triad

00 20

98 19

96 19

94 19

92 19

90 19

19

88

86 19

84 19

82 19

19

80

0

Source: National Science Foundation (2004, Appendix table 4.42)

Figure 6.2

International technology alliances: ownership

However, technology alliances increasingly involve non-Triad economies as well, especially from East Asia (see Figure 6.2). Owing to the increasing homogeneity of technology across countries and firms, alliances have become an important vehicle for technological learning and industrial upgrading in select latecomer countries (Narula and Sadowski, 2002; see also Hobday, 1995). Compared to alliances among firms from Triad economies, those involving firms in developing countries tend to be relatively more equitybased. Equity-based inter-firm relationships lend themselves to the transfer of existing knowledge (cf. Chen and Chen, 2002). Typically, they are asymmetric and may fail to achieve the expectations of the alliance partners for a host of reasons (Kuada, 2002). But, as these firms become more sophisticated, their share in non-equity-based alliances rises along with their graduation from an industrial structure heavily steeped in mature sectors to one in which they increasingly engage with more high-tech activities to which nonequity forms of collaboration are more conducive (Narula and Sadowski, 2002). The comprehensive picture that emerges is thus one in which the Triad economies account for the bulk of alliances, especially in high-tech fields. Unsurprisingly, on the

Competition policy and innovation

119

whole technology alliances therefore mirror the concentration of R&D spending by the OECD’s major firms in fields such as pharmaceuticals and IT (OECD, 2004). Firms from latecomer countries are fast making inroads, especially in more traditional sectors (Hagedoorn, Link and Vonortas, 2000, p.577; Vonortas and Safioleas, 1997). Cooperative agreements are therefore increasingly present in the global economy. This implies that it is important to understand their relevance for innovative activities, along with the treatment they receive in distinct regulatory regimes and industrial policy contexts in both developed and developing countries. 3 The organization of innovative industries: challenges for competition policy Industrial policy and competition policy: substitutes or complements? The relation of industrial policy to competition policy is old and has had surprising twists and turns. Old-fashioned industrial policy aimed at promoting particular firms (picking winners, creating national champions) and old-fashioned competition policy aimed at providing level playing fields for firms. Thus in principle the two sets of policy could clash. Sometimes they did. At times, however, competition has been used to conduct oldfashioned industrial policy, witness, for example, the American Robinson–Patman Act of 1936, the purpose of which was ‘to provide some measure of protection to small independent retailers and their independent suppliers from what was thought to be unfair competition from vertically integrated, multi-location chain stores’ (Clark, 1995, p.1). In pre-World War II Europe, the absence of a competition policy was an integral part of industrial policy. Cartels were seen as an intelligent way of organizing industries, of avoiding ‘ruinous competition’ and of slowing down innovations, creative destruction and chaotic competition. Prominent economists and courts supported these views (Neumann, 2001, pp. 24–7). In recent times, the débâcle of transatlantic trade relations around the merger of McDonnell Douglas and Boeing provides an example of the conflict between the two policies. The two companies announced the merger in December 1996. The EU Directorate General for Competition voiced its concerns about the merger in May 1997, while the US Federal Trade Commission decided to approve it in July.1 The European Commission decided to go ahead and do an independent investigation of these two American companies. To assuage the EU Commission’s concerns, Boeing sent a fax to Brussels in late July suggesting, among other things, to limit the use of exclusive deal contracts with some of the world’s biggest airlines. This was perceived by the EU to make future competition possible between the world’s two largest producers of airliners, American Boeing and European Airbus, and the merger was subsequently approved.2 However, the view differs on each side of the Atlantic as to whether the EU Commission’s decision was a clear case of industrial policy in favour of Airbus or rather a clear competition decision. Trade interventionists, such as Laura Tyson (1992), might see the competition concerns raised by the European Commission as nothing other than a pretext for getting a better deal for its pan-European company, otherwise why should the FTC have approved the deal? Another explanation is possible, however, following the line of thought of Barros and Cabral (1994). Since Boeing and McDonnell Douglas were large net exporters of airliners to Europe, a merger would have different welfare effects in the USA and in the EU. While the resulting producer surplus could counteract the loss of consumer surplus in the USA, leading the FTC to approve the merger from an American point of view, such

120

International handbook on industrial policy

countervailing producer surplus would not be present in the EU. According to this view the European Commission was protecting its consumers against being exploited; it was not protecting Airbus against competition.3 New developments Recent developments in technologies and industrial organization of innovative and hightech industries have altered the way economists and competition authorities think of cooperation between firms, dominance, mergers and market definition. At the same time, their approach towards state aid has changed considerably, not least in the EU. European state aid control attempts to dismantle old-fashioned industrial policy, especially to the extent that the state aid distorts competition (see Chapter 7 by Martin and Valbonesi in this volume. For a useful brief overview of legislative changes in the EU and the USA since the 1980s, see Hemphill, 2003). In addition, the focus of industrial policy has shifted from promotion of particular firms (picking the winners) to facilitation of industrial clusters or innovation networks (Halliday and Seabright, 2001). Sharing of knowledge, that is, information flows between firms, and cooperation in R&D activities seem to be fundamental in creating the dynamic evolution of the markets: new products for the customers and new, improved, more costefficient processes. While competition policy welcomes the general shift in industrial policy, it struggles to find its feet when evaluating agreements, conduct, dominance, mergers and markets in the ‘new economy’. Industrial organization in the ‘new economy’ Traditional industrial economics and competition analysis focus on price competition and static efficiency. The benchmark is often taken to be that of welfare-maximizing perfect competition that ensures the lowest possible price in that a process of entry competes the price down to average costs in the long run. This means that no firm earns supranormal profits in the long run; that consumers get the goods at the lowest possible unit cost; and that firms are forced to be efficient or else quit the industry. Deviations from the benchmark are perceived as lowering welfare and therefore the target of competition policy. Such deviations could consist of increases of price due to collusion or market dominance. Market dominance in turn could be achieved through mergers, predation and other exclusionary practices (exclusive contracts, tying and bundling). Such conduct, however, is not likely to succeed in creating market power unless accompanied by barriers to entry. In turn barriers to new competition may be the result of either government policies and regulation or the very same types of conduct (vertical restraints, mergers, predation and limit pricing). Herein lies the compromise that has been labelled the ‘new industrial organization’ between the Harvard-based structure–conduct–performance school that focused on barriers to entry that were erected by industry itself and the Chicago school that argued that the only barriers to entry that would survive market forces were those created by government. See Martin (1993, ch.1) for a richer discussion. However, especially in the context of rapid innovation, the ‘new industrial economics’ has come under attack from ‘Austrian’ economists. If the rate of technological advance is fast, static welfare may be of less importance than dynamic welfare achieved through new technology, new products or new processes. In fact, giving up static welfare and price

Competition policy and innovation

121

competition as goals of competition policy may be necessary to get the more important innovations. New products are more important than low prices.4 In the following we focus on industries that are characterized by high rates of product or process innovation, driven by significant research and development costs. In addition, network effects could be important. This definition does not exclude old industries such as the pharmaceutical industry or, arguably, airlines (that certainly have a network element to their services). Economic thinking about innovative or dynamic industries has developed rapidly during the last decade. Industrial economists have increasingly come to realize that, while the static welfare standard (for example, consumer surplus or the sum of consumer and producer surplus) that is used in traditional competition analysis may be adequate when analysing monopolization or cartelization of mature industries, it may be inappropriate when analysing industries characterized by significant innovation. While competition policy for mature industries may focus on reducing price and thus profits for firms in the industry (preferably by reducing entry barriers), it would stifle innovation if imitating rivals eliminated all rewards to innovation. Thus, to the extent that innovation is important in a given industry, competition policy must allow supranormal profits and thus prices in excess of marginal and average costs (Audretsch, Baumol and Burke, 2001). Evidently, this is why intellectual property rights (IPRs) and patent laws are in place, but IPRs represent inefficient ways of protecting the investment in R&D in many industries. In those industries, competition authorities may have to allow behaviour such as horizontal cooperation that would be unacceptable in mature industries.5 The first challenge to competition policy is then to identify how important innovation is in a given industry, that is, whether old-fashioned competition policy suffices or whether new thinking is needed. This is difficult, since the assessment will often have to be done ex ante, before the result of the innovation is known.6 The challenge is really to assess whether innovation is more important than price competition, since this entails deciding whether consumers would be better off with a new product that is priced above cost or with an old product that could be priced at (a lower) cost.7 Network effects represent another challenge for competition policy. Network effects (or demand-side scale economies) mean that the value of the good or service to any customer increases with the number of other customers that also use the good. An often-quoted example is that of the telephone: the more people you can call, the more you will value the good. Less direct network effects may arise through the ‘hardware/software’ externality: the larger the circulation of a particular piece of hardware (a game terminal, for example), the more profitable it is to develop software applications (games) for that type of hardware; and the more applications that are developed, the larger the circulation will be. Thus network effects describe some positive feedback mechanism on the user side that imply that there is a tendency that winners take it all or that markets tip in favour of one producer of, for example, hardware (Shapiro and Varian, 1999). This in turn, implies that such a producer could experience rapidly increasing market shares, once consumer expectations have homed in on a particular hardware (Katz and Shapiro, 1986, 1994). Network effects allegedly mean that consumers may be locked in on the wrong (inferior) technology. An often-quoted example is that of the QWERTY keyboard that allegedly was introduced to slow down typing at the time of the mechanic typewriter so that they would not jam (David, 1985). There were competing keyboard layouts at the

122

International handbook on industrial policy

time of adoption, but in the end QWERTY won because all typists had to learn to type on that particular keyboard because that was the keyboard everybody else was using. Even when an allegedly superior Dvorak keyboard was developed, this superior technology was not adopted because of the switching costs of having to learn to type on a new keyboard. However, Liebowitz and Margolis (1990) contested this ‘fable of the keys’, arguing that the evidence that the Dvorak keyboard was superior to QWERTY is not convincing and that consumer lock-in does not represent a serious obstacle to the adoption of a new technology provided that it is sufficiently superior. Thus there is significant disagreement in the economics literature as to whether network effects and consumer lock-in represent problems that government policies need to resolve. Whether owing to advances in technology or network effects, competition in the industries that we think of here may thus involve rapid changes of market position. Pleatsikas and Teece (2001, p.668) argue that precisely because of this instability of market position, competition authorities should adopt a lenient policy towards high-tech industries. ‘Such intense competition generates high risk and requires higher than average returns to compensate. Yet high risk for the investor should translate to lower risk for the anti-trust authorities, as incumbent positions may be fragile even in the short- to medium-term and, frequently, competitive forces are sufficiently powerful to undo monopoly power, should it arise.’ In the following we organize the exposition around four different tasks that are typically carried out by competition authorities: 1.

2.

3.

4.

Co-operative agreements: horizontal agreements are a ‘no-no’ of traditional competition policy but to the extent that network externalities are important for R&D, output, alliances and joint ventures between competitors may be justified. If R&D collaboration affects price competition, the challenge for competition policy is to strike the right balance between dynamic and static welfare. Market delineation: the definition of the relevant market is a pivotal point of many competition cases. The definition of markets in innovative industries must take into account that competition (to some degree) is carried out along different dimensions, viz. innovation and non-price competition in general rather than price competition. Monopolization and abuse of dominance: dynamic markets are typically also dynamic in terms of market shares. Thanks to network effects, markets may tip so that a firm acquires a dominating position in a short span of time, while its competitors are equally quickly reduced to small fringe firms or driven entirely out of the market. To assess whether such an outcome is the result only of (benign) network effects or (also) of exclusionary conduct is a challenge to competition authorities. Mergers and acquisitions: again usual considerations of assessing market shares of merging firms do not make sense if these market shares may change rapidly, for example owing to network effects.

Cooperative agreements between firms Traditionally, competition authorities have been wary of cooperation between firms. Horizontal cooperation, that is, cooperation between providers of substitutes (competitors), was (and still is) seen as an attempt to restrict quantities and raise prices. Vertical cooperation, that is, cooperation between providers of complements in a supply chain, was

Competition policy and innovation

123

initially treated with the same suspicion, but economic analysis has changed that thinking radically: while parties to a horizontal agreement share an interest in raising each other’s prices, parties to a vertical agreement provide complements and would rather have the other party reduce its price. Thus, typically, vertical cooperation increases the efficiency of the supply chain or the distribution channel and it is only when third parties (consumers, rivals, potential rivals) are hurt that competition authorities feel the need to take action. Firms in innovative industries often seem to be interested in forming networks and alliances. These pose a new challenge to competition policy: more often than not, members of a network supply both complements and substitutes. Consider, for example, airline alliances (such as OneWorld, SkyTeam, Star Alliance or Qualiflyer). Each member has a hub and spokes of regional routes, some of which connect to the hubs of the ‘coopetitor’. Such routes that connect hubs are substitutes and the alliance may result in less competition in terms of both price and frequency of the service. However, by organizing in an alliance the customer gets coordinated access to a bigger set of destinations8 since the set of complements has expanded. In evaluating the alliance (or a corresponding merger), the competition authority must trade off the negative impact of reduced competition around the substitutes against the positive impact of increased choice relating to the complements. Evidently, industrialists and management consultants assure us that ‘Alliances are not even distant cousins of cartels. On the contrary, if properly designed, alliances provide companies with their best chance to compete in the global market place by improving skills, gaining access to new markets, and increasing scale. Alliances can make companies and markets more competitive and give companies a unique opportunity to learn’ (Henzler, 1993, p.265). However, cooperative agreements regarding research and development do worry competition authorities. Their concern is that the agreement will not stop at the R&D level but extend to production and marketing, thus creating a cartel. Thus a traditional view is that R&D collaboration is permissible but firms must behave non-cooperatively in the post-innovation market (Jorde and Teece, 1990, p.81; Martin, 2001, ch.5). For this reason, in both the USA and the EU, R&D collaboration is judged to form a potential problem if the participating firms hold a (jointly) dominating position (Neumann, 2001, p.132). In the USA, the National Cooperative Research Act (NCRA) subjects R&D collaboration to a rule of reason;9,10 in the EU, Commission Regulation 2659/2000 provides a block exemption for R&D agreements from the general prohibition of anti-competitive agreements provided that the combined market share of the firms does not exceed 25 per cent. If not, the R&D agreement is subject to a rule of reason in the sense that it must comply with article 81(3) to be granted individual exemption (European Commission, 2000). Japan takes an even more benign view on joint innovation arrangements including joint commercialization (Jorde and Teece, 1990, pp.87–8). However, even if the firms behave non-cooperatively in the post-innovation market, R&D collaboration may affect prices and welfare adversely. Martin (1995) combines a traditional repeated game explanation of product market collusion with the possibility of R&D joint ventures between the oligopolists. If the R&D joint venture makes static competition more profitable to the participating firms, firms that engage in tacit collusion may share an understanding that deviation from the agreed prices or quantities will mean not only a reversion to the Nash equilibrium of the stage game but also the end of R&D collaboration. Since this threat is harsher than Nash reversion alone, this will make it easier

124

International handbook on industrial policy

for the firms to sustain collusion (they will be able to sustain tacit collusion for a lower value of the discount factor). Jorde and Teece (1990) argue that the view of the innovation process of which Martin’s (1995) model may be an example is linear or sequential: research precedes development that is then commercialized. They argue that increasingly today innovation is rather a simultaneous, continuous interplay between the various stages, and that innovation often requires lateral or horizontal linkages in addition to vertical relations. ‘For innovations to be commercialized, the economic system must somehow assemble all the relevant complementary assets and create an interactive and dynamically efficient system of learning and information exchange’ (Jorde and Teece, 1990, p.79). Jorde and Teece (1990) go on to argue, much in the Austrian tradition that is proposed also by Audretsch, Baumol and Burke (2001), that intellectual property rights in many cases are not strong enough to prevent ‘free-riding’ by imitation or reverse engineering, so that it is no surprise that social returns to innovation are greater than private returns. This leads to underinvestment in innovative activities unless R&D joint ventures, alliances or mergers combining many potential competitors are allowed to ‘bolster its market position and its stream of rents by other strategies and mechanisms. These mechanisms include building, acquiring, or renting (on an exclusive basis) complementary assets and exploiting first-mover advantages. We use the term complementary assets to refer to those assets and capabilities that need to be employed to package new technology so that it is valuable to the end user. Broad categories of complementary assets include complementary technologies, manufacturing, marketing, distribution, sales and service’ (Jorde and Teece, 1990, p.83). Thus Jorde and Teece (1990) argue that competition policy should allow alliances,11 joint ventures and horizontal agreements also in commercialization as well as exclusive practices and other conduct that is normally taken to signify monopolization or abuse of dominance. Shapiro and Willig (1990) take an opposing view, arguing that Jorde and Teece (1990) exaggerate the benefits of relaxing competition policy towards R&D joint ventures and underestimate the costs in terms of reduced competition. Adopting a linear or sequential view on the innovation process, they argue that, if anti-trust rules allow collaboration to extend to production, for example as a production joint venture, the joint venture could set the price so high that the effect would be as if the parent companies engaged in a cartel even if they acted independently in distribution and marketing. Hence it is reasonable that competition policy restricts collaboration to the R&D stage. Similarly, they argue that R&D joint ventures may in fact in and of themselves reduce the pace of innovation by preventing a patent race (Ordover and Willig, 1985). In an analysis of US RJVs formed between the mid-1980s and mid-1990s, Vonortas (2000) finds evidence of multi-project and multi-market contact between the mostly large and diversified firms that undertake RJVs. Insofar as they learn about each other’s strategies, firms then have an incentive to collude and a mechanism to police shirking. On the other hand, the participation of nonUS firms in the RJVs, the uncertainty characteristic of the markets on which these RJVs focus, and the turnover of these inter-firm agreements militate against anti-competitive outcomes. Vonortas suggests that the empirical understanding of the overlap between multi-market and multi-project contacts is incomplete. Shapiro and Willig (1990), as well as Brodley (1990), take an opposite view, arguing that Jorde and Teece (1990) exaggerate the benefits of relaxing competition policy towards

Competition policy and innovation

125

R&D joint ventures and underestimate the costs in terms of reduced competition. They conclude that a rule of reason towards the assessment of R&D collaboration is warranted and that this should be made clear in (American) anti-trust legislation. They also acknowledge that the risk parties to joint ventures face of having to pay treble damages in the American anti-trust system might inhibit benevolent R&D collaboration and thus find that the ‘detrebling’ of damages permitted by the NCRA is justified and that it should be extended to production joint ventures. Market delineation in the context of rapid innovation The definition of the relevant market is the starting point (and often also the end point) of almost all competition cases. Traditionally, competition authorities focus on product markets and (then) on geographic markets. The purpose of market delineation is to identify the goods (or services) that exert a competitive pressure on the product under scrutiny. The most used framework is that of the ‘hypothetical monopolist’ or SSNIP (small, significant non-transitory increase of price): if a hypothetical monopolist could profitably increase the price by a small but significant amount (5–20 per cent) for a certain minimum period of time (a year or two), then the product under consideration constitutes its own market; if not, the analysis is extended to include the closest substitute, and the method continues iteratively until the smallest set of goods that satisfy the ‘hypothetical monopolist’ test is found. Very often, however, competition authorities rely on a less vigorous method of quantitative or qualitative indicators supplemented by anecdotal evidence, such as interchangeability of use, (alleged) sensitivity of revenues to price changes, customer assessments, characteristics of the product, the existence of distinct groups of customers, price differentials and specialized distribution channels (Pleatsikas and Teece, 2001, p.673). Pleatsikas and Teece (2001) argue that these traditional ways of delineating product markets are particularly ill-suited in the context of rapid innovation. Price tests such as SSNIP fall foul of gauging the wrong kind of competition, since non-price competition on the performance of the product is much more important than the often non-existent price competition. If customers care about price at all, it is the price/quality or price/performance relationships they value, not price as such. Similarly, price differentials may exist and persist because of significant product differentiation and patents or other protection of proprietory information. Specialized distribution channels are often used to ensure specialized after-sales service or to supply specific complementary assets. Pleatsikas and Teece (2001) recommend the use of a hedonic framework to assess the price/performance relationship along the lines of Hartman et al. (1993) and Teece and Coleman (1998) or the analysis of market shares over time to assess whether these have the dynamic features expected of rapid innovation. Increasingly, competition authorities and industrial economists have come to realize that market delineation in innovative industries must take into account that competition (to some degree) is carried out along different dimensions, viz. innovation and non-price competition in general rather than price competition. Recognizing that the entry barriers to innovation activities may be very different from entry barriers relating to product markets or geographical markets, US anti-trust authorities have invented the concept of ‘innovation markets’, that is, a separate market for innovation efforts consisting of the R&D (and its close substitutes) aimed at a particular product or process innovation

126

International handbook on industrial policy

(Gilbert and Sunshine, 1995; Glader, 2004, p.6). The underlying thinking seems to be that the innovation market should include the various combinations of assets and scientific personnel that have the technological capability of contributing to R&D in a certain area. Jorde and Teece (1990, p.89) similarly argue that market delineation should be modified to take innovation into account by focusing ‘primarily on the market for know-how’, presumably since it is in this market that barriers to entry arise. Geographically they argue that the market a priori should be taken to include the entire world and that anti-trust authorities should lift the burden of proving otherwise. The usefulness of the concept of innovation markets was discussed extensively in the Antitrust Law Journal’s 1995 symposium on ‘a critical appraisal of the “innovation market” approach’ (for a more complete summary, see Glader, 2004, pp.93–7). Hay (1995) argues that, in the end, it is product markets that are important, not innovation markets and, from a legal point of view, Hoerner (1995) concurs that, since R&D is typically not traded in the market, innovation markets are not defined in terms of lines of commerce as required by the Clayton Act. Rapp (1995, p.12) argues that the innovation market approach ‘represents a leap into the unknown, with a potential for harm to economic welfare as great as any potential benefit’ but Glader (2002, pp.86–7) counters that innovation markets serve to supplement the ‘potential competition doctrine’ that asserts that, if entry into a certain product market is probable and imminent, then the potential entrants exert a competitive pressure on the market that must be included in the market definition. However, since R&D is an uncertain and lengthy endeavour, where ‘R&D is directed to a completely new product for which no product market yet exists at all, the potential competition doctrine may seem misplaced. The innovation market approach was designed to remedy this gap’. Abusive practices and innovation Dynamic markets are typically also dynamic in terms of market shares. Owing to network effects or rapid innovation, markets may tip so that a firm acquires a dominating position in a short span of time, while its competitors are equally quickly reduced to small fringe firms or driven entirely out of the market. Competition is Schumpeterian and for the market, rather than traditional price competition that takes place in the market. The Microsoft cases on both sides of the Atlantic show that the assessment as to whether such an outcome is the result only of (benign) network effects and innovation or of exclusionary conduct is a challenge to competition authorities. Evans and Schmalensee (2002) argue that competition policy should allow for considerable short-run market power and sequential monopolies. Predation is even more difficult to analyse in the context of tipping markets, since all competitors that fail to acquire critical mass will be run out of the market anyway. In addition comparing price with variable costs makes little sense when these are zero for practical purposes since first-copy costs or R&D outlays are the important sunk costs; and how to distinguish penetration pricing from predatory pricing? Similarly, bundling or tying by including extra features in a product is part of the innovation process, and should not be treated as per se illegal. To survey the Microsoft cases and the relevant economic literature on those cases would demand a whole book.12 However, the cases are illustrative of a number of problems relating to innovation and network effects and so deserve brief treatment. In 1998, the US DOJ filed a complaint regarding Microsoft relating to its Internet browser, Internet Explorer (IE), which competed with Netscape Navigator. Two years later the District Court found

Competition policy and innovation

127

Microsoft guilty of maintenance of monopolization in the market for Intel-compatible operating systems (OS); attempted monopolization of the internet browser market; and tying of its Windows with IE. The court ordered Microsoft to split up into two vertically separate entities, one relating to OS and one to applications. However, Microsoft filed an appeal, and the Court of Appeal decided in 2001 that Microsoft was guilty of monopolization of the market for OS, but not of attempted monopolization of the browser market and that alleged tying should be analysed under a rule of reason rather than the lower court’s per se rule. The appeals court argued that network effects constitute a significant barrier to entry that protects Microsoft’s dominant position in the OS market: ‘That barrier – the “applications barrier to entry” – stems from two characteristics of the software market: (1) most consumers prefer operating systems for which a large number of applications have been already written; and (2) most developers prefer to write for operating systems that already have a substantial consumer base. . . . This “chicken-and-egg” situation ensures that applications will continue to be written for the already dominant Windows, which in turn ensures that consumers will continue to prefer it over other operating systems’ (US v. Microsoft, p.20). Thus Microsoft was found to have a monopoly position. In addition, the appeals court found that it was guilty of monopolization through restrictive clauses in the contracts with OEMs (prohibiting them from removing system features relating to IE and from changing the appearance of the desk top), through the integration of IE and Windows and through exclusive contracts with Internet access providers such as AOL. Similarly, in the EU case, Microsoft has been accused of integrating its Windows Media Player (WMP) to the detriment of rival Real Player. The European Commission has requested that Microsoft supply a version of Windows without WMP. When Microsoft complied it first suggested that the new version be called ‘Windows XP Reduced Media Edition’, but the European Commission rejected that as unappealing to customers and the parties later settled for ‘Windows XP Professional Edition N’. Microsoft was also requested to license protocols to software developers that would then allow these to develop software that interoperates well with Windows. Microsoft feels that de-tying and other government interference with their design of the bundled products (Windows/IE/WMP) is restricting its freedom to innovate and has set up a lobbying web page by that name: Microsoft Freedom to Innovate Network.13 In a (vertically) related case, the FTC accused Intel of monopolizing the market for general purpose microprocessors by preventing three computer manufacturers (including Compaq) from enforcing their patents. The alleged misconducts included the termination of information flows needed to incorporate Intel processors in a response to patent infringement litigation, for example. The FTC concluded that, by so doing, Intel expropriated the other firms’ intellectual property rights. Intel countered that the conduct would not affect R&D because of cross-licence agreements with other competitors. The case was settled through a consent agreement that prohibited Intel from denying customers access to such information (Glader, 2004, pp.181–5). Mergers, acquisitions and technological change14 The debate on how market structure and concentration affect R&D (and presumably thus innovation) is an old one in industrial economics. While Schumpeter (1942) advanced the hypothesis that big firms or firms with market power are more likely to shoulder costly

128

International handbook on industrial policy

R&D, Arrow (1962) pointed out that monopolists have a smaller incentive to innovate than competitive firms, since monopolists would cannibalize profits earned with old technologies by introducing new ones, while the competitive firm would reap the entire benefit (provided that it could protect the invention from imitation by other competitive firms). Later theories have revealed that oligopolies might have the biggest incentive to innovate (for example Kamien and Schwartz, 1975) and there has been some empirical support to corroborate this finding (Scherer, 1965). However, in a survey of the evidence, Cohen and Levin (1989) conclude that concentration does not seem to have a significant impact on innovative behaviour and performance: ‘Perhaps the most persistent finding concerning the effect of concentration on R&D intensity is that it depends upon other industry-level variables. . . . The conclusion that market concentration may exercise no independent effect on R&D intensity suggests that there may be no Schumpeterian tradeoff between innovation and the ex ante market power conferred by concentration’ (Cohen and Levin, 1989, pp.1076–8). These findings mean that merger policies that aim at regulating the industry structure find little advice of a general nature in the economics literature and that a case-by-case approach is necessary. Rapp (1995) argues that a policy to block mergers by reference to its effect on innovation must rest on two assumptions: (1) that an increase in (R&D) concentration would diminish the amount of R&D; and (2) that the reduction in R&D would in turn diminish innovation. Carlton and Gertner (2002) add to this the assumption that (3) the amount of R&D and/or innovation is not already supraoptimal. The latter possibility is theoretical (Reinganum, 1989) but is dismissed by Katz and Shelanski (2005, p.135)15 as a general, empirical concern. Katz and Shelanski (2005) point to two criteria for evaluating innovation in relation to merger policy: the innovation incentives criterion and the innovation impact criterion. The innovation incentives criterion is of importance when the merger affects the incentives to innovate (positively or negatively). For example, if innovation cannot be protected from imitation by ‘free-riders’ this may lead to a waiting game that postpones innovation unless the parties are allowed to merge. On the other hand, the merger might end product-market competition. This discussion is akin to our discussion of R&D cooperation above and indeed such R&D cooperation may constitute the relevant benchmark (rather than a non-cooperative scenario) for the merger. The advantage of R&D cooperation is that, at least in theory, the companies may be independent in their marketing behaviour, thus preserving product-market competition. The innovation impact criterion is important when innovation alters post-merger product competition (for example through winner-takes-all competition) so that pre-merger market structure is less useful as an indicator of the effects of competition. In this case post-merger competition is not in the market but for the market, but may be fierce nonetheless. Katz and Shelanski (2005, pp.152–3) conclude that a merger policy for innovationdriven markets should realize that a case-by-case approach (rule of reason) is necessary but that competition authorities could prepare an improved basis for decisions through (1) the development of ‘guidelines on drawing inferences of potential product-market competition from evidence on ongoing innovation’; (2) the consideration of merger cases where the concern is with potential (rather than actual) R&D; (3) the acquisition of industry-specific expertise to assist the assessment of welfare trade-offs resulting from mergers in R&D-intensive sectors; and (4) the systematic use of decision and probability

Competition policy and innovation

129

theory to deal with the uncertainty created by different scenarios for innovation. In essence, only detailed industry studies can shed light on the relationships between innovation and firm behaviour. This is easier said than done and, importantly, must be undertaken over time. If a competition authority blocks a merger because of an alleged reduction in socially optimal investment in R&D, the ex ante situation – likely efficiency benefits from a well structured settlement – should be compared to the discounted benefits of R&D consumers enjoy (or not) thanks to the blocked merger. Hence ex post industry evolution is a good guide to whether or not competition policy decisions turned out to be desirable, even if the learning effects that result from it are unlikely to resuscitate deals that were forbidden but with hindsight should not have been, or vice versa; at the very least, they can inform future decisions (Carlton and Gertner, 2002). 4 Conclusion Perhaps it is easiest to begin our summary of the discussion by emphasizing what it does not mean. First, we do not live in an ‘age of innovation’. Of course, ICT, new materials development, nano- and biotechnology produced technologies that were unimaginable (and much fewer existed) until just a few decades ago. But a higher rate of technological change in some areas exists side by side with very little change in others. Those taking a long view might dismiss claims that ours is an era of exceptional technological development as hyperbolic. This means that not everything tried and tested concerning the efficient operation of markets should go into the rubbish bin of history. Hardcore cartels are alive and kicking, and their insidious pursuit has nothing to do with innovation (OECD, 2000). Hence, in many areas of economic activity, old-style competition policy that guards against and sanctions horizontal agreements is very much needed. Second, although we know that, in theory, when R&D externalities are involved or when technology trading between firms promotes the economy-wide diffusion and adoption of knowledge, the long-run welfare effects of competition in innovation may differ spectacularly with short-run effects of competition around prices, in practice it is impossible to accord these trade-offs any degree of certainty. Disagreements in the economics and legal profession about innovation markets illustrate this poignantly. It is particularly evident when competition authorities haul companies to court and both sides use perfectly valid arguments to claim the rule of reason for their side. The crux is that, not only is the significance of innovation markets unclear in an aggregate sense, but case-by-case considerations are controversial as well. This is despite the fact that by definition they make no pretence of judging welfare implications on anything but the merit of the case at hand. And it lies in the nature of innovation that, even if there were general consensus about the desirability of a horizontal agreement among the firms involved and their competitors and anti-trust officials, some or all of them might have to review their position with hindsight. Third, the inherent tension between the dynamic complexity of future technological developments on the one hand and the need to produce rules and guidelines for firm behaviour that are unambiguous and easy to follow on the other does not mean that competition authorities should stick to what they know and stay away from things uncertain until the course of history and theoretical and empirical advances have improved the lacunae in our understanding of what is going on, and why. Too much is at stake in innovation markets to take the chance that what is best for firms will always necessarily be best for consumers as well.

130

International handbook on industrial policy

These three ‘non-conclusions’ raise a number of issues for public policy and academic research. The first relates to the uncertainty innovation introduces to economic activity. Competition policy and industrial policy are means to an end. In abstract terms the end is relatively clear – a dynamically efficient market – although what that actually means in Silicon Valley or Grenoble where the global technology frontier is being pushed is often different from what it means in Bangalore, Gauteng or Brasilia where most rapid technological change is devoted to catching up. But what all these markets have in common is that market prices do not fully reveal the profitability of resource allocations that have yet to materialize. Firms must discover this cost structure to figure out how and where to compete. In the face of uncertainty, this involves taking bets and hedging against risk through pooling resources with other firms. The likely effects of such cooperation are the stuff of competition policy. Overcoming the information externality, especially when inter-firm agreements are not on the cards, is the stuff of industrial policy (Martin and Scott, 2000; Martin, 2003). Section 2 showed that technical alliances are prevalent in the Triad. It also showed that they are on the rise in select latecomer countries, but they do not as yet have a great impact on the developing world. Where competition policy in developed countries grapples with accommodating the strategic responses of firms to uncertainty, enlightened industrial policy in developing countries is a ‘process of discovery’, involving government and the private sector, of what stands in the way of economic change and industrial upgrading, and how it might best be removed (Rodrik, 2004). A caricature of the difference would be that public policy in the North is primarily concerned with the strategic response to uncertainty by the private sector whereas governments of latecomer countries in the South actively help the private sector to internalize information and coordination problems and thus to share the burden of uncertainty. But the parallels are obvious. The difference between the relationship between competition policy and industrial policy on the one hand and technological innovation on the other is that competition policy is necessarily rule-bound. By contrast, industrial policy can afford to be much more eclectic: what works, works. The story of memory chip production in Korea is different from that in Costa Rica, but also from other experiences of technological upgrading in Korea itself. Hence there is no one-size-fits-all. And thus the second issue refers to what one might call reflexive policy regimes. As bad as it is to make decisions that with hindsight turn out to be wrong, in an uncertain world it is impossible to avoid mistakes. The important point is to learn from them. Competition authorities might gain for themselves and improve on their mandate if, after taking a decision on the merits of a horizontal agreement or after a successful court battle, they continued to monitor the firms concerned in the evolving innovation market. Forecasting what happens in markets a few years or decades from now can only be helped by understanding what went on in these markets in the past. Perhaps the relevant directorates should employ a few economic historians among their analysis staff. Finally, the third issue concerns academic research. Industrial organization has pretty much staked out the theoretical issues. Section 3 has shown what they are and where there is disagreement. The verification of which interpretations of innovation markets are more apposite requires a better empirical understanding of where technological trajectories, especially in the new sectors, are heading. Economists and lawyers are unlikely to figure this out unless they receive help from engineers and other technology specialists.

Competition policy and innovation

131

Notes 1. 2. 3. 4. 5. 6. 7.

8. 9. 10. 11. 12. 13. 14. 15.

Katz and Shelanski (2005, p.145) emphasize that the FTC found that an analysis of the effects of the merger on innovation supported the merger. Technologically, McDonnell Douglas had fallen behind their rivals and the acquisition by Boeing meant that their assets would be put to better use. See Aribaud (1997) for a brief European account of the case. A similar line of argument could be applied to the merger between two other American firms, General Electric and Honeywell, that was prohibited by the European Commission in 2001. See Motta (2004, ch.6.6.1) for a comprehensive discussion of this case. See the special issue of the International Journal of Industrial Organization 19(5) on Competition Policy in Dynamic Markets, guest-edited by David Audretsch, William Baumol and Andrew Burke (2001). For an overview of institutions that provide incentives to innovate, see Scotchmer (2004). ‘There is no government which might be able to outguess the entrepreneurs regarding the potential of the future, and there is no government which, like a benevolent dictator, might be able to pursue economic welfare by adopting non-partisan policies’ (Neumann, 1990, p.563). In the EU, a requirement for an R&D collaboration to be exempt from normal competition rules is that it relates ‘to results which are protected by intellectual property rights or constitute know-how, which substantially contribute to technical or economic progress and the results must be decisive for the manufacture of the contract products or the application of the contract process’ (European Commission, 2000, art. 3(4)). For background on EU competition policy with respect to technology transfer and especially licensing agreements, see Peeperkorn, 2003). This may be perceived as a product innovation! A rule of reason allows the court to assess whether a certain business conduct promotes or harms competition on a case by case basis. It is the opposite of a per se rule according to which the conduct is considered harmful without further analysis. The regulation has been liberalized through the National Cooperative Production Amendments of 1990 (H.R. 4611); see Jorde and Teece (1990, p.90). Jorde and Teece (1990, p.85) further argue that ‘The case for planning and industrial policy recedes if a degree of operational and strategic coordination can be attained through private agreements.’ For a very good, brief summary of the US v. Microsoft case, see Motta (2004, ch.7.5). http://www.microsoft.com/freedomtoinnovate/. This section just introduces this area in broad terms. For a more detailed discussion, see Chapter 8 of this volume on Mergers and Concentration Policy by Hans Schenk, and Katz and Shelanski (2005) that also contains an interesting overview of US merger cases. Referring to Griliches (1992) and Jones and Williams (1998).

References Aribaud, J-L. (1997), ‘Summary of the most important recent developments’, EC Competition Policy Newsletter, 3(2). Arrow, K. (1962), ‘Economic welfare and the allocation of resources for invention’, in R.R. Nelson (ed.), The Rate and Direction of Inventive Activity, New York: Princeton University Press. Audretsch, D.B., W.J. Baumol and A.E. Burke (2001), ‘Competition policy in dynamic markets’, International Journal of Industrial Organization, 19(5), 613–34. Barros, P.P. and L. Cabral (1994), ‘Merger policy in open economies’, European Economic Review, 38(5), 1041–55. Brodley, J.F. (1990), ‘Antitrust law and innovation cooperation’, Journal of Economic Perspectives, 4(3), 97–112. Caloghirou, Y., S. Ioannides and N.S. Vonortas (2003), ‘Research joint ventures’, Journal of Economic Surveys, 17(4), 541–70. Carlton, D.W. and R.H. Gertner (2002), ‘Intellectual property, antitrust, and strategic behaviour’, NBER Working Paper 8976, National Bureau of Economic Research. Chen, H. and T-J. Chen (2002), ‘Asymmetric strategic alliances: a network view’, Journal of Business Research, 55, 1007–13. Clark, D.S. (1995), ‘The Robinson–Patman Act: general principles, Commission proceedings, and elected issues’, Speech at Red Lion Inn, San Jose, California, 7 June (http://www.ftc.gov/speeches/other/patman.htm). Cohen, W.M. and R.C. Levin (1989), ‘Empirical studies of innovation and market structure’, in Richard Schmalensee and Robert Willig (eds), Handbook of Industrial Organization, vol. 2. Amsterdam: NorthHolland. David, P.A. (1985), ‘Clio and the Economics of QWERTY’, American Economic Review, 75(2), 332–7. Dunning, J.H. (1997), Alliance Capitalism and Global Business, London: Routledge. European Commission (2000), ‘Commission Regulation (EC) No. 2659/2000 of 29 November 2000 on the application of Article 81(3) of the Treaty to categories of research and development agreements’, Official Journal, L304, 7–12.

132

International handbook on industrial policy

European Commission (2003), Third European Report on Science & Technology Indicators: Towards a Knowledge-Based Economy, Brussels: European Commission. Evans, D.S. and R. Schmalensee (2002), ‘Some economic aspects of antitrust analysis in dynamically competitive industries’, in A.B. Jaffe, J. Lerner and S. Stern (eds), Innovation Policy and the Economy, Cambridge, MA: MIT Press. Gilbert, R. and S. Sunshine (1995), ‘Incorporating dynamic efficiency concerns in merger analysis: the use of innovation markets’, Antitrust Law Journal, 63, 569–601. Glader, M. (2002), ‘Innovation competition and antitrust analysis’, Faculty of Law, Lund University. Glader, M. (2004), ‘Innovation markets and competition analysis – EU competition law and US antitrust law’, Faculty of Law, Lund University. Griliches, Z. (1992), ‘The search for R&D spillovers’, Scandinavian Journal of Economics, 94, 29–47. Hagedoorn, J., A.N. Link and N.S. Vonortas (2000), ‘Research partnerships’, Research Policy, 29, 567–86. Halliday, J. and P. Seabright (2001), ‘Networks good, cartels bad: but how could anyone tell the difference?’, in Swedish Competition Authority (ed.), Fighting cartels – why and how?, Stockholm: Elander Graphics Systems. Hartman, R., D. Teece, W. Mitchell and T. Jorde (1993), ‘Assessing market power in regimes of rapid technological change’, Industrial and Corporate Change, 2, 316–50. Hay, G.A. (1995), ‘Innovations in antitrust enforcement’, Antitrust Law Journal, 64, 7–17. Hemphill, T.A. (2003), ‘Cooperative strategy, technological innovation and competition policy in the United States and the European Union’, Technology Analysis & Strategic Management, 15(1), 93–101. Hemphill, T.A. and N.S. Vonortas (2003), ‘Strategic research partnerships: a managerial perspective’, Technology Analysis & Strategic Management, 15(2), 255–71. Henzler, H.A. (1993), ‘Alliances in Europe: collusion or cooperation’, in J. Bleeke and D. Ernst (eds), Collaborating to Compete – Using strategic alliances and acquisitions in the global market place, New York: Wiley. Hobday, M. (1995), ‘East Asian latecomer firms: learning the technology of electronics’, World Development, 23(7), 1171–93. Hoerner, R.J. (1995), ‘Innovation markets: new wine in old bottles?’, Antitrust Law Journal, 64, 70–73. Jones, C.I. and J.C. Williams (1998), ‘Measuring the social return to R&D’, Quarterly Journal of Economics, 113, 1119–35. Jorde, T.M. and D.J. Teece (1990), ‘Innovation and cooperation: implications for competition and antitrust’, Journal of Economic Perspectives, 4(3), 75–96. Kamien, M.I. and N.L. Schwartz (1975), ‘Market structure and innovation: a survey’, Journal of Economic Literature, 13, 1–37. Katz, M. and C. Shapiro (1986), ‘Technology adoption in the presence of network externalities’, Journal of Political Economy, 94(4), 822–41. Katz, M. and C. Shapiro (1994), ‘System competition and network effects’, Journal of Economic Perspectives, 8(2), 93–115. Katz, M.L. and H.A. Shelanski (2005), ‘Merger policy and innovation: must enforcement change to account for technological change?’, in A.B. Jaffe, J. Lerner and S. Stern (eds), Innovation Policy and the Economy, Cambridge, MA: MIT Press. Kuada, J. (2002), ‘Collaboration between developed and developing country-based firms: Danish–Ghanaian experience’, Journal of Business & Industrial Marketing, 17(6), 538–57. Liebowitz, S. and S.E. Margolis (1990), ‘The fable of the keys’, Journal of Law & Economics, 33, 1–25. Martin, S. (1993), Advanced Industrial Economics, Oxford: Blackwell. Martin, S. (1995), ‘R&D joint ventures and tacit product market collusion’, European Journal of Political Economy, 11, 733–41. Martin, S. (2001), Industrial Organization: A European Perspective, Oxford: Oxford University Press. Martin, S. (2003), ‘The evaluation of strategic research partnerships’, Technology Analysis & Strategic Management, 15(2), 159–76. Martin, S. and J.T. Scott (2000), ‘The nature of innovation market failure and the design of public support for private innovation’, Research Policy, 29, 437–47. Motta, M. (2004), Competition Policy – Theory and Practice, Cambridge: Cambridge University Press. Narula, R. (2003), Globalization & Technology: Interdependence, Innovation Systems and Industrial Policy, Cambridge: Polity Press. Narula, R. and B.M. Sadowski (2002), ‘Technological catch-up and strategic technology partnering in developing countries’, International Journal of Technology Management, 23(6), 599–617. National Science Foundation (2004), Science and Engineering Indicators 2004, two vols, Arlington, VA: NSF. Neumann, M. (1990), ‘Industrial policy and competition policy’, European Economic Review, 34, 562–7. Neumann, M. (2001), Competition Policy – History, Theory and Practice, Cheltenham, UK and Northampton, MA, USA: Edward Elgar. OECD (2000), Hard Core Cartels, Paris: OECD. OECD (2004), Main Science and Technology Indicators, Paris: OECD.

Competition policy and innovation

133

Ordover, J. and R. Willig (1985), ‘Antitrust for high-technology industries: assessing research joint ventures and mergers’, Journal of Law and Economics. Peeperkorn, L. (2003), ‘P Licenses and Competition Rules: Striking the Right Balance’, World Competition, 26(4), 527–39. Pleatsikas, C. and D. Teece (2001), ‘The analysis of market definition and market power in the context of rapid innovation’, International Journal of Industrial Organization, 19, 665–93. Rapp, R.T. (1995), ‘The misapplication of the innovative market approach to merger analysis’, Antitrust Law Journal, 64, 19–46. Reinganum, J. (1989), ‘The timing of innovation: research, development and diffusion’, in R. Schmalensee and R. Willig (eds), Handbook of Industrial Organization, vol. 1, Amsterdam: North-Holland. Rodrik, D. (2004) ‘Industrial policy for the twenty-first century’, Research Working Paper 04–047, John F. Kennedy School of Government, Harvard University, Cambridge. Scherer, F.M. (1965), ‘Firm size, market structure, opportunity, and the output of patented inventions’, American Economic Review, 57, 1097–1125. Schumpeter, J.A. (1942), Capitalism, socialism, and democracy, New York: Harper & Row. Scotchmer, S. (2004), Innovation and Incentives, Cambridge, MA: MIT Press. Shapiro, C. and H.R. Varian (1999), Information Rules – a strategic guide to the network economy, Boston, MA: Harvard Business School Press. Shapiro, C. and R.D. Willig (1990), ‘On the antitrust treatment of production joint ventures’, Journal of Economic Perspectives, 4(3), 113–30. Teece, D.J. (1992), ‘Competition, cooperation, and innovation: organizational arrangements for regimes of rapid technological progress’, Journal of Economic Behavior and Organization, 18(1), 1–25. Teece, D.J. and M. Coleman (1998), ‘The meaning of monopoly: antitrust analysis in high technology industries’, Antitrust Bulletin, 43, 801–57. Tyson, L. (1992), Who’s bashing whom? Trade conflict in high-technology industries, Washington, DC: Institute for International Economics. Vonortas, N.S. (2000), ‘Multimarket contact and inter-firm cooperation’, Journal of Evolutionary Economics, 10, 243–71. Vonortas, N.S. and S.P. Safioleas (1997), ‘Strategic alliances in information technology and developing country firms: recent evidence’, World Development, 25(5), 657–80.

7

State aid to business Stephen Martin and Paola Valbonesi

A group of blind men encountered an elephant, an animal hitherto unknown to them, and sought to understand what manner of creature it was. One touched its tusk, and found it like a spear. One touched its side, and found it like a wall. One touched its trunk, and found it like a snake. One touched a leg, and found it like a tree. One touched its ear, and found it like a fan. The last touched its tail, and found it like a rope. They went on their way, but fell into dispute about the nature of the elephant, each insisting upon the aspect with which he had come in contact, none comprehending the whole. (Indian parable)

1 Introduction The views of economists and policy makers on state aid to business have much in common with the way the blind men of the fable viewed the elephant. State aid is a subject of at least three primarily academic literatures and at least two primarily policy-oriented literatures. The three academic literatures are those on strategic trade policy (section 2), on tax competition (section 3), and on rent seeking (section 4). The two policy literatures are a predominantly legal literature on the treatment of state action under US anti-trust policy (section 5) and a corresponding literature on EU state aid (section 6), which involves both the treatment of state aid under EU competition policy and the impact of state aid on regional convergence.1 Each of these literatures has implications for the others, but these implications are drawn, if at all, only on the margins. Contributions to each literature occasionally acknowledge the existence of the others, but each literature tends to focus mainly on one specific aspect of state aid; none sees the whole elephant. 2 Strategic trade policy The New International Economics (which is no longer new) emphasizes product differentiation and economies of scale as determinants of imports and exports. Competition in markets where products are differentiated and the efficient scale of production is large in relation to market size will of necessity be imperfect. The strategic trade policy literature,2 originating with Brander and Spencer (1985), takes off from the observation that, if competition is imperfect, government action might under some circumstances alter market equilibrium to the benefit of home-country firms. This initial analysis of strategic trade policy considers an international quantity-setting duopoly, with each of two firms based in a different country, selling to consumers located in a third country. The approach is partial equilibrium, eliminates any welfare impact on home-country consumers by assumption, and assumes the two firms behave non-cooperatively (that is, do not collude).3 The case for strategic trade subsidies that comes out of even the seminal model is not strong: by using strategic subsidies, the terms of trade move against the subsidizing country, but its welfare can improve because, with imperfect competition, price exceeds the marginal cost of exports. However, if the other country adopts a similar policy, both are worse off (Brander and Spencer, 1985, pp.95–6).4 134

State aid to business 135 The literature that extends the seminal model does not improve the view. If the product is differentiated and firms set prices instead of quantities, the welfare-improving policy for a single strategically motivated country is to tax its home firm, not to subsidize it. Whether firms set prices or quantities, the assumption that domestic consumers do not participate in the target sector is, in general, false. Consideration of the welfare impact of strategic trade policy on domestic consumers renders the overall welfare effects ambiguous even if only the direct effects of export subsidies on the domestic market for the subsidized good are taken into account.5 There may be indirect effects of export subsidies on domestic consumer welfare. Unless an export-oriented industry uses only inputs that are in excess supply, policy that promotes exports will cause an exporting sector to bid resources away from other segments of the economy. Domestic consumers of those other sectors will suffer indirect welfare losses. In practice, the question that a government contemplating implementation of a strategic trade policy will face will not be whether it should or should not promote exports of a particular industry, but which among several exporting industries, if any, it should promote.6 There is no particular reason to expect that government will be able to pick winners.7 Not only does bilateral trade promotion trap countries in a welfare-reducing prisoners’ dilemma outcome, but the WTO Agreement on Subsidies and Countervailing Measures8 provides for appeals to the WTO that may (a) result in the declaration that unilateral subsidies are illegal and (b) permit countervailing measures unless and until they are removed. This is another reason to think that the implications of the model that by assumption limits subsidies to one country are not likely to be robust. Goldstein and McGuire (2004) document the long history of a Brazil–Canada subsidy battle in support of their respective domestic regional jet producers. The conflict includes state ownership, privatization, production and export subsidies, public support for military technology that generates applications to civil aircraft, and WTO condemnations all around. Neither theory nor case study evidence make a strong argument for the position that strategic trade policy is likely to be welfare-improving. 3 Tax competition9 Theory Oates and Schwab (1991, p.127) succinctly summarize the polar positions in the literature on tax competition: Proponents would have us believe that competition offers both protection from self-seeking bureaucrats and assurances that public goods will be provided efficiently; opponents tell us that competition necessarily leads to inadequate local government budgets, regressive local taxes, and insufficient local taxes for the poor.

The Tiebout (1956) model is one of an economy with (among other assumptions) (a) fully-informed mobile consumers able to choose among a large number of communities, (b) communities that offer different packages of local public services, (c) public services that are not subject to external spillovers. The model has consumers ‘vote with their feet’, moving to the community that offers their preferred bundle of public services and tax rate. Local jurisdictions adjust tax rates to minimize the average cost of providing public services to residents. In equilibrium, the bundles offered are suited to consumers’ preferences for public goods.

136

International handbook on industrial policy

With Oates and Schwab (1988, 1991), in contrast to Tiebout (1956), it is capital rather than labour that is mobile. But the basic model has strong efficiency results (Oates and Schwab 1991, p.140): ●

●

●

First, if communities compete against one another to attract new business investment and jobs, then local taxes are equivalent to user fees. Second, . . . localities in a competitive setting are unable to redistribute capital income through taxation because taxes are nothing more than payments for goods and services received. . . . Third, interjurisdictional competition fosters efficiency.

As Oates and Schwab acknowledge (1991, pp.134–8), the results of the basic model fail if the number of jurisdictions is small, if there are externalities across jurisdictions in the production of public goods, or if information about the benefit/tax policies of different jurisdictions is imperfect. McGuire (1991) argues, in the US context, that the policy implications of the tax competition literature are different for competition among states and for competition among local jurisdictions. The assumption that tax rivalry is among many jurisdictions is much less likely to hold, even approximately, at the state than at the local level.10 Confidence in the realization of the efficiency of outcomes resulting from tax competition among states that is predicted by some models is correspondingly reduced. Janeba (1998) obtains strong efficiency results in a model of tax competition and mobile capital, under the assumption that jurisdictions cannot discriminate in taxes across firms. He writes (1998, pp.150–51): ‘International agreements like GATT or the European Union often establish rules of nondiscrimination. Such agreements are mutually beneficial since deviation leads to a wasteful subsidy race which is worse than nondiscrimination.’ But the mere fact that an agreement is mutually beneficial does not mean that it is a non-cooperative equilibrium. The history of state aid in the European Union is rife with examples of member state aid knowingly granted in violation of state aid rules (see section 6).11 The relevance of the Tiebout ‘invisible foot’ model of tax competition seems dubious at best for the European Union, where, if only for reasons of culture and language, labour is much less mobile than in the United States. It is also fair to note that much state aid in the EU is driven by the sunkenness of capital – capital immobility – in integrating markets and the reluctance of member state governments passively to accept that market integration implies a reduction in the equilibrium number of firms (Martin and Valbonesi, 2005). Sinn’s conclusion about the impact of tax competition in the EU is not an optimistic one (1990, p.501): The losers of tax competition will be those who cannot escape and those who benefit from a large government sector. The first group includes immobile workers and landowners. They are the natural victims of the Tiebout equilibrium, since they will serve as the lenders of last resort to Europe’s impoverished governments. The second group consists of the poor. The poor will lose because governments will no longer be able to maintain their current scales of redistribution.

The controversy over ‘delocalization’ that emerged following the May 2004 accession of ten new EU member states lends credibility to Sinn’s analysis. The delocalization controversy is part and parcel of the ongoing debates over fiscal harmonization and social

State aid to business 137 dumping. These debates do no more than translate the polar positions identified by Oates and Schwab (1991, p.127, quoted above) to the European context. Glaeser’s (2001) topic is location-based tax incentives, which, as he notes (p.1), ‘can be seen as tax rates that are chosen on a firm-by-firm basis’. The analysis of location-based tax incentives thus concerns tax competition when tax discrimination is possible. Glaeser reviews five reasons why local governments might offer tax incentives (pp.2–8): to increase local producer or consumer surplus; to realize agglomeration economies; to compensate firms for expected future tax payments; to offer different tax rates to marginal and inframarginal firms; and in return for personal benefits corruptly paid to local officials. Only the last motive, Glaeser concludes (p.10), gives reason to think that tax incentives will distort firms’ location decisions. Discriminatory tax incentives may also lead to the kind of Prisoners’ Dilemma equilibrium that is generally highlighted in the tax competition literature (p.11):12 The mobility of firms certainly stops some localities from redistributing to the poor. However, the mobility of the rich also stems the ability of localities to redistribute. More to the point, I think that mobility generally means that local redistribution is almost always a bad idea. Tax incentives may certainly limit the ability to engage in local redistribution, but probably that local redistribution should never have gone on in the first place.

Glaeser (2001, p.10) makes an analogy between tax discrimination by a local government and price discrimination by a monopolist. The latter increases net social welfare by eliminating output restriction and the associated deadweight welfare loss. To the extent that tax discrimination balances out other distortions (for example, allows the realization of agglomeration economies) it too improves welfare. But the traditional policy hostility towards price discrimination is not based on the impact of price discrimination on market performance as measured by net social welfare. Policy makers have objected to price discrimination on the ground that price discrimination by an upstream input supplier distorts competition between downstream input users.13 Tax discrimination, as opposed to non-discriminatory tax competition, may similarly distort competition between firms at the same horizontal level, some of which benefit from discriminatorily low tax rates while others do not.14 Evidence The empirical evidence on the effect of tax competition on business behaviour is abundant and its results are ambiguous.15 One problem in assessing this effect stems from the multiple forms of government intervention that contain implicit or explicit subsidy elements.16 Huttin finds a limited impact of a temporary reduction in social employment taxes for firms in the French textile industry (1989, p.489): it had very little effect on employment as a whole, and only some positive effects on investments. . . . The poor efficiency of aid on employment results mainly [from] the fact that a two-year cut in social charges represents a low reduction of the total unit labour cost for hiring an employee (on his average stay in the company). On the investment side, the application of the Plan operated through the liquidity constraints on managers’ decisions, and the positive effect of reducing

138

International handbook on industrial policy

it to increase investment. It appears mainly efficient for companies where the labour cost is a big share of total cost.

Phillips and Goss (1995) carry out a meta-analysis of estimates of tax rate sensitivities surveyed by Bartik (1992). Their estimates of the elasticity of local economic development17 with respect to the tax rate range from 0.216 to 0.346 for studies across states or metropolitan areas, and was much higher (1.25) for studies of jurisdictions within the same metropolitan area. They describe their findings in these terms (1995, p.327): ‘The estimate is at the lower end of the range suggested by Bartik (1992) and generally confirms his view that taxes have a significant but modest impact on economic development.’18 Greenstone and Moretti (2004) compare trends in property values in winning and alsoran countries that competed for the same ‘big ticket’ investment project. They find (p.28) that ‘The results suggest that the successful attraction of a “Million Dollar” plant is on average associated with an increase in property values. Under the assumptions [of the model], this increase in property values may be interpreted as an increase in welfare for the residents of the winner country.’ As they note, this finding does not address the welfare consequences of tax competition at the state or national level. Empirical studies of tax policy in the USA thus find some evidence of a modest ability of lower taxes to shift investment, and some evidence that successful shifting of investment increases local property values. These results are quite consistent with a ‘beggar-my-neighbour’ interpretation of tax competition and selective incentives policies that reduce overall welfare. Empirical studies of the investment behaviour by US firms in foreign markets give some support to the idea that such decisions are influenced by tax differences (among many other factors). Grubert and Mutti (1991, p.290), for example, find that ‘a reduction in the host country tax rate from 20% to 10% is projected to increase U.S. affiliates’ net plant and equipment in the country by 65%’. They note that this response may reflect either the shifting of investment from the USA or from a higher-tax rate foreign market to the lower-tax rate host country. One might expect similar behaviour within the USA, and such a conclusion is explicitly drawn by Hines (1999) who summarizes the results of studies of international business behaviour as indicating (p.319) ‘that investment location and tax avoidance activity are more responsive to tax rate differences than is typically implied by domestic evidence’ and concludes that ‘states offering attractive tax climates will be able to draw business activity away from other parts of the United States. . . . [this] may spur a round of competitive tax reductions at the state level’. 4 Rent seeking Theory There is limited evidence that selective incentives are effective in shifting business investment decisions, scant evidence that they may increase local welfare, and none suggesting that they increase global welfare. One may wonder, then, why tax competition seems to be an enduring feature of the political landscape.19 The answer may lie in rent seeking.20 The literature on rent seeking, the investment of private resources with the object of obtaining privately or publicly sustained market power, can be traced to Tullock (1967).21 The term ‘rent seeking’ is due to Krueger (1974).22 In their work, Tullock and Krueger

State aid to business 139 largely analysed government enforcement of positions of market power in international markets, as did Bhagwati (1982). But the concept of rent seeking quickly became caught up in the debate over the implausibly small estimates of the welfare cost of private exercise of market power for the USA put forward by Harberger (1954). Posner (1975), making the assumptions that there are constant returns to scale in the process of obtaining monopolies, that the activity of obtaining a monopoly is itself competitive, and that the process of obtaining monopolies yielded no socially valuable by-products, presented the result that rent seeking to obtain monopoly would result in complete dissipation of the expected economic profit.23 The case of complete dissipation remains a focal point of this literature, partly as a matter of simplicity: in the case of complete dissipation (Hillman and Samet, 1987, p.63) ‘the social cost associated with contestability of a rent can be inferred from the value of the rent itself, and the detailed and hard-to-come-by information on individual outlays made in the course of the contest becomes unnecessary’. But the subsequent theoretical work has made clear that complete dissipation is by no means a necessary equilibrium result, and experimental results suggest the same view. Baye et al. (1989, 1993) show that a rent-seeking contest for a government grant of monopoly can be modelled as an all-pay auction: all contenders pay (bid) the expense of lobbying decision makers, but only one wins the prize. In the Baye et al. model, dissipation of economic profit in rent-seeking expenses is complete if all contenders have a common value for the prize. If contenders have different values for the prize, dissipation is less than complete. Empirical Two formal specifications of rent seeking appear in the theoretical literature. The lottery specification (Tullock, 1980) makes the probability that a particular contestant receives the prize a positive function of the amount invested in rent seeking by that contestant and a negative function of the amounts invested in rent seeking by other contestants. In the all-pay auction formulation, the contestant who pays the largest absolute amount obtains the prize. Davis and Reilly (1998) report the results of experimental tests of both specifications of rent-seeking behaviour. They find that overdissipation is ‘pervasive’ (pp.110–111): ‘Collectively, the agents tend to dissipate more rents than Nash equilibrium predictions in all auctions – an outcome that diminishes, but does not disappear with experience.’ Anderson and Stafford (2003) report the results of experimental tests of the lottery rent-seeking model, allowing for cost differences across contestants and entry fees. They find (p.208) that ‘mean group expenditure exceeded the mean predicted group expenditure for all treatments in the experiment. Moreover, in the majority of treatments, mean group expenditure exceeded the value of the prize (that is, overdissipation of rents)’. Hazlett and Michaels (1993) analyse the extent of rent dissipation in the US cell phone licence lotteries held between 1986 and 1989. Their results suggest that about one-third of expected economic profit from licences for metropolitan areas was dissipated in rent seeking, and about one-fifth for licences for rural areas. Brook (2005) discusses an extended, and ultimately successful, lobbying campaign by the United Steelworkers of America and leading US steel producers for protection from foreign competition.24 The campaign was financed by a fund that was expected to amount to $10 million, put together by a 5¢ per ton ‘tax’ on steel production.

140

International handbook on industrial policy

Systematic statistics are lacking about lobbying efforts at the US state level. The National Association of State Development Agencies reports that the operating budgets of US state economic development agencies totalled $6.3 billion in 2001, nine tenths of 1 per cent of US GDP. If one takes the complete dissipation case as a focal point, this figure suggests rent-seeking expenditures of the same amount. Under the provisions of the US Lobbying Disclosure Act, lobbyists at the US national level are required to file regular reports indicating who they work for, their income and their expenditures.25 A public interest group makes this data available over the internet (http://www.opensecrets.org/lobbyists/index.asp). Table 7.1 summarizes annual national lobbying expenditures for the period 1997–2000. The 2000 total, $1.5 billion, was about one-quarter of 1 per cent of US gross domestic product. The European Union does not, at the time of writing, mandate the reporting or registering of lobbying activities.26 Over 10 000 lobbyists are estimated to be currently active in Brussels, of which only 141 are recorded in the Society of European Professional Affairs. In March 2005, the Commissioner for Administrative Affairs, Audit and AntiFraud, Siim Kallas, presented an initiative to introduce mandatory transparency rules on lobbying at EU institutions.27 This proposal was immediately rejected by the lobbying profession, in favour of the existing voluntary code. This code attracts only a few registrants and provides no external transparency, all of which suggests that there is much to be said in favour of mandatory lobbying transparency rules at EU institutions. Rent-seeking activity by firms in the EU is assessed by Coen (1997) through a study of 94 large European firms. He analyses different firms’ political preferences in three political areas and concludes that firms ‘are attempting to play a complex political game within the confines of their budget constraint’. This game involves collective lobbying based on strategic alliances with rival firms (for horizontal policy making at EU level) and direct

Table 7.1

Spending lobbying the US federal government($ million)

Sector Agribusiness Communications/Electronics Construction Defence Energy & Natural Resources Finance, Insurance & Real Estate Health Lawyers & Lobbyists Misc. Business Transportation Ideological/Single-issue Labour Other Total

1997

1998

1999

2000

86 154 17 49 143 177 163 13 150 112 73 21 66

119 186 22 49 149 203 165 19 168 115 76 24 69

83 193 24 53 158 214 197 18 193 117 76 24 87

78 204 23 60 159 229 209 16 224 138 85 27 103

1224

1364

1437

1555

Source: Center for Responsive Politics (http://www.opensecrets.org/lobbyists/index.asp).

State aid to business 141 lobbying (for company-specific goals such as state aid and R&D funding); these activities may be mutually reinforcing. The way in which the various options for both collective and direct lobbying are developed is influenced by cost considerations. 5 US policy towards state aid The equivocal evidence about the effects of state and local tax and subsidy competition even from the point of view of a single jurisdiction and the Prisoners’ Dilemma outcome that seems likely to result from unfettered state rivalry has prompted calls for federal regulation.28 There are two lines of development of US law that seem to come into play when federal regulation of state aid is discussed. From an economic point of view, neither is entirely to the point. The US anti-trust state action doctrine California’s 1933 Prorate Act was the basis for a 1940 programme under which the State Agricultural Prorate Advisory Commission administered mandatory limits on the quantities of raisins sold. Brown, a discontented raisin grower, challenged the prorate programme as an unconstitutional interference with inter-state commerce. He prevailed in District Court (Brown v. Parker 39 F. Supp. 895 (1941)), and Parker, the State Director of Agriculture, appealed this decision to the US Supreme Court. In a decision that is generally regarded as the origin of the state action doctrine,29 the Supreme Court reversed the District Court decision on the ground that the Sherman Act had never been intended to apply to the states (317 U.S. 341 at 351): In a dual system of government in which, under the Constitution, the states are sovereign, save only as Congress may constitutionally subtract from their authority, an unexpressed purpose to nullify a state’s control over its officers and agents is not lightly to be attributed to Congress. The Sherman Act makes no mention of the state as such, and gives no hint that it was intended to restrain state action or official action directed by a state. The Act is applicable to ‘persons’ including corporations . . . There is no suggestion of a purpose to restrain state action in the Act’s legislative history.

The Parker v. Brown opinion at least suggests that Congress might adopt laws prohibiting state actions that distort inter-state commerce (317 U.S. 341 at 350): We may assume for present purposes that the California prorate program would violate the Sherman Act if it were organized and made effective solely by virtue of a contract, combination or conspiracy of private persons, individual or corporate. We may assume also, without deciding, that Congress could, in the exercise of its commerce power, prohibit a state from maintaining a stabilization program like the present because of its effect on interstate commerce. Occupation of a legislative ‘field’ by Congress in the exercise of a granted power is a familiar example of its constitutional power to suspend state laws.

A follow-on literature debates the proper scope of the state action doctrine. When contributions to this literature discuss the types of state policy that would be affected by reducing the scope of the state action doctrine, it is invariably state regulation of business, in the broadest sense, that is discussed: public utility regulation, licensing procedures, and the like.30 This literature does not seem to contemplate that the state action doctrine protects decisions a state takes in taxing and subsidizing business. By implication, narrowing the scope of the state action doctrine would limit the ability of a state to direct firms to

142

International handbook on industrial policy

act in ways that would violate the Sherman Act if the actions were the result of joint decisions by firms. It would not necessarily reach tax and subsidy decisions of states that distort inter-state commerce. The commerce clause31 Congressional authority over inter-state commerce forbids states from employing taxes that discriminate against inter-state commerce.32 In this context, ‘discrimination against interstate commerce’ has involved state taxes that disadvantage out-of-state competitors of in-state firms. For example (Hellerstein and Coenen, 1996, p.836):33 West Lynn Creamery arose out of an effort by Massachusetts to aid its struggling dairy industry. To this end, the state imposed a tax on milk dealers for all in-state sales of milk, whether or not the milk had been produced in Massachusetts. The state then placed all tax proceeds in a segregated fund and distributed the fund exclusively to operators of in-state dairy farms.

The distortion in inter-state commerce that arises from discriminatory tax competition and selective economic development subsidies, however, does not favour in-state firms with respect to out-of-state firms. Rather, the discrimination is in favour of the firm that receives the benefit (whether that firm has an initial presence in the granting state or not). The lower cost enjoyed by the receiving firm distorts rivalry between it and in-state rivals that do not receive the benefit. To the extent that all such firms are involved in inter-state commerce, the latter is distorted.34 The tax element of the tax-and-subsidy scheme at issue in West Lynn Creamery was critical to the result (512 U.S. 186 at 199):35, 36 A pure subsidy funded out of general revenue ordinarily imposes no burden on interstate commerce, but merely assists local business. The pricing order in this case, however, is funded principally from taxes on the sale of milk produced in other States. By so funding the subsidy, respondent not only assists local farmers, but burdens interstate commerce. The pricing order thus violates the cardinal principle that a State may not ‘benefit in-state economic interests by burdening out-of-state competitors’.

Hellerstein and Coenen (1996, p.847) make a distinction between taxes and subsidies: 1. 2. 3.

Subsidies, which involve an actual disbursement of public funds, will be subject to greater scrutiny than tax incentives. Tax discrimination is a long-standing concern of U.S. law; the same cannot be said for discriminatory subsidies. The Commerce Clause itself was motivated by concerns over discriminatory taxation, in the form of interstate tariffs that multiplied under the Articles of Confederation.

Hellerstein and Coenen (1996, p.849) also distinguish between general subsidy schemes and selective, one-shot subsidies. Although they acknowledge that discriminatory tax measures and subsidies are (p.868) ‘often indistinguishable’ in their economic effects, they nonetheless take the view that the negative Commerce Clause should not be applied to state subsidies in the way that it is to state tax measures. In stating this view, they emphasize that the form by which subsidy benefits are delivered differs from the form of tax benefits, that form is a traditional concern of the law, and that ‘consideration of a subsidy

State aid to business 143 forces the mind of the public body to consider most pointedly the cost and consequences of moving forward’. Leaving aside the general place of form versus effects in US law, it is clear that anti-trust law has come to emphasize effects rather than form as the vital element.37 The 2004 Circuit Court of Appeals Cuno decision38 involved selective tax breaks in Ohio (386 F.3d 738 at 741): In 1998, Daimler Chrysler entered into an agreement with the City of Toledo to construct a new vehicle-assembly plant near the company’s existing facility in exchange for various tax incentives. Daimler Chrysler estimated that it would invest approximately $1.2 billion in this project, which would provide the region with several thousand new jobs. In return, the City and two local school districts agreed to give Daimler Chrysler a ten-year 100 percent property tax exemption, as well as an investment tax credit of 13.5 percent against the state corporate franchise tax for certain qualifying investments. The total value of the tax incentives was estimated to be $280 million.

The Circuit Court of Appeals condemned the investment tax credit under the negative Commerce Clause.39 The Cuno decision is under appeal to the US Supreme Court, and national legislation has been proposed that seeks to reverse the decision.40 The ultimate fate of the Cuno decision therefore remains to be determined. A direct approach If the scope of the state action doctrine were limited compared with what seems to be its current range, states would be less able to compel business behaviour that would, if undertaken by firms collectively and on their own initiative, violate the Sherman Act. That would not mean that the Sherman Act would apply to state actions taken in pursuance of essential state functions. The negative Commerce Clause, in contrast, directly limits some state actions that distort inter-state commerce. The reach of the negative Commerce Clause is limited to tax measures that favour in-state over out-of-state firms. It does not address distortions in inter-state commerce resulting from the favouring of one in-state firm over other in-state firms. It does not reach selective subsidies at all. From an economic point of view, there is much to be said for a direct approach, that is, national legislation clearly adopting the policy that discriminatory state economic development measures, tax and subsidy, distort inter-state commerce and are prohibited under the commerce clause of the Constitution.41 Such a proposal would not prevent state competition to promote economic development. It would require that such competition be general rather than discriminatory, and would do much to prevent a Prisoners’ Dilemma outcome that leaves all states worse off without promoting the economic development of any state. 6 EU control of state aid: competition policy vs convergence In the European Union the legal basis for development of a system of state aid control is articles 87–9 of the EC Treaty. The system is embodied in a superstructure of guidelines, frameworks and regulations that have been developed over the years and focus on the anticompetitive effects of state aids. In this section we first review this normative system for state aid control and then move on to an analysis of general trends in member state aid (first subsection below), with particular attention to the implications of the accession of ten new member states in May 2004. Second, we outline (second subsection below) the EU’s own subsidy programmes, carried out through the Structural Funds.

144

International handbook on industrial policy

The EU reactive policy on state aid: rules for competition policy State aid control in the Community is based on articles 87–9 of the EU Treaty, which contain the general principle that state aid is incompatible with the common market and give the Commission wide power to investigate potentially illegal aids and to order their repayment. Potential cases of state aid should be ex ante notified to the Commission for clearance and aid cannot be put into effect before it has been cleared (the ‘standstill principle’).42 Article 87 contains a list of mandatory exemptions and examples of discretionary exemptions.43 The general prohibition in 87(1) includes a definition of state aid, develops the idea of avoiding discrimination among undertakings at the national level based on national origin, and aims to promote economic efficiency and market integration by ensuring that competition is not distorted. Mandatory exceptions include state aid having a social character and aid to assist recovery from disasters or exceptional occurrences. Discretionary exceptions include state aid to promote economic development in disadvantaged regions, aid to promote projects of common European interest, and aid to promote culture and heritage conservation. The Commission has adopted a number of frameworks, guidelines and regulations clarifying how discretionary exemption clauses apply, thus implementing a coherent structure of state aid rules across all member states and sectors of the economy. These guidelines, which apply to state aid granted following the notification to the Commission, mainly cover themes under categories presented in 87(3)(a) and (c) and explicitly address the following: – transactions by public authorities, such as state guarantees, furnishing risk capital, public land sales, export credit insurance, fiscal aid;44 – horizontal schemes for research and development aid, environmental aid, rescue and restructuring aid, regional aid; – sectoral aid schemes to shipbuilding, steel, coal, electricity, broadcasting, motor vehicle industry, synthetic fibres, air transport, maritime transport [. . .].

Block exemption regulations, that is, aid permitted without prior notification to the Commission, have been drawn up for employment aid, aid to small and medium-size enterprises and on training. Along with these exemptions, a de minimis rule45 has been amended in an effort to reduce the administrative burden on the member states and on the Commission itself (which ought to be left to concentrate its resources on cases of real importance to the Community) and in order to simplify matters for small and medium enterprises which are usually the beneficiaries of such small amounts of aid. There are then sector-specific conditions for the application of state aid rules to agriculture, fisheries, public transport and services of general interests. As it is concerned with services of general interest it is relevant to stress here that, following the Altmark judgment,46 the European Court of Justice ruled that certain forms of government support for public services did not constitute state aid as defined in the treaties, providing larger scope for supporting private sector delivery of key public services. Other fields where the EU is currently making significant changes are related to the framework for regional aid,47 the framework on rescue and restructuring aid and, in a number of procedures regarding notification forms, standardized reporting, interest rate to be used for recovery of illegally granted aid and rules relating to time limits guidelines on state aids.

State aid to business 145 Current trends Notwithstanding the legal system discussed above and the political priority to the reduction of the amount of state aid declared at the Lisbon European Council in March 2000, state aid continues at a high level in the Community, and continues to differ substantially among member states. According to the Spring 2004 State Aid Scoreboard,48 the overall volume of state aid fell between 1999 and 2002 from €52 billion to €49 billion, reaching the 0.56 per cent of EU GDP in 2002 (or 0.41 per cent excluding fisheries, agriculture and transport). Large disparities in level of aids between member states continue to be recorded: in 2002, Germany, France and Italy granted about 59 per cent of the total EU level of aids; the share of total aid in GDP ranges from 0.25 per cent in the United Kingdom to 1.28 per cent in Finland. About 57 per cent of total aid in the Union in 2002 was directed towards the manufacturing and service sectors, 28 per cent towards agriculture and fisheries, 11 per cent towards coal, and the remainder towards transport or non-manufacturing sectors ‘not elsewhere classified’. Once again, there are significant sectoral differences among member states: aid to coal represents 26 per cent and 23 per cent of the total for Germany and Spain, respectively; in Austria and Finland, the share of aid towards agriculture and fisheries was 66 per cent and 84 per cent, respectively; aid to the manufacturing and service sectors represents 78 per cent of the total in Denmark, 70 per cent in Belgium. These disparities among member states in state aid level and beneficiary sectors are at the bottom line in the analysis on the enlargement process of the European Union. The accession of eight Central and Eastern European countries (the Czech and Slovak Republics, Hungary, Poland, Slovenia, Estonia, Lithuania, Latvia), as well as Cyprus and Malta, was realized on 1 May 2004. Among many other issues, the long process of accession negotiations has focused on obtaining transparency in the levels and aims of existing state aid. A ‘sunset clause’ has been set for the end of April 2007: by that time, the new member states should have ensured compatibility of their subsidies with the EU state aid rules. Accordingly to the Autumn 2004 State Aid Scoreboard, total aid in the new member states, excluding agriculture, fisheries and transport, is estimated at €5.7 billion per year in the period 2000–2003. That corresponds to 1.42 per cent of GDP, and is more that three times the level of EU-15 in 2002. Most of the aid granted by the new member states was awarded by Poland (€2.4 billion), the Czech Republic (€1.9 billion) and Hungary (€0.57 billion). It is worthwhile to observe that the largest part of state aid (78 per cent) in the new member states is given in the form of sectoral aid, potentially the most distortive type of state aid (permitting the survival of less efficient firms and compelling the exit of more efficient firms).49 Favoured sectors include coal, steel, shipbuilding, finance and motor vehicles, where undertakings often receive rescue or restructuring aid. These data highlight a further field of conflict between EU-15 and the new member states, given that sectors receiving aid are almost the same in both the areas.50 EU active policy on regional aid: the Structural and Cohesion Funds The state aid articles of the EU Treaty apply to aid granted by the member states. Thus they do not apply to the EU’s own aid programmes, even though those aid programmes may distort competition and affect trade among the member states in the same way as member state aid.

146

International handbook on industrial policy

The Structural Funds are the Union’s mechanisms for carrying out regional and social policy. They include the European Social Fund (ESF), the European Regional Development Fund (ERDF), the European Agriculture Guidance and Guarantee Fund Guidance Section (EAGGF Guidance Section), the Financial Instruments for Fisheries Guidance (FIFG) and the Cohesion Fund.51 The ESF has its origin in Article 123 of the original EC Treaty; its role was to cushion labour market adjustment costs associated with economic integration. The ERDF was set up in 1975 to promote growth in the less developed regions of the Community. The Cohesion Fund was set up in 1993, in the run-up to economic and monetary union, to channel aid to member states with GNP per capita less than 90 per cent of the Community average. The original qualifying member states were Ireland, Greece, Spain and Portugal.52 With the EU enlargement of 1 May 2004, all the new member states qualified for the Cohesion Fund.53 As its name suggests, the EAGGF Guidance Section is twinned with the agricultural guarantee fund and the Community’s Common Agricultural Policy. Like the ESF, the CAP is anticipated in the Treaty of Rome (Article 40).54 Guidance Section funding, as indicated in Table 7.2, has always exceeded the combined funding of the Structural Funds (the latter, over the period 2000–2006, will account for over one-third of the EU budget). The regulations governing the Structural Funds have been revised repeatedly over the years. The revisions have sought to direct aid to regions and sectors that are less advantaged from the point of view of the EU as a whole, rather than from the point of view of any one member state. The 2000–2006 Structural Funds concentrate on three clearly defined objectives. Regions whose development is lagging behind (Objective 1) receive 70 per cent of the funding. Economic and social conversion in areas experiencing structural difficulties (Objective 2) will take up 11.5 per cent. Promoting the modernization of training systems and the creation of employment (Objective 3) absorbs a further 12.3 per cent of funding. The remainder is spread over four Community Initiatives seeking to address specific problems and the adjustment of fisheries structures outside the Objective 1 regions. The accession of 10 new member states on 1 May 2004 poses a policy dilemma for the Structural Funds (Van der Beek and Neal, 2004, p.587): ‘continuing an active regional policy makes enlargement unattractive to the existing members; switching to the reactive policy makes the EU membership less attractive to the accession countries’. Those

Table 7.2

Appropriations for commitments, billion euros, 1999 prices

2000 2001 2002 2003 2004 2005 2006 Source: EC Commission.

Agriculture

Structural operations

40.9 42.8 43.9 43.8 42.8 41.9 41.6

32.0 31.5 30.9 30.3 29.6 29.6 29.2

State aid to business 147 authors highlighted the fact that the modest active regional aid towards new member states has been not a rational regional strategy and called for an appropriate way to master the issue. Considering possible solutions to this dilemma, it could be reasonable to ask what has been the effectiveness of the EU active regional policy in reducing disparities before accession. Sala i Martin (1996) concludes that since the speed in rate of convergence in Europe is very low and stable (close to about 2 per cent a year in many regions), and given that the level of redistributive effort is different in the different regions, the Structural Funds cannot be very effective.55 7 Conclusion Some versions of string theory hypothesize the existence of 11 space–time dimensions. There seem to be at least as many dimensions to state aid. Most of the dimensions of string theory are imperceptible to the ordinary human senses. While much state aid is also imperceptible, this state of affairs is not rooted in laws of nature. In this regard, the United States and the EU have lessons to offer to one another. At present, the United States has substantial transparency in lobbying expenses at the national level, while the extent of lobbying expenses at the state level and the amounts of aid granted at both the state and the national level are essentially opaque. In the EU, the extent of lobbying expenses at the member state and the Union levels are opaque, while EU efforts have made the extent of member state aid substantially transparent. Transparency would be much improved if the United States would establish federal control of state aid that distorts inter-state commerce, if EU would adopt rules requiring documentation of lobbying efforts at the EU level, and if both jurisdictions were to adopt rules requiring documentation of lobbying efforts at the component (state, member state) level. Increasing transparency in aid delivered in the form of discriminatory tax competition56 would reduce the asymmetric information which usually affects the evaluation of the net benefit of location incentives. In turn, such an increase in transparency would help the selection of aid-seeking companies and reduce artificial inefficient location decisions. A second indication is that the case for selective development aid is a weak one, and that this observation applies to selective aid to home firms for their operations in international markets, to selective aid by the United States’ federal government and by the European Union, and to selective aid granted by regional and local governments. It illbehoves the policy makers of market economies to short-circuit the market mechanism. It may be possible to make an argument for transitional regional and sectoral aid; it seems much less possible to defend operating aid or selective tax breaks that distort market competition. Notes 1. 2. 3.

The latter issues in turn suggest connections to the large literature on foreign aid to less developed countries, which we leave aside in this chapter. On which, generally, see Krugman (1988). The passenger aircraft industry may be one industry that approximately satisfies the assumptions of the strategic trade policy literature; see Baldwin and Flam (1989), Pavcnik (2002) and Irwin and Pavcnik (2004) for references. The strategic trade literature tends also to take market structure as given (one firm in each country) but this is not a necessary characteristic of the approach. If one Cournot duopolist receives a sufficient large subsidy per unit of output, the equilibrium output of the other firm may be driven to zero, thus changing market structure from duopoly to monopoly. See Horstmann and Markusen (1992). Pursuit of this approach has implications for the economic analysis of EU state aid policy.

148 4. 5. 6. 7. 8. 9. 10. 11. 12.

13.

14.

15. 16.

17. 18. 19. 20. 21.

22.

International handbook on industrial policy Essentially the same result appears in the tax-competition literature. Brainard (1994, section 3) demonstrates the ambiguity of the welfare effects of state aid and tariff protection by one of two countries in a model of international trade in a declining industry; Anis et al. (2002) obtain corresponding results for the case in which both countries may engage in strategic trade policy. This remark suggests a connection with the rent-seeking literature. At the risk of being disingenuous, since it is far from being the only example, the track record of French governments in picking losers suggests the contrary. http://www.wto.org/english/tratop_e/scm_e/scm_e.htm. For discussions and surveys of the tax competition literature, see Bradbury et al. (1997), Kenyon (1997), Wilson (1999) and Glaeser (2001). Brueckner and Saavedra (2001) estimate tax rate reaction functions for a sample of 70 Boston-area cities; their slope estimate is positive, implying that strategic interaction does occur and suggesting that the ‘large numbers’ assumption may fail even at the local jurisdiction level. It should also be noted that European Union rules (article 87(3)(a) and (c) of the EC Treaty) provide for exceptions to the general prohibition against state aid. The professional economist is of course entitled to argue that local governments may lack either the information, or the ability to use the information, or the will to implement objective function-maximizing redistribution policies (or all three). On the other hand, the political preferences of the professional economist are entitled to as much weight, but no more, than the preferences of any other citizen (Harberger, 1971, p.785). In US anti-trust, hostility towards price discrimination can be traced to the perceived impact of railroad rate discrimination on rivalry among downstream users of transportation services. The founding fathers of what has become the European Union saw the appearance of a level playing field for competition between firms based in different member states as essential to the overriding goal of political union, and viewed a prohibition of price discrimination as essential to maintaining the appearance of such a level playing field. Lambertini and Peri (2001) show that fiscal incentives, such as fiscal taxes or sectoral state aid, are welfare improving for a country which is in the intermediate phase of globalization: this result is derived by developing the Fujita, Krugman and Venables (1990, ch.16) model of spatial industrial agglomeration and specifically focuses on how the tax policy that maximizes social welfare is related to falling trade costs. Lambertini and Peri’s results suggest that during the transitional phase towards full economic integration, as in the current EU integration process, pressure towards harmonization of tax rates could be suboptimal. For surveys, see Milward and Newman (1989) (who present case studies of tax incentive packages put together for six automobile plants that located in the US Midwest), Bartik (1992, 2002), Bradbury et al. (1997) and Buss (2001). Schwartz and Clements (1999) describe a wide array of government activities that may contain subsidy elements, among which (1) direct government aid to producers or consumers (cash subsidies); (2) government guarantees, interest subsidies to enterprises, soft loans (credit subsidies); (3) reductions of specific tax liabilities (tax subsidies); (4) government equity participation (equity subsidies); (5) government provision of goods and services at below-market prices (in-kind subsidies); (6) government purchases of goods and services at above-market prices (procurement subsidies); (7) implicit payments through government regulatory actions that alter market prices or access (regulatory subsidies). Measures of economic activity considered by the surveyed studies included employment, output and income. Wasylenko (1997) similarly reads the literature as suggesting a tax elasticity of 0.2, suggesting that reducing tax rates by 10 per cent will increase economic activity, variously measured, by 2 per cent. Target industrial incentives seem to continue despite some indications (for example, Hartzheim, 1997) that state and local governments are developing an appreciation for the downside of such policies. For an early survey of the rent-seeking literature, see Tollison (1982). Tullock (1967, p. 228): ‘Generally governments do not impose protective tariffs on their own. They have to be lobbied or pressured into doing so by the expenditure of resources in political activity. One would anticipate that the domestic producers would invest resources in lobbying for the tariff until the marginal return on the last dollar so spent was equal to its likely return producing the transfer.’ Krueger (1974, p. 291): In many market-oriented economies, government restrictions upon economic activity are pervasive facts of life. These restrictions give rise to rents of a variety of forms, and people often compete for the rents. Sometimes, such competition is perfectly legal. In other instances, rent seeking takes other forms, such as bribery, corruption, smuggling, and black markets.

23.

This approach was applied by Cowling and Mueller (1978), in a work that extended that of Harberger (1954). Among many other discussions of the complete dissipation result, see Fisher (1985) and Fudenberg and Tirole (1987).

State aid to business 149 24. 25. 26. 27. 28. 29. 30. 31. 32. 33. 34.

35. 36. 37. 38. 39.

40. 41. 42.

43. 44. 45.

46.

47. 48.

For a discussion of early protection seeking by the US steel industry, see Hillman (1989, pp. 144–7). Organizations spending less than $22 500 on lobbying during a six-month period are exempted from the filing requirements. See Cowles (1996), Pijnenburg (1998) and Greenwood (1999), as well as Coen (1997). Speech, ‘The need for a European transparency initiative’, 17 March 2005. See Burstein and Rolnick (1995) and the June 1996 issue of The Region (http://minneapolisfed.org/pubs/ region/96–06/index.cfm). Parker v. Brown 317 U.S. 341 (1943). Parker v. Brown was foreshadowed by Lowenstein v. Evans 69 F. 908 (C.C.D.S.C. 1895) and Olsen v. Smith 195 U.S. 332 (1904); it was followed in particular by City of Lafayette v. Louisiana Power & Light Co. 435 U.S. 389 (1978). See Donnem (1970), Posner (1974, fn. 12); Handler (1976, p. 6); Areeda (1981). For discussion, see Smith (1999a, 1999b). Other requirements for state taxes are (Cuno v. Daimler-Chrysler, 386 F.3d 738 at 742) that ‘the activity taxed has a substantial nexus [presence] with the taxing State . . . the tax is fairly apportioned to reflect the degree of activity that occurs within the State . . . and . . . the tax is fairly related to benefits provided by the state’. W. Lynn Creamery v. Healy 512 U.S. 186. Similar issues arise in France v. Commission Case 47/69 [1970] ECR 487, [1970] CMLR 351. To the extent that the discriminatory aid has the effect of shifting investment that would have taken place in any case from one state to another, there is no net creation of jobs or increase in post-investment output, compared with the outcome if the aid had not been granted. To the extent that state budget constraints are binding, the immediate effect of forgone tax revenues or subsidy spending is to limit the ability of the state to provide public goods. This limitation is more severe if discriminatory tax competition leads to a Prisoners’ Dilemma outcome in which rival states grant mutually-neutralizing economic development measures. Of course, from an economic point of view, a pure subsidy does distort competition, and the first sentence that is quoted here is incorrect. See also, in footnote 15 of the decision: ‘We have never squarely confronted the constitutionality of subsidies, and we need not do so now. We have, however, noted that “direct subsidization of domestic industry does not ordinarily run afoul” of the negative Commerce Clause’. See, for example, Continental TV v. GTE Sylvania 433 U.S. 36 (1977). Cuno v. Daimler-Chrysler 386 F.3d 738. The facts of the case are also discussed by Enrich (2002). In so doing, the Circuit Court remarked on the subsidy-targeted tax measure distinction (386 F.3d 738 at 746): ‘Although the defendants liken the investment tax credit to a direct subsidy, which would no doubt have the same economic effect, the [Supreme] Court has intimated that attempts to create location incentives through the state’s power to tax are to be treated differently from direct subsidies despite their similarity in terms of end-result economic impact.’ See Zelinsky (2003) for arguments in favour of abandoning the negative Commerce Clause entirely. See Frickey (1996) and Kent and Capehart (2003) for similar views. From a procedural perspective, the grant of a notified state aid, after the Commission investigation, may result ‘positive’, ‘conditional’ or ‘negative’. The Commission can start a procedure of investigation for non-notified aid on its own initiative, a procedure which in the case of a negative judgment results in the recovery of the offending state aid. Article 88 sets out the basic procedural rules regarding the enforcement of article 87, and article 89 is the legal basis for Council regulations in the state aid field. On fiscal state aid, see Nicolaides (2004). As described in the Commission Regulation (EC) No 69/2001, the de minimis rule consists in the definition of a threshold figure below which Article 92 (1) can be said not to apply and – consequently – the aid need no longer be notified in advance to the Commission under Article 93 (3). Actually, the total amount of de minimis aid may not exceed €100 000 over a three-year period and any sector is eligible with the exception of transport, agriculture and export-related activities. The recent ruling by the European Court of Justice in the Altmark case (Case C-280/00, 24 July 2003) provides that compensation for the costs incurred in the discharge of a public service obligation do not qualify as state aid if a number of conditions are cumulatively met: (1) clear public service obligations; (2) pre-established parameters for determining the compensation; (3) no overcompensation; (4) either selection of operator through tender procedure or determination of compensation with reference to costs of a typical, well-run undertaking. Early indications of Commission thinking in this field show the intention to ban aid to large firms altogether except in Objective 1 regions, which post-2007 will be almost exclusively in the new member states of Central Europe and the Baltic states. See http://europa.eu.int/comm/competition/state_aid/scoreboard/. The Spring 2005 State Aid Scoreboard records a slight decline in the level of state aid in relation to GDP, ‘but the underlying trend is more stable than downward’ (Scoreboard 2005, p. 4). Given this evidence, the analysis which follows is based on the

150

49. 50.

51. 52.

53. 54. 55.

56.

International handbook on industrial policy Spring Scoreboard 2004, which permits better data comparisons with the first data on new members’ state aid collected in the Autumn 2004 Scoreboard. See Martin and Valbonesi (2000, Table 5.1) for a discussion. The state aid effect on similar sectors’ structure in different countries is captured by Collie (2000): the author develops a symmetric Cournot duopoly model describing an integrated market where one firm is located in each member state. The result is that the prohibition of state aid may very well increase overall welfare and thus provide a rationale for a ban of state aid. The European Investment Bank and the European Investment Fund provide aid through financial markets, often in coordination with Structural Fund activities. In the period 1999–2001, the Cohesion countries accounted for 10 per cent of total expenditure on state aid to the manufacturing sector, while the same data amount to 76 per cent for the four big economies (Germany, France, Italy and UK). See the EU Third Report on Economic and Social Cohesion, p.129 (http://europa.eu.int/comm/regional_policy/sources/docoffic/official/reports/cohesion3/cohesion3_e n.htm). For the years 2004–6, €18 billion (in 2004 prices) are available, and more than half of the funding has been reserved for the new member states. See the EU Third Report on Economic and Social Cohesion. Giving substance to the CAP was one element in the 1965 ‘empty chair’ crisis, which led to the Luxembourg Compromise. Boldrin and Canova (2001), analysing the evolution of income across EU regions over the last 20 years, find that spending from EU Structural Funds has had no observable impact on regional growth and economic convergence. However, Barry (2003) highlights the case of Ireland, where the interactions of EU structural funding with other elements of internal economic policy have led to positive effects in terms of convergence. This issue should take into account that the cohesion objective is explicitly pursued in the EU – not in the USA – and allows governments to give financial support for relocation from prosperous to disavantaged regions.

References Anderson, L.R. and S.L. Stafford (2003), ‘An experimental analysis of rent seeking under varying competitive conditions,’ Public Choice, 115, 199–216. Anis, A.H., M. Benarroch and Q. Wen (2002), ‘Persistent protection in an international exit game’, Journal of International Economics, 56, 465–87. Areeda, P. (1981), ‘Antitrust immunity for “state action” after Lafayette’, Harvard Law Review, 95, 435–55. Baldwin, R. and H. Flam (1989), ‘Strategic trade policies in the market for 30–40 seat commuter aircraft’, Weltwirtschaftliches Archiv, 125, 484–500. Barry, F. (2003), ‘Economic integration and convergence processes in the EU cohesion countries’, Journal of Common Market Studies, 41 (5), 897–921. Bartik, T.J. (1992), ‘The effects of state and local taxes on economic development: a review of recent research’, Economic Development Quarterly, February, 102–11. —— (2002), ‘Evaluating the impacts of local economic development policies on local economic outcomes: what has been done and what is doable?’, Upjohn Institute Staff Working Paper no. 03-89, November. Baye, M.R., D. Kovenock and C.G. de Vries (1989), ‘The economics of all-pay, winner-take-all, contests’, Texas A&M Department of Economics Working Paper 89-21, November. —— (1993), ‘Rigging the lobbying process: an application of the all-pay auction’, American Economic Review, 83(1), 289–94. Bhagwati, J.N. (1982), ‘Directly unproductive, profit-seeking (DP) activities’, Journal of Political Economy, 90(5), 988–1002. Boldrin, M. and F. Canova (2001), ‘Inequality and Convergence’, Economic Policy, 32, 207–53. Bradbury, K.L., Y.K. Kodrycki and R. Tannenwald (1997), ‘The effects of state and local public policies on economic development: an overview,’ New England Economic Review, March/April, 1–12. Brainard, S.L. (1994), ‘Last one out wins: trade policy in an international exit game’, International Economic Review, 35(1), February, 151–72. Brander, J.A. and B.J. Spencer (1985), ‘Export subsidies and international market share rivalry’, Journal of International Economics, 18(1–2), 83–100. Brook, D.A. (2005), ‘Meta-strategic lobbying: the 1998 Steel Imports case’, Business and Politics, 7(1), article 4, May. Brueckner, J.K. and L.A. Saavedra (2001), ‘Do local governments engage in strategic property-tax competition?’, National Tax Journal, 54(2), 203–30. Burstein, M.L. and A.J. Rolnick (1995), ‘Congress should end the economic war among the states’, The Region March (http://minneapolisfed.org/pubs/ar/ar 1994.cfm).

State aid to business 151 Buss, T.F. (2001), ‘The effect of state tax incentives on economic growth and firm location decisions: an overview of the literature’, Economic Development Quarterly, 15(1), 90–105. Coen, D. (1997), ‘The evolution of the large firm as a political actor in the European Union’, Journal of European Public Policy, 4(1), 91–108. Collie, D.R. (2000), ‘State aid in the European Union: the prohibition of subsidies in an integrated market’, International Journal of Industrial Organization, 18(6), 867–84. Cowles, M.G. (1996), ‘The EU committee of AmCham: the powerful voice of American firms in Brussels’, Journal of European Public Policy, 3(3), 339–58. Cowling K. and D.C. Mueller (1978), ‘The social costs of monopoly power’, Economic Journal, 88, 727–48. Davis, D.D. and R.J. Reilly (1998), ‘Do too many cooks always spoil the stew? An experimental analysis of rentseeking and the role of a strategic buyer’, Public Choice, 95, 89–115. Donnem, R.W. (1970), ‘Federal antitrust law versus anticompetitive state regulation’, Antitrust Law Journal, 39, 950–67. Enrich, Peter D. (2002), ‘Business tax incentives: a status report’, Urban Lawyer, 34(1), 413–28. Fisher, F.M. (1985), ‘The social costs of monopoly and regulation: Posner reconsidered’, Journal of Political Economy, 93(2), 410–16. Frickey, P.P. (1996), ‘The congressional process and the constitutionality of federal legislation to end the economic war among the states’, The Region (downloaded from http://minneapolisfed.org/pubs/region/9606/frickey.cfm, 15 March 2005). Fudenberg, D. and J. Tirole (1987), ‘Understanding rent dissipation: on the use of game theory in industrial organization,’ American Economic Review, 77(2), 176–83. Fujita, M., P. Krugman and A.J. Venables (1990), The Spatial Economy, Cambridge, MA: MIT Press. Glaeser, E.L. (2001), ‘The economics of location-based tax incentives’, Harvard Institute of Economic Research Working Paper 1932, 29 October. Goldstein, A.E. and S.M. McGuire (2004), ‘Political economy of strategic trade policy and the Brazil–Canada export subsidies saga’, World Economy, 27(4), 541–66. Greenstone, M. and E. Moretti (2004), ‘Bidding for industrial plants: does winning a ‘million dollar plant’ increase welfare?’, MIT Department of Economics, Working Paper 04-39, November. Greenwood, J. (1999), ‘The future of EU-level interest representation’, Journal of Communication Management, 4(1), 9–19. Grubert, H. and J. Mutti (1991), ‘Taxes, tariffs and transfer pricing in multinational corporate decision making’, Review of Economics and Statistics, 73(2), 285–93. Handler, M. (1976), ‘The current attack on the Parker v. Brown state action doctrine’, Columbia Law Review, 76(1), 1–20. Harberger, A.C. (1954), ‘Monopoly and resource allocation’, American Economic Review, 44(2), 77–87. —— (1971), ‘Three basic postulates for applied welfare economics: an interpretive essay’, Journal of Economic Literature, 9(3), 785–97. Hartzheim, L.A. (1997), ‘State tax incentives: headed in the right direction’, Journal of State Taxation, 15(4), 51–64. Hazlett, T.W. and R.J. Michaels (1993), ‘The cost of rent-seeking: evidence from cellular telephone license lotteries’, Southern Economic Journal, 59(3), 425–35. Hellerstein, W. and D.T. Coenen (1996), ‘Commerce clause restraints on State business development incentives’, Cornell Law Review, 81, 789–878. Hillman, A.L. (1989), The Political Economy of Protection, Chur: Harwood Academic Publishers. Hillman, A.L. and D. Samet (1987), ‘Dissipation of contestable markets by small numbers of contenders’, Public Choice, 54(1), 63–82. Hines, J.R. Jr. (1999), ‘Lessons from behavioral responses to international taxation’, National Tax Journal, 52(2), 305–22. Horstmann, I.J. and J.R. Markusen (1992), ‘Endogenous market structures in international trade (natura facit saltum)’, Journal of International Economics, 32, 1009–29. Huttin, C. (1989), ‘The effects of state aid on employment in an investment in the French textile and clothing industry’, International Journal of Industrial Organization, 7, 489–501. Irwin, D.A. and N. Pavcnik (2004), ‘Airbus versus Boeing revisited: international competition in the aircraft market’, Journal of International Economics, 64, 223–45. Janeba, E. (1998), ‘Tax competition in imperfectly competitive markets’, Journal of International Economics, 44(1), February, 135–53. Kent, C. and R. Capehart (2003), ‘Chief Justice John Marshall and State competition for economic development’, Journal of State Taxation, 22(1), 72–80. Kenyon, D.A. (1997), ‘Theories of interjurisdictional competition’, New England Economic Review, March/ April, 13–28.

152

International handbook on industrial policy

Krueger, A.O. (1974), ‘The political economy of the rent-seeking society’, American Economic Review, 64(3), June, 291–303. Krugman, P.R. (ed.) (1988), Strategic Trade Policy and the New International Economics, Cambridge, MA: MIT Press. Lambertini, L. and G. Peri (2001), ‘Fiscal incentives and industrial agglomeration’, IGIER Working paper no. 157. Martin, S. and P. Valbonesi (2000), ‘State aid in context’, in G. Galli and J. Pelkmans (eds), Regulatory Reform and Competitiveness in Europe, Vol. 1, Cheltenham, UK and Northampton, MA, USA: Edward Elgar, pp. 176–201. —— (2005), ‘The state aid game’, manuscript, April. McGuire, T.J. (1991), ‘Federal aid to states and localities and the appropriate competitive framework’, in D.A. Kenyon and J. Kincaid (eds), Competition Among States and Local Governments: Efficiency and Equity in American Federalism, Washington, DC: Urban Institute Press, pp. 153–66. Milward, H.B. and H. Hosbach Newman (1989), ‘State industrial packages and the industrial location decision’, Economic Development Quarterly, 3(3), 203–22. National Association of State Development Agencies (2001), ‘NASDA State Economic Development Survey’, CD-ROM. Nicolaides, P. (2004), ‘Fiscal state aid in the EU: the limits of tax autonomy’, World Competition, 27(3), 356–96. Oates, W.E. and R.M. Schwab (1988), ‘Economic competition among jurisdictions: efficiency enhancing or distortion inducing?’, Journal of Public Economics, 35, 333–54. Oates, W.E. and R.M. Schwab (1991), ‘The allocative and distributive implications of local fiscal competition’, in D.A. Kenyon and J. Kincaid (eds), Competition Among States and Local Governments: Efficiency and Equity in American Federalism, Washington, DC: Urban Institute Press, pp. 127–45. Pavcnik, N. (2002), ‘Trade disputes in the commercial aircraft industry’, World Economy, 25(5), 733–51. Phillips, J.M. and E.P. Goss (1995), ‘The effect of state and local taxes on economic development: a metaanalysis,’ Southern Economic Journal, 62(2), 320–33. Pijnenburg, B. (1998), ‘EU lobbying by ad hoc coalitions: an exploratory case study’, Journal of European Public Policy, 5(2), 303–21. Posner, R.A. (1975), ‘The social costs of monopoly and regulation’, Journal of Political Economy, 83, 807–27. Posner, S.P. (1974), ‘The proper relationship between State regulation and the Federal antitrust laws’, New York University Law Review, 49(5), 693–739. Sala i Martin, X. (1996), ‘Regional cohesion: evidence and theories of regional growth and convergence’, European Economic Review, 40, 1325–52. Schwartz, G. and B. Clements (1999), ‘Government subsidies’, Journal of Economic Surveys, 13(2), 119–48. Sinn, H-W. (1990), ‘Tax harmonization and tax competition in Europe,’ European Economic Review, 34(2–3), 489–504. Smith, J.K. (1999a), ‘Use of business tax incentives: part 1’, Journal of State Taxation, 17(4), 1–21. —— (1999b), ‘Use of business tax incentives: part 2’, Journal of State Taxation, 18(1), 1–23. Tiebout, C. (1956), ‘A pure theory of local expenditure’, Journal of Political Economy, 64(5), 416–24. Tollison, R.D. (1982), ‘Rent seeking: a survey’, Kyklos, 35(4), 575–602. Tullock, G. (1967), ‘The welfare costs of tariffs, monopolies and theft’, Western Economic Journal, 5, 224–32. —— (1980), ‘Efficient rent seeking’, in J.M. Buchanan, R.D. Tollison and G. Tullock (eds), Toward a Theory of the Rent-Seeking Society, College Station: Texas A&M University Press, pp. 97–112. Van der Beek, G. and L. Neal (2004), ‘The dilemma of Enlargement for the European Union’s Regional Policy’, World Economy, 27 (4), 587–607. Wasylenko, M. (1997), ‘Taxation and economic development: the state of the economic literature’, New England Economic Review, March/April, 37–52. Wilson, J.D. (1999), ‘Theories of tax competition’, National Tax Journal, 52(2), 269–304. Zelinsky, E.A. (2003), ‘Restoring politics to the Commerce Clause: the case for abandoning the dormant Commerce Clause prohibition on discriminatory taxation’, Benjamin N. Cardozo School of Law, Jacob Burns Institute for Advanced Legal Studies Research Paper no. 56 (http://ssrn.com/abstract_id=366921).

8

Mergers and concentration policy Hans Schenk

1 Introduction Over the last one hundred years, we have seen five merger waves, three of which occurred after World War II. The fifth wave, which had its rising tide from 1995 to 2000, required worldwide investments of no less than about US$12 000 billion. With about US$9000 billion, American and West European firms took the lion’s share (for more details, see Schenk 2002a, 2005). At the time, by way of comparison, acquisition expenditures by American and European firms were about seven times larger than Britain’s annual gross domestic product and more than 20 times Dutch GDP. On average, they amounted annually to about one-fifth of US GDP. Put differently, American and West European investments in mergers and acquisitions were approximately equal to 60 per cent of their gross investments in machinery and equipment (gross fixed capital formation) and they easily outpaced those in research and development (R&D). Business enterprise investments in acquisitions were no less than about eight times higher than business enterprise expenditures on R&D (which amounted to approximately US$1237 billion over the same period). The sheer size of the phenomenon makes it clear that the fate of mergers and acquisitions periodically, that is, during and after a wave, must have a crucial effect on the fate of the economies in which they occur. If they improve the way in which society generates wealth, economies will noticeably benefit, leaving aside the question to which parties the benefits will accrue. If, on the other hand, they do not generate wealth, or even destroy it, then economies will noticeably suffer. One of the ways in which mergers could fail to generate or even destroy wealth would be if they had a negative effect on competition without compensating for this with parallel improvements in the productivity of the parties to a merger or in the quality of their products or services. More precisely, this would be the case if a deterioration of allocative efficiency would not be compensated for by an increase in productive (or internal) efficiency, transactional efficiency and/or dynamic efficiency. Evidently, and given the periodically high incidence of mergers and acquisitions (M&A), the way in which economies appreciate mergers (that is, deal with potential benefits and disadvantages) may be of crucial macroeconomic importance.1 Therefore the present contribution will, first, review the evidence on the performance of mergers. Secondly, it will discuss current merger control regulations and ask whether these are able to deliver the goods; that is, whether they can be expected to further beneficial mergers and block mergers that are a threat to economic welfare. Meanwhile, the discussion on merger performance will raise questions with respect to the determinants of merger. We will discuss this issue to some extent for reasons that will become clear as we proceed. Similarly, it is as yet uncertain how important the macro effects of merger waves exactly are. Yet, in order to assess the usefulness of control regimes and/or changes therein, it matters to know at least the order of magnitude. Space restric153

154

International handbook on industrial policy

tions will only allow us to touch briefly upon the measurement issues involved, so we will refer to recent publications. 2 The performance of mergers It has recently been concluded that ‘there seems to be no support for a general presumption that mergers create efficiency gains’ (Röller et al., 2001, p.37). In the light of evidence not studied by Röller and his co-authors, this conclusion would seem unduly cautious. If half a dozen omitted but essential studies are taken into account as well, it would seem that the general presumption would rather have to be that there would be support for the presumption that (large) mergers do not create efficiency gains. By now, the performance of mergers and acquisitions has been the subject of many dozens of studies, both in terms of real value effects and in terms of shareholder value effects. Almost all of the 60 or so studies discussed and/or cited in Schenk (2002b) have focused on transactions in which large firms (typically large, quoted firms) were involved. By far the most studies have estimated shareholder value effects, mostly using readily available stock market data, and using predicted normal returns as controls. Those studies that estimated real value effects, however, have used more sophisticated data, usually drawn from firm statements, as well as more sophisticated methodologies. They have commonly used size and industry-matched control groups of non-merging firms and/or ceteris paribus extrapolations of pre-merger performance. Although the findings of the various studies are not completely consistent, the general tendencies are clear. Besides, since both shareholder value and real value studies, under certain restrictions, have similar conclusions, these findings must be regarded as rather convincing. Real value Let us begin by discussing some of the real value studies. Dickerson et al. (1997) found for a panel of almost 3000 UK-quoted firms that acquisitions undertaken during the period 1948–77 had a systematic detrimental impact on company performance as measured by the rate of return on assets. Not only was the coefficient on acquisition growth much lower than that on internal growth, there appeared to be an additional and permanent reduction in profitability following acquisition as well. More specifically, for the average company, the marginal impact of becoming an acquirer was to reduce the rate of return relative to non-acquirers by 1.38 percentage points (in the year of the first acquisition). Taking all subsequent acquisitions into account, acquiring firms experienced a relative reduction of 2.90 percentage points per annum. Since the mean return across all non-acquiring firms was 16.43 per cent, this translates into a shortfall in performance by acquiring firms of 2.9/16.43, which is around 17.7 per cent per annum. This finding is not an exception. On the contrary, the most common result of merger performance studies is that profitability and productivity, variously measured, do not improve as a result of merger (see, for example, Meeks, 1977 and Cowling et al., 1980, for the UK; Ravenscraft and Scherer, 1987, for the USA; Bühner, 1991, for Germany; Simon et al., 1996, for the advertising industry). Cowling and co-authors, in particular, broke new ground by estimating changes in unit factor requirements k, where P  k  P0 (1  R ) , I

Mergers and concentration policy 155 in which P0 equals revenue per unit of output, PI is a fixed-weight price index of inputs,  equals total profits from production, and R is total revenue from output produced. A commonsense interpretation of this formula is that P0(1/R) is simply average unit cost. When this is deflated by PI we have average cost after allowing for a change in input prices. The formula will give an exact measure of the change in k from observation to observation; changes in k over time are inversely proportional to efficiency in the sense of the ratio of inputs to outputs. Although the data requirements of this approach are immense, it allows measuring the impact of merger while controlling for input as well as output price changes and for factor substitution. Cowling et al. (1980) conclude that it is impossible to sustain the view that merger is in fact a necessary or sufficient condition for efficiency gain. In many cases efficiency did not improve or in fact decline, while in other cases it improved but no faster than would have been expected in the absence of merger.2 Unfortunately, the Cowling methodology has not been applied since, probably as a result of high data requirements. Still, from most other studies it can be inferred that, if mergers do show profit improvements, these are usually either not merger-specific or due to the creation of additional market power rather than efficiency (Gugler et al., 2003). The lack of merger specificity has, perhaps, been most clearly demonstrated by a large number of real merger effects studies in the banking industry. Before we come back to this industry below, we should notice that the results for another indicator of efficiency also appear pessimistic. For market share growth seems to slow down after a merger as well, while acquired firms lose market share against a control group of firms that remain independent (Mueller, 1986). For instance, among the world’s 18 largest pharmaceutical firms, 11 out of 12 that participated in mergers lost combined market share between 1990 and 1998, whereas all six of those that had not merged gained market share (The Economist, 22 January 2000). Whereas these results are, to a certain extent, consistent with expectations from the Salant et al. (1983) model (merging firms will use output contraction to boost their profits but other firms in the industry will jump into the resulting output gap, thus increasing their market share), this assumes Cournot behaviour and near-equal market shares of the merging firms. Although banking mergers appear to perform slightly better, mainly owing to the fact that banks appear less rationalized as yet than manufacturing firms, the results for manufacturing mergers are similarly reflected in almost all studies of banking mergers. Tichy (2001), who has reviewed some 25 studies of mergers among mostly US banks, concludes that roughly a third have reported positive effects in terms of either rising returns, declining costs, increasing profits or greater efficiency. Neutral effects were reported in slightly more than half of the studies whereas 16 per cent reported negative effects. Thus about two-thirds of the banking mergers investigated in these studies were unsuccessful. However, the positive effects for the remaining cases were typically much smaller than expected. Besides, the studies included many small, locally active banks. Another survey, by Van Rooij (1997), concludes that, even when acquirers are relatively (cost) efficient ex ante, which should create a potential to transfer efficiency to targets, there is hardly any evidence of such opportunities being realized after the merger. Relative to non-merging banks, mergers do not show significant efficiency improvements. Similarly, Akhavein et al. (1997) recall that banks have costs that are typically 20–25 per cent above those of the observed best-practice banks, which would suggest that cost

156

International handbook on industrial policy

efficiency could be considerably improved by merger. Again, however, they notice that such potentials are not systematically realized in practice. By and large, the consensus is that bank M&As at best lead to very little improvement in productive efficiency. Recent confirmation for German banking mergers has been found in Kötter (2005). Exceptions exist, of course, but they mostly pertain to mergers among very small, locally active banks. Although this cannot yet be fully substantiated, the findings look like suggesting that the larger the merging banks, especially when their size is beyond a still quite limited asset size of $10bn, the smaller are the chances of cost improvements. Indeed, as Tichy (1990) has already concluded, for the largest banks in Europe as well as elsewhere, there is no significant relationship between size and profitability, which indicates either absence of market power and efficiency effects or, more plausibly, a compensation of market power gains by decreasing returns to scale (also see De Jong, 1993). Upon reflection, this is what one would expect on the basis of estimates of minimal efficient (or optimal) scale (MES) in banking. Repeatedly, such estimates have ranged between approximately $1bn and $10bn and, more recently (with regulations becoming looser and looser), $25bn. In Europe, Van der Vennet (1996) has reported that optimal banking size from a cost-efficiency point of view would be in the range of €10bn to €100bn in assets. Clearly, the difference from actual practice is enormous, with many of the larger banks having assets beyond, sometimes far beyond, €500bn. Recent deregulation in both the EU and the USA will probably push the MES upwards while technological advance will, as usual, exercise a downward pull, but since the typically found average cost curves are only weakly U-shaped, this is not likely to make the future much different from the past. Some studies of banking efficiency have recently introduced a distinction between internal (or X-) efficiency and profit-efficiency. As a result of merger, a bank may be able to find superior product combinations, for example by moving into higher-valued products like loans instead of securities. Such profit-efficiency effects have only been studied by Akhavein et al. (1997) for US ‘mega-mergers’ undertaken during the 1980s (where these mergers are defined as transactions involving firms with assets in excess of $1bn each) and by Berger (1998) who focused on the early 1990s. Though the evidence therefore is much less definitive, and it remains uncertain how far merging banks generally proceed to a different output mix, their results suggest that the average output mix changes so that profit efficiency is increased by a few per cent. Notice, however, that both samples virtually excluded ‘supermegamergers’ (Berger et al., 1999), a category that has increased in importance recently. Moreover, it is quite uncertain to which extent these changes in product-market combinations were merger-specific. One could easily imagine that similar changes would have been possible without undertaking a merger. Overall, several methodological criticisms may be brought against some of the established types of merger performance studies (see, for example, Calomiris, 1999). Yet the evidence, as we have seen, appears quite consistent across studies of financial as well as non-financial mergers as well as time periods. In fact, the only substantial exception to the findings reported above is a study by Healy et al. (1992) who investigate post-merger cash flow for the 50 largest non-financial US mergers consummated between 1979 and 1984; that is, typically the sort of merger that would be submitted for review to merger authorities. By adopting the same index as Ravenscraft and Scherer (1987) did in the most revealing study to appear before the fifth merger wave (and arguably the best ever), Healy and

Mergers and concentration policy 157 co-authors purported to have refuted the Ravenscraft and Scherer findings. Their results showed that around two-thirds of these mergers had cash flow improvements ex post. However, Healy et al. deflated this index of performance with a market-based asset variable which, as Scherer (2002) and Schenk (2002b) have commented, can imply cash flow/asset performance indicator gains relative to the market even when cash flows are deteriorating relative to those of peer companies, namely if acquiring company market value falls relative to the general market – which, indeed, appeared to be the case. Moreover, it appeared that relatively many assets were sold off after the merger. Upon closer inspection, these assets appeared to have relatively high book values but relatively low sell-off revenues. This clearly suggests cooking of the books, in the sense that some assets may have been artificially inflated in order to prevent high write-ups to goodwill accounts. Sell-offs in this case will result in relatively favourable cash flow/asset performance during the post-merger years. Indeed, when the authors in a later (substantially less well-known) study added acquisition premia to the deflator, results deteriorated significantly. On average, the mergers studied now appeared to be unprofitable and/or insignificantly different from sector indicators. Shareholder value Similar results are obtained when the focus is on shareholder instead of real wealth. A review of 33 earlier studies by Mueller (2003) finds that, while target shareholders usually gain from acquisitions, acquirer shareholders almost always lose, especially in the long run. Generally, the longer the post-merger assessment period, the more negative shareholder returns appear. Usually, positive abnormal returns are only evident for a few days around the event (and even then, only when pre-event build-ups of share prices are underestimated), but taking this as evidence requires a strong belief in the efficient market hypothesis. Another review confirmed Mueller’s findings before testing it for a sample of 110 very large acquisitions – among them the most prolific of the century – undertaken during 1993–2001, thus including years that preceded the beginning of the fifth merger wave and followed its demise (Schenk, 2002b). It found that, for several different models, varying only in terms of event windows, the outcomes were all negative in terms of cumulative abnormal returns (CAR) on the acquirer’s side, running from 3.4 per cent to 8.5 per cent (see Table 8.1). In another study, the effects of 87 much-advertised cross-border acquisitions by Dutch multinationals were investigated. These mergers initially appeared to do better than the benchmark but in due time they went into the red as well (see Figure 8.1). Notice that mergers which were undertaken in the years immediately preceding the fifth merger wave were overrepresented in the second study, whereas the first focused on mergers from this fifth merger wave. The first found negative returns for short event windows while the second would have found only positive returns if the analysis had been constrained to similarly short windows as well. This suggests the possibility of intertemporal (rather than inter-sector) variations in merger performance. Therefore another sample of 100 (European) mergers was divided into five year cohorts (beginning with 1995 and ending with 1999). For 400 post-merger days each, the study revealed that ‘earlier’ acquisitions perform better (or less badly) than ‘later’ acquisitions. Figure 8.2 shows that the 1995 cohort achieved positive results but that all others were in the negative, the 1999 cohort performing worst of all; it saddled its

158

International handbook on industrial policy

Table 8.1

Shareholder value effects of 110 very large acquisitions, 1993–2001 (%)

CAR acquirer T-value* *all T-values significant at 5%

Model 1

Model 2

Model 3

Model 4

Model 5

3.41 4.691

4.48 4.764

5.12 4.422

4.99 3.764

8.48 5.174

Includes Domestic M&As AOL–Netscape Terra Networks–Lycos Verizon–GTE AT&T–McCaw Cellular QWest–US West Compaq–Digital Gener. Electric–Honneywell Pfizer–Warner-Lambert Paramount–Viacom

Starwood–ITT Chevron–Texaco Exxon–Mobil Total Fina Elf–Petrofina Weyerhouser–Willamette Carrefour–Promodes E.On–Viag United Airlines–US Airways

Cross-border M&As Telefonica–Endemol Mannesmann–Orange Deutsche Telekom–Voice-Stream Glaxo–Smithkline Beecham Reed Elsevier–Harcourt Daimler-Benz–Chrysler Unilever–Bestfoods Diageo–Grand Metropolitan Vivendi–US Filter Suez Lyonnaise–Tractebel

Source: Schenk (2005).

10 5 0 1

1

3

5

7

9 11 13 15 17

25 27 29 31 33 35

5 10 15 Source: Schenk (2005).

Figure 8.1

Shareholder returns of foreign acquisitions by Dutch firms, 1990–95

shareholders with an average cumulative loss of almost 25 per cent. Similarly, in a study of about 12 000 (American) acquisitions from 1980 to 2001, Moeller et al. (2003a) found that, while shareholders lost throughout the sample period, losses associated with acquisitions after 1997 were ‘dramatic’. The periodicity found in these studies is consistent with newer work by Carow et al. (2004) who investigate stockholder returns for 520 acquisitions over 14 industry-defined merger waves during the period 1979–98. They find that the combined returns for target

Mergers and concentration policy 159 15 10 5 0 5 10 15 20 25 30

1995

1996

1997

1998

1999

Source: Schenk (2005).

Figure 8.2

Shareholder returns of European mergers, 1995–99 (annual cohorts)

and acquiring shareholders were higher for mergers that took place during the early stages of these waves. Well-performing acquirers all made their acquisitions during these same stages. In conclusion, it would seem that insofar as we are concerned with large mergers, the most robust discriminator of success and failure is intertemporality: the further down the merger wave, the more disappointing economic results become. Private, mostly small, firms, appear to be able to escape from this regularity (Moeller et al., 2003a) but, as has been said above, these mergers typically do not create any competition issue so that they are very unlikely to show up in merger control procedures. The literature, of course, has more on offer (for an exhaustive overview, see Schenk, 2002b), yet most relationships found would seem less relevant for our purposes. For example, it has been found that mergers that have been paid for in cash perform better than those that have been settled by paper (that is, via exchange of stock). Obviously, this suggests that acquiring managers undertake more risky acquisitions if they can load parts of the acquisition risk (concerning the combined unit’s future performance), on the shoulders of target shareholders. However, if the lesson to draw from this would be that paperfinanced mergers should be resisted, then many large mergers would no longer be feasible. During the fifth merger wave, for example, more than 65 per cent of mergers (by deal value) were financed by stock. It has also been suggested that horizontal mergers are less likely to fail than conglomerate mergers. Vertical mergers would promise highest chances for success. Unfortunately, the number of vertical mergers has in most studies been so small that it really is not possible to arrive at robust conclusions that can be generalized. The fact that horizontal mergers, indeed, perform better than conglomerate mergers may have much to do with the greater opportunities for market dominance (Megginson et al., 2003). It remains hard to say which form of merger would dominate in terms of efficiency since the distinctions between them have increasingly become less meaningful as many mergers share horizontal and conglomerate as well as vertical characteristics (Karier, 1993).

160

International handbook on industrial policy

Two more issues that could be of importance in establishing the foundations for a general presumption with respect to merger efficiencies deserve attention. First, it would be interesting to know whether the conclusion drawn above would change if industry type were to be introduced as an additional explanatory variable (that is, breaking this further down than to the level of manufacturing versus services as has been done above with respect to banking). Secondly, it would be important to know whether a firm’s merger experience would be positively related to the rate of merger success (in terms of efficiency). Unfortunately, to our knowledge, there are no reliable statistical studies available with respect to the first question. Although it is known that the sectoral distribution of mergers has changed considerably over time, effect studies concerning mergers of the 1920s were as pessimistic as those concerning mergers of the 1960s or of the 1990s (for the earlier periods, see, for example, Borg et al., 1989; Dewing, 1921; Scherer and Ross, 1990). This would seem to suggest that there is no industry specificity in the prevailing success ratios. Merger experience appears to be no guarantee that a next merger will have a higher chance of success than would be the case for a first acquisition. It should be noted, though, that merger programme research is from a statistical point of view an especially delicate matter as it is very difficult to distinguish intertemporal effects (see above) from programme effects. In view of conflicting findings, the number of studies has also been insufficiently impressive to be able to pass judgment. Thus Haleblian and Finkelstein (1999) mention a few studies – either event studies or case studies – that find no significant association between merger experience and performance, as well as a few that do. More recently, Conn et al. (2005) find little difference between the relative performance effects of single and multiple acquirers. They assess performance impacts in terms of short- and long-run share returns and the profit margin on sales. The empirical analysis makes use of a specially constructed data set covering 1476 UK public firms that carried out over 4000 acquisitions during 1984–98. Of these firms 974 carried out more than one acquisition over the period. The acquirers include the most takeover-intensive companies of the 1980s and 1990s merger waves such as ICI, Allied Lyons and Hanson Trust. Conn et al. (2005) do find, however, that, for multiple acquirers with successful first acquisitions, short- and long-run performance declines significantly with each subsequent acquisition. By implication, the conclusion is that initial merger success is no guarantee at all of subsequent merger success – on the contrary. Finally, and although one might have justified doubts with respect to the specifics of meta-analyses, it is worthwhile to refer to a recent study that assessed the added effects of 93 studies with 852 effect sizes (that is, germane bivariate correlations) with a combined n size of 206 910, where n was derived from adding the number of companies on which each of the 93 studies relied (King et al., 2004). Observed zero-order correlations between the variables of interest were weighted by the sample size of the study in order to calculate a mean weighted correlation across all of the studies involved. The sample included both shareholder and real value studies (with the latter limited to studies of the effects on return on assets, return on equity and return on sales). Abnormal (shareholder) returns for acquiring firms appeared to be only positive and significant at day 0. Except for an insignificant positive effect for an event window of one to five days, all others were negative and significant (for event windows of 6–21 days; 22–180 days; 181 days to three years; and greater than three years). Similarly, all results for acquiring firms’ return on assets (ROA),

Mergers and concentration policy 161 return on equity (ROE) and return on sales (ROS) were either insignificant or negative. Thus, and according to the authors, ‘collectively, these results imply that anticipated acquisition synergies are not realized by acquiring firms. [ . . . ] M&A activity does not create superior post-acquisition performance for acquiring firms’ (King et al., 2004, p.192). Innovation Let us turn now to the evidence on innovation effects. Productivity effects remaining at bay, firms are increasingly rationalizing their mergers by referring to dynamic efficiency effects. Unfortunately, only a few studies have endeavoured to find relationships between merger and innovation. From these it appears likely that mergers and acquisitions have a negative effect on R&D investments, R&D investments relative to the industry average, and R&D output except for some industries, most notably the chemical industry (Scherer, 1984; Ravenscraft and Scherer, 1987). Hitt et al. (1991) studied the effects of 191 US acquisitions on both R&D expenditures and results. Total R&D expenditures were divided by total sales and adjusted for average industry R&D intensity. R&D performance was expressed as the number of patents registered divided by total sales. The results show that the acquisition variable, after size, leverage, return on assets and liquidity were controlled for, was a statistically significant, negative predictor of R&D intensity adjusted for industry. The results for patent intensity were similar, so that it was possible to conclude that mergers do not necessarily have synergetic effects in terms of innovation. In a follow-up study, Hitt et al. (1996) confirmed this for a sample of 250 firms for which R&D data were available for 1985–91. Again, a significantly negative relationship between acquisition and innovation was found. Similar, but mostly insignificant results were found by Hall (1999) in a study of 6000 quoted firms. This somewhat weaker result can probably be explained by the fact that Hall’s sample included many fourth-wave mergers that were strongly leveraged. Indeed, she did find a strong, negative connection between leverage and post-merger R&D investments. According to Hall, her results clearly demonstrate that mergers do not by definition have positive economies of scale or scope effects on R&D. The results applied for classical as well as for so-called ‘R&D-intensive’ industries. Additional arguments would have to draw on indirect evidence, that is, on evidence concerning the relationship between size and innovation. Obviously, mergers lead to increases in scale, so that measuring the effects of scale increases could be valuable as a proxy for the effects of merger. One should be aware, however, of the potential biases of such an approach. Similar to findings with respect to the relationships between degree of diversification and performance (internally generated diversification is much less risky than diversification that has been created by merger: see Schenk et al., 1997), an increase in scale might have positive effects on innovation only when it has been generated internally instead of through merger. In any case, Schmookler’s well-known (1972) study confirmed an even earlier and at the time quite surprising study by Cooper (1964) that had found small firms to be 3 to 10 times more productive in development than large firms. Schmookler subsequently found that firms with more than 5000 employees trail small firms in the number of patented inventions, the percentage of patented inventions used commercially and the number of significant inventions. Anecdotal interview evidence gathered by Cooper had suggested that small firms would be able to hire more productive and cost-effective R&D personnel because of better working and incentive conditions.

162

International handbook on industrial policy

More recent findings would seem consistent with this. Small and medium-sized firms in general appear relatively innovative and/or efficient with respect to innovation (Acs, 1996; Nooteboom and Vossen, 1995; Simon, 1996). For example, in the USA, SMEs as a group have been found to be slightly less than 2.4 times more innovative than large firms when innovativeness is measured as the number of innovations per employee (Acs and Audretsch, 1990). Also it appears that SMEs are able to utilize more effectively the knowledge spillovers from universities (Acs et al., 1994). It should be noticed, though, that the high average relative performance of SMEs conceals variations between industries. According to Acs (1996), small firms produced more than half of all innovations recorded in computers and process control instruments as well as in the engineering and scientific instruments industries. By contrast, in the pharmaceutical preparation and photographic equipment industries, large firms appear to be more innovative. In general, however, there does not appear to be any evidence that increasing returns to R&D expenditures exist in producing innovative output. In fact, with only a few exceptions, diminishing returns are the rule. Summary and implications Most real value studies show that mergers do not have significant economic benefits or even generate economic damage. Post-merger profitability compares unfavourably with pre-merger profitability and/or with what would have been the case without the merger. Shareholder value studies find in a large majority that investors expect acquiring firm performance to decline, especially once they have had the time to assess a merger (or a sequence of mergers) for longer than one or a few days. Of course, changes in profitability or market value are determined by other factors than efficiency alone, most notably by changes in market power. Yet, as Meeks (1977) and Mueller (1989) concluded, since it is unlikely that the market power of merging firms declines after merger, any decline in profitability can be taken to indicate a decline in efficiency. This leaves the analysis open, however, to the possible claim that mergers are a response to increased competition, so that a decline in profitability would be expected anyway. Although such claims cannot be directly studied, the studies that were reported on banking as well as studies such as Cowling et al. (1980) do not leave many hopes. These studies have confirmed that, on average, mergers correlate with deteriorating or lagging profitability beyond what could be expected with a view to current changes in the competitive environment as well as deteriorating or lagging productive efficiency. Thus the evidence reported above must be seen as confidently suggesting that large firms are not good at creating efficiencies of whatever kind through merger. Interestingly, they are not good at creating monopoly rents through merger either, because in that case one would expect to find some superior profitability performance at the least. It is important to notice that the results found would seem to be structural. By this we mean that they are not just a coincidence, not just a feature of mergers in a particular era, or a particular industry. Rather, these results are common and recurring whenever a merger wave develops. Obviously, when most but not all mergers fail to generate profits-cum-efficiency, it becomes of importance to learn which factors are associated with success or failure. Unfortunately, the economics literature has not been able to produce systematic evidence in this respect, except for studies that tracked productivity effects in cases in which specific plants were transferred from one owner to another. These latter studies, though being

Mergers and concentration policy 163 quite limited in terms of numbers, have shown that plants continued to increase productivity after a change in ownership. Some of the studies, however, included asset sales from one subsidiary of a firm to another, as well as mergers between single-plant, thus mainly small, firms. Evidently, and as has already been set out above, such studies are less relevant from the present point of view.3 With respect to ‘real’ mergers, the meta-analysis performed by King et al. (2004), concludes that no significant effects on post-merger performance were found for four frequently investigated variables, viz. type of firm (conglomerate v. specialized); relatedness between target and acquiring firm (in terms of resource or product-market similarity); method of payment (cash v. equity); and acquisition experience. As suggested above, similar conclusions were already drawn by Schenk (2002b), except for method of payment. He finds that mergers that have been paid for in cash usually perform better than those that have been settled with equity (also see Rappaport and Sirower, 1999). The sheer size of some large mergers, however, would make it virtually impossible to find the necessary cash (for example, Vodafone’s acquisition of Mannesmann in 1999 involved an exchange of at least US$175 billion in equity). However, the findings also raise more fundamental questions. If mergers that do not create wealth, or even destroy it, are so common and recurrent, one is led to accept one of two propositions. Either corporate executives are not sufficiently equipped to run the firms they are heading, or they do not care as much about the economic results of their actions as economic theory predicts they should. In more general terms, the question becomes how to explain what I have called ‘the merger paradox’. 3 A theory of purely strategic mergers4 Approaches that rest on economic utilitarianism and methodological individualism have great difficulties in coping with economic subjects that repeatedly engage in modes of conduct that do not maximize economic returns. This is especially the case if, for example, based on historical evidence, it is common knowledge that the chances for wealth creation are systematically small, so that one cannot speak of aberrations, and if the particular modes of conduct are widespread, as is the case with mergers. Two solutions to the merger paradox have meanwhile come to be accepted on a fairly broad scale. Owners and managers Already back in the 1930s, Berle and Means (1932) observed that for joint-stock firms ownership had come to be largely separated from control. This opened the possibility of a conflict of interest between principals (owners) and their agents (managers). Whereas owners were assumed to have as their sole interest the maximization of profits, managers might aim for the maximization of personal utility, for example through steadily pushing for larger size (which was assumed to be positively correlated to managerial income) or for perquisites (which would add to managerial status). Faced with disappointing merger results, agency theory soon proposed that managers were undertaking mergers in order to boost their firm’s size rather than profits, while using up funds that should have been distributed to shareholders. In addition, whereas principals are expected to be risk-neutral since they can diversify their shareholdings across multiple firms, agents are assumed to be risk-averse as their jobs and incomes are inextricably tied to one firm. This would imply that, apart from the

164

International handbook on industrial policy

size effect on income, uneconomic mergers would be undertaken in order to prevent loss of job and/or status. To a certain extent, the empirical evidence is consistent with agency theory expectations. Managerial income and perks, as well as status, indeed, rise with the size of the firm, especially if size has been generated by acquisitions (Schmidt and Fowler, 1990). But mergers may threaten agents’ employment security, as becomes evident from the fact that many CEOs are laid off once the merger wave is over and firms come to realize that many acquisitions have been a waste of funds or have even brought counterproductive results. Moreover, the picture that is depicted of managers is particularly negative. It is somehow hard to believe that the large number of uneconomic mergers should be explained by the fact that managers are disguising and distorting information and misleading or cheating their principals. On the contrary, managers may be just like ordinary people: they may enjoy performing responsibly because of a personal need for achievement, while interpreting responsibility as something that is defined in relation to others’ perceptions. More generally, if the principal–agent relationship is conceived in terms of enlightened self-interest, it may be difficult to decouple an agent’s goals from those of her principal (see Wright et al., 2001). Finally, there is not much evidence that managers would only cheat their principals when they are faced (or are expecting to be faced) with rising profits that should be kept away from them. In fact, many of the most profligate cases of cheating occurred when profits were decreasing rather than increasing (see Brenner, 2002). Hubris Others have tried to explain the apparent paradox by suggesting that hubris may lead managers to expand company size through mergers beyond that which maximizes real shareholder wealth, and/or to disregard dismal experiences with earlier mergers. According to Roll, ‘If there are no aggregate gains in takeover, the phenomenon depends on the overbearing presumption of bidders that their valuations are correct’ (1986, p.200). Such overconfidence may grow when past success (even if this was quite coincidental) leads to a certain degree of arrogance and a feeling of supremacy which in turn leads to overpayment. Indeed, the height of bidding premiums appears to depend on whether the bidders can boast a successful pre-merger record in terms of market-to-book and price-to-earnings ratios (Raj and Forsyth, 2003). Prior success breeds overpayment, smaller bidder returns and higher target returns, thus relative failure. Not only were the premiums paid by hubris firms on average 1.5 times higher, their acquisitions were also paid for with paper in 64 per cent of the cases, whereas the control group managed just 23 per cent. An earlier project by Hayward and Hambrick (1997) had used two more indicators of hubris, viz. the extent of recent media praise for the CEO and the size difference between the CEO’s pay and the other executives’ pay in their firms. They reached similar conclusions.5 Malmendier and Tate (2003) classified CEOs as overconfident when they held company options until expiration. Such CEOs were found to be more likely to conduct mergers while the market reacted more negatively to their takeover bids relative to those of others. In conclusion, it seems evident that both agency and hubris theory, and the latter in particular, would play an important part in an explanation of uneconomic mergers and acquisitions. However, both are static theories. Clearly, in cross-sections empire builders

Mergers and concentration policy 165 as well as hubris CEOs will be found to run the highest risk of merging their firms to the brink of failure. But this cannot account for the fact that empire building and overconfidence become manifest only under particular circumstances. Also they take an individualistic point of view, tacitly assuming that a decision maker’s actions are independent from those of others. Thus, while possibly correct in a substantial number of cases, agency and hubris theories cannot by themselves clarify why mergers should occur in waves.6 What is needed, therefore, is a theory that can explain why firms display uneconomic behaviour (or at least uneconomic mergers) erratically and why they do so at approximately identical intervals. The following section, which draws on Schenk (2002b), proposes such a theory of mergers. Rooted in both agency theory and the theory of games, but interpreted with the comments that have been made in mind, this theory suggests that many mergers are undertaken for purely strategic instead of economic motives. For reasons to be developed, these so-called ‘purely strategic mergers’ call for a special policy approach. According to DiMaggio and Powell (1983), uncertainty or lack of understanding with respect to goals, technologies, strategies, pay-offs and the rest – all of them typical of modern industries – are powerful forces that encourage imitation. Following Cyert and March (1963), they suggest that, when firms have to cope with problems with ambiguous causes or unclear solutions, they will rely on problemistic search aimed at finding a viable solution with little expense. Instead of making decisions on the basis of systematic analyses of goals and means, organizations may well find it easier to mimic other organizations. Most ‘important’ mergers are undertaken by large firms. These firms normally operate in concentrated industries and are usually active in several of those industries at the same time (see, for example, Karier, 1993). In the typical situation of single market or multimarket oligopoly, which involves both interdependence of outcomes and strategic uncertainty, adopting mimetic routines is therefore a likely way of solving strategic decision-making problems. Moreover, organizations with ambiguous or (potentially) disputable goals will be likely to be highly dependent upon appearances for legitimacy. Reputation This latter point is also implied in Scharfstein and Stein (1990). Their model assumes that there are two types of managers, ‘smart’ ones who receive informative signals about the value of an investment (such as a merger), and ‘dumb’ ones who receive purely noisy signals. Initially, neither these managers nor other persons (stakeholders) can identify the types, but after an investment decision has been made, stakeholders can update their beliefs on the basis of two pieces of evidence: whether their agent has made a profitable investment and whether their agent’s behaviour was similar to or different from that of other managers. Given the quite reasonable assumption that there are systematically unpredictable components of investment value, and that, whereas ‘dumb’ managers will simply observe uncorrelated noise, ‘smart’ managers tend to get correlated signals since they are all observing ‘a piece of the same “truth” ’ (Scharfstein and Stein, 1990, p.466), it is likely that the second piece of evidence will get precedence over the first. Since these signals might be ‘bad’ just as well as ‘good’, ‘smart’ managers may, however, all have received misleading signals. Since stakeholders will not be able to assess or even perceive these signals they will refer to the second piece of evidence in assessing the ability of ‘their’ managers. Now, if a manager is concerned with her reputation with stakeholders, it will be natural for her

166

International handbook on industrial policy

to mimic a first-mover as this suggests to stakeholders that she has observed a signal that is correlated with the signal observed by the first-mover, which will make it more likely that she is a ‘smart’ manager. The more managers adopt this behaviour, the more likely it will be that ‘bad’ decisions will be seen as a result of a common unpredictable negative component of investment value. The ubiquitousness of the error, in other words, will suggest that all managers were victims of a ‘bad’ signal. Erring managers will subsequently be able to share the blame of stakeholders with their peers. In contrast, a manager who takes a contrary position will ex ante be perceived as ‘dumb’. She will therefore be likely to pursue an investment opportunity if peers are pursuing that, even if her private information suggests that it has a negative expected value. Thus Scharfstein and Stein’s model explains why conventional wisdom teaches that ‘it is better for reputation to fail conventionally than to succeed unconventionally’ (Keynes, 1936, p.158). Rational herding This result, however, is not generally dependent on reputational considerations. Whereas Scharfstein and Stein’s model is essentially an agency model in which agents try to fool their principals and get rewarded if they succeed, Banerjee (1992) and Bikhchandani et al. (1992) have addressed the imitation phenomenon as a consequence of informational externalities. In these models each decision maker looks at the decisions made by previous decision makers in taking her own decision and opts for imitating those previous decisions because the earlier decision makers may have learned some information that is important for her. The result is herd behaviour, a behavioural pattern in which everyone is doing what everyone else is doing. These models are essentially models which explain why some person may choose not to go by her own information, but instead will imitate the choice made by a previous decision maker. Following Banerjee (1992), suppose that, for some reason, the prior probability that an investment alternative is successful is 51 per cent (call this alternative i1), and that the prior probability that alternative i2 is successful is 49 per cent. These prior probabilities are common knowledge. Suppose further that of ten firms (firms A, B, . . ., J), nine have received a signal that i2 is better (of course, this signal may be wrong) but that the one firm which has received a signal that i1 is better happens to choose first. The signals are of equal quality, and firms can only observe predecessors’ choices but not their signals. The first firm (firm A) will clearly opt for alternative i1. Firm B will now know that the first firm had a signal that favoured i1 while her own signal favours i2. If the signals are of equal quality, these conflicting signals effectively cancel out, and the rational choice for firm B is to go by the prior probabilities, that is, to choose i1. Her choice provides no new information for firm C, so that firm C’s situation is not different from that of firm B. Firm C will then imitate firm B for the same reason that prompted firm B to imitate firm A, and so on: all nine follower firms will eventually adopt alternative i1. Clearly, if firm B had relied fully on her own signal, her decision would have provided information for the other eight firms. This would have encouraged these other firms to use their own information. Thus, from a broader perspective, it is of crucial importance whether firm A’s decision is the correct decision. If it is, then all firms will choose the ‘right’ alternative, but if it is not, all firms will end up with a ‘wrong’ decision. Also the result of this game is depend-

Mergers and concentration policy 167 ent on chance: were firm (B, . . . , J) to have had the opportunity to choose first, things would have come out entirely different. However, when translated into our merger problem, if alternative i2 is set equal to ‘do not undertake a merger’ then A’s action (‘merger’) will always be the first to be observed as a deviation from actual practice, thus prompting firms (B, . . . , J) to respond. The mechanism is especially clear when a first and a second firm have both chosen the same i0 (where the point 0 has no special meaning but is merely defined as a point that is known, that is, observable, to the other firms). That is, the third firm (firm C) knows that firm A must have a signal since otherwise she would have chosen i 0. Firm A’s choice is therefore at least as good as firm C’s signal. Moreover, the fact that B has followed A lends extra support to A’s choice (which may be the ‘wrong’ choice nevertheless). It is therefore always better for C to follow A. The main virtues of Banerjee’s model are (a) that some aspects of herd behaviour can be explained without invoking network externalities, that is, without requiring that a decision maker will actually benefit from imitating earlier decision makers (which would be the case if undertaking some action is more worthwhile when others are doing related things); and (b) that it is possible that decision makers will neglect their private information and instead will go by the information which is provided by the actions of earlier decision makers (or the prior probabilities). Cascades Bikhchandani et al. (1992) use the metaphor of a cascade to stress essentially the same point. The process is depicted as a cascade since, with increasing numbers of decision makers adopting a particular action, it becomes increasingly, that is more than proportionally, likely that the next decision maker will follow suit. According to Bikhchandani et al., a cascade will start if ‘enough’ predecessors have all acted in contradistinction to a subsequent decision maker’s own information and if there is no a priori reason to expect that the signals received by the earlier decision makers are less valuable than the signal received by the subsequent decision maker. The first condition is dependent on the specification of the model. The latter condition is an assumption of the model (but can be adapted by introducing variations in signal strength). Ultimately, the reason that a decision maker will tend to disregard her own information is that she is sufficiently uncertain about the value of her signal to act upon it when faced by the decisions of others. Alternatively, it could be argued that she is simply economizing on the costs that are involved in gathering and processing information. Observing the choices of others and imitating these may be a cheap and helpful alternative in the light of the many uncertainties that are involved with strategic decision making. Regret The models discussed so far make clear that the intricacies of information diffusion in sequential games can cause imitation despite the fact that a follower’s private information would indicate a deviation from the trajectory that seems to have been started. Notice, however, that they are couched in a positive pay-off framework. Furthermore, they make use of binary action sets implying that only correct and incorrect decisions are possible and that a small mistake incurs the same loss as a large mistake. The introduction of a regret framework relaxes these conditions and increases the plausibility of models of herding behaviour. In a seminal series of experiments Kahneman and Tversky (1979)

168

International handbook on industrial policy

found that people systematically violate two major conditions of the expected utility model’s conception of rationality when confronted by risk, viz. the requirements of consistency of and coherence among choices. They traced this to the psychological principles that govern the perception of decision problems and the evaluation of options. Apart from the fact that it appears to matter substantially in which frame a given decision problem is couched (or presented, formulated), even to the extent that preferences are reversed when that frame is changed, choices involving gains are often risk-averse and choices involving losses risk-taking. Thus it appears that the response to losses is more extreme than the response to gains. Kahneman and Tversky’s ‘prospect theory’, of course, is consistent with common experience that the displeasure associated with losing a sum of money is greater than the pleasure associated with gaining the same amount. Consequently, it is likely that the contents of decision rules and standard practices will be ‘biased’ in such a way that they favour the prevention of losses rather than the realization of gains. Thus behavioural norms that carry this property are more likely to be ‘chosen’ as Schelling’s so-called ‘focal points’ (Schelling, 1960). In practice, this will mean that firms are likely to adopt routines that imply a substantial degree of circumspection. A similar degree of circumspection is likely to develop if the decision maker is concerned with the regret that she may have upon discovering the difference between the actual payoff as the result of her choice and ‘what might have been’ the pay-off were she to have opted for a different course of action. Regret in this case may be defined as the loss of pleasure due to the knowledge that a better outcome might have been attained if a different choice had been made. Under conditions of uncertainty a decision maker will modify the expected value of a particular action according to the level of this regret. Minimax regret Dietrich and Schenk (1995), building on Savage (1951) and Loomes and Sugden (1982), have suggested that one way of expressing this is by adopting a minimax regret routine. Let us assume that a decision maker knows the pay-offs for each decision alternative but that she is completely ignorant as to which state of nature prevails. The minimax regret routine then prescribes that she select that strategy which minimizes the highest possible regret assuming that the level of regret is linearly related to the differences in pay-off. The minimax regret criterion thus puts a floor under how bad the decision maker would feel if things went wrong. Moreover, doing so will protect her against the highest possible reproach that can be made by those stakeholders who assess the decision’s utility on the basis of the true state of nature. When put into a framework of competitive interdependence this develops as follows. Given a particular action of firm A which is important enough to be monitored by her peers (rivals), that is, a merger or an acquisition, firm B will have to contemplate what the repercussions for her own position might be. Suppose that there is no way that firm B can tell whether A’s move will be successful or not. A’s move could be genuinely motivated by a realistic expectation that her cost position will improve, or by a realistic expectation that her move will increase her rating with stakeholders or even her earnings. That is, A’s competitiveness position vis-à-vis her peers may be ameliorated as a result of that move, say in terms of a first-mover advantage. But then again, it may not. For example, A’s move may be motivated purely by the pursuit of managerial goals, or it may simply be a miscalculation caused by hubris. What is firm B to do?

Mergers and concentration policy 169 Suppose that A’s move will be successful, but that B has not reacted by imitating that move herself (which we will call scenario ). To what extent will B regret not having reacted? Alternatively, suppose that A’s move will not be successful but that B has imitated it solely inspired by the possible prospect of A’s move being a success (scenario ). To what extent will B regret this when the failure of A’s move becomes apparent? Within a minimax regret framework, it is likely that B’s regret attached to scenario  will be higher than the regret attached to scenario . For in scenario , B will experience a loss of competitiveness, while in scenario  her competitive position vis-à-vis A will not have been harmed. Of course, B could have realized a competitive gain in scenario , had she refrained from imitation, but in terms of the minimax regret model her regret at having lost this potential gain is likely to be relatively small. The implication is that, under conditions of uncertainty, a strategic move by firm A will elicit an imitative countermove by her rivals, even if the economic pay-offs are unknown. The models discussed assume that a decision maker’s pay-offs do not depend on what subsequent decision makers do, so that there is no incentive to cheat in an effort to influence a later player. Moreover, decision makers are not allowed to have heterogeneous values of adoption, yet these models allow us to conclude that it is likely that a decision maker who is using a minimax regret routine will imitate actions of earlier decision makers that are regarded as significant. Thus if, for some reason, a first decision maker within a strategic group has decided to undertake a merger, a second decision maker may follow suit even if her own information suggests otherwise. Evidently, such imitation may lead to cascades that will last very long, if not forever. In a sense, mergers and acquisitions have then become ‘taken-for-granted’ solutions to competitive interdependence. This implies that firms may have become locked into a solution in which all players implicitly prefer a non-optimal strategy without having ready possibilities for breaking away from it. Even if some firms do not adopt minimax regret behaviour, it will be sensible for them to jump on a merger bandwagon too, for cascading numbers of mergers and acquisitions imply that the likelihood of becoming an acquisition target increases. Thus, given the finding that relative size is a more effective barrier against takeover than relative profitability (Hughes, 1993; Dickerson et al., 2003), firms may enter the merger and acquisition game for no other reason than to defend themselves against takeover (also see Greer, 1986). Needless to say, such defensive mergers will amplify the prevailing rate of mergers and acquisitions. The cascade will inevitably stop as soon as (a) the number of potential targets diminishes, which is a function of the intensity of the cascade, and (b) the disappointing merger returns decrease the chances of obtaining the financial means that are necessary for further merger investments. In conclusion, it would seem that the high incidence of non-wealth-creating mergers is not the result of failed implementation techniques, or diverging organizational cultures, as many management scholars would like us to believe. Rather the existence of strategic interdependence under uncertainty, conditioned by the availability of funds, may compel managements to undertake mergers even if these will not increase economic performance. Inertia may prevail for long periods, but as soon as an initial, clearly observable move has been made by one of the major players, it is likely that other players will rapidly follow with similar moves. With multi-market oligopoly omnipresent, and given the increasing weight assigned to stock market performance appraisals, the ultimate result may be an economy-wide merger boom. Eventually, many firms will find themselves stuffed with

170

International handbook on industrial policy Stage 1

Preconditional stage A booming economy provides the necessary means (cash; stock appreciations; borrowing facilities), but is not sufficient

Stage 2

Event stage A single (random) merger ignites the game

Stage 3

Response stage Minimax-regret and defensive routines lead to bursts of merger activity

Stage 4

Depletion stage The merger boom levels off as a result of lacking and/or lagging productivity/profitability gains, and price rises for targets

Stage 5

Recovery stage Reconstitution management sets in (sell-offs, divestitures, demergers; lay-offs)

Stage 6

Normalization stage The pool with targets is refilled

Source: Schenk (2005).

Figure 8.3

Stages of the restructuring wave

acquisitions that were neither meant nor able to create wealth. As a consequence, after the strategic imperatives have receded, firms will start licking their wounds by undertaking corrective actions. In the short run, they are likely to look for cheap and easy alternatives, like economizing on all sorts of expenses (such as labour or R&D). In the medium run they will likely spin off the acquisitions done during the boom, sometimes at great cost. Figure 8.3 depicts the different stages of the restructuring wave. I propose to call such mergers ‘purely strategic mergers’. These are mergers that are intended to create strategic comfort when faced with the uncertain effects of a competitor’s moves, rather than economic wealth (or, for that matter, monopoly rents). It is precisely for this reason that it would be futile to wait on the so-called ‘learning capacities’ of organizations to improve economic merger performance. In a system that is dominated by the few, such purely strategic mergers are simply part of the game and since these mergers on average may only turn out to be wealth-creating by chance, uneconomic mergers will also be the order of the day, more precisely: whenever firms are baiting each other into a merger wave. 4 Merger control Ideally we would want to have a regulatory system that encourages firms to arrange economically beneficial mergers while at the same time discouraging them from undertaking harmful mergers. Notice that this is not quite the same as saying that harmful mergers should simply be forbidden. Accepting harmful mergers might be the price society has to pay for other mergers to succeed. Since economic decision making may involve risk, which certainly is the case if such decision making concerns entrepreneurial or innovative moves, and since it is as good as impossible – as we have seen – to predict which mergers

Mergers and concentration policy 171 will be successful, society may choose to forgo limiting the freedom of firms to undertake mergers. Evidently, in order for this position to be tenable, the harm done by uneconomic mergers must be smaller than the benefits from successful mergers. If, on the other hand, mergers harm the economy more than they benefit it, policy intervention would seem to be warranted. Welfare losses Interestingly, this question has hardly ever been researched. Ravenscraft and Scherer (1987) estimated welfare losses from uneconomic mergers by assuming that profitability movements are reflections of changes in operating efficiency relative to industry norms. They studied the efficiency effects for 634 manufacturing acquisitions done in 1968, 1971 and 1974, separating out those acquisitions that were sold off later from those that were retained. Calibrating the results for 1975–77, they came up with total efficiency losses per annum of between US$2.41bn and US$3.32bn, after having taken into account the gains that were generated by, especially, mergers of equals. Relating these losses to expenditures on mergers in the years mentioned, amounting to about US$57bn (see Schenk, 2002a), gives negative returns on merger investments of between 4 and 6 per cent annually during 1975–7. Total (cumulative) efficiency losses will have been higher, but it is not possible to say by how much. Speculating a bit further on this, one can also take failure estimates as an opening wedge. Repeatedly, these rates have been found to be in excess of 65 per cent, up to as high as 83 per cent (Schenk, 2002b). For the years 1995–9, this would imply that, out of total merger investments of about US$9000bn, roughly between US$5500bn and US$7500bn would have been invested with either zero or negative returns. Including the last year of the rising part of the wave, 2000, adds merger expenditures of more than US$2000bn. In turn, this leads to failure estimates of between US$7000 and US$9250bn. Part of this is a real (efficiency) loss; another part is mainly opportunity cost. Unfortunately, the distribution is unknown so that it is near to impossible to estimate which gains might have been generated by alternative deployments of investment funds. However, applying the same proportions as used above (having eliminated the gains from mergers of equals as we are now exclusively dealing with failure, leading to 5.2 and 6.8 per cent, respectively), we would get estimates of annual efficiency losses of somewhere between US$350bn and US$600bn. The fifth merger wave might therefore have implied efficiency losses to the amount between US$2100bn and US$3600bn, with the lion’s share falling towards the last two years of the previous century. Since these losses would be real efficiency losses rather than numbers that are based upon perceptions of failure, they would seem quite impressive. For example, they amount to between 2.1 and 3.6 per cent of collective and cumulative GDP of those countries that have acquiring firms present in the merger database (see Schenk, 2005). Losses of this size might be large enough to have recessionary impact, certainly for the leading countries, the United States, Great Britain and the Netherlands. Obviously, applying proportions that were valid in the 1970s to the 1990s requires some quite heroic assumptions. Especially in view of the inflated stock prices during the latter part of the fifth merger wave, however, returns on merger investments are likely to have been even worse. The estimations thus far have referred to real (and opportunity) effects. Can we also estimate the macroeconomic impact from the effects mergers had on stock prices? Moeller

172

International handbook on industrial policy

et al. (2003a) have calculated the shareholder value losses for a sample of 12 023 domestic acquisitions within the USA, spanning 1980–2001, at about US$219bn. There were about 44 250 mergers in the USA during the fifth merger wave (1995–2000), so that by extrapolation we would find a total loss in shareholder value of approximately US$800bn due to mergers. The effects were calculated, however, using a short event window (three days) so that the true loss of wealth is underestimated significantly. Again, quite heroic assumptions are necessary. We have found in our studies that three-year losses are typically seven times larger than two-day losses. If it was justified to use this ratio here, the estimate of true wealth losses would amount to roughly US$5500bn. In terms of US GDP, this would be approximately 10 per cent. In a more recent paper Moeller et al. (2003b) demonstrate that almost all of these losses appear during 1998–2001, that is, when the fifth merger wave was peaking. From data supplied to me by the authors, it can readily be seen that 3772 mergers lost US$240bn for the acquiring shareholders, for an average loss of US$64mn per merger. When the gains to target shareholders are subtracted, the net effect appears to be a loss of US$134bn. Again by extrapolation, I find that, for the approximately 36 350 mergers that took place in the USA during 1998–2001, US$1290bn was lost. Evidently, when multiplied by seven, the wealth losses when based on the most recent fifth wave mergers appear dramatic, at about 16 per cent of GDP. It is uncertain, however, what a wealth loss implies in terms of real effects. It is likely that investors, be they firms, institutional investors or private investors, will react in some way to such a dramatic change in wealth but whether such reactions are significant enough to provoke a recession, for example, is an open question. However, if reactions to these losses were as severe as they were to the wealth gains that were realized when the stock market was booming, then the answer would be in the affirmative. Bulmash (2002), for example, estimated that the effect of the wealth gain of US$6000bn over the period 1997–9 on consumption spending in the USA was more than US$120bn in 1998 and 1999. Over 40 per cent of the growth in consumer spending in 1999 was attributable to gains in the stock market in 1997 and 1998. The wealth losses reported above may have had similar real effects. Assuming gain/loss symmetry, the real effect of wealth losses due to merger on American consumer spending would have been between US$110 and US$180bn annually. Generally, however, the regret attached to losses is much higher than the joy attached to gains (Kahneman and Tversky, 1979) so that it is likely that the effect on consumption spending has been much higher. Notice that in order to reach correct conclusions these losses must be added to the efficiency losses estimated above, to opportunity costs and to delays in investment spending. In this respect it should perhaps be recalled that opportunity costs are, in contradistinction to what one might think, important. If funds do not generate wealth, this implies that they do not create economic growth. It could be argued that the billions expended on mergers do not vanish from the economic process. Indeed, it may be that shareholders at the receiving end, instead of creating a consumption bubble, or overindulging themselves in Veblen-type conspicuous consumption (Veblen, 1899), will reinvest their newly acquired pecuniary wealth in investment projects that do create economic wealth. If so, then we would merely have to worry about a delay effect. Still, such an effect may be significant, since an accumulation of delay effects – and this is exactly what is likely to happen during a merger wave – is what is called a recession.

Mergers and concentration policy 173 Competition policy In a market economy there is a presumption that private agents should be free to pursue their own interests as they see fit up to the point at which this pursuit has (significantly) adverse consequences for economic welfare. From the above it would be obvious to conclude that society could benefit from some sort of intervention in the freedom of firms to pursue mergers and acquisitions. Nowadays, indeed, most countries have laws or regulations that require competition authorities to scrutinize certain mergers. However, most such regulations have a line of approach that would not seem to be consistent with the analysis and findings reported in this contribution. The authorities normally do not examine whether a particular merger is likely to have adverse welfare consequences as a result of productive and/or dynamic inefficiencies. Rather, investigations are meant to establish whether a particular merger is likely to affect welfare because it substantially lessens competition (as in the USA) or significantly impedes effective competition (as in the EU).7,8 The pertinent procedures eventually end up in testing whether a merger will substantially increase market concentration (measured by means of either a concentration ratio or the Herfindahl–Hirschman Index) beyond a certain competitive level and, if so, in estimating whether that can be sustained.9 The underlying assumption is that growing beyond a certain threshold may deliver market power, or market dominance, which is consistently defined as the ability profitably to maintain prices above competitive levels, or above longrun average costs, for a significant period of time (see, for example, FTC, 1997). Such market power is expected to result in a transfer of wealth from buyers to sellers or a misallocation of resources. Therefore, what is being ultimately tested is whether a merger is likely to harm economic welfare by affecting allocative efficiency. Apart from the fact that this is not an easy task (some would say an impossible one; see for example, Dewey, 1996) as a result of both conceptual and measurement problems, related to delineating the relevant product and geographic market as well as to estimating demand substitutability, the likelihood of new entry, and the power of potential competition, these procedures overlook the possibility that mergers may be motivated neither by the prospect of monopoly profits nor by the prospect of productive or dynamic efficiency gains. The underlying presumption is that mergers that are not motivated by the prospect of financial gains will not occur or, if they do, will only be short-lived. As we have seen, this presumption cannot be sustained. Taking uneconomic mergers into account has not become easier as competition authorities have changed their focus from the once cherished public interest criterion to efficiency, productivity and contestability considerations (Hess and Adams, 1999).10 In many cases, moreover, it has become a modern tendency to see as the ultimate goal of competition policy the maximization of consumer surplus. This is clearly a much too narrow interpretation of wealth to be able to deal with mergers and acquisitions that might be harmful to the economy by pushing it into recession. Trying to catch the problem in terms of current competition policy principles requires explaining firm behaviour with regard to mergers and acquisitions in terms of pre-existing joint dominance. The fact that some firms, those that are responsible for the lion’s share of merger and acquisition investments, apparently are able to circumvent the absolute powers of the market when recurrently undertaking uneconomic mergers must mean that either common trust in the forces of this market system is unjustified or that some firms

174

International handbook on industrial policy

are somehow tacitly colluding when a merger wave is in statu nascendi. Reappraising mergers in terms of the public rather than the consumer’s interest, however, would seem to be a more elegant line of approach. Moreover, the fact that many competition authorities, among others those from America and Europe, have meanwhile accepted the idea of an efficiency defence indicates renewed sensitivity to the public interest principle. If because of productive or dynamic efficiencies it would be in society’s interest to accept a merger that would be anticompetitive, on the one hand, then it becomes increasingly relevant to wonder why firms can get away with uneconomic mergers, on the other. If economic efficiency is really the pivot in merger control, then it should work both ways. What I call a ‘Full Efficiency Test’ (FET) could be an efficient instrument in this respect. A Full Efficiency Test is a procedure in which a proposed merger is tested not just for allocative effects but for productive and dynamic effects as well. Merger candidates could be required to join an efficiency protocol or prospectus with their request for allowance, that is, a memorandum that demonstrates whether and how productive and dynamic efficiencies will be realized. As it is in the merger candidates’ interest to come up with a plausible efficiency protocol, it will be evident that the burden of proof in productive and dynamic efficiency cases will rest on them. Small mergers, mergers in which both (or all) parties are small or medium-sized, should be excluded from assessment for two reasons. First, mergers among SMEs are less likely to originate from the nonwealth-creating strategies that are played in small number games. Secondly, SMEs are more likely to operate below minimal optimal scale so that there would likely be opportunities for realizing economies of scale. Notice that focusing on all mergers in which a large firm is implicated would also allow an assessment of piecemeal creations of market dominance, or as it is formulated in the USA, the creation of market dominance in its incipiency. Thus mergers that are unlikely to create productive and/or dynamic efficiency would be blocked while mergers that are likely to have positive efficiency effects would be allowed if no negative allocative efficiency effects are to be expected. Two practical points deserve special attention, however. First, at least during the initial years of an FET regime, the numbers of mergers that would have to be investigated may be large, especially since takeovers of SMEs would need assessment too. A pragmatic solution, to be adapted as the FET regime becomes more established, would be to subject to screening only those mergers that are proposed by the very largest firms from the USA and the EU, respectively. For example, Dewey (1996), who has made somewhat similar suggestions, proposes to limit control to those mergers that are proposed by the 500 largest US industrial firms or so. These firms should not be allowed to grow by merger ‘unless it can be shown that the merger is likely to confer some non-negligible consumer benefit through cost reduction, product improvement, increased research outlays or more rapid innovation’ (Dewey, 1996, p.397).11 Mueller (1997), while keeping up a market share criterion, has proposed disallowing mergers that lead to a combined market share of more than 25 per cent in any one market, or to total sales in excess of US$600mn, unless it can be demonstrated that a particular merger is likely to yield substantial efficiency gains in excess of its anti-competitive effects. Dewey’s proposals imply the more radical departure from established practice but a substantial simplification of it as well, as there is no need whatsoever to assess any longer a merger’s impact on competition. An added advantage would be that there is no need either

Mergers and concentration policy 175 to establish whether a merger is horizontal, vertical or conglomerate. While Mueller’s proposals still require estimating the allocative efficiency effects of a merger, though apparently only in cases where firms can come up with a plausible efficiency prospectus, Dewey’s proposals would perhaps fit in best with the gist of the present contribution. The current EU thresholds may be adequate as far as large mergers are concerned, while the acquisitions of SMEs by large firms could be investigated at the level of individual member states. Besides, and given the empirical evidence on mergers, it may be expected that the number of mergers that will have to be vetted will eventually be reduced dramatically once firms are to come up with a plausible efficiency protocol. Secondly, one may rightfully ask whether the present investigating skills of the merger authorities would allow conducting a sufficiently sophisticated FET. As Scherer (1997) has noted, the detailed analysis of plant-specific and product-specific scale economies, and the technological and organizational determinants of productivity, do not belong to the regular bag of tools of an anti-trust economist. Applying an FET would, indeed, require staff with solid management consultancy experience or training. Such skills, according to Scherer, would be expensive to obtain. While Scherer’s doubts may be justified, applying an FET would be less demanding if the burden of proof is, as has been suggested here, on the shoulders of those firms that wish to undertake a merger or acquisition. Also notice that the transparency of merger controls would be increased substantially by using an FET, which should be welcomed both by the firms involved and by those wishing to investigate the decisions taken by the authorities. The chances for opportunistic settlements and regulatory capture would be reduced significantly. Finally, a rather important advantage of the type of merger control proposed here would be that it reduces the complexities of gearing international merger policies to each other. Apart from the jurisdictional questions involved, there is at present much confusion over the interpretations of different merger control procedures and institutional arrangements (see, for example, Waverman et al., 1997). The steadily increasing number of international mergers and acquisitions implies that firms will increasingly have to report their mergers to different national authorities. Moreover, given the internationalization of the world economy, it becomes increasingly possible that domestic mergers will have foreign repercussions. Indeed, the takeovers of McDonnell Douglas by Boeing and of Honeywell by General Electric became the subject of cross-Atlantic confusion as the European Commission decided, for good reasons, to leave its mark on the finalization of the one and to block the other. The transparency of the pertinent discussions would have benefited substantially from an assessment that would have included an FET. 5 Summary and conclusions This contribution has suggested that the ubiquitousness of failed mergers is not surprising since uneconomic mergers seem a natural result of competition among the few. Such competition encourages strategic rather than economic behaviour; that is, behaviour that is not primarily driven by the wish to create wealth but by the behavioural peculiarities of strategic interdependence. It has been argued that, even if only some firms adopt a minimax regret rationale, others will be forced to jump on merger bandwagons for defensive reasons. Under certain conditions the result will be an extremely costly merger wave. Once this becomes evident, firms need to take corrective action. As a consequence, such

176

International handbook on industrial policy

merger waves are followed by periods of restructuring, large-scale divestment and lay-offs. It has been suggested that the sheer size of the problem may be sufficient to provoke economic recessions. Whereas the observed effects are rooted in the high levels of economic concentration that have become typical of modern economies, therefore a matter of competition policy, current merger regulations may not be the preferred means of control. Merger regulations have been designed to prevent as many harmful mergers as possible while preserving economically efficient mergers. As long as the consumer’s interest remains the main vehicle for defining the wealth of nations, however, competition economists as well as authorities will be led away from the most pervasive problematic effect of mergers. Rather, one would want to see competition policy return to its roots by putting the public interest at centre stage. Notes 1. 2.

In this chapter, we will use the terms ‘merger’ and ‘acquisition’ interchangeably, unless noted otherwise. Moreover, the study suggested that mergers and acquisitions have also led to less product variety; an increase in the rate of technological progressiveness was not observable. 3. Since many divestitures were involved, these studies might, in fact, be interpreted as suggesting that demerger is good for productivity. See, for example, Lichtenberg and Siegel (1990) and McGuckin and Nguyen (1995). 4. This section and the next are based on, and partly taken from, Schenk (2005). 5. Interestingly, the CEO’s pay divided by the average pay for other officers was significantly greater for bankrupt firms five years before they failed than for a matched group of survivor companies (Hayward and Hambrick, 1997). 6. Notice that the relationship between merger waves and stock market dynamics is ambiguous. Sometimes stock market cycles correlate with merger waves; sometimes they do not, whereas if they do it remains uncertain what causes what. Evidently, merger waves must be preceded by periods of economic prosperity. But neither historical evidence nor economic logic would lead one to qualify this as the fulfilment of a sufficient condition. 7. US federal merger control rests on the Clayton Act of 1914 and the Celler–Kefauver Act of 1950 and is enforced by the Federal Trade Commission or the Department of Justice (there is no formal division of labour between these two agencies). The Hart–Scott–Rodino Act of 1976 gave both agencies the power to review all mergers above a certain size threshold; owing to a large number of exemptions it is not easy to fix the threshold, but if a party is involved with sales in excess of US$100mn, the merger may have to be submitted for approval. Merging parties may defend their proposal on the basis of efficiency arguments. 8. The European Community Merger Regulation (ECMR) was first adopted in 1989 and took effect in 1990. By now, in about 2800 cases, firms have been required to submit their mergers before the Commission, out of which about 80 were accepted with remedies and fewer than 20 were blocked, three of which were given de facto clearance later by the Court of First Instance. At present, the ECMR applies to mergers in which the parties when taken together have worldwide sales in excess of €5bn while at least two of them have Community-sales in excess of €250mn, unless all parties have more than twothirds of their EU sales within one and the same member state. In the latter case, the merger would be subject to national merger control. In the course of 1998, a revised scheme became effective which adds to the previous conditions that mergers should be notified to the Commission if parties have combined worldwide sales in excess of €2.5bn, as well as sales in excess of €100mn in each of at least three member states, while at least two of them have aggregate sales within the EU of at least €100mn each, provided that at least two of the parties have sales of at least €25mn in the three member states that were mentioned above. The new scheme has become this cumbersome as a result of negotiations between those member states (and the Commission) that preferred lowering the thresholds while others appeared quite satisfied with the 1989 regulation. The decisions of the Court just mentioned caused the Commission to introduce some substantive changes in 2004. Proposed mergers should now receive more extensive economic (rather than juridical) treatments while firms may call upon efficiency benefits if their merger may not be cleared on competition grounds. 9. The HHI is calculated by adding up the squares of the individual market shares of all participants in a particular market, thus reflecting the distribution of market shares, whereas the concentration ratio only reflects the shares of the top n (usually four) firms. Since the HHI gives proportionately greater weight to the largest firms, it is not necessary to know the market shares of all participants.

Mergers and concentration policy 177 10.

11.

According to Hess and Adams (1999), public interest is essentially an idea with a history closely aligned to democratic public administration. Although the failure of political scientists to identify public interest clearly may have contributed to its demise in competition policy, it is clear that public interest considerations are to be distinguished from private or sectional interests and that they make claims to underlying ‘common’ social values as the basis for justification of action. In this respect, also see Brenner (1990) who stresses the importance of a test for the effects of a merger on innovation performance.

References Acs, Z.J. (1996), ‘Small firms and economic growth’, in P.H. Admiraal (ed.), Small Business in the Modern Economy, Oxford: Blackwell, pp.1–62. Acs, Z.J. and D.B. Audretsch (1990), Innovation and Small Firms, Cambridge: MIT Press. Acs, Z.J., D.B. Audretsch and M.P. Feldman (1994), ‘R&D spillovers and recipient firm size’, Review of Economics and Statistics, 76(2), 336–40. Akhavein, J.D., A.N. Berger and D.B. Humphrey (1997), ‘The effects of megamergers on efficiency and prices: evidence from a bank profit function’, Review of Industrial Organization, 12, 95–139. Banerjee, A.V. (1992), ‘A simple model of herd behavior’, Quarterly Journal of Economics, CVII(3), 797–817. Berger, A.N. (1998), ‘The efficiency effects of bank mergers and acquisition: a preliminary look at the 1990s data’, in Y. Amihud and G. Miller (eds), Bank Mergers and Acquisitions, Dordrecht/Boston/London: Kluwer Academic, pp.79–111. Berger, A.N., R.S. Demsetz and P.E. Strahan (1999), ‘The consolidation of the financial services industry: causes, consequences, and implications for the future’, Journal of Banking & Finance, 23, 135–94. Berle, A.A. and G.C. Means (1932), The Modern Corporation and Private Property, rev. edn 1968, New York: Harcourt, Brace, Jovanovich. Bikhchandani, S., D. Hirshleifer and I. Welch (1992), ‘A theory of fads, fashions, custom, and cultural change as informational cascades’, Journal of Political Economy, 100(5), 992–1026. Borg, J.R., M.O. Borg and J.D. Leeth (1989), ‘The success of mergers in the 1920s. A stock market appraisal of the second merger wave’, International Journal of Industrial Organization, 7(1), 117–31. Brenner, R. (1990), ‘Innovations and anti-trust’, in B. Dankbaar, J. Groenewegen and H. Schenk (eds), Perspectives in Industrial Organization, Studies in Industrial Organization, vol. 13, Dordrecht/Boston/ London: Kluwer Academic Publishers., pp.235–57. Brenner, R. (2002), The Boom and the Bubble. The US and the World Economy, London/New York: Verso. Bühner, R. (1991), ‘The success of mergers in Germany’, International Journal of Industrial Organization, 9(4), 513–32. Bulmash, S.B. (2002), ‘A new paradigm: the wealth effect of the stock market on consumption, in a context of interacting bio-systems’, Journal of Socio-Economics, 31, 75–100. Calomiris, C.W. (1999), ‘Gauging the efficiency of bank consolidation during a merger wave’, Journal of Banking & Finance , 23, 615–21. Carow, K., R. Heron and T. Saxton (2004), ‘Do early birds get the returns? An empirical investigation of earlymover advantages in acquisitions’, Strategic Management Journal, 25, 563–85. Conn, R.L., A. Cosh, P.M. Guest and A. Hughes (2005), ‘Why must all good things come to an end? The performance of multiple acquirers’, Mimeo, Centre for Business Research, University of Cambridge. Cooper, A.C. (1964), ‘R&D is more efficient in small companies’, Harvard Business Review, 42, 75–83. Cowling, K., P. Stoneman, J. Cubbin, J. Cable, G. Hall, S. Domberger and P. Dutton (1980), Mergers and Economic Performance, Cambridge: Cambridge University Press. Cyert, R.M. and J.G. March (1963), A Behavioral Theory of the Firm, 2nd edn 1992, Cambridge, MA: Blackwell. De Jong, H.W. (1993), ‘Combination movements in banking’, in H.O. Lenel et al., (eds), Jahrbuch für die Ordnung von Wirtschaft und Gesellschaft, 43, Stuttgart/Jena/New York: Gustav Fischer Verlag, pp.475–99. Dewey, D. (1996), ‘Merger policy greatly simplified: building on Keyes’, Review of Industrial Organization, 11, 395–400. Dewing, A.S. (1921), ‘A statistical test of the success of consolidations’, Quarterly Journal of Economics, 36, 84–101. Dickerson, A.P., H.D. Gibson and E. Tsakalotos (1997), ‘The impact of acquisitions on company performance: evidence from a large panel of UK firms’, Oxford Economic Papers, 49, 344–61. Dickerson, A.P., H.D. Gibson and E. Tsakalotos (2003), ‘Is attack the best form of defence? A competing risk analysis of acquisition activity in the UK’, Cambridge Journal of Economics, 27, 337–57. Dietrich, M. and H. Schenk (1995), ‘Co-ordination benefits, lock-in, and strategy bias’, Management Report 220, Rotterdam School of Management, Erasmus University Rotterdam. DiMaggio, P.J. and W.W. Powell (1983), ‘The iron cage revisited: institutional isomorphism and collective rationality in organizational fields’, American Sociological Review, 48, 147–60.

178

International handbook on industrial policy

FTC (Federal Trade Commission) (1997), 1992 Horizontal Merger Guidelines [with April 8, 1997, revisions to section 4 on Efficiencies], Washington: Federal Trade Commission (ftc.gov/bc/docs/horizmer.htm). Greer, D.F. (1986), ‘Acquiring in order to avoid acquisition’, Antitrust Bulletin, Spring, 155–86. Gugler, K., D.C. Mueller, B.B. Yurtoglu and C. Zulehner (2003), ‘The effects of mergers: an international comparison’, International Journal of Industrial Organization, 21, 625–53. Haleblian, J. and S. Finkelstein (1999), ‘The influence of productive experience on acquisition performance: a behavioral learning perspective’, Administrative Science Quarterly, 44(1), 29–56. Hall, B.H. (1999), ‘Mergers and R&D revisited’, paper presented at the Quasi-Experimental Methods Symposium, University of California, Berkeley, 3–7 August. Hayward, M.L.A. and D.C. Hambrick (1997), ‘Explaining the premiums paid for large acquisitions: evidence of CEO hubris’, Administrative Science Quarterly, 42, 103–27. Healy, P.M., K.G. Palepu and R.S. Ruback (1992), ‘Does corporate performance improve after mergers?’, Journal of Financial Economics, 31, 135–75. Hess, M. and D. Adams (1999), ‘National competition policy and (the) public interest’, NCDS Briefing Paper 3, National Centre for Development Studies, Australian National University, Canberra. Hitt, M.A., R.E. Hoskisson, R.D. Ireland and J.S. Harrison (1991), ‘Effects of acquisitions on R&D inputs and outputs’, Academy of Management Journal, 34(3), 693–706. Hitt, M.A., R.E. Hoskisson, R.A. Johnson and D.D. Moesel (1996), ‘The market for corporate control and firm innovation’, Academy of Management Journal, 39(5), 1084–119. Hughes, A. (1993), ‘Mergers and economic performance in the UK: a survey of the empirical evidence 1950–1990’, in M. Bishop and J. Kay (eds), European Mergers and Merger Policy, Oxford: Oxford University Press, pp.9–95. Kahneman, D. and A. Tversky (1979), ‘Prospect theory: an analysis of decision making under risk’, Econometrica, 47(2), 263–91. Karier, T. (1993), Beyond Competition. The Economics of Mergers and Monopoly Power, Armonk: M.E. Sharpe. Keynes, J.M. (1936), The General Theory of Employment, Interest, and Money, London: Macmillan. King, D., D.A. Dalton, C.M. Daily and J.G. Covin (2004), ‘Meta-analyses of post-acquisition performance: indicators of unidentified moderators’, Strategic Management Journal, 25, 187–200. Kötter, M. (2005), Efficiency and Mergers in German Banking, unpublished PhD thesis, Utrecht University. Lichtenberg, F.R. and D. Siegel (1990), ‘The effects of leveraged buyouts on productivity and related aspects of firm behavior’, Journal of Financial Economics, 27, 165–94. Loomes, G. and R. Sugden (1982), ‘Regret theory: an alternative theory of rational choice under uncertainty’, Economic Journal, 92, 805–24. Malmendier, U. and G. Tate (2003), ‘Who makes acquisitions? CEO overconfidence and the market’s reaction’, Mimeo, Stanford University, Graduate School of Business (draft). McGuckin, R.H. and S.V. Nguyen (1995), ‘On productivity and plant ownership change: new evidence from the longitudinal research database’, RAND Journal of Economics, 26(2), 257–76. Meeks, G. (1977), Disappointing Marriage: A Study of the Gains from Merger, Cambridge: Cambridge University Press. Megginson, W.L., A. Morgan and L. Nail (2004), ‘The determinants of positive long-term performance in strategic mergers: corporate focus and cash’, Journal of Banking and Finance, 28(3), 523–52. Moeller, S.B., F.P. Schlingemann and R.M. Stulz (2003a), ‘Do shareholders of acquiring firms gain from acquisitions?’, Working Paper 9523, NBER. Moeller, S.B., F.P. Schlingemann and R.M. Stulz (2003), ‘Wealth destruction on a massive scale? A study of acquiring-firm returns in the recent merger wave’, Working Paper 10200, NBER. Mueller, D.C. (1986), Profits in the Long Run, Cambridge/New York: Cambridge University Press. Mueller, D.C. (1989), ‘Mergers: causes, effects and policies’, International Journal of Industrial Organization, 7(1), 1–10. Mueller, D.C. (1997), ‘Merger policy in the United States: a reconsideration’, Review of Industrial Organization, 12, 655–85. Mueller, D.C. (2003), ‘The finance literature on mergers: a critical survey’, in M. Waterson (ed.), Competition, Monopoly and Corporate Governance. Essays in Honour of Keith Cowling, Cheltenham, UK and Northampton, MA, USA: Edward Elgar, pp.161–205. Nooteboom, B. and R.W. Vossen (1995), ‘Firm size and efficiency in R&D spending’, in A. Van Witteloostuijn (ed.), Market Evolution. Competition and Cooperation, Studies in Industrial Organization, vol. 20, Dordrecht/Boston/London: Kluwer Academic Publishers, pp.69–86. Raj, M. and M. Forsyth (2003), ‘Hubris amongst UK bidders and losses to shareholders’, International Journal of Business, 8, 1–16. Rappaport, A. and M. Sirower (1999), ‘Stock or cash? The trade-offs for buyers and sellers in mergers and acquisitions’, Harvard Business Review, November–December, 147–58. Ravenscraft, D.J. and F.M. Scherer (1987), Mergers, Sell offs, and Economic Efficiency, Washington, DC: Brookings Institution.

Mergers and concentration policy 179 Roll, R. (1986), ‘The hubris hypothesis of corporate takeovers’, Journal of Business, 59(2), 197–216. Röller, L-H., J. Stennek and F. Verboven (2001), ‘Efficiency gains from mergers’, Mimeo, Centre for Economic Policy Research, London. Salant, S.W., S. Switzer and R.J. Reynolds (1983), ‘Losses from horizontal merger: the effects of an exogenous change in industry structure on Cournot–Nash equilibrium’, Quarterly Journal of Economics, 98, 185–99. Savage, L.J. (1951), ‘The theory of statistical decision’, Journal of the American Statistical Association, 46, 55–67. Scharfstein, D.S. and J.C. Stein (1990), ‘Herd behavior and investment’, American Economic Review, 80(3), 465–79. Schelling, T. (1960), The Strategy of Conflict, Cambridge, MA: Harvard University Press. Schenk, H. (2002a), ‘Mergers: incidence and trends – an international overview’ in J-C. Bartel, R. Van Frederikslust and H. Schenk (eds), Mergers and Acquisitions: Fundamental Aspects (in Dutch), Amsterdam: Elsevier, pp.37–61. Schenk, H. (2002b). ‘Economics and strategy of the merger paradox’, in J-C. Bartel, R. Van Frederikslust and H. Schenk (eds), Mergers and Acquisitions: Fundamental Aspects (in Dutch), Amsterdam: Elsevier, pp.62–133. Schenk, H. (2005), ‘Organisational economics in an age of restructuring, or: how corporate strategies can harm your economy’, in P. De Gijsel and H. Schenk (eds), Multidisciplinary Economics, Dordrecht: Springer, pp.333–65. Schenk, H., J-P. Warmenhoven, M. Velzen and C. Van Riel (1997), ‘The demise of the conglomerate firm’ (in Dutch), Economisch Statistische Berichten, 82(4122), 736–40. Scherer, F.M. (1984), Innovation and Growth: Schumpeterian Perspectives, Cambridge, MA: MIT Press. Scherer, F.M. (1997), ‘Comment on “merger policy in the United States” ’, Review of Industrial Organization, 12, 687–91. Scherer, F.M. (2002), ‘The merger puzzle’, in W. Franz, H.J. Ramser and M. Stadler (eds), Fusionen, Tübingen: Mohr Siebeck, pp.1–22. Scherer, F.M. and D. Ross (1990), Industrial Market Structure and Economic Performance, Boston: Houghton Mifflin. Schmidt, D.R. and K.L. Fowler (1990), ‘Post-acquisition financial performance and executive compensation’, Strategic Management Journal, 11(7), 559–69. Schmookler J. (1972), ‘The size of firm and the growth of knowledge’, in Z. Griliches and L. Hurwicz (eds), Patents, Invention, and Economic Change, Cambridge, MA: Harvard University Press. Simon, H. (1996), Hidden Champions. Lessons from 500 of the World’s Best Unknown Companies, Boston: Harvard Business School Press. Simon, J.L., M. Mokhtari and D.H. Simon (1996), ‘Are mergers beneficial or detrimental? Evidence from advertising agencies’, International Journal of the Economics of Business, 3(1), 69–82. Tichy, G. (1990), ‘Bankengrösse und Effizienz’, Kredit und Kapital, 3, 358–88. Tichy, G. (2001), ‘What do we know about success and failure of mergers?’, Journal of Industry, Competition and Trade, 1, 347–94. Van der Vennet, R. (1996), ‘The effects of mergers and acquisitions on the efficiency and profitability of EC credit institutions’, Journal of Banking & Finance, 20, 1531–58. Van Rooij, M.C.J. (1997), ‘Bank mergers, banking efficiency, and economies of scale and scope: a review of empirical literature’, Research Memorandum WO&E 511/9726, De Nederlandsche Bank, Amsterdam. Veblen, T.B. (1899), The Theory of the Leisure Class: An Economic Study of Institutions, New York: Macmillan. Waverman, L., W.S. Comanor and A. Goto (eds) (1997), Competition Policy in the Global Economy. Modalities for cooperation, London/New York: Routledge. Wright, P., A. Mukherji and M.J. Kroll (2001), ‘A re-examination of agency theory assumptions: extensions and extrapolations’, Journal of Socio-Economics, 30, 413–29.

9

Liberalization and regulation of public utility sectors: theories and practice Tom Björkroth, Sonja Grönblom and Johan Willner

1 Introduction Traditionally, public utilities were seen as associated with a number of potential market failures that justified public intervention. For example, telecommunications, electricity, railways, gas and water supply typically include elements of a natural monopoly. In the USA, intervention usually meant regulated private monopolies, whereas such industries in Europe were nationalized monopolies until the 1980s. The European approach has now shifted towards competition (with and without privatizing the incumbent) and sometimes privatization with regulation if competition is not possible. State-owned companies and public utilities have been privatized for a variety of reasons, including sales revenues and popular capitalism, or as part of a perceived modernization. But the dominant official motive has in general been the belief that public ownership is less cost-efficient (Kay and Thompson, 1986; Vickers and Yarrow, 1988). But some surveys of public and private ownership, such as Borcherding et al. (1982) and Megginson and Netter (2001) conclude in favour of private ownership, while others, such as Millward (1982), Boyd (1986) and Willner (2001, 2003) do not. Also, a number of studies find no improvement in performance after privatization (on Britain and Finland, see for example Martin and Parker, 1997; Florio, 2004; Willner, 2006a). Privatization has in addition mostly affected monopolistic or oligopolistic industries, so the social costs of, for example, imperfect competition may dominate even if costs are reduced (Willner, 1996, 2003). It is therefore often argued that performance improvements after privatization are explained by competition rather than the change of ownership (Vickers and Yarrow, 1988; Parker, 2005). Insofar as competition is superior to a regulated monopoly, privatization is therefore often seen as part of a liberalization process rather than as an end in itself (Newbery, 2001). However, competition in public utilities may be associated with social costs of its own, so there usually exist a number of trade-offs. We proceed as follows. The next section deals with the basic issue of when competition is desirable in a public utility with sunk costs, and identifies the cost savings that are necessary for competition to outperform a public monopoly. The desirability of competition is addressed within a framework of vertical relations and network externalities in section 3. Section 4 deals with X-inefficiency, incomes distribution and quality. Sections 5–7 analyse telecommunications, electricity, water utilities and public transport in Europe. Section 8 contains some concluding remarks. 2 Sunk costs and the desirability of monopoly and competition Liberalization is inspired by the view of perfect competition as superior even to a welfaremaximizing public monopoly, because of lower costs. However, sunk costs may create a 180

Liberalization and regulation of public utility sectors 181 natural oligopoly, and hence a deadweight loss, which depends on only unnecessary duplication if there is free entry, so there is trade-off. Free entry would then imply too many firms (Mankiw and Whinston, 1986). Denote price and quantity by p and x, and the vertical intercept (shift parameter) of a linear demand curve by a; its slope is normalized to 1. Each firm has constant marginal costs c and fixed costs F. Suppose that liberalization means an n-firm Cournot oligopoly indexed by C, but by M if n 1. The total surplus is then TSC 

(n2
2n) (a  c) 2  nF. 2(n
1) 2

(9.1)

Suppose first that F and c are the same in all types of firms, and subtract TSC from the monopoly total surplus TSM. Rearranging and dividing by the monopolist’s gross profits M (a c)2/4 shows that an n-firm oligopoly outperforms a profit-maximizing monopoly only if (n
3)/2(n
1)2 F. In other words, a duopoly and a three-firm oligopoly are inferior to the monopoly if the sunk cost is larger than 27.7 and 18.8 per cent of the monopolist’s gross profits. If there is entry until n n* and net profits are zero, (a c)2/(n*
1)2
F holds true. Rearranging TSC TSM then implies that n* 5 is sufficient for a welfare improvement.1 In other words, while n* 4 does not imply a natural monopoly, competition is then not desirable. It is obvious that n* depends on a, c and F, so changes in demand and technology can make competition in a former natural monopoly viable, but not necessarily desirable. The latter would not mean a natural monopoly in normal usage (where only one firm can survive), but the normative conclusions are the same.2 Liberalization, however, usually means that a public monopoly which can be given wider objectives is replaced by competition. A welfare-maximizing public monopoly with a break-even constraint would outperform a liberalized industry where F 0 even if n  n*, because of economics of scale, unless liberalization reduces costs. To get an idea of the size of these necessary cost reductions, suppose that the public monopoly maximizes welfare subject to a break-even constraint, and that liberalization means free entry and zero profits. The total surplus then consists only of the consumer surplus in both cases, so a higher output means a higher total surplus. Suppose that the fixed costs before and after liberalization are kF and F, where k 1 (on variable cost differences, see Willner, 1996). Compare the output levels using n*
1
(a c)/ √F . It follows that liberalization then improves welfare if c a  c  √F a  2


√  ac 2

2

 kF .

(9.2)

This would require F (a c)2/(k
1)2, but F
(a c)2/(n*
1)2, so k n* would have to hold true before liberalization becomes beneficial. To abolish a public monopoly can in other words improve welfare, but only if costs are thereby significantly reduced. 3 Vertical relations and network externalities Industries like electricity, gas and railways contain elements of a natural monopoly, but also activities that can be separated and exposed to competition. For example, there are

182

International handbook on industrial policy

usually no competing rail networks or electricity grids, but there can in principle be competition in train operation, and in the retail trade and generation of electricity. A vertical relation implies that competition can be associated with additional costs. Vertical separation means profit margins both upstream and downstream if competition remains imperfect, or double marginalization (Perry, 1978) and hence, for example, excessive network access charges (vertical foreclosure). There may also be underinvestments in the network infrastructure. Consider, for example, an upstream natural monopoly with a (large) sunk cost F0. Its output is denoted x and its total cost function is TCx F0
cx, where c stands for its constant marginal costs. The (access) price of the output, which is used downstream, is p. Competition downstream means an n-firm Cournot oligopoly where the output is y. Let zyi denote the total use of x, so that variable costs are zyi p. Total costs TCy also include a sunk cost F that limits the number of firms. The downstream inverse demand function is the same as before, but the price is now r. Profits in a vertically integrated monopoly would be 0 ay y2 czy F0 F, because the variable cost term zyp cancels out. This would yield the output yM (a cz)/2. Note that the consumer surplus is y2/2, so the total surplus is TSM 

3(a  cz) 2  F0  F . 8

(9.3)

On the other hand, under vertical separation and free downstream entry, each oligopolist maximizes i ayi yyi pzyi F, given the upstream monopoly price p. Each downstream firm would produce yi (a zp)/(n
1), and get the net profits i (a  zp)2/(n
1)2 F
0. Solve for n and insert into y nyi to get zy(p) z(a zp √F ), which when inverted yields the monopolist’s inverse demand function. Its profits are 0  a z √Fx  z12x2  cx  F0.

(9.4)

Maximizing (9.4) yields xS z(a cz  √F )/2, which implies the downstream output yS  (a cz √F )/2. There will be nS
yS/ √F firms (insert zpa y  √F into the FOC to get yi  √F ). The monopolist gets 0S (a cz  √F )2/4, so the total surplus is TSS 

3(a  cz  √F) 2  F0 . 8

(9.5)

Comparing TSS and TSM shows that TSS TSM would require (a cz)/2 √F 11/12. However, nS is approximately [(a cz)/2 √F ] 1/2, so this cannot hold true even if there was only one downstream firm. In other words, vertical separation and liberalization are not beneficial unless they are associated with other advantages such as significantly lower costs (see also Vickers, 1995).3 Many public utilities are also characterized by so-called ‘network externalities’. For example, the usefulness of a telephone depends on the number of other subscribers. To see how this affects the desirability of liberalization, we shall introduce endogenous entry into a standard model of network externalities (see Rohlfs, 1974; Shy, 1995).

Liberalization and regulation of public utility sectors 183 Suppose that there is a continuum of the size 1 of potential consumers that are indexed by a variable z
[0, 1], which expresses their willingness to pay (WTP) for a service with the price p. They are ordered so that those with a high WTP are to the left, so that a high z means a low WTP. There are m users, each of whom buys one unit of the service. The net utility of subscribing is then m(1z)p. A negative value would mean no subscription and zero utility. The consumer is then indifferent (m(1z)p 0) for a critical value x of z. This defines the inverse demand, because it also expresses the number of consumers within the continuum who are users, so p x(1x). Each price is associated with two levels of x, but the lower, xL 0.5(0.25p), is unstable and can be understood as the critical mass, whereas x0.5
(0.25p) is stable. When is a duopoly superior to a monopoly? Assuming zero marginal costs and fixed costs F means that a monopolist would maximize  x(1 x)xF, so xM 0.667, pM  0.222 and xM L  0.333. As for the duopoly, each firm maximizes i x(1x)xi F, so the first-order condition is x
xi 2xxi x2 0. Adding the first-order conditions and solving for x yields xC 0.750, pC 0.188 and xC L  0.250. The total surplus is the sum of the area under x(1x) but above the price, and the profits. Note however that px cancels out, so we get the following expression:



x

t(1  t)dt
pxL  F .

(9.6)

xL

A monopoly and a duopoly would yield TSM 0.154 F and TSC 0.161 2F, respectively. The monopolist’s total revenues (gross profits) are 0.148. It follows that the duopoly is superior only if F is smaller than 4.81 per cent of these sales revenues, or 5.08 per cent of the sales revenues in the duopoly (0.141). So the presence of a network externality may increase the threshold for a duopoly to outperform even a profit-maximizing monopoly as compared to the conventional model in section 2. This suggests that both vertical relations and network externalities tend to shift the balance somewhat more in favour of a monopoly. In particular, if the incumbent is in public ownership, liberalization is beneficial only if the cost savings are significant enough to dominate over the double advantages of welfare maximization and consolidation. The models can be used for getting an idea about how large they must be (for details, see Willner, 2006b, 2006c). 4 Endogenous efficiency, distribution and quality Liberalization may indeed often reduce costs (see sections 5–7), but this is not sufficient for outperforming a welfare-maximizing public monopoly if the industry remains imperfectly competitive (see Willner, 1996, 2003). Moreover, less is known about the impact of entry on cost efficiency than on profit margins. Entry can in fact reduce X-efficiency in a principal–agent model (Martin, 1993), and studies from Britain and Italy suggest that competition does not always make privatization more successful (Martin and Parker, 1997; Fraquelli and Erbetta, 2000). An important reason for public utility reform has been to reduce especially the labour costs (Newbery, 2001). But this would also redistribute pay-offs, in which case liberalization would not be Pareto-improving (Vogelsang, 2006). Suppose that a welfaremaximizing public monopoly with user charges is replaced by an unregulated n-firm Cournot oligopoly. The output and price levels are x1, x0, p1 and p0. The (absolute value

184

International handbook on industrial policy

of the) post-privatization price elasticity of demand is . Long-run marginal costs are c and kc under private and public ownership; fixed costs are ignored. The difference (k 1)c depends on wages and working conditions only, so (k1)cx0 represents internal rent capture and is part of the total surplus, like the profits. The total surplus is then higher after privatization only if the price is reduced below kc. It is well known that p1 can be written nc/( n 1), so p1 kc implies k n/( n  1).4 This condition is very stringent if and n are low. For example, if is about 0.35 (as in electricity generation in Finland; see Kopsakangas and Savolainen, 2003), n 5 would require at least a 57 per cent reduction in marginal costs (that is, k 2.33). Reductions in average costs may also be related to quality or maintenance, in which case the nature of the output changes by liberalization. Let m denote the sum of all maintenance mi conducted by the n firms in an industry and let b and g stand for a positive coefficient, so that bm0.5 becomes part of the intercept of a linear demand function, that is p a0
bm0.5 p, and fixed costs are Fi gmi. The Cournot profits given mi are then [(a0
bm0.5 p)/(n
1)]2 gmi, so the quality in a symmetric two-stage equilibrium is m* {(a0 c)b/[g(n
1)2 b2]}2, which is unambiguously decreasing in n. While this is only an example, it suggests that reforms which emphasize just prices may pay insufficient attention to quality or maintenance. 5 Telecommunications Demand and technology have changed so that telecommunications are no longer a natural monopoly as a whole. In the USA, the monopoly status of AT&T/Bell Systems was therefore reduced. In the EU, the regulatory framework has focused on the transition from public monopoly to competition, and now on limiting the incumbents’ market power, changing the definitions of markets when needed. The member states operate in a federal framework, because legislation is initiated and enacted by the European Parliament, European Commission and the Council of Ministers. So called ex ante regulation is also being replaced by ex post application of competition laws (OECD, 2002). Important areas such as the coverage of services, public interest and interconnection have been the targets of regulation. Coverage is closely related to market access, where terms of service, concessions, procedures and conditions of access have been regulated. Safeguarding public interest requires regulatory measures on issues such as universal service. From a consumer perspective, the main issues are consumer rights, privacy and data security. Well functioning markets also require smooth interconnection both of services and between the networks, and regulation of the distribution of frequencies and the allocation of telephone numbers. Competition policy may have to give way to regulation if EU directives impose very rigid ex ante definitions and rules of conduct on issues like significant market power.5 Regulation to address anti-competitive behaviour may be slow, costly and not easily predictable, and might discourage entrants from investing and challenging the incumbent (Pelkmans, 2001). Regulation may on the other hand have favourable effects on issues such as interconnection charges, where competition policy has little effect. In the EU, interconnection charges remained relatively high until the national regulatory authorities decided that the large operators enjoyed significant market power. This led to increased monitoring and price regulation, which reduced prices dramatically from 1999 onwards (Pelkmans, 2001).6

Liberalization and regulation of public utility sectors 185

Privatization Liberalization Both

FIN FRA

AUS

D ITA NZ 1996

1994

1992

1990

1988

1986

1984

1982

1980 Source:

2000

ESP

UK

1998

USA

Boylaud and Nicoletti (2000, pp. 28, 30)

Figure 9.1

The timing of privatization and liberalization

Australia

Market shares of entrants by country

Finland 70

UK

60

New Zealand

50

USA

40 30 20 10 0

Year 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 00 01 19 19 19 19 19 19 19 19 19 19 19 19 19 19 19 19 20 20

Figure 9.2

The development of market shares in the fixed long-distance markets

The incumbent has usually been at least partly privatized before liberalization, as illustrated by Figure 9.1, where the height of the lines illustrates the variation in the extent of state ownership in the long-distance market in 1998. The exceptions are New Zealand and Italy, where privatization and liberalization were simultaneous, and Finland, where the market was liberalized despite a state ownership of 78.8 per cent in 1998. On average, state ownership in the OECD was reduced from more than 77 per cent to 51.2 between 1992 and 1998.7 The market structures for long-distance calls in the OECD (see Figure 9.2) changed only gradually during the first years after liberalization (OECD, 1996; Götz, 2000).

186

International handbook on industrial policy

Finland is again an exception, as the entrant captured a 55 per cent market share during the first year of operation (Nattermann and Murphy, 1998; Björkroth, 2003). As for carrier selection, subscribers are now able to choose between at least five operators in fixed long-distance and international telephony in 12 member states in EU-15, but only in eight in the case of local telephony. In Germany, only 30 per cent of the subscribers had access to alternative providers, the percentage being 58 in Finland. But consumers have not always been active. The highest proportions of subscribers actually using the possibility to choose are found in Italy and Finland (50 and 65 per cent, respectively), the usual percentage being 20–30. (See Commission of European Communities, 2002.) Competition has developed swiftly also within mobile telecommunications. In 2002, a large number of OECD countries had three or more network operators. The rapid diffusion of mobile communications means that there are now more mobile subscribers than in the fixed network in a majority of these countries (OECD, 2002). Number portability has also increased competition. Contrary to some initial scepticism, mobile number portability has led to more equal market shares.8 The providers without a network of their own may also get some bargaining power over network charges, depending on how many customers they get. If costs and call charges are reduced further thereby, the market structure may change more rapidly than expected. After the introduction of competition in the USA, long-distance calls have become cheaper, as can be expected when there is more competition among profit-maximizing operators (Blank et al., 1998; Hausman et al., 1993).9 Prices seem to have fallen also in Europe after prospective and/or effective competition (Brunekreeft and Gross, 2000; Boylaud and Nicoletti, 2000).10 But the industry has also been subject to extensive regulation throughout the OECD (generally in the form of cap regulation in fixed line telephony, but usually no regulatory constraints for mobile retail prices). Even if liberalization may have been more important than the more stringent regulation, technical progress and anticipated competition have also had an impact. For example, prices fell substantially in long-distance voice telephony during the decade before liberalization in Finland (Björkroth, 2003), so actual competition seems to have caused just a final squeeze.11 However, most incumbents were originally public monopolies, so possible excessive profit margins could also have been addressed by finding other sources of public revenues, unless the higher prices are explained by costs rather than mark-ups. Boylaud and Nicoletti (2000) analyse the effect of liberalization and privatization also on productivity in long-distance and mobile cellular services in 23 OECD countries in 1991–7, and find evidence for improved productivity (while ownership had less of an impact), but Daßler, Parker and Saal (2002) find considerable variation in Europe and as yet no conclusive evidence of higher labour or total factor productivity growth. Moreover, a complete welfare assessment would also have to include such benefits of consolidation that are related to vertical relations and network externalities. To observe costs or prices in isolation from service quality may also be misleading. Improved quality can make consumers better off despite higher prices, but the reverse may also be true. Boylaud and Nicoletti (2000) and the European Commission (1996) find signs of improved quality, but Wydell (1999) finds lower consumer satisfaction.12 While his choice of cases to compare is open to criticism, the possibility of lower quality and ‘confusion marketing’ have been pointed out also by van Dam and Went (2001), Stephen (2001) and The Guardian (14 October, 2000).

Liberalization and regulation of public utility sectors 187 Liberalization will probably continue, with challenges in the form of introducing competition in local markets, unbundling the local loop and creating appropriate regulation. To impose and ensure effective competition in this bottleneck is not straightforward, and will have strong secondary effects on the rest of the markets. Another challenge is the fact that mobile telephony may be a natural oligopoly (Valletti, 2003). 6 Electricity The electricity industry as a whole was earlier seen as a natural monopoly, but the oil crisis showed that several producers could supply a network without reducing its stability (IEA, 1999). Also, higher demand and a reduction of the minimum efficient scale of some types of generators (Steiner, 2001) meant scope for several producers. This does not necessitate liberalization but competition in the retail trade was believed to stimulate entry and hence lower prices (Sioshansi and Hamlin, 2004). The supply side of the electricity market consists of generation, transport (high-tension and lower voltage local distribution) and trade. Transport is generally thought of as still being a natural monopoly, at least in the case of the low-voltage distribution network. Before deregulation, vertically integrated public monopolies tended to dominate, so liberalization has required a separation between those activities that can be made subject to competition and those that cannot. The British approach to electricity liberalization inspired The European Commission’s 1996 Electricity Directive (and the 2003 Directive’s even stronger emphasis on liberalization), and also the recommendations by the World Bank and consultant organizations like Coopers & Lybrand (Thomas, 2004a, 2005).13 There has now been liberalization in the electricity supply industry in 51, privatization in 30, and (at least planned) vertical separation in 27 out of 62 countries (Newbery, 2001). In Britain, state-owned companies and subsequently public utilities were in the 1980s usually just divested. But the initially state-owned Central Electricity Generating Board (CEGB) was split up in 1989–90, thus separating transmission and generation, and a wholesale spot market, which now works as a half-hourly wholesale market, was created in order to make generation competitive. Generation, transmission, distribution and supply are now unbundled. All consumers have been able to choose their provider since 1999. Transmission and to some extent retail are still regulated, but supply price regulation was abandoned in 2002. The industry regulator OFGEM monitors the development of competition, whereas the related body Energywatch gives information to consumers about the suppliers. Some companies also provide easy switching through the Internet (Newbery, 2001; Giulietti et al., 2004; Thomas, 2004a). Public ownership still dominates in the Nordic countries, but they are forerunners in commercialization and competition, and in separating generation and transmission (Dugstad and Roland, 2003). The pool Startnett Marked was founded in Norway in 1991, but the need to utilize hydro-power optimally led to trade as early as 1971. Startnett Marked was joined by Sweden in 1996 and renamed Nord Pool, the first international electricity pool. The Nordic countries are still ahead of most regions in terms of crossborder integration (IEA, 2001; Newbery, 2001). Finland, where supply and production were separated from transmission and distribution in 1995 and where also households can choose their provider since 1998, joined the Nord Pool in 1998 (Willner, 2006a), and western Denmark in 1999–2000 (Nord Pool, 2002).

188

International handbook on industrial policy

The Nordic grid companies are regulated public monopolies. In Norway and Denmark, this means financial rules in advance (ex ante regulation) and a revenue cap, which is costly and difficult to implement but gives better consumer protection. There is rate-of-return regulation in Sweden and Finland, where companies are investigated case-by-case, for example, after customer complaints (ex post regulation). This implies stronger incentives for innovation and easier implementation, but possibly also monopoly profits, overinvestments and a biased choice of firms to be investigated (Grønli, 2001). Turning to the results of liberalization, consumer prices were reduced in the UK after privatizing and restructuring the CEGB. Distribution and transmission prices fell by 50 and 40 per cent, respectively, after 1995 (Thomas, 2004a). It seems that cost efficiency and profitability became higher than without reform, but consumer benefits became lower, so the total surplus appears as increased or reduced, depending on whether a ‘pro-privatization’ or a ‘pro-CEGB’ counterfactual is chosen (Newbery, 2001).14 Prices are relatively low also in Germany, but this has benefited industrial customers more than households (Brunekreeft and Keller, 2000; Böge, 2001), and in the Nordic countries.15 However, both public and private plants in Finland have been described as fairly productive, and total factor productivity increased by 5.8 per cent in 1994–6, but not significantly in 1996–8 (Kopsakangas-Savolainen, 2003; Sulamaa, 2001).16 However, the lower prices in Europe may represent a temporary grace period because of the excess capacity that was inherited from the old system (Finon et al., 2004; Thomas, 2004b). To adjust to increasing demand will now cost more, as private companies have up to 50 per cent higher capital costs (Watts, 2001). Capacity and reliability may have a public-good character and may therefore be underprovided by a completely liberalized system (Jiang and Yu, 2004). There must be overcapacity in a truly efficient system because of the difficulty of obtaining adequate forecasts, so the former vertically integrated monopolies may not have represented bad market discipline (Watts, 2001). The social costs of undercapacity may also be higher than in the opposite case (Watts, 2001). The recent major blackouts in Europe and North America, most notably in California, are possibly caused by underinvestments associated with liberalization (LiebDóczy et al., 2003), but also regulatory flaws. For example, there was a cap on the wholesale price, which could be avoided by selling electricity to buyers elsewhere, who immediately resold it to California (Pierce, 2003). The prices of gas and electricity then increased by 1000 per cent in the late 2000 and early 2001. This development was partly caused by companies like Enron, which was accused of manipulating the congestion rules, circumventing price-cap rules and withholding capacity unilaterally. The security issues may have become more complex. For example, an interconnected system of plants under different ownership may break down because of failures in one particular plant (Auriol, 1998). There may be a trade-off between security and economic efficiency and/or environmental concerns (Lieb-Dóczy et al., 2003). Economists and policy makers may have paid too little attention to probabilistic reliability constraints which would increase the optimal price above the long-run marginal production costs (Jiang and Yu, 2004). The liberalization policy may therefore have overemphasized low prices as a social objective; a stronger emphasis on reliability might reduce the role for competition (Moss, 2004).

Liberalization and regulation of public utility sectors 189 The benefits of vertical integration (see section 3) also tend to be significant, in particular in large plants, as follows from studies in the USA and Italy (Kwoka, 2002; Fraquelli et al., 2004). There is also evidence of a natural monopoly in municipal Swiss electric utilities (Filippini, 1998). The social advantages of a monopoly have in fact rarely been assessed before liberalization, and may for example have dominated in the Norwegian electricity distribution (Salvanes and Tjøtta, 1998). The industry, on the other hand, may not in practice be as liberalized as described, because there are signs of increasing regulation and extensive consolidation, so that Europe may become dominated by a few companies from a group now consisting of EDF (France), RWE and E.ON (Germany), ENEI (Italy), Verbund (Austria), Vattenfall (Sweden), Statkraft (Norway) and Fortum (Finland) (Thomas, 2003, 2005).17 The network and competition activities are separated in Britain, but the wholesale market is now dominated by confidential long-term contracts, and the half-hourly market is not as important as intended. Retail competition has favoured industrial customers but not households, who faced increasing prices in 1999–2002. Also all the 14 privatized retailers are now owned by foreign generation companies, despite the ambition towards vertical separation (Thomas, 2003; 2004a).18 To achieve the full benefits of liberalization would also require a more active switching behaviour. For example, consumers in Britain do not switch to a sufficient extent despite simple Internet-based procedures. Giulietti et al. (2004) conclude that the market has not become fully competitive, which may explain why retail prices have not been reduced by as much as the wholesale prices. Such consumer inertia may also partly be explained by (deliberately?) confusing service packages (Thomas, 2003). Liberalization was partly based on the idea of public ownership as inefficient but not necessarily less desirable than an unregulated private monopoly. A couple of studies do indeed suggest that public ownership may be less efficient in the electricity industry (see Bagdadioglu et al., 1996; Kumbakhar and Hjalmarsson, 1998). Hayashi et al. (1987) found public ownership more efficient in the 1960s and less efficient in the 1970s in the USA. Public ownership may be more efficient under cost-of-service regulation, but not under price-cap regulation (Arocena and Waddams Price, 2002). A comparison of plants in 14 countries found public plants and transmission equally efficient in a technical and managerial sense, but often subject to restrictions that lead to a less efficient input mix (Pollitt, 1995). But most studies suggest that cost efficiency in electricity is better or no worse under public ownership in the UK, the USA and Sweden. We refer to Atkinson and Halvorsen (1986), De Alessi (1974), Färe et al. (1985), Foreman-Peck and Waterson (1985), Hausman and Neufeld (1991), Hjalmarsson and Veiderpass (1992), Meyer (1975), Moore (1970), Nelson and Primeaux (1988), Neuberg (1977), Pescatrice and Trapani (1980), Peters (1993), Spann (1977) and Yunker (1975).19 We may conclude that liberalization can mean either disadvantages related to vertical disintegration, diseconomies of scale and reliability, or consolidation, which might increase prices and hand over the strategic decisions to a limited number of very large companies, or a combination of the problems. The old public monopoly system had by contrast an excellent record after World War II in terms of affordability and reliability (Thomas, 2003). Its problems were mainly believed to be related to cost inefficiency, but in light of the evidence cited above it seems that the main reason for reform has been based at least on an exaggeration.

190

International handbook on industrial policy

7 Other cases of liberalization The water and sewerage industry is capital-intensive and its fixed costs are high, so competition in the normal sense is difficult to achieve (Hukka and Katko, 2002). Also the pipe network is usually local, so the local monopoly utilities have often been small. This would suggest that liberalization should mean competition for the market; that is, the company that promises the lowest tariffs gets a monopoly. A large enough number of franchise bidders would then drive down the price. But this would be ineffective if the periods are short, and too weak incentives to invest in the opposite case. Competition in the market, where competing companies serve the same area using the same pipes, might then be more efficient (Foellmi and Meister, 2002), but trade between adjacent utilities may be superior to competition (Foellmi and Meister, 2004). Britain is a forerunner also in the reform of water and sewerage services, which were privatized in England and Wales in 1989. But the industry is still dominated by regional monopolies, and there is no competition for the market (Saal and Parker, 2000). Some competition nevertheless exists in the market in the form of common carriage between adjacent interconnected networks (Foellmi and Meister, 2004). As for price regulation, the independent body OFWAT applies the so-called RPI
K formula, where RPI stands for the retail price index and K depends both on expected efficiency gains and targets for quality-improving investments. Quality is monitored by the Drinking Water Inspectorate (DWI) (Saal and Parker, 2000; Byatt, 2001). France is not a forerunner, but its large private companies are taking part where water services are privatized and regulated elsewhere. There have always existed private water companies in France, and they now serve about 78 per cent of the population. For example, Vivendi is now a big owner that operates also in the UK, the USA, Australia and the Ivory Coast. Such companies have taken over smaller utilities. There is competition for the market, but the concessions are long and easily renewable. The structure has been criticized for being oligopolistic and causing excess profits at the expense of the consumers (Roche and Johannès, 2001; Hukka and Katko, 2002). Public ownership in the form of municipal or supramunicipal companies dominates in the Nordic countries, but some rural districts have had private and non-commercial providers, and there has been private participation in non-core operations. Planning and design, construction and maintenance and many other activities, which dominate the cash flow, have often been competitively tendered long before liberalization became fashionable. Also, unlike Britain, such forms of public–private partnerships have been initiated by the municipalities and are not imposed by the central government as often in the UK (Hukka and Katko, 2002). As for the results of privatization in England and Wales, the industry has made qualityimproving investments. However, the water companies were not allowed by the government to borrow enough funds in the late 1970s and the 1980s (Byatt, 2001; Hukka and Katko, 2002), so the same policy that privatized the industry was also responsible for its underinvestments. Saal and Parker (2000) find no evidence of better efficiency after privatization; Hunt and Lynk (1995) even suggest lower cost efficiency because of lost economies of scope. However, tightened but heavy-handed regulation in 1995 did increase efficiency (Saal and Parker, 2000; Hukka and Katko, 2002). The case for vertical integration may be even stronger in the water than in the electricity industry, so the cost-efficiency gains may have to be even larger than in other indus-

Liberalization and regulation of public utility sectors 191 tries. The British experiences are not therefore fully convincing. Also it seems that the policy makers were unaware of the fact that those studies that compared public and private waterworks in the USA and the UK, that is, Bhattacharyya et al. (1994), Bruggink (1982), Byrnes et al. (1986), Feigenbaum and Teeples (1983), Mann and Mikesell (1976),20 Teeples and Glyer (1987) and Lynk (1993) found that the former were more cost efficient, Crain and Zardkoohi (1978) providing an exception. In the railways, the EU wants to promote competition through its rail directive, but competition in the market has been difficult to achieve because of scheduling problems. Competition has therefore mostly meant competition for the market (Newbery, 2001). The railways in the UK became vertically separated in 1993, and the network infrastructure became a private monopoly, Railtrack. The trains were leased by the franchise bidders. The franchise period is seven years, and there is full protection against entry for the first four years (Newbery, 2001). As in the water industry, there were insufficient investments during the Thatcher era. These may have contributed to the economic difficulties of Railtrack, which was later reorganized as the non-profit maximizing company Network Rail. Many economists are still uncertain about the benefits of the reorganization (Newbery, 2001). For example, it has been reported that changes work less smoothly now that connecting trains may belong to different companies (Beckett, 2002). Also it seems that the separation between train operation and rail infrastructure has led to a 40–50 per cent reduction in track-maintenance staff (Guardian, 3 April 2001), which meant externalized rather than reduced costs.21 While there may be a large variation in the efficiency of publicly run railway monopolies, some research, such as Caves and Christiansen (1980) and Caves et al. (1982) suggests that public ownership as such does not lead to inferior performance. The railways have been reformed to some extent in Sweden as well, partly because of a perceived need to increase public spending efficiency and to increase the use of the system. The intention is to increase competition gradually. There is no British-style independent regulator; the industry is regulated by the competition authority and a new body called Rikstrafiken (Baumstark, 2001). No similar developments have taken place in Finland, where passenger railways are run by a corporatized public monopoly, at least for the moment (Willner, 2006a). Bus transport has also been liberalized in many countries. In Britain, this meant in most regions except for Greater London that the control over schedules and route network was taken over by the operators. Bus transport is labour-intensive, and most contributions except for Kennedy (1995) suggest that private operation is indeed less costly (see summaries in Willner, 2003, 2006a). An early ‘welfare balance sheet’ suggested that staff and passengers usually became worse off, but profits increased. The total surplus increased in some regions and deteriorated in others (White, 1990). Deregulation can under such circumstances mean insufficient incentives to attract customers away from cars by offering low prices (Ireland, 1991). As for service quality, competition and free entry and exit may lead to reduced coordination, instability, confusing changes in schedules and network, and reduced through- and inter-ticketing (Tyson, 1990; White, 1990; Oldale, 1997). The Nordic version has mostly meant authority control over routes and schedules, but journeys are tendered.22 This provides better stability and coordination. With some notable

192

International handbook on industrial policy

exceptions, there has been little privatization, but private operators have often increased their shares of routes. There has been some deterioration in frequency during the last decades, but they cannot be blamed on deregulation as such. Costs have been reduced also in Finland and Sweden, but it seems that these cost reductions have at least partly taken place at the expense of the employees also in Finland (see Willner, 2006a). 8 Concluding remarks As our theoretical model suggests, a combination of privatization and liberalization can improve welfare, provided that costs are reduced (often dramatically) and there are no other disadvantages. But public ownership does not seem to be less efficient in utilities such as water and electricity, where even the reverse may be true. Telecommunications are often seen as a more successful example of liberalization than electricity and in particular railways. Some publicly owned incumbents, on the other hand, were, technically progressive, so significant cost reductions even without reform cannot always be ruled out. Network externalities and, to some extent, economies of vertical integration may in addition motivate some question marks. In electricity, there have been concerns both about excessive market power and the dangers of fragmentation, which may reduce both the advantages of vertical integration and the reliability of the system. As for the water industry, the underinvestment problem might have been addressed without privatization. The same can be said about the railway industry, which has also raised issues about proper maintenance and coordination. Costs have been reduced in bus transport, but partly by making the staff worse off, and in some countries also by reducing the service quality. Notes 1.

2. 3. 4. 5.

6. 7. 8. 9. 10.

The condition n* 4 is not sufficient, because  0 can hold true n
1 n* n
1. A monopoly is superior in most of the interval of F in which n* 2, but not if n* 4. Another way to analyse the desirability of entry is to divide the upper limit of F by pnxn. Suppose that we know the post-liberalization price elasticity (a
nc)/n(a c). We can then conclude ex post that liberalization has increased the total surplus if nF is below [100(n
3)/8n ] per cent of the industry sales revenues. Vogelsang’s (1988) natural monopoly definition emphasizes, however, the social costs of duplication. An unregulated integrated monopoly may therefore be superior to separation and downstream competition, even if the alternative is a differentiated Bertrand oligopoly (Buehler, 2005). The rule k n/( n 1) also applies when focusing on consumer welfare rather than the total surplus, even if (k1)c represents waste. Significant market power (not to be confused with ‘dominant position’ in the competition law) is a key concept in European regulation and means control of more than 25 per cent of the relevant market (which may be far less than what is required for a dominant position). It is used for deciding whether an organization should be subject to specific obligations under the open network provision directives. Regulation may not always be as successful. Technical efficiency has improved after the transition from rate-of-return to incentive regulation in the USA according to Majumdar (1997) and Resende (2000), but not according to Uri (2003). These percentages are, on the other hand, 39 and 31.3 if the average is weighted by each country’s share of the OECD telecom revenues, because of some large full privatizations prior to 1992. Number portability has been available in Singapore since 1997, in Britain and in the Netherlands since 1999. Interestingly, Germany, Finland and especially the USA, a pioneer in liberalizing fixed and mobile communications, have been laggards in implementating number portability. However, the market has been criticized for being oligopolistic, possibly with tacit collusion. Human and physical capital improvements and intervention from the Federal Communications Commission may have been at least as important as deregulation (MacAvoy, 1998; Taylor and Taylor, 1993; Sung, 1998). But despite intensified regulation, abuse of dominant position still exists, and termination and access charges are among the main areas of concern (Boylaud and Nicoletti, 2000).

Liberalization and regulation of public utility sectors 193 11. 12.

13. 14. 15. 16. 17. 18. 19. 20. 21. 22.

As for mobile communications, mobile number portability has intensified competition, thus causing further reductions in charges (Björkroth, 2004). Wydell finds better quality in terms of digitalization and fault rates in monopoly markets, but this seems to be associated with higher costs as well. The overall percentage tariff change between 1991 and 1995 was 14 in France and 23.7 in the UK, but
37 for the Dutch public monopoly (European Commission, 1996). However, production was opened up for new actors in the USA in 1978, before the British reforms, and large consumers were allowed to choose among several providers in Chile from 1982 (IEA, 2001, p.29). The counterfactuals mean either an unchanged CEGB or similar efficiency gains to those after restructuring. However, wholesale prices increased in 2002–3 by 600 per cent because of a drought (Finon et al., 2004). The Nordic countries’ relative success may depend on a tradition of both public ownership and fragmentation, and in Norway on the pre-existence of a wholesale spot market (Thomas, 2004b). There is also evidence from Spain on collusion, interrupted by price wars (Fabra and Toro, 2005). Consolidation and cross-ownership might in the Nordic countries even lead to more market power than in the old system, insofar as profit maximization has replaced wider objectives. It has been argued that the comparative advantages of public ownership may be in distribution rather than generation (Kwoka, 2005). The often cited survey by Borcherding et al. (1982) wrongly cites Mann and Mikesell (1976) as finding private waterworks more efficient. Similar effects were observed in the UK gas industry, where Transco made 1000 engineers responsible for gas-pipe maintenance redundant in 1997. There have been fewer changes in rural bus transport where private operators have always played an important part, in particular in Finland.

References Arocena, P. and C. Waddams Price (2002), ‘Generating efficiency: economic and environmental regulation of public and private electricity generators in Spain’, International Journal of Industrial Organization, 20 (1), 41–69. Atkinson, S. and R. Halvorsen (1986), ‘The relative efficiency of public and private firms in a regulated environment: the case of U.S. electric utilities’, Journal of Public Economics, 29, 281–94. Auriol, E. (1998), ‘Deregulation and quality’, International Journal of Industrial Organization, 16 (2), 169–94. Bagdadioglu, N., C.M. Waddams Price and T.G. Weyman-Jones (1996), ‘Efficiency and ownership in electricity distribution: a non-parametric model of the Turkish experience’, Energy Economics, 18 (1), 1–23. Baumstark, L. (2001), ‘The pioneering Swedish experiment in railway regulation’, in C. Henry, M. Matheu and A. Jeunemaître (eds), Regulation of Network Utilities, The European Experience, Oxford: Oxford University Press. Beckett, F. (2002), ‘The Great Western rip-off’, New Statesman, 8 April, 35–6. Bhattacharyya, A., E. Parker and K. Raffiee (1994), ‘An examination of the effect of ownership on the relative efficiency of public and private water utilities’, Land Economics, 70 (2), 197–209. Björkroth, T. (2003), Engine or Wheels of Our Prosperity? Infrastructure and Economic Growth and Effects of Liberalisation of the Finnish Telecommunications Market, Turku: Åbo Akademi University Press. Björkroth, T. (2004), ‘Mobile number portability – lower charges or increased product variety or perhaps both?’, mimeo, Åbo Akademi University. Blank, L., D.L. Kaserman and J.W. Mayo (1998), ‘Dominant firm pricing with competitive entry and regulation: the case of IntraLATA toll’, Journal of Regulatory Economics, 14 (1), 5–53. Böge, U. (2001), ‘Liberalization of Energy Markets: The German Way’, in C. Henry, M. Matheu and A. Jeunemaître (eds), Regulation of Network Utilities, The European Experience, Oxford: Oxford University Press. Borcherding, T.E., W.W. Pommerehne and F. Schneider (1982), ‘Comparing the efficiency of private and public production: the evidence from five countries’, Zeitschrift für Nationalökonomie, Suppl. 2, 127–56. Boyd, C.W. (1986), ‘The comparative efficiency of state owned enterprises’, in A.R. Negandhi (ed.), Multinational Corporations and State-Owned Enterprises: A New Challenge in International Business, Greenwich, CT and London: JAI Press. Boylaud, O. and G. Nicoletti (2000), ‘Regulation, market structure and performance in telecommunications’, OECD Economics Department Working Papers, 237. Bruggink, T.M. (1982), ‘Public versus regulated private enterprise in the municipal water industry: a comparison of water costs’, Quarterly Review of Economics and Business, 22 (1), 111–25. Brunekreeft, G. and W. Gross (2000), ‘Prices for long-distance voice telephony in Germany’, Telecommunications Policy, 24, 929–45.

194

International handbook on industrial policy

Brunekreeft, G. and K. Keller (2000), ‘The electricity supply industry in Germany: market power or power of the market?’ Utility Policy, 9 (1), 15–29. Buehler, S. (2005), ‘The promise and pitfalls of restructuring network industries’, German Economic Review, 6 (2), 205–28. Byatt, I. (2001), ‘The water regulation regime in England and Wales’, in C. Henry, M. Matheu and A. Jeunemaître (eds), Regulation of Network Utilities, The European Experience, Oxford: Oxford University Press. Byrnes, P., S. Grosskopf and K. Hayes (1986), ‘Efficiency and ownership: further evidence’, Review of Economics and Statistics, 68 (2), 337–41. Caves, D.W. and L.R. Christensen (1980), ‘The relative efficiency of public and private firms in a competitive environment: the case of Canadian railroads’, Journal of Political Economy, 88 (5), 958–76. Caves, D.W., L.R. Christensen, J.A. Swanson and M.W. Thretheway (1982), ‘Economic performance of U.S. and Canadian railroads: the significance of ownership and the regulatory environment’, in T. Stanbury and F. Thompson (eds), Managing Public Enterprises, New York: Praeger. Commission of European Communities (2002), Eighth Report from the Commission on the Implementation of the Telecommunications Regulatory Package, Brussels. Crain, W.M. and A. Zardkoohi (1978), ‘A test of the property rights theory of the firm: water utilities in the United States’, Journal of Law and Economics, 21 (2), 395–408. Daßler, T., D. Parker and D. Saal (2002), ‘Economic performance in European telecommunications, 1978–98: a comparative study’, European Business Review, 14 (36), 194–209. De Alessi, T. (1974), ‘An economic analysis of government ownership and regulation: theory and evidence from electric power industry’, Public Choice, 19 (1), 1–42. Dugstad, E. and K. Roland (2003), ‘So far so good: experiences and challenges in the Scandinavian power market’, International Journal of Regulation and Governance, 3 (2), 135–60. European Commission (1996), ‘Universal service for telecommunications in the perspective of a fully liberalised environment’, Communication To The European Parliament, The Council, the Economic and Social Committee and the Committee of the Regions, DG XIII, Brussels, 12 March. Fabra, N. and J. Toro (2005), ‘Price wars and collusion in the Spanish electricity market’, International Journal of Industrial Organization, 23 (3–4), 155–83. Färe, R., S. Grosskopf and J. Logan (1985), ‘The relative performance of publicly owned and privately owned electric utilities’, Journal of Public Economics, 26 (1), 89–106. Feigenbaum, S. and R. Teeples (1983), ‘Public versus private water delivery: a hedonic cost approach’, Review of Economics and Statistics, 65 (4), 672–8. Filippini, M. (1998), ‘Are municipal electricity distribution utilities natural monopolies?’, Annals of Public and Cooperative Economics, 69 (2), 157–74. Finon, D., T.A. Johansen and A. Midtton (2004), ‘Challenges when electricity markets face the investment phase’, Energy Policy, 32, 1355–62. Florio, M. (2004), The Great Divestiture. Evaluating the Impact of the British Privatizations 1979–1997, Cambridge, MA and London: MIT Press. Foellmi, R. and U. Meister (2002), ‘Product market competition in the water industry: voluntarily nondiscriminatory pricing?’, working paper, EARIE 2002. Foellmi, R. and U. Meister (2004), ‘Enhancing efficiency of water supply – product market competition versus trade’, working paper, EARIE 2004. Foreman-Peck, J.O. and M. Waterson (1985), ‘The comparative efficiency of public and private enterprise in Britain: electricity generation between the world wars’, Economic Journal, 95, 83–95. Fraquelli, G. and F. Erbetta (2000), ‘Privatisation in Italy: an analysis of factor productivity and technical efficiency’, in D. Parker (ed.), Privatisation and Corporate Performance, Cheltenham, UK and Northampton, MA, USA: Edward Elgar. Fraquelli, G., M. Piacenza and D. Vannoni (2004), ‘Cost savings from generations and distribution with an application to Italian electric utilities’, working paper, EARIE 2004. Giulietti, M., J. Otero and M. Waterson (2004), ‘Supply competition and price behaviour in the UK electricity supply industry’, working paper, EARIE 2004. Götz, G. (2000), ‘Der Deutsche Telekommunikationsmarkt zwei Jahre nach der vollständigen Marktöffnung: Eine Bestandsaufnahme aus Sicht der Verbraucher’, paper presented at the 11th ITS European Regional Conference. Grønli, H. (2001), ‘A comparison of Scandinavian regulatory models: issues and experience’, The Electricity Journal, Aug./Sept., 57–64. The Guardian (2000), ‘Ringing up the right numbers’, 14 October, Jobs and Money section, p.2. The Guardian (2001), ‘The breaking point’, 3 April, pp.4–5. Hausman, W.J. and J.L. Neufeld (1991), ‘Property rights versus public spirit: ownership and efficiency of U.S. electric utilities prior to rate of return regulation’, Review of Economic Statistics, 73 (3), 414–23. Hausman, J., T. Tardiff and A. Belifante (1993), ‘The effects of the breakup of AT&T on telephone penetration in the United States’, American Economic Review, 83, 178–84.

Liberalization and regulation of public utility sectors 195 Hayashi, P.M., M. Sevier and J.M. Trapani (1987), ‘An analysis of pricing and production efficiency of electric utilities by mode of ownership’, in M.A. Crew (ed.), Regulating Utilities in an Era of Deregulation, Houndmills and London: Macmillan Press. Hjalmarsson, L. and A. Veiderpass (1992), ‘Efficiency and ownership in Swedish electricity retail distribution’, Journal of Productivity Analysis, 3 (1), 7–23. Hukka, J.J. and T.S. Katko (2002), ‘Water privatisation revisited – panacea or pancake?’, Working paper, Tamper University of Technology. Hunt, L.C. and E.L. Lynk (1995), ‘Privatisation and efficiency in the UK water industry: an empirical analysis’, Oxford Bulletin of Economics and Statistics, 57, 371–88. International Energy Agency, IEA (1999), Electricity Market Reform – an IEA Handbook, Paris: OECD. International Energy Agency, IEA (2001), Competition in Electricity Markets, Paris: OECD. Ireland, N.J. (1991), ‘A product differentiation model of bus deregulation’, Journal of Transport Economics and Policy, 25, 153–62. Jiang, J.N. and J. Yu (2004), ‘Reliability in electricity markets: another binding constraint’, The Electricity Journal, June, 43–53. Kay, J.A. and D.J. Thompson (1986), ‘Privatisation: a policy in search of a rationale,’ Economic Journal, 96 (March), 18–32. Kennedy, D. (1995), ‘London bus tendering: the impact on costs’, International Journal of Applied Economics, 9 (3), 305–17. Kopsakangas-Savolainen, M. (2003), ‘Quantity versus price competition in the Finnish electricity markets’, Finnish Economic Papers, 16 (2), 51–60. Kumbakhar, S.C. and L. Hjalmarsson (1998), ‘Relative performance of public and private ownership under yardstick competition: electricity retail distribution’, European Economic Review, 42, 97–102. Kwoka, J.E. (2002), ‘Vertical economies in electric power: evidence on integration and its alternatives’, International Journal of Industrial Organization, 20 (5), 653–72. Kwoka, J.E. (2005), ‘The comparative advantage of public ownership: evidence from U.S. electric utilities’, Canadian Journal of Economics, 38 (2), 622–40. Lieb-Dóczy, E., A.-R. Börner and G. MacKerron (2003), ‘Who secures the security of supply? European perspectives on security competition and liability’, The Electricity Journal, December, 10–19. Lynk, E.L. (1993), ‘Privatisation, joint production and the comparative efficiencies of private and public ownership: the UK water industry case’, Fiscal Studies, 14 (2), 98–116. MacAvoy, P.W. (1998), ‘Testing for competitiveness of markets for long distance telephone services: competition finally?’, Review of Industrial Organization, 13 (3), 295–319. Majumdar, S.K. (1997), ‘Incentive regulation and productive efficiency in the U.S. telecommunications industry’, Journal of Business, 70 (4), 547–76. Mankiw, N.G. and M.D. Whinston (1986), ‘Free entry and social inefficiency’, Rand Journal of Economics, 17 (1), 48–58. Mann, P.C. and J.L. Mikesell (1976), ‘Ownership and water systems operation’, Water Resources Bulletin, 12 (5), 995–1004. Martin, S. (1993), ‘Endogenous firm efficiency in a Cournot principal–agent model’, Journal of Economic Theory, 59 (2), 445–50. Martin, S. and D. Parker (1997), The Impact of Privatization. Ownership and Corporate Performance in the UK, London and New York: Routledge. Megginson, W.L. and J.M. Netter (2001), ‘From state to market: a survey of empirical studies on privatization’, Journal of Economic Literature, 39 (2), 321–89. Meyer, R.A. (1975), ‘Publicly owned versus privately owned utilities: a policy choice’, Review of Economics and Statistics, 57 (4), 391–9. Millward, R. (1982), ‘The comparative performance of public and private ownership’, in L.E. Roll (ed.), The Mixed Economy, London: Macmillan. Moore, T.G. (1970), ‘The effectiveness of regulation of electric utility prices’, Southern Economic Journal, 37, 365–75. Moss, D.L. (2004), ‘Competition or reliability in electricity? What the coming policy shift means for restructuring’, The Electricity Journal, March, 11–28. Nattermann, P.M. and D.D. Murphy (1998), ‘The Finnish telecommunications market: advantage of local access incumbency’, Telecommunications Policy, 22, 555–69. Nelson, R.A. and W.J. Primeaux (1988), ‘The effects of competition on transmission and distribution costs in the municipal electric industry’, Land Economics, 64 (4), 338–46. Neuberg, L.G. (1977), ‘Two issues in the municipal ownership of electric power distribution systems’, Bell Journal of Economics, 8, 303–23. Newbery, D.M. (2001), Privatization, Restructuring and Regulation of Network Utilities, Cambridge, MA and London: MIT Press.

196

International handbook on industrial policy

Nord Pool (2002), ‘The Nordic power market – electricity power exchange across national borders’, Accessed 13.1.2003 (www.nordpool.no). OECD (1996), Local Telecommunications Competition: Development and Policy Issues, OECD/GD(96)179, Paris. OECD (2002), ‘Competition and regulation issues in telecommunications’, DAFFE/COMP(2002)6, 01-Feb2002. Oldale, A. (1997), ‘Local bus deregulation and timetable instability’, working paper, The Economics of Industry Group, London School of Economics. Parker, D. (2005), ‘The UK’s privatisation experiment: the passage of time permits a sober assessment’, forthcoming in M. Koethenbuergher, H.-W. Sinn and J. Whalley (eds), Privatisation Experiences in the EU, Cambridge, MA: MIT Press. Pelkmans, J. (2001), ‘Making EU network markets competitive’, Oxford Review of Economic Policy, 17 (3), 432–56. Perry, M.K. (1978) ‘Vertical integration: the monopoly case’, American Economic Review, 68 (September), 561–70. Pescatrice, D.R. and J.M. Trapani (1980), ‘The performance of public and private utilities operating in the United States’, Journal of Public Economics, 13 (3), 259–76. Peters, L.L. (1993), ‘For-profit and non-profit firms: limits of the simple theory of attenuated property rights’, Review of Industrial Organization, 8 (5), 623–34. Pierce, R.J. (2003), ‘Market manipulation and market flows’, The Electricity Journal, January/February, 39–46. Pollitt, M.G. (1995), Ownership and Performance in Electric Utilities: The International Evidence on Privatization and Efficiency, Oxford: Oxford University Press. Resende, M. (2000), ‘Regulatory regimes and efficiency in US local telephony’, Oxford Economic Papers, 52 (3), 447–70. Roche, P.-A. and B. Johannès (2001), ‘Regulation in the water and sanitation sector in France’, in C. Henry, M. Matheu and A. Jeunemaître (eds), Regulation of Network Utilities, The European Experience, Oxford: Oxford University Press. Rohlfs, J. (1974), ‘A theory of interdependent demand for a communications service’, Bell Journal of Economics and Management Science, 5 (1), 16–37. Saal, D. and D. Parker (2000), ‘The impact of privatization and regulation on the water and sewerage industry in England and Wales: a translog cost function model’, Managerial and Decision Economics, 21, 253–68. Salvanes, K.G. and S. Tjøtta (1998), ‘A test for natural monopoly with application to Norwegian electricity distribution’, Review of Industrial Organization, 13 (6), 669–85. Shy, O. (1995), Industrial Organization. Theory and Applications, Cambridge, MA: MIT Press. Sioshansi, F. and L.D. Hamlin (2004), ‘Competitive, regulated, or somewhere in between? Where do we go from here?’, Electricity Journal, June, 87–94. Spann, R.M. (1977), ‘Public versus private provision of public services’, in T.E. Borcherding (ed.), Budgets and Bureaucrats, Durham, NC: Duke University Press. Steiner, F. (2001), ‘Regulation industry structure and performance in the electricity supply industry’, OECD Economic Studies, 32, 2001/1. Stephen, A. (2001), ‘Coast to coast with the baby bells’, New Statesman, 20. Sulamaa, P. (2001), Essays in Deregulated Finnish and Nordic Electricity Markets, Helsinki: Taloustieto OY, published for ETLA. Sung, N. (1998), ‘The embodiment hypothesis revisited: evidence from the local U.S. local exchange carriers’, Information Economics and Policy, 10 (2), 219–36. Taylor, W.E. and L.D. Taylor (1993), ‘Postdivestiture long-distance competition in the United States’, American Economic Review, 83, 185–90. Teeples, R. and D. Glyer (1987), ‘Cost of water delivery systems: specification and ownership effects’, Review of Economics and Statistics, 69, 399–408. Thomas, S. (2003), ‘The seven brothers’, Energy Policy, 31, 393–403. Thomas, S. (2004a), ‘Evaluating the British model of electricity deregulation’, Annals of Public and Cooperative Economics, 75 (3), 367–98. Thomas, S. (2004b), ‘Electricity industry reforms in smaller EU countries; experience from the Nordic region’, working paper, Public Services International Research Unit (PSIRU), University of Greenwich. Thomas, S. (2005), ‘The grin of the Cheshire cat’, Energy Policy (forthcoming). Tyson, W.J. (1990), ‘Effects of deregulation on service co-ordination in the metropolitan areas’, Journal of Transport Economics and Policy, 24 (3), 283–95. Uri, N. (2003), ‘The adoption of incentive regulation and its effect on technical efficiency in telecommunications in United States’, International Journal of Production Economics, 86, 21–34. Valletti, T.M. (2003), ‘Is mobile telephony a natural oligopoly?’, Review of Industrial Organization, 22 (1), 47–65. van Dam, J. and R. Went (2001), ‘Liberalizing network utilities: what’s in it for the citizen?’, mimeo, The Netherlands Court of Audit.

Liberalization and regulation of public utility sectors 197 Vickers, J. (1995), ‘Competition and regulation in vertically related markets’, Review of Economic Studies, 62 (1), 1–18. Vickers, J. and G. Yarrow (1988), Privatization: An Economic Analysis, Cambridge, MA: MIT Press. Vogelsang, I. (1988), ‘Regulation of public utilities and nationalised industries’, in Paul G. Hare (ed.), Surveys in Public Sector Economics, Oxford: Basil Blackwell. Vogelsang, I. (2006), ‘Network utilities in the U.S. – sector reforms without privatisation’, in M. Koethenbuergher, H.-W. Sinn and J. Whalley (eds), Privatisation Experiences in the EU, Cambridge, MA: MIT Press and CESifo. Watts, P.C. (2001), ‘Heresy? The Case against deregulation of electricity generation’, The Electricity Journal, May, 19–24. White, P.J. (1990), ‘Bus deregulation: a welfare balance sheet’, Journal of Transport Economics and Policy, 24 (3), 311–32. Willner, J. (1996), ‘A comment on Bradburd: privatisation of natural monopolies’, Review of Industrial Organization, 11 (6), 869–82. Willner, J. (2001), ‘Ownership, efficiency, and political interference’, European Journal of Political Economy, 17 (6), 723–48. Willner, J. (2003), ‘Privatization: a sceptical analysis’, in D. Parker and D. Saal (eds), International Handbook on Privatization, Cheltenham, UK and Northampton, MA, USA: Edward Elgar, pp. 60–86. Willner, J. (2006a), ‘Privatisation and public ownership in Finland’, in M. Koethenbuergher, H.-W. Sinn and J. Whalley (eds), Privatisation Experiences in the EU, Cambridge, MA: MIT Press and CESifo. Willner, J. (2006b), ‘Liberalisation, competition and ownership in the presence of vertical relations’, mimeo, Åbo Akademi University. Willner, J. (2006c), ‘Privatisation and liberalisation in an industry with network externalities’, forthcoming in Economics Letters. Wydell, J. (1999), ‘Does competition in telecoms markets improve the quality of service?, Teligen research report. Yunker, J.A. (1975), ‘Economic performance of public and private enterprise’, Journal of Economics and Business, 28 (1), 60–75.

10 State-owned enterprises, privatization and industrial policy Andrea Goldstein

Introduction This chapter critically assesses alternative explanations of state ownership as a tool for industrial policy.1 It provides indirect evidence on the role that state-owned enterprises (SOEs) have played in the economic growth process of selected OECD and non-OECD countries, on the motives that have underlined the pendulum switch towards privatization, on the conditions under which SOEs have been sold, and on the consequences of this policy choice. In order to highlight major outstanding issues, a separate sub-section presents a rapid analysis of current issues in the case of France, a country that has consistently used state ownership and divestiture as a major tool of microeconomic, and indeed macroeconomic, policy. The links between market structure and efficiency, on the one hand, and the perceived differences between domestic and foreign ownership, on the other hand, are stressed. State ownership in economic development State ownership has characterized most market and, a fortiori, planning economies in the 20th century (Toninelli, 2000). Financial, strategic, political and technological reasons have been advanced to justify either the nationalization of existing private-sector firms or the ex novo creation of SOEs. Wars, in particular, played a major role in explaining the rise of state ownership: the Spanish Civil War, World War I for Germany, World War II for Italy and the USA, States nationalized to increase output and secure self-sufficiency in certain key industries, or to stimulate investment and employment when the end of war left redirected resources unemployed. State ownership of utilities, for its part, has been justified on the grounds of technological conditions leading to natural monopolies, external economies and diverging social and private discount rates. A further motive has been the desire to keep domestic control over monopoly rents produced by exploiting nonrenewable natural resources. On weaker economic grounds, SOEs have been established to build a coalition in support of new regimes and for different distributional aims: to make essential goods and services available, to create new jobs, to reduce geographical concentration of economic power. They were also meant to ensure national independence in strategic industries, in particular in countries run by military regimes; to avoid the danger of unpopular minorities’ eventual control of the economy, such as the Lebanese in West Africa and the Chinese in Southeast Asia; and to support the interests of ruling minorities in other countries, such as the Afrikaaner ‘poor whites’ and Afrikaans capital in South Africa under apartheid. A further explanation for the diffusion of state ownership in semi-peripheral countries, especially in Latin America, was advanced by the dependency literature. While accepting the primacy of private ownership, ‘bureaucratic–authoritarian regimes’ considered that the 198

State-owned enterprises, privatization and industrial policy

199

role of the state had to go beyond that of a subsidiary character to that of supplying the complementary inputs to the process of private capital accumulation (Evans, 1978). Different forms were adopted to organize the SOE sector. The Italian model of state-owned holdings, very successful in the country’s economic catch-up (Barca and Trento, 1997), proved particularly influential, for instance in Egypt (Waterbury, 1992). Elsewhere, the desire of foreign donors and lenders to bypass the public administration, regarded as inefficient, led to the concentration of scarce human resources in a limited number of relatively autonomous SOEs. The establishment of some Brazilian SOEs was indeed suggested in the 1950s by the Joint Brazil–United States Economic Development Commission (Sikkink, 1991). For a long time the links between state ownership and development appeared self-evident, as many governments in late industrializing and developing countries argued that only the public sector could overcome critical ‘bottlenecks’ in terms of physical, financial and human capital investments required by industrialization. Planning and policy interventions, it was held, could successfully substitute for markets and could tame business cycles, an important consideration after the twin shock of the Great Depression and World War II. Shleifer (1998), for whom state ownership has ‘grotesquely’ failed, observes that, in the 1930s and 1940s, ‘even many of the laissez-faire economists focused overwhelmingly on the goal of achieving competitive prices, even at the cost of accepting government ownership in non-competitive industries’ (p.134). During ‘the golden age of intervention’ (the 1950s to 1970s), the European economy grew at very high rates, a striking coincidence in time. Unfortunately, determining the casual relationship – that is, in counterfactual terms, how much higher (lower) would current and (possibly) future per capita GDP have been without state ownership – is exceedingly difficult. It is a fortiori impossible to assess the whole industrial policy, which is usually very complex. Garcia-Blanch (2001) uses an index of SOE, finding that state ownership has a negative but insignificant (and very small) effect on growth. The empirical findings of Gylfason et al. (2001) are stronger, in the sense that the ratio of investment per GDP and the growth of per capita income are both slowed down by an increase in the state sector’s share of total employment, but the authors themselves conclude that ‘even so, the results need to be interpreted with caution in view of the limited data coverage across countries and over time’ (p.447). At the micro level, various studies attempted to compare public and private enterprise performance. On aggregate, in both OECD and non-OECD countries they found SOEs to be ceteris paribus less efficient than the private sector (Shirley and Walsh, 2001). In Italy, for instance, Goldstein and Nicoletti (1996) found that, on the whole, the public enterprise sector has been characterized by a disappointing performance and widespread inefficiency. According to aggregate balance sheet data of public and private non-financial firms, public enterprises have been constantly less profitable than their private counterparts over the 1974–91 period. The comparison with private businesses is particularly telling when comparing performance of firms engaged in the same activity in competitive markets: according to 1991 data on profits, labour income and labour productivity, private firms outperformed public firms in virtually all competitive sectors. Nonetheless, the findings in this literature, while suggestive of the greater productive efficiency and profitability of private ownership, are not compelling. Most analyses focused almost exclusively on cross-sections of US companies operating in heavily

200

International handbook on industrial policy

regulated industries and, at any rate, ‘cost comparisons are rarely straightforward in the absence of a controlled experiment, and no such experiments exist for public enterprise’ (Domberger and Piggott, 1994, p.40). Other analyses rely on description, with somehow cruder post hoc propter hoc inferences. This literature relies heavily on country-specific case studies and is rich in contextual detail, although aggregation and generalization are inherently much more difficult. In some cases, state ownership played a catalyst role in industrial development. In Korea, by buying equity stakes from foreign investors that had initially provided the technological know-how, SOEs set up large integrated complexes under single management (Chesnais and Kim, 1999). Although state ownership was not as widespread as in Latin America, one of the architects of that country’s outstanding development trajectory, SaKong (1993, pp.79–83), considered the high linkage that the public enterprise sector created to the rest of the economy and its substantial contribution to the nation’s overall fixed capital formation as proof of its strategic importance for the economy as a whole. Similarly, ‘it is doubtful that Brazil could have recorded such important economic successes in the 1965–80 period without the results achieved by the public sector enterprises’ (Trebat, 1983, p.9). Project execution capabilities were particularly important in Brazil’s SOEs and public managers that mastered them often moved on to very successful careers in government or in the domestic private sector (Schneider, 1991). What little innovation activity took place in Brazil, expressed as research and development outlays, was also mostly undertaken by SOEs such as CVRD, Embraer and Petrobras (Dahlman and Frischtak, 1993). Nonetheless, various problems emerged almost everywhere. SOEs lack a clear-cut objective of profit maximization, do not face the risk of bankruptcy and the market for corporate control, because of the state’s tight grip through majority stakes, cannot act as an adequate device for disciplining managers. In theory, this concentration of ownership in the hands of a single majority investor could have served to circumvent the collective action problem that impedes small, dispersed shareholders in listed companies from wielding efficient monitoring on corporate managers. This relative advantage of public ownership, however, has often been offset by agency problems stemming from the multiplicity of ties linking the government and ministries with specific competences; parliament, political parties and the management of SOEs, all have different goals, and are possibly at crosspurposes. Hart et al. (1997) showed that state ownership generates only low-powered incentives for investment and cost reduction, whereas private ownership can deliver permanent ones. Moreover, SOEs’ managers were often appointed for political reasons, rather than for their corporate skills. Very often SOEs had become crucial elements for playing patronage politics through jobs and the servicing of constituencies and the appropriateness of public management of productive units was put in doubt. Finally, state ownership was increasingly used not only to foster development of new sectors (‘picking winners’) but also to manage the decline of ‘sunset’ ones (steel, shipbuilding and so on). Examples of these pathologies emerged in different environments. In Egypt, India, Mexico and Turkey, public ownership created its own culture and nurtured powerful associated interests, which blocked efforts to reform or liquidate SOEs (Waterbury, 1992). Even where public ownership had recorded some successes, such as in Brazil, serious problems plagued SOEs – in particular the inability to generate the cash flows needed to keep investing in fixed assets and human capital (Goldstein, 1999). In Italy, transferring control from professional managers and technocrats to politicians served the latter’s

State-owned enterprises, privatization and industrial policy

201

interests, but degenerated into such excesses that SOEs’ corruption was one of the main findings of the 1990s Clean Hands drive (Tolliday, 2000). The dynamics of privatization Since the rise of oil prices in the 1970s, the whole post-war model of economic development has been heavily criticized, and SOEs have not been exempted from this swing in the ideological pendulum from the structuralist to the neoliberal end. In the 1970s, ‘public choice’ economists sought to show that public spending and public bureaucracies grow on an excessive and inefficient scale (Niskanen, 1971), while students of government regulation identified interest group regimes that instigate and preserve inefficient regulatory practices (Stigler, 1971). Summarizing a major research programme on industrialization in seven large developing countries, Little et al. (1970) noted that ‘investment decisions can be efficiently made only in the light of a detailed knowledge of the circumstances of each case, and central planners have not had access to the information which is really required. Nor have they had large enough or sufficiently expert staffs to process the information available’ (p.5). Radicals and Marxists also agreed that capitalist societies require increased state intervention to control their internal contradictions and crises (Baran and Sweezy, 1966; O’Connor, 1973). As far as privatization is concerned, the initially timid progress recorded in some pioneering countries (United Kingdom, Chile, New Zealand) were matched and sometimes even surpassed by new countries joining the privatization bandwagon, first in the developing world (Argentina, Mexico, Malaysia), then in Europe (France, Italy) and finally in transition economies.2 Selling SOEs is part of the political game: it requires the creation of supportive coalitions and it has an impact on the distribution of power resources. Strong vested interests, such as organized labour, the military, parts of the urban bourgeoisie and most economic groups, mount in defence of nationalization. In this respect it is very telling to observe that, even in the two OECD democracies that ‘pioneered’ privatization, implementing this policy was not an explicit objective in the policy makers’ electoral manifestos. British Conservatives were ‘unfavorably disposed towards the nationalized industries’ but ‘the party placed little emphasis on the denationalization and liberalization proposals in the [1979] manifesto’ (Galal et al., 1994, p.43). In New Zealand, the ‘Labour Party stated unequivocally that it had “no plans to sell off any publicly-owned concerns” ’ (Williams, 1990, p.140), something which it then did vigorously (Evans et al., 1996). The case of Chile’s radical conservative experiment is obviously different because the ‘seven modernizations’, including SOE reform, were implemented by an authoritarian government that did not face the threat of losing power through elections. Even there ‘it was not clear at the time [of the coup] what type of change (restoration? revolution?) would be possible’ (Foxley, 1983, p.92). The priority assigned to privatization (and to using for this purpose the resources that, even when politicians do not have to be elected, still have to be spent for implementing policies) resulted from the political leeway gained by the so-called ‘Chicago Boys’, a group of economists expressing strong free market views (see Valdés, 1995, and the review by Barber, 1995). Moreover, privatization is a typical example of a collective action situation: benefits are uncertain and far from immediate, actors more likely to appropriate them are widely dispersed, and beneficiaries of the status quo may then find it easier to organize an opposition than future beneficiaries to create a support group. Finally, in developing countries

202

International handbook on industrial policy

where capital markets are thin and scarcely organized, and where it is thus difficult to realize popular capitalism through public placing of SOEs, the potential of privatization for coalition building is rather remote. Depending on the context, privatizations have been partial or full and different strategies have been used, including block sales, voucher plans and initial public offers on the stock market. The intrinsic allure of free market economics has proved less important a factor for the success of privatizations than more conjunctural conditions (Armijo, 1999). First, the nature and the seriousness of crisis deeply influence both the perceived need of radical changes and the central themes of the political debate. The need for the financial support from international organizations and donors, in particular, plays a strong role in convincing governments to adopt privatization, as shown for instance by Campbell White and Bhatia (1998) in the case of Africa. In addition, IMF conditionality has an important indirect economic benefit, in that investors are willing to pay more for privatized assets in countries that owe the IMF money (and hence that are subject to the policy constraints attached to the loans) (Brune et al., 2004). The reason for this is that investors view IMF conditionality as a signal of credible policy reform. An important issue is therefore the choice of firms to be sold and the method of sale. Understanding the economic goals to be met through divestiture, in particular the choice between an immediate reduction of macroeconomic imbalances and the medium-term achievement of efficiency improvements by liberalization and more effective regulation, is crucial. The least productive enterprises are harder to dispose of, which might make it necessary to concede easy terms well beyond a low price, when not eliminating them altogether. At the same time, countries in dire economic straits need quick ‘leading cases’ to build their reputation vis-à-vis foreign investors. The choice has thus often been to start from the best managed and more profitable SOEs, such as telecommunications, at the risk of making learning mistakes in those very cases where state assets should be more coveted by international investors. Moreover, such industries present technological and tariff complexities, requiring sophisticated tools for asset valuation and regulation that developing countries seldom possess. Timing and organizational structure also mattered. New political incumbents have higher chances of implementing reforms, not only thanks to the ‘post-election honeymoon’, but also because, according to the electoral cycle theory, there are fewer incentives to deter adopting risky and painful actions. Likewise, although the relationship between regimes and economic policies is indeterminate, being able to use emergency powers opens smoother roads on the reforms’ path. Insulated ‘technopoles’ (Domínguez, 1997) are important, especially where governments came to power with the backing of the lower class and need to prove themselves credible in the face of international investors’ fears. Thirdly, given the short-term adjustment costs of economic reforms, it is necessary to introduce compensatory mechanisms. While scholarly attention has been focused on specific target measures to face the reduction of fiscal subsidies to the poor, large business groups that dominate the industrial sector in most developing countries are hit by falling trade barriers, positive real interest rates, and more careful awarding of public credit. In order to win their support, some countervailing measures may be needed, such as setting low (relative to book value) prices (Schamis, 1999). More generally, privatization can be used as a strategy to realign the institutional framework so as to privilege the aims of some goals over the competing aspirations of others (Feigenbaum and Henig, 1994).

State-owned enterprises, privatization and industrial policy

203

The outcome of privatization Studies of privatization in OECD economies report very positive results in firms’ performance (for example, Meggison and Netter, 2001), although in many cases the experience has been for cost-cutting measures to lead to improved performance prior to privatization with little evidence of continued improvements after divestiture (see Goldstein, 2005, on Italy; Reeves and Palcic, 2004, on Ireland). In general these results mirror those for developing countries that also reveal significant increases in profitability, operating efficiency, capital spending, output and employment, which are usually greater in countries with higher per capita income (Boubakri and Cosset, 1998). In less developed regions, the statistical significance of profitability improvements is generally lower. A sample of 16 privatized firms in Africa suggests that efficiency, as well as output measured by real sales, decreased slightly but not significantly, while capital expenditures rose significantly in the post-privatization period (Boubakri and Cosset, 2002). According to Buchs (2003), in sub-Saharan Africa privatization has had a minimal ‘oneoff’ impact on the budget, employment has been adversely affected by privatization, FDI and stock markets have played a limited role in privatization transactions, and enforcement problems have plagued regulation and competition policies, leading to numerous corruption scandals. The evidence on the better productivity and profitability performance of privately owned firms is generally compelling in competitive industries. Counterfactual analyses comparing the performance of private firms with their hypothetical performance had they remained public, or ‘natural experiments’ in which a market is liberalized and some firms are privatized while other (similar) firms are not, also tend to confirm the conclusion that it is the combination of privatization and product market competition that is associated with the best outcomes (Galal et al., 1994). It must be borne in mind, however, that factors such as the type of corporate control (insider, such as former public managers and/or employees, or outsider), the nature of monitoring mechanisms, and the structure of managers’ incentives impinge on post-privatization performance (Gönenç et al., 2001). In the case of transition economies, in particular, empirical tests of the relationship between enterprise performance and ownership generally refute the hypothesis that privatization per se is associated with improved performance (Estrin and Rosevaer, 1999). Economic performance gains only occur with ‘deep’ privatization, that is, when key institutional and agency-related reforms exceed certain threshold levels (Zinnes et al., 2001). The belief that mass privatization per se, by providing powerful incentives for efficient restructuring, would release entrepreneurial endeavours has also proved naïve. The chances of fostering entrepreneurship appear greater in a gradualist environment permitting negotiated solutions to restructuring as opposed to market-driven reforms (Spicer et al., 2000). If privatization is a technical, albeit qualified, success why then does it remain widely and increasingly unpopular? Polls in different continents show that most people perceive divestiture policies to be fundamentally unfair, both in conception and execution.3 In the United Kingdom, railways users blame privatization for delays, lack of comforts and cleanness, and low safety. All over Europe, hostility also surrounds the possibility of privatizing postal and water services. In Africa more people think that ‘civil servants should keep their jobs’ than favour the downsizing of the public bureaucracy and the margin of this pro-state, anti-market sentiment is very wide (70 per cent versus 23 per cent). The margin is particularly large in Ghana, where privatization has advanced further than

204

International handbook on industrial policy

elsewhere in Africa. Finally, an impressive 75 per cent of Latin American consumers deem that privatization has led to worse services. Although the findings in this area are affected by data limitations, Birsdall and Nellis (2002) conclude that, at least in the short run and especially in transition economies, most divestiture programmes have worsened assets and income distribution. This conclusion appears to be rather robust in the case of banks, oil companies and other natural resource producers. In electricity and telecommunications, on the other hand, the poor have tended to benefit from much greater access. In Latin America, the picture is mixed (Ugaz and Waddams Price, 2003). Prices have often risen as a result of reforms, and this has frequently adversely affected low income groups more than others, either in absolute or relative (to income) terms. But most networks (including cell telephony) extended their coverage, and the poor, who have had least coverage to date, are benefiting more than other groups in most cases. Changes in access were crucial, and often dominated the effect of changing prices for those already connected. However, the deficit of connection, particularly for water and sewerage, remains high, as reforms have concentrated on urban areas. In Peru, in particular, water quality has dropped as a result of reforms and tariff increased, although this has been necessary to ensure the viability of the sector, leading to welfare gains (Torero and Pascó-Font, 2003). The requirements of regulatory capitalism Over the past two decades, the emphasis of industrial policy has shifted from direct interventionism to so-called ‘horizontal’ policies (antitrust, contract laws, property rights protection, utilities liberalization, human capital, innovation and research, and so on) aiming at guaranteeing the appropriate framework for enterprise growth and international competitiveness. Globalization has made, in many cases, the national border no longer relevant, as production is organized at the global level and the relevant resources and capabilities are obtained by following new financing and corporate governance norms and rules that are increasingly similar in different countries. All such dimensions are crucial in determining the ultimate outcome of privatization. Utilities regulation In the hitherto natural monopolies that constituted the core of public enterprises (telecoms, energy, water and transport), fast technological advances have substantially weakened the argument that public ownership of integrated monopolies is necessary to ensure investment growth, service quality and price declines in real terms. The substitution of private for public ownership, however, has been accompanied by the emergence of new regulatory challenges as policy makers try to prevent the new private owners from simply pocketing monopoly rents. The nature of the inputs used by the utilities (assets’ specificity) and of the services supplied (non-tradability), means that for privatization to be accompanied by regulatory reform the (prior or simultaneous) development of safeguarding institutions is crucial (Spiller, 1993).4 Where local capital markets remain too small to absorb large-scale privatization, attracting foreign investors is a priority and this precondition is especially relevant. An additional argument for making policy and institutional reforms before involving the private sector is that, rather than speeding up policy reforms, individual transactions may actually delay them by relaxing the resource constraints and creating a false sense of achievement. Further, individual transactions could

State-owned enterprises, privatization and industrial policy

205

complicate reform later if, for example, the government decides to increase competition but finds it too costly because it has to renegotiate existing contracts with independent power producers. In sum, ‘utility regulation has two goals: to encourage investment and to support efficiency in production and use’ (Levy and Spiller, 1994, p.216). Safeguarding institutions must signal policy makers’ commitment not to engage in opportunistic behaviour and reassuring potential and actual investors against the risk of administrative expropriation of their assets. This reduces the regulatory risk and premia on financial markets. Two features are important, regulatory governance and regulatory incentives. The former refers to ‘all the mechanisms that a society uses to restrain government discretionary moves and to solve conflicts between firms and regulators’ (Abdala, 2000), while the latter encompasses specific norms on issues such as market structure, tariffs, access, unbundling, interconnection and universal service obligations. No quick institutional, one-size-fits-all fixes are available. In each country, the evolution, and indeed the performance, of these two sets of variables depend on the institutional endowment, itself composed of many different dimensions, including the interactions between the executive and the legislative power, the ability of the judicial system to uphold property rights and review administrative agencies, the development of administrative capabilities, the pattern of conflict between contending social groups, and the nature of informal norms that tacitly constrain the actions of individuals or institutions. Political and social institutions not only affect the ability to restrain administrative action, but also have an independent impact on the type of regulation that can be designed, and hence on the appropriate balance between commitment and flexibility. In particular, to complement regulatory procedures in a welfare-enhancing way, three mechanisms restraining arbitrary administrative action must be in place (Levy and Spiller, 1994): (a) substantive restraints on the discretion of the regulator; (b) formal or informal constraints on changing the regulatory system; and (c) institutions that enforce the above formal (substantive or procedural) constraints. In the electricity industry, in particular, weaknesses in the regulatory framework have sometimes reduced the benefits of privatization and deregulation. Public authorities must devise sectoral policies that introduce and maintain competition; establish and maintain a sound regulatory framework for the remaining monopolies, public and private; keep transparency in transactions and convince investors that their investments are secure; negotiate, monitor and enforce contracts with private suppliers of management and financing; ensure that resources from privatization sales are put to productive uses; and manage the inevitable political and social tensions that arise as enterprise reforms are implemented, especially the critical issues of foreign ownership and labour lay-offs. There is, however, much less agreement on how to approach the next set of challenges (second-generation issues) for countries facing the consolidation of initial reforms. In general, these issues are related to post-privatization disputes and renegotiations between governments and the private sector and to the mechanisms necessary to promote competition in the reformed industries (Pires and Goldstein, 2001). Corporate governance In view of the goal of creating ‘people’s capitalism’, how to provide appropriate mechanisms of corporate governance that protect small shareholders, while allowing management flexibility to pursue long-term corporate goals? Advances in the theory of

206

International handbook on industrial policy

corporate finance and industrial organization, as well as the ongoing debate over the ‘best’ capitalist model, have made the analysis of corporate governance – the mechanisms whereby economic systems (in their broadest possible sense) cope with the information and incentive problems inherent in financing investments and facilitate the intertemporal transfer of income claims – a burgeoning theme in comparative institutional economics. Key elements of this literature are the sources of financial resources for corporations, the concentration of ownership, the relevance of listing, the role and composition of boards of directors, the rules governing the market for corporate control and the obstacles that they may pose to corporate control activity, the relative importance of the voice and exit mechanisms (boardroom pressures and takeovers, or internal and external control, respectively) in disciplining managers. Links between privatization and corporate governance are of two main kinds: selling SOEs exposes them to takeover and bankruptcy threats, thereby easing the corporate governance problems proper to public ownership, and they provide an opportunity to modify the distribution of ownership rights among different classes of investors, by extending public listing among large firms, increasing the number of small shareholders, and reducing ownership concentration. As mentioned before, if privatization is associated with the transfer of enterprises charged with symbolic values to domestic conglomerates and/or foreign investors it is difficult to find popular support. It is no surprise, then, that creating ‘people’s capitalism’ has ranked highly among governments’ goals. The total market value of privatized firms grew from less than $50 billion in 1983 to almost $2.5 trillion in 1999 and former SOEs are the most valuable companies in most developing countries’ stock exchanges (Meggison et al., 2000). Listed privatized firms generally have a far larger number of stockholders than do capitalization-matched private firms in the same country, although their ownership structure is often unstable. This partly reflects the fact that tackling the issues that remain after ownership is transferred from public to private hands and when no (absolute) majority shareowner emerges has seldom been an overriding concern. Potential improvements in technical efficiency following control transfer may be jeopardized if corporate control is not contestable: in this case, and especially if conduct regulation proves insufficient to open up protected markets, managers can exploit rents accruing from market position without having to worry about the threat of takeovers. Privatization policies should take such elements into account, making it imperative to introduce reforms and redress perceived inefficiencies. Governments have great discretion in pricing the SOEs they sell, especially those being sold via public share offerings, and they use this discretion to pursue political and economic ends. Most experiences, however, suggest the limited power of privatization in changing the modes of governance which are prevalent in each country’s large private companies. Further, those countries which have chosen the mass (voucher) privatization route have done so largely out of necessity and face continuing efficiency problems as a result. OECD countries’ experiments and experiences in this area have proved very influential in developing countries through the circulation of skilled personnel, in particular advisors, consultants and merchant bankers. In Italy, for example, privatization was accompanied by a legislative effort aimed at providing non-controlling shareholders (that is, both individual and collective investors) with more adequate safeguards, and at introducing the necessary conditions to allow them to monitor managers (Goldstein, 2005).

State-owned enterprises, privatization and industrial policy

207

Successive governments were unsuccessful in broadening the number of large private business groups, whereas enhancing the mobility of control to investors outside of the traditional core of Italy’s capitalism was explicitly included among the authorities’ strategic goals. On the other hand, experiences such as those of Britain and Chile underscore the fact that mass sell-offs require the development of new institutional investors, such as pension funds, that may later play an active role in corporate governance.5 Nonetheless, after the steep rise experienced in the immediate aftermath of privatizations, the slow but constant decline in the number of small shareholders in the United Kingdom highlights the difficulties in sustaining people’s capitalism in the longer run. Even in a country whose policies in this area are often taken as a benchmark, ‘control [of privatized companies] is not exerted in the forms of threats of take-over or bankruptcy; nor has it for the most part come from direct investor intervention’ (Bishop et al., 1994, p.11). Corporate governance in remaining SOEs is complicated by the lack of market value to provide constant and direct monitoring and performance evaluation for 100 per cent government-owned firms, the constraints that governments are facing in the management of their shareholding or, somehow inversely, the presence of explicit or implicit guarantees by the government on banking loans, and the scarcity of performance-related components in managers’ remuneration packages. In 2005, the OECD published the Guidelines on Corporate Governance of State-Owned Enterprises as a complement to the OECD Principles of Corporate Governance which have now become a touchstone for excellence in the corporate governance of listed and privately owned companies. Although these Guidelines, like the OECD Principles of Corporate Governance, will be non-binding, they will provide a basis for policy makers to address specific governance issues in the state sector. They will deal with such topics as how ownership should be organized, how boards should be appointed, what responsibilities they should be given, and how to ensure transparency and accountability vis-à-vis both the state, as shareholder, and society as the ultimate owner. The case of France If it is difficult to overestimate the role played by the SOE sector in the post-World War II French economy, over the past two decades privatization has had an equally important impact on industrial restructuring. In 1993, the three largest French groups, ranked by turnover, were public, as were four of the top 10 and 15 of the top 50 (Goldstein, 1996). Out of the 11 most global French groups (in terms of foreign sales over total sales) in 1993, six were public, and so were six of the 12 largest French transnational corporations ranked by foreign assets. Emblematic projects such as Minitel, Concorde, Airbus, TGV and nuclear power were all launched and brought to execution by SOEs. Nonetheless, public ownership and the higher degree of unionization characteristic of SOEs have limited their flexibility in capital operations, as proved, for example, by the failure of Renault in swapping an equity stake with Volvo in the early 1990s, or more recently by the hostility that met Snecma’s attempts to build a pan-European manufacturer of aerospace engines. Corporate governance issues, moreover, have adversely affected the performance of selected SOEs such as Crédit Lyonnais and, most recently, EdF and France Telecom, which had to write off its German UMTS licence.6 In four successive waves since 1986, authorities have sold public assets worth €71.4 bn, of which €57.2 bn have served to refinance SOEs (Agence des Participations de

208

International handbook on industrial policy 6

6 1996

5

5

4

4

3

3

2

2003

2 1

0

0

POL TUR FRA NOR AUT GRC ITA MEX BEL CZE FIN DEU HUN NLD KOR SWE IRL SPA DEN CAN PRT GBR LUX CHE NZL USA AUS

1

Note: Index 0–6 scale from lowest to highest share of public enterprises, index based on the extent of state ownership and (gross) proceeds from privatizations. Source:

OECD (2005).

Figure 10.1

The public enterprise sector in OECD countries

l’Etat, APE; that is, State Ownership Agency, 2004). In 1986–8 the stability of control over divested SOEs was ensured through the so-called noyaux durs and a plethora of cross-shareholdings. In the 1990s, the emphasis shifted towards shareholder capitalism and, under the Jospin government of 1997–2002, the creation of European industrial entities such as EADS (European Aeronautics Defense and Space company) capable of competing head-on with American and Asian rivals. Over this period the public sector almost completely abandoned competitive activities and, leaving aside some minor presences in banking and insurance, is currently centred on energy, transport, postal services and telecommunications. SOE weight in total employment has fallen from 10.5 per cent in 1985 to 5.2 per cent in 2003 (Gandin, 2004). Nevertheless, public ownership remains relatively high (Figure 10.1). The state still holds majority control over some 1500 enterprises in France, with a 2003 turnover of €247.8 bn (
3.2 per cent with respect to 2002) that is equal to 10 per cent of the market-value of non-financial production (APE, 2004). In France, where the portfolio of state holdings is conservatively evaluated to be worth €100 bn, mere accounting separation can hardly provide an effective safeguard against anti-competitive practices. Moreover, while the competition law applies to public-owned companies, the deterrence effect of sanctions may be limited by fines being ultimately paid by the public purse and by the possibility that legislative action will undo the enforcement decision (OECD, 2005). The universal provision of high-quality public services at reasonable and geographically uniform prices is one of the main justifications for maintaining public control in network industries. EC harmonization will make it necessary to adapt the notion of ‘mission de service public’ and the increase in the number of providers may lead to higher price dispersion.7 Moreover, new ways of financing the cost of universal service obligations (USO) must be found to ensure that the provision of these is competitively neutral. In 2004, the government also took steps to corporatize public enterprises supplying electricity and natural gas, in view of a possible partial privatization.

State-owned enterprises, privatization and industrial policy

209

In the most recent period, French authorities have taken a renewed interest in the management of state shareholding as a useful tool for enforcing ‘new-style’ industrial policy. The State Ownership Agency APE, which was established in 2003 as an advisory body (government’s shareholder rights are still mostly exercised by the relevant line ministries) is expected to align corporate governance to the standards of the private sector. In a similar vein, the Caisse des dépôts et consignations – a public financial institution that is the first- or second-largest investor in half of the 40 companies included in the CAC40 index – is taking a more assertive role as an active shareholder.8 Authorities are now planning to use some of the proceeds from their forthcoming privatization programme to create an agency for industrial innovation to invest in the ‘Airbuses or Ariane programmes of tomorrow’. All such signals of renewed industrial activism respond to the worry that French economic performance – in particular, the concentration in low-growing industries – is at least partly hindered by the rising weight of foreign financial investors in domestic firms and by the perceived risk that acquisitions by foreign competitors, no matter how welcome by French managers, may weaken the country’s competitive position. According to Commissariat général du Plan (2004), foreign investors do indeed own a larger percentage of listed companies’ shares in France than elsewhere in continental Europe. The early success of the Air France–KLM merger suggests the benefits of linking cross-border rationalization with privatization (after the deal the government’s stake in the flag carrier fell to 44 per cent, cut to 20 per cent later in 2004). A similarly remarkable turnaround is that of STMicroelectronics since the 1980s, under joint Franco-Italian public ownership although with increasing private ownership. But political meddling remains a risk, as proved by the delays in the consolidation of European shipyards that other governments, as well as some analysts (such as Martre, 2004), impute to Paris’s stubborn defence of French interests. The road ahead Privatization is now so widespread that it is hard to find countries not using the approach; North Korea and Cuba make up the shrunken universe of the resistant as even Myanmar has had a Privatization Commission since 1995. Counting the transition, post-communist states, well over 100 000 medium and large enterprises have been privatized in one form or another in the last two decades. Nonetheless, SOEs are big players in many economies, even after the extensive privatizations of the last 20 years. State ownership is apparent in various sectors, though it is usually most prevalent in utilities and infrastructure, with energy, transport and telecommunication being the most important industries. The role of governments in regulation and in assuring universal access to basic services is found to be essential. Large-scale private sector corporate scandals such as Enron and Parmalat have also made the general public strikingly more sceptical of the virtues of free markets. Looking into the future, if public opinion and policy makers so wish, much remains to be sold. Entire countries and regions – indeed not the smallest – have so far remained largely immune from privatization. China provides the best example in this regard. While the reform period has witnessed a shift from a wholly state-owned industrial sector towards one increasingly dominated by ‘non-state’ enterprises, starting with township and village enterprises (TVEs) and other collectively owned businesses, enterprises either wholly owned or controlled by government entities still account for 30 per cent of

210

International handbook on industrial policy

industrial output (OECD, 2002). While the government provides them with preferential credits and shelter from competition, it also constrains flexibility and their performance is often lacklustre. SOEs still employ nearly half of the urban workforce in the formal sector and concerns for social stability have so far dictated a gradual approach to SOE restructuring. China’s WTO accession is accelerating the pace of the process, but important complementary reforms in the social security system and in training programmes still largely remain on the drawing board. India is another big country where the SOE sector still plays an important role (Naib, 2004). In comparison, Vietnam’s greater success in reducing the size of the SOE sector in the 1990s was due to the introduction of a safety net composed of severance pay and early retirement schemes (O’Connor, 1996). The need to develop social safety nets for those adversely affected by privatization is indeed one of the key challenges for reformists. Likewise, WTO membership and closer integration with the European Union in the context of the Euro-Mediterranean Partnership are reinforcing pressures on MEDA countries to improve their competitive position. Privatization, regulatory reform and the creation of independent regulatory agencies in telecoms are key elements of this reform package for a number of reasons: the direct effects that divestiture receipts and foreign investment flows may have, the indirect contribution of an efficient service sector to the rest of the economy, and the positive externalities of well functioning institutions on the rest of society. In telecoms in five MEDA countries, the institutional endowment is proving a stumbling block on the reform path (Goldstein, 2003). Even in high-income OECD countries, governments must decide the appropriate level of involvement in commercial activities. In New Zealand, for instance, the Treasury (2004) identifies a weak rationale for state ownership in no less than seven enterprises where operations are wholly commercially viable and no regulations to allow disinvestment are needed.9 Where ownership is justified, improved and sometimes increased, government guidance is potentially useful, assuming all projects are required to have a positive net present value. Policy makers will need criteria against which to make decisions and to assist them in considering the impact on economic growth of a project’s externalities. Finally, in industries where externalities and other strategic considerations justify a role for the government, appropriate policy tools must be used to avoid past mistakes. A telling example of the fine line that must be walked is provided by the European defence industry, where state ownership remains important, public demand is necessarily preponderant, and a consolidation must happen swiftly, but without incurring the wrath of competition authorities. Analytical challenges also remain for improving the dialogue between theorists and empiricists (Shirley and Walsh, 2001). Many of the recent theoretical critiques of privatization address deviations from optimal firm behaviour. Not surprisingly, given the difficulty in finding such an ideal firm, the flaws and shortcomings of privatization are magnified in such a framework. Choosing a counterfactual is part of the problem. On the other hand, although in most cases gains can be documented by taking state ownership as such, this may not be ideal since the decision to relinquish control suggests that the utility function of the government has changed. A government that puts higher value on efficiency or sound fiscal policies would operate its SOEs differently as well. Empirical research consists largely of before-and-after comparisons that do not capture any change in government preferences, or controls for changes in markets. An important exception

State-owned enterprises, privatization and industrial policy

211

here is the OECD regulatory reform project, the methodology of which could be replicated for analysing emerging, transition and developing economies. In sum, if the pressures of globalization make it imperative to redefine the function of the state, understanding the development role of different forms of ownership requires focusing on hard analysis and resisting the temptation of ideology. Although privatization’s promise has been frequently oversold (Bouin, 1992), not least by international organizations,10 its ills have also been greatly exaggerated.11 When ownership transfer has been accompanied by market liberalization and proper implementation, in OECD and non-OECD countries alike, consumers and end users have benefited in terms of choice, quality and prices. When talking about regulation, however, the devil is in the details – and the boundary between a good and a bad framework is very thin. If it tilts towards rent seeking and drags with it the pursuit of high and sustainable growth, the risk of provoking a backlash against market reforms is great indeed. The future challenges for the regulatory state therefore remain substantial to ensure maximization of welfare benefits. Notes 1. 2. 3. 4. 5.

6. 7. 8. 9. 10.

11.

I thank Stephen Martin and the editors of this volume for comments on earlier drafts of this chapter. The opinions expressed and arguments employed are my sole responsibility and do not necessarily reflect those of the OECD, the OECD Development Centre or their Members. Possibly the first large-scale sale of a public enterprise was the Volkswagen issue in the early 1960s. Nonetheless, the dramatic fall of the share price after markets were hit by the Cuban missile crisis had a lasting negative impact on equity ownership culture in Germany. See Bratton et al. (2004), Commission Européenne (2002) and Latinobarómetro (2004). Firms enjoy an informational advantage over regulators, so the first-best solution to solve this problem would be to introduce competition. In Chile, for example, the takeover of the country’s dominant electricity utility, Enersis, one of the largest in emerging markets, was stalled for some months in 1998 as pension funds disputed lucrative additional terms that the management had negotiated for themselves based on important agreements concerning the future strategic direction of Enersis that they never told other shareholders about. See ‘A very big French turn-off’, The Economist, 3 July 2004. This is already the case in water services, where local authorities contracted out provision to private companies. See ‘La Caisse des dépôts affiche ses ambitions d’actionnaire’, Le Monde, 16 January 2004. These SOEs had overall 2004 revenues equal to NZ$632 million, that is, roughly a fifth of those of New Zealand Milk’s, the country’s largest private firm. I thank Bob Buckle (Growth Outcome Leader, NZ Treasury) for providing data. The candid admission of two former IMF resident representatives in Eastern Europe is probably valid elsewhere: ‘A major lesson for economists that have had the experience of living with transition is that attention must be paid to establishing the institutional underpinnings of the market. The task of creating market economies has reminded us that markets are embedded in a set of institutions and behavior patterns that economists have normally taken for granted. [. . .] Economists in general have paid no more attention to the institutional structure in which markets work than fish pay to the water they swim in’ (Allen and Haas, 2001, p. 25). In the Porto Alegre Call for Mobilisation, for example, the 2001 World Social Forum said that ‘privatisation is a mechanism for transferring public wealth and natural resources to the private sector. We oppose all forms of privatisation of natural resources and public services. We call for the protection of access to resources and public goods necessary for a decent life’.

References Abdala, M. (2000), ‘Institutional roots of post-privatization regulatory outcomes’, Telecom Development, 24 (8/9). Allen, M. and R. Haas (2001), ‘The transition in Central and Eastern Europe: the experience of two resident representatives’, IMF Staff Papers, 48 (Special Issue).

212

International handbook on industrial policy

APE (Agence des participations de l’Etat) (2004), L’État actionnaire, Paris: Ministère de l’Economie, des Finances et de l’Industrie. Armijo, L. (1999), ‘Balance sheet or ballot box? Incentives to privatize in emerging democracies’, in P. Oxhorn and P. Starr (eds), The Problematic Relationship between Economic and Political Liberalization, Boulder, CO: Lynne Rienner. Baran, P. and P. Sweezy (1966), Monopoly Capital, New York: Monthly Review Press. Barber, W.J. (1995), ‘Chile con Chicago: a review essay’, Journal of Economic Literature, 33 (4). Barca, F. and S. Trento (1997), ‘State ownership and the evolution of Italian corporate governance’, Industrial and Corporate Change, 6 (3), 533–60. Birdsall, N. and J. Nellis (2002), Winners and Losers: Assessing The Distributional Impact of Privatization, Washington: Center for Global Development. Bishop, M., J. Kay and C. Mayer (1994), ‘Introduction: privatisation in performance’, in M. Bishop, J. Kay and C. Mayer (eds) Privatisation and Economic Performance, Oxford and New York: Oxford University Press. Boubakri, N. and J.-C. Cosset (1998), ‘The financial and operating performance of newly privatized firms: evidence from developing countries’, Journal of Finance, 53, 1081–1110. —— and J.-C. Cosset (2002), ‘Does privatization meet the expectations? Evidence from African countries’, Journal of African Economies, 11 (AERC supplement 1). Bouin, O. (1992), Privatisation in Developing Countries: Reflections on a Panacea, Paris: OECD Development Centre. Bratton, M., C. Logan, W. Cho and P. Bauer (2004), ‘Afrobarometer round 2: compendium of comparative results from a 15-country survey’, Afrobarometer, Working Paper, 34. Brune, N., G. Garrett and B. Kogut (2004), ‘The International Monetary Fund and the global spread of privatization’, IMF Staff Papers, 51 (2). Buchs, T.D. (2003), ‘Privatization in Sub-Saharan Africa: some lessons from experiences to date’, mimeo, International Finance Corporation. Campbell White O. and A. Bhatia (1998), Privatization in Africa, Washington, DC: The World Bank. Chesnais, F. and H.-S. Kim (1999), ‘Petrochemicals in Korea and Brazil, in L. Krieger Mytelka (ed.) Competition, Innovation and Competitiveness in Developing Countries, Paris: OECD Development Centre. Commissariat général du Plan (2004), Mondialisation et recomposition du capital des entreprises européennes. Commission Européenne (2002), Etude sur les services d’intérêt économique général auprès des citoyens des 15 pays de l’Union Européenne, DG Marche Intérieur/DG Santé et Protection des Consommateurs. Dahlman, C. and C. Frischtak (1993), ‘National systems supporting technical advance in industry: the Brazilian experience’, in R. Nelson (ed.), National Innovation System, Oxford and New York: Oxford University Press. Domberger, S. and J. Piggott (1994), ‘Privatization policies and public enterprise: a survey’, in M. Bishop, J. Kay and C. Mayer (eds), Privatization and Economic Performance, Oxford and New York: Oxford University Press. Domínguez, J. (ed.) (1997), Technopols: Freeing Politics and Markets in Latin America in the 1990s, University Park, PA: The Pennsylvania State University Press. Estrin, S. and A. Rosevaer (1999), ‘Enterprise performance and ownership: the case of Ukraine’, European Economic Review, 43, 1125–36. Evans, L., A. Grimes, D. Teece and B. Wilkinson (1996), ‘Economic reform in New Zealand 1984–95: the pursuit of efficiency’, Journal of Economic Literature, 34. Evans, P. (1978), Dependent Development, Princeton, NJ: Princeton University Press. Feigenbaum, H.B. and J. Henig (1994), ‘The political underpinnings of privatization: a typology’, World Politics, 46 (2), 185–208. Foxley, A. (1983), Latin American Experiments in Neo-conservative Economics, Berkeley, CA: University of California Press. Galal, A., P. Leroy Jones, P. Tandon and I. Vogelsang (1994), Welfare Consequences of Selling Public Enterprises, Oxford and New York: Oxford University Press for the World Bank. Gandin, I. (2004), ‘Répertoire des entreprises contrôlées majoritairement par l’Etat au 31 décembre 2003’, Insee Résultats Économie, no. 18. Garcia-Blanch, F. (2001), ‘An empirical inquiry into the nature of South Korean economic growth’, Center for International Development at Harvard University, Working Paper, no. 74. Goldstein, A. (1996), ‘Privatisations et contrôle des entreprises: l’expérience française’, Revue Economique, 47 (6), 1309–32. —— (1999), ‘Brazilian privatisation: the rocky path from state capitalism to regulatory capitalism’, Industrial and Corporate Change, 8 (4), 673–711. —— (2003), ‘Institutional endowment and regulatory reform in telecoms: a five-country comparison in the MEDA region’, OECD Development Centre, Working Paper, no. 216.

State-owned enterprises, privatization and industrial policy

213

—— (2005), ‘Privatization in Italy 1993–2003: goals, institutions, outcomes, and outstanding issues’, in Reiner Fehn, Hans-Werner Sinn and John Whalley (eds), Privatization Experiences in the EU, Cambridge, MA: MIT Press. —— and G. Nicoletti (1996), ‘Italian privatisations in international perspective’, Cuadernos de Economia, 33 (100), 425–52. Gönenç, R., M. Maher and G. Nicoletti (2001), ‘The implementation and the effects of regulatory reform: past experience and current issues’, OECD Economic Studies, 32. Gylfason, T., T. Thor Herbertsson and G. Zoega (2001), ‘Ownership and growth’, The World Bank Economic Review, 15 (3), 431–49. Hart, O., A. Shleifer and A. Vishny (1997), ‘The proper scope of government theory and an application to prisons’, Quarterly Journal of Economics, 112 (4), 1127–61. Kikeri, S. and J. Nellis (2004), ‘An assessment of privatization’, The World Bank Research Observer, 19 (1), 87–118. Latinobarómetro (2004), Encuesta Latinobarómetro 2004. Levy, B. and P. Spiller (1994), ‘Regulation, institutions, and commitment in telecommunications’, Proceedings of the World Bank Conference on Development Economics 1993, Washington, DC. Little, I., T. Scitovsky and M. Scott (1970), Industry and Trade in Some Developing Countries: A Comparative Study, Oxford: Oxford University Press for the OECD Development Centre. Martre, H. (2004), ‘EADS, Thales et les mécanos de l’absurde’, Le Monde, 3 December. Meggison, W. and J. Netter (2001), ‘From state to market: a survey of empirical studies on privatisation’, Journal of Economic Literature, 39 (2), 321–89. Meggison, W., J. Netter and M. K. Boutchkova (2000), ‘The impact of privatisation on capital market development and individual share ownership’, Fondazione Eni Enrico Mattei, Nota di lavoro, 53. Naib, S. (2004), Disinvestment in India Policies, Procedures, Practices, Delhi: Sage Publications. Niskanen, W.A. (1971), Bureaucracy and Representative Government, Chicago: Aldine, Atherton. O’Connor, D. (1996), ‘Labour market aspects of state enterprise reform in Viet Nam’, OECD Development Centre, Working Paper, no. 117. O’Connor, J. (1973), The Fiscal Crisis of the State, New York: St. Martin’s Press. OECD (2002), China in the World Economy: The Domestic Policy Challenges, Paris. —— (2005), Economic Survey of France, Paris. Pires, J.-C. L. and A. Goldstein (2001), ‘Agências Reguladoras Brasileiras: Avaliação e Desafios’, Revista do BNDES, 16 (3–42). Reeves, E. and D. Palcic (2004), ‘Privatization policy and enterprise performance: the case of Ireland’, Annals of Public and Cooperative Economics, 75 (4), 525–48. SaKong, I. (1993), Korea in the World Economy, Washington, DC: Institute for International Economics. Schamis, H. (1999), ‘Distributional coalitions and the politics of economic reform in Latin America’, World Politics, 51 (2), 236–68. Schneider, B.R. (1991), Politics within the State, Pittsburgh, PA: University of Pittsburgh Press. Shirley, M. and P. Walsh (2001), ‘Public vs. private ownership: the current state of the debate’, The World Bank, Working Paper, no. 2420. Sikkink, K. (1991), Ideas and Institutions’ Developmentalism in Brazil and Argentina, Ithaca, NY: Cornell University Press. Shleifer, A. (1998), ‘State versus private ownership’, Journal of Economic Perspectives, 12 (4), 133–50. Spicer, A., G. McDermott and B. Kogut (2000), ‘Entrepreneurship and privatisation in Central Europe: the tenuous balance between destruction and creation’, Academy of Management Review, 25 (3), 630–49. Spiller, P. (1993), ‘Institutions and regulatory commitment in utilities’ privatization’, Industrial and Corporate Change, 2 (3), 421–52. Stigler, G. (1971), ‘The theory of economic regulation’, Bell Journal of Economics. The Treasury (2004), New Zealand Economic Growth: An Analysis of Performance and Policy. Tolliday, S.W. (2000), ‘Introduction: enterprise and state in the Italian “Economic Miracle” ’, Enterprise & Society, 1 (2), 241–8. Toninelli, P. A. (ed.) (2000), The Rise and Fall of State-owned Enterprise in the Western World, Cambridge: Cambridge University Press. Torero, M. and A. Pascó-Font (2003), ‘How has privatisation affected Peruvian urban households?’, in C. Ugaz and C. Waddams Price (eds), Utility Privatization and Regulation, Cheltenham, UK and Northampton, MA, USA: Edward Elgar. Trebat, T. (1983), Brazil’s State-Owned Enterprises, Cambridge and New York: Cambridge University Press. Ugaz, C. and C. Waddams Price (eds) (2003), Utility Privatization and Regulation: A Fair Deal for Consumers?, Cheltenham, UK and Northampton, MA, USA: Edward Elgar. Valdés, J.G. (1995), Pinochet’s Economists: The Chicago School of Economics in Chile, Cambridge and New York: Cambridge University Press.

214

International handbook on industrial policy

Waterbury, J. (1992), ‘The heart of the matter? Public enterprise and the adjustment process’, in S. Haggard and R.R. Kaufman (eds), The Politics of Economic Adjustment, Princeton, NJ: Princeton University Press. Williams, M. (1990), ‘The political economy of privatisation’, in M. Holland and J. Boston (eds), The Fourth Labour Government. Politics and Policy in New Zealand, 2nd edn, Auckland: Oxford University Press. Zinnes, C., Y. Eilat and J. Sachs (2001), ‘The gains from privatization in transition economies: is “change of ownership” enough?’, IMF Staff Papers, 48 (Special Issue), 146–70.

PART III RESEARCH, TECHNOLOGY AND INNOVATION

11 Intellectual property rights and market dynamics Fabrizio Cesaroni and Paola Giuri

1 Introduction The modern economy is currently characterized by some emerging phenomena related to the protection of intellectual property. First, technologies are increasingly bought and sold in the market, mostly in sectors in which patent protection is relevant and strong enough. Second, the open science model is more and more diffused in the technology development process of business sectors like (open source) software. Third, scientific discoveries of academic researchers, traditionally diffused in an open science system, are increasingly protected through intellectual property rights and, specifically, patents. These phenomena highlight that two contrasting models, the strong intellectual property rights (IPR) model, and the open science model, are currently diffusing in the technological and economic environment, and shaping in opposite ways the behaviour of different institutional agents and organizations. These two models have been analysed in the economic and policy literature. They have traditionally been applied to alternative institutional contexts. The strong IPR model has been associated with the business environment, while the open science model has been associated with the academic or research system. Nevertheless, more recently, a strengthening of the IPR system has occurred in the public research system, and open science models have been adopted in private sectors like the open source software. This chapter discusses these different models of intellectual property protection and their implications for the innovative activity of firms and economies and the market dynamics. The IPR and the open science models present advantages and drawbacks in terms of incentives to innovate, adoption of technological innovations, and diffusion of technologies in different institutions (large and small firms, public research institutions) and different sectors. This chapter identifies several policy issues related to the conditions under which the two models can be sustainable and beneficial for the economy. Under the strong IPR model, scientific and technological knowledge is protected by exclusive rights (such as patents) granted to the inventors. The public authority of the country where the inventor has applied for its invention ensures the enforcement of the property right. In return for such a protection, information on the invention have to be publicly disclosed, in order to enrich the total body of technical knowledge in the world. Such an ever-increasing body of public knowledge promotes further creativity and innovation from others actors. This model creates strong and direct incentives to innovate, as the owner of the property right retains all the possible returns from an innovation. Either in the case in which the innovation is used in-house to produce products or services to be sold in the final market, or in the case in which the new (scientific and technological) knowledge is sold to downstream users in the market for technologies (Arora et al., 2001), this model prevents freely available knowledge spillovers being generated. Returns from new scientific and technological knowledge are completely privatized at all different stages of technology and product development. 217

218

International handbook on industrial policy

By contrast, under the open science model, inventors are not granted any right for their inventions, nor are those who use the same inventions for subsequent developments. New scientific and technological knowledge is freely available to anyone in the public domain. While the original knowledge cannot be used on an exclusive base, specific applications of the same knowledge can be privatized in some cases. According to this model, new knowledge diffuses more rapidly among agents. Inventors do not extract any direct gain from the new knowledge, but only indirect incentives. On the one hand, their public reputation increases. On the other, and in the long run, they can gain profits from further developments or applications of the original knowledge or from career development. This model is typical of the academic research (Dasgupta and David, 1994), and partly of the software industry (Lerner and Tirole, 2002a, 2004; Gambardella and Hall, 2006). From a social point of view, this model maximizes the social returns from innovations, but requires some forms of coordination among independent agents. Absent any coordination, individual agents might behave as free riders and privatize the public knowledge. Both models present advantages and disadvantages from a private and a public perspective. In some cases, the two models operate simultaneously. A first case is the software industry, where open source software is simultaneously made available with products based on proprietary software (that is, software protected through copyright and without disclosing the source code) or patented software. A second case is the university system, where open science is the predominant model of knowledge production and diffusion, but patenting and licensing of the outcomes of the research is more and more diffused. In both cases, all the relevant issues have not yet been completely assessed and discussed, and policy implications have not been fully explored. Against this background, this chapter focuses on some key and emerging issues on intellectual property protection, and discusses potential implications of the strength or openness of IPRs for the market dynamics. Specifically, the chapter examines the effects of (stronger or weaker) patent protection on large and small firms, and on entry and growth of new technology-based firms. Strengths and weaknesses of both strong IPRs and open science models are then analysed and discussed. The chapter is organized as follows. Section 2 provides a brief introduction on recent trends in intellectual property protection in the USA and Europe. Section 3 discusses a few emerging issues associated with the strong IPRs system: the development of markets for technologies, the use or not of patents, the patenting of upstream inputs like research tools, the increasing usage of litigation and the corresponding (social) costs. Section 4 analyses the system of open science, and specifically considers the case of software and the coexistence of patenting and open science in the software sector (open source software). Section 5 concludes and discusses some policy implications. 2

Recent trends in intellectual property protection: institutional changes and empirical evidence In the last two decades, the system of intellectual property protection has witnessed a few crucial institutional transformations.1 At the same time, patenting trends in different countries and technologies and the patenting behaviour of different types of inventive actors have changed. In essence, several empirical evidences point up an intensification of the patent activities, with differential rates of growth across patent systems, countries, technologies and actors.

Intellectual property rights and market dynamics

219

The main empirical fact that has attracted the interest of several scholars is the unprecedented growth in the demand for protection of intellectual property observed during the 1980s and the 1990s. The number of US patent applications has grown from a range of 40 000–80 000 in the 1980s to at least 120 000 patents in 1995 (Kortum and Lerner, 1999).2 This discontinuity has also been examined by Hall (2004), who has shown a structural break in the time series of the growth of the number of patents in the mid-1980s. The growth rate of the number of patents was boosted from an average of 0.3 per cent before the break in 1984 to an average of 6.9 per cent after 1984. At the country level, the largest explosion in patent activity is observed for US inventors at the US Patent and Trademark Office (USPTO) and in other patent offices, while a similar upturn is not registered for European or Japanese inventors. In Europe, the rate of growth in the 1980s is evident but is not as pronounced as in the United States, while in Japan the rates of growth have been marked since the 1970s. In some technologies, like biotechnology and software, the patenting activity intensified more than the average, but in absolute terms the impact on total growth has been limited (Kortum and Lerner, 1999). Hall (2004) has also demonstrated that the intensification of patenting activities has not been specific to some technologies, but to some sectors (the electrical, electronics, computing and scientific instrument sectors). When looking at the applicant level, it is also worth noting that the growth rate of patents for new patenters has been greater than the growth rate of patents by incumbent patenters (Kortum and Lerner, 1999). During the 1980s and the 1990s, the patent system also witnessed some important institutional changes, which have contributed to the strengthening of patent protection. First, in 1982, the creation of a unified Court of Appeals for the Federal Circuit (CAFC) for patent settlements in the USA replaced the courts of appeal in local circuits. This decision marked a ‘pro-patent’ stance of the courts (Mazzoleni and Nelson, 1998; Kortum and Lerner, 1999; Jaffe, 2000; Hall and Ziedonis, 2001), as desired by many of its advocates (Granstrand, 2005). Kortum and Lerner (1999) have reported that before the creation of CAFC (between 1953 and 1978) only 62 per cent of decisions were in favour of the patent holders, while this share increased to 90 per cent after the creation of the CAFC (from 1982 to 1990). The reversing of decisions of invalidity or non-infringements also increased from 12 per cent pre-CAFC to 28 per cent post-CAFC. Moreover, in some court decisions concerning patent suits the amount of damages to be paid for infringement has been strikingly high. Rivette and Kline (2000) report that Kodak had to pay, after the infringement of Polaroid’s patents on the instantaneous camera, $925 million of damages to Polaroid and $100 million of legal fees for a 14 year-long battle. Kodak was also forced to close a plant valued at $1.5 billion, to fire 700 employees and finally to spend $500 million to buy back 16 million cameras sold over a ten-year period. This accounted for a total of more than $3 billion. In another settlement in 1991, Minolta had to pay $127.5 million to Honeywell for having used a patent, even though not used by Honeywell (Rivette and Kline, 2000). Nevertheless, the effects of CAFC were not immediate after its formation. The impact of CAFC on the favourable legal treatment of patent rights in US courts was not widely publicized until the mid-1980s. During this period, a series of business press articles focused on the strengthening of patent protection (Forbes, 1985, 1986; Fortune, 1986). Second, at the international level, the TRIPS agreement of 1994 represented another relevant step of the pro-patent era. It focused on the setting of international standards

220

International handbook on industrial policy

for IP protection, assisting particularly the interests of multinational corporations with stronger international intellectual property protection, but also engendering higher barriers for developing countries (Granstrand, 2005). Finally, in 1980, the Bayh–Dole Patent and Trademark Amendment Act was adopted. By replacing the specific Institutional Patent Agreements that universities were obliged to set with federal agencies, the new Act allowed unlimited patenting and exclusive licensing of results of federally funded research, and made these activities easier (Mowery et al., 2004). The Bayh–Dole Act contributed to a stronger protection of intellectual property by providing incentives to increasingly protect research outputs to institutions traditionally producing knowledge made freely available in the public domain. Mowery and Sampat (2005) illustrate that the share of US university patents on all US domestic patents has increased from less than 1 per cent at the end of the 1970s to more than 3.5 per cent at the end of the 1990s. However, they report that there is no evidence of a change in patent propensity over time (relative to R&D). A large proportion of the growth of the share of US university patents is covered by biomedical patents, thanks to the expanded opportunities for commercial applications of patents in the related technological fields. Thursby and Thursby (2003) also report some data from the 1991 and 2000 surveys of the Association of University Technology Managers (AUTM, 1999, 2001) showing that, following the Bayh–Dole Act, there has been a remarkable growth in the number of inventions disclosed (84 per cent), number of patent applications (238 per cent), licensing agreements (161 per cent) and royalties from licensing agreements (520 per cent). Furthermore, since the Bayh–Dole Act, Japan and some European countries also introduced legislation ruling on the transfer of intellectual property rights to individual researchers or to institutions with the aim of improving technology transfer from university to industry. An additional reinforcement of the IPR system derives from the expansion of the range of admitted subject matters for patenting undergone in the last two decades. Several authors have reported court decisions of validity of patents covering inventions previously non-patentable, such as biotechnologies, software and business methods (Cohen and Lemley, 2001; Gallini, 2002; Bessen and Hunt, 2004; Hall, 2003; Graham and Mowery, 2003). In 1980, the decision in Diamomd v. Chakrabarty allowed the patentability of genetically engineered bacteria. With the Diamond v. Diehr dispute in 1981, software began to be patentable. Finally, the State Street Bank and Trust v. Signature Financial Group in 1998 opened up the patentability of business methods. Bessen and Hunt (2004) documented a continuous growth in the number of software patents in the USA over the last 25 years. The share of total patents has also jumped from 1.1 per cent in 1976, to 5.2 per cent in 1990, and to 14.9 per cent in 2002. The software sector has also been characterized by the emergence of the open source model of software development, distinguished by disclosure and sharing of innovations (Lerner and Tirole, 2002a). However, the future sustainability of both models and the effects on the rate of patenting or increasing disclosure is still unclear. Issues concerning software will be better explored in Section 4. 3 The effects of strong IPRs on large and small (technology-based) firms Subsequent to the changes in the patent environment described in the previous section, the economic and managerial literature has started to debate the reasons for the explosion

Intellectual property rights and market dynamics

221

of patenting activities by different actors (large and small firms, public research institutions, new firms) and the effects of the strengthening of the patent system on the incentives to innovate by large and small firms, the diffusion of knowledge and the pace of technical progress, and the incentives to invest in the development of second generation technologies. Among the benefits of a strong IPR system, the most frequently cited (and commonly accepted) is that it encourages the entry of new firms and the commercialization of technologies. Firms specializing in the supply of technology-based components but without downstream manufacturing facilities have no alternative means of protecting their technological assets, and earning profits from them, to IPRs. Especially for these firms, but also for larger manufacturing firms involved in technology trade, a system of strong and enforceable IPRs is a necessary condition for their business model. Therefore strong IPRs allow markets for technology to operate, and encourage the division of innovative labour among independent firms and the entry of specialized technology producers. At the same time, increased patent protection is also associated with potential costs. First, in the case of cumulative, complex technologies, if intellectual property rights are assigned to several developers of technological components, the integration of these components in the complex product might be difficult, or at least result in an increase in transaction costs. Hence an excessive fragmentation of property rights is often associated with the so-called ‘tragedy of anti-commons’. Second, strong IPRs might induce firms to protect with patents a larger fraction of technologies, even those technologies that are not actually used in any existing product (sleeping patents), or whose only purpose is to avoid competitors to explore specific technological trajectories (blocking patents). Indeed, stronger IPRs are often associated with a larger share of unused patented technologies. Third, as indicated in the previous section, the pro-patent era has introduced another actor into the patenting arena: universities. The direct consequence of their aptitude for patenting is that the (scientific and technological) knowledge that was previously available free of charge is now traded in the market for technology. Finally, stronger IPRs are also associated with increasing costs of litigations. Apart from the financial resources spent in patent suits and settlements, the most relevant indirect effect is that increased litigation costs reduce the incentives to innovate, especially for new and small firms for which the costs of a suit may be unaffordable. As a consequence, from a social point of view, litigations often result in a transfer of resources from small to large firms. We discuss these issues more extensively in the following sections. Our aim is to provide theoretical and empirical evidence of the benefits and costs associated with a system of strong IPRs. Furthermore, we aim at discussing the effects of such benefits and costs for both large incumbent firms, and small technology-based new entrants. IPRs and markets for technology One of the commonly accepted implications of the existence of (strong) property rights is that they allow technologies and technological knowledge to be exchanged and traded in specialized markets for technology. The functioning of such markets has been empirically documented for several high-tech industries, such as chemicals, biopharmaceuticals, software, semiconductors and electronics. The number of technology transactions amongst independent organizations has grown constantly, especially during the 1990s (Arora et al.,

222

International handbook on industrial policy

2001), while the presence of an active market for technology during the late 19th and 20th centuries has been documented as well in the American glass industry (Lamoreaux and Sokoloff, 1997, 1999). In the chemical processing industry, an upstream sector of independent technology suppliers has emerged since the switch to petrochemicals after World War II (Freeman, 1968; Arora and Gambardella, 1998). New specialized suppliers – the so-called specialized engineering firms (SEFs) – have appeared and have started diffusely to license process technologies to downstream chemical companies in the developed countries, and later in the less developed countries. The market for process technologies in chemicals, however, has also been operated by large chemical corporations, which have been (and still are) involved in licensing their proprietary technologies to smaller chemical and non-chemical firms (Cesaroni and Mariani, 2001; Cesaroni, 2003). Similarly, in the biopharmaceutical industry, a division of innovative labour between new research-intense biotechnology firms and large drug manufacturers has been particularly pronounced during the 1990s and to date (Cockburn et al., 1999). Licensing contracts are the preferred way that specialist suppliers use to provide pharmaceutical companies with new chemical compounds, research services or research tools such as software programs for drug research, testing and so on (Red Herring, 1998). Finally, in the semiconductors and electronics industries, licensing and cross-licensing contracts are a common solution that established firms adopt in order to obtain the rights to use technological components owned by several and dispersed firms for the development of complex products (Grindley and Teece, 1997). The key point of these findings is that the functioning of such markets for technology is strictly linked to the existence of patents and an efficient system of protection of intellectual property rights. On the one hand, the possibility to protect part of a technology with one or more patents allows the same technology to be licensed to third parties. In situations in which a technology holder enters into a licensing agreement with a potential buyer or licensee, the information disclosed from the former can be used by the latter to invent around (Gallini, 1992). However, if patents protect the codified component of the technology, the possibility to invent around declines. The stronger the patent protection, the greater the cost of inventing around, and hence the higher is the potential advantage arising from licensing. The chemical industry provides a clear example in this respect (Arora, 1995, 1997). On the other hand, markets for technology are operated either by large corporations with unused patents in their patent portfolios or by specialized technology suppliers that use patents to protect their core technological assets. In the first case, markets for technologies represent an additional option available to large firms in order to exploit (sell or license) existing unused patent-protected technologies. Usually, firms tend to protect with patents more technologies than they actually use in their core business operations. Apart from processes of technological diversification, unused patents have a strategic motivation (sleeping patents, strategic delays and so on) (Cohen, Nelson and Walsh, 2000). These patents can be used for blocking innovative activity of rivals, or just because technologies and products based on such technologies are not completely developed. As a matter of fact, the share of unused patents in the portfolios of large corporations might be relevant. A recent survey (PatVal-EU survey by Giuri et al., 2005) provides evidence on the use and value of more than 9000 European (EPO) patents of inventors in six European countries (France, Germany, Italy, Netherlands, Spain and United Kingdom). Results from the

Intellectual property rights and market dynamics

223

PatVal-EU survey (Table 11.1) show that the share of patents used (or not) to prevent competitors from investing in specific technological areas (that is blocking patents), or to wait for future uses (that is, sleeping patents), ranges from about 18 per cent, in the case of small companies, to 40 per cent, in the case of large firms and universities. The share of unused patents is almost 45 per cent in other public research organizations. Obviously, these results reflect differences in missions and goals of the institutions, but also show that the loss of potential benefits due to unused patented technologies might be a severe concern for the economy as a whole. In some cases, however, firms might be willing to sell these unused technologies to other firms. In an often cited survey of European, North American and Japanese firms (BTG, 1998), it has been documented that unused technologies, which would have a commercial value if they were sold or licensed, amount to more than 30 per cent of patent portfolios of the surveyed firms. Furthermore, nearly two-thirds of such organizations have a share of unutilized patents, and one in eight have in excess of 1000 patents. Most of them find licensing out attractive, primarily because of potential financial, economic or commercial benefits. Examples of firms that have started to license their unused patent-protected technologies can be found in several industries. For instance, in chemicals, British Petroleum, Union Carbide and many others have been particularly active in licensing (Grindley and Nickerson, 1996; Arora et al., 2001). In electronics and semiconductors, IBM has earned almost $1 billion in 1998 by licensing its patents (Rivette and Kline, 1999) and Texas Instruments has earned royalties of almost $1.8 billion over a five-year period (Grindley and Teece, 1997). In general, many other large corporations from different industries have started to sell their patent-protected technologies. Some of these companies, such as Boeing, Monsanto, Philips and Procter & Gamble, have also used virtual market-places, such as, ‘yet2.com’.3 Apart from large corporations, the most relevant evidence of how patent protection can allow firms to enter the market for technologies is that of small technology suppliers. Table 11.1 Share of patents actually used economically by type of inventors’ employer (per cent) Licensing Licensing Crossonly & used licensing Large firms Medium firms Small firms Hospitals, foundations Public research org. Universities & education Other government org. Other Total Source: Giuri et al. (2005).

Used only

Blocking Sleeping Total (unused) (unused)

3.03 5.38 14.97 35.42 23.19 22.50 16.67 17.02

3.22 3.59 6.90 6.25 5.80 5.00 8.33 8.51

3.03 1.20 3.89 0.00 4.35 5.00 0.00 4.26

49.93 65.62 55.78 16.67 21.74 26.25 41.67 34.04

21.72 13.90 9.62 18.75 10.87 13.75 8.33 12.77

19.06 10.31 8.84 22.92 34.06 27.50 25.00 23.40

100.00 100.00 100.00 100.00 100.00 100.00 100.00 100.00

6.17

3.92

3.06

50.53

18.83

17.50

100.00

224

International handbook on industrial policy

Table 11.1 shows that small firms license almost 26 per cent of their patents (including patents that are only licensed, patents that are simultaneously licensed and used internally and cross-licensed patents). This share is only about 9 per cent for large firms and 11 per cent for medium firms (100–250 employees). The share of licensed patents is the largest for universities and other research organizations. It is also worth noting that small firms use about 15 per cent of their patents ‘only’ for licensing agreements, against 3 per cent for large firms. The core business of small technology-based firms is in fact the development of new technologies (or new technological components) to be sold to other firms located downstream in the value chain. They are not directly involved in manufacturing or marketing, and hence lack appropriate complementary assets needed to exploit their technological competences (Teece, 1986). However, if they can protect their technologies with patents, and if patents are strong enough to prevent risks of infringement or to be invented around, such small technology suppliers can trade their technologies and overcome the lack of complementary assets (Arora et al., 2001). In other words, strong patents are critical to encourage vertically specialized suppliers to enter the market, because appropriate contracts for the exchange of technology and related tacit know-how can be defined. In relative terms, the value of patent protection to these firms is higher than to large established technology holders, which have the capabilities to protect their innovations with alternative means, and compete in the market for products. As demonstrated by Arora and Fosfuri (2003), the presence of a firm in the market for products and the level of competition in this market affect its incentives to license technologies to other firms wishing to enter the same product market. In this case, a revenue effect (the increase in revenues coming from the licensing fees) and a rent dissipation effect (the reduction of revenues deriving from the increased competition in the product market) are compared. If a firm is a monopolist in the product market it will have no incentive to license its technologies, because the rent dissipation effect will never be compensated by incoming licensing fees. By contrast, if there are other competitors in the product market, the rent dissipation effect will be shared across all competitors. The effect on each firm will be marginal, and this will increase each firm’s incentive to license. On the other hand, small firms lacking the complementary production and commercialization capabilities and with no share in the product market have the maximum incentives to license their technologies. Contrarily to large companies, small independent technology suppliers have no alternative possibilities to protect their technologies and appropriate the benefits of their innovations other than by means of patent protection. In turn, the existence of a system of strong intellectual property rights enhances the incentives to innovate of the specialist technology suppliers compared to established companies. As a consequence, entry of such firms into the market for technologies should be facilitated. Empirical evidence in the cases of biopharmaceuticals (Cockburn et al., 1999), chemicals (Arora and Gambardella, 1998) and environmental technologies (Arduini and Cesaroni, 2004) shows that entry in such industries has been characterized by similar dynamics. Hence, from an industry perspective, strong IPRs not only determine an increase in the overall level of investments in R&D and innovations, but also contribute to the locus of innovative activities (Arora and Merges, 2004). By allowing the upsurge of smaller independent suppliers, strong IPRs contribute to the ‘dis-integration’ of industry structures

Intellectual property rights and market dynamics

225

and to an increasing division of innovative labour at the industry level. Consequently, the number of integrated, Chandlerian large firms should reduce accordingly. Universities as suppliers of upstream inputs Similarly to small independent technology suppliers, whose entry into the market for technology has been facilitated by the strengthening of IPRs, universities (and other Public Research Organizations – PROs) can be considered a second, relevant actor in such a market. Similarly to small technology suppliers, their entry into the market has been allowed by the possibility of protecting with patents their research results. University patenting is not a new phenomenon. In the USA, universities and other publicly funded research laboratories have been active in patenting for a long time, since the earliest years of the 20th century (Mowery and Sampat, 2001). However, since the late 1980s, the number of patents from PROs has significantly increased (Jaffe, 2000; Carlsson and Frihd, 2000). Recent data of the National Science Foundation (NSF, 2004) report an escalation of the annual number of university patents from less than 500 in the early 1980s to more than 3000 in the late 1990s (Figure 11.1). As recalled in previous sections, the change in the institutional framework that took place in the USA during the 1980s, with the introduction of the Bayh–Dole Act, has been the main driver of this phenomenon. Accordingly, many universities have modified their mission and their organization. Specific Technology Transfer Offices have been created, with the explicit aim of promoting effective licensing of patented technologies. The successful example of the University of Columbia is a representative case in point (Mowery and Sampat, 2001). As a matter of fact, following the adoption of the Bayh–Dole Act, many universities have started to protect the results of their research activity with patents, and to assign exclusive and non-exclusive licences to private companies. While the broad diffusion of

3 500

Number of patents

3 000 2 500 2 000 1 500 1 000

Total academic patenting

University patenting in the USA, 1982–98

2001

2000

1999

1998

1997

1996

1995

Top 100 universities

Source: NSF (2004).

Figure 11.1

1994

1993

1992

1991

1990

1989

1988

1987

1986

1985

1984

1983

1982

0

1981

500

All others

226

International handbook on industrial policy

Table 11.2

Reasons for patenting by type of inventors’ employer Commercial exploitation

Licensing

Cross- Prevention Blocking licensing from patents imitation

Reputation

Large firms Medium firms Small firms Hospitals, foundations & private research org. Public research org. Universities & education Other government org. Other

3.76 3.91 4.03 3.65

1.96 1.76 2.26 3.58

1.96 1.35 1.32 1.25

3.85 3.87 3.66 3.27

3.08 3.09 2.87 2.58

2.26 2.14 2.14 2.44

3.05 3.48 2.54 3.39

3.00 3.15 2.77 2.8

1.54 1.59 1.15 1.74

2.86 2.85 2.38 3.46

2.05 2.42 2.15 2.54

2.91 2.94 1.23 1.89

Total

3.78

2.06

1.79

3.76

3.01

2.27

Source: European Commission (2005).

research results was (one of) the main concerns of public research organizations, and a general objective of their mission, with the advent of university patenting the diffusion of knowledge from university to industry has been subject to additional barriers. Especially in those scientific fields, such as biomedical and genetic research, where results of scientific research have immediate technological (and commercial) applications, universities have used patents in order to extract the maximum value from them (Mowery et al., 2001). Some of them have considered the possibility of licensing their patent-protected research results as an additional source of funds, capable of balancing the reduction of funds from governments and national agencies. In so doing, they have become active in the market for technologies. As Table 11.1 revealed for the case of EPO patents of European inventors, compared to other institutional actors, universities and public research organizations have the second largest share of patents of respective patent portfolios licensed to other organizations. In general, universities and public research organizations are more involved in patent licensing than industrial firms (Giuri et al., 2005). Table 11.2 also reports the average level of importance (on a scale from 1 to 5) of different reasons for patenting by type of institution (European Commission, 2005). Licensing is considered relatively more important for universities and PROs than for firms. Prevention of imitation and blocking patents is not an important reason for patenting by universities and PROs, while only these actors consider reputation to be quite important. Interestingly, once excluded the peculiar case of individual innovators (that is, innovators who are not employees of any private or public institution), the survey of European patents (European Commission, 2005) reports that universities are also the most active institution to use their patents to create new companies, typically through processes of spin-offs creation (Figure 11.2).4 Taken together, these two figures confirm how, contrarily to firms involved in manufacturing and product commercialization, universities use patents through external exploitation processes.

Intellectual property rights and market dynamics 4.99%

Total

14.63%

Others 9.09%

Other govt. institutions

16.32%

Universities 8.15%

Public research institutions

14.58%

Private research institutions

17.11%

Small companies 5.06%

Medium-sized companies 1.91%

Large companies 0%

227

2%

4%

6%

8%

10%

12%

14%

16%

18%

Source: European Commission (2005).

Figure 11.2 Share of patents that give rise to the creation of new firms, by type of applicant Apart from motivations that induce PROs to adopt a patent licensing strategy, and conditions that motivated governments to approve the Bayh–Dole Act and subsequent patenting of university research results,5 what matters from a policy perspective is the possible implications that university patenting might have for innovative activity of companies located downstream in the innovation process. University patenting represents a radical change in the traditional aptitude of openness and diffusion of scientific knowledge, whose consequences are not completely clear and explored. While some authors (Etzkowitz et al., 2000) advocate an increased industrial finalization of research activity promoted by PROs, and an increased role of PROs in local economic development, others (Florida, 1999; Salter et al., 2000) adopt an opposite perspective. According to this alternative point of view, public research institutions and universities should pay more attention to the training of human capital, and to the longterm development of a freely available stock of knowledge that firms can then deploy for their innovation purposes. The attention given by PROs to patenting is likely to determine a more than reasonable delay in the publication of results, and an alteration of the traditional system of incentives that normally characterize the behaviour of researchers. This change would determine potential damage to the correct functioning of the scientific community, and ultimately result in a reduction of innovation potential for the whole economy, or in a modification of the research agenda of PROs towards more applied research and resolution of specific industrial needs (Dasgupta and David, 1994). In sum, although the primary political motivation for allowing universities to patent their research results has been that of permitting them to assign exclusive licences to private companies, and hence offer such companies adequate incentives to exploit results arising from university research, in the long run this policy might generate negative effects that would reduce (or counterbalance) the short-term advantages. Solutions to mitigate these negative effects, and reduce the trade-off between positive short-term advantages and negative long-term disadvantages are required.

228

International handbook on industrial policy

The problem of ‘anti-commons’ and fragmentation of IPRs If strong intellectual property rights favour the entry into the market for technology of small independent technology suppliers (and universities), and hence result in potential benefits to the economy, major problems arise when technologies are cumulative and systemic. Unfortunately, in several industries this is the case. Two canonical examples in this respect are those of the semiconductor and biopharmaceutical industries (see Hall and Ziedonis, 2001; Heller and Eisenberg, 1998). A complex product based on a cumulative technology is made of several components and subsystems, whose development and production is often demanded by different, independent firms. A system integrator willing to produce the complex product needs to collect all the different components from different producers. However, if property rights for each component have been assigned to the different producers, each producer has the right to exclude others from the use of her component, and hence to reduce the possibilities of the integrator producing the complex product. The larger the fragmentation of property rights among independent organizations, and the larger the scope of each patent, the higher the effort and costs sustained by the system integrator to collect all the property rights needed to produce the complex product.6 This problem gives rise to the so-called ‘tragedy of anti-commons’ (Heller and Eisenberg, 1998). If transaction costs were not a problem, the integrator would simply bargain with each patent holder for the appropriate fee for the licensing of each component. However, in a more realistic world, the presence of transaction costs might induce the potential system integrator not to collect all the property rights. For instance, in order to extract the maximum value of her innovation, the holder of property rights over an important component of the complex product may use her patent to block (‘hold up’) the collection of the entire set of property rights. If there are no useful technological alternatives, and if the scope of the patent is large enough to prevent the integrator inventing around, a ‘hold-up’ problem is likely to occur. Several causes contribute to the emergence of a ‘hold-up’ problem (Arora et al., 2001). The first is the extent of fragmentation of property rights amongst producers of components. When property rights are spread across dispersed patent holders, the incentive to collect all of them declines. Furthermore, the incentives to invest for the development of a complementary technology decline as well, given that each owner of a single component has to incur higher transaction costs if she is to collect all the required property rights to produce the integrated product. The second cause is the capability to find alternatives to the technological components, or to develop the components in-house. If useful alternatives to the components cannot be found, the bargaining position of the patent holder increases, hence increasing the overall level of transaction costs. Indeed, according to the traditional transaction costs literature (Williamson, 1975), such costs increase in the presence of specificity of investments and opportunistic behaviour of the agents. This is the case of holders of property rights over single components seeking to extract as much of the rent as possible. Third, a ‘hold-up’ problem is more likely to arise if holders of patents on key components are non-manufacturing firms (such as specialist technology suppliers or public research institutions). Such organizations lack downstream complementary assets needed to appropriate the returns from inventions, and are more likely to bargain more aggressively for obtaining licensing fees. By contrast, patent holders with market shares in

Intellectual property rights and market dynamics

229

the product market should find cooperative strategies more attractive, and should be less sensitive to the amount of licensing fees. Finally, the fourth cause of ‘hold-up’ problems is the case of sequential innovations. Sequential innovations put the system integrator in a very challenging position. Once the integrator has collected the rights over a share of components, the bargaining power of the patent holders of the remaining components increases drastically. Indeed, the first purchases represent a sunk cost to the system integrator, which she aims at covering in any case, even at the cost of leaving to the remaining component holders the largest part of the surplus generated by the technology. One of the main consequences of the ‘hold-up’ problem is that firms, especially those with large investments in technology-specific assets, will patent more aggressively than otherwise expected (Ziedonis, 2004). The basic idea is related to the expectations of the system integrator to be sued for patent infringements. Given the dispersion of property rights, it is unlikely that a single firm can sign contracts (for the use of external technological components) with each patent holder, and it is much more likely that she will exclude those holders of technological components that are marginal or far from the core of the technology. However, this strategic choice increases the risks of infringement of property rights. Disputes for infringement of patents often end up with a settlement in which the two parts exchange their respective patent portfolios through cross-licensing agreements. The bargaining power of each part increases the larger the extent of the patent portfolio. As a consequence, firms increase ex ante the size of their patent portfolios, expecting to enter into future (cross-)licensing negotiations. In turn, a dynamic process arises: fragmentation of property rights in sectors based on cumulative and complex technologies increases the risks of infringement, given the difficulties in collecting all the property rights spread over several, dispersed agents; in order to mitigate the risks of infringement in subsequent licensing and cross-licensing transactions, firms patent more aggressively; as a consequence of this patenting behaviour, the fragmentation of property rights increases. One of the sectors in which this mechanism is more visible is that of semiconductors and electronics. Empirical studies have demonstrated that, first, the extensive use of licensing and cross-licensing agreements is the direct consequence of large transaction costs that firms face when they try to put together dispersed property rights for the development of complex products (Grindley and Teece, 1997); second, firms that have made large investments in costly and rapidly depreciating facilities are more likely to patent more, in order to reduce the risks of being held up by external patent owners, and to negotiate access to external technologies on more favourable terms (Hall and Ziedonis, 2001). For these companies, an increase in patenting is the result of a strategic concern, rather than simply reflecting an increase in R&D productivity. Even though these results highlight potential social costs associated with a higher propensity to patent, and to assign stronger patent protection, the same authors (Hall and Ziedonis, 2001) suggest a positive effect deriving from the existence of strong IPRs on the sector dynamics, at least in the semiconductors industry. In this industry, stronger patent rights seem particularly critical to design firms, that is, those smaller firms specializing in the design of new components or products not directly involved in manufacturing. Patents are important, not only for defensive reasons and for improving their competitive position vis-à-vis market rivals. More relevant is the possibility of signalling

230

International handbook on industrial policy

their capability to defend specific market niches, and hence attracting additional venture capital funds. Furthermore, recent empirical evidence suggests that the ‘anti-commons’ problem may not be as ubiquitous as predicted by the literature. In the case of research tools in biomedical innovations, after analysing the results of interviews conducted with large pharmaceutical companies, small biotechnology firms, university personnel and patent attorneys, Walsh et al. (2003) suggest that the problem of ‘anti-commons’ is overall considered as manageable. Indeed, the vast majority of respondents declared that they have never encountered a breakdown problem, they have never modified a defined line of research owing to the impossibility of using specific discoveries or research tools (also because ample technological opportunities still exist in this field), and they consider that the difficulty of managing several licensing contracts and of paying the relevant amount of royalties can be overcome. Moreover, the authors suggest possible working solutions to resolve the ‘anti-commons’ problem and related restriction on access of research tools. Besides licensing negotiations, firms might adopt a combination of different solutions, such as inventing around, offshore acquisition of alternative technologies, challenge in court and, eventually, infringement. Apart from these private solutions, however, the problem needs an adequate institutional response made available to all potential users of upstream inputs. In the case of research tools, public databases provided by public consortium might be one of such solutions. In fields different from biomedical research tools, other ‘public’ solutions might be available or likely. For instance, in the case of software, the provision of open source software usually goes along with a Generalized Public Licence (GPL), according to which the producer of an open source program requires that all modifications of the program be subject to the same rules of openness; that is, the source code of all modifications ought to be made available to the public as with the original program (Lerner and Tirole, 2002b). It has been demonstrated that this solution creates a coordination condition between the different program producers and users, and favours the formation of a public equilibrium (Gambardella and Hall, 2006). The social cost of litigation Another negative effect associated with a system of strong intellectual property protection is that the enforcement of the property rights is usually very costly and, in general, the social loss of resources due to litigations in legal disputes is relevant. Furthermore, these costs are unevenly sustained by the different actors (individuals versus firms, large firms versus small firms, firms with different characteristics). In turn, litigation costs may both mitigate the positive effects arising from strong IPRs and reduce the incentives to innovate in the long run, especially in those technological fields where IPR protection is most relevant. Statistics on litigations show that the crude number of patent suits filed in the USA has constantly grown during the last two decades, although with some differences among different technological areas, following the constant increase of patenting over the same period of time. If the growth of patenting is taken into account (that is, if percentages of suits filings per thousand of patents are considered instead of crude number of filings), no trend increase in the filing of suits can be observed. For instance, in a recent study, Lanjouw and Schankerman (2003) analyse the filing rates by different technology fields

Intellectual property rights and market dynamics

231

(drugs, other health, chemicals, electronics, mechanical, computers, biotechnology and miscellaneous) over three time periods: 1978–84, 1985–90 and 1991–95. They show that the propensity for litigation varies among fields. For all the technological fields, and for the whole period 1978–95, the average number of suits filed (per thousand patents) is 19. The lowest rates are for chemicals (11.8 suits filed per thousand patents), electronics (15.4) and mechanical (16.9). The highest rates of filings are in ‘other health’, miscellaneous, computers and biotechnology. Interestingly, computers and biotechnology are newer technological areas compared to the others, hence demonstrating that litigation is mostly expected in those technological fields characterized by greater uncertainty about legal outcomes. However, when time trends are considered, and once the growth in patenting is taken into account, the study shows that no increase in litigation (in relative terms) occurred in any technological field. Furthermore, very few patent suits actually go to trial. Approximately 95 per cent of all patent suits settle either before or during trial. In particular, 78 per cent settle before the pre-trial hearing, an additional 16 per cent settle before trial, and 1 per cent settle during trial. Even though most patent suits end up with a settlement before or during trial, about half of the estimated legal costs of litigation are incurred before the end of the discovery phase (AIPLA, 2001), thus making litigation extremely costly. Moreover, from a social perspective, settlements before trials are likely to lead to collusive outcomes, hence generating a social loss of resources. Estimates of direct costs due to litigation are difficult to provide, and generally result in divergent figures. Relative to the amount of risk, patent litigation costs may spread from half a million dollars to three million dollars (AIPLA, 2001). Relative to the number of claims that are under examination, direct costs are estimated at half a million dollars per claim (Barton, 2000). In aggregate terms, annual patent suits may generate more than four billion dollars in legal fees (Berman, 2002). Furthermore, the time spent in trials is generally very long. The time lag between a suit filing and the final resolution has been estimated at 32 months, on average, in the case of District Court suits (Magrab, 1993). Similarly, other evidence on the length of suits shows that, while patent suits only represent 0.57 per cent of all civil cases in the federal courts, they account for over 9 per cent of those that require a trial of 20 days or more (Berman, 2002). The reason that makes legal costs so relevant is twofold. On the one hand, legal standards associated with patents and patent suits make the environment in the federal courts unfriendly. According to Hall et al. (2004), one of the difficulties opponents in patent disputes find concerns the process of proofs formation. In order to prove a claim invalid, ‘clear and convincing’ evidence has to be provided, while in ordinary civil suits the simple ‘preponderance’ of proofs is required. This difference in standards makes patent suits particularly costly, and creates strong evidentiary barriers that challengers have to face. As a matter of fact, since the creation of CAFC, and subsequent to the rapid increase in patenting in recent decades in the USA, the rate of success of challenges to patent validity has declined from 50 per cent to about 30 per cent (Lemley and Allison, 1998). On the other hand, the ways in which attorneys face patent suits contribute to the increase in direct legal costs (Berman, 2002). First, if compared to the overall monetary value that challengers can gain from the suit, the amount of legal fees associated with each

232

International handbook on industrial policy

patent suit can be considered of marginal relevance. Firms often tend to neglect the importance of such costs, convinced that additional legal costs might generate (more than) proportional litigation advantages. They tend to prefer high-quality lawyers adopting an aggressive and costly approach. Furthermore, attorneys usually prefer firms aiming at avoiding early settlements instead of waiting for trial response. The main reason is not that lawyers consciously avoid early settlements, but that they tend to prefer more aggressive firms, expecting to earn more money from protracted patent litigations. In turn, those firms adopting a less aggressive approach are eventually selected out of the financially competitive market of attorneys. As a matter of fact, according to the study of Berman (2002), the median legal fees for a case of patent litigation through court trial is about two million dollars per side, and it is constantly increasing. Apart from direct costs, patent litigations generate indirect costs whose relevance is even more important for the society as a whole. Often, rather than the actual amount of legal costs associated with litigations, it is the risk of incurring such costs that create negative effects and induce firms to modify their behaviours accordingly. In industries where patent protection is stronger, the risk of being sued for patent infringement and hence incurring the direct legal costs associated with a trial creates a barrier to entry, and prevents firms from investing in innovations. In turn, contrary to policy expectations, a system of strong property rights might reduce the overall level of investments in R&D and innovative activities, especially by small firms that are most exposed to the risks of patent litigations. Since small firms often lack the required financial resources to sustain long and costly litigation causes, they underinvest in those areas where patent protection of large firms is higher. This conclusion is supported by earlier empirical evidence. For example, Lerner (1995) shows that small firms reduce R&D investments in the fields in which the threat of litigation from larger firms is high. Similarly, Lanjouw and Lerner (2001) find that large firms use the instrument of preliminary injunction to discourage R&D by small firms. Preliminary injunction may result in the court decision to prevent both parts involved in a patent litigation using the patent under dispute, sometimes by imposing a stop in the production of goods that incorporate the innovation protected by that patent. If a large firm can manage a similar risk, the lack of adequate financial resources can represent an unaffordable obstacle for smaller firms, especially if their activity is concentrated in a single business. In their study based on 252 trials filed in Europe between January 1990 and June 1991, Lanjouw and Lerner (2001) show how the use of preliminary injunctions is a specific prerogative of large firms, which use this instrument twice as often as smaller firms. Large firms use preliminary injunctions as a means to prevent other firms investing in the same technological areas in which they have been granted patents. Even in the cases in which patent violation is not strongly demonstrable in court trials, small firms may lack the financial resources to file a patent suit. This behaviour allows larger firms to maintain monopolistic control over specific technological areas, and eventually results in higher prices for the products that integrate the technology protected by patents. Apart from possible distortion in prices, however, from the social point of view the most negative effect is that the enforcement process of intellectual property rights undermines the R&D incentives of small firms (Lanjouw and Schankerman, 2003). Moreover, the risk of litigation for small firms is larger and patent suits can be longer with respect to firms with large patent portfolios, because the latter may avoid suits or long settlements through the devel-

Intellectual property rights and market dynamics

233

opment of cooperative agreements after repeated interactions or disputes with other firms (Lanjouw and Schankerman, 2004). Furthermore, in the case of cumulative inventions, excessive protection stated by previous sentences of courts in favour of patent holders in cases of patent infringement may induce subsequent innovators to reduce their investments in that research trajectory (Llobet, 2003). The reason is that, in order to exploit an innovation based on a cumulative technology, the patent holder needs to license that patent to other firms that will build on it and develop subsequent innovations. A strong patent protection, stated by positive previous court sentences or settlements, will provide the patent holder with higher licensing fees: that is, her bargaining power rises, as well as her capacity to extract higher rents from licensing. At the same time, however, a lower number of firms will be willing to invest in that research trajectory and hence to enter into licensing agreements with the original patent holder. The two effects go in opposite directions. However, the stronger the protection given to patents in cumulative technological domains, the higher the likelihood that the second effect dominates the first. The result of previous patent litigations might create barriers to entry, and might discourage smaller improvements that would benefit the patent holder, in particular, and the whole society, in general terms. Apart from direct and indirect costs associated with patent litigation, the enforcement of patent rights has a strong effect on industry dynamics as well, and especially on the role played by smaller firms. The effects of strong property rights and high costs associated with patent litigation are particularly severe for smaller firms. This fact counterbalances the positive effect arising from strong IPRs on such a category of firms. While strong IPRs favour the entry into the market of small firms specializing in the supply of technologies or technological components, the enforcement of strong IPRs mainly punishes the same set of firms. High costs of enforcement lower their incentives to innovate, and prevent (small) firms from entering those technological areas strongly protected by incumbent patent holders. This aptitude is confirmed by empirical evidence. In one of the earliest studies on this issue based on a survey of 376 firms, Koen (1990) shows that, even though many smaller firms were aware of the fact that their patents were being violated, in 55 per cent of cases they were not taking any action to enforce their property rights, because of the high costs (and long time delays) associated with patent enforcement. Similarly, Cohen et al. (2000) show that the stronger motivations identified by small firms for failing to protect their innovations with patents are the costs associated with patent application and enforcement. In sum, although strong intellectual property rights are needed for technology-based small firms to enter the market and to supply technologies to downstream users, the costs associated with the patent system and the enforcement of patent rights mitigate this positive effect and might result in a lower than expected entry of smaller firms. Differences in costs associated with patents in different countries (for example, the USA and Europe) might induce different behaviours in firms operating in different markets. 4 Software technology: open science and patent system Software technology represents a peculiar case for understanding the advantages and drawbacks of different levels of intellectual property protection. In fact, in recent years,

234

International handbook on industrial policy

software technologies have been characterized by both a strengthening of the patent system and a trend towards the sharing and free revealing of the software source code. The traditional means of protecting intellectual property is the copyright institution, associated with proprietary licences that usually limit the use, and deny the possibility to copy, modify and distribute the software. Furthermore, proprietary software is usually protected by trade secrets and therefore the source code is not disclosed. Several scholars are trying to understand the sustainability of different models of IP protection, the reasons for different behaviour of actors, and the implications for the market dynamics. The empirical and historical analysis of software patenting documents that a progressive strengthening of the patent system occurred first in the USA and subsequently in Europe. Software was not patentable, given that patent offices have not until recently allowed the patenting of mathematical formulae and mental processes. The US Supreme Court in Gottschalk v. Benson7 in 1972 did not consent to the patenting of mathematical algorithms. During the 1970s, courts’ decisions followed this line, not allowing patents of ‘pure’ software, and in particular of inventions whose novelty lay exclusively in a computer program. The courts’ position changed in 1981, when, in Diamond v. Diehr,8 a patent for a process of temperature measurement of a machine by using a computer and a known equation was held valid. However, the protection did not cover other, different applications. Software patenting was further strengthened in 1994 with In re Appalat,9 when it was granted a patent for an algorithm applicable to a general purpose computer, without the need of specifying the applications of the computer program. Finally, in 1998, with State Street Bank & Trust v. Signature Financial Group,10 the court allowed business methods patents (Cohen and Lemley, 2001). In Europe, the recent EU Directive on computer-implemented inventions, on which a lively debate is still in progress, is moving towards the patentability of software. The growth of software patents is, however, a precedent to the policy changes just recalled. Recent studies have identified the existence of software patents through search procedures of words and combination of words referring to software (software, computer programs, algorithms and so on) in the title and abstract of patents. Bessen and Hunt (2004) have illustrated a continuous growth in the number of software patents at the USPTO over the last 25 years. The number of patents jumped from 1080 in the 1980s to more than 24 000 in 2002 (Figure 11.3). The website of the FFII software patents groups shows that at the EPO the number of patents is smaller but rapidly growing, especially since 2000.11 Bessen and Hunt (2004) also highlight the characteristics of patents and assignees of US software patents compared to all other patents in 1990–95 (Table 11.3). It is worth noting that assignees are mostly companies, that the share of patents from individuals is lower for software than for other patents, that software patents are on average most frequently cited and have a larger number of claims than other patents. This suggests that software patents have a potential broader application with respect to the rest of patents, which have fewer citations and claims. Moreover, the median of sales is larger for software patents assignees, suggesting a larger company size. Finally, the share of US assignees and inventors is larger for software patents. Contrarily, the statistics on EPO patents report that 46.84 per cent of patents in the overall period have a US applicant, 21.76 per cent a Japanese applicant. Among the European countries, Germany and Great Britain have the largest share of software patents (respectively 4.90 per cent and 4.76 per cent).

Intellectual property rights and market dynamics

235

24 891 25000 21 065

USPTO 20000

EPO

15000 9 000

10000 4 704 5 000 765

1 080

1976

1980

62

2 453 371

4 362

997

5 001

1 326

0 1985

1990

1995

2000

2002

Source: Bessen and Hunt (2004) and http://swpat.ffii.org/patents/stats/country_stat.en.html.

Figure 11.3

Table 11.3

Software patents in USA and Europe

Characteristics of software patents in the USA, 1990–95

Characteristics of the assignee Companies Individuals/not assigned Government US assignee US inventor Median of company sales ($ mn) Characteristics of the patent Average number of forward citations Average number of claims

Software

Other patents

88% 11% 2% 70% 69% 13 382

80% 18% 2% 51% 53% 8940

9.7 16.8

4.6 12.6

Source: Bessen and Hunt (2004).

The fact that large companies are the larger software patent holders seems to suggest that the increasing patentability of software may create high barriers for SMEs in the software sector. This hypothesis is also confirmed by the sectoral distribution of US software patents reported by Bessen and Hunt (2004). Some 75 per cent of software patents are applied by firms in manufacturing sectors (28 per cent in electronics, 24 per cent in mechanical, 9 per cent in instruments), with 25 per cent by firms in non-manufacturing sectors, and only 7 per cent in the software sector. It is also worth emphasizing that software patenting by firms in other sectors may create barriers for small and new firms in the software sector and to the emergence of a specialized market for software technologies. Bessen and Hunt (2004) finally suggested that software patents might not provide additional incentives

236

International handbook on industrial policy

for innovation in this sector. In fact they show that R&D investments were already high before the IP strengthening, and that the annual rate of growth of R&D investments dropped from an average of 16 per cent in the 1980s to 11 per cent in the 1990s. They point out that firms have essentially replaced R&D investments with increasing patenting, and their results are consistent with the previous findings of Hall and Ziedonis (2001) in the semiconductor industry. This sector has been characterized by a ‘patent paradox’; that is, the number of patents substantially increased during the 1980s and the 1990s, although patents in semiconductors are considered the least effective means of appropriating the returns from innovation (CMU survey, Cohen et al., 2000).12 Hall and Ziedonis found that large capital-intensive firms in the semiconductor industry patent for strategic reasons, such as blocking competitors and using patents in cross-licensing negotiations. Large firms increasingly patent latent inventions to expand their patent portfolio, also in order to avoid violating others’ patents and ‘to ensure the freedom to design and manufacture’. From a policy perspective this also implies that patent protection is not essential for the development of the software industry. Furthermore, the hypothesis of stronger incentives for innovation is scarcely supported by the recent growth of the number of developers of free/open source software. A rising number of innovators/developers in fact prefer sharing instead of protecting software knowledge (David, 2003; Lerner and Tirole, 2002a). David (2003) casts doubt on the fact that traditional means for protecting innovations are necessary for new technologies, discussing the fact that new technologies may require different institutional mechanisms and, in particular, mechanisms closer to the open science organization. On the one hand, a balance between IP protection of inventions and open science is considered healthy for technological progress and diffusion and for the creation of new firms. However the trade-off is that strong IP limits the possibility and speed of diffusion and further developments, but may positively act as a relevant signal for new and small firms. The adoption of mechanisms of open science communities such as reputation, priority and cooperation favours the development of scientific inventions. Openness allows an open validation of research methods and results, and may avoid duplication of efforts (see Dasgupta and David, 1994, for specific cases in which inefficient duplications may occur in open science). In contrast, intellectual property protection on research limits the possibility of knowledge access for new and young scientists, and in particular limits the circulation of information and research tools. Moreover, it may delay the publication of scientific research outcomes13 and lead to some of the anti-commons problems that may occur in the supply of upstream inputs and research tools. In the open source software (OSS) system, developers may freely contribute to the development of a software project, whose source code can be freely downloaded and modified according to the licensing scheme attached to the project (for a description and discussion of the OSS licences, see Lerner and Tirole, 2002b; Giuri et al., 2002). The source code is distributed together with the object, binary software, under licence schemes that allow anyone to download the source code, make modifications and redistribute the software either free of charge or for a relatively small fee (Raymond, 1999). The meaning of ‘free software’ is associated with the Free Software Foundation, founded by Richard Stallman in 1983, and the GNU General Public License (GNU/GPL) that establishes very strong users’ rights (Stallman, 1999).14 Later on, the diffusion of open source software increased substantially after the introduction of Linux by Linus Torvalds in 1991, and the

Intellectual property rights and market dynamics

237

creation of the Open Source Initiative (OSI). The OSI launched an Open Source Definition and an OSI Certification Mark applicable to OSS licences. The OSI also introduced additional types of licences, which facilitate the commercial distribution of open source software with respect to the GNU/GPL. The empirical evidence on the diffusion of open source software suggests that the open source model may produce an intensified competition among large software producers as in the cases of Linux versus Windows, or Apache versus Microsoft.15 Apart from large OSS projects like Linux or Apache, many smaller projects have been introduced in the last few years. For example, SourceForge.net, the largest web repository of OSS projects created in 1999, hosts (at the beginning of 2005) 97 950 registered projects and more than a million users.16 However, most of them are composed by a very small number of participants and are characterized by a skewed distribution of activity levels (Giuri et al., 2004). What makes the study of open source software particularly relevant in the debate on strong versus weak IP protection, is that, on the one hand, the open source system of knowledge production shares some of the mechanisms of the open science communities, where individuals signal their qualities by contributing to a common project, by disclosing their innovation, and by refereeing the others’ contribution on a peer review basis. On the other hand, some OSS producers aim at making profits in the business environment, without protecting their intellectual capital through patents or trade secrets. Von Hippel and von Krogh (2003) discuss the open source system as an intermediate model situated between the strong intellectual property protection system (called ‘private’) and the ‘collective action’ model. The latter shares some characteristics of the open science communities and of industrial contexts in which an informal exchange of know-how among producers is often frequent (Allen, 1983; von Hippel, 1987). The cases documented by Allen (1983) and von Hippel (1987) deal with information that is difficult to protect through patents, and where free revealing is more rewarding than maintaining the secret. In the case of the iron industry in the 19th century (Allen, 1983) manufacturers shared information which was critical for improving furnaces’ performance and productivity. In the case of a steel minimill process discussed by von Hippel (1987) the information shared among manufacturers is mainly non-critical and incremental, and the benefit prevails over the cost of sharing the information. The idea discussed by von Hippel and von Krogh (2003) is in fact that a collective action model can emerge as a response to a market failure in the exchange and protection of knowledge. In this case maintaining the secret is not an available option for the manufacturers, since the effectiveness of protection is very scarce. They can therefore either contribute to the common stock of available knowledge or free-ride. For the functioning of a collective action model the private reward from contributing should be larger than the private reward from free-riding. However, the risk of free-riding, that is of waiting for others’ contributions and using them, should be corrected through the provision of the right incentives, such as monetary subsidies or the rule of priority typical of the open science model (Dasgupta and David, 1994). Von Hippel and von Krogh define the open source model as ‘private–collective’ and one of its main characteristics is that the typical innovators are the users of the software produced. For producers the private incentives to innovate and protect are greater since they expect to have a larger target market for the innovation, which is composed of many users.

238

International handbook on industrial policy

By contrast the expected benefits from the innovation for users derive from the internal use, unless licensing the innovation to other users is an available option. For users therefore the loss of profits from revealing is lower than the benefits of internalizing knowledge from a common pool. The open source system is in fact typical of settings in which users’ innovations are relevant for the producers, and therefore the sharing of information among users and producers is also beneficial for the latter (von Hippel, 1988; Harhoff et al., 2003). The other actors for which the open source software system is viable are the manufacturers of complementary goods or services. The cost of revealing the information (through posting messages on the web) and reducing profits by sharing non-core information for the producer is lower than the reward from network effects and from the sales of complementary goods and services. Moreover, in the OSS ‘private collective’ model, contrarily to the collective action model, individual benefits from free revealing, such as learning and enjoyment from programming and contributing to a common project, are available only to participants who reveal the information and not to free-riders (von Hippel and von Krogh, 2003). Gambardella and Hall (2006) emphasize the fact that open source software or the production of scientific inputs may often require downstream complementary investments for appropriating the returns from innovations. The problem is that, if inputs in the production process are freely available, the downstream market may be highly competitive, since entry would not be precluded. Therefore the returns from investments would be very low and not sufficient to cover the complementary investments in downstream assets. In contrast, IP protection of upstream inputs may raise the cost of entry, limit competition and allow downstream actors to obtain rents from their investments. One possible solution could be narrowing the protection only to the innovations of the downstream producers, thus reducing the potential broad blockage of entry of downstream firms if protection is allowed for general knowledge. Therefore the development of open source software can facilitate entry and competition from small firms only when their competitive advantage resides in the complementary assets for the innovation, although with free inputs larger firms may enjoy advantages from larger-scale investments in downstream activities. Finally, for some software technologies, the sharing of software knowledge may facilitate diffusion and the emergence of a standard. For these reasons the protection of software technologies characterized by network externalities may favour the creation of monopolies (Bessen and Hunt, 2004). 5 Conclusions The aim of this chapter was to analyse the characteristics of two alternative systems of IP protection, namely, strong IPRs and open science/open source, and the effects of both on market dynamics, and more precisely on small and large firms, and on entry of technology-based firms into the market. From the analysis carried out so far it is possible to draw some conclusions. First, intersectoral differences are relevant and have to be taken into consideration. While the two models might operate together in some specific sectors (such as software), in others only one of the two seems appropriate. Conditions such as the characteristics of the knowledge base, the presence of economies of scale in research activities, the complexity and cumulativeness of the technology to be developed are crucial in defining which model offers the strongest incentives to innovators. Second, also related to the first issue, complementary assets become crucial. If new tech-

Intellectual property rights and market dynamics

239

nological knowledge requires large investments in complementary assets in order to be fully exploited, the willingness to invest in such assets will be linked to the possibility of using the relevant knowledge on a private basis. However, an open source model might also be appropriate in the case in which business actors mainly compete on the basis of complementary assets. Related to this point, understanding the different behaviour of large and small firms in the markets for technology and in the investments in complementary assets becomes crucial. For small firms focused on the development and commercialization of technologies rather than products, the IPR system seems necessary. In sectors like the pharmaceutical and biotechnological this is the case. On the other hand, the open source model might favour the entry into the software industry of new and small firms that can freely gain access to the technological knowledge and focus on downstream activities. Third, even though clear evidence has not been provided, and the literature is not univocally oriented towards this question, the two models can be expected to generate different incentives in the development of radical versus incremental innovations. Usually, radical innovations require larger investments and are more risky. Inventors will sustain such costs and risks only if they can expect to exploit the innovation on a large market, preferably on a monopolistic basis. If this is the case, only under the strong IPR model can radical innovations be pursued. By contrast, if an open source model operates, the possibility of promoting further developments on new scientific and technological knowledge is spread over several agents, all using the same (public) new knowledge. Hence, in an open source model, not only a rapid diffusion of the innovation should be expected, but also a rapid process of introduction of incremental modifications to the same innovation. Notes 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16.

For a long-term historical account of the evolution and diffusion of the patent system, see Granstrand (2005). For a longer-term evolution of the patent activities see Lerner (2002). See http://www.yet2.com. For an extensive examination of the characteristics of science-based spin-offs spurred by universities and corporations, see the chapter by Giarratana and Torrisi in this volume. These issues are extensively discussed in Chapter 12 of this volume. See also Merges and Nelson (1990) and Scotchmer (1991) for a discussion on the impact of patent scope on first-generation and second-generation innovators in cumulative and science-based innovations. 409 U.S. 63 (1972). 450 U.S. 175 (1981). 33 F.3d 1526 (Fed. Cir. 1994). 149 F.3d 1368 (Fed. Cir. 1998), cert. Denied, 525 U.S. 1093 (1999). http://swpat.ffii.org/patents/stats/country_stat.en.html. Note that the numbers in the two patent systems are not perfectly comparable since they adopted different search procedures. However they provide a clear picture of the size and dynamics of the phenomenon in the two patent systems. The other mechanisms for appropriating the returns to product and process innovation investigated by Cohen et al. (2000) are secrecy, lead time, complementary sales and services, complementary manufacturing, and other legal mechanisms. In Europe, the priority of a patented invention starts from the filing date of the patent. For this reason, US inventors delay the publication of scientific outputs if they intend to patent in Europe. GPL allows users to have access to the source code on condition that they will make the source code freely available (or at the distribution costs which tend to be zero with the Internet) to third parties (copyleft property). Empirical evidence on the web server software is available at http://www.netcraft.com/survey/archive.html. For some models, mainly inspired by the competition between Linux and Windows, see Dalle and Jullien (2001), Saint Paul (2001) and Casadesus and Ghemawat (2003). http://sourceforge.net/; last visited 26 March 2005.

240

International handbook on industrial policy

References AIPLA (American Intellectual Property Law Association) (2001), ‘Report of Economic Survey’, Washington, DC. Allen, R.C. (1983), ‘Collective invention’, Journal of Economic Behavior and Organization, 4, 1–24. Arduini, R. and F. Cesaroni (2004), ‘Development and diffusion of environmental technologies’, in F. Cesaroni, A. Gambardella and W. Garcia-Fontes (eds), R&D, Innovation and Competitiveness in the European Chemical Industry, Dordrecht: Kluwer Academic Publishers. Arora, A. (1995), ‘Licensing tacit knowledge: intellectual property rights and the market for know-how’, Economics of Innovation and New Technology, 4, 41–59. Arora, A. (1997), ‘Patent, licensing and market structure in the chemical industry’, Research Policy, 26, 391–403. Arora, A. and A. Fosfuri (2003), ‘Licensing the markets for technology’, Journal of Economic Behaviour and Organization, 52, 277–95. Arora, A. and A. Gambardella (1998), ‘Evolution of industry structure in the chemical industry’, in A. Arora, R. Landau and N. Rosenberg (eds), Chemicals and Long-Term Economic Growth, New York: John Wiley & Sons. Arora, A. and R.P. Merges (2004), ‘Specialized supply firms, property rights and firm boundaries’, Industrial and Corporate Change, 13(3), 451–75. Arora, A., A. Fosfuri and A. Gambardella (2001), Markets for Technology: The Economics of Innovation and Corporate Strategy, Cambridge, MA: MIT Press. AUTM (1999, 2001), AUTM Licensing Survey (FY 1999 and 2001). Barton, J.H. (2000), ‘Reforming the patent system’, Science, 287, 1933–4. Berman, B. (ed.) (2002), From Ideas To Assets. Investing Wisely In Intellectual Property, New York: John Wiley & Sons. Bessen, J. and R.H. Hunt (2004), ‘An empirical look at software patents’, WP No. 03/17R, Federal Reserve Bank of Philadelphia (available at www.researchoninnovation.org). British Technology Group (BTG) (1998), ‘IPR market benchmark study’ (www.btgplc.com). Carlsson, B. and A.C. Frihd (2000), ‘Technology transfer in United States universities’, paper presented at The Eighth International J.A. Schumpeter Society Conference – The Millennium Conference Change, Development and Transformation: Transdisciplinary Perspectives on the Innovation Process, Manchester (UK), 28 June–1 July. Casadesus-Masanell, R. and P. Ghemawat (2003), ‘Dynamic mixed duopoly: a model motivated by Linux vs. Windows’, Strategy Unit Working Paper 04-012, Graduate School of Business Administration, Harvard University. Cesaroni, F. (2003), ‘Technology strategies in the knowledge economy: the licensing activity of Himont’, International Journal of Innovation Management, 7 (2), June. Cesaroni, F. and M. Mariani (2001), ‘The market for knowledge in the chemical sector’, in B. Guilhon (ed.), Technology and Markets for Knowledge, Boston: Kluwer Academic Publisher. Cockburn, I., R. Henderson, L. Orsenigo and G. Pisano (1999), ‘Pharmaceuticals and biotechnology’, in D. Mowery (ed.), US Industry in 2000: Studies in Competitive Performance, Washington, DC: National Academy Press. Cohen, J.E. and M.A. Lemley (2001), ‘Patent scope and innovation in the software industry’, Columbia Law Review, 89(1), 1–57. Cohen, W.M., R.R. Nelson and J.P. Walsh (2000), ‘Protecting their intellectual assets: appropriability conditions and why U.S. manufacturing firms patent (or not)’, NBER Working Paper 7552, February. Dalle, J. and N. Jullien (2001), ‘Open-source vs. proprietary software’, working paper (http://opensource.mit. edu/online_papers.php). Dasgupta, P. and P.A. David (1994), ‘Towards a new economics of science’, Research Policy, 23, 487–521. David, P.A. (2003), ‘Does the new economy need all the old IPR institutions? Digital information goods and access to knowledge for economic development’, presentation to the UN University WIDER Conference New Economy in Development, Helsinki, 10–11 May. Etzkowitz, H., A. Webster, C. Gebhardt and B.R. Cantisano Terra (2000), ‘The future of the university of the future: evolution of ivory tower to entrepreneurial paradigm’, Research Policy, 29, 313–30. European Commission (2005), ‘Study on evaluating the knowledge economy – what are patents actually worth?’, Report for the European Commission, DG Internal Market, Contract no. MARKT/2004/09/E. Florida, R. (1999), ‘The role of the university: leveraging talent, not technology’, Issues on Science and Technology, XV(4), 67–73. Forbes (1985), ‘A change in the legal climate’, 7 October, 41. Forbes (1986), ‘A weapon at last’, 10 March, 46. Fortune (1986), ‘The surprising new power of patents’, 23 June, 57. Freeman, C. (1968), ‘Chemical process plant: innovation and the world market’, National Institute Economic Review, 74, 931–41.

Intellectual property rights and market dynamics

241

Gallini, N.T. (1992), ‘Patent policy and costly imitation’, Rand Journal of Economics, 23(1). Gallini, N.T. (2002), ‘The economics of patents: lessons from recent U.S. patent reform’, Journal of Economic Perspectives, 16, 131–54. Gambardella, A. and B. Hall (2006), ‘Proprietary vs public domain licensing in software and research products’, Research Policy (forthcoming). Giuri, P., G. Rocchetti and S. Torrisi (2002), ‘Open source software: from open science to new marketing models. An enquiry into the economics and management of open source software’, LEM working paper no. 2002/23, Sant’Anna School of Advanced Studies, Pisa. Giuri, P., M. Mariani, S. Brusoni, G. Crespi, D. Francoz, A. Gambardella, W. Garcia-Fontes, A. Geuna, R. Gonzales, D. Harhoff, K. Hoisl, C. Lebas, A. Luzzi, L. Magazzini, L. Nesta, O. Nomaler, N. Palomeras, P. Patel, M. Romanelli and B. Verspagen (2005), ‘Everything you always wanted to know about inventors (but never asked): evidence from the PatVal-EU survey’, LEM working paper no. 2005/20, Sant’Anna School of Advanced Studies, Pisa. Giuri, P., M. Ploner, F. Rullani and S. Torrisi (2004), ‘Skills and openness of OSS projects: implications for performance’, LEM working paper no. 2004/19, Sant’Anna School of Advanced Studies, Pisa. Graham, S. and D.C. Mowery (2003), ‘Intellectual property protection in the U.S. software industry’, in W. Cohen and S. Merrill (eds), The Patent System in the Knowledge-Based Economy, Washington, DC: National Academies Press. Granstrand, O. (2005), ‘Innovation and intellectual property rights’, in J. Fagerberg, D.C. Mowery and R.R. Nelson (eds), The Oxford Handbook of Innovation, Oxford: Oxford University Press. Grindley, P.C. and J. Nickerson (1996), ‘Licensing and business strategy in the chemical industry’, in R. Parr and P. Sullivan (eds), Technology Licensing: Corporate Strategies for Maximizing Value, New York: John Wiley and Sons. Grindley, P.C. and D.J. Teece (1997), ‘Licensing and cross-licensing in semiconductors and electronics’, California Management Review, 39(2), 8–41. Hall, B.H. (2003), ‘Business method patents, innovation, and policy’, working paper E03-331, Economics Department, University of California, Berkeley. Hall, B.H. (2004), ‘Exploring the patent explosion’, NBER working paper 10605. Hall, B.H and R.H. Ziedonis (2001), ‘The patent paradox revisited: an empirical study of patenting in the U.S. semiconductor industry, 1979–1995’, RAND Journal of Economics, 32(1), Spring, 101–28. Hall, B.H., S. Graham, D. Harhoff and D.C. Mowery (2004), ‘Prospects for improving U.S. patent quality via postgrant opposition’, in A.B. Jaffe, J. Lerner and S. Stern (eds), Innovation Policy and the Economy, vol. 4, Cambridge, MA: MIT Press. Harhoff, D., J. Henkel and E. Von Hippel (2003), ‘Profiting from voluntary information spillover’, Research Policy, 32, 1753–69. Heller, M. and R. Eisenberg (1998), ‘Can patents deter innovation? The anticommons in biomedical research’, Science, 28, 1 May, 698–701. Jaffe, A.B. (2000), ‘The US patent system in transition: policy innovation and the innovation process’, Research Policy, 29, 532–57. Koen, M.S. (1990), ‘Survey of small business use of intellectual property protection: report of a survey conducted by MO-SCI corporation for the small business administration’, MO-SCI Corp., Rolla, Missouri. Kortum, S. and J. Lerner (1999), ‘What is behind the recent surge in patenting?’, Research Policy, 28, 1–22. Lamoreaux, N. and K. Sokoloff (1997), ‘Location and technological change in the American glass industry during the late nineteenth and early twentieth centuries’, NBER working paper 5938, NBER, Cambridge, MA. Lamoreaux, N. and K. Sokoloff (1999), ‘Inventors, firms, and the market for technology: US manufacturing in the late nineteenth and early twentieth centuries’, in N. Lamoreaux, D. Raff and P. Temin (eds), Learning by Firms, Organizations, and Nations, Chicago: University of Chicago Press. Lanjouw, J.O. and J. Lerner (2001), ‘Tilting the table? The predatory use of preliminary injunctions’, The Journal of Law and Economics, 44(2), 573–603. Lanjouw, J.O. and M. Schankerman (2003), ‘An empirical analysis of the enforcement of patent rights in the United States’, in W. Cohen and S. Merrill (eds), Patents in the Knowledge-Based Economy, Washington, DC: National Academies Press. Lanjouw, J.O. and M. Schankerman (2004), ‘Protecting intellectual property rights: are small firms handicapped?’, Journal of Law and Economics, 45–74. Lemley, M.A. and J.R. Allison (1998), ‘Empirical evidence on the validity of litigated patents,’ American Intellectual Property Law Association Quarterly Journal, 185(26). Lerner, J. (1995), ‘Patenting in the shadow of competitors’, Journal of Law and Economics, 38, 463–96. Lerner, J. (2002), ‘Patent protection and innovation over 150 years’, NBER working paper 8977. Lerner, J. and J. Tirole (2002a), ‘Some simple economics of open source’, Journal of Industrial Economics, 2, 197–234.

242

International handbook on industrial policy

Lerner, J. and J. Tirole (2002b), ‘The scope of open source licensing’, NBER working paper 9363 (2002). Lerner, J. and J. Tirole (2004), ‘The economics of technology sharing: open source and beyond’, NBER working paper 10956. Llobet, G. (2003), ‘Patent litigation when innovation is cumulative’, International Journal of Industrial Organization, 21, 1135–57. Magrab, E.B (1993), ‘Patent validity determinations of the ITC: should U.S. district grant them preclusive effect?’, Journal of the Patent & Trademark Office Society, 75(125), 127–35. Mazzoleni, R. and R.R. Nelson (1998), ‘The benefits and costs of strong patent protection: a contribution to the current debate’, Research Policy, 27, 273–84. Merges, R. and R.R. Nelson (1990), ‘On the complex economics of patent scope’, Columbia Law Review, 90 (4), 839–916. Mowery, D.C. and B.N. Sampat (2001), ‘University patents and patent policy debates in the USA, 1925–1980’, Industrial and Corporate Change, 10(3), 781–814. Mowery, D.C. and B.N. Sampat (2005), ‘Universities in national innovation systems’, in J. Fagerberg, D.C. Mowery and R.R. Nelson (eds), The Oxford Handbook of Innovation, Oxford: Oxford University Press. Mowery, D.C., R.R. Nelson, B.N. Sampat and A.A. Ziedonis (2001), ‘The growth of patenting and licensing by U.S. universities: an assessment of the effects of the Bayh–Dole Act of 1980’, Research Policy, 30, 99–119. Mowery, D.C., R.R. Nelson, B.N. Sampat and A.A. Ziedonis (2004), The Ivory Tower and Industrial Innovation: University–Industry Technology Transfer before and after the Bayh–Dole Act, Stanford: Stanford University Press. National Science Foundation (NSF) (2004), ‘Science and engineering indicators’ (available online at http://www.nsf.gov/sbe/srs/seind04/start.htm; last visited 21 March 2005). Raymond, E. (1999), The Cathedral and the Bazaar: Musings on Linux and Open Source by an Accidental Revolutionary, Cambridge: O’Reilly. Red Herring (1998), ‘Chemical attraction: long preferring to work independently, biotech companies are now considering the advantages of consolidation’, May (54). Rivette, K.G. and D. Kline (1999), Rembrandt In The Attic: Unlocking the Hidden Value of Patents, Boston: Harvard Business School Press. Rivette, K.G. and D. Kline (2000), ‘Discovering new value of intellectual property’, Harvard Business Review, January–February, 54–66. Saint-Paul, G. (2001), ‘Growth effects of non-proprietary innovation’, CEPR discussion paper series no. 3069 (http://cepr.org/pubs/dps/DP3096.asp). Salter, A., P. D’Este, K. Pavitt, A. Scott, B. Martin, A. Geuna, P. Nightingale and P. Patel (2000), ‘Talent, not technology: the impact of publicly funded research on innovation in the UK’, SPRU (Science and Technology Policy Research), University of Sussex, Brighton (UK). Scotchmer, S. (1991), ‘Standing on the shoulders of giants: cumulative research and the patent law’, Journal of Economic Perspectives, 2, 29–41. Stallman, R. (1999), ‘The GNU operating system and the free software movement’, in C. DiBona, S. Ockman and M. Stone (eds) (1999), Open Sources: Voices from the Open Source Revolution, Sebastopol, California: O’Reilly. Teece, D.J. (1986), ‘Profiting from technological innovation. Implications for integration, collaboration, licensing and public policy’, Research Policy, 15(6), 285–305. Thursby, J.G. and M.C. Thursby (2003), ‘Intellectual property enhanced: university licensing and the Bayh– Dole Act’, Science, 301, 1052. Von Hippel, E. (1987) ‘Cooperation between rivals: informal know-how trading’, Research Policy, 16, 291–302. Von Hippel, E. (1988), The Sources of Innovation, New York: Oxford University Press. Von Hippel, E. and G. von Krogh (2003), ‘Open source software and the private–collective innovation model: issues for organization science’, Organization Science, 14(2), 209–23. Walsh, J., A. Arora and W. Cohen (2003), ‘Effects of research tool patenting and licensing on biomedical innovation’, in W. Cohen and S. Merrill (eds), Patents in the Knowledge-Based Economy, Washington, DC: National Academies Press. Williamson, O. (1975), Markets and Hierarchies: Analysis and Antitrust Implications, New York: Free Press. Ziedonis, R.H. (2004), ‘Don’t fence me in: fragmented markets for technology and the patent acquisition strategies of firms’, Management Science, 50(6), June, 804–21.

12 New approaches to intellectual property: from open software to knowledge-based industrial activities Nicolas Jullien and Jean-Benoît Zimmermann

1 Introduction Though software intellectual property could not satisfactorily fall into any existing legal framework, all countries have taken the decision to put it into the category of copyright. However, this categorization implies that the double objective of intellectual property protection is not satisfied; that is, on the one hand, to grant the inventor a provisional monopoly for exploiting his invention and, on the other hand, to oblige him to disclose the principles of his invention. Thus the use of the patent system has been increasing in the USA, while it has only been debated in Europe. This use raises other kinds of problems. Hence the alternative model of open source software, based on a very peculiar juridical tool called GPL (general public licence), tends to take on a growing importance. Its main principle is to persuade its adopters to disclose the source code of the concerned programs and of any further improvement if they circulate them, as well as the free circulation of the code under the sole condition of maintaining its ‘open’ character. This does not exclude a possible commercialization of the programs and does not limit ‘open source software’ to a non-marketable sphere. Understanding that, firms have more recently joined the world of cooperative development and of free access to source codes, using two types of strategies. First, firms distribute open software products together with the services that can help to use them, such as training, adaptation to specific cases or contexts, hotline, maintenance, upgrading, and so on, thereby enlarging their user base. Second, a growing number of enterprises have begun to ‘free’ part of their software products with a view to drawing benefits from the potential of development of the free software community or to favouring a large diffusion of a key product, imposing it as a de facto standard and looking for gains from the commercialization of proprietary complementary products. By doing this they introduce a totally different approach to intellectual property within their industrial strategies. However, these strategies do not accommodate well the GPL terms as they were and this behavioural enlargement has led to a juridical enlargement beyond the strict framework of the GPL. A number of ‘hybrid’ licences have been designed in order to control the extent of their openness. Therefore the open software approach does not represent a denial of intellectual property but a new way to manage it. Intellectual property is not rejected through the GPL, authors do not renounce their rights but only the monopoly rent such rights would authorize in a copyright regime. The main legal aspect is that, when a program is declared under GPL licence, any code derived from it or integrating GPL code lines must also be available under GPL licence. Hence the GPL status is ‘contagious’ in the sense that this status attached to any number of lines is automatically transmitted to the whole program 243

244

International handbook on industrial policy

into which they are incorporated. The authors do authorize anyone who wants to make use of their work (modifications, improvements, additional features) under the sole condition that the new product could also circulate freely. Such an approach appears exemplary in the context of a knowledge-based economy, for a growing number of industrial activities, for which the scope of knowledge that has to be mastered appears to be too large for a single even powerful agent. Knowledge as a mutual resource implies a reshaping of the value chain concept, cash flow being drawn from the usage of the knowledge base (services, complementary products), not from the knowledge itself. We shall illustrate this idea with a few examples of knowledge-based activities, including biotechnologies and health. In this chapter, we propose to explain the reasons why the copyright framework has been chosen for computer intellectual property protection, why it does not work very well, but also why a patent system may not be better suited (section 1). We will then consider the alternative model of open source software, based on a new concept of authorship rights as expressed by the GPL. We shall examine the context in which the free software movement appeared and explain the main reasons for its present industrial success (sections 2 and 3). We will defend the idea that this new way of considering intellectual property management echoes debates in other industries, such as biotechnologies (section 4). This has various implications for institutional and public policy that we discuss in the last section (section 5). 2 The nature of computer software and the question of intellectual property At the very beginning of the computer industry, the question of software protection was not relevant. The reason was that software appeared as bearing a program logic, a substitute for the wired logic of hardware, that permitted a bigger flexibility to the machines and gave them the ability to process a plurality of tasks onto an unchanged system architecture.1 With the further emergence of a clear distinction between system and application software, the computer became a universal machine able to be dedicated to any application within the sole limits of its own processing performance. However, as long as application software was developed and supplied by the computer manufacturers themselves, computer programs have been considered as joint products whose supply was an integral part of the marketing arguments for selling a given system. An influential event that led to software being considered as a market good took place as early as the end of the 1960s. This was the decision by US justice, in 1969, to impose the unbundling of the software–hardware IBM’s supply, hence a separate billing of software components. This was the conclusion given to the American federal court of justice in the European suits against IBM in the name of antitrust law and for dominant position abuse. This implied that software products had to reveal their market value and to compete against other computer suppliers’ or independent software developers’ products. Subsequently two important changes that occurred during the 1970s and deepened in the 1980s have significantly contributed to giving software its proper status. The first change is the idea of compatibility, due to Gene Amdhal, supervisor of the IBM 360 series. According to him the software system of a machine could be implemented on a quite different but appropriately structured machine. From this ‘transportability’, operating systems had to be considered as distinct goods likely to be appropriated by other actors than those who conceived and developed them. This disconnection between

New approaches to intellectual property

245

hardware and software components of a system architecture has been reinforced by the take-off of microcomputers during the first half of the 1980s, when IBM charged an external company, called Microsoft, to develop and to supply, under its own licence, the operating system of its brand new ‘personal computer’. The second change was related to application software: the emergence of software houses in the 1970s, as a consequence of the strategy of DEC to supply ‘mini-computers’ as pure data-processing systems without any application program, instead of the former bundling offer of larger computer manufacturers. Computer users could alternatively develop these programs internally from their own skills or sell them from specialized enterprises. By that time the first custommade programs gave rise to a new generation of standardized portable products able to be implemented on a large range of machines at a very low marginal cost. From these interdependent dimensions of historical evolution that gave to software the position of a market good stemmed the need to determine a framework for its intellectual property recognition and protection. In the 1970s, the early studies devoted to this question in Europe as well as in the United States recognized that none of the existing protection frameworks could be considered as totally satisfactory. Nevertheless, they recommended avoiding the long delay required for a new framework design and general adoption at the international level and the risk of a rapid obsolescence as a consequence of a quick and deep technological change (see OTA, 1992). For those reasons, most of the developed countries decided to adopt the copyright law as a common reference for software intellectual property protection, with several variants resulting from national juridical contexts. The question of software intellectual property protection is still not satisfactorily settled by the copyright protection owing to the very specific nature of the software good and its production conditions. First of all, a software product can be considered as an intellectual expression of ideas that are coded by the use of a specific programming language, with its proper vocabulary, syntax and structural rules. For this reason its protection has been considered as falling within the field of copyright. From a practical point of view, however, a software program aims at carrying out a given task, leaning on the resources of the computer it is implemented in. Alternatively, in the case of a system software, the aim is to coordinate the running of the different components of the computer architecture. For this purpose a software product will be ‘translated’ from its explicit expression, called the ‘source code’, written in a given programming language, into a new form directly ‘understandable’ by the machine and very far from human understanding. This new form, obtained through a ‘compilation’ operation, is called the ‘objectcode’; it is the same program but its initial expression is no longer readable. If the source code is not supplied jointly it can only be restored with imperfection and at some significant cost, through a heavy operation of reverse engineering. Implemented on a machine, the program is able to emulate its resources properly for a given task without requiring from the user a precise knowledge of the technical process that is set to work. In that sense it is a technology and should fall within the field of patents. The basic principle of intellectual property protection is to bring an acceptable compromise between granting incentives to the inventor through temporary monopoly rights on the commercial exploitation of his invention and favouring the diffusion of knowledge by compelling him to disclose the principles of his invention (see the Cesaroni and Giuri contribution in the present volume). In the option of software protection through copyright, the problem is that copyright protects a given expression of the ideas

246

International handbook on industrial policy

and not the ideas themselves. Software producers are therefore not obliged to disclose the source-code of the protected programs. Most of the editors do commercialize their software products in the sole form of executable programs. They generally do not reveal the ‘source code’ of the programs, that is, the explicit expression of the program architecture, procedures and algorithms. This appears totally contradictory to the aims of intellectual property protection insofar as the owner of intellectual property is not constrained at all to reveal any information on the working principles of the protected program. American jurisprudence has adopted a quite severe attitude in this regard, strengthening protection, by condemning for copyright infringement any suspected attempt at reverse engineering on copyrighted programs. This, however, does not give any satisfactory answer to the needs for compatibility and interoperability of the different programs likely to be implemented on a given computer. The general European position on this issue has been to oblige the editors to reveal the interface specifications of their programs (the formats and protocols required for exchanging data with them) or to permit a reverse engineering operation with the sole aim of disclosing these interface specifications. This position has progressively induced editors to publish their interfaces spontaneously, thereby avoiding being prosecuted for anti-competitive practices. An alternative solution to this question could be reached by replacing private specifications with public interface standards. The reason is that interface specifications do not correspond to any inventive activity but depend on arbitrary options. Their social value is the sole result of a collective process of adoption. A standardization process therefore appears as a coordination tool aiming at the conciliation between firms’ individual private preferences and collective choices issued through committees or standardization bodies (Farrell and Saloner, 1988). The archetypical illustration is the attempt supported by ISO and IEEE2 in the late 1980s to set up an Open Systems approach in the framework of Unix-based computers, aiming at the definition of non-proprietary interface standards while preserving the variety of proprietary designed architectures (Saloner, 1990). Unfortunately, this approach collapsed because of the supremacy of firms’ strategies aiming at the building and preservation of dominant market positions. In spite of the juridical preference for copyright law expressed in the 1970s, an increasing number of patents for software programs or even simple procedures or algorithms were granted during the 1990s in the United States. In Europe, the European Office of Patents remained on its initial position to grant patents only in the case in which they are an integral component of an industrial device or process. However, the European Commission has more recently submitted a new Directive aiming at the patentability of software. This evolution may have very important consequences in terms of software industry structure and innovation dynamics. On the one hand, a large part of algorithms and procedures that programmers make use of all along their development work has been considered until now as belonging to the public domain and freely available. In the absence of a real state-of-art in the field of computer software programming, patents are granted in a totally arbitrary way, giving private rights for the use of resources that had formerly been shared by professionals without any reference to their origin. On the other hand, a generalization of the patent system would imply a progressive partitioning of knowledge and practices in a domain where innovation is based on cumulativeness and complementarities. In such conditions, a strong regime of intellectual property protection would have dramatic consequences for the dynamics of innovation (Bessen and Maskin, 2000). ‘Entry

New approaches to intellectual property

247

competition and innovation may be easier if a competitor needs only to produce a single better component, which can then hook up the market range of complementary components, than if each innovator must develop an entire system’ (Farrell, 1989). This problem appears particularly crucial nowadays since increasingly complex software products have been designed thanks to modern structural programming methods. Programs are built from the combination of elementary modules into a global architecture. This approach requires both an increasing recourse to a large range of software components, portable and reusable in different contexts, and a growing proximity to the mathematical foundations of programming. This evolution makes the problem of the distinction between public and private property of modules and algorithms more acute. Copyright laws have rejected principles and algorithms from the scope of protection. Patent granting for software components, however, creates a barrier to their usage and contradicts the working mode of the whole community of software developers. The sole actors that will be able to manage such a situation are the large companies that will have the capacity to build a large portfolio of patents. The cost of the defence will be so high for SMEs that an attack for patent infringement may threaten their survival.3 As a matter of fact, the champions for the constitution of patents portfolios over the last ten years are not only software editors but also the firms that are newcomers in the information technology field and that have understood the opportunity to build the basic material for future speculative profits at low cost.4 It is also a real threat for open source software, as illustrated by the SCO case. The SCO Group, born of the merger of Caldera Systems and Santa Cruz Operations, asserts the ownership of part of the Unix codes used in the Linux kernel. It claimed one billion dollars from IBM and sent a letter to 1500 other companies to inform them of the risk they run in the case of their continuing to offer solutions derived from GNU/Linux.5 This progressive but inescapable evolution reveals a fundamental conflict between two opposite conceptions of software development and innovation, depending on whether the core resource of the activity is to be found in the creative potential of developers’ teams or in the monopoly power of the firm that employs them. The basic distinction with traditional industrial activities is that now the main input of the production process is of an informational and cognitive nature. This problem is made more complex by the fact that a growing part of software production is distributed under ‘open’ licences, making this production freely appropriable. Developers belonging to different institutions collaborate to develop such software thanks to the Internet. We will detail the origin of this movement and the reasons for its success in the following sections. 3 The free software initiative: context and goals By the end of the 1980s, the greater part of software production was realized by the private sector and the industry increasingly considered software items as products rather than the outcome of a service. Nevertheless, universities continued to disseminate their production with very liberal licences, such as the BSD licence,6 following the long tradition of public research in the USA. This system, however, had perverse effects, especially for IT professionals. As they did not have access to the source code, they could not adapt the programs to specific configurations or to specific needs. This is the origin of the Free Software Movement, initiated by Richard Stallman, quite a famous researcher at MIT at that time.7 His goal was as

248

International handbook on industrial policy

follows: to develop a whole free operating system (the kernel, but also development tools such as compilers or programming tools, graphic user interface, and so on). In the mind of its creator, ‘free’ means a broader spectrum of ‘rights’:8 ● ● ● ●

‘the freedom to run the program, for any purpose (freedom 0); the freedom to study how the program works, and adapt it to your needs (freedom 1). Access to the source code is a precondition for this; the freedom to redistribute copies so you can help your neighbour (freedom 2); the freedom to improve the program, and release your improvements to the public, so that the whole community benefits (freedom 3). Access to the source code is a precondition for this’.

A new way of managing proprietary rights As stressed in the above quotation, the main difference from private licences is the access to the way the program has been created (the source code) and the freedom to modify this source code. There is no difference from the existing free licences, such as BSD which respects these points. The main innovation lies in the creation of a specific licence, the GPL,9 which grants the ‘rights’ quoted, but which provides the developers with a new tool: the fact that, if someone wants to redistribute a program under GPL, with or without modifications he or she has made, this redistribution must be done under the same terms and conditions. This ensures that the program cannot be closed, as BSD-protected programs can. To remain true to his goal, Stallman resigned from MIT to create a foundation (the Free Software Foundation or FSF, in 1983) in charge of hiring people to develop the free operating system, since as an MIT employee his production would have been owned by MIT.10 Hence the Free Software Movement has been aware of intellectual property rights since its very beginning. GPL does not represent a denial of intellectual property but a new way to manage it. Through GPL-like licences, intellectual property is not rejected, authors do not renounce their rights, but choose to distribute and to keep open their intellectual production. This original way of managing intellectual property has been called by its initiators the ‘Copyleft’ attitude (once again mirroring the juridical reference, the copyright). This action will be analysed in the specific professional context of software production. A specific professional context Stallman has presented his arguments for initiating this project.11 Our analysis of these arguments is as follows. First, it occurred in a specific professional cultural context:12 computer research centres, which were used to collaborate in the USA (private and public ones), from the beginning of computer history, in particular in the exchange of program files. Computer science courses are partly based on the principle that it is more efficient to reuse what exists than to redevelop from scratch. This has been ‘codified’ in the hacker philosophy by Eric Raymond, among others.13 Mainly concentrated into research centres (again, public and private ones, such as IBM or Xerox), this professional culture diffused at the same time that students in computer science were hired by firms. Thus, technically, in terms of productivity, the closure of private programs and the closure of public programs (such as BSD-protected Berkeley Unix) under private distribution were perceived as very inefficient for this category of developers. For the first time

New approaches to intellectual property

249

they lost control of their working tools. This triggered the reaction of creating open, ‘free’ programs and a juridical tool granting that these programs should remain free. In addition, it must be clear that, also in this context, history mattered. A personality, Richard Stallman, has made the difference between a vague feeling of resentment towards the closure of programs in the IT professional community and the construction of a coordinated ‘reply’. He has convictions (in the debate about who has to produce a public good, Stallman states that this public good must remain public as research production, being intellectual production) and he has agreed to renounce his professional situation to defend them. He had the charisma to be respectfully heard by his ‘co-developers’. New technical tools, especially the Internet, have made this campaign possible and have largely facilitated the diffusion of FSF production (via ‘mailing’ lists and ftp sites). Finally, the diffusion of the Internet in non-US research centres first, and to the public later, has increased the number of users and developers of free software. Internet tools have allowed this movement to become a new organizational system to produce software. The consequence: a new way of producing software Internet allows developers to follow the life of a project, reading mailing lists, downloading software and documentation. It has also made it possible for different developers, localized in different places, to co-develop the same software project. This way of working based on voluntary contributions, was first described by Raymond (1999). Each project is led by a core group of developers (the ‘kernel’) which develops most of the source code.14 A larger group follows the development, sometimes reports bugs, proposes corrections (‘patches’) or new developments. An even larger group just uses the program and sometime posts questions on user mailing lists. The organizations differ across projects but some mechanisms always exist to limit the contributions and questions, so that the main developers should not be inundated by peripheral problems.15 Another very important characteristic is that big projects are split into coordinated ‘small’, easier to manage sub-projects. As explained in Demazière et al. (2004) and in Jullien (2001), individual motivations for initiating a Free Software project or for collaborating with an existing project are numerous. For highly skilled developers, it sometimes costs less to develop a program from scratch than to use one which does not exactly meets one’s needs. Once developed, a free publication allows a quicker diffusion and thus quicker feedbacks that are useful for tracking bugs and improving the functions of the program. Following the same principle as the one which guided Stallman’s initiative, developers find it very interesting to be able to adapt open source software to their needs. The cost of finding, correcting or modifying bugs is not high. The developer is also sure that the solution or the program modification will be integrated and maintained in the next generations. Career concerns (pointed out by Lerner and Tirole, 2002) or the social capital a hacker can earn maintaining or contributing to a much used free software also represent motivations for developers. However, developers we have interviewed (Demazière et al., 2004) do not seem to be aware of these motivations since they do not mention them as incentives to contribute to a free software project. However, this provides a key to understanding the rather more surprising involvement of companies in free software development, which will be analysed in the following section.

250

International handbook on industrial policy

4 A historically situated industrial success The situation Free software has been adopted in many commercial offers since the announcement made by IBM in 2001 of investing $1bn in Linux16 development: Novell, buying Ximian and SuSE, has built a new free software-based strategy; Sun has announced it will open-source license its operating system, called Solaris;17 IBM has turned its development tool software Eclipse into open-source; and even Microsoft has released some programs under open-source licence.18 As shown by Figure 12.1, Linux generates significant business. IDC, quoted by KBC Securities in a financial analysis of the Linux distribution producer MandrakeSoft,19 forecasts a market share of 33 per cent for Linux in 2007, in volume. All these studies are centred on the key element in a computer system, namely the operating system. However, open source software products are present in a growing number of markets: PHP is the leading language for the production of dynamic websites; Open Office is today a credible alternative to Microsoft Office and is gaining important clients such as the Ontario Ministry of Education,20 MySQL database software is also increasingly used, particularly in dynamic website design. The Spanish region of Extremadura has even based its IT development policy on free software, editing its own distribution of free software.21 Even if it may appear at odds with the ‘software package’ culture, this success can be explained. An analysis A first explanation First, the actual success is historically situated. Stallman initiated his program in 1984, but the appearance of free software in commercial offers can be dated from the end of the 1990s. This is mainly due to the diffusion of Internet tools, which have been developed in US universities and European research centres and distributed as ‘free’ software (Sendmail server, HTML language, NCSA web server, and so on). At the beginning of the diffusion of the Internet into organizations (firms, administrations), servers

90% 80% 70% 60%

Market share

Growth

50% 40% 30% 20% 10% 0% Unix

Figure 12.1

Linux

Windows

Others (incl. Novell)

Global market share and evolution for operating systems (2003)

New approaches to intellectual property

251

were installed by engineers who had discovered these tools at university when they were students. As, in addition, they often did not have any budgets, they installed them at the lowest cost: ‘free’ software. We may consider that the installed base in universities has initiated a standardization effect, or an ‘increasing return to adoption’,22 to use Arthur’s (1988, 1989) terms. However this does not explain the never-failing diffusion of free software outside the Internet sphere and outside low-cost, low-budget projects, which is today very impressive,23 or the financial reward system which should repay those who have invested in open-source software development and diffusion. This is based on the prevailing features of these open systems, which propose solutions of very good quality at a very low cost to professionals. The quality of the produced software More than mere public research products, free, open-source programs were, first and foremost, tools developed by user-experts, to meet their own needs. These user-experts are behind many software development initiatives (among them Linux, Apache and Samba) and continuously still improve them. The organization24 of production has obtained remarkable results in terms of quality and speed of improvement.25 A recent report by the Coverity company states that ‘Linux is a very good system in terms of bug density’ and may have many times fewer bugs than typical commercial software.26 This is undoubtedly due to the free availability of the sources. This allowed skilled users to test the software programs, to study their code and correct it if they found errors. The higher the number of contributors, the greater the chance that one of them finds an error, and knows how to correct it. Programming rules are also very strict, to make this reading possible.27 All this contributes to guarantee minimum thresholds of robustness for the software. Cooperative work and the fact that the software programs are often a collection of simultaneously evolving small-scale projects also require that communication interfaces be made public and be ‘normalized’.28 Open codes do facilitate the checking of this compatibility, together with the modification of the interfaces if necessary. It is also remarkable to note that, in order to avoid the reproduction of diverging versions of Unix, computer firms have set up organizations which must guarantee the compatibility of the various versions and distribution of Linux.29 All this led Zimmermann (1999, p.166) to argue that, ‘if the solution of the open systems has suffered because of a recent return to the more isolated proprietary systems, the solution offered by open-source programs could contribute to go beyond the limits and contradictions of the systems of intellectual property. It might even lead to a thorough reconfiguration of the software industry’. The use of such programs by companies can be seen as the creation of professional tools to coordinate collectively to create components and software program bricks which are both reliable and, again, ‘normalized’. Up to now, this collective, normalized base has been lacking within the information technology industry (Dréan, 1996). This organization only works with software-literate people and companies that are able to define their needs in terms of software specifications, to evaluate a program download, to install it and to hack it if needed, just as for classical ‘private’ programs. The majority of users are firms, administrations or even single users and are not always competent. That

252

International handbook on industrial policy

is the reason why some members of the open-source movement have been arguing that ‘companies should be created and that this activity should be profitable’ (Ousterhout, 1999, p.45). This is the origin of the first companies which have based their business on free, GPL-protected software. The business Two periods and two kinds of actors in the business development can be distinguished: the newcomers period, with new firms coming with these new programs, trying to propose a new business relationship, and the incumbent present period, where firms see the success of the products and try either to integrate these products into their offers or to adapt the open source concept to their business. The global result is a deep evolution of the market. The first business was to address individual people or companies wanting to test or install Linux or other open-source software, in a context where Internet broadband for individual connection was expensive. Two new types of firm appeared. First, there were ‘distribution editors’, like RedHat, Mandrakesoft or SuSE which proposed the assembling of multiple software programs around one operating system, and developed tools to install and manage these programs. They sold this in CD-Rom, through the usual networks (bookstores, IT retailers). However, nowadays it is not profitable to produce and distribute standard GNU/Linux distributions through this network. The limits of such a positioning lie in the fact that, since these distributors use free products, they can encounter competition, since similar products can be distributed by other producers. Free software does not constitute a source of income but is simply a loss leader for maintenance services. Today, these companies hope to develop resources by proposing services with added value, thereby increasing their ‘ARPU’.30 Doing so, they compete with service companies. Second, service companies, like Linagora in France, propose solutions to organizations based on open-source software. They see their competitive advantage in the quality of the open-source software and the fact that they are easier to adapt and to maintain. Producers can more easily guarantee, through a contract, the reliability of the programs they use, as they can intervene themselves on these software programs. In addition, when programs are GPL-protected, the fact that the software program sources are accessible and that the evolution of these programs is not controlled by a firm can reassure the adopter: the solution conforms and will continue to conform to the standards. It will thus remain interoperable with the other programs he/she uses.But today these firms compete with traditional service companies, or constructors such as IBM, which use the same open software and propose the same services. For the time being, newcomers have not acquired dominant positions in the computer market thanks to these new ways of producing software. This is primarily due to the successful diffusion of the use of open-source software in business rather than to a lack of success of these products. Newcomers face tough competition because incumbents have adopted these open source products, contributing to their development, either participating in existing open-source projects or opening their own production. They participate in existing projects, especially in Internet software (Apache, Sendmail) and operating systems like Linux first for opportunistic reasons. First, some of these programs are dominant in the Internet market. These firms need to contribute to them, to make sure that the programs they produce are well taken into account by the standard, if they want to gain access to the market.31 Second, in some

New approaches to intellectual property

253

markets, challengers see opportunities to sponsor a standard in competition with the dominating standard. This is the case, for instance, in operating systems: some competitors (IBM, HP, Novell now), of SUN in the Unix market, or of Microsoft in the operating systems for the servers’ market, have strong incentives to support Linux. This creates a competitor to these firms at a very low cost and, being GPL-protected, this program cannot be appropriated by a single firm, which would mean coming back to the situation they fight against using Linux. The GPL-protected software appears to be a way to solve the ‘wedding game’ situation when standardization effects are important: actors want to impose their view. Hence a firm which is not a leader may prefer to favour an open, non ‘privatizable’ solution. Such an ‘open’ organization allows the creation of ‘public industrial goods’ (Romer, 1993). Traditional software producers have also understood the interest of this open-sourcebased collective development for the long-run sustainability of the production model. Opening the sources has certain advantages: it allows better feedback from users, as they can propose and develop new functions by themselves, and better bug-tracking, as they can report some bugs. In other words, it is a powerful means to externalize a part of the development costs.32 It is also a good signal to show that the component produced respects norms and standards, especially for the interfaces, as the program code is public. These software producers financed their development costs by selling the product. With GPL-like licences, this is difficult, but not impossible. Some have successfully made the business evolve to service (assistance for the use of the program, maintenance contracts and specific development to adapt the product to clients’ needs). Some examples of such adaptations are Zope corp. or Adacore companies.33 Others are trying to keep the advantage of both worlds with specific licences: the externalization of R&D, as people can have access to the source code, propose some modifications, and the financing system, as the original owner is the only one authorized to integrate and redistribute these modifications and/or to sell the product. For commercial use, some licences hold that a fee must be paid to the original owner (SUN or Microsoft licences are examples of that system). Others have double licensing systems, one allowing free use and access to the source code for non-commercial purposes, the second requiring a fee and restricting modification in commercial situations (this was the MySQL strategy, for instance). Thanks to this innovative way of considering the licensing tool, companies have today a portfolio of strategies to valorize their intellectual property (see Muselli, 2002, for an analysis of the scope of these strategies), even if some doubt may be cast on the efficiency of these open but not completely free licences, since users may not trust that the producer is really willing to keep the sources open and to cooperate.34 The pooling of open bricks should also change the competition on service towards long-term relationships and maintenance services for software programs, instead of the basic installation of private programs. We defend the idea (Jullien, 2003, p.83) that this can encourage firms to improve customer services, which is one of the weaknesses of the computer industry.35 A new way of managing intellectual property and production in knowledge-intensive activities Open source principles cannot be considered as a rejection of intellectual property. In contrast to the paradoxical application of copyright law to software, open source gives a

254

International handbook on industrial policy

better place to authorship. Protection of software by copyright creates a situation where private property and secrecy are reconciled in a total opposition to the foundations of authorship protection (Vivant, 1993). In contrast, open source includes the identification of individual contributors for any part of a given program or further improvements or modifications. First, this contributes to the building of each contributor’s individual reputation and therefore plays a role as individual incentives. Second, the signature of the individual origin of each part of the program is a guarantee of the seriousness of individual contributions. In this way individual identity is not dissolved in a collective purpose and the literacy dimension of the code remains accessible. By connecting a large number of individuals around a common project and without any closure, open software is a way to take advantage of a fantastic potential of distributed skills. Developers not only expect the benefits they will draw from the availability of the public good that constitutes the open source program, but also will benefit from individual learning effects (Foray and Zimmermann, 2001). Learning through contributing to a free software project in the open context of a free software project can be considered as more efficient than learning within the boundaries of an enterprise. This is due to the multiple interactions a developer establishes with a wide variety of programmers using a wide range of methods and programming styles. This ‘learning by interacting’ represents another complementary dimension of individual incentives. Such variety is also a source of quality for programming goals. However, for this purpose, the functioning of an open source community has to satisfy two different levels of constraints that could ensure at the same time decentralization and coherence of the project advancement. First of all, as any complex software project in the classical context of production, the program architecture has to be conceived in terms of modularity and structural sequences. This allows programmers to work on a module development without necessarily being aware of the detailed functioning aspects of the other components that can be considered as black boxes. The development can then be shared in a distributed division of labour and successive steps of the code can be produced either sequentially or simultaneously. This constraint is strong enough to have induced the redesign (and redevelopment) of famous programs such as Apache. The open source development mode generally does not involve a fixed and pre-established division of labour among the developers. Demands for contributions to components’ production or improvement are usually sent out towards a broad and open population of developers in order to raise the probability of getting a good and suitable solution thanks, in particular, to the access to a large variety of skills and approaches. In addition, improvements or modifications can be spontaneously proposed and developers often already have on their own shelves the answer to the question (Von Hippel, 2001). It is thus the responsibility of a core group of developers to establish the relevant orientations of the development, select the contributions, articulate them in the global architecture of the program and supervise a large part of the developments involved in the program structure.36 This is the condition of coherence and quality of the collective effort and it is an essential task of coordination of the community production. For coordination to be effective the core group must be devoid of private economic interests in the project. That is the reason why individuals of the core group are chosen by the sole consideration of their technical competences and are co-opted by the developers’ group alike in the scientific community. In this way the developers’ community

New approaches to intellectual property

255

preserve all the characteristics and merits of an epistemic community: cognitive goal, commonly accepted structural authority, synergy of individual variety, individuals’ knowledge accumulation based on their own experiences and recruitment with regard to the agent’s contribution to the common goal (Amin and Cohendet, 2004). The working mode of open source production implied by the principles of open intellectual property management endows the community with the same coordination efficiency, in terms of collective production consistency, as a formal organization but with the advantage of preserving its flexibility and the wide extent of its resources. It represents a very motivating mode of decentralized coordination very similar to firms’ cooperation networks, but more flexible and never entangled by intellectual property conflicts. Demazière et al. (2004) use the expression of ‘distant community’ referring to the fact that the individuals involved are not only dispersed in geographical or organizational terms but also heterogeneous regarding their individual profiles. First, developers can occupy various professional positions: students, researchers at universities or other research centres, employees of private enterprises linked to open source software, personnel of other enterprises and so on. The balance between open source development activity and remunerated work differs according to the category, inducing a plurality of the juridical and temporal conditions of the contributions. This statutory variety has to be added to the previously mentioned variety of skills, programming styles and methods, but also to the cultural context and ethical dimensions of motivations to contribute to the open source development. However, in spite of this diversity, all the contributors ‘share the feeling of belonging to a specific community motivated by a strong common identity’ (ibid.), that counterbalances the weakness of their direct interactions and their dispersion. The cement of this cohesiveness is clearly the status of intellectual property that is a guarantee against deviations or misappropriation of the collective effort. However, the variety of the open source community members has also to be seen in terms of level of competences regarding the software development activity. Developers have different degrees of involvement, varying from steady contributions in the core group to more occasional ones at the periphery. An open source community, however, does not limit itself to the sole developers and also includes a category of ‘frontier users’ (Kogut and Metiu, 2001, p.23) that are not able to contribute to the software development but have an essential role in terms of source of innovations (Von Hippel, 1988) and above all testing and debugging the base. ‘Hence, it is not modularity that gives open source a distinctive source of advantage, because it too relies on hierarchical development. Rather the source of its advantage lies in concurrence of development and debugging. In spite of its unglamorous nature, maintenance alone represents anywhere 50–80% of the software budget. The largest part of the developer community are not involved with code writing, but with code-debugging’ (Kogut and Metiu, 2001, p.29). The status of intellectual property, as it is conceived in open source software, therefore clearly allows the development and working of a remarkably efficient production process in which all members benefit from the whole community’s efforts, while continuing to manage their proper activities in a decentralized manner. Beyond the sole case of software development, the way of working of such ‘distant community’ could have many applications in other knowledge-intensive fields of activities. Diverse cases have emerged recently that could foreshadow a wider range of applications.

256

International handbook on industrial policy

A first illustration is given by the ‘Wikipedia’ encyclopaedia project.37 The project was born in 2001 in the US Internet company Bomis. It now has more than a million entries, in a very large range of languages: mainly English, but also French (62 000 entries), German (100 000) and Japanese. It is one of the most visited websites in the world. It is supported by gifts and voluntary contributions and a strict post control permits avoidance of content diversion. It is the responsibility of a core group of ‘administrators’ endowed with the trust of the whole set of contributors. The Telabotanica38 project on botany is another example of collective production and sharing of knowledge. Another significant illustration of this need to get out of a strict intellectual property rights protection is given by the recent iCommons initiative. It is an international movement born in the United States and is at the origin of the Creative Commons tool ‘to give authors free tools to enable them to mark their content with the freedom they intend their work to carry, while reserving the rights the author believes must be reserved.( . . . ) It has attracted musicians, academics, authors, film-makers and researchers internationally who want a simpler way to exercise their rights without rejecting the protection of copyright altogether’ (Lessig, 2003, p.7).39 Creative Commons Licences began to be translated and adapted to a wide range of national juridical contexts, including the Netherlands, Taiwan, Australia, Sweden and France. They are now concerned with more than two million creative pieces of works in the world. Levine (2004, p.3) emphasizes the spreading of collectives that ‘operate in concert to accomplish innovation goals that may have great economic significance, and accomplishment of those goals is often their primary raison d’être. The significance is clear: real products are created, services are offered, and economic rents are appropriated or lost’. Such a collective, that Von Hippel and Von Krogh (2003) call the ‘private collective’ innovation model, can be mainly observed today as carrying on their interactions in an Internet framework. However, as outlined by Allen (1983), this type of ‘collective invention’ does not require modern communication tools and can be observed in a more traditional context. In the British iron and steel industry (Cleveland district) in the second half of the 19th century, companies freely revealed to competitors technical information related to blast furnace design in order to allow an experimental incremental advance to improve the efficiency of industrial smokestacks. Osterloh and Rota (2004) also point out two other examples of ‘private collective’ innovation models. With the Homebrew Computer Club formed at Stanford University in 1975, members exchanged ideas and projects in order to explore the potential applications of the already emerging microprocessor technology. Steve Wozniak, Steve Jobs and the other founders of Apple met in the club. Another example they give is of the early developments of flat panel display technology in the late 1960s. Following Spencer (2003), a majority of the firms active in this field have published at least one scientific paper, subsequently sharing a part of their R&D results. These firms are precisely those who produced the highest innovative performance (in terms of the value of patent portfolio). Today, the biggest challenge for intellectual property rights mutualization undoubtedly lies in the field of knowledge-intensive activities where a high level of complexity of the knowledge base has to be mastered and combinatorial aims of information and skills have to be developed. A significant example is the life science industry. Joly and Hervieu (2003) plead for a high degree of knowledge resources mutualization in the field of genomics in Europe, not in order renounce intellectual property but to organize a collective system of

New approaches to intellectual property

257

management of intellectual property. This is expected to reinforce the competitive position of European firms facing US multinationals and, by pooling basic technologies, to avoid innovation clamping. 5 Institutional and public policy consequences The implications of these issues in institutional and public policy have to be understood at different levels. The main level is undoubtedly the aspects of intellectual property status and protection. As noticed by the CONTU (National Commission on New Technological Uses of Copyrighted Works) in the 1970s, it seems difficult to set up a new, well adapted instrument of intellectual property protection. Even if other close domains like databases have been able to give rise to a sui generis framework, the main obstacle seems to lie in the necessity to reach an international consensus within a quite short period of time. The current debate rather appears to rest on the question of the patentability and the deeper institutionalization of the ‘CopyLeft’, GPL and other open-source licences. Intellectual property protection cannot be considered as an aim in itself and has to be treated in general interest considerations (CGP-Piéta, 2004). Currently, the general trend around the world is to strengthen the protection in terms of duration, as of scope. This does not work without any risk. History shows an alternation of strengthening and weakening phases of intellectual property juridical regimes. The present reinforcement, especially visible in the accelerated patent granting in the United States, reflects a relative failure of the patentability criterion that ‘tends to distort competition and threaten to inhibit innovation. Should the occasion arise, this proliferation problem is particularly sharp for enterprises that less often play the role of depositor than the one of user of technologies designed by others. The induced cost has to be understood in terms of a substantial loss of time and juridical expenses’ (ibid.). Concerning software patentability, it seems important to take a clear position at national and European levels in order to clear up the equivocal situation that stems from strong pressures of interest groups that aim to preserve their market-dominating power. Three main aspects of the question have to be taken into account. The first aspect is the recurrent question of the impact of patents’ spectacular increase on innovation dynamics. The aim of intellectual property protection is to foster innovation dynamics both by giving incentives to inventors and by improving the diffusion of ideas. For this purpose, it is important to conceive and mobilize juridical tools that should not become innovation impediments. The second aspect is also related to the principles of protection. It is important to note that software patentability, as it is applied, appears inconsistent with those foundations since the source code remains hidden when the inventor is not obliged to disclose the description of the technical aspects of the protected invention. Lastly, the third aspect is related to what should and what should not be protected. A large part of software components that are daily used by programmers should be considered as non-appropriable in consideration of their character of pure knowledge40 or of common resource. For this purpose a real state of the art would be necessary to determine the character of novelty of a patent claim. It is clear that these questions are not dissimilar from those that can be raised in other domains, especially that of life sciences and technologies, where the question of private knowledge appropriation raises problems of an ethical and cultural nature and more simply of social efficiency. An unbridled patent granting on natural resources gives rise to

258

International handbook on industrial policy

the pirating of less developed countries’ biological resources by large multinational firms.41 The Netherlands, Italy and Norway claimed in 1998, at the European Court of Justice, against the European 98/4/CE Directive on biotechnology patents, grounding their complaint on the argument that this Directive would represent an infringement of the 1973 European Convention on Patents and the 1992 Convention on Biodiversity. This recourse was rejected in June 2001 but the transposition of the Directive in the French jurisdiction will introduce important restrictions by placing genomic sequences in the public domain and by insisting on the strict necessity of the industrial applicability criterion (Desbois, 2004). More recently, the European Office of Patents invalidated, in January 2005, the patent of the American firm Myriad Genetics on the BRCA1 gene of predisposition to breast cancer. With such a patent Myriad Genetics would have benefited from a monopoly position on detection tests at a price three times higher than the current price on the European market. In addition, it would have given Myriad Genetics powerful control over the related work driven by public laboratories in cancer research that would have had to use licences and pay royalties for that. In the field of genomics, the Human Genome Programme, a public international consortium launched in 1998, has sustained an attitude based on the principles of publication and open diffusion of the results for the benefit of the whole scientific community. In France, the researchers demanding support for the ‘Genethon’ committed themselves to immediately publishing their results rather than taking out any patent. However, this line has progressively been contested by the involvement of large private firms supported by huge financial resources (Orsi and Moatti, 2000). This can ‘lead to a de facto vertical integration of academic laboratories into the activities of these firms’ (ibid., p.98), as in the case of the significant agreement signed between one of the most prestigious laboratories of the HGP and the US firm Perkin Elmer for the creation of the joint venture Celera, chaired by Craig Venter. An asymmetrical competition subsequently started between the public consortium HGP and Celera genomics, the later largely using the data produced and published by HGP. In March 2000, US President Clinton and UK Prime Minister Blair jointly declared their position in favour of free access to the whole raw data concerning the human genome. They had in mind to put pressure on Celera without excluding intellectual property protection for the future inventions based on genome knowledge. In February 2001, HGP and Celera simultaneously announced their release of all the human genome finding and published them respectively in ‘Nature’ and ‘Science’. From there, Celera has oriented his business activity towards the use of its knowledge and skills in genomics for the development of new therapies and specific medicine and has established, with that aim in mind, a set of significant strategic alliances.42 The other side of the institutional status of intellectual property is the question of the recognition and acceptability of the CopyLeft principles in the national and European juridical contexts. This includes the treatment of claims for infringement in the cases of abusive appropriation of open-source codes. As an illustration, the French INRIA43 with the CEA and the CNRS44 have settled a new open-source licence called CeCILL – Ce(a)C(nrs)I(nria)L(ogiciel)L(ibre) – in order to offer a GPL-equivalent licence that could underlie contracts consistent with the French law. This initiative carried out by public bodies has also a policy significance related to the feeling of interest in open-source software from those public bodies. Their commitment ‘can reassure some SMEs that would like to adopt those free software products but fear that such a choice could have

New approaches to intellectual property

259

pernicious effects on their own organisation’.45 While the official translation of the GPL is not yet achieved, the CeCILL licence is available in French and in English and this conforms to the so-called ‘Loi Toubon’ of 4 August 1994 that stipulates that contracts implying public bodies have to be written in French. Moreover, CeCILL specifies that, for lack of conciliatory agreement, the potential lawsuits will be treated by Paris courts. This represents an important advantage for French developers and enterprises which do not have to take proceedings into a foreign court. At the European level, a recent report to the European Commission stresses that GPL and other OSS licences do not meet the requirements of the European Union legal framework and therefore there is a need to set up a well-suited OSS licence. ‘The GPL’s major problem is that the right of communication to the public is not provided explicitly amongst the granted rights, and that a clause limits furthermore the granted rights to what is explicitly provided by the licence. Moreover, the GPL is known for being the most viral licence ever, whereas massive spreading through dynamic linkage is not the aim of the European Commission’ (IDA/GPOSS, 2004, p.3). This dimension in favour of a real recognition of the CopyLeft could find expression in the official prerogatives of national, European and even international offices of intellectual property (like the French INPI – National Institute for Intellectual Property – or the international WIPO – World Intellectual Property Organization). They could assume competences in terms of information as well as juridical advice and assistance, as they do for standard intellectual property tools. The second dimension of public policy concern is related to the adoption and use of open-source software in public administrations and services. In France, the prime minister has declared his intention to favour the choice of open-source software in the national administration.46 In the regions, some local governments, such as the Regional Council PACA – Provence Alpes Côte d’Azur – have announced their decision to implement progressively free software in the high school establishments of the region. Such a decision is not only a strong political signal in favour of free software but also a means to increase the new generation’s awareness of the principles of knowledge sharing and cooperative development. In the UK, the Office of Government Commerce announced in 2003 its intention to launch and coordinate large ‘Proof of Concept’ trials of Open Source Software in a range of public bodies in conjunction with IBM and Sun Microsystems. The final report pointed out the direct and indirect aspects of OSS advantages: ‘Apart from reductions in cost of software licences, benefits of Open Source can include cost avoidance through reductions in the replacement cycles of hardware, improved software reliability and security, software platform stability, the ability to tailor and modify the software, easier administration, and greater scalability of hardware platforms’ (OGC, 2004, p.1). Beyond the aspects of costs for public administrations and bodies, the possible turning towards open-source software adoption is thus also the mark of an involvement of the public sector in the process of open-source development, as a very large base for software products debugging and improvements, but also as a potential direct contribution to software development under GPL or equivalent licences. As for IBM investment in Linux, these decisions boost the credibility of such offers and thus favour their diffusion. Finally, the third dimension of public policy is related to the actual support given to open-source software production. On the one hand, this support can take the form of direct finance subsidies to free-software organizations or of putting developers paid by

260

International handbook on industrial policy

public funds at the disposal of open source projects. On the other hand, there are a lot of computer codes that are developed in the context of the public administration or sector and that could be publicized under GPL or equivalent status without any problems of security or confidentiality. This would represent a contribution to the production of the open-source community and, at the same time, avoid problems of duplication of parallel projects in the public sector, while benefiting from the possible improvements and testing of the open-source base. This idea is defended by the Adullact association which tries to organize the collective production and exchange of the solutions produced between public communities in France.47 However, Europe still offers a large variety of attitudes towards OSS use and support. A 2001 study on this topic has classified them, in the countries questioned, into three main groups: – France and Germany are the two innovative leaders in the field (Germany for concrete realisations and guidelines, France for a growing government support to open standards and Open Source); – Spain is an active follower concerning specific departments where competent Open Source advocates have demonstrated the efficiency, the best value for money and the supportability of the solution, and where the installation of a standard specific Linux distribution can provide a scaling effect; – In Belgium, Italy and Sweden, the existing realisations are resulting from individual efforts that are not – until now – actively supported by a government policy. (IDA, 2001, p.7)

6 Conclusions The software industry represents today a remarkable illustration of the fact that any intellectual property system is a compromise between different actors (the producer(s) of a good, their competitor(s), the users, the state, and so on). A good balance between individual incentives to innovate and the maximization of the social utility generated by the diffusion and use of the innovation is very conditional on the technological environment and evolves over time. The traditional intellectual protection framework allows producers to finance the initial phase of the innovation process, anticipating returns in the diffusion phase, in which they expect a legal monopoly to produce and commercialize it. The Open Source movement shows that, for some knowledge-intensive/cumulative innovation, this can be counterproductive, since the diffusion phase allows the producer to finance its initial investment without having a monopoly but using the feedbacks and the joint needs generated by new users and uses. The Open Source movement was made possible at a global level by the existence of the Internet and has been a predecessor to historical occurrences on more local scales. But with the diffusion of the Internet and the fact that research is more and more computer-based, the sharing of knowledge and innovation can be much more efficient than even ten years ago. This could imply the need to stop the present trend of strengthening intellectual property protection regimes. And what happens now in the software industry is going to spread to a large range of other activities and industries. Thus the question of the possible extension of such a model of production, based on knowledge openness and sharing, represents a significant challenge for knowledge-based industries in the near future. We have mentioned in section 4 of this chapter the fields of knowledge-intensive

New approaches to intellectual property

261

activities where the construction of a shared knowledge base is still a matter of fact, drawing advantage from the large diversity of skills and resources mobilized by the contributors for the benefit of all of them and even beyond. Some examples such as the Wikipedia (open encyclopedia http://www.wikipedia.org/) and the Tela-Botanica (http://www.tela-botanica.org/) projects have shown that a collective production of knowledge helped by the Internet is feasible in other contexts than software production. As we saw, this generates problems of free-riding while opening up the opportunity of a higher individual and social efficiency. Hence it seems possible to lay down two main rules to which a model based on knowledge mutualism may conform, in order to aspire to viability. First, business activity should not be based directly on marketing products issued from this knowledge collective production, but rather on using them as a basis for marketable activities: a cognitive input of a productive process, joint services and so on. Second, taking part in this public good production should give rise to direct rewards for contributors (such as learning or a better match to personal needs in the case of OSS) and/or a competitive advantage on related markets. When such conditions can be fulfilled, public action can be useful to favour the success of the model and to select the most appropriate ways according to the prevalent conditions of production and competition in the industry concerned: the direct support of knowledge production (through financial aids and incentives, public contributions and so on); the preference given to open-source products on public markets; the assistance to coordination, training programs, public information and stable juridical and knowledge protection environment. The open-source model could thus apply to a broader range of activities beyond the software sector alone, taking advantage of the huge potential of communication and cooperation drawn from the Internet and the digitization of knowledge production. The key issue for this enlargement actually lies in changing the place of value added production and competition, as in the case of OSS, where commercial profits are mainly generated by joint products and services. This would give rise to a renewed combination of actors’ cooperation and competition (‘coopetition’) and a much higher social efficiency than in the monopoly model based on strong intellectual property protection. Notes 1. 2. 3. 4. 5. 6.

7.

The ‘architecture’ of a computer or a data processing system refers to the way the diverse components of a computing system are organized and assembled from a physical and logical point of view. ISO: International Standards Organisation; IEEE: Institute of Electrical and Electronic Engineers. See ‘Brevets logiciels et Linux: des chiffres qui inquiètent’ (http://www.journalinformatique.com/0408/ 040803_linux.shtml). This is the case of the US company Acacia, formerly a start-up incubator whose sole activity is now to sell licences under the threat of lawsuits. See A.Chassignin, ‘Ces sociétés qui tirent profit des brevets logiciels’ (http://solutions.journaldunet.com/0409/040906_brevets.shtml). See for example Estelle Dumout, ‘Le camp des logiciels libres dénonce SCO dans sa guerre contre Linux’ (http://www.zdnet.fr/actualites/informatique/0,39040745,2135115,00.htm). For ‘Berkeley Software Distribution’ licence. This licence allows anybody to reuse, modify and redistribute the program under its own preferred terms, the argument being that this production has been funded by the public and should be available to the public. SUN (Stanford University Network) used the Berkeley Unix and Stanford network system to construct its first commercial offers. He has created, for example, the first Free software, Emacs editor (1984), that is still used today and is still considered the best design tools program; he also created various compilers, such as GCC C and C++ (1985) that are still the most widely used today.

262 8. 9. 10. 11. 12. 13. 14. 15.

16. 17. 18. 19. 20. 21. 22. 23.

24. 25.

26. 27.

28. 29. 30. 31.

32. 33.

34. 35.

International handbook on industrial policy Taken from the Free Software Organisation (at http://www.gnu.org/philosophy/free-sw.html). General public licence, which protects programs like GNU/ Linux. (See: http://www.gnu.org/ licenses/licenses.html#TOCGPL.) See Stallman (1998). http://www.gnu.org/gnu/thegnuproject.html. We have developed this argument in Jullien (1999). http://www.catb.org/~esr/writings/cathedral-bazaar/hacker-history/;. Hacker being understood in its original acceptation, that is, highly skilled developer. See, for instance, Mocus et al. (2000) for a presentation of the Apache development organization. This can be a problem appearing with the success of a program. If there were no such mechanism, the most productive developers would progressively dedicate more and more time to addressing basic problems and thus losing interest in the development until quitting. See Foray and Zimmermann (2001) for a discussion of this point. For a description of the organization, see Demazière et al. (2004). According to the company, this investment has generated equivalent profit in just a year (see http://news.com.com/2100-1001-825723.html). http://www.theregister.co.uk/2005/01/19/sun_cddl_solaris/. http://solutions.journaldunet.com/0404/040407_microsoft.shtml. http://www.mandrakesoft.com/toprint/KBC-Analyse-MDK-280604-en.pdf. http://news.zdnet.com/2100-3513_22-5227983.html?tag=zdnn.alert. See http://www.linex.org/ and http://www.linex.org/linex2/linex/ingles/index_ing.html for the reasons for such a political choice. Which means that each adoption reinforces the incentive for an adopter to adopt. See, for instance, Netcraft Web server survey (http://news.netcraft.com/archives/web_server_survey.html), or the fact that today Linux appears as the competitor of the Microsoft’s Windows family, not only for analysts, but also for Microsoft. For the first time in its history, this company is doing comparative advertising between its products and Linux (http://www.microsoft.com/mscorp/facts/default.asp). The last success story was the launch of firefox, which, in a few weeks, has perceptibly caused Internet Explorer’s dominant position in the browser market to crumble. Raymond (1998), Lakhani and von Hippel (2003) and Jullien (2001) provide a more extensive analysis of the way open development is structured. See Tzu-Ying and Jen-Fang (2004) for a survey and an analysis of the on-line user community involvement efficiency; Bessen (2002) and Baldwin and Clark (2003) for a theoretical analysis of the impact of opensource software code architecture on the efficiency of open-source development. The latter argue that it may be seen as a new development ‘institution’ (pp.35 ff). As for performance tests, one can refer to http:// gnet.dhs.org/stories/bloor.php 3 for operating systems. The results for numerous comparative evaluations are available on the following sites: http://www.spec.org and http://www.kegel.com/nt-linux-benchmarks. html (the latter mainly deals with NT/Linux). See the 13 December CNet article by Robert Lemos (http://news.zdnet.com/2100-1009_22-5489804.html). Let us remember that the first GPL programs that were created and adopted were development tools (in particular, compilers, interpreters and programming assistance tools). In fact, software programs allowed the normalization of computer languages, which are the minimal common base necessary to communicate.See http://www.gnu.org/prep/standards.html for technical recommendations on how to program GNU software. In the sense that they respect public formats whose evolution is decided collectively. It is the Free Standard Group (http://www.freestandards.org/). Among others, members of this committee include Red Hat, Mandrake, SuSE, VA Software, Turbo Linux, but also IBM, SUN or Dell. ‘Average Revenue Per User’. This term is mainly used in telecommunications and makes it possible to evaluate the profitability of a firm on the basis of the average income generated by a user. The best example of this type of behaviour is undoubtedly the work that SUN has done to develop the Apache free software program. This firm deals with everything that relates to the adaptation of Java to Apache. By making the language it has developed compatible with the market standard, it can hope to sell the Java development tools which it produces (as well as its expertise in this field). See the Microsoft ‘shared source program’ (http://www.microsoft.com/resources/sharedsource/). Zope corp. edits Zope Content management software (http://www.zope.org). Adacore edits the Ada 95 compiler named GNAT (http://www.adacore.com). Ada 95 is a programming language designed for large, long-lived applications – and embedded systems in particular – where reliability and efficiency are essential. Soufron and Sallantin (2005) analyse the increasing variety of free/open source licences as different combinations of four types of requirements: (1) the right to access (to the source code), (2) the right to modify, (3) the right to redistribute, and (4) the right to use. See, on that weakness, the analysis by de Bandt (1995), or Dréan (1996) (pp.276 ff).

New approaches to intellectual property 36. 37. 38. 39. 40. 41. 42. 43. 44. 45. 46. 47.

263

Following Mocus et al. (2000) about the Apache project, ‘despite broad overall participation in the project, almost all new functionality is implemented and maintained by the core group’ (p.268). From ‘Wiki’, meaning fast in Hawaiian, and ‘pedia’ from the Greek ‘paideia’ meaning education. http://www.tela-botanica.org/. Lawrence Lessig is a Professor of Law at Stanford Law School. He also chairs the Creative Commons Project. The notion of ‘pure knowledge’ has to be understood here as opposed to a patentable knowledge that requires an industrial applicability. The Rio de Janeiro Convention on biodiversity in 1992 has pointed out the necessity to preserve the sovereignty of nations over their own biological resources. http://www.celera.com/celera/alliances. National Institute for Research in Informatics and Automatics. The CEA is the French acronym for Atomic Energy Department and the CNRS of the French National Centre for Scientific Research. Gérard Giraudon, chairman for industrial developments and relations at INRIA, in Y. Rocq (2004) ‘Faut-il adopter la licence CeCILL’, Login, no.120, Sept. See, on that topic, the directive published by the French Ministry for infrastructure, transport, spatial planning, tourism and the sea (http://www.equipement.gouv.fr/bulletinofficiel/fiches/Bo200410/A010 0067.htm). http://www.adullact.org/.

References Allen, R. (1983), ‘Collective invention’, Journal of Economic Behavior and Organisation, 4, 1–24. Amin, A. and P. Cohendet (2004), Architectures of Knowledge: Firms, Capabilities and Communities, Oxford: Oxford University Press. Arthur, W.B. (1988), ‘Self-reinforcing mechanisms in economics’, in P.W. Anderson, K.J. Arrow and D. Pines (eds), The Economy as an Evolving Complex System, SFI Studies in the Sciences of Complexity, Redwood City, CA: Addison-Wesley Publishing Company. Arthur, W.B. (1989), ‘Competing technologies, increasing returns and lock-in by historical events: The dynamics of allocations under increasing returns to scale’, Economic Journal, 99, 116–31. Baldwin, C.Y. and K.B. Clark (2003), ‘The architecture of cooperation: how code architecture mitigates free riding in the open source development model’, Harvard Business School, 43. Bessen, J. (2002), ‘Open source software: free provision of complex public goods. Research on innovation’, URL: (http://www.researchoninnovation.org/online.htm#oss). Bessen, J. and K. Maskin (2000), ‘Sequential innovations, patents and imitation’, MIT, Department of Economics, Working papers nos 00–01, January. CGP-Piéta (2004), La propriété intellectuelle dans le contexte de la société du savoir, Les Quatre Pages, no.1, 16 March, Le Plan, Paris: Commissariat Général du Plan. De Bandt, J. (1995), Services aux entreprises: informations, produits, richesses, Paris: Economica. Demazière, D., F. Horn and N. Jullien (2004), ‘Le travail des développeurs de logiciels libres. La mobilisation dans des “communautés distantes” ’, Proximity, Networks and Co-ordination Conference, Marseille, 17–18 June (to be published in Cahiers Lillois d’Économie et de Sociologie). Desbois, D. (2004), ‘Vers une appropriation privative du vivant’, Cadres CFDT, no. 409. Dréan, G. (1996), L’industrie informatique, structure, économie, perspectives, Paris: Masson. Farrell, J. (1989), ‘Standardisation and intellectual property’, Jurimetrics Journal. Farrell, J. and G. Saloner (1988), ‘Coordination through committees and markets’, Rand Journal of Economics, 19(2), Summer, 235. Foray, D. and J.-B. Zimmermann (2001), ‘L’économie du logiciel libre: organisation coopérative et incitation à l’innovation’, Revue économique, 51, October, 77–93. IDA (2001), ‘Study into the use of open source software in the public sector, part 2, use of open source in Europe’ (http://europa.eu.int/idabc/servlets/Doc?id=1973). IDA/GPOSS (2004), ‘Advice report – EUROPEAN COMMISSION / Enterprise Directorate General – IDA/GPOSS – encouraging good practice in the use of open source software in public administrations – report on “Open Source Licensing of software developed by the European Commission (applied to the CIRCA solution)”’ (http://europa.eu.int/idabc/servlets/Doc?id=21197). Joly, P.B. and B. Hervieu (2003), ‘La marchandisation du vivant’, Futuribles, 292, December. Jullien, N. (1999), ‘Linux: la convergence du monde Unix et du monde PC’, Terminal, 80/81, 43–70. Jullien, N. (2001), Impact du logiciel libre sur l’industrie informatique, PhD, Université de Bretagne Occidentale / ENST Bretagne (http://www-eco.enst-bretagne.fr/Etudes_projets/RNTL/documents_universitaires.html). Jullien, N. (2003), ‘Le marché francophone du logiciel libre’, Systèmes d’information et management, 8 (1), 77–100.

264

International handbook on industrial policy

Kogut, B. and A. Metiu (2001), ‘Open source software development and distributed innovation’, Reginald H.Jones Centre working paper nos 01–08, April. Lakhani, K. and E. von Hippel (2003), ‘How open source software works: free user to user assistance’, Research Policy, 32, 923–43. Lerner, J. and J. Tirole (2002), ‘Some simple economics of open source’, Journal of Industrial Economics, 52, 197–234. Lessig, L. (2003), ‘Foreword’, in D. Bourcier and M. Dulong de Rosnay (eds), International Commons at the Digital Age, Paris: Romillat, Coll. ‘Droit et Technologies’. Levine, S. (2004), ‘Towards a grounded theory of collective open source innovation’, Druid Summer Conference on Industrial Dynamics, Innovation and Development, Elsinore, Denmark, 14–16 June. Mocus, A., R. Fielding and J. Herbsleb (2000), ‘A case study of open source software development: the Apache server’, proceedings of International Conference on Software Engineering (ICSE2000), Limerick, Ireland, 4–11 June, 263–72. Muselli, L. (2002), ‘Licences: strategic tools for software publishers?’, in N. Jullien, M. Clément-Fontaine and J.-M. Dalle (eds), New economic models, new software industry economy, Technical report, RNTL (French National Network for Software Technologies) project, pp.129–45. O.G.C. (2004), ‘Open source software trials in government’, Final Report, Office of Government Commerce, (http://www.ogc.gov.uk/embedded_object.asp?docid=1002367). Orsi, F. and J.P. Moatti (2000), ‘Les relations recherche publique / industrie génomique: américanisation ou voie européenne?’, Médecine/Sciences, 14, 94–100. Osterloh, M. and S. Rota (2004), ‘Open source software development – just another case of collective invention?’, SSRN working papers (http://ssrn.com/abstract=561744). O.T.A. (1992), ‘Finding a balance; computer software, intellectual property and the challenge for technological change’, OTA-TCT-527, Washington, DC. Ousterhout, J. (1999), ‘Free software needs profit’, Communications of the ACM, 42(4), 44–5. Raymond, E.S. (1998), ‘The cathedral and the bazaar’, URL (http://www.tuxedo.org/~esr/writings/cathedralbazaar/). Raymond, E.S. (1999), The Cathedral & the Bazaar; Musing on Linux and Open Source by Accidental Revolutionary, Sebastopol, California: O’Reilly. Romer, P. (1993), ‘The economics of new ideas and new goods’, Annual Conference on Development Economics, World Bank, Washington DC. Saloner, G. (1990), ‘Economic issues in computer interface standardisation’, Economics of Innovation and New Technologies, 1, 135–56. Soufron, J.B. and J. Sallantin (2005), ‘The evolution of free/libre and open source software licences : a dynamic model’, Communication to the European Conference on Complex Systems, Paris, 14–18 November. Spencer, J.W. (2003), ‘Firms’ knowledge-sharing strategies in the global innovation system: empirical evidence from the flat panel display industry’, Strategic Management Journal, 24 (3), 217–33. Stallman, R.M. (1998), ‘The GNU project’, Open Sources. Tzu-Ying, C. and L. Jen-Fang (2004), ‘A comparative study of online user communities involvement in product innovation and development’, National Cheng Chi University of Technology and Innovation Management, Taiwan (http://opensource.mit.edu/papers/chanlee.pdf). Vivant, M. (1993), ‘Une épreuve de vérité pour les droits de propriété intellectuelle: le développement de l’informatique’, in L’avenir de la propriété intellectuelle, Paris: Librairies Techniques, Collection ‘Le Droit des Affaires’. Von Hippel, E. (1988), The Sources of Innovation, Oxford: Oxford University Press. Von Hippel, E. (2001), ‘Learning from open-source software’, MIT Sloan Management Review, Summer. Von Hippel, E. and Von Krogh J. (2003), ‘Open source software and the ‘private–collective’ innovation model: issues for organisation science’, Organisation Science, 14(2). Zimmermann, J-B. (1999), ‘Logiciel et propriété intellectuelle: du copyright au copyleft’, Terminal, 80/81, 95–116.

13 Science-based industries and spin-offs Marco Giarratana and Salvatore Torrisi

1 Introduction Science-based and high-technology industries are a major source of economic growth and competitiveness. An important mechanism through which these industries generate benefits for the economy as a whole is new firm formation. The empirical evidence shows that high-tech industries have larger start-up rates compared with other industries (Shane, 2001). In particular, entry of new firms is more likely to occur in ‘entrepreneurial technological regimes’ characterized by high technological opportunities and low entry barriers. However, in industries characterized by high technological change and uncertainty only a few new firms survive and therefore the likelihood of growth and survival is limited. It is important then to understand what are the growth potentialities of different types of new science and technology-based firms (NSTF). The relevance of these issues for industrial policies is demonstrated by the marked differences between Europe and the USA in terms of science-based entrepreneurship and post-entry growth of new start-ups, including those in high-tech industries (Bartlesman et al., 2003). Scholars and policy makers agree that the European ‘innovation deficit’ is in part due to the limited new firm formation in science-based industries. Several recent empirical works focus especially on the links between university and industry. Compared with the US academic system, European universities have limited technology transfer activities and spawn a much smaller number of spin-offs (Debackere et al., 1999; Henrekson and Rosenberg, 2001; Di Gregorio and Shane, 2003). To track this issue, in this chapter we first propose some definitions of NSTFs and the more recent theories and stylized facts. We then focus on spin-offs, that are NSTFs spawned by established firms or academic institutions, and introduce a taxonomy of science and technology-based spin-offs (STSO). We examine the role of parent organizations, firms and academic institutions, and their strategies towards STSOs. We also explore the implications of spin-offs for the incentives of established organization to invest in human capital formation and the impact of IPR and the market for technology on STSOs. Finally, we present evidence on the performance of STSOs and discuss some policy implications. 2 Definitions and theory of technology spin-offs This chapter deals with STSOs that represent a particular type of NSTBFs. Many belong to the category of ‘gazelles’, that is, firms with a high growth rate, between 15 per cent and 20 per cent annual growth of sales or employees (Bhide, 2000). In the USA there are over 15 million entrepreneurs involved in new start-ups in different sectors. Of these about half a million can be defined as ‘technology entrepreneurs’. These are individuals who launch every year about 200 000 new science- and technology-based start-ups (Reynolds, 2000). Before analysing different categories of spin-offs, it is worthwhile to introduce a broad definition of STSOs: new firms that satisfy at least partially the following criteria: (a) 265

266

International handbook on industrial policy

existence of a parent organization (research institute/university or an enterprise where at least one firm founder has been employed as a scientist, administrative or technical staff or student; (b) the new start-up relies substantially on technical knowledge or know-how (not necessarily in the form of licensing or patenting); (c) the parent organization has made an equity investment in the firm or has directly established the new firm. Corporate spin-offs STSOs are new firms founded by individuals or teams who were formerly employed by an established organization. The founders rely on technology developed during their previous occupation and transferred from the parent organization to the new venture. The establishment of an STSO typically involves four agents: the originator of the core technology (an individual or an organization that develops the basic technology to the point where technology transfer can take place); the parent organization that hosts the technology originator; the entrepreneur (or the founding team) who draws on the technology provided by the originator and takes the risk to bring it to the market; and a venture investor. STSOs can be either planned and sponsored by the parent organization or spontaneous start-ups. In the latter case the parent organization does not devote any significant effort to favour technology transfer and, on some occasions, the new venture is discouraged by the parent organization (Roberts and Malone, 1996; Steffensen, Rogers and Speakman, 1999). Spontaneous STSOs may result from market opportunities that the incubator is not willing or able to perceive and pursue for several reasons. One reason is the aim to maintain strategic coherence (Burgelman, 1983) or the need for more cash to devote to the firm’s core business (Garvin, 1983). In the same stream of research, there are several models which analyse the incentives of established firms to deter or tolerate a given rate of STSOs. For instance, Klepper and Sleeper (2000) analyse corporate STSOs as the result of technological and market opportunities created by the innovative activities of established firms.1 When the expected share of employees with entrepreneurial abilities is small the firm decides to take the risk of an STSO. The likelihood of STSO is affected by the scale and scope of R&D activities, by the industry life cycle and the external business environment, for example, skill supply in the region where the firm is located. Similarly, Hellman (2002) has developed a model of spin-offs where the emergence of (sponsored) spin-offs results from the combination of the established firm innovation strategy (exploration of new opportunities versus exploitation of existing ones), the entrepreneurial environment (how easy it is for the new firm to find key resources such as venture capital), and the IPR regime which defines the allocation of IPRs between the employer and the employee. Finally, information asymmetry and agency costs may reduce the established firms’ ability to evaluate the economic value of important innovations that may be pursued by spin-offs. Agency problems are particularly important when an established firm has to evaluate the economic value of new ideas generated by employees which fall outside its core business or core competencies. Differences in the perception of the expected value of an innovation between an employee and her employer lead to frustration and may spur spin-offs. An example of divergences in the expected value of an innovation between the inventor and established firms is represented by Chester Carlson, who founded Xerox after Kodak’s refusal to pursue Carlson’s idea of a new photocopier (Audretsch, 1997). Available empirical evidence suggests that in some countries corporate spin-offs account for an important share of all STSOs. For example, a panel survey of young and

Science-based industries and spin-offs

267

small technology-based firms in Finland shows that established firms represent the most important source of spin-offs while university spin-offs are less frequent (Autio and YliRenko, 1998). Unfortunately, the lack of empirical evidence on corporate spin-offs makes it difficult to obtain accurate estimates of their importance as a source of STSOs. University spin-offs Universities and public research centres represent a significant source of STSO formation. For example, examining a sample of new technology-based firms established in France between 1984 and 1991, Mustar (1997) found that about 40 per cent were founded by scientists employed in public research centres, universities and the system of ‘les grandes écoles’. Similarly, in Ireland and Sweden, a large proportion of STSOs have been founded by former employees of higher institutions (see Storey and Tether, 1998 for a survey). The evidence from Finland suggests that universities or research centres may also offer substantial technological knowledge to non-academic STSOs (Autio and Yli-Renko, 1998). University start-ups typically are spurred by different motivations than corporate spinoffs. For instance, university professors who decide to establish a new firm may be motivated by social recognition, autonomy and control. Unlike many new start-ups spawned by established firms, academic spin-offs, especially in the USA, are often supported by university technology transfer offices (TTO) because they represent a vehicle for commercializing a university’s property rights and demonstrating the contribution of universities to economic growth (Chiesa and Piccaluga, 2000; Nicolaou and Birley, 2003; Feldman, 2003). There are three main factors that have made universities an important source of entrepreneurship especially in the USA. First, the increasing importance of knowledge as an input in several industries. Second, after World War II, the share of federal R&D funding directed to universities increased dramatically (Rosenberg, 2003). Finally, small firms and new start-ups have proved to be more able to exploit commercially the outcome of university research as compared with large firms (Jaffe, 1989). According to estimates based on the US Association of University Technology Managers (AUTM) about 12 per cent of university inventions are exploited by new startups (Di Gregorio and Shane, 2003). These are new firms that have signed a licensing contract with a university TTO. According to AUTM, American universities spawned about 275 spin-offs in 1999.2 This represents a significant increase on the period before 1993, when the average number of university spin-offs was 80. However, the distribution of spin-offs is very skewed, with the majority of universities having none and the most active one having spawned 19 in 1999 (Feldman, 2003). The rate of spin-offs generated by universities or other public research organizations varies across countries as well. For instance, the annual number of spin-offs was 37 in Germany (2001), 67 in Norway (2001) and 390 in the USA (2000): see OECD, 2003. Table 13.1 reports data on patents, licences and start-ups spawned by public research centres and universities in different countries.3 Clearly, there is no correlation between the number of patents and the number of spinoffs. Instead, the correlation between the number of spin-offs and active licences is very strong. This result is in line with recent empirical evidence on European research centres/universities which shows that only some of these centres spawn spin-offs based on proprietary technology (Clarysse et al., 2005).4

268

International handbook on industrial policy

Table 13.1

University and public research centres’ intellectual property and start-ups Stock of patents

Patents granted

Licences

Avg per PRO

(a)

(b)

(c)

(d)

Startups (d)

506 9391 5404 na 114 991 1184 781 682 na

57 832 747 30 28 167 53 64 163 na

46 247 555 36 22 368 475 125 89 4049

7.7 16.5 19.1 1.5 1.4 14.7 17.6 7.8 5.9 24.1

15 56 37 36 51 37 68 11 6 390

Belgium Korea Germany (2) Italy (1) Norway (2) NL (1) Switzerland Spain Japan (1) USA (1)

Notes: (a) stock of patents granted in the last year (b) patents granted in the last available year (c) number of licences negotiated in the last year (d) average number of licences negotiated in the last year per PRO (e) start-ups and spin-offs together (1) year 2000 (2) Public research centres only Source: OECD (2003).

Why do some universities generate more start-ups than others? A study of 101 US universities’ spin-off activity in the period 1994–8 by Di Gregorio and Shane (2003) shows that university policies towards technology transfer are a good predictor of university start-up rates. More precisely, they found that the percentage of royalties distributed to inventors is inversely related to start-up rates. A high share of royalties increases the opportunity costs of establishing a new firm for the inventor. Moreover, university spin-offs are positively related to the practice of taking an equity stake in the licensee firm by the university. This result is also expected, given the liquidity constraints of new firms. As expected, the quality and reputation of a university (measured by the academic rating score of US graduate schools) is also an important predictor of spin-offs even though there are eminent US academic institutions, like Columbia University, which generate only a few start-ups. Instead, the share of industry-supported research, a proxy for the importance of commercially oriented research, has no significant effect on start-ups. This result may be driven by two contrasting effects. On the one hand, by definition, commercially oriented research tends to be closer to commercial exploitation and this should reduce the development costs and the risk for new start-ups. On the other hand, industry-supported research may have more stringent non-disclosure constraints as compared with basic research. 3 A taxonomy of spin-offs Technology entrepreneurship requires a collective effort that is important to overcome the complexity of innovation. Several types of skills are required to produce successful

Science-based industries and spin-offs 269 innovative projects: technical, managerial and financial capabilities (Auerswald and Branscomb, 2003; Blanchflower and Oswald, 1998). In particular, new high-growth firms are started by founding teams endowed with significant human capital and ability to attract resources, from customers to finance (Bhide, 2000). In most cases technological entrepreneurship is a collective action carried out by inventors (that is, former university or corporate researchers), business visionaries (individuals who knows the business environment and champion the new idea) and financiers, who accept the high risk/uncertainty in exchange for high expected returns. NSTF often require management and financial support that can be provided by different institutions: venture capital funds, corporate venture capitalists and private investors (business angels). Besides financial support, these institutions provide assistance (business plans, product and marketing strategies) and contacts with a network of professionals, potential customers, suppliers and skilled labour. As noted by Auerswald and Branscomb (2003), ‘mentoring is as important as funding’ (p.89). These different types of parties have different incentives, capabilities and evaluation criteria. The high uncertainty that characterizes the early-stage technology development and information asymmetry among the parties involved in the collective entrepreneurship makes it difficult to write efficient contracts. Experience, trust and reputation then play an important role. According to recent estimates, the most important financial support to new technology start-ups in the USA comes from established firms that spin out new ventures based on inventions that fall outside their core business: between 31.6 per cent and 47.2 per cent. The second largest source is represented by angel investors (23.9 per cent to 27.9 per cent of total funds) followed by public R&D programmes (22.7 per cent to 29.8 per cent). Professional venture capitalists account for a small share of total sources (2.3 per cent to 8 per cent). Fewer than 500 out of about 200 000 NSTFs receive venture capital ‘seed stage’ funding (Auerswald and Branscomb, 2003).5 To summarize the literature on university and corporation STSOs, we submit a taxonomy based on two dimensions: first, the nature of resources that are provided by the parent organization (managerial versus financial); second, the type of parent organization (academic research institution or established firms). By definition, STSOs receive from the parent organization some degree of scientific or technical knowledge. Even though the level of scientific or technical knowledge may vary across different types of spin-offs, we believe that the main differences among STSOs are determined by the financial and managerial resources inherited. Finance is often referred to as an important constraint for technology entrepreneurs. The financial gap is in part determined by the lack of information about the founders’ capabilities and the quality of the business idea. The lack of initial tangible assets also makes it difficult to obtain capital from traditional financial intermediaries. Even venture capital funds, as mentioned before, accrue to a minority of start-ups and are mostly focused on later-stage firms. Recent empirical evidence shows that both the financial and the human capital of entrepreneurs are important determinants of the hazard of failure (Acs, Carlsson and Karlsson, 1999). To our purposes here, it is not important whether the access to financial capital is determined by the founders’ human capital. It is likely that the decision to provide support to the spin-off is positively correlated to the skills of the founders. But highly skilled employees who start a new firm often do not receive any financial support

270

International handbook on industrial policy

from the parent company, especially when the new firm competes with the latter. Therefore the provision of financial resources is not only determined by the founder’s human capital (and the perceived quality of the business idea) but it is also affected by the market position occupied by the spin-off relative to its parent organization. The second dimension of our taxonomy is represented by the origins of spin-offs: (a) universities or public research centres, and (b) established, usually large, firms. Table 13.2 reports four categories of spin-offs which result from the combinations of the two dimensions discussed so far. A vast body of the literature on spin-off has primarily focused on Type A spin-offs (for example, Louis et al., 1989; Bray and Lee, 2000; Thursby and Thursby, 2002) while the research on the other types is still underdeveloped (Klepper, 2001a; Chesbrough 2003). Type A is the classical example of NSTF spurred by inventions generated in academic and public research labs and commercialized by one of the inventors through a new firm. Financial resources from the parent organization typically take the form of equity stakes (Di Gregorio and Shane, 2003; Feldman, 2003). For this category of entrepreneurs commercial resources most probably are acquired through new entrepreneurs or top managers who join the start-up in its infancy. Type B occurs when the parent company has no incentives to pursue a specific market opportunity originated by in-house R&D activities. Firms with profitable existing lines of businesses usually let other organizations, typically spin-offs, to explore new ideas that are too risky or that could substitute for established products of the firm (Henderson and Clark, 1990). Sometimes, when the market opportunity seems really promising, established firms try to extract rents from new markets by sponsoring new start-ups. This is the case of large high-tech firms like IBM, which has spun off several firms in India (for example, Globarena specialized in IT learning) and Israel (for example, Riverhead software). Type C may fall in the category of involuntary spin-offs or in that of spin-offs that require limited seed financial support. Some involuntary spin-offs occur when the parent company fails to recognize a new business opportunity because of organizational inertia, firm financial inefficiencies and difficulties in managing large and diversified patent portfolios. Christensen (1997) provides many examples of entrepreneurs who have worked for large private research labs that have missed important new market opportunities. Other spin-offs are not sponsored by the parent organization for reasons unrelated to inertia or lack of recognition. This is the case of several new technology-based firms founded in Europe. During the 1990s, several large firms in Europe restructured and downsized their activities. As a consequence, less strategic or marginal operations have been outsourced. This strategy has spurred a wave of technology-based firms which have remained Table 13.2

Spin-off classification Type of parent organization

Main resources received from the parent organization

Academic research centres

Established firms

Finance

Type A

Type B

Managerial

Type D

Type C

Science-based industries and spin-offs 271 marginal and heavily dependent on the initial large customers (Tether and Storey, 1998). Especially in recent years, international outsourcing has become more pronounced and many leading firms in the USA and Europe have started to outsource some R&D activity to third parties or spin-offs in emerging countries such as India and China. Typically, outsourcing concerns relatively low-end R&D activities like new product testing which do not require much ‘up front’ cost and financial resources (Baker and Hubbard, 2003; Giarratana, Pagano and Torrisi, 2005). Type D spin-offs are quite unusual and may take place when academic researchers start a new firm which offers services (such as training or technical consulting) that exploit the reputation of the academic institutions. It is more likely that this type of spin-off is spawned by science parks and incubators, organizations where universities and public research centres often play a key role. In sum, our taxonomy points out different types of spin-offs with different characteristics, incentives and likelihood of surviving. In the following sections we analyse the implications for the parent organization’s human resources policies and IPR management, and for the NSTF growth potentialities. 4 Human resources strategies Spin-offs drain from the parent organization resources such as human capital, technology, business practices, organizational routines and network connections (Christensen, 1997; Chesbrough, 2003). As noted by Chesbrough (2003), resource drain and spin-offs are likely to occur in technological fields that fall outside the existing product lines of the incubator. Therefore spin-offs do not necessarily compete with the parent company. However, this mechanism could generate a trade-off. On the one hand, spillovers and spin-offs may reduce the established firm’s ex-ante incentive to invest in human capital (Davenport et al., 2002; Poldony and Baron, 1997). On the other, in high-tech industries firms are often spurred to invest in R&D and human capital by the dynamism of an innovation-based competition. Given these conditions, established firms face the following set of strategies: (a) to adopt a flexible and decentralized organizational structure to reduce the rate of unused knowledge; (b) to allow for a given rate of spin-off and, possibly, carry out corporate venture capital activities; (c) to prevent spin-offs through non-disclosure agreements and investments in deterrence, for example, by establishing a reputation as a tough or aggressive incumbent. The first strategy is adopted by firms that aim to establish an internal entrepreneurial environment (‘intrapreneurship’). This strategy is constrained by the need to keep corporate coherence under control. The second strategy is quite popular in high-tech industries. A case in point is Xerox. Adobe Systems was a very successful spin-off of Xerox. Adobe’s founders have appropriated the software knowledge that Xerox had created for its workstation ‘Star’ project and exploited it commercially by selling it in the first place to the main Xerox customers, DEC and Apple (Chesbrough, 2003). Finally, an example of the third strategy is represented by Intel. This company is notorious for its aggressive reaction to start-ups established by former employees. A classical example is Tarari, a firm started by Andy Smerik, a former Intel R&D manager (Electronic News, 2002). Smerik was running an Intel R&D division that was responsible for about a half-dozen projects centred on intelligent traffic management. When he decided to start a new firm that aimed to focus on high-speed content processing, he found himself trapped in a long dispute

272

International handbook on industrial policy

with Intel human resource managers. Only at the end of a long bargaining process did he convince Intel’s managers to let the new firm take off. As Smerik claimed, these difficulties are due in part to Intel’s very restrictive IPR policy (‘Intel is religious about its processor operation’, Electronic News, 2002, p.1). The lack of investment in human resources is less severe for spillovers classified as types A and D, for obvious reasons. Academic institutions typically lack many downstream assets that are needed to start and grow a new venture. Finally, type B spin-offs are voluntary initiatives that aim to internalize the spillovers generated by the established firm. For instance, Xerox has taken a stake in several spin-offs (Chesbrough, 2003). 5 Intellectual property and the markets for technology Efficient technology trade contracts are difficult to set up, especially when multiple inventors are involved in cumulative, sequential innovations (Anton and Yao, 2002). The IPR regime can affect the emergence of markets for technology and in some industries, like chemicals and pharmaceuticals, strong protection of inventions has favoured technology trade. As Arora, Fosfuri and Gambardella (2001) pointed out, strong ‘IPRs are key for realizing the full benefits of a market for technologies’ (p.11). In other industries the relationship between the IPR regime and technology trade is more complex. In the software industry, for example, informal know-how trade occurred before a strong IPR regime emerged; however, in the 1980s, an increasingly stronger IPR regime paralleled the emergence of a market for software licences independent of hardware sales (Torrisi, 1998). More recently, in the security software, licensing accounts for an important share of the market (Giarratana, 2004). All in all, formal technology trade, like licensing contracts, has become widely more important across industries in recent years. Nearly 8000 technology licences were signed worldwide during the period 1985–97 (Anand and Khanna, 2000). For some firms the expenditures on licensing in technology are about 12 per cent of total R&D budgets (Arora et al., 2001). More efficient markets for technology should crowd out non-voluntary spin-offs (for example, Type C) because a larger share of proprietary knowledge that is not used by the firm can be licensed (or sold) to third parties or sponsored spin-offs (Type B). By the same token, sponsored spin-offs spawned by universities (Type A) are favoured by a more efficient market for technology. European data provided by Cesaroni and Giuri in the present volume show that universities are the most active institutions in using patents to stimulate new forms of formation, typically through processes of spin-off creation. More generally, the empirical evidence highlights that several universities and public research centres prefer to transfer their technology to spin-offs rather than licensing to established firms or other start-ups. Bray and Lee (2000) have estimated that, in 1996, the US university average value of equity sold in public spin-off companies was larger than the average annual return on traditional licences.6 In line with these findings, a survey of 64 US universities conducted by Thursby and Thursby (2002) shows that university patent applications have grown at the same rate of university spin-offs and a greater rate of licensing contracts. It is worth noting that, often, spin-offs tend to rely on informal technology transfer channels with the parent research institutions rather than formal channels such as patenting and licensing. For example, the new technology-based firms analysed by Autio and Yli-Renko (1998) rely on informal channels such as informal discussion between colleagues. In contrast, patenting and licensing are much less frequent.

Science-based industries and spin-offs 273 As far as university STSOs are concerned, equity stakes in spin-offs are important, especially in the USA, but they do not necessarily substitute for traditional licences to independent firms. As noted by Bray and Lee (2000), ‘when combined with a strong program of traditional licensing, taking equity in start-up companies maximizes the financial return that universities realize from their intellectual property’ (p.386). Why do universities prefer spin-offs to licensing? One reason is the positive reputation effect on the parent research institutions. Universities can use leverage on this reputation to obtain future research grants and to signal the effectiveness of job placement strategies among potential students. 6 Spin-off performance The factors affecting post-entry performance are the subject of a rich body of the literature. As Geroski (1995) has pointed out, ‘entry appears to be relatively easy, but survival is not’ (p.435). It is important, then, to understand which factors help to predict survival and post-entry growth. Several empirical studies which have focused on the start-up size as a predictor of post-entry performance have reached contrasting results (for example, Audretsch, Santarelli and Vivarelli, 1999). Another stream of the literature explores the importance of initial firms characteristics such as founders’ skills and working experience as a predictor of performance (for example, Bhide, 2000). An important issue from a public policy perspective is whether spin-offs have better chances to survive and grow as compared with other start-ups. For instance, Klepper (2001b) has demonstrated that in the US car industry spin-offs from large incumbents have a lower probability of failure and a higher performance compared with other newcomers. Similarly, in the laser industry, spin-offs account for an important market share and the two market leaders have been spawned by earlier entrants (Klepper and Sleeper, 2000). An important reason for the comparative advantage of spin-offs is that the R&D activities conducted by established firms are not a pure public good. Unlike other firms, spin-offs have direct access to key resources produced by the parent firm. They can also benefit from a greater familiarity with managerial practices that can be exploited in the market. Therefore the scale and scope of the parent organization’s activities are key determinants of success. Important differences exist between different types of spin-offs with respect to the probability of survival. The literature suggests that Type A spin-offs have a low probability of surviving (Shane and Stuart, 2002). Despite the knowledge and financial resources inherited by the parent institution, Type A spin-offs often face a high level of technological and market uncertainty, for two reasons. First, they have to devote significant efforts before the ideas developed in the parent’s research environment can be brought to the market. Second, they have to acquire downstream resources and capabilities (such as marketing and general management) to be competitive in the business world. To our knowledge, however, there is no systematic evidence about post-entry performance of university/ public research centres’ SBSOs. A case in point is represented by new science and technology-based firms established by scientists in France which have lower failure rates and greater growth rates than the average firm in the same country (Mustar, 1997). Type B spin-offs may suffer from similar problems, especially when they are originated by industrial R&D laboratories’ basic research. But, unlike Type A spin-offs, this type of spin-off will probably benefit from unintentional spillovers such as management experience and the business network of the parent company.

274

International handbook on industrial policy

Type C spin-offs may include both highly risky firms and relatively safe ventures. For instance, an involuntary spin-off which focuses on new markets most probably cannot rely on the business experience of the parent company. On the contrary, an involuntary spinoff which enters the same market of the parent company will have to face the reaction of the latter. More secure spin-offs are those which result from vertical disintegration and outsourcing. However, the dependence on the parent company as the main source of revenues may hamper the future growth of the spin-off. For instance, a number of Irish firms have been spawned by the local subsidiaries of multinational corporations. However, most of these spin-offs have remained small or have not survived, probably because they have not succeeded in reducing their dependence on the parent company (Giarratana, Pagano and Torrisi, 2005). 7 Implications for public policy and conclusions This chapter has argued that NSTFs in general have a high growth potential compared with other start-ups. And science and technology-based spin-offs have even greater opportunities to grow than other NSTFs. This may explain why, in the 1980s and especially in the 1990s, entrepreneurship policies became more selective and have increasingly supported NSTFs rather than new firm formation in general (Pages, Freedman and Von Bargen, 2003). This does not mean that NSFs have a low failure rate. Innovative activities are bounded by uncertainty and in fact many NSTFs fail. However, even in the case of failure, their ideas and skilled personnel can benefit other firms. Typically, regional and national programmes have tried to stimulate the supply of services such as technology-oriented law firms, venture capital funds, incubators located in science and technology parks, and university technology transfer offices following the examples of the Silicon Valley and the Boston Area (Audretsch, 2003). Among the various public policy initiatives explicitly addressing the formation of new technology-based firms, incubators are probably the most relevant to our purposes. Science parks and business incubators are ‘property-based organizations with identifiable administrative centres focused on the mission of business acceleration through knowledge agglomeration and resource sharing’ (Phan et al., 2005, p.167). These organizations are typically established by universities and public agencies, often in collaboration with other public or private stakeholders such as local business groups and regional bodies. There are over 123 science parks and business incubators in the USA, according to the Association of University Research Parks (AURP) and 850 in the European Union according to the European Commission’s Enterprise Directorate General (see Phan et al., 2005). Even if their activities have not been examined in detail in this chapter, it should be noted that they play a role similar to that of universities and established high-tech firms. The intensity of linkages with the academic system is very heterogeneous and could vary from country to country. For instance, the Business and Technology Centres in Germany and the Technopoles in France promote local development without necessarily or exclusively relying on the science produced by local universities, while in the UK the leading role of universities is a historical attribute. Science parks are important because they spawn NSTFs or attract established firms from elsewhere, such as high-tech multinational firms that are supposed to generate spillovers to domestic firms. Unfortunately, the empirical evidence in this respect is very

Science-based industries and spin-offs 275 sparse and shows that the direct contribution of science parks and incubators to the formation of rapidly growing NSTFs is quite difficult to assess. Compared with the US science parks, European science parks have spawned or attracted a smaller number of firms. Employment growth of on-science parks firms in Europe is also smaller than in the USA (Storey and Tether, 1998). Westhead and Storey (1994) have found that on-science parks NSFTs in the UK have identical failure rates and only marginally greater growth rates than off-science parks. The evidence provided by a recent study of Italian science parks and business incubators leads to different results. Colombo and Delmastro (2002) have compared the performance of Italian firms located in incubators within science parks and business innovation centres with that of a control sample of off-incubator firms. Their results show that on-incubator firms outperform their off-incubator counterparts in terms of growth, innovation and human capital endowment. To our purposes here it is worth noting that the sample of on-incubator firms used by Colombo and Delmastro (2002) includes both university/research centre spin-offs and corporate spin-offs. More recently, public policies have tried to go beyond the supply of services to foster a local ‘eco-system’ characterized by a pool of diversified skills and actors, strong mechanisms for the selection of successful entrepreneurs and business models, and a network of linkages among different players: entrepreneurs, financiers, providers of support services such as legal, marketing, and real estate developers, public officers and policy makers. An example of these policies is Northern Virginia’s Netpreneur program launched in the 1990s to support entrepreneurship in Internet-related sectors in Washington DC (Pages, Freedman and Von Bargen, 2003). Along these streams of analysis, this chapter has analysed the two main sources of entrepreneurship: universities and industrial R&D laboratories. Technology transfer activities of university and public research centres occupy a central position in entrepreneurship policies. As Jaffe (1989) and other scholars have pointed out, universities are an important source of spillovers for firms located in the same region. This is especially the case of small firms which have limited in-house R&D activities. As discussed before, academic spin-offs represent an important channel of technology transfer (Louis et al., 1989; Shane and Stuart, 2002). We have seen that there is substantial empirical evidence showing that different technology transfer policies, such as the distribution of royalty fees between inventors and the university and equity investments in start-ups have an impact upon the rate of university spin-offs (Di Gregorio and Shane, 2003). There are other potentially important dimensions that remain quite unexplored. For instance, what about the effects of incentive schemes that explicitly aim at researchers who have started new firms and try to return to their previous academic positions? In Italy, for example, a potentially important step in this direction is represented by a law passed in 1999 (‘Legge 297/199’) which allows researchers employed by university and public research centres to work temporarily for a business firm. To our knowledge, the implications of this law for the rate of academic spin-offs have not been investigated yet. Another issue which would deserve further analysis is the support of hybrid organizations between universities and private firms.7 Established R&D-intensive firms are the second major source of voluntary or involuntary spin-off. But, as we have seen before, established firms have incentives to prevent spin-offs by imposing ex ante non-disclosure contracts on employees or engaging in costly legal battles against spin-offs. This explains the proliferation of US consulting companies that specialize in intellectual property-based new ventures and patent litigations between

276

International handbook on industrial policy

spin-offs and former employers (Reuters, 2004). It is important to note that a strong intellectual property regime helps established firms to discourage involuntary spin-offs and imitators and more generally to control and manage more structurally their R&D resources. Conversely, in the case of patents, a broad patent range allows inventors to make a large number of claims per patent. This gives excessive monopoly power to established firms and may hamper future innovations and competition. The negative effects of a broad patent range are particularly significant when technical change is characterized by a sequence of interdependent innovations, that is when technical change is cumulative in nature, as in semiconductors, computers and automobiles. This is because the holder of a pioneer patent can control and retard the future development of important complementary inventions (Merges and Nelson, 1990). Policy makers then face a trade-off between the incentives of early inventors and the benefits of a competitive environment that favours subsequent improvements and new firm entry. Policies that impose some limits on the scope of patent protection or prescribe compulsory licensing to allow for the development of complementary inventions may have positive effects on the rate of science and technology-based spin-offs because these are part of the competitive environment and may provide an important contribution to innovation and growth. Drawing on the importance that spin-offs and people mobility have in fast-growing high-tech clusters like the Silicon Valley, some scholars have argued that the legal treatment of mobile high-tech workers should allow weak property rights on information, narrow trade secret scope and limited bounded enforcement of ‘conveyances not to compete’ contracts (Klepper, 2001b). Public intervention to favour the formation of NSTFs can take place through the supply of high-tech entrepreneurs. Founders of NSTFs often have a high level of education and the supply of skilled personnel, especially in science-based fields, may represent an important constraint for the establishment of new technology-based firms. The marked differences across countries in the share of the workforce with a university degree may account for the different rates in NSTF formation. For instance, in Italy the low proportion of people with a university degree in scientific fields may explain the low rate of NSTF formation which is at odds with the typically high overall rate of new firm formation of this country. Public policies should also focus on the propensity of university students to start new business ventures. Some universities such as Chalmers (Gothenburg) and Linkoping in Sweden have stimulated several academic spin-offs. Other experiences, such as the temporary entrepreneurial places (TEPs) at the University of Twente in the Netherlands have also contributed to stimulating the supply of high-tech entrepreneurs even though their effects on the ‘cultural’ gap between universities and business are debatable. Notes 1. Contracting with potential entrepreneurs is costly because of information asymmetry and agency problems (the firm does not know who are the potential entrepreneurs). 2. By university spin-offs the AUTM means all new firms which have been founded by licensing of a university IPR. This definition does not distinguish firms which have been established by university researchers from firms that have been established by non-university employers. Moreover, it may lead to underestimation of university start-ups, since it does not account for new firms established by entrepreneurs who draw on non-protected university inventions (Feldman, 2003). 3. According to the OECD (2003), Public Research Organizations (PRO) include research-performing universities, public research laboratories and other research organizations receiving a substantial share of their funds from public sources. Moreover, OECD distinguished spin-offs (new firms founded by staff from a

Science-based industries and spin-offs 277

4. 5. 6. 7.

PRO) from start-ups (new firms established to develop or commercialize technology licensed from a PRO but without involvement of PRO staff). In some countries universities have started to adopt an IPR policy only recently and this may explain the lack of correlation between spin-offs and university patents. In Italy, for instance, universities started to adopt formal intellectual property policies only in the 1990s (see Baldini, Grimaldi and Sobrero, 2004). Corporate venture capital represents about one-third of VC funds in 2000. Licensing revenues typically come in two parts: (a) a fixed licence issue fee and (b) an annual royalty on sales. Typically equity stakes substitute for the ‘upfront’ licence issue fee (Bray and Lee, 2000). A survey conducted by the OECD shows that public research organizations in some countries have increased the recognition of intellectual property activity in human resources management policies (hiring and career advancement) (OECD, 2003).

References Acs, Z., B. Carlsson and C. Karlsson (1999), ‘The linkages among entrepreneurship, SMEs and the macroeconomy’, in Z. Acs, B. Carlsson and C. Karlsson (eds), Entrepreneurship, Small & Medium-Sized Enterprises and the Macroeconomy, Cambridge, MA: Cambridge University Press, pp.3–42. Anand, B.N. and T. Khanna (2000), ‘The structure of licensing contracts’, Journal of Industrial Economics, 48 (1), 103–35. Anton, J. J. and D.A. Yao (2002), ‘The sale of ideas: strategic disclosure, property rights, and contracting’, Review of Economic Studies, 67, 585–607. Arora, A., A. Fosfuri and A. Gambardella (2001), Markets for Technologies, Boston: MIT Press. Audretsch, D.B. (1997), ‘Technological regimes, industrial demography and the evolution of industrial structures’, Industrial and Corporate Change, 6(1), 49–82. Audretsch, D.B. (2003), ‘Entrepreneurship policy and the strategic management of places’, in D. Hart (ed.), The Emergence of Entrepreneurship Policy, New York: Cambridge University Press. Audretsch, D.B., E. Santarelli and M. Vivarelli (1999), ‘Start-up size and industrial dynamics: some evidence from Italian manufacturing’, International Journal of Industrial Organization, 17, 965–83. Auerswald, P.E. and L.M. Branscombe (2003), ‘Start-ups and spin-offs. Collective entrepreneurship between invention and innovation’, in D.H. Hart (ed.), The Emergence of Entrepreneurship Policy, New York: Cambridge University Press, pp.61–91. Autio, E. and H. Yli-Renko (1998), ‘New, technology-based firms in small open economies – an analysis based on the Finnish experience’, Research Policy, 26, 973–87. Baker, G.P. and T.N. Hubbard (2003), ‘Make versus buy in trucking: asset ownership, job design, and information’, American Economic Review, 93(3), 551–72. Baldini, N., R. Grimaldi and M. Sobrero (2004), ‘Institutional changes and the commercialization of academic knowledge: a study of Italian universities’, CRESCO working paper 11, University of Siena, Italy. Bartelsman, E., S. Scarpetta and F. Schivardi (2003), ‘Comparative analysis of firms’ demographics and survival: micro-level evidence for the OECD countries’, OECD Economic Dept. Working Papers, 348, Paris. Bhide, A.V. (2000), The Origin and Evolution of New Business, Oxford: Oxford University Press. Blanchower, D. and A. Oswald (1998), ‘What makes an entrepreneur?’, Journal of Labor Economics, 16(1), 26–60. Bray, M.J. and J.N. Lee (2000), ‘University revenues from technology transfer: licensing fees vs. equity positions’, Journal of Business Venturing, 15, 386–92. Burgelman, R.A. (1983), ‘Corporate entrepreneurship and strategic management: insights from a process study’, Management Science, 29(12), 1349–64. Chesbrough, H. (2003), ‘The governance and performance of Xerox’s technology spin-off companies’, Research Policy, 32(3), 403–21. Chiesa, V. and A. Piccaluga (2000), ‘Exploitation and diffusion of public research: the case of academic spinoff companies in Italy’, R&D Management, 30, 329–39. Christensen, C. (1997), The Innovator’s Dilemma, Boston: Harvard Business School Press. Clarysse, B., M. Wright, A. Lockett, E. van de Velde and A. Vohora (2005), ‘Spinning out new ventures: a typology of incubation strategies from European research institutions’, Journal of Business Venturing, 20(2), 183–216. Colombo, M.G. and M. Delmastro (2002), ‘How effective are technology incubators? Evidence from Italy’, Research Policy, 31, 1103–22. Davenport, S., A. Carr and D. Bibby (2002), ‘Leveraging talent: spin-off strategy at industrial research’, R&D Management, 32(3), 241–54. Debackere, K., M. Luwel and R. Veugelers (1999), ‘Can technology lead to a competitive advantage? A case study of Flanders using European patent data’, Scientometrics, 44(3), 379–400. Di Gregorio, D. and S. Shane (2003), ‘Why do some universities generate more start-ups than others?’, Research Policy, 32, 209–27. Electronic News (2002), ‘Intel execs branch out on their own’, 48(7), 1–2.

278

International handbook on industrial policy

Feldman, M. (2003), ‘Entrepreneurship and American research universities. Evolution in technology transfer’, in D.H. Hart (ed.), The Emergence of Entrepreneurship Policy, New York: Cambridge University Press, pp.92–112. Garvin, D. (1983), ‘Spin-offs and the new firm formation process’, California Management Review, 25(2), 3–20. Geroski, P. (1995), ‘What do we know about entry?’, International Journal of Industrial Organization, 13, 450–56. Giarratana, M. (2004), ‘The birth of a new industry: entry by start-ups and the drivers of firm growth: the case of encryption software’, Research Policy, 33(5), 787–806. Giarratana, M., A. Pagano and S. Torrisi (2005), ‘The role of multinational firms in the evolution of the software industry in India, Ireland and Israel’, in A. Arora and A. Gambardella (eds), From Underdogs to Tigers: The Rise and Growth of the Software Industry in Brazil, China, India, Ireland, and Israel, New York: Oxford University Press. Hellman, T. (2002), ‘When do employees become entrepreneurs?’, Stanford Graduate School of Business, Research Paper Series, 1770, June. Henderson, R. and K. Clark (1990), ‘Architectural innovation: the reconfiguration of existing product technologies and the failure of established firms’. Administrative Science Quarterly, 35(1), 9–30. Henrekson, M. and N. Rosenberg (2001), ‘Designing efficient institutions for science-based entrepreneurship: lessons for the US and Sweden’, Journal of Technology Transfer, 26, 207–31. Jaffe, A.B. (1989), ‘Real effects of academic research’, American Economic Review, 79(5), 957–70. Klepper, S. (2001a), ‘The evolution of the US automobile industry and Detroit as its capital’, mimeo. Klepper, S. (2001b), ‘Employee startups in high-tech industries’, Industrial and Corporate Change, 10(3), 639–74. Klepper, S. and S. Sleeper (2000), ‘Entry by spinoffs’, mimeo, Carnegie Mellon University, Pittsburgh. Louis, K.S., D. Blumenthal, M.E. Gluck and M.A. Stoto (1989), ‘Entrepreneurs in academe: an exploration of behaviors among life scientists’, Administrative Science Quarterly, 34(1), 110–31. Merges, R. and R. Nelson (1990), ‘On the complex economics of patent scope’, Columbia Law Review, 90(2), 839–70. Mustar, P. (1997), ‘Spin-off enterprises. How French academics create hi-tech companies: the conditions for success and failure’, Science and Public Policy, 24, 37–43. Nicolau, N. and S. Birley (2003), ‘Social networks in organizational emergence: the university spinout phenomenon’, Management Science, 49(12), 1702–25. OECD (2003), Turning Science into Business. Patenting and Licensing at Public Research Organizations, Paris: Organization for Economic Cooperation and Development. Pages, E.R., D. Freedman and P. Von Bargen (2003), ‘Entreprenurship as a state and local development strategy’, in D.M. Hart (ed.), The Emergence of Entrepreneurship Policy, New York: Cambridge University Press, pp.204–59. Phan, P.H., D.S. Siegel and M. Wright (2005), ‘Science parks and incubators:observations, synthesis and future research’, Journal of Business Venturing, 20, 165–82. Poldony, P.J. and J.N. Baron (1997), ‘Resources and relationships: social networks and mobility in the workplace, American sociological review’, Management Science, 62(4), 673–93. Reuters (2004), ‘Patent payoffs prop start-ups’, Daniel Sorid, 12 April. Reynolds, P.D. (2000), ‘National panel study of business start-ups: background and methodology’, Advances in Entrepreneurship, Firm Emergence and Growth, 4, 153–227. Roberts, E.B. and D.E. Malone (1996), ‘Policies and structures for spinning off new companies from research and development organizations’, R&D Management, 26(1), 17–48. Rosenberg, N. (2003), ‘America’s entrepreneurial universities’, in D.M. Hart (ed.), The Emergence of Entrepreneurship Policy, New York: Cambridge University Press, pp.113–37. Shane, S. (2001), ‘Technological regimes and new firm formation’, Management Science, 47(9), 1173–90. Shane, S. and T. Stuart (2002), ‘Organizational endowments and the performance of university start-ups’, Management Science, 48(1), 154–70. Steffensen, M., E.M. Rogers and K. Speakman (1999), ‘Spin-offs from research centers at a research university’, Journal of Business Venturing, 15, 93–111. Storey, D.J. and B.S. Tether (1998), ‘New technology-based firms in the European Union: an introduction’, Research Policy, 26, 933–46. Tether, B.S. and D.J. Storey (1998), ‘Small firms and Europe’s high technology sectors: a framework for analysis and some statistical evidence’, Research Policy, 26, 947–71. Thursby, J.G. and M.C. Thursby (2002), ‘Who is selling the ivory tower? Sources of growth in university licensing’, Management Science, 48(1), 90–104. Torrisi, S. (1998), Industrial Organisation and Innovation. An International Study of the Software Industry, Cheltenham, UK and Lyme, USA: Edward Elgar. Westhead, P. and D.J. Storey (1994), ‘An assessment of firms located on and off science parks in the United Kingdom’, HMSO, London.

14 Science parks and high-tech clustering Jan Annerstedt

1 Introduction: science parks and ‘clusters of competencies’ Science parks as instruments for industrial policy have existed for more than half a century, but there is still no general acknowledgment or broad recognition of their potential or real economic impact. Controversies continue to arise around their costeffectiveness. Are the typical science parks of today effective platforms for high-tech innovation activity? Why do some science parks reinforce and sustain the competitive performance of the companies and institutions that operate on their premises? How is it that a number of parks fail to reach their main objectives? Many well-documented case studies of science parks and similar, managed innovation environments are available. Some case reports depict major success stories, while others show a range of non-intended consequences. There are few systematic studies of longterm results of science park activity, although there are many hundreds of science parks now operating in locations across the world. Critics claim that most science parks do not reach their ultimate set of goals, but that they rather serve secondary objectives such as attracting inward investment, advancing high-tech infrastructure, creating a vehicle for indirect subsidies or raising property values of their neighbourhoods.1 Thus some science parks have become symbols rather than real catalysts of science–technology–economy relations in a modernizing society.2 Policy interest in science parks declined somewhat in the first part of the 1990s. One reason was the ambition in many cities and regions to renew the broader urban structures in which science parks are typically embedded. Urban regeneration activities were recognized as more important determinants of local economic development.3 Nonetheless, over the last 10 years, with the promotion by policy makers of specialized, local ‘clusters’ of firms and supporting institutions as strategic means for industrial policy, science parks have come back into the limelight of the centre stage for industrial policy deliberations. Science parks are even considered the genesis of managed ‘clusters of competencies’. On the other hand, the science parks of today appear in more than 57 varieties. Some truly deserve the term and are successfully integrated into a wider policy realm by stakeholders at the local, regional and even national levels. They could involve a wide range of specialized, knowledge-intensive institutions supporting high-tech business firms as tenants and stakeholders. Other parks are actually well-disguised real estate projects, branded and effectively marketed as science parks. From an industrial policy perspective, it is important to understand how locally managed innovation environments are inserted into the larger economy and society, yet science parks might be viewed as relatively self-contained local entities. They are definitely locally anchored, but competition and collaboration among firms, institutions and individuals do make them into integrated parts of regional and even global levels of the economy. With high levels of uncertainty in innovation processes, it makes little sense for firms to engage in a simple network of collaboration; rather, they should seek to become 279

280

International handbook on industrial policy

stakeholders in clusters of firms and supporting institutions. Clusters are typically a combination of firms specializing in selected competence areas and they depend on close relationships over relatively extensive periods of time. Being more stable than networks in their composition of firms, a cluster could allow for more joint experimentation and the sharing of knowledge and know-how. The members of a cluster with changing competencies could benefit more from close combinations of project partners than firms within a more loosely organized network. Clusters are often confined in scope (firms in one or a few related industries only) and restricted in space (clusters are usually to be found at a local or regional level only).4 Yet, more and more, localized clusters of firms do cooperate with other, similar clusters elsewhere. Clusters are linked among city hubs. This chapter summarizes some of the traits in the evolution of science parks as dedicated areas for developing managed ‘clusters of competencies’. Three generations of science parks are presented. Emphasis is placed on how the third-generation science park is becoming an essential ingredient of the revitalization of cities and of urban and regional renewal. From a policy point of view, a science park could easily be recognized as an instrument for constructing local and regional advantages. 2 What is this thing called science park? A science park is the generic term for ‘an organization managed by specialized professionals, whose main aim is to increase the wealth of its community by promoting the culture of innovation and the competitiveness of its associated businesses and knowledgebased institutions’. This is the broad, authorized definition from the International Association of Science Parks (IASP).5 The term ‘science park’ may be replaced in the definition by research park, technology park or Technopole. The most resourceful science parks will provide an assortment of value-adding services together with high-quality space and facilities. According to the IASP membership criteria, a science park should facilitate the creation and growth of innovation-based companies, for example through incubation or spin-off processes among existing firms. To enable its goals to be met, the science park should also stimulate the flow of technology and related knowledge amongst business firms, universities and other R&D institutions, as well as wider markets. In many cases, partnerships between academic institutions, local and regional governments and the private sector are seen as particularly instrumental. In this chapter, the term ‘science park’ refers to parks with some or all of these innovative resources and characteristics. Accordingly, with this label, there are parks with tenants mainly engaged in various stages of applied research and experimental development. Moreover, some science parks could best be called technology parks, since the majority of tenants are deeply engaged in prototype development and product design and not in applied research as a primary activity; nor are they tapping into a research facility. Some countries apply stricter criteria. In the United Kingdom, for example, science parks without working relations with an R&D facility of some size would not qualify for membership of the national Science Park Association,6 which requires its members (a) to have formal operational links with a university, (b) to encourage on-site formation and growth of knowledge-based businesses; and (c) to manage the transfer of technology and business skills to on-site businesses.

Science parks and high-tech clustering 281 3 Science park value: benefits of ‘clusters of competencies’ During the past ten years or more, in many parts of the globe, market competition, privatization and deregulation have released extraordinary economic and other capabilities while leading to significant differences of innovative capabilities between cities and other urban spaces. Around the world, key business areas are being selected and equipped with state-of-the-art telecommunications gear, forming ‘glocal nodes’ that link up with equivalent areas elsewhere and with the region that surrounds each of them. In economic terms, they are all highly specialized and diverse, while they remain local as well as global in orientation.7 New or young high-tech firms seem to be influenced by a whole range of cluster benefits combined into a well-managed innovation environment. For high-tech industries, the following common assets and competencies seem to drive agglomeration of firms when becoming part of a science park or other managed innovation environment:8 ● ● ● ●

● ● ●

Availability of a range of specialized services in direct support of innovation activity and business incubation. Pools of highly-specialized or multi-skilled human resources. Early identification of new technological (and related market) opportunities. Access to new technologies as well as new products and services through a relatively large number of inventive start-up firms that may offer the benefit of ‘technology spill-overs’ (incl. migration of personnel among firms). Rapid product development due to proximity in problem-solving expertise. Swift reproduction of product innovations. Easy entries to seed and other venture capital.

The special value of a science park or other managed innovation environment for a firm (or for a subdivision or other unit of a firm) may depend on one or several of the assets and capabilities listed above. Business firms strive to achieve cheaper factor costs and easy access to other useful resources. Public policies could help to make science parks and such intermediary institutions effective in providing the right mix of services relevant to the companies and complement firm-specific or cluster-specific assets. 4 Efficiency through proximity and dynamic positioning An underlying argument for science parks and other managed innovation environments is that fast and easy transmissions of new knowledge occur most effectively among actors located close to each other. According to David Wolfe, proximity to critical sources of knowledge, whether they are accessible via public or private research institutions or grounded in the core competencies of leading firms, facilitates the process of acquiring new technical knowledge. This is especially the case when relevant knowledge is just behind the research frontier or may involve a largely tacit, yet accessible, dimension. Knowledge of this nature may be transmitted through a range of well-organized science park services and, maybe even more efficiently, through interpersonal contacts in the park and the cross-sectoral mobility of highly-skilled personnel.9 Stefano Breschi and Franco Malerba argue that this approach overestimates the benefits of physical proximity. Sheer proximity is not sufficient to account for efficient knowledge transfers and spillovers. In their view, the body of research on local knowledge spillovers

282

International handbook on industrial policy

overlooks the broader set of factors and conditions that support the effective transfer of knowledge in clusters. For example: ‘A key feature of successful high-technology clusters is related to the high level of embeddedness of local firms in a very thick network of knowledge sharing, which is supported by close social interactions and by institutions building trust and encouraging informal relations among actors.’10 In other words, the degree to which firms can tap into a common knowledge base at the local level depends on more than just spatial proximity, dynamic positioning and a relative transparency. 5 Communicative interaction and ‘institutional thickness’ Another way of explaining how and why science parks and other institutionally supported innovation environments evolve over time is to delve deep into the park’s embeddedness in the local and regional economy.11 It may be necessary to recognize that there are different levels of embeddedness as firms and institutions mediate their goals and innovation strategies in relation to a science park. A key dimension of embeddedness in a local economy is trust in the form of familiarity (through repeated interactions) and norms that create predictability and trustworthiness among individuals in firms and institutions. Economically successful regions seem to provide also effective institutional and social support to trust building. Trust by interpersonal contacts, reputation and calculation (calculative form of trust) seems to be the dominant coordinating mechanism of innovative networks or communities. Over time an individual’s reputation could be enhanced by reciprocal trading of know-how.12 However, failure can quickly erode one’s reputation. Trust, and the ‘social capital’ that supports it, is a common element of the industrial districts in northern Italy, northern California and elsewhere. Here, ‘social capital’ is defined as the means by which reciprocal exchange can occur in network forms of organization (such as a science park or other managed ‘cluster of competencies’).13 However, the way in which social capital supports trust in the specific local and regional setting differs greatly. In Silicon Valley the foundation of trust seems to be one based on a reputation for technical skills and other competences, measured by performance. The basis for trust even in science-based or high-tech business can also be social or familial connections within business and in the larger community. The importance of networks linking individuals across different companies within a city or region is emphasized in studies of inter-firm networks, characterized by cooperation as well as competition. When a science park contains considerable social capital, it can be considered well embedded in the local economy. A special quality of embeddedness is called ‘institutional thickness’,14 which indicates the presence of R&D firms and other supporting institutions, including non-local actors, the levels of communicative interactions and knowledge sharing in the park and in the neighbouring area and, as a qualitative dimension, the awareness and recognition of being involved in a common enterprise fostered by mutual interests. Collectively, the park’s stakeholders work together to create a context that will appear to be more or less ‘thick’ and therefore, it is assumed, beneficial to those involved. Originally, the notion of ‘institutional thickness’ was created for understanding better local or regional economies of scope, but it is used also to comprehend why science parks may remain attractive places for innovative activities to firms and institutions. The notion may also explain why some science parks are able to foster a continuous interplay between

Science parks and high-tech clustering 283 (a) the local urban setting, which they are part of or even deeply embedded in and (b) globalized innovation processes, which involve non-local actors across sectors, industries and places, yet linked to the park. Here, the park may serve as a ‘thick’ node in global innovation networks that reach firms and institutions through value chains on several continents, while these firms and institutions will still be having a strong local presence. If effective, a well-connected science park or other managed innovation environment may help a city to enforce its role as a regional hub city, linked functionally to other cities with innovative firms and supporting institutions. As a ‘space of flows’ of knowledge, a science park could help position a city or other urban area in relation to global value chains cross-cutting different spatial scales. 15 Recently, John Hagel and John Seely Brown have exposed a ‘hassle factor’ for well-connected companies when transacting with other companies in a science park or other location and, also, across borders. Contrary to the ‘transaction costs’ theory of business collaboration,16 they claim that cooperation and transactions between companies are not simply a form of cost, but could also be an important source of innovation. A science park or other business ecosystem could be made more effective if the ‘productive friction’ among companies and business units is recognized as an accelerator of innovation processes. By a series of examples, they say that successful innovations occurred not because the interactions between companies were seamless, but because ‘the activity at the seams was challenging, stimulating and catalytic’.17 This could happen at any level, among companies in networks as well as among individuals, engaged with one another for problem solving. ‘Productive friction’ arises with misunderstandings and arguments. Time and efforts are consumed before the resolution and learning take place. ‘Too often, in fact, the friction becomes dysfunctional’ and leads to mistrust. Like the effective science park management, Hagel and Seely Brown highlight the need to design performance fabrics that can help companies to work more effectively together in long-term relationships and push each other to accelerate their own and their joint capability building for innovation.18 For a company that needs to network, there are several broad imperatives: focus tightly on distinctive internal capabilities, effectively mobilize the capabilities of other specialized business partners, and harness ‘productive friction’ to accelerate capability building across networks of business partners.19 Turned into a science park advantage, ‘productive friction’ should occur when companies and people with diverse and still appropriate specializations creatively resolve difficult business issues such as a workable design and a well-functioning prototype. More strategically, it will not be enough with appropriate ‘thickness’ of institutional and other support, nor with plenty of chances to reach ‘productive frictions’ across companies and competence areas. To gain full benefits, a science park and the companies it may host must also establish framework conditions and processes to help the people to reflect on the practices emerging from these collaborations, recognize patterns of effective communication and increase awareness of the high-impact solutions across the tenants of the science park. These are not easy tasks to achieve. 6 What is the economic impact of a science park? Except for some combinations of case studies, few researchers have studied science parks comparatively.20 For example, despite the operations of many hundreds of science parks across the world, there is no tradition of specialized, comparative investigations of the

284

International handbook on industrial policy

economic impact of science parks in the local and regional contexts. For example, the development of indicators for depicting types and degrees of the actual impact of transfer of R&D results to commercial applications through science parks is still rudimentary. Much remains to be done to understand the economic benefits – and the costs and deficiencies – of a science park and its tenants in full operation. However, there are many studies of firm performance in science parks, incl. the productivity of R&D investments.21 Apart from studies specifically of companies in science parks, there are three main types of research on the economic impact of science parks: those that examine its entrepreneurs and their venturing activities, those that assess broadly science parks as managed innovation environments, and those that analyse – at the systemic level – science parks and science–economy relations in a city, region and country. Efforts to connect these four research streams into a more comprehensive understanding of the economic impact of science parks have not been rewarding. At present, there is no generally accepted framework to link these levels of analyses of innovation that involve several actors, as there is, for example, for the similar relationship between various R&D and industrial design facilities of multinational firms and their subsidiaries.22 7 The first science park: Stanford Research Park In its modern form, the science park experience has a history of some 50 years. Just after World War II, as Dean of Engineering at Stanford University, Frederick Terman envisioned this part of northern California as a ‘community of technical scholars’, a modern counterpart of the groupings of medieval scholars in European centres of learning such as Heidelberg, Oxford and Paris.23 His vision came down to linking the Stanford campus with the regional economy by an industrial zone, owned by the university, open to business firms and enforced by a set of institutional measures. Terman strived for the formation of an inventive and entrepreneurial technical community, like those medieval communities where there was, in his words, ‘continuous ferment of new ideas and stimulating new challenges’.24 In 1951, the first modern science park was opened, soon to be called Stanford Research Park, which was to connect also with some of the world’s most successful technologyintensive companies. Palo Alto and the neighbouring cities in and around Santa Clara County were later to be branded ‘Silicon Valley’ and become one of the most significant high-tech regions that underpin the global ‘Information Society’ of our time. The thrust of Fred Terman’s original institutional invention (the modern science park) is the managed extension of a research-intensive university or other major R&D centre into the wider economy by academics becoming entrepreneurs, part-time industry experts and even acknowledged members of a high-tech business community. By design and by effective management, such a science and technology park should offer a productive innovation environment in which to grow ideas and build companies next to a research facility and a pool of academic and other experts. Ideally, these professionals should be willing to engage in informal communicative networks and to migrate between R&D institutions and industrial corporations and across sector borders. 8 The science park as a global phenomenon In the USA, the Stanford Research Park example was soon followed by the Cornell Business and Technology Park, New York (1952), the University Research Park (originally

Science parks and high-tech clustering 285 Swearingen Research Park), Oklahoma (1957), the Research Triangle Park, North Carolina (1959) and the Purdue Research Park, Indiana (1960).25 Today, the Research Triangle Park is the biggest in the USA with firms and institutions having some 40 000 employees, while Stanford Research Park accommodates 23 000. The Research Triangle remains the largest science park in the USA with some 2700 hectares or ten times the size of Stanford Research Park. The first wave of investments into science parks and high-tech business areas reached Europe in the early 1960s. At the time, science parks were still conceived as suburban high-tech zones and it would take a quarter of a century before science parks would start to ‘go urban’ and move into the fabric of city centres, while developing unique blends of innovation support services. The European versions of the US science park model were first elaborated and tested in the United Kingdom, yet the generally recognized European pioneer is Sophia-Antipolis of France (1969). Sophia remains Europe’s biggest science park (2300 hectares) with some 24 500 persons employed in its companies and institutions. A study in 2004 of Germany’s 440 municipalities showed that they are home to 362 science and technology parks (broadly defined).26 The majority of these parks would not qualify for full membership of a science park association, since the park’s research base is very limited, if existing at all, but the number of German ‘science parks’ shows that ambitions remain high at the local and regional levels in promoting science–technology–industry relations through such institutional arrangements. Similarly high densities of (broadly defined) science parks are found in the Nordic countries, especially in Finland and Sweden. Europe is now the continent with the highest number of science parks (of all varieties) although many should rather be described as ‘real estate parks with no university affiliation’. The latter variety represents a relatively common ‘low-end’ science park category on most of the other continents as well. By the mid-1980s, well-equipped science parks had become fashionable among industry policy specialists and investors in the OECD countries and remained so for about 10 years, while many hundreds of science parks were established in North America and West Europe and, a bit later, in several parts of Asia and Latin America. The first major science park investment in Japan (Tsukuba Science City) was made in the early 1970s. Japan now has at least 110 science parks in operation. The upsurge of science parks in the largest countries of Asia (China and India) started slowly in the mid-1980s. Officially, China now has about 100 science parks, but there are actually many more established through local initiatives among a large number of cities and provinces. During the past few years, the increase of dedicated high-tech zones, including new science parks, in East Asia led to the decision to hold the 2005 World Congress of Science Parks in Beijing, China. There are now science parks on all continents. Science parks have become a global phenomenon. 9 Science parks as part of the resource-rich urban fabric The traditional science park of the 1950s and 1960s was deliberately separated from the context of a city centre. Many of the parks were literally put into suburban areas or even further away from the economic centre of the city. Some of the original science parks were rural with no urban agglomeration in the neighbourhood. They were designated to remain separate, even detached, from the urban fabric. In retrospect, the suburban

286

International handbook on industrial policy

high-tech zones seem effective for some types of innovation activity, but less effective for others. Even if considered a city-related innovation environment, the suburban science park seems to have experienced serious shortcomings in its diversity, attractiveness of location, sustainability and sociocultural dynamism. What once seemed to be an acceptable tradeoff for the seclusion of the science park in its pioneering days has become a competitive disadvantage – at least for some of the parks. Cities competing for inward investments, resources for research and experimental development, high-tech business firms, and talented entrepreneurs to energize them, are concerned that they cannot provide framework conditions conducive to knowledge-intensive business firms and to talents searching for the best place to live and work.27 A major challenge is the ongoing trend of urbanization of today’s science park model. Across Europe as well as on other continents, the trend is to integrate into the urban spatial setting some of the main functions of the science park and other managed innovation environments. For example, a city’s inherent openness and accessibility, its energy and drive, paired with a broader participation among stakeholders, may require inventive thinking by urban planners, industry policy makers and private investors. Metropolitan cities in Europe, such as Barcelona, Copenhagen and Helsinki, try to exploit public–private partnerships and other modes of governance to the benefit of the new innovation environments and infrastructures which are integrated into the central parts of the cities.28 Fostering effective innovation environments has emerged as a principal policy issue in connection with urban renewal and city development and under pressure from competing city-regions due to economic globalization. At the city hall, the repercussions of this kind of urbanization could be registered across a wide spectrum of socioeconomic processes. These processes might involve science park-related developments such as accessibility of public services to high-tech businesses, management of clustering of competencies, real-estate asset values and the quality of the environment, urban morphology, and place branding and marketing. This will probably call for new policies or new combinations of existing policies as well as new interdisciplinary approaches among specialists. Already, traditional disciplinary boundaries are under increasing pressure to adapt their methodologies and find responses to changing socioeconomic realities. Park managers need to recognize the increasing involvement of new stakeholder groupings. 10 First-generation science parks: ‘science push’ In Europe, the very first generation of science parks grew up as early as the beginning of the 1960s. Inspiration was drawn from Stanford University29 and other US universities that made it more trouble-free for academics to become entrepreneurs next door to their university campus. A first-generation science park is an extension of a university into a dedicated neighbouring area that includes incubating facilities for start-up firms, related business services and, as importantly, pathways into new, research-based technology (and know-how) for potential investors and other business persons. Ideally, the park should operate as a science-based technology zone. The typical mode of governance of a first-generation science park is that of university control through a foundation or limited company, created by the university or a related association. The aim is to enforce wider economic goals and objectives for the univer-

Science parks and high-tech clustering 287 sity and energize the business community around it. The innovation philosophy of a first-generation science park is ‘science push’. The many new ideas stemming from research and experimental development (R&D) should be channelled without difficulty to new firms established within, or aggregated around or in, the neighbourhood of the science park. It is a ‘linear approach’ to innovation, which sees scientific results as raw material for innovative activities among the business firms in the science park. For many reasons, during the past half-century, Stanford Research Park, as the first modern science park role model, has gradually changed its original ‘science push’philosophy to engage in more comprehensive university–industry relations. The park’s management has recognized that there are new market signals that need to be considered.For example, the park now offers affiliated companies means (a) to sponsor joint research projects with Stanford faculty and students, and conduct seminars and workshops on campus or in the park that encourage the exchange of technical information; (b) to invite Stanford faculty to join corporate boards and retain faculty as consultants; (c) to obtain access to the university library systems, and consult Stanford’s Office of Technology Licensing; (d) to benefit from the proximity to offer part-time internships to Stanford students and recruit Stanford graduates. Managed by faculty members, each of the Stanford Affiliates programmes offers individual contact between the representatives of companies and the faculty and students in the programme.30 One could say that these new modes of operating the park are more sensitive to the needs of the business community. 11 Second-generation science parks: ‘market pull’ The second-generation science parks are somewhat more recent institutions of science–economy interaction. In most cases, a second-generation science park or technopole remains an extension of a university (or other major R&D facility) into a dedicated high-tech zone. However, the park could also be a totally separate entity, even located relatively far away from a university. The drive and the decisive energy come from businesses, interested in the creation and growth of innovation-based companies. Managers of second-generation science parks respond to such business needs by making available a mix of high-quality facilities in the park, by streamlining the flow of technology and related knowledge, and by advancing and combining value-adding business services, from early incubation of newborn firms to a variety of spin-off and spinon processes of technological significance to already established firms. The typical mode of governance of a second-generation science park is that of a privately owned company which manages the park. Academic or other research representatives may be involved with local or regional policy makers in shaping rules and regulations for the park tenants and in setting the overall principles for the park’s operations. The innovation philosophy of a second-generation science park is ‘demand pull’. It is market-driven to a higher degree than the first-generation science park, which is less concerned with the early exploitation of scientific results and capabilities than with the final stages of the innovation process. Research results and technoscientific findings are regarded as ‘raw materials’ for the park’s innovating firms.31 12 Third-generation science parks: ‘interactive’ glocal flows As already mentioned, among the current evolution of science (and technology) park generations, a more comprehensive park set-up is being implemented. Called ‘third

288

International handbook on industrial policy

generation science parks’ and located within a vibrant urban community or dynamic region, they are perceived as the quintessence of science–industry–government relations, increasingly functional and specialized along with their participation in local, regional and even global innovation activities. At the same time, this science park generation could become a contradiction in terms, as its management is striving to eradicate the fixed spatial boundaries of the park for it to become truly embedded in the urban fabric as a catalyst for innovation.32 Like its two predecessors, a well-functioning third-generation science park is an organization managed by professionals, experts on innovation support. The objective is to increase the wealth of the whole community that the park serves by promoting science–industry–government relations in a number of ways. But it does not stop with that. In addition, by its more central urban location, a fully-fledged third-generation science park is conditioned to offer more broadly-based and more comprehensive varieties of innovation-related services.33 If managed well, the park will ride rather than just react to the flow of technology and related know-how in the surrounding urban and regional settings. It is more able to create communicative linkages among a broader spectrum of entrepreneurial activities than the two other science park generations. In such contexts, the third-generation science park could serve as a model for the way to foster innovations throughout the economy. It becomes an urban catalyst for innovation that could also influence the broader culture of entrepreneurship in the city-region. With modest efforts only, the park’s tenants will be able to interact directly with various layers of urban and regional activities.34 The typical mode of governance of a third-generation science park is that of longterm public–private partnerships, where strategic decisions regarding the park’s operations are agreed through joint decisions, while the day-to-day operations could be carried out by a highly-specialized limited company. Various stakeholder groups, including neighbourhood representatives, could be involved in preparations for major investments. The innovation philosophy of a third-generation science park is cluster-oriented ‘interactive innovation’. It is both ‘science push’ and ‘market pull’. It departs from an underlying ‘linear model’ of innovation, while making more effective use of the network overlay of communications in university–industry–government relations. In a third-generation science park, innovations tend to appear as outcomes of these functional interactions. Embedded in the urban setting, the third-generation science park may be more complicated to govern. The blurring of borders between activities that take place inside and outside the park may cause difficulties for the park’s management. Involving new stakeholders and wider stakeholder communities may require new methods of leadership, more appropriate structures for governance and new ways to stimulate entrepreneurship and foster the growth of business. 13 New competition spaces: both global and local in orientation In the current era of swift globalization, do regional and local economies really matter? In truly global markets for particular goods and related services, and ever stronger international industry networks and multinational corporations, does the physical localization of a firm or a specialized unit of a firm make a genuine difference? Could geography and local resources influence the chances for survival of a science-based or knowledge-intensive

Science parks and high-tech clustering 289 business firm?35 Do science parks, as locally managed innovation environments, really matter to inventive businesses? More and more, the ability of a firm to compete and prosper in the global economy seems to depend on how it gets into and makes use of local innovation environments that foster technology and talent in an open and tolerant atmosphere, yet in a competitive context. It is about agglomeration of talents, well-trained professionals and other human capital. In a recent study of innovation environments of major cities across Europe and the USA, the ability ‘to attract, retain and develop creative people’36 in cities and regions is considered a key factor for the development of new products and new economic activities. Even if the notion of a ‘creative class’ and the methodology are being questioned, the issues raised by a series of comparative examples remain significant to the understanding of the qualities of urban innovation environments. The most spectacular success story of comprehensive and swift regional growth – with a global competitive edge – is the Silicon Valley of the 20th century. Here, in the Valley’s microcosms of innovative environments, there are specialized ‘technological communities’ overcoming borderlines between universities and businesses with relatively open and reciprocal ties. While the industrial history of Silicon Valley must take into account also its military high-tech roots during the first decades of the 20th century, institutional inventions like Stanford’s science park and related initiatives are considered advantageous, even decisive through the accomplishments of microelectronics and information technology, designed and developed in the Valley. ‘With time, the forums for information exchange in the region multiplied.’ So did inter-firm and inter-sector mobility of engineers and other professionals involved in new technology projects. ‘Silicon Valley’s social and professional networks operate as a kind of meta-organization through which engineers, in shifting combinations, organize technological advance.’37 Such combinations or clustering of competencies in Silicon Valley have aroused controversies among analysts as well as policy makers. Changes over time in the Valley’s competence clusters and in its institutional underpinnings (such as the many research universities and R&D centres) in the Bay Area of California have led to fierce debates about what really causes regional dynamics and spurs world-class competitiveness among companies. The importance of policy support, especially the impact of advanced military procurement programmes among Silicon Valley-based firms (as elements in the federal government’s industrial policy), plus generous government research funding, have been central parts of this debate. The issues of entrepreneurship, migration of specialists and the abundance of seed and venture capital are other factors considered particularly important.38 The explanations for regional success converge around the point that dynamism can be attributed to a regional economic ‘ecosystem’, constituted by a large number of technologically diverse, specialized firms and supporting institutions which interact constantly to accelerate innovation.39 Like other regional economies, Silicon Valley has had its ups and downs. For example, a cyclical downturn following the Internet boom of 1998–2000 led to job losses and company restructurings among sectors seen to be among the main drivers of the region’s economy. Hence there is no guarantee, even for an advanced mega-cluster like Silicon Valley, of being continuously successful, even if the technology remains leading-edge, competitiveness on a par with any other region and access to global markets relatively straightforward.

290

International handbook on industrial policy

If a cyclical downturn is driven also by a more profound shift in the composition of products and related services offered by the companies in the region, this would certainly imply challenges to the region’s firms and supporting institutions. 14 New competition spaces: science parks as intermediaries In today’s world of global industries, location seems highly dynamic.40 Innovations by way of new technology and new organization may incite companies to expand geographically, quickly change their supply base, and shift their mode of operation in a particular city or region. Global competition forces firms to relocate their production, since local factor costs may change dramatically. New national and regional policies, such as trade and industry policies, inward investment policy and regional development policy, may strongly influence the framework conditions for business in a particular region or other locality. After all, for most business firms, the value of a location is likely to change over time. One overall trend seems to be clear: in the early development of an industry, or of a new subfield of an existing industry, firms are more likely to colocate their various facilities for experimental development and early product design. When new products undergo extensive experimentation or when they remain uncertain about inventions that are to pass the test on the market, much could be gained by the firm that shares some special services and business resources with others and with supporting institutions. In general, proximity facilitates communication, lowers transaction costs and smooths the absorption of contextual or location-specific know-how. Such industry clustering may lead to sharply increased returns and a positive path dependence on a particular technology or a limited set of its forceful applications.41 Similarly, science–industry–government relations do not dwell on universal infrastructure and services, but on highly specialized, particular services, often dedicated to selected groups of customers and geared towards their innovative activities. In a nutshell, a full-fledged science park stimulates the flow of technology and related knowledge amongst business firms, local universities and R&D, while at the same time the park provides value-added services. In this sense, the science park is a micro-cosmos of larger ‘clusters of competencies’ such as those of a high-tech industrial zone or a particularly resource-rich urban area, able to facilitate the flow of new technology and related know-how and knowledge. 15 Horizontal company integration: networks of firms The creation of the first generation of science parks in North America and West Europe coincided with a relatively new phenomenon. To become more inventive and more effective in exploiting new technology for new products and processes, many large, hightech corporations started to decentralize internally their innovation environments into semi-autonomous business units and functional teams, while relying more on local suppliers and subcontractors rather than trying to do everything themselves. Particularly for large corporations, which previously would strive to be vertically integrated from raw material to final product, it became attractive to outsource parts of the innovation activity to highly specialized small and medium-sized companies, even start-up firms created by former employees of the large company. A dynamic regional economy will often include firms that both compete and collaborate within the region’s technical community. A region with matching firms that operate in clusters, or are connected through inter-firm innovation networks, will be able to foster

Science parks and high-tech clustering 291 parallel experiments with new technologies, new products and applications. A specialized firm could focus on what it does best and leave other firms to pursue a similar strategy. Hence a network-based regional economy could generate and grow a range of technological and organizational alternatives that will reduce the time and costs for each firm in developing new technology applications and encourage reciprocal innovation. In broader terms, and according to modern followers of Alfred Marshall (on industrial districts or clusters as elements in agglomeration economies)42 and Joseph Schumpeter (on entrepreneurial and other innovation activities),43 new economic spaces (such as new ‘development blocks’ or industrial ‘competence clusters’) can be created and stimulated by a series of combined actions, where entrepreneurship and finance belong to the crucial elements for commercial success. Public policies could help create and foster the necessary framework conditions for technological advancements, innovation and competitiveness. In his examinations of the early growth of the biotech industry of northern California, Martin Kenney identifies the processes by which research results at the universities became commodities, professional expertise migrated among research centres and the biotech firms, risk capital accumulated and was invested, legal norms and rules were promulgated and ‘numerous other relations [were] developed to permit the science to become an industry’ with several nodes of actors and stakeholders, operating in networks.44 In a similar study of the early, multifaceted development of the electronics and infocom industries in Silicon Valley, he sees similar patterns of cluster evolution and the creation of regional economic spaces.45 A tentative conclusion could be that some regions and metropolitan areas are likely to generate more or less ‘organically’ new ‘clusters of competencies’ thanks to their ‘institutional thickness’ (see section 3 on the value of a science park). At the same time, other regions have reached similar densities of cooperating and competing firms and institutions by the designing of comprehensive schemes and implementation of policy-induced strategies for target clusters.46 16 High-tech cluster insights: what is needed by policy makers? Policies in support of high-tech clusters exist in all OECD countries and among many less industrialized countries, even in some of the least resourceful countries. The fact is that the notion of a ‘cluster of competencies’ has become elastic in various policy contexts, also because the insights into cluster composition and cluster behaviour are far from being well developed. As for the science park phenomenon, research on clusters is mostly casebased. Systematic international comparisons are few and limited in scope. Even when the statistical grounding of the cluster concept is well-specified, comparative data are limited. There are exceptions in cluster studies carried out in North America, and some countries in Europe and Asia. For example, based on in-depth interviews with key participants in Canadian ‘clusters of competencies’, a national study47 of cluster development across industries, sectors and regions concluded that at least eight factors should be considered of general significance when developing coherent cluster-related industrial policy. These could be summarized as follows: 1. 2.

size and composition of the actual (or potential) cluster; history of the cluster’s evolution, including key events (intentional and nonintentional);

292 3. 4. 5. 6. 7.

International handbook on industrial policy relationships among firms, and between the cluster’s firms and the R&D infrastructure; geographical or spatial structure of these relationships; role of risk capital (especially that of angel investors and venture capitalists); role of social capital, ‘civic entrepreneurs’ and other associative behaviour; and other factors contributing to (or inhibiting) the growth of the cluster.

In recent years, case studies of clusters have become plentiful, but there are few comparative facts and other details about the eight main factors just listed. Much more comparative information and policy-relevant insights are needed for the policy maker to attain a first-class understanding of commonalities as well as differences in cluster evolution. Without insights into comparable experiences, policy making relating to ‘clusters of competencies’ could, in effect, be shallow and inattentive to anything but branding and marketing. 17 Hub cities and their regions: new global and local trends? The early cluster concepts are drawn from economic and human geography. Most definitions of ‘clusters of competencies’, also those by policy makers, recognize the spatial limitations to the cluster. However, clusters seem to be more successful if they are linked across wider spaces through their companies and supporting R&D institutions. Megaclusters, such as those of Silicon Valley’s IT and electronics industries, cut across wider spaces and reach out to ‘clusters of competencies’ in China, Southeast Asia and India. Highly specialized clusters of this type remain locally anchored, but could also be seen as nodes in transcontinental networks and operated through company-managed value chains, in effect enforced by various degrees of institutional support. Accordingly, the knowledge flows in clusters are not necessarily restricted to the local level. Dynamic clusters, through their firms and supporting institutions, typically engage in relations with other, complementary clusters through sharing of knowledge.48 If the carriers of these dedicated knowledge flows will not be business companies, which they typically are, institutions could help in providing effective interfaces also for businesses. Confronted by knowledge flows across the world, a science park could actively assist in facilitating innovation-oriented company-to-company relations in selected ‘sister parks’ in other regions and even on other continents. Evidence is increasing that the specialized knowledge-base of highly specialized agglomerations of firms and supporting institutions could never remain exclusively local. With globalization, more and more of the local cluster activity is becoming an integrated part of complex chains of global production and distribution. This is also reflected in globalization of engineering and product design and other experimental development. Transnational corporations and a growing proportion of small and medium-sized enterprises are increasingly linked to and even embedded in varieties of specialized ‘clusters of competencies’, located across a continent and around the globe. Still, both types of firms use their presence in a local cluster to gain access to specialized bodies of knowledge, created by the local R&D institutions, or tap into a particular skill-set or unique blend of technical know-how among cluster-related firms.49 For example, the medical device industry is a highly-competitive global business and, yet, an industry that relies on a limited number of advanced, highly-specialized city nodes,

Science parks and high-tech clustering 293 which develop original designs of world-class products accessible to hospital clinics across the continents. According to Tim Sturgeon and his colleagues,50 for these network-based firms to remain competitive, there will be further demands for ‘modular forms’ of management of global value chains, where modularization of inputs into a manufacturing, service or innovation process is an advanced tool for cross-sector or cross-border integration. For John Hagel and John Seely Brown, such modular designs and related management techniques are keys to improved effectiveness of distributed networks of highly-specialized companies that operate across borders to create more inventive products and processes. A challenge for the network-based company will be what they call ‘productive friction’. The management needs to find techniques to accelerate capability building by creating suitable settings for people with diverse, yet appropriate, specializations to resolve innovation issues creatively.51 They claim that ‘productive friction’ may be an advantage, if addressed as a management issue, since friction can help meet the difficult challenges of building capability across multiple enterprises. The changing landscape of global business will amplify these challenges – and the possible rewards for the companies involved. The innovative performance of the individual company could improve rapidly, but so will that of their partners. Companies that master the techniques of ‘productive friction’ for learning with and from each other will become much more attractive as partners because they will deliver more value.52 When setting the stage for production and distribution of new products, there will be a continuous need to adapt these products and related inputs to the management challenges of global value chains. Such modularization is driven by factors such as (a) the complexity of information and knowledge transfer required to sustain a particular transaction across sectors and distances, particularly with respect to product and process specifications; (b) the extent to which this information and knowledge can be codified and, therefore, transmitted efficiently and without transaction-specific investment between the parties to the transaction; and (c) the capabilities of (actual and potential) suppliers in relation to the requirements of the transaction. Hence, what typically gets worked out in globally connected, knowledge-intensive clusters of cooperating companies is the codification schemes, required to create and manage spatially dispersed, yet tightly integrated, production and distribution systems. For highly-inventive business activities, on the other hand, interactions between lead firms and key suppliers would typically resist codification into modules made to fit smoothly into cross-border value chains. Nonetheless, some modularization and other codification will be needed also in the processes of original product design and experimental development. More and more, these innovation processes will take place in cities with special design resources, well connected to other similar cities across the world. In Europe, for example, the major design hubs for garments, furniture and household equipment operate as ‘linked innovation environments’. Cities with advanced education and training programmes for their designers and with effective branding through design fairs, exhibitions and other similar events, also tend to attract more designers, design studios and client companies looking for design services. There tends to be a cumulative effect due to ‘clustering of competencies’. Production of high value-added or low-volume products may also take place in such specialized ‘node locations’,53 creating dynamic regional economies of both scale and scope.

294

International handbook on industrial policy

18 Hub cities and the global flows of knowledge Today’s most advanced hub cities are locations that effectively serve companies and their supporting institutions both locally and as ‘glocal’ bridges between local innovation environments and the global economy.54 They have come to function less as self-contained units of production and innovation and more as nodes in global networks. Hub cities operate less as stand-alone ‘innovation clusters’ and ‘industrial districts’ and more as points of connection and, in particular, economic transaction points that enable, complement and enforce linkages between firms that operate as both local and global economic entities. In the context of city hubs, science parks, technology service centres and other such institutionally supported innovation environments could lubricate the channels of communication, help secure the flow of knowledge and promote organizational intelligence functions and a variety of new applications of scientific, technological and other knowledge. When successful, these innovation environments tend to create and sustain densities of the global knowledge flows while brokering specialized knowledge and know-how. In such a ‘distanciated’ economic organization, characterized by the search for new ideas and inventions and by the promotion of entrepreneurship and innovation,55 hub cities do represent an important globally oriented channel for these flows of knowledge. Accordingly, hub cities could be perceived as prime sites for economic circulation as well as institutional settings for advanced innovation. By involving firms, institutions and other stakeholders, hubs may be able to brand their regional ‘identities’ successfully as major contributors to economically effective links between the local and the global. By so doing – given their relatively multifaceted and sophisticated competitive resources (including innovative capabilities in and among firms) on location and by way of global-scale networks – cities could project themselves strategically as innovative ‘design hubs’, ‘financial hubs’, ‘media hubs’, ‘biomedical hubs’ and so on, thereby creating virtual circles that positively influence their global visibility, improve communication links and attract investors. At the same time this branding activity is also an opportunity to identify, negotiate and communicate the ‘hub values’ among key stakeholders. If effective, the brand is a consistent portrayal of the hub city’s identity, culture, economic and other opportunities as well as other profiling of assets and capabilities. Ideally, the branding is for both internal and external benefits. The bridging of the local and the global is the unique quality of the hub, reflecting the true transition not only to a more knowledge-based economy and society, but also to a new mode of the interplay of local and global communities of interest. Notes 1. 2. 3.

4.

Gower, S. and F. Harris (1995), ‘Evaluating British science parks as property investment opportunities’, Journal of Property Valuation & Investment, 14 (2), 1996, 24–37. There is a variety of assessments of science park projects that point to ‘secondary effects’ of these investments only (with regard to their main, declared objectives). Phillips, Su-Ann Mae and Henry Wai-chung Yeung (2003), ‘A place for R&D? The Singapore science park’, Urban Studies, 40 (4), 707–32. Gertler, M.S., David Wolfe and D. Garkut (2000), ‘No place like home? The embeddedness of innovation in a regional economy’, Review of International Political Economy, 7(4), 688–718. Bunnell, T.G. and N.M. Coe (2001), ‘Spaces and scales of innovation’, Progress in Human Geography, 25 (4), 569–90; see also Sternberg, R. and O. Arndt (2001), ‘The firm or the region: what determines the innovation behaviour of European firms?’, Economic Geography, 77 (4), 364–82. Maskell, Peter and Mark Lorenzen (2004), ‘The cluster as market organisation’, Urban Studies, 41 (5/6), 991–1009.

Science parks and high-tech clustering 295 5. 6. 7.

8.

9. 10. 11. 12. 13. 14. 15. 16.

17. 18. 19. 20.

21.

Definition confirmed by the IASP International Board, 6 February 2002 (www.iasp.ws). Massey, D., P. Qunitas and D. Wield (1992), High-tech Fantasies: Science Parks in Society, Science and Space, London: Routledge, p.14. Graham, S. and S. Marvin (2001), Splintering Urbanism: Networked Infrastructures, Technological Mobilities, and the Urban Condition, London: Routledge; Graham, S. and S. Marvin (1996), Telecommunications and the City, London: Routledge. Cf. also Castells, M. (1996), The Rise of the Network Society, Malden, MA: Blackwell; and Castells, M. (2001), ‘The geography of the Internet: networked places’, in his The Internet Galaxy: Reflections on the Internet, Business, and Society, Oxford: Oxford University Press, pp.238–41. Cf.McKendrick,D.(1998),‘Dispersedconcentration:industrylocationandglobalizationinharddiskdrives’, SanDiego,Universityof CaliforniaatSanDiego,InformationStorageIndustryCenter(Reportno.3);quoted in McKendrick, D.G., R.F. Doner and S. Haggard (2000), From Silicon Valley to Singapore: Location and CompetitiveAdvantage in the Hard Disk Drive Industry,Stanford: Stanford University Press, p.46. Wolfe, D.A. and M.S. Gertler (2004), ‘Clusters from the inside and out: Local dynamics and global linkages’, Urban Studies, 41 (5/6), 1071–93. Breschi, S. and F. Malerba (2001), ‘The geography of innovation and economic clustering: some introductory notes’, Industrial and Corporate Change, 10 (4), 817–33. Phillips, Su-Ann Mae and H.W.C. Yeung (2003), ‘A place for R&D? the Singapore science park’, Urban Studies, 40 (4), 707–32. Hippel, Eric von (1987), ‘Cooperation between rivals: informal know-how trading’, Research Policy, 16, 291–302. Patton, Donald and Martin Kenney (2003), ‘Innovation and social capital in Silicon Valley’, University of California (BRIE), working paper no. 155. Amin, Ash and Nigel Thrift (eds) (1994), Globalisation, Institutions, and Regional Development in Europe, Oxford: Oxford University Press, pp.14–15. Castells, Manuel (1996), The Rise of the Network Society. The Information Age: Economy, Society and Culture, vol. one, Malden, MA: Blackwell; especially ‘The Space of Flows’, ch. 6. Williamson, Oliver E. (1981), ‘The economics of organization: the transaction cost approach’, American Journal of Sociology, 87 (3), 548–77; Williamson, Oliver E. (1991), ‘Comparative economic organizations: the analysis of discrete structural alternatives’, Administrative Science Quarterly, 36, 269–96. See also Coase, Ronald H. (1937), ‘The nature of the firm’, Economica, 4 (1937), 386–405. Hagel, John and John Seely Brown (2005), ‘Productive friction: how difficult business partnerships can accelarate innovation’, Harvard Business Review, 83 (2), 83–91. Hagel, John and John Seely Brown (2005), ‘Productive friction: how difficult business partnerships can accelarate innovation’, Harvard Business Review, 83 (2), 83–91. Hagel, John and John Seely Brown (2005), The Only Sustainable Edge, Cambridge, MA: Harvard Business School Press, p.157. Link, A.N. and K.R. Link (2003), ‘On the growth of U.S. science parks’, Journal of Technology Transfer, 28 (1), 81–5; Link, A.N. and J.T. Scott, ‘U.S. science parks: the diffusion of an innovation and its effects on the academic missions of universities’, Greensboro: University of North Carolina (working paper) is an example of a systematic, comparative study. In addition, there are a number of historical studies of the formation and/or growth of science parks: Castells, Manuel (ed.) (1985), High Technology, Space and Society, London: Sage; Castells, M. and P. Hall (1994), Technopoles of the World: The Making of the 21st Century Industrial Complex, London: Routledge; Saxenian, Annalee (1994), Regional Advantage: Culture and Competition in Silicon Valley and Route 128, Cambridge: Harvard University Press; Luger, M.I. and H.A. Goldstein (1991), Technology in the Garden, Chapel Hill: University of North Carolina Press; Gibb, M.J. (1985), Science Parks and Innovation Centres: Their Economic and Social Impact, Elsevier: Amsterdam; Grayson, L. (1993), Science Parks: An Experiment in High Technology Transfer, London: British Library Board; Löfsten, H. and P. Lindelöf (2002), ‘Science parks and the growth of new technology-based firms: academic–industry links, innovation and markets’, Research Policy, 31, 859–76; Chordà, I.M. (1996), ‘Towards the maturity state: an insight into the performance of French technopoles’, Technovation, 16, 143–52. See Monck, C.S.P., R.B. Porter, P. Quintas, D.J. Stoney and P. Wynarczyk (1988), Science Parks and The Growth of High Technology Firms, North Ryde, New South Wales: Croom Helm; Sternberg, R. (1995), ‘Supporting peripheral economies or industrial policy in favour of national growth? An empirically based analysis of goal achievement of the Japanese Technopolis Program’, Environment and Planning, 12, 425–39; Reynolds, Storey and Westhead (1994), ‘Cross-national comparisons of the variation in new firm formation rates: an editiorial overview’, Regional Studies, 28 (4), 443–56; Cowling and Westhead (1995), ‘Employment change and independent owner-managed high technology firms in Great Britain’, Small Business Economics, 7 (2), 111–40; Cowling, Storey and Westhead (1995), The management and performance of family business in the UK’, Stoy Centre for Family Business, London; Westhead, Paul, Stephen Batstone and Frank Martin (2000), ‘Technology-based firms located on science parks: the applicability of

296

22. 23.

24. 25. 26. 27. 28. 29. 30. 31. 32.

33. 34. 35. 36.

37. 38.

39. 40. 41. 42. 43.

International handbook on industrial policy Bullock’s soft–hard model’, Enterprise & Innovation Management Studies, 1 (2), 107–39; Phana, Phillip H., Donald S. Siegel and Mike Wright (2005), ‘Science parks and incubators: observations, synthesis and future research’, Journal of Business Venturing, 20, 165–82. Wessner, C.W. (ed.) (2002), Capitalizing on New Needs and New Opportunities: Government–Industry Partnerships in Biotechnology and Information Technologies, Washington, DC: National Academy of Sciences. Phana, P.H., D.S. Siegel and Mike Wright (2005), ‘Science parks and incubators: observations, synthesis and future research’, Journal of Business Venturing, 20, 165–82. Fred Terman’s pioneering role is excellently summarized by AnnaLee Saxenian in her paper prepared for the inaugural symposium on The Inventor and the Innovative Society, The Lemelson Center for the Study of Invention and Innovation, National Museum of American History, Smithsonian Institution, 10–11 November 1995. Blakeslee, Sandra (1977, 1995), ‘Want to develop a world center of innovative technology? Its simple. Get yourself a Fred Terman’, The Stanford Observer, pp.3, 8, quoted from Saxenian, ibid. (1995). Link, A.N and K.R. Link (2003), ‘On the growth of U.S. science parks’, Journal of Technology Transfer, 28 (1), 81–5. ‘R&D infrastructure versus science and technology parks: reinforcement or compensation? Benchmarking innovation factors in Germany’, BAW Institute for Economic Research, Bremen (2004). Haselmayer, S. (2004), ‘Why science and technology parks go urban: towards embedded innovation environments’, Urbanistica Informazioni. Graham, S. and S. Marvin (2003), ‘The city as sociotechnical process’, Splintering Urbanism, London: Routledge. Kenney, Martin and Urs von Burg (2000), ‘Institutions and economies: creating Silicon Valley’ in M. Kenney (ed.), Understanding Silicon Valley: Anatomy of an Innovative Region, Stanford: Stanford University Press, pp.218–40. This and previous sections on Stanford University are based on information processed for the author by Charlotte Fung Korff (2005). Saxenian, AnnaLee (1994), Regional Advantage: Culture and Competition in Silicon Valley and Route 128, Cambridge, MA: Harvard University Press. Henton, Douglas (2001), ‘Lessons from Silicon Valley: governance in a global city-region’, in Allen J. Scott (ed.), Global City-Regions: Trends, Theory, Policy, Oxford: Oxford University Press, pp.391–400; Kenney, Martin and Urs von Burg (2002), ‘Paths and regions: the creation and growth of Silicon Valley’, in R. Garud and P. Karnoe (eds), Path Creation and Path Dependence, New York: Lawrence Erlbaum Associates, pp.479–513; Rosenberg, David (2002), Cloning Silicon Valley: The Next Generation High-Tech Hotspots, London: Reuters/Pearson Education. For a comprehensive list of advanced business services to innovation, see Colin Barrow (2001), Incubators: A Realist’s Guide to the World’s New Business Accelerators, Chichester: Wiley. Castells, Manuel and Peter Hall (1994), Technopoles of the World: The Making of 21st Century Industrial Complexes, London: Routledge. Lomi, A. (1995), ‘The population ecology of organizational founding: location dependence and unobserved heterogeneity’, Administrative Science Quarterly, 40, 111–14. Florida Richard and Irene Tinagli (2004), ‘Europe in the creative age’, Carnegie-Mellon Software Industry Center, Pittsburgh, research report, p.5. Cf. also Florida, Richard (2002a), ‘The economic geography of talent’, Annals of the Association of American Geographers, 92, 743–55; Florida, R. (2002b), The Rise of the Creative Class, New York: Basic Books. Both quotes from AnnaLee Saxenian in her paper prepared for the inaugural symposium on The Inventor and the Innovative Society, The Lemelson Center for the Study of Invention and Innovation, National Museum of American History, Smithsonian Institution, 10–11 November 1995. Kenney, M. (ed.) (2000), Understanding Silicon Valley: The Anatomy of an Entrepreneurial Region, Stanford, CA: Stanford University Press; Wolfe, D.A. (ed.) (2003), Clusters Old and New: The Transition to a Knowledge Economy in Canada’s Regions, Kingston: School of Policy Studies, Queen’s University, pp.161–85; Lee, C.-M., W.F. Miller, M.G. Hancock and H.S. Rowen (eds) (2000), The Silicon Valley Edge: A Habitat for Innovation and Entrepreneurship, Stanford, CA: Stanford University Press. Wolfe, D.A. and M.S. Gertler (2004), ‘Clusters from the inside and out: Local dynamics and global linkages’, Urban Studies, 41 (5/6), 1071–93. McKendrick, D.G., R.F. Doner and S. Haggard (2000), From Silicon Valley to Singapore: Location and Competitive Advantage in the Hard Disk Drive Industry, Stanford, Stanford University Press, pp.37–65. Arthur, W.B. (1994), Increasing Returns and Path Dependence in the Economy, Ann Arbor: University of Michigan Press. Marshall, Alfred (1890), Principles of Economics, London: Macmillan. Cf., for example, a second-generation pioneer such as E. Dahmén (1988), ‘ “Development Blocks” in industrial economics’, Scandinavian Economic History Review, 1, 3–14; Dahmén, E. (1986), ‘Schumpeterian

Science parks and high-tech clustering 297

44. 45.

46. 47. 48. 49. 50. 51. 52. 53. 54. 55.

dynamics: some methodological notes’, in R. Day and G. Eliasson (eds), The Dynamics of Market Economies, Amsterdam: North Holland, pp.181–97; Dahmén, E. (1988), ‘Entrepreneurial activity, banking and finance: historical aspects and theoretical suggestions’, The Industrial Institute for Economic and Social Research (IUI), Stockholm. Kenney, M. (1998), ‘Biotechnology and the creation of a new economic space’, in A. Thackray (ed.), Private Science: Biotechnology and the Rise of the Molecular Sciences, Philadelphia: University of Pennsylvania Press, pp.131–43. Kenney, M. and U. v. Burg (2000), ‘Paths and regions: the creation and growth of Silicon Valley’, in R. Garud and P. Karnoe (eds), Path Creation and Path Dependence, New York: Lawrence Erlbaum Associates, pp.479–513. Cf. also Kenney, M. and U. v. Burg (2000), ‘Institutions and economies: Creating Silicon Valley’, in M. Kenney (ed.), Understanding Silicon Valley: Anatomy of an Innovative Region, Stanford: Stanford Business Books, pp.218–40. A pioneer in advanced cluster analysis and an early advisor in cluster development schemes is Chris DeBresson. Cf. his DeBresson, C. (1989), ‘Breeding innovation clusters: a source of dynamic development’, World Development, 17 (1), 1–16. Reported by David Wolfe at the 2003 Copenhagen DRUID conference. Cf. also Wolfe, D.A. and M.S. Gertler (2004), ‘Clusters from the inside and out: local dynamics and global linkages’, Urban Studies, 41 (5/6), 1071–93, p.1083. Batheld, H., A. Malmberg and P. Maskell (2002), ‘Clusters and knowledge: local buzz, global pipelines and the process of knowledge creation’, Copenhagen: DRUID Conference working paper no. 02–12. Wolfe, David A. and M.S. Gertler (2004), ‘Clusters from the inside and out: local dynamics and global linkages’, Urban Studies, 41 (5/6), 1071–93. Gereffi, G., J. Humphrey and T. Sturgeon (2004), ‘The governance of global value chains’, Review of International Political Economy. Hagel, John and John Seely Brown (2005), ‘Productive friction: how difficult business partnerships can accelarate innovation’, Harvard Business Review, 83 (2), 83–91. See also Hagel, John and John Seely Brown (2005), The Only Sustainable Edge, Cambridge, MA: Harvard Business School Press, pp.22–6. Sturgeon, T. (2002), ‘Modular production networks: a new American model of industrial organization’, Industrial and Corporate Change, 11 (3), 451–96; Sturgeon, T. (2003), ‘What really goes on in Silicon Valley? Spatial clustering dispersal in modular production networks’, Journal of Economic Geography, 3, 199–225. Sassen, Saskia (2000), Cities in a World Economy, 2nd edn, Boston: Pine Forge Press; Sassen, Saskia (ed.) (2002), Global Networks, Linked Cities, New York and London: Routledge; Scott, Allen J. (ed.) (2001), Global City-Regions: Trends, Theory, Policy, Oxford: Oxford University Press. Amin, A. and N. Thrift (2002), Cities: Reimagining the Urban, Cambridge: Polity Press, esp. pp.51–77.

15 Technology policy and social policy: how industrial policy applies to health Daniele Paci and Stuart O. Schweitzer

Introduction The purpose of this chapter is to provide a framework of analysis in which industrial policy and health are related to one another. This appears to be an ambitious goal because traditionally they are seen as two completely unrelated worlds. Industrial policy, often designed to promote particular industries, has its own objectives and its own instruments and it is implemented by specific agencies and authorities (antitrust, Ministry of Industry, European Commission and so on), while health policy usually has different objectives and instruments and involves different policy makers and authorities (NIH, FDA, WHO and so on). However, we argue that a convergence between these two concepts could be both possible and potentially fruitful. Following Di Tommaso and Schweitzer (2005), health can be seen as one of the largest industries in most industrialized countries, measured in terms of expenditures, employment and output. This new perspective has important policy implications: it suggests a rethinking of the public policy tools and objectives to be applied in a sector which, over the past few decades, has been strongly influenced by increasing budget constraints. Why should (or should not) industrial promotion be applied to health-related sectors? The answer lies in the essence of industrial policy itself. Therefore we will provide a tentative analysis of a possible definition of industrial policy and we will describe the main features of the health industry sectors, which make them a likely target of public intervention of industrial promotion. Then we will refer to a particular sector of the health system (pharmaceuticals) to analyse the different health and industrial policies affecting such industry, discussing their intended and unintended effects, underlining the point of convergence but also the unavoidable conflicts between health and industrial/economic objectives of pharmaceutical policies. Differences in priority setting may not be easy to overcome, but they can be minimized by making these issues explicit and by analysing and openly discussing their causes and effects. The chapter is organized as follows. In the first section, we will outline the main characteristics of health policy. The second section will try to examine convergence and divergence between industrial and health policy objectives, emphasizing the effects of both kinds of instruments. In the third section we will describe the pharmaceutical industry, which provides an excellent example of the intertwined role of industrial and health policy. Indeed, the fourth and the fifth sections are devoted to the explanation and discussion of some health policies and industrial policies, respectively, in that sector. In the final section we will examine some issues arising from the previous discussion. We have identified five controversial areas in which we think health and industrial policy are closely linked. 298

Technology policy and social policy 299 1 Towards a definition of health policy We can define the goal of health policy as providing the maximum level of health to a population (thus maximizing healthcare outcomes1) subject to budget constraints. The most commonly quoted definition of health is the one provided by the WHO: ‘Health is a state of complete physical, mental and social well-being and not merely the absence of disease’ (WHO, 1948). This definition is particularly encompassing and helps us to understand that health affects several spheres of our life, including social interactions.2 Starting from this definition it appears clearly that health has much to do with economics. There is a remarkable body of literature on the importance of health in the process of economic growth and development. Health is usually considered as a form of human capital (Becker, 1993), having an instrumental role in development, but also becoming part of the definition of development itself if we consider that it deals with the living conditions and the well-being of people (Sen, 1999). But can we consider health as an economic good? Why should one care about budget constraints? Health is perceived as a matter of ‘life and death’, but this view can be misleading or inconsistent with constraints imposed by a world of limited resources. First, health is more about improving quality of life and not only about being alive or not. Therefore health demand is not determined only by needs, but also by preferences, which include considerations of the price of health. Secondly, even if we consider a lifesaving treatment, we should be aware that there could be alternative treatments, and patients choose also by comparing the relative prices (taking into account also invasiveness, time consumption, risks and probability of success) of those substitute interventions. Since the formulation of the Grossman model of the so-called ‘health production function’ (Grossman, 1972), economists have started to think of health as something that each individual can ‘produce’. Even if one accepts the idea that a minimum level of health is a fundamental right of human beings (as stated by the UN Charter of Human Rights and by some commentators),3 the production of health requires resources, therefore health necessarily implies an assessment of the allocation of resources, which are obviously scarce in a finite world. Thus there is a huge debate on the public nature of ‘good’ health. Some authors consider that health is a completely private good. However, there is evidence of healthrelated externalities. A person being healthy has a positive value for the entire society, since, for example, it reduces the risk of transmission of infectious diseases. However, health cannot be considered a pure public good, as it is both excludable and rival, at least in principle. The issues regarding health are even more complex, since health depends in part on individual behaviour. It is often argued that a typical health function depends on two main sets of variables; one set is composed of healthcare system-related variables (number of physicians, availability of drugs, quality of care and so on); the other one is made up of ‘environmental factors’, which is a simplifying term to indicate non-medical, social, economic and physical conditions (income, education, lifestyle, air quality and so on). It is important to stress that there are variables over which people have some control (utilization of medical services, education and life style) as opposed to others that individuals cannot modify with their behaviour. Therefore a huge debate arises over what is the limit of the burden of health costs on society (through public intervention or insurance coverage) when people can deliberately choose to adopt risky behaviour, such as smoking.

300

International handbook on industrial policy

The second half of the 19th century witnessed the beginning of the state’s involvement in the organization and provision of healthcare services. In 1945, thanks to Beveridge and Bevan, the National Health System was established in the UK. Since then, developed countries have begun to put the issue of national protection from health-related risks at the top of their agenda.4 Even in the United States, where there is no National Health System, the intervention of the public sector in the healthcare sector is important. In a narrow sense, the government intervention in the field of healthcare can be described in the light of the ‘public interest’ view (Feldstein, 2001). This view considers legislation as a means to serve the public interest. According to its proponents government has two main objectives: to improve market efficiency and to redistribute resources in a more equitable manner.5 The instruments to achieve these goals are several, but it is possible to identify three main groups of them: expenditures, taxation and regulation. These policy instruments can be applied to either the demand side or the supply side of the market (Barr, 1998; Feldstein, 2002). It is possible to articulate these two broad objectives in four intertwined principles that health policy attempts to satisfy: 1. 2.

3. 4.

efficiency: the best value of services should be delivered for a given amount of resources; equity: all should have access to health services regardless of income and location of residence and provisions are made so that different parts of the population achieve similar health outcomes regardless of life styles, ethnicity or geographical location; choice: patients should have relative freedom when choosing their providers and providers should be responsive to patients’ needs; quality: all necessary services should be provided in the best possible way that would assure consumer satisfaction.

The large majority of governments in the world subscribe to all the above objectives and list them as broad policy priorities. Nevertheless, a closer look makes it clear that the above health policy objectives may contradict each other. For example, it is almost impossible to have a cost-effective system that would guarantee the widest choice as well as satisfy all equity considerations. Therefore compromises have to be reached in order to guarantee an overall acceptable level of healthcare provision. Countries have different priorities: for example, the US healthcare system puts less emphasis on equity, and much more on choice and quality, while European countries usually make the opposite choice. However, the worldwide tendency is to consider cost as the key attribute that, at least partly, determines all the others. In fact, this has become the epicentre of health policy making in all European Union countries, in the USA, in Japan and in all industrialized and developing economies, as the share of gross domestic product (GDP) spent on health has been steadily increasing since the 1970s. Figure 15.1 shows the percentage of GDP that some industrialized countries spend on health. The graph illustrates clearly the point that a large share of the GDP in industrialized countries is devoted to health expenditure, even if notable differences remain between countries, in particular between the G8 countries and the newly industrialized countries (except for Brazil, which spends relatively more than the UK and Japan), in which the population (fertility rates, dependency ratios and so on) and the epidemiological situation

Technology policy and social policy 301 14.6

USA 10.9

Germany France

9.7

Canada

9.6 8.5

Italy

7.9

Brazil Japan

7.9

United Kingdom

7.7

Russian Federation

6.2

India

6.1

China

5.8

Source: WHO (2005).

Figure 15.1 Total (public
private) health expenditures as a percentage of GDP (2002) are still similar to those of the developing world. The United States has the most expensive healthcare system, consuming 14 per cent of its GDP, followed by countries such as Germany and France, whose expenditure are around 10 per cent of GDP. This exerts pressure on health systems, the majority of which are publicly funded and provided. Even in the United States, where nearly half of national health expenditures is funded through private health insurance firms (OECD, 2003), public expenditures are very large (Medicare programme for the elderly and Medicaid programmes for the indigent).6 Therefore health policy and its multidimensional objectives have actually often collapsed into what Di Tommaso and Schweitzer (2005) called the traditional ‘Health Model’,7 with the unique policy imperative of cost containment. 2 Health policy and industrial policy: health as an industry Health can also be seen in a different perspective. Indeed, health is linked to a set of service and manufacturing sectors, which together form one of the largest industries in most industrialized countries in terms of measured expenditures, employment and output. This alternative view of health systems, called ‘the Health Industry Model’ by Di Tommaso and Schweitzer (2005), considers other objectives than health status. This other view suggests that the health system also produces benefits for an economy in terms of economic externalities (mainly through investments in research and development), and it can produce services for the world market and hence produce valuable export earnings. The Health Industry Model views the health industry’s outputs in the framework of a multiproduct production function, where health providers and health manufacturers produce health and other outputs.8 The health sector is made up of a significant number of interconnected industries that produce both health services and manufactured goods, including pharmaceuticals and medical equipment. According to the proponents of this model, the structure of the health industry contains three clusters of industries: healthcare providers (hospitals, nursing homes, ambulatory care clinics, community-based medical practices, home-based care programmes and so on), healthcare financiers (public or

302

International handbook on industrial policy

private) and health manufacturers (pharmaceutical industry, diagnostic and therapeutic medical equipment producers and so on). Therefore policy towards health could be reconsidered, in order to take account of this broad concept of the health industry, as a system of intertwined actors and thus also industrial policy, can be implemented in this field.9 In addition to maintaining a population’s health status at some acceptable level, the health system is a major generator of scientific knowledge through its research and development functions. Increases in health status not only produce social welfare, but also increase the productive capacity of a nation’s workforce both in developing and in industrialized countries (WHO, 2001; IMF, 2004). However, the benefits of a strong health industry are not limited to direct health effects. Scientific knowledge and productive capacity generate benefits throughout the economy. These benefits are usually described as ‘public goods’: they are likely to be underproduced in a free private market since no private firm has a sufficient incentive to invest in knowledge or worker health, given that the benefits extend far beyond its boundaries. Moreover, experiences around the world indicate that there could be additional justifications based on strategic objectives (such as employment and national competitiveness). While in the past strategic sectors have included automobiles and steel, nowadays the sectors that are considered strategic in all industrialized countries (and some emerging countries) include new science-based industries, such as biotechnology and pharmaceuticals. Finally, high-technology industries typically start out small and usually need support, at least in their ‘infancy’ stage. With sufficient protection and public investment, these industries have the possibility to grow and to achieve a critical mass that allows them to be self-sustaining and competitive in the international market. Examples include biotechnology and the new field of nanotechnology. Understanding the multi-objective role of the health industry explains, in part, why policy decisions affecting the size of the health industry are often so debated. There is a conflict between those who see the health industry as responsible only for producing an acceptable level of health (at minimum cost), and those who see broader benefits of a dynamic, scientifically sophisticated health industry. Policies in the field of healthcare were therefore linked to the above-mentioned efficiency criterion in order to contain health expenditures, since in most developed countries healthcare spending generally grows at a rate faster than the overall economy (OECD, 2003). Projections indicate that rates are likely to rise in years ahead because of the ageing of the population and the achieved economic prosperity of some large emerging economies, and because technological innovation provides treatment opportunities not available only a few years ago. The main focus of the following section will be on the pharmaceutical industry. Health policy makers seek to provide good-quality, safe and effective drugs for patients at a cost affordable to patients and to the health system. The issue of the cost of medicines is crucial and has acquired tremendous importance over the past decades, not least because the cost of medicine has been rising steadily as a proportion of healthcare costs. Therefore, health policy makers often aim to control the total cost of pharmaceuticals by imposing price controls either directly or indirectly through a wide assortment of measures. At the same time, the pharmaceutical industry is widely recognized to be one of the most important high-technology industries, creating value added for national economies,

Technology policy and social policy 303 contributing to exports, requiring highly trained employees because of its high R&D content, generating highly paid and stable jobs and contributing to economic growth. Because of its special nature, many countries support the pharmaceutical industry by specific policy actions, which range from (relative) freedom in the pricing of medicines to specific measures providing tax incentives for inward investment, to initiating policy actions that would facilitate the pursuit by the industry of new R&D and which would potentially lead to the discovery of new drugs. 3 The pharmaceutical sector Public attention to the function and performance of the pharmaceutical industry is justified by the importance of drugs for health, the share of pharmaceutical spending on total health expenditures, the level of drug prices, the peculiarities of the process of drug discovery and manufacturing (R&D intensity, mergers and acquisition and so on) and by the important contribution of such an industry to the overall economic system. Certain characteristics of the pharmaceutical market suggest that competition and consumer sovereignty may not suffice to produce either adequately low prices or reasonably safe and effective products. Moreover, the special features of the sector create a monopolistic environment where small and innovative companies find it difficult to flourish. Let us review briefly some of these characteristics. ●

●

●

●

●

The three-tier demand side creates important imperfections in the pharmaceutical market (Schweitzer, 1997). The demand for medicines is not fully controlled by the actual consumer (patient) but by the physician who neither pays for the product nor consumes it. The financial costs for the patient are not clear, since in most cases an insurance system (private or public) bears part or all of the direct costs. Effects of new medicines are uncertain; no medicine is safe a priori, but the level of uncertainty can be reduced through regulated clinical trials. Moreover, decision making is often delegated to a public regulatory agency on the assumption that profit incentives and market competition are unlikely to generate a socially optimal level of information or to produce a socially desirable decision by pharmaceutical companies. It has been argued that R&D raises the problem of determining the economically correct level of profit to increase innovation (for instance, the treatment of R&D expenses, the allowance for risk and so on). Transparency is limited: the market for pharmaceutical products is subdivided into therapeutic groups with a high cross-elasticity of supply and demand and which can be divided into at least three groups according to their technological content: in-patent medicines, generics and over-the-counter drugs (OTCs). There are economies of scale in research, production and marketing. The industry is characterized by high fixed costs (R&D and marketing) and negligible marginal costs. This suggests that, to be profitable, companies must aim to sell their products as widely as possible. This may explain the multinational character of the industry, marketing efforts and the tendency towards size expansion through mergers and acquisitions. The resulting supply side is thought to be oligopolistic, as high profits, prices and promotion expenditures can only exist in an industry characterized by monopoly power.

304

International handbook on industrial policy

4 Health policy towards the pharmaceutical industry As we said above, because of the rising costs of healthcare, mostly due to advances in health technologies, the intensity of services provided as well as the limitations of publicly funded systems to meet an ever-increasing demand, governments in virtually all countries have implemented tight cost containment policies. A primary target of cost containment policies has been pharmaceutical expenditure, as one of the most dynamic parts of rising health costs. There is an increasing concern about pharmaceutical expenditures and the reason, in our view, is twofold. First, pharmaceuticals is a key component of health policy and accounts for a sizable and increasing proportion of health expenditure in most countries. The characteristics of the pharmaceutical markets of some selected OECD countries over time are shown in Table 15.1. The United States is a considerable market (despite pharmaceuticals representing a relatively low percentage of the country’s healthcare budget), but markets in France and Japan are also large, approaching 20 per cent of their healthcare budgets in 1998. Second, pharmaceutical spending can be easily identified and directly compared with other parts of the health system, for instance hospital spending, or the total expenditure on salaries of healthcare professionals.10 Regulatory interventions aimed at containing pharmaceutical spending take different forms: they can be directed at patients, physicians and pharmaceutical companies.11 In the first case, one of the instruments used to lower the demand for drugs by patients is the introduction of co-payments and user charges. There is a wide variety of these interventions across the world and little analysis of their effects. Evidence has come largely from the United States. The RAND Health Insurance Experiment is considered the most rigorous experimental study in existence. It demonstrates that the use of healthcare is reduced by any form of cost sharing compared with none, although the reduced utilization has no impact on health status (Newhouse, 1993). However, quasi-experimental studies on low-income populations (Medicaid patients) have demonstrated reduced use of pharmaceuticals from cost sharing, with associated adverse health and expenditure effects (Soumerai et al., 1991, 1994). Another target of policy instruments is physician prescribing behaviour. The introduction of formularies and guidelines is common in most industrialized countries, as are various forms of incentive to consider cost-effectiveness in the prescription of medicines to patients. A large number of countries give incentives for generic substitution. In some cases, budgetary controls are performed.12 Other measures affect directly the pharmaceutical industry. Controls on the drug industry itself focus mostly on price and only recently on cost-effectiveness. Apart from the UK, where prices are regulated indirectly through a profit scheme,13 all EU countries as well as many developing and industrialized countries have a form of price regulation. In setting prices, these countries usually use therapeutic comparators and the price of products in other markets (external reference pricing). A notable exception is represented by the USA, in which drug prices are freely determined by pharmaceutical companies. Countries, especially in the EU market, increasingly reimburse the average EU price within a therapeutic category. These may reduce price variation across markets but may also induce inflation in generic prices and reduced competition, as all therapeutic prices are driven to similar levels. Moreover, price control schemes that do not control volume are incomplete. Whatever the rigour of price control, its effects on pharmaceutical spending

305

Source:

a

OECD (2002).

data refers to 1997; b data not available.

–b 1.3 1.3

–b 14.1 20.3

France Germany Japan

Notes:

0.6 0.9 0.8 1.0

Percent of GDP

8.1 14.7 13.6 9.3

Percent of total health spending

1987

–b 191 185

90 104 107 186

Spending per capita. in US$

20.5 14.2 22.0

9.9 14.2 14.2 8.8

Percent of total health spending

1.5 1.4 1.4

0.8 1.1 1.0 1.1

Percent of GDP

1992

300 260 281

143 152 167 280

Spending per capita. in US$

Spending on pharmaceuticals in selected OECD countries in selected years

Australia New Zealand United Kingdom United States

Country

Table 15.1

18.5 12.7 17.0

11.6 14.4a 15.9a 10.3

Percent of total health spending

1.7 1.3 1.2

1.0 1.1a 1.1a 1.3

Percent of GDP

1998

391 312 295

239 196 236 428

Spending per capita. in US$

306

International handbook on industrial policy

may be dissipated by an increase in volume, which is determined by physicians and influenced by marketing. In countries such as Germany, the Netherlands, Sweden, Denmark and Norway, a decade of reference pricing has resulted in only short-term savings (Ioannides-Demos et al., 2002). When considering the peculiarities of the pharmaceutical sector, it becomes evident that some form of public intervention is necessary in order to underwrite the objectives of health policy. Most agree that minimum standards of quality, safety and efficacy cannot under any circumstances be compromised. The control functions of the public regulatory agencies for pharmaceuticals are justified by these reasons. In the USA, the Food and Drug Administration control and regulate the market for drugs, which cannot be marketed without its approval, involving a process of review of the candidate drugs through three stages of clinical trials. Similar public agencies are present in Europe (EMEA) or at the national level (for example, the Italian Agenzia del Farmaco). Equity and quality, however, affect public policy more broadly. Indeed, the issue of providing high-quality medicines at an affordable cost has critical policy implications since this is where health policy and industrial policy in pharmaceuticals become intermeshed. 5 Industrial policy issues in the pharmaceutical industry Pharmaceuticals is one of the sectors of the economy in which governments usually intervene most prominently. The objectives of government intervention are various, the most important being correction of market failures, fostering innovation and promoting competition. It is important to note that these objectives are strongly intertwined and are not necessarily mutually consistent. For example, correcting market failures in the field of research means also fostering innovation and providing patients with new treatments for their diseases. Promoting competition in the drug market has the main effect of promoting efficiency, but the resulting lower prices also improve access to medicines. Sometimes, as we will see, these policy objectives appear opposed to one another. This is the case, for example, of the tension between the claims of free access to medicines in developing countries and patent protection, which enables firms to recoup R&D investments. We identify two main areas in which industrial policy intervention is and should be of particular importance. The first is the correction of market failures in pharmaceutical R&D: since social benefits exceed private ones, firms are likely to underinvest in research activities without adequate public intervention. This is also related to the issue of pharmaceutical innovation. The second main field of industrial policy is the promotion of efficiency through competition. 5.1 Market failures in pharmaceutical R&D and the promotion of innovation The pharmaceutical industry is a high-tech sector in which research and marketing play a key role. The process of research is particularly long and uncertain. From the discovery of a new potential molecule to the authority approval the process lasts on average 10 years. Less than 1 per cent of molecules examined in pre-clinical trials reach phase III of clinical trials (Schweitzer, 1997; Phrma, 2003). According to recent estimates the average cost of bringing a new drug to the market is about $800mn including the costs of failed trials (Di Masi et al., 2003). This figure, together with the fact that the cost of copying and replicating drugs is negligible, reveals a typical case of market failure in the production of

Technology policy and social policy 307 knowledge.14 According to Arrow (1962), knowledge bears three problems that give rise to market failure: it is not possible to create a market for knowledge, since the producer cannot fully appropriate it; the process of creation of knowledge is dominated by uncertainty; and there are significant economies of scale in the production of knowledge.15 The market for new knowledge is thus likely to fail, because there is no private incentive to invest in producing it, which means engaging in costly and risky R&D activities, unless there were some measures to let the inventor exclude others from the free availability of knowledge. The pragmatic approach in balancing the need for private incentives to invest in R&D and the need for availability of knowledge is to combine two policy instruments: public financing of basic R&D and patent protection. In the USA, for example, the federally funded biomedical research supported by the National Institutes of Health (NIH) plays a vital role in drug development, feeding into the R&D activity of the private pharmaceutical companies, which operate under a regime of patent protection.16 For example, among all anti-cancer drugs that received FDA approval for marketing in the USA as of 1 January 1996, 50 of 77 products (about 65 per cent) benefited from National Cancer Institute-sponsored clinical trials (National Cancer Institution, 1995). The importance of publicly funded research is greater for basic research or in therapeutic areas where there are no sufficient market incentives to invest in. Biotechnology offers an interesting example of the outstanding role of public research. The first biotech discovery dates back to the 1970s, when S. Choen of Stanford and H. Boyer of UC–San Francisco discovered the basic techniques for recombinant DNA, which became the basis for genetic engineering. There is considerable evidence of the central role played by the university in the early development of the industry, as many biotech companies started out as academic spin-offs from university research. Biotech firms also show the tendency to form geographically localized clusters and these agglomeration dynamics have been explained as university-driven, since the proximity to public research centres has been found to be the major factor underlying biotech clustering (Zucker et al., 1999; Schweitzer et al., 2005). This is because biotechnology stands at the frontier of scientific innovation. It is recognized as the field in which basic science and industry are more closely related, with breakthrough commercially valuable innovations coming directly from basic research (Pavitt, 1998; Prevezer, 1997). Biotechnology poses interesting industrial policy issues. Firstly, interventions could be more effective if implemented at a ‘meso’ level’.17 This is an emerging focus of industrial policy that is neither at the level of the overall economy, nor at the level of individual firms: the agglomeration of firms and institutions within a geographic area, usually within a single industry, into high-technology industrial clusters18. Secondly, it appears clearly that the promotion of biotechnology can be realized indirectly, through interventions in research institutes rather than directly in firms. It is often claimed that the Bayh–Dole Act approved by the US Congress in 1980 has played an important role in this process. However, there is contrasting evidence on the final outcome of such policy instruments that allowed universities to retain intellectual property rights on inventions originating from federally funded research. Some scholars are doubtful about the pivotal role of university patents (Nelson, 2003; David, 2004).19 For sure they cannot be seen as the unique means of promoting innovation. Without an adequate level of public funding for basic research these instruments would be meaningless or they could have a very limited effect.20 Thirdly, the patentability of biotech products poses some issues from a legal point of view

308

International handbook on industrial policy

too. A too broad patenting runs a serious risk of blocking subsequent research lines, creating downstream monopolies. Moreover, there are ethical constraints on the use and patenting of genes and genetically modified organisms. We will discuss patent policy, even if not specifically for biotech products, later in this section. The other case of market failure in which government intervention is required is the case of scarce incentives for pursuing research. This could be true both for orphan drugs and also for the so-called ‘tropical diseases’, which affect exclusively low-income populations in specific areas and that do not have any strategic importance.21 In both cases the expected returns are too low to justify investments, because of the low number of cases or because of the low purchasing power of the patients. The Orphan Drug Act in the USA has provided a good stimulus for R&D in an otherwise unexplored field. Similarly, there are forms of public intervention also to promote research on tropical diseases, through donations and public programmes specifically designed for such diseases. Two of the most recent initiatives are the G8 Global Fund and the Neglected Diseases Initiative (NDI). The other policy instrument used to overcome market failure problems in pharmaceutical R&D is the granting of patent rights. Both products and processes can be patented. Patents are especially important to pharmaceutical inventors, allowing them to capture the potential profits from new products, which otherwise could be copied more easily than products whose production processes can be kept secret, or for which the time and relative expense needed to copy the invention are much higher. Because patents normally cover well-defined chemical molecules, substitution around which could require expensive new clinical trials, pharmaceutical product patents are particularly effective in limiting competitive entry into the production of specific new drugs and hence in permitting the producer to hold prices well above production costs. This is the conclusion of a number of studies which attempted to determine different sectoral propensities to patent. For example, Levin et al. (1987) and Cohen et al. (2000) with their surveys showed that firms in general do not value patents very much to support and stimulate innovation. Pharmaceutical companies are the exception. They considered patents as extremely important. Taylor and Silberston (1973) estimated that the absence of patent protection would have reduced R&D investments by nearly 8 per cent in many industrial sectors. In the pharmaceutical industry the estimated reduction was 64 per cent. Finally, Mansfield (1986) calculated that, over the period 1981–3, on average, 14 per cent of new products would not have been invented without patents. For the pharmaceutical industry the percentage was 65 per cent. Industrial policy intervention in the field of research can have effects which go beyond the boundaries of the pharmaceutical market, since there are positive spillovers on other sectors. For example, biotechnology is closely connected with several industries other than life-science (agriculture, new materials, nanotechnology and so on). In addition, there is a question of what actually is innovation. Empirical research underlines the fact that the pharmaceutical industry spends huge amounts on R&D, but only a small proportion of its R&D outlays are spent on research in radically and fundamentally new areas for the discovery of innovative medicines. Figure 15.2 indicates that only 35 per cent of new drugs approved by the FDA during the 1989–2000 period contain new active ingredients and only 15 per cent has considered priority ‘new molecular entities (NMEs)’ by the regulating agency, which means that they provide significant clinical

Technology policy and social policy 309 MOST INNOVATIVE

15%

Priority NMEs

LEAST INNOVATIVE

20%

8%

46%

11%

Standard NMEs

Priority IMDs

Standard IMDs

Other Drugs

New Active Ingredients

Old Active Ingredients

Source: The National Institute for Health Care Management Research and Educational Foundation (2001), on FDA data

Figure 15.2

Distribution of FDA new drugs approvals, 1989–2000 (total  1035 drugs)

improvement on existing drugs. Therefore, some may argue that the outputs of R&D activities are mainly ‘me-too’ drugs, slight modifications of existing products. Finally, there is a debate about what is the best price to reward innovation. It is not unusual to have prices fixed by the government or by the social insurance scheme. In these cases, governments and social insurance funds within the limits of public budgets arbitrarily set aside a certain amount for innovation and profit, even if the manufacturers’ perception of ‘the innovation mark-up’ and ‘the profit rate’ are quite different. Relevant information asymmetries exist in that it is not in the manufacturers’ interest to reveal their total cost structure to governments/insurance funds; rather, they push for the highest possible price given that the former are profit maximizers and the latter social utility maximizers subject to cost minimization. In some countries, such as the UK, Germany and Denmark, free pricing for new and innovative medicines is allowed. This should give the industry an incentive to conduct innovation on their territory. Indeed, the above three countries account for almost two-thirds of the R&D conducted in the European Union. At the same time, however, all three countries have mechanisms in place that would control the overall spending on pharmaceuticals, including reference pricing for off-patent medicines (Germany and Denmark), profit control for the sales to the National Health Service (UK), practice guidelines (UK, Germany), pharmaceutical budgets for physicians (UK, Germany until 1997) and generic promotion (all three). However, a policy focused on price liberalization would not succeed if other conditions were not met. Attracting pharmaceutical investment requires a sound science base, an adequate and receptive financing environment, a culture that promotes university–industry collaboration and fosters entrepreneurship. These issues are as important as pricing and there is a direct relationship between pharmaceutical industrial structure and the availability of these factors in a number countries (USA, UK, Germany, France and Denmark) and a lack of them in most others. 5.2 Promoting competition in the pharmaceutical markets Patents may have negative effects on another traditional field of industrial policy intervention: competition. To address this issue, the US Congress in 1984 approved the Drug Price Competition and Patent Term Restoration Act (commonly known as the

310

International handbook on industrial policy

Hatch–Waxman Act), a complex legislative compromise that made changes in both the drug laws and the patent laws. It provided a streamlined procedure for approving a generic version of a previously approved drug through use of an Abbreviated New Drug Application (ANDA) that is considerably less burdensome than a standard New Drug Application to the authority (in the US case, the Food and Drug Administration – FDA). At the same time, it extends the patent term for new drugs for up to five years to compensate for delays during regulatory review prior to first commercial marketing of the product. To accelerate the introduction of generic versions as soon as drugs go off patent, it relieves generic manufacturers from infringement liability for conducting clinical tests during the patent term. There is significant evidence of the efficacy of this policy intervention even if the success of the market for generics in the USA could also be linked to the relatively high price of prescription drugs. However, the risk of anti-competitive alliances is always present, as in the case of Hoechst and Andrx for the development of the generic version of Cardizem in 1997. Concerns about competition do not refer only to branded versus generic drug producers. In the last decades, the pharmaceutical market has been involved in several anti-trust cases. Anti-competitive strategies are likely to include mergers and acquisitions or cartels and agreements. However, some additional issues arise in the pharmaceutical market. The pharmaceutical industry is characterized by high fixed costs (research and marketing) and very low variable cost (production and replication), which make it similar to a natural monopoly because of significant economies of scale. Therefore larger companies are likely to be more efficient than smaller ones and thus mergers and acquisitions, while unacceptable in other industries, are often seen positively for overall efficiency in the pharmaceutical markets. The most recent mergers include Pfizer-Pharmacia, which created the largest actor in the market for drugs, and Sanofi–Aventis, which aims to become its major competitor in the near future. Another controversial issue is the concern for ‘potential competition’. In fact many big pharma–small biotech mergers and acquisitions may not be particularly relevant in terms of market power and concentration, but they may limit future competition when biotech companies show a high probability of developing new drugs in the same therapeutic field as the other company involved in the operation. For example, in the case of the acquisition of Genetech by Roche in 1990, the Federal Trade Commission applied this concept of ‘innovation market’. Since both firms were engaged in research on anti-AIDS drugs based on human protein C-4, the US authority allowed the operation but required compulsory licences on any drug eventually produced. 6 Health and industrial policy in the pharmaceutical industry: conflicts and convergences Having discussed health policy and industrial policy in the pharmaceutical sector, we can state that some of the most important situations, where health and industrial policy are likely to be intermeshed, are the ones we now discuss. 6.1 Increasing competition and reducing prices This approach is likely to reduce the pharmaceutical expenditures to the payers (public or private) and thus to increase access and affordability, but it may come at the price of a lower

Technology policy and social policy 311 quality, if quality cannot be adequately controlled. This in turn means reduced health benefits. The pharmaceutical manufacturers also argue that this policy reduces innovation because of the loss of profits that would occur to the innovator. There is no direct evidence of negative consequences of competition (in respect of patent protection) for the pace of innovation. However, there can be serious problems if the price of the new drug is set at a level considered too low to refund R&D expenditure. Feldstein (2002), using US data, shows that it is the price of new drugs that drives the growth of prescription drug spending. Two-thirds of the expenditure increase (13 per cent per year) is found to be generated by new drugs (introduced after 1992) over the 1993–8 period. Therefore, according to these data, the most important reason for rising drug expenditure is the replacement of old drugs by newer treatments even if one could question, as we noted above, which medicines are ‘really’ new. The promotion of competition is also linked to the issue of branding. Brand names are valuable for manufacturers, as they allow them to benefit from a good long-term performance and goodwill. They may also provide some benefits to patients, as they help them recognize good quality. But brand names can also be used (abused) to promote unjustified product differentiation, based, not on any real advantage, but only on marketing practices. Market differentiation tends to increase market prices, as suppliers can compete on perceived quality rather than on prices. Furthermore, brand names may also confuse prescribers, making rational drug use more difficult to attain. Thus generic policies (promoting generic substitution) and setting up reference pricing (in fact, reference reimbursing) mechanisms, are likely to conflict with commercial interests, and perhaps with long-term development, owing to their potential negative impacts on R&D investments and transfer of technology. 6.2 Patent protection The legal temporary monopoly position granted to the inventor of a new drug (or a new production process) is valued as an incentive for R&D which in the long term means a higher probability of finding new, more effective and safer drugs. But monopoly will probably mean higher prices (and expenditure) on medicines. Moreover, strong patent systems seem to benefit industrialized countries at the expense of the less developed, because strong patent protection reduces the feasibility of a national industry developing, based on the easiest strategy of copying existing products. The introduction of patent protection in Italy, in 1978, for example, did not result in increasing pharmaceutical R&D expenditure in the country (Scherer and Weisburst, 1995). The experience of some countries (especially in Southern Europe) seems, however, to indicate that lack of patent protection does not automatically ensure expenditure control, nor does it lead to an internationally competitive generic or research industry in the country. 6.3 Differential or equity pricing A common market-friendly proposal to assure access to medicines of low-income populations is differential or equity pricing. This pricing strategy (sometimes also called Ramsey-pricing) consists in setting different prices for different populations according to their ability to pay. It may have a positive impact on the health of the population that enjoys lower prices, namely developing or low-income countries. However, several

312

International handbook on industrial policy

problems arise in this regard. Market segmentation is a necessary condition for this strategy. If it is not possible to impede parallel imports from lower-price countries to the highprice ones, this drug policy will probably be ineffective, resulting in the long run in a single intermediate price in the global market-place. Moreover, the articulated system of ‘external reference pricing’, adopted by many countries, has the unintended effect of maintaining an international level of prices which is not flexible downwards. Donations to developing countries are usually used instead of differential pricing. However, even in this case, some problems may arise. If the continuity of the donation programme is not guaranteed, the authorities may be forced to pay for the drugs when the donation period expires, because of the strong pressures to avoid discontinuity once a treatment has been provided. Furthermore, the perception that a donation is always a good deal for the recipient country that should not be refused may induce authorities to give priority to programmes that may not be cost-effective, if all the costs required are taken into account. 6.4 Enforcing (minimum) quality standards to improve the effectiveness and safety of treatments Quality control and enforcement have their own costs. For a developing country with high health budgets and a weak industry these costs take two forms: opportunity costs (renouncing the satisfaction of other needs) and negative industrial effects (less production, employment, taxes and so on) Developed countries can use health regulations as non-tariff barriers: trying to impose high, Western-level quality standards on developing countries may be seen as a hidden type of protectionism favouring developed countries’ industries. However, the opposite is also true: developing countries’ companies which would be allowed to produce goods with lower quality and safety standards may be accused of unfair competition and trade in the international market. This suggests that all countries should have a uniform level of quality standards. Therefore there could be a greater role for international organizations to recommend appropriate quality standards taking the needs and interests of developing countries also into account, but without forgetting the protection of patients as the ultimate goal. Though regulation by government agencies is the usual way of assuring product safety and efficacy in most countries, there is a private sector model that is also applied in other markets, and could be applied to pharmaceuticals. The problem with pharmaceuticals is asymmetry of information, in that quality differences (in terms of efficacy and safety) of alternative products are not readily apparent. In many cases they are not even apparent to prescribing physicians. But an alternative to drug licensing, by which only drugs that are ‘proved’ safe and effective are allowed to be sold, is ‘certification’, by which drugs are tested and then labelled in such a way that both physicians and patients can verify their suitability. Certification is widely used in consumer goods markets in the USA, as a number of private laboratories exist that are used by manufacturers to validate advertising claims.22 In these cases manufacturers pay the laboratory to test their product and, in return, are allowed to market their products with the respective ‘seal’ of quality, because they recognize that many consumers are more likely to purchase products that have been certified by a reputable laboratory – and that they are willing to pay the additional cost of this quality-control measure. Of course other consumers may be willing to take the risk of buying products that have not been tested, at a lower price.23

Technology policy and social policy 313 This raises another issue, besides asymmetry of information: that consumers differ in their preference for low price, on the one hand, and quality, on the other. Presumably a product that is engineered and produced to the highest standards will be ‘better’ than products that are not. But they are likely to be more expensive, as well. Do all consumers have identical preferences for quality? Evidence suggests they do not. Consumers widely believe that generic drugs are inferior to branded products, but they are often considerably less expensive. Some patients insist on the branded products (and are willing to pay more for them), while others are satisfied with generic versions. No doubt the proportion in either category varies by type of drug. For example, drugs for minor conditions, like aspirin, may have a larger generic market share than drugs for cancer. The point is that all consumers are not alike and some may be willing to forgo some degree of quality in return for a lower price. Therefore a uniform standard imposed by an agency such as the FDA may reduce social welfare. In this case, another regulatory model that certifies quality may be preferable to the licensing agency that assures a uniformly high level of quality. 6.5 Conflicts in the area of financing Industry is usually interested in and supportive of public financing of pharmaceuticals and low out-of-pocket co-payment, because it increases demand and hence sales, by decreasing price elasticity of demand (Schweitzer, 1997; Folland et al., 2001). However, public financing of drugs may make authorities more sensitive to high prices and increased expenditure, and thus more likely to control these variables. Governments are likely to welcome proposals that reduce the pressure on public health budgets. It must however be considered that such a strategy is likely to affect the accessibility to drugs and finally the health status of the poorer sections of the population. Indeed, the effect of co-payment is twofold: on the one hand it reduces expenses on unnecessary drugs, but on the other hand it can limit the access to needed drugs.24 The design of an optimal public–private mix in the financing of pharmaceutical and health remains an empirical question. When does the market fail? Can the government be more efficient than the market in such cases and is it possible to avoid government failures?25 7 Concluding remarks A new view of healthcare as an industry induces one to rethink the traditional concept of industrial policy in this particular field, merging industrial and health policy into a unique framework: two worlds that showed some overlapping areas but that have traditionally been considered as unrelated to one another. This is evident especially in the pharmaceutical industry. Pharmaceuticals is a domain which is subject to variable treatment by different parts of public policy. Health policy makers, facing tight budgets allocated on the basis of cabinet consensus and other needs in the economy, are interested, among other things, in containing the expanding cost of medicines and, therefore, control their prices in many countries. Almost all countries in the world have implemented pharmaceutical cost containment policies of very different amplitudes and intensity, even if a very few of them have succeeded in curbing pharmaceutical expenditure more than temporarily. Cost containment in the pharmaceutical sector is not simply a matter of balancing the health budget. It has implications for the development of a high-technology industry, with very high value added, highly paid jobs, and high growth.

314

International handbook on industrial policy

We can observe contrasting objectives because of the conflicting interests of health policy (comprising health policy makers) and industrial policy (comprising the pharmaceutical industry and industrial policy makers) but also possible convergences. Notes 1. 2. 3.

4.

5.

6.

7. 8.

9.

10.

The health status at the level of the whole population can be summarized by measures such as life expectancy, infant and child mortality, disability-adjusted life years (DALYs), and so on. For a discussion on the definition of health, see Buchanan (2000). Article 25 (1) of the UN Universal Declaration of Human Rights states: ‘Everyone has the right to a standard of living adequate for the health and well-being of himself and of his family, including food, clothing, housing and medical care and necessary social services, and the right to security in the event of unemployment, sickness, disability, widowhood, old age or other lack of livelihood in circumstances beyond his control.’ It is worth noting that in 1883, Bismarck had introduced the national insurance for work accidents in Germany. In 1911, the first public fund for the poor’s healthcare assistance was created in Japan. Further, in 1912, Lloyd George introduced in the UK the first obligatory health insurance for workers. Nevertheless, all these measures were isolated interventions; they were not part of a general framework of public intervention in the health sector. However, as Feldstein (2001, 2002) noted, there is an alternative view of government intervention, namely the Economic Theory of Regulation, which considers the legislation as the outcome of market-like transactions between different interest groups with their own objectives: legislators aim to be elected, so they exchange favourable regulation for the groups which can assure them of political support. In other countries public financing of healthcare predominates, although private insurance coverage is increasing. Therefore, though the relative proportions of public and private financing differ between European countries, the USA and Japan, most financing programmes in industrialized countries are, in fact, mixed, with a coexistence of public and private funding. The authors state that this traditional view on health considers health services as an input in a production function that has health as the only output. Therefore the responsibility of the health system is to produce an acceptable level of health as efficiently as possible. Di Tommaso and Schweitzer (2005) noted that educational systems are similar to the example of the health industry as an industry that produces broader societal benefits in addition to its stated mission. The educational system indeed produces broader benefits than knowledge, spilling over to improve labour productivity, political awareness and cultural enrichment. However, the health industry is considered particularly interesting in being able to produce broader technological development, as well as health, because of its high-tech nature. In the large literature on the topic, it is possible to identify a number of different definitions of industrial policy. Some authors agree with an encompassing definition (for example Corden, 1997) while other scholars prefer to adopt more focused definitions (see, on this issue, Chang, 1994). For the purposes of this chapter we will adopt an encompassing definition, identifying ‘industrial policy’ as the set of government interventions, aimed to have an impact on the industrial system. This definition embraces all three types of policy listed by Lall and Tuebal (1998): (a) functional policies, which improve market operations in general, for example, policies designed to enhance competitive pressure (competition and antitrust policy); (b) horizontal policies, which cross different sectors, such as generalized incentives to promote R&D and training; (c) selective policies, which are designed to promote particular sectors (for example sector-specific subsidies) or specific firms (for example, the promotion of ‘national champions’, like Italy’s FIAT and France’s Renault automobile companies). It is also worth emphasizing that the concept of industry has changed through years. We used to refer to industry as the manufacturing or secondary sector. Nowadays, it is almost impossible clearly to distinguish manufacturing from services: many services that were incorporated into large firms have been externalized (outsourcing) and now industry and services often work in such a close connection that it would be misleading to think that policies to promote industry cannot be applied to the service sector. For example, the implementation of the so-called ‘scientific tertiary sector’ (services for promoting technology transfer from universities to firms, legal consultants specializing in intellectual property protection and so on) is usually considered as a key ingredient in the promotion of high-tech industries. The outputs of the health industry are both services and products and virtually all of the health industry’s components are interconnected, forming a complex system of manufacturers, insurers and providers. This is a reasonable explanation for efforts to reduce overall health expenditures by reducing expenditures in one sector (for example, hospitals) so often failing. It is often observed that pharmaceuticals, for example, can be both substitutes and complements for other health services (Comanor and Schweitzer, 1995).

Technology policy and social policy 315 11.

12.

13.

14. 15.

16.

17.

18.

19. 20. 21.

Publicly funded (tax- or social insurance-financed) systems, are the main, and in most cases the only, payers of healthcare and are therefore interested in balancing their budgets and avoiding cost explosions, which will then have to be shifted to the consumers either by means of higher taxes or by higher social insurance premia. Indeed, consumers are likely to be resistant to higher tax rates and, in social insurance-financed systems, higher premia mean higher production costs. The increase in insurance premia in Germany, for example, prompted a substantial investment outflow of German companies seeking to relocate in countries where production and insurance costs were lower. There is limited evidence about the effects of primary care budget holding. In the United Kingdom and Ireland, where general practitioners were offered the opportunity to economize on prescribing and to use surplus resources in other practice activities, evaluation showed only short-term effects on drug spending (Stewart-Brown et al., 1995; Walley et al., 2000). Germany, which for three years capped overall pharmaceutical spending, announced in 1991 that the first DM 280 million above the ceiling was to be funded out of the doctors’ remuneration budget. As a result, the number of prescriptions fell by more than 10 per cent in 1992–3, and spending fell by 25 per cent. Although prescribing volume later returned to its former level, more permanent savings may have been achieved as ‘dubious’ products were abandoned (saving DM 1.8 billion) and a shift to generics saved a further DM 350 million (Schulenberg, 1994). The UK Pharmaceutical Price Regulation Scheme (PPRS) regulates profits to a band of 17–21 per cent on historical capital, with 25 per cent variation on either side. Companies are free to set prices, provided their rate of return is within these bands. If profits are higher, the company has to reimburse the National Health Service or reduce profits the next year. If profits are lower, the company is allowed to raise its prices. Knowledge is a good example of a quasi-public good or an impure public good (Stiglitz, 1999), since it is not rival (meaning that the use of knowledge by one person does not diminish its availability for others) but is – at least in principle – excludable. Knowledge, once acquired, can often be easily reproduced by its buyer, and possibly (apart from restrictions imposed by intellectual property rights) sold to others. In addition knowledge, once it is sold, remains also in the hands of the seller, and finally there is a fundamental paradox in the determination of the demand for knowledge: ‘its value for the purchaser is not known until he has the information, but then he has in effect acquired it without cost’ (Arrow, 1962, p.616). Patents provide the exclusive right to the inventor to enjoy the fruits of his innovative activity over a limited period of time, but, in return, the inventor must disclose the details of his invention. The details provided in the patent application make an enormous amount of knowledge freely available. The development of rayon, for example, provided other researchers with enormous information: it demonstrated the feasibility of a synthetic fibre, knowledge which itself was of enormous commercial value and which provides incentives for others to look for other synthetic fibres. Indeed, research in chemicals often consists of looking for slight variations of the original chemical. According to Di Tommaso and Schweitzer (2005) industrial policy in the health industry can be seen from three different points of view, according to the degree of central authority inherent in a policy. The first, the economy-wide, macro view looks at the effect of national government policies in encouraging innovation and high-tech industrial development. The second view, called the micro view, discusses both production of innovation and adoption of technological change by individual health industry actors: firms, universities, hospitals and so on. There is also a third, intermediate (or meso) level, which is neither economy-wide, nor specific to a single actor. This refers primarily to regions and in particular to industrial clusters, defined as geographic agglomerations of firms and institutions operating in a single industry or in a group of complementary sectors. Industrial clusters have existed for centuries (Bianchi, 1996). However, only recently has the literature in the field of industrial organization noted the importance of these clusters in increasing industry efficiency and competitiveness. Most of this literature refers to the industrial sector and not to high-technology industries. Some have suggested that the nature of high-technology industries (for example, the lack of need for physical proximity to inputs or markets), make clustering in this sector obsolete (distance is obsolete). The evidence, however, is strongly to the contrary, and we and our colleagues point out that these clusters are as central to the development of ‘new’, high-technology industries as they were to manufacturing industries. According to these critics’ voices, there could also be serious threats to knowledge diffusion and access to new knowledge with an academic system centred on patents rather than on publications (Nelson, 2003; David, 2004). Other actors are important in the creation of a so-called ‘innovative milieu’, an innovation-friendly environment in which biotech firms can develop and grow. One of them is the venture capital sector (see Di Tommaso et al., 2005 for a discussion of this topic). The first research to find a remedy for malaria was undertaken because of the military campaign in Vietnam in the 1960s. Today, the interest in antimalaral treatement is primarily linked to tourism with tropical destinations.

316 22.

23.

24. 25.

International handbook on industrial policy Examples are the ‘Good Housekeeping Seal of Approval’ that is often applied to kitchen implements, the ‘Underwriters Laboratory’ (UL) seal that is applied to electrical appliances, and the Consumer Union that tests all sorts of consumer goods, ranging from food and clothing to appliances and even automobiles. In the case of Good Housekeeping and UL, manufacturers pay the laboratory to test their product and, in return, are allowed to market their products with the respective ‘seal’ of quality. The liability systems seem to differ between the USA and Europe: certification is more of the ‘selfregulatory’ type in the USA (in the sense that producers have to take care of it) while it is mainly the government which regulates and certifies in Europe, even if self-regulation has become more popular in recent years (for example, ISO 9000; BRC – Technical Standard and Protocol for Companies Supplying Retailer Branded Food Products). However in pharmaceuticals both Europe and the USA use a blend of selfregulation and government regulation, but with little emphasis on the first. Whether more self-regulation could actually improve consumer welfare without significantly lowering safety is an empirical question, unexplored until now. Finally, it has to be noted that in the current system the liability is always with the manufacturer; even if a government regulatory agency (EMEA or FDA) approves a drug, the government is not liable for damages. Similarly to the case of strict formularies, high co-payment is likely to decrease the probability of having a drug used when it is unnecessary (fist-type error), but is also likely to increase the probability of not having a drug used when it is necessary (second-type error). Two critiques of the efficiency approach, that states that government intervention is required to correct market failure, come from the political economy literature and from the government failure literature. The former underlines that the government in reality has different objectives than maximizing the welfare of society. In fact, for example, it can take actions in the interest of small influential groups. The latter points out that, even assuming the ‘benevolence’ of the state, it is not able to correct market failure efficiently, since it usually lacks information and has high bureaucratic costs. Therefore the intervention should take into account the costs of market failure, the cost of government intervention and the cost of correcting government failure (for example improving public sector efficiency). See Chang (1994) for a comprehensive review on the topic.

References Arrow, K.J. (1962), ‘Economic welfare and the allocation of resources for invention’, in R.R. Nelson (ed.), The Rate and Direction of Innovative Activity: Economic and Social Factors, Princeton: Princeton University Press. Barr, N. (1998), The Economics of the Welfare State, Oxford and New York: Oxford University Press. Becker, G.S. (1993), Human Capital, 3rd edn, Chicago: University of Chicago Press. Bianchi, P. (1996), ‘New approach to industrial policies at the local level’, in F. Cossentino, F. Pyke and W. Sengenberger (eds), Local and Regional Response to Global Pressure: The Case of Italy and its Industrial Districts, Geneva: International Institute for Labour Studies, Research Series no. 103, pp.195–206. Buchanan, D.R. (2000), An Ethic For Health Promotion. Rethinking the Sources of Human Well-Being, Oxford and New York: Oxford University Press. Chang, H. (1994), The Political Economy of Industrial Policy, New York: St Martin’s Press. Cohen, M., R.R. Nelson and J.P. Walsh (2000), ‘Protecting their intellectual assets: appropriability conditions and why firms patent and why they do not in the American manufacturing sector’, NBER working paper no. 7552. Comanor, W.S. and S.O. Schweitzer (1995), ‘Pharmaceuticals’, in W. Adams and J. Brock (eds), The Structure of American Industry, 9th edn, Englewood Cliffs, NJ: Prentice-Hall. Corden, W.M. (1997), Trade Policy and Economic Welfare, 2nd edn, Oxford: Clarendon Press. David, P. (2004), ‘Can “open science” be protected from the evolving regime of intellectual property rights protection?’, Journal of Theoretical and Institutional Economics, 160, 1–26. Di Masi, J.A., R.W. Hansen and H.G. Grabowski (2003), ‘The price of innovation: new estimates of drug development costs’, Journal of Health Economics, 22, 151–85. Di Tommaso, M.R. and S.O. Schweitzer (2005), ‘Introduction: why apply industrial policy to the health industry?’, in M.R. Di Tommaso and S.O. Schweitzer (eds), Health Policy and High-tech Clusters Industrial Development. Learning from Innovation in the Health Industry, Cheltenham, UK and Northampton, MA, USA: Edward Elgar, pp.3–16. Di Tommaso, M.R., D. Paci and S.O. Schweitzer (2005), ‘The geography of intangibles, the case of the health industry’, in M.R. Di Tommaso and S.O. Schweitzer (eds), Health Policy and High-tech Industrial Development. Learning from Innovation in the Health Industry, Cheltenham, UK and Northampton, MA, USA: Edwar Elgar, pp.258–78. Feldstein, P.J. (2001), Health Care Economics, 6th edn, Albany, NY: Delmar Publishers. Feldstein, P.J. (2002), Health Policy Issues. An Economic Perspective, 3rd edn, Washington, DC: AUPHA Press. Folland, S., A.C. Goodman and M. Stano (2001), The Economics of Health and Health Care, 3rd edn, Upper Saddle River, NJ: Prentice-Hall.

Technology policy and social policy 317 Grossman, M. (1972), ‘On the concept of health capital and the demand for health’, Journal of Political Economy, 80, 223–55. International Monetary Fund – IMF (2004), The Macroeconomics of Aids, Washington, DC: IMF Publications. Ioannides-Demos, L.L., J.E. Ibrahim and J.J. McNeil (2002), ‘Reference based pricing schemes: effects on pharmaceutical expenditure, resource utilisation, and health outcomes’, Pharmacoeconomics, 20( 9), 577–91. Lall, S. and M. Teubal (1988), ‘ “Market stimulating” technology policies in developing countries: a framework with examples from East Asia’, World Development, 26(8). Levin, R., W. Cohen and D. Mowery (1987), ‘R&D appropriability, opportunity and market structure: new evidence on some Schumpeterian hypotheses’, American Economic Review, papers and proceedings, 75, 20–24. Mansfield, D. (1986), ‘Patents and innovation: an empirical study’, Management Science, 32, 173–81. National Cancer Institution–Division of Cancer Treatment (1995), Program Information Booklet. The National Institute for Health Care Management Research and Education Foundation (2001), ‘The changing pattern of pharmaceutical innovation’, research report, Washington, DC. Nelson, R.R. (2003), ‘The market economy and the scientific commons’, Research Policy, 33, 455–71. Newhouse, J.P. (1993), Free For All, Lessons from the RAND Health Insurance Experiment, Cambridge, MA: Harvard University Press. OECD (2002), OECD Health Data, Paris. OECD (2003), OECD Health Data, Paris. Pavitt, K. (1998), ‘The social shaping of the national science base’, Research Policy, 27, 793–805. Phrma–Association of pharmaceutical manufacturers of America (2003), Pharmaceutical Industry Profile. Prevezer, M. (1997), ‘The dynamics of industrial clustering in biotechnology’, Small Business Economics, 9, 255–71. Scherer, F.M. and S. Weisburst (1995), ‘Economic effects of strengthening pharmaceutical patent protection in Italy’, International Review of Industrial Property and Copyright Law, 26, 1009–24. Schulenberg, M. (1994), ‘The German health care system at the crossroads’, Health Economics, 3(5), 301–4. Schweitzer, S.O. (1997), Pharmaceutical Economics and Policy, New York: Oxford University Press. Schweitzer, S.O. and M.R. Di Tommaso (2005), ‘The health industry model: New Roles for the Health Industry’, in M.R. Di Tommaso and S.O. Schweitzer (eds), Health Policy and High-tech Clusters Industrial Development. Learning from Innovation in the Health Industry, Cheltenham, UK and Northampton, MA, USA: Edward Elgar, pp.17–45. Schweitzer, S.O., J. Connell and F.P. Schoenberg (2005), ‘Clustering in the biotechnology industry’, in M.R. Di Tommaso and S.O. Schweitzer (eds), Health Policy and High-tech Industrial Development. Learning from Innovation in the Health Industry, Cheltenham, UK and Northampton, MA, USA: Edwar Elgar, pp.258–78. Sen, A. (1999), Development as Freedom, Oxford: Oxford University Press. Soumerai, S.B., D. Ross Degnan, J. Avorn, T.J. Mclaughlin and I. Choodnovskiy (1991), ‘Effects of Medicaid drug-payment limits on admission to hospitals and nursing homes’, New England Journal of Medicine, 325(15), 1072–7. Soumerai, S.B., D. Ross Degnan, J. Avorn, T.J. Mclaughlin and I. Choodnovskiy (1994), ‘Effects of a limit on Medicaid drug-reimbursement benefits on the use of psychotropic agents and acute mental health services by patients with schizophrenia’, New England Journal of Medicine, 331(10), 650–55. Stewart-Brown, S., R. Surender, J. Bradlow, A. Coulter and H. Doll (1995), ‘The effects of fundholding in general practice on prescribing habits three years after the introduction of the scheme’, British Medical Journal, 311(9), 1543–7. Stiglitz, J. (1999), ‘Knowledge as a public good’, in E. Kaul, I.E. Grunberg and M.A. Stern, Global Public Goods, New York: Oxford University Press. Taylor, C. and Z. Silberston (1973), The Economic Impact of the Patent System: A Study of the British Patent System, Cambridge, MA: Cambridge University Press. The National Institute for Health Care Management Research and Educational Foundation (2002), ‘Changing patterns of pharmaceutical innovation’ (http://www.nihcm.org). Walley T., M. Murphy, M. Codd, Z. Johnston and T. Quirke (2000), ‘Effects of monetary incentives on primary care prescribing in Ireland: changes in prescribing patterns in one health board, 1990–95’, Pharmacoepidemiology and Drug Safety, 9(7), 591–8. World Health Organization–WHO (1948), Preamble to the constitution of the World Health Organization as adopted by the International Health Conference, New York, 19–22 June 1946; signed on 22 July 1946, entered into force on 7 April 1948. World Health Organization–WHO (2001), Macroeconomics and Health: Investing in Health for Economic Development, Report of the Commission on Macroeconomics and Health Geneva, World Health Organization. World Health Organization–WHO (2005), World Health Report 2005.

PART IV TERRITORY, INDUSTRIAL DEVELOPMENT AND SMALL FIRMS

16 Decentralizing industrial policies: threat or opportunity in developing countries? Leandro Sepulveda1 and Ash Amin

1 Introduction Over the last decade a new approach to industrial policy has been maturing, after a long period of abandonment of state-centred, top-down and product-specific industrial policies. The new approach, stimulated by the revival of traditional industrial districts as well as new patterns of regional revival based on high-tech industries, is bottom-up, associationalist and cluster-based. As an approach that does not radically threaten the market foundations of neoliberal thinking that has become so prevalent around the world, the new decentralized approach to industrial policy has rapidly grown from its modest beginnings into a new policy orthodoxy for dealing with spatial inequality through attempts to boost local economic competitiveness. In this chapter we examine the central tenets of the new policy approach, which combines supply-side intervention with regional boosterism as a new form of industrial policy. We then critically review the applicability of the new approach in the context of countries with fragile or absent macroeconomic stability, based on a detailed case study of a relatively buoyant industrial district in Argentina. We conclude that decentralized policies in the absence of a stable and interventionist macroeconomic policy framework risk becoming, at best, blunt instruments and, at worst, unwitting agents in a broader policy framework that might actually be increasing economic inequality. 2 The new industrial policy approach The shift in the nature of industrial policy has been advocated by different schools of thought which are grouped and referred to here as ‘the new industrial policy’ literature. Geography lies at the heart of the new industrial policy thinking, in the form of hopes pinned on unlocking local resources. More specifically, the emphasis falls on the powers of local agglomeration and inter-firm cooperation, both of which are said to contribute in various ways to increase firm competitiveness as well as local returns. The positive effects stressed include the generation of externalities (Marshall, 1890, 1919; Becattini, 1979, 2002; Brusco, 1982; Porter, 1990; Krugman, 1998) and economies of agglomeration and specialization (Krugman, 1995; Piore and Sabel, 1984; Porter, 1990, 1994, 1998a, 1998b, 1998c; Saxenian, 1994; Scott, 1998; Storper, 1997), the reduction of transaction costs (Scott, 1995; Krugman, 1995) and the stimulation of knowledge creation and innovation (Braczyk et al., 1998; Cooke and Morgan, 1998; Florida, 1995; Maskell and Malmberg, 1999; Storper, 1997). Two broad conceptual traditions drive the thinking on powers of place in the new thinking. The first tradition includes at one end the influential ‘theory of clusters’ proposed by Michael Porter (1990, 1994) and at the other end the ‘new geographical economics’ associated with the international trade economist Paul Krugman (Krugman, 1995, 1998; Fujita et al., 2002). The second tradition brings 321

322

International handbook on industrial policy

together more sociologically informed scholars, especially economic geographers influenced by evolutionary and institutional economists. The ‘new economic geography’ Krugman’s (1995, 1998) recent rediscovery of geography is linked to his analysis of international trade. Understanding the latter requires understanding the reasons why industries tend to concentrate geographically. In Krugman’s vision, firms tend to locate in areas where the supply of inputs is economically convenient and/or demand is larger, such that new firms are constantly added to the agglomerations. This explanation links to the old Marshallian theory of external economies. In his Principles of Economics, Marshall (1890) maintains that advantages emerge from the geographical concentration of specialized firms, including information flows, the pooling of specialized local labour markets, provision of varied intermediate inputs, and firm specialization in different stages and branches of production. Krugman’s economies of scale affect both intermediate and final outputs which give rise to an uneven pattern of market dominance, hence to specialization, productivity gains, and trade (especially intra-industry trade). The ‘new economic geography’ has however been criticized for failing to provide a full explanation of the very source of agglomeration economies (Martin and Sunley, 1996, 1998, 2003; Storper, 1997). Storper (1997) argues that Krugman says little about how clustered intermediate-output producers become specialized, how firms ‘get there’ (notably, producers of final goods), and how the structure of specialist suppliers and demand for proximity are created through market transactions. In Krugman’s geographical assertion, ‘everything reduces rather axiomatically to fully efficient, though imperfectly competitive, clusters of producers who enjoy pecuniary externalities’ (Storper, 1997, p.14). Economies of specialization generated from intense business relations amongst collocated firms represent also a source of productivity gains and growth for Porter’s theory of international competitive advantage (see Porter, 1990, 1994, 2001). In Porter’s view, the success of export firms of a country depends upon the quality of its ‘competitive diamond’, influenced by firm strategy, market structure and rivalry, factor input conditions, demand conditions and related and supporting industries (Porter, 1990). The development of the components of this ‘diamond’ and their effective interaction form the base for firm productivity gains. According to Martin and Sunley (2003), the cluster constitutes Porter’s ‘spatial metaphor’. Initially, clusters were defined as ‘geographic concentrations of [vertically or horizontally] interconnected companies, specialised suppliers, service providers, firms in related industries, and associated institutions (for example, universities, standards agencies, and trade associations) in particular fields that compete but also co-operate’ (Porter, 1998a, p.197). Above all, Porter sees a cluster as ‘a self-reinforcing system’ that stimulates the competitive strategies of clustered firms and hence the ‘competitiveness of the cluster itself’ (Martin and Sunley, 2003, p.14). Porter’s theory of clusters has become a mantra for recent debates on industrial policy and competitiveness (OECD, 1999, 2001). As Martin and Sunley (2003, p.1) put it, the cluster ‘has become the standard concept in the field and policy-makers the world over have seized upon Porter’s cluster model as a tool for promoting national, regional, and local competitiveness, innovation and growth’. Why have Porter’s ideas become so popular? In many ways, the Porterian conceptual framework fills the policy gap created by the collapse of traditional industrial policy, by shedding light on the sources of

Decentralizing industrial policies 323 competitiveness which by the mid-to-late 1980s became ‘one of the central preoccupations of government and industry in every nation’ (Porter, 1990, p.1). His notion of cluster ‘is framed directly in terms of the economics of business strategy’ which offers an ‘accessible’, ‘commonsense’ and ‘business-policy friendly’ approach that is ideologically compatible with mainstream market-oriented principles of economic management (Martin and Sunley, 2003). Redistributional and compensatory aspects (amongst industrial sectors and amongst localities or regions) are not part of the package, and neither are national state-centred industrial policies. Despite the alleged importance assigned to the socioinstitutional factors in cluster formations and dynamics (see Porter 1994, 1998b and 1998c), Porter advances little, however, to elucidate this issue. As theorized below, the socioinstitutional factors constitute a central piece to understand how innovation and learning occur in firm agglomerations; and yet, it remains ‘something of a black box in Porter’s work’ (Martin and Sunley, 2003, p.16). The socioeconomics of colocation The second tradition of thought focuses on social, cultural and institutional factors that underlie the links and interactions between colocated firms. It examines the role of nonmarket relations between economic actors which are considered vital for boosting competitiveness and innovation. These relations have been synthesized through a variety of metaphors to describe the dynamic of places with strong economies of association, including ‘industrial districts’ (Bagnasco, 1988; Becattini, 1979, 1990; Bianchi and Giordani, 1993; Brusco, 1982, 1989; Pyke et al., 1990), ‘milieux innovateurs’ (Camagni, 1991, 2002; Garofoli, 2002), ‘flexible specialization’ (Piore and Sabel, 1984), ‘neoMarshallian nodes’ (Amin and Thrift, 1992), ‘new industrial spaces’ (Scott, 1988, 1998), and ‘learning regions’ (Braczyk et al., 1998; Cooke and Morgan, 1998; Florida, 1995; Maskell et al., 1998; Morgan, 1997). At the core of these concepts lies an interest in ‘the intangible aspect of a territorial or regional economy that underlies innovative, flexible, agglomerations, of both the high- and low-tech variety’ (Storper, 1997, p.5) (italics in original). For Storper, this aspect can be referred to as ‘untraded interdependencies’ or ‘relational assets’ in the form of ‘conventions, informal rules, and habits that coordinate economic actors under conditions of uncertainty’ and actually support ‘trade interdependences’ that underpin competitiveness (ibid.). Untraded interdependencies thus have a market value. However, what is critical here is to identify those assets which, apart from possessing market value, can be transformed into either cheaper, higher quality, or innovative goods and services that ultimately endow firms and local productions with market distinctiveness. There is widespread agreement in this literature that a prime relational asset is ‘knowledge’, specifically ‘tacit’ or ‘embedded knowledge’ (Audretsch, 1998; Maskell et al., 1998; Cooke and Morgan, 1994, 1998; Storper, 1994, 1997). As Howells (2002, p.872) points out, tacit forms of knowledge ‘concern direct experience that is not codifiable via artefacts [ . . . ] it represents disembodied know-how that is acquired via the informal take-up of learned behaviour and procedures’. In contrast to ‘codified’ knowledge, tacit knowledge cannot be easily transmitted or communicated through formal language or acquired via market transactions. Face-to-face contacts, communications and interactivity are seen as the best ways to gain access to, learn and master tacit knowledge. Knowledge which is relationally created, for example, through local networks of

324

International handbook on industrial policy

mutual interdependence facilitates not only the generation of ‘firm-specific’ competences but also ‘place-specific’ competences (Maskell et al., 1998). Such competences are difficult if not impossible for competitors and/or outsiders to imitate. Hence dense local and regional agglomerations of firms operate as generators and repositories of knowledge, thus occupying a strategic place in the supply architecture of the learning and innovation economy. The policy implication of work on the sociorelational nature of learning and innovation is that firms need to craft external links to mobilize and put together varied and often fragmented sources of information and knowledge. The importance given to connectivity and interactivity amongst economic agents has led the new industrial policy literature to think in terms of the quality of systems of innovation (IS). IS are made up of a range of institutional forms, including firms, government agencies, the structure of industry, the availability and quality of R&D institutes, educational and training institutions, physical infrastructure of information and communication, the financial system and professional organizations, labour organizations, the level of education and expertise of the labour force and, critically, the industrial relations of firms with clients, suppliers and rival firms (Lundvall, 1992; Lundvall and Johnson, 1994; Freeman, 1994). To a large extent, however, IS are locally constituted, and influenced by the quality and nature of local labour markets, educational structures, business innovation supports, links between industry and formal research environments, inter-firm conventions, and a host of informal embedded traditions that make up a local way of doing things. Following Scott and Storper (1995, p.513), central policy principles of the new industrial policy thinking include, firstly, ‘context-sensitiveness’. Accordingly, a decentralized industrial policy agenda should be concerned with the embeddedness of industrial activity within specific economic and socioinstitutional spaces. Lambooy (2000, p.28) in this respect explains that agglomeration economics should be ‘connected’ with the ‘socioeconomic infrastructure’ and ‘the behavioural options of people’ in a locality or region. Secondly, it should encourage the ongoing adjustment of local capacities instead of promoting only the imitation and transference of best practice, attending to endogenous knowledge creation and neo-Schumpeterian competition as policy benchmarks. Thirdly, it should be a ‘production systems-oriented’ agenda, focusing on economies of association, instead of focusing exclusively on isolated single firms. A policy should exploit three different levels of associative capacity: intra-firm, inter-firm and firm-institutional set-ups (Cooke and Morgan, 1998). Hence the policy interest in modern ‘intermediary logics’ such as clusters and cooperative networks of firms; brokers and intermediaries, and firmsupporting institutions including service-oriented business associations, centre of services and development agencies. Both of these encourage dialogue and promote, establish and coordinate ‘upstream’ and ‘downstream’ collaborative actions amongst colocated firms and other local economic agents (notably on exchange of information and knowledge, collective strategies for solving shared problems, pool resources, activities of production and R&D, marketing place and commercialization). A new governance consensus The new industrial policy thinking has a third core conceptual foundation, related to the governance structure of locally based competitiveness strategies. Replacing the state as the central actor of the old industrial policy approach is the policy network made up of

Decentralizing industrial policies 325 a plurality of local institutions (including the state) working with each other and in a nondirigiste way. Concepts such as ‘policy network’ (Rhodes, 1990, 1997; Cooke and Morgan, 1993, 1998), ‘forums and mechanisms of collective action’ (Scott, 1998) and ‘developmental coalitions’ (Keating, 1998), to mention but a few, have been coined to capture this conceptual requirement. Cooke and Morgan (1998, p.80) define ‘policy network’ as ‘an informal or semi-formal organizational mechanism consisting of public and private individuals, groups, organizations, and associations whose key discriminating factors is that they interact around specific policies and programmes [. . .] The key is that network participants are “of consequence” to the policy field in question’. A multi-agency logic of collective action is argued to help generate conditions for coordination and regulation of local business activities by providing micro certainty to firms which often operate under conditions of market volatility and macroeconomic uncertainty (Camagni, 1991; Scott, 1998; Storper, 1997). The presence of network animateurs is identified as a key ‘bottleneck’ by the literature on ‘decentralized industrial policy’ (Begg et al., 1995; Begg and Mayes, 2000), since they held the capacity to transform the potential of individual willingness into the political mobilization of a locality or region. The concept of ‘institutional thickness’, coined by Amin and Thrift (1994), opportunely came to identify the institutional determinants of local collective action in support of a network economy. These include a strong presence of business and support institutions; high levels of interaction amongst the institutions in a local area and, as a result of this, the development of ‘sharply defined structures of domination and/or patterns of coalition resulting in the collective representation’; and presence among collective actors of ‘a mutual awareness that they are involved in a common enterprise’ (ibid., pp.14–15). As regards the role of the state in promoting economic development, this has changed from that of rule maker to that of ‘animateur’ (Brusco and Righi, 1989; Boekema et al., 2000; Cooke and Morgan, 1998). The main role of the (local) state is to build up capacities for action and craft networks, ‘by enabling’ and ‘by orchestrating’, through which economic agents ‘are able to collaborate for mutually beneficial ends’ (Morgan and Nauwelaers, 1999, p.14). More efficient and less complex functional structures and regulations, transparent business legislation, and consensus building via inclusiveness in policy decision making and dialogue-oriented policy vocation, are also seen as pillars of the new role played by the state (Amin, 1999; Boekema et al., 2000; Morgan, 1997; Keating, 1998). In turn, private sector associations are expected to be strong, plural, responsive, nonhierarchical and active in articulating the needs and demands of local firms, pooling resources, providing ‘club goods’ directly or indirectly (notably, through the establishment of centres of services), and fostering conditions for collective learning (Best, 1990; Brusco, 1982; Helmsing, 2001; Keating, 1998; Maskell et al., 1998). The private sector, represented by business associations or networks of entrepreneurs, is thus seen as an agent of industrial policy acting either on its own or in partnership with the local state. The new literature argues that neither the national state nor the local state on their own, and even less so decentralized market mechanisms, can cope with the complexity associated with new governance demands (Jessop, 1997; Rhodes, 1997; Scott, 1998). However, because the state continues to be the main coordinator and regulator of economic activity inside national boundaries, the dialogue, synergy and coordination between bottom-up policy assertion and national and supranational policy frameworks and institutional structures become a critical dimension (Bianchi, 1994; Bianchi et al.,

326

International handbook on industrial policy

1998; Hudson, 2001), not least to avoid institutional overlap and duplication of, for instance, ‘decentralized’ national industrial policy initiatives intended to affect different cities or regions (see Begg et al., 1995; Begg and Mayes, 2000). It is also critical to manage and/or maximize the effects of successful locally led industrialization on the whole economy, on the one hand, and on the other hand, to control and reduce the disparities between thriving and lagging behind local economies within the same country. The paradigmatic case of the industrial districts in Europe (Baden-Württemberg and North Rhine-Westphalia in Germany and the Emilia-Romagna in Italy) has demonstrated that the status of such regions as engines of economic growth for a country does not necessarily prevent the emergence and/or deepening of local/regional asymmetries within it. In all these instances, the coordinating role of the state becomes crucial. As with firms, learning is also a vital component for the governance of industrial policy. From an evolutionary standpoint, institutional learning requires minimal thresholds of coordination between local public and private stakeholders for the creation of a local ‘collective subject’. This is particularly relevant in cases of scarcely developed and/or dysfunctional institutional set-ups in which retraining and capacity building at the level of leaders and officials within public and private institutions become a policy imperative, a scaffolding from which a local industrial policy may emerge (Sepulveda, 2004). An equally important baseline condition is the effective management of dissent, which is likely to appear in localities and/or institutions traditionally dominated by hierarchical governance structures, as the latter are put under threat. 3 The developing world and the new industrial policy literature The new industrial policy thinking has gained considerable influence in developing countries, particularly within policy circles concerned with issues of local, urban and regional development, SME policy, clusters, and entrepreneurship (see Altenburg and MeyerStamer, 1999; Bianchi, 1997; Boisier, 2002; ECLAC, 1999a, 1999b; Gatto et al., 1993; Humphrey, 1995; Nadvi, 1992; Pedersen et al., 1994; Schmitz, 1990, 1999; Schmitz and Nadvi, 1999; Sepulveda, 1999, 2001, 2004; World Development, 1999; Yoguel and Boscherini, 1999). For some, this literature was seen as an opportunity for ‘digging up’ the forbidden and forgotten concept of industrial policy after decades of neoliberal banishment. ‘Industrial districts’, ‘local systems of production’, ‘clusters’ and ‘local/regional systems of innovation’ one-by-one became fashionable concepts, especially in those countries with an industrial tradition such as Argentina, Brazil and Mexico in Latin America, India and Pakistan in South Asia, and Egypt and South Africa in Africa, to mention only a few. Critically, policy initiatives inspired by the new industrial policy thinking started to become a reality in several of these countries. There are some fundamental questions that arise regarding the effectiveness of decentralizing industrial policy in developing countries.2 Is it an effective way of boosting competitiveness in these countries, given the high expectations associated with it regarding inter-firm cooperation, localized learning and local institutional capacity? Do localities in developing countries possess even the germ of such attributes? In turn, what does the model offer to those countries marked by severe macroeconomic instability and limited autonomy of economic policy in the face of neoliberal globalization and the rise of powerful international policy regimes? Then, how should this model be read in the long history of policy decision making in developing countries based on uncritical imitation of

Decentralizing industrial policies 327 development models generated in industrialized economies (see Evans, 1987, 1992, 1995; Evans and Rauch, 1997; Maxfield and Schneider, 1997)? Is it yet another policy fashion that is inappropriate for developing countries? In this section, we seek to address these questions through a case study in Argentina.3 Our aim is not to generalize from this case study for all developing countries, but we do believe that our findings are relevant to other countries marked by, on the one hand, macroeconomic instability and, on the other hand, by institutional ‘thinness’, state failure and limited economies of association within the business community. We also believe that, despite evident advances in the literature on new industrial policy in developing countries (see references quoted above), a more comprehensive analysis needs to be done to understand the huge gap existing between the ‘script’ that has been drafted in the developed world (as discussed in the first section of this chapter) and the reality in the developing world. The case of Argentina Argentina is a developing economy with a considerable industrial tradition in Latin America. Estimations based on the last economic census in 1994 suggest that around 90 000 manufacturing firms make up its industrial base (see Table 16.1). Until the early 1970s the manufacturing industry was in fact the engine of growth and job creation in the Argentinean economy (Katz and Kosacoff, 1989; Kosacoff, 2000). Consecutive cycles of stop-and-go and political turmoil brought about programmes of adjustment and structural reforms from the mid-1970s onwards which ended up destroying the industrialization model based on import substitution that had prevailed since the 1930s.4 A new model of development came into existence, one based on the principles of an ‘open and modern’ market economy. As far as the manufacturing industry is concerned, the main consequences of market liberalism were a generalized process of capital concentration, the return to natural resource-intensive productions, and deindustrialization (Chudnovsky et al., 1996; Kosacoff, 1998, 2000). A small group of national Holdings Companies (HCs) and MNCs emerged as the main beneficiaries of the new model, which is largely based on capital-intensive industries including energy, mining, the food/agroindustrial complex, and former state-owned utilities, which often benefit from oligopolistic or monopolistic market positions (Azpiazu et al., 1986; Bisang, 2000; Kosacoff, 2000). The privatization of public utilities, the disappearance of manufacturing state-owned companies, and the hegemonic role reached by the financial capital also define the new model of development. Industrial policy virtually disappeared from the policy agenda. After 30 years of experimentation, the market model has yet to deliver a sustainable Table 16.1 Manufacturing industry in Argentina

Number of firms Employment Source: INDEC (1998).

Census 1974

Census 1985

Census 1994 (inf. for 1993)

126 388 1 525 221

109 376 (13.5%) 1 381 805 (9.4%)

93 156 (14.8%) 1 061 528 (23%)

328

International handbook on industrial policy

pattern of growth. Argentina’s frequent economic and financial crises (notably in 1989 and 2001), and the political upheavals generated, are a testimony of this. In fact, after four years of economic slowdown and financial crisis, the spectacular collapse of the peso–dollar convertibility5 in 2001 marked the tragic entrance of Argentina in the new millennium.6 Although the (former) widespread ‘consensus’ existing around marketdriven policies has been weakened as a result of this failure, this has not brought about sustained efforts against the dominant economic model. One consequence of the convertibility crisis has however been the political opportunity for debate in relation to different pathways for economic development. While this debate remains highly contested, it may create an opening for the micro industrial policy initiatives that emerged in the mid-1990s in the fringes of the Argentine political economy in face of the declining manufacturing base. Micro policy initiatives The rise of local industrial policy initiatives in the core industrial economies attracted some attention in Argentina in the 1990s, leading to some experiments. The cases of ‘Productive Municipalities’ and the City of Rafaela (both in the province of Santa Fe), and the Buenos Aires Institute of Entrepreneurial Development (in the province of Buenos Aires) are the most notable. A productive Municipalities (MP) was set up by the Ministry of Production of the Santa Fe province in 1996. It aims to support the development of SMEs in diverse localities of the province through the provision of business services to local firms in the areas of management, business opportunities, subcontracting and international trade. As a part of its political strategy, MP seeks to encourage a process of place-led decentralization via the establishment of Secretariats of Production by the local governments. These Secretariats are responsible for the transformation of the local states into facilitators and animateurs of local strategies for the SME development and lobbying the provincial government to get support for local firms. The City of Rafaela in Santa Fe province represents the most outstanding case of manufacturing cluster and bottom-up industrial policy in Argentina. Rafaela, a medium-sized city of 80 000 inhabitants located in central Argentina, has successfully developed endogenous competitive advantage on the basis of the coexistence and synergy between the following three factors. Firstly, a dynamic and forward looking group of small and medium-sized entrepreneurs largely associated with the manufacturing industry (mostly metal engineering and food industry firms). Secondly, a proactive, competent and politically homogeneous private sector which is organized through different local business organizations and one major organization that groups the former. Finally, a strong leadership exerted by an also forward looking local government and its operative arm, the Secretariat for Economic Planning (which the Rafaelians refer to as the ‘task force’), which is committed to promoting the local SMEs (Costamagna and Ferraro, 2000; Sepulveda, 1999). During the 1990s, the mobilization of Rafaela’s political and economic interests gave rise to a potent public–private partnership which since then has operated as the platform for the design, setting up and coordination of the local industrial policy. The establishment of the Centre for Entrepreneurial Development (in the mid-1990s) constituted a landmark in the regional industrial competitiveness. This state-of-the-art centre of service provision to firms is the result of the strategic association between the Argentine Industrial Union (the largest national business organization), the Inter-American Development Bank and

Decentralizing industrial policies 329 public/private local stakeholders. The Centre covers a wide range of specialized services including business information, training in different management areas, quality, marketing, ICTs, subcontracting, international trade and technological upgrading. Furthermore, a proactive policy of networking with national and international sources of learning and knowledge was set up by many local institutions (that is, with public and private institutions from Porto Alegre City in Brazil, Piemonte and Emilia Romagna regions in Italy, and Baden-Württemburg in Germany) contributing to Rafaela’s thriving internalization process. Thanks to the city-region’s achievements, Rafaela was soon considered as a ‘quasiindustrial district’ (Gatto et al., 1993), and despite the fact that subcontracting and labour division amongst local firms is considerably lower in Rafaela vis-à-vis the Italian industrial districts, it became a paradigmatic case of best practice for bottom-up industrial development. IDEB and, largely, MPs were in effect inspired by this experience. Less celebrated, the Buenos Aires Institute of Entrepreneurial Development (IDEB) is the largest place-based industrial policy initiative so far set up in Argentina. Established in 1996 by the government of Buenos Aires Province,7 IDEB aims to promote local firms through the decentralized provision of services to the firms, in particular to micro-, smalland medium-sized firms. The IDEB headquarters are located in the provincial capital (City of La Plata) and its operations are decentralized to five regional IDEB centres and 35 local IDEB institutes (originally). The latter are distributed throughout the province, with 13 institutes located in Greater Buenos Aires, including one in the Borough of Tigre, where part of this study was conducted. IDEB is the product of the strategic alliance between the provincial government and core provincial business associations. It aspires to multi-agency governance, although under the political surveillance (and more recently control) of the provincial Ministry of Production which IDEB depends upon financially. IDEB services cover a wide range of areas including management, design and assessment of business plans, quality control in products and processes, technological upgrading, international trade, subcontracting and inter-firm collaboration, and some sector-specific programmes (on tourism). In what follows our aim is to focus, not on the achievements of exemplary cases like Rafaela, but on the example of Tigre, as a way of discussing the difficulties that local industrial policy faces in places where bottom-up energies and top-down guidance are less assured, as is the condition of the vast majority of industrial regions in developing countries. The new industrial policy on test We do not have the space here to discuss the details of industrial structure and performance in Tigre.8 Instead, the analysis will focus on experience measured against the three factors endorsed by the new industrial policy thinking: bottom-up governance, associationalist networking and tacit learning. We find, first, a pervasive top-down approach through which public agencies design, set up and deliver IP services and programmes; second, the absence of a participative and associative business culture amongst firms and institutions; and third, the nature of learning and innovation in the large majority of firms studied differs substantially from that portrayed by the new industrial policy literature. The nature of learning and innovation The new literature on learning and innovation assumes that firms, particularly the SMEs, innovate gradually through ongoing

330

International handbook on industrial policy

incremental improvements, mostly drawing on tacit and experiential knowledge. Firms learn collectively, through their network of business relations, hence innovation and learning is said also to be favoured by spatial proximity. In contrast, the findings from the sample of firms studied reveal that these assumptions hold only partially. The nature of learning in the large majority of firms (80 per cent) is firm-centred, incremental, unsystematic and generally unintended. Innovation activities are reduced to ad hoc strategies of product differentiation through which firms seek to compete in market niches in which import competition can be avoided. Only occasionally it implies a move towards higher quality products. The competitive advantage of firms largely depends upon the entrepreneur’s craft skills, imitative capabilities and practical expertise. As our study reveals, these assets were necessary to ‘stay alive’ (as the firms point out) under such an adverse business environment.9 Furthermore, firms neither associate inter-firm collaboration with learning nor do they see the local institutional set-up as an interface for or facilitator of innovation. Firms operate on the basis of high levels of vertical integration, despite the emerging need addressed by firms to ‘open themselves up’ towards other firms to gain access to complementary assets otherwise inaccessible. Although firms rely on ‘shorter distances’ for securing new sources of information and knowledge, geographical proximity matters only partially for their competitiveness. Tracking rivals was identified as a positive source of learning within the wood furniture firms’ cluster. Small firms learn ‘by hearing’ what the local leading firms do, what they do not do and what they give up doing. The concept of ‘cluster noise’ allows us to conceptualize the circulation of information that occurs informally through proximity-based social relations and that helps firms to reduce their levels of uncertainty without costs (Sepulveda, 2004). In addition, firms benefit from basic or passive externalities existing in the cluster which are largely sector-specific, including Tigre’s local market for furniture and second-hand machines, specialized workforce, training infrastructure (notably technical schools) and the entrepreneurial atmosphere which favours entrepreneurship. Overall, however, as pointed out by one expert on furniture technology, the firms do not visualize the relationship between training or technological upgrading and productive efficiency, as the former strategies do not necessarily assure ‘tangible’short-term sales. Similarly, because firms do not hire professional advice or conduct ex ante market research, they have to rely upon ex post tests of their new products in the market. Therefore, the firms studied deploy ‘defensive’ and ‘reactive’ strategies to cope with the changing business environment. In a context of uncertainty and economic decline, firms see learning and innovation as risky activities of minor relative importance for their competitiveness vis-à-vis cutting the cost of labour, avoiding regulation and tax compliance, and a risk-averse wait-and-see approach to strategic decisions (that is, investment). ‘Always, every two or three years something happens to the Argentine economy that stops you’ was a frequent expression heard in fieldwork in relation to learning and innovation. This illustrates clearly the short-term business horizon under which firms have to operate. Macroeconomic instability also contributes to stimulate vertical integration by raising firms’ transaction costs due to high uncertainty, which ultimately contributes to create low-trust business environments. In addition, instability and uncertainty reinforce the belief that the main restrictions hindering firm competitiveness in Argentina are exogenous to the firms, notably, of a macroeconomic nature (FUNDESUNGS Report, 1999).

Decentralizing industrial policies 331 Where we found some evidence of higher levels of learning and innovation through interactive practices of knowledge searching, exploration and generation, it was based on ‘partner to partner’ market relations with foreign firms (that is, through joint ventures but not collective learning), rather than through local inter-firm links. In this case foreign partners operate as technological ‘gatekeepers’ for domestic firms. After years of selffounded experimentation and visits to other producers abroad, these firms have developed a clear view of the advantages of collaboration. However, it is positively seen only if it involves foreign partners, as they distrust collaboration with local partners, judged to be ‘individualistic’ and ‘opportunistic’. Firm competitiveness largely relies on distanced rather than proximate learning strategies. Institutional set-up The literature on industrial clusters, notably the Italian ones, highlights the role that local governments and public agencies play in providing what Schmitz and Humphrey (1996, p.1861) refer to as ‘a framework in which clusters of SMEs can flourish’, for example, by providing selective services to firms. Our field studies confirm that Argentina’s place-based services and programmes are supplied to local firms regardless of their needs which are associated with issues of size, sector, competences and context. Officials explained that services and programmes are directly ‘downloaded’ from the providers’ headquarters, which are frequently located in Buenos Aires City or other major provincial capitals. A study commissioned by ECLAC (2000) in fact concludes that the local IDEBs operate as a ‘selling point’ of services and not as a ‘reception point’ of the firms’ needs and demands, which would subsequently be used as input for design policy. The case of Tigre illustrates this issue well. Tigre’s institutional set-up is shaped by an important number of public and private institutions (including IDEB) which profess their commitment to supporting the local SME competitive performance through service provision and, in the case of business organizations, political representation. In Tigre, however, institutional ‘presence’ means neither effective support nor representation. Our findings reveal that a top-down approach to service provision, insufficient communication and lack of accurate knowledge between firms and institutions (a key issue to develop institutional density) characterize the nature of the relationships between firms and both public and private institutions. What is more striking is the fact that the majority of the institutions in Tigre do not see this issue as a major concern. A local business association remarked in this sense that, although they supply cost-free services to firms, ‘the entrepreneurs do not make use of them’. The fact that firms see the institutions neither as facilitators nor as partners in their learning process explains in part that the entrepreneurs recurrently perceive the national state, and not the local or provincial one, as the only legitimate (though absent) source of solutions for their problems. There is a rationale behind top-down-type strategies such as IDEB’s approach. Generic or ‘catch-all’ services and programmes are seen as good low-cost solutions compared to their costly custom-specific counterparts, owing to the economies of scale involved in their design and delivery and the larger market that they can capture. Provision of generic services allowed IDEB to quickly attain visibility throughout the Buenos Aires Province by incorporating a large number of firms into the IDEB system, notably SMEs that often do not gain policy support (ECLAC, 2000). Critically, in a context of generalized budgetary restriction, urgent needs and ideological opposition to the very idea of an industrial policy,

332

International handbook on industrial policy

the opportunity cost of a policy approach that does not show immediate results is extremely high. The strategy pursued by different micro industrial policy initiatives was the only one that would allow them to legitimate their existence and continuity in a neoliberal Argentina. In addition, officials from local business associations and IDEB claim that ‘the desire for overnight solutions’ among firms when they are approached blocks any chance for nurturing long-term relationships with them. Long-term relationships are essential from an evolutionary standpoint to develop ongoing support to the firms, particularly in areas such as technological upgrading, and continuing quality and innovation at large. The economic slowdown after 1998 made this demand for overnight solutions more evident. As shown below, this happened at exactly the time when some forward-thinking institutions began to change their corporate strategies in order to improve and exploit their links with the firms. On a positive note, our findings reveal that a slow process of institutional change is in course. Four institutions of Tigre’s system of innovation claim that, since they launched a new networking strategy in the mid-1990s, they have experienced increasing success in developing iterative links with firms (namely, IDEB, the local technical university, a local technical school and the Regional Centre for Technological Innovation in the Timer Industry). They understand the value of a more entrepreneurial approach towards service provision based upon a more specialized and customized supply of services and fast response times to firm requirements (and not on the mobilization of substantial financial resources). This is a considerable innovation for these types of organization in Argentina, and an opportunity for policies inspired by the new thinking. However, it by no means allows us to conclude that this type of soft institutional innovation has facilitated the development of a local system of innovation in Tigre. Likewise, IDEB Tigre, the technological university and the Industrial Union of Tigre (the largest business association in the region) have made substantial progress in identifying common interests and transforming them into specific policy actions aimed at local firms (for example, auto-part SME suppliers). This ‘common interest’ has not however turned into a policy network able to mobilize local stakeholders in order to push forward towards a local industrial policy agenda. So far its main political success has been to have initiated a public debate in relation to the ‘SME issue’ which previously did not exist in the region. State behaviour Aware that most localities and regions in Argentina do not possess the proactive institutional set-ups observed in the Third Italy or Rafaela City, we wanted to know whether in Tigre the local government, the business organizations and other industrial policy stakeholders at least developed joint actions or shared projects to support local firm competitiveness. Findings confirm that the local institutions seldom develop formal joint activities with other local institutions aimed at such ends. Nor do they hold informal contacts to exchange ideas and concerns in this sense. The institutions interviewed explain this gap principally as a lack of discernible motives or reasons to collaborate with other institutions. The local private sector appears as a highly fragmented, if not fractured, collage of interests and organizations which rarely relate to each other. Similarly, the synergy of the local state, represented by the Tigre Municipality and the Municipal Office of Industrial Affairs (in charge of local industrial policy issues), with other local institutions concerned with the manufacturing firms is extremely low. As pointed out by several local institutions, the ‘lack of interest’ that the local government has shown in the local manufacturing industry must be understood in a broader

Decentralizing industrial policies 333 sense. In the absence of one coherent strategy for Tigre’s local economic development, the local state has engaged in an infrastructure-led regeneration programme. Large public investments in infrastructure (bridges, roads, squares, street paving, schools and a new railway station) turned into a political and economical priority for the Tigre Municipality. Public infrastructure is developed alongside large influxes of non-local private investments in real estates since the mid-1990s (particularly in gated communities). As a key informant explains, these investments in real estate will generate enormous council tax revenues for the municipality which is, strictly speaking, ‘Tigre’s master plan’ for local development. The critical point here is that the role played by local government represents one of the biggest obstacles to the emergence of industrial policy initiatives in Tigre. In fact, IDEB, the technological university and the Industrial Union of Tigre argue that directly or indirectly the local government blocks even those policy initiatives which they have launched (notably, on government and real estate procurement). If the question is whether Tigre constitutes a case of governance failure the answer is no. On questioning, firms and private organizations pointed out that recent streamlining regulations and reforms throughout the public administration made Tigre’s bureaucracy more efficient and transparent. It facilitated the influx of external private investments in the 1990s (in real estates, a new casino and a modern funfair) and fostered the development of a friendly business environment for firms in the region. Hence the municipality, despite its lack of explicit policy support to local firms, can in some regards be seen as an example of institutional efficiency and good governance. Tigre also shows a level of success in developing bottom-up decision-making capacity and power as a result of its economic and political achievements.10 Tigre is therefore an example of proactive decentralization with respect to the Buenos Aires Province (to which, politically, it belongs). However, the experience does not correspond to the type of inclusive decentralization to which the new industrial policy literature refers. Like many other charismatic leaders at a local level in Argentina, the mayor in Tigre has developed a hierarchical and highly centralized style of government within the region, with negative effects on any local initiative that does not emerge from or is led by the local government. 4 Conclusion: the missing link between the local and the national So far we have analysed a number of practical difficulties and contextual restrictions that challenge the relevance of the powers of ‘the learning and innovation economy’, ‘economies of association’ and ‘bottom-up’ assertions as a rationale for place-based industrial policy in a particular regional context. The experience of Tigre shows that building the baseline conditions for a decentralized policy framework is demanding and not open to easy or quick policy fixes. Cultures of cooperation mature over decades and only when the economic benefits are tangible, economic organization based on the centrality of learning and knowledge generation remains largely the privilege of prosperous and well-organized economies, while the challenge of converting hierarchical models of governance to heterarchical ones requires a sea change in institutional habits. This is not to deny the potential of decentralized policies (Rafaela is a case in point and Tigre shows signs of possibility) but to qualify the limits and timescale of such an approach. It is also to invite speculation on the broader institutional conditions that are required to support local solutions, or, at the very least, on the national constraints that have to be tackled before a decentralized policy approach can be implemented. We end this chapter with a

334

International handbook on industrial policy

comment on the national policy setting in Argentina in order to draw attention to the enormity of impact that macroeconomic circumstances have in shaping local fortunes, an issue that has been largely ignored in the new policy literature. In Argentina, policy interest in SMEs and in manufacturing industry in general was marginal under the Convertibility model (1991–2001).11 Policy initiatives launched in this sense were mostly the result of a political need triggered by the limited creation of new jobs and the rising ‘structural unemployment’ generated by the model. In practical terms it meant that micro initiatives including IDEB, MPs and the case of Rafaela neither formed parts of a coordinated national strategy of decentralization of industrial policy, nor were they articulated with policy frameworks set up by the main (though unfunded) national public agencies concerned with IP issues (namely, the National Secretariat of Industry and the National Secretariat for the SME, the latter established in 1997). In contrast to neighbouring countries such as Brazil and Chile, Argentina has failed institutionally to establish a public agency specifically devoted to planning or guiding the IP framework. This framework can be characterized as a collage of barely coordinated public agencies, services and programmes; as a chaotic amalgamation and superposition of new and remaining old policies established by different governments while they were in office. Overlap, duplication and institutional complexity at all levels of government and all tiers of spatial responsibility are only a few expressions of such inattention (Gatto, 2003). Without a coherent national strategy for economic development, the regions (provinces) frequently committed their political support only to a small group of traditional large agroindustrial productions and other capital-intensive industries based on the extraction and processing of natural resources (oil, gas, mines and forests), owned by large national HCs and/or MNCs, and with little impact on the whole regional economy. In addition, policy makers inspired by the new industrial policy literature should note that for decades the Argentine political economy has been exclusively led by the issue of macroeconomic stability, and that this is not only the consequence of chronic problems of the balance of payments, inflationary cycles, financial crisis and subsequent adjustment programmes (notably, in 1989 and 2001). Critically, it is also the product of a generalized belief that only a healthy macro economy would bring about the necessary microeconomic adjustments in the manner in which the firms produce and compete. In an economy governed by such principles industrial policy of any sort is regarded as a second-best, if not a needless, policy choice. Our field evidence confirmed this: officials and policy makers from key national public agencies argued that to rectify the macro economy must be the main goal for Argentine industrial competitiveness. A policy maker from the powerful Ministry of Economy claimed that policies aimed at firms whether by size or location would not be needed if Argentina had a ‘normal’ or ‘healthy’ macro economy. There have been some changes in the ideological climate since the tragic events of December 2001. The Argentine state has reactivated some key tools of economic policy (critically, monetary policy). The newly elected government that came into power in mid2003 seems to be more committed to the problems of the real economy (including the manufacturing industry), and society is again beginning to see the so-called ‘national industry’ as a factor of economic progress. However, the new government’s prime aim of economic policy is, once again, trying to consolidate the extremely fragile macroeconomic ‘equilibrium’ achieved towards mid-2003. The fact that, almost across the board, Argentina’s economy is ‘booming’ (The Economist, 2004) and its exports soared to historic

Decentralizing industrial policies 335 records in 2004 (INDEC, 2005)12 will surely revive private trust and boost investment in the manufacturing industry; the wait-and-see behaviour of firms at large will only change once the wide range of political, social and economic uncertainties dominating the current Argentinean economy turn into medium-term certainties (Bonvecchi and Porta, 2003). Against a long history of business and institutional treatment of local structures of competitiveness as secondary to national frames of economic governance, the prospects of a decentralized approach to economic efficiency and interregional inequality look rather fragile. This is specially the case if, as we have shown, the dynamic of such an approach can be neither short-term nor dirigiste, but a slow evolutionary experimentation with associationalist democracy. But, equally importantly, what our evidence has shown is that such an experiment cannot assume that the conditions of success in some of the world’s most dynamic industrial districts and regionalized economies can be treated as a model for best practice to be emulated by actors and regions that face very different historic circumstances. What should a decentralized approach to industrial policy look at? As argued already, our evidence indicates that a lot depends upon the context and the circumstances under which the policy is going to be established. This requires careful evaluation of the thresholds of critical mass necessary for decentralization – in terms of number and types of firms, technological capacities, ‘trajectories’ of industrialization (Humphrey, 1995; Schmitz and Nadvi, 1999), institutional resources, and synergies amongst economic actors – before rushing into policy conclusions. In turn, special attention should be paid to complementarities and conflicts between local or regional economic policy interests and national policy frameworks, in a range of fields of intervention. To improve the productive efficiency of the firms, in developing countries, there is a prime need to transform ‘craftbased’ firms into ‘industrially-organized’ firms, especially for smaller enterprises operating in low-tech industries. A ‘context-sensitive’ provision of real services in areas such as management, marketing, quality control and technological ‘awareness’ (before ‘upgrading’) could help firms to generate business intelligence so as to activate evolutionary learning within them. Generating conditions for ‘access’ to the exploration and exploitation of new sources of tacit and codified learning and practical knowledge (that is, amongst peers, through visits to trade fairs and centres of expertise or via the establishment of joint ventures, nationally and internationally) becomes a crucial policy need, not least so that firms are able to adapt to or to anticipate future business settings. The small minority of firms that have developed competitive advantage via evolutionary learning could become role models of best practice and learning gatekeepers for the rest of the firms. The policy challenge for developing economies in this sense is immense. Hence it is vital to recognize that, without the theoretically ‘old-fashioned’ and ideologically ‘unsound’ financial backing, policy receipts will lack a key ingredient for their realization. In the area of promoting the development of economies of association, the emphasis on distanced or deterritorialized forms of learning (via ‘virtual’ links) should note the fact that the firms (particularly smaller ones) in developing countries rely on locally-based and often institutionalized channels of ‘access’. However, industrial policy could act as a lever for localities and regions to transform themselves into ‘launching pads’ for firms’ distanced learning. In turn, cluster and proximity-based policies work where economies of scale are relevant (provision of specialized information and training, travel to international trade fairs and so on). But this is rare, and one way of building critical mass might

336

International handbook on industrial policy

be via harnessing networking beyond the local/regional boundaries and institutional ‘anchors’; hence ‘relational’ instead of ‘geographical’ proximity is what should really matters for policy design. Finally, our analysis makes it clear that policy actors need to generate means to promote institutional learning, capacity building and good governance throughout the state apparatus and the whole institutional configuration of the industrial policy. Equally, advocates of the new policy thinking should return to the critical debates on macroeconomic policy and its links with industrial development and policy. This link is the key to unlocking the full potential of decentralized industrial policies in the developing world. Notes 1.

2.

3. 4. 5.

6.

Leandro Sepulveda would like to thank the YPF-Repsol Foundation, the Institute of Industry – University of General Sarmiento, the Buenos Aires Institute of Entrepreneurial Development (IDEB) – Tigre Institute, the ECLAC – UN Buenos Aires office, and particularly, Francisco Gatto, for their huge support on this project. My gratitude also goes to the Department of Geography at Durham University and the School of Health and Social Sciences at Middlesex University in the UK. These questions are in addition to a list of critical observations that have recently been raised against the new approach in developed countries (see Amin and Cohendet, 1999, 2002, 2004; Bianchi, 1998; Grabher, 1993; Hudson, 1999; Jessop, 1995, 1997; Lovering, 1999; Markusen, 1996; Oinas, 1999; Staber, 1996). A main source of criticism was the tendency of the new literature to generalize from the experience of a small group of successful stories of bottom-up industrialization, mostly in Western Europe. Prominent cases of industrial districts or clusters were ‘canonized’ and presented as benchmarks for policy action to localities and regions elsewhere around the world regardless of the context and particularities of real places. The exaggerated emphasis on endogenous capacities as the key source of sustainable competitive advantage is a second source of criticism, particularly in light of the increasing globalization of economic activities. Evidence shows that even the most successful cases of ‘place-based’ industrialization (that is, the Italian industrial districts) have experienced trends towards offshore production and outsourcing in lower cost countries, which have helped to weaken their foundations. Continuing success therefore requires an openness towards the influence of external sources of information and knowledge as well as opportunities for learning, which may not be available in the geographical proximity of economic agents. This might include thinking about ‘economies of distance’, for instance in key business aspects such as marketing, subcontracting, training and R&D. To flourish in the market-place the most competitive firms can no longer rely upon established routines, localized tacit forms of knowledge, social capital and incremental innovation only. They need to combine localized competitive assets with codified forms of knowledge and more radical methods of learning, unlearning and innovation which are generally ‘available’ beyond the immediate geography of firms. Firms can in fact link and network virtually or physically (that is, via Internet, teleconferences or travelling abroad) by using multiple forms of proximity, including relational, cultural and cognitive proximity (Amin and Cohendet, 2004). Consequently, traditional business or industrial cultures and institutional set-ups embedded in context-specific social relations can turn into barriers to change. Institutional thickness is never that far from institutional sclerosis, especially in changing economic circumstances. Finally, the new literature fails to consider the relationship between local strategies and wider institutional settings, notably the terms and conditions of macro-economy policy (interest and exchange rates, import duties, competition policy), which is the domain of the national state. Macroeconomic conditions rule the overall business environment in which firms operate and therefore may define the performance of place-based industrialization and industrial policy. Despite this, there has been a manifest lack of conceptual debate in the new literature on the role that the nation state plays (or could play) in promoting decentralized IP schemes. Primary sources of information utilized in this section are based on an intensive survey of firms and institutions in Buenos Aires region by Leandro Sepulveda between March and October 2001 for his doctoral thesis at Durham University. The main programmes of liberal reforms were implemented between 1976–1983 and 1991–2001, under both military and democratically elected governments, respectively. Convertibility: the currency-board scheme established in 1991 that fixed the peso by law and limited the money supply to the stock of hard-currency reserves. From 1990 to 1995, the country’s GDP grew by 5.9 per cent a year; hence the IMF hailed Argentina as the role model for reform in emerging markets. Convertibility’s main success (until its collapse in 2001) was that it put an end to hyperinflation. Indeed, four continued years of economic recession (1998–2002), which ended with an economy 20 per cent smaller, the state highly indebted, the bank system collapsed, the consequent collapse of the

Decentralizing industrial policies 337

7. 8.

9. 10. 11. 12.

peso-dollar convertibility (1991–2001), and the declaration of the largest default in modern world economic history (U$132 bn); the political turmoil of December 2001, which ended with thousands on the streets, the fall of the elected government and the subsequent designation of five nominal heads of state in ten days; and the critical social situation, with 53 per cent of the population falling below the poverty line, 20 per cent no longer affording food in a food export country, and unemployment and underemployment figures affecting around 50 per cent of the workforce, marked the entrance of Argentina into the new millennium (Rock, 2002). Buenos Aires province is the Argentina’s richest province, representing a third of GDP. Tigre is situated 32 km to the north of Buenos Aires centre within Greater Buenos Aires. It represents in part a locality inside a larger urban metropolis (where the large majority of firms are located in Argentina) and also a region outside of it. The reasons that make Tigre an interesting case study for testing on the ground the relevance of the new IP thinking can be summarized as follows: In Tigre (and in neighbouring San Fernando) there is a critical mass of manufacturing firms (630 firms in total) which account for a large part of Tigre’s GDP. An important number of firms, mostly SMEs, belong to the timber and wood furniture industry. For their geographical proximity, number and sectoral composition these firms constitute (broadly speaking) an industrial agglomeration or cluster. Owing to sectoral specificities associated with this industry, a theoretical possibility exists for specialization and division of labour amongst firms. Tigre’s institutional set-up is shaped by a considerable number of public, intermediate and private institutions in theory concerned with local firm competitiveness. Finally, Tigre, unlike many localities, showed a vibrant economic performance during the first 1990s. As regards the methodology, four different types of economic actors were considered: a sample of 21 SMEs (five to 100 employees); 10 local public and private institutions of the Tigre’s region; a sample of 10 provincial and national institutions concerned with industrial policy issues; and five key informants and experts in industrial economy. The interviews were conducted in 2001, during one of the most critical moments of the economic recession. Mr Ubieto (the mayor) and his party (Frente Vecinal de Tigre) have won three consecutive elections with 70 to 80 per cent of the total votes in each local election. The only exception was the ‘automotive regime’ which was intended for the motor vehicle industry. Namely, U$34 453 millions, rising by 17 per cent compared to 2003.

References Altenburg, T. and J. Meyer-Stamer (1999), ‘How to promote clusters: policy experiences from Latin America’, World Development, 29(9), 1693–1713. Amin, A. (1999), ‘An institutionalist perspective on regional economic development’, International Journal of Urban and Regional Research, 23(2). Amin, A. and P. Cohendet (1999), ‘Learning and adaptation in decentralised business networks’, Environment and Planning D: Society and Space, 17, 87–104. Amin, A and P. Cohendet (2002), ‘Knowledge practices, communities and competences in firms’, mimeo, Department of Geography, University of Durham. Amin, A and P. Cohendet (2004), Architectures of Knowledge: Firms, Capabilities and Communities, Oxford: Oxford University Press. Amin, A. and N. Thrift (1992), ‘Neo-Marshallian nodes in global networks’, International Journal of Urban and Regional Research, 16, 571–87. Amin, A. and N. Thrift (eds) (1994), Globalization, Institutions, and Regional Development in Europe, Oxford: Oxford University Press. Audretsch, D. (1998), ‘Introduction’, in D. Audretsch (ed.), Industrial Policy and Competitive Advantage,Volume I, The Mandate for Industrial Policy, Cheltenham, UK and Northampton, MA, USA: Edward Elgar. Azpiazu, D., E. Basualdo and M. Khavisse (1986), El Nuevo Poder Económico en la Argentina de los años 80’, Buenos Aires: Editorial Legasa. Bagnasco, A. (1988), La Costruzione Sociale del Mercato, Bologna: I1 Mulino. Becattini, G. (1979), ‘Dal settore industriale al distretto industriale: alla ricerca dell’unita d’analisi dell’economia industriale’, Rivista di Economia e Politica Industriale, 1, 7–21. Becattini, G. (1990), ‘The Marshallian industrial district as a socio-economic notion’, in F. Pyke, G. Becattini and W. Sengerberger (eds), Industrial Districts and Inter-firm Cooperation in Italy, Geneva: International Studies for Labour Studies. Becattini, G. (2002), ‘Industrial sectors and industrial districts: tools for industrial analysis’, European Planning Studies, 10(4), 483–93. Begg, I. and G. Mayes (2000), ‘The case for decentralised industrial policy’, European Planning Studies, 8(2). Begg, I., M. Landsbury and G. Mayes (1995), ‘The case for decentralised industrial policy’, in P. Cheshire and R. Gordon (eds), Territorial Competition in an Integrating Europe, Aldershot, UK and Brookfield, VT, USA: Avebury.

338

International handbook on industrial policy

Best, M. H. (1990), The New Competition: Institutions of Industrial Restructuring, Cambridge: Polity Press. Bianchi, P. (1994), ‘Industrial strategy in an open economy’, in P. Bianchi, K. Cowling and R. Sugden (eds), Europe’s Economic Challenge. Analyses of industrial strategy and agenda for the 1990s, London: Routledge. Bianchi, P. (1997), ‘Construir el Mercado. Lecciones de la Unión Europea: el desarollo de las instituciones políticas de la competitividad’, Universidad Nacional de Quilmes, Argentina. Bianchi, P. (1998), ‘Requiem for the Third Italy? Rise and fall of a too successful concept’, Entrepreneurship & Regional Development, 10. Bianchi, P. and M. Giordani (1993), ‘Innovation policy at the local and national levels: the case of EmiliaRomagna and its industrial districts’, European Planning Studies, 1, 25–45. Bianchi, P., K. Cowling and R. Sudgen (1998), Industrial policies and economic integration: learning from European experiences, London and New York: Routledge. Bisang, R. (2000), ‘The responses of National Holding Companies’, in B. Kosacoff (ed.), Corporate strategies under structural adjustment in Argentina: responses, by industrial firms to a new set of uncertainties, Oxford: St Antony’s Series. Boekema, F., K. Morgan, S. Bakkers and S. Rutten (2000), Knowledge, Innovation and Economic Growth: The Theory and Practice of Learning Regions, Cheltenham, UK and Northampton, MA, USA: Edward Elgar. Boisier, S. (2002), ‘What if development is really the emergence of a system?’ mimeo, ILPES/ECLAC/UN Santiago, Chile. Bonvecchi, C. and F. Porta (2003), ‘Las condiciones de consistencia micro/macroeconómica’, ECLAC, UN Estudio, 1.EG.33.2 in Estudios Macroeconómicos, March. Braczyk, H-J., M. Heidenreich and P. Cooke (eds) (1998), Regional Innovation Systems: the Role of Governance in a Global World, London: UCL Press. Brusco, S. (1982), ‘The Emilian model: productive decentralisation and social integration’, Cambridge Journal of Economics, 6. Brusco, S. (1989), ‘A policy for industrial districts’, in E. Goodman and J. Bamford (eds), Small Firms and Industrial Districts in Italy, London: Routledge. Brusco, S. and E. Righi (1989), ‘Local government, industrial policy and social consensus: the case of Modena (Italy)’, Economy and Society, 18, 405–24. Camagni, R. (1991), ‘Local “milieu”, uncertainty and innovation networks: towards a new dynamic theory of economic space’, in R. Camagni (ed.), Innovation Networks: Spatial Perspectives, London and New York: Belhaven Press. Camagni, R. (2002), ‘On the concept of territorial competitiveness: sound or misleading?’, Urban Studies, 39(13), 2395–2411. Chudnovsky, D., F. Porta and M. Chidiak (1996), Los límites de la aperture: Liberalización, reestructuración productive y medio ambiente, Buenos Aires: CENIT/Alianza Editorial. Cooke, P. and K. Morgan (1993), ‘The network paradigm: new departures in corporate and regional development’, Environment and Planning D: Society and Space, 11, 543–64. Cooke, P. and K. Morgan (1994), ‘Growth regions under duress: renewal strategies in Baden Württemberg and Emilia-Romagna’, in A. Amin and N. Thrift (eds), Globalization, Institutions, and Regional Development in Europe, Oxford: Oxford University Press. Cooke, P. and K. Morgan (1998), The Associational Economy: Firms, Regions, and Innovation, New York: Oxford University Press. Costamagna, P. and C. Ferraro (2000), Entorno institucional y desarrollo productivo local. La importancia del ambiente y las instituciones para el desarrollo empresarial: El caso de Rafaela, Buenos Aires: ECLAC, UN. ECLAC (1999a), ‘Entorno institucional y desarrollo productivo local. La importancia del ambiente y las instituciones para el desarrollo empresarial. El caso de Rafaela’, mimeo, P. Costamagna, E. Dominguez, C. Ferraro and L. Sepúlveda, ECLAC, UN, Buenos Aires. ECLAC (1999b), ‘Decentralización y apoyo a los sectores productivos en la Provincia de Santa Fé: Evaluación del Programa Municipios Productivos’, ECLAC, UN, Buenos Aires. ECLAC Report (2000), ‘Análisis de la experiencia del Instituto de Desarrollo Bonaerense – IDEB – Documento de Trabajo no 85, ECLAC, UN, Buenos Aires. Evans, P. (1987), ‘Class, state and dependence in East Asia: lessons for Latin Americanists’, in D. Frederic (ed.), The Political Economy of the New Asian Industrialism, Ithaca, NY: Cornell University Press. Evans, P. (1992), ‘The state as problem and solution: predation, embedded autonomy and adjustment’, in S. Haggard and R. Kaufman (eds), The Politics of Economic Adjustment: International Constraints, Distributive Politics, and the State, Princeton: Princeton University Press. Evans, P. (1995), ‘Embedded autonomy: states and industrial transformation’ (Princeton Paperbacks), Princeton: Princeton University Press. Evans, P. and J. Rauch (1997), ‘Bureaucracy and growth: a cross national analysis of the effects of the Weberian state structures on economic growth’, American Sociological Review, 64(5), October. Florida, R. (1995), ‘Toward the learning region’, Futures, 27, 527–36.

Decentralizing industrial policies 339 Freeman, C. (1994), ‘Critical survey: The economics of technological change’, Cambridge Journal of Economics, 18, 463–514. Fujita, M., P. Krugman and A. Venables (2002), The Spatial Economy: Cities, Regions and International Trade, Cambridge, MA: MIT Press. FUNDES-UNGS Report (1999), ‘Los problemas del entorno de negocios: El desarrollo competitivo de las PyMEs argentines’, ed. G. Yoguel and V. Moori-Koening, Buenos Aires: Miño y Dávila Editores. Garofoli, G. (2002), ‘Local development in Europe: Theoretical models and international comparisons’, European Urban and Regional Studies, 9(3), 225–39. Gatto, F. (2003), Estudios de competitividad territorial. In Componente A: Las estrategias productivas regionales, debilidades del actual tejido empresarial, sistema tecnologico, financiero y commercial, ECLAC, UN Estudio, 1.EG.33.5, March, Buenos Aires, Argentina. Gatto, F., R. Ascúa, C. Ferraro and A. Quintar (1993), ‘Rafaela: un cuasi-distrito italiano a la argentina’, Documento de Trabajo no. 35, Programa CFI-CEPAL PRIDE. Grabher, G. (ed.) (1993), The Embedded Firm: On the Socioeconomics of Industrial Networks, London: Routledge. Helmsing, A.H.J. (2001), ‘Externalities, learning and governance: new perspectives on local economic development’, Development and Change, 32, 277–308. Howells, J. (2002), ‘Tacit knowledge, innovation and economic geography’, Urban Studies, 39(5–6), 871–84. Hudson, R. (1999), ‘The learning economy, the learning firm and the learning region: a sympathetic critique of the limits to learning’, European Urban and Regional Studies, 6(1), 59–72. Hudson, R. (2001), ‘What makes economically successful regions in Europe successful?’, Working Paper Series 99, Economic Geography Research Group. Humphrey, J. (1995), ‘Industrial reorganization in developing countries: from models to trajectories’, World Development, 23(1), 149–62. INDEC (1998), Argentina: Statistical Synopsis, Edición 1998, INDEC. INDEC (2005), Informe Económico, Instituto Nacional de Estadísticas y Censos. Jessop, B. (1995), ‘Towards a Schumpeterian workfare regime in Britain? Reflections on regulation, governance, and the welfare state’, Environment and Planning, 27, 1613–26. Jessop, B. (1997), ‘The governance of complexity and the complexity of governance: preliminary remarks on some problems and limits of economic guidance’, in A. Amin and J. Hausner (eds), Beyond Market and Hierarchy: Interactive Governance and Social Complexity, Cheltenham, UK and Lyme, USA: Edward Elgar. Katz, J. and B. Kosacoff (1989), El proceso de industrialización en la Argentina, evolución, retroceso y prospectivas, Buenos Aires: Centro Editor de América Latina. Keating, M. (1998), The new regionalism in Western Europe: Territorial restructuring and political change, Cheltenham, UK and Lyme, USA: Edward Elgar. Kosacoff, B. (1998), ‘Estrategias empresariales en tiempos de cambio. El desempeño industrial frente a nuevas incertidumbres’, Universidad Nacional de Quilmes/CEPAL. Kosacoff, B. (ed.) (2000), Corporate strategies under structural adjustment in Argentina: Responses by industrial firms to a new set of uncertainties, Oxford: St Antony’s Series. Krugman, P. (1995), Development, Geography and Economic Theory, London: MIT Press. Krugman, P. (1998), ‘What’s new about the New Economic Geography?’, Oxford Review of Economic Policy, 14 (2), 7–17. Lambooy, J. (2000), ‘Learning and agglomeration economies: adapting to differentiating economic structures’, in Boekema et al. (eds.) Knowledge, Innovation and Economic Growth: The Theory and Practice of Learning Regions, Cheltenham, UK and Northampton, MA, USA: Edward Elgar. Lovering, G. (1999), ‘Theory led by policy: The inadequacies of the “new regionalism” (illustrated from the case of Wales)’, International Journal of Urban and Regional Research, 23, 379–95. Lundvall, B-A. (ed.) (1992), National System of Innovation: towards a Theory of Innovation and Interactive Learning, London: Pinter. Lundvall, B.A. and B. Johnson (1994), ‘The learning economy’, Journal of Industry Studies, 1, 23–41. Markusen, A. (1996), ‘Sticky places in slippery space: a typology of industrial districts’, Economic Geography, 72, 293–313. Marshall, A. (1890), Principles of Economics, London: Macmillan. Marshall, A. (1919), Industry and Trade, London: Macmillan. Martin, R. and P. Sunley (1996), ‘Slow convergence? Post-neoclassical endogenous growth theory and regional development’, working paper 44, ESRC Centre for Business Research, University of Cambridge. Martin, R.L. and P. Sunley (1998), ‘Slow convergence? The new endogenous growth theory and regional development’, Economic Geography, 74, 201–27. Martin, R.L. and P. Sunley (2003), ‘Deconstructing clusters: chaotic concept or policy panacea?’, Journal of Economic Geography, 3, 5–35.

340

International handbook on industrial policy

Maskell, P. and A. Malmberg (1999), ‘The competitiveness of firms and regions: “Ubiquitification” and the importance of localised learning’, European Urban and Regional Studies, 6, 9–25. Maskell, P., H. Eskelinen, I. Hannibalsson, A. Malmberg and E. Vatne (1998), Competitiveness, Localised Learning and Regional Development. Specialisation and Prosperity in Small Open Economies, London: Routledge. Maxfield, S. and B.R. Schneider (eds) (1997), Business and the State in Developing Countries, Ithaca and London: Cornell University Press. Morgan, K. (1997), ‘The learning region: institutions, innovation and regional renewal’, Regional Studies, 31(5), 491–503. Morgan, K. and C. Nauwelaers (eds) (1999), Regional Innovation Strategies: The Challenge for Less-favored Regions, Regions, Cities and Public Policy Series, London: The Stationery Office. Nadvi, K. (1992), ‘Flexible specialization, industrial districts and employment in Pakistan’, working paper no. 232, ILO World employment Programme, Geneva. OECD (1999), Boosting Innovation: The Cluster Approach, Paris: OECD. OECD (2000), Knowledge Management in the Learning Economy: Education and Skills, Paris: Organisation for Economic Cooperation and Development. OECD (2001), Innovative Cluster: Drivers of National Innovation Systems, Paris: OECD. Oinas, P. (1999), ‘Activity-specificity in organizational learning: implications for analysing the role of proximity’, GeoJournal, 49, 363–72. Pedersen, P.O., A. Sverisson and M.P. van Dijk (eds) (1994), Flexible Specialization: The Dynamics of Smallscale Industry in the South, London: Intermediate Technology. Piore, M. and C. Sabel (1984), The Second Industrial Divide, New York: Basic Books, Inc. Publishers. Porter, M. (1990), The Competitive Advantage of Nations, London: Macmillan. Porter, M. (1994), ‘The role of location in competition’, Journal of Economics of Business, 1(1), 35–9. Porter, M. (1998a), On Competition, Cambridge, MA: Harvard Business School Press. Porter. M. (1998b), ‘Location, clusters and the “new” microeconomics of competition’, Business Economics, 33(1), 7–17. Porter, M. (1998c), ‘Clusters and new economics of competitiveness’, Harvard Business Review, December, 77–90. Porter, M. (2001), ‘Regions and the new economics of competition’, in A. Scott (ed.), Global City Regions, Oxford: Blackwell, pp.139–52. Pyke, F., W. Sengerberger and G. Becattini (eds) (1990), Industrial Districts and Inter-firm Cooperation in Italy, Geneva: International Institute for Labour studies (ILO). Rhodes, R. (1990), ‘Policy networks: a British perspective’, Journal of Theoretical Politics, 2, 293–317. Rhodes, R. (1997), Understanding Governance: Policy Networks, Governance, Reflexibility and Accountability, Philadelphia: Open University Press. Rock, D. (2002), ‘Racking Argentina’, New Left Review, 17, Second Series, Sept./Oct., 54–86. Saxenian, A. (1994), Regional Advantage: Culture and Competition in Silicon Valley and Route 128, Cambridge, MA: Harvard University Press. Schmitz, H. (1990), ‘Small firms and flexible specialisation in developing countries’, Labour and Society, 15 (3). Schmitz, H. (1999), ‘Global competition and local cooperation: success and failure in the Sinos Valley, Brazil’, World Development, 29(9), 1627–50. Schmitz, H. and J. Humphrey (1996), ‘The triple C approach to local industrial policy’, World Development, 24(12), 1859–77. Schmitz, H. and K. Nadvi (1999), ‘Clustering and industrialization: introduction’, World Development, 29(9), 1503–14. Scott, A. (1988), New Industrial Spaces. Flexible Production Organization and Regional Development in North America and Western Europe, London: Pion. Scott, A. (1995), ‘The geographic foundations of industrial performance’, Competition and Change, 1. Scott, A. (1998), Regions and the World Economy: The Coming Shape of Global Production, Competition, and Political Order, Oxford: Oxford University Press. Scott, A. and M. Storper (1995), ‘The wealth of regions: market forces and policy imperatives in local and global context’, Futures, 27, 505–26. Sepulveda, L. (1999), ‘Construyendo política industrial. Enfoque socioeconómico aplicado a un estudio regional’, (MA thesis) Universidad Nacional de General Sarmiento (UNGS) Buenos Aires, Argentina. Sepulveda, L. (2001), ‘Construcción regional y desarrollo productivo en la economía de la globalidad’, Serie Estudios y Perspectives, no. 3, ECLAC, UN, Buenos Aires, Argentina. Sepulveda, L. (2004), ‘The relevance of new industrial policy thinking to Argentina’, (PhD thesis), University of Durham, UK. Staber, U. (1996), ‘Accounting for variations in the performance of industrial districts: the case of BadenWürttemberg’, International Journal of Urban and Regional Research, 20, 299–315.

Decentralizing industrial policies 341 Storper, M. (1994), ‘Institutions of a learning economy’, mimeo, School of Public Policy and Social Research, UCLA, Los Angeles. Storper, M. (1997), The Regional World: Territorial Development in a Global Economy, New York: Guilford Press. The Economist (2004), ‘The long road back. A survey of Argentina’, 5 June. World Development (1999), Special Issue: Industrial Clusters in Developing Countries, 29(9). Yoguel, G. and F. Boscherini (1999), ‘La capacidad innovative y el fortalecimiento de la competitividad de las firmas y el rol del ambiente: el caso de firmas industrials argentines pertenecientes a distintos sistemas locales: Evaluacion de hipotesis en diversos sistemas locales’, (mimeo), UNGS, Buenos Aires, Argentina.

17 The local dimensions of industrial policy Marco Bellandi and Marco Di Tommaso1

1 Introduction This chapter is about the local dimension of industrial policy. The first question is: has industrial policy a local dimension and what does it exactly mean? We will see that a positive answer may be inferred from some theoretical debates (on market and government failures, their remedies and possible policy answers), and from what we may learn from the history of industrial development with reference to a variety of experiences around the world. Industrial policy with a local dimension does not reduce to intervention promoted by local government. If local governments have a central role in defining and implementing local industrial policy, other institutions can also offer an important contribution in this field, for example the national government policies or the international institutions’ interventions. A clearer identification of the local dimension of industrial policy should be related to its focus on local actors and relations: ‘local’ because rooted in localities identified by sets of relations within specific communities of people, firms and institutions. The point is discussed directly in section 5. A second question concerns the domain of Local Industrial Policy (LIP), here defined as the application of general principles of industrial policy to the organization of industry at the just mentioned level of localities. Looking at the wide and complex literature on industrial policy, we consider, in sections 2 and 3, the possible role of LIP in correcting market failures, and in promoting strategic and meta-economic objectives. A related issue concerns the impact of government failures in reducing the domain of LIP, and the possible remedies to such failures. According to some views, local government policy and upper-levels government intervention in favour of local areas can be improved, for example throughout appropriate devolution programmes, mitigating the effect of their traditional failures (section 4). A third general question concerns the requisites for an effective application of LIP. The presence of evolutionary processes rooted at the local level (local actors and local relations) may be seen as the stable basis for the application of LIP. This view is illustrated by two combined sets of results of the literature on industrial districts, local production systems and localities of industry. The results are recalled in sections 6 and 7, and concern, first, the central role of public bases of industrial efficiency and development; and, second, the multi-layered but embedded nature of an effective supply of such bases. The consideration of these aspects opens the way to the definition of industrial policy with a rich local dimension. It is LIP at its real roots. We conclude the chapter (sections 8 and 9) by summarizing and trying to relate the previous discussion to the application of LIP in different types of historical, geographical and sectoral conditions. 2 Market failures and LIP Today the market is considered the main pillar of contemporary capitalism. Building on a suggestion by Adam Smith, economic theory has identified models of the free functioning 342

Local dimensions of industrial policy 343 of the market as the locus where rational agents interact, maximizing their own individual interest. And the relations between some market forms and collective efficiency have been defined in a rigorous if limited way. However, in some circumstances markets fail and in these cases corrective interventions become necessary (Stiglitz, 1988). Market failures, related to conditions such as public goods, externalities, limited competition or incomplete markets, are common phenomena. The first question discussed in what follows is whether LIP has a role in the provision of solutions to some of these failures. 2.1 Public goods In the case of public goods, characterized by non-rivalry and non-excludability, the problem is that, since a good is available without paying for it, there is no incentive for individuals to reveal how much they would be willing to pay and consume. This brings about an underprovision of public goods. Some public goods are crucial for local development (transport, communication, justice and so on) and some public goods have a clear local dimension (knowledge, taste, culture, reputation). The point is discussed thoroughly in section 6. In general terms, it is important to recall that public good considerations should not automatically suggest public intervention, neither are they enough in the case of LIP. Of course in some cases public intervention could be requested: for example where government power of coercion in taxation helps in solving undesirable problems of underprovision of public goods. However, a realistic picture suggests that private answers are quite common and in many circumstances they are successful. In this framework, an updated perspective focuses on the combination of private and public actions. In particular, as we will suggest in the next pages, governments can play an important role in encouraging the development of local relationships among private actors, working on their attitude in relation to the provision of local public goods. 2.2 Externalities The market mechanism leads to undesirable outcomes if negative or positive externalities are relevant: respectively, overproduction of negative spillovers or underproduction of positive ones. Here again LIP may play a specific role. Any community of people, firms and institutions locally concentrated generates spillover effects that are not compensated appropriately through market exchanges. Government (not only local) could intervene in sanctioning local negative externalities, directly providing goods with local positive externalities or indirectly encouraging their production through subsidy policy. This is clearly the case of pollution, as the most quoted example of negative externality. This is also the case of positive externalities related to public goods (such as health, education and research). Government can act less directly by highlighting the benefit of some positive externalities and thus encouraging the development of joint actions. Joint actions may be the appropriate private answer to the externality failure. However joint actions are fragile mechanisms that may need further policy answers: as we will see in the next sections, joint actions tend to produce further externalities, inhibiting desirable paths of growth. 2.3 Restricted competition Non-competitive conditions emerge in markets for several reasons. In some cases monopoly is justified on the ground of efficiency. Natural monopolies are common cases related to the existence of internal economies of scale and a limited size of the market or to the

344

International handbook on industrial policy

scarcity and the specificity of a particular input. The deadweight loss problem should justify public intervention to guarantee the optimal outcome that is, the one that would have resulted in a competitive scenario. Of course here the debate is wide and the positions can be different according to different schools of thought. The continuum of policy answers ranges from public monopoly to contestable-market arguments. The local impact of this kind of market failure poses a specific problem. Concentration of economic and political power in private local hands may be non-desirable also merely in terms of efficiency criteria. The risk is that the dependency of a local community on a limited number of large actors could stop the emergence of other potential actors and make too rigid the socioeconomic relations within the local community. This rigidity may become an obstacle to the necessary flexibility and creativity that open markets seem to require. 2.4 Incomplete markets Another common feature of economies is that important markets are not working. In other words, the existing market structure may fail to provide an adequate range of goods and services. Insurance and loans markets are the typical examples. At local level the gaps may be more or less extended, with important effects. For example, the unavailability of some specific goods and services (spare parts and components, transport, consultancy and so on) makes difficult the development of the local industry. Here again policy solutions range from the public production of some goods and services to less direct interventions that have the goal of encouraging private answer to fill the gaps. 3 Beyond market failures: other rationales for LIP Another rationale in favour of industrial policy refers to arguments that go beyond efficiency and the traditional perspective of conventional economic theory. The academic debate and experience around the world (Chang, 1994, 2002a, 2002b; Stiglitz, 2001; Amsden, 1989, 1994, 2003) show additional justifications based on strategic objectives, such as employment and national competitiveness, or based on what we define as metaeconomic objectives (Di Tommaso and Schweitzer, 2005), such as distribution of wealth and economic opportunity or access to specific goods and services. 3.1 Strategic objectives Important lessons are suggested by the historical experiences of industrial development in successful economies, such as France, Ireland, Italy, Japan, South Korea and China. In many cases in these countries industry has been promoted for reasons that go beyond the correction of market failures. The governments tried to protect infant industries, defend declining sectors, or support ‘national champions’.2 In the past, strategic sectors have included coal, steel, textiles, automobiles and construction. Today, the sectors that are considered strategic in highly performing countries (and in many emerging countries) are new industries, including software, telecommunications, pharmaceuticals, aerospace, genetics or biotechnology (Di Tommaso and Schweitzer, 2005). In these cases, justifications for governmental intervention go beyond issues of market failure discussed above. Here the question is: has local industrial policy also a strategic dimension? The answer is positive, for two reasons. First, localities may define their own industrial strategy. For example, some local communities and their governments decide to encourage the investment in particular sectors or actors, or try to offer solutions for the crises of local actors,

Local dimensions of industrial policy 345 or consider it necessary to foster local ‘championship’ because in some strategic fields the dependency from non-local actors is not desirable. Second, local government, and more generally the local community, may provide bottom-up inputs and suggestions in order to encourage the achievement of non-local industrial policy goals (Bianchi, 1998). 3.2 Meta-economic objectives If we look at industrial development from an historical perspective, it is clear that in many countries – developed, developing and in transition – industry has also been favoured or promoted for ‘meta-economic’ reasons.3 Industry is seen as an ‘intermediate object’ driving the achievement of complex social goals, including for example access to fundamental goods such as knowledge, education or health, specific distribution of wealth, and social or environmental sustainability. Moreover policy intervention could encourage the provision of ‘merit goods’, deter ‘de-merit goods’ and regulate some market-type transactions that are considered not acceptable. Industry is seen as a vehicle to achieve the development of nations, region or localities, where this status of development is a goal defined by each specific community, and not an objective towards which all societies naturally tend (Myrdal, 1970; Seers, 1972; Hirschman, 1981; Sugden et al., 2006; Sweeney, 1999). From this perspective development is a normative concept, evaluated (rather than precisely measured) in terms of sets of quantitative and qualitative parameters, defined according to the values rooted in any specific community. The promotion of industry is a way through which a particular community of people can promote its model of development (Di Tommaso, 2006). Thus industrial policy has also a role to play in promoting development, and local communities have a role in codefining their own policy objectives. This approach has an impact on the relationship between market failures and industrial policy interventions. It gives additional rationales to industrial policy because community-based demands may go beyond market failure corrections. Also the need to respond to market failures can be strongly linked to what the local community defines as valuable. The underprovision of specific public goods, the presence of positive externalities, or the absence of a particular market may be considered acceptable in some locality but not in others. We come back to this perspective in section 7. 4 Government failures, remedies and the role of LIP So far we have illustrated what are the main reasons that justify LIP. However, other considerations reduce its extent. In general terms, government action and industrial policy may be unsuccessful, so the choice seems to be between a failed market (in the traditional meaning but also failing in terms of its capacity to promote strategic and meta-economic objectives) and a failed government intervention. The traditional argument in this field runs as follows: government policy failures may cost more than the benefit the intervention is supposed to offer, and this is why it may be preferable to accept market failure than the consequences of failed government remedies. In what follows in this section we discuss this thesis and offer specific considerations on LIP. 4.1 Sources of government failures Many theories have implicitly assumed that governments are able to pursue their goals. However, theoretical analysis and empirical evidence show that public actors – at local,

346

International handbook on industrial policy

regional, national or international levels – can fail to achieve their objectives for several reasons, or can divert them from public interests to private ones. 4.1.1 Internal forces: focusing on the government as a ‘black box’ A first source of government failure is the disjunction between costs and revenues. Typically the revenues that sustain government interventions are derived from sources unrelated to the cost of production. This is a cause of redundant and rising costs. And if technological possibilities exist for lowering cost functions or raising productivity, these opportunities are likely to be unexploited. A second often mentioned source of government failure is that bureaucracy pursues internal goals that may be different from the ones that it is supposed to have. In the absence of market indicators such as consumer behaviour, market shares and profit-and-loss bottom line, public agencies develop their own parameters to guide, regulate and evaluate their performance. These internal parameters are the ones which provide the motivation behind individual and collective behaviour within the agency and constitute its real agenda. Moreover, it has been said that bureaucrats are not different from the other types of homo oeconomicus in pursuing their own interest. There is no reason to suppose that bureaucrats always behave altruistically. It is more plausible to assume that, since they derive utility from higher salaries, prestige, office perks, promotion and greater power of their bureaux, they look for these results also to the disadvantage of social output. 4.1.2 External forces: focusing on the society The arenas of politics, at international, national and local levels, are where various interests interact and contribute to the definition of general and specific policy through numerous mechanisms. This is the case also of industrial policy. Different groups have different weight and capacity to express their ‘voice’. These differences interfere with the policy capacity of achieving public interest, both during the phase of definition of policy targets (as for example the investment in particular areas, sectors or firms) and more covertly later, during the implementation of the policies. Diversion of governments and bureaucrats from public interest results when mechanisms able to immunize policy against the influence of partial interests do not intervene. 4.1.3 Rent seeking The theory of rent seeking goes even further (Buchanan et al., 1980). Additional losses occur when resources are deviated into unproductive activities by private agents in order to capture rents generated by public intervention. Subsequently, when one agent gets the rent there are no forces (as in the case of the market) able to ensure its dissipation. Thus two problems emerge: (1) the waste of resources during the process of rent seeking that otherwise could be utilized elsewhere in the economy; (2) rent created by the government intervention does not allow free entry and there are no forces leading to its dissipation. 4.1.4 Information problems Among the conditions bringing about government failures, the problem of information is paramount. A state needs information in order to decide among the range of possible policies. Then, once it chooses one specific policy, it needs information again in order to control the implementation of these policies by the

Local dimensions of industrial policy 347 lower-level bureaucrats and to monitor the reaction of the groups at whom the policy is aimed. All these activities of collection and process of information imply a huge amount of resources. Moreover, even if we accept this cost, the information is likely to remain insufficient and the final outcome modest. Another dimension of the information problem is related to the principal–agent asymmetry. There are two kinds of asymmetries involved in government intervention: the one between top decision makers and the lower-level bureaucrats, and the one between the state and the policy target groups. In the first case the supposed asymmetry implies a difficult, if not impossible, top-down control and a general low level of communication and coordination between bureaucrats. The second case leads to a substantial ignorance of that reality that the government would like to correct. Then, if we add that, as recalled above, the bureaucrats have individual preferences not necessarily consistent with collective objectives and that any government agency is guided by its own objectives, failure should be considered at least probable. 4.2 Government failures: remedies and the role of LIP The study of government failures’ remedies has become an important part of the general debate on industrial policy. Before accepting failed markets because governments fail, policy scientists and makers have to investigate how to improve policy efficiency and effectiveness through the elaboration of innovative solutions (Chang, 1994). Remedies are studied also with reference to the local dimension of industrial policy. It is possible to suggest that LIP can play an important role in offering solutions to government failures. The point lies in the understanding that we are not facing only principal–agent problems (Barca, 2004). The problem cannot be reduced to the fact that top bureaucrats are not able to monitor lower colleagues, or that policy institutions have problems in controlling their target entities. On the contrary, the challenge is also how to define the goals of LIP rooting the interventions on local community knowledge and demands. The issue is how to stimulate the identification of the needs and how to guide the policy implementation exploiting the stock of knowledge present (and often hidden) within national governments and local governments and communities: that is, how to identify local scattered poles of knowledge, how to include marginal actors, how to foster the linkages with non-local partners, stimulating joint actions able to define a ‘participated project’ of industrial development. Here LIP plays a role: at the local level, stimulating the actors to share and develop their collective knowledge; at the national level, developing the knowledge of local realities and latent opportunities. This is one of the main lessons learnt from the international literature on successful clusters, for example those characterizing the local development of many industrial districts in Italy in recent decades.4 Here industrial policy appears clearly related to a local dimension of efficiency in organization of industries and development. The survey of that literature helps in the identification of what is meant by ‘local dimension’ of industrial policy and in the definition of conditions of effective application of LIP. 5 Local units of investigation and action The study of the organization of industry in a territory and of processes of industrial development in such territory takes two different units of analysis (Becattini, 1989; Brusco, 1986; Russo, 1996): the industry, with its organization and territorial characters;

348

International handbook on industrial policy

and the locality, with its socioeconomic characters and evolutionary processes. The first unit includes different types: 1.

2.

3.

A generic industry, that is a set of productive activities with common technological and market characters, shared by key producers (entrepreneurs, managers, skilled workers and technicians, service providers), without a necessarily common territorial basis. A generic cluster, as the agglomeration of a set of productive entities and organizations, operating within a field of economic activity, possibly extending to related activities; and explaining the sectoral specialization (or one of the specializations) of the locality where the cluster is located. A local production system, that is, a cluster with systemic characters (added to the characters in (2), where sets of specialized firms run complementary activities; the producers are locally organized by way of both market mechanisms and non-market mechanisms; the organization and development of the cluster is supported by the local fabric of social and civic life.

Also the second unit includes different types. 4. 5. 6.

The territory under a government: a municipality, a county and so on. A locality with a socioeconomic identity, corresponding to a set of contiguous towns, villages, rural areas, with a principal town or city. A locality of industry, that is, a locality characterized in social and economic terms by the presence of one or a few local production systems whose systemic characters are embedded in the local fabric of social and civic life (for example, an industrial district).

Some local production systems are characterized by many specialized and small-tomedium sized firms, others by a dominant role of local or external big firms. Some are characterized by a core of manufacturing competencies, others by a core of R&D and project competencies. Some clusters extend to different localities, some localities do not have clusters but only a mix of parts of many different industries, others are localities of industry. In the loosest applications of the generic concept of cluster, the ‘locality’ may be identified with whatsoever administrative area: a municipality, a county, a region, a nation state. And the cluster may be seen as just the result of converging external forces, in particular strategic and market forces.5 When industrial clustering is explained by external forces only, the result (if any) is industrial organization and development happening in this or that locality. It may be interesting but it does not show levers for an affective application of LIP. Instead, industries that evolve in historical time with the life of a locality present more promising requisites. A strong link between an industry and a set of localities supports (and is supported by) the presence and the evolution of public bases of local industrial efficiency and development. Furthermore the constitution and the development of new industries may result from the overlapping of innovation promoted by the forces rooted in the locality and the strategies of localization led by external forces. The power of action on such public bases supplies a specific lever to LIP. In what follows, we illustrate how the potential domains of LIP, recalled in sections 2–4, include objects to which the specific levers of LIP can be applied. Finally, the literature

Local dimensions of industrial policy 349 we are going to review tends to expand the lesson and provide suggestions also for less favoured economies, where nonetheless the local dimension seems to offer a way out of the dilemma between blocked and dependent development. 6 Public bases of efficiency and development of localities of industry 6.1 Cooperative nexus Industrial districts are prototypical examples of localities characterized by the economic and social prominence of a cluster of small-to-medium sized firms embedded within the social relations of the area (Becattini, 1989, 2004). The model of the industrial district, following a steady path of development, is centred on the identification of a progressive link between the local production system and the locality: 1.

2.

The social core of the cluster is a set of independent producers sharing a bent towards trust in reciprocal exchanges; a diffuse attitude towards productive (against rentseeking) entrepreneurship, proactivity of workers on the job, cooperative joint action between entrepreneurs; and cognitive proximity, that is the similarity of some basic know-how among the producers. The producers are ‘embedded’ locally, in the sense that, living and working in the district, they develop common subjective tracts (the cooperative nexus). They better understand when one of them can be trusted; they look for the acknowledgment of economic success and innovation within the group; and the sharing of familial, civic, scholastic experiences make it easier to understand and compare the ‘mysteries’ of the industry, as Marshall said.6

Even outside district conditions, it is possible to find clusters that have the nature of local production systems.7 What distinguishes them all from a casual agglomeration of business activities, or from a generic cluster, is that each one is kept together and evolving by some sort of local cooperative nexus.8 The productive exploitation of such nexus may result in what Alfred Marshall termed ‘external economies’: that is, relative advantages that independent producers gain from inclusion within a system of organized division of labour.9 Recent developments in the literature have highlighted the need for an architecture of systemic conditions complementing the cooperative nexus as conditions for the realization of external economies. The need and the presence of this architecture enlarge the definition of the public sphere supporting the industrial efficiency at the local level and the development of localities of industry. 6.2 Systemic conditions, market failures and development Reflections on industrial districts, from Marshall on, suggest that external economies come precisely from an appropriate organization of a set of connected activities and business fields, employing an appropriately large set of complementary and substitute human and technical resources. Organization is needed for integrating the productive contribution of many different specialized resources, but organization is not necessarily the result of the planning activity of a single centre of strategic decision making. Organization also results from sharing systemic conditions that have the nature of public goods (see section 2 above) whose net benefits are specific to the system itself.

350

International handbook on industrial policy

Let us consider some types of systemic conditions. 1.

2.

3.

4.

5.

The realization of advantages of specialization demands the presence of technical and commercial standards allowing the connection of the specialized contributions of the producers. The selection of the standards by means of competitive markets implies the opening of many markets (one for each type of possible connection). Alternatively, the standards come from the action of firms with market power, with the ensuing distortion on allocation. One way or the other, the pure market mechanism has difficulties. The realization of the advantages in the access to a large pool of related skills demands an orientation to educational and training programmes, in schools or on the job, for the young and the working people of the locality. These programmes, though differentiated, should present some common bases, supporting paths of accumulation of human capital within the cluster’s activities, the mobility of skilled employees (increasing allocative efficiency), the capacity to tell a supplier what one needs, or to understand from a supplier the correct use of an input and so on. However, individual choices of investment in education and training within a given set of related business and productive fields are hindered by various effects usually external to existing markets: benefits of consistent choices spill over to other agents, risks of lock-in arise if education and training are too specific, and so on. A third example concerns the advantages of diffuse industrial creativity. They come from the interaction of ideas and original approaches within a (fuzzily defined) field of production and business activity. Innovation is strengthened by convergent investments in new components and competencies. But convergence is hampered by uncertainty of results surrounding innovation (Lane and Maxfield, 2005). A fourth example transversally concerns all the types of external economies, referring to the existence of local labour and product markets that favour the allocation of skills and productive capacities, give a point of reference for developing contractual relations within firms and teams of firms, and support the generation and realization of entrepreneurial ideas by specialized producers. A fabric of well working local markets needs the support of various systemic conditions, of the types recalled before. The same fabric may be considered as a systemic condition, per se, being the not automatic result of a multi-layered combination of those other conditions.10 Similar considerations extend to certain types of tangible goods. For example, within a set of complementary activities localized in a district, there may be one (or more) whose technical process is characterized by a facility with relevant indivisibility or relevant network economies (such as an industrial purifier, or a complex web page storing or linking digital services of various types). The products or services can be sold (that is, exclusion is possible), but minimum optimal scale would not allow more than one supplier in correspondence to the demand from the district. One solution is to look for general national or international markets supplying such products or services. However, it may well be the case that the district’s demand has specific characteristics. On the local side, if the essential facility is run as a private monopoly the economies are appropriated by the supplier. The availability of the essential facility without too heavy monopolistic distortions is a systemic condition. This systemic condition has the nature of a specific club good.

Local dimensions of industrial policy 351 6.

Another class of tangible goods, to which systemic conditions are attached, is represented by networks of local roads, gateways to external transportation and communication networks, public green spaces, local levels of basic security and health, and so on. They have a quasi-public nature. Exclusion is difficult, but they may be congested. High quality, easy access and efficiency in provision (keeping at bay the levels of local taxation and/or the opportunity costs in the use of national public funding) contribute to the overall efficiency of the clusters of a locality.

Such specific public goods may or may not be provided within a cluster of independent producers (Bellandi, 2006). If not, market failures arise (see section 2 above): spillover effects are not governed, and relevant gaps in the local structure of the markets are left unattended. An adequate provision of such public goods seems to ask for central state action, but state provision is often standardized and far away from the cluster’s peculiar needs (Goglio, 1999). An alternative is the internalization of an essential facility as the direct business of a firm, but this produces monopolistic distortions within the relevant part of the cluster (Schmitz, 1999). An extreme private solution is the integration of the cluster under the control of a leading firm: many public goods would be privatized and the independent producers would change status to a dependent one. One way or the other, external economies are not (fully) realized and external diseconomies increase. This restricts the possibilities of development open to the local community. However, in a dynamic and lively locality of industry, the cooperative nexus goes together with social and strategic processes, supporting appropriate forms of constitution and adaptation of the intermediate architecture of specific public goods. As we discuss in the next section, the real core of LIP pursuing industrial efficiency and development at the local level is conscious and intelligent action within and upon such processes. 7 Provision of the public bases: failures and remedies 7.1 Informal institutions First of all, specific public goods (corresponding to systemic conditions) may be governed by informal institutions like customs and conventions, emerging organically from the cooperative nexus (Trigilia, 2005). However, customs and conventions change slowly, through incremental adaptation. They tend to fail when higher challenges demand discontinuous adaptation. Similarly, the use of tangible public goods regulated only by means of customs (the ‘commons’) shows heavy dynamic failures. Here we see the importance of an appropriate collective action (Schmitz, 1999). Henceforth the collective efficiency of successful localities of industry seems to be also strongly related to the crucial role of cooperative behaviours among private agents and of strategic action by policy makers. 7.2 Private joint action Specific public goods resulting from joint private action are usually characterized by excludability and by compensation mechanisms (from exchange of a monetary nature to the exchange of information, technology, know-how and human capital). They are club goods, both tangibles and intangibles. Different types of joint action are defined.11 Here too various shortcomings have to be considered, as when (a) actions are directed to collusive behaviour, slowing down social and economic mobility, and allowing the

352

International handbook on industrial policy

constitution of positions of market and social power within the cluster and the locality;12 or (b), vice versa, the goods produced by joint action are too free, and a large access of free-riders curbs the investments in joint actions;13 or (c) information asymmetries and fear of opportunistic behaviour interfere heavily with the distribution of joint action costs and the access to benefits.14 The tuning of joint private actions towards efficiency and development is favoured of course by the presence of private leaders rooted in the locality, and by the social constraints and motivations given to them by a cooperative nexus. However the shortcomings above have, at their origin, a difficult contradiction: that is, the provision of public bases through the negotiation of private interests. The presence of actors having the mandate, the motivation and the strength for identifying and representing the public interest provides a solution in certain conditions. 7.3 At the core of LIP Local public authorities have daily face-to-face contacts with the socioeconomic context and also a reputation to safeguard in that same context. If the rules of the game are appropriate, they seek to use contextual knowledge in order to devise solutions: public goods, fostering the economic and social growth of the community. In doing so their local prestige grows. An effective local policy is nurtured by the availability of the same progressive features of the local social and economic life that support private individual and collective action: the cooperation of many and diffused centres of knowledge, trust and self-help (Dei Ottati, 1994; Trigilia, 2005). Public action may supplement private joint action through measures (a) guaranteeing, in all the relevant actors, the awareness of the full potentialities of joint actions, and inducing them to contribute to the costs associated with collective actions; (b) lowering the transaction costs related to joint actions; (c) limiting private monopolistic distortions. More directly, public action supplies or regulates the supply of public goods strongly linked to the territory, where the limits of the informal institutions and of private joint action are stricter. Some are goods provided for the specific needs of the local production systems: industrial infrastructures, basic professional training and education, general information and easy access to public administration services, and so on. Here industrial efficiency is supported against market failures. But the other rationales recalled in section 3 are at work too, first of all the strategic one. Public action is particularly important when a bifurcation in the path of development is met, and private actions are weakened by high uncertainty and lack of coordination. The perspectives of positive results are enhanced if scenarios are constructed within a public sphere of interests, ideas on the perspectives are discussed by the local society, and leadership is won on the ability to progress in a direction that presents long-run advantages against short-run risks and costs (Garofoli, 2001). Secondly, internal social changes or mutations of international technology or market trajectories may weaken the cooperative nexus itself (Schmitz, 1999). Considering these complex feedback relations, public policy with a rich local dimension has to be conceived within an integrated (that is meta-economic) frame. For example, political intervention ‘on life in the out-factory sphere’ (Brusco, 1996, pp.155–8) has been an ordinary component of action in dynamic industrial districts, contributing to the processes of economic mobility and social cohesion in the local society, to participation and involvement of workers on the job, and to resolution of conflicts within bilateral economic relations (Dei Ottati, 2003).

Local dimensions of industrial policy 353 Finally, both for the strategic and meta-economic reasons, and for more ordinary indivisibility problems, the success of LIP is favoured by appropriate upper-level public action taken to define and administer fair rules for interindustrial games, to supply general infrastructure, or better to facilitate the constitution of regional public goods specific to different common needs and relations between localities, and to deal with general distributive problems.15 When they play deliberately in this perspective, upper-level governments become proactive parts of LIP. 7.4 LIP, local governance and the risks of their failure The fact that the cluster (specifically, the local production system) may not correspond to a well-defined level of the public administration requires actions by a set of public players (mayors of contiguous municipalities, county representatives and so on). Furthermore, as acknowledged above, other important strategic agencies and leaders – within local business associations and clubs, trade unions, important companies, cultural and religious institutions – intervene in the constitution and management of specific public goods. Lack of contacts between the public authorities promoting LIP and those other actors would contradict the core of LIP as industrial policy with a rich local dimension. Not only has the core of LIP to be defined with the help and the discussion promoted among various stakeholders, but its intermediate and operative aims and contents also need the contribution of such a variety of actors. This is encapsulated in the concept of local governance. Cooperation between public and collective agencies is once again helped by the local cooperative nexus, and by its positive relation with more extended institutional pivots.16 Local governance is neither automatic nor easy, however. The availability of a benevolent and enlightened class of civil servants and policy makers is not granted; and, even when in place, this class has not always the right knowledge for acting effectively.17 The action of public players has to be embedded (Evans, 1996), but if some degree of autonomy from private players is not preserved the risk of economic and cultural capture is high (Rodrik, 2004): too many or too strategic public goods may be given the nature of club goods, rent seeking or rent protection for a cartel of local/external players may become pervasive. On the other hand, of course, too much autonomy means separation. A right evolving balance between embeddedness and autonomy is a fundamental step, easier to be said than to be done. The result also depends on the availability of the right people at the right moment. This may depend on random events or on long traditions. However, an increasing consciousness of this fundamental strategic problem gives better chances; and this consciousness may be favoured by the diffusion of centres of high education on local development and industrial policy with a rich local dimension, and related matters.

THE CORE OF LIP, OR INDUSTRIAL POLICY WITH A RICH LOCAL DIMENSION It concerns: – the provision of architectures of specific public goods supporting industrial efficiency and development at the local level.

354

International handbook on industrial policy

The direct field of application being: – the reproduction and the improvement of the systemic conditions supporting efficiency and equity in management, allocation and distribution within industrial clusters’ division of labour. Larger fields of application, in which such industrial policy combines with other types of policy, being: – the reproduction and the improvement of the systemic conditions supporting learning and creativity within old and new industrial clusters (labour, educational, innovation policies); – the reproduction and the improvement of the systemic conditions supporting an open and dynamic embedment of industrial clusters within the life of localities of industry (labour, educational, social welfare, environment policies), that is, those conditions that give to clusters the nature of local production systems; It consists of: – a deliberate programme devoted to the objects above, – whose final aims are the expression of the local governments’ political strategies possibly pushed or pulled by the territorialized strategies of upper level (for example regional) government; – whose intermediate and operative aims are defined and implemented under the responsibility of public agencies with a political mandate, in a ‘conscious’ and ‘intelligent’ way, that is, through adequate processes of local governance, involving both public and private actors, local and nonlocal.

8 Guidelines for an industrial policy with a rich local dimension Being concerned mainly with architectures of specific public goods, the core of LIP has both a horizontal and a vertical nature (Navarro, 2003; Commission, EU, 2004): horizontal because the public core is meant to define an open stage for sustainable combinations between entrepreneurial projects, social cohesion, environmental respect; vertical, because this core should specifically support the development of progressive links between clusters that are or may become local production systems, and localities that are or may become localities of industry. In this last section a synthesis of the various lines of reflection and debates recalled above is illustrated, and a set of guidelines is proposed. The topics are distributed into two sets: applications to developed regions; applications to new industrialized, transition and underdeveloped regions. 8.1 Applications to developed localities of industry If consciousness (of a right balance between autonomy and embeddedness) is somehow supported, a set of general guidelines for knowledge-based policies in developed areas is worth considering: ●

Identifying the spatial borders of the locality and of its inner territorial structure. Defining the explicit or hidden components of the progressive features (cooperative

Local dimensions of industrial policy 355

●

●

●

●

nexus) of the local social structure, their complementary relations and the conflicting ones. Drawing realistic paths of change, in terms of the inner characters of old and new clusters, considering the evolving relations with the locality, and the external constraints and opportunities.18 Being alert to situations of discontinuity and bifurcation (in which multiple and differing types of paths of change become really possible). In that case, gaining a lead on an idea of development, and envisaging a strategy of actions in order to pull the system towards the type of path opted for (for example, Jacobs, 1969; Crouch et al., 2004). Considering that, when a bifurcation is met, there is a specific problem: ‘institutional thickness’ (see Amin and Sepulveda, above) may become ‘institutional fatness’ (Bellini, below).19 Managing a positive presence of big private players (such as MNEs), selecting incentives, considering the high risk of a loss of autonomy against the strategic and economic power of those players.20

8.2 Applications to less developed regions Another sphere of application is the industrial cluster in transition, underdeveloped and developing economies and regions (Bianchi et al., below; Amin and Sepulveda above): ●

●

●

●

The deposits of productive knowledge embedded into the life experiences of a local community, as the result of past activities, are a hidden source of development potentialities, whose public identification is fundamental for an effective LIP (Brusco, 1996). The presence of an international network of native entrepreneurs having emigrated in the past and coming back to invest in the homeland may add to strictly local factors. Making use of those deposits requires windows of opportunity in the external markets and a big entrepreneurial push in the local context. Failure is a probable result, however, if the local community is characterized by a structure which strongly limits cooperative attitudes within and among the firms. This is often the case, for example, with the caste system; with extreme laissez-faire policies destroying the local ‘industrial public sphere’ (Zeitlin, 1989, p.369); with the predominance of subsidiaries of multinational companies when they are oriented to the pure exploitation of local resources (Nadvi and Schmitz, 1999); and with other types of lock-in conditions (see Isaksen and Hauge, 2002, p.42). The utilization of the deposits of productive knowledge as the technological support for the growth of a set of complementary productive activities allows the creation of a virtuous circle, when the production set is the field for a growing local experience of cognitive interaction. Fundamentally the idea, as in the PACA methodology by Meyer Stamer (2004), is that of trying to promote sets of socioeconomic relations consistent with the logic of local development, hidden within the life of a locality. The problem is not only their recognition, but also to break the way for their expansion with some nontraditional action.

356

International handbook on industrial policy The success of new localities of industry in industrializing countries represents in perspective a great challenge to old localities of industry, but also an opportunity for an enlarged set of exchanges and division of labour.21 The opportunity is increased if systemic conditions for international trans-local collaborative relations are built. They need the support of specific trans-local public goods, from technical and communication standards to trust rules and cognitive proximity. These goods are constituted both by strategic action and cross-cultural fertilization. Trans-local collaborative relations do not cancel out the competitive challenge, but they provide an alternative to games led by MNEs playing localities against localities: relocalization instead of delocalization (Rullani, 2003).

●

9 Conclusions In conclusion, LIP has a role to play in correcting markets, governments and governance failures. It has more complex tasks too, that is, to offer strategic and meta-economic public options tied to industry. These options are relevant since industry, as any organized field of evolving division of labour for the production of valuable goods and services (see the Introduction to this volume), is among the main vehicles through which communities of people can promote their models of development. Conceived in narrow economic terms, detached by a social background, deprived of leaders with a wide cultural horizon, LIP tends to be weak at best. At worst it is played in a mercantilist vein, as a game between conflicting localities. If trans-local competition has necessary and positive sides, the mercantilist attitude would instigate rent seeking, curb the constitution of trans-local public goods and damage economic entrepreneurship. Perhaps, at a deep cultural level, it is at odds also with the reproduction of local attitudes to trust in private and collective action. Are correct perspective and good leadership enough? Probably not, since industrial policy with a rich local dimension needs to be nurtured by the availability of (and the relation with) progressive features possibly included within the local social and economic life, and by fruitful feedbacks with larger scale institutional pivots. Sources of strong inertia and close circles lie here. However complexity in the organization of industry and in the social life of localities gives almost always some clues for a conscious and knowledgebased LIP to try to escape vicious circles and to promote virtuous ones. Notes 1.

2. 3. 4.

5.

Although this chapter results from joint work of the authors, Di Tommaso has primarily contributed to sections 2, 3 and 4 and Bellandi to sections 5, 6 and 7. The authors would like to acknowledge help and comments by Silvia Lombardi, Annalisa Caloffi, Francesca Marchetta, Daniele Paci, Lauretta Rubini, Sandrine Labory and Stuart Schweitzer. The usual disclaimer applies. See, for example White (1988), Auty (1994), Bianchi (1998), Nolan (2001), Chang (1994), Amsden (2003). Di Tommaso (2005), Di Tommaso and Schweitzer (2005). For historical perspectives, see Federico and Foreman-Peck (1999), Bianchi (2002), Chang (2002b). Industrial clustering is widely spread in most of the developed and developing countries. See Bellandi and Di Tommaso (2005), Di Tommaso and Schweitzer (2003), Di Tommaso et al. (2004), Di Tommaso et al. (2006), Cooke (2002), Costa Campi and Viladecans Marsal (2002), Enright (2000), Garofoli (2002), Guerrieri et al. (2001), Nadvi and Schmitz (1999), Navarro Arancegui (2003), OECD (2001), Porter (1998), Pyke et al. (1996), Rabellotti (1997), Raines (2002), Sabel and Zeitlin (1997), Solvell et al. (2003), Storper (1997). Forces such as the increasing advantages for companies with international strategies and structures to localize their entities near the nodes of the infrastructures for long-distance transport and communication; the need of proximity for sets of dependent suppliers servicing the just-in-time organization of production facilities of an international company in a locality; the increasing empowerment of local public

Local dimensions of industrial policy 357 administrations in periods of globalization, in particular as regards territorial marketing strategies (see section 3); and the propensity to apply this power within delimited fields of business, both for economizing on coordination costs, and for showing, at election times, clear results and big road signs boasting that ‘our locality is an important home base for that wonderful industry’. 6. A recent book on Marshall’s thinking, including topics such as organization of industry and industrial districts, is Aréna and Quéré (2002). 7. Other examples of strong links between the locality and the production system are the rural localities and the art cities hosting clusters of agribusiness, tourist industries, crafts works, culture-related activities. Mixed conditions, where the link is not so strong but has many points of application, are found in dynamic metropolitan areas with their dense tertiary core and diversified industrial structure (Bellandi and Sforzi, 2003; Maillat, 1998). 8. This is well accepted by important contributors to the international debates on clusters, including – of course the Italian scholars. Nevertheless different uses proliferate. Cooke (2002, p.127) writes on this point: ‘it is clearly no use to define clusters in terms of co-location alone as many studies in fact do’. This is a point that Fabio Sforzi addressed long ago in the Italian debates on industrial districts (Sforzi, 2003). 9. It is possible to classify external economies, according to the generating processes, as specialization, learning and creativity economies, resounding classical Smithian and Marshallian thinking on the advantages of division of labour (Becattini, 1989; Bellandi, 1989). 10. The systemic conditions in the previous examples have the nature of specific public goods. On the relations between public goods and industrial districts see Bellandi (2006), Goglio (1999). An important aspect is the relation between the so-called ‘centres of real services’ and the industrial districts. See Bianchi (1996), Brusco (1992), Bellini, below. 11. Let us recall (a) vertical actions: involving producers and agents with whom the producer is vertically linked by means of upstream links (for example with sub-suppliers) and downstream links (for example with dealers or clients). The cooperation concerns organization of production, production flow, management of inventories, quality control, training of workforce, process development, product development, marketing; (b) horizontal actions between agents competing on some markets, but cooperating, for example, in collective action aimed at input acquisition, common use of special machinery, product development, shared know-how; (c) multilateral actions: institutional bodies (for example business associations) offering a variety of services including, for example, joint lobbying activities, legal consultancy, market research, technical consultancy, training, managerial and financial support, credit assistance, marketing assistance, monitoring of quality, organization of trade fairs and the creation of export consortiums. See also Isaksen and Hauge (2002), and Bellini, below. 12. Collective action is here implemented in order to abuse a dominant position or to foster a permanent barrier to entry (Goglio, 1999). 13. For example, an interactive dialogue between producers and sub-suppliers, leading to improvements in processes or products, brings advantages to other agents who have not taken an active part in the cooperative actions (Nadvi and Schmitz, 1999). 14. There may be conflict concerning the distribution of the costs sustained for carrying out the joint action: costs to be paid beforehand in order to define the contract regulating the agreement (that is, collecting information on the nature of the agreed object and on the other party) as well as costs to be paid after the agreement for monitoring the other party’s behaviour, for sanctioning any possible irregularities. Members contributing to the costs may also be incapable of having complete access to the benefits. 15. Cf. Amin and Sepulveda, above, Isaksen and Hauge (2002), Viesti (2002), Lane and Maxfield (2005), Trigilia (2005). 16. A fruitful multi-level cooperation (for example between a regional government and the leaderships of localities of industry within the region) demands a regional institutional field with a progressive orientation. The spread of multicultural theses in the general arenas of political discourse is needed, countering the risks of heavy top-down ‘dirigism’ or of untrammelled ‘laissez-faire’. See Goglio (1999), Becattini (2004), Crouch et al. (2004), Trigilia (2005). 17. We have already recalled in section 4 the terms of debates on government failures. They extend of course to local governance, with some peculiarities: see Garofoli (2001), Meyer Stamer (2004), Rodrik (2004). 18. On technological constraints and opportunities see, for example, the literature on regional innovation systems (for example Cooke et al., 2004). On the necessity of a customer orientation see, for example, Humphrey and Schmitz (1996). 19. Perhaps it is worth considering some conditions that bring about strong challenges to contemporary mature localities of industry, in particular to industrial districts in western countries: (a) the growth of new industries in countries like China, and the increasing competencies of MNEs to manage global production filières; (b) the difficulties for many traditional districts to react to a jump in the collective competence in incorporating new information and communication technologies, and promoting international

358

20.

21.

International handbook on industrial policy strategies; (c) the difficulties of rejuvenating the pool of local attitudes towards entrepreneurship and trust and of incorporating the fresh energies coming from large sets of immigrants. The contents of specific public goods and services are various, and contribute to shaping a cluster’s structure. Some contents support a more active exchange with transnational companies. A technological capacity not strictly dependent on the (Greek) ‘gifts’ of transnational companies implies that the cluster productive core will be complemented by a set of interlinked technological and scientific knowledge, shared by the main actors through a common scientific–technical language. A central institutional role has to be played by universities and other centres of research, culture and high education safeguarding the public sphere of accumulation and exchange of scientific knowledge. Cf. Belussi (1999), Bianchi (1996), Capecchi (1996), Di Tommaso and Schweitzer (2003), Di Tommaso et al. (2004), Di Tommaso et al. (2006), Schweitzer and Di Tommaso (2005), Cooke (2002), Gilly and Perrat (2005). For example, the challenge that new Chinese industries present to European industries, in particular to those at the core of European industrial districts, is explained by a rich combination of factors: cheap labour; large internal market; state and local policies opening to MNE investments; but also local reserves of entrepreneurship and competence; regional policies supporting the development of industrial clusters and the constitution of large infrastructures of education, research and communication; the influence of networks of overseas Chinese entrepreneurs; and the field supplied by several cases of specialized towns (localities characterized by the presence of an industrial cluster). Within the specialized towns and industrial clusters of Guangdong, for example, some suggest a relatively strong role of developmental local forces (Bellandi and Di Tommaso, 2005).

References Amsden, A. (1989), Asia’s Next Giant: South Korea and Late Industrialization, Oxford and New York: Oxford University Press. Amsden, A. (1994), ‘Why isn’t the whole world experimenting with the East Asian model to develop? Review of the East Asian miracle’, World Development, 22(4). Amsden, A. (2003), The Rise of ‘the Rest’: Challenges to the West from Late-Industrializing Economies, Oxford: Oxford University Press. Aréna, R. and M. Quéré (eds.) (2002), The Economics of Alfred Marshall. Revisiting Marshall’s Legacy, Houndmills: Palgrave Macmillan. Auty, R.M. (1994), Economic Development and Industrial Policy: Korea, Brazil, Mexico, India and China, London: Mansell Publisher. Barca, F. (2004), ‘Stato e sviluppo: un paradigma e un cantiere aperto’, Scienze regionali, 3(3). Becattini, G. (1989), ‘Sectors and/or districts: some remarks on the conceptual foundations of industrial economics’, in E. Goodman and J. Bamford (eds), Small Firms and Industrial Districts in Italy, London: Routledge, pp.123–35. Becattini, G. (2004), Industrial Districts: A New Approach to Industrial Change, Cheltenham, UK and Northampton, MA, USA: Edward Elgar. Bellandi, M. (1989), ‘The industrial district in Marshall’, in E. Goodman and J. Bamford (eds), Small Firms and Industrial Districts in Italy, London: Routledge. Bellandi, M. (2006), ‘A perspective on clusters, localities, and specific public goods’, in C. Pitelis, R. Sugden and J. Wilson (eds), Clusters and Globalization: The Development of Economies, Cheltenham, UK and Northampton, MA, USA: Edward Elgar, forthcoming. Bellandi, M. and M.R. Di Tommaso (2005), ‘The case of specialized towns in Guangdong, China’, European Planning Studies, 13(5), 707–29. Bellandi, M. and F. Sforzi (2003), ‘The multiple paths of local development’, in G. Becattini, M. Bellandi, G. Dei Ottati and F. Sforzi (eds), From Industrial Districts to Local Development. An Itinerary of Research, Cheltenham, UK and Northampton, MA, USA: Edward Elgar, pp.210–26. Belussi, F. (1999), ‘Policies for the development of knowledge-intensive local production systems’, Cambridge Journal of Economics, 23(6), 729–47. Bianchi, P. (1996), ‘New approaches to industrial policy at the local level’, in F. Cossentino, F. Pyke and W. Sengenberger (eds), Local and Regional Response to Global Pressure. The Case of Italy and its Industrial Districts, Geneva: International Institute for Labour Studies, pp.195–206. Bianchi, P. (ed.) (1998), Industrial Policies and Economic Integration: Learning from European Experiences, London: Routledge. Bianchi, P. (2002), La Rincorsa Frenata: L’Industria Italiana Dall’unita Nazionale All’unificazione Europea, Bologna: Il Mulino. Brusco, S. (1986), ‘Small firms and industrial districts: the experience of Italy’, in D. Keeble and E. Wever (eds), New Firms and Regional Development, London: Croom Helm, pp.184–202.

Local dimensions of industrial policy 359 Brusco, S. (1992), ‘Small firms and the provision of real services’, in F. Pyke and W. Sengenberger (eds), Industrial Districts and Local Economic Regeneration, Geneva: International Institute for Labour Studies, pp.172–96. Brusco, S. (1996), ‘Global systems and local systems’, in F. Cossentino, F. Pyke and W. Sengenberger (eds), Local and Regional Response to Global Pressure: The Case of Italy and its Industrial Districts, Geneva: International Institute of Labour Studies, pp.145–58. Buchanan, J., R. Tollison and G. Tullock (1980), Towards a Theory of Rent Seeking Society, College Station: Texas A&M University Press. Capecchi, V. (1996), ‘New technological districts for promoting the quality of life and the environment’, in F. Cossentino, F. Pyke and W. Sengenberger (eds), Local and Regional Response to Global Pressure: The Case of Italy and its Industrial Districts, Geneva: International Institute of Labour Studies, pp.175–94. Chang, H.J. (1994), The Political Economy of Industrial Policy, London: Macmillan. Chang, H.J. (2002a), The Rebel Within: Joseph Stiglitz and the World Bank, London: Anthem Press. Chang, H.J. (2002b), Kicking Away the Ladder: Development Strategy in Historical Perspective, London: Anthem Press. Commission of the European Communities (2004), ‘Fostering structural changes: an industrial policy for an enlarged Europe’ (downloaded from http.//europa.eu.it/comm./enterprise/enterprise_policy/industry). Cooke, P. (2002), Knowledge Economies. Clusters, Learning and Cooperative Advantage, London: Routledge. Cooke, P., M. Heidenreich and H.J. Braczyk (eds) (2004), Regional Innovation Systems. The Role of Governance in a Globalized World, London: Routledge. Costa Campi, M.T. and E. Viladecans Marsal (2002), ‘Economías externas, competitividad y sistemas productivos locales: evidencia empírica e implicaciones para el diseño de politica industriales y de desarrollo local’, in G. Becattini, M.T. Costa Campi and Trullén (eds), Desarrollo local: Teorías y estrategias, Madrid: Cividas Ediciones, pp.245–74. Cowling K., C. Oughton and R. Sugden (1999), ‘A reorientation of industrial policy?’, in K. Cowling (ed.), Industrial Policy in Europe: Theoretical Perspectives and Practical Proposals, London: Routledge. Crouch, C., P. Le Galés, C. Trigilia and H. Voelzkow (eds) (2004), Changing Governance of Local Economies. Responses of European Local Production Systems, Oxford: Oxford University Press. Dei Ottati, G. (1994), ‘Trust, interlinking transactions and credit in the industrial district’, Cambridge Journal of Economics, 18(6), 529–46. Dei Ottati, G. (2003), ‘Local governance and industrial districts’ competitive advantage’, in G. Becattini, M. Bellandi, G. Dei Ottati and F. Sforzi, From Industrial Districts to Local Development. An Itinerary of Research, Cheltenham, UK and Northampton, MA, USA: Edward Elgar, pp.184–208. Di Tommaso, M.R. (2006), ‘Politica industriale: le esperienze, il dibattito teorico e una possibile agenda per le Nazioni Unite’, in M.R. Di Tommaso and S. Giovannelli (eds), Nazioni Unite e sviluppo industriale. Per un intervento di politica industriale nell’ interesse della comunità della nazioni, Milan: Franco Angeli. Di Tommaso, M.R. and S.O. Schweitzer (2003), ‘Why do biotechnology firms cluster? Some possible explanations’, in R. Sugden, R. Cheng and G.R. Meadows (eds), Urban and Regional Prosperity in a Globalised New Economy, Cheltenham, UK and Northampton, MA, USA: Edward Elgar, pp.181–95. Di Tommaso, M.R. and S.O. Schweitzer (eds) (2005), Health Policy and High-Tech Industrial Development. Learning from Innovation in the Health Industry, Cheltenham, UK and Northampton, MA, USA: Edward Elgar. Di Tommaso, M.R., D. Paci and S.O. Schweitzer (2004), ‘Clustering for intangibles’, in P. Bianchi and S. Labory (eds), The Economic Importance of Intangible Assets, Aldershot: Ashgate, pp.73–103. Di Tommaso, M.R., D. Paci, L. Rubini and S.O. Schweitzer (2006), ‘Is distance dead? High-tech clusters analysis and policy perspectives’, in C. Pitelis, R. Sugden and J.R. Wilson (eds), Clusters and Globalisation: The Development of the Economies, Cheltenham, UK and Northampton, MA, USA: Edward Elgar, forthcoming. Enright, M. (2000), ‘The globalization of competition and the localization of competitive advantage: policies towards regional clustering’ in N. Hood and S. Young (eds), Globalization of Multinational Enterprise Activity and Economic Development, London: Macmillan, pp.303–31. Evans, P. (1996), ‘Government action, social capital and development: reviewing the evidence on synergy’, World Development, 24, 1119–32. Federico, G. and J. Foreman-Peck (1999), European Industrial Policy: The Twentieth-Century Experience, Oxford: Oxford University Press. Garofoli, G. (2001), ‘I livelli di governo delle politiche di sviluppo locale’, in G. Becattini, M. Bellandi, G. Dei Ottati and F. Sforzi (eds), Il caleidoscopio dello sviluppo locale. Trasformazioni economiche nell’Italia contemporanea, Turin: Rosenberg & Sellier, pp.213–35. Garofoli, G. (2002), ‘Local Development in Europe. Theoretical Models and International Comparisons’, European Urban and Regional Studies, 9(3), 225–39. Gilly, J.P. and J. Perrat (2006), ‘The institutional dynamics at work in territories: between local governance and global regulation’, in C. Pitelis, R. Sugden and J. Wilson (eds), Clusters and Globalization: The Development of Economies, Cheltenham, UK and Northampton, MA, USA: Edward Elgar, forthcoming.

360

International handbook on industrial policy

Goglio, S. (1999), ‘Local public goods: productive and redistributive aspects’, Economic Analysis, 1, 5–21. Guerrieri, P., S. Iammarino and C. Pietrobelli (eds) (2001), The Global Challenge to Industrial Districts. Small and Medium-sized Enterprises in Italy and Taiwan, Cheltenham, UK and Northampton, MA, USA: Edward Elgar. Hirschman, A.O. (1981), Essays in Trespassing: Economics to Politics and Beyond, Cambridge: Cambridge University Press. Humphrey, J. and H. Schmitz (1996), ‘The triple C approach to local industrial policy’, World Development, 24(12), 1859–77. Isaksen, A. and E. Hauge (2002), ‘Regional clusters in Europe’, European Commission, DG Enterprise – Observatory of European SMEs, no. 3, Brussels. Jacobs, J. (1969), ‘Strategies for helping cities’, American Economic Review, 59(4), 652–6. Lane, D.A. and R. Maxfield (2005), ‘Ontological uncertainty and innovation’, Journal of Evolutionary Economics, 15(1), 3–50. Maillat, D. (1998), ‘Interactions between urban systems and localized productive systems: an approach to endogenous regional development in terms of innovative milieu’, European Planning Studies, 6(2), 117–29. Meyer Stamer, J. (2004), ‘The theory and practice of policy, polity and politics. Around local economic development’, in N. Bellini, E. Giuliani, C. Pietrobelli and R. Rabellotti (eds), The Theory and Practice of Local Development. Opportunities for Cooperation between Mediterranean Regions, Florence: Euromedsys Projet, pp.135–54. Myrdal, G. (1970), The Challenge of World Poverty, New York: Pantheon. Nadvi, K. and H. Schmitz (eds) (1999), World Development, Special Issue on Industrial Clusters in Developing Countries, 27(9). Navarro, L. (2003), ‘Industrial policy in the economic literature. Recent theoretical developments and implications for EU policy’, Enterprise Papers no. 12, Enterprise Directorate-General, European Commission. Navarro Arancegui, M. (ed.) (2003), Economiaz. Revista Vasca de Economía, Número especial: La Política de Clusters en el País Vasco, 53. Nolan, P. (2001), China and the Global Economy: National Champions, Industrial Policy and the Big Business Revolution, London: Palgrave Macmillan. OECD (2001), Innovative Clusters. Drivers of National Innovation Systems, Paris: OECD. Porter, M. (1998), ‘Clusters and the new economics of competition’, Harvard Business Review, November–December, 73–93. Pyke, F., G. Becattini and W. Sengenberger (eds) (1996), Industrial Districts and Inter-Firm Cooperation in Italy, Geneva: International Institute of Labour Studies. Rabellotti, R. (1997), External Economies and Cooperation in Industrial Districts: a Comparison of Italy and Mexico, New York: St Martin’s Press; London: Macmillan. Raines, P. (ed.) (2002), Cluster Development and Policy, Aldershot: Ashgate. Rodrik, D. (2004), ‘Industrial policy for the twenty-first century’, CEPR Discussion Papers. Rullani, E. (2003), ‘Il ruolo dei distretti industriali in Europa: la lezione dell’Italia’, in Formez and TeDis, La governance dei processi di internazionalizzazione dei sistemi produttivi locali, Rome: Formez, pp.13–50. Russo, M. (1996), ‘Units of investigation for local economic development policies’, Economie Appliquée, 1, 85–118. Sabel, C.F. and J. Zeitlin (1997), World of Possibilities: Flexibility and Mass Production in Western Industrialization, Cambridge, MA: Cambridge University Press. Schmitz, H. (1999), ‘Collective efficiency and increasing returns’, Cambridge Journal of Economics, 23(4), 465–83. Schweitzer, S.O. and M.R. Di Tommaso (2005), ‘La Pequenas y mediana empresas en industrias de alta tecnologia: la experiencia de las empresas biotecnologicas en los Estatos Unidos’, in M.D. Parilli, P. Bianchi and R. Sugden (eds), Alta Tecnologia, Productividad y Redes, El Collegio de Tlaxcala A.C., Tlaxcala, Mexico, pp.75–91. Seers, D. (1972), ‘What are we trying to measure?’ in N. Baster (ed.), Measuring Development, London: Frank Cass. Sforzi, F. (2003), ‘Local development in the experience of Italian industrial districts’, in G. Becattini, M. Bellandi, G. Dei Ottati and F. Sforzi (eds), From Industrial Districts to Local Development. An Itinerary of Research, Cheltenham, UK and Northampton, MA, USA: Edward Elgar, pp.157–82. Solvell, O., G. Lindqvist and C. Keetels (2003), The Cluster Initiative, Stockolm: Greenbook. Stiglitz, J. (1988), Economics of the Public Sector, New York: W.W. Norton. Stiglitz, J. (2001), Rethinking the East Asian Miracle, Washington, DC: World Bank. Storper, M. (1997), The Regional World. Territorial Development in a Global Economy, New York: The Guildford Press.

Local dimensions of industrial policy 361 Sugden, R., P. Wei and J.R. Wilson (2006), ‘Clusters, governance and the development of local economies: a framework for case studies’, in C. Pitelis, R. Sugden and J.R. Wilson (eds), Clusters and Globalization: The Development of Economies, Cheltenham, UK and Northampton, MA, USA: Edward Elgar, forthcoming. Sweeney, G. (1999), ‘Sustainable regional development: new dimensions in policy formulation’, in K. Cowling (ed.), Industrial Policy in Europe. Theoretical Perspectives and Practical Proposals, London: Routledge, pp.35–57. Trigilia, C. (2005), Sviluppo locale. Un progetto per l’Italia, Bari: Laterza. Viesti, G. (2002), ‘Economic policies and local development: some reflections’, European Planning Studies, 10(4), 467–81. White, G. (1988), Developmental States in East Asia, London: Macmillan. Zeitlin, J. (ed.) (1989), Economy and Society. Special Issue on Local Industrial Strategies, 18(4).

18 Business support policies Nicola Bellini

1 Introduction This chapter reviews the various analytical approaches to business support policies. Within the general framework of industrial policies, one may identify a set of policies specifically aiming at assisting enterprises or entrepreneurs to develop their businesses successfully and to respond effectively to the challenges of their business, social and physical environment. These policies are different from other policies that, instead of assisting businesses, either regulate or induce (or discourage) a certain kind of behaviour of companies and individuals, although they may be linked to each other not only in strategic, but also in operational, terms. Recent developments show that a large share of industrial policy objectives rely on the appropriate delivery of support policies. This happens especially when SMEs’ policies are concerned with the aim of promoting entrepreneurship, innovation and technology transfer, or internationalization. The relevance of support policies has been recognized in many countries, mainly at local and regional (or state) level. In several countries, however, support policies are dealt with by national programmes and organizations.1 The development of ‘top class business support services’ has been identified as a priority by the European Commission (CEC, 2001). The importance of business support has been repeatedly emphasized by international organizations such as the OECD and UNIDO2 in the field of economic cooperation and development. Undoubtedly the increased relevance of ‘indirect’ industrial policies owes much to the emergence of the neoliberal consensus in contemporary industrial economies. Support policies are often described as ‘being industry-driven and market-defined’3 as an explicit contrast to more ‘dirigiste’ patterns of policy making and in fact these policies are ‘market-conforming’ (that is, they are enacted in a way that exploits market mechanisms), although (contrary to neoliberal commonplaces) they are not necessarily ‘demanddriven’, as we will discuss later in this chapter. At the same time, they are examples of ‘regional experimentalism’. In countries as different as Italy and the USA, support service policies are an integral part of the trend towards the territorial fragmentation of economic policy making and are symbolic of the emergence or re-emergence of subnational governments as ‘a countervailing source of policy intervention and experimentation’ (Shapira, 2001, p.979). Furthermore, support policies are not new in history. Support services for manufacturing companies were established in several European countries during the 19th-century industrialization, promoting local production and the diffusion of new technologies. The Steinbeis Foundation in Baden-Württemberg reminds us of the name of one of the most active and innovative promoters of industrial development, Ferdinand Steinbeis (JäckleWittmann and Alberti, 2001). Furthermore, some agencies have been active for decades: for instance, the Danish Technological Institute was founded in 1906 and the German Rationalisierungs- und Innovationszentrum der Deutschen Wirtschaft (RKW) was estab362

Business support policies

363

lished as Reichskuratorium für Wirtschaftlichkeit in Industrie und Handwerk in 1921. In other cases similar support policies were designed for non-manufacturing sectors, like agriculture. The American ‘manufacturing extension’ programmes borrow the term ‘extension’ from the Agricultural Cooperative Extension Service, started in 1914. ‘Extension’ meant ‘reaching out’; institutions ‘extended’ their resources to meet collective needs. In the next sections I emphasize the meaning and implications of a policy that is implemented through service activities. Three main approaches are then discussed, based on different research paradigms. Contributions deriving from industrial and regional economics, political economy and policy analysis, and management and marketing research are discussed. Finally, the role of evaluation studies is outlined. 2 Policy through services The fundamental character of these policies lies in the fact that they are delivered through services, usually defined as ‘business support services’.4 In more abstract terms, Table 18.1 shows a four-dimensional taxonomy of business support services (Bellini, 2002), based on the user, which may be an individual company or a group (network) of companies; the function: the business support service package typically includes one or more core services, that perform the functions of the three ‘second-generation’ industrial policy instruments discussed below: the creation, selective activation and management of linkages; specialized technical support, that is added to the internal capacity of the client; transfer of knowledge, that enables the customer to interact with the provider and possibly to perform the service independently in the future; the activity: activities may consist of the following: ‘information’, providing analysis, key information for decisions, options, recommendations and so on, ‘assessment’ (for example, check-ups), ‘development’ of new solutions, ‘implementation’ of known solutions, client ‘skill development’ of human resources (training and so on); and finally, contents: business support services may focus on the ‘organization’ of the user company, the ‘technology’ or the relationship with the ‘market’. This character clearly upsets the standard hierarchical relationship between ‘state’ and ‘civil society’, where the sovereign state regulates the behaviour of its citizens or subjects and gives a prize for their socially virtuous behaviour (or imposes a cost on their behaviour, when it is not socially virtuous). A different kind of relationship is established, that is similar to the relation between a ‘service provider’ and a ‘customer/user’. This relationship is shaped by a specific kind of demand, supply interaction, that is often found, for example, in professional services (Dawson, 2000). This is a market exchange that occurs at two levels, as shown by Figure 18.1. Table 18.1

A taxonomy of business support services

User

Function

Activity

Individual

Linking

Information

Organization

Group

Supporting

Assessment

Technology

Development

Market

Knowledge transfer

Implementation Skill development

Contents

364

International handbook on industrial policy

Service provision

explicit exchange implicit exchange

User

Provider

Approval / Payment

Reassurance Prophecy fulfilled

Figure 18.1

Explicit and implicit exchange in business support services

On an explicit level, the service provider transfers intangible resources, such as knowledge, in exchange for approval and satisfaction, which may also (but not necessarily) be made visible by paying a price for the service. Customer satisfaction concerning the service may be translated into political consensus and support for the policy that originated the service activity. On an implicit level, the service provider offers a number of psychological benefits to the customer. By certifying (both technically and socially) the innovative course undertaken by the customer, the service provider reassures him and lowers his uncertainty. In exchange, users behave in a way that ‘fulfils the prophecy’ of the policy maker. The policy effect lies essentially in this circle of ‘self-fulfilling prophecies’, that is a central mechanism of economic and industrial policies in modern capitalism: if companies agree with a policy, they will act in a way that is consistent with the assumptions and with the objectives of the policy. This makes the success of the policy more likely. If the policy succeeds, policy makers will become more credible in their relationship with entrepreneurs and their policies will be even more trusted. Companies will comply even more readily, and so on and so forth. The central role of the service relationship has other implications. Firstly, support policies do not imply a transfer of financial resources from the state to companies (subsidies) or, vice versa, tax breaks. Nonetheless, subsidies may accompany support policies mainly for two reasons: they contribute to increasing their visibility and diffusion and they may help to decrease barriers to access, especially for SMEs. In fact, the use of these services is never free. Even when totally subsidized, it requires some kind of involvement of the user and his resources in the design and/or production of the service. In other words, the beneficiaries of these policies share some of the costs of their ‘production’ and some of the responsibilities for their success. Secondly, support services are context-dependent and must therefore be highly specific in the design, but especially in the delivery. Specificity refers to country (or region)

Business support policies

365

context, industry and level of industrial development. The nature of the target companies largely shapes the support provided. Although these policies are normally aimed at SMEs, relevant distinctions may be introduced between policies aimed at already existing companies in more or less traditional industries and new companies (with further specifications concerning spin-off/high-tech companies originating from research and university). Support may be aimed also at companies established by entrepreneurs belonging to underrepresented social groups (like women or ethnic minorities) or areas (such as rural businesses) and to companies with special social relevance (social enterprises like cooperatives, community businesses and other ‘not for profit’ businesses). Furthermore companies may be targets individually or as groups (districts, clusters, members of industrial associations and so on). 3 The approach of industrial and regional economics Research in industrial and regional economics has very often looked at business support policies as a component of the standard toolbox of industrial development policies, although it has not always been successful in tracing their impact on units of analysis such as regions and industrial sectors. Nonetheless, industrial and regional economics provide some fundamental perspectives on support policies, with important practical implications. While investigating the role of services in economic development and innovation and especially the role played by knowledge-intensive services, scholars of this discipline look into the nature of market failures that ‘justify’ policy intervention. In short, four types of market failures emerge: 1. 2. 3. 4.

inefficiencies and discrepancies that arise via and during the exchange of information (adverse selection issues); services, often information, that are not provided because no return can be made by the private sector (public goods); services that are not provided because insufficient return can be made (mixed goods), and the wider impacts and effects on other businesses that arise from the provision of services to an SME but which are not calculated into the price (externalities) (FSMED, 2002, p.6).

Market failures occur both on the supply side and on the demand side. The private sector providers are predominantly oriented to large firms and small companies are perceived as difficult, risky and scarcely profitable customers. Furthermore, the information needed by small and medium-sized firms is often of a generic character, thereby having public good characteristics. As a consequence, the private sector fails to produce the type of information needed by SMEs. SMEs need more external support than large firms, but they are less able to make effective use of it, less able to search for specialist providers (and tied into the local network of generalist and less qualified providers), less able to specify their needs and, in too many cases, also less open to external advice (because of individualistic and self-centred attitudes). The gap may be difficult to bridge: demand may not search for a supply that is not searching for demand either.5

366

International handbook on industrial policy

A key and sensitive issue concerns the possible market distortion that may be caused by public and subsidized providers and that may damage the operations and the development of the private suppliers. Distortions consist in the subtraction of market shares, but above all in the spreading of incorrect views about the role, impact and pricing of external services among small entrepreneurs. However, support service policies may also have a positive impact on the development of the private sector. Not all support service providers are producing the services they provide. The agencies or companies that deal directly with users are often only brokers of services that are produced by someone else. A service broker may add to the service acquired a greater or lesser amount of value or may simply take care of promotion and distribution. In the USA, according to a survey, manufacturing extension centres, besides providing services that are not supplied by private consultants, play a relevant role in informing SMEs about private consultants and in facilitating access to them. It is therefore argued that local partnerships do not compete directly with private consultants, that they provide assistance that enhances the relationships with private consultants and that they encourage the kind of openness to change that is conducive to a greater and more profitable use of outside assistance (Oldsman, 1997). Regional economics adds an emphasis on the territorial dimension of development and innovation to these arguments. Support policies increase the ‘institutional thickness’ of local and regional economies (Amin and Thrift, 1994) and can be a cornerstone of local and regional innovation systems: there is a positive relationship between the nature and strength of local/regional innovation systems and the knowledge-intensive service endowment. At the same time service providers are instrumental to cluster formation and can work as gatekeepers of the external relations of an area, playing a role in local-to-local cooperation (Bellini and Condorelli, 2004). Moreover, the best response to the above-mentioned market failures takes place at the local and regional level. Business support services are obvious candidates for localized delivery, because of the importance of close and frequent interactions for long-term relations to develop and because of the necessity to make an impact first on the local ‘strong ties’ of SMEs. 4 The view of policy sciences Research in political economy and policy analysis takes several features of business support policies into account. Firstly, these policies clearly fit into the new role of the state in economic development along the paradigm of the ‘associational economy’ (Cooke and Morgan, 1998) and provide an example of ‘governance through policy networks’ (Kickert et al., 1997). Although support policies are often (erroneously) identified with direct provision of support by public agencies, not only does a variety of policy options exist besides the direct provision of support,6 but support provision itself may be managed by a variety of actors, including governmental bodies or independent public agencies, but also public–private partnerships and agencies, consortia and specialized agencies sponsored by industrial associations, and private companies, acting according to government guidelines or within government projects or subsidization schemes. A possible division of labour with the private sector then emerges, making additional and/or rare resources (financial, political, human, relational) available and providing a precious opportunity for collective learning when policies have an experimental character.7 The effectiveness of the policy will depend on the ability to exploit a portfolio of actors,

Business support policies

367

with different endowments in terms of in-house competence and network relations, efficiency and flexibility in the actual operations, legitimization and neutrality. Secondly, support activities must be seen as examples of ‘second-generation’ industrial policy tools (cf. Bruijn and Heuvelhof, 1997), reflecting the idea of power as ability to influence social processes.8 Their main characteristics are the following: ●

● ● ● ● ●

informality (to the extent that it is acceptable within the general legal framework and that it does not induce excessive uncertainty): informality makes flexibility, responsiveness and fine-tuning possible; the possibility of being carried out by several, possibly competing, actors; public–private subsidiarity and a preference, when possible, for indirect actions, managed by other non-governmental bodies;9 serendipity, implying the acceptance – within complex scenarios – of unanticipated effects and opportunities and the readiness to exploit them; a certain degree of acceptance of the risk of failure, as a physiological component of the collective learning process; an absolute lack of coerciveness: not only are there no evident costs in not complying, but participation itself costs, at least in the short term, and may require changes in perceptions, adaptation of individual strategies and postponement of profits.

To sum up, ‘second-generation’ instruments are ‘services’ provided to the members of a policy network, that can be ‘bought’ or not. Second-generation instruments provide companies with inputs that are parallel to the ones provided by traditional instruments. Thus three ‘families’ of ‘second-generation’ innovation policy instruments can be identified (Table 18.2): 1.

opportunities are created and constraints are set to individual behaviour, not through legal instruments, but rather by the management of network relations, either with the outside (gate management) or within the network (brokerage). Management includes activities such as network structuring, (selective) network activation and game management (Klijn and Teisman, 1997);

Table 18.2

‘Traditional’ vs. ‘second-generation’ innovation policy instruments ‘Traditional’ innovation policy instruments

‘Second-generation’ innovation policy instruments

legal instruments (sticks)

management of network relations

additional resources activating individual behaviour

economic incentives (carrots)

technical support to company functions

additional information on the socio-economic and/or technological compatibility of individual behaviour

communicative instruments for moral suasion (sermons)

transfer of information

Input provided: opportunities and constraints on individual behaviour

368 2.

3.

International handbook on industrial policy additional resources are supplied, that can activate (or deactivate) individual behaviour, not through the transfer of financial resources, but rather by giving technical support to a company in a phase of structural adjustment; additional information on socioeconomic and/or technological compatibility is provided, not through formal certifications, but rather by transferring social, economic and technological information that otherwise would not be available to companies and that may derive from the knowledge assets of the policy maker.

Actual support policies will mix these ‘second-generation’ instruments, although one single aspect may often prevail. Traditional and ‘second-generation’ instruments may also be mixed together in reality, for example, when economic incentives may be set up to accompany support services. Thirdly, the nature and dynamics of the underlying sociotechnical constituencies can be investigated.10 The policy networks that underlie support service policies are composed of several groups of ‘holders’ of some quality or resource that entitles them to participate in the network. The ‘selective activation’ (and selective deactivation) of ‘holders’ (or of individuals within ‘holder’ organizations) is necessary or contributes to the realization of a new service project. In other words, among the many potential actors, one needs to decide ‘who should be involved and who should not. The success of this activity depends among other things on the willingness of those who are invited to participate to invest their time and resources in the decision making issue concerned, and the willingness of those who are not invited to stand on the sidelines’ (Kickert and Koppenjan, 1997, p.47; see also Bellini, 2002).11 Also dynamics matter: because of changes in their objectives, visions and perceptions of interests, results and impact, holders may disalign (and possibly realign) themselves with the network over time. 5 Insights from management and marketing Management and marketing research (and in particular, research focusing on the service industry and, business to business (B2B) marketing) provides insights into a number of strategic and operational issues. Firstly, marketing concepts may help in shifting the focus of policy making to the nature and durability of the relationship activated with the target companies. Traditional industrial policy tools were based on an established relationship between ‘government’ and companies, whose nature was essentially hierarchical. When the relation was given, the essence of policy was to get a specific behaviour from target companies. The market metaphor could be applied to industrial policies in the sense that policies (subsidies, protection and so on) could be demanded and bought by companies and paid back with consensus. In marketing terms, policy had a transactional approach. When hierarchical relations disappear, relations themselves are the objective of industrial policy making. In other words, the strategic objective of industrial policies becomes the establishment, selection and retention of relationships with other actors within policy networks. In marketing terms, one shifts from a transactional to a relational approach, strictly analogous to the establishment, selection and retention of relationships of service providers with customers. Secondly, marketing concepts can assist in ‘targeting the right customers’. It is vital for ‘the centres [to] focus on manufacturers who are willing to invest their time, money, and/or

Business support policies

369

human resources to improve their business’ (NIST, 1998). The rationale for this is not different from the emphasis on the ‘lead firms’ that has characterized many policies of economic development (Erickson, 1972). The ‘developmental’ behaviour of the lead firm was expected to generate growth, thanks to the economic linkages that were activated. In the case of business support policies, choosing to focus on lead firms means above all a higher likelihood to succeed. This generates trust and analogous behaviour, through imitation or through network mechanisms. The selection of potential users must be pursued also because in business support services there is a built-in tendency to what may be called the ‘adverse selection’ of customers (Cusmano et al., 2000). This means that providers may be led to select, not the most dynamic, but the weakest and less innovative companies. The reasons for this propensity are twofold: (a) especially when managed by collective or governmental actors, service providers are likely to be subject to political pressures in order to avoid discrimination; (b) the weaker companies in the local economy, especially when under serious competitive threats, may also be the most active in pushing forward their demands and are therefore an easier market for service providers. Target selection requires segmentation. Segmentation must go beyond the traditional and static identification of similar needs and wants, but – consistently with the relational approach to marketing – needs to look at the customers’ strategies and behaviour as they evolve in the interaction with the service provider (Freytag and Clarke, 2001). Segmentation must allow for the identification of ‘beachheads’ and possible market penetration paths.12 Thirdly, the demand dilemma can de defined more precisely. Should support policies be market-driven or driving markets?13 In fact, especially when policies have an innovative content and/or promote innovative behaviour by target companies, the assessment of the potential user’s needs is highly sensitive and difficult. Many interesting analogies exist with the ‘expeditionary marketing’ of new products and services, that renounces maximizing, ex ante, the odds on the introduction of a new service and focuses on incorporating the result of a well-planned learning process (Hamel and Prahalad, 1994; Mohr, 2001). Risks in need assessment are especially high when needs are unknown and the provider can rely only on the ‘weak signals’ of the company.14 In this case conventional wisdom that assumes a knowledge balance between the producer and the customer (as in industrial markets) does not hold. Risks originate on the one hand, from the necessity to adopt a ‘listening attitude’, that is, to avoid biases in the assessment process that may derive from preconceived opinions about problems and solutions or from a weak marketorientation in technology-driven programmes (cf. Hurmerinta-Peltomaki and Nummela, 1998); and on the other hand, from the necessity to interpret and not just accept the customer’s words: the question that the firm raises is often not the question that needs to be answered. Fourthly, marketing research helps to define the issues concerning the management of expectations. Expectations are relevant in the policy networks, where ‘holders’ can have three kinds of expectations about support policies:15 ●

fuzzy expectations frame problems in a schematic, emotional or ideological way, point to generic objectives, look for change, but do not have a clear understanding of what should be done concretely and how. However, these expectations are likely

370

●

●

International handbook on industrial policy to be unrealistic. Holders with fuzzy expectations will end up dissatisfied with any outcome that is not unambiguously positive; explicit expectations are clear both in their content and in the solutions that they imply. They may or may not be openly communicated to the other members of the network and are neither necessarily compatible with theirs nor necessarily realistic; implicit expectations derive from past experiences or are the result of structural factors, of organizational culture and so on. They are not the object of open discussion, because they are considered self-evident or cannot be renounced. (A typical example is the link between competitiveness and employment: it is expected that more competitiveness leads to more employment.) Because these expectations are not made explicit, they are often also not communicated to the other actors. Their feasibility and compatibility with the others’ expectations is not verified. The existence of these expectations becomes evident only when they are not met.

The holders’ perceptions and expectations can and must be managed. Strategies may work either in the social or in the cognitive dimension. They are based on the following: ●

●

● ●

developing procedures that guarantee voice and compensations to all actors and make them willing to accept the costs of cooperation. This will allow them to give up some implicit expectations and will calibrate explicit expectations (that is, make them more realistic and more compatible); furthering reflection, openness to new ideas and critical discussion of the present course, especially when unanticipated consequences of interaction emerge either within or outside the network. This allows focusing fuzzy expectations and revealing implicit expectations; maintaining or increasing social and cognitive variation, by preventing the exclusion of critical actors and favouring the inclusion of new actors with new ideas; or, in extreme cases, restructuring the existing networks or disrupting them and mobilizing new coalitions (cf. Termeer and Koppenjan, 1997, and Klijn and Teisman, 1997).

Target companies also have expectations that will be decisive to determine their final perception of the quality of the support. Their expectations may be either realistic or unrealistic and are again of three basic kinds: fuzzy, explicit and implicit. Two kinds of processes may take place (Figure 18.2). The first may arise unintentionally because of long customer relationships and a large number of repeat purchases, ‘explicit expectations become implicit. Some characteristics or elements of the service become so self-evident that the customer does not actively or consciously even think about them, or consider the possibility that they might not be realized’. The second kind may arise intentionally; through appropriate communication and marketing, the customer’s fuzzy expectations may be made more precise (focusing expectations); the customer’s implicit expectations may be made explicit (revealing expectations); and explicit, but unrealistic, expectations may be made realistic (calibrating expectations) (Ojasalo, 1999, pp.86ff). This approach has several important implications in the case of business support services. Firstly, fuzzy expectations run a high risk of being disappointed. Expectations are normally based on the user’s past experiences, both direct and indirect. However, the more

Business support policies

371

Fuzzy expectations (focusing)

Explicit expectations (calibrating) unrealistic

realistic

(revealing)

Implicit expectations Intentional dynamics

Unintentional dynamics

Source: Ojasalo (1999).

Figure 18.2

A dynamic model of the customers’ expectations

innovative is the service, the less likely are past experiences to contribute to realistic expectations. Expectations may be biased by several factors. In some cases fuzzy expectations can be nurtured by generic and ‘visionary’ policy statements. Customers with fuzzy expectations must be helped by the provider to make their expectations explicit and realistic. This may happen also without the provider’s intervention, as a result of the user’s learning process, but this may lead the user to exit from the relationship or to replace fuzzy expectations with explicit or implicit, but unrealistic ones. Secondly, explicit expectations may be unrealistic, as a result of their innovative or experimental character or because of the unrealistic promises made by managers and politicians. These expectations must be rapidly brought round to realism. Again this may happen autonomously, while the company learns about the service, but what the user will end up understanding as realistic may be felt as insufficient and disappointing. Thirdly, implicit expectations emerge when expectations are not fulfilled. For example, the user may incorrectly expect an increase of profits and realize that innovation does not deliver it in the short run. Implicit expectations are strong when a company purchases the support service as a last resort in order to solve extraordinary problems. As a consequence, it is likely that uncertainties emerge even at early stages and trigger a negative spiral of disappointment and distrust. Implicit expectations should therefore be made explicit as early as possible and whether or not they are realistic must immediately be clarified. Finally, as time goes, by expectations that were once explicit can become implicit and get out of control. For example, a certain level of quality and empathy is taken for granted, but may decline as a consequence of the routinization of the service. 6 The evaluation practice The practice of evaluation of support policies summarizes the intellectual challenges of this policy field.16 Not only does it require reconnecting different analytical perspectives,

372

International handbook on industrial policy

but it faces an impressive amount of technical intricacies. These relate mostly to the limited availability of control groups; the multiplicity and diversity of stakeholders; the mobility of objectives; the heterogeneity of policy initiatives; data collection and reliability; the relevance of long-term, indirect and implicit benefits; the relevance of rival explanations; and the intangible character of benefits. Besides this, simple performance indicators (such as the self-financing ratio or the diffusion rate) are inherently ambiguous and may have perverse effects on support practice. Nevertheless, precisely because of their need for realism, evaluation studies represent nowadays the most promising and the richest ground for further exploration of support policies. A good example is provided by the additionality criterion (cf. OECD, 1995, pp.165f). Additionality surveys try to estimate what the firm’s behaviour would have been in the absence of the policy initiatives. Two kinds of additionality may be identified: (a) the additionality of the policy in terms of increased purchases of business support services; (b) the additionality of the policy in terms of impact on the user company’s performance (increased competitiveness; increased innovation; increased internationalization and so on). This additionality is sometimes considered implicit in the previous one (the company X uses more services, therefore the company X is more innovative, more competitive and so on), but some prudence in this respect is necessary. In the first case (additionality in terms of increased purchase of business support services) tests should evaluate the relative weight of each of the following: ● ● ● ●

no additionality: the firm would have gone ahead with the same or similar service assistance at the same time; full additionality: the firm would not have used services at all; partial additionality (a): the firm would have gone ahead with inferior services (either in quantity or in quality); and partial additionality (b): the firm would have gone ahead with the same or similar service assistance, but later.

In the second case (additionality in terms of impact on the user company’s performance) tests should evaluate the relative weight of each of the following: ● ● ● ●

no additionality: the firm’s performance would have changed in the same way and at the same speed; full additionality: the firm’s performance would not have recorded any change in performance; partial additionality (a): the firm’s performance would have changed less significantly (in quantitative or in qualitative terms); and partial additionality (b): the firm’s performance would have changed at a slower rate.

Additionality should be evaluated in a careful way. For instance, in both configurations it may be considered positively that additionality tests report a limited number of cases of full additionality and mainly cases of partial additionality. This is no paradox. Partial additionality reflects the fact that policies have selected, identified and interacted with companies that were already on a growth path and ready to cooperate with the providers in an effective way. Full additionality should be regarded with concern if it means

Business support policies

373

that policy has interacted with weak companies that are fully dependent on external inputs. Additionality is difficult to prove. Evidence is derived from questionnaires sent to, and other inquiries into, user companies and obviously reflects companies’ perceptions and, in some cases, also their unwillingness to acknowledge external merits, as already mentioned in this chapter. Evidence of no additionality has to be recorded with prudence. Besides the use of control groups, a comparison between providers’ and users’ perceptions can contribute to a more correct judgment (cf. PACEC, 1998). Additionality tests often fail to account for learning effects and for the growth of ‘metacapabilities’, such as ‘learning new processes and skills, open communication throughout the organisation, sharing and leveraging knowledge internally, taking in knowledge from outside the organisation, flexibility and adaptability of organisational structure’ (Dawson, 2000, p.105). These may concern not only the user companies, but also the service providers (especially those with limited know-how in business support service provision, such as a new service centre, a new company, an entrepreneurial association with little previous experience and so on), making them more effective in the future. Attention to learning can lead to reappraisal of negative or mediocre results on specific programmes, if the common experience has strengthened both the ability to cooperate within an area and the individual actors’ capability to formulate goals, communicate them, understand the others’ goals and so on; that is, its ‘social capital’. A second relevant issue concerns the use of market shares as an indicator of the support service penetration. In practice, one uses the percentage of companies (or of SMEs) that have bought the service. Such figures lend themselves to easy political interpretations. The larger the number of companies using the service, the larger is the number of subjects that are involved in the implicit exchange: in other words, the market share matches the share of those who benefit from the policy, appreciate it and presumably are ready to reward policy makers in terms of political consensus. However, figures of market shares for business support services are so far quite disappointing. This seems to reflect a structural problem, not easily modifiable by policies.17 Market shares of business support services seem to be highly sensitive to a variety of factors. Shares are normally higher as a result of a number of factors, including the (de facto) compulsory character of the service, the wider range of service supply (that may include standardized or scarcely innovative services) and the generosity of subsidies. In addition sectoral and geographical variables may matter. For instance, contact patterns may be more intensive in peripheral regions than in urban areas because private services are more developed in the latter areas, while the support agency may be the only partner in the former ones. To sum up, the comparability of market share data seems quite low. Rather than implying that business support services are of marginal importance, low market-share figures seem to be a sign of the limited relevance of the indicator. It is not the only one. It is interesting to note, for example, that there is no evidence of a relationship between participation rates and satisfaction.18 In a relational marketing approach the market share analysis is in fact replaced, at least partially, by the analysis of the customer base. In other words, the customer must be valued not only with regard to its present ‘transactional’ value, but also and mainly with regard to its long-term attractiveness, that is, its ‘relationship value’. By analysing the provider’s portfolio of customers, one aims at ‘giving a name’ to the users. This gives evaluators some crucial information.

374

International handbook on industrial policy

Firstly, this analysis makes it possible to distinguish ‘good’ customers from ‘bad’ customers, that is, to ‘give priority to a limited portfolio of established businesses with the potential and aspiration to grow, establishing and maintaining a long term relationship’ (DTI Business Link Directorate, 1998, p.6). The ‘community of customers’ determines the nature and the level of the service, in terms of both image and substance. Evaluators should then be able to verify the overlap between the customer base and the hard core or the élite of one area’s productive apparatus.19 Secondly, the analysis of the client base allows us to identify and weigh new customers and loyal customers, who guarantee a longer-term commitment to the relationship and are likely to participate more in service production. Among loyal customers it is possible to distinguish between high-involvement relationships (implying coordination of activities, adaptation of resources and interaction among individuals) and long-term, but low-involvement relationships. Based on the portfolio analysis, the customer satisfaction surveys can also shift focus from the satisfaction that may originate from the individual transaction to the ‘relational satisfaction’. In fact a dynamic view of satisfaction is crucial because short-term quality, originating from an individual transaction and generating immediate satisfaction, does not necessarily imply long-term quality, that is based on the realization of the service impact. There may even be trade-offs between short-term and long-term quality. Shortterm quality at the expense of long-term quality may just ‘postpone dissatisfaction’ (Ojasalo, 1999). Many evaluations also reflect the fact that there is a clear relationship between the long-term nature of the relationship and both expectations and quality perceptions. As was noticed for example in a customer survey of the Scottish Enterprise Network: those who have an ongoing relationship with the Network were most likely to be satisfied with the quality of follow up contact compared with those whose contact has been occasional or infrequent. ( . . . ) Generally those describing their contact with the Network as regular were more positive about the assistance provided and more knowledgeable about the Network than other customers with less frequent contact. For the latter, perceptions of the Network and its assistance were positive, albeit more lukewarm than was found amongst the priority group. (George Street Research Ltd., 1999)

The implications on the behaviour of support agencies should also be carefully considered. Instead of building long-term relations with a relatively small group of ‘leader’ companies, market-sensitive suppliers may be pushed to maximize the number of transactions, to ‘deliver the services that firms ask for and are easiest to sell’ (Luria and Wiarda, 1996, p.245) and therefore driven into low value-added markets. A final key issue emphasized by evaluation studies concerns performance and efficiency in business support. Evaluating the relative efficiency of support policies is a complex task not only for technical reasons (their extreme differentiation that prevents comparisons), but also – again – for their important impact on the behaviour of support providers. If support services are judged on the basis of economic efficiency, are we going to get efficient business support services? As pointed out by Shapira (2001, pp.980ff), business support activities must respond simultaneously to three different kinds of rationality: first, a political rationality that sees public–private partnerships as a sustainable approach to industrial and innovation policy;

Business support policies

375

second, a social rationality that aims at enhancing the scope and quality of action. This is done by providing users with services or combinations of services that would not be available through individual initiatives and, above all, by promoting learning and network formation, that is, by investing in social capital and ‘associational economies’; third, an economic rationality that aims at efficiency, economies of scale, reduced overlaps, leveraging of scarce resources, reduced ‘government failures’ and corrected ‘market failures’. The real problem is that in theory, ideal public–private partnerships are ones that concurrently promote the triple goals of economic efficiency, social learning, and political sustainability. In practice, this ideal is hard to achieve. At best, most partnerships optimize between these three goals. If pushed too far in any particular direction, performance on one or both of the other goals suffers. (Shapira, 2001, p.981; emphasis added)

The emphasis on efficiency and performance may have important and negative impacts on the other goals. One strong inconsistency exists, for instance, between the tendency to set quantitative performance goals, defined by the number of clients served, and the emphasis on long-term relationships. Clear evidence of these difficulties is provided by the most popular (and politically appealing) performance measure: self-financing ratios. Self-financing ratios certify the ability of the service provider to collect income from the market instead of depending on the transfers by the government.20 Many have believed this to be a positive incentive to make service centres more efficient and more aggressive with regard to the market of potential users. At the same time, self-financing provides arguments to respond to possible criticisms about the use of public resources. The self-financing objective is tightly linked to the issue of cost sharing (or of the level of subsidization) of business support services. In fact these strategies to boost revenues and improve self-financing (besides increasing fees) risk jeopardizing the mission of support service initiatives in two ways. Service providers are pushed to adopt a commercial attitude and this is bound to reproduce those market failures that public policies were trying to correct (such as preferring larger companies). Moreover, self-financing ratios provide an incentive not to opt for innovative and (from the financial point of view) inevitably risky activities, which would upset the overall self-financing level of the centre. As a result the service mix is biased in favour of the most traditional products, which are more standardized and more efficiently produced. Companies already know these services and demand is stable and reliable. In contrast, innovative services need to ‘create’ their own market and must rely on public funds in order to keep the price sufficiently low. 7 Bringing risk back in Efficiency evaluation must deal with one further and possibly even more fundamental difficulty: the recognition (and acceptance) of the risk of failure. A view of policy making that perceives success or failure as the obvious result of fair or unfair, efficient or inefficient implementation of rules and good practices is inherently averse to the very idea of risk. Administrative, legal and political forces ‘predispose government toward certainty regardless of actual problem conditions’ (Christensen, 1999, p.97ff). In contrast, activities such as support services take place in risky conditions. Of course, risks in business support services may be unduly increased by the biases of policy

376

International handbook on industrial policy

making: the dependence on consensus that favours the adverse selection of customers; the insufficient knowledge of problems and possible solutions; and the wrong time estimates of the policy impact. Other biases are less political in nature and are well known by the literature on strategic forecasting: optimism, wishful thinking; inconsistency; dominance of the impact of recent events over those in the less recent past; availability of information and data; illusory causal relationships; conservatism; and selective perception (Makridakis, 1988). In the recent history of business support services many cases of second-rate results were due to a mix of unrealistic expectations, intellectual arrogance and mediocre planning, looking for easy answers to complex problems, and based on some simplistic imitation of ‘successful’ experiences of other regions or countries. Nevertheless, not only normal business operations but also innovative projects may fail, not because of any specific error, but because of the uncertain evolution of the general conditions. Uncertainty is, then, not the perverse result of political decision making, but an obvious condition (Hellström and Jacob, 2001). Policies share the entrepreneurial risk, they cannot cancel it. Furthermore, risks inherent in policies are higher to the extent that, in the name of subsidiarity, policies consciously choose to act at the margin of the conveniences, as perceived by companies. Industrial policies then add innovation to the system instead of simply reducing the costs of innovation that would in any case take place. But additionality and subsidiarity mean accepting levels of risk and uncertainty that are higher than the ones normally accepted by the ‘rational’ homo oeconomicus. It is a matter of fact that not only evaluation processes, but the whole policy-making process in business support seems uneasy in dealing with failures and is often unable to distinguish failures caused by radical mistakes in objective identification and in management from those that are due to incomplete learning. Notes 1. 2.

3. 4.

5. 6. 7. 8.

For example, the ‘Manufacturing Extension Program’ in the USA and the Small Business Service and the Business Link network in the UK. UNIDO promotes the development of business support services within its activities in the field of Small Business Development and in the clustering projects. The UN Economic and Social Commission for Western Asia (ESCWA), based in Beirut, is also active in promoting business support services in the area and has published directories of support services in some countries. This is the self-defined commitment of the US Manufacturing Extension Program, as stated in its website. Also common is the expression ‘real services’, that is clearly derived from the Italian ‘servizi reali’ (Brusco, 1992; Glasmeier, 1999). In English-language literature and policy documents other expressions, like ‘external assistance’ or ‘external advice’ or ‘business development services’, are frequently found. In the USA, business support services are mostly referred to as industrial or manufacturing ‘extension services’. In other languages business support services are referred to as ‘services de soutien’ or ‘services d’accompagnement’ (French), ‘Unterstützungsdienste’ or ‘Wirtschaftsförderungsdienste’ (German), or ‘servicios de apoyo’ (Spanish). A systematic discussion of the market failure argument is provided by FSMED (2002). Several authors have stressed specific aspects of it, for example, Brusco (1992), Feller (1997), Bryson and Daniels (1998), Bennett and Robson (1999a, 1999b), Chell and Baines (2000). For example, it is possible to focus government intervention on the possibility of improving the functioning of the business service market, qualifying demand, incentivating supply, improving market transparency and so on. A good example is provided by business incubators initiatives; see Grimaldi and Grandi (2005). Traditional policy instruments can be grouped in three ‘families’ (cf. Vedung, 1998): the ‘sticks’: that is, legal instruments, consisting of regulations (orders and prohibitions), limiting the discretion of individuals and of organizations. Regulations are normally associated with the threat of negative sanctions; the ‘carrots’: that is, economic instruments, like financial incentives (subsidies, taxes and tax breaks, fees) and

Business support policies

9. 10.

11.

12.

13. 14.

15. 16. 17. 18.

19. 20.

377

non-financial incentives (for example, the availability of infrastructures, free public services and so on). These make a certain behaviour more or less convenient to the individual, by decreasing or increasing the amount of the necessary resources (money, time, personnel, knowledge); the ‘sermons’: that is, communicative instruments which exert a moral suasion on individuals through a formal and authoritative certification of their social, economic or technological compatibility. All these instruments have in common that they reflect the exercise of power as control over individual actors and that they imply some degree of coerciveness. This is obvious in the first case (sticks), but coercion is not absent in the cases of carrots and sermons, because disagreement is costly: renouncing carrots may give competitive advantages to competing companies; disregarding sermons may be punished in several ways, ranging from social blame to the exclusion from public procurement. ‘In contemporary technology policy, public–private partnerships have emerged as a central organising mechanism for promoting research and development and accelerating the diffusion of innovations’ (Shapira, 2001, p.977). Sociotechnical constituencies are defined as dynamic ensembles of technical constituents (tools, machines and so on) and social constituents (people and their values, interest groups and so on), which interact and shape each other in the course of the creation, production and diffusion (including implementation) of specific technologies (Molina and Kinder, 2000). In particular, policy makers may behave quite independently of the performance and quality of the service providers. Politicians may sustain ailing providers, that still deliver consensus. On the contrary, they may disrupt programmes that are efficiently managed but not in tune with their political priorities: ‘political sustainability and symbolism cannot be ignored: some partnerships that have high political value can maintain, even if less than fully effective in economic and social performance. Paradoxically, even if partnerships seem to demonstrate economic and social outcomes, their continuance can be uncertain if they lack appeal to political values’ (Shapira, 2001, p.982). Many approaches are possible. Basic segmentation criteria may be based on readily available data (organization type, size, industry, location). Others may look for more qualitative or dynamic aspects (growth potential, user strategy, strategies, readiness and capabilities). Others try to detect ‘trigger’ events in order to exploit the windows of opportunity linked to particular events in the life cycle of the potential user company. These events may concern the personal life of the owner/manager or consist of exogenous facts that impose some degree of structural adaptation on the company: for instance, the introduction of new regulations concerning environment or working safety, the liberalization of certain markets, the dismantling of trade barriers, the introduction of international standards, and so on (Bellini, 2002). A market-driven approach implies the acceptance of the market structure as given. A driving markets approach implies a proactive approach towards the composition of the market and/or the role of the players in it (Jawrorski et al., 2000). The crucial issue is that companies (and especially SMEs) are often reluctant or unable to recognize their need for business. According to a recent survey, 54 per cent of European SMEs that do not make use of support services state that this is due to the fact that they do not need external help. Among sole proprietors the share of enterprises not feeling any requirement for external help reaches 60 per cent (Sheikh et al., 2002, p.24). This typology is an elaboration of the conceptual framework provided by Ojasalo (1999, 2001). The most comprehensive efforts to evaluate support policies have concerned the MEP programme in the USA (Shapira et al., 1996; NAPA, 2003; Shapira, 2003) and the Business Link in the UK (Bennett and Robson, 2003). According to NAPA (2003, p.40), MEP interventions taking place annually concern less than 2 per cent of the SME population of the USA (6000 out of 350 000). A survey of Europe’s business support services shows very clearly that the countries with the highest satisfaction rates are not necessarily countries with high participation rates, with the only exception of Ireland. Countries with participation rates higher than the European average, such as the Netherlands, Denmark, the UK, Finland or France, show satisfaction rates close to the European average level or lower. Portugal, Spain and Belgium show high satisfaction rates with participation rates equal to or less than the European average level (Sheikh et al., 2002, p.41). In the case of technology transfer centres operating in the Lombardy region, Cusmano et al. (2000) show that the group of innovative companies (identified through objective criteria, such as patents and the access to regional innovation programmes) and the customer base of the centres in fact do not coincide. For example, original MEP legislation included a so-called ‘Sunset clause’, stating that ‘the Department of Commerce should provide up to 50 percent of the annual capital and operating costs for any centre during its first three years of operation and then funding should be reduced to zero by the end of the sixth year. In keeping with this legislation, the NIST MEP instituted rules allowing for up to 50 percent cost reimbursement in the first three years, declining to 40 percent in the fourth year and one-third in years five and six’ (NIST, 1998).

378

International handbook on industrial policy

References Amin, A. and N. Thrift (1994), ‘Living in the global’, in A. Amin, and N. Thrift (eds), Globalisation, Institutions, and Regional Development in Europe, Oxford: Oxford University Press. Bellini, N. (2002), Business Support Services. Marketing and the Practice of Regional Innovation Policy, Cork: Oak Tree Press. Bellini, N. and F. Condorelli (2004), ‘Peripherality and proximity. Do business support services matter?’, Regional Studies Association conference, Angers 2004 (mimeo, http://www.regional-studies-assoc.ac. uk/events/presentations 04/belliniandcondorelli.pdf). Bennett, R. and P. Robson (1999a), ‘Intensity of interaction in supply of business advice and client impact: a comparison of consultancy, business associations and government support initiatives for SMEs’, British Journal of Management, 10, 351–69. Bennett, R. and P. Robson (1999b), ‘The use of external business advice by SMEs in Britain’, Entrepreneurship & Regional Development, 11, 155–80. Bennett, R. and P. Robson (2003), ‘Changing use of external business advice and government supports by SMEs in the 1990s’, Regional Studies, 37–8, 795–811. Bruijn, J.A. de and E.F. ten Heuvelhof (1997), ‘Instruments for network management’, in W. Kickert, E-H. Klijn and J. Koppenjan (eds), Managing Complex Networks. Strategies for the Public Sector, London: Sage, pp.119–36. Brusco, S. (1992), ‘Small firms and the provision of real services’, in F. Pyke and W. Sengenberger (eds), Industrial Districts and Local Economic Regeneration, Geneva: ILO, pp.177–96. Bryson, J. and P.W. Daniels (1998), ‘Business link, strong ties, and the walls of silence: small and medium-sized enterprises and external business-service expertise’, Environment and Planning C: Government and Policy, 16, 265–80. CEC – Commission of the European Communities (2001), Creating top-class business support services, Commission Staff working paper, SEC(2001) 1937, November, Brussels. Chell, E. and S. Baines (2000), ‘Networking, entrepreneurship and microbusiness behaviour’, Entrepreneurship & Regional Development, 12, 195–215. Christensen, K. (1999), Cities and Complexity. Making Intergovernmental Decisions, Thousand Oaks: Sage. Cooke, P. and K. Morgan (1998), The Associational Economy. Firms, Regions, and Innovation, Oxford: Oxford University Press. Cusmano, L., F. Dissoni and M. Sironi (2000), ‘Selezione avversa e trasferimento tecnologico: un’analisi del centri di servizio alle imprese della Regione Lombardia’, Economia e politica industriale, 105, 19–57. Dawson, R. (2000), Developing Knowledge-Based Client Relationships. The Future of Professional Services, Boston: Butterworth-Heinemann. DTI Business Link Directorate (1998), Personal Business Adviser Service. Policy Guidelines, URN 98/695, Sheffield, November (revised). Erickson, R. (1972), ‘The “lead firm” concept: an analysis of theoretical elements’, Tijdschrift voor Economische en Sociale Geografie, 63 (Nov./Dec.), 426–37. Feller, I. (1997), ‘Manufacturing technology centres as components of regional technology infrastructures’, Regional Science and Urban Economics, 27, 181–97. Freytag, P.V. and A.H. Clarke (2001), ‘Business to business market segmentation’, Industrial Marketing Management, 30, 473–86. FSMED – Foundation for Small and Medium Enterprise Development, University of Durham (2002), Business Support Services and Market Failure, Brussels: European Commission. George Street Research Ltd. (1999), Scottish Enterprise Network Research (mimeo), Edinburgh, July. Grimaldi, R. and A. Grandi (2005), ‘Business incubators and new venture creation: an assessment of incubating models’, Technovation, 25 (2), 111–121. Hamel, G. and C.K. Prahalad (1994), Competing for the Future, Boston: Harvard Business School Press. Hellström, T. and M. Jacob (2001), Policy Uncertainty and Risk. Conceptual Developments and Approaches, Boston: Kluwer. Hurmerinta-Peltomaki, L. and N. Nummela (1998), ‘Market orientation for the public sector providing expert services for SMEs’, International Small Business Journal, 16 (2), 69–83. Jäckle-Wittmann, G. and G. Alberti (2001), Ferdinand Steinbeis und die Gewerbeförderung im Königreich Württemberg, 2nd edn, Stuttgart: Steinbeis-Stiftung für Wirtschaftsförderung. Jawrorski, B., A. Kohli and A. Sahay (2000), ‘Market-driven versus driving markets’, Journal of the Academy of Marketing Science, 28 (1), 45–54. Kickert, W. and J. Koppenjan (1997), ‘Public management and network management: an overview’, in W. Kickert, E-H. Klijn and J. Koppenjan (eds), Managing Complex Networks. Strategies for the Public Sector, London: Sage, pp.35–61. Kickert, W., E-H. Klijn and J. Koppenjan (eds) (1997), Managing Complex Networks. Strategies for the Public Sector, London: Sage.

Business support policies

379

Klijn, E.-H. and G.R. Teisman (1997), ‘Strategies and games in networks’, in W. Kickert, E-H. Klijn and J. Koppenjan (eds), Managing Complex Networks. Strategies for the Public Sector, London: Sage, pp.98–118. Luria, D. and E. Wiarda (1996), ‘Performance benchmarking and measuring program impacts on customers: lessons from the Midwest Manufacturing Technology Centre’, Research Policy, 25, 233–46. Makridakis, S. (1988), ‘Metaforecasting – ways of improving forecasting accuracy and usefulness’, in D. Mercer (ed.), Marketing Strategy: The Challenge of the External Environment, London: SageMercer. Mohr, J. (2001), Marketing of High-Technology Products and Innovations, Upper Saddle River, NJ: PrenticeHall. Molina, A. and T. Kinder (2000), ‘National systems of innovation, industrial clusters and constituency-building in Scotland’s electornics industry’, in R. López-Martínez and A. Piccaluga (eds), Knowledge Flows in National Systems of Innovation, Cheltenham, UK and Northampton, MA, USA: Edward Elgar, pp.15–55. NAPA – National Academy of Public Administration (2003), The National Institute of Standards and Technology’s Manufacturing Extension Partnership Program. Report 1: Re-examining the Core Premise of the MEP Program, Washington, DC. NIST – National Institute of Standards and Technology (1998), Review of Mission and Operations of Regional Centres of the Manufacturing Extension Partnership, Washington, DC: US Department of Commerce, February. OECD (1995), Boosting Business Advisory Services, Paris. Ojasalo, J. (1999), Quality Dynamics in Professional Services, Helsingfors: Swedish School of Economics and Business Administration. Ojasalo, J. (2001), ‘Managing customer expectations in professional services’, Managing Service Quality, 11 (3), 200–212. Oldsman, E. (1997), ‘Manufacturing extension centres and private consultants: collaboration or competition?’, Technovation, 17 (5), 237–43. PACEC – Public and Corporate Economic Consultants (1998), Business Links – Value for Money Evaluation. Final Report, Cambridge, October. Shapira, P. (2001), ‘US manufacturing extension partnerships: technology policy reinvented?’, Research Policy, 30, 977–92. Shapira, P. (2003), ‘Evaluating manufacturing extension services in the United States: experiences and insights’, in P. Shapira and S. Kuhlmann (eds), Learning from Science and Technology Policy Evaluation, Cheltenham, UK and Northampton, MA, USA: Edward Elgar. Shapira, P., J. Youtie and J.D. Roessner (1996), ‘Current practices in the evaluation of U.S. industrial modernisation programs’, Research Policy, 25, 185–214. Sheikh, S., I. Pecher, N. Steiber and E. Heckl (2002), Support Services for Micro, Small and Sole Proprietor’s Businesses. Draft Final Report, Brussels–Vienna: European Commission–Austrian Institute for Small Business Research (IfGH). Termeer, C. and C. Koppenjan (1997), ‘Managing perceptions in networks’, in W. Kickert, E-H. Klijn and J. Koppenjan (eds), Managing Complex Networks. Strategies for the Public Sector, London: Sage, pp.79–97. Vedung, E. (1998), ‘Policy instruments: typologies and theories’, in M.L. Bemelmans-Videc, R. Rist and E. Vedung (eds), Carrots, Sticks & Sermons. Policy Instruments and Their Evaluation, New Brunswick, NJ: Transaction Publishers, pp.21–58.

19 Small and medium-sized enterprise policies in Europe, Latin America and Asia Patrizio Bianchi, Sandrine Labory, Daniele Paci and Mario Davide Parrilli*

1 Introduction This chapter aims at providing an overview of recent SME policies in both developed and developing countries, focusing on the cases of three world regions, namely Europe, Latin America and Asia. We show that policies are rather specific in their implementation, owing to differences in history, in social and political context and in both tangible and intangible resources. However, some general trends can be outlined, such as the focus of policy on SME networking and on support for internationalization. These trends mainly result from the changes in the world political and economic context outlined in Chapter 1 and stylized in the term ‘globalization’. Their principal aspect is the acceleration of the increase in world trade and FDI that has occurred over the last 20 years or so, owing to numerous factors, including essentially the liberalization of trade within GATT rounds and later WTO discussions; the diffusion of information and communication technologies; and the development of new productive systems, replacing in most sectors the Fordist system of production. These factors have implied an increase in worldwide competition, inducing firms in many sectors to define strategies at world level, trying to be present on all markets. Products have been differentiated in order to meet the demands of the consumers in the various markets. Differentiation has been realized via more frequent product renewal. In this context, various authors have shown that the Fordist system of production is generally no longer efficient (see Labory, 1997, for a review; Bianchi, 1991, for an early analysis) and that more flexible systems of production are required. The large firm has been externalizing many phases of the production process in order to increase flexibility. The large, vertically integrated firm of the mass production system has been replaced by a more decentralized firm, that leaves more autonomy and responsibility to the lower levels of the hierarchy and organizes as a network with suppliers and clients. Clusters of SMEs have been able to meet the requirements of the new competitive context very efficiently and effectively, and therefore SMEs have no longer been regarded as unproductive and marginal, but as key elements of industrial development and growth. Another reason for the renewed attention to SMEs that occurred in recent years is the problem of unemployment that politicians, especially in Europe, had to take seriously into consideration. Job creation arises mainly via firm creation and the newly created firms are SMEs that grow progressively. Europe has also been concerned with its technological gap vis-à-vis the USA, that is, its lower rate of innovation and of commercialization of innovations relative to the USA. Innovation is often brought to the market by SMEs that are created to commercialize 380

Small and medium-sized enterprise policies

381

a product resulting from an innovation. Hence attention has also been turned to SMEs for their innovative capacity. Individual SMEs often face a number of market failures, such as the lack of access to information (they do not have the resources to collect and process a large volume of information and knowledge, although ICTs have improved this capacity); a weakness in the face of dominant actors such as large firms and the consequent risk of abuse of dominant position by the latter; a difficulty in attracting or acceding to certain resources such as human capital, since they do not have the ability to offer as interesting and prestigious compensation mechanisms as large firms (SMEs offer lower wages and less prestigious career prospects) or financial capital. Hence the need to support SMEs in their creation and in their development. This chapter provides an overview of SME policies in Europe, Latin America and Asia. It shows that there has been a general tendency of policy to ‘go bottom’, that is, to shift focus to the promotion of SMEs in specific areas, especially in clusters (favouring their clustering and their relationships with local institutions). Integrating such local programmes into a national industrial development strategy is important to ensure the diffusion of spillovers between regions and sectors, so that the local success can benefit the rest of the economy. Attention to SMEs is crucial to avoid a gap between a few large successful firms and a myriad of micro firms that do not manage to grow, as occurred in some developing countries. The chapter is structured as follows. Section 2 makes a brief review of the literature on SMEs and policies for SMEs. The performance of SMEs in Europe, Latin America and Asia is subsequently analysed, in section 3. Section 4 turns to an analysis of the policy approaches that have been and are currently being adopted in these continental areas, while section 5 derives a number of implications and conclusions. The main results are that in Europe, SME policy is implementing a wide variety of actions that affect all the elements of a firm’s competitiveness; while in developing countries SME policies appear generally less complete, implying notably a risk of fracture of the industrial structure. 2 A brief review of the literature From the early 1970s onwards SMEs have been able to join the international process of development of the global economy in a very active form. Following the crisis of the Fordist system several local production systems based upon numerous SMEs have been able to respond to the challenges of the new competition in a very dynamic way, often taking the lead in development, in particular in manufacturing sectors such as footwear, textiles and clothing, tiles, furniture and spectacle manufacturing (Best, 1990; Piore and Sabel, 1984). The literature highlighted the relevant features of the strength of small and mediumsized enterprises in the open market. In particular, the characterization of local production systems, often called ‘industrial districts’ and ‘SME clusters’, has probably been the most successful configuration of SMEs in the international market. In specific contexts (for example, Japan), SMEs have been successful also in combination with large firms, especially when efficient subcontracting chains were constructed (Caddy, 1998). The literature shows that SME networks can take various forms (Asheim, 1994; Markusen, 1996; Porter, 2001; Parrilli, 2004a), from simple agglomerations of similar firms without much cooperation to industrial districts of different forms. The most typical model of industrial districts (called Marshallian industrial districts) are geographical

382

International handbook on industrial policy

clusters characterized by vertical specialization and horizontal cooperation, each firm specializing in one or more phases of the production process, competing with other firms with similar specialization but cooperating with firms with complementary specialization. The districts benefit from dense economic and social relations, embedded in a homogeneous culture, that lead to a collective efficiency (Schmitz, 1995), the reduction in transaction and communication costs, and collective learning (for instance, see Becattini, 1989, 2004; Brusco, 1982; Bellandi, 1989). Industrial districts and SME clusters may benefit from agglomeration economies, which permit the SMEs to exploit, as a group, those scale and scope economies that large firms internalize in their own structure (Sylos Labini, 2000). Joint actions (for example, export and credit consortia) and external economies (for example, free flow of information and labour), that represent the so-called ‘collective efficiency’, are those spontaneous endogenous forces that allow SMEs a flexible response to demand variations and found their competitive position in the open market (Schmitz, 1995). In the past 19 to 20 years the international context has changed substantially thanks to a boosting process of market globalization. The evidence is that, although industrial districts (for example, in Italy) have shown extremely good performance over recent decades, they have incurred difficulties in coping with globalization (Labory, 2002; Mariotti and Mutinelli, 2003). One difficulty has been the competition from developing countries, such as China and India, which benefit from low-cost labour. The strategic answer of most successful districts has been to move to higher quality segments of the market, in the attempt to look for a new way of competing. Since the early 1990s, they have been increasingly trying to adopt a high-road type of development, which is based upon knowledge and innovation and leads to higher remunerations for their goods and services (Cooke, 1996, 2002; Kaplinsky and Readman, 2001). The current tendency in research on clusters is therefore to focus on the new competitive and technological frontiers of the most advanced local production systems as a way to maintain their shares in the globalized market (Pietrobelli and Rabellotti, 2004; Bianchi and Labory, 2004; Bianchi, 1998a; Audretsch, 1998).1 Other recent development schemes are also being explored, such as the establishment of global value chains (Humphrey and Schmitz, 2004; Gereffi and Korzeniewicz, 1994), trans-local and transnational networks of SMEs (Storper, 1997; Bianchi, 1998a, 1998b) and urban trans-sectoral networks and clusters (Fujita et al., 2002; Krugman, 1998; Scott, 1998). As argued by Audretsch (2003), globalization is having two major consequences for SMEs. First, their international activities have been eased; second, their size disadvantage has often turned into an advantage in what he calls ‘the knowledge-based economy’. Evidence on these strategies shows that SMEs manage to conquer external markets via two major strategies: do it alone or ally with one or more leaders in the market, that is, insert into a global value creation process. The recent studies in Schmitz (2004) examine the longterm effects of such insertion on the competitiveness of the SMEs and derive preliminary development strategy implications, especially for clusters in developing countries. In the past 20 years, the success of many SME clusters in developed countries, especially in the Italian case, but also in others, such as Silicon Valley and Route 128 in the USA (Saxenian, 1994; Markusen, 1996), the Baden-Württenberg territory in Germany, South-West Flanders in Belgium and South-East Jutland in Denmark (Musick and Schmitz, 1994), motivated a stream of development economists to analyse national and

Small and medium-sized enterprise policies

383

local production systems in developing countries aiming at adopting the strengths and potential exhibited by SME clusters in industrialized countries (Boomgard et al., 1992; Van Dijk et al., 1994; Nadvi and Schmitz, 1999; Knorringa, 2002; Van Dijk and Sandee, 2002). Hence programmes and policies aimed at boosting the participation of SMEs in the development process of developed and, more specifically, developing countries, have been developed. Organizations such as UNIDO, UNDP, ECLAC, the World Bank and the Inter-American Development Bank (IADB), among others, have been supporting SMEs, with special attention to issues of clustering (Humphrey and Schmitz, 1996; Ceglie and Dini, 1999; Hernandez, 2003; Pietrobelli and Rabellotti, 2004), business development services (Bellini, 2002; Bouchier, 1999; Pietrobelli and Rabellotti, 2002), innovation and entrepreneurship (Kantis et al., 2001; Van Dijk and Sandee, 2002; Kantis, 2004) and financing.2 SME policies are therefore elaborated and implemented in both developed and developing countries. Their aims and measures differ, but with globalization some common policy trends (or at least priorities) are likely to exist. Hence our discussion below of the main SME policies in the three regions we have chosen, after an illustration of their importance using statistical evidence. 3 SME performance in the selected regions The evidence is that everywhere in the world, SMEs are anything but marginal operators. They reach leadership positions in various sectors, including high-tech ones. SMEs, however, are very heterogeneous, ranging from the rural family activity based on survival production to a firm of several hundreds of employees that sells on all world markets and frequently innovates. In this chapter, we define broadly SMEs as firms of fewer than 500 employees, given that the definition of SMEs varies greatly across countries (see Bianchi et al., 2000). This definition is very broad since the small and very small firms (fewer than 20 employees) make up the bulk of the firms in the range. 3.1 Europe In Europe, the vast majority of enterprises are SMEs: in Europe-19 (the EU prior to the 2004 enlargement and the three countries of the European Economic Area, together with Switzerland), in 2003, there were more than 19 million enterprises employing almost 140 million people and 99.8 per cent of these enterprises were SMEs (Observatory of European SMEs, 2003). In addition, most of these SMEs are micro enterprises, since the average number of occupied persons per SME is five (Table 19.1) and the turnover per enterprise and value added per enterprise are much smaller than for large firms. European SMEs export relatively less than the larger firms, especially micro firms which are typically more oriented towards the domestic market. Thus micro enterprises export about 9 per cent of turnover, while large firms export 23 per cent of turnover, the share for all firms being 17 per cent of turnover. Labour productivity increases with firm size. Thus an employee in a micro firm creates on average about 40 000 euros of value added, against 120 000 euros for an employee in a large firm. The new members (the Central and Eastern European Countries or CEECs) have a similar distribution of firms to that of Europe-19, although the average number of employees per firm is lower in these countries than in Europe-19 (see Table 19.1).

384

European Observatory of SMEs, 2003, pp.26, 31.

5 670 10 210 2

CEECs (2001) Enterprises Occupied persons Occupied persons per enterprise

Source:

17 820 55 040 3 440 120 9

Europe-19 Number of enterprises (000s) Employment (000s) Occupied persons per enterprise Turnover per enterprise (1000 euros) Value added per enterprise (1000 euros) Share of exports in turnover (%)

Micro

Table 19.1 SME performance in Europe (2003)

230 4 970 22

1 260 24 280 19 3 610 1180 13

Small

SME

50 5 350 107

180 18 100 98 25 680 8 860 17

Medium

5 950 20 530 3

19 270 97 420 5 890 280 12

Total

10 10 150 919

40 42 300 1 052 319 020 126 030 23

Large

Large

5 970 30 670 5

19 310 139 710 7 1 550 540 17

Total

Small and medium-sized enterprise policies

385

The results from the European Observatory of SMEs (2003) are that although the share of enterprises accounted for by SMEs remained fairly constant over the period 1988–2003, the share of micro enterprises has risen. There is also considerable evidence that new jobs in economies mainly stem from SMEs (see European Observatory of SMEs, 2003). Putting the latter two results together, it follows that new jobs mainly come from micro enterprises, often new enterprises which are bound to grow. SMEs are also important in new member states. Since the beginning of the transition the governments of these countries have promoted structural change aiming at privatization and dissolution of the formerly state-owned sector and at encouraging entrepreneurship. The former measure sometimes implies the creation of several smaller firms rather than one big state-owned firm, while the latter measure is directly aimed at firm creation, hence SME development. During transition, however, SMEs have increased slowly in CEECs, owing to slow improvements in the business environment. The prospect of EU membership has helped these countries formulate adequate policies and receive support for implementing measures. Examining the effects of European SME policy on SME performance is a difficult exercise because of the many factors at play and only qualitative analyses, or analyses restricted to certain territories, and sometimes including other industrial policies, exist. 3.2 Latin America As for other types of industrial policies, evaluation of the effects of the various policies taken is quite rare. In the case of Latin America, the Economic Commission for Latin America and the Caribbean (ECLAC) studied the performance of SMEs over the past 20 years of market liberalization (Peres and Stumpo, 2002; Hernandez, 2003). In particular, Peres and Stumpo (2002) show that, on the whole, SMEs were neither winners nor losers from the 1980s until the mid-1990s. In some cases SMEs have been able to grow (Mexico, Argentina, Chile and so on), in others they have been losing weight (Chile in the 1980s, Uruguay, Venezuela) (see Table 19.2). Table 19.2 shows the result of the study conducted by ECLAC researchers. It reveals that among the main possible determinants of SME performance – market liberalization, industrial production, SME policies and macroeconomic stability – macroeconomic policies of stabilization, specifically economic growth and inflation, are particularly significant in that context. More scattered analyses indicate that the SMEs in Latin America have suffered setbacks in the past few years. This may be explained partly in relation to global recession, which affected the region, starting from Mexico in 1995 and Brazil in 1997 and on to Chile in 1998, Argentina in 2000, and the rest; and in part owing to the lack of systemic support policies that integrate the national production systems through better linkages among the modern and traditional sectors of these economies (Bianchi, 2002; Bianchi and Parrilli, 2002; Dussel Peters, 2003; Parrilli, 2004b). 3.3 Asia The importance of SMEs in Asian economies is well documented (Farruh and Shujiro, 2005)3. Table 19.3 shows the share of SMEs in the total number of firms and the percentage of employment and GDP attributable to small and medium-sized firms.

386

International handbook on industrial policy

Table 19.2

SME performance in Latin America Performance of SMEs

Increase in liberalization

Weight of industry

Macroeconomic conditions












 

Large None Small Small Large Large Large Large Large Small Small

Significant Non-significant Non-significant Significant Non-significant Non-significant Non-significant Non-significant Significant Non-significant Non-significant













Argentina 1984–94 Chile 1990–96 Chile 1981–90 Mexico 1988–93 Colombia 1991–96 Costa Rica 1990–96 Ecuador 1991–96 Peru 1992–94 Brazil 1985–97 Uruguay 1988–95 Venezuela 1990–95

Source: Peres and Stumpo (2000, 2002).

Table 19.3

SME performance in East Asia

Country

No. of SMEs (%)

Employment (%)

99 98 95 97 – 99 99.7 97 90 89

73 60 80 94 18 45 71 58 70 65

China Hong Kong India Indonesia Malaysia Philippines Republic of Korea Singapore Sri Lanka Thailand

Value added/output (%) 60 (Industrial output) – 40 (Industrial output) – 15 (Total output) 28 (Value added) 49 (Value added) 41 (Industrial output) 55 (Value added) 47 (Value added)

Share of SMEs in total exports (%) 50 – 38 11 15 – 40 16 – 10

Source: UNCTAD (2003).

In Asian developing and newly industrialized countries, the contribution of SMEs appears crucial since these firms make up 80–90 per cent of all enterprises, providing more than 60 per cent of the jobs in the private sector (50–80 per cent of total employment). Moreover they account for about 30 per cent of direct total exports and contribute about 50 per cent of sales or value added (Narain, 2003).4 However, data reported in Table 19.3 indicate also some important differences due to country specificities (see the chapter by Lall in this volume). The role of SMEs in Korea has undergone a dramatic transformation in the last 25 years. The share of employment originating in manufacturing SMEs rose from 46 per cent (in 1975) to 69 per cent (in 1997) while that of value added rose from 32 per cent to 47 per cent.5 Korea’s economy, however,

Small and medium-sized enterprise policies

387

remains heavily dominated by large conglomerates: the so-called chaebol (Hong et al., 1999; Nugent and Yhee, 2002). Unlike Korea, Taiwan has generally been considered as the realm of SMEs. Since the early 1980s, the overall size distribution in manufacturing has stayed more or less the same with a heavy preponderance of SMEs (Aw, 2002).6 While employment data show a relatively constant share for SMEs in China, value added data suggest an increase in share of total output from 57 per cent in 1980 to around 66 per cent in 1996. In dynamic terms, two trends appear to be common to the experience of most Asian economies in the last quarter-century: first, SMEs seem to have either held their position or become more important when their role is measured by their share of employment and value added; second, there is some evidence that, within the class of SMEs, the role of very small firms (those with fewer than 10 workers) has tended to decline.7 Another important aspect to take into account in the evaluation of SMEs’ recent performance is their response to the financial crisis of 1997–8. The relative impact of the crisis by size of firm can be judged in part by the effects on capacity utilization and employment. Table 19.4 summarizes some results from a survey sponsored by the World Bank in five crisis-afflicted East Asian countries towards the end of 1998 (World Bank, 2000). It has often been argued that one of the advantages of SMEs is their flexibility. SMEs may therefore be expected to do relatively better under volatile macro conditions than large firms producing more standardized outputs, where reorganizations of the assembly line take time.8 But at the same time it is often thought that SMEs are the first victims of macroeconomic crises. Table 19.4 reflects this ambiguity. In all the five cases considered large firms tended to have higher rates of capacity utilization after the crisis than smaller firms. This result might suggest that large firms either were not as badly affected by the crisis in 1997–8 or were recovering faster.9 The table also shows that, for two of the five cases (Korea and Malaysia), a smaller share of small firms reported fewer workers after the crisis, thus suggesting that they had engaged in less retrenchment of their workforce, an outcome consistent with the shock absorption view. However, two other cases (Thailand and Indonesia) show the opposite result: a relatively higher share of SMEs report fewer workers after the crisis. Table 19.4

Impact of the 1997 financial crisis on East Asian firms

Country

Indonesia Malaysia Philippines Republic of Korea Thailand

Capacity utilization level after crisis

Share of firms with fewer workers after crisis

SMEs

Large

SMEs

Large

51 64 76 71 57

63 73 79 77 67

54 29 50 61 56

45 40 50 80 45

Note: For the purpose of this survey, SMEs are defined as firms with fewer than 150 workers, according to the standard World Bank classification. Source: World Bank (2000).

388

International handbook on industrial policy

4 Analysing SME development policies 4.1 SME policies in Europe In Europe, SMEs have long been regarded as not able to sustain competition in the long term. Hence policy was primarily aimed at helping large firms but with some measures to support firm growth. This policy stance started to be questioned in the 1980s, for a number of reasons: the decline of the Fordist production system and diffusion of the flexible production system which induced large firms to restructure; the success of some SMEs, especially those in clusters; the intensification of globalization (defined as increase in worldwide trade and market integration) implying that firms try to compete on all world markets and organize production (and the whole value creation process) on a global scale. The European SME policy started in 1983 with the adoption of the first Community programme for SMEs. Various integrated programmes have been adopted since then with increasing resources allocated. The European Union considerably extended the scope of its enterprise policy in 1993 with the White Paper on ‘Growth, Competitiveness and Employment’, where the necessity to strengthen the competitiveness of firms and SMEs in particular was stressed. The European enterprise policy primarily aims at creating an environment favourable to firm creation and growth (Article 157 of the EC Treaty), and it is mainly a framework policy, where the main measures are taken at national and local level, but the European level provides information and favours the sharing of experiences among member states in order for the successful policies of one member to be tried out in the other states. The focus is on the regulatory environment (for instance, less costly procedures to start a business), entrepreneurship, access to resources and innovation capacity. The policy is also horizontal in the sense that other policies should take the concern about SMEs into account. More specifically, the European Charter for SMEs provides the framework, in the sense of general principles, for the adoption of measures for SMEs. It was adopted at the beginning of the 21st century and since then each year member states have produced reports on the various measures they have taken towards the objectives of the Charter (10 key areas of SME policies are defined in the Charter, including education and training for entrepreneurship; quicker and less costly business start-up; better regulatory environment; supply of competence; improved on-line access; taxation and financial issues; benefits from the internal market; technological capacity; benchmarking across member states and better representation of the interest of SMEs at EU level). For instance, Table 19.5 summarizes the measures taken by member states in 2004. One particular emphasis of both the European and national SME policies is support for the creation of networks which favours the EU integration process. The project on enterprise clusters and networks (MAP project) was launched by the European Commission in 2002, aiming at providing a framework for dialogue and inter-firm cooperation, as well as for cooperation between SMEs, higher-education and research institutions, public and non-public organizations at local, regional, national and European level. Some forums are thus created that favour the meeting of European entrepreneurs and their exchange of and access to information and knowledge. The European technological programmes, the Framework Programmes that finance research projects, also favour the collaboration of SMEs since finance is offered preferably to projects that are realized by networks of SMEs (together, possibly, with other institutions).

Small and medium-sized enterprise policies Table 19.5

Belgium Denmark Finland Austria UK Greece Italy Ireland Sweden Portugal Germany Spain Neth.

389

SME policies in Europe in 2004 Financial issues

Innov. and tech. capacity

Training of entrepreneurs

Easier start-ups

Better regulation

Better on-line access

Access to EU market

* ** * ** **** * ** ** ** * **

** * * * **** * * ** * ** **

* *** * ** ** ** * *

* ** * * * * *** *

* *

* ***

* ** *

** ** *

* * * **

* ** ** * * *

*

**

* **

* ****

Taxation

*

* * * *

* * * *

* * *

Note: The stars correspond to the type of measures; the higher the number of stars, the higher the number of measures implemented. Source: Elaboration by the authors from information provided in Commission documents (European Commission, 2004b).

The networking strategy is also applied at the territorial level to favour the growth of the less favoured areas, as defined by the structural policies (references); it is used for innovation policy and human resource development in education and research policy; it is adopted for developing small and medium-sized firm clusters (European Commission, 2004a). Actions favouring networking are implemented in all member states, but often not in a structured enough way (European Commission, 2004a). Moreover, all the member countries of the EU support start-up and development of new and small and medium-sized firms. Most financial incentives have been established during the 1980s but, in most of the cases, these direct financial support schemes have been integrated and in several cases replaced by information and counselling services and courses to train the new entrepreneurs to manage their own business. Equally, most of the countries have programmes supported by the Commission that promote the creation of new firms by women, young people and returning migrants (Bianchi, 2000). All the European governments provide public support for exports. Moreover, in this case, there is a clear trend of public policy going from financial aid provided by central government to information and counselling services provided by the local authorities, through specialized promotion agencies. In financing there are a variety of national measures to respond to the evidence that in Europe the financial sector is not instrumental in helping new and small firms to develop. The EU Commission intervenes to stimulate the diffusion of seed capital funds, venture capital programmes, mutual guarantee schemes, cooperatives, mutual societies and associations (Bianchi, 2000). As shown by Martin and Valbonesi in this volume, state aids are still implemented, although these are illegal from a

390

International handbook on industrial policy

Table 19.6

Review of support policies for SMEs in selected Latin American countries Tools

Financial subsidies

Tax breaks

Information & consultancy

Training

Start-up of new enterprises

B, M

M

B, M, P

B, P

Development of subcontracting

M, Ch

M

A, M, Ch, P

M, P

Export support

A, B, Ch, CA, Co

–

A, B, Ch, CA, P, Co

A, B, Ch, P, Co

Development of financial activities

A, B

–

B

–

Employment support

B

A, B

B, P

A, B, P

Professional training & education development

M, CA

–

B, M, CA, P

Technology upgrading

A, B, Ch, CA, Co, M, P

B, Co

A, B, Ch, CA, Co, M, P

Objectives

A, B, M, CA, P, Co A, B, Co, P, CA

Note: A: Argentina; B: Brazil; M: Mexico; Ch: Chile; P: Peru; Co: Colombia; CA: Central America (Guatemala, Nicaragua, Honduras, El Salvador only). Source: Author’s elaboration on the basis of Moori-Konig et al. (2002) (Argentina); Rocha and Da Silva (2002) and Bacic and Teixeira (2004) (Brazil); Bouchier (1999), Dussel Peters (2003), Parrilli (2003) (Central America); Parrilli (2004a) and Alarcon and Stumpo (2001) (Chile); Villamil and Tovar (2002) (Columbia); Garrido (2002) (Mexico); Fairly and Bacha (2002) (Peru).

Treaty point of view and have also been shown to be inefficient from an economic point of view. 4.2 SME policies in Latin America In Latin America the situation is quite different from that of Europe. One reason is that the common regional market has not been working over the past decades (the Association for Latin American Integration, ALADI).10 Many different countries participate in several tiny common markets and free trade areas (for example, Mercosur, Andean Pact, Caricom, Central American Common Market, CACM). Despite this they have not been able to pursue these agreements effectively or to cooperate to boost the competitive position of SMEs in their own region (Peres and Stumpo, 2000). As a result, as in other policy fields, regarding SME policy too, these countries are not coordinated. They have not been able to produce the increasing uniformity of spirit and practice that has been produced in recent years in Europe; in contrast their policies reproduce significantly different approaches. These can be seen in Table 19.6. This table shows, on the one hand, the scarce initiatives taken by governments in the area of tax breaks and administrative simplifications. On the other hand, it shows that the areas of training, information and consultancy are the main fields of support, which are often named as the non-financial ‘business development services’ (BDS).

Small and medium-sized enterprise policies

391

Through a more detailed reading of the regional specificities, we can highlight the features of different approaches to SME development throughout Latin America. First of all is the approach taken by Brazil and Peru and, to a lesser extent, the Central American countries, namely Honduras, Guatemala, El Salvador and Nicaragua. In their case the institutional approach concentrates on the micro-enterprises as the focus of government policy because this is the most numerous and needy sector. Bigger, small and medium enterprises are left on their own. This is shown by the approval of the ‘statute of the micro and small enterprises’ in Brazil in 1999 (Rocha and Da Silva, 2002) as well as the strong effort implemented for this sector by institutions such as the Brazilian Technical Service for micro and small firms (SEBRAE). In Brazil, public agencies are also very active in the areas of export promotion, through the government institution APEX, and high-technologies, through the work of the specific programmes and agencies, such as Softex in the software industry (Bacic and Teixeira, 2004). Peru and Central America show the same strong focus on micro and small enterprises, in particular through important support of the increasing network of non-conventional financial institutions (mainly NGOs), which offer a sustainable micro credit to these kinds of enterprises (Fairly and Baca, 2002; Bouchier, 1999) and strong support of government and international organizations for the provision of business development services, especially in Central America (Dussel Peters, 2003; Parrilli, 2003). An intermediate kind of approach is Chile’s, where SME policy is defined primarily by the Development Corporation (CORFO), in the Ministry of Economy. Its programmes include a very broad array of actions oriented to support networks of suppliers/subcontractors, the association of firms to explore markets ( PROFO), the technological restructuring of firms (FONDEC, FDI), the exporting capacity of firms, and many others. These programmes have arisen as a response of these and other organizations to the demands of specific groups of firms and business associations. The difficulty in this case is the capacity of this Corfo to reach a large number of SMEs; more often its instruments tend to benefit only a reduced number of firms (Bianchi and Parrilli, 2002; Parrilli, 2004). A third relevant approach is taken by Mexico, Argentina and Colombia, which emphasize the need to work with the wider category of small and medium-sized enterprises and to do it in a rather active way. On the one hand, the Mexican governments in the 1990s protected the national market even by raising barriers to imports selectively (Saravì, 1997). In spite of the recent change in policy produced by the Fox administration, the previous approaches had the capacity to defend the competitiveness and the learning process of SMEs and to let them adjust rather gradually to the increasingly open market. Simultaneously, from the early 1990s onwards, Mexico boosted a policy to promote ‘maquiladoras’11 as a way to generate employment so as to boost increasing relations among large transnational companies and national small and medium-sized enterprises (Garrido, 2002; Reinhardt and Peres, 2000). In this way the possibility of securing them a market and of supporting them in a process of learning and technology transfer was promoted. International projects implemented by the UNDP and the National Financial Institution (NAFIN) permitted the development of this kind of initiatives, which are complemented by the creation of an Industrial Subcontracting System and a stock market for subcontracting firms (Garrido, 2002). Colombia’s focus has been mainly on promoting SME access to export markets. The measures and reforms approved in 1994 and 1998 contributed to boosting the exports of

392

International handbook on industrial policy

this sector, which certainly benefited from the work realized by the Institute for Industrial Development (IFI) to equalize the treatment of SMEs to large firms’. Also the creation of Infoexport centres by the national programme Proexport has been useful to promote the insertion of SMEs in the international market. This is also permitted thanks to a set of other measures, which include the creation of a public guarantee system for enterprise access to credit (Villamil and Tovar, 2002). Finally, Argentina represents the country that mostly developed a solid and integrated approach to the development of industrial SMEs, through a whole range of policies and tools aimed at boosting the local production capacity of these firms. Among the most relevant instruments are the Argentinean Technological Fund (Fontar), which operates through its Advising Services and their Technological Linkage Units, which help the producers to understand their own needs and to undertake a significant change in their production structure. Argentina, like Colombia, focused a whole set of programmes and measures on promoting the export capacity of SMEs. In particular, the Programme for Industrial Restructuring for Export (PREX) has been able to finance more than 1000 projects with a total amount of more than 27 million dollars of related finance (Moori-Konig et al., 2002). However, evidence on the effectiveness of such programmes is not clear, and Sepulveda and Amin in this volume point to mixed results generally for Argentina owing to a lack of centralization. 4.3 SME development policy in Asia A common regulatory framework is also lacking in Asia, despite some recent examples of regional agreements, which have a commercial and financial nature (for example, The Asian-Pacific Economic Cooperation – APEC). However, governments across the Asian region have increasingly recognized the importance of SMEs and they have all adopted policy measures specifically to promote these firms.12 There is little doubt that the rapidly growing SME sectors in most Asian countries generally focused on labour-intensive activities that employed more lower-skilled workers than did large enterprises.13 As a consequence, SMEs were, and often still are, considered as a major source of employment creation. However, one of the common mistakes that some Asian countries made was to consider SMEs only as a poverty-reduction instrument and not as a possible source of competitive advantage. Despite some interventions, coherent SME policies have been lacking in Indonesia, Nepal, Vietnam, Philippines and Thailand, although various programmes have been introduced to promote SMEs.14 In this section, we focus on five Asian countries which seem significant to us: the Asian Tigers, which are emerging economies, and two of the major newly industrialized countries, India and China. Table 19.7 shows the main area of intervention in favour of SMEs in these countries. There have been a number of policy initiatives in support of SMEs in the Republic of Korea since 1995, following the Special Measures Law for Regionally Balanced Economic Development and for Fostering Local SMEs (1994). With the Ministry of Trade, Industry and Energy as the apex body of the SME organizational support structure, various institutes have been working to create a favourable environment for entrepreneurship development and to provide assistance in finance, technology development and manpower training. In February 1996, the Small and Medium Business Administration (SMBA) was established to substantially strengthen support to SMEs. The SME Structure Improvement

Small and medium-sized enterprise policies

393

Table 19.7 Policies in favour of small and medium-sized enterprises in some selected Asian countries Objectives

Financial subsidies

Tax breaks

Information and consultancy

Training

Start-up of new enterprises

In, Chi, Kor

In, Chi

In, Chi

In, Chi, Ind, Thai, Kor

Development of subcontracting

In, Chi, Thai

Chi, Thai

In, Thai

In, Thai, Ind

Export support

In, Chi, Kor

In, Chi, Kor

In, Chi, Kor

In

Development of financial activities

In, Chi, Kor, Thai

In, Chi

Employment support

In

In

In

Development of education and profess. training

In, Thai, Ind, Kor

In

In, Indo, Thai, In Kor

In, Thai, Kor

Other

Notes: InIndia; KorKorea; IndIndonesia; Thai Thailand; Chi China. Source: Elaboration of the authors from Ministry of Small Scale Industry of India (2005); Narain (2003); Wang and Yao (2002); Bianchi et al. (2000); Sevilla and Soonthornthada (2000).

Programme, started in 1993, now comes under SMBA and is responsible for planning and coordinating technology programmes, with 5000–6000 SMEs assisted annually. The Indian SME sector (known in the Indian literature as the Small Scale Industry Sector – SSI) has recently emerged as a powerful engine of growth. In order to provide a more focused attention to the development of this sector, the Government of India created a new Ministry of Small Scale Industries & Agro and Rural Industries in October 1999.15 A number of comprehensive and articulated measures have been taken, covering industrial extension services, networking and information sharing, provision of developed sites for construction of warehouses, assistance for domestic marketing as well as exports technical consultancy and financial assistance for technological upgrading.16 The SSI Ministry also promotes entrepreneurship through three national institutes.17 Another remarkable area of intervention is the financing of SMEs. The Small Industries Development Bank of India (SIDBI) is a specialized financial institution of the Central Government of India, established in 1990 for the promotion, financing and development of SMEs. SIBDI is now considered to be one of best development banks in the world (Narain, 2001). Chinese SMEs have flourished as ideological barriers have fallen.18 A relevant number of SMEs were ‘collectives’ (that is, operated by authorities below the county level) managed by individuals who used their semi-public ownership status. These SMEs expanded considerably throughout the 1990s under the second half of the landmark ‘grasp the large, release the small’ (zhua da fang xiao) policy for state-owned firms. Under this policy 80 per cent of the state-owned SMEs changed their ownership status through corporatization, shareholding or private management.19 SMEs still lack access to financial

394

International handbook on industrial policy

resources. Chinese commercial banks are continuing to give priority to large state-owned enterprises because of their implicit government guarantees. SMEs generally do not receive approval to list on China’s domestic stock exchanges, which are still dominated by politically connected state-owned firms.20 However, China has improved the policy environment for SMEs since the ‘SME Promotion Law’ was promulgated in 2002. The central government allocated special funds for the development of SMEs in the period 2002–2003 and offered management training for more than 20 000 SME owners, set up a credit rating system for over 4000 SMEs in five major cities21 and helped nearly 10 000 SMEs in six major cities to start business. Thailand adopted an organic framework of policies for SMEs only after 2000, with the SME Promotion Law and the SME Development Plan. However, these interventions, which include financial subsidies, specific lines of credit, training programmes and measures to foster clustering and networking, seem to be extremely limited (Sevilla and Soonthornthada, 2000). Government support in Indonesia has consisted of limited and uncoordinated interventions. Main areas of intervention were training, creation of networking and cluster formation. Within clustering programmes the emphasis has been on innovation, achieved through training programmes for new entrepreneurs. However, none of these interventions were systematic and recurring, being without any possibility of evaluation and follow-up. This fact, together with the limited budget and the lack of qualified personnel for training programmes, makes the policy measures not particularly effective. A survey found that assistance was relevant and helpful in the case of dynamic clusters (see Sandee, 2002). Such clusters evince a demand for technical and also financial assistance to accommodate growth processes that have been put on track by producers and buyers. Thus traders frequently requested that producers in a woven cloth cluster shift to new designs in high demand. Such sales prospects stimulated producers to participate in training courses organized by the cluster development programme of the ministry. However, the cluster development programme was found not effective in bringing about technological change in the majority of clusters in Indonesia. 5 Key policy aspects across the three regions The previous analysis indicates the main features of SME development in three important geographical areas and the set of policy approaches and measures that governments have been implementing in these areas over time. A few results can be sketched on this basis. 5.1 On the completeness of the policy approach The first indication is the quite different implementation of policies in the selected geographical areas. The European Union shows a very wide and integrated approach to SME development and to SME policies. The framework provided by the various integrated programmes defined at EU level allows the member states to coordinate efforts and to have the possibility to benchmark policy measures. Thus information on the various measures taken by the member states is immediately available and other countries can assess whether some successful policies tried out in some states can fit their needs too. Measures are thus taken mainly at national level but the coordination and exchange of information, together with the analyses provided by the European Commission services on the various problems SMEs are facing and their possible solutions helps member states in their policy choices. Table 19.5 showed some measures taken in a specific year, so it is not an indication

Small and medium-sized enterprise policies

395

of all the measures that member states have taken in the various fields. However, it shows how the reports allow coordination of their actions and perhaps induce member states not to postpone action since they have to report on them each year. The various measures are specific to each country and although they fall under the same general heading, they may represent different modalities of adoption. Thus the European policy framework explained in the Pelkmans chapter leaves member states the freedom of policy making but also obliges them to take measures in certain fields where the EC Treaties commit them to action (competitiveness, environment protection and so on). The EC legislative framework also constrains the set of possible actions (for instance, state aids are forbidden). In contrast, the implementation across Latin America and South, South-East and East Asia (see Tables 19.6 and 19.7) is much more scattered. In particular, implementation in Latin America is much weaker in all respects, although with reference to export promotion and technological upgrading there seems to be a good effort, while in business startups, subcontracting development, development of financial activities and employment support the implementation is very poor. In terms of measures, the use of tax breaks is almost nil in almost all selected countries. In addition, in some areas of Latin America, the emphasis is concentrated upon a lighter approach to SME policies, which is performed through the implementation of programmes for business development services; in other countries, more specifically Brazil, Peru and Central America, the focus tends to be on micro enterprises, which are the less competitive firms in general and the less capable to promote a diffusion of development across the economy. In the case of the selected South, South-East and East Asian countries (India, Indonesia, Thailand, Korea and China), there seems to be a higher public capacity to structure programmes and deliver services to SMEs. However, this capacity is lower than in Europe, despite the fact that these Asian economies are much more governmentled, as the whole debate on the Export-Oriented Industrialization model showed from the late 1980s onwards (Amsden, 1989; Wade, 1990; World Bank, 1993; Rodrik, 1994; Stiglitz, 1998; and Lall, in this volume). In several aspects of industrial development, SME policy was lacking, with special reference to employment support, development of financial activities and export promotion, among others (see Table 19.7). A particular case is that of India. Table 19.7 shows that India implements a quite complete set of SME policies. India has focused strongly on promoting clusters of small firms and supporting the development of high-tech sectors (for example, the software cluster in Bangalore), and this policy seems to have been successful, since Indian IT engineers are now competing directly with, for instance, American ones, and many services of the Silicon Valley are now moving to India (see also the chapter by Arora and Gambardella). However, evidence on the effect of this development on the rest of the economy does not appear to be much explored. These data seem to indicate that the long process that led Europe to form the Union has been quite useful to guide the process of development of SMEs and to adopt a more and more homogeneous approach in SME development policy across the region. Starting from the wide variety of situations and approaches (Parrilli, 2005), a more coherent bulk of SME policies has been promoted, especially after the change in policy stance of the 1990s, permitting these countries to count nowadays on a set of more complete and commonly shared SME policies.

396

International handbook on industrial policy

A similar process could be undertaken in the developing regions although they do not benefit from the advantages of a common political–economic system that drives the development trajectory of the region. The lack of this common framework does not help these regions to speed up processes of homogenization and cooperation in the implementation of a stronger set of SME support policies. Nonetheless, these developing regions, in particular Asia, where China and India alone make up more than two billion consumers, have the size of the market to their advantage and the possibility to exploit their policies on such huge scales. In the case of Latin America, in spite of having a relatively smaller market, the closeness to the United States leaves significant potential to orient SMEs and SME policies towards successful entry into that market, which is also facilitated by the continuous bilateral agreements that are being signed over recent years (CEPAL, 2004). 5.2 On the networking of SMEs A second implication of the analysis performed in this chapter is that everywhere SME policy tends to focus on favouring networking. Europe thus puts strong emphasis on SME networking as a key to SMEs’ competitiveness. Promotion of networking in developing countries has occurred via specific territorial programmes often encouraged by international organizations such as UNIDO. India is a country that has been implementing programmes for the promotion of SME clusters, of which probably the human capital policy outlined in the chapter by Arora and Gambardella has been a key determinant. In fact, the policy of promoting clusters has to be qualified. Many successful experiences of promoting a cluster in a particular region are put forward to justify trying out in another country or territory a productive or innovative cluster. However, there are also many cases of failures of cluster creation. Moreover, in many countries SME policies tend to focus on micro enterprises (the case of Brazil, Peru and Central America, but also of India) or on individual small firms (such as in Chile; see Ministerio de Economia de Chile, 2001), which are taken in isolation and separated from a particular territorial context or local production system. In this way, these countries lose one of the recognized strengths of SMEs, which is their capacity to work together and, by doing so, to generate the agglomeration economies that help them offset the scale economies exploited by larger firms (Schmitz, 1995; Sylos-Labini, 2000). In this respect, the successful European experience, with special reference to the Italian case, shows the importance of considering the size of the firms too. Indeed, the policy focus passed from craft enterprises in the 1940s (see the statute of craft enterprises in 1934: Arrighetti and Seravalli, 1997) to small and medium firms in the 1950s and 1960s (see the law for SME financing, the 1965 law for leasing, and so on: Bertini, 1998) up to the territorial agglomerations, such as industrial districts and SME clusters in the 1980s and 1990s, which culminated in the territorial pacts of 1996 and the local development agencies (for example, SviluppoItalia: Parrilli, 2004a). 5.3 On the possible fracture arising within the production system and a systemic approach to the development of SMEs In general, no systematic evaluation of the policies exist. Regarding developing countries, the only evidence is general and shows that SME policies appear to be useful in that the countries that have adopted consistent policies for SMEs as part of an overall industrial

Small and medium-sized enterprise policies

397

policy have been among the most successful (also see the chapter by Lall on East Asia). However, SME policies in developing countries often lack a systemic approach that would include all the actions or measures within the framework of some integrated programmes. The risk is that the lack of systemic approach may lead to a fracture within the nation’s production system, as occurred in Argentina (see the chapter by Sepulveda and Amin) or in Chile (Parrilli, 2004; Bianchi and Parrilli, 2002) and in many other countries. Programmes developing SMEs in developing countries are often limited in terms of sectors and territory; for instance, a cluster is developed in a particular region but the effect of this development on the rest of the economy might not be as high as expected, implying the emergence of a duality, between the regions that successfully spur the development of clusters and the regions which do not benefit from such policy and remain backward. This kind of fracture can arise also within the production system in a specific region. All in all, although representing considerable structural diversity, due particularly to the dominating presence, or absence, of raw materials and oil and mineral extraction industries, the economies of the emerging countries seem to be characterized by a strong dualism between the few industrial giants, usually directly or indirectly controlled by the state and in some cases by multinational companies, and an enormous variety of small and micro enterprises (Bianchi, 2002; Bianchi and Parrilli, 2002; Parrilli, 2004). The latter may be rural, manufacturing or commercial, and have often emerged from an informal sector in which the various activities are continually mixed (Bianchi et al., 2000). This is true also for Asian countries where recent evidence shows a declining share of micro enterprise in the total number of SMEs. In almost all the emerging and developing countries analysed, manufacturing contributes to the national revenue by providing just over 15 to 20 per cent of total national revenue, leaving agriculture, commerce, the oil and mineral extraction industries and the public sector to account for the remaining 80 to 85 per cent. Therefore, most developing countries are characterized by a large number of small and very small private companies of fewer than 10 employees, that are beginning to be recognized by national governments as important elements, not only because they employ a large part of the working population, but also as an important part of a development that is no longer based on chemical or steel industries, directly connected with the processing of mineral products, or with heavy industry, connected with military production (Peres and Stumpo, 2002; AESMEC, 1998). In many developing countries, SMEs producing consumer goods, often in traditional sectors such as textiles and clothing, or in the more innovative sectors, such as that of pharmaceutical products, are aiming their activities more towards exports. For example, in several Asian countries, such as India, there are important areas of electronics and IT production, but in most cases those firms are insulated with respect to the entire economy, like islands in a wide sea of inefficient craft and small firms, simply devoted to providing means of survival for the local population. Therefore it is important to distinguish between the vast area of small and very small businesses operating essentially on the local market and that small but fundamental group of companies operating on the international market. The first enterprises are important because they constitute an important base of socioeconomic stability and provide the majority of jobs. The second are also important because international competition forces

398

International handbook on industrial policy

them to seek production quality, efficiency and innovation, so that these companies can become catalysts for modernization of the production system. However, these two groups of SMEs risk being separated in a national context which is closed towards international trade, while, as confirmed by various European experiences, they must necessarily be linked in a context of transition (Bianchi et al., 2000; Bianchi and Parrilli, 2002; Parrilli, 2004). The opportunity for increasing the number of companies able to compete on the international level primarily depends on the possibility of allowing the largest number of firms to expand from local production to the international market, enabling them effectively to become a catalyst for the whole production system. At the same time, local business in general needs to be encouraged to grow, so that it can adequately compete against products that are easily imported in a situation of rapid market opening. This operation can be realized only through a systemic approach to the development of the production system, which should be able to create a coherent framework in which innovation policies tend to promote the high value-added production of the most innovative enterprises (pharmaceuticals, electronics and so on), structural policies promote the upgrading in the productivity and competitiveness of SMEs in traditional sectors (such as textiles, footwear and furniture) and linkage policies promote the development of joint initiatives between the first group of enterprises and the second and, in this way, drive the production system of any country to grow as a whole (Parrilli, Bianchi and Sugden, 2005; Bianchi and Parrilli, 2002; Parrilli, 2004). Notes * 1.

2. 3. 4. 5. 6. 7.

8. 9.

Although the chapter results from the joint work of the authors, M.D. Parrilli has primarily contributed to sections 2, 3.2, 4.2 and 5.1; S. Labory to sections 1, 3.1, 4.1, and 5.2; D. Paci to sections 3.3 and 4.3; and P. Bianchi to section 5.3. In general, the new competitive strategy of successful industrial district is double-faced (Parrilli, 2004a). On the one hand, some districts show the emergence of a group of SMEs, which has become the leader in the district, by taking up the role of coordinator and intermediary with the world markets (for example, the case of the tile industry in Sassuolo, Russo, 1989); on the other, some leading local firms opened significant relationships with large firms, of domestic or foreign origin, whereby district firms gain access to the distribution channels of the leader while the leader gains access to the district’s know-how (for example, the shoe industry in Brenta, Rabellotti, 2004). For an extended review of the issue of financing SMEs see, for example, the journal Small Enterprise Development. The comparability of data across different countries suffers from the fact that each country adopts its own definition of SMEs. For a comprehensive review of this issue, specifically for Asian countries, see Bianchi et al. (2000). The data on exports are illustrative of whether SMEs are playing a leading role in the global economy and the fact that specific support measures might be needed to improve their performance. Korean SMEs have also become well established, like Japanese SMEs before them, in two activities often associated with dynamism: exporting and overseas investment. Korean SMEs have accounted for 35 per cent or so on average of total exports from Korea during the last 20 years and have moved more rapidly. Within this relatively stable structure, some subsectors such as textiles and clothing show growth in the employment and output shares of SMEs while others, such as basic metals and chemicals, show some decline. Thailand is one of the few countries in Asia which shares only one of these two trends. Unlike the other cases, the overall share of SMEs in total employment fell, from 60 per cent to 52 per cent. But at the same time, as in other Asian countries, the share of the smallest firms (with fewer than 10 workers) decreased, from 12 per cent to 6 per cent (Wiboonchutikula, 2002). Sunk costs are lower in smaller firms and labour–capital struggles are less frequent. Moreover the reliance on informal financial markets makes SMEs safer in times of financial crisis. Under the non-negligible assumption that firms of both size groups had the same rate of capacity utilization before the crisis. Since we do not have these data, we cannot be definite in our inference.

Small and medium-sized enterprise policies 10. 11.

12.

13. 14.

15. 16.

17. 18.

19. 20. 21.

399

The Latin American Association for Free Trade (ALALC) was created in 1960 and then transformed into the Latin American Integration Association in 1980. The concept of ‘maquiladoras’ generally refers to large, low skill-intensive firms that focus on exportorientated strategies. This form of footloose foreign direct investment is motivated by the possibility to benefit from many sources of comparative advantages such as fiscal incentives, cheap labour and rather flexible environmental regulations, among others. For the purpose of this section, we will consider only some of the most significant Asian emerging economies (the so-called ‘Asian Tigers’), and two of the major newly industrialized countries (namely India and China). We have not included, for example, Japan, since it followed very different development and industrial trajectories, making it difficult to compare its experience with the others in the same region. For example, in 1989 in Taiwan, the average capital/labour ratio was 1.305 in large enterprises and 0.187 in SMEs. Government efforts in the Philippines have focused on promoting livelihood activities by facilitating training and micro credit. This caused micro enterprises to increase in number without any prospect of improving incomes and productivity by establishing linkages with larger enterprises. In Indonesia, small enterprise promotion has been linked to the government-sponsored development of cooperatives to facilitate access to micro credit in rural areas. The Vietnamese experience with the Bank for the Poor (established in 1995 as part of its follow-up to the World Summit for Social Development) was not very successful either. Given the importance of SMEs for the Indian economy, the Ministry was split into two separate ministries, namely the Ministry of Small Scale Industry and the Ministry of Agro and Rural Industries, in September 2001. While most of the institutional support services and some incentives are provided by the Central Government, others are offered by the state governments in varying degrees to attract investments and promote small industries with the objective of enhancing industrial production and of generating employment in their respective states. They are The National Institute of Small Industry Extension Training (NISIET), The National Institute for Entrepreneurship and Small Business Development (NIESBUD) and The Indian Institute of Entrepreneurship (IIE). The vast majority of the SMEs in China are privately owned. SDRC statistics show that, by the end of 2002, China had a total of 2.43 million private firms and 23.77 million owner-operated businesses, offering 81.5 million or about one-third of the total jobs in urban areas. In 2001, non-government sectors made up one-third of China’s gross domestic product (GDP) and 45 per cent of the total capital investment. The State Council Commission to Reform the Economic System managed the process until the 1998 government restructuring, when responsibility was shifted to the State Economic and Trade Commission’s (SETC) new SME office. As a result of the banking system’s continuing preference for lending to large state-owned firms, SMEs usually raise money through indirect financing, primarily from family members or from their local communities. Retained earnings are another key source of financing. The credit guarantee system for SMEs keeps improving. By 30 June 2003, China had established nearly 1000 credit guarantee organizations for SMEs. As a result, the SMEs receiving credit guarantees increased employment by 580 000 and sales revenue by more than 110 billion yuan (over 13 billion US dollars). Moreover, the State Development and Reform Commission (SDRC) has been working in cooperation with commercial banks to provide special loans for SMEs in some cities.

References AESMEC (1998), ‘The role of SME: Asian and European experiences’, Proceedings of the 1998 Asia–Europe SME Conference, Naples, Italy, 28–30 May. Alarcon, C. and G. Stumpo (2001), Las pequeñas y medianas empresas en Chile, Paper de la CEPAL, Santiago. Amsden, A. (1989), Asia’s Next Giant, Oxford: Oxford University Press. Arrighetti, A. and G. Seravalli (1997), ‘Istituzioni e dualismo dimensionale dell’industria italiana’, in F. Barca (ed.), Storia del capitalismo italiano dal dopoguerra ad oggi, Rome: Donzelli Editore. Asheim, B. (1994), ‘Industrial districts, inter-firm co-operation and endogenous technological development: the experience of developed countries’, in UNCTAD (ed.), Technological Dynamism in Industrial Districts, Geneva. Audretsch, D. (1998), ‘Introduction’, in D. Audretsch (ed.), Industrial Policy and Competitive Advantage,Volume I, The Mandate for Industrial Policy, Cheltenham, UK and Lyme, USA: Edward Elgar. Audretsch, D. (2003), ‘Introduction’, in D. Audretsch (ed.), SMEs in the Age of Globalisation, Cheltenham, UK and Northampton, MA, USA: Edward Elgar.

400

International handbook on industrial policy

Aw, B.-Y. (2002), ‘Productivity dynamics of small and medium enterprises in Taiwan’, Small Business Economics, 18(1–3), 69–84. Bacic, M. and L.A. Teixeira (2004), ‘El programa Softex en Brazil’, in H. Kantis (ed.), El proceso emprendedor en America Latina y otras experiencias internacionales, Washington, DC: Banco Inter-Americano de Desarollo (BID). Becattini, G. (1989), ‘Sectors and/or districts: some remarks on the conceptual foundations of industrial economics’, in E. Goodman and J. Bamford (eds), Small Firms and Industrial Districts in Italy, London: Routledge, pp.123–35. Becattini, G. (2004), Industrial Districts: A New Approach to Industrial Change, Cheltenham, UK and Northampton, MA, USA: Edward Elgar. Bellandi, M. (1989), ‘The industrial district in Marshall’, in E. Goodman and J. Bamford (eds), Small Firms and Industrial Districts in Italy, London: Routledge. Bellini, N. (2002), Business Support Services. Marketing and the Practice of Regional Innovation Policy, Cork: Oak Tree Press. Bertini, S. (1998), ‘SMEs, industrial development and competitiveness in Italy’, in AESMEC, Proceedings of the 1998 Asia–Europe SME Conference, Naples. Best, M.H. (1990), The New Competition: Institutions of Industrial Restructuring, Cambridge: Polity Press. Bianchi, P. (1991), Produzione e Potere di Mercato, Rome: Ediesse. Bianchi, P. (1998a), ‘Requiem for the third Italy? Rise and fall of a too successful concept’, Entrepreneurship & Regional Development, 10. Bianchi, P. (ed.) (1998b), Industrial Policies and Economic Integration, London: Routledge. Bianchi, P. (2000), ‘Policies for small and medium-sized enterprises’, in W. Elsner and J. Groenewegan (eds), Industrial Policies after 2000, Dordrecht: Kluuwer Academic Publishers. Bianchi, P. (2002), La Rincorsa Frenata: L’Industria Italiana Dall’unita Nazionale All’unificazione Europea, Bologna: Il Mulino. Bianchi, P. and M.D. Parrilli (2002), ‘Small and medium-sized enterprises: a comparative approach to Latin America and the European Union’, Dept. of Economics, discussion paper no. 6, Ferrara University, November. Bianchi, P., M.R. Di Tommaso and L. Rubini (2000), Le Api Audaci, Milan: FrancoAngeli. Boomgard, J., S. Haggblade, S. Davies and D. Mead (1992), ‘A sub-sector approach to small enterprise promotion and research’, World Development, 20(2), 371–84. Bouchier, S. (1999), Las Pymes en Centroamerica, San Salvador: GTZ. Brusco, S. (1982), ‘The Emilian model: productive decentralisation and social integration’, Cambridge Journal of Economics, 6. Caddy, J. (1998), ‘The changing role of SMEs in Japan’, ASEM Conference, Naples, May. Ceglie, G. and M. Dini (1999), ‘El desarollo de clusters y redes de Pymes en los paises en desarollo: el caso del Nicaragua’, Revista Encuentro, 50. CEPAL (2004), Balance Preliminar de las economìas de América Latina y el Caribe, Santiago. Cooke, P. (1996), ‘Building a twenty-first century regional economy in Emilia-Romagna’, European Planning Studies, 4(1). Cooke, P. (2002), Knowledge Economies. Clusters, Learning and Cooperative Advantage, London: Routledge. Dussel Peters, E. (2003), ‘Condiciones y retos de las MIPYME en Centroamerica’, in R. Hernandez (ed.), Competitividad de las MIPYMES en Centroamerica, Mexico City: CEPAL-GTZ. European Commission (2004a), Final report of the expert group on Enterprise clusters and networks, Brussels, http://europa.eu.int/comm/enterprise/entrepreneurship. European Commission (2004b), Report of the Commission to the Parliament and to the European Council on the implementation of the European Charter for SMEs, COM (2004) 64 final, Brussels. Fairly, A. and L. Bacha (2002), ‘Las pequeñas y medianas empresas en Peru’, in W. Peres and G. Stumpo (eds), Las pequenas y medianas empresas industriales en America Latina, CEPAL, Mexico City: Ediciones Siglo XXI. Farruh, I. and U. Shujiro (2002), ‘Small firm dynamism in East Asia – an introductory overview’, Small Business Economics, 18(1–3), February. Fujita, M., P. Krugman and A. Venables (2002), The Spatial Economy: Cities, Regions and International Trade, Cambridge, MA: MIT Press. Garrido, C. (2002), ‘Las pequeñas y medianas empresas manufactureras en Mexico’, in W. Peres and G. Stumpo (eds), Las pequenas y medianas empresas industrials en America Latina, CEPAL, Mexico City: Ediciones Siglo XXI. Gereffi, G. and M. Korzeniewicz (1994), Commodity Chains and Global Capitalism, Westport, CT: Greenwood Press. Gomes-Casseres, B. (1997), ‘Alliance strategies of small firms’, Small Business Economics, 9(1), 33–44. Hernandez, R. (ed.) (2003), ‘Competitividad de las MIPYMES en Centroamerica’, CEPAL-GTZ, Mexico. Hong, S.-Y., J.H. Park and J.Y. Park (1999), ‘Status and prospects of small & medium enterprises (SMEs) in Korea’, Korea Small Business Institute, Seoul.

Small and medium-sized enterprise policies

401

Humphrey, J. and H. Schmitz (1996), ‘The triple C approach to local industrial policy’, World Development, 24(12), 1859–77. Humphrey, J. and H. Schmitz (2004), ‘Chain governance and upgrading: taking stock’, in H. Schmitz (ed.), Local Enterprises in the Global Economy, Cheltenham, UK and Northampton, MA, USA: Edward Elgar. Kantis, H. (ed.) (2004), ‘El proceso emprendedor en America Latina y otras experiencias internacionales’, Washington, DC: Banco Inter-Americano de Desarollo (BID). Kantis, H., M. Komori and M. Ishida (2001), Entrepreneurship in Emerging Economies: Cases from Latin America and South East Asia, Washington, DC: Banco Inter-Americano de Desarollo (BID). Kaplinsky, R. and J. Readman (2001), ‘Integrating SMEs in global value chains: spreading the gains from globalisation’, IDS Bulletin, 32(3), Sussex University. Knorringa, P. (2002), ‘Cluster trajectories in developing countries’, in M.P. van Dijk and H. Sandee (eds), The Importance of Innovation for Small Enterprise Development in the Third World, Cheltenham, UK and Northampton, MA, USA: Edward Elgar. Krugman, P. (1998), ‘What’s new about the New Economic Geography?’, Oxford Review of Economic Policy, 14(2), 7–17. Labory, S. (1997), ‘The EU car industry: structural changes and challenges for foreign and domestic producers’, in A. Cafruny and P. Peters (eds), EU Foreign Policy: How does it affect the World?, Dordrecht: Kluwer Academic Publishers. Labory, S. (2002), ‘Relazioni tra grandi imprese e PMI. Considerazioni generali con particolare riferimento al settore moda’, in S. Labory and L. Zanni (eds), Il Sistema Moda in Toscana, Florence: Collana IRPET. Mariotti, S. and M. Mutinelli (2003), ‘L’internazionalizzazione passiva dei distretti Italiani’, Economia e Politica Industriale, 119, 139–54. Markusen, A. (1996), ‘Sticky places in slippery space: a typology of industrial districts’, Economic Geography, 72, 293–313. Ministerio de Economia de Chile (2001), Estatuto de la pequeña empresa, Santiago: Department for Small and Medium Enterprises. Ministry of Small Scale Industries of India (2005), Annual Report 2003–2004. Moori-Konig, V., C.A. Ferraro and G. Yoguel (2002), ‘Situacion y politica para las pequeñas y medianas empresas industriales en Argentina’, in W. Peres and G. Stumpo, Las pequeñas y medianas industrias en America Latina, Siglo XXI, Mexico City: CEPAL-Chile. Musick, B. and H. Schmitz (1994), ‘Industrial districts in Europe: policy lessons for developing countries’, World Development, 22(6). Nadvi, K. and H. Schmitz (eds) (1999), ‘World development, special issue on industrial clusters in developing countries’, 27(9). Narain, S. (2001), ‘Development financial institutions’ and commercial banks’ innovative schemes for assisting SMEs in India’, Improving the Competitiveness of SMEs in Developing Countries. The Role of Finance to Enhance Enterprise Development, Geneva: UNCTAD. Narain, S. (2003), ‘Institutional capacity building for small and medium sized enterprises promotion and development’, in UNESCAP – United Nations Economic and Social Commission for Asia and the Pacific (ed.), Investment Promotion and Enterprise Development Bulletin for Asia and the Pacific, Bangkok: United Nations Publications, pp.19–44. Nugent, J.B. and S. Yhee (2002), ‘Small and medium enterprises in Korea: achievements, constraints and policy issues’, Small Business Economics, 18(1–3) 85–119. Observatory of European SMEs (2003), ‘Highlights from the 2003 Observatory’, Observatory of European SMEs, 8. Parrilli, M.D. (2003), ‘Innovation and competitiveness in small furniture industry in Nicaragua’, in M.P. Van Dijk and H. Sandee, The Importance of Innovation for Small Enterprises in the Third World, Cheltenham, UK and Northampton, MA, USA: Edward Elgar. Parrilli, M.D. (2004a), ‘A stage ecletic approach to industrial district development: two policy keys for survival clusters in developing countries, European Planning Studies, 12(8), 1115–32. Parrilli, M.D. (2004b), ‘Integrating the national industrial system: the new challenge for Chile’, Review of International Political Economy, 11(5), 905–25. Parrilli, M.D. (2005), ‘Alta tecnologia, prodcutividad y redes: un enfoque sistemico para el desarrollo de las pequeñas y medianas empresas’, in M.D. Parrilli, P. Bianchi and R. Sugden (eds), Alta tecnologia, prodcutividad y redes: un enfoque sistemico para el desarrollo de las pequeñas y medianas empresas, Mexico City: Colegio de Tlaxcala. Parrilli, M.D., P. Bianchi and R. Sugden (eds) (2005), Alta Tecnologia, Productividad y Redes, Tlaxcala, Mexico: El Colegio de Tlaxcala, A.C. Pederson, P.O., A. Sverrisson and M. Van Dijk (1994), Flexible Specialization: The Dynamics of Small-Scale Industries in the South, ITDG Publications.

402

International handbook on industrial policy

Peres, W. and G. Stumpo (2000), ‘Small and medium-sized manufacturing enterprises in Latin America, World Development, 28(9). Peres, W. and G. Stumpo (2002), Las pequenas y medianas empresas industriales en America Latina, Siglo XXI, Mexico City: CEPAL. Pietrobelli, C. and R. Rabellotti (2002), ‘Business development services in Italy. An empirical analysis of three regional experiences: Emilia-Romagna, Lombardia and Veneto’, Serie Desarrollo Productivo, CEPALECLAC, Santiago del Chile. Pietrobelli, C. and R. Rabellotti (eds) (2004), Upgrading in Clusters and Value Chain in Latin America, Washington, DC: Inter-American Development Bank. Piore, M. and C. Sabel (1984), The Second Industrial Divide, New York: Basic Books, Inc. Publishers. Porter, M. (2001), ‘Regions and the new economics of competition’, in A. Scott (ed.), Global City Regions, Oxford: Blackwell, pp.139–52. Reinhardt, N. and W. Peres (2000), ‘Latin America’s new economic model’, World Development, 28(9), 1543–66. Rocha, C.E. and E. Da Silva (2002), ‘Situacion y politicas para pequeñas y medianas empresas industriales en Brazil’, in W. Peres and G. Stumpo (eds), Las pequenas y medianas empresas industriales en America Latina, CEPAL, Ediciones Siglo XXI. Rodrik, D. (1994), ‘Industrial policy for the twenty-first century’, CEPR Discussion Papers. Sandee, H. (2002), ‘SMEs in South East Asia: issues and constraints in the pre- and post-crisis environment’, in B.C. Lee and C. Harvey (eds), Small Enterprise Development in South East Asia, Cheltenham, UK and Northampton, MA, USA: Edward Elgar, pp.61–82. Saravi, G. (1997), Redescubriendo la Microindustria, Mexico City: FLACSO. Saxenian, A. (1994), Regional Advantage: Culture and Competition in Silicon Valley and Route 128, Cambridge, MA: Harvard University Press. Schmitz, H. (ed.) (2004), Local Enterprises in the Global Economy: Issues of Governance and Upgrading, Cheltenham, UK and Northampton, MA, USA: Edward Elgar. Scott, A. (1998), Regions and the World Economy: The Coming Shape of Global Production, Competition, and Political Order, Oxford: Oxford University Press. Sevilla, R.C. and K. Soonthornthada (2000), ‘SME policy in Thailand: vision and challenge’, Nakhon Pathom: Institute for Population and Social Research, Mahidol University, July. Stiglitz, J. (1998), ‘Towards a new paradigm for development’, ninth Paul Prebisch Lecture, UNCTAD, Geneva, 19 October. Storper, M. (1997), The Regional World: Territorial Development in a Global Economy, New York: Guilford Press. Sylos-Labini, P. (2000), Sottosviluppo: una Strategia di Riforme, Rome: Laterza. UNCTAD (2003), Improving the Competitiveness of SMEs through Enhancing Productive Capacity, Geneva. Van Dijk, M.P. and H. Sandee (eds) (2002), The Importance of Innovation for Small Enterprises’ Development in the Third World, Cheltenham, UK and Northampton, MA, USA: Edward Elgar. Villamil, J.A. and J.A. Tovar (2002), ‘Situacion actual y politicas para pequeñas y medianas empresas industriales en Colombia’, in W. Peres and G. Stumpo (eds), Las pequenas y medianas empresas industriales en America Latina, Ediciones Siglo XXI, Mexico City: CEPAL. Wade, R.H. (1990), Governing the Market, Princeton: Princeton University Press. Wang, Y. and Y. Yao (2002), ‘Market reforms, technological capabilities and the performance of small enterprises in China’, Small Business Economics, 18(1–3) 197–211. Wiboonchutikula, P. (2002), ‘Small and medium enterprises in Thailand: recent trends’, Small Business Economics, 18(1–3), 213–26. World Bank (1993), The East Asian Miracle, Economic Growth and Public Policy, New York: Oxford University Press. World Bank (2000), Asian Corporate Recovery, Washington, DC.

PART V THE USE OF GAME THEORY IN INDUSTRIAL POLICY ANALYSIS

20 The genesis of game theory and its role in industrial economics Luca Lambertini

1 Introduction Nowadays, game theory provides the essential analytical instruments that virtually any professional researcher carrying out theoretical research in the social sciences has to master. Furthermore, predictions and normative indications produced by theoretical models based upon game theory represent the basis for policy analysis and applied studies. This is true in particular of economics, and above all of industrial organization, yet the current state of our profession as industrial economists is significantly different from what it was no more than 20 years, or at most a quarter of a century ago. The work of those who have contributed to build up game theory has changed the way we conceive economics, and therefore also industrial economics and policy, once and for all. The aim of this chapter is to propose a brief outline of the main achievements of static game theory in terms of the main equilibrium concepts, and then proceed to illustrate the route followed by the recent developments of industrial economics. These two elements, in some sense, would represent a standard approach to the matter, and in themselves would contain no particular value added to the state of our knowledge. However, as I hope will become clearer in what follows, this is only a part of the entire story that stems from the early years of game theory. Rather unfortunately, the rest of the story is probably either unknown or, to say the least, unfamiliar to many. This has to do with another field of applied mathematics, namely differential game theory. As a matter of fact, the theory of differential games, which essentially can be considered as an extension of optimal control theory to the case of at least two interacting agents, was being developed over much the same time span as static game theory, and in the very same places, such as RAND Corporation and the Office of Naval Research. By its very nature, this would have represented the obvious toolbox for the analysis of the dynamic evolution of economic systems, including industries. Yet, as we all know, things went differently, the result being that by far the largest part of contemporary theoretical economics builds upon either static or repeated games, neither being properly dynamic as they do not feature any state variables evolving over time. The material proposed in the remainder of this chapter is aimed, amongst other things, at shedding some light on these aspects of the evolution of dynamic game theory and its role, or the lack thereof, in shaping our current knowledge in the field of industrial economics. In this respect I will not be, and I could not pretend to be, exhaustive, my purpose being solely to suggest a possible interpretation of the reasons why differential games did not immediately become the main analytical approach in our discipline. My aim consists rather in suggesting that this could indeed constitute an interesting and productive line for future research. The outline of the remainder of the chapter is the following. Section 2 recollects the main steps of static and repeated game theory through its main equilibrium concepts. 405

406

International handbook on industrial policy

Section 3 describes the evolution of industrial economics into the current theory of industrial organization, and summarizes its main theoretical achievements as well as policy recommendations. Finally, section 4 outlines the sources of differential game theory as a prelude to the next chapter, where the predictive and normative capabilities of static and dynamic game theory are discussed in detail. 2 A matter of milestones This section is devoted to singling out the essential steps made by game theory in order to build up its essential instruments, the equilibrium concepts.1 We begin with Zermelo’s Theorem (Zermelo, 1913), which refers to the outcome of a chess game. The theorem states that either the ‘white’ player or the ‘black’ player wins, or they draw. Moreover, it also says that if both players are equally capable of (a) reconstructing the past history of the game, and (b) forecasting its further evolution at any instant, then the ‘white’ player cannot lose. At least, he must be able to bring the game to a draw. Zermelo’s Theorem is the first application of the axiomatic approach to games, this being a subset of Hilbert’s programme aiming at the complete axiomatization of formal systems. Next we turn to the Minmax Theorem (von Neumann, 1928). Based upon the first formulation of the Fixed Point Theorem (Brouwer, 1910), the minmax equilibrium is the first solution concept for non-cooperative games, although it is applicable only to fixedsum (or zero-sum) games. Indeed, as the original paper was published in German, the potential of the minmax solution became apparent only after the publication of Theory of Games and Economic Behavior (von Neumann and Morgenstern, 1944).2 The cornerstone of game theory, and its applications to social science in general, is the Nash Equilibrium (Nash, 1950, 1951). This solution concept, based upon a later formulation of the Fixed Point Theorem (Kakutani, 1941), is applicable to any zero- or non-zero-sum non-cooperative game and is de facto a generalization of von Neumann’s minmax equilibrium.3 The Cournot (1838) equilibrium is in fact a special case of Nash’s Theorem, and quite a lot of attention has been devoted in the literature to the possible role of Cournot as a forerunner of Nash. In this respect, there exists a lively debate on the paternity of the equilibrium concept. My personal view is that this issue is not very well posed. Leonard (1994), with the aim of showing that Nash cannot be nested into a parental line that goes ultimately back to Cournot, argues against Aumann’s position (Aumann, 1985), according to which the Nash equilibrium concept had always existed in the economic literature, and all that Nash has done is nothing but shaping it into its final and most general formulation.4 My opinion in this regard is that Aumann was far from literally seeing Cournot as a forerunner of Nash. Rather, Aumann may have taken the Platonic view (which, by the way, is common to many mathematicians) maintaining that mathematical concepts have a life of their own and, from time to time, some researcher attains them in a more or less accurate and general form. Nash would be a representative example of the first type, while Cournot, Edgeworth and others are examples of the second type. Be that as it may, over the years it became rapidly clear to the profession that a game could produce a plethora of Nash equilibria. This gave rise to an intensive research effort aimed at defining the so-called ‘refinements’ of the Nash equilibrium concept, such as trembling hand perfection and iterated dominance. To the purpose of the present chapter, it will suffice to draw the reader’s attention to one such refinement, the subgame perfect

The genesis of game theory 407 equilibrium by backward induction (Selten, 1965, 1975). The concept of subgame perfection (or perfectness) allows one to eliminate Nash equilibria based upon non-credible strategies in non-cooperative multi-stage games characterized by perfect information, and it is probably the most widely used equilibrium concept in the current economic literature, not only in industrial economics. The subgame perfect solution by backward induction is closely associated with the extensive form representation of games also known as Kuhn tree (Kuhn, 1953). All of the aforementioned solution concepts apply to non-cooperative (possibly multistage) games characterized by complete information (no matter whether it is perfect or not). The next step deals instead with the issue of solving non-cooperative games under incomplete information: the Bayes–Nash perfect equilibrium (Harsanyi, 1967). This solution concept applies to games of incomplete information, showing that they can be treated as games of complete but imperfect information. A further development, that mainly relies upon the concept of subgame perfection by backward induction, concerns the analysis of repeated games (or supergames), where the constituent game is a prisoners’ dilemma affected by a free-riding incentive that, in the one-shot case, induces the emergence of a Pareto-inefficient Nash equilibrium in (at least weakly) dominant strategies. The main result of this particular stream of research is known as the ‘Folk Theorem’, labelled in this way to account for the fact that it is unclear who coined it first. The earliest formulation of it made use of the infinite reversion to minmax strategies as a deterrence against deviations from the collusive path. The first Folk Theorem in written form is probably Friedman’s (1971), using the Nash reversion, with many follow-ups refining the theorem itself (see Axelrod, 1981; Abreu, 1986; Fudenberg and Maskin, 1986; Segerstrom, 1988, inter alia) through the introduction of (optimal) punishments which are more severe than the Nash behaviour. This brief overview sums up the main steps made by game theory over the decades, by listing the equilibrium concepts that won Nash, Harsanyi and Selten the Nobel Prize in 1994 (see Van Damme and Weibull, 1995). In some sense, to a professional researcher in the social sciences, this is the official story of the evolution of game theory. An analysis of its role in shaping modern industrial economics is proposed in the next section. 3 From industrial economics to the theory of industrial organization The source of industrial economics can be traced back (at least) to Adam Smith, although one usually starts with Cournot (1838), as his is the first instance of a formal analysis of the strategic behaviour of firms. Several aspects of market behaviour are investigated by Walras (1874), Bertrand (1883),5 Edgeworth (1881, 1897), Hotelling (1929) and Chamberlin (1929, 1933). Stackelberg (1934) and Fellner (1949) investigate sequential play between firms, what we would now call ‘a game of perfect information’. Then Bain (1956), Sylos-Labini (1957) and Modigliani (1958) build up the paradigm that will remain the dominant approach to industrial economics for two decades. In the mid-1970s, things begin to change, with game theory starting to filter into industrial economics. In this regard, one may ask: why things went this way? That is, why did not game theory become the dominant approach until the late 1970s, given that the main tools were already available long before? These questions have received several answers, all of which capture different aspects of the issue. According to Schotter and Schwödiauer (1980, p.480):

408

International handbook on industrial policy

By the late 1950s the consensus in the profession was clear: game theory was to be a theory for the small numbers case in economics. As soon as the theory was pigeonholed, however, its popularity waned, for mathematical theorists turned their attention toward the axiomatic analysis of general equilibrium theory.

This amounts to saying that, in Schotter and Schwödiauer’s view, mathematical economics was so busy in completing the construction of the Arrow–Debreu general equilibrium model that there were no resources left for opening the way for the penetration of game theory into mainstream economics or any single part thereof. However, Schotter and Schwödiauer’s assessment should be considered with caution, as they also recognize that game theory rapidly found fertile applications in the field of general equilibrium theory (see Shubik, 1959; Aumann, 1964; Gabszewicz and Vial, 1972). A completely different position is taken by Martin (1993, 2001, 2004), pointing out that the responsibility for the route taken by industrial economics must be attributed to the Chicago school (in particular, to its second generation), which paradoxically should have striven to build up a stark defence of the status quo: Like Pandora, who loosed the ills of the world and found they could not be closed up again, the Second Chicago School evoked formal theory in its contest with the SCP approach, and found it could not close it up again. Faced with the fact that game-theoretic models reproduce, as often as not, the conclusions of the SCP paradigm, the reaction of the Second Chicago School was to reject the use of game theoretic models. (Martin, 2004, p.24)

According to Martin, game theory was adopted to supply formal arguments against the position of the old Chicago school, whereby ‘the model of competitive markets in the long-run equilibrium is sufficient to explain real-world phenomena’ (Martin, 1993, p.9). Once game theory had gained the centre stage, the discipline previously known as industrial economics became first the new industrial economics and then the theory of industrial organization. In summary, its main achievements are essentially three: a.

b. c.

modelling the rational behaviour of firms operating in oligopolistic or monopolistically competitive markets, in terms of entry (and barriers to entry), horizontal and vertical mergers, the choice of productive capacity, product differentiation, R&D and advertising; the formal analysis of implicit collusion; the provision of anti-trust and regulatory prescriptions concerning every aspect listed in (a–b).

This extremely succinct list briefly accounts for what is in fact an extremely wide literature, which consists mainly of contributions based upon applications of static game theory, in the form of either multi-stage or repeated games. Several (if not all) of the aspects of firms’ strategic behaviour in (a) are intrinsically dynamic: accordingly, one could argue that static games, by definition, have a limited explanatory power in these respects. A commonly accepted view maintains that multi-stage games summarize dynamic features that remain de facto behind the curtains, but the solidity of such a statement has been seldom (if ever) proved. That is, for a static game to offer a reliable and fully-fledged characterization of some significant features of firms’ or industries’ dynamics, it must be true (at least) that the subgame perfect equilibrium of such a game grasps the essence of the steady state of the properly dynamic version of the same problem. If

The genesis of game theory 409 this is not the case, then my personal view is that the credibility of any conclusions drawn from a static set-up is, to say the least, questionable. The precise meaning and implications of such a claim, which may well sound too harsh and pessimistic at first, will be explained in detail in the remainder of this chapter as well as in the next chapter. As to point (b), the non-cooperative mechanism sustaining implicit collusion over time hinges upon the assumption that the constituent game (which is a prisoners’ dilemma game) repeats essentially unmodified at every period. This ultimately implies that repeated games are inherently static in nature. All in all, the aforementioned considerations indicate that anti-trust and regulation policies have so far relied upon the conclusions drawn from reduced-form static models that should describe the essential features of firms’ intrinsically dynamic (strategic) behaviour. Surprisingly enough, a specific toolkit for the properly dynamic analysis of firms’ behaviour was being developed at approximately the same time as (if not even earlier than) static non-cooperative game theory, and by researchers working next door to Nash, Kuhn and the other makers of static game theory, in the same research institutions. 4 Differential games at RAND Corporation and the ONR The RAND Corporation (Santa Monica, CA) and the Office of Naval Research (Monterey, CA) were founded respectively in 1945 and 1946 as independent (non-profit) research centres connected with the analysis of the Cold War. The choice of the name ‘RAND’ is connected with the fact that its original location was the R&D plant of Douglas Aircraft at Santa Monica Airport, where the initial Project RAND was set up in December 1945 as a special contract to Douglas Aircraft. In 1947, RAND moved from the Douglas plant at Santa Monica Airport to offices in downtown Santa Monica. In May 1948, RAND was transformed into a non-profit corporation under the laws of California.6 As a matter of fact, many of those instruments that nowadays are so familiar to us were shaped by RAND and ONR researchers7 having in mind, amongst other things,8 how to win a nuclear confrontation with the USSR: Then under the auspices of the U.S. Air Force, RAND was concerned largely with military problems and, to us there, this syllogism seemed incontrovertible: 1) Game theory is the analysis of conflict. 2) Conflict analysis is the means of warfare planning. 3) Therefore game theory is the means of warfare planning. (Isaacs, 1965, p.v)

A detailed account of such activities and their motivation goes far beyond the scope of the present chapter, but two examples will suffice to provide the flavour of it all.9 The first example is familiar to all of us: it is the prisoners’ dilemma game, whereby we usually show to undergraduates taking their first course in microeconomics that, more often than not, Nash equilibria are Pareto-inefficient owing to free-riding incentives. The genesis of the prisoners’ dilemma game, probably the most famous of all, and surely the one that has found the largest number of applications in so many different fields, can be taken as an illustrative example of the source of the impulse driving the development of game theory in the decade following the end of World War II. The game was conceived and carried out experimentally by Merrill Flood and Melvin Dresher in January 1950 at RAND Corporation (see Flood, 1952, 1958). Its immediate application, from the point

410

International handbook on industrial policy

of view of those who financed research activities at RAND, consisted in the analysis of escalation in arms races, although its official version tells a much more comfortable story about two individuals suspected of being jointly responsible for some crime, the district attorney in charge of investigations being, however, unable to obtain any hard evidence in this regard, to be used in court. Therefore, he/she may only rely upon confession by either convict. The free-riding incentive underlying the game drives its equilibrium outcome, whereby both prisoners indeed find it convenient to confess. Clearly, a qualitatively analogous story can be told about two countries examining the opportunity of building up an arsenal of nuclear weapons. The game reflects the same essential properties of the original prisoners’ dilemma, as can be quickly verified without going into any technical details of the game. Intuitively, everybody agrees that a world without nuclear weapons is definitely much better than one where both countries have become nuclear powers. Yet, unfortunately, the equilibrium is precisely the latter as nuclear proliferation is driven by the fact that both countries are well aware that not being endowed with some nuclear deterrence while the other country is, would put them in a very unpleasant position. These aspects were not publicly revealed when, shortly after, Albert Tucker (the supervisor of both Nash and Kuhn) was invited to illustrate the nature and scope of game theory to a non-technical audience (psychologists) at a meeting held in Stanford.10 A similar example is represented by the so-called ‘pursuit–evasion game’, a truly dynamic (that is, differential) game which illustrates the strategic interaction taking place between a thief and a policeman who is trying to catch him. In its simplest formulation, the game unravels over a finite time horizon ending, say, at some instant T which is known a priori to both players. If the policeman does not reach the thief at any instant t T, then the thief ultimately escapes. This game is widely treated in virtually all of the existing textbooks on differential games, entire chapters being devoted to its description (see Intriligator, 1971; Basar and Olsder, 1982, 2nd edition 1995; Clemhout and Wan, 1994; Mehlmann, 1988; and Dockner et al., 2000, inter alia). This, at least in early sources, is done without ever telling explicitly what the story is really about. That is, the matter is in fact the characterization of optimal strategies in a game where the thief is indeed a hostile intruder (for example, an enemy aircraft or missile) while the policeman is the air defence of a country which is being attacked. Given that the aircraft (or the missile) will reach and destroy its objective at time T, the air defence must necessarily intercept it strictly before that instant, that is, at some t T.11 To the best of my knowledge, the first example of a pursuit–evasion game can be found in the famous RAND discussion paper P-257 written by Isaacs (1951), where it is labelled simply as the ‘pursuit game’ (cf. Breitner, 2002, p.116). Rufus Isaacs joined the RAND Corporation in fall 1948. Then, starting from 1949, Isaacs and Bellman discussed new ideas about dynamic games and optimal control theory at RAND seminars whose ultimate outcomes must be identified in Isaacs’s tenet of transition and Bellman’s dynamic programming principle. In fact, the latter can be considered as a special case of the former. The same consideration applies to Pontryagin’s maximum principle.12 The above considerations suffice to prove two related facts. First, the bulk of game theory was not developed to investigate social interaction in general, but rather to deal with problems specifically related to the Cold War and international relations. While it is certainly true to a certain extent that some of the researchers involved in the venture had

The genesis of game theory 411 in mind social as well as economic problems (and John Nash is certainly an example of this kind), one cannot underestimate the strength of the impulse exerted by military applications on the development of both static and differential game theory. Second, this entails in turn that the common view shared by the vast majority of those who deal in various degrees with game theory and its applications is, in a sense, completely mistaken. That is, when reading about game theory at any level, it is very often the case that one is left totally unaware of the true genesis of the discipline, and is led to believe that the first and by far also the widest field of application has been economics. This holds a fortiori for differential game theory, whose earliest application to industrial economics, to the best of my knowledge, is the paper by Clemhout et al. (1971), which unfortunately has been largely disregarded later on in the profession. In this regard, it is also worth stressing that Simaan and Takayama’s (1978) contribution to differential oligopoly theory,13 that has probably created the highest number of follow-ups, was published in Automatica, a journal published under the auspices of IFAC (International Federation of Automatic Control). Therefore one could well say that, for far too long, the analytical power of differential game theory has remained outside the domain of economics in general and industrial economics in particular, dwelling in a sort of no man’s land where, every now and then, a relatively large number of engineers and mathematicians took a detour, being accompanied by comparatively much fewer economists. This may seem an extremely negative or pessimistic view of how things have been unravelling in our profession. This, however, is not entirely true. In a relatively recent past, industrial economics and specifically the theory of managerial firms had undertaken the enterprise of building up a set of models where the dynamics of firms’ behaviour was explicitly taken into account and modelled through the instruments of optimal control theory (Baumol, 1958; Penrose, 1959; Marris, 1963; Williamson, 1966; Solow, 1971; Slater, 1980). Accordingly, what we have to do next is just to revive that line of research in order to explore the potential of a comparatively new approach to relatively old issues, as well as new ones. The next chapter offers an account of the industrial organization literature using the tools of differential game theory, attention being of course confined to the field of industrial economics and policy.14 Notes 1. 2. 3. 4. 5. 6.

For wider and more detailed assessments of the evolution of game theory, see Aumann (1985, 1999) and Myerson (1999), inter alia. The role of Morgenstern should not be underestimated, as acknowledged by von Neumann himself saying, ‘without Oskar I would not have written that book’ (quoted in Samuelson, 2001, p.302). On this, see also Leonard (1995). Although it seems that von Neumann did not (or refused to) acknowledge this fact when Nash went to von Neumann’s office to illustrate the theorem in 1949. For an account of this episode, see Nasar (1998). Concerning this issue, see Daughety (1989) and Martin (2001). Bertrand’s paternity of the oligopoly model with price competition is arguable, as it should be attributed to Launhardt (1885). In this regard, see Dos Santos Ferreira and Thisse (1996), Dos Santos Ferreira (1998) and Lambertini and Mosca (2001). A more detailed account of the history and mission of both institutions can be found in their respective websites, www.rand.org and www.onr.navy.mil. The current scope and mission of ONR are somewhat narrower than those of RAND, at least from an economist’s standpoint. To quote what appears on the ONR website: ‘The Office of Naval Research (ONR) coordinates, executes, and promotes the science and technology programs of the United States Navy and Marine Corps through schools, universities, government laboratories, and nonprofit and for-profit organizations. It provides technical advice to the Chief of Naval

412

7.

8.

9. 10. 11. 12.

13. 14.

International handbook on industrial policy Operations and the Secretary of the Navy and works with industry to improve technology manufacturing processes.’ While RAND publishes the RAND Journal of Economics, which is commonly recognized as a leading journal in the field of industrial economics, the ONR started publishing the Naval Research Logistics Quarterly in 1954, later continued under the heading of Naval Research Logistics, whose primary interest has relatively little in common with economics. In order to give a precise idea of the relevance, scope and quality of research at RAND, it is only fair to recall that among the fellows of RAND’s think-tank, in addition to Albert Tucker, Harold Kuhn, John Nash, Lloyd Shapley and many other founders of static game theory, there were also Richard Bellman, John von Neumann, David Blackwell, Paul Samuelson, John Milnor, Kenneth Arrow, Herbert Simon and many others. Analogous considerations of course hold as for the Office of Naval Research. Among the scientific and technological contributions by RAND researchers, one could list relevant achievements in space systems, as well as important contributions to computing and artificial intelligence. The research carried out by Paul Baran on packet switching yielded the building blocks for Arpanet, Milnet and, ultimately, today’s Internet. For more on these aspects, I refer the interested reader to Mirowski (2002, ch. 6). This episode is reported in Nasar (1998, ch. 12). Recent literature has abandoned this metaphor, to speak openly of the issue at hand. See, for instance, several of the contributions in Petrosjan and Zenkevich (2002), Neck (2003) and Bars and Gyurkovics (2003). Approximately at the same time as Bellman and Isaacs were working at RAND, Pontryagin and many others were indeed developing much the same instruments in the USSR, behind the Iron Curtain. Also, some of their respective results were to be published almost simultaneously, as is the case of differential game theory (Isaacs, 1965; Pontryagin, 1966). See also Isaacs (1954, 1973). The exposition of the basic features of Simaan and Takayama’s model and its later developments (Fershtman and Kamien, 1987, 1990; Cellini and Lambertini, 2004, inter alia) can be found in the next chapter. Optimal control theory and differential game theory have played a relevant role in shaping the debate on the time consistency of optimal policy, initiated by Kydland and Prescott (1977), Calvo (1978) and Barro and Gordon (1983a,1983b). This issue, however, goes well beyond the scope of the present as well as the next chapter.

References Abreu, D.J. (1986), ‘Extremal equilibria of oligopolistic supergames’, Journal of Economic Theory, 39, 191–225. Aumann, R.J. (1964), ‘Markets with a continuum of traders’, Econometrica, 32, 39–50. Aumann, R.J. (1985), ‘What is game theory trying to accomplish?’, in K. Arrow and S. Honkapohja (eds), Frontiers in Economics, Oxford: Blackwell. Aumann, R.J. (1999), ‘Game theory in Israel: looking backward and forward’, ASSET Conference, Tel Aviv, 10–11 October. Axelrod, R. (1981), ‘The emergence of cooperation among egoists’, American Political Science Review, 75, 306–18. Bain, J.S. (1956), Barriers to New Competition, Cambridge, MA: Harvard University Press. Barro, R.J. and D.B. Gordon (1983a), ‘A positive theory of monetary policy in a natural rate model’, Journal of Political Economy, 91, 589–610. Barro, R.J. and D.B. Gordon (1983b), ‘Rules, discretion and reputation in a model of monetary policy’, Journal of Monetary Economics, 12, 101–21. Bars, R. and E. Gyurkovics (eds) (2003), Proceedings of the IFAC Symposium on Control Applications of Optimization (CAO 2003), Oxford: Pergamon Press. Basar, T. and G.J. Olsder (1982), Dynamic Noncooperative Game Theory, San Diego: Academic Press. Baumol, W. (1958), ‘On the theory of oligopoly’, Economica, 25, 187–98. Bertrand, J. (1883), ‘Review’, Journal des Savants, 68, 499–508; republished in English, trans. J.W. Friedman, in A. Daughety (ed.) (1989), Cournot Oligopoly. Characterization and Applications, Cambridge: Cambridge University Press, pp.73–81. Breitner, M.H. (2002), ‘Rufus P. Isaacs and the early years of differential games: a survey and discussion paper’, in L.A. Petrosjan and N.A. Zenkevich (eds), Proceedings of the X International Symposium on Dynamic Games and Applications, St. Petersburg: International Society of Dynamic Games and St. Petersburg State University, vol. I, pp.113–28. Brouwer, L. (1910), ‘Über eineindeutige, stetige Transformationen von Flächen in sich’, Mathematische Annalen, 69, 176–80. Calvo, G. (1978), ‘On the time consistency of optimal policy in a monetary economy’, Econometrica, 46, 1411–28.

The genesis of game theory 413 Cellini, R. and L. Lambertini (2004), ‘Dynamic oligopoly with sticky prices: closed-loop, feedback and openloop solutions’, Journal of Dynamical and Control Systems, 10, 303–14. Chamberlin, E.H. (1929), ‘Duopoly: value when sellers are few’, Quarterly Journal of Economics, 44, 63–100. Chamberlin, E.H. (1933), The Theory of Monopolistic Competition, Cambridge, MA: Harvard University Press. Clemhout, S. and H.Y. Wan, Jr. (1994), ‘Differential games. Economic applications’, in R.J. Aumann and S. Hart (eds), Handbook of Game Theory, Amsterdam: North-Holland, vol. 2, pp.801–25. Clemhout, S., G. Leitmann and H.Y. Wan, Jr. (1971), ‘A differential game model of duopoly’, Econometrica, 39, 911–38. Cournot, A. (1838), Recherches sur les principes mathématiques de la théorie des richesses, Paris: Hachette; translated by N.T. Bacon as Researches into the Mathematical Principles of the Theory of Wealth, New York: Macmillan, 1897; chapter 7 is also in A. Daughety (ed.) (1989), Introduction to Cournot Oligopoly. Characterization and Applications, Cambridge: Cambridge University Press, pp.63–72. Daughety, A.F. (ed.) (1989), Introduction to Cournot Oligopoly. Characterization and Applications, Cambridge: Cambridge University Press. Dockner, E.J., S. Jørgensen, N. Van Long and G. Sorger (2000), Differential Games in Economics and Management Science, Cambridge: Cambridge University Press. Dos Santos Ferreira, R. (1998), ‘Launhardt and Hotelling’, Recherches Economiques de Louvain, 64, 97–109. Dos Santos Ferreira, R. and J.-F. Thisse (1996), ‘Horizontal and vertical differentiation: the Launhardt model’, International Journal of Industrial Organization, 14, 485–506. Edgeworth, F. (1881), Mathematical Psychics, London: Kegan Paul. Edgeworth, F. (1897), ‘La Teoria Pura del Monopolio’, Giornale degli Economisti, 40, 13–31. Fellner, W. (1949), Competition among the Few, New York: Kelly. Fershtman, C. and M.I. Kamien (1987), ‘Dynamic duopolistic competition with sticky prices’, Econometrica, 55, 1151–64. Fershtman, C. and M.I. Kamien (1990), ‘Turnpike properties in a finite-horizon differential game: dynamic duopoly with sticky prices’, International Economic Review, 31, 49–60. Flood, M. (1952), ‘Some experimental games’, Rapporto RAND RM-798, RAND Corporation, Santa Monica, CA. Flood, M. (1958), ‘Some experimental games’, Management Science, 5, 5–26. Friedman, J.W. (1971), ‘A non-cooperative equilibrium for supergames’, Review of Economic Studies, 28, 1–12. Fudenberg, D. and E. Maskin (1986), ‘The Folk Theorem in repeated games with discounting or with incomplete information’, Econometrica, 54, 533–54. Gabszewicz, J.J. and J.-P. Vial (1972), ‘Oligopoly à la Cournot in a general equilibrium analysis’, Journal of Economic Theory, 4, 381–400. Harsanyi, J. (1967), ‘Games with incomplete information played by Bayesian players, Part I, II and III’, Management Science, 14, 159–82, 324–34, 486–502. Hotelling, H. (1929), ‘Stability in competition’, Economic Journal, 39, 41–57. Intriligator, M.D. (1971), Dynamic Optimization and Economic Theory, Englewood Cliffs, NJ: Prentice-Hall. Isaacs, R. (1951), ‘Games of pursuit’, RAND discussion paper P-257, RAND Corporation, Santa Monica, CA. Isaacs, R. (1954), ‘Differential Games, I, II, III, IV’, RM-1391 Reports, 1399, 1411, 1486, RAND Corporation, Santa Monica, CA. Isaacs, R. (1965), Differential Games, New York: Wiley. Isaacs, R. (1973), ‘Differential games: their scope, nature and future’, in A. Blaquiere (ed.), Topics in Differential Games, Amsterdam: North-Holland, pp.1–42. Kakutani, S. (1941), ‘A generalization of Brouwer’s fixed point theorem’, Duke Mathematical Journal, 8, 457–9. Kuhn, H.W. (1953), ‘Extensive games and the problem of information’, in H.W. Kuhn and A.W. Tucker (eds), Contributions to the Theory of Games, vol. 2, Princeton, NJ: Princeton University Press. Kydland, F. and E. Prescott (1977), ‘Rules rather than discretion: the inconsistency of optimal plans’, Journal of Political Economy, 85, 473–92. Lambertini, L. and M. Mosca (2001), ‘Give to Caesar what is Caesar’s. Or, give to Launhardt what we are used to thinking is Bertrand’s’, working paper 413, Dipartimento di Scienze Economiche, Università degli Studi di Bologna. Launhardt, W. (1885), Mathematische Begründung der Volkswirtschaftslehre, Leipzig; trans. B.G. Teubner (1993), Mathematical Principles of Economics, Aldershot, UK and Brookfield, USA: Edward Elgar. Leonard, R.J. (1994), ‘Reading Cournot, reading Nash: the creation and stabilisation of the Nash equilibrium’, Economic Journal, 104, 492–511. Leonard, R.J. (1995), ‘From parlor games to social science: von Neumann, Morgenstern, and the creation of game theory’, Journal of Economic Literature, 33, 730–61. Marris, R. (1963), ‘A model of managerial enterprise’, Quarterly Journal of Economics, 77, 185–209. Martin, S. (1993), Advanced Industrial Economics, Oxford: Blackwell.

414

International handbook on industrial policy

Martin, S. (2001), Advanced Industrial Economics, 2nd edn, Oxford: Blackwell. Martin, S. (2004), ‘Remembrance of things past’, unpublished paper, Purdue University, West Lafayette. Mehlmann, A. (1988), Applied Differential Games, New York: Plenum Press. Mirowski, P. (2002), Machine Dreams. Economics Becomes a Cyborg Science, Cambridge: Cambridge University Press. Modigliani, F. (1958), ‘New developments on the oligopoly front’, Journal of Political Economy, 46, 215–32. Myerson, R. (1999), ‘Nash equilibrium and the history of economic theory’, Journal of Economic Literature, 37, 1067–82. Nasar, S. (1998), A Beautiful Mind, New York: Touchstone Books. Nash, J.F. (1950), ‘Equilibrium points in n-person games’, Proceedings of the National Academy of Science of the USA, 36, 48–9. Nash, J.F. (1951), ‘Non-cooperative games’, Annals of Mathematics, 54, 289–95. Neck. R. (ed.) (2003), Proceedings of the IFAC Symposium on Modelling and Control of Economic Systems (SME 2001), Oxford: Pergamon Press. Penrose, E. (1959), The Theory of the Growth of the Firm, Oxford: Blackwell. Petrosjan, L.A. and N.A. Zenkevich (eds) (2002), Proceedings of the X International Symposium on Dynamic Games and Applications, St. Petersburg: International Society of Dynamic Games and St. Petersburg State University, 2 vols. Pontryagin, L.S. (1966), ‘On the theory of differential games’, Uspekhi Mat. Nauk, 21, 219–74. Samuelson, P.A. (2001), ‘Some game theory anecdotes’, Japan and the World Economy, 13, 299–302. Schotter, A. and G. Schwödiauer (1980), ‘Economics and the theory of games: a survey’, Journal of Economic Literature, 18, 479–527. Segerstrom, P. (1988), ‘Demons and repentance’, Journal of Economic Theory, 45, 32–52. Selten, R. (1965), ‘Spielteoretische Behandlung eines Oligopolmodells mit Nachfragetragheit’, Zeitschrift für di gesamte Staatswissenschaft, 121, 301–24. Selten, R. (1975), ‘Re-examination of the perfectness concept for equilibrium points in extensive games’, International Journal of Game Theory, 4, 25–55. Shubik, M. (1959), ‘Edgeworth market games’, in A.W. Tucker and W.D. Luce (eds), Annals of Mathematical Studies, vol. 4, Contributions to the Theory of Games, Princeton, NJ: Princeton University Press. Simaan, M. and T. Takayama (1978), ‘Game theory applied to dynamic duopoly problems with production constraints’, Automatica, 14, 161–6. Slater, M. (1980), ‘The managerial limitation to the growth of firms’, Economic Journal, 90, 520–28. Solow, R. (1971), ‘Some implications of alternative criteria for the firm’, in R. Marris and A. Wood (eds), The Corporate Economy, London: Macmillan. Sylos-Labini, P. (1957), Oligopolio e progresso tecnico, Turin: Einaudi. van Damme, E. and J. Weibull (1995), ‘Equilibrium in strategic interaction: the contributions of John C. Harsanyi, John F. Nash and Reinhard Selten’, Scandinavian Journal of Economics, 97, 15–40. von Neumann, J. (1928), ‘Zur Theorie der Gesellschaftsspiele’, Mathematische Annalen, 100, 295–320. von Neumann, J. and O. Morgenstern (1944), The Theory of Games and Economic Behavior, Princeton, NJ: Princeton University Press. von Stackelberg, H. (1934), Marktform und Gleichgewicht, Berlin, Heidelberg: Springer-Verlag. Walras, L. (1874), Eléments d’économie politique pure, Lausanne: Corbaz. Williamson, O. (1966), ‘Profit, growth and sales maximisation’, Economica, 33, 1–16. Zermelo, E. (1913), ‘Über eine Anwendung der Mengenlehre auf die Theorie des Schachsspiels’, Proceedings of the Fifth Congress of Mathematicians, 2, 501–4.

21 Differential game-theoretical analysis and industrial policy Roberto Cellini and Luca Lambertini

1 Introduction Even if industrial organization (IO) generally deals with dynamic issues, the most widely used theoretical models are in fact static. This holds for the vast majority of the existing models involving strategic interdependence among agents, and taking a game-theoretical approach. At most, they are multi-stage games, in which the dynamics is implicitly summarized by the backward induction process, or they are repeated static games (as is the case for the various folk theorems employed to investigate collusive behaviour or entry deterrence). A proper dynamic approach in presence of strategic interdependence among firms requires the adoption of differential game theory tools. Differential games are a special class of dynamic games: specifically, they are games in which a set of state variables is introduced, to describe the state of a dynamic system at any particular instant, and its evolution over continuous time.1 In the previous chapter of this Handbook, Lambertini has discussed why differential games have received relatively scant attention from economics and specifically from IO so far. In this chapter we show that, once a proper differential game approach is taken, relevant results may drastically change as compared to the results deriving from a static game approach. More specifically, we focus on a few particular issues, and discuss how a proper dynamic approach affects the resulting industrial policy prescriptions. We focus on five specific problems: (i) the choice between price and quantity competition in oligopolistic markets, in the presence of capacity requirements; (ii) competition versus cooperation in R&D aimed at process innovation in the presence of spillovers in the R&D activity; (iii) horizontal mergers. Furthermore, we quickly deal with the choices concerning investment in (iv) product differentiation and (v) advertising efforts. All of the aforementioned issues are inherently dynamic, in the sense that the time dimension plays an important role in these cases: a serious policy analysis has to take into account the evolution of relevant variables over time. The desirability of this requirement is clearly stated in the available literature. More questionable is whether the available models are really genuine dynamic models. We argue that they are not. In several cases, as we have already mentioned, the time dimension is suppressed, and the analysis is static indeed: this is the case of models that represent the system through a multi-stage game, whose solution is obtained by the standard backward induction procedure. In other cases, the framework is that of a repeated game, in which the repetition of the constituent game may give rise to an equilibrium, different from the equilibrium of the constituent game, along the lines suggested by some version of the Folk Theorem. We aim to show that relevant policy conclusions change once one has taken a really dynamic model approach; that is, a differential game approach. The different conclusions reached by static and differential games are largely due to the fact that the policy maker 415

416

International handbook on industrial policy

can grasp an exhaustive picture of the behaviour of an industry at any instant, if a differential game approach is taken. Hence he/she can design policy intervention accordingly, taking into consideration the evolution of the system, and possibly also the feedback from private agents over time. Here we confine our attention to partial equilibrium models. We assume that the goal of policy makers is simply represented by the maximum social welfare flow (generated by the single market under consideration), measured in present value. Even adopting such a simple policy goal, we will show that the policy prescriptions can differ significantly from those yielded by static game models. The remainder of the chapter is structured as follows. Section 2 introduces the cornerstones of differential game theory. Section 3 proposes a baseline model, which provides a general framework for subsequent discussion. Sections 4 to 7 present specific cases of IO problems and policies. First, we evaluate the social desirability of quantity competition versus price competition (that is, Cournot versus Bertrand competition) in an oligopoly model in which capacity is necessary in order to produce, and it is accumulated over time at some cost. Second, we take into consideration the case in which firms can invest in order to obtain process innovations entailing lower production costs; in this case, we analyse the private and social costs and benefits from R&D cooperation versus competition. Third, we present an analysis concerning the evaluation of horizontal mergers among firms. Finally, we analyse firms’ choices about investment efforts in product differentiation and advertising activities, when such activities take place over time and have long-lasting effects. Section 8 concludes. 2 Elements of differential game theory A proper dynamic approach to analyse strategic interaction among firms over time requires a differential game framework. In this section we introduce the basics of differential game theory: (i) the notions of state and control variables, (ii) the objective functions of players, and (iii) information sets and the related solution concepts.2 2.1 The state variable(s) and the control variable(s) In any dynamic settings, at least one variable changes over time, depending on its past values as well as the players’ choices. This variable is defined as the state variable. An example pertaining to IO may be a firm’s productive capacity or installed capital, which depends upon both the capacity held by the firm in past periods and its current investment decisions. Insofar as there exists strategic interaction among firms, both the optimal investment at any point in time, and the resulting evolution of capacity over time, will clearly depend upon the investment undertaken by all other firms. Define the set of players as P{1, 2, . . . i . . . N}. The actions of players at any time t consists in setting the so-called ‘control’ (or choice) variables. In the aforementioned example, current investment is the control variable of the generic firm i, which must set it optimally over time. Let the game take place over t
(0, T]; the time horizon of the game may well be infinitely long, so that T →
. In the subsequent models, we always consider the case of an infinitely long time period.3 Moreover, let x(t) define a state variable. The state variable might be unique for all players, but it is not necessarily so. In the above example, it is certainly not, because the

Differential game-theoretical analysis

417

accumulation of capacity or capital characterizes every individual firm in the market, possibly in different ways. Put differently, in some cases it is possible to associate each (out of N) state variable with a specific player; in other cases, the number of players is different from the number of state variables, and in some games there is only one state variable (whose dynamics is affected by the choices of all players). Formally, the dynamics of a state variable can be described by the following equation: dx(t)  x(t)  f(x(t),{ui (t)}N i1 ) , dt

(21.1)

where ui denotes the action of player i at time t, that is, the value of the control variable at time t. We can denote by X the set of state variables. The value of the state variables at the beginning of time (t0) is assumed to be known. The behaviour of the state variable(s) over time represents a dynamic constraint for each player. As long as the state variable affects each player’s optimal decisions, and there exists a feedback from the players’ actions to the value of each state variable, strategic interdependence among players emerges. 2.2 The objective function Each player is assumed to have an objective function, to be either maximized or minimized, depending upon the way in which pay-off functions are defined (that is, whether pay-offs denote gains or losses). The function is defined as the discounted value of the flow of pay-offs over time. Define the instantaneous pay-off accruing to player i at time t as i(t) and, for the sake of simplicity, suppose i(t) is a gain (or profit). Of course, the instantaneous pay-off must depend upon the choices made by player i as well as its rivals. Player i’s objective is then 

Max    i (t)etdt,

(21.2)

0

where the factor et discounts future gains. Observe that the discount rate  has no index, owing to the simplifying assumption that all players discount future pay-offs at the same (constant) rate; moreover, notice that we have assumed an infinite time horizon. In order to solve his (dynamic) optimization problem, each player sets the value of his control variable in each period, so that he actually chooses a time path for his control, under the dynamic constraint represented by the behaviour of state variable(s) and the initial conditions. 2.3 Information Differential game theory distinguishes three cases, as concerns the relevant information set available to each player at any date t: 1.

Open-Loop Information (OLI). Common knowledge consists only in the state of the world, that is, the vector of values of the state variables at initial time. At this date, each player sets the path of his control variable (taking into account the expected behaviour of all other players). All decisions are taken at t0, and applied accordingly by players during the whole relevant time span. The solution for each player is

418

2.

3.

International handbook on industrial policy hence given by a differential equation describing the time evolution of the control variable, ui(t)f(t, X(0)). Feedback Information (FI). Players are assumed to know, at any t, the state of the world at t 1, so that the information set at time t can be summarized by the vector of values of the state variables at t1, defined as X(t 1) (or X(t) where  is positive and arbitrarily small, if the game is specified in continuous time, as it is here). Consequently, each player derives a solution rule of the type ui(t)f(X(0), x(t)). Closed-Loop Information (CLI). Under closed-loop (or history-dependent) information, players are assumed to know at date t the whole previous history of the game over the time span [0, t), and the preceding history determines the optimal choice of each player at any instant. In some cases all the relevant information is summarized by the value of state variable(s) at time t.

The above information sets give rise to different behaviour on the part of agents, and consequently to different equilibria, according to which type of information is considered. It is worth mentioning that the same labels also apply to strategies, so that a player is said to adopt a closed- or open-loop or feedback strategy. Moreover, if the game is defined in such a way that strategies depend upon the state but not upon time, they are often referred to as Markovian (see Bas¸ar and Olsder, 1982; Reinganum and Stokey, 1985; Clemhout and Wan, 1994; Dockner et al., 2000). Notice that, apart from the initial conditions, an open-loop strategy is conditioned only on current time, that is, on a minimal amount of information. It has been criticized for being too similar to a static solution, as long as it does not allow for genuine strategic interaction among players over time (see Dockner et al., 2000, p.30). Moreover, in general, a Nash equilibrium under OLI is not subgame-perfect (or strongly-time consistent). However, this solution concept can be defended on the argument that, in several cases, players cannot observe the state of the world over time, and in such a case an open-loop strategy can be taken to design the optimal plan. Moreover, the solution of the game under the open-loop information structure is simpler as compared to more sophisticated solution concepts, and this solution is at least weakly time-consistent in the sense that, if one considers the game over the truncated interval [, T] with 
[0, T], such a solution coincides with the solution of the original game over the same interval, provided agents have played optimally over [0, ].4 In principle, the CLI solution is preferable in that it is subgame-perfect. However, it should be noted that the choice may depend upon the context. Indeed, the main difference between OLI and CLI is that, in the former, players decide by looking at the clock (calendar time), while in the latter, they decide by looking at the stock (the past history of the game). Whether the second picture is more realistic than the first has to be evaluated within the specific environment of the model being used, in relation to the kind of story the model itself tries to account for (Clemhout and Wan, 1994, p.812). In some (a little bit special) cases, the time path followed by all variables coincides under different information structures. 2.4 Equilibrium concepts Exactly as in static games, we may describe strategic interaction between players in different ways. First of all, we may suppose that each player takes all of his opponents’ choices as given, in which case the relevant equilibrium concept, that is, the Nash

Differential game-theoretical analysis

419

equilibrium, can be defined as usual: a set of strategy paths is a Nash equilibrium if each player considers his own action as optimal given the other players’ behaviour (such that this path remains optimal even after observing the rivals’ behaviour). Second, we may consider the Stackelberg equilibrium, where the leader takes the follower’s best reply into account, so that the reaction function of the follower must be inserted into the leader’s problem as an additional constraint. Third, players can cooperate: they can adopt a common objective function defined, for instance, by the sum of individual discounted flows of pay-offs. In the field of industrial organization, this is the case when, for example, firms build up a cartel in order to maximize joint profits with respect to prices, output levels, or investments in R&D to reduce marginal production costs or to introduce new products. Eventually, it is possible to take a social welfare perspective: in such a case, one generally solves a simple optimal control problem, in which the goal is the maximization of the present value of the social welfare associated with the market at any instant over the relevant period. 2.5 Optimization over time Solving a differential game amounts to solving a problem of dynamic optimization with several agents interacting strategically with each other. We are not going into the formal details of dynamic optimization (see Chiang, 1992; Mehlmann, 1988; Bas¸ar and Olsder, 1995). The standard solution is based on the Hamiltonian technique. Generally, a special emphasis is given to the steady state; that is, the configuration of the system in which variables no longer change over time. A ‘steady state equilibrium’ entails two different concepts of equilibrium: (i) it is part of a Nash equilibrium, in the sense that each player is giving his optimal reply to the plans of his opponents, and (ii) the variables (or specific combinations of variables) do not move any longer over time. A steady state equilibrium may exhibit different features as far as its stability is concerned. More precisely, the steady state equilibrium can be as follows: 1. 2. 3. 4.

a stable (unstable) node, when the system non-cyclically converges to (diverges from) that steady state, regardless of where it starts from; stable along a saddle path, when there exists one and only one time path leading to the steady state; a stable (unstable) focus, when the system cyclically converges to (diverges from) from the steady state; a vortex, when the system orbits around the steady state in a perpetual motion.

From a mathematical point of view, the stability property can be studied by considering the Jacobian matrix associated with the dynamic system. From an economic point of view, the stress on the steady state analysis makes sense if the steady state is stable, so that it can be interpreted as the ‘attractor’ of the economic system. 3 The basic model Here we present a model encompassing several contributions. We consider the case in which N single-product firms offer differentiated products over time. At any time t, the inverse demand function for variety i is (see Spence, 1976):

420

International handbook on industrial policy pi (t)  A(t)  Bqi (t)  D(t)  qj (t) , j
i

(21.3)

where D is the symmetric degree of substitutability between any pair of varieties. If D B, products are completely homogeneous; if D0, strategic interaction disappears and firms are independent monopolists. The direct demand function obtains by inverting (21.3): qi (t) 

[B
D(t) (N  2)] pi (t)
D(t)  pj (t) A(t) j
i .  (B
D(t) (N  1)) [B
D(t) (N  1)][B  D(t)]

(21.4)

Suppose that unit variable cost at time t is equal to ci(t) 0. Firm i’s instantaneous profits are i(t) pi (t)qi(t)ci(t)qi(t). The dynamic dimension of the problem can be provided by different factors. In section 4 (and in section 6) we assume that production requires some capacity, which is accumulated over time thanks to specific investments (Kamien and Schwartz, 1979; Fershtman and Muller, 1984; Cellini and Lambertini, 1998). In those sections we postulate A(t)A, D(t)D, ci(t)ci. Put differently, the dynamic nature of the problem is given only by the accumulation of capacity, while reservation prices, marginal costs and product differentiation are taken to be constant over time. In such a framework we investigate the properties of the competition à la Cournot and à la Bertrand and we show how the conclusions regarding the social desirability of different types of competition change as compared to the results deriving from static models. We also analyse the private and social incentive for horizontal mergers among firms (section 6). In section 5 we do not dwell upon the necessity of building up capacity, any such constraints being assumed away. Instead, we assume that specific investment are effective in reducing operative costs, so that ci(t) moves over time thanks to specific investment, that is, R&D activity devoted to process innovation, while we maintain that A(t) and D(t) are constant over time; that is, they are exogenously given parameters. In section 7 we assume constant operative production cost, and we analyse the case where costly investments are aimed at reducing D(t); this means that investments are devoted to increasing product differentiation. Alternatively, the case that Ai(t) moves over time is considered: one example of this type, with firms’ efforts being effective in modifying A(t), is advertising investment, which can enlarge market size. In all cases, we maintain that the ultimate goal of each firm is to maximize its profits (in present value). The goal of a policy maker is assumed to be the maximum market social welfare (in present value), defined as the sum of consumers’ surplus and firms’ profits. 4 Price versus quantity competition Consider the differential game at hand, in which an accumulation process is necessary in order to build capacity.5 Assume that the capacity can be accumulated thanks to specific investment I(t), in a way resembling the capital accumulation in the growth model à la Solow (1956); specifically, the capacity of firm i at time t, denoted by ki(t) evolves over time according to: dki (t)  Ii (t)  ki (t) , dt

(21.5)

Differential game-theoretical analysis

421

where  denotes a depreciation rate.6 Differently from static models dealing with the choice of capacity in two-stage games (see, for example, Beckhan, 1967, Levitan and Shubik, 1972; Kreps and Scheinkman, 1983), in the present framework, capacity is not a choice variable, but it is the result of the choice pertaining to specific investment (and sales) over time. We believe that, in order to analyse the firms’ behaviour concerning the accumulation of productive capacity, the differential game approach is particularly appropriate. Indeed, building capacity needs time; the depreciation of capacity over time is an important factor; the time dimension is important in planning the investment efforts towards capacity accumulation. Furthermore, strategic interaction among firms is an important issue both in the market phase and in the phase when capacity decisions are taken. As to the market competition behaviour, it can be assumed, alternatively, that all firms behave as either quantity setters or price setters. Hence the control variable related to market competition is either qi(t) when all firms are Cournot agents, or pi(t) in the case where firms adopt a Bertrand behaviour. Cellini and Lambertini (2002b) show that the equilibrium outcome of the Bertrand game coincides with the Cournot equilibrium. This amounts to saying that the differential game encompasses the results of the static analysis carried out by Kreps and Scheinkman (1983) in the spirit of the original idea dating back to Edgeworth (1897). The coincidence between the equilibrium allocation under quantity competition and price competition is technically due to the fact that the Bertrand setting obtains through a linear symmetric transformation of the Cournot setting. Therefore the two models are isomorphic and produce the same steady state equilibrium. In contrast, a different equilibrium arises in the game where one firm is a price setter while the other is a quantity setter. In such a case, the transformation that must be operated to obtain the asymmetric model from either the Bertrand or the Cournot model is asymmetric. In particular, the asymmetry affects state and costate equations in such a way that the resulting game is substantially different from the symmetric ones, and does not produce the same equilibrium.7 Steady state equilibrium profits can be used to assess firms’ incentives to adopt price or quantity as the relevant market variable, in the same vein as in Singh and Vives (1984). This comparison highlights that, in the dynamic setting, quantity is no longer a (weakly) dominant strategy, as it is in the static setting. Indeed, it emerges that there exists a parameter region wherein the mixed price–quantity profile is an equilibrium one. Concerning social incentives, the main result is that the level of social welfare in the steady state associated with the asymmetric setting is higher than the steady state social welfare of the symmetric settings where both firms are price setters or both firms are quantity setters. The straightforward implication of this analysis is that the conflict between private and social incentives towards the adoption of prices or quantities may disappear in a proper dynamic model, based on a differential game approach. As a matter of fact, this result modifies the conclusions about the social desirability concerning price or quantity setting, achieved by taking a static game perspective (see, for example, Singh and Vives, 1984). The relevance of this conclusion as concerns the industrial policy implications is self-evident. It is worth remembering that, in the static game, Cournot (Bertrand) competition is better for firms, if goods are substitutes (complements), while Bertrand competition is the

422

International handbook on industrial policy

best setting for a social planner maximizing the social welfare, regardless of the nature of the goods (be they substitutes or complements), as established by Singh and Vives (1984, p.553). In the present model, based on a differential game approach, the symmetric settings are observationally equivalent, so that the social desirability associated with these settings is exactly the same. Conversely, the social welfare is different in the asymmetric framework. This leads us to conclude that the setting providing the larger social welfare in steady state is the mixed setting, where one firm plays à la Cournot and the other one plays à la Bertrand. This conclusion holds when goods are both substitutes and complements. When goods are independent, firms behave as monopolists, and the choice concerning price or quantity is pointless, so that the social welfare associated with the different settings is the same. It is worth stressing, once again, that the welfare considerations refer to the steady state allocations. As one can see, the possibilities are much more articulated than in the static framework considered by Singh and Vives (1984). In particular, it is no longer true that the Bertrand setting is the best one from the social welfare point of view. Thus we have to conclude that, in this case, the static game representation may not be a good simplified version of a dynamic setting. 5 Dynamic R&D with spillover: competition versus cooperation The role of technological spillovers in shaping the incentives to conduct R&D for process innovation has attracted a wide attention in the existing literature on oligopoly theory. The most relevant contributions in this vein are those of d’Aspremont and Jacquemin (1988, 1990), Kamien et al. (1992), Suzumura (1992), Amir (2000) and Cabral (2000), to mention only a few. A general appraisal of the advantages associated with R&D cooperation, and the related policy measures, can be found in Katz and Ordover (1990).8 The above-mentioned contributions share with Brander and Spencer (1983) the concept of R&D as a cost-reducing activity, adding to the Brander–Spencer set-up the possibility of information transmission or technological externalities. The theoretical debate on the private and social advantages generated by R&D cooperation was triggered by an analogous policy debate on the same issue, leading to the National Cooperative Research Act passed in the USA in 1984 (for the EU and Japan, see Goto and Wakasugi, 1988; EC Commission, 1990). Then, following Katz (1986), a large body of literature discussed the theoretical and empirical facets of welfareimproving technology policies based upon two forms of R&D cooperation, namely, R&D cartels and research joint ventures. (A relatively scant attention has been paid to the possibility that any form of R&D cooperation facilitates collusion, either in prices or in quantities. In this regard, see Martin, 1995, Lambertini et al., 1998, 2002, 2003). The model presented by d’Aspremont and Jacquemin (1988, 1990) considers a homogeneous Cournot duopoly, where each firm enjoys a spillover from the rival in terms of the final outcome of R&D activity, in the following sense. To firm i, investing effort ki costs an amount bki2, which captures the presence of decreasing returns to innovative activity, but the total effective R&D contributing to reduce firm i’s marginal cost ci is in fact Ki k
kj , where [0,1] is the technological externality generated from the rival’s investment. Therefore, given a generic initial marginal cost c we have ci  ci  Ki. The model is used to compare two different games: one where firms behave non-cooperatively in choosing both

Differential game-theoretical analysis

423

R&D efforts and output levels, the other where firms form a cartel in the R&D stage, thus choosing R&D investments so as to maximize joint profits in that stage only, while they continue to adopt a Nash behaviour in the market stage.9 Comparing the two set-ups, d’Aspremont and Jacquemin (1988) find that (i) for high spillover levels – specifically,  1/2 – R&D investments (and also clearly cost reduction) are higher under cooperative behaviour, and conversely for low spillovers; (ii) for high spillover levels (that is, (still  1/2), social welfare is higher under cooperative behaviour, and conversely. Unfortunately, they also find that cartel profits are higher than non-cooperative profits when spillovers are low ( 1/2): this yields an undesirable conflict between private and social incentives towards R&D cooperation (or cartelization). The d’Aspremont–Jacquemin model, or some variation of it, has been used to investigate several related issues, such as the possibility of setting up research joint ventures in relation to absorptive capacity (Kamien and Zang, 2000), the efficiency comparison between Bertrand and Cournot behaviour with product differentiation (Qiu, 1997), the endogenization of spillovers (Katsoulacos and Ulph, 1998; Poyago-Theotoky, 1999; Amir and Wooders, 1999, 2003) and the effects of increasing the number of firms in the market (Hinloopen, 2000). A large amount of research has also been carried out on the empirical side (see Lambertini et al., 2004 and the references therein). However, the above-mentioned lack of overlapping between social and private incentives towards cooperation has remained unsolved. The explicit consideration of a differential game framework permits overcoming the aforementioned problem. Cellini and Lambertini (2004c) adopt an explicitly dynamic approach to describe the R&D activity aimed at process innovation, modelled as a differential game whose basic components are as close as possible to the original ones contained in d’Aspremont and Jacquemin (1988). Specifically, the dynamic evolution of the operative cost is assumed to obey the following equation: dci (t)  ci (t)[  ki (t)  kj (t)
], dt

(21.6)

where ki(t) is the R&D effort exerted by firm i at time t, while parameter  measures the positive technological spillover that firm i receives from the R&D activity of firm j and parameter  is a constant depreciation rate measuring the instantaneous decrease in productive efficiency due to the ageing of technology. As in the d’Aspremont and Jaquemin (1988) paper, Cellini and Lambertini (2004c) confine their attention to the alternative cases where firms either behave fully noncooperatively or build up a cartel in R&D investments. In the model, steady state profits and social welfare at the subgame perfect equilibria of the two cases are analysed, finding that, irrespective of the spillover level, R&D cooperation is preferable to non-cooperative behaviour from both a private and a social point of view alike. Intuitively, this result stems from investment smoothing, which is carried out by firms over the time horizon of the dynamic setting, while it is utterly impossible to achieve in a static two-stage game where firms are compelled to invest one-shot in the full amount of resources required to achieve the equilibrium efficiency level of their productive technology. Furthermore the R&D effort consistent with steady state is independent of the market competition type (that is, price competition or quantity competition), but is

424

International handbook on industrial policy

dictated only by technological factors. Both considerations are ruled out by definition in a static setting. Moreover, taking into account the time path of relevant variables along the transition to the steady state, under the different regimes, it is possible to get further results; specifically the possible conflict between individual incentives to R&D and its social efficiency. In any case, the conflict is far from being the rule, as there exist admissible parameter regions wherein private and social incentives indeed coincide. Once again, the static game (or the two-stage game) is far from being an appropriate simplification of a proper dynamic model taking into consideration the time dimension. An additional result worth mentioning refers to the incentive in R&D cooperation in the case where market competition takes place à la Bertrand. Hinloopen (2000) has solved the oligopoly equilibrium with N firms in the static case, finding that both aggregate and individual R&D efforts are non-monotone (first increasing and then decreasing) with respect to the number of firms. The differential counterpart of this model (see Cellini and Lambertini, 2004a) provides a different result: a clearly Arrovian conclusion obtains, since the aggregate R&D effort is everywhere increasing in the number of firms, this being true along the equilibrium path as well as in steady state. The dramatic difference between the results from the dynamic model and the ambiguous conclusions drawn from the static model relies upon smoothing the investment efforts over a long time horizon, a perspective which is ruled out by definition in a static setting. In this respect, the static approach proves to fall short of appropriately accounting for the inherently dynamic nature of research and development which is not captured by multi-stage game modelling. Within this model, which studies optimal R&D investment devoted to process innovation in the presence of market competition à la Bertrand, an additional result concerns the relationship between product differentiation and the investment efforts devoted to process innovation: for a number of firms equal to or larger than three, individual and industry investment are monotonically decreasing in product substitutability; this means that the more similar are varieties, the weaker are investment efforts in R&D aimed at process innovation: in this respect, the result has a Schumpeterian flavour, since the market power given by the product differentiation is associated with larger optimal efforts aimed at process innovation. 6 Horizontal mergers Another relevant case in which a differential game approach leads to different conclusions as compared to a static approach is represented by the regulation of horizontal mergers. The seminal paper by Salant et al. (1983) introduced a famous puzzle about the profitability of horizontal mergers in a Cournot oligopoly with a technology characterized by constant returns to scale, common to all firms. According to their analysis, we should not worry about horizontal mergers unless they involve almost all the firms in the industry. However, their model severely underestimates the incentive towards merging, as shown by Perry and Porter (1985) by introducing fixed costs into the picture. Perry and Porter assume that the aggregate amount of capital is fixed at the industry level, and the associated fixed costs are distributed across firms in proportion to their individual holdings of the capital factor. By doing so, Perry and Porter take into account the intuitive fact that a merger gives rise to a new firm that is bigger than its parts (previously

Differential game-theoretical analysis

425

independent firms) and the new productive unit may produce a larger output at any given marginal costs, as compared to the rivals (as well as the previously independent firms that participated in the merger). Contrary to the conclusions reached by Salant et al. (1983), Perry and Porter conclude that the price increase generated by the merger can often suffice to compensate for the decrease in the output of the merged firm, and therefore may make such a merger a profitable one. The increase in productive efficiency due to the synergies associated with the merger may be sufficiently large to more than compensate for the decrease in consumer surplus. If so, then the merger is also socially convenient. Farrell and Shapiro (1990) interpret mergers as concentration-increasing transfers of industry-specific capital among firms. These transfers affect industry structure, therefore inducing changes in the Cournot equilibrium of the oligopoly market under consideration. Farrell and Shapiro characterize the conditions under which such transfers raise the market price, and examine the social and private incentives to merge, accounting for the fact that mergers alter the distribution of outputs across firms as well as the aggregate output level. They find that small firms typically have insufficient incentives to merge, while large firms have excessive incentives to do so. They also show that a merger is socially more attractive the more concentrated is production among the non-participant firms. The mainstream of the literature dealing with horizontal merger assumes Cournot behaviour to model market interaction. Deneckere and Davidson (1985) consider Bertrand behaviour, and show that any merger is profitable in such a case, since nonmerging firms will match any price increase induced by the merger, as long as prices are strategic complements (that is, products are demand substitutes). Hence it appears that the nature of market competition strongly affects the profitability of mergers. Notwithstanding the fact that mergers heavily affect the evolution of an industry over time, mergers have usually been investigated in static settings. To the best of our knowledge, there exist two relevant exceptions. Relying on Erickson and Pakes (1995), Gowrisankaran (1999) relates merger decision to entry and investment decision in a dynamic (stochastic) model, showing that mergers allow firms to better deal with shocks. In a more recent paper, Dockner and Gaunersdorfer (2001) assess the profitability of horizontal mergers in a dynamic Cournot game with sticky prices, first introduced by Simaan and Takayama (1978) and then further investigated, for example, by Fershtman and Kamien (1987, 1990).10 Dockner and Gaunersdorfer (2001) show that any merger is profitable in such a setting, since the anti-competitive effects due to the merger are strong enough to benefit both the merging and the non-merging firms alike. From the social standpoint, mergers prove to be always detrimental to welfare. In summary, the existing literature stresses the relevance of capital (or productive capacity) in determining whether or not any given merger is profitable and socially efficient. Hence merger analysis represents a natural field of application of dynamic games with capital accumulation to build up productive capacity over time. Cellini and Lambertini (2004b) take such an approach and show that, indeed, the change in the relative (steady state) size of productive plants induced by a merger may play a decisive role in determining whether such a merger is profitable, or socially efficient. However, unlike the existing (static) contributions in the same vein, it is shown that the parameter sets where, respectively, firms find it convenient to merge, and the merger is welfare-increasing, do not intersect at all, irrespective of the capital accumulation dynamics being considered (capital accumulation à la Ramsey or à la Solow). Put differently, it

426

International handbook on industrial policy

cannot happen that a horizontal merger can simultaneously be both privately and socially convenient. The policy conclusion is clear-cut: a regulator concerned with the welfare performance of an industry should prevent firms from carrying out any horizontal merger. Once again, the difference among the results rests on the fact that, in a proper dynamic framework, the individual choice of firms is driven, not only by ‘market competition’ considerations, but also by considerations regarding the optimal amount of intertemporal investment decisions, depreciation factors, and so on. Accordingly, the dynamic model can be interpreted as an extension of the two-stage game of Kreps and Scheinkman (1983), in which, however, it is no longer possible to distinguish between quantity and price competition, as firms’ strategy space is defined over investment efforts. 7 Investment in product differentiation and investment in advertising In this section we mention the most relevant conclusions reached by differential game models that investigate the cases in which the investment effort of firms is devoted to product differentiation, or to advertisement.11 A simple way to investigate product differentiation in the framework at hand is to postulate that D, which captures the symmetric degree of product substitutability, is no longer a paramenter, but a (state) variable, influenced by specific investment carried out by firms over time. The dynamics of the variable can be described by a differential equation of the following type: dD(t) N ),  (D(t),{ki (t)}i1 dt

(21.7)

in which the reduction of D(t) over time can be interpreted as the output of the investment effort of firms.12 It is important to notice that D(t) is common to all firms, so that the effort of any firms to reduce D(t) leads to a beneficial effect for all firms; from this perspective, the reduction of variable D(t) can be interpreted as a pure public good. If the technology able to reduce D(t) is concave (that is, the productivity of the investment effort is decreasing), it is clearly convenient to smooth the investment of product differentiation over time. Moreover, in sharp contrast to the conventional wisdom (see, for example, Tirole, 1988, ch. 7; Eaton and Lipsey, 1989), in this model the private (that is, firms’) incentive to provide product differentiation is lower than the social incentive. This is due to the fact that, in this model, product differentiation is partly a public good; therefore, its benefits spill over completely to all rivals, unlike what we see in so-called ‘address models’ where firms choose locations (as in the literature stemming from Hotelling, 1929) or qualities (as in Shaked and Sutton, 1982, 1983), where the internalized benefit exceeds the external effect. Consequently, setting up an R&D cartel leads to a higher degree of differentiation than in the decentralized case. From the policy making perspective, the market failure originating from the externality and the public good nature of product differentiation could be corrected by an appropriate incentive scheme, supporting the private efforts at investment in product differentiation. A more or less similar conclusion holds with regard to investment in advertising, provided that advertising entails external effects (see Cellini and Lambertini, 2003b). Put simply, consider the case in which the market size of firm i (as measured by Ai) can be enlarged by appropriate investment carried out over time, and assume also that external

Differential game-theoretical analysis

427

effect occurs from advertising investment carried out by all other firms; and assume that the dynamics of variable Ai is described by the following differential equation: dAi (t)  (Ai (t), ki (t), {kj (t)}j
i )  ki (t)
zj
i kj (t)  Ai (t), dt

(21.8)

where kh is the effort in advertising made by firm h, parameter z captures the external effect of the advertising of a firm on the market size of different firms (which can be positive or negative, according to the sign of the effect) and  is a depreciation parameter indicating that market size shrinks as time goes by. Assume also that advertising activities have decreasing marginal productivity, that is, they entail a convex cost. Quantities or prices (depending on the market competition regime) and advertising efforts are the control variables, while market sizes are state variables. Each player chooses the path of her control variables over time, in order to maximize the present value of her profit flow, subject to (i) the motion laws regarding the state variables, and (ii) the initial conditions. Also in this problem a symmetric steady state exists as the solution of the game, and it is possible to evaluate the properties of such a steady state, as compared to the steady state associated with the optimal control problem of a policy maker aiming at the maximum social welfare. The policy implications are straightforward, given the presence of external effects: forbidding collusive agreements among firms when the advertising activity has external effects could turn out to be socially inefficient. 8 Conclusions In this chapter we have focused our attention on the differential game approach to model IO situations in which strategic interaction among actors takes place over time. We have argued that the analysis of such situations by means of a static game approach generally leads to conclusions different from the conclusions derived by a proper dynamic approach. As a consequence, even the resulting industrial policy prescriptions differ, sometimes significantly. Three cases in particular have been highlighted. First, there is the evaluation of price competition versus quantity competition in the presence of accumulation of capacity over time by part of oligopolistic firms; in such a case we have shown that, if all firms compete either in price or in quantity, the competition regime is irrelevant upon the market allocation. This is clearly different from the conclusion obtained by static (or multi-stage) games. Second, we have provided an analysis of the individual and social costs and benefits of cooperation in R&D in the case in which technological spillovers take place; in such a case, a proper dynamic approach permits overcoming the conflict between individual and social incentives found by d’Aspremont and Jacquemin (1988) in a static framework. Third, we have analysed the private and social incentive for firms to merge horizontally. Also in this case, the conflict between the individual and social evaluation of mergers, which seems to appear regularly in a static framework, is far from being the rule in a proper dynamic setting. Summing up, we have provided some theoretical perspectives pointing out that a proper dynamic description of situations evolving over time leads to positive as well as normative conclusions which are rather far from the positive and normative conclusions derived

428

International handbook on industrial policy

from a static game or multi-stage games. The reason for this difference rests on different factors. First, taking into account time explicitly permits us to allow for the fact that players (firms) smooth the time profile of their choice variable over time; this is particularly significant in the presence of decreasing returns to the variable(s) on which decisions are made over time. Second, the steady state amount of several variables is dictated solely by technological conditions, whereas in static models market conditions also affect such variables. Last but not least, there is a technical aspect which also exerts some influence in this respect, namely that the standard backward induction procedure used in static games does not apply to differential ones, where all controls are manoeuvred together at any time during the game. The above-mentioned differences between the results provided by models based on a static versus a differential game approach should suggest particular caution when the policy prescriptions are evaluated, as we have stressed in the foregoing discussion. We believe that a properly dynamic approach would indeed allow us to improve our understanding of industry evolution and consequently our ability to design appropriate policy measures. To this end, of course, a closer cooperation with policy maker is highly desirable. Notes 1.

2. 3. 4.

5. 6. 7. 8. 9. 10. 11. 12.

Here we will limit ourselves to differential games, without considering explicitly dynamic games in discrete time. However, many of the considerations proposed in the remainder of this chapter can be easily extended to dynamic games in discrete time. In this regard, see Bas¸ar and Olsder (1982, 1995), inter alia. This section in based on Cellini and Lambertini (2003a). See also Dockner et al. (2000) or Mehlmann (1988) for a technical introduction to differential games in economics. This may involve no particular loss of generality. In the case of a finite terminal time, the game must be solved taking into account different transversality conditions replacing those used in infinite horizon cases. See below. The limitations affecting open loop solutions are well known (Kydland, 1977; see also Fudenberg and Tirole, 1991, pp.520–36). In principle, it would be preferable to solve a differential game under either CLI or, even better, FI, rather than under OLI. It can be shown that there are classes of games for which the open-loop and the closed-loop solutions coincide (see Reinganum, 1982; Mehlmann and Willing, 1983; Dockner et al., 1985; Fershtman, 1987). For a survey of these classes of games, see Dockner et al. (2000). For an exhaustive discussion of time consistency and subgame perfection in differential game theory, once again see Mehlmann (1988, ch. 4) and Bas¸ar and Olsder (1995). Here we follow Cellini and Lambertini (2001, 2002b). See also Spence (1979) and Reynolds (1987). Alternatively we can imagine that the capacity accumulates thanks to the unsold output, in line with a differential equation of the following type: dki(t)/dt  yi(t)  qi(t)  ki(t), where y denotes the level of production and q the level of sales (see Cellini and Lambertini, 1998). In all cases, the open-loop equilibrium is strongly time-consistent, given that the time path of variables coincide with the path followed under the closed-loop information structure. For an overview of the literature on R&D, see Reinganum (1989). d’Aspremont and Jacquemin (1988) consider also a third case, in which firms cooperate both in the R&D stage and in the market stage; however, we are not interested in this case (the reason being that cartels in the market stage are explicitly forbidden). The same set-up has also been used to analyse other issues, such as the use of quotas and tariffs in models of intra-industry trade. In this regard, see Dockner and Haug (1990, 1991). There exist such a vast number of contributions applying dynamic game theory to advertising that any attempt to give an exhaustive account of this literature would be in vain. In this regard, see Leitmann and Schmitendorf (1978), Erickson (1985), Erickson and Pakes (1995). An overview is in Feichtinger et al. (1994). Cellini and Lambertini (2002a) provide the solution of this game under the open-loop information structure, while in Cellini and Lambertini (2004d) can be found the solution under the closed-loop structure. The two solutions differ, in the sense that the closed loop leads to a lower steady state level of D; that is, it entails a larger investment effort. The economic meaning of the solutions, however, is the same.

Differential game-theoretical analysis

429

References Amir, R. (2000), ‘Modelling imperfectly appropriable R&D via spillovers’, International Journal of Industrial Organization, 18, 1013–32. Amir, R. and J. Wooders (1999), ‘Effects of one-way spillovers on market shares, industry price, welfare, and R&D cooperation’, Journal of Economics and Management Strategy, 8, 223–49. Amir, R. and J. Wooders (2003), ‘Noncooperative versus cooperative R&D with endogenous spillover rates’, Games and Economic Behavior, 42, 183–207. Bas¸ar, T. and G.J. Olsder (1982, 1995), Dynamic Noncooperative Game Theory, San Diego: Academic Press. Beckhan, M. (1967), ‘Edgeworth–Bertrand duopoly revisited’, in R. Henn (ed.), Operations Research-Verfahren, III, Meisenhein: Verlag Anton Hein. Brander, J. and B. Spencer (1983), ‘Strategic commitment with R&D: the symmetric case’, Bell Journal of Economics, 14, 225–35. Cabral, L. (2000), ‘R&D cooperation and product market competition’, International Journal of Industrial Organization, 18, 1033–47. Cellini, R. and L. Lambertini (1998), ‘A dynamic model of differentiated oligopoly with capital accumulation’, Journal of Economic Theory, 83, 145–55. Cellini, R. and L. Lambertini (2001), ‘Differential oligopoly games where the closed-loop memoryless and openloop equilibria coincide’, working paper no. 402, Department of Economics, University of Bologna. Cellini, R. and L. Lambertini (2002a), ‘A differential game approach to investment in product differentiation’, Journal of Economic Dynamics and Control, 51–62. Cellini, R. and L. Lambertini (2002b), ‘Price vs quantity in a dynamic duopoly game with capacity accumulation’, working paper no. 449, DSE University of Bologna. Cellini, R. and L. Lambertini (2003a), ‘Differential oligopoly games’, in P. Bianchi and L. Lambertini (eds), Technology, Information and Market Dynamics, Cheltenham, UK and Northampton, MA, USA: Edward Elgar, pp.173–207. Cellini, R. and L. Lambertini (2003b), ‘Advertising with spillover effects in a differential oligopoly game with differentiated goods’, Central European Journal of Operations Research, 11, 409–23. Cellini, R. and L. Lambertini (2004a), ‘R&D incentives under Bertrand competition: a differential game’, working paper no. 519, DSE University of Bologna. Cellini, R. and L. Lambertini (2004b), ‘Capital accumulation and horizontal mergers in differential oligopoly games’, working paper no. 477, DSE University of Bologna. Cellini, R. and L. Lambertini (2004c), ‘Dynamic R&D with spillovers: competition vs. cooperation’, working paper no. 495, DSE University of Bologna. Cellini, R. and L. Lambertini (2004d), ‘Private and social incentives towards investment in product differentiation’, International Game Theory Review, 6, 493–508. Chiang, A. (1992), Elements of Dynamic Optimization, New York: McGraw-Hill. Clemhout, S. and H.Y. Wan, Jr (1994), ‘Differential games. Economic applications’, in R.J. Aumann and S. Hart (eds), Handbook of Game Theory, Amsterdam: North-Holland, vol. 2, pp.801–25. d’Aspremont, C. and A. Jacquemin (1988), ‘Cooperative and noncooperative R&D in duopoly with spillovers’, American Economic Review, 78, 1133–7. d’Aspremont, C. and A. Jacquemin (1990), ‘Erratum’, American Economic Review, 80, 641–2. Deneckere, R. and C. Davidson (1985), ‘Incentives to form coalitions with Bertrand competition’, RAND Journal of Economics, 16, 473–86. Dockner, E.J. and A. Gaunersdorfer (2001), ‘On the profitability of horizontal mergers in industries with dynamic competition’, Japan and the World Economy, 13, 195–216. Dockner, E.J. and A.A. Haug (1990), ‘Tariffs and quotas under dynamic duopolistic competition’, Journal of International Economics, 29, 147–59. Dockner, E.J. and A.A. Haug (1991), ‘The closed loop motive for voluntary export restraints’, Canadian Journal of Economics, 3, 679–85. Dockner, E.J., G. Feichtinger and S. Jorgensen (1985), ‘Tractable classes of nonzero-sum open-loop Nash differential games: theory and examples’, Journal of Optimization Theory and Applications, 45, 179–97. Dockner, E.J, S. Jørgensen, N. Van Long and G. Sorger (2000), Differential Games in Economics and Management Science, Cambridge: Cambridge University Press. Eaton, B.C. and R.G. Lipsey (1989), ‘Product differentiation’, in R. Schmalensee and R. Willig (eds), Handbook of Industrial Organization, Amsterdam: North-Holland, vol. 1, pp.723–68. EC Commission (1990), Competition Law in the European Communities, volume I, Rules Applicable to Undertakings, Brussels/Luxembourg: EC Commission. Edgeworth, F. (1897), ‘La teoria pura del monopolio’, Giornale degli Economisti, 40, 13–31 reprinted in English as, ‘The pure theory of monopoly’, in F. Edgeworth (1925), Papers Relating to Political Economy, vol. 1, London: Macmillan, pp.111–42.

430

International handbook on industrial policy

Erickson, G.M. (1985), ‘A model of advertising competition’, Journal of Marketing Research, 22, 297–304. Erickson, G.M. and A. Pakes (1995), ‘Markov perfect industry dynamics: a framework for empirical work’, Review of Economic Studies, 62, 53–82. Farrell, J. and C. Shapiro (1990), ‘Horizontal mergers: an equilibrium analysis’, American Economic Review, 80, 107–26. Feichtinger, G., R.F. Hartl and P.S. Sethi (1994), ‘Dynamic optimal control models in advertising: recent developments’, Management Science, 40, 195–226. Fershtman, C. (1987), ‘Identification of classes of differential games for which the open-loop is a degenerated feedback Nash equilibrium’, Journal of Optimization Theory and Applications, 55, 217–31. Fershtman, C. and M.I. Kamien (1987), ‘Dynamic duopolistic competition with sticky prices’, Econometrica, 55, 1151–64. Fershtman, C. and M.I. Kamien (1990), ‘Turnpike properties in a finite-horizon differential game: dynamic duopoly with sticky prices’, International Economic Review, 31, 49–60. Fershtman, C. and E. Muller (1984), ‘Capital accumulation games of infinite duration’, Journal of Economic Theory, 33, 322–39. Fudenberg, D. and J. Tirole (1991), Game Theory, Cambridge, MA: MIT Press. Goto, A. and R. Wakasugi (1988), ‘Technology policy’, in R. Komiya, M. Okuno and K. Suzumura (eds), Industrial Policy of Japan, New York: Academic Press. Gowrisankaran, G. (1999), ‘A dynamic model of endogenous horizontal mergers’, RAND Journal of Economics, 30, 56–83. Hinloopen, J. (2000), ‘Strategic R&D co-operatives’, Research in Economics, 54, 153–85. Hotelling, H. (1929), ‘Stability in competition’, Economic Journal, 39, 41–57. Kamien, M.I. and N. Schwartz (1979), ‘Optimal capital accumulation and durable good production’, Zeitschrift für National Ökonomie, 37, 25–43. Kamien, M.I. and I. Zang (2000), ‘Meet me halfway: research joint ventures and absorptive capacity’, International Journal of Industrial Organization, 18, 995–1012. Kamien, M.I., E. Muller and I. Zang (1992), ‘Cooperative joint ventures and R&D cartels’, American Economic Review, 82, 1293–1306. Katsoulacos, Y. and D. Ulph (1998), ‘Endogenous spillovers and the performance of research joint ventures’, Journal of Industrial Economics, 46, 333–57. Katz, M.L. (1986), ‘An analysis of cooperative research and development’, RAND Journal of Economics, 17, 527–43. Katz, M. and J. Ordover (1990), ‘R&D cooperation and competition’, Brookings Papers, Microeconomics, 137–203. Kreps, D. and J. Scheinkman (1983), ‘Quantity precommitment and Bertrand competition yield Cournot outcomes’, Bell Journal of Economics, 14, 326–37. Kydland, F. (1977), ‘Equilibrium solutions in dynamic dominant-player models’, Journal of Economic Theory, 15, 307–24. Lambertini, L., F. Lotti and E. Santarelli (2004), ‘Infra-industry spillovers and R&D cooperation: theory and evidence’, Economics of Innovation and New Technology, 13, 311–28. Lambertini, L., S. Poddar and D. Sasaki (1998), ‘Standardization and the stability of collusion’, Economics Letters, 58, 303–10. Lambertini, L., S. Poddar and D. Sasaki (2002), ‘Research joint ventures, product differentiation and price collusion’, International Journal of Industrial Organization, 20, 829–54. Lambertini, L., S. Poddar and D. Sasaki (2003), ‘RJVs in product innovation and cartel stability’, Review of Economic Design, 7, 465–77. Leitmann, G. and W.E. Schmitendorf (1978), ‘Profit maximization through advertising: a nonzero sum differential game approach’, IEEE Transactions on Automatic Control, 23, 646–50. Levitan, R. and M. Shubik (1972), ‘Price duopoly and capacity constraints’, International Economic Review, 13, 111–23. Martin, S. (1995), ‘R&D joint ventures and tacit product market collusion’, European Journal of Political Economy, 11, 733–41. Mehlmann, A. (1988), Applied Differential Games, New York: Plenum Press. Mehlmann, A. and R. Willing (1983), ‘On nonunique closed-loop Nash equilibria for a class of differential games with a unique and degenerate feedback solution’, Journal of Optimization Theory and Applications, 41, 463–72. Perry, M.K. and R.H. Porter (1985), ‘Oligopoly and the incentive for horizontal merger’, American Economic Review, 75, 219–27. Poyago-Theotoky, J. (1999), ‘A note on endogenous spillovers in a non-tournament R&D duopoly’, Review of Industrial Organization, 15, 253–62. Qiu, L. (1997), ‘On the dynamic efficiency of Bertrand and Cournot equilibria’, Journal of Economic Theory, 75, 213–29.

Differential game-theoretical analysis

431

Reinganum, J. (1982), ‘A class of differential games for which the closed loop and open loop Nash equilibria coincide’, Journal of Optimization Theory and Applications, 36, 253–62. Reinganum, J. (1989), ‘The timing of innovation: research, development and diffusion?, in R. Schmalensee and R. Willig (eds), Handbook of Industrial Organization, vol. 1, Amsterdam: North-Holland. Reinganum, J. and N. Stokey (1985), ‘Oligopoly extraction of a common property natural resource: the importance of period of commitment in dynamic games’, International Economic Review, 26, 161–73. Reynolds, S.S. (1987), ‘Capacity investment, preemption and commitment in an infinite horizon model’, International Economic Review, 28, 69–88. Salant, S.W., S. Switzer and R.J. Reynolds (1983), ‘Losses from horizontal merger: the effects of an exogenous change in industry structure on Cournot–Nash equilibrium’, Quarterly Journal of Economics, 98, 185–213. Shaked, A. and J. Sutton (1982), ‘Relaxing price competition through product differentiation’, Review of Economic Studies, 49, 3–13. Shaked, A. and J. Sutton (1983), ‘Natural oligopolies’, Econometrica, 51, 1469–83. Simaan, M. and T. Takayama (1978), ‘Game theory applied to dynamic duopoly problems with production constraints’, Automatica, 14, 161–6. Singh, N. and X. Vives (1984), ‘Price and quantity competition in a differentiated duopoly’, Rand Journal of Economics, 15, 546–54. Solow, R. (1956), ‘A contribution to the theory of economic growth’, Quarterly Journal of Economics, 70, 65–94. Spence, M. (1976), ‘Product differentiation and welfare’, American Economic Review, 66, 407–14. Spence, M. (1979), ‘Investment strategy and growth in a new market’, Bell Journal of Economics, 10, 1–19. Suzumura, K. (1992), ‘Cooperative and noncooperative R&D in duopoly with spillovers’, American Economics Review, 82, 1307–20. Tirole, J. (1988), The Theory of Industrial Organization, Cambridge, MA: MIT Press.

PART VI PARTICULAR VIEWS

22 Industrial policy: perspectives, experience, issues Christos N. Pitelis

1 Introduction The objective of this chapter is to provide a bird’s eye view of extant alternative perspectives on industrial (and competition) policies, paying particular attention to the case of the European Union (EU). We also briefly explore the idea that non-classical ideas have currently achieved an almost mainstream status within EU policy circles. We claim this to be good news and (up to a point) a vindication of efforts by contributors to this volume (and to many others) over many years. Section 2 of the chapter deals with definitional issues and extant perspectives. Section 3 discusses practice, sketches an integrative framework, assesses recent EU policies that seem to draw on evolutionary/resource-based ideas, and discusses implications for transition and developing countries, and issues of implementation. Section 4 offers a summary and conclusions. 2 Industrial and competition policies: theories and international experience Definitional issues The term ‘industrial policy’ (IP hereafter) refers to a set of measures taken by a government and aiming at influencing a country’s industrial performance towards a desired objective.1 Competition policies refer to the stance governments adopt towards the role of competition between firms and (in) industries’ economic development, and the measures they take to implement their objectives. Competition policies (CPs hereafter) usually attempt to influence the degree of competition in industries, for example, the food or textile industry. In this context they are part and parcel of a more general category, that of supply-side industrial policies. As all government measures and policies affect industry one way or the other, boundaries between competition/industrial policy and other policies, such as technology policy, regional policy, structural policy, competitiveness policy, and even macroeconomic policy, are not always clear. The closest we can get to a demarcation line is arguably by referring to the government’s own perception of what they aim CP and IP to be, plus an underlying body of theoretical knowledge hopefully informing such perceptions. The government’s objective is assumed to be the improvement of the welfare of its citizens, which is achieved when resources are allocated efficiently, and wealth creation, and especially capture, is taking place at a pace preferably faster than in other countries (improved international competitiveness). Industry is believed to be an important contributor to the wealth creation process for numerous reasons, for example the tradability of its products, its positive link with technology, innovation and productivity growth, and even its close links with services. It follows that a government wishing to improve welfare will be well advised to design measures that lead to efficient allocation of resources and the creation of wealth, that is, to a strong, productive, competitive and therefore wealthy economy. There is universal agreement among economists that competition between firms in industries can be a potent means of facilitating the desired objective. However, views differ as to the 435

436

International handbook on industrial policy

role, the type, the degree and even the nature (including the very definition) of competition. It is not possible or even useful to discuss all these issues in detail in this chapter, but a bird eye’s view of alternative perspectives can facilitate understanding, including the link between competition, industrial and competitiveness policies. Perspectives The neoclassical industrial organization (IO) perspective The dominant perspective on industrial and competition policy today arguably still remains the mainstream (neoclassical) theory of competition, monopoly and IO. In the context of this approach, competition is seen as a type of industry structure. This can be perfect or imperfect. Perfect competition is characterized by the existence of numerous firms, which produce very similar (homogenous) products, full and symmetrically distributed knowledge about firm and industry conditions (demand and cost curves, in particular), and free mobility of resources, for example no barriers to entry and exit of firms in the industry. Under such conditions firms are price takers; they cannot influence the prices, which are determined by the interplay of supply and demand in the industry. In addition, such prices only cover average costs, and there are only normal profits. The opposite to perfect competition is monopoly. Here we have only one firm in the industry, with blockaded entry. When this is the case, a monopolist that maximizes profits will charge a price that is higher than the price of the perfectly competitive firm, by restricting output. As a result, consumers will end up with lower quantities of goods for which they pay higher prices. In addition to this being bad for consumers, it also leads to misallocation of resources, because output is being restricted compared to that of perfect competition. The neoclassical view is that the leading concern of an economy should be to allocate efficiently its scarce resources and in so doing to maximize the welfare of its consumers. In this context monopoly leads to market failure due to the ‘wrong’ type of market structure (thus it is called structural market failure). When such failures exist, it is believed that the government can step in to correct them. However, there are problems. Firstly, in reality, the two polar opposites, monopoly and perfect competition, are recognized to be unrealistic, the most prevalent form of industry structure being some sort of ‘imperfect competition’, such as ‘monopolistic competition’ or more likely ‘oligopolistic competition’. An industry structure is characterized as ‘oligopoly’ when there is interdependence between (usually a relatively small number of) firms. When one firm acts, the other is affected and needs to react. This raises the question of how imperfect competition is linked to resource allocation. A second problem is that the comparison so far between monopoly and perfect competition assumes that these face the same cost and demand conditions, that they have the same information/knowledge, technology, resources and competences, and therefore the same impact on economic performance. If this is not the case, one has to take into account any such differences. The theory of IO has built on the above discussion and developed models of oligopoly. The best known early models are the ‘limit pricing’, the contestable markets and the ‘generalized oligopoly’ models. Various contributors, such as Modigliani (1958), developed the first one in the 1950s. The model recognizes that there exist barriers to entry in an industry, such as economies of large-scale production (scale economies), product differentiation, initial capital requirements, advantages due to differences in the absolute

Industrial policy: perspectives, experience, issues

437

costs that different firms face, and so on. In addition, it is recognized that such barriers cannot fully protect established firms (incumbents) from entry, except in rare cases. In normal cases, firms, which wish to charge the maximum price that does not induce entry, will have to find a lower, ‘limit’ price. This is the price that is low enough not to attract entry. In the limiting case where the only barrier to entry is the scale of output, the limit price is found by simply taking away the minimum efficient scale (MES) of output from the perfectly competitive output. If we assume that entrants will enter at MES in order not to suffer cost disadvantages, the limit price ensures no entry. This is because, even if one firm enters with MES, the resulting output and, therefore, price becomes the perfectly competitive one. However, entrants are attracted by above-normal profits. Realizing that their very action of entering the industry reduces profits to normal, they will choose not to enter. If oligopolistic firms charge limit prices, consumers will pay less than in the case of monopoly, but still more than in the case of perfect competition. However, assume that there are no barriers to entry and exit is costless (there are no sunk, that is, nonrecoverable, costs). In this case, it can be argued that any price departures from average costs will invite entrants, who can then exit costlessly, if prices are driven down. This ‘hit and run’ behaviour by potential entrants will tend to ensure prices that are perfectly competitive, even in the case of oligopoly (Baumol, 1982). If, however, we assume instead that incumbents can somehow blockade entry, it can be shown that, if they also maximize profits and collude to increase prices, they can again achieve monopoly prices and outputs (Cowling and Waterson, 1976). Clearly, whether prices will be competitive (or contestable), limit prices or monopoly prices will depend on the existence of barriers to entry and exit (mobility barriers). These need not be only structural (MES), as assumed by the limit-pricing model. They can also be strategic, namely the result of conscious action by firms to restrict entry. Indeed, limit pricing itself can be seen as such a policy. Instead of reducing prices, however, firms can do other things; they can advertise, innovate, invest in excess capacity and/or produce many apparently competing products (product proliferation) with the express purpose of reducing entry. Depending on the extent and degree of success of such actions, the resulting industry price–output outcome can be anywhere between competition and monopoly! This long detour has actually not helped very much, other than in pointing out that every case is different and that what we need is empirical evidence from real-life cases, as well as some measures of the losses in allocative efficiency due to the prevailing degree of monopoly. To find out the degree of monopoly, economists usually use measures of concentration. If an industry is highly concentrated it is presumed that there is prima facie evidence for reduced competition, possibility for collusion, strategic barriers and high prices. However, the link between concentration and market power in the form of higher profits is questionable; it can be that more efficient firms grow larger (which increases concentration), and are also more profitable. The crucial issue in this context is collusion over prices, and barriers to entry. The first is normally illegal in most countries, and notoriously difficult to identify. Studies on barriers to entry have confirmed their existence and importance. Even they, however, could be seen as an inducement to innovation (see below). To obtain an indication of the importance of monopoly welfare losses at the economywide level, one has to try and measure such losses. There are various ways to do this and a large empirical literature. The results vary greatly, but it is widely believed that some

438

International handbook on industrial policy

static losses do exist, therefore that monopoly is a potentially serious problem (see Scherer and Ross, 1990). As already suggested, such findings fail to account for any difference between perfectly competitive firms and oligopolistic firms with a degree of monopoly power, in terms of efficiency. There are various dimensions to this issue. One refers to differences in costs. Oliver Williamson, for example, has argued that monopolies may have lower cost curves, which implies an efficiency gain vis-à-vis the perfectly competitive industry. This gain should be traded off against any static losses (Williamson’s trade-off). Perhaps most importantly, oligopolistic market structures may be more prone to invent and innovate. This can result in dynamic productivity benefits, which must also be taken into account. The literature on the link between market structure and innovation has failed to detect any significant differences in relative innovative records. In this context, one might be justified in looking at the static losses and concluding that monopoly is a problem and needs to be dealt with by the government. In principle, the government could step in to ensure perfectly competitive markets by, for example, discouraging mergers and acquisitions, encouraging mobility and even breaking up large firms, and so on. If this took place simultaneously in all industries, consumer welfare would be maximized. However, if this is not the case, perfectly competitive structures in one industry but not in others need not improve overall welfare. This is the problem of ‘second best’. In a second best world, what can be the scope of competition policy? Many economists believe that a degree of ‘workable competition’ is still desirable. This could take the form of guarding against the acquisition and abuse of monopoly power. In practice, this can be done by, for example, disallowing a firm from achieving a certain market share and/or pursuing certain restrictive practices, namely practices that restrict competition. Various governments have adopted this view. For example, it is explicitly acknowledged in the Treaty of Rome (original articles 85 and 86) and the antitrust policy of the USA (see Pitelis, 2000). Transaction costs A way through which firms may increase their size and, in this way, industry concentration, is internalization of activities that could potentially be placed at the market. For example, firms may take over their suppliers or distributors rather than dealing with them at arm’s length. There are numerous reasons why firms could thus ‘integrate’, including the pursuit of market power, reduction in the forces of competition (Porter, 1980) and so on. However, one possibility is that firms integrate because market exchange is costly; finding and dealing with other firms can lead to high exchange or transaction costs. Coase (1937), Williamson (1975) and many other economists believe that reducing market transaction costs is an important reason for the existence and boundaries (and therefore, the size) of firms. If increasing firm size is the result of transaction cost reduction, such efficiency gains should also be taken into account by regulatory bodies. A vertical acquisition, for example, could be motivated by efficiency, not market power motives. If this is so, the relevant authorities should take this into account. Overall, such considerations suggest a more lenient attitude towards large oligopolistic firms. An almost ideal scenario would be transaction costs-motivated integration in contestable markets. The observed industry structure would be the result of efficiency and, if oligopolistic, it would still charge competitive prices. In reality, it is unlikely that transaction costs will be the exclusive determinant of firm size and/or that markets will be contestable. Moreover, this approach takes technology

Industrial policy: perspectives, experience, issues

439

and innovation as constant, which is rather limiting. Nevertheless, both views seem to have influenced industrial and competition policy experts, and they should certainly not be ignored. Evolutionary/resource-based perspectives It is arguable that the dominance of the neoclassical view on competition, monopoly and industrial and competition policy is currently under serious threat. This is the result of the emergence and current popularity of an alternative perspective, which can broadly be defined as the evolutionary/resource and competence-based perspective. Within this broad category, there is a diverse group of contributors. However, between them they share the view that competition is not just a type of market structure, and that what is important is not just the efficient allocation of scarce resources, but also the creation and capture of value and wealth, and that the two are not necessarily the same. There is a wide belief that firms are very important contributors to value/wealth creation and capture, and also that each firm is an individual entity, which differs from other firms primarily in terms of its distinct resources, capabilities or competences. The lineage of this perspective is impressive. It can be claimed to include founding fathers, such as Adam Smith (1776) and Karl Marx (1959), and more recently influential economists such as Joseph Schumpeter (1942), Edith Penrose (1959), George Richardson (1972) and Nelson and Winter (1982). In brief, classical economists, such as Smith and Marx, focused on wealth creation, not just resource allocation. They both saw competition as a process, regulating prices and profit rates, not a type of market structure. Smith described the amazing productivity gains through specialization, the division of labour, the generation of skills and inventions within the (pin) factory. Marx suggested there is a dialectical relation between monopoly and competition (whereby competition leads to monopoly and monopoly can only maintain itself through the competitive struggle). Marx focused in addition on competition (conflict) within the factory, and in the society at large, between employers and employees. Building critically on Marx, Joseph Schumpeter described competition as a process of creative destruction through innovations. He saw monopoly as a necessary and just (yet only temporary) reward for innovations. He attributed firm differential performance to differential innovativeness and saw concentration to be the result of such innovativeness. Penrose’s now classic 1959 book on The Theory of the Growth of the Firm, is arguably a glue that can bind all such contributions together. In the book, firms are seen as bundles of resources in which interaction generates knowledge, which releases resources. ‘Excess resources’ are an incentive to management for (endogenous) growth and innovation as they can be put to use at almost zero marginal cost. Differential innovations and growth lead to concentration, which, however, can also be maintained through monopolistic practices. The world is seen as one of big business competition, where competition is god and the devil at the same time. It drives innovativeness, yet it is through its restrictions that monopoly profit can be maintained. Building on Penrose, Richardson (1972) observed that firms compete but also cooperate extensively. Such cooperation is not just price collusion as the neoclassical theory assumes. It lies between market and hierarchy, and occurs when firm activities are complementary but dissimilar (requiring different capabilities).

440

International handbook on industrial policy

Nelson and Winter (1982) have developed ideas currently of import to the resourcebased view. Notable are those of firm ‘routines’, which simultaneously encapsulate firms’ unique package of knowledge, skills and competences, allow firms to operate in an evolving environment with a degree of path-dependent institutionalization that does not necessarily rely on continuous redesign, and pass on the evolving, ‘routines’ to the evolving organization. The focus on the evolutionary and resource-based view on evolution, knowledge and innovation, as well as its ‘systemic’ (as opposed to market) perspective, has arguably facilitated the emergence of a major change in industrial, competition and competitiveness policies, one that emphasizes the knowledge and innovation-producing potential of different institutional configurations. The ‘national’, regional and sectoral systems of innovation approach, the literature on clusters of firms, and the work of Michael Porter (1990) on national competitiveness, draw upon, and relate to, the evolutionary/resourcebased view (see Wignaraja, 2003, for various contributions). There are various other implications of the evolutionary/resource-based perspective on industrial and competition policies. First, the focus on value and wealth creation and capture through differential competitiveness suggests a broader welfare criterion than just the consumer surplus. Second, differential competences provide another efficiency-based reason for concentration. Third, competition as a dynamic process of creative destruction through innovation implies a need to account for the determinants to innovate, when considering the effects of ‘monopoly’, but also more widely. Fourth, competition with cooperation (co-opetition) as in Richardson implies the need to account for the potential productivity benefits of co-opetition, in devising competition policies. Another dimension on competition relates to its strength, and the role of proximity and location. This links to the work of Richardson, but has been developed by Porter (1990), Krugman (1991), Audretsch (1998b), Dunning (1998) and others. For example, Porter claims that local competition is more potent than foreign competition. This may have important implications in devising domestic competition policies; see below. 3 Practice, an integrative framework and some implications International practice and lessons from experience Despite its limitations, the neoclassical perspective has arguably dominated industrial and competition policy thinking in the Western world for many decades. For example, the various antitrust legislation in the USA, as well as the original Articles 85 and 86 of the Treaty of Rome in Europe, seem to be directly informed and influenced by the aforementioned perspective. At least this seems to be the case in theory. The practice has varied from theory, and also between countries and in time. As we have argued elsewhere (Pitelis, 1994), the European policy, for example, can be described as ad hoc, discontinuous and even inconsistent: ad hoc because the theoretical basis of various policies was not clear. A notable example is the ‘national champions’ or ‘picking winners’ policy, which various European countries pursued in the 1960s and 1970s. This involved identifying potentially successful firms or industries and using a number of measures such as subsidies, tax breaks and so on to promote them. It also involved a lenient and even encouraging attitude towards mergers and in cases (often in pursuit of considerations of fairness and distribution) of nationalization of utilities, but also other ‘strategic’ industries. Underlying was the hope that such firms could compete successfully with foreign rivals, thus raising

Industrial policy: perspectives, experience, issues

441

export surpluses and the country’s competitiveness. Evidently, this tended to exacerbate structural market failures, and was also inconsistent with the theoretical pursuit of ‘perfect competition’. The policy was also pursued at a pan-European level, in search of a pan-European company, which could out-compete large American multinationals. It is believed that such policies blunted incentives for protected firms to compete, and gave rise to ‘problematic enterprises’, or ‘lame ducks’.2 After trying to rescue them for a number of years, European governments, led by Mrs Thatcher’s Britain, eventually resorted to deregulation and privatization, as well as a switch of focus to small firms and entrepreneurship. This also resulted in discontinuity of policies, from large firms and the government, to small firms and the market. The approach of Japan, and the so-called ‘tigers’ of the Far East, was different. The policy of, for example, Japan, led by the Ministry of International Trade and Industry (MITI) was not informed by neoclassical economics. It involved a strongly interventionist approach by the government aimed at creating dynamic advantages in sectors of the future. Such sectors were chosen on the basis of being high value-added, high-income elasticity of demand and gradually knowledge-intensive. In such sectors, MITI provided financial and other support and guidance. It regulated the degree of competition (neither too little, nor too fierce) by aiming at an ‘optimum’ number of firms in it, and protected these sectors from foreign competition at the same time. It also paid attention to the benefits of cooperation and the promotion of small and medium-sized enterprises (SMEs) (Best, 1990). Overall, the approach to competition could be described as domestically focused competition balanced with cooperation (co-opetition). The approach of the East Asian ‘tigers’ was similar, albeit some of them, especially Singapore, affected ‘technology transfer’, not through licensing as practised by the Japanese, but through an aggressive inward investment policy. The performance of the Japanese economy and that of the tigers has been very impressive until recently. It is not surprising that commentators attributed this success in part to its approach to competition and industrial policy (as well as to other characteristics of the Far Eastern economies, such as education, equitable distribution of incomes and high saving ratios). To attribute the success of the Far East just to its approach to competition and its interventionist IP, especially given similarly interventionist policies by Western governments in the past, implies either misconceived policies by the latter, and/or a differential degree of (in)competence. This may well be the case, but there is also a second potential argument. In contrast to the West, the Japanese rejected (were thankfully unaware of ?) the neoclassical perspective and favoured an approach that focused on resource creation (not just allocation) through dynamic competition for innovations, growth, productivity and competitiveness. This approach, which seems to combine Schumpeterian and Penrosean ideas with its accompanied focus on production and organization (Best, 1990), may well be a differentia specifica of the Far Eastern approach. It has been associated with major innovations: among others, total quality, ‘just-in-time’, lifetime employment and the coexistence of competition with cooperation (co-opetition). There have been numerous developments in economics and management in recent years (the resource-based perspective, the new international trade theory, endogenous growth theory, new location economics, ‘new competition’ and so on), some discussed above (see also Wignaraja, 2003, for more). These arguably offer some support to the Japanese perspective and policies. In part owing to both of these, recent approaches to competition

442

International handbook on industrial policy

and industrial policies in the Western world have tended to move away from the neoclassical perspective, towards an approach and policies aimed at improving competitiveness at the firm and macro levels. There are various versions of this new approach. The ‘new industrial policy’ approach, for example, retains its neoclassical flavour but emphasizes input, linkages and technology policies as incentive-compatible means of improving firm and industry competitiveness (see Audretsch, 1998a). More general competitiveness models, such as Michael Porter’s, focus on the role of firm clusters and other determinants of competitiveness (Porter, 1990). Cluster policy is seen as a new IP (Porter, 1998), based on co-opetition. The focus by the European Union on education, (soft) infrastructure, technology and innovation and (clusters of) small firms in the late 1990s, represents a move in this direction. An integrative framework The aforementioned discussion suggests the need for a framework on competition and industrial policy, which takes into account the broader considerations presented in this section, importantly the role of industrial competition (and cooperation) as a process of sustainable innovation, value and wealth creation. This requires a discussion of the determinants of sustainable wealth creation. Building on the works presented so far, it can be argued that productivity, and (thus) value/wealth creation within firms, is determined by firms’ organization and structures (infrastructure) and strategy, their human resources (labour, management, entrepreneurs), innovativeness and unit cost economies (of scale, scope, learning growth, transactions, external, and economies of diversity and pluralism). At the meso and national levels these same factors, plus a supportive regional and a supply side-friendly macro policy and facilitatory institutional context, can act positively on value creation (see Pitelis, 1998). Industrial and competition policies can be derived within this context. Industrial, competition and cooperation (co-opetition) can positively influence all determinants of productivity and value creation. All the same productivity enhancements may lead to differential advantages that can be used to restrict competition. The need for a domestic competition and cooperation (co-opetition) policy thus arises from the need not to thwart the beneficial effects of co-opetition on productivity and value creation. Firm coopetition strategies (for example, firm clusters) that enhance productivity should not be stopped, but facilitated in this context. Non-productivity enhancing forms of cooperation (like collusion), instead, should be forcefully discouraged. The same is true for other restrictive business practices. Mergers and acquisitions should be examined on a case-by-case basis, as they may have productivity-enhancing attributes, but may also lead to market power, which can eventually stifle incentives for innovation and productivity. Pluralism and diversity should be encouraged, as they provide benchmarks for comparison and thus information. Institutional changes facilitating a productivity-enhancing culture and ideology and productivity-compatible legal frameworks should be aimed at. Industrial and competition policies should be compatible with macroeconomic and other policies, but also be supported by a facilitatory institutional context. Douglass North (1981) has shown the importance of institutions and institutional change in reducing transaction and transformation costs and increasing productivity and growth.

Industrial policy: perspectives, experience, issues

443

Institutions, but also culture, attitudes and ideology, can be hugely important factors in economic organization. Governments can be a potent catalyst in institutional change, as they possess a monopoly of force and the ability to legislate and regulate. Devising a facilitatory framework is part and parcel of industrial and competition policy. The neoclassical ‘market failure’ theory of the state assumes the institutional context is given. The possibility to vary it implies a more proactive role for the state. In this context, the state should not just intervene when markets fail. Rather, it should legislate and regulate proactively, so that markets, firms and itself should fail less, see below. In sum, our integrative framework points to the need for a broader conceptual framework for industrial and competition policy, to account for the role of innovation, cooperation, institutions and knowledge creation through diversity and pluralism. The need for tough competition policy measures that discourage the emergence and exploitation of market dominance is maintained and even strengthened in this framework; but the focus is not just on consumer welfare, but on overall productivity, value and wealth creation. Despite apparent similarities to the neoclassical perspective, the one developed here is claimed also to provide some new and different insights on domestic competition policy, which, in addition, are most relevant for developing countries, see below. The EU in the 2000s Particularly interesting, possibly impressive and, up to a certain point, a vindication of the efforts of contributors to this volume (among many others, of course) in the past 25 years or so, is the shift of the EU to a non-neoclassical, arguably evolutionary/resource-based approach to its industrial policy in the early 2000s. First of all, and importantly, the very term ‘industrial policy’ has returned, following years of ‘disrepute’ and a focus on ‘horizontal measures’. Related to this, the ‘sectoral’ element has also resurfaced. Last, but not least, recent EU policy reads very much like a synthesis of various contributions to the European Network of Industrial Policy (EUNIP), over many years, and quite similar to what has been argued above (see also Pitelis, 1998, 2001). We focus on two EU documents here: first, ‘Industrial Policy in an Enlarged Europe’ (2002), and second, ‘Fostering Industrial Change: Industrial Policy in an Enlarged Europe’ (2004). Besides the focus on enlargement, both documents use the term ‘industrial policy’. This is significant. The major themes of the 2002 document are as follows: (a) industry matters! (exclamation mark added), (b) enlargement is an opportunity, (c) sustainability matters, (d) horizontal policy measures need to be applied in response to specific sectoral needs! (exclamation mark added), and (e) policies need to contribute to competitiveness. The objective of the 2004 document is for ‘industrial policy’ to accompany the process of industrial change (‘deindustrialization’). Proposed ‘actions’ include a ‘regulation framework’, ‘synergies of policies’ and a ‘sectoral dimension’. The importance of industry, ‘deindustrialization’, ‘competitiveness’, the ‘sectoral dimension’, synergies of policies and systemic view, regulation and sustainability are all well known and accepted themes within the contributions of this book and many of its readers. Indicatively, these are discussed, among many others, in Pitelis (1994, 1998, 2001) and Pitelis and Antonakis (1998). In this context, EU policies in the new millennium are more in line than ever before with the evolutionary/resource-based view. This is both good news and, in our view, of more relevance to ‘developing countries’ than

444

International handbook on industrial policy

‘Washington Consensus’-type approaches (see below). They are, therefore, to be welcomed. Nevertheless, the implications of the evolutionary/resource-based view and the integrative framework we have sketched here go a lot further than current EU projects. Sustainability requires an analysis of the relationship between private and public governance, the implications of the extant power structures, recognition of the ‘hierarchy of agencies’ (shareholders, firms, nations, communities, the globe) and the potential need for ‘global governance’. These issues are beyond the scope of this chapter (but see Pitelis, 2004, for an early attempt to address them). For our purposes here it suffices to note that current EU policy is very much in the right direction, of use to the EU itself and to developing countries, a partial vindication of contributions in the evolutionary/resource-based perspective, and needing to explore further the implications of its themes and objectives on ‘global governance’. Clearly, the issue of implementation is also important. In what follows, we make some observations on a particular aspect of implementation, that of ‘cluster’ policies. Developing and transition countries and some issues of implementation A problem with much of our discussion concerning the role of competition and industrial policies is that it presupposes the existence of well-functioning markets, an entrepreneurial class and the existence of a structural and institutional framework, which facilitate the implementation of chosen policies. This is evidently not the case in developing countries which need to create firms, markets, entrepreneurship, competitive advantages and, in addition, to ensure the existence of competitive forces. In the early stages of economic development it is common for countries to rely on some sort of support for domestic industry. This is in line with theoretical arguments of the infant industry type, and can be of benefit to a country, provided it is not permanent. When it is, the usual problem of disincentives manifests itself. Firms in protected industries will lack incentives to innovate. Related to this, they will tend to focus on the domestic market, which will seem to be more profitable. This will not expose firms to international competition, reducing further these incentives to innovate so as to become internationally competitive. Domestic and international competition, and also export rivalry, can be useful complementary means of providing incentives to firms to improve efficiency and innovate (see Kikeri et al., 1992). However, it is widely agreed that international competition cannot serve as an adequate substitute for domestic competition (see Porter, 1990; World Bank, 1998). In addition, to expose domestic firms to (foreign) competition, in the absence of the prerequisites discussed above, could have detrimental effects on domestic industry. This is a fundamental issue for developing countries. Generally, developing and transition economies are typically characterized by highly concentrated industries, large state-owned sectors and firms, and firms (public or private) operating in industries protected by various barriers, thus insulated from the forces of competition. The issue of industrial and competition policies should therefore be seen within this context, as follows: develop institutions and capabilities which help you adopt industrial and competition policies which expose state-owned firms to competitive forces, while at the same time facilitating the creation of new SMEs which, alongside existing firms, ‘co-opete’ between themselves (for example, in the form of clusters) and with the larger firms to increase productivity and competitiveness. Implementing industrial and competition policies requires setting up requisite authorities. Here the selection of competent, knowledgable and independently minded individuals

Industrial policy: perspectives, experience, issues

445

is crucial. They should aim at ensuring sustainable competitiveness, an important determinant of which is competition. They should not be ‘captive’ to business or other interests, and should coordinate with regulatory bodies and other authorities, domestically and internationally. They should be able to devise clearly articulated and transparent rules, concerning the acquisition and abuse of dominant positions and vehicles for their attainance, such as mergers and acquisitions, as well as restrictive practices. At the same time they should recognize that competition policy is not a panacea for competitiveness; competitiveness depends significantly on the other determinants of productivity discussed in section 1. All these issues are closely interlinked. This implies the need for a systemic approach to industrial and competition policy that tries to address simultaneously the issues of doing, while also addressing the prerequisites, such as capability building. Building on our earlier discussion, industrial and competition policies should not only be linked to the degree of competition in industries, but should aim at improving productivity and efficient resource allocation. A prerequisite to achieving this is to encourage inter- and intra-firm competition so as to nurture conditions favourable to the creation of new ideas, techniques, products, processes, organizational and institutional forms and, moreover, to exploit best for this purpose the information-providing (and enhancing) attributes of economic organizations, notably markets, firms, states and people at large. Competition policy should provide incentives, support, mechanisms and institutions for achieving productivity and competitiveness, for example through linkages, joint inputs and resource mobility (Audretsch, 1998a). They should address ‘state’ capture by sectional interests – in part through striving for conditions of contestability in private and (up to a point) political markets3 – and a plurality of institutional and organizational forms, including, for example, support for (clusters of) SMEs (see below). Pluralism can also enhance the generation and use of new knowledge. The exact measures which need to be taken to achieve the above can vary according to the conditions prevailing in every country. For example, the recognition of the benefits of cooperation, and therefore the need to ensure competition and cooperation, suggest the need for measures facilitating the ‘clustering’ of SMEs (see Best, 1990; Best and Forrant, 1996). ‘Clusters’ of SMEs can also be a potent source of indigenous development for less favoured regions (LFRs), countering a dependence on TNCs, and can themselves be a determinant of inward investment (see box below).

CLUSTERS AND CO-OPETITION Scale is essential for efficiency through economies of scale and learning. M&As are often an easy way to do this and to rationalize, especially in declining sectors. A minimum efficient scale is also essential to compete in export markets, often a sine qua non for small developing countries. An alternative to acquiring the benefits of scale can be through clustering. Theory and (our own) experience suggest the following good practice. 1. Explore the possibilities of clustering. When these are present, try to develop clusters.

446

International handbook on industrial policy 2. Encourage M&As of SMEs where clustering is not possible, and benefits from scale are required for competitiveness. 3. Promote competition through gradually removing barriers to domestic competition, import competition and export rivalry. Regulate any (natural) monopolies.

For the successful implementation of industrial and competition policies, a crucial factor is the institutional framework. For North, ‘the central issue of economic history and of economic development is to account for the evolution of political and economic institutions that create an economic environment that induces increasing productivity’ (1991, p.98). Also the analysis of institutions and institutional change ‘offer the promise of dramatic new understanding of economic performance and economic change’ (North, 1991, p.111). This is particularly important for developing countries. Examples of required institutional measures include the delineation and enforcement of property rights and a pluralism of organizational forms and ownership structures, which exploit existing, and generate new, knowledge through economies of pluralism. Important also is an attempt to promote attitudes, values and generally culture conducive to innovativeness, thus to productivity, growth and convergence. All these are easier said than done. A way through which these can be achieved is with the government assuming the role of a catalyst, by identifying and implementing, in close cooperation with the private sector, changes proposed by those nearer to the action, such as the private sector itself. Such bottom-up policies exploit dispersed knowledge and also promote subsidiarity and democracy. Precise actions, however, should be based on an analysis of each particular case. This is beyond the scope of this chapter, but the following methodology can be proposed:4 first, a consensually agreed upon theoretical framework; second, an audit of the external (international) environment; third, an audit of the internal (national) environment; fourth, deciding the direction of the strategy; fifth, its dimensions; sixth, the required actions; seventh, addressing the issues of prerequisites, resources and mechanisms for implementation; eighth, feasible actions; ninth, control-evaluation; and tenth, new actions for implementation. To conclude, industrial and competition policy should focus on the nurturing of institutions, mechanisms and organizations, which foster dynamic efficiency, productivity and value creation, competition and cooperation (co-opetition) other than price collusion.5 Concerning the degree of competition per se, measures taken to promote domestic competition should work alongside import competition and export rivalry. When steps are taken to support domestic industry, these should not be allowed to consolidate and become anti-incentives for improved competitiveness. A clear phasing out strategy should be in place. Measures to remove barriers to mobility are essential in this respect, as there seems to be the need to coordinate entry and exit policies (Frydman et al., 1999). In this context, mergers should be discouraged when there is risk of monopoly power and (strategic) barriers to entry, and only encouraged between SMEs. Co-opetition should be of the type described in the case of clusters. Clustering can and should be seen as a new form of competition and industrial policy (Porter, 1998). Clusters can provide locally based development and also be an attraction to inward investment. They are of the utmost importance in that they could allow for the simul-

Industrial policy: perspectives, experience, issues

447

taneous removal of the constraints outlined at the beginning of this section, in that they facilitate entrepreneurship, decentralization and locationally specific advantages, simultaneously. Overall, in their complex interrelationships, the exploitation of knowledge through the existence of a plurality of institutional and organizational forms, the benefits of coopetition also arising from these and appropriate competition policies, the related amelioration of the problem of state capture, and the parallel exploitation of the benefits of clustering, can enhance productivity and development. To summarize, industrial and competition policies should aim at enhancing the benefits from competition and cooperation (co-opetition), (for) innovation, productivity and value creation. As noted, such benefits can fail to materialize when firms use market power to restrain competition. In this context, mergers that can lead to the acquisition of monopoly power should be discouraged, as well as restrictive and collusive practices. When support of industry is adopted for infant industries and related reasons there should be a clear and well understood phasing out clause. Protectionism can be a most potent disincentive for firms to become internationally competitive (Kikeri et al., 1992). 4 Summary and conclusions We have discussed industrial and competition policies in theory and in practice, with an eye to drawing lessons from theory and experience on conceptual and implementation issues pertaining to these topics, of use to policy makers. In brief, we have found out that industrial and competition policies were often motivated by neoclassical ideas, which are currently challenged by alternative views. In practice, competition policies varied between and within countries and were often inconsistent with their alleged objectives. Current EU policies represent a step in the right direction, and a partial vindication of evolutionary and resource-based ideas. They are of use to the EU itself and to transition and developing countries. Developing countries often lack the competences required for successful implementation of industrial and competition policy. In addition, it is crucial for these countries to develop markets, entrepreneurship and competences, including institutional and social capital. Industrial and competition policies, in general, and in developing countries, should be seen within the broader context of enhancing productivity and competitiveness. At the theoretical level, domestic competition policy should aim at maximizing the net benefits from co-opetition. In this context, privatization can be a means of addressing problems of government failures, introducing competition and enhancing incentives for productivity. Establishing a regulatory framework can be important for the potential benefits of privatization to realize; as it is to learn from experience. It is important to observe that the road to productivity and competitiveness is not one-way. Developing countries should exploit the informational benefits from the existence of a plurality of institutional and organizational forms, to include clusters and state-owned enterprises. Theory and history suggest there are no panaceas. Current EU policies are a step in the right direction, of potential help to the EU and to policy makers in the developing and transition economies. Notes 1. Industry refers in the main to manufacturing. This, however, tends to recede, given an emerging fuzziness for the boundaries between manufacturing and services.

448

International handbook on industrial policy

2. However, it would be dangerous to generalize. For example, it is questionable that, for example, the European car industry could have developed in the absence of government support. 3. There can be too much contestability in public sector markets, in that it can increase the susceptibility of politicians, bureaucrats and so on to pressures by organized interest groups, leading to regulatory capture. 4. This is based on the author’s own experience with policy making in Greece, where he has coordinated the ‘Future of Greek Industry Project’, a consensus-based, bottom-up industrial strategy, orchestrated by the government and supported by the major social partners (see next section). 5. This need not exclude (threats to) protectionism, either, both in support of such players and as a means of ensuring fair and open trade.

References Audretsch, D.B. (1998a) (ed.), Industrial Policy and Competitive Advantage. Volume 1: The Mandate for Industrial Policy, Cheltenham, UK and Lyme, USA: Edward Elgar. Audretsch, D.B. (1998b), ‘Agglomeration and the location of innovative activity’, Oxford Review of Economic Policy, 14(2), 18–29. Baumol, W. (1982), ‘Contestable markets: an uprising in the theory of industry structure’, American Economic Review, 72, 1–15. Best, M. (1990), The New Competition: Institutions for Industrial Restructuring, Oxford: Polity Press. Best, M. and R. Forrant (1996), ‘Creating industrial capacity: Pentagon-led versus production-led industrial policies’, in J. Michie (ed.), Creating Industrial Capacity, Oxford: Polity Press. Coase, R.H. (1937), ‘The nature of the firm’, Economica, 4, 386–405. Cowling, K. and M. Waterson (1976), ‘Price–cost margins and market structure’, Economica, 43(171), 267–74. Dunning, J.H. (1998), ‘Location and the multinational enterprise: a neglected factor?’ Journal of International Business Studies, 29(1), 1st quarter, 45–66. European Commission (2002), ‘Industrial policy in an enlarged Europe’, COM(2002) 714 final, 11.12.2002, Brussels. European Commission (2004), ‘Fostering industrial change: industrial policy in an enlarged Europe’, COM(2004) 274 final, 20.04.2004, Brussels. Frydman, R. (1999), ‘When does privatisation work? The impact of private ownership on corporate performance in the transition economies’, The Quarterly Journal of Economics, November, 1153–91. Kikeri, S., J. Nellis and M. Shirley (1992), Privatisation: The Lessons of Experience, Washington, DC: The World Bank. Krugman, P.R. (1991), ‘Increasing returns and economic geography’, Journal of Political Economy, 99, 183–99. Marx, K. (1959), Capital, London: Lawrence and Wishart. Modigliani, F. (1958), ‘New developments on the oligopoly front’, Journal of Political Economy, 66(3), 215–32. Nelson, R.R. and S.G. Winter (1982), An Evolutionary Theory of Economic Change, Cambridge, MA: Belknap/Harvard University Press. North, D.C. (1981), Structure and Change in Economic History, London and New York: Norton. North, D.C. (1991), ‘Institutions’, Journal of Economic Perspectives, 5(1), 97–112. Penrose, E. (1959), The Theory of the Growth of the Firm, Oxford: Basil Blackwell. Pitelis, C.N. (1994), ‘Industrial strategy: for Britain in Europe in the world’, Journal of Economic Studies, 21(5), 2–92. Pitelis, C.N. (1998), ‘Productivity, competitiveness and convergence in the European economy’, Contributions to Political Economy, 17, 1–20. Pitelis, C.N. (2000), ‘A theory of the (growth of the) transnational firm: a Penrosean perspective’, Contributions to Political Economy, 19, 71–89. Pitelis, C.N. (2001), ‘Industrial strategy’, in M. Warner (ed.), International Encyclopedia of Business and Management, 2nd rev. edn, London: Routledge/ITBP, vol. 2, pp.2026–44. Pitelis, C.N. (2004), ‘(Corporate) governance, (shareholder) value and (sustainable) economic performance’, Corporate Governance: An International Review, 12(2), 210–23. Pitelis, C.N. and N. Antonakis (eds) (1998), International Competitiveness and Industrial Strategy, Athens: Tipothito (in Greek). Porter, M.E. (1980), Competitive Strategy: Techniques for Analyzing Industries and Competitors, New York: The Free Press. Porter, M.E. (1990), The Competitive Advantage of Nations, Basingstoke: Macmillan. Porter, M.E. (1998), ‘Clusters and the new economics of competition’, Harvard Business Review, 76(6), 77–90. Richardson, G.B. (1972), ‘The organisation of industry’, Economic Journal, 82(327), 883–96. Scherer, F.M. and D. Ross (1990), Industrial Market Structure and Economic Performance, 3rd edn, Boston: Houghton Mufflin. Schumpeter, J. (1942), Capitalism, Socialism and Democracy; 5th edn, 1987, London: Unwin Hyman.

Industrial policy: perspectives, experience, issues

449

Smith, A. (1976), An Inquiry into the Nature and Causes of the Wealth of Nations, (reprinted in R.H. Campbell and A.S. Skinner (eds); first pub. 1776), Oxford: Clarendon Press. The World Bank (1998), Competition Policy in a Global Economy: An Interpretive Survey, Washington, DC: The World Bank. Wignaraja, G. (ed.) (2003), Competitiveness Strategy and Industrial Performance in Developing Countries: a Manual Policy Analysis, London: Routledge. Williamson, O.E. (1975), Markets and Hierarchies: Analysis and Antitrust Implications: a Study in the Economics of Internal Organization, New York: The Free Press.

Index advertising, investment in 426–7, 437 Africa 80–81, 84, 202, 203–4 agglomerations 19, 59, 322, 323, 324, 381, 382, 396 Amin, Ash 19, 255, 321–41, 355, 366 Annerstedt, Jan 279–97 anti-trust policies European Union 7–11, 16, 19, 23 and innovation 116, 126 and monopoly regulation 7–11 United States 7, 8, 19, 125, 141–2, 143, 438, 440 Argentina decentralization 22, 23, 328, 335 deindustrialization 327 economic crises 328, 330, 334, 385 employment 327 entrepreneurship 328, 329, 332 industrial clusters 328, 330, 331 industrial policy 326, 327–36 innovation 329–31, 332 institutional set-up 331–2 manufacturing 327, 328, 332–3, 334 micro policy initiatives 328–9, 334 MNCs 327 networking 329 privatization 201, 328, 332 regeneration 333 SMEs 328, 329–30, 331, 334, 385, 386, 390, 391, 392, 397 state intervention 332–3, 335 Arora, Ashish 28–44, 217, 221, 222, 224, 228, 272, 396 Australia 185, 190, 208, 256, 305 Austria 58, 124, 145, 189, 208, 389 balanced development theory 18 Bangladesh 30 Belgium 6, 58, 145, 208, 260, 268, 382, 389 Bellandi, Marco 342–61 Bellini, Nicola 362–79 Bertrand competition 420, 421–2, 423 Bianchi, Patrizio 3–27, 323, 325, 326, 345, 355, 380–402 biopharmaceuticals 222, 224, 228, 230 biotechnology 70, 258, 302, 307–8, 344 genome research 18, 258, 344 patents 219, 220, 230, 231 Björkroth, Tom 180–97

Brazil economic crises 385 Embraer 82 employment 33 entrepreneurship 34 exports 82 GDP 33 healthcare 300, 301 human capital 34–6, 39 human capital to USA 35–6, 39 industrial policy 326, 334 IT graduates 34, 35, 40 privatization 207 R&D 200 SMEs 386, 390, 391, 395, 396 software sector 23, 32–6, 41, 391 state ownership 199, 200, 207 subsidy battle with Canada 135 Bulgaria 57 business incubators 274 business support services 55–6, 64, 81 additionality tests 372–3 customer satisfaction 104, 364, 374 evaluation of 371–5 expectation management 369–71 industrial and regional economic approach 365–6 institutional thickness 366 knowledge transfer 364, 373 management and marketing 368–71, 373 market failures 365, 366, 375 policy 363–5 policy science view 366–8 and private sector 366–7, 374–5 and risk 375–6 segmentation 369 and SMEs 362, 364, 365, 366, 373 specificity of 364–5 subsidies 364 taxonomy of 363–4 Canada 56, 135, 208, 291–2, 301 capitalism 3–7, 9, 204–7, 211, 342–3 Cellini, Roberto 415–31 Cesaroni, Fabrizio 217–42 chemical industry 4, 9, 37, 54, 67, 70, 222, 224, 231 Chile 82, 201, 209, 334, 385–6, 390–91, 396–7 China

451

452

International handbook on industrial policy

competition 82, 382 electronics industry 72 employment 33 entrepreneurship 34 exports 82, 84 GDP 33 growth rates, high technology 87 healthcare 301 human capital 34–6, 39 human capital to USA 35–6, 39 ICT 70 industrial policy 16, 66, 81, 82, 344 IT graduates 34, 35, 40 manufacturing 80, 82, 83, 84, 85–6, 87 privatization 208, 209–10 R&D 72, 271 science parks 285 SMEs 386, 387, 393–4, 395, 396 software sector 23, 32–6, 41 state-ownership 209–10 tariffs 62 technology spin-offs 271 Colombia 386, 390, 391–2 comparative advantages theory 28, 29–30, 39, 104 competition 3, 4, 5, 8, 15, 20, 21, 28–9, 67, 79–80, 257, 310 advantage theory 322–3 cooperative agreements 122–5 developing countries 64, 79, 101, 312, 326, 444–7 European Union 8, 10–14, 23, 46, 48–55, 58, 60, 62–70, 72, 116, 119–20, 126–7, 136–7, 143–7, 173–4, 183–4, 187, 191, 208–10, 441, 443 and free trade 79–80 and game theory see game theory and globalization 15, 79 imperfect 180, 436 and industrial policy 16, 119–20, 306, 435–40, 442 and innovation 101, 115–33, 439, 440 integrative framework 442–3 international 7, 22, 64 and licensing 224 and market abuse 106 market delineation 122, 125–6 mergers and acquisitions 122, 127–9, 153, 173–5 monopolization 122, 126–7, 128 neoclassical perspective 436–8, 439, 440, 442, 443 network effects 121–2, 126, 127 perfect 436, 438, 441 pharmaceuticals 309–11

and production 19, 100, 103, 120 public utility sectors 180–82, 189, 190, 191, 205 restricted 343–4 and science parks 288–90 and SMEs 381, 382, 396, 397–8 software industry 237 and state intervention 120 tax 135–8 versus cooperation 422–4 and wage levels 100 welfare effects 137, 435–7, 440 copyright 234, 245, 246–7 see also intellectual property rights Costa Rica 386 Cournot equilibrium 406, 407, 420, 421–2, 423, 425 Czech Republic 145, 208 De Bandt, Jacques 98–111 decentralization 19, 22, 23, 104, 321–41, 328, 335 deindustrialization 52, 54, 66, 88, 98–9, 327, 443 delocalization of production 98, 100–101 Denmark 145, 187–8, 208, 286, 306, 309, 362, 382, 389 developing countries 4, 79, 258 and competition 64, 79, 101, 312, 326, 355–6, 444–7 decentralization 321–41 donation programmes 312 entrepôt strategies 88 entrepreneurship 355 exports 80, 85–6, 87 free markets 79, 85 health policy 312 human capital 79 import substitution policies 82, 89 and industrial clusters 355–6 industrial development 18, 20, 23, 79–97, 118, 199, 326–33, 444–7 manufacturing 80–81, 87 MNCs 86, 88–9, 355, 356 networking 396 privatization 201–2, 206 R&D 89 specializations 4, 30, 31 state ownership 199, 202, 206 technology alliances 87, 118 trade barriers 202, 444 TRIPS agreement 220 see also individual countries Di Tommaso, Marco 298, 301, 342–61 diversification 28, 31–2

Index East Asia APEC 392 economic crises 83, 387 employment 33, 37, 387, 392, 441 exports 81, 82, 84, 85, 92 FDI 81, 86, 89, 90, 92, 93 ICT 70 industrial policy 9, 16, 20, 22, 23–4, 66, 79–97, 107, 109, 326, 344, 441 interventionism 81, 88, 89, 92 manufacturing 80–87 protectionism 81 R&D 4, 38, 67, 71, 72, 81, 86, 88–91, 90, 271, 422 science parks 88, 285 SMEs 85, 86, 90, 381, 385–7, 392–6, 441 software industry 4, 23, 32–41, 256, 395 specialization 30, 88 state ownership 200, 208–10 subsidies 81, 89, 90 technology alliances 115, 117–18 see also individual countries EC Treaty article 4 60 article 87 69, 143, 144 article 89 51, 144 article 123 146 article 157 52–3, 55, 60, 64, 388 article 295 49 ECSC Treaty 50 see also European Union economic crises 83, 328, 330, 334, 385, 387 and economic renaissance 16–20 Ecuador 386 education 39, 42, 68–9, 90 Egypt 199, 200, 326 El Salvador 390, 391 electricity 187–9, 204, 205 electronics industry 4, 72, 222, 229 employment East Asia 33, 37, 386, 387, 392, 441 European Union 13, 33, 55, 63, 68, 146 Latin America 33, 327 mobility 35–6 and SMEs 380, 383–5 United States 33, 37 entrepreneurship 34, 291, 326, 355, 385, 392 business support services see business support services European Union 34, 54, 66, 68, 265 failures in 30–31 United States 7, 8, 18, 34, 265, 267, 275 entry barriers 54, 66, 120, 125, 126, 127, 235, 437 environmental technologies 70, 224

453

European Council 54 European Union accession countries and state aid 145, 146–7 agglomeration effects 59 agriculture 50, 67, 140, 144, 145, 146 air transport 12, 59, 70 Altmark case 144 anti-trust policies 7–11, 16, 19, 23 Bangemann Reports 13, 49, 54, 63–4 banking sector 9, 10, 156 Barcelona European Council 68 BEST programme 13 ‘better regulation’ principle 60 biotechnology 70, 258 Bolkestein directive 56 brain drain 67–8, 69 business incubators 274, 275 business support services 55–6, 64, 362–3 cars 70, 145 chemical industry 9, 54, 67, 70 Christopherson report 58, 63 clothing quotas 51 coal industry 69, 145 Coal and Steel Community (ECSC) 50 Commission Regulation 2659/2000 123 Common Agricultural Policy 50, 146 competition 8, 10–14, 23, 46, 48–55, 58, 60, 62–70, 72, 116, 119–20, 126–7, 136–7, 143–7, 173–4, 183–4, 187, 191, 208–10, 441, 443 Competitiveness Advisory Group 54, 64 consumer protection 56 convergence 143–7 cooperative federalism 60, 66 CopyLeft principles 258, 259 crisis cartels 62 cross-cutting issues 51–9 defence 72, 140, 210 deindustrialization 52, 54, 66, 443 delocalization 136–7 Directive 98/4/CE 258 EC Treaty see EC Treaty ECJ 50, 60, 62, 144 economic integration 4, 45, 46, 48, 52 education 42, 68–9 eEurope programme 70–71 electricity sector 187–8, 189 employment 13, 33, 55, 63, 68, 146 energy security 52 enterprise policies 46, 54 entrepreneurship 34, 54, 66, 68, 265 entry barriers 54, 66 environmental technology 70 ESPRIT programme 49, 70 Essen European Council 58

454

International handbook on industrial policy

Euratom Treaty 13, 71 EUREKA programme 71 Euro-Mediterranean Partnership 210 europeanization 55, 57 Eurovignette system 59 Financial Perspectives 68 fiscal policy 52, 64, 65, 66 Framework Programmes 13, 18, 46–8, 50, 56–7, 64, 67, 71, 184, 388, 443 GEANT university network 59 and globalization 14, 382 growth 38–9, 53, 57, 63, 66, 71 Growth, Competitiveness and Employment White paper 58, 63 harmonization 48, 49, 56, 208 healthcare 300, 301, 304 High Level Groups 53 human capital 34–6, 38, 68 i2010 programme 70, 71 ICT 10, 23, 55, 59, 70–71 India, software outsourcing 36–9 industrial policy 11–14, 19–24, 45–78, 107, 144, 209, 325, 326, 344, 347, 438, 440, 443–4 industrial revolution 6, 8–9 Information Society 63–4, 71 infrastructure 50, 56–9 innovation 13, 18, 19, 47, 49, 52, 53, 68, 265, 389 Innovation Scoreboard 68 internal market regime 48–9, 59–60 interventionist policy 45, 48, 53, 67, 69, 70, 199 investment schemes 57, 58 Japanese MNCs in 37 liberalization 48, 49, 53, 55–6, 59 Lisbon strategy 13, 54, 55, 66, 67, 68, 71, 145 lobbying activities 140 Maastricht Treaty 13, 53, 56, 58, 63, 64 manufacturing industry 54, 145 MAP project 388 market failures 61, 62, 67, 441 mergers and acquisitions 11–12, 50, 52, 71, 153, 156–60, 173–5 MNCs 33, 37, 39, 42, 55, 274 Moliter report 64 national champions 440 network industries 48, 49, 53, 55–6, 58, 62, 66, 71, 396 oil crisis 9, 11 outsourcing 55 patents 48, 72, 219, 222–3, 226, 232, 234–5, 243, 246, 256, 258–60, 311 pharmaceuticals 304, 306, 309, 311

privatization 10, 12, 49, 57, 62, 199, 201, 203, 206–9, 286, 441 protectionism 64, 69, 70 public capital and economic growth 57–8 public utility liberalization 64, 144, 180, 183–92, 208 quota system 51, 62, 69 R&D 13, 51–3, 64–9, 71–2, 90, 123–4, 256, 309, 422 rail transport 58–9, 70, 191 REACH 54, 67, 70 regional aid 50, 144, 145–7 rent seeking 140–41 Research Area 67 restructuring fund 47, 49 Risk Capital programme 68 Round Table of Industrialists 57 Sapir report 67 science parks 274, 275, 280, 282, 285, 286, 290, 291 service sector 55–6, 145 shipbuilding 12, 50, 69, 70, 145, 209 Single Market Programme 11, 12, 50, 71 SLIM programme 13, 64 SMEs 13, 23, 52, 56, 66, 68, 144, 175, 258–60, 380–85, 388–90, 394–6 software services 23, 32–9, 33, 41, 55, 70–71, 234–5, 243, 246, 252, 256, 258–60 start-up finance 68, 389 state aid to industry 11, 12, 49–51, 53, 62, 136, 143–7, 389–90 state ownership 11, 12, 47, 49, 198–201, 206–10 steel sector 9, 12, 49, 50, 69, 145 Structural and Cohesion Funds 145–7 structural policies 12–13, 14, 23, 66 subsidies 49, 69, 70, 71–2 tariffs 11, 51, 62 taxation 52, 136–7 technology alliances 115, 117–18 technology spin-offs 267, 270–71, 274 technology-based industries 70–72, 115, 117–18, 265, 282 telecommunications 10, 59, 70, 184, 186, 207 TENs (TransEuropean Networks) 58, 59, 63 textiles sector 9, 12, 70 tourism 55 trade policy 51, 55, 66, 69 training 144, 146 Treaty of Rome see Treaty of Rome university spin-offs 267, 268, 273, 274, 275, 276 venture capital 68, 389 water and sewerage 190–91 and WTO members 51

Index FDI 15 East Asia 81, 86, 88, 89, 90, 92, 93 European Union 9, 57, 209 Latin America 82 Finland bus transport 192 electricity 187–8, 189 privatization 208, 286 public utility liberalization 180, 184, 185, 186, 187–8, 189, 192, 208 railways 191 science parks 285, 286 SMEs 389 state aid to industry 145 technology spin-offs 267 telecommunications 185, 186 firms business support services see business support services colocation 323–4 cooperative R&D 123–4, 128–9 game theory see game theory and knowledge accumulation 17, 440 lead behaviour 369 management, traditional 103–4 mergers and acquisitions see mergers and acquisitions productivity levels 29–30, 102, 322, 442 size and transaction costs 438–9 technology alliances 115–19 technology-based 220–33 see also industrial policy; production Folk Theorem 407, 415 France anti-trust 11 biotechnology 258 competition 11, 208, 209 corporate governance 207 electricity 189 FDI 209 genomics 258 healthcare 301, 304 industrial policy 60, 62, 107, 209, 344 monopolies 10–11 patents 222, 256, 258–9, 260 pharmaceuticals 304, 305 political meddling 209 privatization 201, 207–9 public utility liberalization 185, 189, 190, 208 science parks 285 SMEs 258–9 software industry 252, 256, 258–9, 260 state aid to industry 145 state-ownership 207–9

455

telecommunications 185, 207 trade unions 207 university spin-offs 267, 273 water and sewerage 190 GALILEO satellite 59 Gambardella, Alfonso 28–44, 218, 222, 224, 230, 238, 396 game theory advertising investment 426–7 competition versus cooperation 422–4 differential 409–11, 415–31 equilibrium concepts 418–19 essential instruments 406–7 and industrial organization 407–9 mergers, horizontal 165, 424–6 optimization over time 419 price versus quantity competition 420–22 product differentiation 426–7 RAND Health Insurance Experiment 304, 405, 409–11 technological spillovers 422–4 GATT 15, 51, 62, 136 genome research 18, 258, 344 Germany anti-trust 9 business incubators 274 business support services 362–3 electricity 188, 189 employment 33 entrepreneurship 34 GDP 33 healthcare 301 industrial policy 60, 326 industrial revolution 6, 8–9 mergers and acquisitions 154, 156 oil crisis 9 patents 222, 234, 260 pharmaceuticals 305, 306, 309 privatization 208 public utility liberalization 185, 186, 188, 189, 208 R&D 90, 309 science parks 285 SMEs 382, 389 software sector 33, 260 state aid to industry 145 telecommunications 185, 186 university spin-offs 267, 268 Ghana 203–4 Giarratana, Marco 265–78 Giuri, Paola 217–42 globalization 15, 38, 54, 79–80, 381, 382 new economy 28–44 new industrial policies 14–16, 21–2, 23

456

International handbook on industrial policy

and production activities 100, 103, 204, 290 science parks 290, 292–3, 294 and technology alliances 117, 288–9, 290, 292–3 Goldstein, Andrea 198–214 Greece 57, 146, 208, 389 Grönblom, Sonja 180–97 Guatemala 390 health policy conflicts and convergences 310–13 definition of 299–301 developing countries 312 economic benefits 301, 302, 304 Grossman model 299 health as an industry 301–3 health insurance 304 and industrial policy 298–317 pharmaceutical industry see pharmaceuticals pharmaceuticals 304–6 quality of life 299, 302 R&D 302, 303 Honduras 390, 391 Hong Kong 85–8, 386 human capital 17, 21, 22, 34–6, 68, 81 and health policy 299 immigrants as (US) 35–6, 38, 39–40 mobility 23, 38, 39 and SMEs 381, 396 Hungary 145, 208 IBM 223, 244–5, 250 ICT 10, 15, 18, 23, 55, 59, 70–71, 162 IMF 202 import substitution policy 18, 82, 88, 89 incentive schemes 106, 108, 137, 138 India competition 38, 82, 382 employment 33, 37 entrepreneurship 34 GDP 33 human capital 34–6, 38 human capital to USA 35–6, 38 import substitution 82 industrial policy 66, 81, 326 IT graduates 34, 35, 40 MVA (manufacturing value added) 82 NASSCOM 34 outsourcing to 36–40, 395, 396 R&D 38, 271 science parks 285 SMEs 86, 386, 393, 395, 396 software sector 4, 23, 32–40, 395 specialization 30, 39

state ownership 200, 210 technology spin-offs 270, 271 wage levels 40 Indonesia 81, 386, 387, 392, 393, 394, 395 industrial clusters 17, 19, 21, 32, 120, 321–8, 330–31, 347–8, 351–3, 355–6, 442 and science parks 279–97, 307 SMEs 380, 381–3, 388, 395, 396, 397, 444, 445–7 technology-based firms 279–97 industrial policy 3, 4, 23, 80–82, 90 in 21st century 29–32 balanced development theory 18 business support services see business support services clusters see industrial clusters colocation socioeconomics 323–4 and competition policy 16, 119–20, 306, 435–40, 442 decentralization of 19, 22, 104, 321–41 and diversification 28, 31–2 East Asia 9, 16, 20–24, 66, 79–97, 107, 109, 326, 344, 441 entry barriers 120, 235, 437 European Union 11–14, 19–24, 45–78, 107, 144, 209, 344, 347, 438, 440, 443–4 game theory see game theory and globalization 14–16, 21–2, 23 governance consensus, new 324–6 and health policy see health policy horizontal 46, 47, 49–50, 53, 56, 58, 60, 62–9, 70, 107, 144, 204, 388 and informational service activities 98–111 integrative framework 442–3, 444 international practice 440–42 Latin America 326, 327–8, 334 local dimensions of see industrial policy, local national champions 344, 440 need for 103–6 neoclassical perspective 436–8, 439, 440, 442, 443 in new economy 120–22, 322–3 new forms of 106–10, 321–6, 442 old 3–27, 60–62, 69, 98–9, 119 and oligopolies 436–7, 438 path dependency 17 perspectives 435–49 pharmaceuticals 306–10 policy networks 325 resource-based perspectives 439–40, 444 science parks see science parks second-generation 363, 367–8 sectoral and specific 46, 47, 50–51, 53, 69–72, 107, 144

Index specialization see specialization and state aid see state aid to industry and state ownership see state ownership state role 19–24, 98 transaction costs 438–9 transition to 3–27 and welfare 435–7, 440 see also firms industrial policy, local competition, restricted 343–4 cooperative nexus 349, 351, 352, 353, 354–5 developing countries 355–6 dimensions of 342–61 efficiency bases 349–51, 352, 354 externalities 343, 349, 350, 351 failure risks 353–4 and government failures 345–7 guidelines 354–6 infrastructure 351, 353 investigation and action units 347–9 joint action 347, 351–2 localities of industry 349–51, 352, 354, 355 and market failures 342–5, 349–51, 352 markets, incomplete 344 meta-economic objectives 345, 352, 353 principal-agent asymmetry 347 production systems, local 110, 348 public base provision, failures and remedies 351–4 public goods 343, 349, 351, 352, 353 rent seeking 346, 353 specialization 350 and state intervention 344, 345–7, 352 systemic conditions 349–51, 354, 356 training programmes 350, 352 industrial revolution 4–7 infant industry protection theory 18 innovation 17, 21, 22, 23, 37, 90, 106–7, 257, 289, 323–4, 329–32 abusive practices 126–7 agency problems 266 business support services see business support services and competencies 31 and competition 101, 115–33, 439, 440 cooperative schemes 106, 348 entry barriers 125, 126, 127 European Union 13, 18, 19, 47, 49, 52, 53, 68, 265, 389 industries, organization of 119–29 markets 125–6, 438 mergers and acquisitions 127–9, 161–2 networks 120, 123 and new economy 120–22 and production 106–7, 108–9

457

quality 324 and science parks see science parks and SMEs 106, 232 and users 38 see also intellectual property rights; product differentiation; R&D intellectual property rights 5, 19, 20, 23, 121, 124 anti-commons 221, 228–30 GPL (General Public Licence) 20–21 institutional changes 218–20 institutional and public policy consequences 257–60 life science industry 256–7 and litigation costs 221, 230–33 and market dynamics 217–42 new approaches to 243–64 protection 245–6, 258 and R&D 236 and software industry 233–8, 244–7, 253–7, 272 strong, and technology-based firms 220–33 technology markets 221–5 technology-based firms 221–5 and technology-based firms 220–33, 272–3 transaction costs 228 TRIPS agreement 219–20 see also patents interventionism 45, 48, 53, 67, 69, 70, 81, 88, 89, 92, 199 Ireland deregulation 191 employment 33 entrepreneurship 34 GDP 33 human capital 34–6, 38 human capital to USA 35–6, 38 industrial policy 344 IT graduates 34, 35, 40 MNCs 33, 42, 274 privatization 203, 208 SMEs 389 software sector 23, 32–6, 39, 41 state aid to industry 146 structural policies 12–13 technology spin-offs 274 university spin-offs 267 Israel 23, 32–6, 33, 38, 39, 42, 270 Italy air transport sector 12 banking sector 9, 10 biotechnology 258 business incubators 275 business support services 362 chemical sector 9

458

International handbook on industrial policy

Clean Hands Drive 201 competition 10, 183 education 42 electricity 189 FDI 9 film industry 19 and globalization 382 healthcare 301 ICT 10 industrial clusters 331 industrial policy 19, 60, 325, 344, 347 industrialization 9–10 innovation 19 IRI 10, 12 Legge 297/199 275 monopoly regulation 10 patents 222, 258, 260, 311 pharmaceuticals 306, 311 privatization 10, 12, 199, 201, 203, 206–7, 208 public utility liberalization 183, 185, 186, 189 raw materials 10 science parks 275, 282 SMEs 382, 389, 396 social capital 282 software sector 260 state aid to industry 145 state ownership 198, 199, 200–201, 206–7, 208 steel sector 9, 12 structural policies 13 technology-based industries 282 telecommunications 185, 186 textile industry 9 university spin-offs 267, 268, 275 Japan competition 441 employment 33 entrepreneurship 34 GDP 33 growth rates 38–9 healthcare 300, 301, 304 ICT 71 India, software outsourcing 38 industrial policy 9, 16, 107, 109, 344, 441 innovation 123 MITI 9, 441 MNCs 37, 39 national champions 9 patents 219, 223 pharmaceuticals 304, 305 R&D 67, 90, 422 science parks 285

SMEs 381, 441 software sector 33, 38 technology alliances 115, 117–18 university spin-offs 268 Jullien, Nicolas 243–64 knowledge of production 23, 102–5, 106, 107–8 transfer, business support services 364, 373 knowledge-based economy 14, 17, 21, 22, 29–30, 70, 382 see also ICT; software Korea 4, 81, 86, 87 chaebol 89, 90, 387 competition 89, 130 economic crises 387 employment 386 industrial policy 344 interventionism 81, 89, 92 manufacturing 85–6, 87 privatization 89, 208 R&D 4, 81, 86, 89, 90 reverse engineering 89 SMEs 86, 386–7, 392–3, 395 software industry 34 state ownership 200, 208 university spin-offs 268 Labory, Sandrine 3–27, 380–402 Lall, Sanjaya 15, 79–97 Lambertini, Luca 405–14, 415–31 Latin America automotive industry 82 comparative advantage 82 competition 82, 390 economic crises 83, 328, 330, 334, 385 exports 82, 84 FDI 82 import substitution policy 18, 82 industrial policies 23, 80–81, 82, 326, 327–36 liberalization 81, 83 manufacturing 80–84, 327, 328, 332–3, 334 privatization 201, 204, 207, 208, 209, 328, 332 R&D 82, 200 SMEs 23, 328–31, 334, 385–7, 390–92, 395–7 software industry 23, 32–6, 41, 391 software sector 23, 32–6, 391 state ownership 198–9, 200, 201, 207, 208, 209 taxation 390 technological skills 82 training 390 wage levels 82 see also individual countries

Index liberalization 15, 79, 81, 83, 85 European Union 48, 49, 53, 55–6, 59 neoliberalism 16–17, 79, 86, 326 public utility sectors see public utility sectors Lorentzen, Jochen 115–33 Luxembourg 208 Maastricht Treaty 13, 53, 56, 58, 63, 64 Malaysia 81, 201, 386, 387 manufacturing East Asia 80–87 European Union 54, 145 Latin America 80–84, 327, 328, 332–3, 334 SMEs 381 market abuse, and competition 106 market failures business support services 365, 366, 375 European Union 61, 62, 67, 441 and industrial policy, local 342–5, 349–51, 352 and monopoly 436 pharmaceuticals 306–9 SMEs 381 marketing 15, 368–71 Martin, Stephen 120, 123, 124, 130, 134–52, 180, 183, 408, 422 mergers and acquisitions 11–12, 50, 52, 71, 104, 115, 119 agency theory 164, 165 banking 155–6 cascades 167, 169 competition policy 122, 127–9, 153, 173–5 and concentration policy 153–79 control 170–75 efficiency gains 154–5, 174 game theory see game theory horizontal 424–6 and hubris 164–5 and innovation 127–9, 161–2 market power 155, 156, 173, 174 merger experience 160–61 minimax regret 168–70 owners and managers 163–4 performance of 154–63 pharmaceutical industry 310 rational herding 166–7 real value 154–7 regret framework 167–8 reputation 165–6 shareholder value 154, 157–61, 163, 172 SMEs 162, 174, 175 strategic theory 163–70 supermegamergers 156 uneconomic 164, 165, 169–70, 171–2, 173–4 welfare losses 171–2, 173

459

Mexico economic crises 385 export market 82, 84 industrial policy 326 maquiladoras 391 MVA (manufacturing value added) 82, 83, 84 NAFTA 82, 84 privatization 201, 208 SMEs 385, 386, 390, 391 state ownership 200, 208 Microsoft 19, 126–7, 245 Middle East 80–81, 84, 210 MNCs developing countries 86, 88–9, 327, 355, 356 European Union 37, 39, 55 and science parks 274–5 United States 257 Møllgaard, Peter 115–33 Monmax Theorem 406 monopoly 7–11, 139, 224, 436 abuse 17 and competition 122, 126–7, 128 and economies of scale 343–4, 352 and efficiency 343, 437–8 and entry barriers 437 hypothetical monopolist 125 public utility sectors 180–81, 182–3, 184, 188, 189, 190, 191, 198, 204 UK 8, 10 welfare losses 437–8 Myanmar 209 Nash equilibrium 406–7, 409, 418, 419, 423 Nepal 392 Netherlands biotechnology 258 mergers and acquisitions 157, 171 patents 222, 256, 258 pharmaceuticals 306 privatization 208 SMEs 389 university spin-offs 268, 276 network industries 16, 19, 20, 21, 106, 280 European Union 48, 49, 53, 55–6, 58, 62, 66, 71, 396 networking schemes 104, 105, 115–16 and competition 121–2, 126, 127 innovation 120, 123 production 104, 105, 110 new economy evaluation of 4–7 and globalization 28–44 industrial organization in 120–22 New Zealand 185, 201, 208, 305

460

International handbook on industrial policy

Nicaragua 390, 391 Norway 187–8, 189, 208, 258, 267, 268, 306 OECD 90, 98, 119, 185, 186, 203, 207, 210, 211, 285, 291 see also individual countries oil crisis 9, 11, 17 open science model 217, 218, 233–8, 243 open source model 20 outsourcing 36–9, 41, 104, 270–71, 395, 396 Paci, Daniele 298–317, 380–402 Pakistan 30, 326 Parrilli, Mario Davide 380–402 patents 20, 121, 161, 217, 218, 219, 221, 222, 260 blocking 221, 223 and copyright 234, 245, 246–7 cumulative inventions 233 European Union 48, 72, 219, 222, 223, 226, 232, 234–5, 243, 246, 256, 258–9, 260, 311 ‘hold-up’ problems 228–9 infringements 229, 247 licensing 222–4, 225–7, 228–30, 236, 243–4, 247–8 litigation costs 230–33 and natural resources 257–8 paradox 236 pharmaceutical industry 308, 309–10, 311 protection 222, 223–4, 233, 238, 244, 276 and R&D 308 sleeping 221, 223 software industry 219, 220, 233–8, 243, 244, 257, 276 TRIPS agreement 219–20 United States 72, 143, 219–20, 223, 225–6, 230–32, 234–5, 244–6, 257–8, 275–6, 309–10 and universities 221, 224, 225–7 unused 222–3 see also intellectual property rights Pelkmans, Jacques 13, 45–78, 184 Peru 204, 386, 390, 391, 395, 396 pharmaceuticals 72, 162, 222, 301–2, 303–13, 344 brand names 311 competition promotion 309–11 European Union 304, 305, 306, 309, 311 health policy towards 304–6 industrial policy issues 306–10 market failures 306–9 patents 308, 309–10, 311 pricing strategy 311–13 R&D 306–9, 311

Philippines 81, 386, 387, 392 Pitelis, Christos N. 435–49 Poland 145, 208 Portugal 12–13, 57, 146, 208, 387, 389 privatization 16, 210 and business support services 366–7, 374–5 European Union 10, 12, 199, 201, 203, 207–9, 286, 441 Latin America 201, 204, 207, 208, 209, 328, 332 outcome of 203–4 public utility sectors 204–5, 206 unpopularity of 203 see also state ownership product differentiation 15, 17, 21 investment in 426–7, 437 see also innovation production ‘bottom-up’ approach 109–10, 446 competence poles 110 competency development 108 and competition 19, 100, 103, 120 complexity 109 control revolution 101 delocalization 98, 100–101 diversity of 101–2, 134 and economic intelligence 104, 108 and globalization 100, 103, 204, 290 and incentive schemes 106, 108–9 and innovation 106–7, 108–9 of knowledge 23, 102–5, 106, 107–8 local 110, 348 networking schemes 104, 105, 110 new conditions and modes 99–103 organizational set-up 5, 19, 105–6, 109 resource allocation 107–8, 109 returns 102 state intervention 109 systems analysis 104–5 training programmes 104, 107, 108 and wage levels 100–101 see also firms productivity levels 29–30, 102, 203, 322, 383, 442 protectionist policies 22–3, 64, 69, 70, 81, 89, 91 Prussia 6 public utility sectors bus transport 191 competition 180–82, 189, 190, 191, 205 corporate governance 205–7 efficiency, distribution and quality, endogenous 183–4, 189, 190 labour costs 183–4 liberalization 64, 144, 180–97, 204–6, 208

Index monopolies 180–81, 182–3, 184, 188, 189, 190, 191, 198, 204 railways 191–2 and state ownership 198 sunk costs 180–81, 186, 190–91 vertical relations and network externalities 181–3 water and sewerage 190–91 see also electricity; telecommunication water and sewerage R&D 15, 17, 18, 31, 89, 104, 106, 144, 222, 324 cooperative agreements 123–4, 128–9 East Asia 4, 38, 67, 71, 72, 81, 86, 88–91, 90, 271, 422 European Union 13, 51–3, 64–9, 71, 72, 90, 123, 124, 256, 309, 422 game theory see game theory health policy 302, 303 and intellectual property rights 236 Latin America 82, 200 and patents 308 pharmaceuticals 306–9, 311 and science parks see science parks with spillover 422–4 technology alliances 115–19, 254–6 technology-based firms see technologybased firms United States 67, 72, 90, 123–6, 220, 256, 267, 269, 307, 422 see also innovation regional development 19, 21, 22 rent seeking 138–41, 346, 353 Romania 57 Russia 301 Schenk, Hans 153–79 Schweitzer, Stuart O. 298–317 science parks communicative interaction 282–3 and competition 288–90 East Asia 88, 285 economic impact 22, 283–4 efficiency of 281–2 European Union 274, 275, 280, 282, 285, 286, 290, 291 first (Stanford Research Park) 284, 286, 287 first-generation 286–7, 290 and globalization 290, 292–3, 294 and high-tech clustering 279–97, 307 institutional thickness 282–3 as intermediaries 290 and MNCs 274–5 and networks 290–91 policy decisions 291–2

461

purpose of 280 resource-rich urban fabric, part of 285–6 second-generation 287 sister parks 292 third-generation 287–8 United States 274, 275, 282, 284–5, 286–7, 289, 290, 291 see also networks semiconductors 90, 228, 229, 236 Sepulveda, Leandro 321–41 Singapore 34, 81, 85–6, 87, 88–9, 92, 386 SMEs 18, 326, 348, 349, 380–402 business support services see business support services clusters 380, 381–3, 388, 395, 396, 397, 444, 445–7 collective learning 382 and competition 381, 382, 396, 397–8 development of 396–8 East Asia 85, 86, 90, 381, 385–7, 392–6, 441 and economic crises 387 and employment 380, 383–5 and entrepreneurship 385, 392 entry barriers 120, 235 European Union 13, 23, 52, 56, 66, 68, 144, 175, 258–60, 380–85, 388–90, 394–6 flexibility 387 and globalization 381, 382 and human capital 381, 396 and innovation 106, 232 international support for 383 and knowledge-based economy 382 Latin America 23, 328–31, 334, 385–7, 390–92, 395–7 manufacturing sector 381 market failure 381 mergers and acquisitions 162, 174, 175 micro firms 383–5, 391, 395, 396, 397 networking 396 and patent infringement 247 performance 383–7 policy analysis 388–94 productivity 383 specialization 382 and subcontracting 381, 391 technology-based 220–33, 236 transnational networks 382 software industry 28–9, 42, 344 business development 252–3 collective production 254–6 competition 237 copyright 243, 245–7, 253–4 Creative Commons Licensing 256 East Asia 4, 23, 32–41, 256, 395

462

International handbook on industrial policy

European Union 23, 32–9, 33, 41, 55, 70–71, 234–5, 243, 246, 252, 256, 258–60 free software 247–51, 254 and intellectual property rights 233–8, 244–7, 253–7, 272 Latin America 23, 32–6, 41, 391 licensing 247–8, 253, 256, 258–9, 272 new methods of production 249 open science model 218, 233–8, 243 Open Source Initiative 236–8 open-source programs 251–5, 258, 259–60 operating systems 250–57 outsourcing 36–40, 395, 396 patents 219, 220, 233–8, 243, 244, 257, 276 software quality 251–2 United States 23, 33, 35–6, 220, 234–5, 243, 244–5, 246, 256, 275–6, 289–91 South Africa 80, 326 South Asia see East Asia Spain education 42 FDI 57 infrastructure 57 patents 222, 260 privatization 208, 286 public utility liberalization 183, 185, 208 science parks 286 SMEs 389 software sector 260 state aid to industry 145, 146 structural policies 12–13 telecommunications 185 university spin-offs 268 specialization 28, 29, 30, 39, 88, 322, 323, 350, 382 and entrepreneurship 30–31 industrial policy, local 350 SEFs 222 spin-offs 31–2 Sri Lanka 386 Stackelberg equilibrium 419 Stallman, Richard 236, 247–9 state aid to industry 88, 134–52 European Union 11, 12, 49–51, 53, 62, 136, 143–7, 389–90 market failure 443 rent seeking 138–41 strategic trade policy 134–5 tax competition 135–8 United States 141–3 state intervention 3, 4, 5, 7, 89, 109, 201, 332–3, 335 and competition 120 health policy see health policy and industrial development 20–24, 98

and industrial policy, local 344, 345–7, 352 subsidies 16, 19 state ownership 198–214 developing countries 199, 200, 202, 206 East Asia 200, 208–10 in economic development 198–201 European Union 11, 12, 47, 49, 198–201, 206–10 Latin America 198–9, 200, 201, 207, 208, 209 and national independence 198 privatization dynamics 201–2, 209 privatization outcome 203–4, 206 public utility sector see public utility sector regulatory capitalism requirements 204–7, 211 steel industry 9, 12, 49, 50, 69, 139, 145, 237 structuralist theory 18 subcontracting, and SMEs 381, 391 subsidies business support services 364 East Asia 81, 89, 90 European Union 49, 69, 70, 71–2 state intervention 16, 19 strategic trade 134–5 United States 142 Sweden bus transport 192 electricity 187–8, 189 patents 256, 260 pharmaceuticals 306 privatization 208 public utility liberalization 187–8, 189, 191, 192, 208 R&D 90 railways 191 science parks 285 SMEs 389 software industry 256, 260 university spin-offs 267, 276 Switzerland electricity 189 public utility liberalization 189 SMEs 383 university spin-offs 268 Taiwan 4, 22, 34, 81, 86, 87, 89, 90–92, 256, 387 tariffs 11, 51, 62, 89, 90, 91 taxation 52, 135–8, 141–2, 143, 343, 390 technology alliances 288–9, 290, 292–3, 422–4 R&D 115–19, 254–6 technology spin-offs 267, 270, 271, 274 technology-based firms

Index business incubators 274, 280 corporate spin-offs 266–7, 270, 272–4, 275, 275–6, 280 entrepreneurship 266–7, 268–9, 270, 271, 274, 275 European Union 70–72, 115, 117–18, 282 high-tech clustering 279–97 human resource strategies 271–2 and intellectual property 220–33, 272–3 involuntary spin-offs 270–71, 274 licensing contracts 272 new entrants 265, 266, 269–70, 271–2, 273, 281, 302 performance 273–4 and public policy 274–6, 302–3 science parks see science parks SMEs 220–33, 236 spin-offs 265–78, 280, 281, 291 United States 265, 269, 273, 282 university spin-offs 267–8, 269, 270, 271, 272–3, 274, 275, 280, 291 telecommunications 8, 10, 16, 59, 70, 184–7, 202, 204, 207, 210, 281 Thailand 81, 386, 387, 392, 393, 394, 395 Torrisi, Salvatore 265–78 trade unions 207 training programmes 104, 107, 108, 350, 352, 390 Treaty of Rome 50 Article 40 146 Articles 85 and 86 (now 81 and 82) 11, 438, 440 Articles 90 and 92 12 Euratom Treaty 13 see also European Union Turkey 81, 200, 208 UK anti-trust 10 Broadcasting Act 10 bus transport 191 and Commonwealth 10 competition 10, 183, 187, 191 electricity 187, 188, 189 genomics 258 healthcare 300, 301 in Hong Kong 88 industrial revolution 4–9 information sector 10 mergers and acquisitions 154, 160, 171 monopoly regulation 8, 10 OFGEM 187 OFWAT 190 patents 222, 234, 256 pharmaceuticals 304, 305, 309

463

privatization 10, 201, 203, 207, 208, 441 public capital 58 public utility liberalization 180, 183, 185, 187, 188, 189, 190–91, 208 R&D 256, 309 railways 191 science parks 275, 280, 285 SMEs 389 software sector 259 state aid to industry 145 technology alliances 115 telecommunications sector 10, 185 university spin-offs 274 water and sewerage 190–91 UN 22, 57 United States air transport sector 8 anti-trust policies 7, 8, 19, 125, 126, 141–2, 143, 438, 440 AT&T case 8 Bayh-Dole Act 115, 220, 225–6, 227, 307 biotechnology 258, 291, 307 business incubators 274 business support services 362, 363, 366 CAFC 219 car industry 273 chemical industry 37 Clayton Anti-Trust Act 8, 126 commerce, inter-state 142–3 comparative advantage 38 competition 8, 67, 119–20, 126–7, 174, 257, 310 Creative Commons Licensing 256 Cuno case defence 72 Diamond v. Diehr 220, 234 education 39 electricity 188, 189 employment 33, 37 Enron 188, 209 entrepreneurship 7, 8, 18, 34, 265, 267, 275 GDP 33 genomics 258 Gottschalk v. Benson 234 Hatch-Waxman Act 310 healthcare 300, 301, 304 human capital, immigrants as 35–6, 38, 39–40 ICT 18, 70, 71, 162 In re Appalat case 234 incentive policies 138 India, software outsourcing 36–40, 395, 396 industrialization 7 infrastructure 56 innovation 125–6, 162, 257, 289

464

International handbook on industrial policy

IT graduates 35 Lobbying Disclosure Act 140 mergers and acquisitions 153–6, 158, 163, 171–5 MNCs 257 monopoly regulation 7, 8, 139 National Cooperative Research Act 115, 123, 125, 442 Netpreneur program 275 ‘New Economy’ 18 Orphan Drug Act 308 outsourcing 36–40, 41, 395, 396 Parker v. Brown 141 patents 72, 143, 219–20, 223, 225–6, 230–32, 234–5, 244–6, 257–8, 275–6, 309–10 pharmaceuticals 72, 162, 304–7, 309–10, 311, 312 privatization 208 production returns 102 Prorate Act, California 141 public utility liberalization 180, 184–6, 188–91, 208 quality certification 312 R&D 67, 90, 123–6, 220, 256, 267, 269, 307, 422 RAND Health Insurance Experiment 304, 405, 409–11 rent seeking 139–40 Robinson-Patman Act 119 science parks 274, 275, 282, 284–5, 286–7, 289, 290, 291 SEFs 37 Sherman Anti-Trust Act 7, 8, 141–2, 143 SMEs 162, 267, 382 social capital 282 software sector 23, 33, 35–6, 220, 234–5, 243–6, 256, 275–6, 289–91 specialization 31–2 Stanford Research Park 284, 286, 287 state aid to industry 141–3

State Street Bank Trust v. Signature Financial Group 220, 234 steel industry 139 subsidies 142 taxation 136, 138, 141–2, 143 technology alliances 115, 117–18 technology spin-offs 271, 273, 275–6 technology start-ups 269 technology-based industries 265, 269, 273, 282 telecommunications sector 8, 184, 185, 186 university spin-offs 267, 268, 273 wage levels 40 water and sewerage 190, 191 Wikipedia 256 university spin-offs 21, 267–76, 280, 291 and patents 221, 224, 225–7 Uruguay 385, 386 Valbonesi, Paola 134–52 Venezuela 385, 386 Vietnam 210, 392 wage levels 40, 82, 88, 100–101 Washington Consensus 16, 79, 81, 82, 84, 444 water and sewerage 190–91 welfare effects and competition 137, 435–7, 440 industrial policy and 435–7, 440 mergers and acquisitions 171–2, 173 rent seeking 139 strategic trade policy 135 Willner, Johan 180–97 World Bank 187 WTO 15, 51, 79, 82, 210 Subsidies and Countervailing Measures Agreement 135 Zermelo’s Theorem 406 Zimmermann, Jean-Benoît 243–64