Resources, Technology and Strategy (Routledge Advances in Management and Business Studies)

Resources, Technology and Strategy

Over recent years, the Resource-Based Perspective (RBP) has become a dominant force in strategic thinking. Under the influence of RBP, managers have become more aware of their ability to control their destinies through the accumulation, and efficient use, of resources including production capabilities, marketing prowess, finance and organisational culture. Technology is undoubtedly one of the most important factors in the strategic management of resources. As the rate of change has accelerated in recent decades, firms have been challenged to use technological capabilities as weapons in their battles for market share. This book brings together experts from Europe, North America and Asia to consider the strategic relationship between technology and other resources. Much of Resources, Technology and Strategy is devoted to a theoretical examination of RBP, assessing its strengths and weaknesses. Case studies reveal the importance both of having and not having strong technological capabilities in settings as diverse as the US semiconductor industry, small family manufacturing firms in Hong Kong and state owned enterprises in China. Overall, this book demonstrates the usefulness of RBP for the analysis of strategic problems. It shows the value of economic theory in providing a rigorous framework for considering the consequences of technological change in a dynamic framework. Aimed at graduate specialists in the fields of strategic management and technology management, this book will also be a valuable resource for practitioners in the field. Nicolai J.Foss is Professor of Economic Organisation at the Copenhagen Business School, Denmark. He has edited a number of books and readers on management and economic organisation with Routledge such as Towards a Competence Theory of the Firm, and Economic Organisation, Capabilities and Coordination. Paul L.Robertson is Professor of Management at the University of Wollongong in Australia. He is the co-author of Firms, Markets and Economic Change: A Dynamic Theory of Business Institutions, and Authority and Control in Modern Industry, both published by Routledge.

Routledge Advances in Management and Business Studies 1 Middle Managers in Europe Yves Frédéric Livian and John G.Burgoyne 2 Marketing Apocalypse Eschatology, escapology and the illusion of the end Edited by Stephen Brown, Jim Bell and David Carson 3 Relationship Marketing in Professional Services A study of agency-client dynamics in the advertising sector Aino Halinen 4 Job Design and Technology Taylorism vs Anti-Taylorism Hans D.Prujt 5 Regulation and Organisations: International Perspectives Edited by Glenn Morgan and Lars Engwall 6 Information Technology, Organisations and People Transformations in the UK retail financial services sector Jeff Watkins 7 HRM, Technical Workers and the Multinational Corporation Patrick McGovern 8 The Internationalization of Small to Medium Enterprises The Interstratos Project Edited by Rik Donckels, Antti Haahti and Graham Hall 9 Neo-Industrial Organising Renewal by action and knowledge formation in a project- intensive economy Rolf A.Lundin, Hans Wirdenius, Eskil Ekstedt and Anders Soderholm 10 Perspectives on Public Relations Research Edited by Danny Moss, Dejan Vercic and Gary Warnaby 11 Resources, Technology and Strategy Explorations in the resource-based perspective Edited by Nicolai J.Foss and Paul L.Robertson

Resources, Technology and Strategy Explorations in the resource-based perspective

Edited by Nicolai J.Foss and Paul L.Robertson

London and New York

First published 2000 by Routledge 11 New Fetter Lane, London EC4P 4EE Simultaneously published in the USA and Canada by Routledge 29 West 35th Street, New York, NY 10001 Routledge is an imprint of the Taylor & Francis Group This edition published in the Taylor & Francis e-Library, 2005. “To purchase your own copy of this or any of Taylor & Francis or Routledge’s collection of thousands of eBooks please go to www.eBookstore.tandf.co.uk.” © 2000 Nicolai J.Foss and Paul L.Robertson for selection and editorial matter All rights reserved. No part of this book may be reprinted or reproduced or utilised in any form or by any electronic, mechanical, or other means, now known or hereafter invented, including photocopying and recording, or in any information storage or retrieval system, without permission in writing from the publishers. British Library Cataloguing in Publication Data A catalogue record for this book is available from the British Library Library of Congress Cataloging in Publication Data Resources, technology, and strategy/[edited by] Nicolai J.Foss and Paul L.Robertson (Routledge advances in management and business studies) Includes bibliographical references and index. 1. Industrial management. 2. Strategic planning 3. Technological innovations—Management. 4. Human capital I. Foss, Nicolai J., 1964–. II. Robertson, Paul L. III. Series HD31.R464 2000 658–dc21 99–32605 CIP ISBN 0-203-98225-8 Master e-book ISBN

ISBN 0-415-21585-4 (Print Edition)

Contents

List of tables and figures

vii

The contributors

ix

1

Introduction: resources, technology and strategy NICOLAI J.FOSSPAUL L.ROBERTSON

1

2

Equilibrium vs. evolution in the resource-based perspective: the conflicting legacies of Demsetz and Penrose NICOLAI J.FOSS

11

3

Knowledge and capabilities: a new view of the firm J.STANLEY METCALFEANDREW JAMES

31

4

Synthesising resource-based, evolutionary and neoclassical thought: resource-advantage theory as a general theory of competition SHELBY D.HUNT

53

5

A processual account of innovative capabilities PETER W.ROBERTS

79

6

Product and process architectures in the management of knowledge resources RON SANCHEZ

99

7

Building innovative assets and dynamic coherence in multi-technology companies JENS FRØSLEV CHRISTENSEN

121

8

Dynamic complementarities and technology acquisition TOMI LAAMANENERKKO AUTIO

149

9

Economic organisation and the accumulation of rent-yielding assets KIRSTEN FOSS

171

Capabilities and vertical disintegration in process technology: the case of semiconductor fabrication equipment RICHARD N.LANGLOIS

193

10

vi

11

Technological capabilities and the strategies of small manufacturing firms: the case of Hong Kong TONY F.YUPAUL L.ROBERTSON

219

12

Technological capabilities and foreign direct investment in China YIZHENG SHI

243

Index

261

Tables and figures

Tables 4.1 Foundational propositions of Perfect Competition and Resource-Advantage theory 6.1 Comparison of product definition, design and development in creating conventional versus modular product architecture 6.2 Know-how, know-why and know-what forms of knowledge 7.1 Combinations of categories of innovative assets and their contributions to competitive advantage 7.2 The relationship between categories of innovative asset dynamics, and the dominant features of innovative assets 8.1 Life cycle stages of the acquired companies cross-tabulated with industry sectors 8.2 Regression equations of three regression analyses: 1 with all variables that correlated with the compound measure of success; 2 with same variables as in 1 but with a backward elimination procedure with a significance level of 0.10; 3 with same variables as in 1 but with a forward elimination procedure with a significance level of 0.05 10.1 Top ten semiconductor equipment suppliers, 1979 and 1989 10.2 Top ten semiconductor equipment suppliers, 1996 12.1 Foreign direct investment in China (US$ million) 12.2 Foreign direct investment in China by country of origin 12.3 Hong Kong imports from China (US$ million) 12.4 Technological assets and FDI strategies of firms

60 105 114 126 133 158 161

196 197 249 250 251 257

Figures 4.1 4.2 5.1 5.2 5.3 5.4

A schematic of the Resource-Advantage theory of competition Competitive position matrix Hierarchy of firm capabilities Series of positional advantages Non-deterministic representations of innovation Innovative capabilities: a reconciliation

58 59 80 83 90 91

viii

6.1 Modular product architecture as coordinating mechanism for modular process architectures 6.2 Loosely coupled learning processes for developing know-how, know-why and know-what knowledge 8.1 A process model on the evolution of technology-based company networks and dynamic complementarities with the assumption that a fundamental discontinuity starts the evolution 8.2 Distribution of the studied acquisitions according to the industry; N=111 8.3 Size distribution of the companies in the sample 8.4 Results of the regression analysis with the component and compound measures of success 8.5 An illustration of the influence of complementarities on the sales revenue of the acquired company 10.1 A parallel-processing configuration 10.2 A hypothetical modular integrated-processing system (after Burggraaf 1989)

111 114 154

157 157 161 166 200 201

The contributors

Erkko Autio is Professor of Industrial Management, Department of Strategy and International Business, Helsinki University of Technology. Jens Frøslev Christensen is Professor of Technology Strategy, Department of Industrial Economics and Strategy, Copenhagen Business School. Kirsten Foss is Associate Professor (technology strategy, economic organisation), Department of Industrial Economics and Strategy, Copenhagen Business School. Nicolai J.Foss is Professor (economic organisation), Department of Industrial Economics and Strategy, Copenhagen Business School, and director of the RESPECT research group. Shelby D.Hunt is the J.B.Hoskins and P.W.Horn Professor of Marketing, Texas Tech University, College of Business Administration, Area of Marketing. Andrew James is Research Fellow, PREST (Policy Research in Engineering, Science and Technology), University of Manchester. Tomi Laamanen is Professor of Industrial Management, Department of Strategy and International Business, Helsinki University of Technology. Richard N.Langlois is Professor of Economics and Management, Department of Economics and Department of Management, University of Connecticut. J.Stanley Metcalfe is Professor and Director of CRIC (Centre for Research on Innovation and Competition), University of Manchester. Peter W.Roberts is Assistant Professor, GSIA, Carnegie-Mellon University. Paul L.Robertson is Professor, Department of Management, University of Wollongong. Ron Sanchez is Professor, Institute for Management Development, Switzerland. Yizheng Shi is Assistant Professor, Department of Marketing, Hong Kong Baptist University. Tony F.Yu is Research Fellow, School of Economics and Management, University College, University of New South Wales.

x

1 Introduction Resources, technology and strategy Nicolai J.Foss and Paul L.Robertson

Since the end of the 1980s there has been a notable change in the way that many academics in strategy research conceptualise the strategy process and how they think about strategic content. The strategy process is increasingly seen as beginning in a modest and ‘introverted’ way, by analysing the firm’s portfolio of resources, rather than with such broad questions as, ‘What is our corporate mission?’ or ‘What businesses are we in?’ The content side of strategising is also increasingly cast in terms of resources. Thus, ‘Our strategy is to get maximum market share in markets x, y and z’ has given way to variants along the lines of ‘Our strategies in markets x, y and z aim at more fully sharing resources a, b and c’, or ‘Our strategy is to stretch existing resources and create new ones so that we are not trapped by blurring industry boundaries.’ There is clear, if unsystematic, evidence that this change is also taking place in managerial practice. Much of this reorientation is due to the breakthrough in academic as well as practical strategy thinking of what is often referred to as ‘the resource-based perspective’ (henceforth, ‘the RBP’), which reaches back to the classic work of Edith Penrose (1959) and Philip Selznick (1957), but has only emerged as a strong contender on the strategic management scene in the mid-1980s with the work of Birger Wernerfelt (1984), Richard Rumelt (1984), Jay Barney (1986) and others.1 The key ideas in the RBP are that successful firms possess heterogeneous collections of resources, that these varied collections of resources allow firms to implement different strategies, that different strategies yield different returns (which may be interpreted as rents accruing to the underlying resources), and that successful strategies and their associated return streams are sustainable to the extent that they are prohibitively costly to imitate. This is the perspective that all of the contributors to the present book begin from, before moving on to extend or criticise it in various ways. In little more than a decade, the RBP has emerged as arguably the dominant contemporary approach to strategy (content) research—as perhaps the new orthodoxy in strategy research. The perspective’s appeal to academics would seem to be a matter of combining relative analytical rigour with apparent managerial relevance. Thus, there are good reasons for the success of the RBP; however, there are certainly also reasons to hold one’s breath and curb enthusiasm, primarily because there are many unresolved problems and issues in need of clarification.

2 NICOLAI J.FOSS AND PAUL L.ROBERTSON

Most conspicuously, perhaps, there is a considerable amount of terminological ambiguity, with various resource-based theorists using concepts such as ‘resources’, ‘competencies’, ‘capabilities’, etc. to refer to what are seen as strategic assets. This problem may be overcome as certain terminological standards gradually become dominant in the community of resource-based theorists; it is essentially a minor problem. However, there are also deeper issues in need of clarification and there are potentially conflicting insights. A conspectus of some existing problems with the RBP would include: • The isolated resource problem. There is a tendency in the RBP to analyse some resources in isolation from others, so that systems effects and complementarities may be lost from the view of the analyst. Moreover, although no single resource may be central to a firm’s strategy, the interplay between several resources may very well yield rents (Porter 1996; Robertson 1996). • The environment problem. The RBP is overly ‘introspective’ (Porter 1994) and has a tendency to neglect the environment or only incorporate it implicitly under the rubric of such broad competitive forces as ‘the threat of imitation’. Whereas the structureconduct-performance (SCP) approach that dominated industrial economics for several decades assumed that the resources available to firms were homogeneous and that all important influences on performance could be traced to external factors such as market structure, the resource-based perspective tends to downplay the importance of external variables. In common with the SCP, however, the RBP takes demand as given rather than as a dynamic factor that firms can manipulate strategically. • The resource application problem. Given that resources are central in the RBP, the actual application of resources in production has received scant analytical attention. • The resource organisation problem. The RBP tells us very little about how resources are best organised (Williamson 1994). For example, if the resources in question are human resources, the services those resources yield are dependent on a host of determinants such as incentives, monitoring and culture that have been investigated in organisational economics and organisational behaviour studies. To date, the RBP has shown little interest in these matters. • The resource creation problem. The RBP has concentrated overwhelmingly on the analysis of existing resources, and has given remarkably little attention to the creation of new resources. As a result, there is a distinctly retrospective character to the RBP, which may threaten its managerial relevance (Foss et al. 1995). Some authors, such as Kay (1993) who criticises ‘wish-driven strategies’, contend that firms cannot easily adapt or augment their existing resources to embrace new strategic opportunities and are thus trapped by their histories. Pralahad and Hamel (1990), on the other hand, are more optimistic about the ability of firms to gather new resources in response to changes in their ‘strategic intent’, but are vague on how this can be accomplished. There are other problems with the perspective, such as the lack of solid empirical work, but the problems above are major ones, and arguably the major ones. All the contributions to this book explicitly or implicitly grapple with these problems and suggest various

INTRODUCTION 3

remedies. For example, a number of contributors point out and discuss the fact that there is a lack of a clear and coherent treatment of dynamics in the RBP. Thus, the RBP does not in its present version(s) theorise in any convincing way about the mechanisms underlying the creation of new resources—what we have called ‘the resource creation problem’. This problem is perhaps particularly troublesome for the future evolution of the RBP, since dynamics (broadly conceived) is all the rage in the strategy and organisational behaviour fields these days, as witness the recent enthusiasm about ‘hypercompetition’, ‘organisational learning’, ‘the knowledge-creating company’, etc. It is true that some RBP theorists have tried to grapple with these issues, typically under the banner of ‘the core competence approach’, but this has only been done—it is fair to say—by substantially sacrificing rigour. On the other hand, many of the more economics-orientated contributions to the RBP exhibit a certain degree of rigour, but do not treat dynamics in any detail. The divide may be seen as largely a matter of whether one seeks to address and include dynamic—or better, evolutionary—factors, or instead relies on standard economic theory. It is a choice, in short, between equilibrium and evolution, with ‘evolution’ and ‘evolutionary’ referring here to whether such concepts as irreversibility (e.g., in the form of path-dependence and learning) and novelty (e.g., in the form of unanticipated innovations) are included in the analysis at some level. In practice, dynamism and evolution are at the heart of strategic behaviour. Few managers are content with equilibrium; instead, they deliberately attempt to upset existing market positions. The goal of strategy research should therefore be to find an approach that treats evolutionary developments with something approaching the rigour that neoclassical economists bring to equilibrium situations. Although this is a tall order, it is the only way to generate a strategy literature that is both usable and analytically respectable. As the title implies, our primary focus in this volume is on technological change. In particular, we are concerned with the ways in which a firm’s resources are related to the product and process technologies that it adopts, and with the changes in other resources that may be needed to deal with changes in the firm’s internal and external technological environments. In contrast to Solow (1956) and most other neoclassical economists, we treat technological change as being endogenous to firms, but unlike the ‘New Growth Theorists’ (e.g., Romer 1986, 1990, 1993, 1994), we are searching for a fine-grained analysis of the motivation that underpins strategic behaviour at the firm level. The implied argument in most of the contributions is that, in order to account for the emergence and maintenance of the systematic heterogeneity among firms that is a basic premise of the RBP, and to derive managerial lessons with respect to issues such as resource building and corporate renewal, it is necessary to develop insights into the endogenous creation of resources. In the present state of analysis, heterogeneity is simply asserted, and dynamic and normative issues relating to endogenous heterogeneity are largely neglected. What the RBP needs, we suggest, is more agreement that these dynamic issues are crucial but should be approached in a more precise and analytical way than at present. If this does not happen, there is a real danger that the RBP may split even more deeply, first, into a formal, stark, abstract branch strongly inspired by economics and gradually losing contact with

4 NICOLAI J.FOSS AND PAUL L.ROBERTSON

managerial reality, and, second, into an increasingly loose and free-wheeling branch where almost anything goes on the analytical level. Technological resources and capabilities2 can be viewed from several angles, many of which demonstrate the validity of the ‘problems’ that we have cited. The resources of any given firm can be broken down into a number of classes, both current and potential, with differing strategies calling upon different classes or combinations of classes for their success. As Edith Penrose (1959) emphasised, for example, administrative or managerial talent is one of the most important resources that a firm may have. She believed that the greatest factor behind the growth of firms is the presence of ‘slack’ administrative skills that results from managers learning to master recurrent problems. As learning occurs, managerial time is freed up to address new challenges. Among the other areas in which resources are important are production, marketing, research and development, raw material procurement, organisational culture and finance. Slack may arise from resources actually held at the moment (for example, excess manufacturing capacity) or from resources that can be tapped if needed (an ability to borrow funds to build and equip a new factory). In some cases, strategies may take advantage of resources that a firm has or can tap, but in other cases strategy may be formulated around a need to sidestep a relative weakness that a firm faces: a company that cannot afford the manufacturing and marketing investment required to become a full-line producer may decide instead to target a niche market. The value of each type of resource depends on the nature of the technology strategy being pursued. Both ‘the isolated resource problem’ and the ‘environment problem’ are important in the application of the RBP to technological questions. Although Bill Gates’s (1999) recent book gives the impression that Microsoft attempts to find ‘digital’ solutions to all problems, it is generally misleading to think that a firm whose operations are at all complex can follow a single technology strategy. Despite the emphasis in many publications on high-technology operations, mostly involving microchips, such strategies are not appropriate for the core operations of many businesses, which remain resolutely low-to-middle tech. Historically, systemic technological change has always progressed slowly and unevenly, with backward and forward linkages taking many years to develop, giving rise to the ‘reverse salients’ described by Hughes (1992). Lateral linkages have, if anything, frequently been slower. Although improvements may have been ‘in the air’ (Marshall 1961) and the common currency of thought in particular industrial districts, the spread of technological analogies across space and industries has often taken decades.3 In some cases, this has been a result of the economic logic of the situation. It would have made no more sense to build the entire European or American railway networks in a single push in the 1840s than equipping every house in Europe or the United States with a fibre optic connection would make today. Frequently, however, this slow diffusion of technologies also reflects inadequacies of knowledge and perception. Managers are generally not well acquainted with developments in other industries, and even when they have some knowledge they may not appreciate parallels between their own operations and those elsewhere.

INTRODUCTION 5

As a consequence, technological unevenness is to be expected across industries, but unevenness is also the rule across resources and capabilities within the same industry and within the same firm. Instead of industries or firms following high-tech or low-tech strategies, they use a combination of technologies that change at uneven rates, and the overall effects of change cannot be accurately gauged if attention is focused on the technological trajectory of only one or two resources. In areas such as furniture manufacturing or food processing, for example, in which both product and process technologies remain comparatively unsophisticated, state of the art resources in sales and inventory management may be employed to great competitive advantage. Concentration of analysis on a single, isolated resource can obscure the significance in many situations of a mix of technologies, each of which is appropriate to a different important resource. Furthermore, as industry structure and firm performance may interact as a result of technological change from outside the industry, an introspective approach to technology strategy can overlook vital interconnections between internal and external resources. Lowand medium-technology industries are embedded in networks of resources owned by others. For example, a change in the costs and speed of transport available to firms in a manufacturing industry can make it feasible for each firm to serve a larger market geographically, opening the way to enhanced economies of scale and other efficiencies without any change having occurred in the technology available within the manufacturing industry itself. Alternatively, the increase in scale could lead to the adoption of new technologies in the core industry, developments that would not have been efficient if the available markets were smaller. The upshot could well be a change in market structure, with an increase in overall sales but a reduction in the number of firms, as operators in isolated areas lose the protection provided by high transport costs. The analytical problems associated with an introspective approach become even clearer if, not unreasonably, one assumes that many uses of technological resources are ‘socially constructed’ (Bijker et al. 1987) in the sense that they reflect ‘social and political negotiation among a variety of groups’ (Ceruzzi 1998), including customers, regulatory bodies and other stakeholders in addition to technologists. Under these circumstances, the emerging strategic value of technological resources is far from linear, but reflects feedback loops as engineers and scientists are influenced by others in a complicated process. But, while it may be dangerous to ignore these complexities, they are also difficult to incorporate smoothly in a rigorous analytical framework. Dealing with complications of this order is one of the greatest challenges to the usefulness of the RBP. ‘The resource organisation problem’ is often associated with ‘the isolated resource problem’. Neoclassical economists have long acknowledged that the choice of a technology depends on the relative costs of a range of factors, including labour and raw materials as well as capital equipment. In an evolutionary situation, path dependency can increase the complexity of technological choice. As Robertson and Alston (1992) have shown, for instance, the ability of a firm to adopt a new technology is in part a function of power relationships within the firm and how receptive its existing workforce is to change. Technological change can be competence destroying, competence enhancing or competence neutral (Tushman and Anderson 1986) from the standpoints of both the firm

6 NICOLAI J.FOSS AND PAUL L.ROBERTSON

and its workers. Developments that are beneficial to the firm as a whole may nevertheless harm the interests of selected groups of workers. If these workers occupy strategic positions in the firm’s production process, they can bargain to gain a larger share of the benefits flowing from technological change than employers had planned on when investing in new equipment, or even block change altogether. In both cases, further investment in technological change would be discouraged. To be realistic, the RBP has to be able to contend effectively with these sorts of developments that can affect the strategic use of resources. Finally, to achieve its maximum degree of analytical usefulness, the RBP needs to cope both with the dynamic effects of technological change and with any accompanying need for new complementary resources. D’Aveni’s (1994) picture of firms operating under conditions of ‘permanent revolution’ contrasts strongly with the stance taken by RBP theorists such as Barney (1991, 1997) and Peteraf (1993), who emphasise the role of ‘sustained competitive advantage’ in strategic policy formulation. The concept of ‘hypercompetition’ that D’Aveni presents highlights the role of change as an on-going strategic weapon, while the equilibrium-based wing of the RBP tends to view change as a one-off lunge for superiority and domination. Realistically, there is no more reason to expect that most firms will operate virtually forever in highly turbulent environments—that the oligopolistic accommodation that is generally said to characterise the mature phase of the product life cycle has been replaced by hyper-competition—than there is to believe that most firms have any hope at all of securing Ricardian rents into the indefinite future. Nevertheless, D’Aveni’s view of hyper-competition does point up the dangers involved in embracing the equilibrium approach of Barney and Peteraf. In a world of multiple sources of innovation, it is quite simply useless for most firms to search for strategies that will yield sustained competitive advantage. Emphasis on incremental change and differentiation is more feasible for the majority of companies. This, in turn, often entails concentration on building complementary resources to take better advantage of the opportunities offered by innovation in any single area, since, if firms are unable to package their core competences suitably, they risk substantial problems in appropriating the returns from their areas of strength. An analytical and retrospective fixation on equilibrium by RBP theorists can yield only limited returns. What is required is a framework that can explain both rapid and slow change across both wide and narrow fronts. Many of the contributions to this volume refer to technology studies and innovation management. Most take a basic evolutionary approach, but try to retain rigour by discussing the endogenous creation of new resource in the context of established approaches, such as evolutionary economics (Metcalfe and James (chapter 3), Nicolai Foss (chapter 2), Hunt (chapter 4)), Austrian economics (Roberts (chapter 5)), and work on the theory of the firm (Kirsten Foss (chapter 9)). In the process of grappling with dynamic issues, most contributions also deal with other important problems in the RBP. For example, in chapter 6 Sanchez provides an extensive discussion of modular products and processes and thus illuminates ‘the resource application problem’, that is, the problem in the RBP that the actual application of

INTRODUCTION 7

resources in production receives comparatively little attention. Similarly, in chapter 11, Yu and Robertson examine how resources are actually applied by small manufacturers in Hong Kong, given the strengths and weaknesses of these firms in comparison with their international competitors. Relatedly, a number of contributors—such as Kirsten Foss in chapter 9 and Langlois in chapter 10—address ‘the resource organisation problem’. For example, Kirsten Foss draws on recent work on property rights to address the organisation of learning in firms and combines property rights ideas with resource-based ideas to give a rationale for the existence of firms that does not hinge on considerations relating to the reduction of opportunism but rather emphasises that firms may be superior (relative to markets) mechanisms for conducting experiments (broadly conceived) in a cost effective manner. ‘The isolated resource problem’, that is, the tendency in the RBP to treat resources as if they were free-standing entities, is also dealt with in a number of contributions. For example, Christensen’s discussion of what he calls ‘dynamic coherence’ (chapter 7) is a strong implicit denial of the meaningfulness of treating resources in an atomistic manner, since dynamic coherence is a property of a cluster of resources. Much the same point is made in Laamanen and Autio’s discussion of the role of dynamic complementaries in technology acquisition (chapter 8). Yu and Robertson (chapter 11) examine how strengths in organisational resources provide compensation for weaknesses in technological resources in Hong Kong, as well as assessing the limits of that compensation. Also relating to this problem are the contributions that in various ways deal with the interaction between firms and their environments, such as Langlois’ examination of vertical disintegration in the US semiconductor industry (chapter 10). In chapter 12, Shi looks at the transfer of technological resources from the perspectives of both the donors and the recipients. He describes one central strategy that the People’s Republic of China has used to generate new technological resources to promote economic development and analyses the differing motives of high-tech and low-tech foreign firms for undertaking direct investment in the PRC. The contributions to this book deal with core problems in the emerging resource-based perspective and aim to advance its problem-solving capacity and increase the range of issues to which it is relevant. Thus, the RBP is applied to topics relating to economic organisation, innovation and small-firm strategies. Taken as a whole, the contributions represent a significant advance with respect to the analytical tools the RBP may legitimately employ. In particular, they demonstrate the advantages of developing market process approaches to the RBP, such as Austrian and evolutionary economics. Notes 1 The essential contributions to the RBP are collected in Foss (1997). A recent collection of papers is Montgomery (1995). 2 In a useful distinction, Amit and Schoemaker (1993) define resources as ‘stocks of available factors that are owned or controlled by the firm’, while capabilities ‘refer to a firm’s capacity to deploy Resources, usually in combination, using organizational processes, to effect a desired

8 NICOLAI J.FOSS AND PAUL L.ROBERTSON

end’. In general we adhere to this distinction but, as capabilities are clearly also a resource, we will use the term resources when we refer to resources and capabilities in combination. 3 See the chart relating percentage of consumer use and years since the introduction of important innovations in Gates (1999:118). The chart also indicates that the diffusion of some recent technologies such as the personal computer and the internet has occurred more quickly than the spread of the radio or the VCR. In some cases, however, the measurements depend crucially on the definition of the innovation. The slow spread of the VCR, which was initially introduced for commercial purposes in 1952, is attributable to the need to develop rather different technologies for home usage (Graham 1986). Once this was accomplished, diffusion was quite rapid.

References Amit, R. and P.J.H.Schoemaker (1993) ‘Strategic Assets and Organizational Rent’, Strategic Management Journal 14:33–46. Barney, J.B. (1986) ‘Strategic Factor Markets’, Management Science 32:1231–41. ——(1991) ‘Firm Resources and Sustained Competitive Advantage’, Journal of Management 17:99– 120. ——(1997) Gaining and Sustaining Competitive Advantage, Reading, MA: Addison-Wesley. Bijker, W.E., T.P.Hughes and T.Pinch (1987) The Social Construction of Technological Systems, Cambridge, MA: MIT Press. Ceruzzi, P.E. (1998) A History of Modern Computing, Cambridge, MA: MIT Press. Collis, D.J. and C.A.Montgomery (1995) ‘Competing on Resources: Strategy in the 1990s’, Harvard Business Review July–August:118–28. D’Aveni, R.A. (1994) Hyper-competition: Managing the Dynamics of Strategic Maneuvering, New York: The Free Press. Foss, N.J. (1996) ‘Whither the Competence Perspective?’, in Nicolai J.Foss and Christian Knudsen (eds), Towards a Competence Theory of the Firm, London: Routledge. ——(1997) Resources and Strategy: A Reader, Oxford: Oxford University Press. Foss, N.J., C.Knudsen and C.A.Montgomery (1995) ‘An Exploration of Common Ground: Integrating Evolutionary and Strategic Theories of the Firm’, in C.A. Montgomery (ed.), Resource-Based and Evolutionary Theories of the Firm, Boston, MA: Kluwer Academic. Gates, B. (1999) Business @ the Speed of Thought: Using a Digital Nervous System, New York: Warner. Graham, Margaret B.W. (1986) The Business of Research: RCA and the VideoDisc, Cambridge: Cambridge University Press. Hughes, T.P. (1992) ‘The Dynamics of Technological Change: Salients, Critical Problems, and Industrial Revolutions’, in G.Dosi, R.Rianetti and P.A.Toninelli (eds), Technology and Enterprise in a Historical Perspective, Oxford: Clarendon Press. Kay, J. (1993) Foundations of Corporate Success: How Business Strategies Add Value, Oxford: Oxford University Press. Mahoney, J.T. (1995) ‘The Management of Resources and the Resource of Management’, Journal of Business Research 33:91–101. Mahoney, J.T. and J.R.Pandian (1992) ‘The Resource-Based View within the Conversation of Strategic Management’, Strategic Management Journal 13: 363–80. Marshall, A. (1961) Principles of Economics, 9th (Variorum) edn, London: Macmillan. Montgomery, C.A. (ed.) (1995) Resource-Based and Evolutionary Theories of the Firm, Boston, MA: Kluwer Academic.

INTRODUCTION 9

Nelson, R.R. and S.G.Winter (1982) An Evolutionary Theory of Economic Change, Cambridge, MA: The Belknap Press. Penrose, E.T. (1959) The Theory of the Growth of the Firm, Oxford: Oxford University Press. Peteraf, M.A. (1993) ‘The Cornerstones of Competitive Advantage: A Resource-Based View’, Strategic Management Journal 14:179–91. Porter, M.E. (1994) ‘Toward a Dynamic Theory of Strategy’, in R.P.Rumelt, D.E. Schendel and D.J.Teece (eds), Fundamental Issues in Strategy, Boston, MA: Harvard Business School Press. ——(1996) ‘What is Strategy?’, Harvard Business Review November–December: 61–78. Prahalad, C.K. and G.Hamel (1990) ‘The Core Competence of the Corporation’, Harvard Business Review May–June:79–91. Reve, T. (1990) ‘The Firm as a Nexus of Internal and External Contracts’, in M.Aoki, B.Gustafsson and O.E.Williamson (eds), The Firm as a Nexus of Treaties, London: Sage. Robertson, P.L. (1996) ‘Competences, Transaction Costs and Competitive Strategy’, in N.J.Foss and C.Knudsen (eds), Towards a Competence Theory of the Firm, London: Routledge. Robertson, P.L. and L.J.Alston (1992) ‘Technological Choice and the Organization of Work in Capitalist Firms’, Economic History Review 45:330–49. Romer, P.M. (1986) ‘Increasing Returns and Long-Run Growth’, Journal of Political Economy 94: 1002–37. ——(1990) ‘Endogenous Technological Change’, Journal of Political Economy 98: S71–S102. ——(1993) ‘Two Strategies for Economic Development: Using Ideas and Producing Ideas’, Proceedings of the World Bank Annual Conference on Development Economics 1992 (Washington: International Bank for Reconstruction and Development): 43–91. ——(1994) ‘The Origins of Endogenous Growth’, Journal of Economic Perspectives 8(1):4–22. Rumelt, R.P. (1984) ‘Towards a Strategic Theory of the Firm’, in R.B.Lamb (ed.), Competitive Strategic Management, Englewood Cliffs, NJ: Prentice Hall. Selznick, P. (1957) Leadership in Administration, Berkeley, CA: Harper and Row. Solow, R.M. (1956) ‘A Contribution to the Theory of Economic Growth’, Quarterly Journal of Economics 70:65–94. Tushman, M.L. and P.Anderson (1986) ‘Technological Discontinuities and Organizational Environments’, Administrative Science Quarterly 31:439–65. Wernerfelt, B. (1984) ‘A Resource-Based View of the Firm’, Strategic Management Journal 5:171– 80. Williamson, O.E. (1994) ‘Strategizing, Economizing, and Economic Organization’, in R.P.Rumelt, D.E.Schendel and D.J.Teece (eds), Fundamental Issues in Strategy, Boston. MA: Harvard Business School Press.

10

2 Equilibrium vs. evolution in the resourcebased perspective The conflicting legacies of Demsetz and Penrose Nicolai J.Foss

Introduction1 This chapter is taken up with discussing, among other things, the sources of what has now become generally known as ‘the resource-based perspective on strategy’ (henceforth, ‘the RBP’). Although the chapter should have some appeal to intellectual historians, it is not an exercise in the doctrinal history of the strategy discipline per se. The primary motivation of the chapter is rather to understand the present condition of the RBP, and to speculate on its possible future paths of development, in terms of the historical conditioning that some crucial contributions have imposed on the perspective. In other words, it is an examination of intellectual path-dependence2 in the context of strategic management and the theory of the firm, but also an attempt to suggest briefly how resource-based scholars may avoid some less fortunate future paths of development. Thus, at the general level the chapter sets the stage for much of the discussion in the rest of this book, and links up directly with the argument of some of the other chapters (particularly Metcalfe’s and James’ and Roberts’) by arguing that the RBP should take more seriously its Penrosian heritage and adopt an explicit process mode of analysis. It is well known that the RBP is conventionally (e.g., Mahoney and Pandian 1992; Knudsen 1996) traced back to the seminal, but for a long time neglected, contribution by Edith Penrose in The Theory of the Growth of the Firm (1959). However, in its modern manifestation the RBP may conveniently (if admittedly also somewhat arbitrarily) be dated to the year 1984, which was the year of publication of Birger Wernerfelt’s ‘A Resource-Based View of the Firm’ and Richard Rumelt’s ‘Towards a Strategic Theory of the Firm’. These papers were quickly followed by a spate of important work by such writers as Barney, Montgomery, Dierickx, Cool, Amit and others,3 and the following decade of strategy research is difficult to characterise as anything other than a wildfire of interest in resources, capabilities, competencies, etc. Thus, in little more than a decade, the RBP has emerged as arguably the dominant contemporary approach to strategy (content) research, indeed as perhaps the new orthodoxy in mainstream strategy research. The perspective’s appeal to academics may be explained by its relative4 ability to combine analytical rigour with apparent managerial relevance more successfully than alternative approaches (Foss 1996b).

12 NICOLAI J.FOSS

However, as has recently been pointed out (Schulze 1994; Mahoney 1995; Foss 1996a), the approach is far from being homogeneous. As an illustration, one may compare the enormously successful 1990 Harvard Business Review paper, ‘The Core Competence of the Corporation’, by C.K.Prahalad and Gary Hamel, and the 1982 Bell Journal of Economics paper, ‘Uncertain Imitability: An Analysis of Interfirm Differences under Competition’, by Stephen Lippman and Richard Rumelt. Nothing seems to relate these two papers, apart from a shared emphasis on firm resources. Everything else is different, and not just because they were written for different audiences. However, both papers are usually seen as important contributions to the RBP. Generalising this observation, we may point to a broader schism in contemporary resource-based thought of which these two papers are merely representatives—a schism that is both thematic and disciplinary. Thus, while those resource-based contributions that have taken their cues from Prahalad and Hamel are ‘soft’, in that they are taken up with issues learning, innovation, competence building, entrepreneurship, vision, etc. in a nonformal way, there is also a set of resource-based contributions that explicitly rely on equilibrium economics and, perhaps accordingly, do not treat the more dynamic phenomena that are centre stage in the ‘softer’ approach. Instead, the interest centres around the distribution of returns in economic equilibrium. In this chapter, I shall talk about ‘Mark I RBP’ (the equilibrium, economics-orientated version) and ‘Mark II RBP’ (the process-orientated version). Rather than further diagnosing this schism, I shall in this chapter dig a little deeper and be concerned with (some of) its causes. The claim here is that ultimately we need to look at the prehistory of the RBP in order to understand present tensions within the RBP, and in order to speculate on how the perspective may develop in the future. Many contributions have discussed the prehistory of the RBP5 and there is considerable consensus on the issue. According to this consensus, the basic resource-based insights were present in the work of Edith Penrose (1959) in particular, but also to some extent in the work of Philip Selznick (1957) and Alfred Chandler (1962). More impetus was given in the work of Kenneth Andrews (1971), but then resource-based ideas were temporarily swept aside by an all-consuming interest in industry analysis as a foundation for strategy. The revival of those resource-based ideas that had been present in what is essentially the mainstream of American strategy thinking was then undertaken in the 1980s by the younger resource-based theorists mentioned above. In this chapter I link up with this received view by discussing the work of Penrose as an important source of inspiration for the RBP. However, I add to the received view by discussing the work of UCLA economist Harold Demsetz as an influence on a par with that of Penrose.6 In fact, Demsetz’s influence may have been even stronger on the emerging RBP than Penrose’s and also more direct: it is almost certain that Demsetz influenced the early important resource-based theorists, such as Rumelt and Barney, through their shared institutional affiliation with the University of California at Los Angeles, and many of Demsetz’s ideas are directly reflected in the work of these important scholars. However, more than strict doctrinal history is involved: I also argue that the Demsetzian influence has tended to lockin the intellectual development of the RBP. Specifically, Demsetz’s influence meant that

EQUILIBRIUM VS. EVOLUTION IN THE RESOURCE-BASED PERSPECTIVE 13

the basic discipline underlying the development in the early 1980s of the RBP was equilibrium economics, albeit of a sophisticated kind. This has resulted in the RBP (Mark I), exemplified by the work of Rumelt and Barney. In many respects, the RBP (Mark I) is in conflict with the RBP (Mark II), exemplified by the work of Prahalad and Hamel, which I here contend manifests the process-orientated (or, if you like, evolutionary) influence stemming from Penrose.7 In order to place the discussion in context, I begin by providing a brief sketch of the resource-based approach(es), before moving on to argue that • not only Penrose but also Demsetz should be seen as dominant sources of inspiration for RBP scholars, • these two crucial influences hold different and even conflicting views of the economic process, and • they helped found different research areas and research approaches within the RBP. On the basis of this discussion, it is then argued that • the conflicting legacies of Penrose and Demsetz threaten the coherence of the RBP, • a central problem in the RBP is the lack of understanding of the process of resourcecreation which tends to give the perspective a retrospective character, but • work on technological innovation and change, framed in the broader theoretical context of Austrian and evolutionary economics, may help remedy this shortcoming, as argued by several contributors to this book. The resource-based approach: a brief sketch Simplifying somewhat, we may say that there are two main research themes in the RBP, namely, first, analyses of the conditions for sustained competitive advantage, and, second, diversification studies. They are presented seriatim in the following. Competitive advantage The resource-based analysis of (sustained) competitive advantage may be seen as starting out from two basic empirical generalisations, namely (1) that there are systematic differences across firms in the extent to which they control resources that are necessary for implementing strategies, and (2) that these differences are relatively stable. The basic structure of the RBP emerges when these two generalisations are combined with fundamental assumptions that are to a large extent derived from economics. Among these assumptions are (3) that differences in firms’ resource endowments cause performance differences, and (4) that firms seek to increase their economic performance. The overall managerial implication is that firms may secure a strong performance by building or otherwise acquiring certain endowments of resources. More generally, the overall objective that informs the RBP is to account for the creation, maintenance and renewal of

14 NICOLAI J.FOSS

competitive advantage in terms of the resource side of firms. The fundamentals of the resourcebased analysis of the conditions for sustained competitive advantage are basically simple (Peteraf 1993): in order that resources yield a sustained competitive advantage, they should meet four basic criteria: • Heterogeneity, i.e., in lieu of efficiency differences across resources, there cannot be any differences in the rents firms earn (in fact, there cannot be any rents at all). This indicates that resource heterogeneity, leading to efficiency differences and therefore rents, is a basic necessary condition for competitive advantage.8 • Ex ante limits to competition, i.e., resources have to be acquired at a price below their discounted net present value in order to yield rents. Otherwise future rents will be fully absorbed in the price paid for the resource (Demsetz 1973; Barney 1986; Rumelt 1987). • Ex post limits to competition, i.e., it should be difficult or impossible for competitors to imitate or substitute rent-yielding resources. As Dierickx and Cool (1989) clarify, there are in successful firms a number of mechanisms at work that often make it hard for competitors to copy the sources of competitive advantage of a successful firm. For example, there may be ‘causal ambiguity’, which means that competitors confront difficulties ascertaining precisely how a bundle of resources contributes to success. • Imperfect mobility, i.e., the resource should be relatively specific to the firm. Otherwise, the superior bargaining position that is obtained from not being tied to a firm can be utilised by the resource (or the resource’s owner) to appropriate the rent (or at least a large portion of the rent) that the resource helps create. In other words, the key questions to ask here are who captures value from the resource and how may the firm capture more value from this resource? Several things are noteworthy about this basic analysis. First, it explicitly draws on economics, more precisely on basic, equilibrium price theory.9 For example, in connection with ex ante limits or barriers to competition, these limits are evaluated relative to a full-information, competitive equilibrium (Barney 1986). It is clearly the case that equilibrium assumptions play a key role in many contributions to the RBP. This is the case in Peteraf (1993), in which the concept of Ricardian rent is developed using efficiency differences across firms under competitive equilibrium as a benchmark. And it is also the case in Barney (1986), in which the finance concepts of strong and weak efficiency are (implicitly) used to elucidate the reasoning behind the concepts of perfect factor markets and factor market imperfections. Indeed, the very concept of sustained competitive advantage is often defined in equilibrium terms: it is that advantage which lasts after all attempts at imitation have ceased. This has the implication, unfortunately, that sustained competitive advantage has no meaning outside equilibrium. Moreover, the above analysis actually tells us very little of direct value for understanding the more dynamic and managerial aspects of competitive advantage, such as how to build new resources, coordinate existing ones, etc.

EQUILIBRIUM VS. EVOLUTION IN THE RESOURCE-BASED PERSPECTIVE 15

Diversification Although the basic story is much refined now, the basic RBP analysis is not novel, as Penrose (1959) essentially laid the foundations here (her work is discussed in more detail later). What is new relative to Penrose, however, is (1) the point that diversification may in turn help in building new resources (cf. Markides and Williamson 1994), (2) the insight that it is necessary to bring transaction costs into the story (Teece 1980), (3) the more rigorous form the argument has now taken, and (4) the many empirical tests that have been carried out (e.g., Montgomery and Wernerfelt 1988). Diversification studies may arguably be where the resource-based approach has had its greatest impact. The commonly accepted theory of diversification, in both economics and strategy research, is roughly the resource-based theory (Montgomery 1994). The basic story is that firms gradually accumulate excess resources as a (non-intended) consequence of their normal operations. Tasks become routinised and this releases human resources, such as managerial resources; some physical resources are indivisible, which means that they may not be fully exploited in their present use; etc. In principle, these resources could be traded over markets; however, the presence of transaction costs often hinders trading excess resources. As Teece (1980) clarifies, this is particularly likely to be the case if the resources in question are knowledge resources. An important implication of the theory is that firms earn decreasing average rents as they diversify more widely (Montgomery and Wernerfelt 1988). Varieties of the resource-based approach: equilibrium or evolution? As already indicated, there is substantial evidence, such as the complete absence of terminological agreement, that the RBP is far from being a coherent perspective. It is a set of contributions published since about 1984 (plus some important precursors to these contributions) that share important basic themes. But, apart from that, the contributions that may be seen as constituting the RBP are undeniably quite heterogeneous in terms of, for example, which disciplines they draw on (economics, sociology, psychology, decision theory). The argument here is that it makes sense to distinguish two different versions of the RBP, the RBP (Mark I) and the RBP (Mark II), and that existing differences between these are to a very large extent a matter of whether one seeks to address and include dynamic— or better, evolutionary—factors, or instead relies on standard economic theory.10 It is a choice, in short, between equilibrium or evolution, ‘evolution’ and ‘evolutionary’ here being meant to refer to whether such concepts as irreversibility (e.g., in the form of pathdependence and learning) and novelty (e.g., in the form of unanticipated innovations) are included in the analysis at some level (cf. Loasby 1991; Foss et al. 1995).11 While various dynamic phenomena (innovation, organisational learning, resource accumulation, competence building, the development of mental models of the management team, etc.) come first in the RBP (Mark II) (Prahalad and Hamel 1990;

16 NICOLAI J.FOSS

Hamel and Heene 1994; Teece and Pisano 1994), statics comes first in the RBP (Mark I). That is to say, in the latter approach one begins by clarifying and examining the conditions that must obtain in order for resources to yield rents in equilibrium (e.g., Barney 1986, Peteraf 1993 or Wernerfelt 1995), before going on to discuss, for example, the renewal of competitive advantage (supposing one ever gets this far). Thus, process issues enter subsequently, or are simply suppressed. In the RBP (Mark II), equilibrium is at most a theoretical benchmark; an abstraction without any apparent practical value. If this analysis is correct, resource-based scholars would at least to some extent seem to be caught on the horns of a dilemma, whether to emphasise realism and perhaps sacrifice some analytical rigour and clarity, or vice versa, and to confront difficult questions, such as the issue of how to combine process analysis (including firm growth) and sustainability of competitive advantage. In the next section, I explore some sources of the schism in resource-based thought by looking at the contributions of two crucial precursors of the RBP, Edith Penrose and Harold Demsetz. Two crucial precursors: Penrose and Demsetz Edith Penrose on the theory of the growth of the firm It is a commonplace that many of the great works of economics have been interpreted in widely different ways, and normally in both a mainstream, neoclassical way and in a nonneoclassical way.12 This is also the case with Edith Tilton Penrose’s (1914–96) major work, The Theory of the Growth of the Firm. Thus, Paul Rubin (1973) rationally reconstructed Penrose’s work in terms of finding the best solution to the dynamic optimisation problem of balancing the development of new resources (using existing resources) and the use of existing resources directly in production. On the other hand, Penrose’s work has been heavily cited by heterodox economists (e.g., Loasby 1991),13 and her foreword to the third edition of The Theory of the Growth of the Firm (published in 1995) leaves little doubt that her sympathies were with more heterodox strands of economics. As she noted there: ‘One of the primary assumptions of the theory of the growth of firms is that “history matters”; growth is essentially an evolutionary process and based on the cumulative growth of collective knowledge, in the context of a purposive firm’ (1959 [1995]:xiii). The basic reasoning of the 1959 book is very well known and will be only briefly summarised: Firms are collections of productive resources that are organised in an administrative framework which partly determines the amount and type of services that the resources yield. As they proceed with their productive operations, firms in Penrose, particularly the management team, acquire increased knowledge of the services that may be obtained from resources. The (related) results of such learning processes are, first, the expansion of the firm’s ‘productive opportunity set’ (the opportunities that the firm’s management team can see and can take advantage of) and, second, the release of managerial excess resources that can be put to use in other, mostly related, business areas.

EQUILIBRIUM VS. EVOLUTION IN THE RESOURCE-BASED PERSPECTIVE 17

Since the opportunity costs of excess resources are zero, there will be a strong internal incentive for such diversification. Because the firm’s expansion to a large extent builds on its ‘inherited’ resources, and because there ‘is a close relation between the various kinds of resources with which the firm works and the development of the ideas, experience and knowledge of its managers and entrepreneurs’ (Penrose 1959:85), this expansion will tend to take place in areas of competence that are close to the firm’s existing areas of competence. Undeniably, the basic skeleton of some of Penrose’s ideas may be cast in the language of equilibrium and (dynamic) optimisation characteristic of mainstream economics. For example, one may argue that at any given point of time there is a set of product market applications (business areas) that maximise the rents on the firm’s existing resources and correspond to an organisational equilibrium (à la Montgomery and Wernerfelt 1988), and a part of the optimisation problem is the information costs that the firm’s management team confronts (Casson 1997). However, in Penrose’s own view, her theory constituted a powerful critique against certain aspects of the neoclassical theory of the firm (if not necessarily against neoclassical economics in general). In the neoclassical theory of the firm, there is ‘no notion of an internal process of development leading to cumulative movements in any one direction’ (1959:1), a notion that is absolutely crucial for understanding firm development. Rather, growth is simply a matter of adjusting to the equilibrium size of the firm. But if services are produced endogenously (and continuously) through various intra-firm learning processes involving increased knowledge of resources, ‘new combinations of resources’ (p. 85) and an expanding productive opportunity set, there is no equilibrium size. There is clearly what we today would recognise as a Schumpeterian (change ‘from within’) and Veblenian (cumulative causation) flavour to such arguments. But it is more than a matter of dressing up arguments in fancy Schumpeterian garb. Penrose’s basic vision of the competitive process in general, and of the firm in particular, is disequilibriumorientated and subjectivist,14 and, normally overlooked, it stresses entrepreneurship, flexibility, change and uncertainty. ‘In the long run’, Penrose (p. 137) explains, the profitability, survival and growth of a firm does not depend so much on the efficiency with which it is able to organise the production of even a widely diversified range of products as it does on the ability of the firm to establish one or more wide and relatively impregnable ‘bases’ from which it can adapt and extend its operations in an uncertain, changing and competitive world. Thus, seemingly paradoxically, flexibility is just as much a message of the analysis as specialisation is. The paradox vanishes when it is realised that specialisation is specialisation in terms of the underlying resource-base (rather than products) and that such specialisation may be fully consistent with reacting to new business opportunities. In fact, as Penrose makes clear, there may be a considerable option value associated with a specialised resourcebase:

18 NICOLAI J.FOSS

A firm is basically a collection of resources. Consequently, if we can assume that businessmen believe there is more to know about the resources they are working with than they do know at any given time, and that more knowledge would be likely to improve the efficiency and profitability of their firm, then unknown and unused productive services immediately become of considerable importance, not only because the belief that they exist acts as an incentive to acquire new knowledge, but also because they shape the scope and direction of the search for knowledge. (p. 77) Thus, firm development is essentially an evolutionary and cumulative process of ‘resource learning’ (Mahoney 1995), in which increased knowledge of the firm’s resources both helps create options for further expansion and increases absorptive capacity (Cohen and Levinthal 1990). Therefore, a major focus of The Theory of the Growth of the Firm lies in the application of resources, something that has been missed by many resource-based theorists who only consider the issues of the terms at which resources were acquired (Barney 1986) and/or whether they are protected (Peteraf 1993), but forget that it is the actual application, and not the mere possession, of resources that creates revenue (Spender 1994). To sum up, Penrose’s seminal work is indeed a founding contribution to the RBP. However, her basic, and too often overlooked, themes—flexibility in an uncertain world, organisational learning as an evolutionary discovery process, path-dependency, the vision of the management team, entrepreneurship, etc.—do not seem to square easily with the RBP (Mark I), that is, the version of the RBP which utilises equilibrium constructs and builds directly on price theory. And it is indeed the contention here that this strand of the RBP finds its most important source of inspiration, not in the work of Penrose, but rather in Chicago-UCLA price theory, notably as represented by the work of Harold Demsetz. Harold Demsetz on industrial economics The work of Harold Demsetz (b. 1930) has fallen within a number of economic subdisciplines. Thus, he is (with Ronald Coase and Armen Alchian) a pioneer in the development of the theory of property rights and the theory of the firm, and an important contributor to the theory of industrial organisation. He is often thought of as an important member of the Chicago school of antitrust analysis, although most of his career has taken place at the University of California, Los Angeles. Although there is a considerable degree of coherence to Demsetz’s whole œuvre, it is primarily in his capacity as a contributor to industrial organisation economics that I shall consider him here. Much of Demsetz’s work (see, in particular, Demsetz 1974) in this area has been concerned with critically discussing doctrines developed by economists associated with the so-called ‘Structure-Conduct-Performance’ school in industrial organisation (Bain 1959; Scherer 1980). According to this school, there is a strong causal flow from the basic structure of an industry (e.g., number of firms, entry-barriers), to their conduct (e.g., firms’ pricing policies), to performance (e.g., how large is the deadweight welfare loss). Specifically, Demsetz has subjected conventional thinking on entry-barriers and on the

EQUILIBRIUM VS. EVOLUTION IN THE RESOURCE-BASED PERSPECTIVE 19

link between industry structure and performance to critical scrutiny. These critical discussions have all been based on the conviction that models that do not feature information costs—costs of search, of processing information, of communication, etc.— are likely to distort seriously our understanding of the industrial landscape, and in particular to lead policy analysis astray (Demsetz 1969). On the other hand, including information costs in economic models, Demsetz argues, will reveal that many of the practices15 that have traditionally been condemned as monopolistic abuses are in reality efficiency-enhancing arrangements (Demsetz 1982). Thus, Demsetz was probably the first economist to develop an understanding of barriers to entry as essentially informational in nature (and to argue that this understanding should influence antitrust policies). For example, advertising has often been singled out as an important (strategic) entry-barrier. And, in fact, in a model that does not feature information costs, it is difficult to rationalise advertising as anything else than an instrument that is used in the pursuit of monopolistic advantages. But an information cost perspective allows for the understanding that advertising and brand loyalty are rational responses to an underlying scarcity of information. Thus, the real entry-barrier is not the advertising, but rather the information costs and it is not at all clear that antitrust authorities should be concerned about these costs. This focus on information asymmetries and costs as the real entry-barriers is clearly related to the overall resource-based idea that the primary barriers that hinder the equalisation of rents across firms are informational in nature.16 But there are many other similarities. In order to elucidate these, I shall quote extensively from a single paper, namely Demsetz’s 1973 article, ‘Industry Structure, Market Rivalry, and Public Policy’. It is here that we encounter the most explicit anticipations of what would eventually become the RBP.17 The paper is taken up with discussing the observation that a concentrated industry structure is often accompanied by high returns. Basically, there are two hypotheses that may account for this. The first one is that presented by the SCP paradigm: a high degree of market concentration eases coordination among oligopolists and thereby the setting of a price approximating the profit-maximising monopoly price. The other hypothesis is essentially an efficiency hypothesis, according to which the coexistence of high returns and high concentration in an industry is caused by more efficient firms growing at the expense of their smaller rivals, contributing to an increase in concentration, and earning higher returns than these smaller rivals. Demsetz tests these two hypotheses in an ingenious way, finds support for the efficiency hypothesis, and draws the appropriate policy conclusions. At the beginning of the paper, Demsetz launches one of his favourite themes, namely that economists are prone to seeking monopoly explanations for virtually all deviations from perfect competition. However, in a world of uncertainty and positive information costs, many of these deviations reflect not monopolistic practices, but efficient responses to scarcity. For example, advertising and credit rationing are rational practices in a world of positive information costs. More importantly, the presence of information costs, uncertainty and less than fully mobile factors may imply that ‘a differential advantage in

20 NICOLAI J.FOSS

expanding output develops in some firms’ (Demsetz 1973:1). And the returns (rents) that such differential advantages may yield need not be eliminated soon by competition. It may well be that superior competitive performance is unique to the firm, viewed as a team, and unobtainable to others except by purchasing the firm…The firm may have established a reputation or goodwill that is difficult to separate from the firm itself…Or it may be that the members of the employee team derive their higher productivity from the knowledge they possess about each other in the environment of the particular firm in which they work, a source of productivity that may be difficult to transfer piecemeal. (p. 2) Note the emphasis placed on heterogeneity, on different resource-bundles as the sources of heterogeneity and therefore differential efficiencies that in turn are the basis for differential competitive advantages. Note also the emphasis placed on team-effects, such as the learning effects that arise from the continuity of association between inputs.18 All this is as clear an anticipation of the resource-based emphasis on heterogeneity as the basic condition of competitive advantage as one could wish for. But there is more, for Demsetz has also, in the same article, interesting things to say about what are essentially the resource-based conditions of ‘ex post limits to competition’ and ‘ex ante limits to competition’. With respect to the former, Demsetz says, in connection with a discussion of the emergence of superior efficiencies, that One such enterprise happens to ‘click’ for some time while others do not. It may be very difficult for these firms to understand the reasons for this difference in performance or to know which inputs to attribute the performance of the successful firm. It is not easy to ascertain just why G.M. and I.B.M. perform better than their competitors. The complexity of these organisations defies easy analysis, so that the inputs responsible for success may be undervalued by the market for some time. (p. 2) In other words, firms may enjoy long-lived rents because would-be imitators confront difficulties ascertaining ‘just why’ some firms perform better than others, difficulties that are explicitly traced to ‘complexity’. In the last sentence of the above quotation, Demsetz also signals that the valuation of factors on their relevant markets influences returns. He goes on to observe that inputs are acquired at historic cost, but the use made of these inputs, including the managerial inputs, yields only uncertain outcomes. Because the outcomes are surrounded by uncertainty and are specific to a particular firm at a particular point in its history, the acquisition cost of inputs may fail to reflect their value to the firm at some subsequent time. By the time their value to the firm is recognised, they are

EQUILIBRIUM VS. EVOLUTION IN THE RESOURCE-BASED PERSPECTIVE 21

beyond acquisition by other firms at the same historic costs, and, in the interim, shareholders of the successful or lucky firm will have enjoyed higher profit rates. (p. 2) By implication, competitive advantage can only be obtained from resources that are acquired at a price below their discounted present value—the essential point in an important resource-based paper by Barney (1986) (see also Rumelt 1987 for the same reasoning). Moreover, certain ex ante barriers to competition, notably the presence of information costs on input markets, imply that divergences may arise between the prices of resources and the discounted present value of those resources. Sufficient evidence has now been presented to allow us to infer that Demsetz should indeed be reckoned as among the important precursors of the RBP, particularly with respect to the analysis of the conditions of sustained competitive advantage, the first key research theme within the RBP. In this respect, he is more important than Penrose, for she does not really inquire into these conditions in her 1959 book.19 An important issue is whether these ideas were first developed by Demsetz, and then independently discovered by later resource-based strategy theorists. Or, is there a more direct route through which Demsetz’s ideas may have spread? In the following, I argue that there may in fact have been such a direct route. Jay Barney, who was one of the prime movers behind the emergence of the resourcebased approach in the 1980s, recently argued that the rational reconstruction approach to the history of the RBP, according to which the development of the RBP can be dated back to Selznick and Penrose and progressing rather smoothly from there, is simply a ‘myth’ (Barney 1995). Instead, Barney argued that the modern RBP largely owes its origin to the interaction, mainly at UCLA, between such economists and strategy scholars as William Ouchi, Michael Porter, Richard Rumelt, Oliver Williamson, Sidney Winter and Barney himself. Only subsequently came the realisation that much of the early work of Selznick, Penrose, Chandler and Andrews anticipated modern resource-based thought. Barney’s critique is a welcome warning about too eagerly ascribing to older writers views that, only by twisting facts, can they be seen as anticipating, but it also leaves out a number of important considerations.20 Space is too limited here, however, to criticise Barney’s account in detail. Instead, I shall focus on his emphasis on the UCLA environment. From the interaction at UCLA emerged two seminal contributions that came to play a founding role for the emerging RBP (Mark I) in the 1980s. The first was Lippman and Rumelt’s 1982 paper, ‘Uncertain Imitability: An Analysis of Interfirm Differences under Competition’, in which they demonstrate that it is possible to sustain an equilibrium with firms that earn different returns (rents) (because they have different productive efficiencies) as long as imitation barriers hinder the equalisation of rents across firms. The intuition of this paper is pure Demsetz, and the paper may be seen as formalisation of key ideas in, for example, Demsetz (1973). The other seminal paper is Barney’s 1986 Management Science article on ‘Strategic Factor Markets’, in which he argues that imperfections in input markets are a necessary condition for competitive advantage; otherwise, the discounted present value of resources will be fully capitalised in their acquisition price. Again, this is a

22 NICOLAI J.FOSS

restatement (and extension) of reasoning that was present much earlier in Demsetz’s work.21 These facts should be combined with the emphasis given in Barney’s account to the UCLA environment and with the fact that it was in the UCLA economics department that Demsetz taught while Barney and Rumelt occupied positions in UCLA’s Anderson Graduate School of Management. Perhaps there was no direct personal influence, but only an exposure on the part of Barney and Rumelt to a more general UCLA style of doing and thinking of economics, a style that was also represented by Benjamin Klein and Armen Alchian.22 But the connections are too obvious to be merely spurious. Implications for the resource-based perspective Taking stock It is time to take stock of the above discussion. I have argued, first, that there are two key themes in the RBP, the analysis of sustained competitive advantage and the analysis of diversification. Second, it has been argued that the RBP actually exists in two different versions, Mark I and Mark II, and that the difference between these is largely a difference in terms of the extent to which dynamic factors are treated, as in the underlying analytical frameworks (equilibrium vs. evolution). Third, I have traced the key themes of the RBP and the two different types of theorising existing within the RBP to the work of the two crucial precursors, Penrose and Demsetz. Thus, Demsetz’s influence not only manifests itself in the equilibrium style of analysis pursued by RBP (Mark I) theorists, but also is manifest in the way that the theme of sustained competitive advantage is handled within the RBP. Penrose’s entirely different and non-neoclassical, non-equilibrium emphasis on learning, vision, entrepreneurship, flexibility, etc., on the other hand, is clearly manifest in the RBP (Mark II), that is, the work that has to a large extent taken its cue from the work of Prahalad and Hamel. In other words, Demsetz and Penrose’s seminal and widely different works have laid the foundations for diverging paths of development within the RBP. This may be interesting as a matter of intellectual history; but what does it matter to the future development of the RBP, not to mention practical concerns? Implications The Demsetz influence on the RBP (Mark I), I have argued, helped align strategy and economic equilibrium. As Spender (1993:42) noted in a related context, ‘The notion of rents is simply a way of bringing the homogeneity of economic thought together with the heterogeneity of the real world.’ For example, if information costs are positive, we can have an equilibrium with firms of different efficiencies and rents (and therefore different competitive advantages), and we can perform the usual comparative static exercises in this setting (Demsetz 1973, 1989b; Lippman and Rumelt 1982). Moreover, equilibrium, in the eyes of writers such as Barney, is a useful benchmark, one that can be used for

EQUILIBRIUM VS. EVOLUTION IN THE RESOURCE-BASED PERSPECTIVE 23

analysing factor market imperfections and sustained competitive advantage. The latter, recall, is defined as the advantage that lasts after all attempts at imitation have ceased (Barney 1991:102). So a sort of zero imitation, Nash equilibrium is utilised as a yardstick to define and understand (sustained) competitive advantage. While it cannot be denied that in some ways the alignment of equilibrium and strategy has proven fruitful, it should also be recognised that it made difficult the incorporation of the Penrosian legacy with its emphasis on organisational learning, entrepreneurship, etc. in the more formal, economics-inspired body of resource-based thought. The result has been the emergence of what has here been called the Mark I and Mark II versions of the RBP, with the latter addressing the more dynamic issues of resource-creation, but doing so in such broad and sometimes diffuse terms that their real contribution to the furthering of the RBP may be questionable. As a result of this dichotomisation of resource-based research, there is clearly a lack of a clear and coherent treatment of dynamic factors: while the RBP (Mark II) does address dynamic issues, it does so in rather diffuse and incoherent terms, and while the RBP (Mark I) is clear and coherent, there is no real treatment of dynamics. Therefore, the RBP does not in its present version(s) generate an adequate theory of the mechanisms underlying the creation of new resources, a feature that tends to give the perspective a distinctly retrospective orientation. More specifically, the perspective cannot adequately frame questions relating to corporate renewal, organisational learning, resource building, etc. It is true, of course, that the RBP (Mark I) can to a limited extent frame such questions by treating, for example, capabilities for corporate renewal as rare, hard to imitate, etc., so that these capabilities are seen as strategic resources. But this ex post analysis is clearly begging the normative and practical issue of how firms may build such capabilities. This problem is arguably particularly troublesome for the future evolution of the RBP, for dynamics (broadly conceived) is all the rage in the strategy (and organisational behaviour) field(s) these days, as witnessed by the recent enthusiasm for ‘hyper-competition’, ‘organisational learning’, ‘the knowledge-creating company’, etc. The underlying problem in this context is that there is no clear conceptual model of the endogenous creation of new resources to be found in the RBP. The same critique that Penrose directed against the neoclassical theory of the firm is also applicable to the RBP: there is ‘no notion of an internal process of development leading to cumulative movements in any one direction’ (1959: 1). Thus, while Demsetz (1973), Lippman and Rumelt (1982) and Barney (1986) provide a theory of rents in equilibrium, they actually tell us very little about how the heterogeneous conditions underlying differential rents arise. Clearly, this has something to do with the role of equilibrium and the restrictive behavioural assumptions that normally accompany equilibrium models (such as admitting only maximising rationality) in this sort of work. As many writers, including Penrose, have pointed out, an overly firm commitment to equilibrium and optimisation may seriously impede the development of models of endogenous change. Learning, innovation and entrepreneurial discovery activities by definition involve novelties in the sense of the acquisition or creation of novel knowledge, and such novelties are hard to force into an equilibrium straitjacket (Loasby 1991). Thus, one important reason why the RBP lacks a

24 NICOLAI J.FOSS

clear model of the endogenous creation of resources may simply be that equilibrium economics of the Demsetz variety has been such an important force in the development of the RBP. Instead, theorists have fallen back on what we may call ‘Big Bang Theories of Competitive Advantages’ (Spender 1993:45): competitive advantage is rationalised by pointing to an initial, unexplained event, such as sheer luck. Remedies What the RBP needs, in the view taken here, is more agreement that the dynamic issues featured in such contributions as Prahalad and Hamel (1990) are crucial, that they should be approached in a more precise and analytical way, that economics may prove useful, but also that equilibrium and maximisation may not always be the best tools for framing strategic issues. If this agreement does not obtain, there is a real danger that the RBP may split even more visibly, first, into a formal, stark, abstract branch strongly inspired by economics and gradually losing contact with managerial reality, and, second, into an increasingly loose and free-wheeling branch where almost anything goes on the analytical level. In general, it has been argued, there is a need for bringing process issues more directly into the focus of the RBP, and much of the neglect of such issues has to do with the influence of equilibrium economics on the RBP. Accordingly, it appears to be natural to turn towards economic theories that address process/disequilibrium issues, or what may be called ‘market process theories’. This is the approach taken by several of the contributors to this book (for example, Peter Roberts), who rely on arguments from Austrian economics and emphasise the quality of entrepreneurial alertness (Kirzner 1973). Like Kirzner today, Penrose argued that ‘the decision to search for opportunities is an enterprising decision requiring entrepreneurial intuition and imagination and must precede the “economic” decision to go ahead with the examination of opportunities’ (1959:34). In such a view, it is misleading to reduce competitive advantage to luck or asymmetric information and to think of sustainability as a matter of the persistence of rents in equilibrium. Rather, sustainability becomes (also) a matter of continuous alertness to a stream of disequilibrium opportunities for profit. However, as argued elsewhere in more detail (Foss 1996a), it is ironic that Austrian economics, as a theory about the market process, has so very little to say about the arguably most important constituent element of the market process, namely the firm. A processual approach that spans several levels of analysis, including notably that of the firm, is evolutionary economics, which Metcalfe and James in this volume argue is able to further the RBP (see also Montgomery 1995; Foss 1996a; Teece et al. 1997). For example, evolutionary economics and the RBP (in both versions) are both characterised by emphasising the fundamental heterogeneity of firms as a necessary starting-point for theorising, but, in contrast to the RBP (at least in its Mark I version), evolutionary economics endogenises the sources of heterogeneity. For example, evolutionary economists have cultivated an advanced understanding of the mechanisms of technological

EQUILIBRIUM VS. EVOLUTION IN THE RESOURCE-BASED PERSPECTIVE 25

change, an insight that may both help develop a more refined resource-based analysis of the environment and increase our understanding of the process of creation of new resources through innovation. Thus, notions such as ‘technological paradigms’, ‘regimes’, ‘technology systems’, etc. provide an understanding of the environmental forces that change the distribution of returns over time. Conclusions Any theoretical perspective carries with it an open horizon in the sense that we cannot exactly know in advance how the perspective will fare with respect to future problemsolving. However, where a perspective will go is constrained by where it has been in the past. In this chapter, I have argued that the influences of two central precursors of the RBP have resulted in a split within the RBP in an economics-orientated and equilibriumbased version (Mark I), which reflects the influence of Harold Demsetz, and a disequilibrium-orientated version, which owes much more to the influence of Edith Penrose (Mark II). Furthermore, the argument has been that unless a sort of alignment between these two different versions of the RBP is accomplished, there is a risk that they will develop even more strongly in their own distinct directions, the equilibrium branch becoming increasingly formal (possibly mathematical) and orientated towards mainstream economics, while the process-orientated branch will increasingly draw on ‘softer’ ideas and disciplines. As several contributions to this book indicate, this alignment may arguably be accomplished by drawing on Austrian and/or evolutionary economics, and by relating to work on organisational learning and technological innovation and change. Notes 1 This chapter is essentially an extension of the reasoning in Foss (1996a), and repeats some of the insights and conclusions of that paper. The comments of Jay Barney, Peter Earl and Paul Robertson on earlier versions are gratefully acknowledged. All remaining errors, obscurities, etc. are entirely my responsibility. 2 In analogy to the work of David (1985) and Arthur (1989). 3 The important contributions are reprinted in Foss (1997). 4 Relative to other streams in strategic management. 5 Conner (1991) is without any doubt the most meticulous discussion; see also Mahoney and Pandian (1992), Foss (1996a) and Knudsen (1996). 6 Admittedly, the Demsetz influence is less visible than the Penrose influence; for example, Demsetz is quoted less frequently by RBP scholars. Thus, the admittedly speculative part of my story is that Demsetz to some extent has been a ‘sleeping partner’ in the evolution of the RBP. 7 Because I want to concentrate on the Demsetz vs. Penrose story, I disregard the other possible reasons for intra-RBP heterogeneity, such as the greater practice connection of the RBP (Mark II) as contrasted with the more academic orientation of the RBP (Mark I). See Spender (1993) for interesting reflections on the history of the strategy field that highlight the

26 NICOLAI J.FOSS

8

9 10

11

12 13 14

15 16 17 18 19 20 21

22

tensions between academic ambitions and relevance for managerial practice. Similarly, I disregard other possibly important influences, such as the work of Schumpeter and the Austrians. An alternative formulation put forward by Barney (1991) is that with homogeneous resources, all firms can implement the same strategies; hence, no firm can differentiate itself from other firms, and nobody will have a competitive advantage. More specifically, as I later argue, it is the Chicago-UCLA brand of price theory on which the basic resource-based analysis of sustained competitive advantage is based. The ‘two resource-based approaches’ argument is also made by Schulze (1994), Mahoney (1995) and Foss (1996a), who all point to the static/dynamic distinction in rationalizing this argument. It should be mentioned that while all evolutionary theories are dynamic, not all dynamic theories are evolutionary (for example, economic growth theory is not evolutionary). The use of the word ‘evolutionary’ in this paper is deliberately broad, and does not carry any necessary connotations to the biological analogies that were so harshly criticised by Penrose (1952). Think of the work of Marx, Marshall and Keynes, to mention just the most obvious. It has even been argued that the economist whose work lies closest in many respects to Penrose’s is the idiosyncratic, but often brilliant, Thorstein Veblen (Foss 1998). Penrose’s subjectivism is particularly apparent in her adoption of Kenneth Boulding’s concept of ‘the image’: ‘the environment is treated…as an “image” in the entrepreneur’s mind of the possibilities and restrictions with which he is confronted, for it is, after all, such an “image” which in fact determines a man’s behaviour’ (1959:5). In other words, the environment is basically ‘enacted’, to use Weick’s terminology. See also Roberts (1997) for a critique of the inability of resource-based scholars to come to grips with these aspects of Penrose’s work. Notably various ‘vertical restraints’. For example, causal ambiguity, as in Lippman and Rumelt (1982). It should be noted that this paper is often cited in contributions to the RBP, for example Conner (1991). Demsetz (1988) elucidates this and builds a theory of the firm on this basis. This later paper has also become a standard reference in the RBP literature. On the other hand, Demsetz has had very little to say about Penrose’s major theme, namely firm growth through efficient diversification. For example, it neglects David Teece’s (1980) and Birger Wernerfelt’s (1984) role and the fact that both Wernerfelt and Teece in their early papers explicitly draw on Penrose’s work. ‘What directly prompted the writing of that paper, however, was an earlier paper by Rumelt and Wensley, “In Search of the Market Share Effect”: In this paper, Rumelt and Wensley argued that there is a market for market share, and that this market is quite efficient. They have a sentence in that article that says something like “this argument, of course, depends on rational expectations”. My 1986 Management Science article was an effort to understand the implications of this sentence’ (Barney, pers. comm.). That this is so is indicated by the following comment by Professor Barney: ‘In terms of the Demsetz connection, I knew him, but not well. Both Dick [Rumelt] and Bill Ouchi knew him better. Kathleen [Conner] knew him pretty well…We were certainly aware of his work…Despite this, I would not say that Demsetz had a strong personal influence on those of us who were at UCLA. In fact, if anyone had this influence, it was Armen Alchian… While the personal Demsetz connection was not there (at least for me), there is no doubt

EQUILIBRIUM VS. EVOLUTION IN THE RESOURCE-BASED PERSPECTIVE 27

that the equilibrium approach used by Demsetz and Alchian was a very strong influence at UCLA…if economically oriented faculty wanted to have any credibility at all at UCLA, they had to do equilibrium kind of analyses because of the standards set by Demsetz, Alchian and Klein. That was absolutely clear. This is one reason why I never found Penrose all that helpful’ (Barney, pers. comm.).

References Andrews, Kenneth (1971) The Concept of Corporate Strategy, 1993 edn, Homewood, AL: James-Irwin. Arthur, W.Brian (1989) ‘Competing Technologies, Increasing Returns, and Lock-In by Historical Events’, Economic Journal 99:116–31. Bain, Joe (1959) Industrial Organization, New York: John Wiley and Sons. Barney, Jay B. (1986) ‘Strategic Factor Markets’, Management Science 32:1231–41. ——(1991) ‘Firm Resources and Sustained Competitive Advantage’, Journal of Management 17:99– 120. ——(1995) ‘The Resource Based View: Evolution, Current Status, and Future’, handout for a presentation, Third International Workshop on Competence-Based Competition, Ghent, Belgium, November, 1995. Casson, Mark (1997) Information and Organisation, Oxford: Oxford University Press. Caves, Richard E. (1980) ‘Industrial Organization, Corporate Structure, and Strategy’, Journal of Economic Literature 18:64–92. Chandler, Alfred D. (1962) Strategy and Structure, 1990 edn, Cambridge, MA: MIT Press. Cohen, Wesley M. and Daniel Levinthal (1990) ‘Absorptive Capacity: A New Perspective on Learning and Innovation’, Administrative Science Quarterly 35:128–52. Cohen, Michael D. and Lee S.Sproull (eds) (1996) Organizational Learning, London: Sage. Conner, Kathleen R. (1991) ‘A Historical Comparison of Resource-Based Theory and Five Schools of Thought within Industrial Organization Economics: Do We Have a New Theory of the Firm?’, Journal of Management 17:121–54. David, Paul. A. (1985) ‘CLIO and the Economics of QWERTY’, American Economic Review 75:332– 7. Demsetz, Harold (1969) ‘Information and Efficiency: Another Viewpoint’, in Demsetz (1989a). ——(1973) ‘Industrial Structure, Market Rivalry, and Public Policy’, Journal of Law and Economics 16:1–10. ——(1974) ‘Two Systems of Belief about Monopoly’, in Demsetz (1989a). ——(1982) ‘Barriers to Entry’, in Demsetz (1989a). ——(1988) ‘The Theory of the Firm Revisited’, reprinted in Oliver E.Williamson and Sidney G.Winter (eds) (1993) The Nature of the Firm, Oxford: Basil Blackwell. ——(1989a) Efficiency, Competition, and Policy, Oxford: Basil Blackwell. ——(1989b) ‘The Indivisibility Rent Theory of Measured Oligopoly Profit’, in Demsetz (1989a). Dierickx, Ingemar and Karel Cool (1989) ‘Asset Stock Accumulation and the Sustainability of Competitive Advantage’, Management Science 35:1504–11. Foss, Nicolai J. (1996a) ‘Whither the Competence Perspective?’, in Nicolai J.Foss and Christian Knudsen (eds), Towards a Competence Theory of the Firm, London: Routledge. ——(1996b) ‘Research in Strategy, Economics, and Michael Porter’, Journal of Management Studies 33:1–24. ——(1997) Resources and Strategy: A Reader, Oxford: Oxford University Press.

28 NICOLAI J.FOSS

——(1998) ‘The Competence-Based Approach: Veblenian Ideas in the Modern Theory of the Firm’, forthcoming in Cambridge Journal of Economics. Foss, N.J., C.Knudsen and C.A.Montgomery (1995) ‘An Exploration of Common Ground: Integrating Evolutionary and Strategic Theories of the Firm’, in C.A. Montgomery (ed.), Resource-Based and Evolutionary Theories of the Firm, Boston, MA: Kluwer Academic. Hamel, Gary and Aimée Heene (1994) Competence-Based Competition, New York: John Wiley and Sons. Kirzner, Israel M. (1973) Competition and Entrepreneurship, Chicago: University of Chicago Press. Knudsen, Christian (1996) ‘The Competence Perspective: A Historical View’, in Nicolai J.Foss and Christian Knudsen (eds), Towards a Competence Theory of the Firm, London: Routledge. Lippman, Stephen A. and Richard P.Rumelt (1982) ‘Uncertain Imitability: An Analysis of Interfirm Differences under Competition’, Bell Journal of Economics 13:418–38. Loasby, Brian J. (1991) Equilibrium and Evolution, Manchester: Manchester University Press. Mahoney, Joseph T. (1995) ‘The Management of Resources and the Resource of Management’, Journal of Business Research 33:91–101. Mahoney, Joseph T. and J.R.Pandian (1992) ‘The Resource-Based View within the Conversation of Strategic Management’, Strategic Management Journal 13:363–80. March, James G. (1991) ‘Exploration and Exploitation in Organizational Learning’, Organization Science 2:71–87. Markides, Constantinos C. and Peter J.Williamson (1994) ‘Related Diversification, Core Competences and Corporate Performance’, Strategic Management Journal 15:149–65. Montgomery, Cynthia A. (1994) ‘Corporate Diversification’, Journal of Economic Perspectives 8:163– 78. ——(1995) Resource-Based and Evolutionary Theories of the Firm, Boston, MA: Kluwer Academic. Montgomery, Cynthia A. and Birger Wernerfelt (1988) ‘Diversification, Ricardian Rents, and Tobin’s q’, Rand Journal of Economics 19:623–32. Nelson, Richard R. and Sidney G.Winter (1982) An Evolutionary Theory of Economic Change, Cambridge, MA: The Belknap Press. Penrose, Edith T. (1952) ‘Biological Analogies in the Theory of the Firm’, American Economic Review 42:804–919. ——(1959 [1995]) The Theory of the Growth of the Firm, Oxford: Oxford University Press. Peteraf, Margaret A. (1993) ‘The Cornerstones of Competitive Advantage: A Resource-Based View’, Strategic Management Journal 14:179–91. Prahalad, C.K. and Gary Hamel (1990) ‘The Core Competence of the Corporation’, Harvard Business Review May–June:79–91. Rubin, Paul (1973) ‘The Expansion of Firms’, Journal of Political Economy 81:936–49. Rumelt, Richard P. (1984) Towards a Strategic Theory of the Firm’, in Richard B. Lamb (ed.), Competitive Strategic Management, Englewood Cliffs, NJ: Prentice Hall. ——(1987) ‘Theory, Strategy, and Entrepreneurship’, in David J.Teece (ed.), The Competitive Challenge, Cambridge, MA: Ballinger. Rumelt, Richard P., Dan E.Schendel and David J.Teece (eds) (1994) Fundamental Issues in Strategy: A Research Agenda, Boston, MA: Harvard Business School Press. Sanchez, Ron (1993) ‘Strategic Flexibility, Firm Organization, and Managerial Work in Dynamic Markets: A Strategic Options Perspective’, Advances in Strategic Management 9:251–91. Scherer, Frederick M. (1980) Industrial Market Structure and Economic Performance, Boston, MA: Houghton-Mifflin. Schulze, William (1994) ‘The Two Schools of Thought in Resource-Based Theory’, Advances in Strategic Management 10:127–52.

EQUILIBRIUM VS. EVOLUTION IN THE RESOURCE-BASED PERSPECTIVE 29

Selznick, Philip (1957) Leadership in Administration, Berkeley, CA: Harper and Row. Spender, J.-C. (1993) ‘Business Policy and Strategy: A View of the Field’, unpublished ms. ——(1994) ‘Organizational Knowledge, Collective Practice and Penrose Rents’, International Business Review 3:353–67. Teece, David J. (1980) ‘Economies of Scope and the Scope of the Firm’, Journal of Economic Behavior and Organization 1:233–47. Teece, D.J. and G.Pisano (1994) ‘The Dynamic Capabilities of Firms: An Introduction’, Industrial and Corporate Change 3:537–56. Teece, David J., Gary Pisano and Amy Shuen (1997) ‘Dynamic Capabilities and Strategic Management’, in Nicolai J.Foss (ed.), Resources, Firms, and Strategies, Oxford: Oxford University Press. Tushman, Michael L. and William L.Moore (eds) (1988) Readings in the Management of Innovation, Cambridge, MA: Ballinger. Weick, Karl E. and Frances Westley (1996) ‘Organizational Learning: Affirming an Oxymoron’, in Stewart R.Clegg, Cynthia Hardy and Walter R.Nord (eds), Handbook of Organization Studies, London: Sage. Wernerfelt, Birger (1984) ‘A Resource-Based View of the Firm’, Strategic Management Journal 5:171– 80. ——(1995) ‘Resource-Based Strategy in a Stochastic Model’, in Cynthia A. Montgomery (ed.), Resource-Based and Evolutionary Theories of the Firm, Boston: Kluwer Academic. Williams, Jeffrey A. (1992) ‘How Sustainable Is Your Competitive Advantage?’, California Management Review 34:29–51.

30

3 Knowledge and capabilities A new view of the firm J.Stanley Metcalfe and Andrew James

Introduction In recent years there has been a remarkable flowering of ideas on the theory of the firm, the principal purpose of which is to break open the black box which normally constitutes the firm and to seek to understand its behaviour in terms of its internal properties, that is to say its capabilities. There are several admirable reviews and commentaries on this literature (Montgomery 1995; Foss and Knudsen 1996; Foss 1997) so we do not intend to repeat their findings here. Rather our purpose is fourfold: to explore the wider significance of the capabilities view in terms of a number of important controversies in relation to the role of firms; to attempt to construct a simple framework in which the main features of the capabilities view can be clearly expressed; to provide some necessary supplements to the capabilities approach in relation to the notions of rationality and knowledge, particularly technological knowledge; and to draw some implications in relation to the role of mergers and acquisitions in enhancing a firm’s capabilities. This last theme is particularly significant for us because it involves a movement away from treating capabilities as if they can be generated only within the firm. We should also note here that we shall speak of the firm and its constituent business units interchangeably, thus sweeping under the carpet important distinctions between capabilities at the corporate and at the business level. For the present, this issue remains unfinished business. The framework within which these ideas are developed is evolutionary, and in the conclusion we summarise briefly our understanding of the link between capabilities and economic evolution. The significance of the capabilities theory of the firm In this section, we propose to review a number of controversial issues on which a capabilities perspective can shed some important light. In turn, these are the distinctive nature of the firm, the sources of competitive advantage and the nature of competition, the theory of endogenous growth, and the institutional systems approach to the development of technology.

32 J.STANLEY METCALFE AND ANDREW JAMES

The nature of the firm There is a long controversy in the economics literature concerning why firms exist and how they differ as organisations from markets as institutions. Typically the contractual perspective which developed from the seminal work of Coase (1937) either explains the existence of firms in terms of the difficulty of writing fully specified contracts between the collaborators in the enterprise, so creating scope for opportunistic behaviour, or emphasises the difficulties in monitoring and evaluating the contributions of the different members of the firm (Alchian and Demsetz 1972). In essence, firms are the consequence of specific market failures in the provision of information which prevent the firm being assembled as a particular, fluid nexus of contracts. The capabilities perspective takes us in a quite different direction. It not only explains why firms and markets are distinctively different, it also explains why firms differ one from another, even in closely related lines of business, something a contracts perspective cannot do. What the capabilities view underpins is the view of the firm as a transformation process not an exchange process: put simply, what is distinctive about the firm is that it determines what is produced and how and to what end (McNulty 1984). Moreover, on this view ‘firms’ become the necessary prerequisites for the existence of markets and, in many cases, firms provide the information on prices as well as product qualities which market institutions then diffuse to potential customers. Indeed, it is perhaps the false perception of the typical economic transaction involving a homogeneous product exchanged between anonymous parties with prices set by extra-market institutions that has done most to promote the view of the firm as merely a substitute for market arrangements. The crucial point to comprehend at this stage is that firms and markets as categories are complements not substitutes. The existence of the one depends on the existence of the other: each is the response to the underlying, fundamental phenomena which are specialisation and the division of labour. It is perfectly sensible to imagine a Crusoe-like economy of entirely self-sufficient individuals who neither barter nor trade, but this is not our world. What we have are markets and independent producers. However, it is also clear that specialisation is not itself sufficient to explain the modern firm. We need something else. This problem is put into sharper focus if we follow McNulty’s suggestion and ask why individuals find it beneficial to collaborate in teams? Why does not everybody form his own individual firm? The capabilities-based answer is that collaboration yields productive outcomes which are more highly valued by their customers compared to the outcomes achieved when individuals act in isolation. In short, the integrated capabilities of teams have been selected for by market processes. While the contractual perspective on the firm produces negative reasons for the existence of firms, the capabilities perspective relies upon positive arguments. When individuals collaborate in teams they do not simply exchange productive services, for such exchanges can be handled by market transactions. Team members, however, interact in such a way as to make available qualitatively superior productive services. As we explore below, this is because the knowledge of the team is greater than the sum of the knowledge of the isolated individuals through interaction: the

KNOWLEDGE AND CAPABILITIES 33

firm is a creator of knowledge which is shared between its members to greater productive effect. Thus firms do not simply use available factor services; they transform these services through more roundabout teamwork (Young 1928) by exploiting the gains in knowledge which come from interaction. It is upon the non-linear combination and transformation of factor services that the capabilities approach ultimately depends, and combination and transformation are not possible without organisation. Indeed this is precisely the point made by Penrose (1959) when she distinguished the resources available to the firm from the productive services which could be extracted via managerial knowledge. Resources are accessible through market contracts, but productive services are not. They are created by activities within the firm. Of course, this still leaves open the question of the boundaries of the firm, and of what is within those boundaries, and what is without and necessarily the subject of market relationships (Richardson 1972). Here a perspective which identifies the comparative costs of competing organisational arrangements certainly has an important contribution to make since market selection will, in the long run, result in the dominance of those forms of enterprise which offer the most favourable cost and quality combinations. But this does not explain the existence of firms. Rather, it contributes to an explanation of the different ways in which they operate, and of the way in which some governance forms acquire dominance by market processes. The capabilities approach and the transactions cost approach to the firm are to us complementary and distinctively different. There are two further aspects of this line of reasoning which should be commented upon. As we shall see below, the capabilities view places special emphasis upon knowledge and the accumulation of knowledge both within the firm and in market arrangements between firms. Now economists have developed good reasons why markets in information are unlikely to be efficient (Arrow 1961) and it would be quite natural to see the firm as an institution necessary for the production of specific proprietary knowledge when there is no market alternative. A second aspect of knowledge is that it must be acquired in indivisible amounts; one cannot produce with an incomplete understanding of the production process or sell with an incomplete understanding of the relevant customers. On the one hand this makes it impossible to deduce the marginal contribution of any one individual’s understanding or idea since without the rest of the complementary knowledge its value is undefined. On the other hand, capabilities are created by indivisible investments and their exploitation is always subject to increasing returns. Penrose (1959) understood this well when she drew attention to the twin roles of management in both creating managerial processes and implementing these processes. But creation is a discrete activity: once the new opportunity has been discovered and the associated activities defined, that portion of management effort is free to be deployed in new areas of investigation. In this way the creativity of the firm in developing new lines of activity becomes contingent upon the division of managerial resources between exploration and exploitation (Dosi and Marengo 1994). Moreover, because of increasing returns, there is pressure to make use of the management team’s accumulated stock of intangible assets to the greatest degree consistent with the other constraints on the expansion of the firm.

34 J.STANLEY METCALFE AND ANDREW JAMES

Thus, to Penrose, firms are simultaneously creative and growing. Not surprisingly, a static contracts viewpoint tells us very little about such matters. Competition and competitive advantage If the exploitation of capabilities gives rise to increasing returns, this presents a serious difficulty for the equilibrium theory of competition and the firm. But here the capabilities approach makes its second major contribution for it provides the basis, not for competition as a state of affairs, but for competition as a process of change; a process which depends in evolutionary fashion upon firms being different one from another. This change in perspective is profound. First it requires us to abandon methodological essentialism and treat firms in an industry not as collections of identical entities but as distinct individuals (Nelson 1991; Kogut and Zander 1992). One asks not what they have in common but how they are different one from another. This leads naturally to the question of why they are different, which becomes an issue in the differential development of capabilities (Teece et al. 1994), that is, the differential ability to innovate considered in the broadest terms. All of this connects directly with evolutionary theories of the competitive process, in which differences in the behaviour of firms create particular characteristics or competitive traits which give selective advantage in the marketplace, that is, relative to the requirements of users and the offerings of rival firms. In turn these selection advantages are reflected in the profitability of the firm and the rate at which it can expand its market position relative to its rivals. Increasing returns do not undermine this process, but enhance its operation. What prevents the ultimate concentration of production in a dominant firm is to be found only in the differential capability enhancing efforts of the other rival firms (Downie 1958; Metcalfe 1998). One immediate consequence of this view is to identify competitive advantage not as an attribute of a firm in isolation but as a consequence of the market interaction of rival firms with different competitive traits. Of course, the possible competitive traits are many in number, and they certainly are not reducible to technological differences; organisation of production and organisation of relationships with customers are equally significant. In supporting the dynamic basis of competition, a capabilities perspective can usefully distinguish three broad classes of capabilities: capabilities in relation to the conduct of existing activities; capabilities in relation to the growth of those given activities through investments in productive capability and market position; and capabilities in relation to the development of all the firm’s activities, including the introduction of new products and processes, entry into new market areas, and the entering into of relationships with other firms through alliances, joint ventures or acquisitions. It is the latter capabilities in relation to acquisition that are of particular significance for this chapter. Endogenous growth and innovation systems The third aspect of a capabilities perspective to which we draw attention is the underpinning it provides for theories of endogenous growth and economic policy in

KNOWLEDGE AND CAPABILITIES 35

relation to economic growth. But contrary to the macroeconomic literature on endogenous growth, the perspective shifts to the microeconomic level. Endogenous growth becomes a matter of the accumulation of capabilities in firms and the growth in the economic application of those capabilities. It involves structural change in a fundamental way between broad sectors and within sectors as firms compete for market advantage. We can certainly measure the growth outcome of these processes at the macroeconomic level but we cannot comprehend them at the macroeconomic level. Thus the capabilities perspective provides us with the material to develop an understanding of economic growth very much in the tradition of Adam Smith and Alfred Marshall. One implication of this is also to provide a framework for growth policy. One of the important themes here is that firms do not develop capabilities solely through their own isolated efforts; rather they are connected within a matrix of other capability supporting institutions. This is precisely the perspective which is developed in the innovations systems literature (Nelson 1993; Edquist 1997; McElvey 1996; Lundvall 1992; Carlsson 1995) and it tells us that firms develop their capabilities by interacting with customers, suppliers and other knowledge and skill generating institutions in the public and private sectors. It matters whether or not a firm can engage in discussion with intelligent customers and suppliers and with other high-quality knowledge generating institutions. Indeed a capability to interact in this way is surely a major determinant of long-term competitive advantage. The significance of this for growth and innovation policy is that it is the operation of this wider set of capability supporting institutions that provides the conduit for policy initiatives: to build an appropriate capabilities support infrastructure and to ensure that firms connect with it via networks or whatever becomes the central policy concern. Having sketched several of the fields of enquiry which are illuminated by the capabilities approach to the firm, it is time to explain more clearly its nature. A necessary prelude to this is to consider some of the issues raised by rationality and the accumulation of technological knowledge. Rationality, imagination and innovation One of the issues we must confront at the outset is the relationship between firm behaviour and rational decision making. This is a complex matter and no serious student of the innovation process has ever put much faith in the extreme idea of rationality, which is otherwise so prevalent in modern economy theory. Let us start by accepting a far less restrictive version of rationality as reasoned behaviour, the directed, intentional behaviour of firms seeking competitive advantage by committing resources to innovation. Such behaviour is motivated, it involves striving (Winter 1984) and it entails no more than decision making and thus choice which seeks to do the best with the resources at the command of the firm or innovating business unit. The arguments in favour of reasonable rationality as distinct from Olympian rationality are, we believe, compelling, and they help us understand why firms come to develop in such widely different ways. Olympian rationality, it will be recalled, requires a full

36 J.STANLEY METCALFE AND ANDREW JAMES

specification of all the relevant options from which choices are to be made, and a full evaluation of how each option contributes to the objectives of the rational decision maker, including options whose consequences lie entirely in the future. In the field of innovation neither Olympian prerequisite seems remotely plausible: where innovations are concerned neither what is considered best nor the resources considered to be available are free from ambiguity. When problems become too complex to be well defined let alone solved analytically, one is inevitably reduced to judgement and the guiding hand of experience (Lane et al. 1995). Complexity implies that there is no single metric to reduce choices to a comparable basis, that there is computational indeterminacy in translating between options and objectives, and that there is an ineluctable range of ignorance which means that decisions can only be made on the basis of expectation and hunch. To the extent that optimisation occurs at all, i.e., that there is reasonable behaviour, it takes place within local domains in which the choice set and the evaluation of that choice set are historically contingent. Of itself this is quite enough to underpin the diversity of firm behaviour which is the chief empirical characteristic of the development process in modern capitalism. But there is so much more to bounded rationality than the notion that it is simply rationality with extra constraints. The much more fundamental issue is the question of the exercise of imagination and the construction of the choice sets for change. Just what are the options for development? Langlois (1990), for example, along with many others in the Shackle (1961)/Simon (1955) tradition, has pointed out the fundamental problems that arise in compiling the list of potential choices, problems relating to the imagination required to construct the relevant choice set. Here it is important to remember that bounded rationality is too often interpreted negatively, as an unfortunate and unavoidable deviation from comprehensive rationality. This misleads greatly, for it hides the positive side of boundedness, namely the freeing of the imagination to conceive new configurations of knowledge. Liberated from the constraints of received wisdom, one can genuinely innovate (Loasby 1991). This is perhaps why technology so often develops ahead of an understanding of the relevant laws of nature or understanding of human behaviour. Vincenti’s (1997) account of the realisation of supersonic flight in advance of any scientific theory, indeed in the face of theoretical predictions that it might be impossible, is a good case in point. Perhaps we underestimate the extent to which genuine innovation is contingent upon ignorance. Within evolutionary epistemology this issue is closely connected with the question of ‘where guesses come from’ (Stein and Lipton 1989), a problem in the generation of variety and the subsequent selection of options from that range of possibilities (Campbell 1987). Within this literature two themes are particularly important: the processes which give rise to the winnowing of ideas in a pre-adaptive sense; and the limitations on the generation of possible ideas via principles of guided variation. Stein and Lipton place particular emphasis on mechanisms for pre-adaptation, those heuristics and focusing devices which shape the hypotheses that are candidates for testing. They suggest that the heuristics can be of three general types: abstract rules or general principles of inquiry;

KNOWLEDGE AND CAPABILITIES 37

concrete rules, specific ways of generating hypotheses in a discipline context; and exemplars, canonical solutions or concrete models of solutions to specific problems. This emphasis on diversity of innovative behaviour raises the question of whether there exists a true model of the reality of any given innovative context, or whether there are as many divergent representations as there are diverse behaviours of firms. Wilson (1990) tackled this question by distinguishing two approaches: where the generation of ideas is based on some model reality, and when it is based on what he call adaptive imaginary representations or fictional worlds which simplify complex reality and provide a set of instructions (algorithms) on how to behave in particular circumstances. He placed all models of rationality, including bounded rationality, in the first group. However, a fictional world limits and motivates behaviour and should be judged by its operational effectiveness, not by its relation to the truth—in the scientific sense appropriate to a model of reality. Thus if models of reality are ‘scientific’, adaptive imaginary representations are ‘technological’ in that, like engineering knowledge, they are tested by their utility, accumulated by trial and error, and are the epitome of rule-guided behaviour. It seems to us that managerial ideas about what a firm is capable of are adaptive imaginary representations rather than models of reality and in this they equate to technology and engineering not science. History matters because history shapes the knowledge which is extracted from particular items of information. Hence the reason why firms so often get it wrong in their appraisal of new technology or market opportunities is that they appraise them through a distorting mirror, the mirror of their past which in all likelihood stands ready to shatter in the face of significant innovation. In Schumpeter’s theory of innovation (1912), it will be remembered the entrepreneur is precisely distinguished by an ability to take leaps into the unknown, to break with the past, to act although ignorant of the consequences of action. Brian Arthur (1994) expresses the matter perceptively when he emphasises the resort to inductive reasoning and pattern recognition in the face of complexity. New information is evaluated by reference to previous experience; it cannot usually be evaluated by theory and this gives to the knowledge-generation process an inherently path-dependent character. This is a natural point to turn to the mental frameworks which guide our innovative imagination, and in particular to these frameworks in relation to the accumulation of technological knowledge. The structure of the firm’s knowledge In a fundamental sense all the activities of the firm are premised upon a certain state of knowledge, which immediately suggests a connection between the capabilities perspective and the way in which recent scholars have come to view the accumulation of practical knowledge. To a substantial degree this literature is concerned with the accumulation of technological knowledge and one of its purposes is to draw out the difference between science and technology in terms of the relative lack of theory dependence of the latter and its reliance upon practical falsification criteria. Now the operational and strategic knowledge of the firm is in this regard similar to technological knowledge. It is dominated

38 J.STANLEY METCALFE AND ANDREW JAMES

by questions of procedure and the appropriate tests are strikingly simple. Either the knowledge results in profitable activities or it does not; the truth of the knowledge in a scientific sense is not the issue. It is thus worthwhile to explore in more detail the nature of technological knowledge and draw the implications for capabilities. The central point about this perspective is the notion of structured knowledge and the associated constraints on the rate and direction in which new knowledge is acquired. For the firm this implies limitations on what it can do and limitations on how it can develop. Here, the idea of technological paradigms first articulated by Dosi (1982) building on Kuhn (1962) is instructive. Put succinctly, a paradigm is a structured set of concepts, facts, and the relations between them. In particular, a technological paradigm constrains the rate and direction of technological change in both positive and negative terms. That is, the paradigm influences the positive steps in certain technological directions, while avoiding steps towards other uncharted waters. It defines an opportunity set, acting as a focusing device which raises productivity in the search for technical improvement precisely because it limits the questions which are asked. The associated trajectories have an ex post existence as a realised pattern of technical development, and an ex ante existence in terms of the beliefs, expectations and intentions which shape the day-to-day activities of technologists and their managers. The general thrust of this paradigmatic perspective is cognitive. To paraphrase Vincenti, it concerns what technologists know and how they come to know it. Such a paradigm indicates fruitful directions for technological change, defines concepts of progress, establishes tests to judge performance and has a powerful exclusion effect on the collective thinking of engineers, technologists and the organisations they represent. A technological paradigm builds cumulatively by suggesting a sequence of puzzles sometimes guided by theory but often solved entirely empirically. A progressive technology generates many performance enhancing puzzles, and in this routine aspect it is akin to normal science. However, technological design and development is more concerned with puzzle solving than with hypothesis testing, more with verifying what works than with theoretical falsification. A technological puzzle is solved when the performance standards of an artefact are improved or become more predictable, not when a puzzle solution yields a better understanding of a natural phenomenon (Vincenti 1990). Moreover, technologies involve practice as well as knowledge, and this is why it is necessary to investigate the development of a technology in three dimensions: the growth of codified knowledge, the acquisition of tacit skills, and the development and application of product and process artefacts (Layton 1974). What Dosi has meant, in more general terms, by technological paradigm applies also to the more focused concept of a normal design configuration (Georghiou et al. 1986). Indeed, a paradigm will normally have within it a range of design configurations for specific groups of closely related artefacts. The paradigm is a broad, aggregate concept, for some purposes too broad to explain effectively exactly what technologists within specific firms focus upon in their day-to-day activities. We therefore prefer the narrower concept of a design configuration which relates to a specific group of artefacts and their

KNOWLEDGE AND CAPABILITIES 39

production process. By a normal design configuration, as applied to engineering technology, we mean the set of fundamental design concepts, i.e., ‘the general shape and arrangement that are commonly agreed to best embody the operational principle’, which in turn implies the understanding of how a device works, how the parts of which a device is made concur together in achieving a wanted purpose (Vincenti 1990:208–9). These lower levels of technology define more precisely the frameworks within which engineers and technologists think. They share the general features of the paradigm, but differ significantly in important details. For example, within the paradigm of electricity generation, there were two competing design configurations, one based upon direct current and the other on alternating current. Sufficiently similar, yet sufficiently different. Likewise, the diesel and petrol engines are different design configurations within the internal combustion engine paradigm and water jet and air jet looms are different design configurations within a paradigm for weaving cloth. Betablockers have a design configuration different from calcium antagonists and ACE inhibitors within the cardiovascular drugs paradigm. We note also that Nelson and Winter have employed the phrase technological regime as an equivalent to technological paradigm, comprising a collection of design configurations each one defining a specific realisation of the technology. One further implication of the piecemeal, trial and error nature of technological development and the imperfect dependence on scientific knowledge is that the limits of any technology are often only poorly understood. Consequently, when the operation of the artefacts strays beyond the bounds of normal operation, the outcome is frequently disastrous and we have, in Constant’s words, a functional failure. Such failures are not predictable in any rational sense but they certainly change the prevailing ‘world’ view of the relevant community of practitioners (Petroski 1994). The closer we get to the firm the more technology is specific; the articulation process, though drawing substance from the same body of latent knowledge, leads to differentiated artefacts. Such a process generally leads to the creation of sets of artefacts characterised by differences in performance and evolving along different paths. Kuhnian scientific paradigms overlap with the concept of technological paradigms. Insights drawn from a given scientific paradigm may underpin development in a number of technological paradigms. Conversely, the development of technological paradigms poses puzzles for the underlying science so that, as a general proposition, science is stimulated by technology and stimulates technology. It is a two-way street (Rosenberg 1990). What is important about this argument is that epistemic variations, which all developments within the firm are at some point, while they may necessarily be blind in that their economic performance can only be anticipated, are not random (Campbell 1987). The mental frameworks within which innovative ideas arise are heavily shaped by past experience and by the limits imposed by the particular sets of heuristics in play. The search for novelty cannot be random, for if it were it would fall victim to the tyranny of combinatorial explosion, there being too many possible combinations of human ideas to imagine and evaluate. If new ideas are to come, they come because search is directed to

40 J.STANLEY METCALFE AND ANDREW JAMES

limited areas of the set of possibilities. In this sense innovations are never entirely novel, but are always prefigured in some of their dimensions. What is true of technology, we suggest, is applicable in a broader sense to the knowledge which defines the productive opportunity of the firm in its many dimensions. Technology, in the broad sense of production and use, is, of course, a major component of that opportunity, and hence the set of productive services that the firm develops may be expected to have the features of design configurations as outlined above. But the same is true of the firm’s knowledge of its markets. Hence, we suggest, what is distinctive about each firm is its particular theory of business, those interrelated concepts, facts and expectations which constitute the knowledge base of the firm. When interpreted as the firm’s design configuration, they provide some interesting insights. The development of the firm’s knowledge proceeds within bounds. Variation is guided variation, as the firm is not capable of comprehending all changes in its environment. Rather, its comprehension is localised since the firm is to a considerable degree locked into its view of the world, from which escape may be very difficult. Perhaps its beliefs are its most important category of sunk capital, with all that implies for the willingness and ability to change. After all, economists would not normally consider that they could with any facility transform themselves into surgeons. Why should it be any different for specialised organisations? This is to say not that radical change does not occur within the firm, but rather that it is unusual and that it is attempted more often than it is achieved. At this point the link with capabilities becomes clear for, as we suggest, the capabilities of any business unit are formed around its particular set of theories of business. It provides the basis for a shared view of the world and provides the lens through which that world is perceived by the firm. As with all cognitive frameworks, this permits progress by limiting progress; by illuminating some avenues of development, others are necessarily closed off, as Loasby has persuasively argued (Loasby 1991). The capabilities approach Having outlined relevant aspects of bounded rationality and the growth of knowledge, we turn to a knowledge-based view of the firm. The firm is defined by what it knows; its knowledge determines what it can do and how. In recent literature, this has become known as the capabilities view of the firm and it is of considerable interest in any study of the sources and conduct of innovation-related activities, for the essence of a change of capability is that it requires a change of knowledge within the firm (Kogut and Zander 1992). However, within this burgeoning literature there exists a good deal of conceptual ambiguity to which we now turn. In the following sections, we make a number of broad distinctions which can be summarised by the statement that the market performance of the firm depends upon the combination of its capabilities with its strategic objectives or intent. That is to say, given capabilities can be deployed in a variety of ways; they are a latent basis for action. We see capabilities as Penrosian bundles of productive services, derived from resources, and

KNOWLEDGE AND CAPABILITIES 41

articulated by routines in specific organisational contexts, formal and informal. Let us explore these concepts in more detail. Resources Several recent contributions to the management literature have focused upon the differences in the resources commanded by different firms. Barney (1991), for example, emphasises the importance of firm resource heterogeneity as a basis for competitive advantage and distinguishes in broad fashion between the physical, human and organisational resources of the firm. Many of these resources can be acquired or disposed of in market transactions, and therefore cannot be a basis for differential advantage. To be the basis of differential advantage, the bundle of resources must be deployed to create differences in effectiveness and efficiency, there must be no strategic substitutes available to rivals, and the resource bundle must not easily be imitated. What is it that transforms separable marketable inputs into differentiated, unique and integrated resource bundles? The answer is the knowledge of the productive opportunity at the firm’s disposal because it is this knowledge that is the key to capability. What is it that sustains the uniqueness of a particular capability? Imperfect mobility in imperfect resource markets is one reason for imperfect imitability, but more significant is the complexity of the knowledge links between particular resource bundles and competitive advantage. These give rise to observational difficulty and causal ambiguity (Dierickx and Cool 1989), as reinforced by the social complexity of the resource bundles when viewed internally, in terms of organisational structures, and externally, in terms of customer supplier linkages. Amit and Schoemaker (1993) have explored similar themes but they emphasised the imperfect tradability of those resources, particularly knowledge, which give strategic advantages and the consequence that they must be accumulated internally. This, together with the specialised nature of the intangible assets, means that the latter are sunk and open to sudden obsolescence, and that they are a basis for inertia in responding to new opportunities. Of course, this is one of the implications of the paradigm view of knowledge—that it accumulates within boundaries and is not infinitely malleable. It is clear from this resource-based perspective that a great deal of the competitive value of knowledge depends on matters of organisation and activity. Organisation Consider now the matter of organisation, the firm as a collection of skilled individuals, each with his own knowledge, aptitude and motivation, brought together in teams which are focused around activities (Penrose 1959; Richardson 1972). These activities may be grouped in different functional ways—production, research and development, marketing, etc.—but for our purposes it is useful to define them in relation to three broad groupings, concerned respectively with the operation of current activities, the growth or replication of current activities, and the development of qualitatively different operations through innovation.

42 J.STANLEY METCALFE AND ANDREW JAMES

Of course, any effective team is much more than simply the sum of the individuals contained within it and the knowledge and skills that they possess as individuals. First, the individuals operate with the help of other complementary productive assets (Teece 1987) and, second, the individuals are co-ordinated in that they work jointly on tasks. The organisation of the firm creates its distinctive signature by coordinating the internal division of labour and, equally importantly, by transforming the knowledge and skills of the individual members into a collective, unitary capability, which McKelvey (1982) has suggestively called a competence pool. A moment’s reflection on the nature of any team sport will illustrate the point perfectly, and reinforce the idea that an important management task is the effective choice of individuals and motivation and allocation to teams. The organisation is the framework for learning and creativity that further forms the context in which any new knowledge is acquired. In terms of our previous discussion, it is the framework within which the firm’s distinctive adaptive imaginary representations of its world emerge. The same teams of individuals placed in different organisational context are expected to know different things, while a change in the individual members of the organisation may equally be expected to change what it knows and how it learns. Thus what individuals know and can achieve depends on their context. In this way we like to sum up the organisation as an operator transforming the individual into the collective; it is a device for coordinating the division of labour in the generation of knowledge and skill, a device which gives the firm or business unit its individual and indissoluble character. If one accepts this argument about organisation, one is of course in danger of underemphasising the role of the individual. This would be a mistake, since firms are composed of individuals and indeed any serious discussion of entrepreneurship emphasises that point (Hughes 1971; Douglas 1987; Millard 1990). So this line of argument is meant not to deny the importance of the individual but rather to emphasise that individuals always operate in an organisational context. However distinctive the individual may be, context is still important. Clearly, in closely coupled organisations there is far less scope for individual creative behaviour than in more open, loosely coupled organisations. It is a familiar theme in the management literature as to where one draws the balance, close coupling being beneficial for efficiency but loose coupling being beneficial for innovation and development (Loasby 1991). Routines The third strand of the capabilities approach is to explore the nature of the distinctive organisational resource or organisational operator in terms of its sets of decision rules or routines (Dosi and Marengo 1994). Each routine relates to some task, cognitive or physical, within a specific activity, and provides a template, the instructions for action according to the circumstances prevailing (Nelson and Winter 1982). But routines do not operate in isolation and the important point is that the firm’s behaviour depends upon the interaction between its bundle of routines. Capability is something which relates to the

KNOWLEDGE AND CAPABILITIES 43

integration and joint operation of routines. Hence it is not usually possible to link specific behaviours and competitive characteristics to single routines. Arrow (1974) has usefully developed this idea by specifying the firm’s internal language or communication code as the element which shapes the integration of routines. Who talks to whom with what frequency on what topics and with what authority, both internally and externally, with suppliers, customers and the wider community with which the firm interacts, is a rather simple-minded way of specifying this integrative function of an organisation. However, no organisation is simply reducible to its set of routines. For one thing, routines vary considerably in the specificity of the action they stimulate and in the nature of the information by which they are triggered. Certain routines may brook no deviation from a performance template and be entirely automatic in operation, being invoked purely by habit. Other routines may admit a great deal of interpretative flexibility so that the outcome varies with the individual or team performing the routine, and the other teams with which they interact. This is particularly likely with respect to the sets of highlevel routines, such as those which guide innovative activities, since their purpose is to encourage and accommodate creativity. Equally important is the fact that routines do not denote purpose or intent, the strategic objectives of the firm. Strategic intention Of themselves, capabilities are inert. What brings them to life is the sense of purpose embodied in the firm’s theory of business. This is a view which has been stressed by Prahalad and Hamel (1990) when they emphasise the importance of strategic intent and the failure of a resource-based perspective to account for differences in firm performance which reflect, they argue, a mismatch between resources and ambitions. In short, productive opportunities are conjectured connections between capabilities and intentions (Loasby 1991, 1994), while capabilities are organised packages of resources and routines. The capabilities in relation to any given bundle of resources are contingent on the firm’s purpose, as embodied in its theory of the business opportunity and in its theories of how to articulate that opportunity. But capabilities are not static. They are accumulated, learnt by working with the design configuration, by exploring the opportunities that it represents. They evolve; they are joint products of the firm’s activities with its proper outputs. Consequently, the relation between the knowledge of the firm and its capabilities is double-edged: on the one hand knowledge of what to do and how to do it defines capability, while on the other, the execution of a capability in the normal course of business develops the knowledge base of the firm. Hence all business units are specialised and the source of that specialisation is their design configuration. Moreover, their patterns of development arise within the localised context of a specific theory of business, each associated with its own set of practices and standard solutions to problems and each with its own external support system of users, suppliers and other practitioners within an appropriate network.

44 J.STANLEY METCALFE AND ANDREW JAMES

Differential and quasi-sustainable rents One of the central themes of the capabilities literature is that differences between firms in their behaviours are a basis for generating differential and quasi-sustainable rents (Peteraf 1993). Often this is treated in a wholly static context in terms of the origins of differential earnings and the reasons they are not eliminated by market arbitrage. However, this hides a more fundamental dynamic implication of differential capabilities, namely the uses to which rents can be put to sustain and enhance competitive advantages, giving rise to the possibility of virtuous circles of firm development. This is one reason why it is important to distinguish between those capabilities which generate the current rents and those which develop that rent-earning capacity. Much of the literature on capabilities (e.g., Dierickx and Cool 1989) emphasises a link between rent-earning capacity and the lack of tradability of the collective capability. Individuals may come and go, physical assets may be acquired or sold, but given ownership, the unitary capability is not traded. One thing this tells us is that the organisation must possess a memory, or tradition of practice embodied in its routines, so that losses in personnel can be matched with new employees who can be trained in the firm’s routines. Indeed Penrose pointed out that one of the major constraints on the rate of growth of the firm is the rate at which it can absorb and train new staff of all kinds, that is to say, the rate at which they can be integrated into the capability. Thus an individual may be difficult to replace when she leaves one organisation and of much lower effectiveness when working in another organisation. The importance of context again. Now we want to emphasise that it is with respect to the knowledge of the firm that non-tradability becomes particularly significant. For well-known reasons, markets in knowledge, be it managerial or technological, are highly imperfect and this reinforces the bias towards the growth of capability through internal accumulation. The tacitness of much of the knowledge, its indivisibility in use, the uncertainty of its values in different contexts, its proprietary nature, and the fact that much of what is known is jointly produced by the firm’s activities, and indeed decays if the activity ceases, means that firmspecific capabilities cannot easily be augmented through trade and market relationships. This does not mean that the firm does not gather information from external sources. Quite the contrary, it will typically maintain a multiplicity of external sources, some arising in the normal course of trade while others reflect contractual commitments to generate new information. Rather it means that information is not knowledge and only becomes knowledge through the firm’s internal transformation processes. That is to say, the transformation of information into knowledge and action depends on organisation. To repeat, this is where uniqueness of capability resides: competing firms operate in similar contexts, they receive similar flows of information, yet they know different things about their world. In short, the foundation of the capabilities perspective is the firm as a generator of differential knowledge (Teece et al. 1994).

KNOWLEDGE AND CAPABILITIES 45

Merger, acquisition and the development of capabilities However, there is an important qualification to this argument, for collective capabilities can be traded. While they may be extremely difficult to imitate and while they may cease to function if broken into their constituent elements, nonetheless they can be traded via the purchase of the embodying business units (Teece and Pisano 1994). This provides us with new insights into the market for corporate control which is so often presented as a negative discipline on wayward managers. Instead we can see mergers and acquisitions as an integral aspect of the process of business experimentation and capability generation. Since this is a risky activity, this market also provides a route for the disposal of failed attempts to build capabilities, or, indeed, successful attempts which no longer fit with the firm’s strategic intent. Through merger and acquisition, firms seeking to build particular capabilities may obtain them more quickly through the market for corporate control. Acquisition of another firm or one of its business units may allow a firm to access capabilities that are new to that firm. It may allow the internalisation of capabilities previously accessed through external means such as collaborations or joint ventures. Equally, ownership of particular capabilities may be used to prevent competitors accessing particular capabilities, thereby allowing the exploitation of capability rents. However, from our previous discussion it will be clear that this route to capability building has its own difficulties and dangers. From this perspective an acquisition is more than simply the transfer of ownership of a collection of physical artefacts and codified technical information. Rather, it brings together the routines and distinctive capabilities of two teams of people generated by very different managerial and organisational processes which have been built up over time in response to the unique histories of those organisations. These capabilities are deeply rooted in values and norms associated both with the firm’s technological knowledge and with its goals and reasons for existence—its strategic intent (Leonard-Barton 1992; Hamel and Prahalad 1994). As we have seen, the capabilities literature emphasises the complex, context-dependent and causally ambiguous nature of capabilities and the tacit and social character of productive knowledge (Foss 1996). These characteristics of capabilities have implications when they are traded in the market for corporate control and raise interesting questions for firms about how to identify and value what is being acquired, how to integrate the capabilities of two organisations after acquisition, and how to align the acquired capabilities with the strategic intent of the new owner. We focus here on three classes of problems related to the nature of capabilities which arise during the acquisition process. The causally ambiguous nature of capabilities The organisation-specific, tacit and causally ambiguous aspects of capabilities and the routines that generate them have implications for the identification of suitable acquisition targets, their valuation and the conduct of due-diligence. Much of the knowledge that

46 J.STANLEY METCALFE AND ANDREW JAMES

underlies a firm’s managerial and organisational processes is tacit knowledge of the organisation and of the complexity of the links between routines and outcomes, which for the external observer may give rise to observational difficulty and causal ambiguity (Lippman and Rumelt 1982; Dierickx and Cool 1989; Metcalfe 1996). What features of an organisation are responsible for its success and failure may be far from clear, may well not be fully understood—even by the organisation itself—and, consequently, may be rather difficult to identify and evaluate from outside. Typically one must rely on outcome measures which may indicate little about the nature and operation of the underlying capabilities. Such causal ambiguity means that the process of search and evaluation is likely to be characterised by asymmetric knowledge and uncertainty. As a consequence, a key issue for management during the acquisition process is how to manage the associated risks. The fragility of organisational knowledge and routines Organisational knowledge and the management and organisational processes that underpin it can be fragile and, where poorly handled, post-acquisition management can destroy the very attributes that the company has sought to acquire. If, as Foss (1996) suggests, much productive knowledge not only is tacit but is an emergent aspect of the interaction of many agents, then this has important implications for the acquisition process. In acquiring a business unit, the new owner is acquiring its routines for generating new technological knowledge. Two important points arise here that are related to the interdependencies among activities and the agents participating in those activities. First, some of these routines relate to internal knowledge generation and may be firm-specific. The business unit’s relationship with other parts of the firm (for instance, the production department of another business unit within the firm) may be important to its innovative capacity, and acquisition of the single business unit rather than the whole company may cause such linkages to be lost. Second, the generation of new technological knowledge is likely to involve a class of routines related to relationships with suppliers, universities, competitor companies, professional associations and other external sources of knowledge. Linkages with external (and indeed internal) sources of technological knowledge may be informal and often specific to individuals. Such knowledge-generating routines are likely to be fragile and, at least to some degree, the loss of key individuals during the acquisition process may destroy or reduce the effectiveness of such routines. If acquisition of the target’s distinctive capabilities is important, then the acquirer must seek to protect and enhance those distinctive processes if the acquisition is going to realise its potential. Where poorly handled, the post-acquisition stage can destroy the very capacity that the acquirer has sought to buy. Differences between acquirer and target The distinctiveness of firms raises interesting questions about how to integrate the capabilities of two organisations after acquisition and how to align the acquired capabilities

KNOWLEDGE AND CAPABILITIES 47

with the strategic intent of the new owner. A firm’s distinctive way of looking at the world represents a powerful source of interfirm differences. This ‘cognitive framework’ is part of the organisation’s taken-for-granted reality and may go beyond surface values to ‘preconscious’ and ‘invisible’ basic assumptions about the nature of the organisation, its innovation process and its business environment (Schein 1984; Leonard-Barton 1992). The cognitive framework is based on shared history and experience and on the accumulated decisions and events of its corporate history. As such, the distinctive identity of firms may not change easily and the bringing together of such organisations may be the source of tension and conflict after acquisition. The causally ambiguous nature of capabilities means that the relative value of firm distinctiveness may be highly contested, not least when post-acquisition changes threaten the status quo. How the acquirer and the acquired respond may depend very much upon their self-image and the value they place on retaining their distinctive identities. Clearly, if acquisition of the target’s distinctive capabilities is important, then the acquirer must seek to protect the distinctiveness of the acquired organisation rather than to impose its distinctive style. This also suggests that acquisition may require changes in the routines and cognitive framework of the acquirer as well as of the acquired organisation if the acquisition is to be successful. From a capabilities perspective, the weight placed on the importance of capturing distinctive capabilities and organisation style is an important factor of the post-acquisition management approach adopted by the acquirer. In this respect, the way that a company chooses to manage the business that it has acquired reflects certain assumptions about the feasibility and desirability of combining capabilities (James 1997). All of this suggests that the way in which the integration of capabilities is approached will be the key to the outcome of a mergers situation. Finally, it will be apparent from this discussion that the valuation of marketable business units is a non-trivial task. Granting for the sake of argument that the capital market is efficient, it still only follows that all that is valued is the income stream associated with the current managerial team. The capability is not valued directly and so it is not difficult to see the relevance of the literature which identifies a presumption of poor post-acquisition performance, a presumption that only the target shareholders will gain, and a presumption that targets will be overpriced (Gaughan 1994). The valuation of targets is a matter of judgement, judgement which may be more seriously flawed the more dissimilar are the capabilities and theories of business of the two firms, and the less the acquirer factors into its calculations the possible costs of integrating the two sets of capabilities. Given an efficient market, it is obvious that the purchase of a target will result in its rents being translated into capital costs for the acquiring firm. Hence, rent-generating capacity can only be sustained if the acquirer believes it can improve the target’s capabilities over and above those of the existing management. Thus, the root cause of mergers and acquisitions is to be found in beliefs that the capabilities of the acquirer can improve the capabilities of the target or establish new capabilities for the joint enterprise to an extent which is not public knowledge. Once again we see the significance of tacitness, complexity and causal ambiguity.

48 J.STANLEY METCALFE AND ANDREW JAMES

Concluding remarks Where does this discussion fit? We have argued that the capabilities perspective on the firm is distinguished by its relation to the dynamics of knowledge accumulation in market economies. From this foundation it can make important contributions to the theory of economic growth, to the theory of competition and competitive advantage, and to the theory and practice of innovation policy. In respect to this latter point we believe that it is important to recognise that capabilities are not developed solely in isolation, but depend on the firm’s interaction with the wider world of which mergers and acquisitions provide one important example. However, the deeper significance of this approach lies in its bearing upon the evolutionary dynamics of economic change. Economic evolution depends upon the continued generation of novelty and thus upon the continued development of the capabilities of firms and the effective selection of the differences between firms by market processes. It is here that the capabilities approach has so much to offer in helping to understand how firms come to be different and how those differences are developed differently through time. How capabilities matter thus depends on the selective properties of markets and, as we claimed at the outset, markets and firms are complements, not substitutes. Acknowledgements This chapter has two antecedents. It is a substantial development of the Honeywell/ SWEATT Lecture delivered at the Centre for Technological Leadership at the University of Minnesota in April 1996 (Metcalfe 1996), and it builds upon joint work with Luke Georghiou on technology strategy in mergers and acquisitions. We are grateful to the support of the Engineering and Physical Sciences Research Council via its Technology Management Initiative, to Rod Coombs for useful comments and to Sharon Dalton for substantial help in preparing the draft. References Alchian, A. and H.Demsetz (1972) ‘Production, Information Costs and Economic Organisation’, American Economic Review 62:777–95. Amit, R. and P.J.H.Schoemaker (1993) ‘Strategic Assets and Organizational Rent’, Strategic Management Journal 14:33–46. Arrow, K. (1961) ‘Economic Welfare and the Allocation of Resources to Invention’, in R.Nelson (ed.), The Rate and Direction of Inventive Activity, New York: NBER. ——(1974) The Limits of Organisation, New York: Norton. Arthur, W.B. (1994) ‘Inductive Reasoning and Bounded Rationality’, American Economic Review 84: 406–11. Barney, J. (1991) ‘Firm Resources and Sustained Competitive Advantage’, Journal of Management 17: 99–120.

KNOWLEDGE AND CAPABILITIES 49

Campbell, D.T. (1987) ‘Blind Variation and Selective Retention in Creative Thought as in Other Knowledge Processes’, in G.Radnitzky and W.Bartley (eds), Evolutionary Epistemology, Theory of Rationality and the Sociology of Knowledge, New York: Open Court. Carlsson, B. (ed.) (1995) Technological Systems and Economic Performance, Dordrecht: Kluwer Academic. Coase, R. (1937) ‘The Nature of the Firm’, Economica 4:386–405. Dierickx, I. and K.Cool (1989) ‘Asset Stock Accumulation and Sustainability of Competitive Advantage’, Management Science 35:1504–14. Dosi, G. (1982) ‘Technological Paradigms and Technological Trajectories’, Research Policy 11:147– 62. Dosi, G. and L.Marengo (1994) ‘Some Elements of an Evolutionary Theory of Organizational Competencies’, in R.W.England (ed.), Evolutionary Concepts in Contemporary Economies, Ann Arbor: University of Michigan Press. Douglas, S.J. (1987) Inventing American Broadcasting 1899–1922, Baltimore: Johns Hopkins University Press. Downie, J. (1958) The Competitive Process, Duckworth, London. Edquist, C. (1997) Systems of Innovation: Technologies, Institutions and Organizations, London: Pinter. Foss, N.J. (1996) ‘Capabilities and the Theory of the Firm’, Revue d’Economie Industrielle 77:7–28. ——(1997) Resources, Firms and Strategies, Oxford: Oxford University Press. Foss, N.J. and C.Knudsen (eds) (1996) Towards a Competence Theory of the Firm, London: Routledge. Gaughan, P. (1994) Readings in Mergers and Acquisitions, London: Blackwell. Georghiou, L., J.S.Metcalfe, M.Gibbons, T.Ray and J.Evans (1986) Post-innovation Performance, London: Macmillan. Hamel, G. and C.K.Prahalad (1994) Competing for the Future, Cambridge, MA: Harvard Business School Press. Hughes, T.P. (1971) Elmer Sperry, Baltimore: Johns Hopkins University Press. James, A.D. (1997) ‘Effective Management of Technology in Mergers and Acquisitions—A Capabilities Perspective’, in D.F.Kocaoglu, T.R.Anderson, K.Niwa, D.Milosevic and M.J.Gregory (eds), Innovation in Technology Management: The Key to Global Leadership, Proceedings of the Portland International Conference on Management of Engineering and Technology, Portland, Oregon, USA, July 27–31, 1997. Kogut, B. and U.Zander (1992) ‘Knowledge of the Firm, Combination Capabilities and the Replication of Technology’, Organization Science 3:383–96. Kuhn, T. (1962) The Structure of Scientific Revolutions, Chicago: University of Chicago Press. Lane, D., F.Malerba, F.Maxfield and L.Orsenigo (1995) ‘Choice and Action’, Santa Fe Institute (Working Paper 95–01–004). Langlois, R.N. (1990) ‘Bounded Rationality and Behaviourism: A Clarification and a Critique’, Journal of Institutional and Theoretical Economics 146:691–5. Layton, E.T. (1974) ‘Technology as Knowledge’, Technology and Culture 15(1):31. Leonard-Barton, D. (1992) ‘Core Capabilities and Core Rigidities: A Paradox in Managing New Product Development’, Strategic Management Journal 13:111–25. Lippman, S. and R.Rumelt (1982) ‘Uncertain Imitability: An Analysis of Inter-Firm Differences in Efficiency under Competition’, Bell Journal of Economics, Autumn: 418–38. Loasby, B.J. (1991) Equilibrium and Evolution: An Exploration of Connecting Principles in Economics, Manchester: Manchester University Press. ——(1994) ‘The Organization of Knowledge and the Organization of Industry’, mimeo, University of Stirling. Lundvall, B.A. (1992) National Systems of Innovation, London: Pinter.

50 J.STANLEY METCALFE AND ANDREW JAMES

McKelvey, M. (1996) Evolutionary Innovations, Oxford: Oxford University Press. McKelvey, W. (1982) Organizational Systematics, Berkeley: University of California Press. McNulty, P. (1984) ‘On the Nature and Theory of Economic Organization: The Role of the Firm Reconsidered’, History of Political Economy 16:223–53. Metcalfe, J.S. (1996) Technology Strategy in an Evolutionary World, Honeywell/ SWEATT Lectures in Technological Leadership, University of Minnesota, 11 April. ——(1998) Evolutionary Economics and Creative Destruction, London: Routledge. Millard, A. (1990) Edison and the Business of Innovation, Baltimore: Johns Hopkins University Press. Montgomery, C. (1995) Resource-Based and Evolutionary Theories of the Firm, Dordrecht: Kluwer Academic. Nelson, R. (1991) ‘Why Do Firms Differ, and Why Does It Matter?’, Strategic Management Journal 14:61–74. Nelson, R. (ed.) (1993) National Systems of Innovation: A Comparative Study, Oxford: Oxford University Press. Nelson, R., and S.Winter (1982) An Evolutionary Theory of Economic Change, Cambridge, MA: Harvard University Press. Penrose, E. (1959) The Theory of the Growth of the Firm, Oxford: Basil Blackwell. Peteraf, M.A. (1993) ‘The Cornerstone of Competitive Advantage: A Resource Based View’, Strategic Management Journal 14:179–91. Petroski, H. (1994) Design Paradigms, Cambridge: Cambridge University Press. Prahalad, C.K. and G.Hamel (1990) ‘The Core Competence of the Corporation’, Harvard Business Review May–June: 79–91. Richardson, G.B. (1972) ‘The Organization of Industry’, Economic Journal 72:883–96. Rosenberg, R. (1990) ‘Why Firms Do Basic Research (With Their Own Money)’, Research Policy 19: 165–74. Schein, E.H. (1984) ‘Coming to a New Awareness of Organisational Culture’, Sloan Management Review, Winter:3–15. Shackle, G.L.S. (1961) Decision, Order and Time in Human Affairs, Cambridge: Cambridge University Press. Schumpeter, J.A. (1912; Eng. trans. 1961) The Theory of Economic Development, New York: Oxford University Press (first English translation 1934). Simon, H.A. (1955) ‘A Behavioural Model of Rational Choice’, Quarterly Journal of Economics 69:99– 118. Stein, E. and P.Lipton (1989) ‘Where Guesses Come From: Evolutionary Epistemology and the Anomaly of Guided Variation’, Biology and Philosophy 4:33–56. Teece, D.J. (1987) ‘Capturing Value from Technological Innovation: Integration, Strategic Partnering and Licensing Decisions’, in H.Brooks and B.R.Guide (eds), Technology and Global Industry, Washington DC: National Academy Press. Teece, D.J. and G.Pisano (1994) ‘The Dynamic Capabilities of Firms: An Introduction’, Industrial and Corporate Change 3:537–56. Teece, D.J., G.Pisano and A.Shuen (1994) ‘Dynamic Capabilities and Strategic Management’, Strategic Management Journal 18:509–33. Vincenti, W.G. (1990) What Engineers Know and How They Know It, Baltimore: Johns Hopkins University Press. ——(1997) ‘Engineering Theory in the Making: Aerodynamic Calculation “Breaks the Sound Barrier”’, Technology and Culture 38:819–51. Wilson, D.S. (1990) ‘Species of Thought: A Comment on Evolutionary Epistemology’, Biology and Philosophy 5:37–62.

KNOWLEDGE AND CAPABILITIES 51

Winter, S. (1984) ‘Schumpeterian Competition in Alternative Technological Regimes’ in R.Day and G.Eliason (eds), The Dynamics of Market Economies, Amsterdam: North Holland. Young, A. (1928) ‘Increasing Returns and Economic Progress’, Economic Journal, 28: 527–42.

52

4 Synthesising resource-based, evolutionary and neoclassical thought Resource-advantage theory as a general theory of competition Shelby D.Hunt

Introduction Consider the two versions of the contractual theory of the firm. In the nexus of contracts view, one finds maximising managers who face moral hazard. Therefore, firms result from the incentive problems that arise when team-production is combined with opportunism (Jensen and Meckling 1976). In the incomplete contracting view, one finds boundedly rational managers who face opportunism. Therefore, firms arise from efforts to secure the rents that flow from transaction-specific assets (e.g., Williamson 1985). As Foss (1993) points out, however, both contractual perspectives imply a static analysis in which ‘technology—as well as the menu of inputs and outputs more generally—is given through some process that is historically and logically prior to the issue of the organisation of economic activities’ (p. 131). Indeed, he notes that taking the transaction as the analytical unit and viewing ‘the firm’ as nothing more than shorthand for a set of contracts among transacting individuals result in the firm’s strategy as a meaningless concept and ‘it comes as no big surprise that…nexus of contracts theorists…call for the abandonment of the concepts of “the entrepreneur” and “the firm”, respectively’ (p. 130). In contrast, Foss (1993) argues for an evolutionary, competence perspective, which views the firm as a repository of tacit knowledge. Associated with the works of Demsetz (1988), Klein (1988), Langlois (1986) and Prahalad and Hamel (1990), Foss (1993) traces the competence perspective to the long-neglected work of Penrose (1959). She not only maintained that firms are pools of intangible resources but, with prescience, pointed out: ‘Businessmen commonly refer to the managerial group as a “team” and the use of this word implies that management in some sense works as a unit’ (Penrose 1959:15). Building on Penrose’s (1959) seminal work, Carlsson and Eliasson (1991), Eliasson (1990a, 1990b), and Pelikan (1988, 1989a, 1989b) have developed the competence perspective. Following Pelikan’s (1989b) treatment, Foss (1993:134) characterises the epistemic and economic content of the concept of competence: 1 competence is an information capital which tells its owner how to understand and use information to solve economic problems, how to make economic decisions and how to expand existing competence further;

54 SHELBY D.HUNT

2 it is agent-specific and it is tacit; 3 it is distributed asymmetrically over individuals and firms; and 4 stocks of economic competence are difficult to measure and compare with each other. From this perspective, what gives birth to firms and determines their boundaries? Foss (1993) argues that it is the idiosyncrasy of entrepreneurship: ‘Utilisation of entrepreneurial competence…requires a firm with the entrepreneur as residual claimant’ (p. 136, italics in original). Furthermore, the tacit nature of competence implies that firms integrate because, at times, ‘it is practically impossible for…integrating firms to convey—at least at reasonable cost—information to their suppliers about what precisely they want from them’ (p. 138). Therefore, ‘integration arises from the largely noncommunicable, idiosyncratic, and non-contractible nature of competence’ (p. 139). Langlois and Robertson (1995) further develop the competence perspective in their evolutionary theory of firm capabilities. Identifying the capabilities of an organisation or network of organisations as the repertoire of routines that an organisation (or network) possesses, they contend: ‘Those business organisations that can create and utilise superior capabilities tend to perform better’ (p. 3). For them: Our basic argument is that firms and other types of organisations consist of two distinct but changing parts. The first part, the intrinsic core, comprises elements that are idiosyncratically synergistic, inimitable, and noncontestable. That is, the capabilities in the intrinsic core cannot be duplicated, bought, or sold, and they combine to generate unique outcomes that are more valuable than the outcomes that the core elements could produce separately. The remainder of the organisation consists of ancillary capabilities that are contestable and may not be unique. (p. 7, italics in original) Langlois and Robertson (1995) point out similarities between their theory and the resource-based view in strategic management. Indeed, ‘although influenced by the work of economists, the capabilities approach finds a more welcome home in the literature of strategic management’ (p. 5). Foss (1993) also points out that the evolutionary, competence perspective parallels, in many respects, strategic management’s resourcebased theory of the firm. Even though many resource-based works seem, to him, ‘to be more associated with University of Chicago thinking…which basically saw industry structure as reflecting efficiency outcomes’, he believes that ‘the resource-based view of the firm is broadly consistent with evolutionary theory’ (Foss 1993:132). Paralleling the efforts in evolutionary economics, strategic management theorists are developing both a competence view of the firm and a theory of competence-based competition (Prahalad and Hamel 1990; Hamel and Heene 1994; Sanchez, Heene, and Thomas 1996; Heene and Sanchez 1997; Sanchez and Heene 1997). Viewing competence as ‘an ability to sustain the coordinated deployment of assets in a way that helps a firm achieve its goals’ (Sanchez et al. 1996:1), strategic management theorists anticipate that a

SYNTHESISING RESOURCE-BASED, EVOLUTIONARY AND NEOCLASSICAL THOUGHT 55

theory of strategy based on competences would contribute to explaining, among other things, how the continuous renewal of firms’ capabilities determines the changing sources of competitive advantage (Hamel and Heene 1994). Such a new theory of strategy would be part of a theory of competence-based competition that would be dynamic, systemic, cognitive and holistic (Sanchez et al. 1996:11). As do Foss (1993) and Langlois and Robertson (1995), competence theorists in strategic management acknowledge that their view adopts much from the dynamic capabilities and resource-based views of the firm, while at the same time ‘striving to build a conceptual framework capable of addressing key dimensions of strategic competition not adequately addressed by those perspectives’ (Sanchez et al. 1996:1).1 Exploring how evolutionary and resource-based theories of the firm are related, Foss, Knudsen and Montgomery (1995) identify commonalities and differences. In common, both (1) trace their heritage to Penrose (1959), (2) stress the importance of firm diversity, (3) recognise the epistemic content of their key concepts (e.g., routines, capabilities, competencies and resources), and (4) focus on efficiency (rather than market power) to explain firm performance. As to differences, whereas evolutionary theory is processorientated and focuses on heterogeneous routines within firms to explain industrylevel technological evolution and competition, much of resource-based theory is staticequilibrium-orientated and focuses on heterogeneous resources among firms to explain firm-level sources of competitive advantage. Because they believe the two approaches can complement each other, they call for a synthesis of evolutionary and resource-based thinking. A recently developed theory of competition attempts the synthesis called for by Foss, Knudsen and Montgomery (1995). Indeed, this theory, labelled ‘resource-advantage theory’ (hereafter R-A theory), not only adopts the resource-based theory of the firm, but is avowedly evolutionary in nature (Hunt 1995, 1997a, 1997b, 1997c, 1998, 1999a, 1999b, 1999c; Hunt and Morgan 1995, 1996, 1997). Specifically, using Hodgson’s (1993) taxonomy, R-A theory is an evolutionary, phylogenetic, non-consummatory, disequilibrium-provoking, process theory of competition (Hunt 1997a).2 This chapter addresses two questions. First, to what extent is R-A theory’s resourcebased view of the firm consistent or inconsistent with the evolutionary, competence perspective? Is it truly a synthesis? I argue that R-A theory fully accommodates the competence perspective by viewing competences as higher-order resources that are comprised of distinct packages or bundles of basic resources. In brief, R-A theory views competences as socially complex, interconnected packages of tangible basic resources (e.g., specific machinery) and intangible basic resources (e.g., the skills and knowledge of specific employees and specific organisational policies and procedures) that fit coherently together in a synergistic manner. These distinct packages of basic resources enable firms to produce valued market offerings efficiently and/or effectively. Because competences are likely to be heterogeneously distributed among rivals, they can produce marketplace positions of competitive advantage for those competitors that differentially possess them. Because competences are highly immobile, rivals in positions of competitive disadvantage cannot, in general, purchase them in the factor markets. Because competences are tacit,

56 SHELBY D.HUNT

causally ambiguous and socially complex, the time required by rivals to imitate an advantage competitor holding them may be long. Indeed, because competences are comprised of interconnected basic resources, the absence of a key basic resource can thwart competitors’ efforts at imitation. Therefore, the superior financial performance brought about by firms occupying marketplace positions as a result of a comparative advantage in competences can endure through time, despite the concerted efforts of rivals. Second, because both R-A theory and its theory of the firm are evolutionary, how do they relate to such static-equilibrium theories as perfect competition? Drawing on the works of Sellars (1963), Dosi (1991), Foss (1994), Nelson (1995) and Levy (1996), I argue that R-A theory and the competence perspective, rather than being mutually exclusive with perfect competition, incorporate it. In brief, I argue that the structure and foundational propositions of R-A theory represent the descriptively realistic, general case of competition, of which the structure and foundational propositions of perfect competition are a special case. Furthermore, I argue that under certain circumstances (that this article explicates) the process of R-A competition will gravitate towards static equilibrium and, therefore, perfect competition will be ‘close enough’ to predict well. As a consequence, R-A theory explains the predictive successes of perfect competition and preserves the cumulativity of economic science. I begin the analysis with a brief overview of R-A theory. An overview of resource-advantage theory Being evolutionary, R-A theory views (1) innovation and organisational learning as endogenous, (2) firms and consumers as having imperfect information, and (3) institutions and public policy as affecting economic performance. Specifically, firms and resources are proposed as the heritable, durable units of evolutionary selection, with competition for a comparative advantage in resources constituting the selection process. Because the selection process focuses on the locally fitter, not the maximally fittest, R-A theory is non-consummatory and, therefore, accommodates path dependencies (Hunt and Morgan 1996). Thus, though R-A competition is a process that is moving, it is not moving towards some ideal point (such as a Pareto-optimal, general equilibrium). My overview of R-A theory focuses on its pedigree, structure and empirical evidence. The overview follows closely R-A theory’s development in Hunt (1995, 1997b). Pedigree The pedigree of R-A theory traces to several different literatures. First, it traces to the historical tradition (Chandler 1990) and the resource-based theory of the firm (Wernerfelt 1984), which itself traces to Penrose (1959). This theory holds that resources are the tangible and intangible entities available to the firm that enable it to produce efficiently and/or effectively a market offering that has value for some market segment(s). Furthermore, firms are theorised to be combiners of heterogeneous, imperfectly mobile

SYNTHESISING RESOURCE-BASED, EVOLUTIONARY AND NEOCLASSICAL THOUGHT 57

resources that are historically situated in space and time. Consistent with the view in institutional economics (De Gregori 1987; Ranson 1987) that the most important firm resources are intangibles, the ‘resource-based view’ has been significantly developed by Lippman and Rumelt (1982), Barney (1986, 1991, 1992), Dierickx and Cool (1989), Prahalad and Hamel (1990), Reed and DeFillippi (1990), Collis (1991, 1994), Conner (1991), Grant (1991), Peteraf (1993), Barney and Hansen (1994), Black and Boal (1994), Brumagim (1994), Lado and Wilson (1994), Schendel (1994), Schoemaker and Amit (1994), Montgomery (1995) and Madhok (1997). Second, R-A theory draws on marketing’s heterogeneous demand theory (Chamberlin 1933; Alderson 1957, 1965). This theory holds that, because intra-industry demand is significantly heterogeneous, different market offerings are required for different market segments in the same industry. Third, R-A theory draws on competitive advantage theory (Alderson 1957, 1965; Clark 1961; Porter 1985). In this theory, marketplace positions of competitive advantage (disadvantage) determine superior (inferior) financial performance. Thus, firms can have (1) an efficiency advantage, i.e., more efficiently producing value, (2) an effectiveness advantage, i.e., efficiently producing more value, or (3) an efficiency-effectiveness advantage, i.e., more efficiently producing more value. Fourth, R-A theory draws on evolutionary economics (Marshall 1898; Schumpeter 1950; Nelson and Winter 1982; Langlois 1986; Hodgson 1993). Evolutionary economics views competition as a selection process, a struggle. It is this process of competition that produces innovation, creative destruction, increases in productivity and economic growth. Fifth, R-A theory draws on ‘Austrian’ economics (Hayek 1935, 1948; Mises 1949; Kirzner 1979). Stressing the importance of entrepreneurship and economic institutions, Austrian economics views competition as a process of competitive rivalry in which information is dispersed and tacit. Therefore, competition is a knowledge-discovery process. Sixth, R-A theory draws on socioeconomics and institutional economics (De Gregori 1987; Ranson 1987; Etzioni 1988; North 1990). R-A theory recognises that societal institutions, such as laws, customs, taboos, traditions and moral codes, produce order by structuring political, economic and social interaction. The kind of order produced by societal institutions influences the process of R-A competition and, thereby, influences such outcomes as productivity and economic growth. Indeed, many scholars believe that non-wealthy, market-based economies lack certain productivity-enhancing societal institutions, particularly those that promote trust (Arrow 1972; Phelps 1975; Gambetta 1988; North 1990; Harrison 1992; Fukuyama 1995). R-A theory contributes to explaining (1) the presence of non-wealthy, market-based economies on the basis that evolutionary, competitive processes are embedded within differing institutional frameworks and (2) the productivity-enhancing capability of institutions that promote trust on the basis that societies having moral codes based on deontological ethics have reduced transaction and transformational costs (Hunt 1997b).

58 SHELBY D.HUNT

Figure 4.1 A schematic of the Resource-Advantage theory of competition Note: Competition is the disequilibrating, ongoing process that consists of the constant struggle among firms for a comparative advantage in resources that will yield a marketplace position of competitive advantage and, thereby, superior financial performance. Firms learn through competition as a result of feedback from relative financial performance ‘signalling’ relative market position, which, in turn signals relative resources. Source: Hunt (1997b).

Although R-A theory draws on the previously cited streams of literatures, it is not precisely the same thing as any of the works in its pedigree. For example, R-A theory, among other things, explicitly adds an analysis of the firm’s environment and a theory of demand to extant versions of resource-based theory. An examination of R-A theory’s structure shows other differences. The structure of R-A theory Figures 4.1 and 4.2 provide a schematic depiction of R-A theory’s key constructs, and Table 4.1 shows its foundations. Because R-A theory draws heavily on Austrian economics and the Schumpeterian tradition in evolutionary economics, (1) innovation and organisational learning are endogenous to R-A competition, (2) firms and consumers have imperfect information, and (3) entrepreneurial competence and institutions affect economic performance. Because R-A theory incorporates marketing’s heterogeneous demand theory, intra-industry demand is viewed as significantly heterogeneous as to consumers’ tastes and preferences. Therefore, different market offerings are required for different market segments in the same industry. Because it adopts a resource-based view of the firm, firms are theorised to be combiners of heterogeneous, imperfectly mobile resources. Combining the resource-based view of the firm with heterogeneous demand and imperfect information results in diversity in the sizes, scopes and levels of profitability of firms. This diversity exists not only across industries but for firms within the same industry. R-A theory stresses the importance of market segments, a comparative advantage (disadvantage) in resources, and marketplace positions of competitive advantage

SYNTHESISING RESOURCE-BASED, EVOLUTIONARY AND NEOCLASSICAL THOUGHT 59

Figure 4.2 Competitive position matrix

Note: The marketplace position of competitive advantage identified as cell 3 results from the firm, relative to its competitors, having a resource assortment that enables it to produce an offering for some market segment(s) that (a) is perceived to be of superior value and (b) is produced at lower costs. Source: Hunt (1997b).

(disadvantage). Market segments are intra-industry groups of consumers whose tastes and preferences for an industry’s output are relatively homogeneous. (The ultimate segment is, of course, a segment of one.) Resources are the tangible and intangible entities available to the firm that enable it to produce efficiently and/or effectively a market offering that has value for some market segment(s). Because many of the resources of firms within an industry are significantly heterogeneous and relatively immobile, some firms will have a comparative advantage and others a comparative disadvantage in efficiently and/or effectively producing market offerings that have value for particular market segments. Just as international trade theory recognises that nations have heterogeneous, immobile resources, and it focuses on the implications of comparative advantages in societal resources on trade, R-A theory recognises that many of the resources of firms within the same industry are significantly heterogeneous and relatively immobile and it focuses on the implications of comparative advantages in firm-level resources on competition. Analogous to nations, some firms will have a comparative advantage and others a comparative disadvantage in efficiently and/or effectively producing particular market offerings that have value for particular marketing segments. Specifically, when firms have a comparative advantage (disadvantage) in resources, they will occupy marketplace positions of competitive advantage (disadvantage), as shown in Figure 4.1 and further explicated in the nine marketplace positions in Figure 4.2. Marketplace positions of competitive advantage (disadvantage) then result in superior (inferior) financial performance. Innovation plays a key role in R-A competition. R-A theory distinguishes between proactive and reactive innovation. The former is innovation by firms that, though motivated by the expectation of superior financial performance, is not prompted by specific

Note: The foundational propositions of R-A theory are to be interpreted as descriptively realistic of the general case. Specifically, P1, P2, P5 and P7 for R-A theory are not viewed as idealised states that anchor end-points of continua. For example, P1 posits that intra-industry demand in most industries (i.e., the general case) is substantially heterogeneous, not perfectly heterogeneous. In contrast, P1 for perfect competition assumes the idealised state of perfect homogeneity. Source: Hunt (1997b).

Table 4.1 Foundational propositions of Perfect Competition and Resource-Advantage theory

60 SHELBY D.HUNT

SYNTHESISING RESOURCE-BASED, EVOLUTIONARY AND NEOCLASSICAL THOUGHT 61

competitive pressures. Indeed, proactive innovation is genuinely entrepreneurial in the classic sense of entrepreneur, i.e., in the sense of spotting new opportunities and subsequently developing market offerings (Kirzner 1979). In contrast, reactive innovation is directly prompted by the learning process of firms’ competing for the patronage of specific market segments. R-A theory maintains that it is not only entrepreneurial, proactive innovation that makes competition dynamic, but also reactive innovation. In reactive innovation, as the feedback loops in Figure 4.1 show, firms learn through competing as a result of the feedback from relative financial performance signalling relative market position, which, in turn, signals relative resources. When firms competing for a market segment learn from their inferior financial performance that they occupy positions of competitive disadvantage (cells 4, 7 and 8 in Figure 4.2), they attempt to neutralise and/or leapfrog the advantaged firm (or firms) by acquisition of their rival’s resource and/or by reactive innovation. Reactive innovation includes imitating the resource, finding (creating) an equivalent resource, or finding (creating) a superior resource. Here, ‘superior’ implies that the innovating firm’s new resource enables it to surpass the previously advantaged competitor in terms of either relative efficiency, or relative value, or both. Hence, reactive innovation promotes economic change. Firms occupying positions of competitive advantage (cells 2, 3 and 6 in Figure 4.2) can continue to do so if (1) they engage in proactive innovation, (2) they continually reinvest in the resources that produced the competitive advantage, and/or (3) rivals’ acquisition and reactive innovation efforts fail. Rivals will fail (or take a long time to succeed) when an advantage-producing resource either is protected by such societal institutions as patents or it is causally ambiguous, socially complex, highly interconnected, tacit, or has time compression diseconomies or mass efficiencies. Competition, then, is the constant struggle among firms for a comparative advantage in resources that will yield marketplace positions of competitive advantage for some market segment(s) and, thereby, superior financial performance. The nature of competitive processes and how well they work (e.g., how effectively competition produces economic growth) are significantly influenced by five environmental factors: the societal resources on which firms draw, the societal institutions that form the ‘rules of the game’ (North 1990), the actions of competitors and suppliers, the behaviours of consumers, and public policy decisions. The empirical evidence The empirical evidence for R-A theory focuses on the issue of firm diversity raised by Nelson (1991). Both within and across industries, firms differ dramatically in size, scope and financial performance. This diversity, though an anomaly for neoclassical theory, is expected under R-A theory (Hunt and Morgan 1995; Hunt 1997b). For R-A theory, five factors account for firm diversity. (1) Every firm is a unique entity in time and space (Chandler 1990). As a result of their history in obtaining and deploying resources, firms will differ from their rivals. (2) Different assortments of resources may be equally efficient or effective in serving some market segments, which will lead to firms

62 SHELBY D.HUNT

of varying sizes and scopes. (3) Heterogeneous demand implies that servicing different market segments in the same industry will lead to firms with different sizes and scopes, e.g., ‘niche’ marketers. (4) Firms tend to conduct activities in-house, rather than contract them out, when they constitute, or are part of an assortment of resources that constitutes, a competence. Therefore, because firms will be more hierarchical on those dimensions that constitute competences, differences in competences will promote diversity. (This fourth factor promoting firm diversity is, of course, precisely the point that Foss (1993) elaborates in detail in his discussion of the role of the competence perspective in evolutionary economics.) The fifth factor promoting firm diversity is that sometimes firms servicing some market segment have a comparative advantage in resources that competitors cannot imitate, find substitutes for, or leapfrog with an entirely new resource. Under such circumstances there will be firm diversity in the very important area of financial performance. Because the standard defence of neoclassical theory is that it makes accurate predictions and, therefore, either its assumptions are ‘close enough’ to reality or competitive processes work ‘as if‘its assumptions were true (Friedman 1953), diversity in financial performance provides an area for testing directly the merits of R-A versus neoclassical theory. If firms are viewed as combiners of homogeneous, mobile resources by means of a standard production function and intra-industry demand is viewed as homogeneous, then the variance in financial performance across firms and their business units must result from such industry factors as barriers to entry. This, of course, is the standard view of the neoclassical, structure-conduct-performance model in industrial organisation economics (Bain 1956). Empirically, therefore, if its assumptions are ‘close enough’ or competitive processes work ‘as if’, then neoclassical theory predicts that ‘industry effects’ should explain most of the variance in firms’ performance and ‘firm effects’ should explain very little. In contrast, if firms are viewed as combiners of heterogeneous, imperfectly mobile resources and intra-industry demand is viewed as heterogeneous, then firm effects should dominate industry effects. Thus, R-A theory and neoclassical theory make empirically testable, opposite predictions. Schmalensee’s (1985) investigation of the industry effects versus firm effects issue used Federal Trade Commission line-of-business data for 1975 and found industry effects accounting for 20 per cent of the variance in business-unit return on assets and corporate effects to be not significant. Rumelt (1991), however, pointed out that Schmalensee’s use of only one year’s data not only confounded stable industry effects with transient annual fluctuations but also made it impossible to separate overall corporation from individual business-unit effects. When Rumelt (1991) supplemented Schmalensee’s 1975 data with FTC data for 1974, 1976, and 1977, he found that, whereas industry effects explained only 8 per cent of the variance, corporate and business-unit effects explained 2 per cent and 44 per cent respectively. Supporting Rumelt’s findings, Roquebert, Phillips and Westfall (1996) found industry, corporate and business-unit effects to be 10, 18 and 37 per cent respectively (resulting in ‘total firm’ effects of 18+37=55 per cent). Notably, their sample was much larger (over 6,800 corporations), had a broader base (over 940

SYNTHESISING RESOURCE-BASED, EVOLUTIONARY AND NEOCLASSICAL THOUGHT 63

SIC, 4-digit categories), and (unlike FTC data) included both small and large corporations. The finding of large-scale studies that firm effects dominate industry effects strongly supports R-A theory’s premises: firms are better viewed as combiners of heterogeneous, imperfectly mobile resources and intra-industry demand is better viewed as significantly heterogeneous. Contrasted with neoclassical theory, R-A theory makes the correct prediction. Therefore, using Friedman’s (1953) reasoning, either R-A theory’s assumptions are closer or competitive processes work as if its assumptions were true. The competence perspective and R-A theory As the overview indicates, R-A theory views firm competences as resources and one of five factors that, collectively, explain the phenomenon of firm diversity. Competences are resources because a resource is defined as any tangible or intangible entity available to the firm that enables it to produce efficiently and/or effectively a market offering that has value for some market segment(s). Departing from neoclassical theory, resources are not restricted to a firm’s tangible assets or even to tangible assets plus human capital (see Table 4.1). Rather, resources can be categorised as financial (e.g., cash reserves and access to financial markets), physical (e.g., plant, raw materials and equipment), legal (e.g., trademarks and licences), human (e.g., the skills and knowledge of individual employees, including, importantly, their entrepreneurial skills), organisational (e.g., controls, routines, cultures and competences—including, importantly, a competence for entrepreneurship), informational (e.g., knowledge about market segments, competitors and technology), and relational (e.g., relationships with competitors, suppliers and customers). Note that even intangible entities can be resources, as long as they have an enabling capacity. Note also that a resource need not be owned by a firm, but only available to it. For example, a strategic alliance between A and B may contribute to their efficiency and/ or effectiveness and, thus, constitute a resource for both. Yet neither A nor B owns the resource, for neither can sell it. Therefore, R-A theory views a competence as a higherorder resource that consists of a distinct package of basic resources. In particular, competences are viewed as socially complex, interconnected combinations of tangible basic resources (e.g., specific machinery) and intangible basic resources (e.g., specific organisational policies and procedures and the skills and knowledge of specific employees) that fit coherently together in a synergistic manner. Competences, then, play a major role in enabling firms to produce efficiently and/or effectively valued market offerings. Competences take centre stage in R-A theory because many resources are proposed to be both significantly heterogeneous across firms, i.e., asymmetrically distributed, and imperfectly mobile, i.e., not commonly, easily or readily bought and sold in the factor markets. For example, an organisational competence for building high-quality automobile engines is not something that an organisation can ordinarily purchase in the marketplace. Thus, when a firm has a competence (or any other resource) that is rare among competitors, it has the potential for producing a comparative advantage for that firm.

64 SHELBY D.HUNT

Focusing on either an existing market segment or one newly identified through entrepreneurial activity, a comparative advantage in resources exists when a firm’s resource assortment enables it to produce a market offering that, relative to the extant offerings by competitors, (a) is perceived to have superior value and/or (b) can be produced at lower cost (Figures 4.1 and 4.2). Competences, then, are distinct resources because they are entities that contribute as distinct packages to the firm’s ability to produce efficiently and/or effectively market offerings that have value for some market segment (s). As such, competences—like tangible assets—can be deployed by managers in their strategies. For R-A theory, the role of management in a firm is to recognise and understand current strategies, create new strategies, select preferred strategies, implement (or manage) the strategies selected, and modify strategies through time. Such strategies involve, at the minimum, the identification of (1) market segments, (2) appropriate market offerings, and (3) the resources required to produce the offerings. Strategies that yield positions of competitive advantage and superior financial performance will do so when they rely on those resources in which the firm has a comparative advantage over its rivals. Sustained, superior financial performance occurs when a firm’s comparative advantage in resources continues to yield a position of competitive advantage despite the competitive actions of rivals. Because of the tacit nature of the knowledge underlying competences, they are precisely the kind of resource that managers can (and should) deploy in their quest for superior financial performance. For example, Teece and Pisano (1994:141) highlight the importance of ‘dynamic capabilities’, which they define as ‘the subset of the competences/ capabilities which allow the firm to create new products and processes, and respond to changing market circumstances’. Like Foss (1993), Teece and Pisano (1994:140) note that any firm with organisational competences ‘cannot be usefully modeled as a nexus of contracts’. They argue that a firm should develop its strategy by taking into close account its ‘managerial and organisational processes, its present position, and the paths available to it’ (p. 541). By ‘position’, they refer to the firm’s ‘current endowment of technology and intellectual property, as well as its customer base and upstream relations with suppliers…[and] its strategic alliances with competitors’ (p. 541). R-A theory therefore, merges with Teece and Pisano (1994) by delineating precisely the process by which a competence that enables firms to respond creatively to changing market conditions can be successful in achieving superior financial performance. To conclude this section, R-A theory accommodates the evolutionary concept of firm competence by placing it within an evolutionary theory of competition. If we consider again how Foss (1993) characterises the epistemic and economic content of competence and compare it with the specific version of the resource-based view adopted by R-A theory, there appears to be total congruence. Both view a competence as (1) constituting information capital for solving economic problems, (2) being agent specific and tacit, (3) being distributed asymmetrically (heterogeneously) over individuals and firms, and (4) being difficult to measure. Finally, both the competence perspective and the resourcebased view in R-A theory are evolutionary.

SYNTHESISING RESOURCE-BASED, EVOLUTIONARY AND NEOCLASSICAL THOUGHT 65

Therefore, both the evolutionary, competence view and R-A theory, by sharing a common epistemic and economic content, share a common problem. How do these theories relate to neoclassical, static-equilibrium, perfect competition? The competence perspective, R-A theory and perfect competition On the issue of how neoclassical, static-equilibrium theories relate (or should relate) to evolutionary, process theories, no consensus has emerged. For example, Dosi (1991:5) suggests that, on many issues separating the evolutionary and orthodox approaches, ‘there is no clear discontinuity…but some fuzzy continuum’. Consistent with this continuum view, Nelson (1995) argues that evolutionary theories should incorporate the empirical truth-content of their neoclassical counterparts. For example, in reviewing evolutionary growth theory, he cautions that ‘for evolutionary theory to have credibility these predictions had better be similar [to neoclassical theory], because any broad growth theory needs to be consistent with the basic empirically documented broad features of economic growth as we have experienced it’ (p. 71). On the other hand, Hodgson (1992) seems closer to viewing evolutionary and neoclassical theory as being mutually exclusive. He argues that even though evolutionary theory is not a ‘panacea’, it does ‘offer a metaphor for economics that is superior to the mechanistic model’ because, among other things, it can ‘incorporate the temporal as well as the moral aspects of all economic activity’ (p. 760). Nonetheless, he asks: ‘assuming it is possible to build a superior economic framework for economic theory, would it subsume neoclassical theory within itself as a special or limiting case?’ (p. 749). And, after reviewing the deficiencies of neoclassical theory, he concludes: ‘It is argued here that the reform of economics is not a question of adding additional dimensions to neoclassical economic theory. A theoretical revolution is required at the core of economics itself’ (p. 760). What, then, is the relationship between R-A theory, its competence perspective and perfect competition? Meritorious, I suggest, is the philosophy of science approach of Sellars (1963) and Levy (1996). They argue that science progresses best when a new theory can satisfactorily explain the explanatory and predictive successes of its predecessor. In so doing, a new theory incorporates its predecessor, which has the highly desirable benefit of preserving the cumulativity of a discipline.3 Similarly, after arguing that neoclassical economics is becoming increasingly open to fresh insights, Foss (1994: 16) advances the ‘speculative conjecture’ that evolutionary and neoclassical economics will eventually have a ‘kind of consistency relation…in which a more encompassing theory is constructed that can incorporate previously unrelated theories as special cases’. The thesis defended here is that (1) R-A theory and its foundations represent the general case of competition and (2) perfect competition and its foundations are a special case. Therefore, (3) R-A theory incorporates perfect competition, which (4) preserves the cumulativity of economic science. Empirical evidence provides a prima facie case for this thesis.

66 SHELBY D.HUNT

Recall the massive study of the determinants of business-unit return on assets by Roquebert, Phillips and Westfall (1996). Their sample of over 6,800 corporations came from a total of 940 SIC, 4-digit industries (representing 94 per cent of the universe of 1, 005 4-digit categories). The finding that firm effects dominate industry effects supports viewing the process identified by R-A theory as the general case of the process competition. Therefore, because a theory is derived from its assumptions, the evidence supports viewing each of R-A theory’s foundational postulates (Table 4.1) to be either descriptively realistic or, at least, ‘close enough’ (Friedman 1953). I now argue that the foundational propositions of perfect competition are, indeed, special cases of R-A theory and, consequently, R-A theory incorporates perfect competition by showing when the latter’s foundations are ‘close enough’ to predict. Demand, information and resource characteristics How should foundational propositions P1, P2, P5 and P7 in Table 4.1, concerning demand, information and resource characteristics, be interpreted? Note that each assumption could be viewed as an idealised state that anchors an end-point on a continuum. That is, demand (P1) can be conceptualised as a continuum with perfect homogeneity and perfect heterogeneity as idealised anchor-points. Similar continua could be conceptualised for information and its cost (P2 and P5) and for the homogeneityheterogeneity and mobility-immobility of resources (P7). However, whereas perfect competition is customarily interpreted in the idealised, anchor-point manner, in no case is R-A theory to be interpreted as the anchor-point opposite perfect competition. Rather, each foundational proposition of R-A theory is proposed as the descriptively realistic general case. Therefore, intra-industry demand (P1) is to be interpreted for R-A theory as substantially heterogeneous. Similarly, information for both firms (P5) and consumers (P2) is substantially imperfect and costly. Likewise, many, but not all, resources (P6) are substantially heterogeneous and immobile. As an example, consider footwear (SIC 314). R-A theory views consumers’ tastes and preferences for footwear to be substantially heterogeneous and constantly changing. Furthermore, consumers have substantially imperfect information concerning footwear products that might match their tastes and preferences, and obtaining such information is often costly in terms of both time and money. The implication of heterogeneity is that few —if any—industry markets exist: there are only market segments within industries. There is no ‘market for shoes’ (SIC 314), nor are there even separate markets for women’s shoes (SIC 3144) and for men’s shoes (SIC 3143). Even though all consumers require footwear and one can readily identify a group of firms that manufacture shoes, there is no shoe-industry market. That is, the group of firms that constitute the shoe industry do not collectively face a single, downward-sloping demand curve—for such an industry demand curve would imply homogeneous tastes and preferences. R-A theory maintains that, to the extent that demand curves exist at all, they exist at a level of (dis)aggregation that is too fine to be an ‘industry’. For example, even if there were a men’s walking shoe market, one certainly would not speak of the men’s walking

SYNTHESISING RESOURCE-BASED, EVOLUTIONARY AND NEOCLASSICAL THOUGHT 67

shoe industry. The fact that intra-industry demand is substantially heterogeneous in most industries (even at the 4-digit SIC level) contributes to R-A theory’s ability (and neoclassical theory’s inability) to make the correct prediction as to the diversity in business-unit financial performance. Likewise, the fact that intra-industry demand is relatively homogeneous (‘close enough’) in at least some commodity-type industries, e.g., gold ores (SIC 1041), contributes to explaining those special cases where perfect competition predicts well. Human motivation Now consider how perfect competition and R-A theory treat human motivation (P3). Whereas perfect competition assumes self-interest (utility) maximisation, R-A theory proposes that such self-interest seeking is constrained. In particular, R-A theory maintains that self-interest seeking is constrained by a moral code. I argue that R-A theory incorporates perfect competition on the grounds that, whereas the descriptively realistic R-A theory allows for different people (and different peoples) to have different moral codes, perfect competition restricts itself to the special case where everyone has one particular moral code—ethical egoism.4 In general, the moral codes that guide individual behaviour draw on two traditions in moral philosophy: deontology and teleology (Beauchamp and Bowie 1988). Because deontological codes focus on specific actions or behaviours and teleological codes focus on consequences, the former stress the inherent rightness-wrongness of a behaviour and the latter emphasise the amount of good or bad embodied in a behaviour’s consequences. Deontologists believe that ‘certain features of the act itself other than the value it brings into existence’ make an action or rule right or wrong (Frankena 1963:14). Moral codes based on deontology will emphasise the extent to which a behaviour is consistent or inconsistent with such deontological norms as those proscribing lying, cheating, deceiving or stealing and those prescribing honesty, fairness, justice or fidelity. Accordingly, they emphasise duties, obligations and responsibilities to others. Teleologists, on the other hand, ‘believe that there is one and only one basic or ultimate right-making characteristic, namely, the comparative value (nonmoral) of what is, probably will be, or is intended to be brought into being’ (Frankena 1963:14). Whereas deontological codes must address the difficult issue of conflicting norms, those emphasising teleological factors must grapple with the problem of which stakeholders are to be valued. Moral codes adopting utilitarianism hold that an act is right only if it produces for all people a greater balance of good over bad consequences than other alternatives (i.e., ‘the greatest good for the greatest number’). Even though it focuses on consequences, because utilitarianism demands that decision-makers consider an act’s consequences on all stakeholders, it shares common ground with deontology’s emphasis on duties and responsibilities to others. In stark contrast, codes adopting ethical egoism hold that an act is right only if the consequences of the act are most favourable for the individual decision-maker—a view directly opposed by deontological ethics.

68 SHELBY D.HUNT

For example, whereas an employee guided by deontological ethics believes padding expense reports is wrong because it violates norms proscribing cheating and stealing, an egoist believes padding is right or wrong depending on the relationship between the benefits of padding versus the likelihood of getting caught and the resulting punishment. Therefore, though both deontologists and egoists may abstain from padding expense reports, they do so for different reasons. These different reasons have efficiency and effectiveness consequences for both firms and societies (Hunt 1997b). Etzioni (1988) points out three conceptualisations of ‘utility’ and ‘utility maximisation’ in contemporary economics. The first, ‘P-utility’, is the self-interest maximisation of ethical egoism. P-utility directs all actions towards the pursuit of pleasure or the avoidance of pain: P-utility has longstanding philosophical and psychological foundations, it provides a major explanatory concept and generates testable hypotheses…To the extent that it is hypothesised that the pursuit of P-utility is a major explanatory factor, the hypothesis is clearly valid…To argue that people are pleasure-driven…surely explains a good part of human behavior. (pp. 28, 34) However, it is clear, as Etzioni (1988) documents, that people do not maximise P-utility. Indeed, many people do act unselfishly on many occasions, as shown by scores of empirical studies on genuine acts of altruism, on economic decisions related to such behaviours as saving, on activities related to public goods, on free riders, on the cooperative behaviours in prisoner’s dilemma experiments, and on the voting behaviours of citizens. Because people are not guided solely by ethical egoism, if ‘utility’ is meant as P-utility, then the utility maximisation thesis in the neoclassical model is clearly false. Etzioni (1988) discusses two means by which neoclassicists attempt to save utility maximisation in the face of its falsity. First, many turn utility into a tautology. Thus, acts of altruism are ‘explained’ by suggesting that the pleasure of the person who benefits from such acts has become a source of the doer’s pleasure, part of his or her utility. However, ‘saving’ the utility maximisation thesis by making it a tautology destroys its scientific status, for there are no conceivable behaviours that could possibly even count against the thesis. Second, some treat utility maximisation as an ex post rank ordering of preferences that makes the mathematical equations work out. However, science, unlike mathematics, requires substantive concepts. Moreover, even though neoclassicists can ‘save’ utility maximisation by turning the concept into either a tautology or an empty mathematical abstraction, when they turn to giving advice to policy-makers they tend to revert to treating P-utility maximisation as a substantive, verified thesis. As an example of treating P-utility maximisation as a substantive thesis, consider the atomised actors in transaction cost economics (TCE). They not only maximise self-interest, but do so with opportunistic ‘guile.’ Indeed: ‘Economic man…is thus a more subtle and devious creature than the usual self-interest seeking assumption reveals’ (Williamson 1975:155). Although TCE acknowledges that not all economic agents engage in

SYNTHESISING RESOURCE-BASED, EVOLUTIONARY AND NEOCLASSICAL THOUGHT 69

malfeasance, universal opportunism is assumed because it is ‘ubiquitous’ (Williamson 1981b: 1550) and ‘opportunistic types cannot be distinguished ex ante from sincere types’ (Williamson 1975:17; italics added) or, at the very least, ‘it is very costly to distinguish opportunistic from nonopportunistic types ex ante’ (Williamson 1981b:1545; italics added). Therefore, for Williamson, without the assumption of opportunism, ‘the study of economic organisation is pointless’ (p. 1545). Indeed, even though he acknowledges that ‘to craft credible commitments…is to create functional substitutes for trust’ (1994:1), he maintains that ‘the study of economic organisation is better served by treating economic organisation without reference to trust’ (1993:19). In contrast, R-A theory’s position that human behaviour is motivated by constrained self-interest seeking implies that consumers, owners and managers are constrained in their self-interest seeking by their moral codes, i.e., by considerations of what is right, proper, ethical, moral or just plain appropriate. In turn, individuals’ moral codes are shaped, but not determined, by societal institutions (see Figure 4.1). Because codes of behaviour are shaped by institutions, R-A theory is embedded (Granovetter 1985). Because such codes are not determined by institutions, R-A theory adopts a moderately socialised view of human behaviour. Indeed, R-A theory does not deny that individuals may act opportunistically. Rather, constrained self-interest seeking implies that opportunism is not assumed to be universal. That is, R-A theory maintains that the extent to which people behave opportunistically in various contexts is a research question to be explored and explained—not presumed a priori. Therefore, a major implication of constrained selfinterest seeking is that trust may exist among people who share a moral code. Firm’s objective and firm’s information Because it maximises the self-interests of the owner, perfect competition assumes the special case that owner-managed firms profit maximise (P4). Profit maximisation (or wealth maximisation, i.e., the maximisation of the net present value of future profits) occurs under conditions of, again, the special case of perfect and costless information about product markets, production techniques and resource markets (P5). In contrast, RA theory proposes that the firm’s primary objective is superior financial performance (P4), which it pursues under conditions of imperfect (and often costly to obtain) information about extant and potential market segments, competitors, suppliers, shareholders and production technologies (P5). For R-A theory, superior financial performance is indicated by such measures as profits, earnings per share, return on investment and capital appreciation. Here, ‘superior’ equates with both ‘more than’ and ‘better than’. It implies that firms seek a level of financial performance exceeding that of some referent. For example, the referent can be the firm’s own performance in a previous time-period, the performance of rival firms, and industry average, or a stockmarket average, among others. Affecting the process of competition, both the specific measure and the specific referent will vary somewhat from time to time, firm to firm, industry to industry and culture to culture.

70 SHELBY D.HUNT

Firms are posited to pursue superior financial performance because superior rewards— both financial and non-financial—will then flow to owners, managers and employees. Superior financial performance does not equate with ‘abnormal profits’ or ‘rents’ (i.e., profits differing from the average firm in a purely competitive industry in long-run equilibrium) because R-A theory views industry long-run equilibrium as such a rare phenomenon that ‘normal’ profits cannot be an empirical referent for comparison purposes. Furthermore, the actions of firms that collectively constitute competition do not force groups of rivals to ‘tend towards’ equilibrium. Instead, the pursuit of superior performance implies that actions of competing firms are disequilibrating, not equilibrating. As Langlois (1986:152) points out, though economic ‘agent[s] prefer more to less all things considered’, this ‘differs from maximizing in any strong sense’. Similarly, though RA theory posits that firms seek superior financial performance, I argue that the general case is that they do not ‘strong sense’ maximise because: (1) managers lack the capability and information to maximise (Simon 1979), (2) managers’ self-interests may diverge from those of owners, i.e., the ‘agency problem’, and (3) financial performance is constrained by managers’ views of morality. Perfect competition, of course, recognises none of these factors. Although some versions of neoclassical theory ‘relax’ the idealised assumptions of perfect competition and acknowledge factors 1 and 2, factor 3 remains unacknowledged in the undersocialised, neoclassical tradition. In contrast, R-A theory recognises that, at times, some managers resist cheating or opportunistically exploiting their customers, suppliers and others because they believe that such self-interest maximising behaviours would violate their duties or responsibilities, their sense of rightness or wrongness, their moral codes. Evolutionary processes and equilibrium For all of the first eight foundational propositions shown in Table 4.1, I argue that R-A theory’s foundations are descriptively realistic of the general case of competition and perfect competition’s assumptions are special cases. However, proposition 9, the issue of competitive dynamics, provides an opportunity to explicate the argument in detail. Under what set of circumstances will the process of R-A competition (which is disequilibriumprovoking, with innovation endogenous) result in perfect competition (which is equilibrium-seeking, with innovation exogenous)? Consider the following scenario. First, assume that a set of firms producing an offering for a particular market segment within an industry has been competing according to R-A theory. Therefore, because of resource heterogeneity (say, different levels of a key competence), the firms are distributed throughout the nine marketplace positions in Figure 4.2. Some firms, because of their comparative advantage in resources, are enjoying superior returns; others have inferior returns; and still others have parity returns. Next assume that, through time, both disadvantaged and parity firms gradually learn how the advantaged firms are producing their offerings more efficiently and/or effectively and successfully imitate them by acquiring or developing the requisite resources. For

SYNTHESISING RESOURCE-BASED, EVOLUTIONARY AND NEOCLASSICAL THOUGHT 71

example, they gradually develop competences equivalent to the advantaged firms. Then assume that, even though all firms seek superior financial performance, no firm can acquire, develop or create new resources (e.g., developing a new competence) that will enable it to produce a market offering more efficiently or effectively than any other firm. That is, for some reason or set of reasons, all competition-induced innovation stops, both proactive and reactive. Consequently, all competition-induced technological change stops. Under these economic conditions, then, the resources of all firms serving this market segment become homogeneous and there will be parity resources producing parity offerings. Next assume that the tastes and preferences of consumers in all other market segments served by the firms in this industry shift towards the original segment. Industry consumer demand will then become relatively homogeneous. Suppose further that consumers’ tastes and preferences remain stable throughout a significant period of time and that consumers become very knowledgeable about the relative homogeneity of firms’ offerings. There will then be parity resources producing parity offerings, which results in all firms having parity marketplace positions (cell 5 in Figure 4.2). Suppose further that firms have accurate information about competitive conditions and there are no institutional restraints preventing them from producing their market offerings in the profit-maximising quantity. Under these economic circumstances, the industry experiences no endogenous technological change, firms become price-takers, and a static equilibrium theory of competition, such as perfect competition, applies. That is, there will be parity resources producing parity offerings, which results in parity marketplace positions and parity performance (see Figure 4.1). The industry has now become a candidate for ‘industry effects’ to dominate ‘firm effects’ in empirical studies, for collusion and barriers to entry to become viable explanations for any industry-wide, superior financial performance, and, in general, for industry-level theoretical analyses to be appropriate. Next assume that the preceding process occurs in every industry in an entire economy. Then, if this set of economic circumstances persists through time, all competition-induced technological change ceases, all endogenous technological progress stops, all endogenous growth ceases, and a long-run, general equilibrium theory applies (such as Walrasian general equilibrium). In such an economy, growth comes only from exogenous sources, including those sources (e.g., government R&D or a state planning board) that might develop innovations that result in exogenous technological progress.5 Note that the preceding analysis began with the process of R-A competition for a market segment and sketches the special economic circumstances that must prevail for the competitive process to result in a static-equilibrium situation in an industry. Among other conditions, it showed that a very important circumstance is that all endogenous innovation must stop (or be stopped). Such a stoppage might come as a result of collusion, complacency, institutional restrictions, governmental fiat or lack of entrepreneurial competence. The analysis then sketched the special circumstances for a long-run general equilibrium to develop and, again, it showed that all endogenous technological progress in all industries—hence, all endogenous economic growth—in an economy must cease.

72 SHELBY D.HUNT

Therefore, perfect competition with (or without) Walrasian general equilibrium may be viewed as a special case of R-A theory. R-A theory relates to perfect competition in the same way that Newtonian mechanics relates to Galileo’s Law: the former incorporates the latter. Conclusion Many theorists now advocate the competence perspective of the firm (Langlois 1986; Klein 1988; Pelikan 1989a, 1989b; Carlsson and Eliasson 1990; Prahalad and Hamel 1990; Foss 1993; Teece and Pisano 1994; Langlois and Robertson 1995). Foss (1993), for example, puts forth an evolutionary, competence explanation for vertical integration. In this explanation, firms are viewed as repositories of tacit knowledge and they integrate because they have difficulty in communicating their needs precisely to potential suppliers. This chapter shows how the resource-advantage theory of competition accommodates the competence perspective and, thus, synthesises evolutionary economics with resourcebased theory. In R-A theory, competences are viewed as higher-order, socially complex, interconnected combinations of tangible and intangible basic resources that fit coherently together in a synergistic manner and, in doing so, contribute to enabling firms to produce valued market offerings efficiently and/or effectively. Because competences are likely to be heterogeneously distributed among rivals, they can produce marketplace positions of competitive advantage for those competitors that differentially possess them. Because competences are highly immobile, rivals in positions of competitive disadvantage cannot, in general, purchase them in the factor markets. Because competences are tacit, causally ambiguous and socially complex, the time required by rivals to imitate an advantaged competitor holding them may be long. Indeed, because competences are comprised of interconnected basic resources, the absence of a key basic resource can thwart competitors’ efforts at imitation. Therefore, the superior financial performance brought about by firms occupying marketplace positions as a result of a comparative advantage in competence can endure through time despite the concerted efforts of rivals. If R-A theory accommodates the evolutionary, competence perspective of the firm, what is the relationship between R-A theory, its competence perspective and perfect competition? I argue that R-A theory incorporates perfect competition on the grounds that the foundational propositions of R-A theory represent the general case of competition and the assumptions of perfect competition are special cases. Furthermore, not only does R-A theory predict well in the case of firm diversity in financial performance, but it also explains why perfect competition predicts well (when in fact it does so) by showing the special circumstances that must prevail for the process of R-A competition to gravitate towards static equilibrium. R-A theory shows that perfect competition will be close enough when, among other things, all competition-induced technological change ceases and, therefore, all endogenous technological progress stops and all endogenous economic growth ceases. Importantly, therefore, R-A theory is a general theory that incorporates

SYNTHESISING RESOURCE-BASED, EVOLUTIONARY AND NEOCLASSICAL THOUGHT 73

the extant explanatory and predictive successes of the special case of perfect competition. In so doing, it preserves the cumulativity of economic science. Notes 1 See Hunt (1999b) for how competence-based competition relates to R-A theory. 2 Hodgson (1993:39–51) provides the most detailed analysis yet of the characteristics of the various kinds of evolutionary theories possible in economics. His taxonomy distinguishes developmental theories that focus on ‘stages’ from genetic theories that concentrate on a ‘set of fairly durable human entities’ and a ‘detailed causal explanation’ of their interactions. Within genetic theories, he distinguishes ontogenetic theories that focus on ‘a set of given and unchanging’ entities from phylogenetic theories that focus on the ‘complete and ongoing evolution of the population, including changes in its composition’. He then distinguishes phylogenetic, consummatory theories that permit never-ending evolution. Although not all phylogenetic theories sort through natural selection, the key requirements for those that do so, he maintains, are that they must have units of selection that are ‘fairly durable’ and ‘heritable’ and that there must be a selection process that involves a ‘struggle for existence’ that ‘encompasses a renewable source of variety and change’, where the struggle results in the survival of the ‘fitter’, not necessarily the ‘fittest’. Therefore, Hunt (1997a) defends the position that R-A theory is a phylogenetic, non-consummatory, evolutionary theory of competitive firm behaviour. 3 The classic example of incorporation, of course, is that Newtonian theory (which maintains that the acceleration of two masses increases as they approach each other) incorporates Galileo’s Law of Descent (which assumes that acceleration is constant between two bodies) and, thereby, explains all the predictive successes of Galileo’s Law. Simply put, if d is the distance of a body from the surface of the earth and D is the radius of the earth, Galileo’s Law predicts well for most falling objects because the ratio d/D is—as in Friedman (1953) —‘close enough’ to zero that assuming g to be constant in S=1/2 gt2 is nonproblematic. Therefore, the foundations of Newtonian theory are such that they incorporate Galileo’s Law as a special case. 4 The discussion here follows the analysis in Hunt (1997b). 5 See Hunt (1997c) for the argument that R-A theory can provide a theoretical foundation for endogenous growth models, such as those of Romer (1990, 1993, 1994), Stokey (1991) and Young (1993).

References Alderson, W. (1957) Marketing Behavior and Executive Action, Homewood, IL: Richard D.Irwin. ——(1965) Dynamic Marketing Behavior, Homewood, IL: Richard D.Irwin. Arrow, K.J. (1972) ‘Gifts and Exchanges’, Philosophy and Public Affairs 15:343–561. Bain, J.S. (1956) Barriers to New Competition, Cambridge, MA: Harvard University Press. Barney, J.B. (1986) ‘Strategic Factor Markets, Expectations, Luck and Business Strategy’, Management Science 32:1231–41. ——(1991) ‘Firm Resources and Sustained Competitive Advantage’, Journal of Management 17:99– 120.

74 SHELBY D.HUNT

——(1992) ‘Integrating Organizational Behavior and Strategy Formulation Research: A ResourceBased Analysis’, in P.Shirvastava, A.S.Huff and J.E.Dutton (eds), Advances in Strategic Management, Greenwich, CT: JAI. Barney, J.B. and M.H.Hansen (1994) ‘Trustworthiness as a Source of Competitive Advantage’, Strategic Management Journal 15:175–90. Beauchamp, T.L. and N.E.Bowie (1988) Ethical Theory and Business (3rd edn), Englewood Cliffs, NJ: Prentice Hall. Black, J.A. and K.B.Boal (1994) ‘Strategic Resources, Traits, Configurations and Paths to Sustainable Competitive Advantage’, Strategic Management Journal 15: 131–48. Brumagim, A.L. (1994) ‘A Hierarchy of Corporate Resources’, in P.Shrivastava, A.S.Huff and J.E.Dutton (eds), Advances in Strategic Management, Greenwich, CT: JAI. Carlsson, B. and G.Eliasson (1990) ‘The Nature and Importance of Economic Competence’, Working Paper no. 294, IUI, Stockholm. Chamberlin, E. (1933) The Theory of Monopolistic Competition, Cambridge, MA: Harvard University Press. Chandler, A.D. (1990) Scale and Scope: The Dynamics of Industrial Capitalism, Cambridge, MA: Harvard University Press. Clark, J.M. (1961) Competition as a Dynamic Process, Washington, DC: Brookings Institution. Collis, D.J. (1991) ‘A Resource Based Analysis of Global Competition’, Strategic Management Journal 12:49–68. ——(1994) ‘How Valuable are Organizational Capabilities?’, Strategic Management Journal 15:143– 52. Conner, K. (1991) ‘A Historical Comparison of Resource-Based Theory and Five Schools of Thought within Industrial-Organization Economies. Do We Have a New Theory of the Firm?’, Journal of Management 17:121–54. De Gregori, T.R. (1987) ‘Resources Are Not; They Become: An Institutional Theory’, Journal of Economic Issues 21(35):1241–63. Demsetz, H. (1988) ‘The Theory of the Firm Revisited’, Journal of Law and Economics 4:141–62. Dierickx, I. and K.Cool (1989) ‘Asset Stock Accumulation and Sustainability of Competitive Advantage’, Management Science 35:1504–11. Dosi, G. (1991) ‘Some Thoughts on the Promises, Challenges, and Dangers of an “Evolutionary Perspective” in Economics’, Journal of Evolutionary Economics 1:5–7. Eliasson, G. (1990a) ‘The Firm as a Competent Team’, Journal of Economic Behavior and Organization 13:275–98. ——(1990b) ‘Business Competence, Organizational Learning and Economic Growth’, Working Paper no. 264, IUI, Stockholm. Etzioni, A. (1988) The Moral Dimension: Toward a New Economics, New York: The Free Press. Foss, N.J. (1993) ‘Theories of the Firm: Contractual and Competence Perspectives’, Journal of Evolutionary Economics 3:127–44. ——(1994) ‘Realism and Evolutionary Economics’, Journal of Social and Evolutionary Systems 17(1): 21–40. Foss, N.J., C.Knudsen and C.A.Montgomery (1995) ‘An Exploration of Common Ground: Integrating Evolutionary and Strategic Theories of the Firm’, in C.A. Montgomery (ed.), Resource-Based and Evolutionary Theories of the Firm: Towards a Synthesis. Norwell, MA: Kluwer Academic. Frankena, W. (1963) Ethics, Englewood Cliffs, NJ: Prentice Hall. Friedman, M. (1953) Essays in Positive Economics, Chicago: University of Chicago Press.

SYNTHESISING RESOURCE-BASED, EVOLUTIONARY AND NEOCLASSICAL THOUGHT 75

Fukuyama, F. (1995) Trust: The Social Virtues and the Creation of Prosperity, New York: The Free Press. Gambetta, D. (1988) Trust: Making and Breaking Cooperative Relationships, New York: Basil Blackwell. Granovetter, M. (1985). ‘Economic Action and Social Structure: The Problem of Embeddedness’, American Journal of Sociology 91(3):481–510. Grant, R.M. (1991) ‘The Resource-Based Theory of Competitive Advantage: Implications for Strategy Formulation’, California Management Review 33:114–33. Hamel, G. and A.Heene (eds) (1994) Competence-Based Competition, New York: John Wiley and Sons. Harrison, L.E. (1992) Who Prospers?: How Cultural Values Shape Economic and Political Success, New York: Basic Books. Hayek, F.A. (ed.) (1935) Collectivist Economic Planning: Critical Studies on the Possibilities of Socialism, London: Routledge. ——(1948) Individualism and Economic Order, Chicago: University of Chicago Press. Heene, A. and R.Sanchez (eds) (1997) Competence-Based Strategic Management, New York: John Wiley and Sons. Hodgson, G.M. (1992) ‘The Reconstruction of Economics: Is There Still a Place for Neoclassical Theory?’, Journal of Economic Issues 26(3):749–67. ——(1993) Economics and Evolution. Ann Arbor: University of Michigan Press. Hunt, S.D. (1995) ‘The Resource-Advantage Theory of Competition: Toward Explaining Productivity and Economic Growth’, Journal of Management Inquiry 4:317–32. ——(1997a) ‘Resource-Advantage Theory: An Evolutionary Theory of Competitive Behavior?’, Journal of Economic Issues 31 (March):59–77. ——(1997b) ‘Resource-Advantage Theory and the Wealth of Nations’, Journal of Socio Economics 26 (4):335–57. ——(1997c) ‘Evolutionary Economics, Endogenous Growth Models and Resource-Advantage Theory’, Eastern Economic Journal 23(4):427–41. ——(1998) ‘Productivity, Economic Growth, and Competition: Resource Allocation Resource Creation?’, Business and the Contemporary World 9. ——(1999a) ‘The Strategic Imperative and Sustainable Competitive Advantage: Public Policy Implications of Resource-Advantage Theory’, Journal of the Academy of Marketing Science 27. ——(1999b) ‘The Competence-Based, Resource-Advantage, and Neoclassical Theories of Competition: Towards Synthesis’, Strategic Management Review 1. ——(1999c) A General Theory of Competition: Strategy, Resources, Productivity, Economic Growth, Thousand Oaks, CA: Sage. Hunt, S.D. and R.M.Morgan (1995) ‘The Comparative Advantage Theory of Competition’, Journal of Marketing 59:1–15. ——(1996) ‘The Resource-Advantage Theory of Competition: Dynamics, Path Dependencies, and Evolutionary Dimensions’, Journal of Marketing 60:107–14. ——(1997) ‘Resource-Advantage Theory: A General Theory of Competition?’, Journal of Marketing 60:74–82. Jensen, M.C. and W.H.Meckling (1976) ‘Theory of the Firm: Managerial Behavior, Agency Costs, and Ownership Structure’, in J.B.Barney and W.G.Ouchi (eds) (1986) Organizational Economics, San Francisco: Jossey-Bass. Kirzner, I.M. (1979) Perception, Opportunity and Profit: Studies in the Theory of Entrepreneurship, Chicago: University of Chicago Press. Klein, B. (1988) ‘Vertical Integration as Organizational Ownership’, Journal of Law and Economics 4: 201–13.

76 SHELBY D.HUNT

Lado, A.A. and M.C.Wilson (1994) ‘Human Resource Systems and Sustained Competitive Advantage: A Competency-Based Perspective’, Academy of Management Review 19:699–727. Langlois, R.N. (1986) Economics as a Process, Cambridge: Cambridge University Press. Langlois, R.N. and P.L.Robertson (1995) Firms, Markets, and Economic Change: A Dynamic Theory of Business Institutions, London: Routledge. Levy, G.B. (1996) ‘Theory Choice and the Comparison of Rival Theoretical Perspectives in Political Sociology’, Journal of the Philosophy of Social Science 26 (March): 26–60. Lippman, S.A. and R.P.Rumelt (1982) ‘Uncertain Immitability’, Bell Journal of Economics 13:418– 38. Madhok, A. (1997) ‘Cost, Value, and Foreign Entry Mode’, Strategic Management Journal 18:39–62. Marshall, A. (1898) ‘Distribution and Exchange’, The Economic Journal 8:37–59. Mises, L. (1949) Human Action: A Treatise on Economics, New Haven, CT: Yale University Press. Montgomery, C.A. (ed.) (1995) Resource-Based and Evolutionary Theories of the Firm: Towards a Synthesis, Norwell, MA: Kluwer Academic. Nelson, R.R. (1991) ‘Why Do Firms Differ, and How Does it Matter?’, Strategic Management Journal 12:61–74. ——(1995) ‘Recent Evolutionary Theorizing about Economic Change’, Journal of Economic Literature 33:48–90. Nelson, R.R. and S.G.Winter (1982) An Evolutionary Theory of Economic Change, Cambridge, MA: The Belknap Press. North, D.C. (1990) Institutions, Institutional Change, and Economic Performance, Cambridge: Cambridge University Press. Pelikan, P. (1988) ‘Can the Imperfect Innovation System of Capitalism be Outperformed?’, in G.Dosi, C.Freeman, R.R.Nelson, G.Silverberg and L.Soete (eds), Technical Change and Economic Theory, London: Pinter. Pelikan, P. (1989a) ‘Evolution, Economic Competence and Corporate Control’, Journal of Economic Behavior and Organization 12:279–303. ——(1989b) ‘Markets as Instruments of Evolution of Structures’, Working Paper no. 237, IUI, Stockholm. Penrose, E.T. (1959) The Theory of the Growth of the Firm, Oxford: Basil Blackwell and Mott. Peteraf, M.A. (1993) ‘The Cornerstones of Competitive Advantage: A Resource-Based View’, Strategic Management Journal 14:179–91. Phelps, E.S. (1975) Altruism, Morality, and Economic Theory, New York: Russell Sage Foundation. Porter, M.E. (1985) Competitive Strategy, New York: The Free Press. Prahalad, C.K. and G.Hamel (1990) ‘The Core Competence of the Corporation’, Harvard Business Review May–June:79–91. Ranson, B. (1987) ‘The Institutionalist Theory of’ Capital Formation’, Journal of Economic Issues 21 (September):1265–78. Reed, R. and R.J.DeFillippi (1990) ‘Causal Ambiguity, Barriers to Imitation, and Sustainable Competitive Advantage’, Academy of Management Review 15:88–117. Romer, P.M. (1990) ‘Endogenous Technological Change’, Journal of Political Economy 71–102. ——(1993) ‘Two Strategies for Economic Development: Using Ideas and Promoting Ideas’, in Proceedings of the World Bank Annual Conference on Development Economics 1992, New York: World Bank. ——(1994) ‘The Origins of Endogenous Growth’, Journal of Economic Perspectives (Winter): 3–22. Roquebert, J.A., R.L.Phillips and P.A.Westfall (1996) ‘Markets versus Management: What “Drives” Profitability?’, Strategic Management Journal 17:653–64. Rumelt, R.P. (1991) ‘How Much Does Industry Matter?’, Strategic Management Journal 12:167–85.

SYNTHESISING RESOURCE-BASED, EVOLUTIONARY AND NEOCLASSICAL THOUGHT 77

Sanchez, R. and A.Heene (eds), Strategic Learning and Knowledge Management, Chichester: John Wiley and Sons. Sanchez, R., A.Heene and H.Thomas (eds), Dynamics of Competence-Based Competition, Oxford: Elsevier Pergamon. Schendel, D. (1994) ‘Introduction to Competitive Organizational Behavior: Toward an Organizationally-Based Theory of Competitive Advantage’, Strategic Management Journal 15: 14. Schmalensee, R. (1985) ‘Do Markets Differ Much?’, American Economic Review 75: 341–50. Schoemaker, P.J.H. and R.Amit (1994) ‘Investment in Strategic Assets: Industry and Firm-Level Perspectives’, in P.Shrivastava, A.S.Huff and J.E.Dutton (eds), Advances in Strategic Management, Greenwich, CT: JAI. Schumpeter, J.A. (1950) Capitalism, Socialism, and Democracy, New York: Harper and Row (1st edn 1942). Sellars, W. (1963) Science, Perception, and Reality, London: Routledge and Kegan Paul. Simon, H.A. (1979) ‘Rational Decision Making in Business Organizations’, American Economic Review 69:493–512. Stokey, N. (1991) ‘Human Capital, Product Quality, and Growth’, The Quarterly Journal of Economics (May):587–616. Teece, D. and G.Pisano (1994) ‘The Dynamic Capabilities of Firms’, Industrial and Corporate Change 3:537–56. Wernerfelt, B. (1984) ‘A Resource-Based View of the Firm’, Strategic Management Journal 5:171– 80. Williamson, O.E. (1975). Markets and Hierarchies, New York: The Free Press. ——(1981a) ‘The Economics of Organization: The Transaction Cost Approach’, American Journal of Sociology 87 (November):548–77. ——(1981b) ‘The Modern Corporation: Origins, Evolution, and Attributes’, Journal of Economic Literature 19:1537–68. ——(1985) The Economic Institutions of Capitalism, New York: The Free Press. ——(1993) ‘Opportunism and Its Critics’, Managerial and Decision Economics 14: 94–107. ——(1994) ‘Transaction Cost Economics and Organization Theory’, in N.J.Smelser and R.Swedberg (eds), The Handbook of Economic Sociology, Princeton: Princeton University Press. Young, A. (1993) ‘Invention and Bounded Learning by Doing’, Journal of Political Economy (June): 443–72.

78

5 A processual account of innovative capabilities Peter W.Roberts

Introduction This chapter revisits innovative capabilities in light of insights developed within the Austrian school of economics. It discusses the extent to which one may recognise the existence of an innovative capability when future means-ends combinations are indeterminate, and which specific elements of an innovative capability are susceptible to competitor imitation. With theoretical roots extending back to the 1950s (Penrose 1959), the resource-based view of the firm has become a dominant strategic paradigm in the 1990s. Within resourcebased theory, firms are ultimately interested in achieving sustained superior financial performance (Hunt and Morgan 1995). Firms are modelled as bundles of productive resources and capabilities which combine to yield sustainable competitive advantages— and therefore sustained superior financial performance—for some but not others (Wernerfeldt 1984; Barney 1991). Under conditions of competitive parity—wherein all firms compete with similar sets of capabilities—a firm should expect to earn only normal returns. Because (by assumption) firms compete with undifferentiated attributes and offerings, they subject one another to competitive pressures that drive superior returns on to normal levels. Against this back-drop, it is straightforward to see how a positional capability that is both valuable and rare may deliver a competitive advantage, and therefore superior performance at a point in time. A capability is valuable if it contributes to delivering a product or service with attributes that are valued by customers. Rareness assures that competitive pressures do not eliminate the excess returns that otherwise flow to the valuable capability. This captures the static dimension of resource-based theory. Consistent with Schumpeter’s (1934) observation that the economic system does not react passively to the presence of superior profits, resource-based theory also comprises an important dynamic dimension (Foss 1997). Unless otherwise impeded, the imitation of a rare positional capability makes it increasingly common as time passes. As the degree of rareness diminishes, competition from similarly endowed firms erodes the superior returns that otherwise flow to a positional capability. This imitation dynamic ensures that, unless one can identify a set of isolating mechanisms (which protect positional capabilities from

80 A PROCESSUAL ACCOUNT OF INNOVATIVE CAPABILITIES

Figure 5.1 Hierarchy of firm capabilities

imitation—Rumelt 1984, 1987), a firm should not expect to sustain a competitive advantage into the long run by relying on a single positional capability. If a firm cannot rely on a given positional capability to deliver a sustainable competitive advantage, it must continually innovate in order to establish new positional capabilities as the competitive significance of older ones depreciate (Rumelt 1984, 1987; Grant 1991; Brumagim 1994; Collis 1994; Henderson and Cockburn 1994). In the development of its theoretical foundations, resource-based theory has been effectively compared and contrasted with different economic approaches (Conner 1991; Mahoney and Pandian 1992). However, notwithstanding numerous references to Schumpeter’s (1934) process of creative destruction (e.g., Rumelt 1984, 1987; Barney 1991; Conner 1991; Grant 1991; Mahoney and Pandian 1992; Amit and Schoemaker 1993), few scholars address the commonalties between resource-based theory and the more processually orientated Austrian school of economics (for exceptions, see Jacobson 1992; Foss 1997). Such an oversight—not surprising given the marginal positions occupied by Austrian economists relative to the mainstream of economic thought—is allowing resource-based theory to develop in the direction of the static-equilibrium frameworks that characterise orthodox (cf., Nelson and Winter 1982) economic theory. This is particularly problematic as formal recognition of imitation and innovation within the dynamic dimension of resource-based theory suggests that a fundamental theoretical concern should be with the unfolding market process, and not with equilibrium constellations of outputs and prices (Foss 1997). This shortcoming is evident in current thinking on innovative capabilities. Collis (1994) proposes that a hierarchy of capabilities exists within firms, flowing from positional capabilities, through innovative capabilities, and up to creative capabilities (see Figure 5.1). Positional capabilities perform the basic activities of the firm. As one moves up the hierarchy, dynamic capabilities deliver dynamic improvements to the activities of the firm, while creative capabilities are responsible for delivering novel strategic insights. Brumagim (1994) suggests a similar hierarchy comprising production/maintenance

PETER W.ROBERTS 81

resources, administrative resources, organisational learning resources (supporting innovation and change) and strategic vision resources. Each of these contributions is traceable to Nelson and Winter’s (1982) evolutionary perspective, which juxtaposes the static routines in a firm and the higher-order innovation routines which are responsible for transforming static routines over time. In discussing the long-run competitive significance of innovative capabilities, Collis (1994) transports the above imitation discussion to the next level of the capabilities hierarchy, arguing that sustainability is not assured even for firms possessing valuable innovative capabilities because these too are subject to competitor imitation. Following this logic through successively higher-order capabilities, he argues that any search for a truly sustainable competitive advantage is pre-empted by successive rounds of higher-order imitation. An Austrian perspective on market process (which links an innovative capability with entrepreneurial discovery) reveals a critical flaw in this line of thinking stemming from the symmetric treatment of positional and innovative capabilities. The value of an existing positional capability is determined within known means-ends frameworks. Its existence is premised on an understanding of the linkages between a given set of firm activities (the means) and a desired set of outcomes (the ends) (Teece et al. 1997). The fact that such capabilities are currently in place suggests that knowledge of the underlying means-ends framework is also in place. Consistent with this, Leonard-Barton (1992:113, emphasis added) defines ‘a core capability as the knowledge set that distinguishes and provides a competitive advantage’. At the same time, the imitation by competitors of an existing positional capability also takes place within the context of known means-ends frameworks. Imitation occurs as the means-ends knowledge that previously resided solely in one location (the advantaged firm) is diffused across firm boundaries. Because the knowledge of the means-ends framework that underpins a positional capability has already been discovered, it is non-problematic to discuss the existence of a positional capability at a point in time. And, because the (pure-form) imitation of such a capability also occurs within existing knowledge structures, one can confidently predict the imitative behaviour of competitors. However, the status of an innovative capability is not so clear-cut. The value of future positional capabilities is tied to yet-to-be-discovered means-ends frameworks. In a non-deterministic depiction of the competitive process, the specific means-ends frameworks that underpin future positional capabilities are not fully knowable at the present time. Although such possibilities can be imagined within the current competitive configuration, they are by no means determined by it (O’Driscoll and Rizzo 1996). It is only during the course of the unfolding market process that entrepreneurial firms discover (ex post) valuable positional capabilities (Kirzner 1973). Non-determinism is an essential feature of any valid account of the competitive process because a firm’s knowledge and expectations (which drive entrepreneurial discovery) are continually revised with the accumulation of its specific experiences; firms experience genuine learning as time elapses. In light of this, an Austrian perspective on the market process questions the extent to which an innovative capability can fully exist in any meaningful sense. This is not to say

82 A PROCESSUAL ACCOUNT OF INNOVATIVE CAPABILITIES

that a historical analysis cannot uncover relationships between certain firm attributes and the propensity of a firm to make discoveries (see Henderson and Cockburn 1994; Cohen 1995). However, the results of such historical analyses are very much time and place dependent and do not generalise to the future in a non-problematic way (Kirzner 1976; Amit and Schoemaker 1993).1 If an innovative capability is given meaning solely by its ability to generate a stream of future positional capabilities, but if these future capabilities are not fully knowable ex ante, then one must question what aspects of an innovative capability exist at any point in time, and further what aspects are subject to competitor imitation. The next section outlines a resource-based view of sustained superior financial performance, highlighting its dynamic dimension and the distinction between positional and innovative capabilities. The following section briefly summarises an Austrian perspective on the market process.2 Central to this perspective is an emphasis on the subjective elements of economic action and a concomitant aversion to deterministic accounts of the market process. The third section of the chapter extends the Austrian perspective to assess critically the current treatment of innovative capabilities within resource-based theory. The final section suggests that there may be two interrelated components of an innovative capability. Dynamic routines (with which firms search for solutions within prescribed problem spaces) operate in a relatively deterministic fashion and are dependent on existing (dynamic) means-ends frameworks, while entrepreneurial alertness reflects genuine discovery and is characterised by qualitatively different properties. Capabilities and sustained superior financial performance Resource-based scholars theorise that competitive advantages (and superior financial returns) flow to firms that possess or control valuable and rare positional capabilities. The term positional capability refers to ‘a firm’s capacity to deploy resources, usually in combination, using organisational processes, to effect a desired end’ (Amit and Schoemaker 1993:35). This usage is consistent with Grant (1991:120), who notes that ‘the capabilities of a firm are what it can do as a result of teams of resources working together’. Strategic positional capabilities are valuable in the sense that they contribute to the provision of a good or service with attributes that customers ultimately value. They are also rare in the sense that they are not available in sufficient supply that competition among firms drives abnormal profits to zero (Peteraf 1993). A competitive advantage at a point in time is kept logically distinct from a sustainable competitive advantage by the threat of imitation. Unless otherwise impeded, imitators emerge to compete away excess profits. In the extreme, such imitation is immediate and the profit associated with a positional capability vanishes as soon as it arises. However, two factors might sustain a competitive advantage: friction in the competitive process or the establishment of outright monopoly positions (Schumpeter 1934). Resource-based scholars have embraced this imitative dynamic. The most obvious example is Rumelt’s (1984:567; 1987) discussion of isolating mechanisms, which are factors ‘that make

PETER W.ROBERTS 83

Figure 5.2 Series of positional advantages

competitive positions stable and defensible’ (e.g., patents, information impactedness or firm reputation). Related references are found in Barney’s (1991) discussion of resource inimitability, Peteraf’s (1993) discussion of ex post limits to competition, Reed and DeFillippi’s (1990) discussion of causal ambiguity (see also Lippman and Rumelt 1982), and Williams’ (1992) discussion of sustainability regimes. In his recent book on hypercompetition, D’Aveni (1994) demonstrates (pictorially) the relationship between imitation and the sustainability of competitive advantage at the firm level (Figure 5.2). Figure 5.2 introduces what have come to be called innovative capabilities. If the rareness of a positional capability is adversely affected by imitation, then firms wishing to sustain an advantage must establish new positional capabilities as the strategic significance of existing ones diminishes: in industries where competitive advantages based on differentiation and innovation can be imitated…firms must be concerned not with sustaining existing advantages, but with creating the flexibility and responsiveness that permits them to create new advantages at a faster rate than the old advantages are being eroded by competition. (Grant 1991:131) Similar references are made by Rumelt (1984, 1987), who stresses the significance of a firm’s entrepreneurial ability, Brumagim (1994), who discusses organisational learning resources as contributors to a firm’s sustained superior performance, and Collis (1994: 144), who notes that ‘capabilities researchers are, therefore, now searching for the organisational structures and behaviours that will generate effective product introduction’. Finally, Henderson and Cockburn (1994) relate the innovative propensities of US pharmaceutical firms to factors such as component competence (unique disciplinary expertise) and architectural competence (the ability to access and integrate various types of expertise).

84 A PROCESSUAL ACCOUNT OF INNOVATIVE CAPABILITIES

Consistent with Figure 5.2, innovative capabilities enhance the sustainability of competitive advantage at the firm level by generating a series of positional capabilities over time. However, as imitation plays a critical role in the sustainability of advantages related to positional capabilities, some scholars allude to the deleterious effect of imitation vis-à-vis the sustainability associated with innovative capabilities. Lengnick-Hall (1992: 400) treats these two types of capabilities symmetrically, suggesting that ‘if the innovation process or the outcomes of innovation are difficult to copy, effective corporate entrepreneurship becomes an increasingly important ingredient in sustaining competitive advantage’. Collis (1994) is more explicit in his analysis of sustainability, suggesting that it is non-problematic transporting the imitation argument from positional to innovative capabilities. A more processual account of sustainable competitive advantage would suggest that the value of an innovative capability need not diminish after it has been imitated. This would only be true in the unlikely instance of a one-to-one correspondence between a set of firm attributes at some point in time and the innovative outcomes that it subsequently generates. Then, diffusion of this capability across competitors would lead to the simultaneous introduction of similar new-to-the-market products or processes, and each would not enjoy the temporary monopoly position that delivers above-normal profitability. It is more likely that the proliferation of such capabilities (to the extent that they are associated with enhanced innovative propensity) will lead to a faster rate of introduction of differentiated innovative offerings. In other words, ‘given uncertainty, the ex post results of entrepreneurial activity will necessarily be resource heterogeneity’ (Rumelt 1984:561). Note the correspondence between this and debates surrounding the race models of innovation that are found in the industrial organisation economic literature. In race models, it is assumed that competing firms chase the same innovative output, the arrival time of which is determined by the amount of, and timing of R&D investments. In their analysis of pharmaceutical industry competition, Cockburn and Henderson (1994:484, emphasis added) conclude that ‘the assumption that competition for a single “prize” is probably not a useful one in this context…the industry is characterised by substantial spillovers of knowledge and similar research can lead to related but significantly different outcomes’. Similarly, Comanor (1994:13) discusses the relationship between the technology that is employed to discover new drugs and that which is embodied in each specific product when he refers to the change in technology that accompanied the movement into biotechnology research and development: ‘the new technology refers not to the production of individual products, but rather to the process by which new products are discovered’. Finally, Temin (1979) corroborates this view by painting the following picture of innovation within the drug industry: ‘the drug industry was transformed by a research technology, rather than by a production technology. This method of doing research and finding new drugs could not be patented. It was an approach—a series of questions—rather than a specific method; and it produced a variety of outcomes, not a single product’ (p. 436). Each of these passages alludes to the separability of an innovative capability and the specific innovative outputs that it generates. This separation is given enhanced clarity with

PETER W.ROBERTS 85

a more processual orientation to resource-based theory. In the following, we ask a pair of related questions. First, to what extent is it possible to collapse the means-ends frameworks that underpin future positional capabilities on to the present in a deterministic fashion? In other words, to what extent does it makes sense to recognise the existence of an innovative capability? Second, to what extent are innovative capabilities subject to imitation? Before addressing each of these questions, we first provide a brief summary of the Austrian school of economics. An Austrian perspective on the market process The Austrian school of economics is differentiated from its more orthodox counterparts in its conception of markets as processes that are in perpetual motion, rather than as equilibrium constellations of quantities and prices. It is difficult to distil the broad range of Austrian economic research into a few workable ideas. With this in mind, the following paragraphs discuss two interrelated themes that are prevalent within Austrian economics: the subjective nature of economic decision making, and the aversion to deterministic accounts of the market process. In respect of the former, we borrow from Kirzner’s (1973, 1992) description of the market process driven in large part by entrepreneurial discovery on the part of market participants. In the latter respect, we rely on O’Driscoll and Rizzo’s (1996) subjective conception of the passage of time. An adherence to praxeology—or ‘the fundamental axiom that…individuals engage in conscious actions toward chosen goals’ (Rothbard 1976:19)—forces Austrian economists to understand events from the perspectives of the economic actor. Because of this, Austrian economics has been branded ‘subjective economics’. More concretely, Austrian economists trace all theorising about economic activity through the minds of the individual decision-makers, which ultimately shape the actions they take (Dolan 1976). Those minds, which embody knowledge, expectations and preferences, are both indeterminate and unpredictable (Kirzner 1976). Moreover, those minds are capable of learning from accumulated experience. As such, economics cannot be reduced to the search for generalisable, law-like predictions about economic life. Focusing on competitive markets, Kirzner (1973) theorises that the market process is driven by two different classes of decisions. The first are the familiar economising decisions, whereby firms employ known resource combinations in increasingly efficient ways to achieve given ends. The second class of entrepreneurial decisions encompasses those acts of discovery whereby novel resource combinations (or capabilities) are derived to satisfy previously unidentified ends. Consideration of both classes of decisions has strong implications for an equilibrium-based view of markets (Kirzner 1973). It is the entrepreneur who discovers novel resources, resource combinations and resource uses, creating new markets and fundamentally transforming existing ones. And, whereas economising decisions may push a market towards an equilibrium, the continuous stream of entrepreneurial decisions assures that that equilibrium is never actually reached (Lachmann 1976).

86 A PROCESSUAL ACCOUNT OF INNOVATIVE CAPABILITIES

There is a fundamental association between a processual view of markets and beliefs about the flow of knowledge into and through competitive markets. Markets are kept in perpetual disequilibrium by the mutual presence of ignorance and discovery (Kirzner 1992). In conceding ignorance, one accepts that knowledge of current and especially future profitable opportunities is never complete. In the absence of complete knowledge and perfect foresight, entrepreneurs are ever alert for previously unheeded profitable opportunities. This entrepreneurial discovery is driven by the evolution of individuals’ specific knowledge bases and their expectations of future competitive configurations. This point is critical because, from one period to the next, individuals learn and their specific experiences accumulate (often in unexpected ways). Because economic decisions are based on what is currently known and what is expected to transpire in the future, the constellation of market decisions are in a constant state of revision. By emphasising this evolution of knowledge and expectations, Austrian economists allow for genuine learning and therefore indeterminacy with respect to future market outcomes.3 Moreover, ‘genuine learning is not merely the result of a determinate processing of what is already known. It must go beyond those narrow confines and include unpredictable shifts in the method of processing itself’ (O’Driscoll and Rizzo 1996:24). Respecting the evolving subjectivity of market participants, theorists must be wary of attempts to cast market dynamics as deterministic processes. Such a tack implies that all future competitive activity is explicitly derived from current competitive conditions. As argued by O’Driscoll and Rizzo (1996), any deterministic account of market dynamics is problematic as it imposes a spatial representation on to a temporal phenomenon. Sequential time periods are treated as adjacent vectors on a map, and functional representations of dynamic processes are akin to sets of directions for moving about this map. Such an account is misleading as it does not allow market participants to learn in unpredictable ways. Genuine learning (as manifested in entrepreneurial discovery) injects new knowledge into the competitive landscape—knowledge that did not exist prior to its discovery. Because this sort of learning modifies experience (knowledge) bases as well as specific expectations about future events, the passage of time (in and of itself) brings transformations that cannot be handled within simple spatial representations. With these insights, Austrian economists suggest a critique of the current treatment of innovative capabilities within resource-based theory. Before proceeding, however, it is important to impose some theoretical constraints on this subjective depiction of the market process. Within some variants of Austrian economics, appreciation of individual subjectivity and the concomitant aversion to deterministic accounts of market process are pushed to the extreme. Some theorists argue that a true appreciation of subjectivity must lead to scepticism about any purported regularities within the market process. An extreme subjectivist position (advocated by the likes of Brian Loasby) suggests that future economic activity is not only non-deterministic, it is also fundamentally random and irregular. While the epistemological debates relating to this extreme subjectivist stance are beyond the scope of this chapter, it is important to present a tenable middle-ground position that justifies stopping short of the extremely pessimistic stance advocated by Loasby (and others).

PETER W.ROBERTS 87

Kirzner (1992) suggests that Austrian economics occupies a middle ground between extreme deterministic and subjectivistic accounts of markets. On the one hand, mere recognition that individuals engage in entrepreneurial discovery mandates against the deterministic extreme. On the other hand, Kirzner (1992) and O’Driscoll and Rizzo (1996) both emphasise that forward-looking and intentional actions on the part of entrepreneurs are only made intelligible in light of perceived regularities within the market process. A subjective account of the market process respects that entrepreneurs are alert for previously unheeded profit opportunities, which come in the form of potential competitive advantages. The above description of the static dimension of resource-based theory suggests that a competitive advantage is supported by a positional capability that is both valuable and rare. The former criterion links a capability with the wants of some customer group, while the latter suggests a linkage with a cohort of competitors. Therefore, in order for a firm to anticipate a potential future positional capability, the future customer and competitor attributes must be (at least to a degree) predictable. This is not to say that the entrepreneur never errs in his/her predictions. However, there must be sufficient systematic tendencies within the competitive system to allow a potential entrepreneur to predict future customer and competitor attributes: ‘events that occur during the market process are subject to powerful constraining and shaping influences. These influences tend to reflect the relative urgencies of different consumer preferences, and the relative scarcities of different productive resources’ (Kirzner 1992:34). If these future attributes were truly random, then it would be meaningless to discuss the active pursuit of future competitive advantage. The next section outlines a perspective on innovative capabilities that is consistent with this middle-ground view of the market process. On the one hand, we formally recognise the presence of dynamic search routines (Nelson and Winter 1982), which drive a firm’s deliberate attempts to develop enhanced productive routines over time (by improving either a product or a service offering, or the process by which an existing offering is produced). These dynamic routines allow us to recognise the predictable and pathdependent aspects of evolutionary dynamics. However, dynamic routines, while constraining, do not fully determine the future path of the firm or the market context in which it competes. To varying degrees, firms engage in entrepreneurial discovery which generate options (and therefore paths) that are not wholly determined by pre-existing conditions. The mutual recognition of both aspects of an innovative capability (as well as their interaction) argues against both the deterministic and subjectivistic extremes: ‘the process of entrepreneurial learning is neither determinate nor random’ (O’Driscoll and Rizzo 1996:38). Innovative capabilities revisited In a recent exposition of an Austrian school of strategy, Jacobson (1992:803) notes that ‘Austrian thought and the resource-based theory of the firm…exhibit a great many similarities.’ At the same time, the two schools are thought to diverge in some important respects. In particular, ‘the resource-based perspective does not focus on the market

88 A PROCESSUAL ACCOUNT OF INNOVATIVE CAPABILITIES

process but rather makes use of equilibrium concepts and analysis’. An examination of the resource-based literature reveals a dominant emphasis on orthodox economic-type analysis which stresses equilibrium modelling (Barney 1986, 1991; Conner 1991; Peteraf 1993; Wernerfelt 1984). However, there is also an important (albeit somewhat underemphasised) component of resource-based theory that captures the spirit of the Austrian emphasis on entrepreneurial discovery and the unfolding market process (Schultze 1994; Foss et al. 1995; Foss 1997). In it, firms are challenged to be alert for opportunities to develop their resource bases in order to compete effectively into the future (Grant 1991).4 Building on the more processual aspects of resource-based theory, this section critiques the current treatment of innovative capabilities before proposing an integrative approach that recognises two different components of an innovative capability: dynamic routines and entrepreneurial alertness. In Austrian terminology, a positional capability exists within the context of a known means-ends framework. The means are a specific combination of resources and the ends are a linked set of outcomes that are valued by some group of customers. A firm with a unique valuable positional capability may be the only one possessing the knowledge of the meansends framework. Or, it may have (at least in the short run) proprietary access to some or all of the required resources that comprise that capability. In either case, the proprietary aspect of the capability gives this firm a competitive advantage. Over time, active current and potential competitors are expected to overcome these knowledge or resource asymmetries and erode the strategic significance of the proprietary positional capability. Because imitating firms are also working with an established means-ends framework, such acts of imitation are precisely the economising decisions referred to by Kirzner (1973). As such, it is non-problematic within an Austrian perspective on market process to infer the existence of, and the imitation of, a positional capability. The conflict between the Austrian and resource-based perspectives becomes evident when attention shifts to innovative capabilities. The current treatment within the resource-based literature implies a strong symmetry between positional and innovative capabilities. However, because the value of an innovative capability is derived from future positional capabilities, it is very much reliant on entrepreneurial discovery. It is therefore inappropriate to model innovative capabilities as if the future means-ends frameworks that give them value are already in place. Rather, they must be modelled in the context of the unfolding market process within which ignorance is eradicated through the process of discovery. The means-ends frameworks that underpin future positional capabilities are discovered with the passage of time by firms that are alert to previously unheeded profit opportunities. These entrepreneurial efforts are driven by what the firm knows and what it expects will unfold with the passage of time. However, each of these elements is subject to continuous revision as the firm augments its stock of knowledge and revises its expectations based on cumulated experience. With this in mind, one may question what portion of an innovative capability is vulnerable to imitation by competing firms. Subject to the usual caveats regarding the difficulties inherent in imitating intangible knowledgebased assets (Kogut and Zander 1992; Liebeskind 1996), it is conceivable that a

PETER W.ROBERTS 89

competing firm might copy an incumbent firm’s accumulated stock of knowledge at some point in time. This would entail replicating the individual skills, as well as the organisational routines that are embedded within the focal firm (Nelson and Winter 1982). Figure 5.3 provides a simple representation of a firm working towards the realisation of Outcome (1) at some future point in time. By definition, everything that is known by a firm at any point in time has been discovered (i.e., based on established means-ends frameworks). It is therefore theoretically possible for another firm to imitate that stock of knowledge and join the focal firm at Point A. The upper frame of Figure 5.3 illustrates a case in which there is no indeterminacy about the future and therefore there are no divergent expectations about which outcomes will be desired by customers and about what technological paths will deliver those outcomes. In such a case, a firm that has its existing stock of knowledge duplicated will find itself at parity with other firms as they chase this single outcome along a single path. Just as the previous comments by Cockburn and Henderson (1994), Comanor (1994) and Temin (1979) cast doubt upon the empirical robustness of race models of innovation, Austrian economists cast further doubt on the assumption that, in an unfolding competitive market, the precise nature of the future outputs of an innovative capability are knowable at the time an innovative capability is imitated. Only in a deterministic world —where future means-ends frameworks are knowable in advance—can one talk about a one-to-one correspondence between an innovative capability at some point in time and the subsequent positional capabilities that it generates. In the words of O’Driscoll and Rizzo (1996:5) there is ‘no stable endpoint toward which the process must lead, nor a single path it must follow’. Consider the lower two panels of Figure 5.3. The middle panel abstracts away any technological uncertainty. Here, even if one assumes that a competitor can fully replicate the knowledge base of a focal firm at some point in time, one cannot assume that each firm will have identical expectations of what customers might desire in the future. As such, even with similar knowledge bases, the innovative behaviour on the part of each firm may be directed towards different conceived outcomes. Such complexity is amplified in the lower panel of Figure 5.3, which allows for technological (in addition to market) indeterminacy. Here, even two firms that have similar expectations about desired future outcomes may choose different technological routes to their target destinations. Finally, this processual account is made even more complex by the fact that firms update their knowledge bases (as well as their expectations) as their respective experiences accumulate (Hayek 1937). Therefore, one should expect that even in the extreme case of perfect knowledge imitation at some point in time, two firms will have increasingly divergent knowledge bases—and therefore different innovative capabilities—over time as their idiosyncratic innovative investment decisions cumulate. This point is echoed by Hunt and Morgan (1995:9), who state that ‘each firm in an industry is a unique entity in time and space as a result of its history’. One may take this critique a step further and question the extent to which it makes sense to acknowledge innovative capabilities taking a processual perspective on innovation and entrepreneurship. An innovative capability is novelty-creating. Assuming less-than-

90 A PROCESSUAL ACCOUNT OF INNOVATIVE CAPABILITIES

Figure 5.3 Non-deterministic representations of innovation

perfect knowledge of future means-ends frameworks, a novel positional capability cannot be fully known in advance of its discovery. One cannot, therefore, draw a definitive connection between an innovative capability and its subsequent innovative outputs. Therefore, if an innovative capability is defined with reference to future valuable and rare positional capabilities, it quite simply cannot fully exist. With this in mind, one may point to firm attributes that have meaningful dynamic properties and which may be considered a subset of innovative capabilities. According to Teece and Pisano (1994:541), ‘dynamic capabilities are the subset of the competencies/ capabilities which allow the firm to create new products and processes, and respond to

PETER W.ROBERTS 91

Figure 5.4 Innovative capabilities: a reconciliation

changing market processes’. The latter part of this definition suggests that some firms are more flexible than others and that flexibility aids in the pursuit of sustainable competitive advantages. To the extent that certain organisational features are consistent with the ability to change as circumstances dictate, such responsiveness may indeed be a key component of an innovative capability. However, one must clearly articulate the source of change to which the firm is responding. Responsiveness might be associated with more effective imitation (as with a fast-follower strategy). In such instances, the acts of discovery that deliver the novel means-ends frameworks occur outside the firm in question. Without any follow-on discovery on the part of the firm, such a dynamic capability will only deliver the firm into a world of competitive parity. In other words, responsiveness by itself only makes a firm a more effective economiser. References have also been made to a firm’s dynamic routines, e.g., new product development routines (Nelson and Winter 1982). A scenario without indeterminacy may still contain uncertainty. This is represented in the upper panel of Figure 5.3 by the physical separation between Point A and the future realisation of Outcome (1). Successful firms must work out the uncertainty that characterises the problem space connecting the current vision of some future output and its ultimate realisation. In such cases, firms with more refined and/or more developed search routines might be advantaged in the pursuit of a given technology/demand combination. Again in reference to Figure 5.3, these dynamic routines affect the pace at which a firm moves along the prescribed path towards Outcome (1). The dynamic search routines that propel firms down a given path whose endpoint is characterised by expectations about a desired end, as well as appropriate means, may indeed be established ex ante. However, the same cannot be said about the complementary expectations that firms hold about what these means and ends should be. This suggests that search routines are considerably more valuable in the presence of entrepreneurial discovery because any given search routine has more strategic significance if it is invoked uniquely within a more promising problem space (see Figure 5.4). And, while the specific search technologies that characterise a firm’s dynamic routines exist and may be imitated, the entrepreneurial alertness that gives it added competitive significance cannot.

92 A PROCESSUAL ACCOUNT OF INNOVATIVE CAPABILITIES

The joint recognition of dynamic routines and entrepreneurial alertness moves the current treatment of innovative capabilities away from the nihilistic subjectivist extreme outlined in the previous section. Just as dynamic routines are given enhanced strategic significance in the presence of entrepreneurial alertness, so too is discovery enhanced by prevailing dynamic routines. Ultimately, the value of an innovative capability derives from its ability to deliver future positional capabilities that support competitive advantage. In addition to discovering an opportunity, the alert firm must also come to possess and/ or control the specific resources that will comprise the new positional capability once the underlying knowledge has been discovered (Lengnick-Hall 1992). Therefore, firms have some incentive to be alert for opportunities that make use of their current static and dynamic routines. At the same time, we have seen that a firm’s entrepreneurial actions are conditioned by its evolving knowledge base and set of expectations, each of which is affected by its (unique) accumulated experiences. Because a firm’s dynamic search routines affect the nature of these experiences, it is likely that a firm’s discoveries will be affected by its dynamic routines. These arguments each suggest a propensity for a firm to discover opportunities that are ex post related to its current set of positional capabilities and dynamic routines. However, this does not imply a deterministic causal link to future positional capabilities. To see this point, consider the emerging dynamic capabilities perspective within which firms are propelled through time by the confluence of their specific processes, paths and positions (Teece et al. 1997). The linkages between this and the discussion surrounding Figure 5.3 are quite obvious. A firm at Point A is propelled along some path by its built-in competitive processes (or dynamic routines). The current discussion augments the dynamic capabilities framework by insisting that the paths ahead of a firm are not fully established ex ante. Nor will they be revealed with passive attention to existing market data. Rather, the firm must be alert to currently unheeded opportunities (it must engage in entrepreneurial discovery) before these paths may become evident. Moreover, as the specific knowledge and expectations that define these future paths are under a constant state of revision, these paths may not become fully evident until very near the end of the process. At the very least, this Austrian refinement confirms a strong and active component for firm managers in creating a firm’s potential (Hamel and Prahalad 1994). At the same time, the fundamental inimitability of entrepreneurial alertness suggests that firms that do uncover these latent opportunities may expect the sustainable competitive advantage that Collis (1994) suggests is so fleeting. Concluding remarks Given the comparative oversight of ideas developed within the Austrian school of economics, the resource-based view of the firm has come to treat positional and innovative capabilities in more-or-less symmetric fashion in theorising about sustainable competitive advantage. A more processual account of market evolution suggests that innovative capabilities should be analysed in a manner that is consistent with entrepreneurial discovery. Formal recognition that the means-ends frameworks that

PETER W.ROBERTS 93

underpin future positional capabilities are not fully known in advance of their discovery severs any direct connection between a given set of firm attributes and the subsequent pattern of innovative outputs. Such recognition calls into question what particular aspects of an innovative capability may be said to exist and therefore subject to competitor imitation. While it is conceivable (although not always possible) that a firm may copy another firm’s stock of knowledge at a point in time, it is less obvious how it might imitate that firm’s specific expectations about future market or technological outcomes. As such, to the extent that entrepreneurial discovery is driven by a firm’s specific knowledge and expectations, it is hard to envision the complete imitation of an innovative capability. This concern is amplified if we consider evolution in the context of subjective time. In this respect, the series of specific actions of a firm (whether they are associated with perceived competitive successes or failures) leads to revisions in its knowledge base and its expectations about future events. Because these actions will be firm-specific (both in content and context), an imitating firm will find it impossible to maintain an identical knowledge base as time unfolds. The most extreme subjective position on innovative capabilities is that they cannot exist at all. Because of its unique properties, entrepreneurial alertness cannot be classed among the firm’s stock of capabilities: the value of it is necessarily dependent on a stream of future acts of discovery. Because of the indeterminacy surrounding future means-ends frameworks, it is impossible to assign value to or transact, trade or transfer such entrepreneurial alertness. With this in mind, it is useful to envision a firm’s stock of dynamic search routines as a valuable (albeit incomplete) component of its innovative capability. By applying these routines, the firm purposefully searches within specified problem spaces for solutions to current problems. However, it is critical that one does not come to see these dynamic routines as fully comprising an innovative capability. Such a deterministic account would not allow for the prospect of genuine learning within which firms discover opportunities that were not directly inferable from existing market data. Therefore, a robust account of innovative capabilities should formally recognise both dynamic routines and entrepreneurial alertness. We close the chapter with a call for a broader discussion of the implications of Austrian economics (and its emphasis on the evolving market process) for strategy research. One issue that follows from this discussion relates to the conditions that might move this more processual account of innovative capabilities in the direction of the race model depicted in the upper panel of Figure 5.3. Dosi (1982:148) defines a technological paradigm as ‘an outlook, a set of procedures, a definition of the relevant problems and of the specific knowledge related to their solution’, and a technological trajectory as ‘the direction of advance within a technological paradigm’ (p. 148). In a forward-looking sense, the presence of such paradigms might very well channel the expectations of the various market participants, collapsing a myriad of conceivable future states of the world on to a more finite set. In the extreme, expectations may be so aligned as to deliver the race-type behaviour illustrated in the upper panel of Figure 5.3. It is important to stress that technological paradigms are purportedly social, as much as economic or technological phenomena. To the extent that such social phenomena may influence ex ante perceptions of

94 A PROCESSUAL ACCOUNT OF INNOVATIVE CAPABILITIES

the competitive and technological landscape, they have a rightful place within an Austrian depiction of the market process. However, it would be erroneous to suggest that observation of a pattern in a series of historical data equates to an ex ante constraint on the entrepreneurial activity of market participants (Henderson 1995). In fact, Schumpeter (1934) argues that the innovator role is distinct from other economic roles precisely because it embodies the propensity to work against perceived wisdom at a point in time when the value of the proposed endeavour may not be completely obvious. To the extent that technological paradigms reflect this perceived wisdom, entrepreneurs (or innovators) may expect considerable gain off the prevailing technological trajectory. Formal recognition of the different expectations that underlie this alertness to new possibilities distinguishes the Austrian conception of an innovative capability. It is also precisely what makes such capabilities fundamentally inimitable. Another point for consideration relates to the potential interplay between evolutionary and Austrian economics. In a recent article, Foss (1997) reviews the potential contributions of both Austrian and evolutionary economics to the development of resource-based theory, stating a strong preference for the latter. However, this chapter suggests that this need not be an either/or proposition. Foss (1997) is quite correct in insisting that theoretical plurality is not a sufficient condition for accepting contributions across a wide range of contributing sources, especially when there are fundamental contradictions across contributors. However, Nelson and Winter (1982) do recognise the contributions of Austrian economics in the development of their evolutionary perspective. The discussion in this chapter alludes to the potential for dynamic routines (and evolutionary construct) to operate alongside entrepreneurial alertness and to a strong potential for meaningful interplay across the two components of innovative capabilities. Moreover, a combined Austrian-evolutionary perspective is also consistent with the often observed ex post path dependencies in capability development. In light of this, future research should further explore the implications of a joint Austrian-evolutionary perspective on sustainable competitive advantage. There is no question that the development of an Austrian resource-based research agenda (which is necessarily broader than the above two examples would suggest) represents a strong challenge, given the difficulties inherent in moving from a more established research paradigm (Montgomery 1995). However, it will undoubtedly improve the dynamic aspects of strategy research, and is therefore worth the effort. Notes 1 For example, consider that at least two of the pharmaceutical firms that would have demonstrated above-average innovative propensity during Henderson and Cockburn’s (1994) sample time period (Merck and Glaxo) have recently experienced significant declines in the rate of radical drug innovation. 2 Although Austrian economics addresses issues that are broader than those relating to innovative capabilities economic theorising, this chapter focuses on contributions that describe the role of entrepreneurship in the unfolding market process.

PETER W.ROBERTS 95

3 Indeterminacy differs from uncertainty in that the latter assumes a given set of future states with a knowable probability distribution describing the likelihood that each state will occur, while indeterminacy implies that the set of possible future states is in no way knowable with any degree of confidence. 4 Resource-based scholars underemphasise the role played by entrepreneurial competence in Penrose’s (1959) thesis. Alertness to opportunities is a necessary precondition for the exploitation of underutilised resources: ‘the decision to search for opportunities is an enterprising decision requiring entrepreneurial intuition and imagination and must precede the “economic” decision to go ahead with the examination of opportunities for expansion’ (Penrose 1959:34).

References Amit, R. and P.Schoemaker (1993) ‘Strategic Assets and Organisational Rent’, Strategic Management Journal 14:33–46. Barney, J. (1986) ‘Strategic Factor Markets: Expectations, Luck, and Business Strategy’, Management Science 35:1231–41. ——(1991) ‘Firm Resources and Sustained Competitive Advantage’, Journal of Management 17:99– 120. Brumagim, A. (1994) ‘A Hierarchy of Corporate Resources’, Advances in Strategic Management 10A: 81–112. Cockburn, I. and R.Henderson (1994) ‘Racing to Invest? The Dynamics of Competition in Ethical Drug Discovery’, Journal of Economics and Management Strategy 3:481–519. Cohen, W. (1995) ‘Empirical Studies of Innovative Activity’, in P.Stoneman (ed.), Handbook of the Economics of Innovation and Technological Change, Oxford: Blackwell Publishers. Collis, D. (1994) ‘How Valuable are Organisational Capabilities?’, Strategic Management Journal 15: 143–52. Comanor, W. (1994) The Pharmaceutical Industry. Working Paper, University of California, Santa Barbara and Los Angeles. Conner, K. (1991) ‘A Historical Comparison of Resource-Based Theory and Five Schools of Thought within Industrial Organization Economics: Do We Have a New Theory of the Firm?’, Journal of Management 17:121–54. D’Aveni, R. (1994) Hypercompetition: Managing the Dynamics of Strategic Manoeuvring, New York: The Free Press. Dolan, E. (1976) ‘Austrian Economics as Extraordinary Science’, in E.Dolan (ed.), The Foundations of Modern Austrian Economics, Kansas City: Sheed and Ward. Dosi, G. (1982) ‘Technological Paradigms and Technological Trajectories’, Research Policy 11:147– 62. Foss, N. (1997) ‘The Resource-Based Perspective: An Assessment and Diagnosis of Problems’, Scandinavian Journal of Management, forthcoming. Foss, N., C.Knudsen and C.Montgomery (1995) ‘An Exploration of Common Ground: Integrating Evolutionary and Strategic Theories of the Firm’, in C. Montgomery (ed.), Resource-Based and Evolutionary Theories of the Firm: Towards a Synthesis, Boston, MA: Kluwer Academic. Grant, R. (1991) ‘The Resource-Based Theory of Competitive Advantage: Implications for Strategy Formulation’, California Management Review Spring:114–35. Hamel, G. and C.K.Prahalad (1994) Competing for the Future, Boston, MA: Harvard Business School Press.

96 A PROCESSUAL ACCOUNT OF INNOVATIVE CAPABILITIES

Hayek, F. (1937) ‘Economics and Knowledge’, Economica 3:33–54. Henderson, R. and I.Cockburn (1994) ‘Measuring Competence? Exploring Firm Effects in Pharmaceutical Research’, Strategic Management Journal 15:63–84. Hunt, S. and R.Morgan (1995) ‘The Comparative Advantage Theory of Competition’, Journal of Marketing 59:1–15. Jacobson, R. (1992) ‘The Austrian School of Strategy’, Academy of Management Review 17:782–807. Kirzner, I. (1973) Competition and Entrepreneurship, Chicago: University of Chicago Press. ——(1976) ‘On the Method of Austrian Economics’, in E.Dolan (ed.), The Foundations of Modern Austrian Economics, Kansas City: Sheed and Ward. ——(1992) The Meaning of Market Process: Essays in the Development of Modern Austrian Economics, London: Routledge. Kogut, B. and U. Zander (1992) ‘Knowledge of the Firm: Combinative Capabilities, and the Replication of Technology’, Organization Science 3:383–97. Lachmann, L. (1976) ‘On the Central Concept of Austrian Economics: Market Process’, in E.Dolan (ed.), The Foundations of Modern Austrian Economics, Kansas City: Sheed and Ward. Lengnick-Hall, C. (1992) ‘Innovation and Competitive Advantage: What We Knowand What We Need to Learn’, Journal of Management 18:399–429. Leonard-Barton, D. (1992) ‘Core Capabilities and Core Rigidities: A Paradox in Managing New Product Development’, Strategic Management Journal 13:111–25. Liebeskind, J. (1996) ‘Knowledge, Strategy, and the Theory of the Firm’, Strategic Management Journal 17 (Special Issue):93–107. Lippman, S. and R.Rumelt (1982) ‘Uncertain Imitability: An Analysis of Interfirm Differences in Efficiency under Competition’, Bell Journal of Economics 13:418–38. Mahoney, J. and J.Pandian (1992) ‘The Resource-Based View within the Conversation of Strategic Management’, Strategic Management Journal 13:363–80. Montgomery, C. (1995) ‘Of Diamonds and Rust: A New Look at Resources’, in C. Montgomery (ed.), Resource-Based and Evolutionary Theories of the Firm: Towards a Synthesis, Boston, MA: Kluwer Academic. Nelson, R. and S.Winter (1982) An Evolutionary Theory of Economic Change, Cambridge, MA: Harvard University Press. O’Driscoll, G. and M.Rizzo (1996) The Economics of Time and Ignorance, New York: Routledge. Penrose, E. (1959) The Theory of the Growth of the Firm, New York: John Wiley and Sons. Peteraf, M. (1993) ‘The Cornerstones of Competitive Advantage’, Strategic Management Journal 14: 179–92. Reed, R. and R.DeFillippi (1990) ‘Causal Ambiguity, Barriers to Imitation, and Sustainable Competitive Advantage’, Academy of Management Review 15:88–102. Rothbard, M. (1976) ‘Praxeology: The Methodology of Austrian Economics’, in E. Dolan (ed.), The Foundations of Modern Austrian Economics, Kansas City: Sheed and Ward. Rumelt, R. (1984) ‘Toward a Strategic Theory of the Firm’, in R.Lamb (ed.), Competitive Strategic Management, Englewood Cliffs, NJ: Prentice-Hall. ——(1987) ‘Theory, Strategy and Entrepreneurship’, in D.Teece (ed.), The Competitive Challenge: Strategies for Industrial Innovation and Renewal, Cambridge, MA: Ballinger. Schultze, W. (1994) ‘The Two Schools of Thought in Resource-Based Theory: Definitions and Implications for Research’, Advances in Strategic Management 10A: 127–51. Schumpeter, J. (1934) The Theory of Economic Development, Cambridge, MA: Harvard University Press. Teece, D. and G.Pisano (1994) ‘The Nature and Importance of Economic Competence’, Industrial and Corporate Change 3:537–56.

PETER W.ROBERTS 97

Teece, D., G.Pisano and A.Shuen (1997) ‘Dynamic Capabilities and Strategic Management’, Strategic Management Journal 18:509–34. Temin, P. (1979) ‘Technology, Regulation and Market Structure in the Modern Pharmaceutical Industry’, Bell Journal of Economics 10:429–46. Wernerfelt, B. (1984) ‘A Resource-Based View of the Firm’, Strategic Management Journal 5:171– 80. Williams, J. (1992) ‘How Sustainable Is Your Competitive Advantage?’, California Management Review Spring:29–51.

98

6 Product and process architectures in the management of knowledge resources Ron Sanchez

Introduction In both theory and practice, effective management of knowledge resources is increasingly recognised as critical to an organisation’s ability to derive economic value from its uses of both tangible and intangible resources. Because no one can realistically hope to manage well that which one does not understand in at least fundamental respects, the task of managing knowledge resources begins with the challenge of developing a conceptual framework for identifying and defining the specific knowledge resources that are important in helping a given organisation carry out its ‘strategic intent’ (Hamel 1989a). Given some kind of framework that enables identification of the strategically useful knowledge resources available to an organisation, the next challenge is to develop effective approaches to using existing knowledge in leveraging current organisational competences and to creating and acquiring new knowledge in building new competences (Sanchez and Heene 1996a, 1996b). The challenge of creating more effective management systems for identifying, acquiring or accessing, coordinating and augmenting strategically important knowledge is thus an essential dimension of the ‘contest between managerial cognitions’ that characterises competence-based competition (Sanchez et al. 1996). Foss and Eriksen (1995) have characterised technology as consisting of both firmspecific capabilities and non-proprietary ‘industry capabilities’. In the vocabulary of competence-based strategic management, technology may therefore be either a ‘firmspecific’ or a ‘firm-addressable’ resource (Sanchez et al. 1996). Managing technologybased knowledge resources requires making some difficult but strategically important decisions about where to position a firm in an ‘external resources and capabilities space’ (Foss and Eriksen 1995:45). A firm chooses a resource and capability position close to those of other firms when it pursues technology and product strategies that access employees and knowledge that exist as a firm-addressable ‘industry capability’ because other firms are also using similar resources. Conversely, a firm may distance itself in resources and capability space by staking out a technology strategy and product strategy that requires idiosyncratic, firm-specific technology resources (or at least resources that are not usually available to competing firms). Within this framework, a close positioning

100 PRODUCT AND PROCESS ARCHITECTURES

of firms in the external resources and capabilities space may bring certain benefits in the form of network externalities (discussed below), but the potential to earn rents within an industry depends on finding some dimension(s) in which a firm can distance itself from its competitors by developing and using firm-specific resources and capabilities. Some important benefits may accrue to firms that are participants in an industry or a cluster of firms within an industry that share technology resources. Saxonian (1991) and others have argued that use of common technologies may achieve sufficient scale within an industry that firms can begin to specialise in various aspects of a technology, leading to greater technological efficiencies through specialisation of labour. In addition, users of common technologies often constitute a technology network that facilitates ‘information exchange and joint problem solving between firms’ (Saxonian 1991:410), a notion which reflects Dosi’s (1982) observation of the use of standardised methods for solving technological problems within a ‘technological regime’ (Nelson and Winter 1982). Farrell and Saloner (1985) have also suggested that the use of standards within a technological regime appear to accelerate technological progress in an industry using that regime. This chapter seeks to extend these conceptualisations about technology as both a firmspecific and a firm-addressable resource by investigating the roles of product and process architectures in the strategic management of technology-based knowledge resources. The following discussion proposes that both the products which an organisation creates and the processes which it follows in creating, producing and supporting its products have architectures which greatly influence the structures of much of the organisation’s strategically useful knowledge about technologies and markets. Clearly defining an organisation’s product and process architectures is argued to be a critical step both in clarifying the nature of the technological knowledge resources an organisation has or has access to, and in delineating the relationships between an organisation’s knowledge resources—which I refer to as an organisation’s technological knowledge architecture (Sanchez 1996a). I also suggest that alternative product and process architectures an organisation may adopt have greatly different potentials for enabling identification of both existing organisational knowledge resources and opportunities for building strategically important new knowledge. In particular, the use of modularity in creating product and process architectures is argued to lead to greater clarity in identifying key current organisational knowledge resources and to more effective targeting of strategically useful organisational learning. Some of the major implications of introducing the concepts of product, process and knowledge architectures into an analysis of technology resources are these: 1 Adding an architectural dimension to our characterisation of technology brings into focus the structure of a technology regime and the forms in which technological specialisation can occur within that structure. The increasing use of standard product, process and knowledge architectures by a group of firms creates opportunities for firms to specialise in developing and producing standardised components for products, in providing standardised processes as manufacturing and

RON SANCHEZ 101

product support services, and in applying standardised expertise in solving problems that arise in developing new products and processes within industry architectures (Sanchez and Mahoney 1996). 2 The theory of competence-based competition has sought to move beyond representations of resources in which the value of a resource is treated as independent of its context of use. The competence perspective emphasises that the economic value of resources can only be realised when they are coordinated in a given use. Thus, value may be created (or rents generated) not just through the possession of superior resources, but through superior coordination and targeting of resources (Sanchez et al. 1996). Standard product, process and knowledge architectures used in an industry create information structures that, in effect, embed coordination of industry participants’ uses of resources (Sanchez 1996a) by defining standard inputs and outputs for components in product designs, standard processes for manufacture and support, and standard knowledge applied to problem solving. Thus, in addition to the gains in efficiencies that can result from the specialisation of technological resources within a technological regime, the adoption of industry standard architectures for products, process and knowledge can lead to gains in efficiencies by reducing the costs and difficulties of coordinating the use of specialised technological resources. 3 As Foss and Eriksen (1995:59) have suggested, ‘some measure of coordination of asset-accumulation processes often takes place, to the benefit of participating firms’ when technology development within an industry leads to standardisation. As this chapter proposes, standard—and especially modular (Langlois and Robertson 1992)— product and process architectures provide frameworks that may help a firm to clarify its knowledge about products and processes, and thereby help to identify opportunities for acquiring or building (‘accumulating’) useful new knowledge resources. 4 If firms have differing abilities to use product and process architectures to define opportunities to leverage existing technological knowledge more effectively and to identify and acquire strategically valuable new learning, then a firm’s ability to use and create knowledge resources within the framework of industry standard architectures becomes an important dynamic dimension to be added to the industry resources and capabilities space suggested by Foss and Eriksen (1995). In effect, a firm’s potential for generating rents will be determined not solely by the position it chooses in the resources and capabilities space, but also by its differential ability to use and create technological knowledge in whatever position it occupies. This suggests that the choice to be made in the resources and capabilities space is not one in which the benefits of close positioning must be strictly traded off against the benefits of distancing. A firm that has a superior ability to learn within an industry standard architecture shared with other firms, for example, may be able to reap the efficiency benefits of using industry-level knowledge and other resources that result from close positioning in the shared architecture, but at the same time may be able to generate rents by accumulating firm-specific knowledge about how to use (coordinate and target) resources more effectively within an industry architecture. In this way, architectures

102 PRODUCT AND PROCESS ARCHITECTURES

may function as a key mechanism for coordinating interactions between firmaddressable resources available from the technology regimes of industries and internally developed, firm-specific technology resources. 5 Product and process architectures provide a framework that clarifies the ways in which organisational knowledge and learning can occur in three modes: know-how, know-why and know-what (Sanchez 1996a). Analysis of the architectures associated with technology regimes therefore provides a framework for finer characterisation of firms’ differential abilities to use and acquire various forms of knowledge, and for identifying different kinds of organisational learning strategies. 6 Competing product and process architectures within industries may have inherently greater or lesser efficiencies as coordination mechanisms. Comparisons of architectures may therefore help to explain differential performance levels among groups of closely positioned firms (strategic groups). For example, one could compare the performance of firms using the Windows product architecture with the performance of firms using the Macintosh product architecture. Further, since firms in one product market also compete against firms offering substitute products, the relative efficiencies or inefficiencies of the product and process architectures in common use in competing industries may help to explain differential performance levels of those industries. For example, the relative performance of competing energy industries may result in part from the relative efficiencies of the industry architectures for providing electricity vs. architectures for providing natural gas. In such comparisons, performance may be evaluated in both static and dynamic efficiency terms (i.e., architectural efficiencies in using and in creating new resources). This discussion is organised in the following way. The first section defines product architectures and distinguishes modular product architectures from other kinds of product architectures. The next section defines process architectures and, drawing on the notion that ‘products design organisations’ (Sanchez and Mahoney 1996), elaborates the ways in which product architectures used within an organisation constrain the feasible process architectures an organisation can adopt in creating and producing its products. The discussion then explains how use of modular product architectures enables adoption of modular process architectures. The third section explains the ways in which organisational knowledge becomes structured around product and process architectures. The clarity with which an organisation can determine the current states of its component-level and architecturallevel knowledge within its knowledge structures determines an organisation’s ability to identify opportunities to develop specific new knowledge resources of strategic value to the organisation. By virtue of their greater decomposition into ‘loosely coupled’ systems of components, both modular product architectures and modular process architectures make possible more loosely coupled knowledge domains, which in turn makes possible more precise identification of specific knowledge resources at both component and architectural levels (Sanchez and Mahoney 1996). The improved ability to distinguish current states of

RON SANCHEZ 103

knowledge and their modes of interaction within modular architectures provides a frame of reference which facilitates greater precision in identifying opportunities for acquiring strategically valuable new forms of knowledge—i.e., for more effective strategic targeting of organisational learning. The fourth section proposes that organisations have different kinds of knowledge within their knowledge architectures and suggests how adoption of modular product and process architectures improves the ability of an organisation to leverage and create know-how, knowwhy and know-what forms of knowledge (Sanchez 1996a, 1996b). Consequently, organising knowledge resources within a modular architecture may be a superior approach to mediating the technological and market uncertainties facing organisations, by enhancing organisational learning capabilities. The fifth section concludes with some observations on the increasing prevalence of modular product and process architectures in rapidly evolving areas of knowledge. I suggest that one explanation for this observed association is that modular knowledge architectures create substantial positive network externalities that act to accelerate learning processes within architectural alliances of firms or within industries. However, since the creation and use of modular architectural infrastructures that accelerate learning are often ‘taken for granted’ by practitioners in rapidly evolving areas of technology and market development, the critical role of modularity in facilitating both firm-level learning processes and industry-level technological progress is often overlooked by management researchers interested in understanding knowledge resources and the processes that create these key resources. 1 Product architectures Most products—including service and software products as well as assembled goods— perform several discrete functions that combine together to provide the set of functionalities that distinguish one product from another in the marketplace. After determining the desired package of functionalities or ‘bundles of product attributes’ (Bogner and Thomas 1996) to be provided by a new product, product designers create a new product design by decomposing the desired set of product functionalities into a system of functional components whose individual functions collectively interact to provide the overall functionalities desired in the product (Sanchez 1996b). A product architecture is created by (a) decomposing a product design into a system of functional components (Henderson and Clark 1990), and (b) fully specifying how individual components will interact with other components in that system of components (Sanchez 1995; Sanchez and Mahoney 1996).1 The component interface specifications in a product architecture define, for example, how one component may be physically connected to another (the attachment interface), how power is to be transferred between components (the transfer interface), how signals will be exchanged between components (control and communication interfaces), the spatial location and volume a component may occupy (spatial interfaces), and various ways in which the functioning of one component

104 PRODUCT AND PROCESS ARCHITECTURES

may generate heat, magnetic fields or other environmental effects that must be accommodated by other components (environmental interfaces) (Sanchez 1994). In this sense, a product architecture is a more foundational concept in technology than a ‘product family’ targeted at a market segment (Meyer and Utterback 1993), because some kinds of product architectures enable families of products to be leveraged from a single architecture, while others do not. Architectures are also broader than ‘platforms’ targeted at a market niche (Meyer and Utterback 1993), because they may enable leveraging of product variations for several niches or even several market segments (Sanchez 1996b; Sanchez and Mahoney 1996; Sanderson and Uzumeri 1997). There are two fundamentally different architectural approaches to defining component interfaces in a product design, which will be distinguished here as conventional and modular approaches to creating product architectures. The essential differences between conventional and modular approaches to defining, designing and developing new product architectures are summarised in Table 6.1. As suggested in Table 6.1, the conventional product design process usually relies on extensive marketing research to suggest the specific product functionalities, performance levels and maximum price that will appeal most broadly to a targeted set of customers. Given a set of marketing-determined optimal product attributes, the objective of product designers is to create an optimal product design that provides the desired product attributes at the lowest possible cost or the highest possible level of product performance within a specified cost constraint. The product architectures created by a conventional product design process are typically complex designs in which technically separable functional components have been combined into integrated assemblies of components to increase performance and/or to lower costs. Development of such designs is often timeconsuming and costly, however, because making changes in individual component designs during development to improve the overall product design may require compensating changes in the designs of many closely interrelated or ‘tightly coupled’ (Orton and Weick 1990) components. In contrast to conventional product architectures, a modular product architecture is intentionally designed to permit the ‘substitution’ of different versions of functional components (Garud and Kumaraswamy 1993) for the purpose of creating product variations with different bundles of functionalities, features and performance levels. In effect, the task of modular product design is to create a flexible product architecture or ‘platform design’ (Wheelwright and Sasser 1989; Sanderson and Uzumeri 1997) for introducing component variations that create product variations for serving a range of market requirements. To create a flexible product architecture for leveraging product variety (Sanchez 1995), modular product design follows a discipline of avoiding ‘designing in’ tight interdependencies among component designs and instead seeks to create loosely coupled component designs in which any design variation in one type of component (within a specified range of variation) will work in conjunction with any design variation in another type of component (within its specified range of variation). The most familiar example of this property of modular architectures is probably the personal computer,

Table 6.1 Comparison of product definition, design and development in creating conventional versus modular product architecture

RON SANCHEZ 105

106 PRODUCT AND PROCESS ARCHITECTURES

which generally allows ‘plug and play’ system configuration of various hard disk drives, memory cards, monitors, etc. Although the focus in this chapter is on the impacts of product architectures on the knowledge structures an organisation develops, it is worth noting here that the ‘designedin’ flexibility of modular product architectures has stimulated their adoption in a growing number of industries where their flexibilities are being put to several strategically important uses (Sanchez 1995, 1996a, 1996b): • Increased product variety. The ability to ‘mix-and-match’ components in a modular product architecture enables leveraging of greater product variety than can be obtained from a conventional product architecture. In addition to the familiar example of the personal computer, this property of modular product architectures is now being used strategically to offer expanded ranges of product variations in industries as diverse as automobiles, consumer electronics, home appliances, object-orientated software, banking services and power tools (Langlois and Robertson 1992; Sanchez 1994; Sanchez and Mahoney 1996). • Faster introductions of upgraded products. Modular product architectures may be designed to accommodate both currently available components and technologically improved components that are expected to become available within the intended lifetime of the product architecture. The ability to introduce better components directly into a modular product architecture without having to make design changes in the product architecture enables the speedy introduction of upgraded product models as soon as improved components become available. As a result, competitive product strategies driven by rapid product improvements become possible for firms that are able to create robust modular product architectures which let them introduce new components directly into their existing product architectures. Sony’s use of a modular product architecture for its introduction of the HandyCam video camera, for example, enabled Sony to introduce four upgraded models within a 22-month period by quickly incorporating improved components into the HandyCam product architecture (Sanchez and Sudharshan 1993). • Lower design, production, distribution and service costs. Leveraging several product variations from a single modular product architecture may substantially reduce design costs per product variation introduced. Economies of scale can also be realised when ‘common components’ that can be used across product variations are produced in large numbers. When the component variations that differentiate an individual product model can be added at a ‘late point’ in the production or distribution process, the reduced variety of parts used in product assembly and inventories reduce production and distribution costs per product variation introduced. General Electric’s different stove models, for example, are differentiated by cooktops and control panels added in the last stages of the assembly process, while high-speed production lines assemble the common components used in all GE’s stove models. Taking this concept one step further, Hewlett-Packard differentiates its ink-jet printers by adding a power supply to suit local electrical requirements and a local language instruction manual in its regional

RON SANCHEZ 107

distribution centres around the world (Lee et al. 1993). Modular product architectures may also significantly change product service requirements by enabling the ‘design-in’ of self-diagnostic capabilities and easily replaced modular components that radically lower field service costs. 2 Process architectures In its most basic sense, a system is a structure of parts that interact with each other to some degree. Simon (1981) argues that hierarchical decomposability2 is an organising principle of complex systems, whether the system be a product design or an organisation design. Henderson and Clark (1990) and others have observed that organisations tend to organise their knowledge creation and application processes in groups whose activities are focused on the functional components in their product designs. In other words, the structure of the decomposition of component parts within a product design will tend to be reflected in the structure of the decomposition of the activities of the organisation that creates and produces the product. Adding the observation that the specific tasks that must be performed in developing and producing a product are largely determined by the specific product architecture and component designs a firm adopts, Sanchez and Mahoney (1994, 1996) propose that although organisations ostensibly design products, in a fundamental sense ‘products design organisations’. They further suggest that the use of conventional vs. modular product architectures within an organisation imposes very different sets of constraints on the feasible organisational structures—or process architectures —an organisation can adopt in creating and producing its products. Product designs can be fundamentally distinguished by the degree to which the architecture of each product has been decomposed into tightly coupled versus loosely coupled (Orton and Weick 1990) components.3 Conventional engineering design methodology for optimising product designs to meet defined market preferences usually results in product designs composed largely of tightly coupled components combined in integrated assemblies to improve performance or reduce costs. As a consequence, processes for developing products composed of tightly coupled components typically require intensive communication and coordination across all component development groups, since a change in the design of one component during development may require extensive compensating changes in many other components being developed by other groups. As a further consequence, product designs composed of tightly coupled components will typically require tightly coupled development processes, in the sense that continual exercise of overt managerial authority will be needed to coordinate and adjudicate across highly interdependent component development processes. The exercise of overt managerial authority, in turn, will typically require an authority hierarchy of the sort found within the boundaries of a single firm or within a quasi-integrated group of suppliers of the sort often observed in Japan (Sanchez and Mahoney 1996). Tightly coupled component designs may also necessitate tightly coupled production and service activities that must be carried out within a single authority hierarchy.

108 PRODUCT AND PROCESS ARCHITECTURES

A modular product architecture, in contrast, achieves the loose coupling of components by specifying the input and output interfaces between components to allow for a range of component variations to be used within the product architecture. Moreover, when these ‘flexible’ modular component interface specifications are standardised—i.e., not allowed to vary over some intended period of time—they create a stable information structure of required component input and output specifications (Sanchez and Mahoney 1996). The stable information structure of the component interface specifications in a modular product architecture provides a vehicle for embedded coordination of loosely coupled component development (and possibly production and service) processes (Sanchez 1995), because the component development task of one group is not affected by decisions made in another development group, as long as all development groups conform to the input and output requirements for components specified in the modular product architecture. Thus, modular product architectures make possible modular process architectures in which development tasks can be carried out concurrently and autonomously by potentially widely distributed networks of development resources, without need for continual exercise of managerial authority. 3 Knowledge architectures Taking an information processing perspective, Malone, Yates and Benjamin (1987) propose that the difficulty of coordinating processes for creating and producing a product varies directly with the complexity of the description of the product. Adopting an organisational learning perspective, the discussion in this section makes an analogous argument that the difficulty of learning about components and their interactions increases with the complexity of the descriptions of components and their interactions within a product architecture. From this perspective, product architectures with complexly interrelated, tightly coupled components limit the ability of an organisation to distinguish clear cause-and-effect relationships within and between individual components, thereby limiting an organisation’s ability to develop componentlevel and architectural-level knowledge about component behaviours. Further, a limited ability to define clearly what it does and does not know about component behaviours and interactions also limits an organisation’s ability to identify specific new kinds of knowledge about components and their interactions which could be of strategic value to its product creation capabilities. Similarly, an organisation is limited in its ability to learn about its process capabilities when those capabilities are embedded in complexly interdependent activities in its process architecture. By virtue of their decomposition into loosely coupled components and activities, however, modular product and process architectures create loosely coupled knowledge domains which both facilitate learning about specific components or activities within those domains and make possible greater precision in identifying an organisation’s knowledge at both component and architectural levels of understanding. An improved ability to distinguish current states of component and architectural knowledge improves the ability

RON SANCHEZ 109

of an organisation to perceive and define opportunities for acquiring strategically valuable new forms of knowledge, enabling more effective targeting of organisational learning. To understand these interrelationships of product and process architectures and organisational learning more fully, it is useful to recall von Hippel’s (1994) finding that complex problem-solving processes in organisations create ‘sticky information’ that is difficult to define precisely or to extract from the ‘locus of problem-solving’ in which it is embedded. One explanation why some forms of learning create ‘sticky information’ is that problem-solving involving many complexly interdependent variables creates learning that is likely to be highly context-specific—and therefore difficult to articulate in more general terms that would enable its application to other contexts. To the extent that an organisation’s problem-solving leads to context-specific learning rather than more generalised learning that can be applied to other contexts, however, its knowledge resources are more limited in their strategic value (Sanchez 1997), and, recognising this, an organisation may tend not to deviate from the set of circumstances in which its ‘sticky’ knowledge is embedded. Thus, complex product and process architectures with tightly coupled components and activities are likely to generate context-specific ‘sticky information’ whose value as a knowledge resource in other contexts may be very difficult to discern, including the applicability of knowledge gained in the context of one project to other product development projects within the same organisation. The way in which modular product and process architectures create knowledge architectures (Sanchez 1996a) composed of loosely coupled knowledge domains is suggested by the model in Figure 6.1. At the centre of Figure 6.1 is a modular product architecture represented by the system of loosely coupled components whose ranges of permissible interactions (represented by the arrows linking components) are fully specified by a modular product architecture. Within an organisation, these specifications of permissible component interactions may be articulated (Sanchez 1997) in the form of printed ‘interface documents’ (Sanchez and Collins 1997) and/or of design protocols in CAD systems that constrain the ways in which components may be interrelated by product designers (Sanchez 1996c). These interface specifications and protocols represent ‘design rules’ that within some firms explicitly managing their knowledge resources are often supported by archival documents that give detailed statements of an organisation’s knowledge about how components of different types behave and interact in a product architecture. Creating an interface specification or design protocol that allows the substitution of component variations into a product architecture requires a high level of understanding about component interactions, i.e., substantial knowledge about component behaviours and their interactions at the architectural level. Thus, creating complete interface specifications forces a firm to document the state of a firm’s technical understanding about the components that make up its products and their interactions—in effect creating a ‘balance sheet’ of the firm’s knowledge resources that can be used in creating comparable kinds of products. Thus, just as lowering the quantities of parts in the production flow through a factory brings into clear view the bottlenecks and limitations in various activities of a production process, trying to create fully specified component

110 PRODUCT AND PROCESS ARCHITECTURES

interfaces in a modular product architecture reveals any ‘bottlenecks’ or limitations in an organisation’s knowledge of how the components in its products behave and interact. Thus, creating modular architectures becomes a vehicle not only for clearly defining what a firm knows about components and their interactions (i.e., its component and architectural level knowledge as defined in its interface documents or design protocols), but also for identifying specific forms of new architectural knowledge which would improve the ability of the firm to create more flexible and robust modular architectures. The loose coupling of component designs in a modular product architecture also improves the ability of component developers to learn about the behaviours of individual components through developing alternative component designs. By creating a stable information structure of standardised specifications governing input and output relationships between components within a modular product architecture, the complexity and uncertainty of the design context for each component may be considerably reduced (Sanchez and Tarondeau 1997), which in turn reduces the difficulty of learning about cause-and-effect relationships within each type of component. Thus, the loose coupling of component designs in a modular product architecture facilitates the discovery of less context-specific, more generalisable knowledge about individual components, which in turn leads to improved ability to identify and interrelate an organisation’s componentlevel knowledge in creating architectural-level knowledge about how components will interact in a product architecture. Support for the proposition that loose coupling of components within modular architectures enhances component-level and architectural-level learning is now appearing in a variety of forms. The modular architecture of the Boeing 777, for example, facilitated involvement of key airlines as partners in improving the component-level designs of cabin reconfiguration systems, engine servicing systems, baggage and freight loading systems, and other components whose development processes had been largely decoupled from development processes for other major components. The adoption by GE Fanuc Automation of modular open-systems architectures for its industrial controllers has enabled hundreds of component developers around the world to carry out loosely coupled, concurrent development processes creating improved components that ‘plug and play’ directly in GE Fanuc product architectures (Sanchez and Collins 1997). Moreover, studies by Christensen and Bower (1996) and Christensen (1997) show that firms developing hard disk drives optimised for the specific requirements of large customers have generally failed to create the next generation of improved disk drives, which have typically been developed by firms creating disk drives for modular ‘open architecture’ computers. One explanation for these findings that is consistent with the propositions made here is that firms developing disk drives optimised to meet the specific requirements of a large customer’s product architecture are creating relatively tightly coupled components with a more complex, tightly coupled learning environment that makes it more difficult to conceptualise, develop and test new component designs based on significantly different component technologies. Firms developing disk drives for a modular industry architecture, on the other hand, have a well-defined and bounded set of input and output specifications within which to learn about possible new component

Figure 6.1 Modular product architecture as coordinating mechanism for modular process architectures

RON SANCHEZ 111

112 PRODUCT AND PROCESS ARCHITECTURES

technologies and designs that would lead to ‘generational’ levels of improvement in disk drive performance. For comparable reasons, adoption of a modular process architecture may enhance learning about loosely coupled activities within an organisation and about the desirable modes of interaction between loosely coupled units of activity. As the model in Figure 6.1 suggests, a modular product architecture provides a coordinating mechanism for achieving the loose coupling of an organisation’s design, production, distribution and customer support activities. As suggested in Figure 6.1, the design processes that lead to the creation of a product architecture are defined in terms of the specific design capabilities of the organisation. For example, an organisation may create a ‘design library’ (Sanchez and Mahoney 1996) of well-understood component designs with defined behaviours that can be used with confidence in a product architecture; the organisation may also define the design procedures it is confident can be used to create new component designs that can then be added to the design library. Further, the capabilities of the organisation’s available production processes may be defined and incorporated into design rules that clarify the freedom designers have to create new component variations that the organisation can readily produce. Similarly, the limitations which an organisation’s distribution capabilities impose on the size, weight and other characteristics of a new product may be defined and incorporated into design rules. Current customer or user capabilities may also be defined and represented in the design rules governing the creation of new components and new product architectures. The explicit definition of an organisation’s capabilities in various process activities and of the resulting kinds of interactions which can feasibly be undertaken between those activities is directly analogous to defining component behaviours and their permissible modes of interaction in a modular product design. Once an organisation’s process capabilities and their feasible modes of interaction are articulated in this manner,4 the individual activities within an organisation become loosely coupled in the sense that decisions and changes made within one area of activity do not affect other areas of activity, so long as those decisions and changes stay within the range of input and output capabilities defined for that activity by the process architecture. Thus, within the limits of acceptable change defined by an organisation’s modular process architecture, learning about processes within the knowledge domains of its loosely coupled activities becomes considerably less complex than trying to learn about those activities when they are tightly coupled (often in unspecified ways) to other activities in the organisation. Just as a modular product architecture allows the quick leveraging of potentially many product variations by ‘mixing and matching’ new component variations within the product architecture, a modular process architecture enables the mixing and matching of existing process capabilities to reconfigure variations in organisational activity. Similarly, the explicit articulation of an organisation’s process capabilities creates a frame of reference for identifying new process capabilities that would improve an organisation’s systemic ability to respond to and take advantage of new technologies or market opportunities. Thus, creating an explicit modular process architecture helps an

RON SANCHEZ 113

organisation to perceive and more precisely define opportunities for strategically important learning. The next section discusses how creating modular product and process architectures creates a modular knowledge architecture that facilitates organisational learning. 4 Organisational learning within modular knowledge architectures The preceding discussion has tried to explain how the creation of modular product architectures enables the loose coupling of modular processes for developing, producing and supporting products, how the creation of modular process architectures leads to a greater level of clarity in identifying and defining an organisation’s capabilities and their interactions, and how this process of clarifying organisational capabilities facilitates organisational perceptions of opportunities for strategically useful learning that will improve its modular product and process architectures. This section introduces the concept of a modular knowledge architecture, which is defined here as a knowledge structure composed of loosely coupled knowledge domains. The discussion below considers the ways in which the modular knowledge architectures that result from adoption of modular product and process architectures improve processes for creating specific kinds of knowledge resources in an organisation. Management researchers are now beginning to develop new concepts for representing different kinds of organisational knowledge (Grant 1993; Sanchez 1996a, 1997; Wright 1997). The following discussion elaborates on the notion that organisational knowledge may usefully be represented by categories of know-how, know-why and know-what (Sanchez 1996a, 1997) by showing how modular knowledge architectures may facilitate learning in these three categories of knowledge. Table 6.2 provides a summary of know-how, know-why and know-what forms of organisational knowledge. A firm’s know-how is its ‘practical understanding’ about the current state of a system (Simon 1981), where the system the organisation understands refers to its product or process architectures. Know-how knowledge enables a firm to continue operating its current systems, like leveraging product variations from its existing product architecture or running its production line within its existing process architecture. Knowhow knowledge is useful in maintaining control of current product designs and production systems over some limited range of variations in internal or external conditions. Know-why is ‘theoretical understanding’ of the principles governing the functioning of a system (Simon 1981). Know-why enables an organisation to change an existing system or to create a new system. Within product architectures as systems, know-why knowledge is the theoretical understanding of why specific components in a product architecture interact as they do. Similarly, within a process architecture as a system, know-why is understanding of the principles that determine how processes may be configured to work together. Know-why knowledge enables a firm to make significant changes in existing product or process architectures or to develop new product or process architectures. To

114 PRODUCT AND PROCESS ARCHITECTURES

Table 6.2 Know-how, know-why and know-what forms of knowledge

Figure 6.2 Loosely coupled learning processes for developing know-how, know-why and know-what knowledge

create modular product or process architectures that permit substitution of many component variations requires well-developed know-why levels of knowledge about components or processes and their interactions. Know-what is ‘strategic understanding’ of the purposes to which specific forms of knowwhy and know-how knowledge might be applied in creating product and process architectures (Sanchez 1996a, 1997). Know-what enables managers to make conjectures about what new kinds of products and processes a firm might develop and what the likely competitive impacts of specific new products or process capabilities might be. In essence,

RON SANCHEZ 115

know-what knowledge is what enables managers to imagine and define feasible new kinds of products and processes for developing, producing and marketing products. Figure 6.2 suggests ways in which modular knowledge architectures facilitate knowhow, know-why and know-what forms of organisational or industry level learning processes. Within the modular knowledge architectures derived from modular product and process architectures, know-how learning may take place within loosely coupled component and process knowledge domains. Learning how to develop better components and better process capabilities within current product and process architectures is made easier and more efficient (Sanchez and Mahoney 1996) by intentionally decoupling knowhow level learning from the next higher level of know-why learning. By using knowhow level learning within current product and process architectures, organisations may lower the cost of exploring near-term market preferences by rapidly and cheaply introducing new product variations to test market reactions (Baldwin and Clark 1994; Sanchez and Sudharshan 1993) and may thereby accelerate learning about alternative component technologies through low-cost experimentation with product variations (Sanchez 1991, 1996b). By using modular architecture to reduce the complexities and constraints that arise when introducing new component technologies or product market objectives into current product and process architectures, know-why learning can be focused on identifying and creating new knowledge about technologies and markets that will be useful in defining and developing the next generation of modular product and process architectures. Know-why learning can be channelled into efforts to understand the range of new technology and market possibilities that may be mediated most advantageously through alternative possibilities for next-generation modular product and process architectures of an organisation or an industry. Similarly, know-what learning can be focused on long-term learning about technology and market trends. Long-term know-what learning occurs as a firm (or firms in an industry) tries to define future modular product and process architectures that might mediate technology and market trends most advantageously. Although many organisations try to forecast technology and market trends, relatively few use those forecasts as inputs for conjecturing future product and process architectures. However, such a conjectural exercise may be a very effective way to give content and focus to the ‘corporate imagination’ (Hamel 1989b) and to defining specific targets for ‘stretch and leverage’ of organisational capabilities (Prahalad and Hamel 1993). The basic proposition motivating Figure 6.2 and its attendant discussion is that learning at all three levels of know-how, know-why and know-what may be facilitated by the reduction of complexity at each level of knowledge, and such a reduction in complexity is made possible by the loose coupling of knowledge domains within modular product and process architectures. Recent empirical support for the learning benefits of creating loosely coupled knowledge domains within modular knowledge architectures is offered by two case studies of organisations using modular product and process architectures in knowledge-intensive businesses.

116 PRODUCT AND PROCESS ARCHITECTURES

Post (1997) has studied the use of modular software architectures by Baan Company, a rapidly growing global software firm, and concludes that modularisation of Baan’s software products not only makes possible a flexible modular organisation structure, but also improves Baan’s ability to define and create new kinds of software products (knowwhat learning), to create the new kinds of architectures for those new kinds of products (know-why learning), and rapidly to develop new application program modules that exploit current software architectures to greatest effect (know-how learning). Lang (1997) analysed the product development strategy of ARM Ltd, a small microprocessor design firm that nevertheless maintains a strong position in a global network of large semiconductor firms. ARM Ltd carries out successfully a strategy that focuses on developing and applying know-what and know-why knowledge in defining and developing modular microprocessor designs, while licensing its designs to large semiconductor firms with knowhow knowledge about semiconductor production. Thus, the loose coupling of knowledge domains within the semiconductor industry makes it possible for some firms to focus on building certain forms of knowledge (like design know-why and market knowwhat) while others build up other kinds of knowledge (like production know-how). 5 Conclusions Modular product and process architectures are increasingly prevalent in rapidly developing areas of technology and markets (Sanchez and Mahoney 1996). One explanation for this observed association may be that the adoption of modular product and process architectures within an industry creates modular knowledge architectures with substantial positive network externalities in organisational learning. In effect, use of modular knowledge architectures may create for all firms in an industry some efficiencies in building and leveraging knowledge. As firms in an industry join together to create industry standard specifications for components interactions, they thereby create a loose coupling of component designs that in turn creates a loose coupling of knowledge domains about components. A loose coupling of component knowledge domains can lead to faster, more efficient learning processes by firms focused on developing new components within an industry. Thus, adopting component-level standards within an industry may be a critically important mechanism for improving firm-level learning processes that, in the aggregate, may significantly accelerate the development of the industry’s technological regime (Nelson and Winter 1982). One observation in support of the proposition that adoption of industry standards for components and processes accelerates industry learning processes is that product markets with rapidly developing technologies often have well-established de facto or industrysponsored standards for at least key component interfaces and often even for all the component interfaces in the industry’s product architectures. Examples include the extensive hierarchy of interconnectivity standards in the telecommunications and personal computer industries and the use of Microsoft’s Windows as the de facto standard architecture for personal computer applications software.

RON SANCHEZ 117

Modular product, process and knowledge architectures are ‘industry resources’ that are often ‘taken for granted’ by practitioners in rapidly evolving product markets and are rarely addressed directly by management researchers in studies of technological learning. Perhaps the role of modular architectures in creating knowledge infrastructures that accelerate organisational learning processes has yet to attract attention from management researchers that is commensurate with the impacts of modular architectures on technological learning in both firms and industries. Notes 1 In more complex products, a product design may first be decomposed into a set of subsystems which are then decomposed into individual functional components (which are then further decomposed into subassemblies of parts and then into individual parts). Automobile companies, for example, typically decompose new product designs into a number of subsystems (like a passenger compartment climate control subsystem), each of which is further decomposed into major functional components (like a thermostatic control, an air conditioning compressor and a heater). Since the process of decomposition into subsystems typically involves a further decomposition into individual functional components, for simplicity I refer only to decomposition of product designs into components in this discussion. 2 Hierarchical decomposability, as used in this chapter, is essentially a structural concept. In this use, hierarchy has a broader meaning than that usually imputed in the organisational economics and strategic management literature (e.g., Williamson 1975; Mahoney 1992; Casson 1994), where hierarchies are characterised by an authority relation that governs the system. In the modular systems discussed in this paper, hierarchical decomposition and coordination are accomplished through the information structure (Sanchez and Mahoney 1996) of a product architecture rather than through the explicit use of an authority hierarchy as a coordinating mechanism. 3 Simon (1981) makes a similar distinction by referring to interdependent vs. ‘semiindependent’ elements of a system. 4 Like the articulation of component and architectural knowledge in a modular product architecture, articulation of an organisation’s process architecture may take the form of process interface documents or computer-based process design protocols.

References Baldwin, Carliss and Kim B.Clark (1994) ‘Modularity-in-Design: an Analysis Based on the Theory of Real Options’, Harvard Business School Working Paper, Boston, MA 02169, USA. Bogner, William C. and Howard Thomas (1996) ‘From Skills to Competences: The “Play-Out” of Resource-Bundles across Firms’, in R.Sanchez, A.Heene and H.Thomas (eds), Dynamics of Competence-Based Competition, Oxford: Elsevier Pergamon. Casson, Mark (1994) ‘Why Are Firms Hierarchical?’, Journal of the Economics of Business 1(1):47–76. Christensen, Clayton (1997) ‘Patterns in the Evolution of Product Competition’, European Management Journal 15(2):117–27.

118 PRODUCT AND PROCESS ARCHITECTURES

Christensen, Clayton and Joseph L.Bower (1996) ‘Customer Power, Strategic Investment, and the Failure of Leading Firms’, Strategic Management Journal 17:197–218. Dosi, Giovanni (1982) ‘Technological Paradigms and Technological Trajectories’, Research Policy 11: 147–62. Farrell, J. and G.Saloner (1985) ‘Standardization, Compatibility, and Innovation’, Rand Journal of Economics 16:70–83. Foss, Nicolai and Bo Eriksen (1995) ‘Competitive Advantage and Industry Capabilities’, in Cynthia Montgomery (ed.), Resource-Based and Evolutionary Theories of the Firm: Towards a Synthesis, Boston, MA: Kluwer Academic. Garud, Raghu and Arun Kumaraswamy (1993) ‘Changing Competitive Dynamics in Network Industries: An Exploration of Sun Microsystems’ Open Systems Strategy’, Strategic Management Journal 14:351–69. Grant, Robert M. (1993) ‘Organizational Capability within a Knowledge-Based View of the Firm’, School of Business Administration Working Paper Series, STRAT-2277–03–1293, Georgetown University, Washington, DC 20057. Hamel, Gary (1989a) ‘Strategic Intent’, Harvard Business Review 67:63–76. ——(1989b) ‘Corporate Imagination and Expeditionary Marketing’, Harvard Business Review 69:81– 92. Henderson, Rebecca M. and Kim B.Clark (1990) ‘Architectural Innovation: The Reconfiguration of Existing Product Technologies and the Failure of Established Firms’, Administrative Science Quarterly 35, March:9–30. Lang, John W. (1997) ‘Leveraging Knowledge across Firm Boundaries: Achieving Strategic Flexibility through Modularization and Alliances’, in R.Sanchez and A. Heene (eds), Strategic Learning and Knowledge Management, Chichester: John Wiley and Sons. Langlois, Richard N. and Paul L.Robertson (1992) ‘Networks and Innovation in a Modular System: Lessons from Microcomputer and Stereo Components Industries’, Research Policy 21:297–313. Lee, H.L., C.Billington and B.Carter (1993) ‘Hewlett-Packard Gains Control of Inventory and Service through Design for Localization’, Interfaces 23:1–11. Mahoney, Joseph T. (1992) ‘Organizational Economics within the Conversation of Strategic Management’, Advances in Strategic Management 8:103–55. Malone, Thomas W., Joanne Yates and Robert Benjamin (1987) ‘Electronic Markets and Electronic Hierarchies’, Communications of the ACM 30(6):484–97. Meyer, Marc H. and James Utterback (1993) ‘The Product Family and the Dynamics of Core Capability’, Sloan Management Review 34(3):29–47. Nelson, Richard and Sidney Winter (1982) An Evolutionary Theory of Economic Change, Cambridge, MA: The Belknap Press. Orton, J.Douglas and Karl E.Weick (1990) ‘Loosely Coupled Systems: A Reconceptualization’, Academy of Management Review 15(2):203–23. Post, Henk A. (1997) ‘Modularity in Product Design, Development, and Organization: A Case Study of Baan Company’, in R.Sanchez and A.Heene (eds), Strategic Learning and Knowledge Management, Chichester: John Wiley and Sons. Prahalad, C.K. and Gary Hamel (1993) ‘Strategy as Stretch and Leverage’, Harvard Business Review March–April:75–84. Sanchez, Ron (1991) ‘Strategic Flexibility, Real Options, and Product-Based Strategy’, Ph.D. dissertation, Massachusetts Institute of Technology, Cambridge, MA 02139. ——(1994) ‘Towards a Science of Strategic Product Design: System Design, Component Modularity, and Product Leveraging Strategies’, Proceedings of the Second International Product

RON SANCHEZ 119

Development Management Conference on New Approaches to Development and Engineering, May 30– 31, 1994, European Institute for Advanced Studies in Management, Brussels, Belgium. ——(1995) ‘Strategic Flexibility in Product Competition’, Strategic Management Journal, 16, Summer Special Issue:135–59. ——(1996a) ‘Strategic Product Creation: Managing New Interactions of Technologies, Markets, and Organizations’, European Management Journal 14(2):121–38. ——(1996b) ‘Integrating Technology Strategy and Marketing Strategy’, in H.Thomas and D.O’Neal (eds), Strategic Integration, Chichester: John Wiley and Sons. ——(1996c) ‘Quick-Connect Technologies for Product Creation: Implications for CompetenceBased Competition’, in R.Sanchez, A.Heene and H.Thomas (eds), Dynamics of CompetenceBased Competition, Oxford: Elsevier Pergamon. ——(1997) ‘Managing Articulated Knowledge in Competence-Based Competition’, in R.Sanchez and A.Heene (eds), Strategic Learning and Knowledge Management, Chichester: John Wiley and Sons. Sanchez, Ron and Robert P.Collins (1997) ‘Competing in Modular Markets’, Working Paper, Department of Industrial Economics and Strategy, Copenhagen Business School, DK-1366 Copenhagen, Denmark. Sanchez, Ron and Aimé Heene (1996a) ‘A Systems View of the Firm in Competence-Based Competition’, in R.Sanchez, A.Heene and H.Thomas (eds), Dynamics of Competence-Based Competition, Oxford: Elsevier Pergamon. ——(1996b) ‘Competence-Based Strategic Management: Concepts and Issues for Theory, Research, and Practice’, in A.Heene and R.Sanchez (eds), Competence-Based Strategic Management, Chichester: John Wiley and Sons. Sanchez, Ron, Aimé Heene and Howard Thomas (1996) ‘Towards the Theory and Practice of Competence-Based Competition’, in R.Sanchez, A.Heene and H. Thomas (eds), Dynamics of Competence-Based Competition, Oxford: Elsevier Pergamon. Sanchez, Ron and Joseph T.Mahoney (1994) ‘The Modularity Principle in Product and Organization Design’, Office of Research Working Paper no. 94–0157, University of Illinois, Champaign, IL 61820. ——(1996) ‘Modularity, Flexibility, and Knowledge Management in Product and Organization Design’, Strategic Management Journal 17, Winter Special Issue: 63–76. Sanchez, Ron and D.Sudharshan (1993) ‘Real-Time Market Research: Learning-by-Doing in the Development of New Products’, Marketing Intelligence and Planning 11, August:29–38. Sanchez, Ron and Jean-Claude Tarondeau (1997) ‘Modular Product Design and the Management of Uncertainty in Innovation and Product Development’, Working Paper, Graduate School of Management, University of Western Australia, Nedlands, WA 6907. Sanderson, Susan Walsh and Mustafa Uzumeri (1997) Managing Product Families, Chicago: Irwin. Saxonian, A. (1991) ‘The Origins and Dynamics of Production Networks in Silicon Valley’, Research Policy 20:423–37. Simon, Herbert (1981) The Sciences of the Artificial, Cambridge, MA: MIT Press. Teece, David (1992) ‘Foreign Investment and Technological Development in Silicon Valley’, California Management Review 34:88–106. von Hippel, Eric (1994) ‘Sticky Information and the Locus of Problem Solving’, Management Science 40:429–39. Wheelwright, Steven C. and Earl Sasser, Jr. (1989) ‘The New Product Development Map’, Harvard Business Review May–June:112–22. Williamson, Oliver E. (1975) Markets and Hierarchies: Analysis and Anti-trust Implications, New York: The Free Press.

120 PRODUCT AND PROCESS ARCHITECTURES

Wright, Russell W. (1997) ‘Tangible Integration versus Intellectual Codification Skills: A Comparison of Learning Processes in Developing Logic and Memory Semiconductors’, in R.Sanchez and A.Heene (eds), Strategic Learning and Knowledge Management, Chichester: John Wiley and Sons.

7 Building innovative assets and dynamic coherence in multi-technology companies Jens Frøslev Christensen

Introduction While a large empirical research effort has been devoted to studying the dynamics of the individual innovation process, on the one hand, and the dynamics of the firm’s product or business diversification, on the other, the dynamics of the technology base has attracted much less research interest. However, empirical studies (Pavitt et al. 1989; Granstrand et al. 1990; Oskarsson 1993; Granstrand et al. 1997) show that the trend for technology diversification has been even more pronounced—at least within technology-intensive firms —than the trend for product diversification. In other words, the expansion of the firm’s base of skills or capabilities in terms of technical fields is greater than the expansion of the firm’s portfolio of products. This indicates a strong underlying tendency for increasing diversity of the technology base that, again, seems to have provided a strong impetus stimulating the rising managerial and academic attention directed at technology management and strategy in recent years. Thus, the evolution of the multi-technology company and its efforts to manage the increasingly diverse technology base has made it even more necessary than before to make a distinction between lower-order innovative/ technical capabilities and higher-order competences in managing technology. So far one critical aspect of higher-order management of technology has received little systematic attention in the literature, namely the competence in pursuing dynamic coherence of the technology base. This chapter discusses the dynamics of assets (capabilities and competences) for technological innovation and the conditions for establishing a corporate technology base that balances the difficult trade-off between exploitation of existing and exploration of new innovative assets and commercial territories. This issue is related to recent contributions on competence-based competition and strategy, as primarily conceived within the resource-based theory of the firm (Hamel and Heene 1994; Iansiti and Clark 1994; Teece and Pisano 1994; Sanchez et al. 1995; Foss and Knudsen 1996; Sanchez et al. 1996; Heene and Sanchez 1997; Teece et al. 1997). The chapter is divided into two parts. The first part presents a framework for analysing innovative assets in both a static and dynamic/processual sense. This framework is in turn used to identify some critical aspects of managing the technology base of diversified multi-

122 JENS FRØSLEV CHRISTENSEN

technology companies. A conceptual distinction is suggested between (1) innovative/ technological versus complementary assets, (2) generic versus firm-specific assets, and (3) domain-specific versus integrative assets. It is argued that this simple categorisation is useful in determining the strategic value of innovative as well as other firm assets. While these categories represent a static or stock perspective on innovative assets, the dynamic or flow dimension of innovative assets is then discussed. A taxonomy of ‘innovative asset dynamics’ is proposed that is inspired by work done by Doz (1994), Sanchez et al. (1995) and Markides and Williamson (1994). The concepts and arguments are illustrated by case material on innovation in different companies. The second part of the chapter analyses the technology base of the firm defined as the firm’s portfolio of innovative assets. The purpose is to apply and adapt the notion of corporate coherence—as recently introduced by Teece et al. (1994) and further elaborated from a competence-based perspective by Christensen and Foss (1997) and Foss and Christensen (1999). While these earlier papers address the overall theme of corporate coherence, this chapter takes a more narrow focus on the coherence of the technology base and especially on (1) interdependencies between assets in the technology base, and (2) the trade-off balance between incremental versus explorative forms of changes in the technology base. The chapter ends up with two case illustrations: the case of 3M illustrates a highly dynamic and coherent technology base, while the case of the former ICI illustrates a fragmented technology base. Resources, capabilities and competences for technological innovation From a competence-based perspective the firm may be defined in terms of its constituent resources, capabilities and competences. In order to provide an adequate specification of the distinction between resources on the one hand and competences/capabilities on the other, I suggest, in line with Grant (1991), that capabilities/competences constitute the ‘experience base’ of the firm—and therefore have to be (largely) accumulated internally— whereas resources (physical, intangible or financial) are tradable at ‘strategic factor markets’ (Barney 1986; Christensen 1996a). I use the terms ‘capability’ and ‘competence’ as follows: A capability is a lower-order functional or inter-functional, operational or technical capacity that may be further subdivided into specific (individual) skills or specialised (team-based) capabilities.1 A competence is a higher-order capacity of the firm management to organise, mobilise, coordinate and integrate resources and capabilities to create value and competitive advantage. The term ‘asset’ is used to comprise both (tradable) resources, (technical) capabilities and (managerial) competences. Different categories of assets may be identified. Three broad pairs of asset categories or dimensions are discussed that seem particularly relevant in specifying the strategic value of a firm’s technology base:

BUILDING INNOVATIVE ASSETS AND DYNAMIC COHERENCE 123

1 innovative assets (Christensen 1995, 1996b) versus complementary assets (Teece 1986); 2 generic versus firm-specific assets; 3 domain-specific assets versus integrative assets. These categories of assets (which may be further specified in terms of resources, capabilities and competences) are here discussed in ‘stock terms’, while their dynamic or ‘flow’ features will be explored in the following section. Re 1. Innovative assets versus complementary assets. Innovative assets are required to produce technological innovation, that is, new products (goods and services) and processes. They consist of competences to organise and direct the inventive and innovative activities, as well as resources (i.e., technical skills and equipment, and financial resources) and technical capabilities implement the process in practice. The total portfolio of innovative assets constitutes the technology base of the firm which will be analysed in terms of its coherence in the next section.2 As Teece (1986) has argued, successful commercial exploitation of technological innovation normally requires access to assets that are complementary to innovative assets, such as for example manufacturing, distribution, marketing or after-sales service. While innovation necessarily involves in-house assets of some kind, the required complementary assets do not necessarily have to be sourced in-house. They may comprise resources that can be acquired from the marketplace. Complementary assets may, of course, also be subject to innovative efforts and thus become innovative assets, but in this chapter the notion of innovative assets only includes capabilities and competences for product and process innovation, and not for innovation in, for example, marketing or after-sales service, even if such innovation also may require technological development.3 Innovative/technical capabilities can be classified in (at least) two different ways: according to their constituent technical fields and scientific disciplines, and according to their functional role in the innovation process. The former way represents the conventional ‘technology mapping’ approach that does not, however, in itself provide any insights into the innovative potential for the firm of the technological or scientific assets identified. Elsewhere (Christensen 1995) I have proposed a more differentiated ‘innovative asset’ framework that distinguishes four generic categories of innovative capabilities (or sub-functions) that transcends the conventional notion of R&D: • scientific research capabilities comprising capabilities in (a) basic research and (b) applied/industrial research; • process innovative capabilities for (a) ‘hardware’ innovation, (b) the systemic and organisational development of manufacturing operations, logistics and quality management, and (c) plant layout; • product innovative application or design capabilities; • aesthetic design capabilities.

124 JENS FRØSLEV CHRISTENSEN

Thus, innovative capabilities may be characterised by specifying both the technologies or scientific disciplines and the innovative sub-functions involved in innovative activities, and the interconnectedness between different technical fields and innovative sub-functions.4 While innovative/technical capabilities involve the practical capacity to implement technological innovation, innovative competences involve the capacity to manage and organise technological innovation, and to develop new innovative assets. A critical aspect of innovative competence is the ability to establish the proper coupling and integration of teams or individual experts reflecting different innovative/technical capabilities. More specifically, innovative competence refers to procedures for generating, evaluating and selecting innovative ideas, the ability to identify and recognise relevant technological opportunities and link these to business opportunities, and the ability to manage innovative projects and build absorptive capacity and external relations. Re 2. Generic versus firm-specific assets. Generic assets are based on well-established paradigms and boundaries of knowledge structures, routines and heuristics, i.e., generic scientific disciplines and technological fields, or the conventional wisdom associated with business functions such as marketing, or individual value-chain activities such as inbound logistics or procurement. Generic assets may largely consist of resources that can be sourced from the marketplace, and capabilities and competences to innovate based on established paradigms. On the contrary, firm-specific assets are idiosyncratic to the individual firm and hard to imitate for competitors. Consequently firm-specific assets are either capabilities or competences which have been largely accumulated internally, or intellectual property assets (such as patents) that are subject to effective protection. While generic assets tend to comprise knowledge or skills which are either broadly distributed among firms or individuals (i.e., with specific educational background) or embodied in artefacts that are accessible at the market, firm-specific capabilities or competencies will tend to be relatively complex and contain significant tacit elements. Re 3. Domain-specific assets versus integrative assets. Domain-specific assets5 are assets with a specialised applicability as linked, for instance, to one narrowly defined technical discipline or one business function. Integrative assets6 (which in my terminology always will comprise competences) are based on cross-disciplinary, cross-functional or otherwise boundary-spanning activities that transcend established boundaries of domain-specific assets. The managerial competences involve the coordination and interaction between teams in different parts of the organisation (as well as outside the organisation), each possessing more domain-specific (specialised) capabilities and skills. Thus integrative assets tend to be more complex than domain-specific assets—with respect both to the managerial competence side (since it involves the management of more teams and their interaction, and not only one) and to the technical capability side (since it involves both specialised team activities and coordination between teams possessing skills of different natures). Although not necessarily so, integrative assets may tend to be more firm-specific (and to a larger extent based on tacit, complex knowledge) than may domain-specific assets. By implication, integrative assets tend to be more difficult to imitate and therefore provide a more powerful base of competitive advantage than domain-specific assets may do. Iansiti

BUILDING INNOVATIVE ASSETS AND DYNAMIC COHERENCE 125

and West (1997) use the notion of ‘technology integration’ to signify a firm’s competence in choosing the right technologies among a large number of technological options and make them work well together. They argue that this ‘technology integration’ competence —at least in the computer industry—is now more important than the capability to create the technologies themselves: The U.S. companies that prevailed in the computer industry in the 1990s abandoned the traditional R&D model [the linear model, JFC] and created a radically new one. They did not stop conducting basic research, but they did shift much of the focus of their research efforts to applied science, and they turned to an increasingly diverse base of suppliers and partners—universities, consortia, and other companies—to help generate technological possibilities. In addition, they formed tightly knit teams of expert integrators—people with extensive backgrounds in research, development, and manufacturing—to develop new generations of major products and processes. (Iansiti and West 1997:12) Shortening product life cycles and expanding numbers of technology options underlie this type of R&D approach in which the possible firm-specific element rests on the technologyintegrating and not the technology-creating capacity. If integration is fairly simple and the applied technologies are ‘very’ generic (easy to access), possible lead times from innovation will be short. However, if life cycles are short anyway, the important thing is to assure speedy R&D processes that can lead you either first to the market, or at least to remain part of the innovative race. Innovative assets may be more or less domain-specific versus integrative, and integrative assets may be more or less based on generic versus firm-specific assets. Most industrial innovation requires the coupling of two or more innovative capabilities in terms of innovative sub-functions (i.e., product and process innovative capabilities) as well as technical domains (i.e., mechanical and electronic engineering capabilities). However, since new domain-specific (and possibly generic) disciplines often emerge from the combination and integration of two or more disciplines that were previously distinct (i.e., mechatronics or optoelectronics), I will only term innovative assets integrative to the extent that the firm is able innovatively to cross established boundaries. Table 7.1 illustrates how different combinations of the proposed categories of innovative assets may affect the competitive positions of an innovative firm. The capacity to accumulate innovative/technical capabilities and exploit the interfaces between them may constitute the basis of sustainable competitive advantage and, eventually, produce lock-in dilemmas or rigidities (Leonard-Barton 1992). These interfaces may involve different teams and departments within the firm or between divisions or firms. In terms of transaction cost theory we may speak of inter-asset specificity or co-specialisation between different innovative assets (Christensen 1995, 1996b), and between innovative assets and complementary assets (Teece 1986; Christensen 1996b). The capacity successfully to exploit and explore co-specialisation

126 JENS FRØSLEV CHRISTENSEN

Table 7.1 Combinations of categories of innovative assets and their contributions to competitive advantage

requires not only relevant innovative and technical capabilities, but also managerial or organisational competences that are directed towards, for example, the creation of incentives for cooperation between different teams, the recruitment of the utmost talent to be linked into innovative activities, and the strengthening of the organisation’s absorptive capacity (Cohen and Levinthal 1990). Innovative assets with integrative and firm-specific features will generally provide the basis for a more significant competitive advantage and longer lead time than innovative assets of a domain-specific and/or generic nature, because integrative, firm-specific assets will tend to be particularly difficult to imitate or replicate by rivals compared with both integrative/generic assets and domain-and firm-specific assets.7 In high-technology industries complementary assets are likely only to provide a secondary source of competitive advantage and prolong the lead time of the innovation, but only to the extent that the complementary assets contain either firm-specific or integrative elements. In other words, complementary assets of a purely generic nature are not likely to contribute to competitive advantage. Integrative features with respect to complementary assets may relate to the couplings between different types of complementary assets (for instance specific linkages between marketing and after-sales service) as well as to the couplings between innovative and complementary capacities (for example the development of a distribution system that specifically fits a newly developed product) (see Christensen 1996b). An example of a company that has demonstrated a sustainable competitive strength as a result of building and maintaining integrative assets is Grundfos, a Danish manufacturer of pumps and pump systems, and for years the global leader within the market for circulation pumps. Two categories of innovative capabilities have played a critical role in establishing this competitive advantage: (1) advanced capability in product application design, including—during recent years—the capacity to handle the integration of mechanical and electronic engineering in product construction; (2) the similarly advanced capability in process development, including the in-house capacity to design machinery and production systems exclusively dedicated to the particular product categories in question. However, without the managerial competence in continuously harmonising and co-specialising these capabilities, the strong market position is not likely to have resulted. It is moreover

BUILDING INNOVATIVE ASSETS AND DYNAMIC COHERENCE 127

evident that new generations of pumps take advantage of the very competitive dedicated complementary assets, especially the highly efficient operational manufacturing capacity of Grundfos (Christensen et al. 1998). The failure of EMI to keep a competitive grip on its CT-scanner innovation may be interpreted as a case of an invention that was based on integration of rather generic technologies and for which commercialisation was sought by means of hastily built complementary assets of primarily generic nature. This interpretation finds support in the Harvard Business School case on EMI and the CT-scanner (Bartlett 1983): ‘Although computerised tomography represented a conceptual breakthrough, the technologies it harnessed were quite well known and understood. Essentially, it linked X-ray, data processing, and cathode ray tube display technologies in a complex and precise manner’ (p. 3). EMI was then doing business in music (Capitol Records and Angel), theatres, hotels and defence-electronics (Business Week 1980). The CT-scanner was seen as an opportunity to diversify into the fast-growing medical electronics field. By the end of 1974, eighteen months after the first EMI scanner was installed, the first two competitive scanners were announced. One of them was the Acta-Scanner, which, unlike the EMI scanner, was designed to scan the body rather than the head. ‘Although specifications on scan time and image composition were identical to those of the EMI scanner, the $298, 000 price tag gave the Acta-Scanner a big advantage’ (Bartlett 1983: 1). The other one, the DeltaScan offered by Ohio Nuclear, part of Technicare, represented an even more formidable challenge for EMI, since it presented clearly superior features on a unit price much lower than the EMI scanner. The first competitive scanners resembled the EMI scanner very closely, demonstrating that the wall of patents that EMI had established did not provide any long-term protection. Even if court decisions did uphold some of EMI’s patent claims, Teece (1986:198) states that ‘once the product was in the market it could be reverse engineered and its essential features copied’. By the end of 1975 six new competitors had entered the business, one of them General Electric (GE) who took the infant technology into its third generation. According to Teece (1986), Technicare and GE already possessed both the technological capabilities to develop CT-scanners, and, contrary to EMI, the complementary assets required to commercialise the scanner. Moreover they were experienced and well-reputed producers and marketers of medical equipment. GE had very powerful marketing and after-sales service assets, including a 300-person sales force and a service network of 1,200 in the US (Bartlett 1983:1). By 1978 Technicare had taken over market leadership, which was in turn soon overtaken by GE (Teece 1986:198). For two years EMI’s scanner business had recorded losses of $56 million, and in 1979 the EMI scanner business was acquired by Thorn Electric Industries, who subsequently sold it for $37.5 million to GE (Business Week 1980: 10). In the context of the terminology discussed above, the case can be interpreted to show that EMI’s failure to reap significant returns from its CT scanner innovation was primarily due to lack of adequately firm-specific innovative assets and lack of integrative, complementary assets, exactly the types of assets that made GE a winner despite being a latecomer.

128 JENS FRØSLEV CHRISTENSEN

The dynamics of innovative assets In recent extensions of the resource-based theory of the firm there has been explicit focus on the dynamic dimension of capabilities and competences (Doz 1994; Hedlund 1994; Henderson 1994; Markides and Williamson 1994; Teece and Pisano 1994; Sanchez and Thomas 1995; Sanchez et al. 1995; Teece et al. 1997). Inspired by Sanchez et al. (1995) and Doz (1994), I suggest a distinction between three categories of asset dynamics:8 1 Maintaining or leveraging assets. Maintaining assets (in order to avoid the erosion of existing resources, capabilities and competences) is mostly assured by using them for objectives which do not challenge their validity. Leveraging assets means applying existing assets to current or new market opportunities that require quantitative changes in the stocks of resources underlying the assets (for example machinery or staff) or transference of the assets for use in other parts of the organisation or in other organisations. 2 Improving assets along existing trajectories, based primarily on existing (or gradually renewed) teams. This happens through learning and searching processes which stretch but do not fundamentally challenge the core knowledge on which the assets are based. 3 Building new assets. This necessarily involves new (or partly new) directions of learning and searching and usually also regroupings of existing teams or the building of new teams with skills that are new to the firm. Markides and Williamson (1994) relate the issue of asset or competence dynamics9 to the strategy of related diversification. They distinguish four types of potential advantages from related diversification incurred by different types of asset and competence development across different strategic business units: ‘asset amortisation’ (the potential to reap economies of scope through asset sharing); ‘asset improvement’ (the potential to use a competence to help improve an existing strategic asset, that is, a market-specific asset that underpins a firm’s competitive advantage; ‘asset creation’ (the potential to use a competence to create a new strategic asset); and ‘asset fission’ (the potential to build a new competence in the course of building a new strategic asset, a competence which will in turn help to improve existing assets). Below, these categories are specified or reinterpreted to account for innovative assets in terms of capabilities and competences and of ‘assets’ in terms of internally developed products (goods and services) and processes (equipment, production systems, plant layout, etc.). The taxonomy of asset dynamics suggested above is used and specified both at the higher-order competence level and at the lower-order capability level, since there is not necessarily full congruence across the two levels. • Innovative asset amortisation reflects the firm’s ability to reap commercial benefits from internally developed products and processes through:

BUILDING INNOVATIVE ASSETS AND DYNAMIC COHERENCE 129

(a) market launching of products—that is, mobilising the appropriate complementary assets, or the exploitation of newly developed processes—that is, initial implementation and subsequent exploitation of the learning curve; (b) relaunching the same product (or basic design) in other markets than the initial one, or in the same market but at a later period,10 or using existing processes for other productive uses than those originally targeted. Innovative asset amortisation does not involve any form of innovative asset development (improvement or building), with respect either to managerial competences or to the underlying capabilities. It may, however, require some quantitative leverage, and elements of asset transfer, for example from one department or business unit to another. Moreover, it may involve the development (improvement or building) of complementary assets—if these are not generic, and thus cannot be fully accessed from the marketplace. • Innovative asset improvement reflects the incremental innovation process that results in incremental product or process innovations, for example new models or line extensions within a given product family, or refinements or limited renewal of existing process equipment. Incremental innovation may be incurred by existing competences and capabilities, but it is likely that only the most marginally incremental innovations do not require some elements of capability improvements. Existing (or marginally improved) complementary assets are usually sufficient to commercialise and implement incremental innovations. • Innovative asset creation or recombination reflects the capacity of existing or improved assets to initiate and generate innovation of a more radical kind. This may involve the ‘creative invention’ and the successive recombination of existing capabilities and resources. It is likely, however, that most significant innovations based on innovative asset creation require some elements of capability improving and building within the context of existing or improved managerial competences. Even if the existing complementary assets (or some of them) may be helpful in the process of launching and implementation of such radical innovations, it is likely that some of these assets have to be improved, and others have to be built from scratch. • Innovative asset building reflects the ability to build new innovative competences and capabilities prior to, or in the course of, constructing a distinct new product or process. We may distinguish three types of innovative asset building depending on the effects on the firm’s existing tech nology base: (a) ‘asset cannibalising’ effect, (b) ‘asset adding’ (or neutral) effect, or (c) ‘asset coupling’ (or synergistic) effect. This type of innovative asset building will below be termed ‘innovative asset fission’ corresponding to Markides and Williamson’s term ‘asset fission’.

130 JENS FRØSLEV CHRISTENSEN

Re (a). Innovative asset building with ‘cannibalising effect’ is the least attractive type. Here new products or processes based on new assets make older products, processes and/or capabilities obsolete. The risk of cannibalisation sometimes provides the firm with an incentive not to innovate along certain trajectories unless they are forced to by competitors.11 Re (b). Innovative asset building with ‘neutral effect’ reflects the generation of new assets that are added as ‘stand-alone’ assets with neither destructive nor synergistic effects on other assets within the technology base. Re (c). ‘Innovative asset fission’ implies that the newly built asset not only leads to new products or processes, but also helps to revitalise the firm’s existing products and processes (or some of them) by entering synergistic linkages with other capabilities in the firm. This type of innovative asset building is the most attractive in terms of strategic value for the firm. Innovative asset fission involves not only asset building but also asset integration, that is, aligning the new assets with the old assets. Innovative asset fission usually also incurs a substantial reshuffling of at least part of the complementary asset structure. By using the case of Sony’s Walkman (the original innovation and successive new models) I shall first illustrate the distinction between ‘innovative asset improvement’, ‘innovative asset creation’ and ‘innovative asset building’. Thereafter I shall illustrate ‘innovative asset fission’ by the role of adhesive technologies for the dynamics of innovation in a Danish medical device company, Coloplast. Sony’s initial Walkman, presented to the market in 1979, was certainly not a technological breakthrough (Nayak and Ketteringham 1986; Morita 1991; Sanderson and Uzumeri 1995). Rather it represented a radical ‘market innovation’ that required neither any significant improvement of existing capabilities nor the building of new ones. In the terminology presented above, the original Walkman was the result of ‘innovative asset creation and recombination’ involving: 1 the creative idea of listening to music without disturbing others, or linking an existing portable stereo tape recorder (the Pressman that was developed in the Tape Recorder Division) that could only play not record, with lightweight portable headphones that were under development in the Headphone Division (Nayak and Ketteringham 1986; Morita 1991); 2 the combination and integration of components that were already fully or nearly developed in Sony: the stereo cassette recorder based on compact, high performance (the Pressman), stereo circuitry and the small lightweight headphones (Sanderson and Uzumeri 1995); 3 industrial/aesthetic design or styling of the Walkman. During the 1980s Sony introduced about 250 models of the Walkman in the US. Most of these represented incremental innovations in the form of (1) improvements of existing features in the Walkman or implementation of new features (e.g., Dolby noise reduction) that were widely used in other consumer electronics products and could be implemented

BUILDING INNOVATIVE ASSETS AND DYNAMIC COHERENCE 131

with minor redesign of the electronic circuits (Sanderson and Uzumeri 1995:770), (2) minor rearrangements of existing features and cosmetic redesigns of the external case, involving primarily marketing-orientated design-features. The latter accounted for 85 per cent of the 250 models. These incremental model changes reflect processes of ‘innovative asset improvement’ that contribute only to maintain, not further to improve or develop, innovative asset. The 1980s were not, however, exclusively marked by incrementalism. Sanderson and Uzumeri identify five significant technological innovations with respect to the Walkman product family, implying what they term ‘generational design changes’ involving either the development of new electromechanical product platforms, or the creation of new key components. In both cases these innovations required substantial technical novelty, significant technical design efforts, and in some cases also major changes in the manufacturing processes (1995:769). These generational design changes reflect processes of ‘innovative asset building’. To judge whether innovative asset building can be further categorised as ‘innovative asset fission’, one needs to have some broader insights in the development of a company’s technology base and product portfolio. The building of adhesive assets (both specific types of adhesives and the underlying technical capabilities) in the Danish manufacturing company Coloplast can provide such a case. As a medical device company, Coloplast has undergone a remarkable growth dynamics during the last decade (number of employed grew from around 800 in the mid-1980s to around 2,500 in the mid-1990s; net turnover grew from D.kr 400 million to 1,700 million, equivalent to about $250 million). From its establishment in 1957 and until the end of the 1970s Coloplast exclusively focused on ostomy appliances. Coloplast actually pioneered the disposable self-supporting ostomy bag, creating a market which would subsequently attract major competitors. An adhesive based on zinc-oxide attached the bag to the abdomen. In 1979 Coloplast established its second business field, incontinence appliances (urine bags, etc.) for people with bladder control problems. This related diversification made it possible to apply the adhesive on products other than ostomy bags and, moreover, make better use of the existing sales force, that was already in contact with customers within the incontinence field through the sales of ostomy bags. During the 1980s two new fields were added to the product portfolio: wound management products (particularly hydrocolloid plasters for moist wound healing), and breast prosthesis products. Also these diversifications were motivated by the wish to use the sales force more effectively as well as to find further applications for Coloplast’s adhesive technologies. Coloplast has continuously been committed to build and improve its capabilities within adhesives. The building of the hydrocolloid adhesive capability, which contrary to the zinc-oxide adhesive is a biotechnology-based adhesive, was to start with primarily intended to provide more skinfriendly features for existing and new ostomy appliances, but subsequently this new adhesive technology was applied in the new business areas that Coloplast diversified into: incontinence appliances, wound healing products and breast prostheses. The fact that Coloplast possessed the hydrocolloid adhesive technology was one important rationale and springboard for successful diversification into these fields.

132 JENS FRØSLEV CHRISTENSEN

Today, the most important patents within hydrocolloid technology have expired and most competitors have access to this technology. Thus, the technology has become generic and not any longer a source of competitive advantage. In recent years Coloplast has been investing large R&D efforts in building a new adhesives capability which has been applied in newly developed ostomy appliances (termed Assura) that also embody improvements other than the adhesive. The key feature of this product line, however, is the unique patented skin protective seal, where two adhesives have been combined for greater user benefits. Coloplast’s technology base also comprises other technologies such as coating and die casting technologies, but capabilities within adhesives come closest to a position as a core competence, the building and improving of which has paved the way for increasingly sophisticated and competitive product improvements as well as for diversification into new product application areas. The building of adhesive technologies has played the role of ‘innovative asset fission’. How does this typology of innovative asset dynamics correspond to the distinction between generic and firm-specific assets and between domain-specific and integrative assets? As suggested above, innovative and complementary assets marked by integrative and firm-specific features tend to provide a stronger safeguard against rivals than assets of a generic and domain-specific nature. This may imply that innovative asset improvement based on firm-specific and integrative assets (strong inter-asset specific couplings) may provide a stronger appropriability regime than innovative asset creation based primarily on generic (or domain-specific) assets. Innovative asset fission necessarily implies strong integrative features, in two respects. First, building new assets means—at least to some extent—transcending existing capabilities and paradigms. Second, the ‘fission’ metaphor implies the challenge of integrating the new and the old assets, and this process certainly implies building integrative assets. If successful, innovative asset fission may be a powerful vehicle for building sustainable competitive advantage. If failing, it may turn into an economic disaster, since it gives rise to substantial sunk cost and locks the firm into a deadend trajectory (Kim and Kogut 1996), the course of which can be highly difficult and costly to change. The technology base of the firm should ideally assure an appropriate trade-off balance between the three latter categories of innovative asset dynamics: ‘innovative asset improvement’, ‘innovative asset creation’ and ‘innovative asset building’ (with primary focus on ‘innovative asset fission’)—that is, the types of dynamics that require the active use and development of innovative assets. The efficient exploitation or ‘amortisation’ of innovative assets is primarily the responsibility of operational agents, or functions within the firm other than the R&D function (although intensive interaction between these agents is normally required in the initial phases). It concerns the mobilisation of the complementary assets required for appropriating the maximum rent from a given technological innovation (Teece 1986). Each of the four dynamics may provide the firm with a competitive advantage. However, the time horizon tends to differ across the categories. Generally ‘innovative asset amortisation’ provides the firm with a short-term advantage (although this may be

BUILDING INNOVATIVE ASSETS AND DYNAMIC COHERENCE 133

Table 7.2 The relationship between categories of innovative asset dynamics, and the dominant features of innovative assets

transformed into a longer-term advantage in the case of relaunching of products or finding and exploiting new uses for existing processes). The time horizon tends to be progressively increasing when moving down through the categories ‘innovative asset improvement’, ‘innovative asset creation’, to ‘innovative asset fission’. Table 7.2 illustrates the relationship between the categories of innovative asset dynamics and the dominant features of the innovative assets involved (generic versus firmspecific, domain-specific versus integrative) on the one hand, and the dominant type of asset dynamics (maintaining, improving, etc.) on the other. Taken together, ‘innovative asset amortisation’ and ‘innovative asset improvement’ constitute the reproductive dynamics of innovative assets in which given resources and capacities are basically reproduced or only developed incrementally and within limited margins and a relatively short time horizon. ‘Innovative asset creation’ and ‘innovative asset fission’ constitute the explorative dynamics of innovative assets that provide the longterm basis for competitive advantage by creating new managerial competences (with respect to the management of innovation, or the designing and implementing of a technology strategy) and capabilities (i.e., in new technical fields or innovative subfunctions) that promote the capacity for producing more radical product or process innovation. The reproductive dynamics tend to be based on experiential learning that is strongly pathdependent, such as ‘learning by doing’ (Arrow 1962) and ‘learning by using’ (Rosenberg 1982). Explorative dynamics tend to be based on experimental learning that reflects a more intensive and risk-taking search process. This distinction between experiential and experimental learning seems identical with the distinction between exploitation and exploration (March 1991; Levinthal and March 1993) and moreover seems to have close

134 JENS FRØSLEV CHRISTENSEN

similarities with the distinction between single-loop and double-loop learning (Agyris and Schön 1978). The dynamic coherence of the technology base As previously noted the technology base of the firm is defined as the total portfolio of innovative assets within the firm. While the analysis at the level of distinct competences and capabilities (and couplings between them) is relevant when focus is directed at particular innovation processes, analysis at the level of the technology base as a whole is relevant not only to the management of innovation and technology of the firm, but also to the general (business-or corporate-level) strategic management. The rest of this chapter will operate at the level of the technology base. The purpose is first to apply and adapt the notion of corporate coherence (as introduced by Teece et al. 1994, and further analysed by Christensen and Foss 1997, and Foss and Christensen 1999) to the technology base of the corporation, by drawing on the conception of innovative assets as specified above, and second, to analyse the dynamic dimension of the technology base by using the terminology of asset dynamics presented here. Finally, the implications of the notion of ‘dynamic coherence’ (Christensen and Foss 1997) for the management of the technology base is indicated and related to two cases. One is 3M, showing highly dynamic coherence; the other is the former ICI, showing a fragmented technology base. The coherence of the technology base Recently, the concept of coherence has been applied to the firm or corporate level to signify the relatedness among the businesses in multi-product corporations (Teece et al. 1994): ‘A firm exhibits coherence when its lines of business are related, in the sense that there are certain technological and market characteristics common to each’ (p. 4). From a competence-based perspective, coherence should be analysed at the level of capabilities and competences, and not at the level of products or businesses (Christensen and Foss 1997). From this perspective, coherence (or the degree of coherence) signifies a firm’s ability to exploit and explore assets in an economically efficient and effective way. The coherence of the technology base relates to the interdependencies and synergies between innovative assets (and associated resources, capabilities and competences) of the firm. The coherence of the technology base may apply to two sub-levels of analysis: coherence between the elements of the technology base (‘local coherence’), and coherence between the technology base and the broader firm or corporate context (‘contextual coherence’).12 The local coherence perspective refers to the level of integration and interconnectedness within the technology base between the different innovative competences and capabilities as defined in terms of technical fields and innovative sub-functions. A high degree of local coherence reflects a technology base comprising a varied portfolio of innovative assets, and characterised by the exploitation and exploration of a high level of inter-asset linkages and synergies in innovative efforts. The local coherence is not a function of the diversity

BUILDING INNOVATIVE ASSETS AND DYNAMIC COHERENCE 135

of the technology base (i.e., the more diverse the less coherent). To exhibit coherence in a highly diverse technology base in a diversified corporation is much more demanding than to exhibit coherence in a specialised firm with a narrow technology base.13 Thus, to pursue coherence of the technology base as a strategic objective is especially demanding in a diverse multi-technology context and reflects the capacity to generate economies of scope (asset sharing and transfer) and synergies or economies of co-specialised assets within the technology base. This involves couplings between technical capabilities in different fields (for example software engineering, digital signalling and laser technology) or between different innovative capabilities (for example scientific research, product innovative application and aesthetic design) (cf. Christensen 1995, 1996b). In other words, the technology base exhibits local coherence to the extent that the constituent innovative assets show integrative features. The 3M innovation of the Post-it adhesive—as outlined by Nayak and Ketteringham (1986)—provides a good example of an innovation based on couplings and synergies between different fields of expertise in a diverse corporate technology base. In the course of a ‘Polymers for Adhesives’ programme in 3M’s Central Research Laboratories in the mid/ late 1960s, Spence Silver became aware of a new family of monomers, which he thought contained potentials as ingredients for polymer-based adhesives. He became engaged in an open-ended research effort that led him to identify an adhesive with the ‘curious charm of hanging around without making a commitment’ (Nayak and Ketteringham 1986:77). Trying to think up an application for the adhesive, Silver visited every division at 3M and demonstrated the properties of the adhesive. Not until five years later, in 1974, came the right idea. Some time after having seen Silver show his adhesive, Arthur Fry, a chemist who had just joined a new venture team in Commercial Tape, got the idea of using it for a ‘better bookmark’. Fry mixed up some adhesive and paper, and tried different concentrations. However, the adhesive, when moved on to paper, peeled off on to everything it touched. If it was put down on something and pulled apart, the adhesive could stay with either side. That property had to be changed. Two other members of the new venture team in Commercial Tape were put on the case and invented a way to stick the Post-it down. The next challenge was to find a way to produce the Post-it. The engineers in 3M’s Commercial Tape Division were accustomed to tape which is produced and packaged in rolls. Post-it, which has glue applied selectively to one side of the paper, had to be produced in big sheets, not in rolls, and these sheets had to be laid together and cut into smaller sizes. This required the invention of entirely unique machines. ‘Within two years, Fry and 3M’s mechanical engineers had tinkered their way to a series of machines that, among other things, coated the yellow paper with its substrate, applied adhesive, and cut the sticky paper into little square and rectangular note pads. All of the machines are unique and proprietary to the company—and they are the key to the Post-it notes’ consistency and dependability’ (p. 72). Fry sums up the potential advantages associated with innovation in multi-technology firms in a clearcut way: In a small company, if you had an idea that would incorporate a variety of technologies and you had to go out and buy the equipment to put those together,

136 JENS FRØSLEV CHRISTENSEN

you probably couldn’t afford it, or you’d have to go as inexpensively—or as small —as possible. At 3M, we’ve got so many experts and so much equipment scattered around that we can piece things together when we’re starting off. We can make the adhesive and some of the raw materials here, and do one part over here, and another part there, and convert a space there and make a few things that aren’t available. (p. 41) I shall later return to some of the organisational factors in 3M that promotes a very coherent (and synergistic) technology base. The contextual coherence (or ‘fit’) concept refers to the correspondence between the technology base and the firm’s other assets (the complementary assets, the operational and infrastructural firm context, business and corporate level strategy, management procedures and processes, and culture). Also this dimension of coherence is much more complex in large diversified corporations than in small firms. A ‘bad fit’ may be characterised by rivalry over financial and human resources, and mutual mistrust and lack of effective communication between the ‘innovative asset agents’ and the ‘operational asset agents’ or the top management. A ‘good fit’ is not necessarily marked by harmonic relations and agreement between parties. But possible tensions and conflicts are dealt with in ways that promote synergy and a common sense of contributing to the same general course of direction. Part of the contextual coherence issue can be analysed in terms of couplings between innovative assets and complementary assets (Teece 1986; Christensen 1996b). EMI’s innovation of the CT-scanner provides an example of a bad contextual coherence. The invention and development of the CT-scanner in EMI’s Central Research Laboratories reflected—as mentioned earlier in this chapter—the mobilisation of unique innovative capability, but the innovation did not fit the contextual or complementary assets of the EMI corporation. In short, EMI was a music, entertainment and defence-electronics company and did not possess the ability to manage the ‘innovative amortisation’ of a medical equipment innovation. As Teece (1986) has put it, EMI lacked the complementary assets (in marketing, manufacturing, after-sales service, etc.) to appropriate the rent from a unique innovation. Thus, EMI had to withdraw from the market and let other companies with a more appropriate endowment of complementary assets exploit the huge commercial rents from its innovation (Bartlett 1983). The dynamics of ‘innovative asset amortisation’ represents the acid test of the contextual coherence of the technology base: a good contextual coherence reflects the situation where both top management and the operational parts of the firm (outside the technology base) have strong incentives and either ‘ready’ capabilities to implement the ‘amortisation’ process, or capacities to build—or access from external sources—the appropriate assets. This, of course, presupposes a continuous dialogue between ‘the technology base agents’, ‘the complementary assets agents’ and top management.

BUILDING INNOVATIVE ASSETS AND DYNAMIC COHERENCE 137

The dynamic dimension of the technology base The technology base in a diversified corporation is not necessarily very coherent, and the corporate headquarters may not pursue coherence with respect to the corporation’s technology base. In very decentralised diversified companies in which financial control represents the primary means of exerting corporate influence on the divisions or business units, there will tend to be no or decreasing corporate coherence of the technology base, and coherence may only be pursued at the divisional or business-unit level. This is especially the case for diversified companies in industries which are not very dynamic in terms of technological innovation, or in companies that are so large and diverse that attempts to pursue corporate coherence of the technology base are likely to fail or only succeed at very high bureaucracy and agency costs (Goold and Campbell 1987; Chandler 1994). Below, I shall assume that coherence with respect to the technology base is a strategic objective (whether at the corporate or divisional level) and discuss the dynamics of coherence by applying the proposed conceptual framework. The coherence of the technology base is specified according to its reproductive versus explorative dynamics of the constituent innovative assets, and I propose a simple analytical distinction between reproductive, explorative and dynamic coherence. Reproductive coherence refers to a technology base that is dominated by reproductive dynamics, especially ‘innovative asset improvement’ implying incremental product or process innovation, but also activities to support the ‘innovative asset amortisation’ for which ‘complementary asset agents’ tend to have primary responsibility. Likewise, the dynamics of innovative assets are dominated by reproductive features (maintaining, leveraging and improving existing assets rather than building new assets). The technology base is directed at serving the existing portfolio of product or business lines, and interdependencies focus on exploiting economies of scope (in terms of sharing and transferring existing and partly idle assets) and experiential learning along existing technological trajectories. A technology base that shows reproductive coherence will tend to show strong contextual coherence. On the other hand, such a technology base lacks the stretch and explorative capacities necessary to manoeuvre in the longer term in dynamic competitive environments (Hamel and Prahalad 1993). It is also likely that a strongly reproductive orientation of the technology base will imply a movement towards an increasingly generic and domain-specific (rather than firm-specific and integrative) asset profile, and consequently increasing difficulties in generating competitive advantage. Reproductive coherence is conducive to a market-orientated, incremental technological development and operational efficiency that may provide a competitive edge in the shorter to medium term. Only in industries marked by limited technological opportunities or a stable, conservative user climate, reproductive coherence may provide the firm with sustainable competitiveness (Sanchez and Thomas 1995). In by far the most technology-intensive industries the competitive arena is more dynamic, and if firms continuously engage in exploitation of existing assets to the exclusion of exploration, they become locked into trajectories that gradually come to diverge from those dominating the industry.

138 JENS FRØSLEV CHRISTENSEN

Explorative coherence refers to a technology base that is dominated by explorative efforts (‘innovative asset creation’ and ‘innovative asset build ing’) targeted at radical innovations or radical technological upgrading of existing products or processes on the one hand, and integrative efforts to align the new and the old—but not obsolete—capabilities and competences (‘innovative asset fission’) on the other hand. Interdependencies are here more directed at explorative, synergistic learning and searching which provide the firm with new strategic options in the form of new technological as well as market opportunities (Hamel and Prahalad 1993; Sanchez and Thomas 1995). ‘Innovative asset fission’ is particularly difficult to manage. This type of innovative dynamics involves the building of new innovative assets in the course of producing radical innovation; and these new assets will in turn help improve or revitalise existing products or processes in the firm. This latter type of alignment dynamics is precisely what makes ‘innovative asset fission’ a coherent kind of explorative dynamics and not an autonomous, free-floating type of dynamics that adds new stand-alone assets to the existing technology base or, worse, cannibalises existing assets. The managerial difficulties consist in (1) identifying a promising course of asset building and radical innovation, (2) establishing a synergistic alignment of the new and the old assets, and (3) assuring contextual coherence with respect to the new assets. If there is not a thorough consideration for the contextual coherence, the explorative dynamics can easily fail when it comes to commercialisation or asset amortisation. This will be the more grave if the newly acquired assets do not succeed in revitalisation of existing products and processes.14 However, if successfully managed, ‘innovative asset fission’ will imply the building of firm-specific and integrative assets on which competitive strength will be based. If firms, however, engage in explorative dynamics to the exclusion of reproductive dynamics for extended periods it will become increasingly difficult for the technology base to remain coherent, especially with respect to the contextual dimension, and operational efficiency will tend to erode. To quote March (1991), such firms ‘exhibit too many underdeveloped ideas and too little distinctive competence’ (p. 71). The explorative bias is probably more likely in companies with enduring (1) strong commitment to scientific research and not exclusively development activities, (2) large central R&D labs in which the staff is primarily engaged in long-term projects and not incremental or operational tasks. Du Pont—with its long list of successful breakthrough innovations—is a case in point (Hounshell and Smith 1988). EMI has also shown a remarkable commitment to explorative innovation. However, the company has demonstrated a poor record in exploiting its own technological breakthroughs (Nayak and Ketteringham 1986). This was not only the case with the CT-scanner: ‘EMI failed to pursue an early, superior television technology and profited little from the postwar TV explosion. Their researchers, with Hounsfield leading, built the first large all-transistor computer in 1955, years ahead of any other company’s technology. EMI built twenty-four large machines selling at £250,000 each. They were very successful, until faster transistors curtailed their design life. But EMI hesitated to follow through to the next generation and so missed out on the computer revolution of the 1960s and 1970s’ (p. 125). Thus, it seems that EMI has shown a powerful capacity for ‘innovative assset building’,

BUILDING INNOVATIVE ASSETS AND DYNAMIC COHERENCE 139

but has not been able to ‘amortise’ the innovative assets (at least not the ones mentioned here) and create synergistic effects (innovative asset fission) from its ‘innovative asset building’. Dynamic coherence involves a continuous balancing of the trade-off between the reproductive and explorative dynamics of the technology base. The proper balance is likely to be changing over time, often with cyclical features: after a period of extensive and risky investments and innovative efforts in which the balance favours the explorative side, there will mostly come a period of consolidation and exploitation where the reproductive side takes the lead. However, neither the reproductive nor the explorative side should become heavily dominating for longer periods of time. Sustainable competitive advantage can only be achieved if the management of the corporate technology base continuously strives for the right balance between reproductive and explorative dynamics. To sum up, while reproductive and explorative coherence reflect ‘unbalanced’ types of firm dynamics which will tend to lead to economic trouble (especially in technologydynamic industries), dynamic coherence reflects the ideal type of balanced dynamics of the firm that will tend to provide a sound basis for sustainable competitive advantage. Dynamic coherence implies the capacity: • • • • •

to exploit existing ground efficiently (the reproductive dimension), to explore new territory (the explorative dimension), to strike the ‘right’ balance between the two, to align the old and the new assets (resources, capabilities and competences), and to pursue not only local but also contextual coherence of the technology base.

The long-term evolution of large industrial corporations (such as IBM, Rank Xerox, Philips and General Motors) indicates that exploratively or dynamically coherent corporations often emerge into a position of reproductive coherence, and this may happen without any clear economic warning signals in due time for managerial action. However, once reproductive coherence has turned into an open lock-in crisis, this may trigger a (long) period of reflection (diagnosis, potential solutions), experimentation and trial-anderror in innovative efforts, and power struggles within the corporation. This may lead into a state of either non-coherence or explorative coherence that endangers the life of the corporation, or at least the previous competitive position. But the outcome may also be a new coherence that provides both renewed momentum for the corporation and a dynamic fit with the long-term industry trajectory. There may be a dramatic difference in the evaluation of corporate coherence when viewed from a product/business perspective (as done—at least empirically—by Teece et al. 1994) as compared to a resource-or technology-based perspective (as done in this chapter). A very diverse portfolio of products which seem to be unrelated in terms of customers, distribution channels and ‘industry closeness’ may be strongly coherent from a technology-based perspective. However, the opposite situation may also occur: a diverse

140 JENS FRØSLEV CHRISTENSEN

port-folio of products that are strongly market-related and belong to the same industry or closely related industries may be based on a fragmented or non-coherent technology base. Below, two cases will illustrate the points discussed: 3M provides a case of a very coherent technology base which for decades has demonstrated powerful dynamic features (sometimes with an explorative bias). ICI, on the other hand, illustrates the serious consequences of not being coherent, or more precisely, developing two clusters of businesses between which there were only weak linkages. 3M: a case of a coherent, synergistic technology base 3M possesses a very diverse product portfolio and a very coherent technology base. It was founded as a sandpaper manufacturer (Minnosota Mining & Manufacturing) that around the turn of the century failed to mine corundum, an abrasive used to make sandpaper scratch. When, moreover, the company was outcompeted in trying to sell its own sandpaper using corundum from the East, it was forced to innovate or die (Mitchell 1991: 773). Some of the early innovations were an abrasive cloth for metal finishing and the Wetordry sandpaper, but since then 3M has brought thousands of new products to the market. ‘In dreaming up businesses as diverse as “Post-it” notes, magnetic tape, photographic film, pressure-sensitive tapes, and coated abrasives, the company has brought to bear widely shared competencies in substrates, coatings, and adhesives and devised various ways to combine them. Indeed, 3M has invested consistently in them. What seems to be an extremely diversified portfolio of businesses belies a few shared core competencies’ (Prahalad and Hamel 1990:12). 3M has a remarkable record for innovation and new business creation. Its $14 billion of sales is derived from some forty divisions. More than 80 per cent of 3M’s 60,000-item product portfolio is linked to precision coating technologies, and probably more than 90 per cent works with polymers (Goold et al. 1994:104). ‘Nearly every product made by 3M has been developed by its technical staff in its divisional laboratories. On average, 3M’s laboratories produce many more successful products than its rivals. Much of this success is due to the parenting influence of 3M’s technical function’ (p. 199). The following exposition of 3M’s ‘innovation machine’ is based on the case-study done by Goold et al. (pp. 199–205). 3M has around 150 laboratories involved in some 1,500 new product development programmes: ‘Divisional technical directors report to divisional chief executives. But they also have dotted-line relationships with sector technical directors and the central technical function. The sector technical director has overall responsibility for the technical development of the sector and works closely with the central technical function. Good relations between technical people at the center, the sector level, and the divisions are essential to the smooth working of the 3M system’ (p. 200). Several factors contribute to improve the quality of laboratory management: • The central technical function helps to develop and select technical directors, and the candidates will all be internal technical people.

BUILDING INNOVATIVE ASSETS AND DYNAMIC COHERENCE 141

• The central technical function manages a peer audit process, in which technical directors from other laboratories visit a laboratory and write a report on their findings. These audits help to accelerate learning in R&D management, and the learning flows both ways, from those doing the audit to those being studied and vice versa. • Procedures and cultural norms have become institutionalised in 3M (‘3M ways of doing things’). One example is the 15 per cent rule that permits technical people to spend 15 per cent of their time pursuing their own ideas, as long as they are productrelated (Mitchell 1991:176). • All technical staff in 3M are members of the Technical Forum which is like a professional association with regular meetings to discuss the ‘hot-test technical areas’ or the ‘hottest management issues’. The central technical staff has an important role in ensuring that technology is shared between 3M’s 150 laboratories. A Technology Council made up of researchers from the various divisions regularly gets together to exchange information (Mitchell 1991:180). Divisions are expected to share knowledge and manpower, and informal informationsharing sessions spring up everywhere in 3M (p. 172). ‘This commitment to share is backed by a strong 3M policy that technology belongs to the company not the divisions’ (Goold et al. 1994:201). Goold et al. (p. 203) point to the following distinctive parenting characteristics of 3M: • 3M is a highly decentralised company, but care is taken to decentralise to units that are large enough to be self-standing and, therefore, capable of funding their own research. On the other hand, 3M attempts to prevent divisions from getting so large that individual innovations do not count. 3M’s strong belief in divisional autonomy is countervailed by the corporate requirements of a high level of innovativeness and interdivisional knowledge-sharing. • 3M is strongly committed to nurture its technology and innovation-led culture. This is illustrated by (a) its long-standing target of having 25 per cent of sales from new products launched in the past five years (recently upgraded to 30 per cent of sales from new products launched in the past four years); (b) its target to spend around 7 per cent of sales on R&D; (c) its support for the individual innovator: The Golden Step awards (for US employees) and Pathfinder awards (for international employees) are given to individuals and teams who achieve important product developments. In 1983 3M created the Genesis grant, which gives individual researchers up to $50,000 to carry their projects past the idea stage. As many as ninety grants each year are awarded by a panel of technical experts. (Mitchell 1991:177) Clearly, 3M shows dynamic coherence—perhaps in periods biased towards explorative coherence—in terms of its technology base. This issue is addressed by Mitchell in summing up his story on 3M:

142 JENS FRØSLEV CHRISTENSEN

The result of all this interconnection is an organic system in which the whole is really greater than the sum of its parts. It’s no coincidence that 3M is never mentioned as a possible breakup candidate. Bust it apart, sever the interconnections, and 3M’s energy would likely die. Even if a raider decided to leave it intact, an unfamiliar hand at the helm might send the company off course. (Mitchell 1991:180) 3M may be moving towards a more reproductive type of coherence as some of its main product markets are maturing: In the future, 3M faces the challenge of learning how to compete effectively in more mature market sectors, where manufacturing costs, volumes, and prices are key to success, rather than technology. In the past, growth from new products has made it possible for 3M to deemphasise markets as they matured and the products became commodity items. But in tapes, and more recently in computer diskettes, 3M has decided to stand and fight. It is an alien strategic battle, one that does not play to 3M’s historic parenting strengths. It will be interesting to see whether 3M will succeed in these businesses. (Goold et al. 1994:205) ICI: a case of a fragmented technology base Until the demerging in 1992, Imperial Chemical Industries (ICI) was a fullline chemical company with a wide variety of products. It was established in 1926, when four leading British companies merged with the goal to counter the dominance of Germany’s IG Farben (Owen and Harrison 1995:134). From the beginning the underlying rationale for growth was technology and vertical integration, and R&D ensured a continuing flow of new products (pp. 133ff.). The increasingly complex product portfolio remained biased towards heavy chemicals that had been the core of the founding companies. Pharmaceuticals became a separate division in 1957. During the 1970s the chemical industry began to mature, new entrants challenged the established leaders, and growth rates slowed, as did the flow of innovations (pp. 133ff.). ‘The fierce recession of 1980 to 1981 shocked ICI out of its inertia…ICI went through a painful contraction that included the loss of some 50,000 jobs. The size of the headquarters was cut back, and the number of executive directors on the main board was cut by half, while the product divisions acquired still more autonomy’ (p. 135). The central research laboratory was closed, and the divisions became responsible for their own R&D (p. 138). ICI sought new sources of growth during the 1980s to offset the poor growth and performance of the traditional commodity chemicals. ICI tried to push the portfolio towards higher added value specialty chemicals, advanced materials, pharmaceuticals and electronics (Goold and Campbell 1987:92). However, as Pettigrew (1985) has documented, progress was slow towards these objectives. According to Goold and Campbell (1987:103), the centre’s low-profile role in strategy development may have

BUILDING INNOVATIVE ASSETS AND DYNAMIC COHERENCE 143

inhibited the emergence of a more substantial thrust. During much of the 1980s several of the divisions—especially the commodity-orientated divisions such as Fibres, Organics and Petrochemicals, failed to make a satisfactory return despite major cost reductions, and these divisions were combined with the other bulk chemical interests in a single group, ICI Chemicals and Polymers Group. Goold and Campbell maintain that ‘if these divisions had been independent companies, without support and cash flow from elsewhere in ICI, the need to survive might have dictated even more radical change’ (p. 103). By the end of the 1980s the complexity of the business portfolio had rather increased than decreased: on the one hand, ICI had moved into new areas, such as specialty chemicals, while on the other hand remaining in most of the existing commodity fields in which international competitiveness was deteriorating (Owen and Harrison 1995:135). ICI consisted of about twenty divisions—largely independent profit centres— operating in fifty to sixty businesses, and according to Goold and Campbell (1987:87) the degree of inter-divisional overlap in terms of common technologies, resources, markets, skills and potential for synergies from coordination was generally low. During the 1980s corporate strategy was characterised by ‘strategic control’ style implying that even if the corporate centre was concerned with business-unit planning, there was a strong belief in autonomous profit-responsible business units. Rather than advocating particular ways ahead, the ICI centre focused on establishing planning processes, reviewing business-unit proposals, and exercising tight financial control (pp. 86 and 88). The corporate management team had ‘been willing to sacrifice some centrally managed synergies to create a greater sense of independence in the divisions and businesses (p. 93). At the start of the 1990s, ICI faced a takeover threat because the value of its best businesses (especially pharmaceuticals) was not reflected in the price of its stocks. The whole company was worth less than the sum of its parts (this and the following is based on Owen and Harrison 1995). In May 1991 Hanson, the British conglomerate with a reputation for taking over badly managed companies, announced that it had acquired 2.8 per cent of ICI. Even though it became clear that Hanson would not launch a bid, the episode demonstrated the urgent need to raise shareholder value. An outsider, John Mayo, was brought in to provide an analysis of the situation and suggest solutions. Mayo, in fact, made an analysis of the coherence of ICI: ‘He built up a picture of which businesses shared science and technology in common, what synergy existed among them, and how far that synergy was being exploited’ (p. 137). ‘Mayo concluded that the key to successful restructuring was recognising a technological fault line within ICI. Pharmaceuticals and other bioscience-related activities fell to one side of the fault line; the traditional chemical businesses fell to the other. The fault line divided two coherent groups of businesses that could be managed as separate companies. Within each group—but not across them— there was the potential for mutual support and interdependence’ (p. 137). There were no or only weak connections between pharmaceuticals and most of the chemical businesses. With two exceptions: ‘One was agrochemicals, a bioscience-based, research-intensive business that, like pharmaceuticals, depended on the intimate intermingling of biology and organic chemistry. In research and development, they enjoyed a clear synergy’ (p. 138). The other was the dye business: ‘Its manufacturing base in fine organic chemicals provided

144 JENS FRØSLEV CHRISTENSEN

the building blocks for many pharmaceutical and agrochemical products’ (p. 138). Also with respect to the chemical group, the shared technical capabilities and managerial competences provided the basis for a coherent cluster of businesses: ‘Most of the chemical businesses used large plants for high-volume production, and they had similar needs in chemical engineering and project management. In marketing, too, the differences between the chemical and bioscience businesses was clear’ (p. 138). In July 1992 the plan for a demerging of ICI was approved. The cluster of bioscience businesses became Zeneca, and the cluster of chemicals formed the new ICI. ICI seems to have been an innovative company showing dynamic coherence in the first decades after it was founded. In the 1960s and 1970s ICI was still relatively coherent, but increasingly with a reproductive bias and slowing growth. Efforts to build new growth businesses around pharmaceuticals, agrochemicals, etc., created a new cluster of businesses among which dynamic or explorative coherence prevailed. However, overall coherence was weakened, and the fragmentation between the dynamic pharmaceuticals and biotechnology cluster and the traditional (reproductive) bulk chemicals cluster paved the way for the demerging of ICI. To some extent ICI succeeded in building new innovative assets. However, ICI never succeeded in aligning these new assets with the traditional chemical assets and thus creating an overall process of ‘innovative asset fission’. Concluding remarks This chapter has addressed the issue of innovative dynamics, both at the level of innovative assets, and at the aggregated level of the technology base of the multi-technology firm. The technology base of the firm is a primary locus of strategic capabilities and competences, and to analyse the coherence of the technology base from a competencebased perspective may help to provide guidance for effective management of the technology base, and thus contribute to creating a robust paradigm for management of innovation and technology strategy. But moreover, the focus at the coherence of the technology base from a firm or corporate perspective may also contribute to the development of a ‘coherent’ theory of overall business and corporate strategy (for an explorative attempt see Christensen and Foss 1997; Foss and Christensen 1999). Within the resource-based literature the internal coherence problematic has so far primarily been addressed empirically in terms of product-portfolio coherence and its implications for strategic diversification. The resource-and competence-based theory of the firm and its strategic ramifications needs a better understanding of asset and competence coherence along the lines proposed in this chapter, and future research should be directed towards a better understanding of the managerial and organisational implications of pursuing dynamic coherence of the firm and its technology base.

BUILDING INNOVATIVE ASSETS AND DYNAMIC COHERENCE 145

Notes 1 This notion of capabilities is close to the notion of routines as used by Nelson and Winter (1982) and further discussed by Cohen et al. (1996). I prefer the term capability because it signifies not only the capacity for replicative or recurring behaviour but also the capacity to search and innovate. 2 The important issue of the boundaries of the firm’s technology base (that is, to what extent and under which conditions may innovative assets located outside the firm be considered part of the firm’s technology base) will not be addressed in this chapter. 3 The concept of autonomous versus systemic innovation, proposed by Teece (1988), may be specified in terms of innovative and complementary assets: autonomous innovation only requires product or process innovation (or innovation in one complementary asset), while systemic innovation requires sequential or parallel innovation in products and/or processes as well as in one or more complementary assets (or value chain activities). 4 Of course, in the manifestation of innovative assets there are strong elements of organisational skills involved. Likewise, there are both internal learning dynamics and external learning, or absorptive capacity, involved. I have dealt more thoroughly with these issues elsewhere (Christensen 1996a). 5 I am grateful to Rod Coombs for suggesting the term ‘domain-specific’. 6 The use of the notion of integrative assets or competences is inspired by the similar notion, ‘integrative capability’ (Henderson 1994), ‘dynamic capability’ (Teece and Pisano 1994) and ‘technology integration’ (Ianseti and West 1997). 7 Using the notion of appropriability regime (Teece 1986) we can say that integrative assets mostly reflect a stronger ‘natural’ appropriability regime than domain-specific assets— whether generic or firm-specific. 8 Asset dynamics may involve both functional/technical capabilities and managerial competences. 9 They do not make the distinction between capability and competence, but use the term competence in much the same way I use the term asset. 10 For example, the relaunching of ‘Brylcreme’ in the 1980s to exploit a wave of nostalgia and revitalisation of young men’s hair fashion of the 1950s. 11 This issue has been more broadly analysed by Anderson and Tushman (1990) in terms of technological discontinuities within given product areas. They distinguish competenceenhancing discontinuities from competence-destroying discontinuities that make existing competences obsolete. They find that incumbent firms have a better chance of managing their way through competence-enhancing discontinuities, while they have a significant risk of failing to manage a competence-destroying transformation because they are locked into core competences that are obsolete and therefore cannot guide the transformation. 12 Coherence may also be related to the degree of ‘fit’ between the elements of the technology base and the requirements of the competition (termed ‘external coherence’ in Christensen 1998). However, I shall refrain from discussing this issue here. 13 Coombs and Richards (1991a, 1991b) provide some illustrative evidence on the complexity of promoting coherence with respect to the technology base in very large diversified companies. 14 In the EMI CT-scanner case ‘innovative asset fission’ would have implied that the technical capabilities that were built in the process of developing the CT-scanner subsequently played an active role in revitalising other businesses in EMI than the scanner business

146 JENS FRØSLEV CHRISTENSEN

References Agyris, C. and D.Schön (1978) Organizational Learning, Reading, MA: Addison-Wesley. Anderson, P. and M.L.Tushman (1990) ‘Technological Discontinuities and Dominant Designs: A Cyclical Model of Technological Change’, Administrative Science Quarterly December:604–33. Arrow, K.J. (1962) ‘The Economic Implications of Learning by Doing’, Review of Economic Studies 29 (80):155–73. Barney, J. (1986) ‘Strategic Factor Markets: Expectations, Luck and Business Strategy’, Management Science 42:1231–41. Bartlett, C. (1983) ‘EMI and the CT Scanner’, Harvard Business School Case Services, Harvard Business School. BusinessWeek (1980) GE Gobbles a Rival in CT Scanners, no. 2,637 May 19:40. Chandler, Jr., A.D. (1994) ‘The Functions of the HQ Unit in the Multibusiness Firm’, in R.P.Rumelt, P.E.Schendel and D.J.Teece (eds), Fundamental Issues in Strategy: A Research Agenda, Boston, MA: Harvard Business School Press. Christensen, J.F. (1995) ‘Asset Profiles for Technological Innovation’, Research Policy 24:727–45. ——(1996a) ‘Analyzing the Technology Base of the Firm: A Multi-Dimensional Resource and Competence Perspective’, in N.J.Foss and C.Knudsen (eds), Towards a Competence Theory of the Firm, London: Routledge. ——(1996b) ‘Innovative Assets and Inter-Asset Specificity—A Resource-Based Approach to Innovation’, Economics of Innovation and New Technology 4:193–209. ——(1998) ‘Management of Technology in Multiproduct Firms’ in R.Coombs, K. Green, A.Richards and V.Walsh (eds), Technological Change and Organization, Cheltenham: Edward Elgar. Christensen, J.F. and N.J.Foss (1997) ‘Dynamic Corporate Coherence and Competence-Based Competition: Theoretical Foundations and Practical Implications’, in A.Heene and R.Sanchez (eds), Competence-Based Strategic Management, Chichester and New York: John Wiley and Sons. Christensen, J.F., M.O.Nielsen and M.Iversen (1998) ‘Strategy, Structure and Management of Technology at Grundfos’, CISTEMA Working Paper, Copenhagen Business School. Cohen, M.D., R.Burkhart, G.Dosi, M.Egidi, L.Marengo, M.Warglien and S. Winter (1996) ‘Routines and Other Recurring Action Patterns of Organizations: Contemporary Research Issues’, Industrial and Corporate Change 5:653–98. Cohen, W.M. and D.A.Levinthal (1990) ‘Absorptive Capacity: A New Perspective on Learning and Innovation’, Administrative Science Quarterly 35:128–52. Coombs, R. and A.Richards (1991a) ‘Technologies, Products and Firms’ Strategies. Part 1—A Framework for Analysis’, Technology Analysis and Strategic Management 3(1):77–86. ——(1991b) ‘Technologies, Products and Firms’ Strategies. Part 2—Analysis of Three Cases’, Technology Analysis and Strategic Management 3(2):157–75. Doz, Y. (1994) Managing Core Competency for Corporate Renewal: Towards a Managerial Theory of Core Competencies, Working Paper, Paris, INSEAD. Foss, N.J. and J.F.Christensen (1999) ‘A Market-Process Approach to Corporate Coherence’, Managerial and Decision Economics (forthcoming). Foss, N.J. and C.Knudsen (eds) (1996) Towards a Competence Theory of the Firm, London: Routledge. Goold, M. and A.Campbell (1987) Strategies and Styles: The Role of the Centre in Managing Diversified Corporations, Oxford: Basil Blackwell. Goold, M., A.Campbell and M.Alexander (1994) Corporate-Level Strategy: Creating Value in the Multibusiness Company, New York: John Wiley and Sons.

BUILDING INNOVATIVE ASSETS AND DYNAMIC COHERENCE 147

Granstrand, O., C.Oskarsson, N.Sjöberg and S.Sjölander (1990) ‘Business Strategies for New Technologies’, in E.Deiaco, E.Hörnell and G.Vickery (eds), Technology and Investment: Critical Issues for the 1990s, London: Pinter. Granstrand, O., P.Patel and K.Pavitt (1997) ‘Multi-Technology Corporations: Why They Have “Distributed” Rather than “Distinctive Core” Competences’, California Management Review 39 Summer:8–25. Grant, R.M. (1991) ‘The Resource-Based Theory of Competitive Advantage: Implications for Strategy Formulation’, California Management Review 33 Spring:114–35. Hamel, G. and A.Heene (eds) (1994) Competence-Based Competition, New York: John Wiley and Sons. Hamel, G. and C.K.Prahalad (1993) ‘Strategy as Stretch and Leverage’, Harvard Business Review March–April:75–84. Hedlund, G. (1994) ‘A Model of Knowledge Management and the N-Form Corporation’, Strategic Management Journal 15, Special Issue:73–90. Heene, A. and R.Sanchez (eds) (1997) Competence-Based Strategic Management, New York: John Wiley and Sons. Henderson, R.M. (1994) ‘The Evolution of Integrative Capability: Innovation and Cardiovascular Drug Discovery’, Industrial and Corporate Change 3:607–30. Hounshell, D.A. and J.K.Smith Jr. (1988) Science and Corporate Strategy: Du Pont R&D, 1902–1980, Cambridge: Cambridge University Press. Iansiti, M. and K.B.Clark (1994) ‘Integration and Dynamic Capability: Evidence from Product Development in Automobiles and Mainframe Computers’, Industrial and Corporate Change 3(3): 557–605. Iansiti, M. and J.West (1997) ‘Technology Integration: Turning Great Research into Great Products’, Harvard Business Review May–June:69–79. Kim, D.-J. and B.Kogut (1996) ‘Technological Platforms and Diversification’, Organization Science 7: 283–301. Leonard-Barton, D. (1992) ‘Core Capabilities and Core Rigidities: A Paradox in Managing New Product Development’, Strategic Management Journal 13:111–25. Levinthal, D.A. and J.G.March (1993) ‘The Myopia of Learning’, Strategic Management Journal 14: 95–112. March, J.G. (1991) ‘Exploration and Exploitation in Organizational Learning’, Organization Science 2:71–87. Markides, C.C. and P.J.Williamson (1994) ‘Related Diversification, Core Competences and Corporate Performance’, Strategic Management Journal 15:149–65. Mitchell, R. (1991) ‘Masters of Innovation: How 3M Keeps its New Products Coming’, in J.Henry and D.Walker (eds), Managing Innovation, London: Sage. Morita, A. (1991) ‘Selling to the World: The Sony Walkman Story’, in J.Henry and D.Walker (eds) Managing Innovation, London: Sage. Nayak, P.R. and J.M.Ketteringham (1986) Breakthroughs!, New York: Rawson Associates. Nelson, R. and S.Winter (1982) An Evolutionary Theory of Economic Change, Cambridge, MA: The Belknap Press. Oskarsson, C. (1993) Technology Diversification: The Phenomenon, Its Causes and Effects, Göteborg: Chalmers University of Technology, Department of Industrial Management and Economics. Owen, G. and T.Harrison (1995) ‘Why ICI Chose to Demerge’, Harvard Business Review 72:133– 43. Pavitt, K., M.Robson and J.Townsend (1989) ‘Technological Accumulation, Diversification and Organization in UK Companies 1945–1983’, Management Science 35:81–99.

148 JENS FRØSLEV CHRISTENSEN

Pettigrew, A. (1985) The Awakening Giant: Continuity and Change in Imperial Chemical Industries, Oxford: Basil Blackwell. Prahalad, C.K. and G.Hamel (1990) ‘The Core Competence of the Corporation’, Harvard Business Review May–June:79–91. Rosenberg, N. (1982) Inside the Black Box: Technology and Economics, Cambridge: Cambridge University Press. Sanchez, R., A.Heene and H.Thomas (1995): ‘Towards the Theory and Practice of CompetenceBased Competition’, in R.Sanchez, A.Heene and H.Thomas (eds), Theory and Practice in Competence-Based Competition: From Industry Studies to a New Theory of Competitive Dynamics, London: Elsevier Pergamon. Sanchez, R., A.Heene and H.Thomas (eds) (1996) Dynamics of Competence-Based Competition: Theory and Practice in the New Strategic Management, Oxford: Elsevier Science. Sanchez, R. and H.Thomas (1995) ‘Strategic Goals’, in R.Sanchez, A.Heene and H. Thomas (eds), Theory and Practice in Competence-Based Competition: From Industry Studies to a New Theory of Competitive Dynamics, London: Elsevier Pergamon. Sanderson, S. and M.Uzumeri (1995) ‘Managing Product Families. The Case of the Sony Walkman’, Research Policy 24:761–82. Teece, D.J. (1986) ‘Profiting from Technological Innovation: Implications for Integration, Collaboration, Licensing and Public Policy’, Research Policy 15:285–305. ——(1988) ‘Technological Change and the Nature of the Firm’, in G.Dosi, C.Freeman, R.Nelson, G.Silverberg and L.Soete (eds), Technical Change and Economic Theory, London: Pinter. Teece, D. and G.Pisano (1994) ‘The Dynamic Capabilities of Firms: An Introduction’, Industrial and Corporate Change 3:537–56. Teece, D., G.Pisano and A.Shuen (1997) ‘Dynamic Capabilities and Strategic Management’, Strategic Management Journal 18:509–33. Teece, D.J., R.Rumelt, G.Dosi and S.Winter (1994) ‘Understanding Corporate Coherence. Theory and Evidence’, Journal of Economic Behavior and Organization 23:1–30.

8 Dynamic complementarities and technology acquisition Tomi Laamanen and Erkko Autio

Introduction In the resource-based view, the firm grows because it possesses distinctive, costly-to-copy resources (Barney 1991) that open ‘productive opportunities’ (Penrose 1959) for it to exploit. As the firm ‘sees’ and entrepreneurially exploits these opportunities, new business areas are opened, and the firm needs to undergo an administrative reorganisation to manage efficiently the new businesses created. In the resource-based view, the firm thus constitutes a ‘coherent administrative framework’ (Ghoshal et al. 1997) that manages internal and external resources or competencies for growth. Companies have several ways to develop resources and competencies. They may develop these internally, or they may end up acquiring them from outside the firm boundaries. External competence acquisition can be carried out, for example, by acquiring a patent or a licence, or by entering into a collaborative arrangement, such as a development joint venture. In some cases, technological competencies can be acquired by buying a new, technology-based company that has developed and possesses the required competencies. It is important to understand the role of the different ways to source technological competencies externally. In this respect, company acquisitions represent an interesting and less researched area. The purpose of this chapter is to study the external sourcing of technological competencies empirically by investigating new, technology-based companies. New, technology-based companies are defined here as companies owned and run by an entrepreneur or a group of entrepreneurs, established as separate legal entities in order to develop and commercialise new technologies. They possess technologies new to the acquiring companies. Even though these companies typically are young and innovative, the definition also includes, for example, more mature family-owned companies with technologies not held previously by the acquiring companies. The acquisition of such small companies provides tangible examples of what happens when new competencies are infused into the administrative framework of the acquiring organisation. In what follows, previous research on internal competence development and external competence acquisition is briefly discussed. The model and the hypotheses put forward in this chapter are introduced. The empirical part contains a multiple case study on the

150 TOMI LAAMANEN AND ERKKO AUTIO

acquisition of competencies through the acquisition of new, technology-based companies. An analysis of the company acquisitions on country, industry, company and competence levels is presented. We end with a discussion of the results and conclusions. Definitions and review of previous research on technology acquisition Despite the growing number of writings related to competence-based strategic management (Heene and Sanchez 1997; Sanchez and Heene 1997), there is still some conceptual confusion in defining the main concepts. Some scholars distinguish between atomistic competencies and higher-order organising principles that can develop and modify atomistic competencies. Henderson and Cockburn (1994) call such higher-order competencies architectural competencies while Teece et al. (1992) call them dynamic capabilities. Sanchez et al. (1996a) view capabilities as lower-level competencies. To avoid confusion, it is necessary to define the main concepts used in this chapter. • Company acquisition. When discussing the acquisition of companies, the word acquisition is used instead of the word merger. We focus on the acquisition of small firms by large firms, so this is justified. Acquisition requires the transfer of 50 per cent of voting power, or effective control of the acquired firm, to the acquirer. Acquisition of parts of companies and separable assets is not included. • Competence acquisition. Competence acquisition can take place in connection with company acquisitions, but also through other means. A company acquisition may result in competence acquisition even if the competencies possessed by the acquired company are not the main motive for the acquisition. Competencies can be acquired also by other means than through company acquisitions. • Competence. Competence comprises local abilities and knowledge that are fundamental to day-to-day problem-solving. Competencies are embedded in an organisation’s knowledge structures. Following Henderson and Cockburn (1994), organisational competencies are defined as also comprising the resources required to use them. For example, distribution is an organisational competence. The distribution competence comprises the physical distribution system as well as the ability to coordinate and operate it. • Resource. Following Barney (1991), resources are defined as comprising physical resources, human resources and organisational resources. Resources comprise stocks of available factors of production owned and controlled by the company. Resources also comprise assets linked with different types of competencies and capabilities. • Technology. According to a polytrophic definition, technology includes a physical part called technoware, articulated knowledge called infoware, an organisational part called orgaware and individual skills called humanware. It includes also a tacit component (Ramanathan 1996). • Dynamic complementarity. The concept of dynamic complementarity (Rothwell 1983) is defined in the context of company acquisitions and collaboration to mean the

DYNAMIC COMPLEMENTARITIES AND TECHNOLOGY ACQUISITION 151

complementary activities or resource profiles of the involved companies. To be complementary, activities need to match functionally. The attribute dynamic means that activities are performed through the complementary parts. There are only a few studies that have systematically compared internal development, collaborative arrangements, and company acquisitions as modes of competence acquisition. The studies of Harrigan (1986), Lengnick-Hall (1991) and Roberts and Berry (1985) are among the few to provide some comparison of the three modes. Harrigan discusses the three modes of competence acquisition from the perspective of joint ventures. She identifies conditions for joint venture success and develops hypotheses concerning other modes of governance. Lengnick-Hall analyses the modes of competence acquisition as three designs for corporate innovation. She ends up recommending different designs for corporate innovation in connection with different generic strategies and technology domain strategies. According to Lengnick-Hall, acquisitions are most advantageous when the company has adopted a defender strategy combined with cost leadership and differentiation strategies. Taking the perspective of business entry, Roberts and Berry also discuss the three modes of competence acquisition. They develop a matrix that incorporates the market familiarity and technology familiarity dimensions of a new business area. Depending on the two dimensions, Roberts and Berry propose different market entry strategies. According to them, companies are fully equipped to undertake all the aspects of business development within the ‘base’ and ‘familiar’ sectors. The most valuable potential entry strategies in these sectors include internal development, licensing and acquisitions. For the entry to the less familiar sectors, Roberts and Berry recommend joint ventures. They label company acquisitions in less familiar sectors as educational acquisitions. The body of research comparing modes of competence acquisition presents relatively general evidence concerning the benefits and drawbacks of the different modes. Research papers on individual modes go somewhat deeper into the dynamics of the modes in question. Unfortunately, these research papers often seem to overlook the interrelationships and similarities between the different modes. For example, company acquisitions are often referred to by researchers of collaborative arrangements as the end of a continuum with the highest intensity of collaboration or as the most integrated form of collaboration, as when Quélin (1997) lists the need for ‘integration of the involved parties and resources’ as one drawback in company acquisitions when compared with collaborative arrangements. In reality, company acquisitions do not necessarily require close integration. The companies involved in an acquisition can have much less intensive operative collaboration than companies engaged in intensive technology transfer. Rather, the degree of integration of the acquired company is determined by the acquisition situation and by the characteristics of the acquiring and the acquired organisations (see, for example, Haspeslagh and Jemison 1992). Dimensions other than integration can play a more important role in deciding whether to acquire a company or to enter into a collaborative arrangement. For example, the need to keep the collaborative partner out of the reach of competitors may sometimes trigger

152 TOMI LAAMANEN AND ERKKO AUTIO

an acquisition. After the acquisition, the collaboration can continue as before, but competitors can now be efficiently excluded from the relationship. Another reason for acquiring a partner may be that the collaboration entails large idiosyncratic investment. Companies may not be as willing to invest in technological competence development in a collaborative relationship as in internal development. The fear of hold-up problems may cause a company to acquire its collaborative partner. Asymmetries between the collaborative partners may also result in a situation where the larger partner has invested large sums of money in the collaboration. The smaller collaborator may not be able to invest as much and as a result gradually ends up being owned by the larger company. Research focusing on company acquisitions does not seem to have emphasised competence dynamism and learning in a way similar to the research on collaborative arrangements (see, for example, Doz 1996). Many of the results concerning the dynamism of collaborative arrangements are, however, directly transferable to company acquisition research. The view of technological competence acquisition as a complex, multi-mode phenomenon which we have adopted in this chapter may provide a new way to look at company acquisitions. It may help to bring together research in the fields of technology management and company acquisitions. The competence perspective can further research on company acquisitions by improving the focus on variables on the competence level of analysis. It would also seem to extend the scope of existing research on company acquisitions from the acquisition transactions and the integration processes to the collaboration processes before and after the acquisition. An understanding of the importance of creating suitable conditions for competence transfer in company acquisitions has emerged in recent research on company acquisitions (Haspeslagh and Jemison 1994). The initial conditions create the prerequisites for successful company acquisitions while the acquisition process eventually realises the acquisition outcome. Success in company acquisitions is often linked to the technological success of the acquisition or the success in transferring competencies between the organisations. This can be seen to be particularly true in competence-motivated company acquisitions, but also applies in other cases. Even though the motives for company acquisitions can be diverse, the acquiring and the acquired company typically engage in some form of competence transfer. This competence transfer can take place as early as the pre-acquisition stage or during the post-acquisition stage. On the macro level, the view of the different modes of competence acquisition as bundles of competence acquisitions provides a further opportunity, namely, to increase the understanding of the role of new, technology-based companies in the technological development of industries. This relates to the Schumpeterian models of industry development. On the other hand, it also provides an opportunity to examine the influence of large companies in the technological development of industries. The competence acquisition approach to company acquisition research provides an internal view of industry evolution in contrast to the external view that could be created by looking predominantly at industry structure. Contrary to common belief, many company acquisitions are not end results of a systematic search. As Gulati (1995) has noted, company acquisitions also take place in a

DYNAMIC COMPLEMENTARITIES AND TECHNOLOGY ACQUISITION 153

social context: companies are often acquired because of previous familiarity, or because of geographical proximity, rather than because of competence complementarity. This is one reason why corporate coherence is often so difficult to preserve. Companies carrying out a large number of acquisitions may develop distinctive competence in carrying out the operative competence transfer, acquisition negotiation and post-acquisition integration. However, the acquiring companies sometimes lack higher-level acquisition competences. Despite the ability to implement acquisitions effectively, companies may not have sufficient ability to decide which and how many companies to acquire. As a result, the acquiring companies may fail to develop sufficient corporate coherence or dynamic corporate coherence (Teece et al. 1994; Christensen and Foss 1996). The question of corporate coherence relates to the question of what companies to acquire, rather than to how they should be acquired (Sanchez 1997). This chapter starts from the presumption that dynamic complementarities provide a central element when analysing the strategic logic of competence-based company acquisitions. It is assumed that improved dynamic corporate coherence is achieved by attempting to maximise the benefits of dynamic complementarities. Based on this assumption, we have developed a model of company acquisition activity for empirical testing. Hypothesised model Industries follow their own development trajectories. Industry-level development trajectories are intertwined with the development trajectories of other industries and the national-institutional environment. As the industry competence base develops, the industry structure develops as well. Boundedly rational companies position themselves in relation to industry and competence evolution in a way that strives to maximise value created using the information and capabilities available. In positioning themselves, companies can use different modes of competence acquisition. They can also carry out interventions in the industry (by forming strategic groups and alliances, for example) to change industry structure. The different mixes of the modes of competence acquisition and intervention in the industry can be seen as realisations of the competitive strategies of companies. From the resource-or knowledge-based perspective, one important factor influencing modes of collaboration is the existence of dynamic complementarities between large and small companies. Dynamic complementarities may provide companies with incentives for collaboration. The existence of synergistic potential, be it complementary or overlapping competencies and resources, may also provide companies with incentives to acquire other companies. The dynamism catalysed by dynamic complementarities between small and large firms can be hypothesised to take place as shown in the process model in Figure 8.1. The process model integrates some of the principal ideas on industry and technology development by Utterback and Abernathy (Utterback and Abernathy 1975; Utterback 1994) and Tushman and Anderson (Tushman and Romanelli 1985; Tushman and

154 TOMI LAAMANEN AND ERKKO AUTIO

Figure 8.1 A process model on the evolution of technology-based company networks and dynamic complementarities with the assumption that a fundamental discontinuity starts the evolution

Anderson 1986; Anderson and Tushman 1990), applying them to innovation networks. Utterback and Abernathy identify three main phases in industry evolution: the fluid phase, the transitional phase and the specific phase. The rate of product innovation in an industry or product class is highest during its formative years, in the fluid phase. In the transitional phase, the rate of product innovation slows down, and the rate of process innovation speeds up. Some industries may enter the specific phase when both the rate of product and the rate of process innovation goes down. Technological change can be seen to follow an evolutionary pattern in which a breakthrough or a discontinuity initiates a time of variation and selection. This period of variation and selection culminates in a dominant design followed by a period of incremental progress. The period of incremental progress can again be broken by another discontinuity. As a technology matures, the number of competing product designs commonly decreases. There is an increasing need for large capital expenditures owing to the need to invest in production and the development of sales and distribution channels. Since new, technology-based companies often do not possess sufficient resources, a need arises to establish collaborative arrangements with established large companies: the dynamic complementarities between large and small companies increase in relevance. New, technology-based companies can provide innovative new technologies, while large

DYNAMIC COMPLEMENTARITIES AND TECHNOLOGY ACQUISITION 155

established companies can provide the necessary resources for the commercialisation of the new technologies. The purpose of the empirical part of this chapter is to analyse to what extent empirical data support the hypothesised model. According to the model, dynamic complementarities play an important role in attracting large and small technology-based companies to collaborate with each other. According to the model, dynamic complementarities have the greatest influence when the main emphasis in industry evolution shifts from predominantly innovative activities to production, marketing and distribution dominated activities. According to the model, new, technology-based company acquisitions should become more common in the mature phases of the technology life cycle. Empirical sample The company acquisition cases were selected from a comprehensive database maintained by a major Finnish business magazine, Talouselämä. This magazine has kept records of all company acquisitions that have taken place in Finland since 1980. To control for variation caused by economic cycles (the volume of company acquisitions is strongly affected by these), the economic half-cycle from 1987 to 1995 was selected as the period of study. Out of the total of 4,531 acquisitions in Finland during the period 1987–95, 109 were initially selected for closer inspection. Four selection criteria were used. The first criterion was to select companies that had been sold by entrepreneurs. The second was to select companies that based their business on the exploitation of new technological competencies. This was determined on the basis of the description of the acquired company in the business press. The third criterion was to select companies that were acquired by other companies. The fourth was to select companies that were not mere capacity expansions for the acquiring company. To be included in the sample, it was required that the acquired companies could have provided some new technological competencies for the acquiring company. The presidents, vice presidents or chief technology officers of the acquiring companies were contacted to check the conformance of the initial sample with the sample selection criteria. Carrying out a more detailed screening and a final selection of the acquisition cases in the first interviews improved the validity of the sample. This was feasible in Finland where there are not too many large companies in technology-based industries. For a more detailed description of the case selection process, see Laamanen (1997). As a result of the additions and deletions, the total number of cases included in the study increased somewhat from the initial selection of 109 companies. The response rate of the study was 111/119, or 93 per cent. Due to their small proportion, the missing cases are not considered to cause significant bias in the sample. Even though the 111 cases represent only 2.4 per cent of all the 4,531 acquisitions in Finland during the period, the study provides good coverage of the population where the seller was an entrepreneur, the acquired company based its business idea on technological competencies, the company was acquired by another company, and the company acquisition was not motivated by

156 TOMI LAAMANEN AND ERKKO AUTIO

capacity expansion. Since the population of company acquisitions that met the selection criteria was comparatively small, a survey of the whole population was possible. A pie chart representing the industry distribution of the sample is shown in Figure 8.2. Nearly one-third of the acquisition cases, 29 per cent, took place in the electronics industry; 25 per cent of the acquisition cases took place in information technology sectors. Approximately 18 per cent of the acquisition cases represented the metals and machinery sectors. These industry sectors represent a relatively small proportion of all sectors where company acquisitions have taken place during the chosen time period. On the other hand, in some of the industry sectors in the sample, the studied acquisition cases are relatively representative. For example, the electronics acquisitions studied in this research represent 19 per cent of all the acquisitions in the electronics industry sector in Finland during the period. Similarly, the information technology acquisitions represent 19 per cent of all the acquisitions in information technology sectors during the period under study. The studied population of 111 company acquisition cases is clearly different from the overall population of 4,531 company acquisitions in the Talouselämä database. In addition to different industry sector distribution, there are also other systematic differences. For example, the acquired com-panies in the sample are smaller than those in the overall population. The average sales revenue of the overall population of acquired companies in the database was 130 MFIM compared to 24.6 MFIM in the sample (FIM 1=US$0.20). Similarly, the average number of employees in the overall population of acquired companies was 163 employees compared to 47 in the sample of competence-motivated acquisitions (Figure 8.3). Empirical analysis Acquisition behaviour in different phases of industry evolution According to the hypothesised model, dynamic complementarities play an important role in attracting large and small technology-based companies to collaborate with each other and subsequently to acquire each other. On the basis of the new, technology-based company acquisition cases analysed, this would seem to be the case. In all the company acquisition cases, there was at least one complementarity present. In 92 per cent of the 111 studied cases, at least three complementarities were identified. In 68 per cent of the 111 studied cases, at least four complementarities were identified. According to the hypothesised model, dynamic complementarities play the most important role when the main emphasis of industry evolution shifts from predominantly innovative activities to production, marketing and distribution dominated activities. According to the model, new, technology-based company acquisitions should become more common in the mature phases of the industry technology life cycle. Since the technology life cycle phase is difficult to operationalise in practice, the maturity of the acquired companies is used as an approximation. As technological competencies are

DYNAMIC COMPLEMENTARITIES AND TECHNOLOGY ACQUISITION 157

Figure 8.2 Distribution of the studied acquisitions according to the industry; N=111

companyspecific, the maturity of the acquired companies themselves represents a reasonable approximation of the maturity of the company-specific competencies. It should be pointed out that the competencies acquired by acquiring the new, technology-based companies can be mature even though the competencies themselves are required to be new to the acquiring company. The maturity distribution of the sample companies is shown in Table 8.1. Maturity is operationalised on the basis of the cash flow curves, not on the basis of the ages of the companies. The overall distribution provides support for the model: the differences between the different stages are statistically significant (p<0.001; Chi-square). Industry differences can be detected. The main emphasis in acquisitions in the metal and machinery sectors is on mature companies. This may be due to the fact that the acquired companies are mostly production-based. In contrast, the main emphasis in acquisitions carried out in the electronics industry seems to be on start-up companies. In the field of information technology, most of the companies have been acquired in the growth and maturity stages. Engineering company acquisitions represent acquisitions of technologically mature companies. The largest number of company acquisitions is found in the categories of growing and mature companies. Despite the somewhat larger number of acquisitions in the more mature categories, new, technology-based companies in many industries seem to have been acquired in all

158 TOMI LAAMANEN AND ERKKO AUTIO

Figure 8.3 Size distribution of the companies in the sample Table 8.1 Life cycle stages of the acquired companies cross-tabulated with industry sectors

phases of their life cycles. It thus seems that dynamic complementarities between companies are also important motives for acquisitions earlier than hypothesised in the model. Furthermore, pre-acquisition collaboration between the acquired and acquiring companies does not seem to be common or even a precondition for successful activation of dynamic complementarities: new, technology-based companies can also become encapsulated within the acquiring companies. This encapsulation helps to preserve the

DYNAMIC COMPLEMENTARITIES AND TECHNOLOGY ACQUISITION 159

beneficial dynamic complementarities between the acquired and acquiring companies by stimulating them to collaborate. While complete integration typically destroys the dynamic complementarity nature of the collaboration, some degree of integration seems to be needed to activate the complementarities. Success factors of technology-motivated acquisitions In addition to the behavioural explanatory nature of the model, it was hypothesised that different dynamic complementarities have different implications for the success of the new, technology-based company acquisitions. To check this, an analysis of the success factors of the acquisitions was carried out. In the study, company acquisition success was measured as a compound function of the strategic success, technological success, market success and financial success of the acquisition. Each dimension of success was rated using a scale from • 1 to +1. A successful case was rated as +1. An unsuccessful one was rated as • 1. Otherwise, the company acquisition was rated as ‘0’ on the success dimension in question. The degree of technological success was measured by the successful development of a new product generation, successful transfer of technology from the acquired company to the acquiring company, successful transfer of technology from the acquiring company to the acquired company, or a successful combination of the technologies of the acquiring and the acquired company. Market success was measured according to the growth of the acquired company. If the acquired company or the acquired technology area provided by the acquired company had grown substantially compared with its size at the time of acquisition, it was considered to be a market success. Substantial decreases in the size of the acquired company or technology area were considered market failures. Financial success was determined according to the profits and losses generated. A compound measure of success was developed by summing up the different measures of success for each case. Regression analysis with backward elimination was used to further analyse factors correlating with different component measures of company acquisition success. The results of the normal regression analysis and regression analyses with backward elimination and forward selection are shown in Table 8.2. To improve our understanding of the variables related to success, similar regression procedures were repeated for the variables that were found to correlate with the component measures of success. The results of the regression analyses with the component measures are summarised below. • When technological success is the dependent variable, the main independent variables after the forward selection procedure remain technological complementarity, transfer of competencies and acquisition of science-based assets. The level of explanation is 0.27. • When market success is the dependent variable, the main independent variables after a forward selection procedure are international scope of the acquiring company and distribution complementarity. The level of explanation is 0.29.

160 TOMI LAAMANEN AND ERKKO AUTIO

• When financial success is the dependent variable, the main independent variables after a forward selection procedure are intensity of communication, international scope of the acquiring company, distribution complementarity, previous contacts between the companies and finance complementarity. The coefficient of the finance complementarity variable is again negative. The level of explanation for the model is 0.35. • When strategic success is the dependent variable, the main independent variables after a forward selection procedure are proactive sales motive of the acquired company, intensity of communication between the companies, relatedness of the acquired company and finance complementarity. The coefficient of the finance complementarity is negative. The level of explanation of the model for strategic success is 0.43. Combining the results of the five regression analyses, a model of new, technology-based company acquisition success can be constructed based on both the component measures of success and the compound measure of success, as shown in Figure 8.4. In the compound model, the regression analysis eliminates collinearity between the variables in the component models. The results of the regression analyses seem intuitively appealing. Despite the relatively high level of correlation between the different component measures of success, the regression equations explaining the different component measures differ. It seems that the independent variables contribute to a different extent to the component measures of success. The view of success as a concept consisting of a number of components would seem to provide additional value-added for the analysis. Main contribution of the analysis This chapter was initiated from a hypothesised model of company acquisition activity. The model and a set of hypotheses were looked at both qualitatively and quantitatively based on 111 new, technology-based company acquisition cases. A new model of competence acquisition through company acquisitions was proposed. From an analysis of the 111 new, technology-based company acquisition cases, the following summary can be made. • The hypothesised model of industry evolution does not receive full support from the empirical analysis. Technological competencies are diverse and company specific. The maturity of competencies is context specific. The use of competence maturity or even company maturity as a process dimension is clearly not applicable. New, technologybased companies seem to be acquired no matter how young they are. Contractual and options-based explanations should be used to complement the complementarity and competence-based explanations. • Acquisitions for complementary competence reasons seem to lead to the encapsulation of the acquired company. Altogether 90 of the 111 studied companies preserved their identity, they became encapsulated within the acquiring company. In an encapsulation, the beneficial complementarities between the acquiring and the acquired company can be preserved after the acquisition. Complementarity activation would even seem to

Table 8.2 Regression equations of three regression analyses: 1 with all variables that correlated with the compound measure of success; 2 with same variables as in 1 but with a backward elimination procedure with a significance level of 0.10; 3 with same variables as in 1 but with a forward elimination procedure with a significance level of 0.05

DYNAMIC COMPLEMENTARITIES AND TECHNOLOGY ACQUISITION 161

162 TOMI LAAMANEN AND ERKKO AUTIO

Figure 8.4 Results of the regression analysis with the component and compound measures of success

require some degree of operative integration. This makes the complementarity preservation explanation less valid, while the contractual explanation becomes more valid. Companies acquire to create options and to be able to invest safely in something they own. • Previous contacts, collaboration or equity investment do not seem to be prerequisites for acquisitions. The existence of previous contacts, collaboration or equity investment does not seem to affect the approach selected for integration. Geographic distance, on the other hand, seems to affect new, technology-based company acquisition behaviour. Certain complementarities seem to be more central in some industries than in other industries. Despite some of the industry differences in the importance of the different complementarities, there do not seem to be complementarities that would dominate the evolution of all industry sectors. • Many process-related variables correlate with the different measures of company acquisition success. The process-related variables that were found to have a statistically significant correlation with the compound measure of success include the seller’s proactive motives, previous contacts with the acquired company, communication intensity between the companies after the acquisition, level of operational integration, and the compound measure of integration. The variables representing the approach to integration do not exhibit any significant correlations with the variables representing the previous involvement between the companies. • Many actor-related variables correlate with the different measures of company acquisition success. The acquirer’s nationality was found to correlate with acquisition success. The few acquisitions carried out by foreign companies were clearly more successful than the large number of acquisitions carried out by Finnish companies. Other actor-related variables that were found to correlate with the compound success to a statistically significant extent include the research and development intensity of

DYNAMIC COMPLEMENTARITIES AND TECHNOLOGY ACQUISITION 163

the acquiring company, the international scope of the acquiring company, the degree of dispersion of the acquiring company structure and the degree of business area relatedness. • Many competence-related variables correlate with the different measures of company acquisition success. Statistically significant positive correlations were found between the compound measure of success and the transfer of technological competencies, the degree of systemness of the technological competencies, the degree of patentability, the compound measure of appropriability and the degree of knowledge structuring. Distribution complementarity was found to dominate overall success, while technological complementarity was found to dominate technological success. • Regression analysis with backward elimination or forward selection was used to examine further the variables correlating with success. Regression analysis with backward elimination was used to identify five core variables explaining the variance in the compound measure of success: overall industry trend, seller’s proactive motives, business area relatedness, distribution complementarity and financial complementarity. Technological success could be explained by three core variables: technological complementarity, transfer of technological competencies and acquisition of science-based assets. Market success could be explained by two core variables: distribution complementarity and the international scope of the acquiring company. Financial success could be explained by five core variables: distribution complementarity, the international scope of the acquiring company, previous contacts, communication intensity and finance complementarity. Finally, strategic success could be explained by four core variables: finance complementarity, communication intensity, relatedness of the acquired company, and seller’s proactive sales motive. The empirical analysis also shows that the behaviour of new, technology-based company acquisitions is apparently influenced by varying conditions in different countries, industries and companies as well as by the varying nature of competencies. As expected, the same factors also seem to influence the role and importance of new, technology-based company acquisitions in industry technological evolution. Depending on the industry logic, acquisitions of new, technology-based companies can be used to inject new technological competencies into the administrative framework of the acquiring company. In such a case, the degree of technological complementarity and the integration strategy employed are important. They are needed to preserve the technological competencies and the capability to develop further the acquired competencies. New, technology-based companies can also be used as a source of codified technologies for the distribution system of the acquiring company. New, technology-based companies can be used to get new production facilities, technologies and skills, or new design and engineering skills. New, technology-based companies can be used to get industry-specific knowledge. Finally, new, technology-based company acquisitions can be used as financial investments into technology options. Under certain conditions, the acquisition of new, technology-based companies provides a more attractive alternative to obtain new technological competencies than other modes of competence acquisition. Most large multinational companies continuously monitor

164 TOMI LAAMANEN AND ERKKO AUTIO

their core business areas. Such companies can focus their technological competence acquisition activity on the technologies potentially closest to their core business areas. The research base of an individual country does not affect the competence acquisition practices of multinational companies in as constraining a way as it affects the acquisition practices of domestic companies. Multinational companies with extensive distribution systems have an additional value creation potential through the global replication of the technological competencies of the acquired companies. In small economies such as Finland, there seem to be many opportunities to create value by replicating domestically developed technological competencies internationally. The science and technology system of Finland has been geared to developing new technologies, but the domestic markets are relatively small and there is no real infrastructure for getting these technologies into the international arena. An international distribution resource would seem to be a prerequisite for growth. International distribution complementarity probably explains why acquisitions by multinational companies were more successful than those by domestic companies. A gravitation model describing the company acquisition behaviour of the acquiring companies seems to emerge from this study. The largest companies with an extensive international scope carry out acquisitions of closely related companies even across long geographic distances. The companies acquired across a shorter geographic distance seem to be on average less related to the core business areas of the acquiring companies, resulting in a smaller value creation potential derived from complementarity optimisation. The effect of geographical distance (paradoxically, the shorter the distance, the smaller the complementarity effect) may be due to ‘gravitational pull’ generated by proximity: companies are drawn together from a closer distance more easily than from a longer distance. When a company carries out an acquisition from a long distance, there is usually some additional force—some strong complementarity—bringing the companies together, resulting in a greater success of the acquisition. At the industry level, it is interesting to note the effects of differing characteristics of industries and the possible influences of the industry-level business logic on the behaviour leading to the acquisition of new, technology-based companies. Earlier, we hypothesised that there would be some industry-specific dynamic complementarities that would dominate the development of specific industries (Laamanen and Autio 1996). It was hypothesised that the industry-specific dynamic complementarities affect the propensity for collaboration between large and small firms in an industry. The balances of dynamic complementarities between these firms were assumed to vary in different industries. The different balances of dynamic complementarities were assumed to lead to different size distributions of firms in different industries. It was further assumed that the different dynamic complementarities would affect the nature of collaboration, and the length and probability of success of the collaboration as well as the eventual company acquisition. In this study, the logic concerning the dominance of the dynamic complementarities and the length of collaboration do not receive full support in the cases: companies are often acquired directly without a period of collaboration. The complementarities are not, however, necessarily destroyed. The companies can become encapsulated. The

DYNAMIC COMPLEMENTARITIES AND TECHNOLOGY ACQUISITION 165

encapsulation of the acquired companies can be used to preserve and to activate the dynamic complementarities. It is often safer to acquire the potential collaborative partner if there is a danger that some other company will also approach the new, technology-based company. Furthermore, a collaboration in technology development may require resources that would not be invested in a collaborative relationship but that could be invested in a business that is fully owned. We also found certain complementarities that can be observed more often in some industries than in others. For example, in the metals and machinery industry, product and production complementarities are particularly important. In the field of electronics, technological complementarity is particularly important. On the whole, there are several complementarities in different industries that are important, but none of the complementarities seem really to dominate acquisition activity. Contrasting the complementarities to the different measures of company acquisition success, dominating complementarities can be identified. Distribution complementarity seems to dominate compound success, while technological complementarity seems to dominate technological success. The dominance of both complementarities seems to take place irrespective of the actual industries. These findings do not seem to provide very strong support for the hypothesis on the existence of the industry-specific dominant dynamic complementarities. Technological and distribution complementarities are not independent of each other or of the integration process. The interrelationship can best be illustrated with an example. Among the 111 new, technology-based company acquisition cases studied, typical growth patterns can be identified. In a typical case, if an acquiring company had distribution complementarity and only wanted to buy existing technological competencies, the growth path of the acquired company resembled an inverted U-shaped curve. The sales revenue of the acquired company grew after the acquisition. If there was no further development of the acquired technological competencies, the sales of the business area often went down when the acquired technological competencies matured. A wrong approach to integration sometimes produced the same result. The development of a new product generation failed. A wrong kind of approach to integration could be either too tight or too loose integration. In the case of tight integration, the people of the acquired company sometimes lost interest in the further development of the technological competencies. In the case of loose integration, the people of the acquired companies sometimes developed products that were too sophisticated and became too expensive to sell. The distribution complementarity seems to have facilitated the fastest growth with the initial technological competencies of the acquired company. If there was sufficient technological complementarity or the integration was taken care of well enough, another product generation commonly emerged, continuing the success of the first product generation. Without a distribution complementarity, all the other complementarities seem to have resulted only in an average short-term growth of the acquired companies. Even technological success does not necessarily solve the problem of slow growth. In some of the company acquisition cases, subsequent investment in distribution system building has gradually started to increase the sales revenue of the acquired business area. Some of the typical growth paths are shown in Figure 8.5.

166 TOMI LAAMANEN AND ERKKO AUTIO

Figure 8.5 An illustration of the influence of complementarities on the sales revenue of the acquired company

It seems that to achieve long-term success in new, technology-based company acquisitions both the encapsulation of the acquired company and a significant degree of distribution complementarity are required. On a more general level, the finding seems to reflect the competence integration taking place within companies. The findings of this study seem to indicate that in addition to focusing on technological competencies, it is necessary to focus on the integration of the technological and market competencies and assets. Reverting to the conceptualisations of the innovation processes by Rothwell (1994), it can be seen that our finding is similar to the finding that the success of a new product development process is not solely dependent either on the technological characteristics of the product or on market characteristics. In addition to many other factors, success is dependent on a suitable combination of both. Technological competence transfer is connected to the question of competence integration. The long-term success of technological competence transfer is dependent on the commercial success created with the technological competencies. If there is no commercial success in a reasonable time period, the transferred technological competencies are commonly abandoned. The core issue would seem to be competence integration, not competence transfer. Looking at competence integration in the acquiring company, the encapsulation of the acquired company can be seen as a method for creating a suitable environment, a capsule, for gradual competence integration. This is necessary for preserving dynamic complementarities and for facilitating the combination of the competencies of both companies. The encapsulation can be used to preserve the acquired company’s sometimes

DYNAMIC COMPLEMENTARITIES AND TECHNOLOGY ACQUISITION 167

crucial relationships to the environment and to preserve the culture of the acquired company. Competence integration can take place between different technologies. Technology-market competence integration can also take place. The encapsulation preserves the competencies as the companies jointly develop a new-generation technological competence that integrates the competencies of both companies. (For a similar kind of encapsulation in a different context, see Burgelman and Sayles 1986.) The employees of the acquiring organisation have time to adjust during the encapsulation period when the competencies are being integrated. The not-invented-here atmosphere does not emerge, since both companies participate in the competence integration. The technological competencies, product offerings, service, process or system that are developed jointly can then form a basis for melting down the capsule around the acquired company. Technological competence integration provides a basis for developing the distinctiveness of the competence. The technology-market competence integration provides a basis for leverage. Conclusions There are many points connecting our findings with those of earlier studies. Sanchez et al. (1996a), Chiesa and Barbeschi (1994) and Doz (1994) all present conceptual discussions of competence building and competence leverage as important strategic management tools for managing competencies. Doz and Hamel (1995) discuss motives of competence acquisition and competence leverage in collaborative arrangements. This paper would seem to confirm the applicability of the concepts also to the company acquisition context. Christensen and Foss (1996) discuss the implications of the innovative asset integration inside a company. In this chapter, competence building, innovative asset integration and competence leverage are empirically confirmed to take place in new, technology-based company acquisitions. Mechanisms for competence building and leverage are identified. The discussion of the technological competence integration relates as well to the concept of absorptive capacity (Cohen and Levinthal 1990). In this paper, a simple operationalisation of absorptive capacity for the company acquisition context is presented. The absorptive capacity is operationalised as technological complementarity. The findings concerning competence dependency in connection with competence leveraging corroborate the findings of studies on resource dependency in organisational relationships (Pfeffer and Salancik 1978; Van den Bosch and Elfferich 1991). Our findings concerning the encapsulation of the acquired company relate to the understanding of routines as the organisational memory and the organisational culture (Pautzke 1989; Starbuck 1992; Helleloid and Simonin 1994) that directly affect organisational learning capabilities (Oberschulte 1994). The findings of this chapter concerning the encapsulation would seem to confirm that making abrupt changes in culture can cause adverse results from a competence perspective. In some of the acquisitions, this was evidenced as the inability to develop a new product generation. The preservation of cultural identity would seem to be related to the further technological development capabilities of the acquired companies.

168 TOMI LAAMANEN AND ERKKO AUTIO

The relatedness of the acquired business areas was found to correlate positively with the acquisition success. This supports the hypothesis of the resource-based approach to strategy. A related use of specialised resources provides the highest rent for the resources. Barney and Turk (1994) contradict this in their study on company acquisitions. They point out that in efficient competitive markets, the acquisition of related resources does not provide any additional value creation potential. The value creation potential is already taken into account in the acquisition price. Barney and Turk hypothesise that the main value creation potential comes from the assimilation of unrelated resources into the operation of the company. The point made by Barney and Turk does not, however, invalidate the results of this study. The markets for new, technology-based companies and technological competencies in general cannot be regarded as efficient and competitive. For that reason, even higher-value creation potential can be achieved in connection with the new, technology-based companies by acquiring related technological competencies and combining them into the acquiring company (see also Capron 1996 and Capron et al. 1995). Unrelated acquisitions require more investment into assimilating and developing technological competencies. Additional investments involve an element of additional uncertainty. Further investment is commonly also required to expand the existing distribution system to meet the needs of the new, technological competencies. In some cases, a totally new distribution system may need to be developed. References Anderson, P. and M.L.Tushman (1990) ‘Technological Discontinuities and Dominant Designs: A Cyclical Model of Technological Change’, Administrative Science Quarterly 35:604–33. Autio, E. (1993) Technology Transfer Effects of New, Technology-Based Companies: An Empirical Study, Helsinki University of Technology, Research Report 1993–1. ——(1995) Symplectic and Generative Impacts of New, Technology-Based Companies, Helsinki University of Technology, Research Report 1995–1. Barney, J.B. (1991) ‘Firm Resources and Sustained Competitive Advantage’, Journal of Management 17:99–120. Barney, J.B. and T.A.Turk (1994) ‘Superior Performance from Implementing Merger and Acquisition Strategies: A Resource-Based Analysis’, in G.von Krogh, A.Sinatra and H.Singh (eds), The Management of Corporate Acquisitions: International Perspectives, London: Macmillan. Burgelman, R.A. and L.R.Sayles (1986) Inside Corporate Innovation, New York: The Free Press. Capron, L. (1996) ‘Mechanisms of Value Creation through Mergers and Acquisitions: A Revision of the Traditional Explanations’, Conference paper, The 16th Annual International Conference of the Strategic Management Society, Phoenix. Capron, L., P.Dussauge and W.Mitchell (1995) Appropriation of Resources within Horizontal Mergers and Acquisitions: An International Empirical Study, HEC Working Papers, CR 542/1995. Chiesa, V. and M.Barbeschi (1994) ‘Technology Strategy in Competence-Based Competition’, in G.Hamel and A.Heene (eds), Competence-Based Competition, New York: John Wiley and Sons. Christensen, J.F. and N.J.Foss (1996) Dynamic Corporate Coherence and Competence-Based Competition, Center for Inter-disciplinary Studies in Technology Management CISTEMA, Management CISTEMA, Copenhagen Business School, Technical University of Denmark: no. 1.

DYNAMIC COMPLEMENTARITIES AND TECHNOLOGY ACQUISITION 169

Cohen, W.M. and D.A.Levinthal (1990) ‘Absorptive Capacity: A New Perspective on Learning and Innovation’, Administrative Science Quarterly 35:128–52. Dierickx, I. and K.Cool (1989) ‘Asset Stock Accumulation and Sustainability of Competitive Advantage’, Management Science 35(12):1504–11. Doz, Y. (1994) Managing Core Competency for Corporate Renewal: Towards a Managerial Theory of Core Competencies, Working Paper no. 23, Paris, INSEAD. ——(1996) ‘The Evolution of Cooperation in Strategic Alliances: Initial Conditions or Learning Processes’, Strategic Management Journal 17:55–83, in A.Barnett and R.A.Burgelman (eds), A Evolutionary Perspectives on Strategy, Special Issue. Doz, Y. and G.Hamel (1995) The Use of Alliances in Implementing Technology Strategies, Working Paper no. 22, Paris, INSEAD. Ghoshal, S., M.Hahn and P.Moran (1997) An Integrative Theory of Firm Growth: Implications for Corporate Organization and Management, Working Paper, London Business School. Gulati, R. (1995) ‘Social Structure and Alliance Formation Patterns: A Longitudinal Analysis’, Administrative Science Quarterly 40:619–52. Harrigan, K.R. (1986) Managing for Joint Venture Success, Lexington, MA: Lexington Books. Haspeslagh, P.C. and D.B.Jemison (1992) ‘Managing Acquisitions: Creating Through Corporate Renewal’, Academy of Management Review 18:370–4. ——(1994) ‘Acquisition Integration: Creating the Atmosphere for Value Creation’, in G.von Krogh, A.Sinatra and H.Singh (eds), The Management of Corporate Acquisitions: International Perspectives, London: Macmillan. Heene, A. and R.Sanchez (eds) (1997) Competence-Based Strategic Management, Chichester: John Wiley and Sons. Helleloid, D. and B.Simonin (1994) ‘Organizational Learning and a Firm’s Core Competence’, in G.Hamel and A.Heene (eds), Competence-Based Competition, Chichester: John Wiley and Sons. Henderson, R. and I.Cockburn (1994) ‘Measuring Competence? Exploring Firm Effects in Pharmaceutical Research’, Strategic Management Journal 15:63–84. Laamanen, T. (1997) The Acquisition of Technological Competencies through the Acquisition of New, Technology-Based Companies, Helsinki University of Technology. Laamanen, T. and E.Autio (1996) ‘Dominant Dynamic Complementarities and TechnologyMotivated Acquisitions of New, Technology-Based Firms’, International Journal of Technology Management 12(7/8):769–86. Lengnick-Hall, C.A. (1991) ‘A Conceptual Framework for Evaluating Designs for Corporate Innovation’, Journal of Engineering and Technology Management March: 197–227. Oberschulte, H. (1994) Organisatorische Intelligenz: Ein integrativer Ansatz des organisatorischen Lernens, Mering: Rainer Hampp Verlag. Pautzke, G. (1989) Die Evolution der organisatorischen Wissensbasis: Bausteine zu einer Theorie des organisatorischen Lernens, Munich:Verlag Barbara Kirsch. Penrose, E. (1959) The Theory of the Growth of the Firm, Oxford: Basil Blackwell. Pfeffer, J. and G.R.Salancik (1978) The External Control of Organizations, New York: Harper and Row. Quélin, B. (1997) ‘Appropriability and the Creation of New Capabilities through Strategic Alliances’, in R.Sanchez and A.Heene (eds), Strategic Learning and Knowledge Management, Chichester: John Wiley and Sons. Ramanathan, J. (1996) ‘Technological Capability Assessment of a Firm in the Electrical Sector’, Technocation 16:561–88. Roberts, E.B. and C.A.Berry (1985) ‘Entering New Businesses: Selecting Strategies for Success’, Sloan Management Review 26:57–71.

170 TOMI LAAMANEN AND ERKKO AUTIO

Rothwell, R. (1983) ‘Innovation and Firm Size: The Case of Dynamic Complementarity’, Journal of General Management 8(6):5–25. ——(1994) ‘Towards a Fifth-Generation Innovation Process’, International Marketing Review 11(1): 7–31. Sanchez, R. (1997) ‘Managing Articulated Knowledge in Competence-based Competition’, in R.Sanchez and A.Heene (eds), Strategic Learning and Knowledge Management, Chichester: John Wiley and Sons. Sanchez, R. and A.Heene (eds) (1997) Strategic Learning and Knowledge Management, Chichester: John Wiley and Sons. Sanchez, R., A.Heene and H.Thomas (1996a) ‘Introduction: Towards a Theory of CompetenceBased Competition’, in R.Sanchez, A.Heene and H.Thomas (eds), Dynamics of CompetenceBased Competition: Theory and Practice in the New Strategic Management, Oxford: Pergamon. Sanchez, R., A.Heene and H.Thomas (eds) (1996b) Dynamics of Competence-Based Competition: Theory and Practice in the New Strategic Management, Oxford: Pergamon. Starbuck, W.H. (1992) ‘Learning by Knowledge-Intensive Firms’, Journal of Management Studies 29: 713–40. Teece, D.J. (1986) ‘Profiting from Technological Innovation: Implications for Integration, Collaboration, Licensing, and Public Policy’, Research Policy 15:285–305. Teece, D.J., G.Pisano and A.Shuen (1992) Dynamic Capabilities and Strategic Management, Working Paper, University of California, Berkeley. Teece, D.J., R.P.Rumelt, G.Dosi and S.G.Winter (1994) ‘Understanding Corporate Coherence: Theory and Evidence’, Journal of Economic Behavior and Organization 23:1–30. Tushman, M.L. and P.Anderson (1986) ‘Technological Discontinuities and Organizational Environments’, Administrative Science Quarterly 31:439–65. Tushman, M.L. and E.Romanelli (1985) ‘Organizational Evolution: A Metamorphosis Model of Convergence and Reorientation’, Research in Organizational Behavior 7:171–222. Utterback, J. (1994) Mastering the Dynamics of Innovation: How Companies Can Seize Opportunities in the Face of Technological Change, Boston, MA: Harvard Business School Press. Utterback, J. and W.Abernathy (1975) ‘A Dynamic Model of Process and Product Innovation’, Omega 33:639–56. Van den Bosch, F.A.J. and P.C.Elfferich (1991) Make, Buy, or Cooperate Decisions with Respect to the Acquisition and Exploitation of Technical Knowledge and Organizational Issues, Management Report Series, no. 94, Erasmus Universiteit, Rotterdam School of Management.

9 Economic organisation and the accumulation of rent-yielding assets Kirsten Foss

Introduction The title of this chapter may appeal both to economists of organisation and to resourcebased scholars. However, they are likely to entertain divergent expectations about its contents. Thus, for those mainly interested in economic organisation, expectations probably will be directed towards a question such as, ‘How do we secure incentives to invest in accumulating assets with a given and well-known rent-earning capacity?’ The interesting aspects of this question would depend on the assumed information cost constraints on contracting for such an investment. On the other hand, for those economists and management scholars mainly interested in resource-based perspectives on firms, the title probably would be interpreted as a label covering discussions of issues such as: ‘How may rent-earning assets be accumulated as outcomes of the activities taking place within the firms?’, ‘How may differences in historic circumstances and accumulation capacity between firms explain the rents generated from those assets?’ and ‘How may such assets best be explored?’ I may disappoint both parties. The title conceals an unorthodox idea that to my knowledge has not been present in the economics of organisation, or in the resourcebased approach. The idea is that organising transactions within the boundaries of the firm may reduce the cost of conducting experiments relative to a situation in which such experimentation would have to take place across markets. With ‘experiments’ I have something much broader in mind than the activities performed by the R&D function. Those I have in mind are the many trials and errors involved in setting up a smoothly running production system consisting of many interdependent specialised tasks. Of course, such experimentation is only needed if there is uncertainty with respect to the best way of operating technically interdependent production systems. Owing to such ‘technological uncertainty’, as we may call it, firms start different kinds of experiments and follow different paths of learning. This, in turn, may explain long-lasting differences between firms, thus establishing a link to the analysis of competitive advantage. Moreover, from the resource-based perspective (Wernerfelt 1984; Dierickx and Cool 1989; Barney 1991; Peteraf 1993), we know that the assets that are likely to be sources of sustainable rents and above-normal profits are the outcomes of ambiguous and

172 KIRSTEN FOSS

uncertain assets accumulation processes. The results of path-dependent learning processes would often seem to qualify here. Uncertainty in any deep (e.g., Knightian or Shacklian) sense is not an assumption that characterises the majority of modern contractual theories of firms (e.g., Williamson 1985; Hart 1995). Indeed, I shall argue that most of the contractual theories of firms are counterparts to what Stephen Littlechild (1986) has characterised as the ‘neoclassical’ and the ‘Austrian’ market models—models that encompass risk, information costs and ignorance, but not uncertainty. That is to say, it is essentially taken for granted that the best uses of productive assets are already known. Information costs may make contracting over all the relevant characteristics of assets impossible or may constrain contingent contracting and cause ex post contractual hazards. Economic organisation then mainly is a question of providing the transacting parties with proper incentives so that joint surplus can be maximised. Such relatively weak constraints on rationality allow the researcher to pinpoint exactly what kind of organisation of transactions is optimal, given the information costs constraints introduced in the model. The kinds of contractual problems which can be solved by firms leave room for only very limited differences between firms. Firms, of course, may solve incentive problems related to investments in rent-earning assets, but they are not vehicles for generating such assets (Langlois and Foss 1999). The dynamic transaction cost theory developed by Langlois and Robertson (1995), as well as Coase’s (1937) original work (as I interpret this), are exceptions to the rule. I interpret these explanations of firms as being based more on a counterpart to what Littlechild (1986) calls a radical subjectivist market model—a model which allows for much greater degrees of uncertainty and thus larger differences between firms. Relying on radical subjectivist arguments gives a richer picture of what firms are and what they do. It opens the door more fully for issues of entrepreneurship, experimentation, learning and coordination in an uncertain world. This brings me back to the title of this chapter. The content of the chapter is really an argument for the importance of uncertainty in explaining the existence of firms. The organisation of transactions within the boundaries of firms thus depends on the transaction costs involved in reducing uncertainty by accumulating knowledge from conducting systematic experiments. This conception of firms also provides a bridge between contractual and resource-based perspectives on firms because, as indicated, the knowledge accumulated from experiments may be a source of long-lasting rents. In the second section, I discuss contractual theories of the firm, and I argue that they all, except for dynamic transaction cost theory and Coase’s original story, are based on market models encompassing at most ignorance. The third section provides a reinterpretation of the Coasian notion of firms and it is argued that the Coasian firm reduces the transaction costs associated with conducting experiments needed to discover the true opportunity costs from different uses of productive assets. In the fourth section, I introduce the resource-based view of firms and explain why accumulation of knowledge on how to organise production may be a rent-earning asset.

ECONOMIC ORGANISATION AND THE ACCUMULATION OF RENT-YIELDING ASSETS 173

Contractual theories of firms: a perspective from the theory of market processes In a prominent modern manifestation, contractual theory takes the transaction as the unit of analysis and proceeds to ask why more complex forms of contracts (Williamson 1985) replace anonymous market exchange in some situations. One of the fundamental assumptions of contractual theories is that there is an efficiency rationale behind transacting parties’ choice of economic organisation. A choice between market transactions and transactions organised within firms therefore depends on the efficiency gains from moving transactions within the realm of firms (or vice versa). Contractual theories embrace a relatively broad spectrum of theories about different sources of the costs of transacting and this is reflected in the many different explanations of firms as solutions to contractual problems. Within this broad strand of literature one can identify at least two distinct traditions (Foss 1994). The first is mainly static, taking most inputs and outputs as given. The second one is more dynamic; for example, slightly more provision is made for uncertainty. Representative of the first, static tradition is the nexus of contract and measurement cost approaches (Alchian and Demsetz 1972; Jensen and Meckling 1976; Fama 1980; Barzel 1982, 1985, 1997; Cheung 1983). Formal principalagent theory may also be seen as belonging to this camp. Representative of the slightly more dynamic tradition is the literature on governance mechanisms (Williamson 1975, 1985) and incomplete contracts (Grossman and Hart 1986; Hart 1991; Hart and Moore 1990). In addition to the streams mentioned by Foss (1994), there is the dynamic transaction cost theory developed by Langlois and Robertson (1995). In the perspective taken here, I see a correspondence between the different views of the market process and the different views of the contractual process represented in (1) the static contractual theories, (2) the more dynamic contractual theories, and (3) the dynamic transaction cost approach. For example, is there a dynamic contracting process, so that renegotiations are allowed for (as in Williamson’s work), or is everything basically settled from the beginning (as in agency theory)? The framework suggested by Stephen Littlechild (1986) for distinguishing between different economic theories of the market process helps clarify the differences between the various contractual theories. More important, the framework helps to pinpoint the possible ways in which the different contractual theories may explain long-lasting differences in firm profits. The framework consists of three ‘ideal type’ models of market processes, distinguished by ‘how the decision makers perceive of the world, how these perceptions change over time, how additional information may be sought, and how the decision maker can limit his exposure to uncertainty’ (Littlechild 1986:27). Using these criteria, Littlechild identifies the following three ideal models: 1 the neoclassical model (Frydman 1982), 2 the Austrian model (Kirzner 1973), 3 the radical subjectivist model (Shackle 1972).

174 KIRSTEN FOSS

The static contractual theories resemble the neoclassical market process model in assuming ‘that the form the future can take is known in advance’ (Littlechild 1986:28; emphasis in original). The static market model corresponds to the measurement cost (Cheung 1983; Barzel 1997), agency (Jensen and Meckling 1976) and team production (Alchian and Demsetz 1972) theories of firms. In the static transaction cost theories, the future holds no surprises, just as in the neoclassical market model. This means that contracting is comprehensive and that all uses of assets are well known to economic agents. Not all relevant information is available to the economic agents, but by searching they may obtain more. Information may, for example, be lacking on valued attributes of assets (Barzel 1997) or, in the case of humans, the effort they exert (Alchian and Demsetz 1972; Jensen and Meckling 1976). Prices therefore may not perfectly reflect value, creating problems of excessive sorting, inefficient levels of performance and underinvestment in durable production assets. Some of these problems may be creating the organisation called the firm. For example, in order to realise economies of scale many individuals may be needed to work on the same piece of equipment. However, high information costs make it difficult to determine in advance how much the operation of each individual contributes to the wear and tear of the equipment. This creates a situation where there will be insufficient incentives to invest in maintenance. According to Barzel (1997), one way of controlling such problems is to use a fixed wage contract in which workers are remunerated for their time rather than their output. But since a worker who receives a fixed wage for a fixed period of time has no incentives to identify the tasks needed for an effective operation of equipment, managers have to specify and monitor the task to be performed. Voilà! The firm arises and solves a market failure problem. A similar problem arises when team production is involved (Alchian and Demsetz 1972). In this case, costs of measuring effort may create externalities, as each team member has an incentive to shirk, thereby reducing his cost of contributing, while only shouldering part of the costs in terms of lower output. Again, the solution is to set up an organisation in which a specialised monitor is appointed with the rights to meter effort, to receive the residual income from these activities, to alter membership of the team and to sell all these rights. In both of these cases, a firm (a specialised monitor) is an efficient solution relative to markets. There are two ways in which differences in firm profit ability can emerge. First, under market contracting, it is argued, the monitor will not be able to know so much about the inputs (at the same costs) as in firm-like organisation.1 This must (plausibly) imply that the monitor may more easily learn the relevant characteristics of a careful or hard-working individual when he is associated with the team for a longer period rather than for a short period. Differences in monitors’ abilities to learn how to monitor most effectively may explain why some monitors earn an above-normal profit (a Ricardian rent) on their effort. However, if the ability to perform low-cost monitoring cannot be passed on to a successor, there can be no long-lasting superior profits to the firm (in the sense of lasting much longer than a ‘generation’). Second, a situation with less than fully mobile factors, or private information about the cost or benefits of realising different

ECONOMIC ORGANISATION AND THE ACCUMULATION OF RENT-YIELDING ASSETS 175

plans, or even about the likelihood of possible outcomes, may imply that ‘a differential advantage in expanding output develops in some firms’ (Demsetz 1973:1). In fact, above normal returns are possible even if all possess the same information about the likelihood of different outcomes so that the expected value of assets is perfectly reflected in their prices. Different managers will ‘gamble’ for different outcomes. For some managers the outcome turns out to be the most preferred one and they have acquired the assets for less than the realised value. In such cases, explains Demsetz, ‘[b]y the time their value to the firm is recognised, they are beyond acquisition of other firms at the same historic costs, and, in the interim, shareholders of the successful or lucky firm will have enjoyed higher profit rates’ (p. 74). What Foss (1994) has called a more dynamic transaction cost approach seems closer to what Littlechild has termed ‘the Austrian market process model’. Here the problem is neither uncertainty nor risk, but ignorance: ‘“Tomorrow” is a vector of which the agent knows some components but not others; he or she knows there will be other components, but not what they will be. Consequently, the agent cannot form a probability judgement of the likelihood of their occurring’ (p. 29). Thus, in such a world, tomorrow may bring about ‘discoveries’ of improved materials or techniques, and contracting will be incomplete. Such discoveries may influence the value of rights over assets in ways which were not foreseen at the time of contracting. The allocation of rights to determine the use of the assets thus becomes crucially important. As Littlechild points out: ‘It is now relevant to consider which party is best able to “predict” and respond to such unexpected change—or, perhaps, which party is most optimistic or apprehensive about the possibility of such a development’ (p. 35). Some agents may have natural ownership advantages, in the sense that they are better able to discover better ways of using assets. But rights to use assets may also be valuable to the original owner, because it is the best way of securing oneself against hold-up where transactions involve specific assets. It is this latter problem that is at the centre of the more dynamic transaction costs theories. Thus, Williamson (1975, 1985), Klein, Crawford and Alchian (1978) and Hart (1995) all put great emphasis on how uncertainty makes it difficult to enforce one’s rights over assets. According to this literature, unforeseen events may alter the relative value of assets, or specific properties of valuable assets have to be discovered and therefore cannot be included in contracts. In the context of specific assets, this may result in ex post transaction costs as contracts have to be renegotiated, and/or in the lack of proper incentives to undertake efficient transaction specific investments. Again firms arise to solve market failures. For Williamson (1975, 1985), the choice between market or firm depends on the extent to which transactions are characterised by asset specificity, frequency and uncertainty, given opportunism and bounded rationality (and, hence, incomplete contracting). The assumption of ‘uncertainty’ is, according to Williamson, no hindrance to researchers in forming a predictive theory of economic organisation, since all that is needed is ‘that the factors responsible for differences among transactions be identified and explicated’ (1985:52). In other words, one must assume that if the critical characteristics

176 KIRSTEN FOSS

are well known to and easily identifiable by economic agents, they will in fact tend to choose the optimal governance structures.2 The only problem is how agents know the kind of uncertainty they are facing. Certainly, it cannot be fundamental uncertainty, since in that case decision-makers would have no expectations of possible disturbances in the conditions surrounding the transaction. It must be that the individual, as described by Littlechild, at least ‘knows some components [of tomorrow’s vector of events] but not others’ or that ‘he or she knows there will be other components, but not what they will be’. This could be a situation in which decision-makers vaguely anticipate the possibility of major changes in prices, which could require renegotiations of contract terms. Owing to the ‘uncertainty’ surrounding such negotiations, agents may expect strategic behaviour (misrepresentation of information, withholding of information, etc.) as parties try to capture rents from specific assets. In this case, transactions are best organised within the boundaries of firms. In Williamson’s scheme, firms are ‘ultimate courts of appeal’, where top management possesses rights to decide the uses of assets, and where opportunism may result in termination of contracts leaving the contracting party without access to complementary physical assets.3 Now, within the hierarchy, managers may reduce incentives to take advantage of information asymmetries by switching from the high-powered incentives of payment for output (characterising market transactions) to the low-powered incentives of fixed payment for time. This, however, creates a need for more monitoring of effort and other incentive schemes relative to contracts based on high-powered incentives. In the Williamsonian scheme, firms therefore could differ with respect to the internal structure and incentive schemes and thereby differ in profits, owing to more effective means of economising on control and measurement costs. However, if, given the assumed ‘uncertainty’, it is possible to predict the optimal governance structures for different kinds of transactions, why should it not also be possible to predict the optimal internal organisation of firms? This would leave only slight differences between firms facing the same contingencies. In other words, it is hard to rationalise firm differences within this body of thought. In what Hart (1995) calls the property right theory of firms, the joint surplus is assumed to be known with accuracy, and ignorance is only present as an assumed lack of ability to specify the exact nature of the object over which one contracts. Since contracts cannot be enforced, an agent will have greater incentives to undertake a transactionspecific investment in his human capital if he also possesses residual use rights over complementary physical assets. This is because he then avoids being held up by an owner of complementary physical assets for a share of the residual income his investment can produce.4 The physical assets over which a legitimate owner has formal residual use rights define firms. In the Hart (1995) model, firms could differ with respect to the specific and quasi-rent-earning assets they hold, and differences in luck or private information may explain differences in the profits they can generate from these assets. Firms reduce uncertainty in only one dimension: they provide a guarantee for a certain share of the residual income of investments in sunk cost human capital. Firms do not enable the

ECONOMIC ORGANISATION AND THE ACCUMULATION OF RENT-YIELDING ASSETS 177

discovery of the nature of the object over which one contracts; such discoveries take place independently of whether or not firms exist. The last market process model is the radical subjectivist model. It differs from the Austrian one by introducing new alternatives between which agents may choose; it stresses ‘the imagination needed to create the alternatives between which decisions are made, and hence the inevitable uncertainty associated with the outcome of decisions’ (Hart 1995: 29). This notion of the market process, allowing for genuine entrepreneurship, creativity and experimenting, is akin in some ways to the dynamic transaction costs theory presented in Langlois and Robertson (1995). In their perspective, firms are seen primarily as responses to problems of coordinating interdependent tasks, typically arising from systemic innovations. Faced with fundamental uncertainty and asymmetric information, the firm ‘can more cheaply redirect, coordinate, and where necessary create the capabilities necessary to make the innovation work’ (Langlois and Robertson 1995:3). It is the capacity to coordinate activities rather than ownership that is at the centre in this conception of firms. Individual firms are characterised by their intrinsic core of resources which are ‘idiosyncratically synergistic, inimitable, and non-contestable’ and therefore not transferable across markets. Firms may also control a number of ancillary capabilities, which they have either created in-house or acquired over markets. In the short run, the boundaries of the firm depend on the relative production costs of a firm relative to other firms and the transaction costs involved in securing such ancillary capabilities over markets. Langlois and Robertson (1995) argue that if firms were operating in an environment similar to that envisioned in the Austrian market model, transaction costs would gradually be reduced as individuals learn how to detect shirking and moral hazard and because they learn about the contingencies relevant to contracts. Therefore, in the long run, one should expect a more narrow scope of the firm. On the other hand, one should also expect a gradual improvement of a firm’s capabilities as more activities become routinised and this would result in an expansion of its boundary. However, other firms also improve their capabilities and these are easily accessed through market contracts. This reduces the advantages from in-house production. Now, owing to the routinisation of activities, firms may be faced with a situation in which they have excess capacity relative to the full use of their capabilities (see also Penrose 1959). If costs of transacting in capabilities were low, firms would sell some of this capacity. However, as Langlois and Robertson point out, ‘[o]ne of the principal determinants of the appropriate form of a business institution is the nature of economic change that institution must confront’ (1995:3)—a situation which only arises in an environment similar to that envisaged in the radical subjectivist market model. Economic change arises as economic agents discover new profit opportunities. However, market contracting may not be an efficient means for taking advantage of a new profit opportunity and especially not if the entrepreneur who discovers the opportunity has to contract for adaptations of complementary capabilities to his needs. The entrepreneur may then find it too costly to inform owners of these capabilities of the idea and to persuade them to invest in such an endeavour, one for which not much is known about the vector of possible outcomes.

178 KIRSTEN FOSS

Dynamic transactions, which consist of ‘costs of persuading, negotiating, coordinating, and teaching outside suppliers’ create a need for control over assets. In the dynamic transaction costs theory, there is clearly room for prolonged differences between firms. Firms primarily differ with respect to their intrinsic core of capabilities (routines, culture and other answers to coordination game situations). The intrinsic core changes as they learn more about the nature of tasks they undertake and as they discover new profit opportunities. In relatively stable environments, there may be some overlap in what firms are good at, but never a complete overlap. Within the theory, persistent differences in profits between firms could be seen as being based on slight differences in the context in which solutions to coordination games have evolved. Moreover, firms may differ in their ability to cope with uncertainty; that is, they may be in control of the right capabilities at the right time so that they can quickly take advantage of emerging profit opportunities. So far I have not discussed the seminal contribution by Coase (1937) to the theory of the firm. In the following section I present a reinterpretation of this contribution, in which I argue that firms exist because they reduce the costs of discovering the unknown or not well-understood components of the vector of events which make up ‘tomorrow’. In particular, they discover the best way of coordinating interdependent tasks. Moreover, I argue that the vector of events to be discovered gradually changes as firms see opportunities for economic gains from further specialisation in production, thus creating new coordination problems. This interpretation may in many ways resemble that presented in the dynamic transaction costs perspective. There are, however, a number of important differences that I would like to point out before I turn to the reinterpretation of ‘The Nature of the Firm’. First, in line with Coase (1937), I conceive of firms and also markets as alternative means of coordinating activities and thus the uses of assets. This view of markets seems to me to be neglected in the static and more dynamic versions of contractual theories. Within these perspectives, markets are primarily treated as vectors of prices, and competition is conceptualised as a relatively cheap way of disciplining the users of resources, given that transaction costs (or rather measurement costs) are not too high. The dynamic transaction costs theory seems to have adopted a similar view of the market. Here costs of transacting are all discussed in terms of market failure due to high measurement costs or costs of ensuring against breach of contracts. The role of the market and the relevance of price signals in directing uses of resources seem to be taken for granted except in two cases. The most obvious is in the case of inventions where no prices can indicate the potential benefits of the endeavour and where judgement of the value of investments has to be based entirely on the information provided by the innovator. However, once the invention has proved its worth, these dynamic transaction costs will vanish—except for the costs of transferring the capabilities now developed within the innovating firm. The less obvious case is where it is assumed that prices cannot provide solutions to coordination games and that capabilities therefore substitute for price coordination. However, not all solutions to coordination games evolve within boundaries of firms as defined by the unified ownership of assets; so we still need a theory which will

ECONOMIC ORGANISATION AND THE ACCUMULATION OF RENT-YIELDING ASSETS 179

explain when solutions to coordination problems are most efficiently reached within firms. The Coasian notion of markets and firms differs from the dynamic transaction costs perspective in one important aspect: it is assumed that prices are not given, but have to be discovered. Prices as information reflecting the opportunity costs of different uses of assets can be discovered either in a market process or by managers within the boundaries of firms. Firms and markets are thus alternative modes of coordination, and the substitution of one for the other depends on the costs of using the price system relative to managed coordination. Coase points out when coordination by management is likely to be more efficient than pure market coordination. Second, following Coase (1937), I will argue that sometimes the discovery of the relevant (shadow) prices requires experimenting with different ways of organising activities. Ownership of assets and labour contracts can reduce costs of performing such experiments. The outcomes of such experiments result in the accumulation of information on important interdependencies in productive activities and components as well as on different ways of sequencing activities. This knowledge is an important part of firms’ capabilities, along with norms, codes of conduct and routines, which also guide behaviour and solve coordination problems. However, there are important differences between experimentally accumulated knowledge and norms, codes and some routines, in that a stock of experimental knowledge is more easily produced within a firm as defined by unified ownership over assets, while the evolution of norms, codes of conduct, etc. may depend more on the duration of the relationship, and may thus develop between firms as well as inside firms. A reinterpretation of the ‘nature of the firm’ In this section, I interpret Coase’s analysis of the nature of the firm in terms of property rights theory as developed by Coase (1960), Alchian (1965), Cheung (1983) and Barzel (1997). I argue that Coase’s notion of firms can be viewed as a solution to problems of coordination in situations where use rights over assets cannot be perfectly specified and allocated in manners which ensure the functionality of technologies. Such situations may occur because individuals have only limited computational capacity (bounded rationality), making it too difficult for them to specify user rights in ways that completely solve problems of interdependencies, or because they face uncertainty in the sense that they lack ability to imagine ‘the alternatives between which decisions are made’ (Littlechild 1986:29). This kind of uncertainty (which characterises the radical subjectivist market model) arises when inventions may change the set of alternatives between which economic agents can choose and thus also the structure of prices. However, I argue that increased specialisation can also be a source of such uncertainty. This is because technological interdependencies between subdivided tasks typically only become apparent after the act of specialising. Specialisation then may result in problems such as bottlenecks and uneven development of components for which new solutions have to be imagined. Following Coase (1937, 1991), I argue that one of the reasons why

180 KIRSTEN FOSS

managed coordination and ownership of assets may be advantageous relative to price coordination is because the former reduces costs of learning in the coordination of technological interdependent tasks.5 Coase on the nature of the firm Coase (1937) uses coordination costs to explain why, in an economy of specialised production, markets and firms coexist as alternative modes of coordination. The reason for the existence of firms, he explains, is that there are costs of using the price mechanism, and that ‘[t]he most obvious cost of “organizing” production through the price mechanism is that of discovering what the relevant prices are’ (p. 21; emphasis added). With high market costs (later termed transaction costs), the market mode of coordination is substituted by a firm mode of coordination based on managerial decisions. But as pointed out by Coase, the advantages of the firm mode over the market mode diminish as marginal costs of coordination increase with more tasks being coordinated within the boundaries of a firm, and this leaves room for competition between firms for the coordination of specialised tasks. While Coase (1937) does not explicitly suggest that uncertainty is a reason why there may be costs of discovering the relevant prices, uncertainty seems to play an important role in explaining the need of open-ended contracts, such as employment contracts. According to Coase an employment contract is preferred if, ‘owing to the difficulty of forecasting, the longer the period of the contract is for the supply of the commodity or service, the less possible, and indeed, the less desirable it is for the person purchasing to specify what the other contracting party is expected to do’ (p. 21). Stated in the terminology of property rights theory (Barzel 1997), there are high costs of specifying the valued attributes of assets in all future states, and this results in rights over valued attributes of assets being left unspecified. When the coordination between inputs is influenced by changes in external contingencies which cannot be specified ex ante, continuous redirection of resources and replanning taking advantage of the dimensions of time and place of assets will be necessary in order to avoid bottlenecks and other incidences of misallocation. High costs of discovering the relevant prices constitute a necessary but not a sufficient factor in explaining why firms emerge, since ‘this cost may be reduced but it will not be eliminated by the emergence of specialists who will sell this information’ (Coase 1937: 21). Firms exist only if there are also ‘costs of negotiating and concluding a separate contract for each exchange transaction which takes place on a market’ (p. 21). An arbitrageur holding stocks of assets makes his money from superior knowledge about the value of the unspecified attributes of time and place (Hayek 1945; Barzel 1997). The employment contract may be interpreted as providing a stock of labour services, which, within limits (Simon 1951), can be allocated to different uses by the direction of an arbitrageur in response to unforeseen contingencies. Now, arbitrageurs need to bear the cost of stocks only if they cannot appropriate the benefits of their knowledge of time and place by selling this information. Two factors may

ECONOMIC ORGANISATION AND THE ACCUMULATION OF RENT-YIELDING ASSETS 181

explain why it is not always feasible to sell information about time and place dimensions of assets. First, there is the well-known problem of information as a public good which, if revealed before the transaction, cannot be protected from capture (Arrow 1962). Second, negotiations may take a longer time than direction by orders, and because of this, opportunities for profitable action may be forgone (Littlechild 1986). In ‘The Nature of the Firm’, Coase uses the employer-employee relationship as the archetype of the firm where managers’ rights to direct resources within certain limits fill in the holes in the open-ended employment contracts. More recently Coase (1991) has remarked that already at the time when he wrote ‘The Nature of the Firm’ he was aware that the analogy between the employment contract and the firm could give an incomplete picture of the nature of the firm. He points to an amendment to the original article, when he draws attention to a lecture note from 1934 in which he states that ‘a full firm relationship will not come about unless several such contracts are made with people and for things which cooperate with one another’ (Coase 1991:64). This amendment can be interpreted to mean that managerial decisions fill the holes of open-ended contracts in cases where coordination of large numbers of factors, which cooperate with each other, is required. This is exactly the situation with technological interdependencies between many tasks. To sum up, firms only exist both if there are high costs of discovering the relevant prices and if these costs cannot be reduced by contracting for this information. The latter case may be particularly relevant where interdependencies between many resource owners make it costly to rearrange tasks to take advantage of new information on states of the world. Firms then save transaction costs by substituting many independently determined contingent contracts for managed directions. Coase (1937), like virtually all other later theorisers on the firm, takes the costs of coordinating various tasks as well as the extent of specialisation in the economy as given, and proceeds to analyse why not all transactions among specialised agents are coordinated either in firms or in open markets. However, costs of coordination between tasks may crucially depend on the degree of specialisation. Self-management of more tasks may be an alternative to specialisation, which reduces overall costs of production in cases where coordination between many specialised tasks proves very costly. The degree of specialisation therefore depends on the marginal costs of coordinating increasingly specialised tasks and the marginal benefits from specialisation. Specialisation in production, technological uncertainty and the role of firms As Adam Smith pointed out in The Wealth of Nations, specialisation in production is one of the main sources of productivity improvements. Specifically, he ascribed productivity gains to improvements in a worker’s ability to perform a task as it is repeated more often, the time that is saved from avoiding having to switch from one task to another, and an improved ability of workers to identify labour-saving innovations. However, there are also costs of specialisation.

182 KIRSTEN FOSS

Specialisation may, for example, result in bottlenecks where complexity and interdependent activities make it difficult to specify how best to sequence various activities, where the introduction of more specialised tools and equipment creates capacity utilisation problems because of technical indivisibilities, or where agents’ innovativeness results in an uneven development of tools, equipment and components. Basically these problems arise when those who deliver parts or carry out activities are not aware of the need for mutual adjustment. From a property rights perspective, these problems can be ascribed to imperfectly specified rights over assets as production tasks are subdivided. This is because it is difficult to specify all valued dimensions of assets prior to specialisation, since many of the valued dimensions of assets only become apparent from experimenting with the use of assets. Even if important dimensions can be specified, it may still be difficult to allocate these rights in ways that ensure the best use of assets. This may, for example, be the case with the time and place dimension of assets where non-optimal allocations result in excess stocks of intermediate products or in idle assets. In fact, with a great deal of interdependence in a complex system, the best time and place to use an asset depend on the specification of the uses of all other assets that are needed in the production. Solving problems that arise from technological interdependencies is an important source of innovative improvements6 (as pointed out by Rosenberg 1976 and Sahal 1981). However, such innovations emerge not because of increased specialisation, but because of learning in coordination. The question then arises: what institutional set-up best provides for experimentation and accumulation of experience in coordination? I argue that the Coasian firm provides a low-cost way of discovering solutions to coordination problems of bottlenecks and uneven development of components relative to pure market transactions. The Coasian firm arises not only in response to problems of adapting to unforeseen outside contingencies but also as a response to technological uncertainty of the sort described here. In fact managed coordination is important even if there are no unforeseen contingencies which require adjustment in tasks. For managed direction of resources to be efficient, it is required that managers are at least as qualified in discovering the relevant prices (that is, finding the highest valued uses of assets) as independent contractors would be.7 Otherwise, costs of transacting may be saved at the expense of efficiency in the use of resources. If managers are better able to determine the valuable uses of resources compared to other agents, managers have a natural ownership advantage over resources. Such an advantage explains the single-person firm, but not necessarily why managers hire employees who are prepared to take orders within certain limits in order to take advantage of this knowledge. ‘Managers’ could as well rent the labour time of other agents in return for the exercise of a certain well-specified task. The argument here is that managers stand a good chance of acquiring superior information about the relevant prices of rights over assets which make up a complex technology. From the literature on incremental innovations, it is apparent that the solution to problems of bottlenecks and uneven development in components is based on learning by doing in production and development (Rosenberg 1976; Sahal 1981). This

ECONOMIC ORGANISATION AND THE ACCUMULATION OF RENT-YIELDING ASSETS 183

experience from learning by doing is probably more easily accumulated within the boundaries of firms. One of the reasons why one could expect this learning to be less costly within the boundaries of firms may be that managers who hold residual rights over assets, and in particular the rights to redefine and reallocate specific rights, are able to conduct experiments without continuously having to renegotiate contracts which have more or less unforeseen outcomes. This saves time and ink-costs.8 Managers then are able to create ‘controlled’ experiments in which they only change some aspects of the tasks in order to trace the effects of some specific rearrangements of rights. Setting up a controlled experiment may be more difficult across boundaries of firms when, owing to high information costs, it is difficult to specify all the tasks that must not be changed. Coordinating interdependent tasks within the boundaries of a firm may provide managers with a more complete picture of the nature of interdependencies—information which is important not only in relation to eliminating bottlenecks, but also in relation to avoiding problems of uneven development of components by setting up interface standards and other more permanent solutions. I have argued that allocating residual use rights in the hands of managers saves transaction costs in experimentation. This provides a reason for the use of labour contracts but it does not provide an argument for why full ownership (as compared to rental arrangements) of assets may be advantageous. Such an argument is required if we are to understand how firms can reduce the transaction costs involved in experimentation. Three reasons can be given. First, full ownership saves on the dynamic transaction costs of convincing suppliers of equipment that this equipment will not depreciate more in value from the experimentation than what can be expected from ordinary well-known uses of the equipment.9 Second, experimentation may reveal that the equipment has a higher value than that reflected in its price. Ownership makes it possible to reap the benefits of such discoveries. Finally, one may expect problems of assets specificity, since experimentation is likely to create human capital that is relatively specific to the equipment that has been used. So far I have argued that, relative to markets, firms may economise on the transaction costs of learning the best way of coordinating technological interdependent systems. Now, once a firm has discovered how to coordinate some specialised tasks, there may be little advantage from managed direction relative to market transacting, and coordination by order contracts would substitute for coordination by management.10 However, such specialisation between firms would give way to economic gains from further specialisation in tasks, and this in turn would create new uncertainty and new opportunities for reducing coordination costs by experimenting. In other words, there will be an ongoing process of specialisation in tasks, learning in coordination and specialisation between firms, and new ways of coordination will continuously be imagined by managers/ entrepreneurs. In the introduction to this chapter I indicated that the reinterpretation of the Coasian firm could provide a bridge between the contractual theories of firms and the resource-

184 KIRSTEN FOSS

based perspective. In the next section I discuss the nature of such a bridge. Focus is on why managerial knowledge can be a source of sustained competitive advantage to firms. The resource-based perspective: the nature of rent-earning resources The term ‘resource-based perspective’ refers to a number of recent publications in strategic management, taking their lead from Wernerfelt (1984) and Rumelt (1984). The resourcebased perspective is primarily a theory of firm-specific resources as sources of competitive advantage. Firms are perceived as bundles of more or less unique resources, competencies and capabilities which taken together determine firm profitability and the direction of profitable growth. The important theoretical as well as normative questions in the resource-based perspective are: when do firms earn long-lived rents from their valuable resources, when are rent-earning resources likely to have productive value in excess of their hire price, and how may firms best capture the rent-earning potential of valuable resources?11 As argued by Foss (1996), the resource-based perspective encompasses a more formal version, building (at least to some extent) on equilibrium analysis (e.g., Lippman and Rumelt 1982; Barney 1986; Dierickx and Cool 1989; Peteraf 1993) and a more appreciative and process-orientated version, as presented by, for example, Prahalad and Hamel (1990). In spite of its comparative static type of analysis, the more formal version of the resource-based perspective also features arguments related to causal ambiguity and history-specific learning processes as important explanatory factors of differences between firms with respect to the resources they control. One may say that the resource-based perspective, though firmly based in economic theory, tends to rely on real-time explanations of how differences emerge between firms with respect to the resources they control. One of the distinguishing features of resource-based theory is its emphasis on the specific characteristics of resources rather than on market conditions, which are common to all firms in an industry. In fact, a central assumption in the resource-based perspective is that sustained above-normal profit is linked to resources through their rent-earning potential. Barney (1986, 1991) listed five necessary and sufficient conditions that will ensure the firm sustainable economic advantages from its resources. In order to yield sustainable rents, resources have to be: 1 valuable, that is, they enable a firm to implement strategies that exploit opportunities or neutralise threats in the environment (Barney 1986,1991); 2 rare, and therefore in a position to generate Ricardian rents; 3 imperfectly imitable, making them sources of long-lasting rents; 4 not substitutable by other resources; 5 obtained at a price below their value, so that firms can earn above-normal return on their resources.

ECONOMIC ORGANISATION AND THE ACCUMULATION OF RENT-YIELDING ASSETS 185

If there is a bridge between a contractual and a resource-based theory of firms it must be because the economic organisation of activities has implications for the creation of assets that fulfil the above characteristics. I explained above how a Coasian firm may provide a benefit (i.e., improved efficiency) relative to pure market transacting by reducing costs of systematically experimenting with the coordination of production activities, and how such experimentation may lead to improved knowledge of how to coordinate specialised interdependent tasks (a production capability). Therefore, in order to search for a bridge between resource-based and contractual theories, I focus attention on the accumulation management knowledge of coordination. The question to be discussed is whether management knowledge in coordination can be a source of rent and long-lasting competitive advantage. The rent-earning capacity of management knowledge in coordination Knowledge of how to organise production is valuable to society if it reduces costs of producing valued goods. However, this does not explain why organisers of firms invest in productive assets in order to accumulate such knowledge. In other words, investments in the organisation of firms must be motivated not by efficiency considerations but by profit opportunities. I argued above that one of the reasons why firms emerge is to reduce cost of experimentation when faced with uncertainty as to how best to coordinate activities. Precisely because of the uncertainty, we expect organisers of firms to hold different expectations of the profit opportunities of different uses of the same types of productive assets. In order to discover these opportunity costs they conduct different experiments. Now, resource heterogeneity is a necessary but not a sufficient condition for longlasting competitive advantage. The resources also have to be inimitable and acquired at a price below their value. The first condition hinges on how easily the relevant resources can be replicated and here the concepts of ‘isolating mechanism’ and ‘resource-position barrier’ (Rumelt 1984; Wernerfelt 1984, respectively) are the essential theoretical ones for explaining the sustainable rents. An isolating mechanism/resource-position barrier may make the source of rent strictly inimitable or uncertainly imitable, depending on whether or not the resource is protected by enforceable rights to the exclusive use of the unique resource or by barriers consisting of: (1) cognitive constraints to imitation of success, (2) time disadvantages to imitators, or (3) economic disadvantages to imitators. Resources such as patents on inventions or ownership of a rich mineral deposit may be strictly inimitable while such resources as production or market experience, routines and culture may be inimitable owing to the above-mentioned imitation barriers. Below I discuss how these different imitation barriers may apply to assets stocks of management knowledge in coordinating productive assets.

186 KIRSTEN FOSS

Cognitive constraints to imitation of success First, there is causal ambiguity acting as a cognitive constraint to blockage imitation. Causal ambiguity is ‘the basic ambiguity concerning the nature of the causal connections between actions and results’ (Lippman and Rumelt 1982:420). The ambiguity disguises which factors are responsible for superior performance (Reed and DeFillippi 1990). This is an important simultaneous source of a barrier to imitation and to above-normal profit on accumulated assets. The latter is important, since even if valuable rent-earning resources can be protected from imitation, they may not necessarily be sources of absolute advantage in terms of high profits. According to Barney (1986), the unique and valuable resources also have to be underevaluated (acquired at a price below their value). Barney argues that firms only possess underevaluated resources because strategic factor markets12 are imperfect. Resources obtained in a perfect strategic factor market—that is to say, a factor market in which the price and discounted value of the resource coincide—will not earn above-normal profit, even if the implemented strategies create an imperfect product market. Firms that are either lucky or have superior insight may gain a competitive advantage by obtaining valuable resources at a price below their value.13 However, Lippman and Rumelt (1982) argue that uncertainty (ambiguity) may account for undervalued resources. In a formal model, they show that an ‘irreducible uncertainty’ which creates different expectations regarding the creation of a new product function may act as a barrier to the elimination of all profit opportunities in production activities. They further prove that an expectation equilibrium can exist where no firms expect to gain from further investments in imitative activities and where not all rents are competed away. The Lippman and Rumelt model is based on a model similar to the Austrian market model and for differences to be long lasting in such a setting uncertainty has to be irreducible. Irreducible uncertainty, though, does not explain why firms emerge. In the reinterpretation of the Coasian firm, as I have developed it above, technological uncertainty is an important factor explaining the existence of firms. I argue that, by experimenting in managed coordination, firms reduce this uncertainty. Therefore, the above expectation equilibrium may only consist of very few firms with very complex and interdependent production systems. However, this result depends strongly on the assumption about the kind of uncertainty facing firms. Time disadvantage to imitators Dierickx and Cool (1989) have pointed out that many of the resources over which firms have control are not acquired on factor markets, but are ‘the cumulative result of adhering to a set of consistent policies over a period of time’ (p. 1506). This applies to, for example, investments in reputation for quality, toughness in retaliation and R&D capabilities, as well as to investments in management knowledge of coordination. Such resources are often unique and also inimitable because of time disadvantages to imitators. This may be due either to history-specific path-dependent resource accumulation or to

ECONOMIC ORGANISATION AND THE ACCUMULATION OF RENT-YIELDING ASSETS 187

time barriers to resource accumulation. Organisational culture, for example, is intrinsically bound up with a firm’s unique history and heritage (Barney 1986, 1991) which cannot be copied. Time barriers to accumulation may be caused by time compression diseconomies, asset mass efficiencies and interconnectedness of asset stocks (Dierickx and Cool 1989). These disadvantages to imitators arise because it takes time to accumulate knowledge and because efficiency in accumulation depends on prior accumulated resources or the stock of complementary resources.14 Accumulation of management knowledge in coordination may be characterised both by history specificity (as when firms start different kinds of experiments), time compression diseconomies and asset mass efficiency (as when new knowledge is more easily accumulated when building on a stock of existing knowledge), and by interconnectedness of asset stocks (as when there are complementarities between learning taking place in different areas). In response to Dierickx and Cool, Barney (1989) has pointed out that the strategic factor market argument applies also to assets accumulated within the boundaries of firms, since there are also costs of internal accumulation of assets. Therefore, only in the face of uncertainty may internally accumulated assets produce above-normal returns. As regards the accumulation of management knowledge of coordination, firms may have different expectations of how much to invest in order to discover the best way of coordinating some specific tasks, and they may start with different kinds of experiments; this explains differences in their stock of management knowledge. However, once ‘the solution’ is found it may be very difficult for the successful firm to avoid diffusion of such knowledge. In fact, part of the success is being able to transmit this knowledge to suppliers. The long-lasting benefits of being ahead in accumulating management knowledge have to be found in the successful firm’s ability to move down the learning curve at a faster pace than competitors by gradually improving coordination and by continuously taking advantage of new opportunities for specialisation of tasks as they apply their accumulated stock of management knowledge to new problems. In other words, sustained competitive advantage of economic organisation of production activities within the boundaries of firms depends on the continuous introduction of new opportunities for taking advantage of accumulated management knowledge. As in a world resembling that of the radical subjectivists’ market model, these continuous opportunities are created endogenously to the production and learning process taking place in firms. Economic disadvantage to imitators As some firms move faster down the learning curve than their competitors they enjoy at least temporary profits, as they are protected by the mechanism I have called ‘economic disadvantages to imitators’. Wernerfelt (1984) focuses on resource-position barriers as isolation mechanisms. These barriers protect the kinds of resources where a first-mover advantage translates into a barrier to entry.

188 KIRSTEN FOSS

Conclusion I have argued that the Coasian firm may be interpreted as a solution to contractual problems that arise with increased specialisation. Specifically, the Coasian firm reduces the transaction costs involved in accumulating assets of knowledge of coordination. Now, once firms have learned how to coordinate interdependent activities, management costs must be reduced and this gives way for more specialisation and more learning. Whether or not some of the activities previously organised within the boundary of a firm are spunoff depends on the costs of transferring this knowledge,15 as explained in dynamic transaction costs theory (Langlois and Robertson 1995), as well as on the benefits from releasing more managerial resources to take advantage of further specialisation. It is the opportunities for economic gain from further specialisation that continuously reintroduce technological uncertainty and the need for management. With continuous uncertainty, there is room for continuous differences between firms with respect to their accumulated management knowledge of coordination, as well as for possibilities for long-lasting competitive advantages from such a stock of knowledge. Notes 1 However, Alchian and Demsetz (1972) do not really explain why the service of monitoring is more efficiently acquired in a firm rather than by market contracting. They state: ‘All of these tasks [measuring output performance, apportioning rewards, observing the input behaviour of inputs] are, in principle, negotiable across markets, but we are presuming that such market measurement of marginal productivity and job reassignments are not so cheaply performed for team production’ (p. 782). For this reason market competition for a place on the team does not sufficiently discipline team members. Potential team members cannot detect teams where shirking is a serious problem and even if they did find such a team they would have the same incentives to shirk as the person they replaced given the high cost for other team members in metering each other’s effort. 2 Evolutionary explanations provide an alternative to intentional explanations of differences in governance structure. In fact, Williamson suggests the existence of some evolutionary mechanism, which is responsible for the predominance of some modes of governance relative to others, depending on the circumstances under which the transaction takes place. However, he is not very explicit about the links between contractual relations, of which firms may have many, and the selection process. In addition, it does seem to be a rather static selection model (close to the Alchian (1950) selection model, where all firms are assumed be identical with respect to their propensity to grow) he has in mind; one in which variation in governance structure is given from the outset along with the potential contractual hazards which might make one governance structure more efficient than another. Moreover, it is necessary that there is a persistent search for ways of improving contractual arrangements. See Winter (1971) for a discussion of evolutionary process as a mechanism of global optimisation. 3 As explained by Williamson (1985:155): ‘To be sure, divisions also engage in obfuscation and cover-up against internal auditors. Division managers cannot, however, take the physical

ECONOMIC ORGANISATION AND THE ACCUMULATION OF RENT-YIELDING ASSETS 189

4

5 6

7

8 9

10 11

12 13

14

15

assets they have accumulated through cost overruns and flee. Termination with and without assets makes a difference.’ If he enters into a team of other individuals holding specific human capital, the optimal ownership structure is one where the person best able to influence the income stream generated by all assets holds the rights over physical assets, since this ownership structure ensures the greatest gains from investments in human capital. It should be mentioned that problems that arise from technological interdependencies may not be the sole reason for the existence of the Coasian firm. Problems of bottlenecks and uneven development of components exist even with selfsufficiency, since individuals producing for their own needs may be unaware of how best to carry out an activity or to develop the technologies they use. Specialisation in production, however, is likely to magnify the problems. Coase (1937) mentions ‘increasing opportunity costs due to the failure of entrepreneurs to make the best use of the factor of production’ (p. 23) as one of the factors which set a limit to the efficient size of a firm. He also assumes that ‘the costs of losses through mistakes will increase with an increase in the spatial distribution of the transactions organised, in the dissimilarity of the transactions, and in the probability of changes in the relevant prices. As more transactions are organised by an entrepreneur, it would appear that the transactions would tend to be either different in kind or in different places’ (p. 25). Managers, in other words, have limited capacity to ‘discover the relevant prices’ and this increases mistakes as more and more dissimilar transactions are organised in a firm. In this connection wage contracts may be an efficient way of sharing risks from experimenting. This is essentially a measurement cost argument. If suppliers were convinced that they could easily discover if the equipment had depreciated more in value from its use in experimentation they would not object to this use. Managed direction could still be advantageous in cases where adaptation of interdependent production systems to unforeseen contingencies was called for. The latter question may imply some overlap between the contractual and the resource-based perspective, since the economic organisation of transactions in assets with rent-earning potential certainly may have implications for the realisation of rents. When access to rentearning resources is secured—for example, through acquisition—the rents they yield may in principle be captured in numerous ways, such as selling the resource on a market, renting it or using the resource in-house (Dierickx and Cool 1989). In practice, it is the latter option that has primarily interested resource-based researchers. A strategic factor market is defined as ‘a market where the resources necessary to implement a strategy are acquired’. Wernerfelt (1984) identifies imperfections with the different levels of bargaining power possessed by firms. Barney’s line of argument also shows that superior bargaining power can be a source of above normal profit only if the resources underlying the bargaining power were acquired at a price below their value. Dierickx and Cool (1989) argue that ‘to the extent that new product and process developments find their origin in customer requests or suggestions…it may be harder to develop technological know-how for firms who do not have an extensive service network’ (p. 1508). Some of the coordination knowledge may be embodied in equipment and made more easily transferable. Firms may do this purposively to make their knowledge imitable or transferable

190 KIRSTEN FOSS

in order to release management resources for purposes where they expect higher pay-off of that factor.

References Alchian, Armen A. (1950) ‘Uncertainty, Evolution, and Economic Theory’, in Alchian (1977) Economic Forces at Work, Indianapolis: Liberty Press. ——(1965) ‘Some Economics of Property Rights’, in Alchian (1977) Economic Forces at Work, Indianapolis: Liberty Press. Alchian, Armen A. and Harold Demsetz (1972) ‘Production, Information Costs, and Economic Organization’, American Economic Review 62:772–95. Arrow, Kenneth J. (1962) ‘Economic Welfare and the Allocation of Resources for Invention’, in The Rate and Direction of Inventive Activity: Economic and Social Factors, Princeton, NJ: Princeton University Press. Barney, Jay B. (1986) ‘Strategic Factor Markets: Expectations, Luck, and Business Strategies’, Management Science 32:1231–41. ——(1991) ‘Firm Resources and Sustained Competitive Advantage’, Journal of Management 17:99– 120. Barzel, Yoram (1982) ‘Measurement Costs and the Organization of Markets’, Journal of Law and Economics 25:27–48. ——(1985) ‘Transaction Costs: Are They Just Costs?’, Journal of Institutional and Theoretical Economics 25:27–48. ——(1997) Economic Analysis of Property Rights, 2nd edn, Cambridge: Cambridge University Press. Cheung, Stephen N.S. (1983) ‘The Contractual Nature of the Firm’, Journal of Law and Economics 26: 1–22. Coase, Ronald H. (1937) ‘The Nature of the Firm’, Economica 4:386–405, reprinted in Oliver E.Williamson and Sidney G.Winter (eds) (1991) The Nature of the Firm, Oxford: Oxford University Press. ——(1960) ‘The Problem of Social Cost’, Journal of Law and Economics 3:1–44. ——(1991) ‘The Nature of the Firm: Origin, Meaning, Influence’, in Oliver E. Williamson and Sidney G.Winter (eds), The Nature of the Firm, Oxford: Oxford University Press. Demsetz, Harold (1973) ‘Industrial Structure, Market Rivalry, and Public Policy’, Journal of Law and Economics 16:1–10. Dierickx, Ingemar and Karel Cool (1989) ‘Asset Stock Accumulation and Sustainability of Competitive Advantage’, Management Science 4:169–84. Fama, Eugene F. (1980) ‘Agency Problems and the Theory of the Firm’, Journal of Political Economy 88:288–307. Foss, Nicolai J. (1994) ‘The Two Coasian Traditions’, Review of Political Economy 6: 37–61. ——(1996) ‘Whither the Competence Perspective’, in N.J.Foss and C.Knudsen (eds), Towards a Competence Theory of the Firm, London: Routledge. Frydman, Roman (1982) ‘Towards an Understanding of Market Processes’, American Economic Review 72:652–68. Grossman, Sanford and Oliver Hart (1986) ‘The Costs and Benefits of Ownership: A Theory of Vertical and Lateral Integration’, Journal of Political Economy 94:691–719. Hart, Oliver (1991) ‘Incomplete Contracts and the Theory of the Firm’, in Oliver E. Williamson and Sidney G.Winter (eds), The Nature of the Firm, Oxford: Oxford University Press. ——(1995) Firms, Contracts, and Financial Structure, Oxford: Clarendon Press.

ECONOMIC ORGANISATION AND THE ACCUMULATION OF RENT-YIELDING ASSETS 191

Hart, Oliver and John Moore (1990) ‘Property Rights and the Nature of the Firm’, Journal of Political Economy 98:1119–58. Hayek, Friedrich A.von (1945) ‘The Use of Knowledge in Society’, in Hayek (1948), Individualism and Economic Order, Chicago: University of Chicago Press. Jensen, Michael C. and William Meckling (1976) ‘The Theory of the Firms: Managerial, Agency Cost and Ownership Structure’, in Jay B.Barney and William G. Ouchi (eds) (1988) Organizational Economics, San Francisco: Jossey-Bass. Kirzner, Israel M. (1973) Competition and Entrepreneurship, Chicago: University of Chicago Press. Klein, Benjamin, Robert G.Crawford and Armen A.Alchian (1978) ‘Vertical Integration, Appropriable Rents, and the Competitive Contracting Process’, Journal of Law and Economics 21:297–326. Langlois, Richard N. and Nicolai J.Foss (1999) ‘Capabilities and Governance: The Rebirth of Production in the Theory of Economic Organization’, forthcoming in KYKLOS. Langlois, Richard N. and Paul L.Robertson (1995) Firms, Markets and Economic Change, London: Routledge. Lippman, S. and R.P.Rumelt (1982) ‘Uncertain Imitability: An Analysis of Interfirm Differences in Efficiency under Competition’, Bell Journal of Economics 13:418–38. Littlechild, S.C. (1986) ‘Three Types of Market Process’, in R.N.Langlois (ed.), Economics as a Process, Cambridge: Cambridge University Press. Penrose, Edith T. (1959) The Theory of the Growth of the Firm, Oxford: Basil Blackwell. Peteraf, Margaret (1993) ‘The Cornerstones of Competitive Advantage’, Strategic Management Journal 14:179–91. Prahalad, C.K. and Gary Hamel (1990) ‘The Core Competence of the Corporation’, Harvard Business Review 66 May–June:79–91. Reed, Richard and Robert J.DeFillippi (1990) ‘Causal Ambiguity, Barriers to Imitation, and Sustainable Competitive Advantage’, Academy of Management Review 15:88–102. Rosenberg, Nathan (1976) Perspectives on Technology, Cambridge: Cambridge University Press. Rumelt, Richard P. (1984) ‘Toward a Strategic Theory of the Firm’, in R.P.Lamb (ed.), Competitive Strategic Management, Englewood Cliffs, NJ: Prentice Hall. Sahal, Devendra (1981) Patterns of Technological Innovation, Reading, MA: Addison-Wesley. Shackle, George L.S. (1972) Epistemics and Economics, Cambridge: Cambridge University Press. Simon, Herbert A. (1951) ‘A Formal Theory of the Employment Relationship’, Econometrica 19: 293–305. Wernerfelt, Birger (1984) ‘A Resource-Based View of the Firm’, Strategic Management Journal 5:171– 80. Williamson, Oliver E. (1975) Markets and Hierarchies, New York: The Free Press. ——(1985) The Economic Institutions of Capitalism, New York: The Free Press. Winter, Sidney G. (1971) ‘Satisficing, Selection and the Innovating Remnant’, Quarterly Journal of Economics 85:237–61.

192

10 Capabilities and vertical disintegration in process technology The case of semiconductor fabrication equipment Richard N.Langlois

Introduction Alfred Chandler’s influential book Scale and Scope (1991) has added new weight to the Schumpeterian proposition that the large vertically integrated firm was at the vanguard of economic growth in the late nineteenth and early twentieth centuries (Schumpeter 1950: 82). A number of writers have taken the message of Schumpeter and Chandler to be that economic capabilities are always best created within the framework of large firms enjoying internal economies of scale and scope; as a consequence, industrial competitiveness depends crucially—and perhaps even exclusively—on fostering internal capabilities (Lazonick 1991). For a time at least, this argument found special application in the arena of high technology, including semiconductors. Until recently, it was common to hear that the decentralised, entrepreneurial American industry ultimately would prove no match for the large firms of Asia, and that the United States must somehow consolidate and bolster the internal capabilities of its firms, with government help if necessary (Florida and Kenney 1990; Ferguson 1985, 1990). There is certainly much wisdom in the Schumpeter-Chandler view, especially as an antidote to the naive adulation paid to the model of atomistic competition in neoclassical theory and policy. Nonetheless, there is reason to think that an overemphasis on the internal creation and management of economic capabilities can be equally unhealthy. Networks of decentralised firms, including those networks often derogated as ‘markets’, can also be repositories and generators of economic capabilities. To insist on vertical integration—or, for that matter, on agglomerations of small ‘flexibly specialised’ producers (Piore and Sabel 1984; Best 1990)—as a universal prescription misses the subtlety and historical idiosyncrasy of industrial evolution (Langlois and Robertson 1995). A crucial issue in the comparison among the institutions of industrial organisation is the ability of those institutions to generate technological progress. At the risk of oversimplifying matters somewhat, we might say that the relative merits of firms and networks hinge on whether innovation is systemic or autonomous (Teece 1986). When innovation is systemic, there is reason to think that a firm-like structure will prove more conducive to rapid technological progress. This is so because systemic innovation requires simultaneous change in many different stages of production, and common ownership of

194 RICHARD N.LANGLOIS

complementary stages lowers the transaction costs of persuasion and coordination (Silver 1984; Langlois and Robertson 1995). By contrast, there is reason to think that networks— including archetypical ‘markets’ as one extreme—may have advantages when innovation is autonomous, that is, when technological change in one stage of production can proceed without requiring corresponding changes in other parts of the system. In this case, the advantages of the firm in persuasion and coordination are outweighed by Smithian economies of specialisation and by the ability of networks to access a larger and more diverse pool of relevant capabilities (Langlois 1992; Langlois and Robertson 1992). But the systemic or autonomous character of innovation is not necessarily a ‘given’ factor. Rather, it depends on the design structure or architecture (in the broad sense) of the underlying technology.1 Recently, scholars have begun to emphasise the importance for competitiveness of technological architecture. For example, Morris and Ferguson (1993) insist that it is architecture that ‘wins technology wars’. And, in their view, a modular product architecture has the potential to throw off far greater competitive rents than does a ‘traditional’ or interconnected architecture. As Baldwin and Clark (1997) describe it, a modular system requires ‘visible design rules’ of three types. The first is an architecture (in the narrow sense) that specifies how the system is partitioned into modules. The second is a set of interfaces that specify how the modules will connect together and communicate. The third is a set of standards that test a module’s conformity to the design rules or test the module’s performance relative to other modules.2 For brevity, many writers refer to all three elements as ‘the standards’. Such standards allow a large degree of autonomy in the modules, which not only can but actually must rely on information that is hidden from other modules in the system (Cox 1990). This allows the modules to make use of local information without having to communicate that information elsewhere in the system, and it also prevents other parts of the system from interfering with a module’s operations in what is inevitably an unpredictable way. The benefits of modularity in cutting through the spaghetti tangle of interactions in a complex system has long been recognised (Simon 1962; Alexander 1964). From an economic perspective, these benefits come on both the demand side and the supply side (Langlois and Robertson 1992). One effect of standard ‘interfaces’ among components is to lower the costs of assembly, both by lowering transaction costs and by reducing the optimal scale of the assembly function. As a result, assemblers (who may be the system users themselves in some cases) can more cheaply tailor a system to the user’s exact requirements. On the supply side, modular systems can lower production costs by enlisting specialisation in the cause of innovation. Management scholars have begun to argue that modularity in technological design implies or causes modularity in the structure of the organisations that use and produce the technology (Sanchez and Mahoney 1996; Sanchez in this volume). Morris and Ferguson (1993:93) claim that, in what they refer to as the Silicon Valley Model, ‘the structure of the firm itself mirrors the technical architectures it uses’. This may well be so. But as Langlois and Robertson (1992) point out, the degree of modularity in the design of the underlying system also affects the boundaries of the firm. The principal effect of a modular

CAPABILITIES AND VERTICAL DISINTEGRATION IN PROCESS TECHNOLOGY 195

system may be to give advantage to small vertically specialised firms rather than to diversified firms, even those organised in a ‘modular’ way. After all, coordination through market interfaces is the most fundamental and important example of modular organisational design (Hayek 1945). Ultimately, modularity breaks the barrier of the boundaries of the firm, bringing to bear both a larger volume and a wider diversity of capabilities than even the largest of organisations could cheaply marshal. We can also think of the role of modularity in terms of economies of scope. Following in the tradition of Edith Penrose (1959), Teece (1980, 1982) has argued that economies of scope arise when an organisation has excess capabilities that can usefully be applied to activities similar to the ones in which it is already engaged. To the extent, for example, that a software firm reuses pieces of existing code in writing new packages, it partakes of internal economies of scope: it can write several programs more cheaply per program than could separate organisations that had to write each package from scratch (Cusumano 1991). Analogously, an open modular standard helps create external economies of scope (Langlois and Robertson 1995:5). By making public a common way for the components to fit together, a modular standard allows many distinct organisations to ‘reuse’ the same knowledge. Each organisation can concentrate on changes that improve a component without having to reinvent other components or the architecture that connects them.3 With such external scope economies, a collection of separate organisations can produce a system more cheaply than can a single organisation that has to generate all the relevant capabilities internally. In software, some computer scientists are suggesting that objectorientated programming may lead to inter-firm reuse of code as an alternative to the internal scope economies of the ‘software factory’ (Cox 1990). The examples of the stereo, the PC and even software conjure up the image of a modular system as applying to final-consumer goods. Moreover, the literature on the learning curve often implies that in the realm of manufacturing and process technology, systemic innovation—fine tuning the production process—is the norm. This chapter is an attempt to explore further the concept of a modular system by examining an intriguing case in which these preconceptions do not hold. In the semiconductor equipment industry, especially that branch concerned with so-called cluster tools, a modular manufacturing system may be emerging that will have important implications not only for the structure of that industry but also for international competitiveness. Case study: semiconductor fabrication equipment Overview The integrated circuit was very much an American invention, developed independently but simultaneously by researchers at Texas Instruments (TI) and Fairchild in 1959. As an integrated-circuit industry grew out of the discrete-transistor industry, American firms dominated, both in the fabrication of the chips themselves and in the manufacture of the equipment to make chips. In the early days, semiconductor firms developed much of their

196 RICHARD N.LANGLOIS

Table 10.1 Top ten semiconductor equipment suppliers, 1979 and 1989

Note: Dollars in millions. US=US firm; J=Japanese; E=European. Source: VLSI Research, cited in US General Accounting Office (1990).

own process equipment, often in collaboration with firms in the scientific-equipment industry. Gradually, a distinct semiconductor equipment industry emerged. In 1979, nine out of the top ten firms were American. With the rise of Japanese IC fabrication in the 1980s and the loss of American market share in dynamic random-access memories (DRAMs), American dominance in semiconductor equipment also declined. By 1989, only four of the top ten were American, and only Applied Materials remained among the top five (see Table 10.1). Between 1980 and 1988, worldwide sales of equipment for lithography, chemical vapour deposition (CVD) and ion implantation quadrupled; during the same period, the American share fell from 75 to 49 per cent, while the Japanese share rose from 18 per cent to 39 per cent (Department of Commerce 1991). The Japanese success was most pronounced in lithography equipment (dominated by Canon and Nikon), automatic test equipment, and assembly and packaging equipment. The decline in American pre-eminence in semiconductor equipment generated much the same angst as the better-known decline in American market share in DRAMs. A number of groups, including the National Advisory Committee on Semiconductors, issued dire warnings (NACS 1990). And Sematech, the government-industry consortium, quickly began defining much of its role as helping to reverse the fortunes of the American equipment industry (Robertson 1991). The diagnosis of the equipment industry’s problems was similar to that for the semiconductor industry as a whole: the American industry suffers from excess ‘fragmentation’ and insufficient vertical integration. In one of the few academic examinations of this industry, a study by the Berkeley Roundtable on the International Economy (BRIE) concluded that with regard to both the generation of learning in production and the appropriation of economic returns from such learning, the US semiconductor equipment and device industries are structurally disadvantaged relative to the Japanese. The Japanese have evolved an industrial model that combines higher levels of

CAPABILITIES AND VERTICAL DISINTEGRATION IN PROCESS TECHNOLOGY 197

Table 10.2 Top ten semiconductor equipment suppliers, 1996

Note: Dollars in millions. US=US firm; J=Japanese; E=European. Source: VLSI Research, cited in Erkanat (1997).

concentration of both chip and equipment suppliers with quasi-integration between them, whereas the American industry is characterised by levels of concentration that, by comparison, are too low and [by] excessive vertical disintegration (that is, an absence of mechanisms to coordinate their learning and investment processes). (Stowsky 1989:243, original emphasis) By 1992, however, the situation had changed markedly. American firms regained the lead in market share in semiconductors generally, albeit only slightly at first (Pollack 1992). This turnaround was propelled by the growth of the personal computer market, whose fundamental building block—the microprocessor—was an American strong suit (Langlois and Steinmueller 1999). American firms also increased their attention to manufacturing quality in response to the Japanese challenge. As the fortunes of the American semiconductor industry began to reverse themselves, and with the growth of non-Japanese production in Asian countries (notably Korea) that lack an indigenous equipment industry, the fortunes of the American equipment industry also began to turn up. By 1992, Applied Materials had overtaken its Japanese rivals to become the largest semiconductor equipment firm in the world (see Table 10.2). More significantly, American firms are strong in a technological approach that promises to become increasingly important as integrated circuits become more densely packed. The flexible-cluster tool segment of the equipment market may grow from $359 million in 1995 to $1.44 billion in 2000 (Anonymous 1996). And long-run prospects may be even better: as I will suggest below, the modular cluster tool concept may prove to be a technological trajectory ultimately applicable to the entire semiconductor production process. If American firms do succeed in this developing market, it will be because they have indeed managed to develop ‘mechanisms to coordinate their learning and investment processes’. But, without minimising the importance of improved cooperation among firms, those mechanisms may not involve significant increases in concentration or vertical integration. Cluster tools are rapidly becoming a modular system, driven by standards

198 RICHARD N.LANGLOIS

now emerging. And modular standards are a mechanism to coordinate learning without integration. Moreover, by reducing the minimum efficient scope of the firm—if I may make that analogy with the concept of minimum efficient scale—cluster tool standards may also reduce the costs of miscoordinating investment, thereby moving the function of explicit coordination within the range of those mild forms of quasi-integration that come under the heading of ‘corporate partnering’. From batch to single-wafer processing In order to understand these trends, and the issues of modularity and standard-setting that attend them, one needs to look more closely at the semiconductor production process.4 The traditional approach to the mass production of semiconductors has been batch processing. Silicon wafers, each containing what will become many separate chips, move through the various steps in batches, queuing up when necessary in work-in-process (WIP) inventories. For example, a large vertical furnace may process more than 100 wafers at a time. All high-volume ‘fabs’—as semiconductor production lines are now invariably called —currently use batch techniques, and these have worked well through the current DRAM generations. The alternative to batch processing is single-wafer processing. Instead of processing many wafers simultaneously, single-wafer systems process one wafer at a time. This is analogous to the continuous-throughput techniques that have largely supplanted batchprocessing approaches in the chemical industries.5 Although years of learning currently place the overall economies squarely on the side of conventional batch, there are advantages to single-wafer processing that are emerging and are likely to become increasingly significant as semiconductor line widths decrease below 0.5 microns.6 These advantages fall under three headings: • reduced cycle time; • greater atmospheric control and uniformity; and • the potential for real-time monitoring. Cycle time is the time from when blank wafers enter the production system to when completed wafers emerge and are ready for assembly and packaging. In a batch system, output rates may be high, but so is cycle time. Consider the analogous problem of washing a kitchen-full of dirty dishes. Using a dish-washer is a batch process; washing by hand is a continuous process. Loading the dishwasher may ultimately have a larger ‘throughput’, but the first clean plate reaches the cupboard more quickly with hand-washing. Batch semiconductor processing is like running dishes sequentially through many different dishwashers with many different capacities. This creates a queuing problem, and the wafers must often sit around in WIP inventories while waiting to form a batch of the appropriate size for the next process step. By contrast, single-wafer systems push only a single wafer through at a time (putting aside parallel processing steps), but the progress of that single wafer is not materially slowed waiting for other wafers to be ready. One of the

CAPABILITIES AND VERTICAL DISINTEGRATION IN PROCESS TECHNOLOGY 199

advantages of single-wafer, then, is that chips suffer less degradation waiting in WIP inventories. Those inventories are stored in normal (albeit hyper-clean) atmosphere within the clean room.7 This allows oxygen to attack and oxidise the wafers, producing a ‘black silicon’ that can reduce yield, the fraction of total chips processed that actually work. Other process sequences are sensitive to moisture in the atmosphere. The effects of atmospheric degradation become increasingly significant as line widths get smaller. In addition, WIP inventories are subject to other kinds of oxidations, to polymer deformation of resists, and to ordinary dust contamination and handling breakage. In addition to reduced queuing time, single-wafer systems can also speed throughput because it simply takes less time to process a single wafer than a batch of wafers. This is so for physical reasons: it takes more time to heat up or cool down a large batch than a single wafer, for example. A single-wafer system may also be more easily controlled in a number of respects. This means that the wafer spends less time in the machinery, an important source of lower cycle time. And, quite apart from problems of degradation, faster cycle time means that the first chips get to market more quickly, which can significantly affect ultimate demand by making it more likely that engineers will choose the chip in a systems design. Even for standardised chips like DRAMs, lower cycle time is important because profits are highest earlier in the product life cycle. Also, and perhaps most importantly, reduced cycle time means potentially faster learning by doing, since it permits production engineers to see more quickly the full effect on a wafer of all the process steps and allows them to adjust the process for all subsequent wafers (rather than for subsequent batches of wafers). The single-wafer approach also has benefits for atmospheric control that go beyond limiting degradation from waiting time in WIP queues. With a single-wafer system, one can more easily integrate or cluster together sequential process steps within a controlled atmosphere. This helps to eliminate cleaning steps that would otherwise be necessary if the wafers were exposed to air between steps. Moreover, large batch tools, such as diffusion furnaces, cannot maintain uniformity of temperature and other parameters across all the wafers in the batch, a problem that becomes increasingly important as line widths diminish. By processing only one wafer in a chamber at a time, single-wafer tools can achieve much greater process uniformity.8 A long-term benefit of single-wafer processing may be the ease with which the wafers can be monitored and tested in real time rather than at discrete testing steps. Such monitoring would provide a steady stream of data for operators to use in detecting problems quickly and for process engineers to use in uncovering bottlenecks and fine-tuning the system. This would include improved manufacturing-process documentation and more reliable ‘recipe downloading’, the process of programming process steps. Moreover, the real-time aspect of the data makes it possible to engage in closed-loop control, that is, to test and adjust the process as it is happening rather than to wait until a step is finished, test, and then adjust subsequent runs. In the long run, the single-wafer approach may lead more easily to overall factory simulation, including linking to computer-aided design and engineering (CAD/CAE). Many of these benefits remain on the horizon, of course. Single-wafer processing is a technological trajectory that, for some processes at least, arguably promises lower costs

200 RICHARD N.LANGLOIS

Figure 10.1 A parallel-processing configuration

than batch processing in the long run, but which is not yet nearly as developed as the batch approach. Nonetheless, single-wafer tools are making inroads. Such tools have got a foot in the door of otherwise batch fabs by offering what people in the industry call an ‘enabling technology’. That is, single-wafer tools are beginning to sell because they perform certain specific functions in the fab better than the alternatives. These functions are controlled-atmosphere steps, and today typically include dielectric planarisation, the smoothing of certain layers on the chip, and intermetal connection, the tricky business of making electrical contacts among the various levels of circuitry in a chip.9 These process functions are becoming increasingly important as chips become more sophisticated. Introducing a single-wafer step into a batch fab instantly creates a bottleneck, of course, since throughput of the fab is limited to the throughput of the single-wafer step. The obvious answer is to replicate the bottleneck stage in a parallel-processing configuration. The need for parallel processing was the original motivation for common-platform tools (see Figure 10.1). Rather than having, say, four separate stand-alone process chambers, each with its own separate wafer loading and unloading facilities, one could mount the four chambers on a common platform and use a common wafer-handling mechanism to move wafers to and from the various chambers and input-output loadlocks. Some early

CAPABILITIES AND VERTICAL DISINTEGRATION IN PROCESS TECHNOLOGY 201

Figure 10.2 A hypothetical modular integrated-processing system (after Burggraaf 1989)

tools that took this approach were General Signal’s Drytek Quad (1985) and Applied Materials’ Precision 5000 (1987). From the common-platform configuration, however, it becomes an easy step to sequential rather than parallel processing. Instead of running the same process in all four chambers, one could run different processes, using the wafer handler to move the wafers from one to the other within a controlled atmosphere. This was the genesis of the integrated cluster tool (see Figure 10.2). Such tools also became enabling technologies in some areas, since they obviated removal of the wafers from the controlled atmosphere, preventing deterioration and eliminating cleaning steps. The Precision 5000 is an archetype of a successful cluster tool platform. Of the 800–900 machines shipped by 1992, about 40 per cent were configured for integrated or sequential processing, with the remainder configured for parallel processing.10 The Drytek and Applied machines, among others, are closed proprietary systems. The chambers reside on a central platform or ‘mainframe’ and are linked by a centralised control and communications architecture that uses a closed proprietary standard. Moreover, the chambers themselves are ‘dumb’, relying on commands from the central controller. By contrast, modular cluster tools—or simply modular tools—comprise selfcontained ‘smart’ modules, each possessing its own computer and its own piping. The modules are tied together not by a central controller but by a set of open interconnect and control standards. The modules conform to a mechanical interface standard, which governs the placement and dimensions of the modules and handlers, and to various

202 RICHARD N.LANGLOIS

communications standards, which govern the way the decentralised computers talk to each other over a network. Modular tools represent an alternative not only to proprietary cluster tools but also, in the long run, to the concept of stand-alone tools within a normalatmosphere clean room. Genuinely independent modules for all fabrication steps could be linked and combined so that, in the limit, the wafer never leaves the controlled internal environment of the system. (Indeed, the modules need not be single chambers but could themselves be proprietary multichamber tools—so long as they conform to the standards in their dealings with the other modules.) All the modules would be tied together in a computer network, providing a real-time database for monitoring and engineering improvement. This is the ultimate vision of single-wafer processing, what some call the ‘pipeline fab’. Although the ultimate vision is still far in the future, there have already been a number of efforts in its direction. By the early 1990s, Texas Instruments had constructed a complete single-wafer fab line on an experimental basis. The project, called the Manufacturing Management Science and Tech nology (MMST) program, was funded by the US Air Force and the Defense Advanced Projects Research Agency (DARPA), which were anxious to make certain that American companies possessed all the capabilities necessary for flexible production lines to make a variety of low-volume chips for defence purposes (Iversen 1989). TI’s mandate was to produce commercial equipment, even though the company did not want to get into the business of producing cluster tools (even for internal use) and would at most license some aspects of the technology.11 It does seem clear that the single-wafer approach is well suited to low-volume applications. A more difficult question is whether, when and to what extent single-wafer approaches will be competitive for high-volume mass-production applications. It is significant in this light that IBM has apparently committed itself to the single-wafer approach in its 16M DRAM fab at Essex Junction, Vermont. The line clusters together lithography, dry etching, CVD and other ‘hot’ processing, and deposition-etch sequences. These integrated stages are coupled by intelligent robots that transport the wafers among the ‘islands of automation’. IBM’s explicit goals in taking the single-wafer approach are to reduce the complexity of the production process and to decrease cycle time, thus speeding ‘process, circuit, and defect learning’ (Bergendahl et al. 1990). In most applications, however, it seems likely that a hybrid approach—mixing batch and singlewafer tools—will emerge and persist for some time. The economics of modular cluster tools The most forceful driver of cluster tool adoption in the short term will be technical superiority, the role of such tools as an enabling technology. There are a number of potential areas in which cluster tools play this role. But perhaps the hottest growth area for such tools is in the deposition of Tungsten for intermetal connection, a procedure that requires the low temperatures and controlled atmosphere of a cluster tool (McLeod 1990; Winkler 1991). But this function can be accomplished by closed proprietary tools operating as isolated stations within a batch-processing environment. The speed with

CAPABILITIES AND VERTICAL DISINTEGRATION IN PROCESS TECHNOLOGY 203

which genuinely modular cluster tools penetrate the market will depend on a number of factors: economies of standardisation and specialisation; flexibility; reliability; and organisational issues. Specialisation and standardisation One of the most prominent potential benefits of an open modular cluster tool standard is the ability it offers users to mix and match components—process modules, wafer handlers, etc.—from different manufacturers in a way that takes advantage of the ‘best of breed’. A user might mix a CVD module from manufacturer A with an etch module from manufacturer B and a wafer handler from manufacturer C, all assembled and guaranteed by system integrator D, who might add in some off-the-shelf components like valves and controller software. If, however, manufacturer E produces a CVD module that is innovative or otherwise superior in the eyes of the user, that module could replace module A in the package. In this way, the user does not have to rely on the capabilities of any single firm, which may not be at the cutting edge in all technologies.12 Of course, one might list advantages to production within a single firm that possesses significant economies of scale and scope. Producing self-contained modules may require duplication of parts—piping, controllers, etc.—that could be shared on a closed system. Moreover, a dominant manufacturer of a closed proprietary system could conceivably overcome the disadvantages of having less-than-best-of-breed technology by producing at high volume, thus reducing costs and gaining learning-by-doing benefits. Indeed, critics of American manufacturing deride the US penchant for cutting-edge technology as a fatal obsession. The Japanese semiconductor industry has succeeded despite a conservative approach that prefers refining existing techniques to trying out new technology. In the end, the matter is an empirical one. But there are a number of considerations that weigh in on the side of the modular approach. As I suggested above, one effect of common standards is to create external economies that have much the same benefits as the internal economies of a large firm. Ultimately, economies of scale and scope, whether internal or external, are typically economies of knowledge reuse.13 In the case of modular systems, the existence of standards allows a component maker to use knowledge about other parts of the system without actually possessing that knowledge. By thus lowering the minimum efficient scope of the firm, a modular system encourages competition without duplication. That is, firms competing to supply the parts of a modular system need to generate a much smaller amount of overlapping non-innovative knowledge. An anecdote may convey the spirit of this idea. During the meetings to develop cluster tool standards (on which see below), organisers conducted an informal poll of firms present to gauge which types of businesses were represented. When the participants were asked which of them represented robotics firms, nearly all raised their hands.14 In the absence of standards, the separate competitors were all forced to generate their own internal capabilities in robotics, even though many if not most would have preferred to buy a wafer-handling system if one had been available cheaply. The availability of a compatible

204 RICHARD N.LANGLOIS

third-party wafer handler is an external economy of scope that substitutes for the reusable internal capabilities in robotics that a firm might generate itself. Lowering the minimum efficient scope of the firm can lower costs in two ways. First, the standardisation of parts permits economies of scale in the production of those parts. A prime example of this is in the area of valves, which control the flow of gases between chambers. Before the development of standards, all tools used their own idiosyncratic valve designs. Outside suppliers would craft each valve to the user’s specifications. The dominant firm in the business is VAT of Liechtenstein, which is noted for the quality of its product. Since the promulgation of standards, however, a standard valve has emerged, making valves more a commodity and less a specialty item. American firms like High Vacuum Apparatus (HVA) and MDC Vacuum Products have begun to take business away from VAT, and valve prices have fallen dramatically. Another area in which standardisation is lowering costs is software. With the development of communications and control standards, an increasing number of aspects of the control software can be handled by standard packages provided by firms like Thesis, GW Systems, Realtime Performance and Techware Systems. This may not be the most important source of cost savings, however. A possibly more significant effect is the lowering of the costs of the modules, handlers and other major parts. Even though these are not likely to become commodities or to be produced in volumes that will generate significant economies of mass production, increased competition among component makers will nonetheless tend to spur the search for cost economies.15 What about the internal scope economies of a centralised architecture? It seems very likely that such economies, if important at all, disappear very quickly with increasing system complexity. Cluster tools are becoming decentralised for much the same reasons that networks of personal computers and work stations are replacing mainframes and minicomputers: the rapidly decreasing costs of decentralised computing power. Indeed, at least one maker of modular cluster tools uses controllers for its modules that are essentially off-the-shelf IBM PC-compatible personal computers without keyboards and monitors. Moreover, there is reason to think that makers of closed proprietary systems would have switched to distributed architectures for their future tools even in the absence of a movement towards standards. Furthermore, the best-of-breed argument can be extended from the initial configuration of the machine to changes in configuration over time. One of the benefits of a modular system is the ability to change components as the user’s needs change or as component technology improves (Langlois and Robertson 1992). In the case of semiconductor fabrication equipment, user needs are driven by the obsolescence of the devices being produced. In the last few years, the life cycle of leading-edge chips has fallen to two or three years. Producing a new generation of ICs in an existing fab means scrapping perfectly good production equipment. With a modular cluster tool approach, however, one can keep parts of the system (the wafer handler at the very least and perhaps some of the modules) while scrapping only the obsolete modules. In this way, cluster tools extend the economic life of the user’s capital equipment (McNab 1990). Even closed proprietary systems take a ‘platform’ approach, in which common components

CAPABILITIES AND VERTICAL DISINTEGRATION IN PROCESS TECHNOLOGY 205

serve as the basis for an upgradable family of tools (Monkowski 1991). But a standardised modular approach would go beyond the proprietary platform, freeing users from the technological trajectory of any one company. A related factor in favour of modular tools is the blurring of the line between prototype equipment and production equipment. When a user wants to adopt a new technology for some production step, it typically first tries the process on a demonstration or development line. If the new technology is to be adopted in a batch production line, however, the production model may have to be quite different from the test model, which would have been designed to produce only small sample batches. In a single-wafer system, the production model will be much closer to—or, in the limit, identical with— the test model, since both will have been designed for continuous processing rather than batch. Modularity makes the transition even easier, since both the test line and the production line would conform to the same standards. The modular approach thus eases the always difficult transition from development lab to production line, which in turn speeds the rate of technological improvement. Flexibility If modular tools allow the flexibility to change process technology cheaply when IC generations change, then it also arguably allows flexibility to change the production process cheaply for other reasons. For example, one might be able to switch a fab to the production of a different kind of chip by changing only a few modules. Even with a modular system, of course, such change can be costly. Production engineers insist on ‘qualifying’ process tools, that is, subjecting them to a rigorous set of tests and measurements, before they can be put into service. Even bolting on a few new modules requires that the system be requalified, at least in part, and that takes time. But the process of changeover in a modular system is surely cheaper than making the equivalent change to a completely new dedicated tool set. This flexibility effectively creates economies of scope at the level of IC production. The value of such economies will depend on the demand for standardised chips. Increasingly, fabs around the world cannot fully utilise capacity producing only DRAMs for a saturated market. Thus, there remain few lines that do not produce more than one kind of chip. For example, a DRAM line might also occasionally produce application-specific integrated circuits (ASICs), which require a quite different sequence of process steps. The smaller the production run, the greater the value of flexibility, an observation that underlies the widespread belief that cluster tools (and single-wafer systems in general) will find their earliest economical uses in low-volume specialty fabs. As I suggested above, the question is whether—or, more likely, when—the cost performance of single-wafer techniques will ultimately overtake that of batch techniques at high volumes. Even if batch techniques continue their mass-production superiority for some time, the modular single-wafer approach may nevertheless become important by providing a flexible-manufacturing alternative (Steinmueller 1992). Whatever the ultimate shape of the cost curves at high output levels, it is clear that minimum efficient scale sets in much

206 RICHARD N.LANGLOIS

sooner for an IC producer using flexible single-wafer technology. This may make it economical to produce small volumes of specialised chips rather than large volumes of generic chips, effectively shifting competition in ICs to non-price margins (where the American industry has always felt most comfortable) and away from the price margin (where the Americans have done less well). An alternative to specialised chips is the ASIC, which is a generic logic chip that can be mass produced at low cost and then customised by the user in one of a number of ways. However, such chips perform less well than chips customised at the mask level (that is, chips customised from the start), and their range of application is limited. The success of the flexible-production alternative will depend on the cost-performance characteristics of customised chips (relative to ASICs) and on the cost-performance characteristics that chip users demand. Even though defence procurement, an important source of demand for high-performance, low-volume chips, is likely to diminish, the increasing sophistication of electronic equipment in general may be a favourable trend for the customised-chip business. Reliability The making of semiconductor chips is a capital-intensive business. But unit costs per wafer are typically much more sensitive to factors like throughput, yield and tool reliability than to the purchase price of the tool (Carnes and Su 1991). And reliability, in particular, has been a weakness of cluster tool systems. Part of the disaffection with such devices in Japan and Korea evidently stems from unhappy experiences with unreliable tools in the past (Burggraaf 1991:66). In a continuous-throughput system, a breakdown in any link in the chain can shut down the whole line. If we think of a cluster as a linear sequence of process steps, then the reliability of the whole system goes up as the power of the number of linked steps. For a cluster to achieve 90 per cent reliability over some time interval, for example, each step of a six-step process would need a reliability of 98.3 per cent. At the same time, of course, the existence of parallel-processing steps would create redundancy. If a cluster contained two chambers running process A and one running process B (perhaps because process A takes twice as long as B), the system could limp along if one of the A chambers were to go down. Indeed, some have suggested that, in a modular system, a fab could have spare chambers waiting in the wings to replace a downed chamber; critics wonder, however, whether this would save much time, since the tool would have to be requalified even if the spare chamber had been tested off-line (Newboe 1990:86). On the other side of the ledger, the economic costs of failure in a cluster module can be much smaller than the analogous costs in a stand-alone batch tool. A vertical furnace holding 100 wafers, each worth as much as $10,000 in potential microprocessor chips, puts $1 million at risk. By contrast, a malfunctioning cluster chamber risks only one wafer. In the end, however, the reliability of cluster tools will depend on the quality of the product, as measured by the process results those tools achieve. As in other industries, the

CAPABILITIES AND VERTICAL DISINTEGRATION IN PROCESS TECHNOLOGY 207

American semiconductor equipment industry has been better at product development than at product reliability. Part of the reason for this has been the limited internal capabilities of many American firms. With the advent of standards, however, smaller firms will tend to specialise in particular technologies, allowing a kind of capability deepening. By producing, for example, only one specific module—rather than a module plus complementary technology—a company would be better able to ‘debug’ its product and improve reliability.16 Organisational issues Economists have always understood, at least in principle, that the ‘product’ a firm sells is not often just the unencumbered item that leaves the premises in a shipping crate. Even a manufacturer of a physical product is often ultimately the provider of a service, and the physical item is necessarily bundled with less tangible informational and other services. This is certainly true in the case of semiconductor manufacturing equipment. Here the bundled services include both information and a guaranteeing function. A manufacturing system must fit in with a user’s production line, and it must work properly and consistently. When it fails to work, it must be fixed promptly; moreover, the user must be confident that it will indeed be fixed promptly. And the user and the supplier must communicate information to ensure the continued refinement and improvement of the technology. A large firm with significant internal capabilities can provide these ancillary services. Applied Materials is the American prototype of such a firm, although other firms, like Lam Research and Novellus, come close. Such a firm possesses not only the majority of skills necessary to fabricate the machinery it sells, it also possesses complementary capabilities in repair and customer service, including the ability to gather information to improve the product. Reputation is another important complementary asset, since it provides a guarantee to customers that promised ancillary services, especially onsite repair, will be reliably provided. In this respect, a modular system provided by a network of firms would seem to be at a disadvantage. If the modular approach is to succeed, the role of the system integrator is crucial. A system integrator is an organisation that packages the products of a number of suppliers—chambers, wafer handler, etc.—and provides the necessary ancillary services, including the guaranteeing function. In the absence of standards, the system integrator would almost certainly have to provide most of the technology itself, since working with others would require sharing of proprietary information in a way that could generate transaction costs. With standards, however, the spillover of proprietary knowledge from one firm to another is minimised. This would allow the system-integration function to be provided through the market. The integrator would work with the customer to tailor a system; would work with suppliers (itself probably included) to produce the system; and would provide the necessary service guarantees. This means that the integrator would need to have a reputation of value significant enough to act as a hostage (Williamson 1985).

208 RICHARD N.LANGLOIS

The most likely possibility is that a lead equipment maker would act as system integrator. It would work with other manufacturers of similar size, but would alone provide the ultimate guarantees. In the parlance of the industry, this is called taking ownership of the system. In some cases, as in the ongoing collaboration between Novellus and Lam Research (Rice 1991; Holden 1992b), two similar-sized companies might team up and share ownership, so long as the customer can be made satisfied with the arrangement. In the semiconductor equipment industry, as in many other American industries, such cooperative arrangements are becoming a pronounced trend. It is also conceivable that independent third parties might become integrators. Fearing that too much proprietary information would leak across even the anonymous boundaries of the standard architecture, at least one industry official suggested early on that aerospace companies might be appropriate candidates.17 What seems to be happening, however, is that new companies will arise to fill the integrator role. One such is CVC Products, whose Connexion cluster tool offers the buyer a range of options, almost all from thirdparty vendors (Holden 1992a). Another possibility is that the users—the semiconductor manufacturers—might themselves act as system integrators. This is especially likely in the case of large users (like TI and IBM) that have substantial capabilities of their own in processing technology. The emergence of standards The process by which standards are emerging in the cluster tool industry is rather different from those of well-documented cases like the QWERTY keyboard (David 1985; Liebowitz and Margolis 1990), the VHS videocassette recorder (Cusumano et al. 1992), the IBM-compatible personal computer (Langlois 1992), or the 33-rpm LP record (Robertson and Langlois 1992). In all of those cases, standards emerged through a competition or ‘battle of the standards’ among alternatives originally offered as proprietary schemes. A standards battle did once threaten in the cluster tool industry, and such a battle may yet take place. But the origins of the standards in this case were, if not exactly ‘spontaneous’, then at least far more grass-roots and collaborative in character. The Modular Equipment Standards Architecture (MESA) was the result of the work of an ad hoc organisation comprising the bulk of firms in the cluster tool and related industries. The significant exception was Applied Materials, by far the largest firm in the business, whose Precision 5000 platform briefly offered an alternative to MESA.18 Quickly, however, the MESA committee was folded into Semiconductor Equipment and Materials International (SEMI), the equipment makers’ trade group, becoming the Modular Equipment Standards Committee (MESC). Applied Materials is now a member, albeit somewhat grudgingly. And, although Applied Materials’ strategy for its future tools is still unclear, it appears that MESC has indeed been established as the industry standard. The story begins in 1989. Commercial cluster tools had been on the market for only two or three years, but a number of firms, each considerably smaller than Applied Materials, were either in the market or planning to enter (Burggraaf 1989). On 30 March 1989, a group of representatives from several Bay Area companies congregated at a motel

CAPABILITIES AND VERTICAL DISINTEGRATION IN PROCESS TECHNOLOGY 209

in Fremont to begin what would become a rapid-fire series of meetings. Present at the first meeting were representatives of eleven companies,19 including the CEOs of four of those companies.20 In many ways, the cooperation among these firms was a startling change from the individualist go-it-alone culture supposedly characteristic of the industry. From another point of view, however, the cooperation was made possible precisely by the cultural network of Silicon Valley and its web of personal contacts among engineers and marketers in many distinct firms. For example, General Signal Thinfilm Company was something of a cheerleader in the drive towards standards, having pushed the idea since early 1988 through informal discussions and ‘partnering’ efforts with other vendors.21 The ad hoc group adopted the name MESA as its banner, and put forward this mission statement: ‘Develop technically sound, common, non-proprietary interface standards which the US equipment industry can utilise to individually and collectively offer the best available choice of automated, interchangeable, integrated tools.’22 The group worked feverishly over the ensuing weeks to develop a draft standard. The first goal was to standardise the mechanical interface of future cluster tools, that is, the physical connection between the wafer handler and the modules. This included such parameters as the size and shape of the port and the valve flanges, their height above the floor, and the reach of the robot arm. On 2 and 3 May 1989, Sematech held a technical workshop on cluster tools in Dallas, Texas. Part of Sematech’s mission is to develop industry standards, which were on the agenda for this workshop. The MESA members who attended used the occasion to present the work they had been doing and to offer MESA as an industry standard. Applied Materials also had representatives at the meeting. Quickly, Applied responded to the MESA offer by proposing to ‘open up’ the Precision 5000 architecture as an alternative to MESA (Winkler 1989a). A battle of the standards appeared to be in the offing. Applied’s argument for adopting the Precision 5000 as a standard was that, with an installed base in 1989 of some 300 machines and 1,000 chambers, the 5000 represented a proven technology (Morgan 1989). The MESA standards, Applied argued, were being created in the abstract rather than as a response to customer needs. Proponents of MESA argued back that a standard is necessarily abstract and conceptual, and that the members of the group were in any case bringing to bear a lot of collective experience with customers (Dunn 1989). In Dallas, the MESA committee invited Applied to attend MESA meetings, specifically a congregation later in the month at Semicon West, one of the large industry trade shows that SEMI sponsors around the world. Applied declined, and suggested that it would not become involved in standard setting unless MESA were somehow brought under the umbrella of SEMI. The Semicon meeting attracted considerable attention from other equipment firms, who began joining MESA in large numbers. Sematech, which, as a quasigovernmental organisation, could have affected the choice of standard by throwing its weight in either direction, chose neutrality instead (Winkler 1989b). Ultimately, however, the threatened battle of the standards did not emerge. For one thing, MESA and Applied were quickly united institutionally when, at a meeting at Semicon East on 12 September 1989, the MESA group voted to join SEMI, becoming

210 RICHARD N.LANGLOIS

reconstituted as MESC. As a member of SEMI, Applied was effectively a member of MESC and eligible to vote on proposed standards.23 There is also arguably a technical reason why a standards battle did not really emerge. Although the 5000’s mechanical interface could in principle have served as a standard comparable to the MESC mechanical standard, the Applied machine is otherwise a closed, non-modular system.24 Since the goal of MESA was total modularity, including a decentralised communications architecture, the 5000 could have been ‘opened up’ only by being redesigned. MESC has unquestionably become the standard of the non-Applied world. A large number of equipment vendors quickly began offering MESC-compatible systems or modules (Dorsch 1991b). The economic benefits of a standard for a network of small firms proved overwhelming, and Applied’s machine was not a genuine alternative. This was so for technical as well as strategic reasons: Applied’s Precision 5000 was not designed in a particularly modular fashion, and thus could not serve as the basis for a standard architecture. Rather than a battle of the standards, the current situation might best be thought of as a battle of alternative development paths: the closed system of Applied Materials, with its significant internal economies of scale and scope, and the open modular system of the competitive fringe, driven by external economies of standardisation. So far, the market has chosen Applied over MESC, although the latter may be gaining ground. Lam Research, which made its mark in etching equipment before diversifying and which has adhered to MEC standards, is now the fourth largest semiconductor equipment firm in the world (see Table 10.2). The interesting choices will arise when fabs begin the move to 12 inch (300 mm) wafers in the next few years.25 Because of the larger wafer size, the physical dimensions of the mechanical interface—both MESC’s and Applied’s—will have to change. Applied has several options in designing its 300 mm cluster tools. In 1994, Applied agreed to conform to the MESC communications standard—that is, the standard that regularises the hardware and software controlling the modules—but industry observers do not expect the company to adhere to MESC standards for their mechanical interfaces (Dorsch 1994). The most likely possibility is thus that competition will remain a matter of alternative technological trajectories, not of rival standards. To the extent, however, that new Applied machines are sufficiently open systems, it is possible that the company could offer their interface standards as an alternative, creating a genuine battle of the standards. A third, and perhaps least likely, possibility would be for Applied to offer a fully MESCcompatible machine. If the company takes the last route, it would find itself in a position similar to that of IBM in personal computers in 1981, that is, in a position to become the lead player in the provision of a single-standard modular system.26 In a market in which external economies are significant, this can be a strategically desirable alternative (Langlois and Robertson 1992).

CAPABILITIES AND VERTICAL DISINTEGRATION IN PROCESS TECHNOLOGY 211

Implications for competitiveness In talking with industry sources and academic experts, one develops something like the following stylised picture of competition in semiconductors. The Japanese semiconductor industry is built around a commitment to manufacturing. This means mass production; a conservative philosophy that prefers incremental change and reliability to cutting-edge technology; and a large pool of production engineers at the fab level engaged in finetuning the production process. By having appropriated the market for DRAMs, the Japanese have taken the lead in the technology that drives advances in lithography and therefore in manufacturing technique. Apart from Nikon and Canon, which are themselves large multidivisional firms with a core competence in optics, the Japanese equipment industry is largely controlled by the semiconductor firms themselves, either through vertical integration (in the case of Hitachi) or, more usually, through equity positions in what is described as a keiretsu structure. The manufacturers and their suppliers thus work in close coordination (Stowsky 1989). By contrast, the American semiconductor industry has been better at product development than mass production, focusing more on logic and specialty chips and largely abandoning memories. This choice reflects in part the different structure of user demand in the United States, where computer and military applications dominated (Langlois and Steinmueller 1999). The Americans have focused less on manufacturing, and tend to rely more on their suppliers for process expertise. The fragmented equipment industry does not coordinate well with manufacturers, trusting to public information and market forces for its investment decisions. The Japanese package worked well, at least until recently, and it has often been portrayed as a model for Americans to emulate. Although there is no doubt that American semiconductor makers must continue to increase their focus on manufacturing, it is not clear, however, that the Japanese approach has proven a durable and universally applicable model. For one thing, the strategic value of cutting-edge DRAM production is less obvious than it used to be. Most major companies now crank out some form of mass-produced chip as a technology driver, and the DRAM market has thus become a saturated lowprofit one. The Japanese are beginning to feel the downside of the DRAM strategy, especially in view of the rise of Korean and Taiwanese production (Iritani 1996). Samsung is now the world’s largest DRAM producer, and many in the industry feel that Korean firms, with newer and less conservative process technology (much of it American), have gained cost leadership over Japanese firms. Moreover, Americans have improved their manufacturing capabilities, and the focus on product development may have had its positive side. Even if flexible production of specialty chips does not completely replace the mass production of generic chips, the days of the ‘capacity race’ as the key element in semiconductor competition are probably numbered (Steinmueller 1992). And the Japanese have not proved able to crack American dominance in microprocessors, falsifying the predictions to the contrary that were widespread just a few years ago (Ferguson 1985; Reich and Mankin 1986). In addition to being a product with higher value added than DRAMs, microprocessors are also arguably process drivers: while DRAMs do drive

212 RICHARD N.LANGLOIS

lithography technology, microprocessors require more complicated intermetal connection, and tend to drive a number of critical mid-process technologies.27 What does this imply for competitiveness in the equipment industry? Much of the loss of American market share in that industry was the result of the movement of semiconductor production to Japan, where, with the notable exception of Applied Materials, American firms were not well able to compete. To the extent that manufacturing is recaptured in the United States and moves to other parts of the Far East (like Korea, Taiwan or Singapore) where there is no indigenous equipment industry, the American industry may benefit. Japanese manufacturing has been well served by a user-orientated equipment industry willing to tailor systems to precise needs. But one may wonder whether the closeness of that relationship, widely held to have been a great advantage so far, may not turn into something of a disadvantage in a world dominated by an open standard architecture. Japanese manufacturers themselves employ far more production engineers than in a typical American fab, shifting the locus of process knowledge away from the supplier and towards the manufacturer. This may not amount to a ‘deskilling’ of the equipment makers. But there is a perception among Americans that their Japanese counterparts are much more under the thumb of the manufacturers, who dictate profit margins and in some cases use the suppliers as little more than contract assemblers. American firms prefer the autonomy of selling ‘guaranteed’ systems, that is, systems for which the supplier provides the bulk of technical knowledge and service. There is some move in this direction even in Japan, where service contracts (a substitute for close working relationships with suppliers) are on the rise (Dorsch 1991a). Japanese equipment firms, including Hitachi, Anelva, Kokusai, Ulvac, Tokyo Electron and ASM Japan, are developing cluster tool systems of one sort or another, and the Japanese are closely monitoring the development of the MESC standards (Inaba 1989; Achiwa and Shankar 1990). It is certainly conceivable that Japanese equipment firms may become significant players in a future market for MESC-compatible systems. So far, however, the Japanese style has been distinctly proprietary,28 even if the recent turmoil in the Japanese capital markets may make some standardisation more appealing in the future. In any event, however, the traditional relationship between manufacturer and supplier in Japan will not confer any special benefits on the Japanese equipment industry in the world of standards, since a modular system is to a large extent a substitute for proprietary finetuning to user needs. Thus, if American firms can create reliable equipment, and especially if best-of-breed does become an important selling point, the innovativeness and product-development skills of the American industry may become a competitive advantage —if not in Japan then at least in the US and in Asian markets outside Japan. Acknowledgements I gratefully acknowledge the generosity of the Alfred P.Sloan Foundation, who funded this research through the Consortium on Competitiveness and Cooperation and the University of California, Berkeley. Special thanks go to Professor Richard R.Nelson of

CAPABILITIES AND VERTICAL DISINTEGRATION IN PROCESS TECHNOLOGY 213

Columbia for help and encouragement. This research has benefited from discussions with a number of people in industry and academia. I would particularly like to thank Vaughn Akins, Jeff Benzing, Sue Billat, Brian Chapman, Bob Doering, Marie-Thérèse Flaherty, Peter Hanley, Jim Harper, Jackie Marsh, David Reis, Tom Seidel, William Spencer, May Su, David N.K.Wang, Eric Winkler and George Zilberman. John Dunn from General Signal Thinfilm Corporation gave me copious help with both documents and interviews. Paul Robertson, Amy Shuen, Ed Steinmueller and Sam Wood all provided helpful comments on an earlier draft. Marc Langlois put me on to this idea in the first place and provided help along the way. Karl Almond gained me access to the Lurie Library of the Rogers Corporation. And Brad Andrew, Mike Everett, David March and Brian Kench provided research assistance. I alone bear responsibility for any errors, of course. Notes 1 In fact, the systemic or autonomous nature of innovation is neither entirely exogenous nor driven solely by technology. The structure of organisation helps shape the pattern of innovation, which in turn influences the subsequent structure of organisation. In short, a theory of organisational structure is properly part of an evolutionary theory of social institutions (Langlois 1993). 2 Langlois and Savage (1999) refer to these two kinds of standards as coordination standards and normative standards. See also David (1987). 3 Henderson and Clark (1990) discuss the concept of ‘architectural innovation’, in which the components remain the same but the way in which they are put together changes. In a modular system, by contrast, the components change without affecting the integrating architecture. 4 A good basic reference on semiconductor fabrication is Van Zant (1990). For a brief discussion, see Langlois et al. (1988:13–16). 5 This is an analogy one hears often in this industry. Indeed, it is more than just an analogy, as wafer fabrication involves a series of what are basically chemicalengineering processes. 6 A micron is a thousandth of a millimetre. Finer line widths allow more dense packing of a chip. Line widths of 20 microns were typical in the early 1970s, falling to 2 to 4 microns in the mid-1980s, and to less than 1 micron today; 4M DRAMS have line widths of around 0.8 microns, 16M DRAMs require line widths of about 0.5 microns, and 64M DRAMs require widths of 0.33 microns or less. 7 There is an alternative for batch processing in which the wafers are stored in a controlled cassette environment using such gases as nitrogen and argon instead of ambient atmosphere. These techniques are not in widespread use, however. 8 Actually, it is not necessary to process only one wafer in a chamber at a time. So-called semibatch systems can also achieve high uniformity with a continuous-throughput system that processes several wafers at a time. The Novellus Concept One, for example, is a CVD tool with a lazy susan holding seven wafers. It is ultimately a single-wafer system, however, as the wafers are fed in and removed one at a time. Each wafer is exposed to one-seventh of the deposition process at each turn of the carousel, in effect increasing uniformity by averaging. 9 One normally thinks of a simple integrated circuit as like a microscopic printed-circuit board of great complexity. In fact, the most complicated modern chips are like several distinct

214 RICHARD N.LANGLOIS

10 11 12

13

14 15 16 17 18

19

20 21

printed-circuit boards sandwiched together and connected in appropriate places by metal plugs. A 32-bit microprocessor chip, for example, may have four such ‘metalisation layers’. Interview with Peter Hanley, then of Applied Materials, 12 March 1992. Interview with Bob Doering of Texas Instruments, 20 February 1992. In 1994, TI did in fact license large portions of the MMST technology to CVC Products (Anonymous 1994). It is not strictly true that a maker of a closed proprietary system cannot draw on external capabilities. For example, Applied Materials has had agreements with other firms to supply modules that ‘bolt on’ to the Precision 5000. But the closed, centralised and idiosyncratic character of the 5000 makes that sort of transaction more costly than in a world of standards. The parties to such an agreement must transfer far more knowledge to one another than would be necessary with an open modular standard. And some of that knowledge may be proprietary, leading to transaction costs of the hold-up variety popularised by Williamson (1985) and others. Moreover, a significant part of the knowledge generated by the bolt-on firm would be inapplicable to projects with other system manufacturers. Indeed, exactly this situation has arisen in the case of the Precision 5000. Peak Systems has sued Applied over a bolt-on system that the latter contracted for. Peak claims that Applied illegally appropriated Peak’s technology and that they kept Peak’s resources focused exclusively on Applied’s proprietary platform, thus making it virtually impossible for Peak to serve other customers. What makes it possible to produce the 200,000th of a particular model of car as cheaply as (or even more cheaply than) the 100,000th is that most of the knowledge used to produce number 100,000—embedded in the machines, personnel and organisational rules of the firm —can be reused to produce number 200,000. The same is true when the company produces a new model car: much of the knowledge gained on the old model can be reused on the new model. But we call the first case an economy of scale—in the sense of the learning curve if not in the strictly neoclassical sense—and the second case an economy of scope. For an elaboration of some of these ideas, see Langlois (1999). I am indebted to John Dunn for this story. Of course, competition will operate on the quality and product-innovation margins as well, which is just a way to restate the best-of-breed argument. This is the argument of G.Dan Hutcheson, president of VLSI Research, as quoted by Rice (1991). Bob Doering of Texas Instruments, quoted by Newboe (1990). A better historical analogy for the MESA/MESC standards might be the efforts of the Society of Automotive Engineers, led at first by Howard E.Coffin of the Hudson Motor Car Company, to standardise numerous parts used in the early automobile industry (Epstein 1928:41–3). Between 1910 and 1920, the SAE reduced the number of types of steel tubing from 1,600 to 210 and the number of standards of lock washer from 800 to 16. Throughout the initial period of standardisation, until the early 1920s, most interest was shown by the smaller firms, who had the most to gain. The larger firms such as Ford, Studebaker, Dodge, Willys-Overland and General Motors tended to ignore the SAE and relied instead on internally established standards (Thompson 1954:1–11). AG Associates, Drytek, General Signal Corporation, General Signal Thinfilm Company (GSTC), Genus, Lam Research, Matrix, Novellus, Peak Systems, SSI and Silicon Valley Group. Minutes of 30 March 1989 meeting. The CEOs were Ed Dohring of GSTC, Roger Emerick of Lam Research, Robert Graham of Novellus and Tim Stultz of Peak. Private communication with John Dunn, 8 June 1992.

CAPABILITIES AND VERTICAL DISINTEGRATION IN PROCESS TECHNOLOGY 215

22 Talking paper by Jeffrey C.Benzing of Novellus for the Sematech workshop on cluster tools, 2 May 1989. 23 When the MESC mechanical interface standard eventually came to a vote in June 1990, Applied voted against it (Winkler 1990). They cited technical reasons: that the robot’s proposed size and reach would reduce reliability. MESC proponents dispute this. 24 In fact, proponents of the MESA/MESC standards argue that even the 5000’s mechanical interface would have been unsatisfactory for a standard, since the opening through which the wafer passes in the 5000 is too restrictive to accommodate a wide variety of modules. Applied’s own newer tools use a different size opening from that of the 5000 (John Dunn, private communication, 8 June 1992). 25 The move to 300 mm (12 inch) wafers is expected by the turn of the century, and will require a new generation of equipment. 26 IBM’s situation was different only to the extent that, rather than adopting an existing public standard, it created a new bus and operating system. But the bus was in fact almost identical in concept to the S-100 bus already in use. Moreover, the MS-DOS operating system was virtually a clone of the then dominant CP/M operating system, which would have had to have been updated anyway for the move from the 8-bit 8080 chip to the (quasi) 16-bit 8088 microprocessor (Langlois 1992). 27 An advanced 32-bit microprocessor would typically require four levels of metalisation, whereas even the most advanced DRAM needs at most two. 28 One exception is Sony, which acquired Materials Research, a leading American maker of MESC-compatible systems. Materials Research retains considerable autonomy, however, and continues to operate as an American firm (Weinig 1990; Whiting 1991).

References Achiwa, Kunio and Krishna Shankar (1990) ‘Deposition and Cluster Tool Equipment Trends in Japan’, Solid State Technology July: 31–2. Alexander, Christopher (1964) Notes on the Synthesis of Form, Cambridge, MA: Harvard University Press. Anonymous (1994) ‘CVC and TI Agree on CVD Technology’, Surface Modification Technology News 4 (4): April. ——(1996) ‘Semiconductor Equipment Market Notes’, Integrated Circuits International 1 September. Baldwin, Carliss and Kim Clarke (1997) ‘Managing in an Age of Modularity’, Harvard Business Review 75:84–93. Bergendahl, Albert S., David V.Horak, Paul E.Bakeman and David J.Miller (1990) ‘Cluster Tools, Part 2:16Mb DRAM Processing’, Semiconductor International September: 94–102. Best, Michael (1990) The New Competition: Institutions of Industrial Restructuring, Cambridge, MA: Harvard University Press. Burggraaf, Peter (1989) ‘Integrated Processing: The 1990s Trend’, Semiconductor International June: 65–9. ——(1991) ‘MESC: Is It Working?’, Semiconductor International October: 66–70. Carnes, Ross, and May Su (1991) ‘Long Term Cost of Ownership: Beyond Purchase Price’, in Proceedings of the 1991 IEEE/SEMI International Semiconductor Manufacturing Science Symposium. Chandler, Alfred D. (1991) Scale and Scope: The Dynamics of Industrial Capitalism, Cambridge, MA: Harvard University Press.

216 RICHARD N.LANGLOIS

Cox, Brad M. (1990) ‘Planning the Software Industrial Revolution’, IEEE Software November: 25– 33. Cusumano, Michael (1991) Japan’s Software Factories: A Challenge to U.S. Management, Cambridge, MA: MIT Press. Cusumano, Michael, Yiorgios Mylonadis and Richard Rosenbloom (1992) ‘Strategic Maneuvering and Mass-Market Dynamics: The Triumph of VHS over Beta’, Business History Review 66(1): 51– 94. David, Paul A. (1985) ‘Clio and the Economics of QWERTY’, American Economic Review 75:332–7. ——(1987) ‘Some New Standards for the Economics of Standardization in the Information Age’, in Partha Dasgupta and Paul Stoneman (eds), Economic Policy and Technological Performance, Cambridge: Cambridge University Press. Dorsch, Jeff (1991a) ‘Closing the Deal: How the Vendor-Customer Relationship Has Changed’, Electronic News, 20 May. ——(1991b) ‘Cluster Tools Edging Toward Plant’, Electronic News 8 July. ——(1994) ‘Applied Materials Nods to MESC Com Standard’, Electronic News 25 July. Dunn, John D. (1989) ‘MESA Musings, Part II’, Electronic News 10 July. Epstein, R.C. (1928) The Automobile Industry: Its Economic and Commercial Development, Chicago: A.W.Shaw. Erkanat, Judy (1997) ‘Gear Suppliers Show Gains in Down Market’, Electronic News 24 February. Ferguson, Charles H. (1985) ‘American Microelectronics in Decline: Evidence, Analysis, and Alternatives’, VLSI Memo no. 85–284, Microsystems Research Center, Massachusetts Institute of Technology. ——(1990) ‘Computers and the Coming of the U.S. Keiretsu’, Harvard Business Review July– August: 55–70. Florida, Richard and Martin Kenney (1990) The Breakthrough Illusion, New York: Basic Books. Hayek, F.A. (1945) ‘The Use of Knowledge in Society’, American Economic Review 35:519–30. Henderson, Rebecca M. and Kim B.Clark (1990) ‘Architectural Innovation: The Reconfiguration of Existing Product Technologies and the Failure of Established Firms’, Administrative Science Quarterly 35:9–30. Holden, Daniel (1992a) ‘Cluster Tools Enhanced; May Soon Fabricate DRAMs, ASICs’, Electronic News 29 June. ——(1992b) ‘Lam, Novellus Will “Cluster”’, Electronic News 8 August. Inaba, Minoru (1989) ‘Japanese Firms Back New Multiprocessing Systems’, Electronic News 22 May. Iritani, Evelyn (1996) ‘Japan’s Chips Are Down, Not Out’, Los Angeles Times 10 June: D1. Iversen, Wesley R. (1989) ‘Could This Be the Factory of the Future for Low-Volume Chip Production?’, Electronics January. Langlois, Richard N. (1992) ‘External Economies and Economic Progress: The Case of the Microcomputer Industry’, Business History Review 66(1):1–50. ——(1993) ‘Orders and Organizations: Toward an Austrian Theory of Social Institutions’, in Bruce Caldwell and Stephan Böhm (eds), Austrian Economics: Tensions and Directions, Dordrecht: Kluwer Academic. ——(1999) ‘Scale, Scope, and the Reuse of Knowledge’, in Sheila C.Dow and Peter E.Earl (eds), Economic Organization and Economic Knowledge: Essays in Honour of Brian J.Loasby, Cheltenham: Edward Elgar. Langlois, Richard N., Thomas A.Pugel, Carmela S.Haklisch, Richard R.Nelson and William G.Egelhoff (1988) Microelectronics: An Industry in Transition, London: Unwin Hyman.

CAPABILITIES AND VERTICAL DISINTEGRATION IN PROCESS TECHNOLOGY 217

Langlois, Richard N. and Paul L.Robertson (1992) ‘Networks and Innovation in a Modular System: Lessons from the Microcomputer and Stereo Component Industries’, Research Policy 21:297– 313. ——(1995) Firms, Markets, and Economic Change: A Dynamic Theory of Business Institutions, London: Routledge. Langlois, Richard N. and Deborah A.Savage (1999) ‘Standards, Modularity, and Innovation: The Case of Medical Practice’, in Raghu Garud and Peter Karnøe (eds), Path Creation and Dependence, Hillsdale, MI: Lawrence Erlbaum Associates. Langlois, Richard N. and W.Edward Steinmueller (1999) ‘The Evolution of Competitive Advantage in the Worldwide Semiconductor Industry, 1947–1996’, in David C.Mowery and Richard R.Nelson (eds), The Sources of Industrial Leadership, New York: Cambridge University Press. Lazonick, William (1991) Business Organization and the Myth of the Market Economy, New York: Cambridge University Press. Liebowitz, S.J. and S.E.Margolis (1990) ‘The Fable of the Keys’, Journal of Law and Economics 33:1– 25. McLeod, Jonah (1990) ‘The Tungsten Battle’, Electronics November: 29–30. McNab, Keith T. (1990) ‘Cluster Tools, Part I: Emerging Processes’, Semiconductor International August: 59–63. Monkowski, Joseph R. (1991) ‘Issues in Manufacturing: Economical Innovation’, Microelectronics Manufacturing Technology June. Morgan, James C. (1989) ‘MESA Musings’, Electronic News 19 June. Morris, Charles R. and Charles H.Ferguson (1993) ‘How Architecture Wins Technology Wars’, Harvard Business Review March–April: 86–95. National Advisory Committee on Semiconductors (1990) Preserving the Vital Base: America’s Semiconductor Materials and Equipment Industry, Washington, DC: Government Printing Office. Newboe, Betty (1990) ‘Cluster Tools: A Process Solution?’, Semiconductor International July: 82–8. Penrose, Edith T. (1959) The Theory of the Growth of the Firm, Oxford: Basil Blackwell. Piore, Michael J. and Charles F.Sabel (1984) The Second Industrial Divide, New York: Basic Books. Pollack, Andrew (1992) ‘U.S. Chip Makers Stem the Tide in Trade Battles with Japanese’, The New York Times 9 April: A1. Reich, Robert B. and Eric Mankin (1986) ‘Joint Ventures with Japan Give Away our Future’, Harvard Business Review March–April: 78–86. Rice, Valerie (1991) ‘Semi Equipment Makers: Partner or Perish’, Electronic Business 20 May: 34. Robertson, Jack (1991) ‘Sematech Refocusing for CIM, Software?’, Electronic News 29 July. Robertson, Paul L. and Richard N.Langlois (1992) ‘Modularity, Innovation, and the Firm: The Case of Audio Components’, in Frederick M.Scherer and Mark Perlman (eds), Entrepreneurship, Technological Innovation, and Economic Growth: International Perspectives, Ann Arbor: University of Michigan Press. Sanchez, Ron and Joseph T.Mahoney (1996) ‘Modularity, Flexibility, and Knowledge Management in Product and Organizational Design’, Strategic Management Journal 17:63–76 (Winter Special Issue). Schumpeter, Joseph A. (1950) Capitalism, Socialism, and Democracy, 2nd edn, New York: Harper and Brothers. Shankar, Krishna (1990) ‘Cluster Tools: A $2.2 Billion Market by 1994?’, Solid State Technology October: 43–4. Silver, Morris (1984) Enterprise and the Scope of the Firm, London: Martin Robertson.

218 RICHARD N.LANGLOIS

Simon, Herbert A. (1962) ‘The Architecture of Complexity’, Proceedings of the American Philosophical Society 106:467–82, reprinted in H.A.Simon (1981) The Sciences of the Artificial, 2nd edn, Cambridge, MA: MIT Press. Steinmueller, W.Edward (1992) ‘The Economics of Flexible Integrated Circuit Manufacturing Technology: A Framework and Appraisal’, Review of Industrial Organization 7(3–4):327–49. Stowsky, Jay S. (1989) ‘Weak Links, Strong Bonds: U.S.-Japanese Competition in Semiconductor Production Equipment’, in Chalmers Johnson, Laura Tyson and John Zysman (eds), Politics and Productivity: The Real Story of Why Japan Works, Cambridge, MA: Ballinger. Teece, David J. (1980) ‘Economies of Scope and the Scope of the Enterprise’, Journal of Economic Behavior and Organization 1:223–47. ——(1982) ‘Towards an Economic Theory of the Multiproduct Firm’, Journal of Economic Behavior and Organization 3:39–63. ——(1986) ‘Profiting from Technological Innovation: Implications for Integration, Collaboration, Licensing, and Public Policy’, Research Policy 15:285–305. US Department of Commerce (1991) National Security Assessment of the U.S. Semiconductor Wafer Processing Equipment Industry, Office of Industrial Resource Administration, Bureau of Export Administration. US General Accounting Office (1990) Sematech’s Efforts to Strengthen the U.S. Semiconductor Industry, Report to the Committee on Science, Space, and Technology, House of Representatives. GAO report RECD-90–236. Van Zant, Peter (1990) Microchip Fabrication, 2nd edn, New York: McGraw-Hill. Williamson, Oliver E. (1985) The Economic Institutions of Capitalism, New York: The Free Press. Winkler, Eric (1989a) ‘Group, Applied Offer Gear Interface Standards’, Electronic News 8 May. ——(1989b) ‘Fight Grows on IC Process Architecture’, Electronic News 29 May. ——(1990) ‘MESC Link for Cluster Tools Approved; Applied Votes No’, Electronic News 4 June. ——(1991) ‘The Great Tungsten Cook-off’, Upside April: 23–34.

11 Technological capabilities and the strategies of small manufacturing firms The case of Hong Kong Tony F.Yu and Paul L.Robertson

Introduction Although small firms1 predominate in countries throughout the world, they have received little consideration from resource-based theorists. The general emphasis on large firms is understandable in that the weight of each in the economy is often substantial and many of the industries populated by large firms are among the most ‘modern’ in advanced economies. On the other hand, the combined weight of small firms is also substantial. More significantly, small firms may be viewed as the locus of many of the more dynamic features of modern economic growth. Not only do large firms in general begin as small ones, but regions specialising in high-technology products, such as Silicon Valley, frequently have high concentrations of technologically advanced small concerns. Small firms in the ‘Third Italy’ similarly operate at high levels of efficiency by international standards even when they employ lower levels of technology. Finally, small firms dominate the economies of the Newly Industrialised Countries (NICs) of Asia whose growth records have been among the best in the world in recent decades. In particular, while large chaebol have been important in Korea and multinational firms in Singapore, the growth of the economies of Hong Kong and Taiwan has been driven substantially by the success of small firms. But, while the success of Hong Kong’s manufacturing sector in the post-war period is well documented (see, for example, Szczepanik 1958; Riedel 1974; Chen 1979, 1988; Chau 1993), none of these economic studies looks critically at the role of technological capabilities in the firm strategies that form the foundation of Hong Kong’s industrial prosperity. By contrast, in the 1970s and early 1980s, management studies such as those by Espy (1970), Kwok (1978) and Cheung (1982) noted that manufacturing firms in Hong Kong adopted such strategies as small-scale operations, short-term planning horizons, low-capital-intensive production processes and imitation. However, these works predate resource-based and evolutionary theories. Though Chau (1974), Wong (1988a, 1988b) and Redding (1990) attempt to explain the organisational structure and competitive strategies of Hong Kong’s Chinese firms in terms of culture and familism, it is only in recent years that Yu (1997) and Cheah and Yu (1996) have interpreted the competitive strategies of Hong Kong’s manufacturing firms from an entrepreneurial perspective and

220 TONY F.YU AND PAUL L.ROBERTSON

that Robertson and Yu (1997) have analysed the success of small Chinese family firms in terms of trust. Adding to these insights, in this chapter we examine the technological capabilities and competitive strategies of manufacturing firms in Hong Kong. An inquiry into the technological characteristics of small firms necessarily involves consideration of diverse factors. As with their larger counterparts, the degree of technological sophistication shown by small firms varies greatly, from the lowest to the very highest levels. Moreover, the capabilities that a firm possesses in any particular area are valuable only when used in conjunction with other capabilities to which the firm has access (Robertson 1996). This is a matter of great importance to small firms, whose own capabilities are almost by definition narrower than those possessed by larger concerns. Thus the management of technological capabilities by a small firm requires as well the careful use of other capabilities in order to find the balance of the firm’s total capabilities that generates the highest returns. Accordingly, any analysis of the technological capabilities of small firms must also analyse the firms’ overall use of the total resources available to them. We begin with a discussion of the role of technological capabilities in the strategies of small firms in general. Even though most of the small firms in Hong Kong are small family concerns that follow low-to medium-technology strategies, we also examine the use of high technology by small firms in other regions in order to assess the long-run viability of the small family enterprises that now dominate several Asian NICs. We then examine the resources available to Hong Kong’s manufacturing firms, and analyse the relationship between the firms’ resources and the development of their capabilities. Using the electronics industry as an illustration, we demonstrate how small Hong Kong firms have used their capabilities to compete successfully in world markets. Finally, we discuss the changing competitive strategies that Hong Kong firms have used in recent years to adapt actively to changes in their environments. We conclude that dominance by small family firms, in contrast to other types of small firms, may actually retard the rate of technological progress and may therefore be inimical to long-term economic growth in Hong Kong. Flexible low-technology strategies as a means of sustaining growth There is widespread agreement that most Chinese family firms2 engaged in manufacturing activities, whether they are located in China or abroad, tend to follow low-to mediumtechnology strategies (Department of Foreign Affairs and Trade 1995; Berger and Lester 1997; Chiu et al. 1997; Enright et al. 1997; Wu 1997). The many ‘original equipment manufacturers’ (OEMs) of Hong Kong and Taiwan are examples of firms that compete on the basis of cost and quality to turn out goods under contract that bear the brands of other firms, most of them multinationals located in Japan, Europe or North America. OEMs and other small family manufacturing firms with headquarters in Hong Kong have been highly important in generating economic growth both in the former Crown Colony and in the People’s Republic of China (PRC), where firms owned in Hong Kong now employ

TECHNOLOGICAL CAPABILITIES AND SMALL MANUFACTURING FIRMS 221

eight times as many workers as in Hong Kong itself (Berger and Lester 1997:24). Although there are large managerial firms in Hong Kong that use the most up-to-date technologies (Enright et al. 1997:34–40), these are generally in service industries. By contrast, Hong Kong-owned manufacturing firms are widely known for their avoidance of high-technology strategies and their very limited commitment to independent R&D activities. They follow ‘hustle’ strategies which, in the words of Enright, Scott and Dodwell (1997:25), emphasise flexibility in responding at high speed to new and emerging trends. These strategies are aimed at beating competitors to the punch by recognizing new trends more quickly, and capturing a high margin of profit by being the first to respond, but then moving on quickly in response to new developments in the same or in a different industry. As Amsden (1997) has pointed out, this flexibility rests in part on efficient project management, which allows change-overs to be accomplished quickly and cheaply, but more is involved. Firms practising ‘serial mass production’, in which they turn swiftly from one mass-produced line of output to another, also need good sources of information in order to respond quickly to new trends, information that can be gathered through alertness in scanning their environments and good connections with people who operate more closely with the ultimate markets in which the output of the OEMs and other small firms is sold. But, despite their success, Hong Kong manufacturing firms following hustle strategies face some inherent weaknesses. For one thing, they are sometimes too remote from their ultimate customer bases to respond quickly to changing demand. Second, because they frequently cannot use highly dedicated equipment, they must save money on other inputs in order to meet the stringent cost conditions imposed by the larger firms for whom they act as contractors. As noted, some of these savings come from avoiding investments in R&D, but for the most part they entail access to cheap labour and property. There are now serious doubts as to whether hustle strategies can succeed indefinitely. Rising labour and property costs have already driven much manufacturing industry from Hong Kong proper, as firms have instead expanded their operations in Guangdong and other parts of the PRC. As labour costs increase in Guangdong, however, some experts are already beginning to question whether Hong Kong firms will continue to be able to migrate to other, progressively cheaper, regions. To avoid what they perceive as a deadend, the contributors to the Berger and Lester (1997) book, particularly Amsden (1997), have emphasised the need for Hong Kong-owned manufacturers to adopt more sophisticated technology strategies that are less dependent on low cost inputs. Trust, knowledge and the technology strategies of small manufacturing firms Firms that shift to a new technology strategy, however, generally require different capabilities. In order to explore the possibility of small family-owned manufacturing firms

222 TONY F.YU AND PAUL L.ROBERTSON

making the transition successfully, in this section we develop a framework that contrasts the capabilities needed by small family manufacturing firms following hustle strategies with those of small firms undertaking strategies involving the mastery of more sophisticated technologies. Firms as coordinating organisations Before proceeding, it is useful to establish why firms exist. In theory, operations could be conducted through a series of market transactions between independent individuals. In reality, however, a wide variety of organisational forms coexist, in which (depending on circumstances) an organisation combines its internal operations with the purchase of a portion of its inputs and the sale of its outputs—both intermediate and final—in the marketplace. In the tradition of Coase (1937) and Williamson (1985), the choice between internalisation and market-based transactions is determined by the level of transaction costs. When the costs of opportunistic behaviour, bounded rationality, asset specificity or other factors reach a certain level, firms find it cheaper to undertake given operations internally. Otherwise, they rely on the greater efficiency of markets. Important though they may be, transaction costs are not the only factor that causes firms to internalise operations. As resource-based theories have demonstrated (Barney 1991; Conner 1991; Mahoney and Pandian 1992; Peteraf 1993), firms possess collections of capabilities, or competences, which are forms of knowledge that allow specific actions to be undertaken with greater or lesser degrees of efficiency. To an extent, these capabilities may be supplemented with others purchased through the market, but in other cases a particular firm possesses a unique set of capabilities that is inimitable in the sense that the same set is unavailable in the marketplace (Penrose 1959). The irreducible core of a firm is the set of capabilities that is idiosyncratic (i.e., is unique and cannot be purchased) and synergistic, because this permits the firm to operate in ways that other combinations of resources cannot imitate (Langlois and Robertson 1995:9–13). The existence of firms can therefore be attributed to two qualities. One, the level of transaction costs, is largely a function of the presence or absence of trustworthy behaviour in market transactions (Casson 1995; Fukuyama 1995). The second quality is the availability of certain types of capabilities that form the basis of the internalised operating knowledge of the firm. The relative importance of these two qualities depends on the types of activities in which the firm engages. For our purposes, however, we distinguish primarily between two types of knowledge that derive from idiosyncratic synergy within a given firm. The first involves activities that require the sorts of knowledge associated with internalised goodwill trust (that is, a high level of trust among members of the same firm, which we discuss in detail below), while the second centres on activities that require a high degree of sophisticated technological knowledge. Although some internalised goodwill trust and technological knowledge is required by all firms, the importance of goodwill trust is relatively greater for firms that follow hustle strategies that entail a high degree of flexibility in switching products but rely on technological knowledge that is readily available in the marketplace, while innovative3 firms must place a higher degree of

TECHNOLOGICAL CAPABILITIES AND SMALL MANUFACTURING FIRMS 223

emphasis on the synergistic use of technological knowledge. As a consequence, we argue that a non-innovative firm is more likely to be able to utilise family relationships to establish trust within the firm, while firms with the need for more esoteric knowledge must often be based on non-family relationships even though, on the face of it, this could entail the acceptance of a lower level of trust among members of the firm. Entrepreneurship provides another perspective on the relative importance of trust and idiosyncratic synergy. Entrepreneurship in small firms can be usefully divided into two categories.4 The first, which we term ‘Innovative Entrepreneurship’, is closely related to the traits outlined by Schumpeter (1934) in his discussion of innovation and growth. Entrepreneurs in this category base changes upon entirely new concepts in technology or organisation. ‘Arbitraging Entrepreneurship’, on the other hand, derives from the Austrian concept of alertness as developed by Kirzner (1973). In contrast to Innovative Entrepreneurs, Arbitraging Entrepreneurs do not initiate new techniques. Instead, Arbitraging Entrepreneurs make more efficient use of existing resources by creating new combinations of known resources or by bringing together buyers and sellers who might otherwise have been unaware of each others’ existence.5 Both types of entrepreneur contribute significantly to change, but the set of capabilities necessary for each to thrive may be quite different. The knowledge bases of different types of trust In her discussion of trust between suppliers and customers, Sako (1992:37–40) distinguishes among three types of trust. The first, contractual trust, occurs when each of the parties to a transaction ‘adheres to specific written or oral agreements’. This indicates acceptance of a ‘universalistic ethical standard, namely that of keeping promises’. The second category is competence trust, that is, trust in the technical and managerial competence of one’s trading partner. Competence trust holds a higher priority than contractual trust in the sense that a firm would be reluctant to do business with others who lack competence no matter how honest or otherwise reliable they might be. Finally, goodwill trust is defined as ‘mutual expectations of open commitment to each other’, or ‘the willingness to do more than is formally expected’ (pp. 37–8).6 When we add the concept of a firm as a repository of idiosyncratic synergy, Sako’s distinctions also illuminate behaviour within firms. At this level, contractual trust becomes less important, not because the employees of a firm do not need to adhere to specific agreements, but because, particularly among co-owners or partners of small concerns, more is required. The smooth functioning of any firm is impaired if it is necessary to draw up fully specified contracts or there is a high probability of opportunistic behaviour among partners. Instead, a firm should align the incentive structures of its owners so that goodwill trust exists. When this occurs, in the words of Prescott and Visscher (1980:460), a ‘firm is a structure within which agents have the incentive to acquire and reveal information in a manner that is less costly than in possible alternative institutions’.

224 TONY F.YU AND PAUL L.ROBERTSON

Broadly speaking, small manufacturing firms following low-technology, highly flexible strategies need high degrees of both contractual and goodwill trust among the members of the firm. In practice, taken together these two forms of trust amount to a form of knowledge in that members of the firm know to a considerable extent what other members know and also know that they can rely on each other to act in certain ways with only limited amounts of instruction and supervision, at least for common and oftenrepeated activities. Given structural uncertainty and constraints, firms must utilise their stocks of knowledge to solve everyday problems (Berger and Luckmann 1966; Schutz 1967; Berger and Berger 1976). These stocks of knowledge derive from experience. As agents experience and learn, their stocks of knowledge grow. Using this argument, it follows that, when feasible, organisational efficiency can be facilitated if the parties involved possess similar stocks of knowledge. Establishing a network or entering into transactions with others who share the same stock of knowledge is more likely to yield convergent expectations than when there are several distinct knowledge bases. Similarly, Luhmann (1979: 37) has argued that trust relationships, which are a means to tackle uncertainty, find fertile ground in social contexts with the same kinds of culture. Accordingly, there is an incentive for entrepreneurs to initiate a joint venture with their family members because there is considerable overlap in their individual stocks of knowledge and their expectations. In a family, members are raised with the same culture as their parents. In fact, the family is the first and most important point of socialisation for an individual (Berger and Berger 1976:56–76). Furthermore, it may be easier to raise start-up capital from family members. Relatives may be in a better position to assess each other’s abilities than strangers are, as well as more inclined in a sentimental sense to contribute to the welfare of other family members. Thus, it is not surprising that family business is the most significant and universal form of business organisation. Not only are the economic dynamics of East Asian nations said to be built upon family business (Wong 1988a, 1988b; Redding 1990; Whitley 1992), but at least 85 per cent of all businesses in the USA are family owned (Bork et al. 1995).7 In doing business, family members take their experiences for granted and accept them as real. A family firm is therefore, to use Berger and Luckmann’s term (1966), a social construction. The knowledge possessed by members of a family firm forms an important component of the firm’s core capabilities. Although the presence of trust is an important element in operating family businesses, for neoclassical economists who assume perfect competition, trust has no role to play in economic transactions. As a result, they find it difficult to explain satisfactorily the importance of family firms. But, as Arrow (1975: 24) has correctly stated, ‘virtually every commercial transaction has within itself an element of trust. It can be plausibly argued that much of the economic backwardness in the world can be explained by the lack of mutual confidence.’ In his discussion of markets and hierarchies, Williamson (1985: 405– 6) elaborates Arrow’s insight that a society with a higher level of trust requires less vertical integration than one with a low degree of trust, in which opportunism flourishes. Casson (1995:11) argues that, ‘it is in the high-trust culture that complex

TECHNOLOGICAL CAPABILITIES AND SMALL MANUFACTURING FIRMS 225

interdependencies between firms can be sustained by arm’s-length contracts and within each firm the owner can rely on the loyalty and integrity of employees.’ He concludes that it is high-trust, rather than low-trust cultures, that sustain industrial structures based on a large number of small, highly productive, firms.8 This trust, however, takes place in more than one arena and assumes more than one form. As we have shown, trust within an organisation often results from a shared family background, which has generated a common knowledge base. Trust between organisations may involve a wider network of associates. In order to economise on the costs of searching out reliable associates, people have a tendency to deal with those whose characteristics allow one to make certain behavioural assumptions. In the wider world, these may be not family members but people who belong to other affinity groups. In Chinese communities, greater trust is bestowed on alumni of the same school or people from the same village, about whom more is known, than on others who are personally unknown.9 Small firms adopting high-technology strategies, however, require greater amounts of technologically specialised internal knowledge. In the discussion that follows, we argue that, although all firms need both forms of knowledge to some extent, the balance of types of knowledge required by high-and low-technology small manufacturing firms, respectively, tends to differ. This is because the need for internalised esoteric technological knowledge (competence trust) increases directly with a firm’s technological sophistication. For a firm employing higher-level technologies, this overrides to a degree the importance of goodwill trust, which fosters knowledge of the expected behaviour of one’s bosses and subordinates and a high probability that routine activities will be undertaken satisfactorily without high instructional and control costs. Firms operating with established and well-known technologies can place great emphasis on the reduced agency and other transaction costs resulting from higher degrees of internalised goodwill trust, but firms operating closer to the technological frontier must rely on appropriately trained or educated scientific or technical personnel, even when they do not belong to groups traditionally considered to be trustworthy. If the range of technological capabilities that a firm needs to possess is both narrow and relatively common, i.e., if it intends to pursue a low-to medium-technology strategy and to engage in Arbitraging Entrepreneurship, the firm can be organised primarily in order to maximise contractual and goodwill trust internally. As a high degree of goodwill trust is most likely to be found among family members,10 firms that do not need esoteric technological capabilities can most easily and cheaply create a well-functioning team by drawing on the talents of, and giving allegiance to, immediate relatives. This does not imply that such firms lack strong capabilities altogether. On the contrary, they may be highly competitive precisely because they possess important capabilities in non-technological areas. Because their production and other processes are well known and often simple to communicate, firms engaged in Arbitraging Entrepreneurship do not, in general, call upon idiosyncratic technological synergy to a great extent. In fact, centralised commands may be so routine and well understood by others in the firm that they can be delivered tacitly.

226 TONY F.YU AND PAUL L.ROBERTSON

The high levels of internalised goodwill trust that these low-to medium-technology familyowned manufacturing firms have often led directly to enhanced flexibility since directions from the leader of the firm (who is generally also the leader of the family) are followed implicitly. When more advanced technology strategies are pursued, however, the required capabilities are difficult to locate and a high degree of idiosyncratic technology synergy is needed. This is often true in the case of Innovative Entrepreneurship such as hightechnology projects, where competence trust11 becomes overwhelmingly important. Goodwill trust remains valuable, of course, but a high-technology project will generally fail in the absence of suitable technological capabilities even if there are substantial reservoirs of goodwill. As finding people with esoteric talents may be very hard, especially when it is also necessary to establish a synergistic relationship, would-be Innovative Entrepreneurs with a vision, but only part of the required capabilities to realise it, may not themselves possess the resources to put together a suitable firm. This can be further complicated by the need to find financing. Whereas family members or members of other affinity groups may be willing to trust each other because of shared perceptions and interests that are already inter-twined, potential partners who have esoteric talents may not be willing or able also to contribute financial resources.12 In such situations, Innovative Entrepreneurship can be facilitated greatly by the presence of a broader institutional structure, such as the availability of venture capitalists (Bygrave and Timmons 1992), that allows innovators both to raise capital and to gain easier access to potential partners who possess suitable talents. The major thrust of these two types of strategies can be expressed both positively and negatively. From a positive perspective, firms based on the internalised goodwill trust that can derive from close relationships, as in a family, may consciously choose to adopt flexible strategies that make good use of their existing capabilities, even though these are often also low-technology strategies. Similarly, firms with high levels of the appropriate capabilities may consciously adopt high-technology strategies despite the low levels of flexibility that these may permit. To express matters more negatively, it may be the case that a group such as a family, which wants to exploit the advantages arising from its existing high level of mutual goodwill trust, may be obliged by the limited technological knowledge and talents of the group to adopt a relatively low-technology strategy because it does not want to become dependent on the expertise of outsiders. The technology dynamics of small manufacturing firms We have established that, in a static sense, both small family firms and higher-technology small firms can rely on their respective capabilities to generate potentially viable strategies. The question remains, however, as to whether small family manufacturing firms are likely, as a rule, to be able to improve their relative technological positions without undermining their other capabilities. The arguments we have presented suggest that substantial technological upgrading within the context of small family firms may be very

TECHNOLOGICAL CAPABILITIES AND SMALL MANUFACTURING FIRMS 227

difficult because the capabilities needed by the two types of technology strategies are in a sense orthogonal. Small manufacturing firms that depend on hustle strategies and flexibility for their survival require the sort of internalised goodwill trust, based on a common knowledge of the behaviour of other members of the firm, that a family firm provides. Families, however, are unlikely in most cases to have internal access to esoteric technological capabilities upon which more highly sophisticated technologies draw. Firms that successfully follow strategies involving higher technologies generally have multiple founders without any family connections (Robertson and Yu 1997).13 But if outsiders with the requisite technical knowledge are recruited in substantial numbers, as may be needed to improve the general technological competence of the firm, speed and flexibility could be undermined. In other words, switching to a higher-technology strategy could be competence destroying (Tushman and Anderson 1986). The problem that we highlight here is similar to the historical argument that scholars such as David S.Landes and Alfred D.Chandler Jr. have mounted to explain the failure (as they see it) of British and French family firms to expand as rapidly, and to embrace new technologies as enthusiastically, as German and American firms did after 1870.14 Landes and Chandler contend that the unwillingness of families to surrender control over their firms to any significant degree, even to salaried executives, limited their ability to acquire new knowledge and to raise the funds necessary to build capital-intensive manufacturing facilities.15 Similarly, Fukuyama (1995) contends that modern Chinese (and, more contentiously, Italian) family firms are restricted in their size and technological scope by the lack of generalised trust in Chinese (and Italian) society, which limits the circle of people with whom one is willing to collaborate to family members and, perhaps, members of affinity groups. If this is true, then the organisation of Chinese firms along family lines might be considered more of a pis aller than as a strategy based on the positive embrace of the capabilities that arise in family environments. In the absence of a better legal system and other institutions to promote trust among a wider circle of people, Chinese family firms would therefore be highly reluctant to take the risks that the exploitation of more sophisticated technologies could entail. This, in turn, implies that the ability of their economies as a whole to expand would indeed be compromised if firms could no longer gain access to cheap supplies of other inputs to compensate for their general-purpose equipment and lack of technological knowledge. Structural uncertainty and the adaptive response of Hong Kong’s industry The development of manufacturing industry in Hong Kong in recent decades illustrates both the advantages and the potential pitfalls associated with a reliance on internalised goodwill trust and a de-emphasis on technological capabilities. In a small open economy with a unique political structure16 like Hong Kong’s, manufacturers must cope with an extremely volatile economic environment. Because of the high degree of dynamic uncertainty17 that they face, manufacturing firms in Hong Kong need to be alert to change, to adopt strategies and to acquire capabilities that allow them to maintain a high level of

228 TONY F.YU AND PAUL L.ROBERTSON

flexibility and adaptability. One way to achieve organisational flexibility and adaptability is to keep the size of the enterprise small. Therefore, it is not surprising to find that Hong Kong’s manufacturing sector is dominated by small firms, whose activities are largely a response to buyers’ orders, or to orders to act as subcontractors for either their larger counterparts or their peers (Ho 1992:111). A report by the Hong Kong Government Industry Department (1993:17) reveals that the average size of manufacturing firms in Hong Kong fell from approximately fifty-five persons per establishment in 1950 to 14 persons per establishment in 1992. The data clearly show that Hong Kong’s manufacturing sector is dominated by small-and mediumsized establishments. The average size of Hong Kong’s industrial undertakings, which was already small by international standards, actually became smaller during a period of rising prosperity. Although it is sometimes argued that the dominance of small-and medium-sized enterprises is a characteristic of the early stages of industrialisation, this does not seem to have applied in Hong Kong (Ho 1992:112). Whereas in Western business history, Chandlerian firms have competed by developing gigantic corporations (Chandler 1962, 1977, 1990), we suggest that small manufacturing establishments in Hong Kong have become highly competitive in part by remaining small. From firm resources to capabilities In the resource-based literature (for example, see Christensen 1996:113), three general types of firm resources are identified: physical, financial and intangible. Most small factories in Hong Kong employ fewer than ten unskilled workers to produce low-valueadded commodities for export. In terms of physical resources, these firms own no ‘strategic assets’ (Kay 1993:14), since unskilled labour and production plant can be easily acquired in the market. However, they do possess two distinctive intangible assets: (1) organisational resources in the form of the Chinese family business18 and (2) human resources in the form of Hong Kong’s style of entrepreneurship.19 The first factor contributes to the firms’ competence in maintaining flexibility, while the second factor is associated with the firms’ competitive strategies. Together, they constitute a vital part of the capabilities of Hong Kong’s manufacturing sector.20 The combination of small firm size and Hong Kong’s version of Chinese culture has constituted an important part of the resource base of a typical manufacturing firm in the Territory. They contribute to the core capabilities that allow a firm to maintain flexibility, deal with uncertainty and absorb foreign technologies. More importantly, such capabilities, originating as they do in a deep knowledge of Chinese culture, are difficult for overseas firms to imitate: when firms in Hong Kong are closely linked by family and other social ties, it is hard for potential foreign entrants to recognise these ties, thus increasing their costs of gaining access to industry capabilities.21 Internationally, this means that Hong Kong’s industry is protected to a degree by resource-specific barriers. If these capabilities are properly maintained,22 firms in Hong Kong may benefit from the possession of sustainable competitive advantages.

TECHNOLOGICAL CAPABILITIES AND SMALL MANUFACTURING FIRMS 229

Sources of flexibility The flexibility of Chinese manufacturing firms springs from two sources, namely internal and external capabilities. Internal capabilities: the organisational structure of the Chinese family firm

The ability to maintain flexibility is partly attributable to the organisational features of the Chinese family firm. These features can be described as internal capabilities (Robertson 1996:82–4). Wong (1988b:81–2) remarks that ‘in the range of personal relations used by Chinese entrepreneurs to maintain business trust, family ties are particularly important’. In his view, trust among family members can facilitate quick decision making and secrecy, making the family unit adaptable to rapidly changing situations. Furthermore, in terms of communication and internal coordination, members of a Chinese family firm traditionally believe that what is not said is more important than what is said. Though this style of communication requires some guessing and thus increases transaction costs, it does have advantages, notably flexibility, because it permits the senior person in a firm the luxury of unaccounted-for changes of direction and other forms of arbitrary interference (Redding 1990:163). Furthermore, this mode of communication has important implications for competition. Interactions between employees and employers in the form of a paternal relationship enhance mutual understanding and expectation, thereby facilitating the learning process. Codified knowledge can be easily transferred to others with very little training, but tacit knowledge, characterised as know-how, is often difficult to transfer. In this view, tacit knowledge can only be transferred by long periods of interaction in a group. The communication process within the Chinese family has the advantage, therefore, of facilitating the transmission of tacit knowledge. Moreover, such patterns of communication are difficult to imitate, particularly by overseas competitors, and thus again contribute to the sustainable competitive advantage of firms in Hong Kong. External capabilities: subcontracting and flexibility

The integration of small firms into large-scale industrial networks is accomplished through subcontracting and has been significant for the economic development of Hong Kong. Klein (1977:179) notes that subcontracting has important advantages ‘not only for static efficiency, but also for dynamic efficiency. The smaller the units in which competition can take place, the easier it is to enter an industry; and with easier entry, competition is more likely to thrive.’ Subcontracting has been extensively practised in Hong Kong’s manufacturing sector. Sit and Wong (1989:154) report that roughly onethird of the small-and medium-sized manufacturing establishments are linked up with other local factories through subcontracting. In some extreme cases, a ‘hollow’ or ‘network’ firm (Teece and Pisano 1994) may have no production site at all, but only an office for administrative purposes, with all jobs contracted out to others (Yu 1997:77). Most of the firms in Hong Kong have used subcontracting to tackle labour shortages,

230 TONY F.YU AND PAUL L.ROBERTSON

insufficient capacity and seasonal fluctuations, allowing them a higher degree of flexibility and adaptability to changing economic conditions. When engaged in subcontracting, small family firms in Hong Kong utilise personal networks to acquire orders (Sit and Wong 1989:155).23 Little use of impersonal channels like advertising or sales promotion has been recorded. Hence, small firms manoeuvre within a small circle of business friends, rather than needing to reach out to the larger market. Moreover, their business practices reveal a high degree of goodwill trust. Agreements between trading parties are mostly made verbally on an informal basis (Sit and Wong 1989: 154–5). Only one-third of the small-and medium-sized firms reported that their subcontracts were formalised in written documents. By using personal networks, small family firms are able to handle small orders, meet seasonal demands and deliver goods in a very short period (Yu 1997:73–4). This has also enabled small family firms in Hong Kong to gain a degree of competence trust by revealing to overseas buyers that they can efficiently produce unsophisticated labour-intensive commodities. Eriksen and Mikkelsen (1996:18) argue that, ‘if the firm’s core competences are flexible and they have a high productive capacity, then these two dimensions are likely to be positively associated with the value of the firm’. Chinese firms in Hong Kong generally fulfil the first condition. However, in order to maintain their flexibility, they restrict the size of their establishments. As a result, their productive capacity cannot be expanded even if demand for their services increases. The rents of these companies are thus confined to those derived largely from their entrepreneurial alertness and organisational flexibility, rather than from access to economies of scale. Competitive strategies of Chinese family firms We have argued that the Chinese family business system is capable of exercising flexibility, but we have yet to explain and demonstrate how such assets enable these firms to compete successfully in the world market. This process is illustrated in the electronics industry. In Hong Kong, with small firms and very flexible capacity, electronics manufacturers can easily adapt to the changing needs of the market, especially to price slashing warfare and keen competition. For example, they can switch their product lines from electronic calculators to notebook PCs or to cellular phones in a matter of months (Engardio and Gross 1992:67). According to a survey by the Hong Kong Government Industry Department (1991, Section IV-58), about 95 per cent of the electronics firms can develop products in less than twelve months, while 52 per cent of them can do so in less than six months. According to Thomson Lam, the President of Porro Technologies Ltd, the average product life cycle for most of the electronic goods was approximately eighteen months. If Hong Kong’s producers react slowly, they risk being left with stockpiles of unsold items (Wilson 1991:18). Most Hong Kong manufacturers have adopted hustle strategies that are based on a short-term view of production. The importance of flexibility, rapid reactions and shortterm planning horizons is attested to by Mr King-Tin Luk, the founder of Luks Electronics Ltd. According to Mr Luk, in terms of technology, Hong Kong’s small-and medium-sized

TECHNOLOGICAL CAPABILITIES AND SMALL MANUFACTURING FIRMS 231

firms could not initially compete with Japanese TV manufacturers. For large Japanese firms like Hitachi, however, huge capital commitments restricted flexibility and adaptability. By contrast, Hong Kong’s small-and medium-sized firms have been highly flexible and able to adapt quickly to changing market conditions. For instance, given the existing plant structure, they have been able to reshuffle their production lines from making transistor radios to making electronic watches within three months, or from electronic calculators to black and white TVs within six months, or from black and white TVs to colour TVs within nine months (Economic Reporter 1981:19). An international comparison by Sung (1987:48) reports that, while it took Hong Kong producers only three months to introduce a new electronics product, the Japanese took five months and the Americans eight months. Technology strategies: imitation and followership One source of uncertainty faced by firms in a small open economy comes from the emergence of innovative new products or processes in international markets. Bolton (1993:17) remarks that the best way to tackle this kind of uncertainty is to be an imitator. In her view, imitation can be a clever competitive strategy involving investment, creativity and insight. Imitative firms exploit the success of others. Although they have not invented a product or a service, they have perfected and positioned it. Imitators do not aspire to leap-frog their competitors, but are content to follow the innovators at a distance. In other words, imitation is strategic followership (Hagedoorn 1989: 11; Freeman 1982:179; Bolton 1993:12). On many occasions, imitators even purposely delay adopting a new product or practice (Yu 1997:15). Bolton (1993:12) shows, for instance, that Matsushita’s low-cost strategy in the consumer electronics business was built upon being a second-mover. The company deliberately arrived late in the marketplace, waited and watched until consumers accepted a rival’s new product, and only then started to produce a large volume of standardised improved products which it sold at lower prices. Most manufacturing firms in Hong Kong similarly arbitrage profit opportunities through imitation24 (Espy 1970; Kwok 1978; Davies et al. 1993; Cheah and Yu 1996; Yu 1997). In Hong Kong’s electronics industry, most firms are small, have very limited resources and are unable to undertake any original research and development. In 1973, expenditure on R&D in Hong Kong’s electrical and electronics industry was estimated at HK$4,953, 000, comprising only 0.51 per cent of value added (Sit et al. 1979:11).25 More recently, the transfer of technology from overseas through formal arrangements, with the exception of foreign manufacturers who are generally supported by their overseas parent firms, has still not been widely practised by Hong Kong manufacturers. Only five local electronics manufacturers reported having technical agreements with overseas companies (Hong Kong Government Industry Department 1991, Appendix IV-21). Instead, most firms focused on the production of mature and low-value-added consumer products. Cheung’s study (1982) reveals that approximately 70 per cent of the small electronics firms in Hong Kong seldom or never initiated product design. They copied or modified other products rather than designing new models. Similarly in Yu’s survey (1997), nearly 62 per cent of

232 TONY F.YU AND PAUL L.ROBERTSON

the electronics manufacturers copied and modified other products in the market. About 18 per cent of the firms in the survey copied exactly from other manufacturers without any modification. In total, 80 per cent of the respondents pursued an imitative strategy; only 20 per cent of them aimed to design new products. The adoption of hustle strategies using imitative technologies is illustrated by the dramatic development of a local Chinese firm, VTech. According to Allan Wong, the owner of VTech, the company’s strategy has focused on the three catchwords of ‘novel’, ‘quick’ and ‘cheap’. First, VTech has striven to improve product design, creating new models that are much more efficient than the originals from overseas. Second, the company has tailored products quickly to meet market needs. Wong claimed that he could make decisions within minutes and get things started. Third, the improved products were priced more cheaply than their rivals’. Evaluating VTech’s success, Wong remarks that ‘to manufacture a computer is very easy, but to design it cost-effectively is the name of the game’. To implement these three principles, the company constantly reviews the marketplace for emerging trends and opportunities and seeks to ‘adapt leading-edge technologies in order to create exciting and inventive consumer products’. Wong admits that only a minor portion of the R&D money has been spent on new product design, while the bulk has gone into redesigning and improving products as well as making them easier and cheaper to produce. Overall, the company has adopted a defensive technological strategy (Yu 1997, Case 2, Appendix II). Imitation and learning in international subcontracting: the experience of original equipment manufacturers The integration of small firms into the international industrial network, which is significant for the economic development of Hong Kong, is done through subcontracting (Yu 1997). An important form of international subcontracting in Hong Kong has been the Original Equipment Manufacturer (OEM) business. Despite the growth in the design skills of some local firms, OEM business still accounts for a large proportion of total electronics output in Hong Kong (Hobday 1995a:1176).26 The strategy of OEM businesses is best summarised by the statement, ‘We make it, you sell it’ This insight, made by Mr Lap Lee, the founder of Termbray Electronics Co., encapsulates the successful strategy of most Hong Kong manufacturing exporters. This strategy shuns grand marketing plans and multi-million-dollar brand promotions. Instead, local producers let the overseas buyers bear the risk of selling the products. If a firm engages in OEM business, it does not need to promote its own brands. Many of the electronics firms in Hong Kong have no brand products of their own. A survey by the Hong Kong Government Industry Department (Appendix IV-11, 18) showed that, in 1991, 71 per cent of Hong Kong’s electronics producers were either solely or partially involved in OEM business. Only 8 per cent of the firms were solely involved in manufacturing their own branded products. Moreover, manufacturers mainly had their products distributed through overseas buying offices or importers. Not much effort was

TECHNOLOGICAL CAPABILITIES AND SMALL MANUFACTURING FIRMS 233

devoted to marketing activities, as firms took a passive approach to the marketing of their products. This was particularly true for small local electronics firms. Imitation, organisational learning and technological catching up Manufacturing firms in latecomer economies may not behave as pure imitators forever. Cyert and March (1963) stress that a firm is an adaptive rational system. It is possible for firms to adapt actively to their environments through learning. The phenomenon, in which firms in latecomer economies can catch up with first movers, can be explained by the fact that learning and imitation can lead to innovation. Manufacturers do not just imitate others, they pick and choose from different models and combine them into a novel product. This means that imitators are not merely copycats but innovative producers using the examples of others. At first, producers may struggle to imitate the contents of a certain product. Gradually, however, they may take a portion of a design from another product and copy small aspects of the styles or functions of still others. Ultimately, this mixture becomes a single unique product scarcely recognisable as an imitation at all (Phares 1988:153). Explaining how East Asian manufacturing firms have learned to innovate in electronics, Hobday (1995a:1189) notes that ‘latecomer innovation began with incremental improvements to the manufacturing process. As competences were learned, minor innovations to product designs were made and eventually, some new products were offered to the market.’ The improved models manufactured in follower countries often outperform the originals. Having begun by imitation, for example, Japanese manufacturers went on to master Western technology in selected areas. Specifically, after the technological revolution in semiconductors occurred in the US, Japanese manufacturers initiated a phase of adaptive response, encompassing imitation, product improvement and mass production. After a period of time, Japan eventually emerged as an important source of innovations in its own right (Kingston 1977:11). Hence, catching up means that imitating firms eventually sell more sophisticated goods with better designs, greater reliability and longer durability, but at a cheaper price (Woronoff 1986:116). A successful product of VTech—educational toys and telecommunications equipment —illustrates how manufacturers in Hong Kong compete with American and European producers (Yu 1997:191–2). In 1982, Texas Instruments of the US successfully created three popular electronic educational toys to teach children spelling, pronunciation and numbers. Each sold for US$35. Once they appeared on the market, VTech immediately imitated these products. More importantly, it combined all three functions in one toy called ‘Play Tech’, which sold for only US$30. The VTech product soon gained a 60 per cent share of the US market. This case illustrates the advantages of being a latecomer in exploiting commercial potential (Ernst and O’Connor 1989:10).

234 TONY F.YU AND PAUL L.ROBERTSON

The future of low-technology flexible strategies in Hong Kong While some commentators (Enright et al. 1997) are optimistic about the ability of Hong Kong firms to continue to follow successful strategies based on their current capabilities and others that could be acquired relatively easily, other observers are less sanguine. Despite the recent success of small family manufacturing firms in Asian NICs, some analysts are pessimistic about their future prospects. While low-and medium-technology small firms frequently have superior capabilities that allow them to respond with flexibility and speed to changing market conditions, they must rely heavily on factors which not only are beyond their control, but which may even be affected perversely by their very success. The adoption of low-and medium-technology strategies and relatively non-dedicated capital equipment in markets that are highly cost sensitive can force firms to depend on cheap sources of other inputs to maintain their competitive advantage. But the prices of inputs such as labour tend to rise as the number of small firms increases, ultimately undermining their viability. Although they face high labour and rental costs in Hong Kong, manufacturers in the territory can relocate their production to other developing economies to take advantage of cheaper production costs in these regions. Whether these moves can be expected to succeed indefinitely, however, is uncertain. For example, wages in parts of mainland China that have received substantial investment from Hong Kong have already begun to rise quickly (Berger and Lester 1977). Moves further afield are of course possible, but the special cultural affinity that facilitates the operations of Hong Kong firms in other parts of South China would then be absent. Thus, it may not be possible to escape high factor costs forever, in which case low-technology strategies would lose their viability. Even the independence of OEMs and other firms supplying technologically unsophisticated inputs to larger companies may be illusory. Hatch and Yamamura (1996) argue, for example, that firms throughout Asia are likely to find themselves at a permanent disadvantage in relation to the Japanese firms that they supply. Although acting as contractors for technologically advanced Japanese multinationals may lead to increased access to newer technology for firms in Hong Kong, Thailand, Indonesia or China, the technologies of the Japanese firms (their core competences) will advance even more rapidly. As a consequence, Hatch and Yamamura (1996) contend that, rather than catching up, Asian firms outside Japan will become ever more backward in a relative sense. Firms that follow strategies of flexibility may be less severely affected by their inferior positions than will firms that use dedicated equipment and suffer more from excessive asset specificity because the former group is better placed to change from one customer or patron to another as conditions evolve. Nevertheless, small manufacturing firms following hustle strategies remain vulnerable to events beyond their control as long as they rely for their success on low factor costs and on product technology supplied by others. Thus Amsden and other contributors to the Berger and Lester (1997) volume believe that the technological capabilities of Hong Kong firms must be upgraded quickly if the

TECHNOLOGICAL CAPABILITIES AND SMALL MANUFACTURING FIRMS 235

manufacturing sector there is to continue to prosper even with access to cheap labour and property in Guangdong or further afield. To be sure, they recommend not that these firms adopt high-technology strategies, but rather that they position themselves more firmly in the medium-technology range. Among other things, by improving their technological capabilities, small firms would become less dependent on overseas firms for product and process technologies because they could generate more of their technology through their own R&D activities. Even so the question remains as to whether small family manufacturing firms are likely, as a rule, to be able to improve their relative technological positions without undermining their other capabilities. If their strategic flexibility depends on maintaining close personal relationships, then the introduction of new technologies, especially if they are highly capital intensive, may help to break down those relationships because of increased requirements for technical expertise and capital that cannot be met within current family networks. To the extent that family members insist on retaining control and on using existing authority patterns within their firms, they may be reluctant to hand vital functions over to outsiders who have greater technological sophistication or to financial institutions that are willing to provide the funding that is needed. If this is the rule, then a switch from current low-to medium-technology strategies to ones that call upon better developed technological capabilities will be difficult to bring off. Conclusion The technology strategies of small firms are rooted both in their technological capabilities and in their capabilities in other areas. If, as we contend, certain of the capabilities that enable small family firms to function efficiently by adapting quickly to environmental changes may in fact be in conflict with the capabilities needed by high-technology firms, then the ability of small family manufacturing firms in places such as Hong Kong to upgrade their technologies must be impaired. The implications of this problem extend beyond firms, however, to the level of national economies in parts of Asia. Singapore, for example, has adopted a national strategy to encourage operations of multinational firms in order to bring new technologies into the country. As Yizheng Shi (in this volume) shows, the PRC is also encouraging large foreign companies to set up operations at the same time as it is sponsoring the spread of smaller low-technology firms based in Hong Kong. Whether these efforts, and others in Taiwan and in Hong Kong itself, can lead to substantial technological upgrading remains to be seen. If our analysis is correct, the transfer of technology from large, sophisticated firms to the small firms that dominate the indigenous manufacturing sectors in Hong Kong, Taiwan and Singapore is problematical.27 Hatch and Yamamura (1996) attack the argument that Asian development as led by Japan can be likened to a V-shaped formation of geese in which the outliers will eventually catch up to the point that they achieve a straight line. They believe that, if there is any change, the Japanese leaders are pulling further ahead of the others by keeping close control over their modern technologies. East and Southeast Asian subcontractors for

236 TONY F.YU AND PAUL L.ROBERTSON

Japanese multinationals, including family firms, are generally supplied with detailed instructions but otherwise kept in the dark on the technological significance of their work. Other factors, such as those outlined by Fukuyama (1995), include inadequate legal systems that leave property rights unclear in many parts of Asia (although not in Hong Kong or Singapore). The dangers that we have outlined here, however, although also of an institutional nature, cast doubt on the efficiency of flows of knowledge, given the way in which most firms are currently organised. Chiu et al. (1997) suggest, for instance, that the seepage of technological knowledge from multinationals in Singapore to indigenous firms may be slow, and that the two sets of firms operate almost as separate sectors of a dual economy. If Chinese family firms in general prove inadequate to absorb and generate improved technologies, then the policies adopted by governments in the PRC, Hong Kong and Taiwan may need to go well beyond simply making foreign technologies available if they are to achieve the intended modernisation of their economies.28 In addition, these countries will need to foster the growth of new institutions that reduce the risks (and the perceived risks) of interaction among members of their communities, and between their own nationals and foreigners, so that technological capabilities are shared more freely and more widely. Notes 1 As used here, ‘small firms’ are small and medium-sized enterprises with fewer than 300 employees. 2 There are various definitions of ‘family’ firms (Litz 1995). The one we employ covers firms that are managed and financially controlled by members of a family no more distantly related than first cousins. We therefore include patriarchical or matriarchical firms dominated by a founder but employing descendants, firms run by siblings and those run by first cousins. In the latter two cases, the owner-managers are usually children or grandchildren of the founder or founders. 3 As we use it, the term ‘innovative firms’ means firms that undertake the production of goods and/or the use of processes that, while not necessarily at the technological frontiers, involve the acquisition of knowledge that is difficult to master and probably in short supply in the environment at large. By implication, the technological capabilities needed for innovative strategies may be hard, or even impossible, for members of a particular family or narrowly defined affinity group to acquire. 4 This distinction is discussed more fully in Cheah (1992, 1993), where the types of behaviour are termed respectively ‘Schumpeterian’ and ‘Austrian’ entrepreneurship. 5 White (1976:1) has argued that Kirzner does not distinguish arbitrageurship from entrepreneurship. 6 Barney and Hansen (1994) classify trust as weak form, semi-strong form and strong form. Semi-strong form trust corresponds to contractual trust, and strong form trust to goodwill trust, but Barney and Hansen do not present a category that centres on competence as such. 7 We do not contend that the external (environmental) and internal characteristics of firms in Chinese and Western cultures are the same. However, we do contend that, in practice, these characteristics all translate into capabilities and can be compared fruitfully within a resource-based framework.

TECHNOLOGICAL CAPABILITIES AND SMALL MANUFACTURING FIRMS 237

8 But, as Fukuyama (1995) argues, trust can be found at a number of levels. Reliance on trust within narrow circles, such as families or affinity groups, may indicate a low level of societal trust, perhaps owing to the absence of a reliable legal system that allows strangers to deal with each other on a codified basis. 9 Such relationships are, of course, common in many cultures. Chang and Chang (1994), for example, discuss the role of what they call the FARS (family, alumni, regionalism and state/ government) system in the management of Korean chaebol. A similar recognition that it is possible to reduce search costs by dealing with people whom one already knows and has found to be trustworthy underpins Granovetter’s (1985) concept of ‘embeddedness’. In Western nations, groups such as the Quakers have achieved substantial reputations for mutual aid. The contributors to Chirot and Reid (1997) examine the role of Chinese communities in the economic development of Southeast Asia. 10 The likelihood that family members will provide superior goodwill trust is based not only on the higher degree of common outlook often found among relatives, or on the lower search costs that result from being able to draw on people who are already well known. Goodwill trust is also likely to be enhanced because the incentives of family members may be more closely aligned than those of non-relatives. In addition to family honour, family members often use each other as safety nets in case of emergency. If the prosperity of an entire family is dependent on the functioning of a single firm, all of the members of the family will want to ensure that the firm prospers because, as individuals as well as collectively, they will have no good fallback position if the firm fails. 11 Strictly speaking, what innovative situations require is both a high degree of technical and managerial competence, to keep the direct costs of operating low, and a high degree of competence trust, to reduce transaction costs. 12 Although family financing may offer greater control and stability, it is also more risky to the family. Therefore, the acceptance of outside capital may be more than a last resort employed when family funds are insufficient. 13 Recent detailed treatments of the operations of high-technology firms in the United States include Florida and Kenney (1990), Roberts (1991), Rosegrant and Lampe (1992) and Saxenian (1994). Tully (1996) provides a good up-to-date summary. 14 Landes (1949, 1951) and Chandler (1990). Chandler’s contentions concerning the limitations of family firms have recently been challenged by a number of scholars. See, for example, Schumann (1998), the articles in the ‘Special Issue on Family Capitalism’ of Business History edited by Jones and Rose (1995), and in Cassis et al. (1995). 15 Chandler (1990) claims that the development of modern firms in the USA and Germany before 1940 was based on investments in ‘manufacturing, marketing, and management’ that British family firms were reluctant to undertake. It should be noted that his sample concentrates on the 200 largest manufacturing firms in the three countries at various dates, and that the problem that he is attempting to explain is why the largest British firms were so much smaller and were concentrated in different industries than were large American and German firms. 16 According to the British-Sino Agreement of 1984, Hong Kong can retain its capitalist system for fifty years. 17 See O’Driscoll and Rizzo (1985). 18 Chau (1974) argues that manufacturing enterprises in Hong Kong are principally family affairs, and Wong (1988) has used the term ‘entrepreneurial familism’ to denote the dynamics of Chinese family businesses. In Wong’s view, entrepreneurial familism involves ‘the JIA (family) as the basic unit of economic competition’ (Wong 1988a:142–3). The

238 TONY F.YU AND PAUL L.ROBERTSON

19 20 21 22 23 24 25

26 27 28

application of familism extends to financial loans, management technique and marketing strategies. The family provides the impetus for innovation and support for risk-taking. Langlois (1995:255) remarks ‘that entrepreneurship too is an economic capability consisting in part of tacit knowledge’. For a discussion of the concept of industry capabilities, see Foss and Eriksen (1995:43–69). Similarly, extensive corruption in a nation increases transaction costs for foreign companies that are not familiar with the local art of bribery (see Chau 1993:18). The term ‘properly maintained’ means that firms continue to invest in their strategic assets to generate rents. Shared behavioural norms are among the industry capabilities which may yield rents (see Foss and Eriksen 1995:15). This has been termed the ‘Austrian School of Strategy’ (Jacobson 1992) or Kirznerian entrepreneurship (Yu 1997). The exchange rates (US$/HK$) between the US and Hong Kong are: 1973:5.150; 1977:4. 662; 1980:4.976; 1985:7.791; 1990:7.790; 1995:7.736 (Asian Development Bank 1991, Key Indicators of Developing Asian & Pacific Countries, p. 147). For a detailed discussion of the role of OEM in latecomer technological learning, see Hobday (1995b). Similarly the transfer of high-level technology to the relatively dynamic small township enterprises in the PRC is open to question. Wu (1997) and van Kemenade (1997) both look at the PRC, Hong Kong and Taiwan within a single economic context.

References Amsden, A.H. (1997) ‘Manufacturing Capabilities: Hong Kong’s New Engine of Growth?’, in S.Berger and R.K.Lester (eds), Made by Hong Kong, Hong Kong: Oxford University Press. Arrow, K.J. (1975) ‘Gifts and Exchanges’, in E.Phelps (ed.), Altruism, Morality and Economic Theory, New York: Sage. Barney, Jay B. (1991) ‘Firm Resources and Sustained Competitive Advantage’, Journal of Management 17:99–120. Barney, Jay B. and Mark H.Hansen (1994) ‘Trustworthiness as a Source of Competitive Advantage’, Strategic Management Journal 15:175–90 (Special Issue on ‘Competitive Organizational Behavior’). Berger, P. and B.Berger (1976) Sociology: A Biographical Approach, revised edition, Harmondsworth: Penguin. Berger, S. and R.K.Lester (eds) (1997) Made by Hong Kong, Hong Kong: Oxford University Press. Berger, P. and T.Luckmann (1966) The Social Construction of Reality, New York: Anchor Books. Bolton, Michele K. (1993) ‘Imitation versus Innovation: Lessons to Be Learned from the Japanese’, Organizational Dynamics Winter: 30–45. Bork, David, Dennis T.Jaffe, Sam H.Lane, Leslie Dashew and Quentin G.Heisler (1995) Working with Family Businesses, San Francisco: Jossey-Bass. Bygrave, W.D. and J.A.Timmons (1992) Venture Capital at the Crossroads, Boston, MA: Harvard Business School Press. Cassis, Youssef, François Crouzet and Terry Gourvish (eds) (1995) Management and Business in Britain and France: The Age of the Corporate Economy, Oxford: The Clarendon Press.

TECHNOLOGICAL CAPABILITIES AND SMALL MANUFACTURING FIRMS 239

Casson, Mark (1995) Entrepreneurship and Business Culture: Studies in the Economics of Trust, Cheltenham: Edward Elgar. Chandler, A.D. (1962) Strategy and Structure, Cambridge, MA: MIT Press. ——(1977) The Visible Hand: The Managerial Revolution in American Business, Cambridge, MA: The Belknap Press. ——(1990) Scale and Scope: The Dynamics of Industrial Capitalism, Cambridge, MA: The Belknap Press. Chang, Chan Sup and Nahn Joo Chang (1994) The Korean Management System: Cultural, Political, Economic Foundations, Westport, CN: Quoram Books. Chau, L.C. (1993) Hong Kong: A Unique Case of Development, Washington, DC: World Bank. Chau, Sik Nin (1974) ‘Family Management in Hong Kong’ in Lee Nehrt et al. (eds), Managerial Policy, Strategy and Planning for Southeast Asia, Hong Kong: Chinese University of Hong Kong. Cheah, H.B. (1992) ‘Revolution and Evolution in the Entrepreneurial Process’, Proceedings of World Conference on Entrepreneurship, 11–14 August 1992, Singapore. ——(1993) ‘Dual Modes of Entrepreneurship: Revolution and Evolution in the Entrepreneurial Process’, Creativity and Innovation Management 2:241–51. Cheah, H.B. and Tony F.Yu (1996) ‘Adaptive Response: Entrepreneurship and Competitiveness in the Economic Development of Hong Kong’, Journal of Enterprising Culture 4(3):241–66. Chen, Edward K.Y. (1979) Hyper-growth in Asian Economies, London: Macmillan. ——(1988) ‘The Economic and Non-Economics of Asia’s Four Little Dragons’, inaugural lecture, University of Hong Kong, Supplement to the Gazette, 35:21 March. Cheung, Fu Keung (1982) ‘A Study of Management Practices of Small-Scale Electronics Industry in Hong Kong’, MBA thesis, University of Hong Kong. Chirot, D. and A.Reid (eds) (1997) Essential Outsiders: Chinese and Jews in the Modern Transformation of Southeast Asia and Central Europe, Seattle: University of Washington Press. Chiu, S.W.K., K.C.Ho and T.-L.Lui (1997) City States in the Global Economy: Industrial Restructuring in Hong Kong and Singapore, Boulder, CO: Westview. Christensen, J. (1996) ‘Analysing the Technology Base of the Firm: A Multidimensional Resource and Competence Perspective’, in N.J.Foss and C.Knudsen (eds), Toward A Competence Theory of the Firm, London: Routledge. Coase, Ronald H. (1937) ‘The Nature of the Firm’, Economica (N.S.) 4:386–405. Conner, Kathleen R. (1991) ‘A Historical Comparison of Resource-Based Theory and Five Schools of Thought within Industrial Organization Economics’, Journal of Management 17:121–54. Cyert, Richard M. and James G.March (1963) A Behavioral Theory of the Firm, Englewood Cliffs, NJ: Prentice Hall. Davies, Howard, Jenny Ling and Freda Cheung (1993) ‘Product Design and the Location of Production: Strategic Choice in Hong Kong Manufacturing’, Academy of International Business 3: 1–14. Department of Foreign Affairs and Trade (1995) Overseas Chinese Business Networks in Asia, Canberra: Commonwealth of Australia. Economic Reporter (1981) ‘Hong Kong’s Electronics Industry Moves toward High Value Added Production’, 16 September: 19–20 (text in Chinese). Engardio, Pete and Neil Gross (1992) ‘Asia’s High-Tech Quest’, International Business Week 30 November: 64, 67. Enright, M.J., E.E.Scott and D.Dodwell (1997) The Hong Kong Advantage, Hong Kong: Oxford University Press.

240 TONY F.YU AND PAUL L.ROBERTSON

Eriksen, Bo and J.Mikkelsen (1996) ‘Competitive Advantage and the Concept of Core Competence’, in N.J.Foss and C.Knudsen (eds), Towards a Competence Theory of the Firm, London: Routledge. Ernst, Dieter and David O’Connor (1989) Technology and Global Competition: The Challenge for Newly Industrialising Economies, Paris: OCED. Espy, John Lee (1970) ‘The Strategies of Chinese Enterprises in Hong Kong’, unpublished DBA thesis, Harvard University. Florida, Richard L. and Martin Kenney (1988) ‘Venture Capital-Financed Innovation and Technological Change in the USA’, Research Policy 17:119–37. ——(1990) The Breakthrough Illusion: Corporate America’s Failure to Move from Innovation to Mass Production, New York: Basic Books. Foss, N.J. and B.Eriksen (1995) ‘Industry Capabilities and Competitive Advantage’, in C.A.Montgomery (ed.), Resource-Based and Evolutionary Approaches to the Firm, Boston, MA: Kluwer Academic. Freeman, Christopher (1982) The Economics of Industrial Innovation, 2nd edn, London: Frances Pinter. Fukuyama, F. (1995) Trust: The Social Virtues and the Creation of Prosperity, New York: The Free Press. Granovetter, Mark (1985) ‘Economic Action and Social Structure: The Problem of Embeddedness’, American Journal of Sociology 91(3):481–510. Hagedoorn, John (1989) The Dynamic Analysis of Innovation and Diffusion: A Study in Process Control, London: Pinter. Hatch, W. and K.Yamamura (1996) Asia in Japan’s Embrace: Building a Regional Production Alliance, Cambridge: Cambridge University Press. Ho, Yin Ping (1992) Trade, Industrial Restructuring and Development in Hong Kong, London: Macmillan. Hobday, Mike (1995a) ‘East Asian Latecomer Firms: Learning the Technology of Electronics’, World Development 23(7):1171–93. ——(1995b) Innovation in East Asia: The Challenge to Japan, Cheltenham: Edward Elgar. Hong Kong Government Industry Department (1991) Techno-Economic and Market Research Study on Hong Kong’s Electronics Industry 1988–1989, Hong Kong: Hong Kong Government Printers. ——(1993) Hong Kong’s Manufacturing Industries, Hong Kong: Hong Kong Government Printers. Jacobsen, Robert (1992) ‘The Austrian Theory of Strategy’, Academy of Management Review 17:782– 807. Jones, G. and M.B.Rose (eds) (1995) ‘Family Capitalism’, Special Issue of Business History. Kay, J. (1993) Foundations of Corporate Success, Oxford: Oxford University Press. Kingston, William (1977) Innovation, London: John Calder. Kirzner, I.M. (1973) Competition and Entrepreneurship, Chicago: University of Chicago Press. Klein, Burton H. (1977) Dynamic Economics, Cambridge, MA: Harvard University Press. Kwok, Peter (1978) ‘Small Manufacturing Business in Hong Kong’, M.Phil. thesis, University of Hong Kong. Landes, D.S. (1949) ‘French Entrepreneurship and Industrial Growth in the Nineteenth Century’, Journal of Economic History 9:45–61. ——(1951) ‘French Business and Businessmen in Social and Cultural Analysis’, in E.M.Earl (ed.) Modern France, Princeton, NJ: Princeton University Press. Langlois, R.N. (1995) ‘Do Firms Plan?’, Constitutional Political Economy 6:247–61. Langlois, R.N. and Paul L.Robertson (1995) Firms, Markets and Economic Change, London: Routledge.

TECHNOLOGICAL CAPABILITIES AND SMALL MANUFACTURING FIRMS 241

Litz, Reginald A. (1995) ‘The Family Business: Toward Definitional Clarity’, Academy of Management Journal, Best Papers Proceedings, 100–4. Luhmann, Niklas (1979) Trust and Power, New York: John Wiley and Sons. Mahoney, J.T. and J.R.Pandian (1992) ‘The Resource-Based View within the Conversation of Strategic Management’, Strategic Management Journal 13:363–80. O’Driscoll, Gerald P.Jr. and M.Rizzo (1985) The Economics of Time and Ignorance, Oxford: Blackwell. Penrose, Edith T. (1959) The Theory of the Growth of the Firm, Oxford: Blackwell. Peteraf, Margaret A. (1993) ‘The Cornerstones of Competitive Advantage: A Resource-Based View within the Conversation of Strategic Management’, Strategic Management Journal 14:179–91. Phares, E.Jerry (1988) Introduction to Personality, 2nd edn, London: Scott, Foresman and Co. Prescott, E. and Visscher, M. (1980) ‘Organizational Capital’, Journal of Political Economy 88:446– 61. Redding, S.G. (1990) The Spirit of Chinese Capitalism, Berlin: de Gruyter. Riedel, James (1974) The Industrialisation of Hong Kong, Tubingen: J.C.B.Mohr (Paul Siebeck). Roberts, Edward B. (1991) Entrepreneurs in High Technology, New York: Oxford University Press. Robertson, P.L. (1996) ‘Competences, Transaction Costs and Competitive Strategy’, in N.J.Foss and C.Knudsen (eds), Towards a Competence Theory of the Firm, London: Routledge. Robertson, Paul L. and Tony F.Yu (1997) ‘Innovation Strategies and Firm Organisation: Trust, Capabilities and Small Business’, paper presented at the International Australia-Singapore Forum held at the University of Western Sydney Macarthur, 10–12 December. Rosegrant, Susan and David R.Lampe (1992) Route 128: Lessons from Boston’s High-Tech Community, New York: Basic Books. Sako, Mari (1992) Prices, Quality and Trust, Cambridge: Cambridge University Press. Saxenian, AnnaLee (1994) Regional Advantage: Culture and Competition in Silicon Valley and Route 128, Cambridge, MA: Harvard University Press. Schumann, D. (1998) ‘Buddenbrooks Revisited: The Firm and the Entrepreneurial Family in Germany during the Nineteenth and Early Twentieth Centuries’, in P.L. Robertson (ed.), Authority and Control in Modern Industry, London: Routledge. Schumpeter, J.A. (1934 [1961]) The Theory of Economic Development, New York: Oxford University Press. Schutz, Alfred (1967) The Phenomenology of the Social World, Evanston, IL: North-western University Press. Sit, Victor and Siu-Lun Wong (1989) Small and Medium Industries in an Export-Oriented Economy: The Case of Hong Kong, Centre of Asian Studies, University of Hong Kong. Sit, Victor, Siu-Lun Wong and Tsin-Sing Kiang (1979) Small Scale Industry in a Laissez-Faire Economy: A Hong Kong Case Study, Centre of Asian Studies, University of Hong Kong. Sung, Yung-Wing (1987) ‘Flexibility and Hong Kong’s Competitiveness’, Hong Kong Economic Journal Monthly 119 February. Szczepanik, Edward (1958) The Economic Growth of Hong Kong, London: Oxford University Press. Teece, David J., and Gary Pisano (1994) ‘The Dynamic Capabilities of Firms: An Introduction’, Industrial and Corporate Change 3(3):537–56. Tully, Shawn (1996) ‘How to Make $400,000,000 in Just One Minute…’, Fortune International May 27:74–6. Tushman, M.L. and P.Anderson (1986) ‘Technological Discontinuities and Organizational Environments’, Administrative Science Quarterly 31:439–65. Van Kemenade, W. (1997) China, Hong Kong, Taiwan, Inc.: The Dynamics of a New Empire, New York: Knopf.

242 TONY F.YU AND PAUL L.ROBERTSON

White, L.H. (1976) ‘Entrepreneurship, Imagination and the Question of Equilibrium’, reprinted in S.Littlechild (ed.) Austrian Economics, vol. III, Cheltenham: Edward Elgar. Whitley, Richard (1992) Business Systems in East Asia: Firms, Markets and Societies, London: Sage Publications. Williamson, Oliver E. (1985) The Economic Institutions of Capitalism, New York: The Free Press. Wilson, Joel (1991) ‘Hong Kong Board Maker Sets Ambitious PC Goals’, Electronics Business Asia April:37–8. Wong, Siu-Lun (1988a) ‘The applicability of Asian family values to other socio-cultural settings’ in P.Berger and M.Hsiao (eds), In Search of an East Asian Development Model, Oxford: Transaction Books. ——(1988b) Emigrant Entrepreneurs: Shanghai Industrialists in Hong Kong, Hong Kong: Oxford University Press. Woronoff, Jon (1986) Asia’s ‘Miracle’ Economies, New York: M.E.Sharpe. Wu, C. (1997) ‘Hong Kong and Greater China: An Economic Perspective’, in W.I. Cohen and L.Zhao (eds), Hong Kong under Chinese Rule: The Economic and Political Implications of Reversion, Cambridge: Cambridge University Press. Yu, T.F. (1997) Entrepreneurship and Economic Development in Hong Kong, London: Routledge.

12 Technological capabilities and foreign direct investment in China Yizheng Shi

Introduction The desire for international technology transfer was probably the most important consideration when China’s leadership adopted its open door policy in the late 1970s. However, this was not the first attempt of China to acquire advanced technology from abroad. In fact, since the establishment of the People’s Republic, the acquisition of foreign technology has been the locomotive of China’s drive towards industrialisation that is the top priority in her economic development strategy. In the past, however, technology acquisition by China was exclusively through the channel of arm’s length transactions. Before the reform of 1979, technology transfer in China was confined to imports of foreign technology by state owned enterprises (SOEs) (Zhao 1995). The channel of transferring technology through foreign direct investment firms did not exist, since foreign ownership of the means of production in China was contrary to the ruling ideology at that time. The attitude of China’s leadership towards foreign investment has changed dramatically since the adoption of the open door policy. The key component of the open door policy has been the encouragement of foreign direct investment (FDI). The underlying motivation for this policy is to gain access to advanced technologies. For this purpose, preferential policies were formulated by the Chinese government to encourage ‘technologically advanced’ foreign investment by transnational corporations (TNCs) from developed countries (UNCTC 1988). It is hoped that TNCs from developed countries will transfer state of the art technology to China through their subsidiaries. In reality, FDI ventures in the manufacturing sector during this period can be divided into two groups. The first mainly includes subsidiaries set up by large TNCs from developed countries. This type of FDI firm accounts for only a small proportion of foreign firms in China. The second group of FDI ventures consists of numerous small firms, which are mainly investments made by firms from the newly industrialised economies (NIEs) of East Asia, particularly Hong Kong (Huang and Shirai 1994; Hsieh 1994). The purpose of this chapter is to analyse the relationship between firm-specific technological advantages and the foreign direct investment strategies of firms with different technological capabilities. Specifically, the aims of the chapter are: (1) to identify

244 TECHNOLOGICAL CAPABILITIES AND FOREIGN DIRECT INVESTMENT IN CHINA

and discuss the advantages of firms with different technological capabilities in investing in China as a developing country; and (2) to analyse how technological capabilities of firms affect their foreign investment activities to exploit local conditions better in such areas as resources, market demand and government policies in developing countries. Since FDI in China made by Hong Kong firms constitutes the largest part of FDI from NIEs, in the analysis in the following sections Hong Kong firms will be used to represent small firms from NIEs in general. A theoretical framework of foreign direct investment to developing countries In the literature, foreign direct investment (FDI) and international technology transfer into developing countries are interwoven with the theory of transnational corporations (TNCs) of developed countries (Chen 1993). This is mainly because TNCs in developed countries produce and possess nearly all of the modern technology that is the most important advantage of TNCs in expanding their operations overseas (Allouche and Pogorel 1995). Owing to the different nature of technology from conventional commodities, the behaviour of TNCs in their international operations is normally analysed on the basis of imperfect market competition and transaction costs economics. The focus of analysis is on the factors that determine the strategies of TNCs of developed countries in expanding their businesses into developing countries (Graham 1995; Chen 1996). In international production theory, much of the analysis of FDIs centres on the ability and incentives of TNCs to invest in developing countries. The works of Caves (1982) and Dunning (1981) may be taken as representative of the theory of the determinants of FDI and of the international involvement of TNCs from developed countries. In these works, the resources and market potential of host countries (location-specific advantages) provide the primary incentive for TNCs to invest in developing countries. It is the state of the art technology owned by TNCs that underlies their ability to exploit these location-specific advantages by producing abroad even though TNCs face many disadvantages, caused mainly by distance, when investing in developing countries. The advantages of TNCS have been classified under two headings. The first is the technological advantage that derives from their proprietary technologies, such as patented technology and tacit know-how. The second group comprises synergistic advantages including vertical and horizontal production integration, economies of scale and internal financial markets. Synergistic advantages are mainly due to the large size of TNCs. Following on from this approach, it is understandable that FDI is used by TNCs as the channel to carry out international production. FDI internalises the process of technology transfer across borders, so that the risk of leaking advanced technology to potential competitors can be reduced and synergistic advantages can be fully exploited (Jalilian 1996). This approach is useful in explaining the FDI activities of large TNCs from developed countries. They are large conglomerate firms enjoying the benefits of globalised production, and they possess most advanced technological capabilities that they do not

YIZHENG SHI 245

want to leak to others. However, the case of FDI in China on the part of small firms from NIEs and firms from Hong Kong in particular is very different. First of all, these small firms normally do not possess advanced technological capabilities, but tend instead to use standardised technologies. Therefore, they have no obvious technological advantages, especially compared with large TNCs. Furthermore, according to international production theory, firms that use standardised technologies and that are small tend to use externalised channels to license out their technologies to earn rents (Saggi 1996; Narula 1996; UNCTC 1988:13; Davidson and McFetridge 1985). Second, because they are small in size in comparison to large TNCs, there are fewer and weaker synergistic advantages that can be exploited by these small firms. Therefore, the approach applied to the international production of TNCs is not very adequate to explain the factors affecting the investment decisions and performance of small NIE firms. In fact, firms from NIEs, especially those from Hong Kong, are more willing to invest in mainland China than are TNCs from industrial countries, although they do not possess a comparative advantage in advanced technologies. Because they are less technologically advanced, they do not benefit from the preferential policies formulated by the Chinese government to attract large TNCs from developed countries. Nevertheless, they are normally quite successful. Although accurate data measuring the performance of FDI firms is not widely available owing to the problem of transfer pricing, it can be inferred that they are successful in that they continue to invest and operate in mainland China. Moreover, the small FDI firms have made significant contributions to the economic growth of mainland China. Those provinces that have attracted more FDI from NIEs enjoy a higher rate of economic growth than other provinces. Therefore, small FDI firms must have an advantage in technological capabilities that allows them to utilise locationspecific advantages with investment strategies different from those of the large TNCs of developed countries. Although a large part of the international production literature concentrates on the FDI activities of large TNCs in developed countries, some scholars have discussed the role played by multinational corporations of NIEs in international production. E. Chen (1983) and Thoburn et al. (1990), for example, have analysed the outward investment by Hong Kong firms. They maintain that, when compared to local firms, Hong Kong firms have extensive experience in manufacturing labour-intensive products and marketing them in developed countries. Furthermore, owing to increasing labour costs in Hong Kong, relatively cheap labour in neighbouring areas on the mainland is attractive to them. But although these research findings shed light on the FDI activities of small NIE firms, the authors neither differentiate investments made by large TNCs from developed countries from those by small firms of NIEs nor analyse them in a comparative way. Based on this discussion, one can build a theoretical model to explain foreign direct investment to developing countries made by firms possessing different technological capabilities. In this model, two factors that are not included in traditional international production theory are taken into consideration. The first is the outward investment of firms from NIEs. Foreign direct investment made by firms from the newly industrialised economies to their neighbouring regions has been an increasingly striking phenomenon in

246 TECHNOLOGICAL CAPABILITIES AND FOREIGN DIRECT INVESTMENT IN CHINA

recent years that has made a great contribution to the economic growth in East Asia. The second factor is the government policy of developing countries trying to attract, select and direct inward investment activities to achieve the economic growth goals of the host countries. In this model, there are two sets of variables: the technological and organisational capabilities of investing firms, and the location-specific conditions of host developing countries. These variables are basically the same as the technological, synergistic and locational advantages of TNCs in traditional international production theory (Dunning 1981). However, the investing firms in this model are classified into two groups. The first group includes huge TNCs from developed countries. Their capabilities include state of the art technologies that are produced purposefully with heavy investment in R&D and owned and controlled by the TNCs to give them a competitive advantage over their rivals. Their heavy investments in innovation are motivated by the expectation of monopolistic profits gained from concentrating on specific market niches. These competitive market positions result from their proprietary technology and will not be eroded, at least in the short run. The second group analysed here consists of relatively small firms from NIEs. Their competitive assets are their skills in using standardised technology and their experience in organising labour-intensive production. In modern society, technological change has two sources: invention based on scientific breakthrough, and the improvement and adoption of existing technology to the ever changing production environment based on experience. Although radical innovation of new products and processes based on expensive R&D is the engine of technological change, technology as the solution to production problems means the prompt adoption of specific knowledge which is mainly learned and accumulated through practice (Pavitt 1985:5). Therefore, an important part of technology is differentiated, tacit, ever changing and often location-specific knowledge accumulated in practice. Small firms from NIEs have often gained their skills and experience of using standardised and labour-intensive technology through export-orientated production in the take-off period of the NIEs. The differing technological capabilities of these two groups of firms undertaking foreign investment determine that their investment strategies in developing countries will be different. This is because, with their different technological capabilities, the locationspecific advantages they are able to exploit will be different. They will choose those strategies that will allow their technological capabilities to fit best with local conditions. In the literature, a number of objectives of producing abroad by FDIs have been identified: to acquire natural resources, to exploit local resources of land and labour, to reduce production costs (net of transportation costs), to penetrate local markets by bypassing other possible entry barriers, and to capture economies of scale or benefit from synergistic effects (Dunning 1993). The argument in this chapter concerning investment motives is that investing firms will have one particular objective as their main aim in producing in a particular location, and that this objective must match the special characteristics of their technological capabilities. This is because, in this way, they will be able to utilise fully their technological capabilities

YIZHENG SHI 247

to give them a competitive advantage in operating in a foreign location. With state of the art technology, TNCs have the ability to produce differentiated products which host developing countries cannot make by themselves. Therefore, it is likely that market entry is the main objective of TNCs when investing in developing countries, especially those countries with huge domestic markets. By producing in host developing countries, investing TNCs not only enjoy all the benefits of producing locally, but are also better positioned than rival TNCs in the same business that rely on imports. Their technological capabilities are normally capital intensive and require highly skilled labour, particularly in the manufacturing sector. Consequently, cheap inputs of unskilled labour and land may not be important considerations in their international production strategy. For small firms from NIEs, however, their technological capabilities offer them advantages in using standardised, labour-intensive and sometimes second-hand production facilities to produce goods efficiently for export. Therefore, the exploitation of cheap labour and land to reduce their production costs is probably the primary motive of small firms from NIEs that relocate their production facilities to neighbouring regions. By doing this, they are able to maintain their competitiveness in the world market, which has been gradually eroded owing to the increase of labour and land costs in NIEs since 1980. As their technology is not very advanced and their products are not very sophisticated, they have to compete with local producers for local demand. As a result, conquering local markets in developing countries is probably not the main priority in their foreign investment strategy. On the other hand, they have accumulated capabilities in serving markets in developed countries by such means as responding promptly to changes in demand, being able to imitate and produce new products inexpensively, and having the skill to market their products internationally. These are the distinctive competences of small firms from NIEs. Through establishing production units in developing countries and maintaining marketing and designing functions at the parent companies they are able to internalise and better utilise the advantages arising from these special resources. In order to attract, select and direct inward foreign investment, developing countries usually use various policy measures to provide incentives or disincentives for foreign firms. Common policy measures include low tariffs on capital goods and imported inputs, tax reductions and holidays, foreign currency balance requirements, export subsidies, protection of intellectual property, access to local markets, preference in government procurement, localisation of final product requirements, access to local bank loans and local resources, etc. Developing countries use these policy measures to achieve various and sometimes conflicting objectives such as attracting foreign capital, encouraging technology transfer, promoting exports and generating employment. Government policies of host developing countries towards foreign investment are important to foreign investors, because they affect the cost structure of investing firms (Guisinger 1986). Facing these policy measures, investing firms try to enjoy incentives and avoid disincentives. TNCs from developed countries use their superior capabilities in advanced technologies to exploit policy incentives for technology transfer. In line with their comparative resource advantages, however, small firms from NIEs exploit preferential policies that aim at increasing local employment, promoting exports and speeding up

248 TECHNOLOGICAL CAPABILITIES AND FOREIGN DIRECT INVESTMENT IN CHINA

economic growth. In this sense, different types of technological capabilities possessed by foreign firms are their specific ownership advantages that allow them to take advantage of local market potential and cheap production inputs within the constraints of host country government policies. Foreign firms with different technological capabilities can utilise those local conditions within the policy constraints to generate their location-specific advantages. Foreign direct investment in China In an effort to absorb the internationally mobile factors of capital and technology from abroad, the Chinese government has promulgated a number of laws and regulations concerning foreign investment since the joint venture law in 1979. It has also provided policy incentives, which include mainly tax reductions and domestic market access, to encourage and channel foreign direct investment to priority sectors and in special economic zones. Specifically, the government has formulated special preferential policies to attract foreign firms that are willing to transfer advanced technology or have the ability to export (Liu 1995). As a result, the record of foreign investment since the adoption of the open door policy in 1979 has been quite impressive. From 1979 to 1995, a total of 256, 793 FDI contracts with a total agreed investment of US $395.9 billion were signed. During this period, the actual investment was US $133.2 billion. As Table 12.1 shows, the development of foreign direct investment in China can be divided into three phases. During the period 1979 to 1983, the scale of foreign investment was small. After 1983, owing to the preferential policies, actual investment increased steadily until 1991. After 1991, as the Chinese government further accelerated its pace towards a market economy, foreign investment increased very rapidly. Between 1991 and 1995, 227, 744 investment contracts were concluded, which accounted for 88.7 per cent of the total number of foreign investment contracts since 1981 (SSB 1996:597). The top four sources of foreign investment since 1988 are shown in Table 12.2. It can be seen from the table that Hong Kong and Macao together accounted for the majority of foreign investment. Because of its geographical and cultural proximity and the increase in its domestic labour and land costs, Hong Kong is always the largest investor in China. Another two major sources of foreign investment are Japan and the USA. Their investment in China accounts for 6 to 10 per cent of total foreign investment each year. Since 1990, Taiwan has also become a major investor in the mainland. An examination of the available information on sources of foreign investment suggests that China has not been successful in attracting foreign investment from industrial countries (SSB, various issues). Although Japan, the USA and European countries led the list of foreign investors worldwide, they together accounted for less than one-third of actual foreign investment in China. Foreign investment by industrialised countries in China is very small compared with their total overseas investment. Foreign direct investment in manufacturing in China takes several different forms of contract including equity joint ventures in which the rights and interests that each party

YIZHENG SHI 249

Table 12.1 Foreign direct investment in China (US$ million)

Source: SSB (1996:597).

possesses are proportional to its share of capital; cooperative ventures in which the right and interests are stipulated in the contract; and wholly foreign owned firms. Collectively, they are called sanzi qiye, which literally means ‘the three types of foreign firms’. In addition, there are other more flexible arrangements, such as assembly and processing agreements, and compensation trade. In these flexible forms of foreign investment, the rights and interests of foreign parties are negotiable. Foreign firms normally control the supply of materials and the marketing of final products, while the Chinese party can earn processing fees and acquire equipment without paying foreign exchange in advance. Taken together, the three major types of firms with some degree of foreign investment account for the dominant proportion of FDI in China (SSB 1996:598). Characteristics of foreign investment from Hong Kong Motivations for FDI in China from Hong Kong It is well recognised that FDI in China made by small firms from NIEs is mainly intended to reduce their production costs. This is particularly true for small manufacturing firms from Hong Kong (Thoburn et al. 1990). As a result of rapid growth during the 1970s, firms in Hong Kong have found it desirable to relocate their production facilities to neighbouring regions where the costs of labour and land are much lower, so that they will be able to keep their labour-intensive products competitive in international markets. The economic reform policy of China in the early 1980s opened its door to those firms seeking resources of cheap labour and land on the mainland. Membership in same ethnic group, social and cultural similarities, and kinship links shorten the business distance between Hong Kong and China, even though the latter was isolated from the world market for

250 TECHNOLOGICAL CAPABILITIES AND FOREIGN DIRECT INVESTMENT IN CHINA

Table 12.2 Foreign direct investment in China by country of origin

Source: SSB various issues.

YIZHENG SHI 251

Table 12.3 Hong Kong imports from China (US$ million)

Source: Cheung (1997) Table 1, based on Census and Statistic Department, Hong Kong Review of Overseas Trade, various issues.

three decades. The investment of Hong Kong manufacturing firms in China, and in particular in Guangdong Province that is adjacent to Hong Kong, is mainly in the labourintensive industries that account for most of Hong Kong’s export goods (Eng and Lin 1996). As a result of the influx of FDI by Hong Kong firms to exploit cheap labour and land and thereby prolong the lives of their labour-intensive export businesses, Guangdong, where more than 80 per cent of FDI comes from Hong Kong, has become a huge processing plant for Hong Kong companies. The major sectors that Hong Kong firms have invested in include the traditional labour-intensive export industries such as textiles and garments, plastic goods, electronic goods, watches and toys (Thoburn et al. 1990; Cheung 1997). Almost all of the Hong Kong firms that have invested in the mainland have relocated their production facilities but retained their input sourcing and output marketing functions at their headquarters in Hong Kong. This kind of business arrangement is referred to in Chinese as ‘firms with two ends abroad’. The scale of such processing business is very large: by the end of 1995, there were 16 million people in China working for FDI firms. Among these, nearly one-third were in Guangdong (Zhang 1996). As is shown in Table 12.3, the value of goods made in China and exported by Hong Kong firms to world markets increased very fast; by the end of 1993, nearly 90 per cent of the imports of Hong Kong from China was re-exported. The technological capabilities of Hong Kong manufacturing firms Most Hong Kong manufacturing firms are by no means technologically advanced by the standards set by the Chinese policy makers for foreign investors. According to these standards, only those firms that can produce goods that Chinese local firms cannot produce are technologically advanced (UNCTC 1988:24). The technology used by Hong Kong firms, however, is generally standardised and mature (Davies 1995:688). The goods they produce can mostly be made by local firms as well. Moreover, the equipment they

252 TECHNOLOGICAL CAPABILITIES AND FOREIGN DIRECT INVESTMENT IN CHINA

relocate to China is usually second hand (Wu 1995; Sun 1996; Cheung 1997). What, then, are the technological advantages of Hong Kong firms in their successful operations of FDI on the mainland? The answer to this question must be based on the recognition that technological advantages of firms in international production do not always derive from state of the art technology and patented products generated by expensive R&D activities. Technological advantages of firms may also come from such tacit aspects of technology as the cumulative experience and skills required to adapt specific knowledge to a specific environment. A distinctive capability of many Hong Kong firms is their skill and experience in using and improving general purpose and labour-intensive facilities to produce goods for the market in the developed countries where demand changes very quickly and quality requirements are high (Eng 1996). The most notable feature of Hong Kong manufacturing firms is their small scale. The average employment of an establishment in the manufacturing sector is between 10 and 20 workers (HKID 1993:18–20). In contrast to mass production undertaken by large TNCs in industrial countries, Hong Kong-owned firms normally produce in small batches or fill subcontracting orders (Eng 1996), using standard and general purpose equipment to perform different types of production. Since their technology is mature, and easy and cheap to acquire, entry barriers are low (Davies 1996:688). Consequently, competition is keen and firms have to develop skills in using low-technology equipment efficiently, adaptively and innovatively if they are to survive. Second, not long ago, labour costs in Hong Kong were low relative to capital costs. An abundant labour supply of migrants from the mainland and weak trade unionism in the 1960s contributed greatly to Hong Kong’s economic growth. In 1992, the labourintensive sectors of clothing, electronics, watches, plastic goods, etc. contributed threequarters of the domestic exports of Hong Kong. As a result, Hong Kong firms have gained experience in using unskilled labour (Davies 1996; Eng 1996) and they also are able to control the quality of output by using experienced workers as supervisors. Third, although their equipment is general purpose and their labour force is largely unskilled, Hong Kong manufacturing firms have learned to adjust swiftly to changes in the world market. In the early stage of industrialisation, most Hong Kong manufacturing firms were subcontractors who supplied overseas companies in developed countries. In this way, they learned how to meet requirements for high quality and prompt delivery. This has allowed them to develop the ability to imitate the best foreign practices and forced them to be flexible in order to serve different clients. Consequently, Hong Kong manufacturing firms are now adaptive, flexible and imitative. The skills and experience that Hong Kong manufacturing firms have gained in producing labour-intensive goods for international markets have become valuable assets when they invest in operations on the mainland. Although they do not enjoy the preferential policy measures for ‘technologically advanced FDI’, their specific technological and organisational capabilities enable them to adopt various strategies to take advantage of local resources of labour and land.

YIZHENG SHI 253

The strategies of FDI firms from Hong Kong By relocating production facilities to China, Hong Kong FDI firms are able not only to prevent the competitiveness of their labour-intensive products from eroding, but also to increase their scale of production because of the highly elastic supply of cheap local resources that they can tap. Moreover, the parent companies can gain further through specialisation in product design and marketing. This strategy matches the Chinese government’s policy of encouraging exports through the so-called Hong Kong connection (Panagariya 1995). This allows Hong Kong investors to benefit from preferential policy measures encouraging export-orientated FDI and to enjoy the cooperation of local governments. There are several reasons for the concentration of Hong Kong FDI in southern China. First of all, the policy towards foreign direct investment in southern China is much more flexible than in the north. This flexible policy creates a favourable environment for FDI activities. Furthermore, the four original special economic zones (SEZ) are all located in the south. Among them, three are in Guangdong Province, which is just across the border from Hong Kong. FDI firms in SEZs enjoy even more highly preferential treatment. Second, short distance is very important for outward processing by FDI firms that have ‘two ends abroad’. It greatly reduces transportation costs and makes it easier for them to guarantee delivery times. Finally, most Hong Kong business people are migrants from the mainland, especially from the south. Similarities in dialect and culture and local connections certainly help them in the management of labour-intensive production (Chen 1996; Wu 1995:158). Because of their generally small size and the export-orientated nature of their operations, manufacturing firms from Hong Kong normally seek support from local governments at the bottom end of the bureaucratic hierarchy. As a result of reforms, economic decision making is decentralised to the provincial and municipal levels. Local governments have much more authority over economic activities than before, especially in the case of SEZs. They also control the supply of key resources such as land, bank loans, raw materials and electricity. In addition, the possible local Chinese partners of Hong Kong firms in most cases are small enterprises under the supervision of governments at local level. From the viewpoint of Hong Kong firms, the inexpensive labour and land and flexible policy treatment that they seek can both be obtained at the local level. But while outward processing firms established by Hong Kong companies rely heavily on unskilled labour in production lines and manual jobs to reduce their labour costs, they offer high salaries to attract qualified engineers from SOEs to help them to adapt their technology to the local production environment (Wu 1995:141). This practice succeeds mainly because general purpose and standardised equipment is often used. From the standpoint of the local governments, Hong Kong investment brings to their jurisdictions badly needed capital, management skills, and access to foreign markets. More importantly, these production factors come preassembled as firms, that is as bundles of productive capabilities that enhance economic growth through increases in employment and exports. The contribution of FDI to the economic growth of different regions is

254 TECHNOLOGICAL CAPABILITIES AND FOREIGN DIRECT INVESTMENT IN CHINA

shown by recent increases in the economic power of Guangdong Province, In 1978, when the new policy started, Guangdong was only ranked fifth among provinces for exports and sixth for per capita income. Nowadays, it is ranked first in all important economic indices (SSB 1996). FDI by TNCs from industrial countries The development of FDI by TNCs in China The development of investment in China undertaken by TNCs from industrialised countries since the early 1980s is closely related to their opportunities to enter Chinese domestic markets. The investment of TNCs in China may be divided into three stages. In the first stage, in the first half of the 1980s, only a small number of large TNCs invested, primarily to gain experience. According to a survey undertaken by Z.Wang (1996:10– 15), ten major German companies and twenty major Japanese companies have established 231 subsidiaries or equity joint ventures in China. Among these, only twelve establishments (or 6.2 per cent of the total) were set up before 1987. During this period, there were two important obstacles to the investment of TNCs in China (UNCTC 1988:75–90). First, foreign firms were required to balance their own foreign exchange. Consequently, they were prevented from entering the domestic market of China, since Chinese currency was not convertible at that time. Second, in an environment of underdeveloped legislation regarding foreign investment, many TNCs worried that they would not be able to get sufficient legal protection against arbitrary treatment by local government officials regarding contract approval and enforcement. The later 1980s to the early 1990s was the second stage, during which China further opened up its market to foreign firms and began to adopt a more flexible and pragmatic attitude towards foreign investment by TNCs. The promulgation of the Provisions for Encouragement of Foreign Investment (the twenty-two articles) at the end of 1986 is an indication of this change in attitude. According to the twenty-two articles, foreign firms which are export-orientated or are technologically advanced are provided further preferential treatment by the Chinese government. Specifically, once foreign firms are designated as possessing advanced technology, they can sell their products in the local market as substitutes for similar imported goods. This is the so-called market-fortechnology policy intended to encourage TNCs from industrial countries to set up their subsidiaries in China. As a result, TNCs increased their investment in China. From 1987 to 1992, 16.9 per cent of the total was accounted for by thirty German and Japanese TNCs that set up thirty-nine FDI firms. After 1992, as the Chinese leadership decided to open up the economy further and introduced more reform measures, it became clear that foreign firms would have better access to the local market. As a result, TNCs dramatically increased their investment in China. Among the 231 investment projects of the thirty German and Japanese companies, 180 (or 77.9 per cent of the total) were established during the period from 1993 to 1995.

YIZHENG SHI 255

The main reason for the soaring investment of TNCs after 1992 was that the local currency became partially convertible and the local market was further opened to firms undertaking foreign investment. The strategies of TNCs entering the Chinese market through FDI One of the motivations for TNCs to engage in international production is to gain a substantial market share in the host countries that may not be available through direct exporting (UNCTC 1988:76). This is also true in the case of China. In a survey carried out by Wu (1995:173–7), most foreign firms indicated that entry to the local market was one of their most important objectives in investing in China. Nevertheless, firms from industrial countries also gave the protection of their technological capabilities as an important reason for using the internalised mode of FDI to take advantage of their technology. This consideration also explains why TNCs were reluctant to invest in China in the 1980s, when transferring technology to Chinese state owned enterprises through arm’s length transactions was the major channel for TNCs to exploit their technological advantages (Conroy 1992). The technology imported by China was mainly the standardised technology of assembly lines and production facilities for mature products, especially durable household electrical goods. As a result, Chinese firms still lacked the technological capability to produce key components for these products. More importantly, the imported technologies soon became outdated, since local firms were unable to improve them (Shi forthcoming). In the 1990s, when the Chinese market became more accessible to foreign firms, TNCs began to invest in China on a large scale. Their main objective in setting up subsidiaries in China has been to enter the local market while preventing their technology from leaking to their competitors and local partners. On the other hand, since China is still an economy dominated by the state, especially in the key industrial sectors, the cooperation of state enterprises and central industrial ministries is very important in dealing with the bureaucracy and in permitting TNCs to obtain preferential policy treatment. Therefore, a major strategic reason for TNCs to set up ventures in China is to use fully the market-fortechnology policy. First of all, joint ventures with majority foreign shares or wholly foreign firms are the main forms of FDI firms set up by TNCs, allowing them to exploit their technological capabilities without losing control over them. For example, in the elevator industry, the five largest enterprises in China are all equity joint ventures with foreign majority shares. All products are sold in the domestic market with foreign brandnames, and these firms have captured 60 per cent of the market share. In the pharmaceutical industry, among the thirteen largest joint venture firms, seven are controlled by foreign firms and five are 50 per cent foreign owned. Only one firm is controlled by a Chinese partner (Y.Wang 1996). Second, advanced technology and famous brandnames are the main vehicles for TNCs to gain market share in China. Many TNCs have transferred state of the art technology to their Chinese subsidiaries to compete with local firms. Since the technology used by local

256 TECHNOLOGICAL CAPABILITIES AND FOREIGN DIRECT INVESTMENT IN CHINA

firms began to be imported in the 1980s, the products of TNCs have captured increasingly larger market shares. For example, Motorola has gained more than 70 per cent of the pager and mobile phone market. Although the colour television production capacity of local producers is larger than the national demand, TNCs such as Philips, Samsung and Matsushita have established several ventures and manufacture a large share of the local market for high-grade products. Third, many TNCs have as their joint venture partners Chinese enterprises that imported their technology in the 1980s. These domestic enterprises usually have reasonably substantial technological capabilities and are supported by the Chinese government. Cooperating with their previous technology suppliers allows local firms to gain the ability to produce the key components, which in turn increases their competitiveness in the local market. From the viewpoint of TNCs, these local firms possess all the necessary production inputs except for the technology needed for the key components, which naturally becomes the contribution of foreign firms to joint ventures. For example, Siemens, Sanyo and Matsushita have established many joint venture firms employing this approach (Z.Wang 1996:31). Finally, TNCs usually have good relationships with central industrial ministries which are eager to obtain advanced technology from TNCs to fulfil their technological development goals. Therefore, TNCs usually have the chance to participate in the key technological projects provided for in the industrial plan of the state. Consequently, TNCs are normally the big winners from the market-for-technology policy. This is because, by participating in the state planned technological projects, TNCs get the support of government departments. The Chinese partners are usually large firms under the control of industrial ministries. They are well equipped, have capable technical personnel and possess wide marketing channels. For example, Matsushita has established a joint venture with the Ministry of the Electronics Industry in Dalian by participating in a project to localise VCR tape deck production. Since this project had a high priority in the plan, the contract conditions were preferential and the products were sold to local VCR assemblers (Z.Wang 1996:23–4). Discussion and conclusions This study has compared the FDI strategies of large corporations that own state of the art technology with those of small firms using standardised technology. It argues that both of these types of firms possess their own firm-specific technological capabilities that provide competitive advantage when they invest in China as a developing country. In Table 12.4, the different FDI strategies of TNCs from industrial countries and of small firms from NIEs are compared. The two types of technology possessed by these two groups of firms are very different in nature. When they invest in China, TNCs use as their main resources their famous brandnames and differentiated products based on R&D intensive new technology. They can therefore take advantage of the preferential government policy that aims to attract ‘technologically advanced’ FDI. Because of the state of the art technology that they own, TNCs are in a strong position to negotiate better conditions for their FDI

YIZHENG SHI 257

Table 12.4 Technological assets and FDI strategies of firms

projects in China. Normally, they are permitted to enter the Chinese market through the import substitution scheme. They usually cooperate with central industrial ministries, since their projects are relatively large. In contrast with TNCs from industrial countries, small firms from NIEs such as Hong Kong companies normally do not possess patented technology. They use mature and standardised technology and general purpose equipment. Their technological capabilities are their experience in imitating promptly and producing cost effectively new products and in organising labour-intensive production with unskilled workers. Other advantages include the flexibility and adjustability of their operations, and their cultural and social similarities with the local society. Consequently, when the production costs of their traditional export-orientated businesses increase at home they are able to explore opportunities to relocate their production facilities overseas. This is because they are more efficient than local competing firms owing to their superior technological capabilities. Since FDI firms of this type promote exports that generate foreign exchange and create jobs for unskilled surplus labour, they are welcomed by the host country

258 TECHNOLOGICAL CAPABILITIES AND FOREIGN DIRECT INVESTMENT IN CHINA

government. In the case of China, where the economic system has been decentralised, small foreign firms with labour-intensive technologies usually obtain support from local governments and cooperate with small local firms. The fact that these two groups of foreign investors are operating in different segments of the Chinese economy shows that both their tacit experience in using standardised technology and their ability to generate R&D-intensive new technology are important technological capabilities when firms engage in international production through FDI. China, in common with many developing countries, has underdeveloped rural areas where there is surplus labour and the technological infrastructure is poor. Therefore, labour-intensive technology is probably highly appropriate. More importantly, these areas need capital and management skills to promote economic growth. Relocated second-hand equipment does not represent technological progress if it is not sent to a place where labour productivity and wages are extremely low. When it is relocated to an underdeveloped area, however, and combined with improved management skills, it can increase third-factor productivity dramatically. It is because FDI firms are bundles of standardised technology that they are capable of bringing technological progress to the host country in a cost effective manner. Their skills in using standardised technology and their experience in organising labour-intensive production are the technological resources that Hong Kong firms possess when they compete with local firms. As far as large TNCs are concerned, they have the capabilities to generate and maintain control over R&D intensive technology. Therefore, it is well recognised that their technological advantages are due to their strong positions in the international market for highly differentiated products. That local firms cannot manufacture these products contributes to the technological advantage of TNCs in international production. In the case of China, proprietary technology is an additional advantage of TNCs in their FDI activities. By promising to transfer their advanced technology to local firms, TNCs get concessions on FDI constraints from the government and obtain more favourable terms in comparison to other foreign firms without proprietary advanced technological capabilities. Advanced proprietary technology based on expensive R&D is therefore not the only source of technological advantage of firms engaging in international production through FDI. The tacit aspect of technology, that is, experience and skills in applying technology in specific circumstances, is of equal importance. Small firms from NIEs are relatively weak in generating state of the art technology, but they can use their capabilities in using standardised technology in a cost effective way. Of course TNCs also possess the tacit part of technology that helps to generate and to use more codified and R&D-intensive technology more effectively. Nevertheless, their major advantages in international production arise from their ownership of advanced technology. From the viewpoint of the host developing countries, the critical issues are how these two different FDI activities will affect local technological development and how host country governments can promote technology transfer through foreign direct investment, given the importance of tacit aspects of technology.

YIZHENG SHI 259

References Allouche, Jose and Gerard Pogorel (1995) Technology Management and Corporate Strategies: A Tricontinental Perspective, Amsterdam, New York and Oxford: Elsevier Science, North Holland. Caves, R.E. (1982) Multinational Enterprise and Economic Analysis, Cambridge: Cambridge University Press. Chen, Chien-hsun (1996) ‘Regional Determinants of Foreign Direct Investment in Mainland China’, Journal of Economic Studies 23(2):18–30. Chen, E.K.Y. (1983) ‘Multinational from Hong Kong’, in S.Lall (ed.), The New Multinationals: The Spread of Third World Enterprises, London: John Wiley and Sons. ——(1993) Transnational Corporations and Technology Transfer to Developing Countries, London: Routledge. Cheung, Gui-yin (1997) Hong Kong’s Outward Processing Investment in China: Its Implications on Hong Kong Economy, Working Paper Series no. 47, Center for Asian Pacific Studies, Lingnan College, Hong Kong. Conroy, A. (1992) Technological Change in China, Paris: OECD. Davidson, W.H. and D.G.McFetridge (1985) ‘Key Characteristics in the Choice of International Technology Transfer Mode’, Journal of International Business Studies 16, Summer: 165–75. Davies, Howard (1995) ‘High IQ and Low Technology: Hong Kong’s Key to Success’, Long Range Planning 29(5):684–90. Dunning, John (1981) International Production and the Multinational Enterprise, London: Macmillan. Dunning, John (ed.) (1993) The Theory of Transnational Corporations, London: Routledge. Eng, I. (1996) ‘Flexible Production in Late Industrialization: The Case of Hong Kong’, Economic Geography 72:26–43. Eng, I. and Y.Lin (1996) ‘Seeking Competitive Advantage in an Emergent Open Economy: Foreign Direct Investment in Chinese Industry’, Environment and Planning A 28:1113–38. Graham, Edward M. (1995) Foreign Direct Investment in the World Economy, International Monetary Fund Working Paper 95/59:22. Guisinger, Stephen (1986) ‘Host Country Policies to Attract and Control Foreign Investment’, in John Dunning (ed.) (1993), The Theory of Transnational Corporations, London: Routledge. Hsieh, Wen-jen (1994) ‘Technology Transfer to China: A Path to Growth Different from Japan’s’, Rivista Internazionale di Scienze Economiche e Commerciali 41(9): 797–812. Huang, Dongpei and Sayuri Shirai (1994) Information Externalities Affecting the Dynamic Pattern of Foreign Direct Investment: The Case of China, International Monetary Fund Working Paper 94/44: 18. Jalilian, H. (1996) ‘A Theory of Foreign Investment: Possibility, Modes and Timing’, Managerial and Decision Economics 17(3):331–7. Liu, Xiangdong (ed.) (1995) Zhongguo Duiwai Jingji Maoyi Zhengce Zhinan (Guidebook of China Foreign Economic and Trade Policies), Beijing: China Economic Management Press. Narula, Rajneesh (1996) ‘Multinational Investment and Economic Structure: Globalisation and Competitiveness’, Studies in International Business and the World Economy, vol. 4, London and New York: Routledge. Panagariya, Arvind (1995) ‘What Can We Learn from China’s Export Strategy?’, Finance and Development 32(2):32–4. Pavitt, K. (1985) ‘Technology Transfer among the Industrially Advanced Countries: An Overview’, in N.Rosenberg and C.Frischtak (eds), International Technology Transfer, New York: Praeger.

260 TECHNOLOGICAL CAPABILITIES AND FOREIGN DIRECT INVESTMENT IN CHINA

Saggi, Kamal (1996) ‘Entry into a Foreign Market: Foreign Direct Investment versus Licensing’, Review of International Economics 4(1):99–104. Shi, Yizheng (forthcoming) Chinese Firms and Technology in the Reform Era, London: Routledge. SSB (State Statistics Bureau) (1996) China Statistics Yearbook 1995, Beijing: China Statistics Publishing House. Sun, Nansheng (1996) ‘Sanzi Qiye Jishu de Fenxi yu Duice (The Analysis and Solution to the Problems in Technology Transfer of Foreign Invested Firms)’, Guoji Maoyi Wenti (International Trade Issues) 6:31–5. Thoburn, J., H.M.Leung, E.Chau and S.H.Tang (1990) Foreign Investment in China under the Open Policy, Aldershot: Avebury. Wang, Yonggui (1996) ‘Liyong Waizi zhong yi Shichang huan Jishu Boxi (The Analysis of the Marketfor-Technology Policy in Attracting Foreign Investment)’, Guoji Maoyi Wenti (International Trade Issues) 9:50–3. Wang, Zhile (1996) Zhuming Kuaguo Gongsi zai Zongguo de Touzi (The Investment of Famous Transnational Corporations in China), Beijing: China Economic Press. Wu, Nengquan (1995) Guangdong ‘Sanzi’ Qiye Xiaoji Fengxi (Performance Analysis of FDI Firms in Guangdong), Guangzhou: Zhongshan University Press. Zhang, Jingsheng (1996) ‘Yingjin Waizi Zhanlue (The Strategy to Attract Foreign Investment)’, Guoji Maoyi Wenti (International Trade Issues) 10:11–16. Zhao, Hongxin (1995) ‘Technology Import and Its Impacts on the Enhancement of China’s Indigenous Technological Capability’, Journal of Development Studies April, 31(4):583–603.

Index

3M 121, 135–9, 139–5 Abernathy, W. 153 abstraction: in RBP 23 accumulation 46, 100, 170–9, 182, 186 acquisition: 29, 34, 44–9; company 148–73; competence 148–8; technology 6, 148–73, 242 Acta-Scanner 126 activities: firm 80 adaptability 227, 229 adaptive imaginary representations 36, 42 administrative skills: slack 3 advertising 18–19 after-sales service 122, 125, 127 agency: problem 69 Alchian, A.A. 22, 174, 179 alertness: entrepreneurial 81, 93–6 alliances 34, 64 Alston, L.J. 5 altruism 68 ambiguity: causal 82 Amit, R. 7, 9, 41, 56 amortisation 127–1, 132–6, 136–42 Amsden, A.H. 220, 234 Anderson, P. 153 Andrews, K. 11 Anelva 212 antitrust: analysis 18–19

Applied Materials 195–3, 206–19 architecture 97–19; knowledge 99–2, 102, 108–18; process 99–4, 106–19; product 99–19; technological 193 Arrow, K. 42, 224 Arthur, B. 37 Asia 191, 196, 218, 234; East 223, 232, 235, 242, 245; Southeast 235 ASIC (application-specific integrated circuits) 205–13 ASM Japan 212 assets 120, 170–9, 174–3, 178–7, 182–90, 185, 186–5, 221, 228; accumulation 100, 170; amortisation 127–1, 132–6, 136–42; building 127, 129, 132, 137, 138; complementary 121–5, 126, 128–2, 131– 5, 136–40; creation 127–3, 132–6, 137; definition 121; domain-specific 121–7, 131–6; dynamics 127–7; firm-specific 121–8, 131–6; fission 128–6, 138–2, 143; generic 121–8, 131–6; improvement 127–3, 132–6, 137; innovative 120–48; integrative 121–8, 131–6; knowledge-based 88; rent-yielding 170, 172, 176, 172; tangible 62; transaction-specific 51

261

262 INDEX

Baldwin, C. 193 Barbeschi, M. 166 Barney, J. ix, 5–6, 9, 12, 20–4, 40, 55, 82, 149, 167, 186–4 barriers to entry 18, 70, 187, 246, 251 Barzel, Y. 173, 179 batch processing 197–6, 202, 205 behaviour 68–68; innovative 89; organisational 83 Benjamin, R. 108 Berger, S. 220, 223–3, 234 Berry, C.A. 150 Black, J.A. 56 Boal, K.B. 56 Bolton, M.K. 231 Bower, J.L. 111 brand: loyalty 19 Brumagim, A.L. 56, 80, 83 business: theory of 43, 46–9 Canon 210 capabilities 29, 42–4, 53–5, 77, 81–7, 93, 112– 15, 115, 120, 122–6, 133–7, 138, 144, 153, 177–5, 184, 191–212, 219, 221–1 225, 228, 234; causal ambiguity of 44–7, 47; collective 43–6; core 80; corporate renewal 23; creative 79–2; definition of 7, 121; distinctive 46; dynamic 54, 64, 79, 89; economic 191; exploitation of 33; firm specific 97–1; industry 97, 228; innovative 77–94, 123, 125–9, 129; internal 191, 203, 206; learning 102; personal 92; positional 77–5, 86–92; process 112, 113;

productive 34; rare 77, 81–4, 86; self-diagnostic 106; technological 218–45, 242–68; theory of the firm 29–5, 40–9; valuable 77, 86 capital 154, 223, 246–8, 251, 253, 258; equipment 5; human 62, 176, 183; information 64 Carlsson, B. 51 Casson, M. 237 Caves, R.E. 243 Chandler, A.D. 11, 191, 226–6 change: endogenous 23 Chau, L.C. 218 Cheah, H.B. 218 Chen, E. 244 Cheung, F.K. 218 Cheung, S.N.S. 179 Chicago: price theory 18; University of 18, 53 Chiesa, V. 166 China: People’s Republic of 7, 220, 233–5, 242– 68; see also special economic zones; state owned enterprises Chinese firms 218–9, 227–45 Chiu, S.W.K. 235 Christensen, C. 111, 121, 133, 166 citizens 68 Clark, K. 193 cluster tools 196, 203, 205, 206, 208; flexible 196; integrated 200; modular 196, 200–20; proprietary 201, 204 Coase, R. 31, 171, 177–90, 185–3, 187, 221 Cockburn, I. 83, 88, 94, 149 codes: moral 57, 67, 68–1 coherence: contextual 134, 136–42; corporate 121, 133, 139, 152; dynamic 6, 133–43, 141, 143–8;

INDEX 263

explorative 137–2; local 134–8, 138, 141; technology base 121–5, 133–48 collaboration 31, 44, 150–6, 153, 164–70, 166 collective 42 Collis, D.J. 56, 79–2, 92 Coloplast 129–4 Comanor, W. 83–6, 88 communication 42, 229 competence 2, 51–5, 55, 64, 70–3, 120, 123, 133, 138, 149, 184, 221–1; acquisition 148–8; architectural 83; building 11, 15; component 83; core 2, 6, 139; definition 121; destruction 5; differences 61; enhancement 5; entrepreneurial 58; innovative 128–2; internal 148; organisational 97, 125, 149; perspective 51–6, 62–7, 71, 100, 134, 144; pool 42; technological 148 competencies: see competence competition 29, 33–5, 47, 56–8, 59, 65, 68, 70–4, 80–4, 85, 92, 191, 203, 210–19, 219, 225, 229, 243, 251–2, 255–6; competence based 54, 97, 100; hyper- 2, 5, 23, 82; limits to 13, 20, 82; perfect 19, 55, 58, 64–8, 68–4, 224; see also dynamic routines competitive advantage 5–6, 12–15, 21–3, 23, 29, 33–5, 40–2, 43, 47, 54–6, 58, 59, 64, 71, 79–5, 86–9, 124, 128, 132, 137, 170, 185, 186–5, 244–6, 256; Big Bang Theories of 23; differential 20; equilibrium 14; process 17, 33; sustainable 77, 89, 92, 94, 125, 132, 138; theory 56

competitor 63, 68; actions of 61; alliances with 64; imitation 77, 80–3, 92 complementarity: dynamic 6, 150, 152–63, 164–72 complexity 35, 37, 41, 45, 47, 106, 108, 111, 115 components 102–14, 116 conceptual framework 97 Conner, K. 56 construction: social 5 consumer 55, 57–9; behaviour 61; taste 66, 70 context 42, 43, 100, 108, 111, 134, 136 contractual: perspective 31, 33, 51; theories 171–86, 183–2 Cool, K. 9, 56, 186–4 coordination 101, 108, 171, 178–8, 183, 185– 5, 221–1; of resources 14, 100 core competence approach 2 corporate: coherence 121, 133, 152; control 44; entrepreneurship 83; imagination 115; renewal 2, 23 costs 106; information 16, 18–20, 22, 171, 173, 180, 183; sunk 132; transaction 14, 32, 68, 125, 171–90, 187, 193, 206, 224, 243 Crawford, R.G. 174 creative destruction 79 creativity 33, 42, 79, 128, 176 credit rationing 19 culture: organisational 2 customer 5, 41, 222; base 64; relationships with 33–5; support 111 customs 57

264 INDEX

Cyert, R.M. 232 D’Aveni, R.A. 5, 82 decision: making 35; theory 14 decomposition 103, 106 DeFillippi, R.J. 56, 82 DeltaScan 126 demand 66 Demsetz, H. 9–12, 15, 18–24, 51, 174, 187 deontology 57, 67–9 design: configuration 38–39, 43; costs 106; product 103, 106–9, 154, 231; protocol 109–12, 112; rules 193; technological 38 destruction: creative 79 determinism 81, 84–8; non- 81, 86–9, 89, 92–5 development: internal 16, 148, 150, 152 Dierickx, I. 9, 13, 56, 186–4 differences: systematic, between firms 12–13 diffusion of technologies 4, 7 discovery 85, 88, 89; entrepreneurial 80, 84, 86–9, 91–4 distribution 110, 122, 149, 154, 156, 163–9, 167; costs 106 diversification 14, 120, 127, 144 Dodwell, D. 220 Dosi, G. 37–9, 55, 64, 99 Doz, Y. 121, 166 DRAM (dynamic random-access memory) 195, 197–5, 202, 205, 210–19, 213 Du Pont 138 Dunning, J. 243 dynamic: basis of competition 33; capabilities 54, 64, 79, 89–3; coherence 6, 133–43, 141, 143;

complementarities 6, 150, 152–63, 164– 72; efficiency 101; explorative 132–6, 137, 138–3; factors in RBP 15, 22–4, 77–1; of innovative assets 127–6; of knowledge 47; optimisation problem 15; reproductive 132–6, 137–2, 143; routines 81, 87, 89–6 economic organisation 170–8 economics 15, 23–6, 34, 65, 71; Austrian 7, 12, 23–6, 56, 79–3, 84–88, 92– 6, 171, 173–1, 176, 186; equilibrium 11, 23–6; evolutionary 6–7, 12, 24, 54, 56, 58, 71; heterodox 16; industrial organisation 18; institutional 56; mainstream 16, 24, 79, 87; neoclassical 2, 4, 15–16, 23, 61–3, 64–7, 69, 171, 173, 191, 224; non-neoclassical 15, 22; Schumpeterian 17; standard 2, 15; subjectivist 84, 173, 176–4, 179, 186; transaction cost 68–68; Veblenian 17 economies: of scale 191, 203, 243, 246; of scope 194, 203–12 effects: firm 62, 65, 70; industry 62, 65, 70; learning 20; system 1; team 20 efficiency 14, 54, 229; differential 20 egoism 67–9 Eliasson, G. 51 EMI: CT-scanner 126–30, 136, 138–2 employment contract 173, 180–8, 183 enabling capacity 63 encapsulation 157, 164–3

INDEX 265

endogenous: change 23; creation of resources 3, 6, 23–5; growth 29, 34, 71; innovation 55, 71; organisational learning 55; production of services 16; technological progress 70–4 ends 77, 80, 91; see also outcomes Enright, M.J.E. 220 entrepreneur 51, 59, 85, 87, 94, 148, 154–60, 183, 228 entrepreneurship 11, 17–18, 22–4, 53, 56, 59, 63, 80–3, 83–7, 87, 92, 171, 176, 191, 219, 222, 225, 230; Arbitraging 225; corporate 83, 85, 87, 89–3, 93 entry: barriers to 18, 70, 187, 246, 251 environment 24; problem 1, 3 epistemology 64, 86; evolutionary 36 equilibrium 2, 15–16, 23–6, 69–3, 84–7, 87, 184; approach in RBP 5–6, 9–26, 54, 64, 79; competitive 14; economic 11–12, 22; Nash 22; organisational 16; Pareto-optimal 55; price theory 13; Walrasian 70–3 Eriksen, B. 97, 100, 230 Espy, J.L. 218 essentialism: methodological 33 ethics 67–9, 69; deontological 57, 67–9; teleological 67 Etzioni, A. 68 evolution 2, 5, 85, 93; economic 29; industrial 191; market 92; technological 54, 154, 163 evolutionary:

approach in RBP 6, 9–26, 29, 51–72, 80, 94; discovery process 18; dynamics 47; economics 6–7, 12, 24, 54, 56, 58, 71; epistemology 36; factors in RBP 15; theories of competitive process 33 exogenous: technological progress 70 expectations 85–8, 88, 91–5, 223 experience 35, 46, 86, 88 experimentation 133, 138, 170–8, 178–6, 182– 99, 185, 186; business 44 exploitation 133, 137; of capabilities 33 exploration 33, 133, 137–1 external: resources 4, 97–1, 148; technological environment 2 factor: imperfections 22; market 14, 55, 63, 121; services 32 Fairchild 195 family 222–3, 226, 228–8; firms 219–45 Federal Trade Commission 62 Ferguson, C.H. 193–1 fictional worlds 36 financial success 159, 160, 163 firm, the 29–47, 51–4, 55, 58, 64, 71, 173, 221–1, 232; activities 80; diversity 54, 61; effects 62, 65, 70; entrepreneurial 81; information 68; nature of 29–4, 178–8; objective 68; reputation 82; small 218–45; theories of 29, 51, 54, 173, 184 fixed wage contract 173

266 INDEX

flexibility 17, 22, 89, 104, 116, 191, 205–13, 220, 225–9, 233 foreign direct investment 242–68 formalism: in RBP 23–6 Foss, N.J. 7, 45, 51–5, 55, 61, 64, 71, 94, 97, 100, 121, 133, 166, 172, 174 fragmentation 121, 139, 141–8, 196 Fry, A. 135 Fukuyama, F. 227, 235 functionality 102–7

Government Industry department 227, 231–1; small manufacturers in 6, 218–45 Hughes, T.P. 4 human: motivation 67–68 humanware 150 hustle strategies 220, 222, 226, 230, 231, 234 hyper-competition 2, 5, 23, 82 hypothesis: testing 38

Gates, B. 3, 7 GE (General Electric) 126–30 General Motors 20, 138 General Signal Thinfilm Company 208 German companies 226, 253–4 Grant, R.M. 56, 81–4 growth: economic 47; policy 34 Grundfos 125–9 Gulati, R. 152 GW Systems 204

Iansiti, M. 124 IBM 20, 138, 202, 204, 207, 210 ICI (Imperial Chemical Industries) 121, 139, 141–7 ideas: generation of 36 IG Farben 141 ignorance 36, 85, 88, 171, 174 imagination 35–8; corporate 115; see also adaptive imaginary representations imitation 1, 55, 71, 77–6, 88, 91–5, 124–8, 148, 185, 218, 231–2; barriers 21, 185–3; imperfect 41; perfect 89; resource 59 imperfect mobility 13 indeterminacy 93–6 individuals 42, 64, 85–8, 88 Indonesia 233 inductive reasoning 37 industrial organisation: theory of 18 industry: effects 62, 65, 70 information 43, 66, 68, 109, 152, 175–3, 178, 181; capital 64; costs 16, 18–20, 22, 171, 173, 180, 183; impactedness 82; imperfect 55, 58; processing perspective 108; sticky 108–11; structures 100

Hamel, G. 2, 12, 22, 23, 43, 51, 56, 166, 184 Hansen, M.H. 56 Hanson 143 Harrigan, K.R. 150 Harrison, T. 143 Hart, O. 174, 176 Harvard Business School 126 Hatch, W. 233, 235 Henderson, R. 83, 88, 94, 149 heterogeneity 13, 20; of firms 23; resource 40, 59, 185 heterogenous demand 61; theory 56, 58 High Vacuum Apparatus 204 history 36, 44, 46, 61, 81, 89, 184, 186; corporate 46 Hitachi 210, 212, 230 Hobday, M. 232 Hodgson, G.M. 54, 64, 72 Hong Kong 6, 242–4, 248–63, 258;

INDEX 267

infoware 149 innovation 11, 23, 24, 35–8, 39–2, 58, 79, 94, 120–48, 154, 156, 166, 182, 222, 225; autonomous 193; endogenous 55, 71; networks 153; policy 34, 47; proactive 59; race models of 83, 88; reactive 59; routines 80; systemic 191–193; systems 34; technological 12, 24, 121–30, 132; unanticipated 15; see also innovative innovative: assets 120–48; capabilities 77–94, 123 institutional: economics 56; systems approach 29 institutions: economic 56; societal 57, 59–2 integration 150–6, 157, 160–8, 164–2 vertical 71, 141, 191, 196, 210, 224, 243 intellectual property 64, 123 intention: strategic 43, 44, 46, 97 interfaces 103–6, 109–12, 193 internal: capabilities 191, 203, 206; development 16, 148, 150, 152; properties of the firm 29; resources 148; technological environment 2 international: trade theory 59 intra-firm learning processess 16 introspective: RBP as 1 irreversibility 15 isolated resource problem 1, 3, 5–6 isolating mechanisms 79, 82 Jacobson, R. 87

James, A. 23 Japan 107, 230, 248, 253–4; multinationals in 219, 233, 235; semiconductor industry in 196, 203, 206, 210–20 joint ventures 34, 44 judgment 35, 38 Kay, J. 1 keiretsu 210 Ketteringham, J.M. 135 Kirzner, I. 84–8, 88, 222 Klein, B. 22, 51, 174 know-how 101–4, 113–17, 229, 243 knowledge 29–47, 55, 80, 85, 88, 89–5, 101– 4, 106, 149, 160–8, 180, 221–5, 235; accumulation of 32, 171–9, 179, 187; architectures 99–2, 102, 108–15; based view of the firm 40; codified 38; -creating company 2, 23; differential 43; discovery process 56; firm’s 37–40, 43; fragility of 45; generation 34, 37, 45; management 183, 185–4; market 39, 43; new 86; organisational 101–4, 113; paradigm view of 41; resources 14, 97–2, 109, 113; tacit 43–7, 51, 64, 71, 124, 229; technological 29, 34, 37, 45, 100, 222, 226–6 know-what 101–4, 113–18 know-why 101–4, 113–18 Knudsen, C. 54 Kokusai 212 Korea 218; semiconductor industry in 196, 206, 211 Kuhn, T. 37, 39 Kwok, P. 218 Laamanen, T. 155 labour 5, 178, 244, 246, 257; costs 220, 234, 233, 251, 253;

268 INDEX

division of 31, 41–3 Lado, A.A. 56 Lam Research 206–15, 209 Landes, D.S. 226 Lang, J.W. 116 Langlois, R.N. 35, 51–5, 69, 171–9, 176–4, 187, 194 learning 11, 15, 22–4, 86–9, 100, 102, 138, 151, 183–1, 187; by doing 133, 182, 198, 203; by using 133; costs 179, 183; effects 20; in firms 6; genuine 85, 93; 116; intra-firm 16; organisational 2, 15, 18, 23, 24, 42, 55, 58, 80, 83, 99, 101, 108, 112–19, 167; paths 170–8; resource 17; technological 116 Lengnick-Hall, C. 83, 150 Leonard-Barton, D. 80 Lester, R.K. 220, 234 Levy, G.B. 55, 65 Lippman, S. 11, 21–4, 56 Littlechild, S.C. 171–9, 174–2, 179 Luhmann, N. 223 Luks Electronics Ltd 230 Macao 248 Madhok, A. 56 Mahony, J.T. 106 Malone, T.W. 108 management: role 63 managerial: authority 107; processes 32, 64 manufacturers: small 6 manufacturing 122; industry 4 March, J.G. 138, 232 Mark I RBP 11–12, 15, 21–4, 24 Mark II RBP 11–12, 15, 22–4

market 31, 47, 85, 178, 193, 221, 233, 243, 246, 258; -based economies 57; future 92; offerings 63–5; performance 40; power 54; position 34, 57; preferences 115; process 23, 32, 47, 79–3, 84–9, 93, 172; segments 58, 59–2, 63, 68, 71, 103; share ix, 254; structure 1, 4; success 159, 160, 163; trends 113–17 marketing 3, 41, 122–6, 125, 127, 154, 156, 232, 246, 252, 256; heterogeneous demand theory 56, 58; research 103 Markides, C.C. 121, 127, 129 Marshall, A. 34 materials: raw 3, 5 Matsushita 231, 255–6 maximisation 23 Mayo, J. 143 McKelvey, M. 42 McNulty, P. 31 MDC Vacuum Products 204 means 80, 91; -ends framework 77, 80–3, 84, 87–89, 92–5 mechanisms: isolating 79, 82 memory 43; organisational 167 mergers 29, 44, 46–9 MESA (Modular Equipment Standards Architecture) 207–17 MESC (Modular Equipment Standards Committee) 207–18, 212 Metcalfe, J.S. 23 Microsoft 3, 116 Mikkelsen, J. 230 mobility: imperfect 13, 41 modular: processes 6, 100–91; products 7, 100–19

INDEX 269

modularity 99–19, 193–1, 197, 201–20 monopoly 18–19, 82–5, 245 Montgomery, C.A. 7, 9, 54, 56 moral: codes 57, 67, 68 Morris, C.R. 193–1 motivation: human 67–68 Motorola 255 multinational companies 164, 219, 233, 235, 244; see also transnational corporations multi-technology companies 120–48 Nayak, P.R. 135 Nelson, R. 39, 55, 61, 64, 80, 94 New Growth Theorists 2–3 newly industrialised: countries 218–8, 233; economies 242, 244–7, 249, 256 niche market 3, 103, 245 nihilism 91 Nikon 210 Novellus 206–15 novelty 15, 23, 39, 47, 85, 89 O’Driscoll, G. 84–9 Ohio Nuclear 126 oligopoly 5 open door policy: in China 242, 247, 249 opportunism 51, 68, 175, 221, 223–3 opportunity 91–5; productive 148; technological 123 organisation 41–3, 43; economic 170–8; memory 43; style organisational: behaviours 83; competences 97, 125, 149; knowledge 101–4; learning 2, 15, 18, 23, 24, 55, 58, 80, 83, 99, 101, 108, 112–19, 167; processes 45, 64; routines 88;

structures 41, 83 orgaware 150 original equipment manufacturers 219–9, 231, 233 outcomes 80, 88–4 outputs 79; innovative 84 Owen, G. 143 paradigms 38; scientific 39; technological 24, 37–39, 93–6; view of knowledge 41 partnering 208 patents 59, 82, 123, 126, 148, 160, 185 path 91–4: dependence 5, 15, 18, 37, 55, 87, 94, 133; intellectual 9; prescribed 91 pattern recognition 37 Pelikan, P. 51 Penrose, E.T. ix, 3, 9–12, 14–18, 21–4, 24, 32, 40, 43, 51, 54, 56, 95, 194 perfect competition 19, 55, 58, 64–8, 68–4 performance 38; financial 57, 68–3, 77, 81–6; levels 101, 103, 104; market 40 personal computers 104, 194, 196, 204, 210 Peteraf, M.A. 5–6, 56, 82 Pettigrew, A. 142 pharmaceutical industry 83–6, 94, 255 Philips 138, 255 Phillips, R.L. 62, 65 philosophy: moral 67; of science 65 Pisano, G. 64 platform 103; design 104 policy: growth 34; innovation 34, 47; public 57, 61 Porro Technologies Ltd 230 Post, H.A. 115 power:

270 INDEX

relationships within firms 5 Prahalad, C.K. 2, 11–12, 22, 23, 43, 51, 56, 184 praxeology 84 price theory 13, 18 prices 79, 178–9; maximum 103 prisoner’s dilemma 68 problem; agency 69; dynamic optimisation 15; environment 1, 3; isolated resource 1, 3, 5; resource application 1, 6; resource creation 1; resource organisation 1, 5–6 process: analysis 15; architectures 99–4; artefacts 38; issues 23; new 34, 89; technology 2, 191–212 procurement 123 product: architectures 99–8; artefacts 38; market applications 16; new 34, 89; portfolio 130, 144; strategy 97; technologies 2 production 3, 41; costs 106; organisation of 33 productive: capabilities 34; opportunity 16, 41 profit 68 progress 38 property: intellectual 64, 123; rights 6, 18, 180, 235 psychology 14 puzzle: solving 38

Quélin, B. 150 race models of innovation 83, 88, 93 Rank Xerox 138 rational reconstruction 21 rationality: 34–8, 171; bounded 35, 40, 51, 152, 175, 179, 221; comprehensive 35; maximising 23; Olympian 35; reasonable 35 reality: models of 36 Realtime Performance 204 reasoning: inductive 37 Redding, S.G. 218 Reed, R. 56, 82 regime: technological 39, 99–3, 116 regulatory bodies 5 renewal: corporate 2, 23 rent: capability 44; differential 43; earning 170–8, 176, 184–2; in equilibrium 23; quasi-sustainable 43; Ricardian 5, 14, 174, 184 representation: spatial 86 reputation 82, 97, 206 research and development 3, 41, 83, 122, 124, 131–5, 138, 140, 142, 160, 170, 186, 220, 231, 234, 245 resource: accumulation 15; -advantage theory 54–72; application problem 1, 6; base 7, 87; building 3; bundles 20, 41, 43, 54–6, 77, 184; characteristics 66; creation problem 1–2; firm-addressable 97–1, 101; firm-specific 97–1, 101;

INDEX 271

heterogeneity 40, 185; learning 17; organisation problem 1, 5–6; see also resources resources 1, 54, 80, 97, 246; administrative 80; advantage producing 59; application of 17; assortments of 61; basic 55, 63, 71; competences as 62; complementary 5–6; coordination of 14; creation of 3, 6, 12, 14–15, 23, 24, 58; definition of 7, 121, 149; excess 14, 16; existing 1, 15; external 4, 97, 148; financial 63, 225, 227; human 1, 14, 40, 63, 149; immobile 59; informational 63; internal 4, 148; isolated 4–5; knowledge 14, 97–2, 109, 113; legal 63; maintenance 80; managerial 14, 16; organisational 40, 63, 80, 83, 149; physical 14, 40, 63, 149, 228; portfolio of ix; production 80; relational 63; societal 57, 61; technological 4–5, 7, 99; see also resource responsiveness 89 reverse salients 4 revolution: permanent 5 rights 174, 176, 179, 182–90, 185; property 6, 18, 180, 235 Rizzo, M. 84–9 Roberts, E.B. 150 Robertson, P.L. 5, 53–5, 171–9, 176–4, 187, 194, 219 Romer, P.M. 72 Roquebert, J.A. 62, 65

Rothwell, R. 165 routines 42–8, 53–5; dynamic 81, 87, 89–6; innovation 80; organisational 77; search 91–5; static 80 Rubin, P. 15 Rumelt, R. ix, 9–12, 21–4, 56, 62, 82–5, 184– 3 Sako, M. 222–2 Samsung 211, 255 Sanchez, R. 118, 121, 149, 166 Sanderson, S. 130 Sanyo 256 saving 68 Saxonian, A. 99 Schendel, D. 56 Schmalensee, R. 62 Schoemaker, P.J. H. 7, 41, 56 Schumpeter, J.A. 37, 57, 77–1, 152, 191, 222 science 36–39; philosophy of 65 Scott, E. 220 Sellars, W. 55, 65 Selznick, P. ix, 11, 21 Sematech 208–17 semiconductor: fabrication equipment 195–20; industry 6, 194–20 service: costs 106; productive 40 Shackle, G.L. S. 35 shareholders 68 Shi, Y. 234 Siemens 256 Silicon Valley 194, 208, 218 Silver, S. 135 Simon, H.A. 35, 106 Singapore 211, 218, 234–5 single-wafer processing 197–12 Sit, V. 229 skills 34, 38, 41, 55, 88, 120, 251, 258 slack administrative skills 3 small firms 218–45, 245, 256

272 INDEX

Smith, A. 34, 181, 193 social construction 5, 224 socialisation 68–1, 223 societal: institutions 57, 59–2; resources 57, 61 socioeconomics 56 sociology 15 soft: approach in RBP 11, 24 Solow, R.M. 2 Sony 106, 129–3 space 61; resources and capabilities 97–3 spatial representation 86 special economic zones 252–3 specialisation 17, 31, 43, 99, 125, 179–90, 186– 5, 191–194, 202–12; technological 100 Spender, J.-C. 22 stakeholders 5, 67 standardisation 100–3, 116, 193, 202–12, 206– 20, 244–5 state owned enterprises 242, 253 Stokey, N. 72 Stowsky, J.S. 196 strategic intention 43, 44, 46 structure-conduct-performance approach 1, 18– 19, 62 subcontracting 229–9, 231, 252 subjectivism 17, 84, 86–9, 91, 93, 171, 173, 176–4, 179, 186 Sung, Y.-W. 231 suppliers 34, 41, 68, 222; actions of 61; relations with 64 sustainability regime 82 taboos 57 Taiwan 218–8, 235, 248; semiconductor industry in 211 Talouselämä 154–60 target: of acquisition 45–9 team 41–3; -effects 20; members 32;

production 51, 173 Technicare 126, 127 technological: capabilities 218–45; environments 2; evolution 54; paradigms 24, 37–39, 93–6; paths 88–1; progress 70–4, 102; regime 39, 99–3, 116; success 157–4, 160, 163, 164 technologies: common 99; diffusion of 4, 7; process 2, 4; product 2, 4 technology 36–39: acquisition 6; base 120–5, 130, 132–48; development of 29; diversification 120; high- 3–4, 7, 224–5, 234; integration 124; low- 4, 7, 219–9, 224–5, 233–4; medium- 4, 224–5; production 68, 84; research 84; strategy 97, 219–35, 233–4; systems 24; transfer 242–3, 247, 255, 258 technoware 149 Techware Systems 204 Teece, D. 64, 121–5, 126, 133, 139, 149, 194 teleology 67 Temin, P. 84, 88 Termbray Electronics Co. 231 Texas Instruments 195, 201–9, 207, 232 Thesis 204 Thoburn, J. 244 Thorn Electric Industries 127 time 61, 84, 88, 89, 132, 180; subjective 93 Tokyo Electron 212 tools: cluster 196–20 tradability 43 trade: international 59

INDEX 273

tradition 43, 57 training 43 transaction 32, 221, 224; costs 14, 32, 68, 125, 171–90, 187, 193, 206, 221, 224, 243; specific assets 51; typical economic 31 transnational corporations 242–7, 253–8; see also multinational companies trust 57, 68, 221–8; competence 222, 224–4, 230; contractual 222–2; goodwill 222–6, 229 Turk, T.A. 167 Tushman, M.L. 153 Ulvac 212 uncertainty 17, 171–9, 174–4, 179–7, 185–3, 223, 227–7, 231; technological 89, 170, 182, 187 United States 248; family businesses in 223, 226; multinationals in 219; semiconductor industry 7, 196, 203, 206– 20, 232 University of California, Los Angeles 18, 21 utilitarianism 67 utility 68 Utterback, J. 153 Uzumeri. M. 130 VAT 204 vertical integration 71, 141, 191, 196, 210 224, 243 Vincenti, W.G. 36, 38 von Hippel, E. 108 voting behaviour 68 Vtech 231–2 Walrasian: general equilibrium 70 Wang, Z. 253 Wernerfelt, B. ix, 9, 184–2, 186, 188 West, J. 124 Westfall, P.A. 62, 65 Williams, J. 82 Williamson, O.E. 174–2, 187, 221, 224

Williamson, P.J. 121, 127, 129 Wilson, D.S. 36 Wilson, M.C. 56 Winter, S. 39, 80, 94 Wong, S. 218, 228–8 Wu, N. 254 Yamamura, K. 233, 235 Yates, J. 108 Young, A. 72 Yu, T.F. 218–8, 231 Zeneca 143