New Frontiers in the Economics of Innovation and New Technology
New Frontiers in the Economics of Innovation and New Technology Essays in Honour of Paul A. David Edited by
Cristiano Antonelli Professor of Economics and Director of the Department of Economics at the University of Torino, Italy
Dominique Foray Professor of Economics of Innovation and Director of the College of Management of Technology at the Ecole Polytechnique Fédérale de Lausanne, Switzerland
Bronwyn H. Hall Professor of Economics at the University of California at Berkeley, Research Associate at the National Bureau of Economic Research, Cambridge, Massachusetts, and the Institute of Fiscal Studies, London
W. Edward Steinmueller Professor of Information and Communication Technology Policy, SPRU – Science and Technology Policy Research, University of Sussex, Falmer, Brighton, UK
Edward Elgar Cheltenham, UK • Northampton, MA, USA
© Cristiano Antonelli, Dominique Foray, Bronwyn H. Hall and W. Edward Steinmueller, 2006 All rights reserved. No part of this publication may be reproduced, stored in a retrieval system or transmitted in any form or by any means, electronic, mechanical or photocopying, recording, or otherwise without the prior permission of the publisher. Published by Edward Elgar Publishing Limited Glensanda House Montpellier Parade Cheltenham Glos GL50 1UA UK Edward Elgar Publishing, Inc. 136 West Street Suite 202 Northampton Massachusetts 01060 USA
A catalogue record for this book is available from the British Library Library of Congress Cataloguing in Publication Data New frontiers in the economics of innovation and new technology: essays in honour of Paul A. David/edited by Cristiano Antonelli . . . [et al.] p. cm. Includes bibliographical references and index. 1. Innovative technology–Economic aspects. 2. Diffusion of innovations– Economic aspects. I. Antonelli, Cristiano. II. David, Paul A. HC79.T4N472 2005 338’.064–dc22 2005046147
ISBN-13: 978 1 84376 631 5 ISBN-10: 1 84376 631 0 Printed and bound in Great Britain by MPG Books Ltd, Bodmin, Cornwall
Contents vii
List of contributors PART I 1
The economics of innovation: between and Cristiano Antonelli, Dominique Foray, Bronwyn H. Hall and W. Edward Steinmueller
PART II 2
3
4
5
GENERAL INTRODUCTION
PATH DEPENDENCE IN TECHNICAL CHANGE
Competing technologies, technological monopolies and the rate of convergence to a stable market structure Andrea P. Bassanini and Giovanni Dosi
23
Path dependence, localised technological change and the quest for dynamic efficiency Cristiano Antonelli
51
A history-friendly model of innovation, market structure and regulation in the age of random screening of the pharmaceutical industry Franco Malerba and Luigi Orsenigo Path dependence and diversification in corporate technological histories John Cantwell
6
Is the world flat or round? Mapping changes in the taste for art G.M. Peter Swann
7
Waves and cycles: explorations in the pure theory of price for fine art Robin Cowan
PART III 8
3
70
118 158
188
THE ECONOMICS OF KNOWLEDGE
Learning in the knowledge-based economy: the future as viewed from the past 207 W. Edward Steinmueller v
vi
9
Contents
The economics of open technology: collective organisation and individual claims in the ‘fabrique lyonnaise’ during the old regime Dominique Foray and Liliane Hilaire Perez
239
10
Measurement and explanation of the intensity of co-publication in scientific research: an analysis at the laboratory level 255 Jacques Mairesse and Laure Turner
11
Epistemic communities and communities of practice in the knowledge-based firm Patrick Cohendet and Ash Amin
296
Markets for technology: ‘panda’s thumbs’, ‘calypso policies’ and other institutional considerations Ashish Arora, Andrea Fosfuri and Alfonso Gambardella
323
The key characteristics of sectoral knowledge bases: an international comparison Stefano Brusoni and Aldo Geuna
361
12
13
PART IV THE DIFFUSION OF NEW TECHNOLOGIES 14
Uncovering general purpose technologies with patent data Bronwyn H. Hall and Manuel Trajtenberg
15
Equilibrium, epidemic and catastrophe: diffusion of innovations with network effects Luís M.B. Cabral
16
Technological diffusion under uncertainty: a real options model applied to the comparative international diffusion of robot technology Paul Stoneman and Otto Toivanen
PART V 17
427
438
POSTSCRIPT
An appreciation of Paul David’s work Dominique Foray
Index
389
471
475
Contributors Ash Amin, University of Durham, UK Cristiano Antonelli, University of Turin, Italy Ashish Arora, Heinz School of Public Policy and Management, Pittsburgh, USA Andrea P. Bassanini, OECD, Paris, France Stefano Brusoni, University of Bocconi, Milan, Italy Luís M.B. Cabral, New York University, USA John Cantwell, Rutgers University, Newark, USA Patrick Cohendet, Université Louis Pasteur, Strasbourg, France Robin Cowan, Universiteit Maastricht, the Netherlands Giovanni Dosi, Sant’Anna School of Advanced Studies, Pisa, Italy Dominique Foray, Ecole Polytechnique Fédérale, Lausanne, Switzerland Andrea Fosfuri, Universidad Carlos III, Madrid, Spain Alfonso Gambardella, University of Bocconi, Milan, Italy Aldo Geuna, SPRU – Science and Technology Policy Research, University of Sussex, UK Bronwyn H. Hall, University of California, Berkeley, USA Liliane Hilaire Perez, Conservatoire National des Arts et Métiers, France Jacques Mairesse, Institut National de la Statistique et des Etudes Economiques, France Franco Malerba, University of Bocconi, Milan, Italy Luigi Orsenigo, University of Bocconi, Milan, Italy W. Edward Steinmueller, SPRU – Science and Technology Policy Research, University of Sussex, UK Paul Stoneman, Warwick Business School, UK vii
viii
Contributors
G.M. Peter Swann, Nottingham University Business School, UK Otto Toivanen, Helsinki School of Economics, Finland Manuel Trajtenberg, Tel Aviv University, Israel Laure Turner, Ecole Nationale de la Statistique et de l’Administration Economique, France
PART I
General Introduction
1. The economics of innovation: between and Christiano Antonelli, Dominique Foray, Bronwyn H. Hall and W. Edward Steinmueller 1
PRELUDE
The birth and infancy of innovation economics, as a specific area of academic study and empirical research, and one that had applications to real problems of industrial growth, firm strategy and economic policy, occurred at the same time that the boundaries of economic theory were expanding horizontally in the 1950s and 1960s. During that period, economic theory began to offer a detailed and interpretative framework for the study of economics, assimilated the Keynesian heresy as a structural variation in behavioural modelling, and foraged for new areas of research in which to apply its analytical categories and its powerful mechanisms of systematic analysis. This newly found confidence in the methodological validity of individual decision-making and marginal calculations was based upon rigorous assumptions about the rational behaviour of agents. The addition of a new, large macroeconomic area based upon the paradigm of general economic equilibrium, encouraged economists to venture still further into new areas of application. The extraordinary heuristic capacity of economic theory supported bold claims of explanatory competence and scientific primacy for the field. It is not surprising that those years are sometimes spoken of as economic imperialism. The new territories that economic theory colonised included the economics of health and education, of risk and insurance, of uncertainty and information, and of marriage and the family. Economics and related disciplines claimed to be able rationally to assess choice, in all of its manifestations, by processing the alternatives through the machinery of finance, probability and opportunity cost. At the same time there was a progressive specialisation of competencies. Areas which up to then had been thought of as interchangeable, such as international and regional, developed and developing economies or emerging and mature industries, became distinct.
3
4
General introduction
There are many reasons why this phase can be metaphorically referred to as the . Just as the Greek heroes started their campaign of conversion of the barbarians to the superiority of Greek culture and ideology by winning the Golden Fleece, economists were convinced that their insights had become universal. In particular, their insights could be applied to the origin and spread of innovation and, consequently, to the issues of technical progress, science, the university and, even, the creative process. The result was that economists asked themselves not only what creates a nation’s wealth, but also why are certain countries more innovative than others or why do certain historic periods appear to be more fertile and dynamic than others? The analysis of what determines innovative activity, identified as a specific form of economic action, was distinguished from the analysis of the effects of introducing innovation. In this way, innovation economics invaded the territory that had already been subdivided by industrial economics, the theory of the firm, regional economics while not ignoring important points of international economics and, above all, public economics. The claim of scholars of innovation to the ownership of the Golden Fleece was certainly at hand: the rate of growth of economic systems and their share of the international market, wage differentials and rates of profit were increasingly linked to innovative capacity. Innovative capacity was (and still is) perceived as one of the fundamental source of a nation’s wealth and more specifically the ever changing, if not increasing, differences in that wealth. The wealth of knowledge and the abundance of recipes, diagnoses and therapies, deriving from the economics of innovation have tempted many of its followers to advocate the , a contest for the capital cities waged from what is still often regarded as the hinterland. Studying and investigating the multifaceted field of innovation economics has, in fact, more than once unearthed some awkward and significant results; results that cannot easily be ignored and that are, at least in some circles, embarrassing for claimants to the Golden Fleece of a true knowledge of the means and ends by which economic results are produced. This is true, above all, in the dominant theory that the imposing expedition had advanced and, indeed, financed. Technological knowledge does not appear to be an exogenous flow into the economic system, as had been assumed for the purpose of ‘simplifying’ analysis. Neither does it appear that the choice of factors of production or individuals’ preferences are necessarily reversible. The workings of the whole economic system seem to be far from such Newtonian physics. Instead, the emerging model of economic life is one suffused with nonergodic dynamic elements, where the path taken shapes the destination
The economics of innovation
5
reached, where rate and direction are interdependent and where causes are entangled with effects in uncomfortable new patterns. Correspondingly, examining the origins and spread of innovation has made it possible to understand clearly how incomplete and indeed inaccurate it is to assume that economic agents act in a perfectly rational way with the corollaries of perfect cognition and predictive capacity. There is a notable absence in the real world of the race of super-rational agents endowed with rational expectations and able, thereby, to anticipate a whole range of future actions including innovation and new patterns of consumers’ preference. But an even fiercer battle remains to be waged. For, if technological knowledge is no longer exogenous, but instead strongly influenced by the unfolding of economic life, the same hypotheses regarding the workings of the market must also be questioned. Firms do not limit themselves to adjusting outputs to prices, but also struggle to survive or race to supremacy through innovation. In such a world, it can be understood that it is no longer possible to imagine a single general economic equilibrium, it is necessary to speak more of a range or, perhaps, even more precisely, of a series of possible general economic equilibria. When the evolution of consumers’ preferences are also recognised as being endogenous and dependent upon experience, it seems equally doubtful that we are living in the best available world, and that we might at the same time know how long or how far our journey might be to a better one. As innovative capacity is strongly influenced by the processes of accumulation of technical knowledge, the direction that we take in attempting to add to this accumulation affects what destination can be reached. And through this uncertainty, choice, imagination and inspiration are reintroduced into economic life. The weapons for the are by now ready. Will our heroes manage to complete the journey which returns them to Athens? And, above all, will Athens be able to appreciate the new language, the plates and gold and silver that the Argonauts will bring with them? Will Athens be able to recognise that its past certainties were limited and fragile, and were, above all, based on static and incoherent assumptions, although they were the source of great daring and farsightedness? Economics of innovation enters the twenty-first century unable to decide between the modest yet reassuring temptation to consolidate long-held certainties and the far bolder and more tenuous goal of rewriting the model. Regardless, it has begun its return from the interior to the country of its birth, eager to show that new truths are to be discovered that are less certain but more complex and plausible than those made by the forefather conquerors. This is the context into which the lifelong contribution of Paul David can be better appreciated and valued. David has built essential pathways
6
General introduction
in both directions; in the as well as in the . The contribution of David has been fundamental in the to building economics of innovation as a new discipline and area of specialist expertise and competence. In the same fruitful career, Paul David has contributed some of the most powerful analytical tools by which the may be organised. In this introduction we review David’s contributions to three important areas in the economics of innovation: the economics of knowledge, the role of path dependence in the evolution of economic equilibria and the diffusion of new technologies. We have necessarily been selective in our choice of contributions to discuss, lest this introduction become longer than the volume it introduces! Nevertheless we hope that our brief surveys will give the reader an idea of his many faceted contributions to this literature, at the same time that it places the chapters in the context in which they were written.
2
THE ECONOMICS OF KNOWLEDGE
While ‘knowledge’ in a very broad sense has always been at the heart of Paul David’s work, at both microscopic and macroscopic levels, he came to explore this topic more systematically in the early 1990s. Starting with the analysis of the peculiar properties of knowledge and information as an economic good, he proceeds to the historical and normative analysis of resource allocation mechanisms in the field of knowledge production and distribution and, more generally, socio-economic institutions that can be relied upon to produce, mediate and use knowledge efficiently. The intellectual journey of Paul David in this field consists of a systematic exploration of the three-dimensional space in which ‘knowledge-products’ are distributed.1 The first dimension of that space is the continuum between secrecy and full disclosure; the second is the spectrum of asset ownership status ranging from legally enforced private property rights to pure public goods; and the third is the dimension along which codification appears at one extreme and tacitness at the other. How do institutions, technologies and economic factors determine the location of knowledge in this three-dimensional space, and what are the implications of this on allocative efficiency in the domain of knowledge production and use? These two questions are raised repeatedly in Paul David’s work. The argument that knowledge and information have the properties of public goods creates the theoretical framework in which these questions are addressed.
The economics of innovation
7
Knowledge, the Lighthouse and the Economist David has made it very clear that saying a good (for example, knowledge) is a public good, on the basis of the properties of non-excludability and non-rivalry, does not imply that this good must necessarily be produced by the state, that markets for it do not exist or that its private production is impossible. It simply means that, considering the properties of the good, it is not possible to rely exclusively on a system of competitive markets to efficiently guarantee its production. Considering the example of the lighthouse as a paradigmatic case of a public good, David shows that the fact that the lighthouse service was once provided by the private sector in the UK, as Coase documented in a famous paper, does not mean that the lighthouse is a private good.2 Private markets function in this case because an agent is granted local monopoly on the right to collect a tax in exchange for the service provided. In the same way, the creation of a private monopoly on new knowledge (a patent) enables the market to produce that good. But in both cases the remedy is imperfect, for the owner of the monopoly will not supply the ‘light’ (of the lighthouse or of the knowledge) at a price (harbour tax or royalties) equivalent to the negligible cost of making these goods available to additional users (the marginal cost of use of existing knowledge is nil, as it is in the case of using the harbour’s lighthouse).3 What are the consequences of such a clear and strict position on the economic nature of knowledge? In the domain of the production of new ideas (a very broad domain ranging from scientific discovery and technological and engineering innovations to intellectual creation), the fundamental problem for the allocative efficiency of competitive markets arises from the externalities that exist because of the public good nature of ideas. It is, therefore, crucial to analyse the historical emergence and the allocative efficiency of various kinds of institutions devised to correct or to provide alternatives to such market failures. The 3 Ps figure (public procurement, patronage and private property) then provides a framework for such an investigation.4 From the New Economics of Science . . . In this direction, the main contribution of David (notably with Partha Dasgupta) has been the rigorous and systematic exploration of the economics of ‘open science’ – including both a detailed analysis of the norm of openness which is ‘incentive-compatible with a collegiate reputational reward system based upon accepted claims to priority’,5 and an analysis of the historical emergence of this institution.6 Both analytical and modelling approaches show how efficient such a system is, for it ensures the rapid and
8
General introduction
complete diffusion of new knowledge while preserving a certain level of incentive.7 Moreover, complete disclosure functions as a sort of ‘quality assurance’ in so far as published results can be reproduced and verified by other members of the community. Given the fact that by most measures the productivity of scientific research as organised under these institutional principles has been outstanding, it is clear that the economics of open science provides a framework to study and assess other kinds of institution with similar positive effects on the growth of knowledge. The economics of free/libre open source software is, for example, a direct extension of the work done on science. Today, David is in the forefront of research on this particular class of social systems in which high rates of innovation are correlated with rich spillovers, implying that private agents do not always rely on exclusivity and excludability mechanisms to capture private benefits from their intellectual creative work. . . . to the Comedy of the Commons Owing to the peculiar features of knowledge, the production of knowledge has the potential to create a ‘combinatorial explosion’. This good is difficult to control and can be used and reused infinitely to produce other knowledge which is in turn non-excludable, non-rival and cumulative, and so forth. In many cases knowledge is also deliberately disclosed and organised in order to facilitate its access and reproduction by others. All these processes give rise to the creation and expansion of a ‘knowledge commons’. ‘Knowledge commons’ are not subject to the classic tragedy of the commons, a parable describing the case where exhaustible resources (such as a pasture or a shoal of fish) are subject to destruction by unregulated access and exploitation. Knowledge may be used concurrently by many, without diminishing its availability to any of the users, and will not become ‘depleted’ through intensive use. As David recently wrote, contradicting the American poet Robert Frost’s elegy to New England civility, good fences do not make good neighbours: ‘information is not like forage, depleted by use for consumption; data sets are not subject to being “overgrazed” but instead are likely to be enriched and rendered more accurate and more fully documented the more researchers are allowed to comb through them’.8 The properties of non-excludability, non-rivalry and cumulativeness have features akin to quasi-infinite increasing returns. Thus, the commons is not tragic, but comedic, in the classical sense of a story with a happy ending. Managing and protecting the ‘knowledge commons’ requires social regulations that are entirely different from the social arrangements used to regulate ecological systems of exhaustible
The economics of innovation
9
resources. In this respect, Paul David has devoted a great deal of time and intellectual creativity to conceive of remedies to the current tendency to strengthen intellectual property systems, especially as they apply to scientific research and scientific databases. ‘Knowledge on Line’: the Economics of Learning-by-doing Learning-by-doing has been a key form of technical change analysed by Paul David in his studies in economic history, allowing him to address the issues of localised learning, the importance of history, and the policy implications of supporting infant industries (see the section on path dependence).9 Apart from these works on economic history and path dependence, the most interesting contribution of David related to learning-by-doing is perhaps his emphasis on the fact that the economics of learning-by-doing appears to be an area in which the conflict between static and dynamic efficiency is particularly important. There is a tension between the normal performance expected in the course of ordinary operations and the learning aspect: In most instances of learning-by-doing, the feedback from experience to inferred understanding is severely constrained. The doers have limited facilities for accurately observing and recording process outcomes, or for hypothesizing about the structure of the processes they are trying to control. Advances in knowledge that are empirically grounded upon inferences from trial-and-error in a myopic control process cannot be a big help when they are restricted in both the number of trials they can undertake, and the states of the world they can imagine as worth considering.10
In a similar vein, David has been a pioneer in building the concept of experimental or explicitly cognitive learning-by-doing.11 Such a process consists in performing experiments during the production of goods or services, or in the case considered by David and Sanderson, the production (or non-production) of children. Through these experiments new options are spawned and variety emerges. This is learning based on an experimental concept, where data is collected so that the best strategy for future activities can be selected. Technical and organisational changes are then introduced as a consequence of learning-by-doing. In other words, explicitly cognitive learning-by-doing consists of ‘on line’ experiments. The possibility of moving on to explicitly cognitive learning in activities other than ‘craft trades’ represents an important transition in the historical emergence of the knowledge-based economy. As long as an activity remains fundamentally reliant on learning processes that are procedures of routine adaptation and leave no room for deliberate planning of experiments during economic
10
General introduction
activity, the gap between those who deliberately produce knowledge and those who use and exploit it remains wide. When an activity moves on to higher forms of learning where the individual can plan experiments and draw conclusions, knowledge production becomes far more collectively distributed. The Increasing Use of Codification . . . Paul David has been a leader in the analysis of the economic significance of knowledge codification both at the macro and micro levels. Codified knowledge serves inter alia as a storage depository, as a reference point and possibly as an authority.12 As such, codification is a state in which knowledge is presented (improving memory, learning and communication), a tool for constructing new knowledge and a means to facilitate co-ordination. In pointing out the important distinction between knowledge that is codifiable (in the sense of articulable) and that which actually is codified, and in focusing analytical attention upon the endogenous boundary between what is and what is not codified at a particular point in time, David and colleagues have helped to persuade economists to ‘colonize’ this new area (previously left open to other social sciences) with their tools and concepts. At the macroeconomic level, Abramovitz and David have analysed knowledge codification as both the driving force behind the expansion of the knowledge base and its favourite form; in short, the most salient characteristic of modern economic growth.13 . . . and the Trap of ‘Taciturnity’ Investment in codification is suboptimal due to public good problems and high fixed costs. Any lack of attention paid to complementary components of a codified knowledge base (continuity of languages, software enabling access to older files) runs the risk of irremediably altering the codified knowledge base and diminishing private and social returns from codification investments. One policy failure can, therefore, be seen in the lack of provision of incentives to codify resulting in the building of excessive stocks of knowledge that are left in a ‘tacit’ form. These accumulations of tacit knowledge are distributed throughout the private sector in a way that makes them more costly to locate, to appraise and to transfer. Policies that encourage strategies of excessive tacitness by some firms thus tend to reduce incentives for other firms to invest in searching for existing codified solutions to their scientific and technological problems. A result may be excessive insularity and a waste of resources. In lowering the private rate of return on
The economics of innovation
11
monitoring and information searches over a wide field, such strategies and policies will reinforce excess concentration of research and development (R&D) in certain areas, and contribute to the under-utilisation of existing stocks of knowledge. ‘Taciturnity’ (a word proposed by Paul in this context) describes a deliberate failure to express feasibly codifiable information for strategic or cost reasons and may therefore create private and social inefficiencies.14 Information and Communication Technologies as a Knowledge Instrument and a Mischievous Break A central theme in Paul David’s work on knowledge is the information and communication technology (ICT) revolution in so far as it involves technologies geared to the production and dissemination of knowledge and information. Apart from the productivity paradox issue, which is mainly a ‘diffusion story’ (see section 4 in this chapter), David has developed systematic analyses of the ways in which ICTs interact with organisational and institutional changes to profoundly transform the organisation of economic activities dealing with knowledge creation. E-science is a particular field which warrants in-depth investigation in order to evaluate how far the system of knowledge production and use is transformed through the full realisation of the potential of ICTs as knowledge instruments. Scientists are now reaching the step of building and using a comprehensive virtual federation of resources. ‘Comprehensive’ means that the extent and nature of the resources available through the cyber-infrastructure (people, data, information, computational tools, specialised instruments and facilities) could approach functional completeness for a specific community of practice. However, institutional and organisational issues have to be addressed so that old institutions and organisations are not an impediment to the full and efficient deployment of resources and potentialities in an e-science environment.15 Paul David is certainly a great ‘techno-optimist’ and places strong hope in the advances of new ICTs to improve the ways people are organising their private and professional activities. However, he also knows very well that any success in collaboration and interaction with colleagues is contingent on ‘emotional trust’, namely, a sense of shared identity and familiarity. And this is not going to emerge spontaneously from long-distance collaboration even if mediated through the best of the present ICT infrastructure. This is why Paul applies to himself a practical recommendation for enhancing trust in geographically distributed teams, which is to increase travel early in the history of the project and to travel again each time emotional trust requires some further support!
12
General introduction
Public–Private Interactions and the Transferability of Knowledge Direct transfers of knowledge between academic science communities and the proprietary R&D organisations of the private business sector are especially problematic to institutionalise. This is because the coexistence of two reward systems within any single organisation makes the behaviours of the participants difficult to anticipate, and tends to undermine the formation of coherent cultural norms which promote co-operation among team members. The main issue here is to maintain a proper balance between, on the one hand, the requirements of openness and autonomy of investigation (as these are required for the rapid growth of the stock of knowledge) and, on the other, the need for delays and restrictions upon the full disclosure of all new information (which facilitate the appropriation of economic returns needed to sustain investment in expanding the knowledge base).16 David is a key player in the policy debate on the current transformations of the relationships between science, technology and economic performance, and on the various institutional mechanisms to be designed to get a better protection of the public domain of knowledge from further encroachments by the domain of private property rights. A constant argument in his work is that the basic rationale of intellectual property law depends on an independent public domain containing a stock of freely accessible information. That shared collection of basic knowledge provides the building blocks for new inventions.17 How will it be possible to maintain and expand this collection of freely accessible basic knowledge in the long term? What kind of mechanisms should be designed to preserve the intellectual commons, given that they must be ‘incentive compatible’ with the private allocation of resources to inventive and innovative efforts? These are the ‘big questions’ in his exploration of the future economic organisation of a well-functioning science and technology system.18 A General Insight on this Work As a rationale for his development of such a rich repertoire of works and studies, there is in Paul David’s work a sense of inadequacies and erroneous development in the passage between the understanding that economists have gained about the very detailed mechanisms of knowledge production and distribution, and the stylised facts encapsulated in the formalities of macroeconomic models. This is perhaps why he continuously probes deeper in the microanalysis of detailed resource allocation processes in various areas of the economics of knowledge (open science, open source,
The economics of innovation
13
proprietary R&D, knowledge transfer, intellectual property rights), mobilising for such purposes not only the microeconomics of innovation but also historical analysis, the sociology of science, legal studies and the managerial and organisational literature. However, the most important feature is that such deep ‘drilling’ into very complex structures always leads to the production of very rich stylised facts that have the destiny to inform the rest of the economic profession interested in the determinants of innovation, productivity and economic growth – the foundations of the macroeconomy. Sometimes David himself uses these stylised facts to enrich his own macroeconomic works on economic growth. The value of such application is perhaps best illustrated in the recent set of studies on the economic history of the American macroeconomic growth, published with the late Moses Abramovitz.19 The chapters in this volume by Steinmueller, Foray and Hilaire Perez, Mairesse and Turner, Arora, Fosfuri and Gambardella, Cohendet and Amin and Brusoni and Geuna are all contributions to this field in the various perspectives marked out by David on the economics of information and knowledge.
3
PATH DEPENDENCE
During the twentieth century, economists were finally able to combine the work of Vilfredo Pareto and Marie-Ésprit Léon Walras into the theory of general competitive analysis. The assumptions employed by Kenneth Arrow and Gerard Debreu were numerous, giving rise to both contemporary critiques and rejoinders by those with alternative propositions concerning the fundamental features of market-based economies and actors. The theory of ‘path dependence’ is often, incorrectly, taken as being among these ‘alternative’ approaches to understanding the nature of competitive equilibrium. It is, instead, an effort to address an issue that general competitive analysis leaves unresolved. Namely, what are the consequences of moving between different equilibria as new technologies emerge and are integrated into the market system? For a scholar of economic history, such as Paul David, answering this question was of significance not because ‘history matters’ for that is too easy a target. The more important question is how history might matter.20 David has provided an important answer to this question by developing the modern theory of path dependence, among his most widely cited and least well understood contributions to the modern vocabulary of economic analysis. The theory of path dependence is an attempt to portray the consequences that may arise when the processes of growth and accumulation
14
General introduction
result in reaching one equilibrium rather than another – to explain in a rigorous way how history might matter. At first glance, it would seem that path dependence contests the basic result of general competitive analysis – the achievement of a unique equilibrium through Walrasian and other market-clearing mechanisms that accords with Pareto’s statement of social welfare criteria. The important point is that this equilibrium is at a single point in time.21 It does not account for the unanticipated arrival of the new technologies or other shocks and disturbances that cannot be anticipated and therefore incorporated in the trading equilibrium. When moving between different points in time sufficiently distant to provide an opportunity for intervening events, like the arrival of new technology, a new equilibrium will be reached. Thus it is possible to achieve an array of equilibrium positions in the economy over time depending upon the order, nature and timing of intervening arrivals of technology or other shocks. The foregoing observations only serve to establish the existence of multiple equilibriums that might be reached depending upon the different order, nature and timing of intervening variables. The evolution of economies is thus subject to random disturbances and hence fundamentally unpredictable. This conclusion may be reached from a number of analytical directions and is, itself, unremarkable. As long as the class of all equilibria remains open for further exploration, the economy is not path dependent and, from a probability theory viewpoint, the processes of traversing the various equilibria is ergodic (without memory). What distinguishes the class of ‘path dependent’ equilibriums is that they foreclose reaching other classes of equilibriums. This is a much stronger and more controversial conclusion than arguing that the path of economic development is uncertain.22 The nature of the foreclosure is that the costs of switching to another path involve a substantial diminution of welfare for those that might benefit from the change. This welfare reduction is so large that it makes the decision to ‘switch’ irrational. The key point, however, is that if past movements had led in a different direction, a path not taken, a superior equilibrium could have been reached. Recognising that ‘paths not taken’ may have created an entirely different set of economic outcomes is a direct challenge to the interpretation of the succession of competitive equilibriums either as unambiguously progressive and leading to the ‘best of all achievable worlds’ (subject to the constraints of resource endowments and technology) or the near equivalent, the idea of progress as an ergodic process in which better outcomes will come in time.23 The foreclosure of some equilibria amounts to a ‘lock in’ of actors at or near the vicinity of a particular equilibrium. The most straightforward
The economics of innovation
15
means for generating such lock-in effects is through local feedback and the most prominent example of such local feedback effects are the network externalities that accompany the widespread adoption of a particular technology or method of organisation. Network externalities that substantially lower the costs of one technology or method of organisation raise the costs of switching to alternatives. To illustrate this phenomenon David has employed a number of examples including the case of the QWERTY keyboard where he argues that the local feedback occurred through the decisions of typists to learn ‘touch typing’ and that the ‘lock in’ was reinforced by the ‘quasi-irreversibility’ (high costs of switching) of this skill.24 In his 1985 QWERTY article, David chose a cosmological metaphor to convey the potential significance of path-dependent processes, comparing these processes to the ‘dark stars’ that influence the ordering of our universe. In the intervening years, the significance of the less obvious constituents of our economic universe is receiving more attention. In particular, the role of adoption externalities in information technologies including software have not only been recognised, but have been incorporated in strategy. In considering the role of technical compatibility standards in accelerating the diffusion of new technologies, David invented new meanings for the narrow windows of opportunity for intervention in adoption processes, conjured the spectre of angry orphans that would be created as standards tipped against their (incorrect) choices of the emerging dominant technology, and reanimated the Cyclops ‘blind giant’ – government groping to find a path after the ‘white heat’ of technology had rendered its previous competences obsolete.25 The possibility of ‘premature’ standardisation with the consequence of social welfare losses has become an important component of policy deliberations and a recognised source of ‘market failure’. David and colleagues have provided a series of analyses of technological history in which pathdependent processes play a central role. These include the history of electrical power,26 data communication standards27 and the mechanisation of corn harvesting.28 These contributions as well as his commentaries on the nature of path dependence realise one of David’s central ambitions – to advance the frontier of understanding how history matters in economic analysis. The chapters in this volume by A. Bassanini and G. Dosi, C. Antonelli, F. Malerba and L. Orsenigo, J. Cantwell, P. Swann and R. Cowan not only push the frontier of studies of path dependency forward, they also demonstrate the attractive force of the path dependence as a method for economic analysis.
16
4
General introduction
DIFFUSION
A superficial reading of economics would suggest that the process of exploiting technological and market opportunities is nearly instantaneous – every delay in exploiting opportunity constitutes an opportunity to make gains at the expense of rivals. If a firm cannot reap the first-mover advantage, it must strive to be a ‘fast second’ and woe betide those firms or countries that are laggards in the race towards higher levels of prosperity. The study of economic history provides a useful antidote to this febrile frenzy. Through the lens of economic history one can see that even highly promising developments such as the industrial revolution take decades to spread from one country to another, that innovation often requires new people as well as new machines and that new methods of economic and social organisation displace traditional methods slowly through processes of cumulative change and adaptation. Thus the systematic study of the rate and direction of technological change involves understanding how different economic actors take up or adopt changes that become available. Paul David has contributed to our understanding of the diffusion of innovations through both theoretical and empirical investigations from the very beginning of his career. While the English language use of the term diffusion in relation to knowledge can be found in the writings of Richard Price (1723–91) and James Madison (1751–1836), the modern study of the diffusion of innovations involves an effort to understand why all who might benefit from it do not instantaneously adopt an innovation. One answer is that the diffusion of information and knowledge is uneven, another is that the value of a new innovation depends upon the characteristics of the adopter, and still a third is that the nature of innovation improves over time, raising the benefit of adoption. Zvi Griliches’ pioneering exploration of the diffusion of hybrid corn involves all three sources of explanation.29 David’s study of the mechanisation of reaping in the antebellum Midwest made two very important contributions to this emerging literature.30 First, he positioned the study as an effort to offer an alternative to the existing explanation in which mechanical reaping was simultaneously the consequence of a shortage of labour required to fully exploit the rise in wheat prices during the 1850s, the extension of the planting of wheat in the American Midwest and the growing prosperity of the wheat farmers. As David observes, this explanation falls well short of explaining why an individual farmer with a particular size farm would choose to replace the more labour-intensive cradle technology with the mechanical reaper. Second, David undertook a comprehensive review of ‘what changed’ and what remained static over the period during which reaper use expanded, reducing
The economics of innovation
17
the model of innovation diffusion to the question of the threshold farm size that would benefit from adopting the new technology. This second contribution was an important advance in the study of diffusion. It explicitly set out the comparison between the technology being replaced and the innovative technology – establishing a path that remains fruitful in the analysis of diffusion of new technologies. David’s approach to the economic analysis of diffusion is best understood by considering the ‘naive’ or ad hoc approach to new technology adoption processes. Many ad hoc approaches simply dismiss theory and take the econometric estimation of a logistic diffusion curve as an application of the ‘well accepted diffusion curve’. Others employ the metaphor of information distribution as contagion, in which a simple specification of the percolation of information through a social network provides a basis for a logistic diffusion curve. Such approaches bypass consideration of the mechanisms of adjustment through learning, improvements of competing technologies and the effects of network externalities that play important roles in David’s later works on diffusion.31 The economic theory of diffusion is often seen by other social scientists as being overly deterministic. Alternative specifications that rely upon cognitive limitations and processes of social negotiation such as those involved in opinion formation are, however, not inherently less deterministic. Theories that rely upon factors that are non-observable before the ‘trial’ of an actual diffusion process are, in essence, explanatory rather than predictive theories. Studies of diffusion based upon these theoretical foundations are necessarily either retrospective – one can only deduce facts about cognitive limits or the percolation of knowledge through social networks in reference of actual experience – or such examinations are speculative – newer technologies may be taken as ‘similar’ to older ones providing a means to make a speculative linkage between past and future experience. The advantage of economic theories of diffusion is that they direct the attention of researchers to the margin, the ordering of individual adopters at the boundary between adoption and non-adoption. As time moves on, examining the shifting margin provides the opportunity to sort out whether the factors responsible for adoption are essentially structural – that is, whether changes in price, quality or the growing availability of complements bound and order the underlying heterogeneity of the adopting population – or whether the margin, the boundary between non-adoption and adoption, is shaped by dynamic processes of change, such as learning, within the non-adopting population. In his famous ‘underground classic’,32 David extended the theory of diffusion, reconciling his own work with that of the other students of the diffusion process such as Griliches and Mansfield. The chapters in this
18
General introduction
volume by B.H. Hall and M. Trajtenberg, P. Stoneman and O. Toivanen and L. Cabral further extend the basic framework building upon the foundations laid by Paul David and other ‘first-generation’ scholars of the modern economic theory of diffusion.
NOTES 1. 2. 3. 4. 5. 6. 7. 8.
9.
10. 11.
12. 13.
14. 15. 16.
P.A. David and D. Foray, ‘Accessing and expanding the science and technology knowledge base’, STI, Review, 16, 1995, 13–68. R. Coase, ‘The lighthouse in economics’, Journal of Law and Economics, 17 (2), 1974, 357–76. P.A. David, a reply to ‘Lock and key’, Economic Focus, The Economist, 18 September 1999, unpublished draft, 2000. P.A. David, ‘Knowledge, property and the system dynamics of technological change’, Proceedings of the World Bank Annual Conference on Development Economics 1992, Washington, DC: World Bank, 1999. P. Dasgupta and P.A. David, ‘Towards a new economics of science’, Research Policy, 23 (5), 1994, 487–521. P.A. David, ‘Common agency contracting and the emergence of open science’, American Economic Review, 88 (2), 1998, 15–21. P.A. David, ‘Communication norms and the collective cognitive performance of “invisible colleges”’, in G. Navaretti, P. Dasgupta, K. Mäler and D. Siniscalco (eds), Creation and Transfer of Knowledge, Heidelberg: Springer, 1998, pp. 115–63. P.A. David, ‘Digital technologies, research collaboration and the extension of protection for intellectual property in science: will building “good fences” really make “good neighbors”?’, IPR Aspects of Internet Collaborations, Final Report, EUR 19456, European Commission, 2001. P.A. David, ‘Learning-by-doing and tariff protection: a reconsideration of the case of the ante-bellum United States cotton textile industry’, in P.A. David, Technical Choice, Innovation and Economic Growth, Cambridge: Cambridge University Press, 1975, pp. 95–173. P.A. David, ‘Path dependence and varieties of learning in the evolution of technological practice’, in J. Ziman (ed.), Technological Innovation as an Evolutionary Process, Cambridge: Cambridge University Press, 1999, pp. 118–33. P.A. David and W. Sanderson, ‘Making use of treacherous advice: cognitive progress, Bayesian adaptation and the tenacity of unreliable knowledge’, in J. Nye and J. Droback (eds), Frontiers of the New Institutional Economics, Oxford: Academic Press, 1997, pp. 305–66. R. Cowan, P.A. David and D. Foray, ‘The explicit economics of knowledge codification and tacitness’, Industrial and Corporate Change, 9 (2), 2000, 211–54. M. Abramovitz and P.A. David, ‘American macroeconomic growth in the era of knowledge-based progress: the long run perspective’, in S. Engerman and R. Gallman (eds), The Cambridge Economic History of the United States, vol. 13, Cambridge: Cambridge University Press, 2000, pp. 1–92. P.A. David, paper delivered at the last TIPIK Conference, Strasbourg, Université Pasteur, March 2001. P.A. David and M. Spence, Towards Institutional Infrastructure for E-Science, Oxford: Oxford Internet Institute, 2003. P.A. David, D. Foray and W.E. Steinmueller, ‘The research network and the new economics of science: from metaphors to organisational behavior’, in A. Gambardella and F. Malerba (eds), The Organisation of Inventive Activity in Europe, Cambridge: Cambridge University Press, 1999, pp. 303–42.
The economics of innovation 17. 18. 19.
20.
21. 22. 23.
24. 25. 26.
27.
28. 29. 30. 31.
19
P.A. David and B.H. Hall, ‘Heart of darkness: public–private interactions inside the R&D black box’, Research Policy, 29, 2000, 1165–83. P.A. David, ‘The political economy of public science: a contribution to the regulation of science and technology’, in L. Smith (eds), The Regulation of Science and Technology, London: Palgrave, 2001, pp. 33–57. M. Abramovitz and P.A. David, ‘American macroeconomic growth in the era of knowledge-based progress: the long run perspective’, in S. Engerman and R. Gallman (eds), The Cambridge Economic History of the United States, vol. 13, Cambridge: Cambridge University Press, 2000, pp. 1–92. David’s first ruminations on this question occur in the extended introduction of Technical Choice, Innovation and Economic Growth: Essays on American and British Experience in the Nineteenth Century, Cambridge and New York: Cambridge University Press, 1975. (Second edition, forthcoming in 2004 from Cambridge University Press.) At a point in time does not mean timeless; economic actors may well have expectations (rational or otherwise) about the future. Their expectations are, however, circumscribed by a set of state spaces that are subject to unanticipated change. P.A. David, ‘Path dependence and predictability in dynamic systems with local network externalities: a paradigm for historical economics’, in D. Foray and C. Freeman (eds), Technology and the Wealth of Nations, London: Pinter, 1992, pp. 209–31. The critics of the idea of path dependence seem particularly alarmed by this prospect, see P.A. David, ‘Path dependence, its critics and the quest for “historical economics” ’, Evolution and Path Dependence in Economic Ideas: Past and Present, edited by P. Garrouste and S. Ionnides, Cheltenham: Edward Elgar, 2001, pp. 15–40. P.A. David, ‘Clio and the economics of QWERTY’, American Economic Review, 75 (2), May 1985, 332–37. P.A. David, ‘Some new standards for the economics of standardization in the information age’, in P. Dasgupta and P.L. Stoneman (eds), Economics and Technological Performance, Cambridge: Cambridge University Press, 1987, pp. 206–39. P.A. David, ‘The hero and the herd in technological history: reflections on Thomas Edison and “The Battle of the Systems” ’, in P. Higgonet, D. Landes and H. Rosovsky (eds), Favorites of Fortune: Technology, Growth, and Economic Development Since the Industrial Revolution, Cambridge, MA: Harvard University Press, 1991, pp. 72–119; P.A. David and J.A. Bunn, ‘Gateway technologies and the evolutionary dynamics of network industries: lessons from electricity supply history’, in A. Heertje and M. Perlman (eds), Evolving Technology and Market Structure, Ann Arbor, MI: University of Michigan Press, 1990, pp. 121–56. P.A. David and W.E. Steinmueller, ‘The ISDN bandwagon is coming – but who will be there to climb aboard? Quandaries in the economics of data communication networks’, Economics of Innovation and New Technology, 1 (1 & 2), Fall 1990, 43–62; P.A. David and D. Foray, ‘Dynamics of competitive technology diffusion through local network structures: the case of EDI document standards’, in L. Leydesdorff and P. van den Besselaar (eds), Evolutionary Economics and Chaos Theory, London: Pinter, 1994, pp. 63–77. P.A. David, ‘The landscape and the machine: technical interrelatedness, land tenure and the mechanization of the corn harvest in Victorian Britain’, in D.N. McCloskey (ed.), Essays on a Mature Economy, London: Methuen, pp. 145–205. Z. Griliches, ‘Hybrid corn: an exploration in the economics of technical change’, Econometrica, 25 (4), 1957, 501–22. P.A. David, ‘The mechanization of reaping in the ante-bellum Midwest’, in H. Rosovsky (ed.), Industrialization in Two Systems: Essays in Honor of Alexander Gerschenkron, New York: Wiley and Sons, 1966, pp. 3–39. For example, P.A. David, ‘Technology diffusion, public policy, and industrial competitiveness’, in R. Landau and N. Rosenberg (eds), The Positive Sum Strategy: Harnessing Technology for Economic Growth, Washington, DC: National Academy Press, 1986, pp. 373–91; P.A. David and P.L. Stoneman, ‘Adoption subsidies vs. information provision
20
32.
General introduction as instruments of technology policy’, Economic Journal, 96 (Supplement), March 1986, 142–50; P.A. David and T.E. Olsen, ‘Equilibrium dynamics of diffusion when incremental technological innovations are foreseen’, Ricerche Economiche (Special Issue on Innovation Diffusion), 40 (4), October–December 1986, 738–70. P.A. David, ‘A contribution to the theory of diffusion’, Stanford University Center for Research in Economic Growth Research Memoranda Nos 71, 72, 73, June 1969, mimeograph.
PART II
Path Dependence in Technical Change
2. Competing technologies, technological monopolies and the rate of convergence to a stable market structure* Andrea P. Bassanini and Giovanni Dosi 1
INTRODUCTION
In this chapter we address the dynamics of diffusion of different technologies competing for the same market niche. The stylised fact at the origin of this work is the observation that a stable empirical pattern of market sharing between competing technologies with no overwhelming dominant position rarely occurs in markets with positive feedbacks.1 For example, even in the case of operating systems, which is often quoted as a case of market sharing, Apple MacIntosh has never held a market share larger than 1/5 (a partial exception being the submarket of personal computers for educational institutions). This fact has also triggered suspicion of market inefficiencies: technological monopolies may prevail even when the survival of more than one technology may be socially optimal (Katz and Shapiro, 1986; David, 1992). Think for example of the competition between Java-based architectures and ActiveX architectures for web-based applets: given that with any of the two paradigms the standard tasks that can be performed are different, the general impression of experts is that society would benefit from the survival of both. In turn, from the point of view of interpretation of the processes of diffusion of new products and technologies, it is acknowledged that, in many modern markets, they are characterised by increasing returns to adoption or positive feedbacks. This has partly to do with supply-side causes: the cumulation of knowledge and skills through the expansion of markets and production usually reduce the hedonic price of both production and consumption goods, thus increasing the net benefit for the user of a particular technology. The Boeing 727, for example, which has been on the jet aircraft market for years, has undergone constant modification of the design and improvement in structural soundness, wing design, payload 23
24
Path dependence in technical change
capacity and engine efficiency as it accumulates airline adoption and hours of flight (Rosenberg, 1982; Arthur, 1989). Similar observations can be made for many helicopter designs (Saviotti and Trickett, 1992) as well as for electric power plants designs (Cowan, 1990; Islas, 1997). Supply-side causes of this type have received some attention in the economic literature for quite a while. However, in the last 15 years a great deal of attention has been devoted also to demand-side positive feedbacks, so-called network esternalities or (more neutrally) network effects (Katz and Shapiro, 1994; Liebowitz and Margolis, 1994). For example, telecommunication devices and networks (for instance, fax machines), as a first approximation, tend not to provide any utility per se but only as a function of the number of adopters of compatible technologies with whom the communication is possible (Rohlfs, 1974; Oren and Smith, 1981; Economides, 1996). The benefits accruing to a user of a particular hardware system depend on the availability of software whose quantity and variety may depend on the size of the market if there are increasing returns in software production. This is the case of video cassette recorders (VCRs), microprocessors, hi-fi devices and in general systems made of complementary products which need not be consumed in fixed proportions (Cusumano et al., 1992; Church and Gandal, 1993; Katz and Shapiro, 1985; 1994). A similar story can be told for the provision of post-purchase service for durable goods. In automobile markets, for example, the diffusion of foreign models has often been slow because of consumers’ perception of a thinner and less experienced network of repair services (Katz and Shapiro, 1985). Standardisation implies also saving out of the cost of investment in complementary capital if returns from investment are not completely appropriable: in software adoption firms can draw from a large pool of experienced users if they adopt software belonging to a widespread standard, thus de facto sharing the cost of training (Farrell and Saloner, 1986; Brynjolfsson and Kemerer, 1996). Moreover product information may be more easily available for more popular brands or, finally, there may be conformity or psychological bandwagon effects (Katz and Shapiro, 1985; Banerjee, 1992; Arthur and Lane, 1993; Bernheim, 1994; Brock and Durlauf, 1995). Katz and Shapiro (1994) in their review of the literature on systems competition and dynamics of adoption under increasing returns distinguish between technology adoption decision and product selection decision. The former refers to the choice of a potential user to place a demand in a particular market. Relevant questions in this case are the conditions for an actual market of positive size, the notional features of a ‘socially optimal’ market size and the conditions allowing penetration of a new (more advanced) technology into the market of an already established one (Rohlfs, 1974; Oren and Smith, 1981; Farrell and Saloner, 1985; 1986; Katz
Competing technologies, technological monopolies
25
and Shapiro, 1992). For example, purchasing or not a fax or substituting a compact disc player for an analogical record player are technology adoption decisions. Conversely product selection refers to the choice between different technological solutions which perform (approximately) the same function and are therefore close substitutes. Relevant questions here are whether the market enhances variety or standardisation, whether the emerging market structure is normatively desirable and what is the role of history in the selection of market structure (Arthur, 1983; 1989; Katz and Shapiro, 1985; 1986; David, 1985; Church and Gandal, 1993; Dosi et al., 1994). Choosing between VHS or Beta in the VCR market or between Word or Wordperfect in the word-processor market are typical examples of product selection decisions. This work is concerned with the dynamics of product selection. To explain the stylised fact recalled above we analyse properties of a fairly general and nowadays rather standard class of models of competing technologies, originally suggested by Arthur (1983) and Arthur et al. (1983) and further explored by Arthur (1989), Cowan (1991) and Dosi et al. (1994), among others. This class of models will be presented in details in section 2. Despite mixed results of some pioneering works on the dynamics of markets with network effects (for example, Katz and Shapiro, 1986), unbounded increasing returns are commonly called for as an explanation of the emergence of technological monopolies. Usually the argument is based on the results of the model set forth by Arthur (1989). For instance, Robin Cowan summarises it in the following way: If technologies operate under dynamic increasing returns (often thought of in terms of learning-by-doing or learning-by-using), then early use of one technology can create a snowballing effect by which that technology quickly becomes preferred to others and comes to dominate the market. Following Arthur, consider a market in which two types of consumers adopt technology sequentially. As a result of dynamic increasing returns arising from learning-by-using, the pay-off to adopting a technology is an increasing function of the number of times it has been adopted in the past. Important with regard to which technology is chosen next is how many times each of the technologies has been chosen in the past. Arthur shows that if the order of adopters is random (that is, the type of the next adopter is not predictable) then with certainty one technology will claim the entire market. (Cowan, 1990: 543, italics added)
It will be shown in the following that this statement does not always hold. Unbounded increasing returns to adoption are neither necessary nor sufficient to lead to the emergence of technological monopolies. As proved in the next section, strictly speaking, Arthur’s result applies only when returns are linearly increasing and the degree of heterogeneity of agents is, in
26
Path dependence in technical change
a sense, small. Moreover it cannot be easily generalised further: some meaningful counter-examples will be provided. More generally the emergence of technological monopolies depends on the nature of increasing returns and their relationship with the degree of heterogeneity of the population. Given a sufficiently high heterogeneity amongst economic agents, limit market sharing may occur even in the presence of unbounded increasing returns. The bearing of our analysis, in terms of the interpretation of the empirical evidence, stems from the results presented in section 3: in essence, we suggest that the observation of the widespread emergence of monopolies is intimately related to the properties of different rates of convergence (to monopoly and to market sharing respectively) more than to the properties of limit states as such. It will be shown that a market can approach a monopoly with a higher speed than it approaches any feasible limit market shares where both technologies coexist. Following a line of reasoning put forward by Winter (1986), our argument proceeds by noticing that when convergence is too slow the external environment is likely to change before any sufficiently small neighbourhood of the limit can be attained. The result that we obtain, based on some mathematical properties of Generalised Urn schemes,2 is general for this class of models. The empirical implication is that among markets with high rate of technological change and increasing returns to adoption, a prevalence of stable monopolies over stable market-sharing should be observed. The applications of Arthur’s result have gone far beyond the dynamics of competing technologies and typically extended to the role of history in selecting the equilibrium in any situation wherein complementarities are relevant. The analysis of industry location patterns is a case to the point (for example, Arthur, 1990; Krugman, 1991a; 1991b; Venables, 1996). As James Rauch puts it: In Arthur’s model, firms enter the industry in sequence. Each firm chooses a location on the basis of how many firms are there at the time of entry and a random vector that gives the firm’s tastes for each possible location. If agglomeration economies are unbounded as the number of firms increases, then as the industry grows large, one location takes all but a finite set of firms with probability one. (Rauch, 1993: 843–4, italics added)
The implications of our results extend to this domain of analysis as well. The remainder of the chapter is organised as follows. Section 2 reviews standard models of competing technologies and provides counter-examples to Arthur’s main result. Section 3 establishes our main results on rate of convergence to a stable market structure and builds upon that an alternative explanation for observable patterns of dynamics between competing technologies. Section 4 briefly summarises the results.
Competing technologies, technological monopolies
2
27
COMPETING TECHNOLOGIES REVISITED: ARE UNBOUNDED INCREASING RETURNS SUFFICIENT FOR THE EMERGENCE OF TECHNOLOGICAL MONOPOLIES?
The class of competing technology dynamics models that we consider shares the two basic assumptions that adopters enter the market in a sequence which is assumed to be exogenous, and that each adopter makes its adoption choice only once. More than one agent can enter the market in each period (for example, Katz and Shapiro, 1986) but in order to simplify the treatment we abstract from this complication. The simple theoretical tale that underlies these models can be summarised as follows. Every period a new agent enters the market and chooses the technology which is considered best suited to its requirements, given its preferences, information structure and the available technologies. Preferences can be heterogeneous and a distribution of preferences in the population is given. Information and preferences determine a vector of pay-off functions (whose dimension is equal to the number of available technologies) for every type of agent. Because of positive (negative) feedbacks, such as increasing (decreasing) returns to adoption, these functions depend on the number of previous adoptions. When an agent enters the market it compares the values of these functions (given its preferences, the available information, and previous adoptions) and chooses the technology which yields the maximum perceived pay-off. Which ‘type’ of agent enters the market at any given time is a stochastic event whose probability depends on the distribution of types (that is, of preferences) in the population. Because of positive (negative) feedbacks, the probability of adoption of a particular technology is an increasing (decreasing) function of the number of previous adoptions of that technology. More formally we can write a general reduced form of pay-off functions of the following type: → → ij( n (t)) hi(aij, n (t)), where j D, D is the set of possible technologies, i S; S is the set of possible types; → n (t) is a vector denoting number of adoptions for each technology at time t (n j(t)) is the number of adoptions of technology j at time t); → a i represents the network-independent components of agent i’s preferences (a ji identifies a baseline pay-off for agents of type i from technology j), and hi() is an increasing (decreasing) function that can differ across agents capturing increasing (decreasing) returns to adoption. Information and expectations are incorporated in hi(). If, at time t, an agent of type
28
Path dependence in technical change
i comes to the market, it compares the pay-off functions, choosing A if and only if:3 → → iA( n (t)) arg max{ ij ( n )} jD
(2.1)
Equation (2.1) can be seen as describing an equilibrium reaction function. Consequently, strategic behaviours (including sponsoring activities from the suppliers of technologies) are not ruled out by the foregoing formalisation. In the remainder of this chapter we assume that the order of agents entering the market is random, hence i(t) can be considered as an iid sequence of random variables whose distribution depends on the distribution of the population of potential adopters. Under this assumption, the dynamics of the foregoing model can be analysed in terms of generalised urn schemes. Consider the simplest case where two technologies, say A and B, compete for a market. Let us denote A’s market share with X(t). Given the relationships between (a) total number of adoptions of both technologies n(t)
t 1 nA(0) nB(0), (b) the current market share X(t) of A, and (c) number of adoptions of one specific technology, ni(t), i A, B, that is, nA(t)
n(t)X(t), the dynamics of X(t) is given by the recursive identity X(t 1) X(t)
t(X(t)) X(t) . t nA(0) nB(0)
Here t (x), t 1 are random variables independent in t such that
t(x)
1 with probability 0 with probability
f (t, x) , 1 f (t, x)
and t() is a function of market shares dependent on the feedbacks in adoption. f(t, x) equals the probability that (2.1) is true when X(t) x and is sometimes called urn function. Denoting t(x) E( t(x))
t(x) f(t, x) with t(x) we have X(t 1) X(t)
[ f (t, X(t)) X(t)] t(X(t)) . t nA(0) nB(0)
(2.2)
Provided that there exist a limit urn function f() (defined as that function f() such that f(t,.) tends to it as t tends to ) and the following condition is satisfied
t 1 t1
sup
| f (t, x) f (x)| ,
x|0,1|R(0,1)
(2.3)
Competing technologies, technological monopolies
29
where R(0,1) is the set of rational numbers in (0,1), limit market shares attainable with positive probability can be found by analysing the properties of the function g(x) f (x) x lim f (t, x) x. t→
Particularly, treating g(x) in the same way as the right-hand side of an ordinary differential equation, it is possible to show that the process (2.2) converges almost surely to the set of stable zeros.4 The foregoing formal representation is employed for every result of the present chapter. In some cases, equation (2.1) can be expressed directly in terms of shares rather than total numbers: in this case f(.,.) is independent of t and (2.3) is easily verified. The foregoing formal model can be better visualized by looking at some well-known example. Consider for instance the celebrated example of the VCR market. JVC’s VHS and Sony’s Beta were commercialised approximately at the same time. According to many studies (see Cusumano et al., 1992; Liebowitz and Margolis, 1994), none of the two standards has ever been perceived as unambiguously better and, despite their incompatibility, their features were more or less the same, due to the common derivation from the U-matic design. The relevant decisions were likely to be sequential. First, a consumer chooses whether or not to adopt a VCR – technology adoption decision in Katz and Shapiro’s terminology. Then, once the adoption decision has been made, it devotes its mind to choose which type of VCR to purchase – product selection decision – (in general it can be expected that most of the consumers buy one single item and not both). Network effects in this market come mainly from increasing returns in design specialisation and production of VCR models (so that historically all firms specialised just in one single standard) on the supply side, and from increasing returns externalities and consequent availability of home video rental services on the demand side (Cusumano et al., 1992). Despite technical similarities between the two standards, preferences are likely to be strongly heterogeneous, due mainly to a brandname-loyalty in consumer behaviour, which was in fact exploited (especially by JVC) through original equipment manufacturers’ (OEM) agreements with firms with well-established market shares in electronic durable goods. The size of VCR market is sufficiently large (hundreds of millions of sold units) to make it approximable by the abstract concept of an infinite capacity market. Therefore, the asymptotic dynamics of this market may be meaningfully analysed through the asymptotics of generalised urn schemes.
30
Path dependence in technical change
Many other markets display somewhat similar characteristics (for instance, spreadsheets, wordprocessors, computer keyboards, Pc hardware, automobiles and so on). In particular, in many markets product selection can be assumed to sequentially follow technology adoption decisions.5 The fact that decisions are sequential suggests that product selection decisions might be dependent on market shares rather than on the absolute size of the network. In this case the urn scheme would be even more simplified, with the urn function independent of t. Arthur (1983; 1989) considers a pay-off function of the following type: n (t)) aij r(n j(t)), ij( → where j A, B, iS, S is the set of possible types [in the simplest case, considered also in the above quotation from Cowan (1990), S {1,2}], and r is an increasing function (common for every agent) capturing increasing returns to adoption. If, at time t, an agent of type i comes to the market, it compares the two pay-off functions choosing A if and only if: n (t)) iB( → n (t)), iA( → that is aAi r(nA(t)) aBi r (nB(t)).
(2.4)
Suppose that the selection of which type of agent enters the market at time t is the realisation of an iid random variable i(t). Thus (2.2) implies that the agent coming to the market chooses A with probability P(A(t)) F0(r(nA(t) r (nB(t))), where F() denotes the distribution function of (t) aBi (t) aAi (t). From these considerations Arthur’s main theorem was derived: Theorem 1 (Arthur (1989), Theorem 3) If the improvement function r increases at least at rate 0 as nj increases, the adoption process converges to the dominance of a single technology, with probability one. The proof of the theorem offered by Arthur is based on theorem 3.1 of Arthur et al. (1986). In fact it is easy to check that in this case, whatever the distribution of aj is, the limit urn function f() is a step function defined in the following way:
Competing technologies, technological monopolies
1 f (x) F(0) 0
if if if
x 12 x 12 x 12
31
(2.5)
A generaliszed urn scheme characterised by an urn function such as (2.5) converges to {0,1} with probability 1.6 However theorem 3.1 of Arthur et al. (1986) is not applicable here because condition (2.3) does not hold in this case. Actually the urn functions are defined by: ft(x) F(r(x(t nw nb)) r((l x)(t nw nb))). Moreover for tK0, t even, they are such that ft(0) 0, ft(l/2) F(0), ft(l) 1 and they are continuous in a left neighbourhood (which depends on t) of 1/2; therefore sup x[0,1]R(0,1)
| ft(x) f(x)| min{F(0), 1 F(0)}
which is constant with respect to t.7 Even though Arthur’s proof is wrong, the theorem is right and an ad hoc proof can be constructed by showing that nA(t) nB(t) is a timehomogeneous Markov chain with two absorbing barriers (Bassanini, 1997, proposition 2.1, provides a complete proof along these lines). However this result strictly depends on the fact that the function r() is asymptotically linear or more than linear. Arthur’s result is not generalisable to any type of unbounded increasing returns. Both in the case of increasing returns that are diminishing at the margin and in the case of heterogeneous increasing returns it is possible to find simple examples where convergence to technological monopolies is not an event with probability 1. Let us illustrate all this by means of two straightforward counterexamples. Example 1 Let us assume that increasing returns have the common sense property that the marginal contribution to social benefit of, say, the 100th adopters is larger than that of, say, the 100 000th and that this contribution tends d d2 asymptotically to zero; formally this means that dn j f (n j) 0, dn j 2 f (n j ) 0 and limn j →dnd j f (n j ) 0 (this class of functions has been considered by Katz and Shapiro, 1985). Focusing on the case set forth by Robin Cowan in the above quotation, let us assume that there are only two types of agents (i 1, 2) and two techn (t)) aij r (n j(t)), nologies. Recall Arthur’s pay-off functions (2.4), ij ( → and assume that r(·) s log(·) is a function (which is common for every agent:
32
Path dependence in technical change
s is a constant) that formalises unbounded increasing returns to adoption. Agent i chooses technology A if and only if iA( → n (t)) iB ( → n (t)). By taking the exponential on both sides and rearranging we have: X(t) 1 i i es (aB(t) aA(t)). 1 X(t)
(2.6)
The function of the attributes of agent’s type which is on the right-hand side can be considered a random variable because, as discussed above, i(t) is a random variable. Moreover such agent characteristics are iid because i(t) is iid. Denoting the random variables on the right-hand side with (t), from (2.6) we have that the adoption process can be seen as a generalised urn scheme with urn function given by: (2.7)
f(x) F,(x/(l x)),
where F(·) is the distribution function of (t). Because i(t) takes just two values (1, 2), also (t) takes just two values: (t)
es (aB a A) 1 2 2 es (aB a A) 1
1
1
1
with probability with probability
where we have assumed without loss of generality that a1B a1A a2B a2A. Thus F is by construction a step function with two steps:
0 if y es (a B a A) 1 1 1 2 1 2 F(y) if es (a B a A) y es (a B a A) 1 2 2 1 if y es (a B a A) 1
1
1
Therefore, taking into account (2.7), we have that the urn function has two steps and is defined in the following way:
es (a B aA) 1 1 1 1 es (aB a A) 1
0
x
if
1
es (a B a A) es (a B aA) 1 1 1 1) x 2 2 . (a a 1 es B A 1 es (a B aA) 1 2 2 es (aB aA) x 1 2 2 1 es (aB a A) 1
f (x) if 1
1
if
1
1
1
If the following condition is satisfied es (aB aA) es (aB aA) 1 1 1 1 2 2 1 es (a B a A) 1 es (a B aA) 1
1
1
1
2
2
2
2
Competing technologies, technological monopolies
33
the urn function has five fixed points, three of which are down-crossing, therefore there is a set of initial conditions (that imply giving both technologies a chance to be chosen ‘at the beginning of history’) for which market sharing is asymptotically attainable with positive probability.8 The i i above condition imply that the ratios (e aA)1r(e aB)1r are sufficiently different between the two types. In other words there might be sufficient heterogeneity among agents to counterbalance the effect of increasing returns to adoption.9 ˇ
Example 2 Consider now pay-off functions of this type: n (t)) aj rj n j, j( → where rj, aj, j A, B, are bounded random variables which admit density. Such functions allow agents to be heterogeneous also in terms of the degree of increasing returns which they experience. By applying (2.1), dividing pay-off functions by total number of adoptions, and rearranging we have that A is chosen if and only if: X(t)
rB aB a A . A rA rB (t n (0) nB(0))(rA rB)
(2.8)
Denoting the random variables on the right-hand side with (t), from (2.8) we have that the adoption process can be seen as a generalised urn scheme with urn function f(t, x) F(t)(x), where F(t)(·) is the distribution function of (t). Now suppose that rA and rB are highly correlated and both have bimodal distributions very concentrated around the two modes, in such a way that the distribution of rA/rB is also bimodal and very concentrated around the two modes too. Furthermore suppose that the two modes are far away from each other. To fix the ideas say that for a percentage of the population rA/rB, is uniformly distributed on the interval [1/(1 b), 1/(1 a)], while for a percentage 1 of the population rA/rB is uniformly distributed on the interval [1/(1 d), 1/(1 c)], with 0 abcd. First, let us consider the case of aj 0, j A, B. F is by construction independent of t, implying the following urn function:
0 b 1 a (y a) f(x) F(t)(x) (1 )[d 1 c (y c)] 1 ˇ
if if if if if
y a ay b by c cy d yd
34
Path dependence in technical change
If bc, then there are three stable fixed point of f(x) and, as stated above, it can be shown that there is a set of initial conditions (that imply giving both technologies a chance to be chosen ‘at the beginning of history’) for which market sharing is asymptotically attainable with positive probability. If aj 0 but has bounded support and admits density, then condition (2.3) applies and the same argument holds: in fact, relying on the fact that rj , aj are bounded it is easy to show that supx[0,1]R(0,1)| f (t, x) f (x)| Kt, where K0 is a constant. The essential ingredient of this example is that the distribution of rA/rB is bimodal and very concentrated around the two modes. The argument has nothing to do with the particular (and extreme) distributional form assumed above: following the same constructive procedure adopted here it is easy to build examples with any other distributional form. The only requirement is that the two modes are sufficiently distant. In other words the only requirement is a sufficient degree of heterogeneity in the population to counterbalance the pro-standardisation effects of increasing returns to adoption. The two examples above show that the degree of increasing returns needs to be compared to the degree of heterogeneity. Unbounded increasing returns that are diminishing fast at the margin are not sufficient to generate asymptotic survival of only one technology, provided that agents are not completely homogeneous (see also Farrell and Saloner, 1985; 1986, for early models with homogeneous agents that lead to the survival of only one technology). Even more interesting, when heterogeneity is so wide that there are agent-specific increasing returns, the emergence of technological monopolies is not guaranteed even with returns that are linearly increasing. To summarise, the foregoing examples show that if preferences are sufficiently heterogeneous and/or increasing returns to adoption are less than asymptotically linear, then Arthur’s result cannot be generalised and variety in the asymptotic distribution of technologies can be an outcome with positive probability. From the point of view of empirical predictions, at first look, the foregoing results might sound, if anything, as a further pessimistic note on ‘indeterminacy’. That is, not only ‘history matters’ in the sense that initial small events might determine which of the notional, technologically attainable, asymptotic states the system might ‘choose’: more troubling, the argument so far suggests that, further, the very distribution of the fine characteristics and preferences of the population of agents might determine the very nature of the attainable asymptotic states themselves. Short of empirically convincing restrictions on the distribution of agents (normally unobservable) characteristics, what we propose is instead an interpretation of the general occurrence of technological monopolies (cum increasing returns of some ˇ
ˇ
35
Competing technologies, technological monopolies
kind) grounded on the relative speed of convergence to the underlying (but unobservable) limit states.
3
RATE OF CONVERGENCE IN ONE-DIMENSIONAL MODELS OF COMPETING TECHNOLOGIES
In the example of the VCR market, as well as for many other markets, the possibility of predicting limit market shares depends on the feasibility of formalising the structure of the market in question in terms of a specific urn function. Heterogeneity of preferences, the degrees of increasing returns, the type of expectations, price-policies of producers, all affect the functional form of the urn function. As mentioned before, the goal of this chapter in general and of this section in particular is to provide some general asymptotic results that can be used as guidance for the interpretation of the empirical evidence on emergence of dominant designs. Propositions 2, and 4 suffice for the task. Together they imply the relevant statements on the rate of convergence to technological monopoly or to a limit market share where both technologies coexist.10 Furthermore the analysis that follows applies even in the absence of a clear pattern of increasing returns to adoption. In essence, in the presence of constant returns to adoption, the urn function would be completely constant but the following theorems would still hold. As above, denote the urn function with f(·,·); the following proposition gives a first result on the rate of convergence to 0 and 1. Proposition 1 Let 0 and c1 be such that eventually f(t,x)cx
for x(0, )
( f(t,x))1 c(1 x) for x(1 ,1)). Then for any (0,1 c) and 0 limt→P{t1 c X(t) |X(t) → 0} 1 (limt→P{t1 c [1 X(t)] |X(t) → 1 1), where X(·) stands for the random process given by (2.2). The proposition is proved in the appendix (section 5).
(2.9)
36
Path dependence in technical change
A similar result can be expressed in terms of variances (L2 convergence): Proposition 2 Let 0 and c 1 be such that eventually (2.9) holds. Then for any (0,1 c)
limt→t2(1 c) P{X(t) → 0}
X(t)2dP 0
{X(t)→0}
(limt→t2(1 c) P{X(t) → 1}
(X(t) 1)2dP 0
{X(t)→1}
The proposition is proved in the appendix (section 5). Notice that proposition 2 states that the rate of convergence of the mean square distance from the limit market share is of the order of 1/t as t → , conditional to the fact that the process is actually converging to 0 or 1. Roughly speaking it defines the rate of convergence of mean square errors when the process converges to a technological monopoly. If the set of limit market shares that the process can reach with positive probability contains only these two points, proposition 2 implies a similar statement in terms of the unconditional mean square distance from the limit market share. One would like to derive a counterpart of proposition 1 and 2 for the case of market sharing, whenever this can be attained with positive probability. For a differentiable f(·) at 0(1), (2.9) holds with c arbitrarily close to d d dx f (0) (dx f (1)).We can easily derive a similar result for a differentiable f(·) independent of t from the following conditional limit theorem for the generalised urn scheme. Theorem 2 (Arthur et al., 1987) Let (0,1) be a stable root of f(x) x 0 and f(·) is differentiable at with d dx f () 12. Then for every y ( , ) ˇ
ˇ
lim P
t→
√
t
d 1 2dx f () [X(t) ] y|X(t) → (y), 2(1 )
where (·) stands for the Gaussian distribution function having zero mean and variance 1. From this theorem, we can give an even better characterisation of the lowest possible convergence rate for a limit market share where both technology coexist that can be attained with positive probability. Indeed, the next proposition follows immediately:
37
Competing technologies, technological monopolies
Proposition 3 Let (0,1) be a stable root and let [ f(x) ](x )k(x )2
for
x( , ), x
take place for some 0 and k1/2. Then for every , 0 lim P{t12 |X(t) | |X(t) → 0} 0.
t→
Differentiability of the urn function at the limit point is a highly demanding restriction, as well as the fact that the urn function has to be independent of t. As said before, several actual markets can present overwhelming problems of formalisation. Consequently, it may be impossible to check these conditions, albeit intuitively there is no reason why differentiability should matter. Conversely, we can obtain a general result in terms of L2 convergence that suffices to the task: Proposition 4 Let (0,1) be such that [ f (t, x) ](x ) k(x )2
x( , ), x (2.10)
for
˛
˛
takes place eventually for some 0 and k1. Then for every 0 lim t 2min{1 k, 12} P{X(t) → } ˛
t→
˛
˛
(X(t) )2dP 0.
{X(t)→}
The proposition is proved in the appendix (section 5). Propositions 2 and 4 show that convergence to 0 and 1 can be much faster (almost of order 1/t as t → ) than to an interior limit (which can be almost of order 1 √t only).11 Here t stands for the number of adoptions to the urn. That is, we are talking about relative rates (the ideal time which is considered here is the time of product selection choices). This result is, however, stronger than it may seem at first glance. In fact it has also implications for the patterns of product selection in ‘real’ (empirical) time where plausibly the speed of the market share trajectory depends also on technology adoption decisions. There is much qualitative evidence and some econometric results [for example, Koski and Nijkamp, 1997) showing that technology adoption is at the very least independent of market shares if not enhanced by increasing asymmetry in their distribution. Thus a fortiori we can conclude that there is a natural tendency of this class of processes to converge faster to 0 or 1 rather than to an interior limit. The explanation is that the variance of t(x), which characterises
38
Path dependence in technical change
the level of random disturbances in the process (2.2), is f(t, x)(1 f(t, x)). Under condition (2.9) this value vanishes at 0 and 1 but it does not vanish at (0,1), being equal to (1 ), under condition (2.10). Notice also that in example 1 c 0 and in example 2 c 0. As shown in the previous section, the urn function can have any shape and there is no reason to believe that problems characterised by 0 and 1 as the only stable points are the only ones that we can expect. Therefore, in principle, an asymptotic outcome where both technologies survive should be observable with positive frequency in real markets. As discussed in the previous section the tendency to converge to market sharing or technological monopolies is an outcome induced by the relative impact of heterogeneity of preferences and increasing returns to adoption. What tendency is realised depends on which of the two prevails. Notice, however, that the prevalence of one of the two factors is not always predictable ex ante even for a nearly omniscient agent fully aware of all fundamentals of the economy: in the examples of the previous section both type of outcomes are possible, but which one is realised depends on the actual sequence of historical events that lead to it. In this type of models, in general, when multiple asymptotic equilibria are attainable, history plays a major role in the selection of the actual one.12 If asymptotic patterns were observable, the results of the previous section would imply that we should observe both stable market sharing and technological monopolies. However, for the interpretation of empirical stylised facts, the point where the process eventually would converge may be irrelevant. Indeed, the rate of change of the technological and economic environment can be sufficiently high that one can always observe diffusion dynamics well short of any meaningful neighbourhood of the limit it would have attained under forever constant external conditions. So while it is true that a convergent process should generate a long-lasting stable pattern, the time required to generate it may be too long to actually observe it: the world is likely to change well before convergence is actually attained. In a sense these changes can be viewed as resetting the game to its starting point. On the basis of the propositions of this section we notice that convergence to technological monopolies tends to be much faster (in probabilistic terms) than to any stable market sharing where both technologies coexist, because of the intrinsic variability that market sharing carries over. Thus the empirical prediction of these results can be stated as follows: in markets with increasing returns to adoption and a high rate of technological change we expect to observe a prevalence of both unstable market sharing (persistent fluctuations in the market shares) and stable technological quasimonopolies as compared to stable patterns of market sharing. The reason
Competing technologies, technological monopolies
39
for this is that technological monopolies can be easily attained in a reasonably short time, that is, sufficiently before any significative change in the underlying basic technological paradigms takes place (Dosi, 1982). Finally note that the observation of the frequent emergence of different monopolies in different related markets (for example, different geographical areas) does not contradict our empirical predictions. Of course, it is trivially true that, with mutually independent markets, different trajectories could emerge in different markets as if they were different realisations of the same experiment. In a related paper (Bassanini and Dosi, 1999a) we show that the foregoing results can be extended also to the case when markets are interdependent: not contrary to the intuition, it is the balance between local and global feedbacks which determines whether the system converges to the same or different monopolies in every market. However, even though at high level of aggregation a system of different local monopolies looks like a stable market sharing, it is shown there that it has the same rate-of-convergence properties of a ‘univariate’ system converging to a monopoly.
4
CONCLUSIONS
This Chapter has reassessed the empirical evidence on prevalence of technological monopolies over market sharing in the dynamics of competing technologies. First, we have argued that the dominant explanation in the literature, namely that unbounded increasing returns can be identified as the factor responsible for this pattern, does not always hold. Brian Arthur’s results – we have shown – hold only when increasing returns to adoption are linear or more than linear and the degree of heterogeneity of agents is small. The presented counter-examples suggest that asymptotic patterns of the dynamics of competing technologies depend on the relative impact of (unbounded) increasing returns and the degree of heterogeneity of the population of adopters. Second, given all this, we propose, however, that in a market with high technological dynamism, no interesting predictions can be made by simply looking at theoretical asymptotic patterns. If convergence is too slow the environment may change before the limit can be actually approached. Conversely, developing upon some mathematical properties of Polya urns, we show that convergence to technological monopolies tends to be (in probabilistic terms) much faster than to a limit where both technologies coexist, the empirical implication being that in markets with high turnover of basic technologies, a prevalence of technological monopolies over stable market sharing is likely to be observed.
40
Path dependence in technical change
APPENDIX For the purpose of exposition, to keep the notation simple, all proofs are exposed for generalised urn scheme involving time-independent urn functions. They can be easily repeated for the general case. Proof of Proposition 1 Consider only the first case – convergence to 0. Without loss of generality we can assume that P{X(t) → 0} 0. Indeed the theorem, being a statement about the convergence rate to 0, does not make any sense if X(·) does not converge to 0. Let Z(·) be a conventional urn process with cx as the urn-function and the same initial numbers of balls n nw nb. Then
EZ(t 1) 1
1 c EZ(t), t 1, t n
and consequently t 1
c 1
1 c EZ(t) EZ(1) t 1 j 1 (1 j n ) EZ(1)e
˛
˛
j 1
1 (n j)
n t 1
c 1
EZ(1)e
˛
˛
n 1
1 x dx
EZ(1)(n n t 1 1)1 c EZ(1)tc 1[1 ot(1)],
where ot(l) → 0 as t → . Hence from Chebychev’s inequality P{t1 c Z(t) } → 0
as
t→
(2.11)
for every (0,1 c) and 0. For arbitrary (0, ) and v0 there is N depending on these variables such that P{{X(t) → 0} {X(s) , s N}} v where A B (A\B)(B\A). Also since Z(t) → 0 with probability 1 as t → , we can choose this N so large that ˛
˛
P{{X(t) → 0} {X(s) , Z(s) , s N}} v. To prove the theorem it is enough to show that P{t1 c X(t) , X(t) → 0} → 0,
(2.12)
41
Competing technologies, technological monopolies
or, taking into account that v in (2.12) can be arbitrary small, that P{t1 c X(t) , X(s) , Z(s) , s N} → 0.
(2.13)
However P{t1 c X(t) , X(s) , Z(s) , s N
X ySN
P
t1 c X(t) , X(s) , |X(N ) y Z(s) , s N
P{X(N) y},
(2.14)
X(1)(n i) where SX t { n t 1 , 0 i t 1} is the set of values that X(t) can attain (not necessarily with positive probability). Due to lemma 2.2 of Hill et al. (1980), there exists a probability space such that Z(·) dominates X(·) on the event Z(t), tN, providing that these processes start from the same point. Therefore, for any y SX N
P{t1 c X(t) , X(s) , Z(s) , s N | X(N) y} ˛
P{t1 c Z(t) ,
˛
X(s) , Z(s) , s N | X(N) y}. ˛
˛
However for every y SX N P{t1 c Z(t) , X(s) , Z(s) , s N} P{t1 c Z(t) } → 0 as t →by (2.11). Thus (2.14) is a sum of a finite number – namely N – of terms each converging to zero. This completes the proof. Proof of Proposition 2 As before, consider only the first case – convergence to 0. Let Z(·) be a conventional urn process with cx as the urn-function and the same initial numbers of balls n nw nb. Then E [Z(t 1)2 |Z(t)] Z(t)2 n 2 t (c 1)Z(t)2 (n 1 t) 2 (c 1) 2Z(t)2 (n 1 t)2c(1 c) 2Z(t) 2
42
Path dependence in technical change
and
E [Z(t 1)2] 1
2(1 c) 1 c E [Z(t)2], t n (t n)2
t 1,
and consequently 2(1 c) E [Z(t)2] E [Z(1)2] t 1 j 1 [1 j n oj(1 j)] ˛
t 1
E [Z(1)2]e
2(c 1)
j 1
1 ( n j)
n t 1
E [Z(1)2]e 2(c 1) n 1
1 x dx
n 1 2(1 c)
E [Z(1)2[t2(c 1)]1 ot(1)],
E [Z(1)2]( n t 1)
where tsot(1/ts) → 0 as t → . Hence: t2(1 c )E [Z(t)2] → 0
as
t→
(2.15)
for every (0,1 c). For arbitrary (0, ) and v0 there is N depending on these variables such that P{{X(t) → 0} {X(s) , s N}} v, where A B (A\B)(B \A). Also since Z(t) → 0 with probability 1 as t → , we can choose this N so large that ˛
˛
P{{X(t) → 0} {X(s) , Z(s) , s N}} v.
(2.16)
To prove the theorem it is enough to show that t2(1 c )
X(t)2dP → 0,
{X(t)→0}
or, taking into account that v in (2.16) can be arbitrary small, that
t2(1 c )
X(t)2dP → 0,
(2.17)
{X(s) , Z(s) , s N}
However, t2(1 c )
X(t)2dP
{X(s) , Z(s) , s N}
t 2(1 c )
X y SN
{X(s) , Z(s) , s N, X(N) y}
X(t)2dP
(2.18)
Competing technologies, technological monopolies
43
X(1)(n i) where SX t { n t 1 , 0 i t 1} is the set of values that X{t) can attain (not necessarily with positive probability). Due to lemma 2.2 of Hill et al. (1980), there exists a probability space such that Z(·) dominates X(·) on the event Z(t), tN, providing that these processes start from the same point. Therefore, for any tN and y SX N
t2(1 c )
X(t)2dP
{X(s) , Z(s) , s N, X(N) y}
t2(1 c )
Z(t)2dP
{X(s) , Z(s) , s N, X(N) y}
However, for every y SX N t2(1 c )
Z(t)2dP
{X(s) , Z(s) , s N, X(N) y}
t2(1 c )E [Z(t)2] → 0 as t →by (2.15). Thus (2.18) is a sum of a finite number – namely N – of terms each converging to zero. This completes the proof. Proof of Proposition 4 The proof is based on the following lemmas:13 Lemma 1 Let f(.) be the urn function of the process Z(t) such that [ f(x) ](x )k(x )2 for some k1 and (0,1) and f(x)[l f(x)]0. Then limt→ dt
limt→ E(Z(t) )2, where
(n t) 1 if 2(1 k) 1 0 dt K (n t) 1 log(n t) if 2(1 k) 1 0 (n t) 2(1 k) if 2(1 k) 1 0 where K is a constant term.
44
Path dependence in technical change
Proof Consider the process (2.2) and write n nw nb, then: E [(Z(t 1) )2]X(t)] (Z(t) )2 n 2 t [ f (Z(t)) Z(t)](Z(t) ) (n 1 t)2 [ f (Z(t)) Z(t)]2 (n 1 t)2 f (Z(t))[1 f (Z(t))]. Setting t E(Z(t) )2, from the assumptions of the lemma, taking into account that f(Z(t))[l f(Z(t))], and that [ f (Z(t)) Z(t)](Z(t) ) [ f (Z(t)) ](Z(t) ) (Z(t) )2, we have
t 1 t 1
2(1 k) n t (n t)2
Thus t 1 t
1 i 1
2(1 k) n i
t
t
(n i)2j i 1 1 t
t
i 1
2(1 k) n j
Since
t
j i 1
1
1 2(1 k) ( n j )
e 2(1 k)j i 1 (1 it) n i
e 2(1 k)[log(n t 1) log(n i)](1 it)
n i n t
2(1 k)
(1 it),
where it and it are small terms (ot(1)) not necessarily non-negative,14 then t 1 t
n 1 n t
2(1 k)
(1 1t)
(n t) 2(1 k)
t
(n i) 2 2(1 k)(1 it). i 1
Since, in terms of asymptotic behaviour, t
i 1
(n i) 2 2(1 k)
1 2(1 k) 1 2(1 k) 1 (n t)
log(n t)
if 2(1 k) 1 0 if 2(1 k) 1 0
Competing technologies, technological monopolies
45
we have that
1 1 if 2(1 k) 1 (n t)
2(1 k) 1 0 t 1 K (n t) 1 log(n t) if 2(1 k) 1 0 , 1(n t) 2(1 k) if 2(1 k) 1 0 which implies the statement of the lemma. Set
f ( ) if x f(x) f (x) if x , f ( ) if x t
j (X (t)) X (t) , t n X (1) X(1).
X (t 1) X (t)
Then, with probability 1, X (t) → as t → . Also by lemma 1, t2min{1 k,12} E [(X (t) )2] → 0 as t → .
(2.19)
As for proposition 2, we can ignore the case when X(t) does not converge with positive probability. We have to show that as t → t2min{1 k,12}
(X(t) )2 dP → 0,
{X(t)→}
For every 0 there is t() such that P{{X(s) → } {|X(t) | , t t()}} . Since can be arbitrarily small, (2.20) holds if and only if t2min{1 k,12}
{|X(s) | , st()}
(X(t) )2dP → 0,
However, t2min{1 k,1 2} ˛
˛
(X(t) )2dP
{|X(s) | , s t()}
t2min{1 k1 2} ˛
˛
(X(t) )2dP,
ySt() {|X(s) | , s t(), X(t()) y}
(2.20)
46
Path dependence in technical change
i) where St St {X(1)(n n t 1 , 0 i t 1} is the set of values that X(t) and X(t) can attain (not necessarily with positive probability). Notice that for any tt() and y St()
(X(t) )2dP
{| X(s) |, s t (), X(t()) y}
(X (t) )2dP
{| X (s) | , s t (), X (t()) y}
This follows from the fact that f(x) and f(x) are the same for x [ , ]. However, t2min{1 k,1 2} ˛
˛
(X (t) )2dP
y St() {|X(s) | , s t(), (t()) y} X
t2min{1 k,1}2
(X (t) )2dP
{X (t)→}
t2min{1 k,12} E [(X (t) )2] → 0, as t → by (2.19). This completes the proof.
NOTES *
1. 2.
3. 4.
The views expressed here cannot be ascribed to the Organisation for Economic Co-operation and Development (OECD) Secretariat or its Member Countries. We are indebted to Yuri Kaniovski for very helpful suggestions. We thank also Brian Arthur, Robin Cowan, Klaus Desmet, Judith Gebauer, Michael Horvath, Andrea Prat, Aldo Rustichini, Valter Sorana, and participants to the 3rd workshop on Economics with Heterogeneous Interacting Agents, Ancona, Italy, May 1998, and to the Conference on Economic Models of Evolutionary Dynamics and Interacting Agents, Trieste, Italy, September 1998, for their comments. Financial support from International Institute for Applied Systems Analysis (IIASA), Banca Nazionale del Lavoro (BNL), Italian Research Council (CNR), and Italian Ministry of University and Scientific Research (MURST) is gratefully acknowledged. All errors are ours. See, for example, the empirical literature on dominant designs (for a survey, cf. Tushman and Murmann, 1998). Throughout this chapter we label the generalisation of Polya urn schemes set forth by Hill et al. (1980) as generalised urn scheme. That generalisation is the most popular in economics but obviously it is not the only possible one (see, for example, Walker and Muliere, 1997). We assume that, if there is a tie, agents choose technology A. Qualitatively, breaking the tie in a different way would not make any difference. A convenient review of analytical results on generalised urn schemes can be found in Dosi et al. (1994). The reader is referred to that volume for the results that are not proved in this chapter. Particularly, X() converges almost surely, as t tends to infinity, to the set of appropriately defined zeros of the function g(x) f (x) x. However, since we are not
47
Competing technologies, technological monopolies
going to restrict ourselves to the case when g() is a continuous function, we need some standard definition concerning equations with discontinuous functions. For a function g() given on R(0,1) and a point x [0,1] set a(x, g) inf{y } R(0,1)lim infk→ g(yk), k
a(x, g) sup{y } R(0,1)lim supk→ g(yk). b
where {yk} is an arbitrary sequence converging to x. Then the set of zeros A(g) of g() on [0,1] is defined by the following relation A(g) {x [0,1]: [a(x, g), a(x, g) 0}. Note that for a continuous g(·) this definition gives the roots of the equation g(x) 0 in the conventional meaning. One particular class of attainable singleton components comprises the downcrossing or stable ones, that is, the points where f(x) x changes its sign from plus to minus. More precisely, R(0,1) is said to be stable if there exists 0 such that for every (0, ) inf
(*)
[ f(x) x](x ) 0.
|x |c
If R(0,1) is stable then X(·) converges to with positive probability for some initial → combination n (0). If in addition to (*) f (x) (0, 1)
for all x R(0, 1), →
5. 6. 7. 8. 9.
10. 11.
12. 13. 14.
then it converges with positive probability to for any initial combination n (0). Finally if the urn function does not have touchpoints and the set A(g) with g(x) f (x) x is composed only of singleton components then almost surely the process converges to the set of stable components. For instance, in the data set of Computer Intelligence InfoCorp employed by Breuhan (1996), more than 80 per cent of the firms in the sample report using a single wordprocessing package. See note 4 above, or Dosi et al. (1994), theorems 1 and 3. To be precise Arthur (1989) quotes also Arthur et al. (1983), though there the properties are stated only as yet-to-be-proved good sense conjectures. See note 4 above, or Dosi et al. (1994), theorem 2. Cowan and Cowan (1998) acknowledge this role of heterogeneity, although only for models where interactions are local. They suggest that many models from other scientific disciplines can be adapted to show market-sharing survival as a result of local interaction effects, and they provide one such example, although restricted to linear returns. Notice that, provided that inequalities (2.9) and (2.10) are eventually satisfied for any t, propositions 1, 2 and 4 hold even if the inequality (2.3) does not hold as may happen when agents are assumed to be forward looking. If returns are constant, the result of proposition 3 and 4 simply becomes the well-known textbook result on rate of convergence of the sample mean and its variance. Bassanini and Dosi (1999b) shows that 2min{l k, 1/2} is also an upper bound to the rate of convergence to an interior limit, therefore proposition 4 could be written in an even stronger way, although not necessary for the task of the present chapter. For a general discussion on this point see also Dosi (1997). We are indebted to Yuri Kaniovski for suggesting us the line of the following proof. The line of reasoning here is the same as for the proof of proposition 1 and 2.
48
Path dependence in technical change
REFERENCES Arthur, W.B. (1983), ‘On competing technologies and historical small events: the dynamics of choice under increasing returns’, IIASA Working Paper WP-83-90 (reprinted in W.B. Arthur (1994) Increasing Returns and Path-Dependence in the Economy, Ann Arbor, MI: University of Michigan Press, 1983). Arthur, W.B. (1989), ‘Competing technologies, increasing returns and lock-in by historical events’, Economics Journal, 99, 116–31. Arthur, W.B. (1990), ‘ “Silicon Valley” locational clusters: when do increasing returns imply monopoly?’, Mathematical Social Sciences, 19, 235–51. Arthur, W.B. and D. Lane (1993), ‘Information contagion’, Structural Change and Economic Dynamics, 4, 81–104. Arthur, W.B. (1983), Y. Ermoliev and Y. Kaniovski, ‘Generalised urn problem and its applications’, Cybernetics, 19, 61–71. Arthur, W.B., Y. Ermoliev and Y. Kaniovski (1986), ‘Strong laws for a class of pathdependent urn processes’, Proceedings of the International Conference on Stochastic Optimization, Lecture Notes on Control and Information Sciences, 81, 187–300. Banerjee, A. (1992), ‘A simple model of herd behaviour’, Quarterly Journal of Economics, 107, 797–817. Bassanini, A.P. (1997), ‘Localized technological change and path-dependent growth’, IIASA Interim Report IR-97-086. Bassanini, A.P. and G. Dosi (1999a), ‘Heterogeneous agents, complementarities, and diffusion of technologies: do increasing returns imply convergence to international monopolies?’, in D. Delli Gatti, M. Gallegati and A. Kirman (eds), Market Structure, Aggregation and Heterogeneity, Berlin, Springer. Bassanini, A.P. and G. Dosi (1999b), ‘Competing technologies, technological monopolies and the rate of convergence to a stable market structure’, Laboratory of Economics and Management Working Paper no. 3, Sant ‘Anna School of Advanced Studies. Bernheim, B.D. (1994), ‘A theory of conformity’, Journal of Political Economy, 102, 841–77. Breuhan, A.L. (1996), ‘Innovation and the persistence of technological lock-in’, unpublished manuscript. Brock, W.A. and S.N. Durlauf (1995), ‘Discrete choice with social interactions I: theory’, NBER Working Paper no. 5291. Brynjolfsson, E. and C.F. Kemerer (1996), ‘Network externalities in microcomputer software: an econometric analysis of the spreadsheet market’, Management Science, 42, 1627–47. Church, J. and N. Gandal (1993), ‘Complementary network externalities and technological adoption’, International Journal of Industrial Organization, 11 239–60. Cowan, R. (1990), ‘Nuclear power reactors: a study in technological lock-in’, Journal of Economic History, 50, 541–67. Cowan, R. (1991), ‘Tortoises and hares: choice among technologies of unknown merit’, Economic Journal, 101, 801–14. Cowan, R., and W. Cowan (1998), ‘Technological standardization with and without borders in an interacting agents model’, unpublished paper. Cusumano, M.A., Y. Milonadis and R.S. Rosenbloom (1992), ‘Strategic maneuvering and mass-market dynamics: the triumph of VHS over Beta’, Business History Review, 66, 51–94.
Competing technologies, technological monopolies
49
David, P. (1985), ‘Clio and the economics of QWERTY’, AEA Papers and Proceedings, 75, 332–7. David, P. (1992), ‘Heroes, herds and hysteresis in technological theory: Thomas Edison and the battle of systems reconsidered’, Industrial and Corporate Change, 1, 129–80. Dosi, G. (1982), ‘Technological paradigms and technological trajectories’, Research Policy, 11, 142–67. Dosi, G. (1997), ‘Opportunities, incentives and the collective patterns of technological change’, Economics Journal, 107, 1530–47. Dosi, G., Y. Ermoliev and Y. Kaniovski (1994), ‘Generalized urn schemes and technological dynamics’, Journal of Mathematical Economics, 23, 1–19. Economides, N. (1996), ‘The economics of networks’, International Journal of Industrial Organization, 14, 673–99. Farrell, J. and G. Saloner (1985), ‘Standardization, compatibility and innovation’, Rand Journal of Economics, 16, 70–83. Farrell, J. and G. Saloner (1986), ‘Installed base and compatibility: innovation, product preannouncements, and predation’, American Economic Review, 76, 940–55. Hill, B.M., D. Lane and W. Sudderth (1980), ‘A strong law for some Generalized urn processes’, Annals of Probability, 8, 214–26. Islas, J. (1997), ‘Getting round the lock-in in electricity generating systems: the example of the gas turbine’, Research Policy, 26, 49–66. Katz, M.L. and C. Shapiro (1985), ‘Network externalities, competition, and compatibility’, American Economic Review, 75, 424–40. Katz, M.L. and C. Shapiro (1986), ‘Technology adoption in the presence of network externalities’, Journal of Political Economy, 94, 822–41. Katz, M.L. and C. Shapiro (1992), ‘Product introduction with network externalities’, Journal of Industrial Economics, 40, 55–84. Katz, M.L. and C. Shapiro (1994), ‘Systems competition and network effects’, Journal of Economic Perspectives, 8, 93–115. Koski, H. and P. Nijkamp (1997), ‘The installed base effect: some empirical evidence from the microcomputer market’, unpublished manuscript. Krugman, P. (1991a), ‘History vs. expectations’, Quarterly Journal of Economics, 106, 651–67. Krugman, P. (1991b), Geography and Trade, Cambridge, MA: MIT Press. Liebowitz, S.J. and S.E. Margolis (1994), ‘Network externality: an uncommon tragedy’, Journal of Economic Perspectives, 8, 133–150. Oren, S. and S. Smith (1981), ‘Critical mass and tariff structure in electronic communications markets’, Bell Journal of Economics, 12, 467–87. Rauch, J.E. (1993), ‘Does history matter only when it matters little? The case of cityindustry location’, Quarterly Journal of Economics, 108, 843–67. Rohlfs, J. (1974), ‘A theory of interdependent demand for a communication service’, Bell Journal of Economics, 5, 16–37. Rosenberg, N. (1982), Inside the Black Box, Cambridge: Cambridge University Press. Saviotti, P.P. and A. Trickett (1992), ‘The evolution of helicopter technology’, 1940–1986, Economics of Innovation and New Technology, 2, 111–30. Tushman, M.L. and J.P. Murmann (1998), ‘Dominant designs, technology cycles, and organizational outcomes’, in B. Staw and L.L. Cummings (eds), Research in Organizational Behaviour, 20, Greenwich, CT: JAI Press.
50
Path dependence in technical change
Venables, A.J. (1996), ‘Localization of industry and trade performance’, Oxford Review of Economic Policy, 12 (3), 52–60. Walker, S. and P. Muliere (1997), ‘Beta-Stacy processes and a generalisation of the Polya-urn scheme’, Annals of Statistics, 25, 1762–80. Winter, S.G. (1986), ‘Comments on Arrow and Lucas’, Journal of Business, 59, S427–34.
3. Path dependence, localised technological change and the quest for dynamic efficiency1 Cristiano Antonelli 1
INTRODUCTION
The quest for the conditions of dynamic efficiency can be considered one of the key aims and scopes of much contemporary work in economic theory. Neoclassical economics has provided an elaborate and sophisticated framework to understand the conditions for static efficiency. In that context, growth and development are the consequences of exogenous changes in the shapes of the utility functions, in the characteristics of the technology and in the actual conditions of the demography and of the natural resources. The theory of economic growth elaborated in that context does not address the actual causes of growth. It is limited to analysing the complementary conditions in terms of rates of growth in the supply of labour and savings, that make it possible for exogenous growth to take place. The notion of path dependence elaborated by Paul David provides one of the most articulated and comprehensive frameworks to move towards the analysis of the conditions that make it possible for an economic system to generate and exploit endogenous growth. Path dependence is an essential tool to move from the analysis of static efficiency and enter into the analysis of the conditions for dynamic efficiency. The notion of path dependence proves to be especially attractive for European economists raised in a tradition that considers growth and change rather than equilibrium as the relevant object of analysis and, hence, values historic time and philological investigations as basic tools to study the dynamics of social events. The pages that follow not only try and articulate the wealth of the contribution of Paul David, but also are an attempt show how fertile and stimulating his framework is. The rest of the Chapter is structured as follows. Section 2 explores in details the notion of path dependence and identifies its basic ingredients. 51
52
Path dependence in technical change
Section 3 shows how the different combinations of the basic ingredients lead different types of path dependence, especially when the analysis is applied to understanding the introduction and diffusion and technological innovations. Section 4 outlines the implications for the analysis of the conditions of dynamic efficiency. The conclusions summarise the analysis.
2
THE INGREDIENTS OF PATH DEPENDENCE
According to Paul David, path dependence is an attribute of a special class of dynamic processes. A process is path dependent when it is non-ergodic and yet it is subject to multiple attractors: ‘systems possessing this property cannot shake off the effects of past events, and do not have a limiting, invariant probability distribution that is continuos over the entire state space’ (David, 1992a: 1). As a matter of fact, historic analysis and much empirical evidence in economic growth, and specifically in the economics of innovation and new technologies, confirm that these characteristics apply and are most relevant to understanding the laws of change and growth of complex systems. Path dependence provides a unique and fertile analytical framework able to explain and assess the ever-changing outcomes of the combination and interplay between factors of continuity and discontinuity, growth and development, hysteresis and creativity, routines and ‘free will’, that characterise economic action in a dynamic perspective which is able to appreciate the role of historic time. The notion of ergodicity (Figure 3.1) is quite complex and deserves careful examination. A process is ergodic when its initial conditions have no influence on its development and eventual outcomes. The general equilibrium framework of analysis is typically ergodic, although the analysis of the competition process and its building blocks such as the theory of costs and of the firm are based upon short-term conditions where some costs are fixed and their irreversibility bears major consequences on the outcome of the interaction among firms in the marketplace. When a process is non-ergodic, its initial conditions have an effect on its development and on the final outcome. Past dependence is an extreme form of non-ergodicity. Historic, as well as social and technological, determinism fully belongs to past dependence. Here the characteristics of the processes that are analysed and their results are considered to be fully determined and contained in their initial conditions. In the economics of innovation, past dependence has often been practised: the epidemic models of diffusion of innovations and the notion of technological trajectory are typical examples of technological and social determinism. As such, these
53
Figure 3.1
Typology of dynamic processes
Stochastically convergent
Ergodic processes
Deterministically convergent and globally stable
Past dependence
Path dependence
Non-ergodic processes
54
Path dependence in technical change
models are non-ergodic and fully past dependent. The process takes place within a single corridor, defined at the outset, and external attractors cannot divert its route, nor can the dynamics of the process be altered by internal factors. Path dependence differs from past dependence in that irreversibility plays a role together with the initial conditions of a process. Its development and the final result, however, are shaped by the influence of local externalities and especially local and internal feedbacks. Path dependence builds upon the mix of the notions of irreversibility, local externalities and feedbacks. In so doing, path dependence can be considered at the border between fully ergodic processes and fully non-ergodic processes. In the former, history does not matter; in the latter, only history matters. Path dependence differs from deterministic past dependence in that irreversibility arises from events along the path, and not only the initial conditions may play a role in selecting from among the multiplicity of possible outcomes. The analysis of a path-dependent stochastic system is based upon the concepts of transient or ‘permanent micro-level’ irreversibilities, creativity and positive feedbacks. The latter self-reinforcing processes may work through the price system, or they may operate through non-pecuniary externalities. The conceptualisation of stochastic path dependence can be considered to occupy the border region between the view of the world in which history enters only to establish the initial conditions after which the dynamics unfolds deterministically, and the conceptualisation of historical dynamics in which one ‘accident’ follows another relentlessly and unpredictably. Path dependence provides economists with the tools to include historical forces in their analysis without succumbing to naive historical determinism. When path dependence applies, history matters, but together with other factors. The sequence of steps becomes a relevant issue in path dependence. At each step, in fact, the direction of the process can be changed because of the influence of new forces and attractors. The full understanding of a path-dependent process requires a detailed analysis of the sequence of the steps that have been made, and of the interactions between the effects of irreversibility, local externalities and feedbacks. Irreversibility pushes towards a trajectory where the initial conditions are replicated and command the direction. Local externalities and feedbacks may exert a diverting effect. The dynamic interplay between these elements shapes the actual characteristics of the process, its direction at each stage, and bears an influence on the following stages. At each stage the balance between such dynamic forces may differ and, hence, so may the direction of the process. In path dependence the sequential interplay between the past-dependent effects of irreversibility and the diverting effects of local externalities and
Path dependence and the quest for dynamic efficiency
55
feedbacks plays a key role and qualifies the conditions of path dependence itself as an interface between ergodic and non-ergodic approaches to economics. The careful identification of the stages of each dynamic process and of the changes in their alignment, and therefore in the interplay between ergodic and non-ergodic forces, and the consequent localisation and explanation of the points of inflection, bifurcation or change in direction and intensity, becomes a major task and a fertile area for investigation. Let us analyse briefly the three main ingredients of path dependence: irreversibility, feedbacks and local externalities. Irreversibility clearly is the prime ingredient and causal factor of the non-ergodic dynamics. Irreversibility consists in the lack of dynamic malleability of production factors as well as tastes and preferences, reputation and routines. When irreversibility applies it is difficult and costly to change a given set of conditions; irreversibility engenders specific costs – the costs of switching from one condition to another. The induced, and hence endogenous, change of both production and utility functions provides here a major contribution to understanding the dynamic feature of path dependence. Technological knowledge and preferences are endogenous to the economic system as they are the result of the creative reaction of agents to a mismatch between the actual state of products and factors markets and their original plans and expectations. Building upon such expectations, agents embedded in historical time have made plans and commitments upon which irreversible decisions have been taken. The mismatch between the products and factors markets, conditions as planned and built into the irreversible decisions which have been taken, and their actual conditions, is the basic inducement factor and focusing device that leads to creative reactions. The induced technological change tradition of analysis provides the glue between irreversibility and creativity (Arrow, 2000; Ruttan, 1997; 2001). Feedbacks are the most important ingredient of path dependence. Creativity and reactivity are key factors in understanding path dependence. Path dependence assumes that agents are able to react to the changing conditions of their environment, not only by adjusting prices to quantities, and vice versa, but also, and mainly, by changing their technology as well as their preferences and tastes. The contribution of the prospect theory according to which the more agents are exposed to frustration the less risk adverse they are, is most relevant in this context (Kahneman and Tversky, 1979; Witt, 1998). Path dependence assumes that knowledge and preferences are the result of a dynamic process which is influenced by the initial conditions and yet is open to a variety of local attractors that shape the characteristics of the
56
Table 3.1
Path dependence in technical change
Converging and diverging forces in past and path dependence
Diverging forces in path dependence
Converging forces in past dependence
Collective learning Changes in factors’ prices Local externalities and feedbacks
Internal learning Creativity Irreversibility
learning process and its eventual outcome. Learning is possible mainly, although not exclusively, when it is based upon repeated actions in a welldefined contextual set of techniques and relationships in doing and in using. Hence, learning is mainly local in that it is limited by the scope and the perimeter of the expertise acquired within a given technology, productive conditions and organisational sets. The cumulated effects of localised learning by doing and learning by using consist in higher levels of competence and, eventually, the opportunity to generate new knowledge. All changes to the given set of techniques, production conditions and relationships are likely to engender relevant opportunity costs, in terms of missed opportunities for learning and hence for acquiring higher levels of competence and generating new technological knowledge (David, 1993; 1994). While irreversibility exerts a past-dependent effect that keeps the process within well-defined corridors, local externalities and local feedbacks exert a path-dependent effect which may diverge and push the process away from the initial direction (Table 3.1). The locality of both externalities and feedbacks is clearly central in path dependence. The understanding of the local span of influence of externalities and feedbacks, as opposed to a general or global one, paves the way to appreciate the role of the variety of local contexts and, therefore, to understanding the existence of multiple equilibria and hence multiple directions and intensities of the process. The same forces at play at the outset of the process can lead to different paths and different outcomes when applied in different contexts and in different conditions. Much systematic evidence and stylised classifications find here a clear explanatory framework: the same technology applied in different regions can prove to be more or less effective. Relative factor prices may exert a strong diverging effect that pushes each firm away from the initial trajectory. The path of followers cannot replicate that of leaders. The stages of economic growth for each player differ according to the global context in which they take place. The specific characteristic of technological knowledge as a collective activity has major implications here, because of its low levels of appropriability, excludability and divisibility, articulated in complexity, cumulability
Path dependence and the quest for dynamic efficiency
57
and fungibility, in its production and in its use. Technological knowledge cannot be produced by any single agent in isolation. Technological knowledge is itself both an input and an output; as such it is the product of the continual recombination and internal and external knowledge inputs. In turn, knowledge output spills within well-circumscribed ‘commons’ and this has major effects on the conduct of the firms able to access the technological commons. Relative prices of production factors act as a powerful factor in selection and discrimination favouring some technologies and actions against others. Proximity among firms in regional, product and technological spaces is most important to access technological spillovers. When the local conditions change, an issue of co-evolution and covariance between the internal dynamic conditions dictated by static and dynamic irreversibilities and the external irreversibilities emerges. The co-evolution of the local environment and the dynamic characteristics of the processes in place can determine major discontinuities and drastic changes in the path (Gould, 2002).
3
TYPES OF PATH DEPENDENCE IN THE ECONOMICS OF INNOVATION
The identification of the basic ingredients of path dependence and the appreciation of their diverse and complementary roles in shaping the dynamics of path-dependent processes make it possible to classify different forms of path dependence. The evidence of the economics of innovation is especially rich in this context. Two basic types of factors in path dependence can be identified when attention is focused on the location of the engine for growth: whether internal or external to each agent, and when the framework is applied to understanding the introduction of new technologies or, rather, their diffusion. Internal factors play a major role in past dependence. The latter takes place when the path along which the process takes place is mainly determined by the interplay between the irreversibility of production factors and the conditions for localised learning, internal to each firm. The introduction of a new technology is determined here by the induced creativity of agents facing unexpected and possibly adverse emerging conditions mainly determined by the lack of flexibility engendered by irreversibility. Creativity, however, is stimulated so as to compensate for the mismatch. Positive feedbacks are at work. Creativity builds upon localised learning and, hence, is better able to implement the techniques originally in place. The notion of path dependence elaborated by Paul David (1975) belongs to this case: firms are induced to introduce a new technology by their
58
Path dependence in technical change
internal characteristics in terms of irreversibility, and follow a path of technological change shaped by their internal characteristics in terms of learning processes and acquired competence. The analysis of path-dependent innovation identifies the conditions that explain the localised character of technological change, such as the mix of irreversibility, induced innovation, local externalities and local endowments. Here the notion of localised technological change, first introduced by Atkinson and Stiglitz (1969) and subsequently elaborated by Paul David (1975), provides important complementarities. The mismatch between the irreversibility of the tangible and intangible stock of sunken inputs and the actual conditions of both factors and products markets, affected by the continual introduction of unexpected innovations in the system, is the prime engine. This mismatch induces the creative reaction and the eventual introduction of new localised technologies. Myopic, but creative, agents introduce technological changes that are localised by their knowledge base built upon localised learning processes, the switching costs stemming from irreversible production factors and the external conditions of factors markets. While the rate of introduction of new technologies is induced by the mismatch between the irreversibility of production factors and the actual conditions of the markets, the direction of the new technologies is induced by the relative prices of production factors. The access conditions to technological spillovers and external knowledge available in the local pools of collective knowledge help explaining the localised direction of introduction of new technologies along a welldesigned technological, technical, institutional and industrial path. Such access conditions, in fact, affect in depth the actual results of the induced innovation activity of each agent and, hence, the incentives to innovate. External factors play a role when path dependence is determined by conditions that are external to firms, but internal to the system. Firms are induced to innovate and to follow a well-defined technological path by conditions that are found in the markets for products and factors, as well as in the behaviours of consumers and competitors. A clear case for external path dependence emerges when the role of external knowledge is taken into account and its contribution to the actual span of technological alternatives is appreciated. The understanding of the role of technological externalities in the generation of new technological knowledge and in the introduction of new technologies makes it possible to stress the role of systemic path dependence (Antonelli, 1999; 2001). Firms searching for a possible reaction to unexpected events help each other with localised spillovers and reciprocal knowledge transfer that build upon a local knowledge common based on the competence and experience acquired in learning by doing and learning by using. The direction of the process of introduction
Path dependence and the quest for dynamic efficiency
59
of new technologies of each agent here is seen as the result of the collective knowledge available locally. External factors are also at work in path dependence when local factors markets affecting the introduction and selection of new technologies are appreciated. The productivity and costeffectivity of a new technology is influenced by composition effects (Antonelli, 2003). Composition effects are the outcome of the sensitivity of output to the relative scale of each single factor, rather than to the scale of the bundle of production factors. They are positive when the relative scale of most productive factors is augmented and that of the least productive factors is reduced. In general, because of composition effects, the larger the productivity of the factor which is more widely used and the lower the productivity of the factor which is less used, the larger are the effects of any changes in the relative levels of factor costs. When the most productive factor is cheaper and hence its use is more intensive, and the least productive factor is most expensive and hence its use is least intensive, production costs are lowest. The higher the growth of total factor productivity stemming from the introduction of a given technology, the higher is the output elasticity of the productive factor which locally is most abundant. Different agents, rooted in different regions, with different endowments and hence different conditions of their local factors markets may react with similar levels of creativity to similar changes in their current conditions, introducing new technologies with marked differences in terms of factors intensity not only because of the effects of internal localised learning and the access conditions to the local pools of collective knowledge, but also because of the powerful consequences of composition effects. Here composition effects act as an inducement factor that explain the direction of the introduction of new technologies rather than their diffusion. Irreversibility and the consequent switching costs matter also in a quite different context. Irreversibility, in fact, exerts important effects on the selective diffusion of rival innovations. New technologies are sorted out not only by their absolute levels of efficiency, but also with respect to their complementarity and compatibility with the installed stocks of fixed and irreversible production factors. Firms with important stocks of fixed capital and irreversible competence of their employees, which attach great value to their customer base and to the relationships with their providers of intermediary inputs, will select the new technology which is not only more productive but also more compatible and more easily integrated within the existing production process and within the network of relations in place. External factors such as relative prices in the factor markets and levels of compatibility and interoperativity between different products sold by
60
Table 3.2
Path dependence in technical change
Types of path dependence Path-dependent innovation
Path-dependent diffusion
Factors internal to firms
Switching costs Localised internal learning
Factors external to firms
Localised knowledge commons Composition effects and relative factors’ prices
Complementarity Increasing returns to production Network externalities on supply and demand Composition effects and relative factors’ prices
different firms play an important role in shaping the choice among rival technologies, and in so doing play a role in path-dependent diffusion. The identification of path-dependent diffusion, as distinct from pathdependent innovation, becomes relevant in this context and can be found in the path-breaking analysis of Paul David (1975). Path-dependent innovation can be defined as the set of explanations that make it possible to understand why firms are actually better able to innovate within a limited set of techniques. Path-dependent diffusion is the class of explanatory analysis that makes it possible to understand why firms adopt some technologies, possibly in the proximity of existing technologies, both internal and external to each firm, rather than any possible technological innovations. More specifically, path-dependent diffusion assumes that technological innovations have been already introduced and that some rivalry and substitutability exists between them and the existing ones, as well as among them. Path-dependent innovation instead explains how and why firms innovate. The matrix of Table 3.2 provides a synthesis of the basic argument. Path dependence has contributed substantially to understanding the diffusion of innovations, often bounded by the analysis of the adoption of single innovations, isolated with respect to their own historic context of introduction. In the analysis of the determinants of path-dependent diffusion, the distinction between internal and external factors is again relevant. The notion of path dependence elaborated by Brian Arthur (1989) and Paul David (1985) clearly contributes the analysis of diffusion: new technologies are sorted out mainly by the effects of increasing returns to adoption at the system level. In their analysis, the selection and eventual diffusion of new technologies is path dependent in that it is influenced by the timing of their sequential introduction, which in turn affects their relative profitability of adoption as shaped by the powerful consequences of positive feedbacks consisting of the interplay between network
Path dependence and the quest for dynamic efficiency
61
externalities on the demand side and increasing returns in production. Here the choice of the new technology is shaped by the factor and products markets’ conditions. The diffusion of new and better technologies can be delayed or barred by the lack of compatibility with the internal and irreversible characteristics of potential adopters. Here internal factors play the key role in explaining path-dependent diffusion. The durability and irreversibility of internal factors such as the capital stock, but also the skills of human capital, the location in a given space and the relationships, will determine adoption of new technologies with customers and suppliers. For a given supply of new and rival technologies, composition effects act as powerful selection devices and the diffusion of technologies will be influenced by the local conditions of factors markets. Labour-intensive technologies will diffuse faster in labour-abundant countries, and capitalintensive technologies will be adopted in capital-abundant technologies. The adoption of new technologies that are characterised by high levels of output elasticity of labour, but small shift effects, might be delayed for ever in capital-intensive countries. Too much emphasis has been put on the effects of path-dependent diffusion in terms of ‘lock in’. Technological ‘lock in’ is a possible outcome of path-dependent diffusion, although new waves of better technologies, possibly introduced by other competitors may eventually break the technological resilience. The real key point is in fact not the ‘lock-in’ effects but rather the ‘lockout’ effects: in path dependence firms are induced to change their current state of affairs by some unexpected events they cannot cope with – by means of traditional price-quantity adjustments – because of irreversibilities and constraints, on the one hand, and the opportunities for the introduction of new technologies, on the other. Such dynamics are fuelled by irreversibility and shaped by the changing effects of local externalities and feedbacks, within a path.
4
THE ENGINE OF GROWTH: PATH DEPENDENCE AND LOCALISED TECHNOLOGICAL CHANGE
The different forms of path dependence that have been identified so far can be considered complementary components of a broader dynamic framework which make it possible to understand the engine of growth and the conditions for dynamic efficiency. This can help in understanding that the basic aim of path dependence is to provide a first, and yet quite elaborate, framework to address the key
62
Path dependence in technical change
issues of the analysis of the conditions for dynamic efficiency. In so doing, path dependence provides a framework which is not necessarily in conflict with static analysis and in general with the quest for static efficiency. Path dependence addresses a different set of problems and a different set of conditions. In the standard static framework growth is exogenous and ergodic. Growth takes place when and if exogenous changes concerning the technology, the preferences and the distribution of natural resources and demography take place. The direction of change can be both positive and negative; the system is able to adjust automatically to any new set of exogenous conditions. This context does not change, even when assumptions about the variety and heterogeneity of agents are taken into account. Low levels of irrationality, at least on the supply side, are also compatible. Irrational conduct of firms will lead to suboptimal performances and, eventually, to forced exit. Rational conduct will lead to survival in the marketplace. The notion of Brownian movement has been successfully borrowed from physics to characterise such a micro-dynamic context. Entry and exit of firms, as well as changes in the levels of output and possibly the exclusion of suboptimal conducts, are the prime engine of such a Brownian movement. New firms enter the market attracted by the gap between prices and costs margins, and firms leave the market when prices fall below costs. Firms are not supposed to be able to change their technology and, hence, their production functions. At best, these firms are able to influence the position and slope of supply curves as a consequence of their entry and exit. In the approach which builds upon the path-breaking contributions of Paul David, and further elaborated so far, growth is endogenous and path dependent. Growth is primarily the result of the endogenous changes in technologies and tastes, hence in production and utility functions, which take place because of the creativity and reactivity of agents. Agents are characterised both by high levels of irreversibility and by high levels of creative capability. Irreversibility exposes agents to substantial losses and rigidities when their plans are not fulfilled and their expectations are not realised. Creativity, however, makes it possible to consider, next to entry and exit and changes in output levels, the introduction of innovations as the other possible reaction to the mismatch between expectations and the actual conditions of factors and products markets. Growth is the possible outcome of a system exposed to the continual mismatch between expectations and actual events, and yet is able to organise and structure its creativity so as to change the technology and the
Path dependence and the quest for dynamic efficiency
63
psychology. When the creative reaction is not appropriate and consistent, the system is kept within the standard conditions of the static equilibrium. In this context it is clear that innovation feeds innovation. When and if the creative reaction to any mismatch is conducive to the actual introduction of new products and new processes, a self-sustaining process of growth and innovation takes place. The introduction of innovations, in fact, is itself the factor that changes the conditions of equilibrium of both products and factors markets. The stochastic character of the growth process becomes clear in this context. The actual levels of creativity engendered by the mismatch and the actual levels of technological change introduced depend upon a number of complementary conditions. The mismatch is a necessary but non-sufficient condition to induce the successful introduction of new technologies. Only in special circumstances do all the conditions apply and are conducive to a strong and positive creative reaction in terms of fast rates of introduction of new and highly productive technologies. Even hyper-rational agents who have access to Olympic rationality cannot be expected to foresee the outcomes of the innovative process. When innovation is taken into account, the mismatch between expectations and actual factors and products market conditions is bound to take place with varying levels of intensity and different gaps. At each round the mismatch can lead to the generation of new technologies or decay into an equilibrium adjustment process leading to the conditions of static efficiency. The conditions that are conducive to the actual introduction of innovations become, clearly, the central focus of the quest for the conditions of dynamic efficiency. The understanding of the relationship between the amount of entropy within the system, that is, the amount and the distribution of the mismatch between expectations and the actual conditions of product and factors’ markets, and the amount of creativity, is a first and central area of concern. Low levels of mismatch can be easily absorbed by firms that can simply adjust prices to quantities and vice versa with low levels of attrition. High levels of entropy, however, are also likely to endanger the actual capability of firms to react properly and to be actually able to introduce successful innovations. Empirical analysis in this area is still lacking and can provide the basic ground upon which theoretical investigations can subsequently be elaborated. The conditions that affect the levels of creativity and reactivity of agents, for given levels of entropy, play a major role in assessing the dynamic efficiency of an economic system. The organisation of firms in terms of hierarchical structure and decision-making is most important both with respect to their effects in terms of accumulation of competence
64
Path dependence in technical change
and technological knowledge, to their capability to actually convert tacit knowledge into timely innovations and, most important, to their ability to organise a timely and successful reaction to the changing conditions of their business environment. The debate about the effects of centralised versus decentralised decisionmaking, in other words, finds a new context of understanding which values the dynamic characteristics of firms and the timing of their reactions. Decentralised decision-making favours the levels of creativity and reactivity for agents facing unexpected changes in products and factors markets. The size of firms and the structure of hierarchical decision-making, however, favour the accumulation of knowledge and its timely conversion into technological innovations. The conditions for entry and mobility in the products markets are key factors in this context. The conditions of entry and the general conditions for entrepreneurship are most important. Barriers to mobility across markets also play a role as far as they limit the capability of incumbents to take advantage of their localised knowledge and the scope for technological fungibility and technological convergence. The notion of competition as a discovery process here is consistent with the quest for the conditions for dynamic efficiency. The conditions for the effectiveness of the creativity are the third layer of analysis. The provision of knowledge externalities and the general conditions for the generation of new technological knowledge here become central. The access to knowledge in general is the key factor that make it possible for creativity to be valorised and to lead to actual solutions in terms of higher productivity. The institutions of intellectual property rights and quality of the public knowledge infrastructure are major factors in this context. The quality and the density of communication channels among learning agents and the access conditions to the flow of technological spillovers and technological interactions are also essential. The organisation of financial markets, the access to credit and financial resources at large are most important in that they make it possible for new firms to identify the new markets and take advantage of the opportunities for generating localised technological change. The institutions of labour markets also play a major role. Seniority systems favour the accumulation of competence and make it easier to generate new technological knowledge. Mobility of qualified experts embodying high levels of tacit knowledge and technological competence across firms, however, is conducive to better dissemination and circulation of technological knowledge. Finally, the conditions that favour the working of the path-dependent engine of growth come into question. In this approach the variety of actors
Path dependence and the quest for dynamic efficiency
65
and interacting markets matters (Metcalfe, 1997). Firms are induced to innovate when their myopic expectations do not match the actual markets. Firms, in other words, react to all changes to their myopic expectations, not only with changes in the price–output mix, but also with the introduction of new technologies. The larger the variety of firms, the larger are likely the mismatches and, especially, the variety of alternatives that are elaborated and tested by the firms in the marketplaces. Hence, the larger are the chances that technological innovations that are actually superior and able significantly to increase the general efficiency of the system are likely to be introduced. Moreover, it is clear that the new technologies reflect the specific context of action. Such a context includes firms active in a broad industrial structure, which includes regions and countries with significant heterogeneity in both technologies and endowments. The larger the more differentiated the competitive arena, the larger the incentives to introduce innovations and to sort the most productive ones: globalisation is likely to be the cause rather than the consequence of accelerated rates of introduction of technological change. In a population of heterogeneous agents, rooted in a heterogeneous economic space, with different local factors markets, different needs and preferences of consumers, and different pools of collective knowledge, all discrepancy between the expectations upon which irreversible action is taken and the actual conditions of the factors and products markets is likely to engender the localised introduction of new technologies. The capability of agents actually to generate successful innovations will vary according to the stock of localised learning they can mobilise and the access conditions to the external pools of collective knowledge. The diffusion of such technologies will vary according to their characteristics in terms of output elasticity and shift effects. Increasing returns to adoption on the demand side and the rates of adaptation of consumers to the new products will play a major role in shaping the diffusion of new product innovations. The introduction and diffusion of innovations, however, is likely to affect the conditions of the product and factors markets upon which the expectations of other agents had been built. New discrepancies arise and new feedbacks are likely to take place with an endless process of creative reaction. It is clear that the wider the heterogeneity within the system, the larger the chances that expectations do not match the actual conditions and, hence, the faster the rates of introduction of new technologies and, possibly, the faster the rates of growth. The dynamic complementarity and interdependence among the creative efforts of each agent is likely to play a major role in assessing the eventual outcome. When the innovative efforts of agents happen to be co-ordinated
66
Path dependence in technical change
so as to add complementary bits of new technological knowledge, new technological systems can emerge. New gales of innovation are introduced, and growth, at the system level, can take momentum.
5
CONCLUSIONS
The notion of path dependence provides a unique framework within which to analyse the conditions for dynamic efficiency. Path dependence provides, in fact, an extraordinary and articulated set of interrelated notions, which make it possible to go beyond the analysis of the static conditions for general equilibrium and efficiency to apply. Path dependence builds upon a few simple basic elements: irreversibility and historic time, innovation viewed as a creative reaction, and local externalities and feedbacks. These four elements are put in place and integrated into basic economics, so as to provide a basis upon which a fully articulated post-Walrasian approach can be elaborated – one where the central role of the markets as mechanisms for the creation and distribution of incentives is recognised and emphasised. In such an approach, however, the welfare attributes of equilibrium are no longer valid. Equilibrium itself is questioned. Sequences of possible equilibria can be identified and traced. The investigation of the conditions that make it possible for firms to convert the entropy of the systems, as determined by the continual mismatch between expected and actual market conditions, into new and better technologies, and hence to feed a rapid and effective growth, is the basic object of investigation and problematic core of an approach that builds upon the notion of path dependence. Path dependence provides a framework, which makes it possible to go beyond the deterministic and static world of the Walrasian equilibrium. A new landscape springs up, one where the dynamic outcomes of the interactions of agents in the market places cannot be fully anticipated. A variety of possible outcomes can be predicted, as well as their sequence and their dynamic relationship. Walrasian equilibrium is but one of the many possible outcomes, as well as growth. Growth is the positive, and stochastic, result of a system of interactions where agents are able to react to the shortcomings of the mismatch between expectations and the actual conditions of the products and factors markets changing the equilibrium conditions of the system. A new scope for economic policy is now available. The goal is clearly the build-up and maintenance of a social, institutional and economic environment, which is conducive to fully appreciating, valorising and making effective the innovative reactions of myopic but creative and learning agents.
Path dependence and the quest for dynamic efficiency
67
From the viewpoint of economic analysis, path dependence proves to be far less hostile and conflicting with neoclassical analysis than is often assumed. As a matter of fact, path dependence seems even to provide important opportunities to rescue relevant portions of the received tradition, especially when the distinction between short-term and long-term analysis is appreciated. The definition of the interfaces of compatibility and incompatibility between the conditions for static efficiency and the conditions for dynamic efficiency becomes an interesting area for theoretical analysis. The consequences of the interactions of agents in markets, however, now include the generation and use of new and better technologies, rather than the single gravitation around a stable and unique attractor: the notion of complex dynamics seems to provide a suitable framework to pursue such analysis.
NOTE 1. This work is the result of systematic redrafting, polishing and refining, since its first presentation at the Conference in Honour of Paul David at the Accademia delle Scienze in Torino, in May 2000. The comments of many, including Paul David’s, are acknowledged. It makes explicit many intellectual debts and tentative recombinations: in so doing it provides evidence on the attempt to introduce incremental knowledge along a well-defined path sufficient to become a clear example of a deliberate effort to stand on the shoulders of a giant. The present version of the paper is also the text of the ‘Laudatio’ for the award of the Laurea Honoris Causa in Communication Studies to Paul David for his contributions to the economics of innovation, held at the Università degli Studi di Torino, 12 May 2003.
BIBLIOGRAPHY Antonelli, Cristiano (1995), The Economics of Localized Technological Change and Industrial Dynamics, Boston, MA, and Dordrecht, Kluwer. Antonelli, Cristiano (1999), The Microdynamics of Technological Change, London, Routledge. Antonelli, Cristiano (2001), The Microeconomics of Technological Systems, Oxford, Oxford University Press. Antonelli, Cristiano (2003), The Economics of Innovation, New Technologies and Structural Change, London, Routledge. Arrow, Kenneth J. (2000), ‘Increasing returns: historiographic issues and path dependence’, European Journal of History of Economic Thought, 7, 171–80. Arthur, Brian (1989), ‘Competing technologies increasing returns and lock-in by small historical events’, Economic Journal, 99, 116–31. Arthur, Brian (1994), Increasing Returns and Path Dependence in the Economy, Ann Arbor, MI, Michigan University Press.
68
Path dependence in technical change
Atkinson, Anthony B. and Stiglitz, Joseph E. (1969), ‘A new view of technological change’, Economic Journal, 79, 573–8. David, Paul A. (1975), Technical Choice Innovation and Economic Growth, Cambridge, Cambridge University Press. David, Paul A. (1985), ‘Clio and the economics of QWERTY’, American Economic Review, 75, 332–7. David, Paul A. (1987), ‘Some new standards for the economics of standardization in the information age’, in Dasgupta, Paul and Stoneman, Paul (eds), Economic Policy and Technological Performance, Cambridge, Cambridge University Press. David Paul A. (1988), ‘Path dependence: putting the past into the future of economics’, mimeo, Department of Economics, Stanford University. David, Paul A. (1990), ‘The dynamo and the computer: a historical perspective on the productivity paradox’, American Economic Review, (P&P) 80, 355–61. David, Paul (1992a), Path dependence in Economic Processes: Implications for Policy Analysis in Dynamical System Contexts, Torino, Fondazione Rosselli. David, Paul A. (1992b), ‘Heroes herds and hysteresis in technological history’, Industrial and Corporate Change, 1, 129–79. David, Paul A. (1993), ‘Knowledge property and the system dynamics of technological change’, Proceedings of the World Bank Annual Conference on Development Economics, Washington, DC, World Bank. David, Paul A. (1994), ‘Positive feedbacks and research productivity in science: reopening another black box’, in Granstrand, Owe (ed.), Economics and Technology, Amsterdam, Elsevier. David, Paul A. (1997a), ‘From market magic to calypso science policy: a review of Terence Kealey’s “The economic laws of scientific research” ’, Research Policy, 26, 229–55. David, Paul A. (1997b), Path Dependence and the Quest for Historical Economics: One More Chorus of the Ballad of QWERTY, University of Oxford Discussion Papers in Economic and Social History, Number 20. David, Paul A. (1998), ‘Communication norms and the collective cognitive performance of “Invisible Colleges” ’, in Barba Navaretti, Giorgio et al. (eds), Creation and the Transfer of Knowledge: Institutions and Incentives, Berlin, Springer-Verlag. David, Paul A. (2000a), ‘Path dependent learning, and the evolution of beliefs and behaviours’, in Pagano, Ugo and Nicita, Antonio (eds), The Evolution of Economic Diversity, London, Routledge. David, Paul A. (2000b), ‘Path dependence and varieties of learning in the evolution of technological practice’, in Ziman, John (ed.), Technological Innovation as an Evolutionary Process, Cambridge, Cambridge University Press, ch. 10. David, Paul A. (2001), ‘Path dependence, its critics, and the quest for “historical economics” ’, in Garrouste, Pierre and Ioannidis, Stravos (eds), Evolution and Path Dependence in Economic Ideas: Past and Present, Cheltenham, Edward Elgar. Gould, Steven J. (2002), The Structure of Evolutionary Theory, Cambridge, MA, Harvard University Press. Kahneman, Daniel and Tversky, Amos (1979), ‘Prospect theory: an analysis on decision under risk’, Econometrica, 47, 263–91. Metcalfe, John S. (1997), Evolutionary Economics and Creative Destruction, London, Routledge. Pfenninger, Karl H. and Shubik, Valerie R. (eds) (2001), The Origins of Creativity, Oxford, Oxford University Press.
Path dependence and the quest for dynamic efficiency
69
Ruttan, Vernon W. (1997), ‘Induced innovation evolutionary theory and path dependence: sources of technical change’, Economic Journal, 107, 1520–29. Ruttan, Vernon W. (2001), Technology Growth and Development: An Induced Innovation Perspective, Oxford, Oxford University Press. Witt, Ulrich, (1998), ‘Imagination and leadership: the neglected dimension of an evolutionary theory of the firm’, Journal of Economic Behavior and Organization, 35, 161–77.
4. A history-friendly model of innovation, market structure and regulation in the age of random screening of the pharmaceutical industry1 Franco Malerba and Luigi Orsenigo2 1
INTRODUCTION
In this chapter we present a discussion of the long-term history of the evolution of the pharmaceutical industry and a related ‘history-friendly’ model of the random screening period. This is in line with Paul David’s major interests (two of many) of linking history and modelling. Pharmaceuticals constitutes an ideal subject for history-friendly analysis. Pharmaceuticals are traditionally a highly research and development (R&D) intensive sector, which has undergone a series of radical technological and institutional ‘shocks’. However, the core of leading innovative firms and countries has remained quite small and stable for a very long period of time, but the degree of concentration has been consistently low, whatever level of aggregation is considered. In addition, these patterns of industrial dynamics are intimately linked to two main factors: the nature of the processes of drug discovery and the fragmented nature of the relevant markets. Specifically, innovation processes have been characterised for a very long time by a low degree of cumulativeness and by ‘quasi-random’ procedures of search (random screening). Thus, innovation in one market (a therapeutic category) does not entail higher probabilities of success in another one. Moreover, pharmaceuticals represents a case where competition is less dissimilar to the model of patent races. Understanding if these intuitive factors can indeed explain the observed patterns of industrial dynamics and articulating the mechanisms through which they exert their impact are in themselves interesting challenges; the more so, if this model is compared with the analysis of the computer industry. The comparison might allow for some generalisations about the determinants of the relevant 70
A history-friendly model of innovation
71
similarities and differences in the patterns of industrial evolution across industries. In addition, the pharmaceutical industry, ever since its inception, has been deeply affected by a large variety of institutional factors and policies, ranging from patents, different forms of regulation (procedures for product approval, price controls, and so on), organisation of the public research systems, and so on. From this perspective, pharmaceuticals constitute an ideal case for studying the differential impact and the working of alternative policies. The chapter is organised as follows. Section 2 provides a historical account of the evolution of the pharmaceuticals. Section 3 introduces the main theoretical issues that are raised by the previous historical account and presents the model. Section 4 examines some alternative runs and section 5 concludes.
2
INNOVATION AND THE EVOLUTION OF MARKET STRUCTURE IN THE PHARMACEUTICAL INDUSTRY: AN OVERVIEW
The patterns of development of the pharmaceutical industry have been analysed extensively by several scholars. In what follows, we rely especially on the work by Chandler (1990; 1999), Galambos and Lamoreaux (1997), Galambos and Sewell (1996), Galambos and Sturchio (1996), Gambardella (1995), Henderson et al. (1999), Orsenigo (1989) and Schwartzman (1976). But actually, this account of the history is largely drawn from Henderson et al. (1999) and Pisano (1996). In very general terms, the history of the pharmaceutical industry can be analysed as an evolutionary process of adaptation to major technological and institutional ‘shocks’. It can be usefully divided into three major epochs. The first, corresponding roughly to the period 1850–1945, was one in which little new drug development occurred, and in which the minimal research that was conducted was based on relatively primitive methods. The large-scale development of penicillin during the Second World War marked the emergence of the second period of the industry’s evolution. This period was characterised by the institution of formalised in-house R&D programmes and relatively rapid rates of new drug introduction. During the early part of the period the industry relied largely on so called ‘random’ screening as a method for finding new drugs, but in the 1970s the industry began a transition to ‘guided’ drug discovery or ‘drug development by design’ a research methodology that drew heavily on advances in molecular biochemistry, pharmacology and enzymology. The third epoch of the industry has its roots in the 1970s but did not come to full flower
72
Path dependence in technical change
until quite recently as the use of the tools of genetic engineering in the production and discovery of new drugs has come to be more widely diffused. 2.1
Early History
The birth of the modern pharmaceutical industry can be traced to the midnineteenth century with the emergence of the synthetic dye industry in Germany and Switzerland. During the 1880s, the medicinal effects (such as antiception) of dyestuffs and other organic chemicals were discovered. It was thus initially Swiss and German chemical companies such as Ciba, Sandoz, Bayer and Hoechst, leveraging their technical competencies in organic chemistry and dyestuffs, who began to manufacture drugs (usually based on synthetic dies) later in nineteenth century. For example, salicylic acid (aspirin) was first produced in 1883 by the German company, Bayer. In the US and the UK, mass production of pharmaceuticals also began in the later part of the nineteenth century. However, the pattern of development in the English-speaking world was quite different from that of Germany and Switzerland. Whereas Swiss and German pharmaceutical activities tended to emerge within larger chemical producing enterprises, the USA and the UK witnessed the birth of specialised pharmaceutical producers such as Wyeth (later American Home Products) Eli Lilly, Pfizer, Warner-Lambert and Burroughs-Wellcome. Up until the First World War German companies dominated the industry, producing approximately 80 per cent of the world’s pharmaceutical output. In the early years the pharmaceutical industry was not tightly linked to formal science. Until the 1930s, when sulfonamide was discovered, drug companies undertook little formal research. Most new drugs were based on existing organic chemicals or were derived from natural sources (for example, herbs) and little formal testing was done to ensure either safety or efficacy. 2.2
The ‘Random Screening’ Period
World War II and wartime needs for antibiotics marked the drug industry’s transition to an R&D-intensive business. With the outbreak of the Second World War, the US government organised a massive research and production effort that focused on commercial production techniques and chemical structure analysis. More than 20 companies, several universities, and the Department of Agriculture took part. The commercialisation of penicillin marked a watershed in the industry’s development. Due partially to the technical experience and organisational capabilities accumulated through the intense wartime effort to develop penicillin, as well as to the recognition
A history-friendly model of innovation
73
that drug development could be highly profitable, pharmaceutical companies embarked on a period of massive investment in R&D and built largescale internal R&D capabilities. At the same time there was a very significant shift in the institutional structure surrounding the industry. Whereas, before the war, public support for health-related research had been quite modest, after the war it boomed to unprecedented levels, helping to set the stage for a period of great prosperity. This period was a golden age for the pharmaceutical industry. Research and development spending literally exploded, and with this came a steady flow of new drugs. Drug innovation was a highly profitable activity during most of this period. During the early 1980s, double-digit rates of growth in earnings and return-on-equity were the norm for most pharmaceutical companies and the industry as a whole ranked among the most profitable in the USA and in Europe. A number of structural factors supported the industry’s high average level of innovation and economic performance. One was the sheer magnitude of both the research opportunities and the unmet needs. In the early post-war years, there were many physical ailments and diseases for which no drugs existed. In every major therapeutic category – from painkillers and anti-inflammatories to cardiovascular and central nervous system products – pharmaceutical companies faced an almost completely open field (before the discovery of penicillin, very few drugs effectively cured diseases). Faced with such a ‘target rich’ environment but very little detailed knowledge of the biological underpinnings of specific diseases, pharmaceutical companies invented an approach to research now referred to as ‘random screening’. Under this approach, natural and chemically derived compounds are randomly screened in test tube experiments and laboratory animals for potential therapeutic activity. Pharmaceutical companies maintained enormous ‘libraries’ of chemical compounds, and added to their collections by searching for new compounds in places such as swamps, streams and soil samples. Thousands of compounds might be subjected to multiple screens before researchers honed in on a promising substance. Serendipity played a key role since in general the ‘mechanism of action’ of most drugs were not well understood. Researchers were generally forced to rely on the use of animal models as screens. Under this regime it was not uncommon for companies to discover a drug to treat one disease while searching for a treatment for another. Since even the most productive chemist might find it difficult to synthesise more than a few compounds over the course of a week, researchers tended to focus their attention on synthesising variants of compounds that had already shown promising effects in a screen, but that might not be ideally suited to be a drug. Any
74
Path dependence in technical change
given compound might have unacceptable side effects, for example, or be very difficult to administer. The ‘design’ of new compounds was a slow, painstaking process that drew heavily on skills in analytic and medicinal chemistry. Several important classes of drugs were discovered in this way, including most of the important diuretics, many of the most widely used psychoactive drugs and several powerful antibiotics. While chemists working within this regime often had some intuitive sense of the links between any given chemical structure and its therapeutic effect, little of this knowledge was codified, so that new compound ‘design’ was driven as much by the skills of individual chemists as it was by a basis of systematic science. Random screening worked extremely well for many years. Several hundred new chemical entities (NCEs) were introduced in the 1950s and 1960s, and several important classes of drug were discovered in this way. However, the successful introduction of NCEs has to be considered as a quite rare event. Indeed, estimates suggest that, out of all new compounds that were discovered only one in 5000 reached the market. So, the rate of introduction has been of the order of a couple of dozens per year, and concentrated in some fast-growing areas such as the central nervous system, cardiac therapy, anti-infectives and cytostatics. Innovative new drugs arrived quite rarely but after the arrival they experienced extremely high rates of market growth. In turn, this entailed a highly skewed distribution of the returns on innovation and of product market sizes as well as of the intrafirm distribution of sales across products. So a few ‘blockbusters’ dominate the product range of all major firms (Matraves, 1999; Sutton, 1998). As is well known, however, new products do not ensure profits. Rents from innovation can be lost through competition unless ‘isolating mechanisms’ are in place to inhibit imitators and new entrants. Indeed, for most of the post-war period, pharmaceutical companies (particularly those operating in the USA) had a number of isolating mechanisms working in their favour. Several of these mechanisms, including the strength of intellectual property protection and the nature of the regulatory regime for pharmaceutical products, were institutional in origin and differed significantly across national boundaries. We discuss these types of mechanisms in more detail below. However, it is important to note that the organisational capabilities developed by the larger pharmaceutical firms may also have acted as isolating mechanisms. Consider, for example, the process of random screening itself. As an organisational process, random screening was anything but random. Over time, early entrants into the pharmaceutical industry developed highly disciplined processes for carrying out mass screening programmes. Because random screening capabilities were based on internal organisational processes and tacit skills, they were difficult for potential
A history-friendly model of innovation
75
entrants to imitate and thus became a source of first-mover advantage. In addition, in the case of random screening, spillovers of knowledge between firms was relatively small since, when firms essentially rely on the law of large numbers, there is little to be learned from the competition. Moreover, entirely new products (NCEs) only capture a part of innovative activities. ‘Inventing around’ existing molecules, the introduction of new combinations among them or new ways of delivering them, and so on, constituted a major component of firms’ innovative activities broadly defined. Thus, competition centred not only on new product introductions but also on incremental advances over time, as well as on imitation and generic competition after patent expiration (allowing a large ‘fringe’ of firms to thrive). Processes of generation of new markets and of diversification across product groups was followed by processes of incremental innovation, development of therapeutic analogues, imitation and licensing. Fastexpanding markets allowed for the steady growth of both the first-comer and other early innovators. The successful exploitation of the economic benefits stemming from innovation also required the control of other important complementary assets, particularly, competencies in the management of large-scale clinical trials, the process of gaining regulatory approval, and marketing and distribution, which also have acted as powerful barriers to entry into the industry. As a consequence, throughout its history, the industry has been characterised by a significant heterogeneity in terms of firms’ strategic orientations and innovative capabilities. Indeed, ever since its inception, other firms specialised not in R&D and innovation, but in the imitation/inventing around, production and marketing of products often invented elsewhere and sold over the counter. This group of firms included companies like Bristol-Myers, Warner-Lambert, Plough, American Home Products as well as almost all the firms in countries like France, Italy, Spain and Japan. Conversely, the ‘oligopolistic core’ of the industry has been composed of the early innovative entrants, joined after World War II by a few American and British firms, which maintained over time an innovation-oriented strategy. The isolating mechanisms discussed previously, combined with the presence of scale economies in pharmaceutical research, and marketing, may help to explain the dearth of new entries prior to the mid-1970s. Indeed, many of the leading firms during this period – companies like Roche, Ciba, Hoechst, Merck, Pfizer, and Lilly – had their origins in the ‘pre-R&D’ era of the industry. At the same time, until the mid-1970s only a small number of new firms entered the industry, and even less entered its ‘core’. At the same time, the industry was characterised by quite low levels of concentration, at the aggregate
76
Path dependence in technical change
level (the pharmaceutical industry) but also in the individual sub-markets like, for example, cardiovascular, diuretics, tranquillizers, and so on. 2.3
The Advent of Molecular Biology
Beginning in the early 1970s, the industry also began to benefit more directly from the explosion in public funding for health-related research that followed the Second World War. From the mid-1970s on, however, substantial advances in physiology, pharmacology, enzymology and cell biology – the vast majority stemming from publicly funded research – led to enormous progress in the ability to understand the mechanism of action of some existing drugs and the biochemical and molecular roots of many diseases. This new knowledge had a profound impact on the process of discovery of new drugs. First, these advances offered researchers a significantly more effective way to screen compounds. In the place of the request ‘find me something that will lower blood pressure in rats’ pharmacologists could make the request ‘find me something that inhibits the action of the angiotensin II converting enzyme in a test tube’. In turn, the more sensitive screens made it possible to screen a wider range of compounds, triggering a ‘virtuous cycle’ in that the availability of drugs whose mechanisms of action was well known made possible significant advances in the medical understanding of the natural history of a number of key diseases, advances which in turn opened up new targets and opportunities for drug therapy. These techniques of ‘guided search’ made use of the knowledge that a particular chemical pathway was fundamental to a particular physiological mechanism. But until quite recently the new knowledge was not used in the design of new compounds that could be tested in such screens. The techniques of ‘rational drug design’ are the result of applying the new biological knowledge to the design of new compounds, as well as to the ways in which they are screened. If, to use one common analogy, the action of a drug on a receptor in the body is similar to that of a key fitting into a lock, advances in scientific knowledge in the 1970s and 1980s greatly increased knowledge of which ‘locks’ might be important, thus making the screening process much more precise. However, organic chemists were still forced to rely on random screening or on the elaboration of existing compounds in their search for new drugs since they had no guidance as to what appropriate ‘keys’ might look like. More recently, an improved understanding of molecular kinetics, of the physical structure of molecular receptors, and of the relationship between chemical structure and a particular compound’s mechanism of action has greatly increased knowledge of what suitable ‘keys’ might look like. Chemists are now beginning to be able to ‘design’ compounds that might have particular therapeutic effects.
A history-friendly model of innovation
77
These techniques were not uniformly adopted across the industry. For any particular firm, the shift in the technology of drug research from ‘random screening’ to one of ‘guided’ discovery or ‘drug discovery by design’ was critically dependent on the ability to take advantage of publicly generated knowledge (Gambardella, 1995; Henderson and Cockburn, 1996) and of economies of scope within the firm (Henderson and Cockburn, 1996). Smaller firms, those farther from the centres of public research and those that were most successful with the older techniques of rational drug discovery, appear to have been much slower to adopt the new techniques than their rivals (Gambardella, 1995; Henderson and Cockburn, 1996). There was also significant geographical variation in adoption. While the larger firms in the USA, the UK and Switzerland were amongst the pioneers of the new technology, other European and Japanese firms appear to have been slow to respond to the opportunities afforded by the new science. These differences had significant implications for the industry’s later response to the revolution in molecular biology. This transition was in mid-course when molecular genetics and recombinant DNA technology opened an entirely new frontier for pharmaceutical innovation. The application of these advances initially followed two relatively distinct technical trajectories. One trajectory was rooted in the use of genetic engineering as a process technology to manufacture proteins whose existing therapeutic qualities were already quite well understood in large enough quantities to permit their development as therapeutic agents. The second trajectory used advances in genetics and molecular biology as tools to enhance the productivity of the discovery of conventional ‘small molecule’ synthetic chemical drugs. More recently, as the industry has gained experience with the new technologies, these two trajectories have converged. The advent of ‘biotechnology’ had a significant impact both on the organisational competencies required to be a successful player in the pharmaceutical industry through their impact on the competencies required to discover ‘conventional’, small molecular weight drugs and on industry structure in general. In the USA, biotechnology was the motivating force behind the first large-scale entry into the pharmaceutical industry since the early post-Second World War period. The first new biotechnology start-up, Genentech, was founded in 1976 by Herbert Boyer (one of the scientists who developed the recombinant DNA technique) and Robert Swanson, a venture capitalist. Genentech constituted the model for most of the new firms. They were primarily university spin-offs and they were usually formed through collaboration between scientists and professional managers, backed by venture capital. Their specific skills resided in the knowledge of the new techniques and in the research capabilities in that area.
78
Path dependence in technical change
Their aim consisted in applying the new scientific discoveries to commercial drug development, focusing on two main directions: diagnostics, on the basis of monoclonal antibodies, and therapeutics. Genentech was quickly followed by a large number of new entrants. Entry rates soared in 1980 and remained at a very high level thereafter, favoured also by the availability of venture capital and by the gradual establishment of a very favourable climate concerning patenting. Patents constituted a major problem. Particularly in the early stages, the relevant knowledge was to a large extent of a generic nature and could in principle be codified. Since the product of new biotechnology firms (NBFs) was essentially scientific results, patents were therefore crucial requisites for the private appropriation of the profits generated by innovations. Yet considerable confusion surrounded the conditions under which patents could be obtained. These hurdles were gradually overcome, in the direction of granting ample concessions to industry. In particular, in 1980 Congress passed the Patent and Trademark Amendments Act (Bayh-Dole Act), which in effects liberalised and actually encouraged the pursuit of patent protection for inventions funded by government agencies. On the other hand, again in 1980, the US Supreme Court ruled in favour of granting patent protection to living things (Diamond v. Chakrabarty) and in subsequent years, a number of patents were granted establishing the right for very broad claims (Merges and Nelson 1994). Despite the high rates of entry, it took several years before the biotechnology industry started to have an impact on the pharmaceutical market. The first biotechnology product, human insulin, was approved in 1982, and between 1982 and 1992, 16 biotechnology drugs were approved for the US market. As is the case for small molecular weight drugs, the distribution of sales of biotechnology products is highly skewed. Three products were major commercial successes: insulin (Genentech and Eli Lilly), tPA (Genentech in 1987) and erythropoietin (Amgen and Ortho in 1989). By 1991 there were over 100 biotechnology drugs in clinical development and 21 biotechnology drugs with submitted applications to the Food and Drug Administration (FDA) (Grabowski and Vernon, 1994; Pharmaceutical Manufacturers Association, 1991): this was roughly one-third of all drugs in clinical trials (Bienz-Tadmor et al., 1992). Sales of biotechnologyderived therapeutic drugs and vaccines had reached $2 billion, and two new biotechnology firms (Genentech and Amgen) have entered the club of the top eight major pharmaceutical innovators (Grabowski and Vernon, 1994). However, the large majority of these new companies never managed to become a fully integrated drug producer. The growth of NBFs as pharmaceutical companies was constrained by the need to develop competencies in different crucial areas.
A history-friendly model of innovation
79
First, it was necessary to understand better the biological processes involved with proteins and to identify the specific therapeutic effects of such proteins. Companies, in fact, turned immediately to produce those proteins (for example, insulin and the growth hormone) which were sufficiently well known. The subsequent progress of individual firms and of the industry as a whole was, however, predicated on the hope of being able to develop a much deeper knowledge of the working of other proteins in relation to specific diseases. Yet, progress along this line proved more difficult than expected. Second, these companies lacked competencies in other different crucial aspects of the innovative process: in particular, knowledge and experience of clinical testing and other procedures related to product approval, on the one hand, and marketing, on the other. Thus, they exploited their essential competence and acted primarily as research companies and specialised suppliers of high-technology intermediate products, performing contract research for and in collaboration with established pharmaceutical corporations. Collaboration allowed NBFs to survive and – in some cases – to pave the way for subsequent growth in many respects. First, clearly, it provided the financial resources necessary to fund R&D. Second, it provided the access to organisational capabilities in product development and marketing. Established companies faced the opposite problem. While they needed to explore, acquire and develop the new knowledge, they had the experience and the structures necessary to control testing, production and marketing. Indeed, large established firms approached the new scientific developments mainly from a different perspective, that is, as tools to enhance the productivity of the discovery of conventional ‘small molecule’ synthetic chemical drugs. There was enormous variation across firms in the speed with which the new techniques were adopted. The adoption of biotechnology was much less difficult for those firms who had not made the transition from ‘random’ to ‘guided’ drug discovery. For them, the tools of genetic engineering were initially employed as another source of ‘screens’ with which to search for new drugs. Their use in this manner required a very substantial extension of the range of scientific skills employed by the firm; a scientific workforce that was tightly connected to the larger scientific community and an organisational structure that supported a rich and rapid exchange of scientific knowledge across the firm (Gambardella, 1995; Henderson and Cockburn, 1996). The new techniques also significantly increased returns to the scope of the research effort (Henderson and Cockburn, 1996). In general, the larger organisations who had indulged a ‘taste’ for science under the old regime were at a considerable advantage in adopting the new techniques compared with smaller firms. On the contrary, firms that had
80
Path dependence in technical change
been particularly successful in the older regime and firms that were much less connected to the publicly funded research community were much slower to follow their lead. The embodiment of the new knowledge was in any case a slow and difficult process, because it implied a radical change in research procedures, a redefinition of the disciplinary boundaries within laboratories and, in some cases, in the divisional structure of the company as well. Collaborative research with the NBFs and with universities allowed these companies, in any case, to get access to the new technology and to experiment in alternative directions. The advantages stemming from these interactions could be fully exploited, however, only through the contextual development of in-house capabilities, which made it possible to absorb and complement the knowledge supplied by external sources (Arora and Gambardella, 1992). Collaboration with universities, NBFs and internal research were, indeed, strongly complementary. Thus, a dense network of collaborative relations emerged, with the startup firms positioned as upstream suppliers of technology and R&D services, and established firms positioned as downstream buyers who could provide capital as well as access to complementary assets. Networking was facilitated by the partly ‘scientific’, that is, abstract and codified nature of the knowledge generated by NBFs (Gambardella, 1995), which made it possible, in principle, to separate the innovative process in different vertical stages: the production of new scientific knowledge, the development of this knowledge in applied knowledge, and the use of the latter for the production and marketing of new products. In this context, different types of institutions specialised in the stage of the innovative process in which they were relatively more efficient: universities in the first stage, NBFs in the second stage and large firms in the third. A network of collaboration between these actors provided the necessary co-ordination of the innovative process. The new firms acted as intermediaries in the transfer of technology between universities – which lacked the capability to develop or market the new technology – and established pharmaceutical firms that lacked technical expertise in the new realm of genetic engineering but that had the downstream capabilities needed for commercialisation. However, substantial costs remained in transferring knowledge across different organisations, especially for the tacit and specific component of knowledge. Moreover, the innovative process still involved the effective integration of a wide range of pieces of knowledge and activities, which were not ordered in a linear way and might not easily be separated (Orsenigo, 1989). Thus, the processes of drug discovery and drug development still required the integration of different disciplines, techniques, search and experimental procedures and routines, which were not generally separable and codified.
A history-friendly model of innovation
81
Hence, firms were encouraged to pursue higher degrees of vertical integration. Thus, some NBFs tried to vertically integrate downstream to production and marketing, becoming fully-fledged pharmaceutical companies and directly challenging incumbents. Some of the latter tried to vertically integrate upstream, gaining full mastery of the new knowledge. In recent years, moreover, a significant process of consolidation has begun, with both mergers between biotechnology firms (for example, Cetus and Chiron), as well as the acquisition of biotechnology firms by pharmaceutical corporations (for example, Genentech and Hoffman LaRoche). Ernst and Young (1990) reported that the large majority (71 per cent of the companies surveyed) of the NBFs expected to be acquired by a large firm or to merge with another NBF. However, since knowledge is still fragmented and dispersed, and since the rate of technological change is still very high, no single institution is able to develop internally in the short run all the necessary ingredients for bringing new products to the marketplace. Each NBF, in effects, represents a possible alternative approach to drug discovery and a particular instantiation of the opportunities offered by the progresses of science. New generations of NBFs have been created which adopt different approaches to the use of biotechnology in the pharmaceutical industry. Large established corporations continue, therefore, to explore these new developments through collaborative agreements. 2.4
Institutional Environments
The proliferation of NBFs was essentially an American (and partly British) phenomenon. The development of the biotechnology segment in Europe and Japan lagged considerably behind the USA and rested on the activities of large established companies. The British and the Swiss companies moved earlier and more decisively in the direction pioneered by the large US firms in collaborating or acquiring American startups. But those firms that had smaller research organisations, were more local in scope or were more orientated towards the exploitation of well-established research trajectories – in short, those firms that had not adopted the techniques of ‘rational’ or ‘guided’ drug discovery – have found the transition more difficult (Gambardella, 1995; Henderson and Cockburn, 1996): almost all the established French, Italian and Japanese companies – but also the German giants – have been slow to adopt the tools of biotechnology as an integral part of their drug research efforts. More generally, ever since the mid-1970s the American, British and Swiss companies appear to have gained significant competitive advantages vis-à-vis European firms, including the Germans. And traditionally the
82
Path dependence in technical change
continental European (except Germany and Switzerland) and Japanese industries have been much less orientated towards innovation than to strategies based on imitation, production and marketing mainly for the domestic market. While the reasons for these differentiated patterns of evolution are still controversial, institutional factors seem to have played a decisive role. Indeed, from its inception, the evolution of the pharmaceutical industry has been tightly linked to the structure of national institutions. The pharmaceutical industry emerged in Switzerland and Germany, in part, because of strong university research and training in the relevant scientific areas. Organic chemistry was literally invented in Germany by Professor Justus Leibig, and German universities in the nineteenth century were leaders in organic chemistry. Basel, the centre of the Swiss pharmaceutical industry, was the home of the country’s oldest university, long a centre for medicinal and chemical study. In the USA the government’s massive wartime investment in the development of penicillin, as we discussed above, profoundly altered the evolution of American industry. In the post-war era, the institutional arrangements surrounding the public support of basic research, intellectual property protection, procedures for product testing and approval, and pricing and reimbursement policies have all strongly and directly influenced both the process of innovation and the economic returns (and thus incentives) for undertaking such innovation. We now turn to a brief review of these four key areas. Public support for health-related research Nearly every government in the developed world supports publicly funded health-related research, but there are very significant differences across countries in both the level of support offered and in the ways in which it is spent. In the USA, public spending on health-related research took off after the Second World War and it is now the second largest item in the federal research budget after defence. Most of this funding is administered through the National Institutes of Health (NIH), although a significant fraction goes to universities. Both qualitative and quantitative evidence suggests that this spending has had a significant effect on the productivity of those large US firms that were able to take advantage of it (Henderson and Cockburn, 1996; Maxwell and Eckhardt, 1990). Public funding of biomedical research also increased dramatically in Europe in the post-war period, although total spending did not approach American levels. Moreover, the institutional structure of biomedical research evolved quite different in continental Europe as opposed to the USA and the UK. For example, in continental Europe biomedical research was mainly concentrated in national laboratories rather than in medical
A history-friendly model of innovation
83
schools as happened in the USA and the UK. These differences in the levels and sources of funds, along with a number of other institutional factors, have interacted in continental Europe to create an environment which, in general, not only produces less science of generally lower quality, but also one in which science is far less integrated with medical practice. Intellectual property protection Pharmaceuticals has historically been one of the few industries where patents provide solid protection against imitation (Klevorick et al., 1987). Because small variants in a molecule’s structure can drastically alter its pharmacological properties, potential imitators often find it hard to work around the patent. Although other firms might undertake research in the same therapeutic class as an innovator, the probability of them finding another compound with the same therapeutic properties that did not infringe on the original patent could be quite small. However, the scope and efficacy of patent protection has varied significantly across countries. Both the USA and the majority of the European countries have provided relatively strong patent protection in pharmaceuticals. In contrast, in Japan and in Italy, until 1976 and 1978 (respectively), patent law did not offer protection for pharmaceutical products; only process technologies could be patented. As a result, Japanese and Italian firms tended to avoid product R&D and to concentrate instead on finding novel processes for making existing molecules. Procedures for product approval Pharmaceuticals are regulated products. Procedures for approval have a profound impact on both the cost of innovating and on firms’ ability to sustain market positions once their products have been approved. As in the case of patents, there are substantial differences in product approval processes across countries. Since the early 1960s most countries have steadily increased the stringency of their approval processes. However, it was the USA, with the Kefauver-Harris Amendment Act in 1962, and the UK, with the Medicine Act in 1971, that took by far the most stringent stance among industrialised countries. Germany and especially France, Japan and Italy have historically been much less demanding. In the USA, the 1962 Amendment Act introduced a proof-of-efficacy requirement for approval of new drugs and established regulatory controls over the clinical (human) testing of new drug candidates. Specifically, the Amendments required firms to provide substantial evidence of a new drug’s efficacy based on ‘adequate and well controlled trials.’ As a result, after 1962 the FDA shifted from a role as essentially an evaluator of evidence
84
Path dependence in technical change
and research findings at the end of the R&D process, to an active participant in the process itself (Grabowski and Vernon, 1983). The effects of the Amendments on innovative activities and market structure have been the subject of considerable debate (see, for instance, Chien, 1979, and Peltzman, 1974). They certainly led to large increases in the resources necessary to obtain approval of a new drug application (NDA), and they probably caused sharp increases in both R&D costs and in the gestation times for NCEs, along with large declines in the annual rate of NCE introduction for the industry and a lag in the introduction of significant new drugs therapies in the USA when compared with Germany and the UK. However, the creation of a stringent drug approval process in the USA may also have helped create an isolating mechanism for innovative rents. Although the process of development and approval increased costs, it significantly increased barriers to imitation, even after patents expired.3 The institutional environment surrounding drug approval in the UK was quite similar to that in the USA. As in the USA, the introduction of a tougher regulatory environment in the UK was followed by a sharp fall in the number of new drugs launched into Britain and a shakeout of the industry. A number of smaller weaker firms exited the market and the proportion of minor local products launched into the British market shrunk significantly. The strongest British firms gradually reoriented their R&D activities towards the development of more ambitious, global products (Thomas, 1994). In other European countries, procedures for products approval were less stringent. This allowed the survival of smaller firms specialised in the commercialisation of minor domestic products. The structure of the health-care system and systems of reimbursement Perhaps the biggest differences in institutional environments across countries was in the structure of the various health-care systems. In the USA, pharmaceutical companies’ rents from product innovation were further protected by the fragmented structure of health-care markets and by the consequent low bargaining power of buyers. Moreover, unlike most European countries (with the exception of Germany and the Netherlands) and Japan, drug prices in the USA are unregulated by government intervention. Until the mid-1980s the overwhelming majority of drugs were marketed directly to physicians who largely made the key purchasing decisions by deciding which drug to prescribe. The ultimate customers – patients – had little bargaining power, even in those instances where multiple drugs were available for the same condition. Because insurance companies generally did not cover prescription drugs
A history-friendly model of innovation
85
(in 1960, only 4 per cent of prescription drug expenditures were funded by third-party payers), they did not provide a major source of pricing leverage. Pharmaceutical companies were afforded a relatively high degree of pricing flexibility. This pricing flexibility, in turn, contributed to the profitability of investments in drug R&D. Drug prices were also relatively high in other countries that did not have strong government intervention in prices, such as Germany and the Netherlands. In the UK, price regulation left companies to set their own prices, but a global profit margin with each firm was negotiated which was designed to assure each of them an appropriate return on capital investment including research. The allowed rate of return was negotiated directly and was set higher for export-oriented firms. In general, this scheme tended to favour both British and foreign R&D-intensive companies which operated directly in the UK. Conversely, it tended to penalise weak, imitative firms as well as those foreign competitors (primarily the Germans) trying to enter the British market without direct innovative effort in the UK (Burstall, 1985; Thomas, 1994). On the contrary, in Japan, France and Italy price regulation was organised in such a way to protect the domestic industry from foreign competition and offered little incentive to ambitious innovative strategies (Henderson et al. 1999; Thomas, 1994). In more recent times, the introduction of cost containment policies in almost all countries has led to profound changes in these systems and intense debates about the efficiency of alternative systems in resolving the trade-off between lower prices and incentives for innovation.
3 3.1
THE MODEL Challenges for a History-friendly Model
As was discussed in section 2, there are several important conceptual issues that are raised by an analysis of the evolution of the pharmaceutical industry. In particular, we mentioned three of them: the relationships between the properties of the regimes of search, the nature of markets, the patterns of competition and the evolution of market structure; the relationships between science and innovation; and the role and the impact of alternative forms of public policy and regulation. In this chapter, we will restrict the analysis to the era of random screening and we will address a subset of these issues. Here, the thrust of the story can be summarised as follows. Firms compete to discover, develop and market new drugs for a large variety of
86
Path dependence in technical change
diseases. They face a large space of – at the beginning – unexplored opportunities. However, the search for new promising compounds is essentially random, because the knowledge of why a certain molecule can ‘cure’ a particular disease and of where that particular molecule can be found is limited. That is to say, the role of ‘science’ here is modest. Thus, firms explore randomly the ‘space of molecules’ until they find one which might become a useful drug and on which they obtain a patent. The patent provides protection from imitation for a certain amount of time (patent duration) and over a given range of ‘similar’ molecules (width of patents). After patenting, firms engage in the development of the drug, without knowing how difficult, time-consuming and costly the process will be and what the quality of the new drug will be. Then, the drug is sold on the market, whose size is defined by the number of potential patients. Marketing expenditures allow firms to increase the number of patients they can access. At the beginning, the new drug is the only product available on that particular therapeutic class. But other firms can discover competing drugs or imitate. Indeed, firms are characterised by different propensities towards innovation, on the one hand, and imitation and marketing, on the other. Innovators will therefore experience a burst of growth following the introduction of the new drug, but later its revenues and market shares will be eroded away by competitors and imitators. Since discovery of a drug in a particular therapeutic class does not entail any advantage in the discovery of another drug in a different class (market) – except for the volume of profits they can reinvest in research and development – firms will start searching randomly again for a new product everywhere in the space of molecules. Firms’ growth will then depend on the number of drugs they have discovered (that is, in diversification into different therapeutic categories), on the size of the markets they are in, on the number of competitors, and on the relative quality and price of their drug vis-à-vis competitors. Given the large number of therapeutic categories and the absence of any form of cumulativeness in the search and development process, no firm can hope to be able to win a large share in the overall market, but – if anything – only in specific therapeutic categories for a limited period of time. As a result, the degree of concentration in the whole market for pharmaceuticals will be low. However, a few firms will grow and become large, thanks essentially to diversification. Market structure is also likely to be heavily affected by institutional variables. Essentially, in the previously recounted history and in the model this is due to the patenting regime, to the strictness of the procedures for product approval and to the forms of price regulation. Here we look only at the former two variables, leaving the analysis of price controls to future exercises.
A history-friendly model of innovation
87
Patents constitute a fundamental means for the appropriability of the economic benefits of innovation. Weak patenting regimes reduce the viability of innovative strategies vis-à-vis imitative, marketing-based strategies. Thus, the strengthening (weakening) of the degree of protection offered by patents should increase the rate of innovation and generate higher (and more persistent) degrees of concentration. However, the relationship between the tightness of the patenting regime and the rate of innovation is not necessarily a linear one. Beyond (below) certain maximum (minimum) levels, further increases (decreases) of the protection provided by patents may have little effect. The introduction of more stringent procedures for product approval has often been indicated as leading to a reduction of the number of new drugs, to an increase of their quality, to the exit of smaller firms and to the growth of more innovative companies. Countries characterised by a more lenient approach in this respect are reported to have been losing competitiveness in the long run, especially as far their ability to innovate is concerned. 3.2
The Topography of the Model
In this section we describe the basic structure of the model. The technological and market environment in which pharmaceutical firms are active is composed by several therapeutic categories (TCs). Each TC has a different economic dimension according to the number of potential customers. This economic size is expressed by the total potential sales (VTC) and it is exogenously given in the model. In our model there are n therapeutic categories TC, each of which has a specific VTC. VTC is set at the beginning of each simulation, it is a random number drawn from a normal distribution [VTC N( V, V)]. VTC grows in every period at a certain rate, ranging randomly between 0 and 2 per cent. Firms active in a certain TC get a share of VTC equal to their market share. Within each TC there are a certain number M of molecules, which firms aim to discover and which are at the base of pharmaceutical products that later are introduced in the market. Each molecule has a certain quality Q that could be visualised as the ‘height’ of a certain molecule (see Figure 4.1). In most of the cases (70 per cent), Q has a value equal to zero. In the other 30 per cent of cases, it has a positive value, drawn from a normal distribution [Q N( Q, Q)]. Figure 4.1 depicts the ‘landscape’ in term of therapeutic areas and molecules that firms face. Firms do not know the ‘height’ Q of a molecule. Once they engage in a search process in a specific therapeutic category, they may ‘discover’ a molecule or not. In case they do, firms start a research process (see below): they obtain a patent only if the molecule has a positive quality. Molecules whose
88
Q
Path dependence in technical change
TC1
Figure 4.1
TC2
TCn
Therapeutic categories and molecules
quality does not pass the FDA quality check will not give an origin to their correlated products. In the standard run, this quality threshold (FDA) is set at fixed value. A patent has a specific duration (pd) and width (w). Once patent duration expires, the molecule becomes free for all the firms. A patent gives a firm the right to extent the protection also on the molecules situated in the ‘neighbourhood’ – as defined by the width – of the molecule that has been patented. The protection in the neighbourhood of the existing patent by the innovator has major consequences on the search process of competing firms. Competing firms are in fact blocked in the developments of potential molecules near the patented one. Once the patent has been granted, the firm can start the development of the product based on that molecule. If product development is successful, the product gets an economic value PQ. The value of the product PQi is a function of the value of the molecule Qi. That is: PQi (1 )Qi
(4.1)
where i 1, 2 . . . 150 for each TC and U[ .25, .25]. Each product gives a certain level of utility to consumers (see section 2.3 for a discussion of demand). The value of the product influences the consumers demand for such drug. 3.3
The Firms
3.3.1 The basic features of firms The industry is populated by f firms. Each firm has a budget B, initially equal for all firms. Firms are characterised by three activities – search, research and marketing – but with different intensity in each activity. Some
A history-friendly model of innovation
89
firms, in fact, may want to spend relatively more on research and less on marketing; other firms the contrary. In the model, the marketing propensity of the firms, , is characterised by a share of the budget randomly set in the interval [0.2–0.8]. Relatedly the search–research propensity by firms is characterised by a share of the budget that complements the share of the propensity to marketing, which is (1 ). Thus, the firms’ budget B is divided among search, research and marketing activities according to the specific propensities discussed above: Resources for search (BS): (1 )!B Resources for research (BR): (1 ) (1 !)B Resources for marketing (BM): B where ! (randomly drawn from a uniform distribution ranging between 0.05 and 0.15) is invariant and firm specific. 3.3.2 Innovators and imitators Firms are heterogeneous also in another respect: they can be innovators or imitators. The propensity to research determines whether a firm is an innovator or an imitator. If a firm has a propensity to research (1 ) greater than r (a random number from a uniform distribution ranging from 0 to 1 in each period), then the firm is an innovator. If the firm has a propensity to research (1 ) lower than r, then it will be an imitator. In this way, innovative or imitative nature is a time dependent variable of the specific firm: while the propensity to research is given initially and does not change over time, r does. 3.3.2.1 Innovative search activities If firms are innovators, they look for new molecules. The amount of money invested in search activities, BS, determines the number (approximated to the nearest integer) X of therapeutic categories TCs which are explored by a firm during its current project: X BS/Drawcost
(4.2)
where Drawcost is a parameter fixed in our simulations. This simple specification implies that the number of TCs explored is linear with that fraction of budget BS that any one firm allocates for search processes. If X is lower than 1, then the firm is assumed to be able to explore only one TC. 3.3.2.2 Imitative search activities If a firm follows an imitation strategy, after having drawn a certain number of TCs (as defined in equation (4.3)), it looks for an existing molecule which is free (and thus not protected by
90
Path dependence in technical change
a patent any more). Imitators rate molecules according to their quality Qi which is, however, only imperfectly known. They select the molecule with the highest ‘perceived’ quality, Ri . R is also a measure of the probability of choosing a molecule, that is a function of Q: Ri (1 )Qi
(4.3)
where i 1, 2 . . . 150 for each TC and U[ .25, .25]. Hence, high-quality molecules will be more frequently picked up by imitators, generating a congestion effect. 3.3.2.3 Research activities By ‘research activities’ we mean product development. Both innovators and imitators do research. If the molecule is potentially interesting (that is, it has a quality Q greater than zero), the firm starts a development project, using the budget BR. Remember that firms do not know the quality of the drug. Over time, given the research budget BR the firm progresses towards the full development of the drug (attaining the value Q of the drug). That is to say, firms have to ‘climb’ Q steps in order to develop a drug having a quality Q. Each step implies a unitary cost CS. Thus, the total cost of developing a drug having a quality Q is equal to CSQ. In each period a firm pays CS. However, firms differ in the speed of their development process: higher speed implies higher costs. In each period, the progress that a firm makes towards Q (SP) is randomly drawn, and ranges from 1 to 5. Firms that move ahead faster in their research per period, pay more for each unitary step the unitary cost CS of each step increases as SP increases according to the following relationship: i SP
CS
(Cur
i) i 1
SP
(4.4)
where Cur is the cost of a single step for a firm that has a SP equal to 1 (that is, it progresses with 1 step each period: only in this case CS Cur). CS for imitative firms is set ¼ of the CS of innovating firms. In our simulations, Cur is equal to 15 and is fixed for every firm. With its research resources, a firm may be able to reach Qi. In this case it starts the commercialisation of the product. Otherwise, if Q is too ‘high’ for the resources BR of the firm, the project fails with the firm. Moreover, a product must have a minimum quality, already defined as FDA, to be allowed to be sold in the marketplace. Below this value the drug cannot be
A history-friendly model of innovation
91
commercialised and the project fails. Only when research activities are over (that is, after a product has been created), will the firm start another process of search. 3.3.2.4 Marketing activities As previously mentioned, if a firm reaches Qi, it has to launch a ‘new born’ product on the market. The firm has a budget BM available for that purpose. BM is divided by the firm in two parts, with shares h and 1 h (equal for all firms). BM h defines the marketing investment AjTL for the product j, where TL refers to the launch of the product. It is spent only once, at the moment of the launch of the product. Marketing expenditures yield a certain level of ‘product image’ for the consumers. This level of the ‘image’ is eroded with time at a rate equal to eA in each period. In addition, the firm will profit from a marketing spillover from its previous products kj. The level of the ‘image’ Ajt in period t is given by:
Ak for t TL Ajt Ajt 1·(1 eA) Akfor t TL 1, . . .,100 Ajt AjTL
kj
(4.5)
kj
BM (1 h) defines the total yearly marketing expenditures that will be spent over TA periods. This factor captures the firms’ attempt of keeping the level the ‘image’ over time. Yearly marketing expenditures YAt will therefore be: YAt
(1 h)BM TA
t TL, . . ., TL TA
(4.6)
In our simulations, h is equal to 0.5, erosion (eA) 0.01 and TA 20. 3.3.3 Utility, demand and market share Decisions to buy a specific drug depend on several factors, which together yield a specific ‘merit’ to each drug j. The value of this ‘merit’, Uj, is given by: Ujt PQ aj · (1mup)b · Acjt · YAdjt
(4.7)
PQj is the economic value of the drug as in equation (4.1). mup is the desired rate of return that each firm wants to obtain from its drug. Ajt is the product ‘image’ derived by the marketing investment for that product and YAjt is the yearly marketing expenditure for the product j, as already
92
Path dependence in technical change
defined. Exponents a, b, c are drawn from uniform distributions and are specific to each therapeutic category; on the contrary d is equal in all markets. Finally mup is double for innovative products (mupinn) compared with that applied to imitative products (mupimi). The market share MSfj of the firm f for the product j in each TC is then proportional to its relative merit as compared to other competing drugs in the same TC and it is given by: MSfj
Ufj
TC
(4.8)
UTC
Please keep in mind that a firm may have more than one product in a TC. Thus its market share in a TC is the sum of the market shares for all its products. 3.3.4 Budget and accounting Revenues of firm f for product j are fj; because firm f may have more than one product, total revenues ( tot) are the sum of revenues obtained from all the products of the firm, whatever TC has been explored: tot
ij (MSfj · VTC ) fj
fjk
k
f 1, . . .,30; k 1, . . .,50
(4.9)
In each period, the excess gross profits, that is, the difference between revenues, the current costs search, development and launch of the new products (in the periods when these activities take place) and the yearly expenditures on marketing, accumulate in an account that is used as a budget to finance search, research and marketing investment when a new project is started. The division of the budget among the different activities follows the firmspecific parameters ( and !) already defined in the section 3.3.1. Firms exit the market when their budget falls to zero.
4 4.1
THE SIMULATION RUNS The Dynamics of Market Structure and Innovation
In the standard case (Standard), 30 firms with the same budget start their innovative and imitative activities. They differ in their propensity to research and marketing, and in their rate of progress. These firms search in a space composed of 50 therapeutic areas; each of them has 150 molecules (see the Appendix for the full set of parameters).
A history-friendly model of innovation
93
The dynamics of innovation and market structure in this industry is as follows. Innovative firms start doing their innovative activities by searching over the therapeutic categories. Some of them will succeed to find a positive-quality molecule; they patent it and start doing research on that molecule. If they succeed in completing their research, they start doing marketing activities for the support of their sales in the market. Demand for the product will go to those firms that have higher product quality, higher initial marketing investments and higher marketing yearly expenditures, with different importance depending on the specific therapeutic category. Firms with larger demand will obtain higher market shares and higher profits. Once the patent expires, the molecule becomes available on the market for imitative activities. Imitators do not have to do any search for new molecules. Rather, they look around among existing available molecules whose patents have expired. They choose a molecule whose quality is higher than a given value. Then they start doing their research on that molecule, paying lower costs, and they proceed as in the case of innovators (with the exception that, once their research is over, they do not obtain a patent). Then they start doing their marketing activities. 4.2
The Standard Run
The dynamics of market structure and innovation has been examined for 100 periods and the values shown in the figures are averages over 100 runs. The standard runs (Standard) show that in each therapeutic category concentration (in terms of the Herfindahl index) is quite high and a firm monopolizes the market (Figures 4.2 and 4.3). However, the overall market concentration is very high at the beginning, and then drastically declines as more firms discover more molecules in various therapeutic areas (Figure 4.4). Selection is intense and 10 out of the 30 initial firms exit the market (Figure 4.5). In the long run, almost all the therapeutic areas are discovered (Figure 4.6). In each therapeutic area there are an increasing number of products (Figure 4.7a) and firms (Figure 4.7b). At the beginning there are only innovative products, and then also imitative products (Figure 4.8a). With the passage of time, the share of innovative products on the total products in the market (what we have called the Innovation Index) declines from 1 to approximately 0.6 at the end of the run (Figure 4.9). On average each firm is increasingly present in more than one therapeutic category. They reach around 17 at the end of the run (see Figure 4.10). Finally, the value of the molecules discovered by firms over the total potential value of the market (called Performance Index) increases with time and reaches 11 percent of the total potential value of the whole market (Figure 4.11).
94
1
4
Figure 4.2
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1
Herfindahl index in each TC
Standard
Sim 2
Sim 3
Sim 4
7 10 13 16 19 22 25 28 31 34 37 40 43 46 49 52 55 58 61 64 67 70 73 76 79 82 85 88 91 94 97 100 Time
95
1
Figure 4.3
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1
15
22
29
Share of leader in each TC
8
Standard
36
43
Sim 2
50 Time
Sim 3
57
64
Sim 4
71
78
85
92
99
96
1
Figure 4.4
0
0.05
0.10
0.15
0.20
0.25
0.30
0.35
0.40
0.45
0.50
0.55
0.60
6
16
21
26
31
36
41
Standard
Herfindahl index – total market
11
46
Sim 2
Time
51
56
Sim 3
61
66
Sim 4
71
76
81
86
91
96
97
1
5
Figure 4.5
0
10
20
30
40
50
60
70
80
90
10 0
13
17
Surviving firms
9
21
25
33
41
Standard
37
45
Time
53
Sim 2
49
57
Sim 3
61
65
73
Sim 4
69
77
81
85
89
93
97
98
1
4
Figure 4.6
0
15
30
45
60
Number of discovered TCs
Standard
Sim 2
Sim 3
Sim 4
7 10 13 16 19 22 25 28 31 34 37 40 43 46 49 52 55 58 61 64 67 70 73 76 79 82 85 88 91 94 97 100 Time
99
1
4
Average number of products in each TC
Standard
Sim 2
Sim 3
Sim 4
7 10 13 16 19 22 25 28 31 34 37 40 43 46 49 52 55 58 61 64 67 70 73 76 79 82 85 88 91 94 97 100 Tim e
Figure 4.7a
0
5
10
15
20
25
30
35
40
100
1
4
Average number of firms in each TC
Standard
Sim 2
Sim 3
Sim 4
7 10 13 16 19 22 25 28 31 34 37 40 43 46 49 52 55 58 61 64 67 70 73 76 79 82 85 88 91 94 97 100 Tim e
Figure 4.7b
0
3
6
9
12
15
18
21
24
27
30
101
Imitative Standard
Innovative Standard
Total number of innovative and imitative products
1 4 7 10 13 16 19 22 25 28 31 34 37 40 43 46 49 52 55 58 61 64 67 70 73 76 79 82 85 88 91 94 97 100 Time
Figure 4.8a
0
50
100
150
200
250
300
102
Imitative Sim 2
Innovative Sim 2
Total number of innovative and imitative products
Time
1 4 7 10 13 16 19 22 25 28 31 34 37 40 43 46 49 52 55 58 61 64 67 70 73 76 79 82 85 88 91 94 97 100
Figure 4.8b
0
200
400
600
800
1000
1200
1400
103
Innovative Sim 3 Imitative Sim 3
Total number of innovative and imitative products
1 4 7 10 13 16 19 22 25 28 31 34 37 40 43 46 49 52 55 58 61 64 67 70 73 76 79 82 85 88 91 94 97 100 Time
Figure 4.8c
0
50
100
150
200
250
300
104
1
4
Figure 4.8d
0
20
40
60
80
100
120
140
160
180
Total number of innovative and imitative products
7 10 13 16 19 22 25 28 31 34 37 40 43 46 49 52 55 58 61 64 67 70 73 76 79 82 85 88 91 94 97 100 Time
Innovative Sim 4 Imitative Sim 4
105
1
Figure 4.9
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1.0
6
16
21
26
31
41
Standard
36 Time
51
Sim 2
46
61
Sim 3
56
Innovation Index (share of innovative products on total products)
11
71
Sim 4
66
76
81
86
91
96
106
Number of TCs in which firms are active
Standard
Sim 2
Sim 3
Sim 4
1 4 7 10 13 16 19 22 25 28 31 34 37 40 43 46 49 52 55 58 61 64 67 70 73 76 79 82 85 88 91 94 97 100 Time
Figure 4.10
0
5
10
15
20
25
107
1
4
Standard
Sim 2
Sim 3
Sim 4
7 10 13 16 19 22 25 28 31 34 37 40 43 46 49 52 55 58 61 64 67 70 73 76 79 82 85 88 91 94 97 100 Time
Figure 4.11 Performance Index (aggregated quality of discovered molecules over total available quality)
0
0.05
0.10
0.15
0.20
0.25
0.30
0.35
108
Path dependence in technical change
These results are the outcome of very different microdynamics in each therapeutic class. Results (not reported here) show that in few therapeutic areas are there no firms at all. Interestingly enough, in spite of some differences, the Herfindahl index is always rather similar. This shows that a leader soon emerges and that the rest of the firms remain rather marginal in that therapeutic area. Similarly, the number of products (not reported here) may range from zero to more than 10. We have reported, on the other hand, the number of innovative and imitative products in each therapeutic area: in most of the areas an innovative product corresponds to an imitative product (see Figure 4.12a). 4.3
Alternative Runs
The standard set reproduces some stylised facts of the pharmaceutical industry related to the level of concentration and the relationship between innovators and imitators. We then did three counterfactual exercises regarding the role of the initial number of firms, the changes in patent protection and the approval procedures in terms of quality. 4.3.1 Increase in the number of firms What would have happened if the number of firms would have been much higher, growing from 30 to 100 (Simulations 2)? As one would expect, concentration in each therapeutic categories would have been lower, and the overall concentration even lower. Nearly all the therapeutic areas would have been discovered, with a higher number of firms in the market and in each therapeutic category. On average, there would have been a greater number of innovative and imitative products in each therapeutic category. Moreover the value of the discovered molecules would increase to 40 per cent of the total potential value. 4.3.2 The extension of time of patent protection What would have happened if patent protection would be extended from 20 periods to 60 periods (Simulations 3)? Interestingly enough, concentration in each therapeutic area would not increase, because of its already high level. Only the number of firms in each therapeutic area slightly declines. On the contrary, the number of imitative products drastically decline, which in turns brings up the Innovation Index and reduces the number of TCs in which firms are active. 4.3.3 Increase in the stringency of approval procedures We also did some other runs on the increase in the stringency of approval procedures (Simulations 4). We increased the quality check for obtaining a
109
0
1
2
3
4
5
6
7
0
Figure 4.12a
Imitative products
1
2
3 4 Innovative products
Innovative and imitative products in each TC at last period
Standard
5
6
7
110
0
5
10
15
20
25
30
0
Figure 4.12b
Imitative products
2
4
6
8 Innovative products
Innovative and imitative products in each TC at last period
Sim 2
10
12
14
16
111
0
0.5
1.0
1.5
2.0
2.5
0
Figure 4.12c
Imitative products
1
2
4
Innovative products
3
Innovative and imitative products in each TC at last period
Sim 3
5
6
7
112
0
0.5
1.0
1.5
2.0
2.5
3.0
3.5
4.0
4.5
5.0
0
Figure 4.12d
Imitative products
0.5
1.0
1.5
2.0 Innovative products
Innovative and imitative products in each TC at last period
Sim 4
2.5
3.0
3.5
4.0
A history-friendly model of innovation
113
patent from 25 to 50. As expected, compared to the standard simulation, concentration increases and the total number of therapeutic areas discovered by firms decreases. Similarly, also the number of therapeutic classes in which firms are involved also decreases and the Performance Index is lower. Finally, when the single therapeutic classes are considered (Figure 4.12d), the differences across classes is higher.
5
CONCLUSIONS
In this chapter in honour of Paul David we merged two different types of analysis which lie at the core of his interests: a historical analysis regarding the long-term evolution of the pharmaceutical industry (and focusing on a wide range of factors such as the search process of firms, the role of institutions and regulation) and a modelling analysis, attempting to model the dynamics of market structure and innovation in pharmaceuticals in a history-friendly way. The history tried to examine three different periods: the first based on relatively primitive methods, the second based on ‘random screening’ as a method for finding new drugs and the third on ‘drug development by design’. The model focused on the random screening period. It has been able to replicate the long-term evolution of the industry in terms of concentration and the relationship between innovative and imitative products in the industry. In a companion paper (Malerba-Orsenigo, 2002) we have examined also the period of drug development by design. Some counterfactual exercises show that changes in the number of firms matter negatively for concentration and positively for the discovery of new products. On the contrary, the increase in the length of patent protection matters positively for the share of innovative products in the market, and much less for concentration. Finally, an increase in the stringency of approval procedures has positive effects on concentration and negative effects on the overall total quality of the products discovered by firms.
114
Path dependence in technical change
APPENDIX Parameter
Symbol
Mean of normal distribution of positive quality molecules Standard deviation of normal distribution of positive quality molecules FDA quality check threshold Mean of normal distribution of TCs value Standard deviation of normal distribution of TCs value Total number of firms (standard run) Total number of TCs Total number of molecules in each TC Initial budget for every firm Drawcost in search activities [see eq. (4.2)] Standard width of the patents Standard patent duration Exponent of product quality (PQ) (see Utility function [eq.(4.7)]) Exponent of inverse of rate of return (1/mup) (see Utility function [eq. (4.7)]) Exponent of launch marketing expenditures (A) (see Utility function [eq. (4.7)]) Exponent of yearly marketing expenditures (YA) (see Utility function [eq.(4.7)]) Desired rate of return for innovative products Desired rate of return for imitative products
Q
50
Q
20
FDA V V
Value
25 3000 500
f n M B Drawcost w pd a
30 50 150 3000 10 000 10 20 [1.2–1.4]
b
[1.0–1.2]
c
[0.1–0.2]
d
0.1
mupinn mupimi
0.2 0.1
NOTES 1. 2.
3.
This chapter is a revised version of the paper presented at the conference in honour of Paul David ‘New Frontiers in the Economics of Innovation and New Technologies’, Turin, 20 and 21 May, 2000. We thank Luca Berga, Christian Garavaglia, Marco Gazzola and Nicola Lacetera for their invaluable contribution to the development of the model. Wesley Cohen and Bronwyn Hall have provided useful and constructive comments. We thank the support of the Italian CNR and of the 40 percent programme of the Italian Ministry of University and Research (MIUR). Until the Waxman-Hutch Act was passed in the USA in 1984, generic versions of drugs that had gone off patent still had to undergo extensive human clinical trials before they could be sold in the US market, so it might be years before a generic version appeared even once a key patent had expired. In 1980, generics held only 2 per cent of the US drug market.
A history-friendly model of innovation
115
BIBLIOGRAPHY Arora, A. and Gambardella, A. (1992), ‘Complementarity and external linkage: the strategies of the large firms in biotechnology’, Journal of Industrial Economics, 37 (4), 361–79. Arora, A. and Gambardella, A. (1994), ‘The changing technology of technical change: general and abstract knowledge and the division of innovative labor’, Research Policy, 23, (5), 523–32. Arrow, K.J. (1962), ‘Economic welfare and the allocation of resources for invention’, in Nelson, R.R. (ed.), The Rate and Direction of Inventive Activity. Economic and Social Factors, Princeton, NJ, Princeton University Press, pp. 609–25. Bienz–Tadmor, B., Di Cerbo, P., Tadmor, G. and Lasagna, L. (1992), ‘Biopharmaceuticals and conventional drugs: clinical success rates’ Bio/Technology, 10, 521–5. Bottazzi, G., Dosi, G., Lippi, M., Pammolli, F. and Riccaboni, M. (2001), ‘Innovation and corporate growth in the evolution of the drug industry’, International Journal of Industrial Organization, 19 (7), 1161–87. Breschi, S., Malerba, F. and Orsenigo, L. (2000), ‘Technological regimes and Schumpeterian patterns of innovation’, Economic Journal, 110, 388–410. Burstall, M.L. (1985), The Community Pharmaceutical Industry, Luxembourg: Office for the Official Publications of the European Communities. Chandler, A.D. (1990), Scale and Scope: The Dynamics of Modern Capitalism, Cambridge, MA, Bellknap Press of Harvard University Press. Chandler, A.D. (1999), ‘Paths of learning: the evolution of high technology industries’, manuscript, forthcoming. Chien, R.I. (1979), Issues in Pharmaceutical Economics, Lexington, MA, Lexington Books. Comanor, W.S. (1986), ‘The political economy of the pharmaceutical industry’, Journal of Economic Literature, 24, 1178–217. David, P.A., Mowery, D.C. and Steinmueller, W.E. (1992), ‘Analyzing the economic payoffs from basic research’, Economic of Innovation and New Technology, 2 (4), 73–90. Ernst and Young (1990), Biotech 91: A Changing Environment, San Francisco: Ernst and Young. Galambos, L. and Lamoreaux, N.R. (1997), ‘Understanding innovation in the pharmaceutical industry’, paper presented at conference on ‘Understanding Innovation’, Baltimore, MD, June. Galambos, L. and Sewell J.E. (1996), Network of Innovators. Vaccine Development at Merck, Sharp & Dohme and Mulfor, 1895–1995, Cambridge, Cambridge University Press, Galambos, L. and Sturchio, J. (1996), ‘The pharmaceutical industry in the twentieth century: a reappraisal of the sources of innovation’, History and Technology, 13 (2), 83–100. Galambos, L. and Sturchio, J. (1998), ‘Pharmaceutical firms and the transition to biotechnology: a study in strategic innovation’, Business History Review, 72, 250–78. Gambardella, A. (1995), Science and Innovation in the US Pharmaceutical Industry, Cambridge, Cambridge University Press.
116
Path dependence in technical change
Gambardella, A., Orsenigo, L. and Pammolli, F. (2000), ‘Global competitiveness in pharmaceuticals: a European perspective’, report prepared for the Directorate General Enterprise of the European Commission, EPRIS Working Paper n. 3. Garavaglia, C., Lacetera, N., Malerba, F. and Orsenigo, L. (2002), ‘Project diversification in pharmaceuticals in the random screening paradigm: a simulation model’, working paper. Grabowski, H. And Vernon, J. (1983), The Regulation of Pharmaceuticals: Balancing the Benefits and the Risks, Washington, DC, AEI. Grabowski, H. and Vernon, J. (1994), ‘Innovation and structural change in pharmaceuticals and biotechnology’, Industrial and Corporate Change, 3, (2), 435–50. Henderson, R. (1994), ‘The evolution of integrative competence: innovation in cardiovascular drug discovery’ Industrial and Corporate Change, 3 (3), 607–30. Henderson, R. and Cockburn I. (1996), ‘Scale, scope and spillovers: the determinants of research productivity in drug discovery’, Rand Journal of Economics, 27 (1), 32–59. Henderson, R., Orsenigo, L. and Pisano, G.P. (1999), ‘The pharmaceutical industry and the revolution in molecular biology: exploring the interactions between scientific, institutional and organizational change’, in Mowery, D.C. and Nelson, R.R. (eds), The Sources of Industrial Leadership, Cambridge, Cambridge University Press. Klepper, S. (2002), ‘Firm survival and the evolution of oligopoly’, Rand Journal of Economics, 33 (1), 37–61. Klevorick, A., Levin, R., Nelson, R. and Winter, S., (1987), ‘Appropriating the returns from industrial research and development’, Brookings Papers on Economic Activity, 3, 783–820. Malerba, F. and Orsenigo, L. (2001), ‘Towards a history-friendly model of innovation, market structure and regulation in the dynamics of the pharmaceutical industry: the age of random screening’, CESPRI Working Paper. Malerba, F. and Orsenigo, L. (2002), ‘Innovation and market structure in the dynamics of the pharmaceutical industry and biotechnology: towards a historyfriendly model’, Industrial and Corporate Change, 11 (4), 667–704. Malerba, F., Nelson, R., Orsenigo, L. and Winter, S. (1999), ‘History-friendly models of industry evolution: the computer industry’, Industrial and Corporate Change, 8 (1), 3–40. Malerba, F., Nelson, R., Orsenigo, L. and Winter, S. (2001), ‘Competition and industrial policies in a “history-friendly” model of the evolution of the computer industry’, International Journal of Industrial Organization, 19, 635–64. Matraves, C., (1999), ‘Market structure, R&D, and advertising in the pharmaceutical industry’, Journal of Industrial Economics, 48 (2), 169–94. Maxwell, R.A. and Eckhardt, S.B. (1990), Drug Discovery: A Case Book and Analysis, Clifton, NJ, Humana Press. McKelvey, M. (1995), Evolutionary Innovation: The Business of Biotechnology, New York, Oxford University Press. Merges, R. and Nelson, R.R. (1994), ‘On limiting or encouraging rivalry in technical progress: the effect of patent scope decisions’, Journal of Economic Behavior and Organization, 25, 1–24. Nelson, R.R. (1959), ‘The simple economics of basic scientific research’, Journal of Political Economy, 67 (2), 297–306.
A history-friendly model of innovation
117
Nelson, R. and Winter, S. (1982), An Evolutionary Theory of Economic Change, Cambridge MA, Belknapp Press of Harvard University Press. Orsenigo, L. (1989), The Emergence of Biotechnology, London, Pinter. Orsenigo, L., Pammolli, F. and Riccaboni, M. (2001), ‘Technological change and network dynamics’, Research Policy. 30 (5), 485–508. Peltzman, S. (1974), Regulation of Pharmaceutical Innovation: The 1962 Amendments, Washington, DC, Washington American Enterprise Institute for Public Policy. Pharmaceutical Manufacturers Association (1991), 1989–1991 Statistical Report, Washington, DC: Pharmaceutical Manufacturers Association. Pisano, G. (1991), ‘The governance of innovation: vertical integration and collaborative arrangements in the biotechnology industry’, Research Policy, 20, 237–49. Pisano, G. (1996), The Development Factory: Unlocking the Potential of Process Innovation, Cambridge, MA, Harvard Business School Press. Scherer, F.M. (2000), ‘The pharmaceutical industry’, in Culyer, A.J. and Newhouse, J.P. (eds), Handbook of Health Economics, vol. 1, Amsterdam, Elsevier, pp. 1297–336. Schwartzman, D. (1976), Innovation in the Pharmaceutical Industry, Baltimore, MD, Johns Hopkins University Press. Sutton, J. (1998), Technology and Market Structure: Theory and History, Cambridge, MA, MIT Press. Thomas, L.G. (1994), ‘Implicit industrial policy: the triumph of Britain and the failure of France in global pharmaceuticals’, Industrial and Corporate Change, 3 (2), 451–89.
5. Path dependence and diversification in corporate technological histories John Cantwell1 1
INTRODUCTION
David (1985; 1991; 1993) has applied the notion of path dependence in the context of a continuing impact of individual micro decisions and actions undertaken at certain key historical junctures that represent ‘windows’ of structural change, on the subsequent emergence and evolution of a wider network of actors. Path dependence is hence defined as non-ergodic, in the sense that the particular historical events that occur at these stages of structural transformation, and the precise sequencing of those events, has a lasting effect on the asymptotic probability distribution of outcomes (see also Arthur, 1994). This formulation of path dependency confers greater precision on the phrase that ‘history matters’. This chapter applies David’s concept to the large firm viewed as a network of actors and activities, and in particular as a network of various types of technological effort that promote innovation and learning across the different parts of the firm. The chapter also follows David’s (1994) suggestion that human organisations and institutions act as the ‘carriers of history’, owing to the fact that within the context of the form and functioning of large firms, path dependence plays an especially crucial role. A similar argument is advanced by Cantwell and Fai (1999), supported by empirical historical evidence that the composition of the products or markets of the largest firms tends to have shifted more markedly over time than have their profiles of corporate technological specialisation. Thereby large firms represent repositories of competence or expertise (Winter, 1988), and so provide some historical continuity to the composition of productive and technological capabilities of a society, amidst the sometimes more radical changes that are observed through innovation at the product level. Thus, the argument can be extended to the path dependency observed among national groups of large firms when the profile of their capabilities is considered collectively, or in the enduring character of specific alternative national systems of innovation (Cantwell, 2000). 118
Path dependence and diversification
119
However, as David (2001) has contended, there is a productive tension within evolution (whether of a biological or a social and cultural kind) between the path dependency that is associated with the constraints on development imposed by the structure carried forward from past events (or the current contents of the gene pool), and the teleological principle of natural selection according to inclusive fitness, which may lead to departures from the existing structure in accordance with the newer requirements of the current environment. This argument suggests that there may be both strong and weak versions of path dependency, in terms of differences in the longevity of the effects that are engendered by past events. In the strong kind of path dependency positive feedback effects from some initial events dominate, providing a ‘lock-in’ phenomenon as in the QWERTY illustration (David, 1985), while in the weaker kind negative feedbacks (or new structural shocks) from the selection environment play a greater role in moving a system away from its starting point rather more easily or quickly. The implication is that there may be a variety of individual paths, and indeed with stochastic processes the strength of the observed path dependency is likely to include a chance element. In this chapter corporate technological trajectories are treated as being generally path-dependent, but with a continual drift. Cantwell and Fai (1999) had observed path dependency in the profiles of corporate technological specialisation of 30 large firms over the 60-year period from 1930 to 1990, in that these profiles tended to persist over time, even if they were subject to gradual or incremental change. However, that study also found some evolution and diversification in the patterns of corporate technological capabilities over time. This chapter looks in greater depth at the specific historical paths followed by four of these large firms in their corporate technological trajectories, and over the longer period 1890–1995. In doing so more can be said about the precise nature of the evolution that occurred, and about the character of changes in corporate technological diversification. What is more, the question can be addressed of whether, if profiles of corporate technological capabilities tend to persist over periods of 60 years (such that the specificities of a firm’s primary technological origins can still be identified as being present 60 years later), does the same hold for periods as long as 100 years or more? The four firms chosen were two of the world leaders in each of the chemical and electrical equipment industries respectively. The origins of most of the present leaders in the chemical industry can be traced back to the end of the nineteenth century and even further back in the case of the German companies (Beer, 1959). The leading German firms, which were the international pioneers of in-house corporate research and development (R&D), enjoyed great early success, none more so than Bayer (Haber, 1971). The
120
Path dependence in technical change
most prominent companies, including Bayer, merged their operations into IG Farben between 1925 and 1945, before reconstituting themselves individually again in the post-Second World War era (as Hoechst, BASF and Bayer). Their continuing presence in an international leader capacity over 100 years later constitutes a remarkable demonstration of the perseverence of technological prowess. Only sightly less remarkably, Du Pont had been the first USA company to open in-house research facilities in the chemical industry, and having risen to a leading world position in the inter-war period, also has a record of success that traces back over 100 years. Likewise, in the electrical equipment industry, the largest firms in the USA and Europe trace their origins back to the nineteenth century. In terms of the longevity and the significance of their technological contributions, perhaps the two best known of these are General Electric and American Telephone and Telegraph (AT&T), whose research histories have been well documented (Reich, 1985), but both of which continue to be world leaders to the present day. The focus of attention here is on the long-term paths of technological development of these companies, based on the proposition that history matters, in the sense that the technological characteristics of such large companies (during the period under consideration) were heavily influenced – and constrained – by the type of technological activities that they or their predecessors had carried out in the past. This notion of organisational continuity can be supported with reference to David’s (1994) explanation of the role of historical experience in forming mutually consistent expectations that facilitate co-ordination without the need to rely perpetually on centralised direction, and the role of the interrelatedness that tends to develop among the constituent elements of complex human organisations, as well as by the earlier concept of the central place of organisational routines as representing embedded experience in the course of evolutionary social learning (Nelson and Winter, 1982). In order to compare the evolutionary paths in the sectoral composition of the innovative activity of firms over time we require a quantitative measure of their technological activities. This chapter uses patents granted in the USA to Du Pont, IG Farben (and later Bayer), General Electric and AT&T as a measure of the extent and the spread of the technological achievements of these companies. We contend that patents may be used with relatively good confidence as a proxy measure of the rate and direction of the technological change of these companies, active as they all are in science-based industries. The chapter is divided into five sections. The next section introduces the data to be used in the analysis, discusses the suitability of using patent statistics as a measure of corporate technological activities and briefly reviews the methodology adopted. Section 3 looks at the evolution of technological
Path dependence and diversification
121
capabilities at Du Pont and IG Farben (and later Bayer) through the use of corporate patent data, and section 4 conducts an equivalent analysis of General Electric and AT&T. In the final section some summary cross-firm comparative measures of corporate technological diversification are presented and assessed, and conclusions are drawn with respect to the distinctive technological paths followed by each company.
2
THE DATA AND METHODOLOGY
The Companies Selected The research presented in this chapter, based on a study of four firms, is part of a wider project on long-term patterns of technological change (over a period of more than a century) of the largest US and European industrial companies. For the purposes of the wider project, two types of information have been collected manually from the US Index of Patents and the US Patent Office Gazette. First, all patents were recorded that were assigned to a selection of large US-owned and European-owned firms between 1890 and 1968. From 1969 onwards equivalent information has been computerised by the US Patent and Trademark Office (USPTO). The firms selected for the historical patent search were identified in one of three ways. The first group consisted of those firms which have accounted for the highest levels of US patenting after 1969; the second group comprised other US, German or British firms which were historically among the largest 200 industrial corporations in each of these countries (derived from lists in Chandler, 1990); and the third group was made up of other companies which featured prominently in the US patent records of earlier years (a method that proved most significant for a number of French firms that had not been identified from other sources). In each case, patents were counted as belonging to a common corporate group where they were assigned to affiliates of a parent company. Affiliate names were normally taken from individual company histories. In all, the US patenting of 857 companies or affiliates was traced historically; together these comprise 284 corporate groups. Owing to historical changes in ownership, 17 of the affiliates were allocated to more than one corporate group over the period as a whole. Where patents have been assigned to firms, the inventor is normally an employee of the company or is directly associated with it in some other way, but occasionally independent individual inventors do choose to assign their patents to firms (Schmookler, 1966). Assignments by independent individuals were more common in the nineteenth century but, at least from the inter-war years onwards, the
122
Path dependence in technical change
typical assignor was a prominent member of a corporate research laboratory, or some other similar in-house company facility. Although it is normally difficult to trace these named individuals in secondary sources on the firms concerned (as they are not usually also senior managers), the location of assignors can be checked against business history sources on the international location of activity in particular firms. Such checks on a selection of large firms have confirmed that whenever a location has been responsible for significant numbers of patents being assigned to a company, that firm did indeed have some in-house facility in the location in question at the relevant time. Companies checked in this fashion include various US firms active abroad and European companies in the USA (Stocking and Watkins, 1946; Beaton, 1957; Wilkins, 1974; 1989; Chandler, 1990) including IG Farben and its predecessors (Plumpe, 1990), Du Pont and ICI (Hounshell and Smith, 1988), Courtaulds and British Celanese (Coleman, 1969), and AT&T, General Electric and the British GEC (Jones and Marriot, 1971; Reich, 1985). The six firms granted the largest volume of US patents historically were, in descending order, General Electric, AT&T, Westinghouse Electric, IG Farben, RCA and Du Pont (for the years 1890–1947, the details for which are given in Tables 5.3 and 5.6 below). So as to be able to compare the longrun trends in technological specialisation of the two leading firms in each of two broadly defined industries – electrical equipment and chemicals – in this chapter GE, AT&T, IG and Du Pont are the companies selected for closer study. For the purposes of data continuity in the case of the earlier historical years, the founders of IG – namely, Bayer, BASF, Hoechst and Agfa – are treated together collectively prior to the formation of IG Farben in 1925. With the break-up of IG Farben after 1945, for the purposes of the post-war period attention is directed instead to the leading member of what had been the IG group, namely, Bayer. To construct a measure of technological specialisation of firms the first step is to devise a classification of fields of technological activity, which is derived from the USPTO patent class system. Fortunately, as these classes change, the USPTO reclassifies all earlier patents accordingly, so the classification is historically consistent. This study uses the classification scheme that was in operation at the end of 1995, which is then applied backwards in time. Every patent was classified by the USPTO under at least one such class and sub-class. Although patents can be assigned to more than one field, the primary classification was used in all cases. Various broad categories of technological activity were derived by allocating classes or subclasses to common groups of activity. Patents granted to the companies included in the study were classified in this manner to a total of 23 technological sectors for each industrial group, representing the principal areas of
Path dependence and diversification
123
development in each of these industries respectively. For the wider project, patents have been allocated to one of the 56 fields of technological activity set out in Table 5.1. However, not all these fields are important for the firms of a given industry, so for this study some of the less significant fields are grouped together in each industry, such that the sectoral composition across 23 areas is specifically designed to suit the analysis of patterns of specialisation in the chemical and electrical equipment industries separately. The particular disaggregation chosen for each industry is shown in Table 5.2. Patent Statistics as a Measure of Technological Activities Patent statistics present a potentially very rich source of empirical evidence on questions related to technological change (see Scherer et al., 1959; Schmookler, 1966; Pavitt, 1985; 1988; Griliches, 1990). The learning process which generates accumulated capability in companies relies on inputs of new knowledge and inventions, and so long as the pattern of knowledge requirements thus reflects the underlying distribution of technological competence across firms, corporate patents may be used as a proxy for the underlying pattern of technological change, and not merely as a direct measure of inventions. It is argued that US patent data provide the most useful basis for international comparisons, given the common screening procedures imposed by the US Patent Office (Pavitt and Soete, 1980; Soete, 1987; Pavitt, 1988). Additionally, as the USA is the world’s largest single market, it is likely that firms (especially large ones), will register for a patent there after patenting in their home countries. It is also reasonable to assume that such foreign patents registered in the USA are likely to be on average of higher quality or significance. United States patents reveal to which firm each patent was granted, and with which type of technological activity the patent is associated. Looking within the innovating firms themselves, the hypothesis here of path dependence and persistence in the profiles of corporate technological specialisation comes not so much from the characteristics of the knowledge generation process (R&D) itself, but from the structure of downstream learning and problem-solving in and around production, which calls for the creation of specialised knowledge inputs in specific fields. Thus, our use of patent statistics regards them as a measure of inputs (into innovation, the creation of new commerical products and processes) and not outputs (from R&D); that is, codified knowledge inputs into the processes of problemsolving and learning in production, through which technological competence is created. Of course, this does imply that there may still be potential problems with an input-based classification scheme derived from the patent class system, given the way in which technologies from different disciplinary
124
Table 5.1 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44
Path dependence in technical change
The classification of 56 fields of technological activity Food and tobacco products Distillation processes Inorganic chemicals Agricultural chemicals Chemical processes Photographic chemistry Cleaning agents and other compositions Disinfecting and preserving Synthetic resins and fibres Bleaching and dyeing Other organic compounds Pharmaceuticals and biotechnology Metallurgical processes Miscellaneous metal products Food, drink and tobacco equipment Chemical and allied equipment Metalworking equipment Paper-making apparatus Building material processing equipment Assembly and material handling equipment Agricultural equipment Other construction and excavating equipment Mining equipment Electrical lamp manufacturing Textile and clothing machinery Printing and publishing machinery Woodworking tools and machinery Other specialised machinery Other general industrial equipment Mechanical calculators and typewriters Power plants Nuclear reactors Telecommunications Other electrical communication systems Special radio systems Image and sound equipment Illumination devices Electrical devices and systems Other general electrical equipment Semiconductors Office equipment and data processing systems Internal combustion engines Motor vehicles Aircraft
125
Path dependence and diversification
Table 5.1 45 46 47 48 49 50 51 52 53 54 55 56
(continued) Ships and marine propulsion Railways and railway equipment Other transport equipment Textiles, clothing and leather Rubber and plastic products Non-metallic mineral products Coal and petroleum products Photographic equipment Other instruments and controls Wood products Explosive compositions and charges Other manufacturing and non-industrial
Table 5.2 The relationship of the 23 fields of technological activity used for each industry, to the original 56 sector classification Field 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23
Chemical industry 56 sector codes included
Electrical equipment industry 56 sector codes included
2 3 4 5 6 7 8 9 10 11 12 13, 14 16 15, 17–30 33–41 42–47 48 49 50, 54 51 52, 53 55 1, 31, 32, 56
2–12, 55 13 14 24 15–23, 25–30 33 34 35 36 37 38 39 40 41 42, 43 44–47 48 49 50, 53 51 52 53 1, 31, 32, 56
126
Path dependence in technical change
foundations may be integrated, and given some arbitrariness in the division between certain patent classes. We have tried to alleviate this difficulty by devising a classification scheme that groups together patent classes that are the most technologically related, as described above. So while Schmookler (1966) used patents as a direct measure of invention as such and others (Scherer, 1983; Bound et al., 1984) have since used them as an indirect (output) measure of R&D inputs, the patents granted to the largest industrial firms are used here instead as an indirect (input) measure of the pattern of technological change in these companies. In this sense patents represent knowledge inputs into the corporate learning processes that give rise to changes in production methods, the creation of which knowledge has generally been tailored to the problem-solving agenda of such learning in production. This is a valid inference so long as the knowledge requirements of the learning processes by which firms generate accumulated capabilities reflect the profile of those resultant technological competences across types of innovative activity. Just as the location of inventors that assigned patents to each firm has been checked against the known location of corporate research facilities as mentioned above, so too the sectoral distribution of corporate patenting has been checked against the more qualitative or archival evidence of business history sources on the equivalent firms. Again, we have found an approximate matching between the quantitative patterns of patenting and the qualitative accounts of the primary fields of R&D and productive expertise of the same firms (as described in Reich, 1985; Hounshell and Smith, 1988; Plumpe, 1995). In one respect what is done here is to provide a greater formalisation of propositions on the evolution of the composition of technological specialisation and the degree of technological diversification that can already be found descriptively in business history stories. Without fully reviewing the literature on the use of patent statistics, it may be worth mentioning two of the problems that have been raised. First, the fact that companies do not patent all their inventions and, therefore, any comparison may be biased in favour of those firms which rely more on patenting relative to secrecy; and, second, the fact that those inventions which do get patented differ in their economic and commercial significance. With regard to the first problem, there is evidence which suggests that differences in the propensity to patent are more significant when comparing firms from different industries (Scherer, 1983). We may assume that companies in the chemical industry (or alternatively in the electrical equipment industry) have a similar attitude towards patenting, allowing for a comparison of their technological capabilities based on their patenting activity. However, we have to consider that even within a company the propensity to patent varies between technological fields. Therefore, for the
Path dependence and diversification
127
comparative analysis here an indicator of relative specialisation is used rather than absolute numbers of patents, as explained below. The second question regarding the difference in quality of patented inventions may not necessarily be a significant obstacle when assessing the breadth of a company’s innovative capabilities. Since patents are not only issued on the most significant inventions but also on other related discoveries, they are useful in the analysis of general trends in the sectoral composition of technological activities (which may cover a wider spectrum than important invention alone). For instance, a major shift in the level of a company’s patenting in a particular sector, relative to the aggregate level, is likely to indicate a shift in the focus of its technological efforts. Consequently we may conclude that despite differences in the ‘quality’ of individual patents, comparisons between rival companies across the entire distribution of their respective patenting can shed light on the composition or spread of the technological expertise of firms. Other evidence confirms the suitability of patent data as a measure of corporate technological effort, particularly in inter-firm comparisons within an industry (Griliches, 1990). They are available, they go back over 100 years and allow for a technological classification at a greater level of detail than any other measure of technological activity (such as R&D statistics). The business histories of the firms studied here can also be cited to show how patenting mattered to them as part of their strategy, and with respect to the construction of technology exchange arrangements with other large firms in their respective industries (see Cantwell and Barrera, 1998). Reich (1977) discusses the importance of patents to the struggle to control radio that included both GE and AT&T. Plumpe (1990) shows that the level of IG Farben’s patent applications in Germany was consistent with its research and development (R&D) activities, which gives us an indication of the company’s reliance on patents. In addition, IG Farben’s negotiations with Standard Oil, which led to the establishment of a joint venture in the USA, together with the activities of the US subsidiary of IG, General Aniline and Film, confirm how these patents had to be extended to the US market. However, since we will be using data on patents granted to these companies in the USA, it should be allowed that Du Pont, being a US company, was more prone to patent in its home country than would have been IG Farben. This is another reason for using relative rather than absolute numbers. The Indicators of Corporate Technological Specialisation Derived from Patenting Arguments such as those just considered, and other issues that have been well documented, demonstrate the need for caution in the use of patent
128
Path dependence in technical change
statistics (Basberg, 1987; Pavitt, 1988; Griliches, 1990; Archibugi, 1992; Patel and Pavitt, 1997; 1998). However, a number of the difficulties in the use of patent data have been avoided in our approach through relevant disaggregation and the construction of an appropriate index. Inter-industry differences in patenting propensity are reduced as the chapter deals with intra-industry comparisons only, although admittedly the industrial groups are defined very broadly (but so, too, is the span of activity of the very large companies under consideration). It is recognised that intersectoral (across technological fields) or inter-firm differences in the propensity to patent may arise, but these are controlled for here by the use of the Revealed Technological Advantage (RTA) index (Cantwell, 1989; 1993; Cantwell and Andersen, 1996; Patel and Pavitt, 1997; 1998). The RTA is an indicator of a firm’s technological specialisation across a spectrum of technological activity relative to that of other firms in the same industry. The RTA of a firm in a particular field of technological activity is given by the firm’s share in that field of US patents granted to all companies in the same industrial group, relative to the firm’s overall share of all US patents assigned to all firms in the industry in question. If Pij denotes the number of US patents granted in a particular industry to firm i in technological activity j, the RTA index is defined as: RTAij (Pij iPij (jPij ijPij)) The index varies around unity, such that a value in excess of one shows that the firm is specialised in that field of activity in relation to other firms in its industrial group. In this manner inter-sectoral differences in the propensity to patent are normalised in the numerator of the RTA index, and inter-firm differences are normalised in the denominator. There still remains the possibility of intra-firm and intra-sectoral differences in the propensity to patent, but it is likely that the respective variances of these two factors are systematically lower than the inter-firm and inter-sectoral differences. The degree of technological diversification of the firm is measured by the inverse of the coefficient of variation of the RTA index, CVi, across all the relevant fields for the firm. Therefore, for firm i in each period considered, the proxy DIVi for technological diversification will be the reciprocal of the CVi. In particular:2 DIVi 1CVi RTA RTA i
i
where RTAi is the standard deviation and RTAi is the mean value of the RTA distribution for the firm i.
Path dependence and diversification
3
129
TECHNOLOGICAL DEVELOPMENT AT IG FARBEN AND DU PONT
A central proposition here is that, to a great extent, the research traditions of the companies which joined in the formation of IG Farben in 1925 regulated the subsequent technological development of the new company, and are likely to be reflected as well in the path of its descendants. The major German chemical companies such as Bayer, BASF and Hoeschst had their origins in the dyestuffs industry, one of the most dynamic chemical sectors at the turn of the century. These companies managed to bridge the gap between industrial and academic chemistry by incorporating and organising research within the dye workplaces. Consequently, they had accomplished great innovative and technical achievements prior to their amalgamation into IG Farben. Certainly in the first part of the twentieth century, the technological supremacy of these companies was still unchallenged (Beer, 1959). These German companies had traditionally dominated the world dyestuffs market, including in the USA. Furthermore they had established barriers to entry which proved very difficult to circumvent. These seemed to be the case when, with the outbreak of the First World War, the US market was cut off from German imports. Lacking a research tradition in this area, American companies found it almost impossible to replicate the broad range of dyes which the Germans had introduced. It seems that after the First World War, dyestuffs had nevertheless lost relative importance, and the German chemical companies expanded into other areas of activity, such as nitrogenous fertilisers, plastics, photographic products and synthetic materials (Haber, 1971). However, the 1926 figures for IG Farben’s research staff show that as much as 76 per cent of the labour employed at the company’s laboratories was still concentrated in dyestuffs and dyeing processes research; with about 9 per cent of the research staff engaged on the newer fields of pharmaceuticals, synthetic fibres and photographic chemicals (Plumpe, 1990). Other technologies gaining ground on dyestuffs included cleaning agents and other compositions (coatings and plastics in particular), and coal and petroleum research (the oil from coal hydrogenation process).3 Yet the fastest-growing technologies from the 1930s were those grouped under synthetic resins and fibres. By the 1930s the chemical industry seems to have built upon organic chemistry to move into synthetic materials, photographic chemistry and petrochemicals. By contrast, Du Pont’s origins had been in the explosives business. Its first research laboratory (which opened in 1902) had been established to deal with the problems inherent in the manufacturing process. However,
130
Path dependence in technical change
a threat to the company’s monopolistic position in the smokeless powder market seems to have prompted the need for diversification. This move was influenced by the company’s excess capacity following the expansion of production during the First World War, and for which alternative uses would have to be found in peacetime. With cash available after the war, and based on its experience with cellulose technology (used in the manufacturing of explosives), the company expanded into related areas mainly through the acquisition and further refinement of firms and technologies (Hounshell and Smith, 1988). Following this strategy, by the time of IG Farben’s merger, Du Pont had broadened its areas of activity. In addition to explosives, the company had diversified into paints (Duco being a major innovation), silk, leather and cellulose. As with explosives, these products were based on nitrocellulose technology. The company had bought its way into the production of cellophane and developed moisture-proof cellophane, a successful new product. In addition, Du Pont had finally accomplished the production of dyestuffs following a major struggle to circumvent the barriers established by the German companies (which involved the government’s establishment of a higher import tariff in the years following the war). Research and development expenditures in 1926 give us an idea about the extent of this diversification – roughly 15 per cent of research spending was into explosives, 25 per cent in general chemicals, 27 per cent in paints and related chemicals, 19 per cent in dyestuffs and 6 per cent in plastics (Hounshell and Smith, 1988). Comparing the evolution of the German companies with that of Du Pont prior to 1926, it appears that their technological strategies had been different (Dornseifer, 1989; 1995). Whereas the German firms had relied more on internal generation of innovation (through their strong inhouse research), Du Pont had relied more on external sources, that is, acquisitions. Lacking a research tradition in areas outside its core technology, Du Pont took advantage of the strengths of its organisation to incorporate those acquired technologies and develop them further (Chandler, 1990). By contrast, IG Farben’s predecessors had continued to rely on the internal dynamism of the research organisations that they had built. Table 5.3 presents evidence on US patents granted to the largest chemical firms. In absolute terms (see last row) US patenting activity in the chemical sector experienced sustained growth throughout 1890 to 1947, with the partial exception of the period 1920–24, when there was a fall attributable to the effects of the First World War. The position of the German dyestuffs companies (here amalgamated under IG Farben) is even more remarkable than the high shares tell us, given that, as mentioned above, US companies had a higher propensity to patent their innovations in the USA. Indeed, Du Pont lagged behind in second place until the late
131
56.22 n/a n/a n/a 7.17 1.26 7.53 4.76 1.82 0.40 n/a 5.80 n/a n/a 96.75
1890–1919
11.37 n/a n/a n/a 15.19 5.03 15.68 13.81 1.54 2.11 n/a 5.04 n/a n/a 82.05
1920–24
28.86 n/a n/a n/a 8.99 2.37 7.58 6.69 1.86 3.08 n/a 5.20 n/a n/a 77.60
1925–29
30.75 n/a n/a n/a 11.61 5.12 3.66 6.87 4.36 6.37 n/a 3.38 n/a n/a 81.84
1930–34
18.74 n/a n/a n/a 19.13 3.47 4.73 4.34 4.07 8.10 n/a 3.33 n/a n/a 78.12
1935–39
5.58 n/a n/a n/a 17.46 4.94 4.16 2.35 2.61 3.89 n/a 2.89 n/a n/a 66.78
1940–47
19.81 n/a n/a n/a 15.27 4.09 5.05 4.44 3.13 4.93 n/a 3.56 n/a n/a 76.26
Average 1890–1947
Percentage share of US patenting of large chemical and petrochemical companies
Chemical firms IG Farben Hoechst Bayer BASF Du Pont Dow Chemical Union Carbide Allied Chemical ICI British Celanese Celanese Corp Swiss IG Ciba Geigy Sandoz Total chemicals %
Table 5.3
n/a 8.47 8.39 6.02 7.95 7.10 1.41 0.08 2.66 n/a 0.34 n/a 5.06 0.72 80.74
1991–95
2.64 4.01 5.57 2.79 9.17 6.04 3.59 1.50 2.87 0.87 0.86 n/a 4.25 0.88 74.19
Average 1890–1995
132
4 518
Total number
1 231
4.14 n/a 0.24 17.95
1920–24
2 692
5.91 n/a 0.63 22.40
1925–29
8 079
5.68 n/a 1.47 18.16
1930–34
11 398
4.90 n/a 4.00 21.88
1935–39
17 906
9.45 n/a 5.38 33.22
1940–47
45 824
6.45 n/a 3.42 23.74
Average 1890–1947
37 033
n/a 3.32 3.62 19.26
1991–95
343 703
n/a 5.20 3.52 25.82
Average 1890–1995
Source:
US Index of Patents, Patent Gazette, and computerised data from the US Patent and Trademark Office.
Note: Prior to 1926 IG Farben’s and ICI’s figures correspond to the aggregate of their predecessor companies; a similar procedure applies to Swiss IG. n/a not applicable.
0.77 n/a 0.13 3.25
1890–1919
(continued)
Standard Oil NJ Exxon Shell Total oil %
Oil firms
Table 5.3
Path dependence and diversification
133
1930s. This overwhelmingly commanding position of IG Farben (in what follows this name also applies to its predecessors prior to 1925) was particularly apparent in the period prior to the First World War. However, the figures for the 1920–24 and 1940–47 periods may be somewhat underestimated, since after the two wars the Alien Property Custodian confiscated thousands of German patents, which had not yet been assigned to the relevant companies. Nevertheless, note that, although IG Farben was dissolved in 1945, the relevant period allows for the inclusion of two additional years (to 1947) during which the confiscated American subsidiaries of IG Farben were still patenting. The Evolution of Technological Capabilities in IG Farben and Bayer Table 5.4 presents the cross-field RTA index over time for the IG group and latterly for Bayer. As explained earlier, the RTA index shows the firm’s share of patenting in a given technological field (among patents granted to all large firms in the chemical industry), relative to its equivalent share of total chemical industry patenting in all fields for each period considered. It can be seen that, in spite of the fact that IG Farben’s research remained heavily concentrated in organic chemistry (as that of its predecessor companies had been) as shown by an RTA value consistently above 1, changes began to occur to its profile of technological specialisation in other parts of the cross-sectoral distribution. There was even some decline in related bleaching and dyeing processes after the merger that constituted the IG group in 1925. Looking back to the turn of the twentieth century, it can be seen that Table 5.4 provides evidence that is entirely consistent with what is already known about the early development of the chemical industry, and of the large German firms in the chemical industry in particular. That their histories were essentially coincident at that stage reflects the dominant position in the industry of the German leaders. The industry began with the development of artificial dyestuffs in the 1870s, from which it moved into photographic chemicals and synthetic fibres in the early years of the twentieth century. The predecessors of IG Farben led the way, as is shown by the RTA values above 1 in 1890–1919 in Table 5.4. In the 1920s and 1930s the IG group diversified further into fertilisers (listed here under agricultural chemicals), and pharmaceuticals. Following the formation of IG in 1925 synthetic resins and fibres steadily gained in importance in the group’s profile of technological effort, the RTA rising from 0.45 in 1925–29 to 1.14 in 1940–47. The latter may partly reflect the company’s research into synthetic rubber which led to the development of PVC (Freeman, 1982). In addition, intensification of research in photochemicals (from 1.92 to 2.65
134
Distillation processes Inorganic chemicals Agricultural chemicals Chemical processes Photographic chemistry Cleaning agents and other compositions Disinfecting and preserving Synthetic resins and fibres Bleaching and dyeing proc. Other organic compounds (including dyestuffs) Pharmaceuticals Metals Chemical and allied equip. Mechanical engineering nes. Electrical equipment nes. Transport equipment
Technological sectors
0.80 0.84 2.40 0.42 7.21 0.14 0.00 0.34 4.38 1.96 1.55 0.72 0.18 0.29 0.00 0.00
0.00 1.58 1.53 1.33 0.92 0.49 0.38 0.13 0.07 0.34
1920– 24
0.34 0.58 0.47 0.50 1.45 0.64
1890– 1919
1.14 0.57 0.64 0.52 0.07 0.34
0.00 0.45 1.44 1.30
1.01 0.96 0.90 0.88 1.92 0.59
1925– 29
0.93 0.60 0.53 0.36 0.33 0.53
0.00 0.96 0.82 1.33
0.36 1.16 1.30 0.63 2.40 0.81
1930– 34
IG Farben
1.28 0.46 0.44 0.25 0.61 0.78
0.00 0.90 0.94 1.35
0.60 0.73 0.99 0.68 3.00 0.63
1935– 39
0.97 0.36 0.39 0.20 0.93 0.00
1.71 1.14 1.27 1.33
0.58 0.62 0.22 0.61 2.65 0.52
1940– 47
Table 5.4 Evolution of patterns of technological specialisation at IG Farben, 1890 to 1995
0.96 0.45 0.47 0.24 0.28 0.39
0.17 0.70 1.16 1.48
0.42 0.97 0.92 0.59 1.95 0.56
Cumul. 1890– 1947
0.86 0.40 1.11 0.38 0.37 0.37
0.33 1.41 1.01 1.23
1.33 0.81 1.96 0.56 2.48 0.57
1991– 95
1.31 0.48 0.90 0.53 0.38 0.51
0.41 1.41 0.88 0.91
0.74 0.48 1.53 0.61 3.26 0.51
Cumul. 1947– 95
Bayer
135
Source:
as for Table 5.3.
Textiles, clothing, leather Rubber and plastic products Non-metallic mineral and wood products Coal and petroleum products Professional and scientific instruments Explosive compositions and charges Other manufacturing
0.00 0.42 0.39 0.00 0.48 0.00 0.00
0.00 0.11 0.26 1.17 0.06 0.07 0.10
0.00
0.00
1.08 1.56
0.00 0.29 0.09
0.17
0.00
0.95 1.62
0.00 0.49 0.43
0.17
0.06
1.58 2.62
0.00 0.44 0.50
0.89
0.00
2.72 1.88
0.00 0.31 0.31
0.21
0.05
1.30 1.46
0.00 0.29 0.32
0.26
0.00
0.38 0.98
0.00 0.76 0.58
0.20
0.01
0.29 2.39
0.42 0.78 0.56
136
Path dependence in technical change
over the same period) seems to have been associated with developments in related product areas, namely an involvement in photographic equipment (here included under the other scientific instruments category, the RTA for which was well above 1 throughout 1925–47). In this case technological diversification was linked to product diversification into the photographic sector, as represented by the Agfa branch of the group, this and other parts of the IG group having developed the complementary technological expertise in photographic chemistry. One noteworthy diversification feature of the interwar period is that coal and petroleum research reached an all-time high in the period following the merger (particularly from the late 1930s). This confirms other evidence about IG Farben’s renewed interests in coal liquefaction, which had been initiated at one of the company’s predecessors (Beer, 1959). While this may have originally been commercially inspired in the 1920s and early 1930s when other chemical firms followed suit in investing in the apparently promising field of oil from coal, most other large firms that had experimented with this possibility divested from the area during the 1930s as the early promise was disappointed. This helps to explain how IG’s RTA in the field rose from 1.08 in 1925–29 to 2.72 in 1940–47. The reason seems fairly clear – by this stage IG Farben had become obliged to assist in the Nazi war effort, and as is well known the principal German military weakness was that it lacked its own oil supply. Perhaps the most remarkable part of the IG story comes from updating the account to consider the subsequent profile of technological specialisation of its largest successor company, Bayer. In this context we focus on a comparison of the pattern of technological specialisation of the IG Farben group as a whole over 1890–1947 with that of Bayer over 1947–95, but with an eye on the position of Bayer in the most recent sub-period, 1991–95, to see whether there are any substantive shifts in the pattern of specialisation that appear to have emerged in the latest years. A rationale for comparing two very long-term periods of 1890–1947 and 1947–95 is itself the belief that the knowledge base of firms is cumulative and incremental (Nelson and Winter, 1982; Rosenberg, 1982; Cantwell, 1989). What comes out of this comparison is indeed a picture of the most remarkable continuity. The oldest strength in dyestuffs has held up and, indeed, even reasserted itself most recently (an RTA of 1.23 in 1991–95), while the new developments that had emerged early in the twentieth century, and had been either reinforced or established as a strength in IG Farben in the inter-war period, remain evident in Bayer today. Thus, consider the RTA values in 1947–95 in synthetic resins and fibres (1.41), photographic chemistry (3.26) and instruments (2.39); pharmaceuticals also stands at 1.31, but this advantage has been eroded recently (0.86 in
Path dependence and diversification
137
1991–95). In contrast, the inter-war specialisation in oil-related chemicals has disappeared (an RTA in the Bayer years of 0.29), driven as it was more by external military demands than by the specificities of internal capabilities and commercial logic. In contrast, what had been for IG in the 1930s the newly emergent strength in fertilisers and agricultural chemicals has successfully reappeared in Bayer (an RTA of 1.53 in 1947–95, and 1.96 in 1991–95). The Evolution of Technological Capabilities at Du Pont Looking at the evolution of Du Pont’s technological specialisation (Table 5.5), through to 1947 the company preserved its founding strength in the field of explosives. However, given its incredibly high degree of focus on explosives (an RTA of 11.16 in 1890–1919), it was almost inevitable that growth would take the form of to some extent moving away from its original research activities in and around explosives and allied technologies (chemical and distillation processes, cellulose, silk and leather with textile applications – with RTAs in 1890–1919 of 1.76, 5.40 and 13.49 respectively), initially into rubber and plastics (2.70 in 1890–1919), and then into synthetic resins and fibres (3.05 in 1925–29). The emergence and subsequent rise of the development of synthetic materials in the 1920s and 1930s seems to be for Du Pont the most striking phenomenon of the inter-war period. This confirms other evidence regarding the company’s research into polymers which eventually led to the discovery of nylon, the first synthetic fibre and Du Pont’s most successful product.4 By contrast with IG Farben, organic chemistry was never an area of comparative strength at Du Pont, and it emerged to respectability in this field (an RTA of 0.99 in 1935–39) only after the First World War. Since this was not a traditional area of research for the company, its moderate catching up partly reflects the company’s co-operative agreement with the ICI (after 1929), a central objective of which was to enable these partners to better match the industry leader, IG Farben (see Cantwell and Barrera, 1998). A sustained position in rubber and plastic technologies through the inter-war years reflects in part the rise in Du Pont’s research in polymers leading to the discovery of neoprene (Hounshell and Smith, 1988). An interesting feature hidden by aggregation in Table 5.5 is the growth of the development of textile and clothing machinery (within mechanical engineering), which may have spun off from the interest in synthetic fibres, and particularly nylon. What is most striking from a comparison with IG Farben is that while Du Pont began with a much more focused spectrum of technological specialisation early in the twentieth century than that of IG, reflecting its later
138
Distillation processes Inorganic chemicals Agricultural chemicals Chemical processes Photographic chemistry Cleaning agents and other compositions Disinfecting and preserving Synthetic resins and fibres Bleaching and dyeing proc. Other organic compounds (including dyestuffs) Pharmaceuticals Metals Chemical and allied equip. Mechanical engineering nes. Electrical equipment nes. Transport equipment
Technological sectors
0.00 0.77 0.39 0.63 0.77 0.72 1.98 1.15 0.05 3.60
0.00 0.96 3.16 3.04 0.14 6.75
1.80 0.63 0.00 1.25 0.00 1.51
1920– 24
0.00 0.00 0.00 0.18
5.40 0.49 0.00 1.76 0.00 1.98
1890– 1919
1.32 0.92 1.40 0.38 0.08 2.16
0.00 3.05 0.32 0.85
3.24 0.71 0.00 0.81 2.47 2.48
1925– 29
1.43 0.81 0.87 0.75 0.19 0.00
2.82 2.29 0.75 0.82
1.08 0.85 1.03 0.76 0.52 1.71
1930– 34
0.73 0.68 0.99 0.59 0.45 0.51
0.00 1.39 0.73 0.99
0.87 1.16 0.49 1.26 0.50 1.52
1935– 39
0.72 0.74 0.82 0.61 0.64 0.38
1.64 1.53 0.85 0.87
0.82 1.16 1.04 1.20 1.06 1.33
1940– 47
Table 5.5 Evolution of patterns of technological specialisation at Du Pont, 1890 to 1995
0.85 0.78 1.03 0.73 0.29 1.00
1.09 1.81 0.73 0.80
1.08 0.87 0.78 1.20 0.97 1.61
Cumul. 1890– 47
0.27 1.10 0.86 1.57 2.56 0.78
1.05 1.06 0.49 0.70
2.80 0.91 0.57 1.26 1.82 0.90
1991– 95
0.29 1.03 1.06 1.23 1.05 1.69
0.62 1.19 0.82 0.78
1.48 1.35 1.18 1.15 1.50 1.21
Cumul. 1947– 95
139
Source:
as for Table 5.3.
Textiles, clothing, leather Rubber and plastic products Non-metallic mineral and wood products Coal and petroleum products Professional and scientific instruments Explosive compositions and charges Other manufacturing
3.60 0.95 1.75 0.95 0.36 4.02 3.47
13.49 2.70 1.04 0.54 1.35 11.16 8.34
3.84
5.34
0.43 0.20
0.00 1.23 1.39
1.43
4.50
0.69 0.24
0.00 0.97 2.14
1.29
1.87
0.97 0.25
1.48 0.97 1.39
0.88
1.78
0.75 0.32
0.38 1.47 1.64
1.69
3.14
0.78 0.33
1.15 1.43 1.77
0.95
0.56
1.74 1.00
6.25 2.79 1.88
1.09
2.70
1.56 0.71
1.26 1.70 1.68
140
Path dependence in technical change
start and smaller size, it had become notably more diversified than IG by the 1940s. This again can be related to the strategy of acquisitions rather than internal growth. Over 1890–1947 as a whole, Du Pont held a technological specialisation among others in cleaning agents and compositions (an RTA of 1.61), chemical processes (1.20), synthetic resins and fibres (1.81), disinfecting and preserving (1.09), textiles and leather (1.15), nonmetallic mineral products (1.77), rubber and plastic products (1.43), distillation processes (1.08) and, of course, in explosives technologies (3.14). Within chemical processes, the most important technologies were coating processes, adhesive bonding and chemistry, and electrical and wave energy. Cleaning agents and other compositions included paints and lacquers, with ‘Duco’ (the trademark product) being its most important innovation in this field (Hounshell and Smith, 1988). However, despite a path that emphasised technological diversification at Du Pont compared to greater consolidation among the IG group, Du Pont’s corporate technological trajectory from 1900 to 1947 was in many respects even more coherent in its direction than that followed by IG. The early moves away from explosives mainly represented the development of nitro-cellulose technologies – cellulose (used in the manufacturing process of explosives), silk and leather, chemical processes (including coating and bonding) and paints. From here came the development of cellophane. Then, having worked with rayon and cellulose acetate textile fibres, the company was well placed to diversify into (and to lead) the synthetic fibre revolution (Hounshell and Smith, 1988). Much of this structure was then preserved in the post-war period. In 1947–95 (and in 1991–95) Du Pont retained an RTA greater than one in distillation processes, chemical processes, synthetic resins and fibres, textiles, clothing and leather, rubber and plastic products. However, although this remained true also of explosives in 1947–95 as a whole (an RTA of 2.70), it is striking that by 1991–95 this original primary source of strength had finally lapsed (the RTA value standing at 0.56). Conversely, to a greater extent than in the post-war experience of Bayer, new strengths have emerged in Du Pont in 1947–95 in mechanical processes (an RTA of 1.23 in 1947–95, and 1.57 in 1991–95), electrical equipment (1.05 in 1947–95 and 2.56 in 1991–95) and coal and petroleum products (1.56 in 1947–95 and 1.74 in 1991–95). The latter emergent strength in oil-related chemicals may be somewhat ironic given Bayer’s post-war retreat from that field, but in Du Pont it can be traced back to the development of polymer intermediates that derived from the long tradition in explosives (Hounshell, 1995). So perhaps some residue of Du Pont’s path-dependent history stemming from its beginnings in explosives remains through to the present day after all.
Path dependence and diversification
4
141
TECHNOLOGICAL DEVELOPMENT AT GE AND AT&T
As has been seen already in the case of chemicals, the science-based industries that began towards the end of the nineteenth century were characterised by growth through horizontal diversification into technologically related fields. However, vertical diversification into related mechanical fields mattered too, as in the case of IG’s move into photographic equipment, and Du Pont’s move into textile machinery. Nevertheless, horizontal science-related diversification was the essential theme of corporate technological trajectories in the chemical industry. In comparison, the development of vertically integrated systems was relatively more important in the electrical equipment industry, broadly defined. The electrical equipment industry focused from the outset on the design of complex and interrelated technological systems, of which certain components might lie either ahead (salients) or behind (reverse salients) the general front of development at any point in time, whereby progress in other parts of the overall system is either facilitated or constrained (Hughes, 1983; 1989; Aitken, 1985). While in the chemical industry the largest German companies were the world leaders, in the electrical industry that role was taken up by the largest US firms. The firms examined here were associated with the founding of the two central planks of the industry – namely, electrical lighting, power, traction and related machinery (in the person of Edison, whose role is discussed by David, 1991), and the telephone (in the person of Bell). General Electric was formed in 1892 from the merger of the Edison General Electric Company and the Thomson-Houston Electric Company, while AT&T was originally a subsidiary of American Bell set up in 1885, which with financial reorganisation in 1899 became the holding company for the entire Bell group (Reich, 1985). The original overlap between these two branches of the industry lay in electrical devices and systems, and in some general machinery. This overlap illustrates how the leading firms in this industry were concerned to establish integrated systems, and not the (perhaps interconnected range of) more narrowly defined products that were typical in the chemical industry. The connection between the two segments of the industry became much sharper from the inter-war years onwards with the development of the radio, and subsequently the television. The radio was the primary focus of growth in the electrical equipment industry in the inter-war period, and both parts of the industry made critical contributions to this new area of development. Table 5.6 shows the comparative patenting records of the leading firms in the electrical equipment industry. It shows how AT&T (Bell) was the leading research organisation and the dominant corporate patenter in the
142
as for Table 5.3.
13 057
Total Number
Source:
49.71 20.17 23.72 1.66 0.05 2.43 0.45
1890– 1919
4 883
21.52 34.65 37.64 1.72 0.43 1.52 0.43
1920– 24
9 524
25.21 34.26 29.85 2.54 0.78 2.45 1.57
1925– 29
12 003
28.53 25.96 21.44 8.86 2.07 4.12 2.87
1930– 34
12 005
27.21 20.32 15.24 19.37 1.39 3.62 6.38
1935– 39
18 503
26.97 19.81 19.09 19.50 5.95 1.39 2.36
1940– 47
Percentage share of US patenting of large electrical equipment companies
General Electric AT&T Westinghouse Elect. RCA ITT Siemens AEG
Table 5.6
70 025
30.90 24.02 22.45 10.78 2.31 2.59 2.54
Average 1890– 1947
19 184
23.58 13.58 8.00 0.30 2.61 14.40 0.26
1991– 95
241 069
25.26 16.94 15.20 9.71 5.11 7.26 1.27
Average 1890– 95
Path dependence and diversification
143
1920s (which made it the highest ranked patenting firm in any industry), but that at other times – before the First World War and once again from the 1930s onwards – this role fell to GE. The early 1920s saw a substantial growth and diversification of research at AT&T, which gave rise to the establishment of Bell Telephone Laboratories in 1925, the formation of which gave a further impulse to R&D in the company (Maclaurin, 1949; Nobel, 1979; Reich, 1985). Westinghouse Electric was not far behind the big two, while RCA was set up by GE in 1919 to take over the assets it had acquired from American Marconi in the nascent radio industry. By 1921 Westinghouse Electric and AT&T had also entered into a partnership with RCA in the radio industry, which therefore, of course, was also a partnership with GE in the development of radio systems (Reich, 1977). During the post-war period while GE has remained the industry leader the others have gradually declined (and in the case of AT&T have been broken up), and large firms from other countries have caught up. The German firm, Siemens, is a prime example, but there are also the newer Japanese companies that have not been considered here. The Evolution of Technological Capabilities at GE Given the complex systems nature especially of that part of the electrical business in which GE was active, and given what is known about the breadth of interest of GE in particular from the beginning across the range of electric lighting, power and transport components, it is hardly surprising that the evidence on the spread of GE’s technological specialisation set out in Table 5.7 shows that it was far more diversified in its range of expertise in 1890–1919 than were our other large companies. Although this high initial span of corporate technological diversification is, indeed, partly attributable to the systems nature especially of this segment of the electrical equipment industry from the start, it may also have something to do with relatively wide inventive interests of Edison, and his comparative advantage as an inventor in the design and construction of electromechanical systems (Reich, 1985). This is akin to David’s own (1991) assessment of the significance of the individual personality of Edison as an influence on historical paths, in this case helping to establish GE as an innovator across a broad front, which has remained as a feature of GE’s capabilities (compared to its major competitors) through to the present day. This having been said, GE embarked on a further technological diversification in the final decade of the 1890–1919 period (Reich, 1985), and this is encompassed in the aggregation of the years 1890–1919 into a combined period for the purposes of comparison with other companies (for which change was concentrated in the inter-war years). The earliest strengths in
144
1.09 2.14 1.43 1.69 1.25 0.16 0.46 1.60 0.07 1.55 1.37 1.21 0.00 0.20
1.14 1.21 0.89 1.61 1.41 0.04 0.33 0.00 0.19 1.30 1.01 1.14 0.00 0.13
Chemicals and pharmaceuticals Metallurgical processes Miscellaneous metal prod. Electrical lamp manuf. equip. Other mach. and ind. equip. Telecommunications Other electrical communication systems Special radio systems Image and sound equipment Illumination devices Electrical devices and systems Other general elect. equip. Semiconductors Office equipment, computers, and other data processing
1920– 24
1890– 1919
Technological sectors
0.60 0.37 1.62 1.33 1.68 0.79 0.75
0.93 1.15 1.03 1.27 0.89 0.13 0.84
1925– 29
0.24 0.43 1.51 1.26 1.52 1.28 0.61
1.04 1.12 0.98 1.42 0.93 0.11 1.11
1930– 34
0.28 0.11 1.37 1.08 1.56 0.64 0.18
1.19 1.39 1.24 1.75 1.41 0.11 0.80
1935– 39
0.20 0.16 1.02 1.20 1.45 0.48 0.18
2.09 1.21 1.12 1.54 1.15 0.17 0.62
1940– 47
0.24 0.23 1.28 1.15 1.44 0.61 0.30
1.40 1.21 1.03 1.46 1.22 0.11 0.64
Cumul. 1890– 1947
Table 5.7 Evolution of patterns of technological specialisation at General Electric, 1890 to 1995
0.64 0.28 0.90 0.59 0.88 0.33 0.47
2.48 1.89 0.97 0.45 1.60 0.17 0.54
1991– 95
0.31 0.30 1.12 1.06 1.43 0.46 0.44
1.82 1.35 1.36 1.06 1.26 0.16 0.56
Cumul. 1947– 95
145
Source:
as for Table 5.3.
Motor vehicles and engines Other transport equipment Textiles, clothing and leather Rubber and plastic products Building materials Coal and petroleum products Photographic equipment Other instruments and controls Other manuf. and non-ind.
1.73 1.10 2.01 1.01 1.16 2.01 1.44 1.25 1.37
1.55 0.77 0.00 0.62 1.39 0.00 0.00 1.28 0.91
1.26 0.97 3.97 0.91 1.72 3.97 0.57 1.04 0.79
1.11 1.07 0.00 0.50 0.95 2.04 0.98 1.13 1.20
1.38 1.33 1.05 1.02 1.79 1.47 0.46 1.14 1.58
1.69 1.05 1.48 0.55 1.58 1.01 0.56 1.10 0.85
1.92 1.34 1.03 0.76 1.35 1.66 0.72 1.16 1.13
0.75 2.30 0.00 2.07 2.27 2.83 0.20 0.88 2.08
1.50 1.44 0.96 1.22 1.55 1.91 0.41 0.90 1.33
146
Path dependence in technical change
electric lighting, power and traction are here reflected in RTA values greater than 1 for 1890–1919 in illumination devices, electrical lamp manufacturing equipment, general electric equipment, other machinery and industrial equipment, vehicle (components) and other transport equipment. Other areas of GE technological advantage are best related to developments that came in the early part of the twentieth century. An RTA greater than unity in metallurgy can be related to the development of ductile tungsten filaments for incandescent electric lighting, while that in chemicals may relate to advances in heat insulation and refrigeration (Reich, 1985). What is most striking about GE’s subsequent corporate technological trajectory is its high degree of path dependency and persistence, even across the very wide front of capabilities over which it operated. Almost all the primary fields of advantage in 1890–1919 were also areas of advantage for GE in 1940–47 (including illumination devices, electric lamp manufacture, chemicals and pharmaceuticals, metallurgical processes, other machinery and industrial equipment, other general electrical equipment, motor vehicles and engines, other transport equipment, building materials and coal and petroleum products). Perhaps even more remarkably, GE continued to enjoy an RTA greater than 1 in all these fields in 1947–95! However, in its original core field of lighting it had declined by 1991–95. Thus, by this last period, GE’s RTA in illumination devices had fallen to 0.90, while in electric lamp manufacturing equipment it had dropped as low as 0.45. No doubt lighting is not as central to new development in the electrical equipment industry now as it once was, but given the overall forcefulness of GE’s technological path dependency over a wide range of activities, this retreat from its historical origins is noteworthy, even though it has taken the best part of a century to reach that turning point away from the past. The Evolution of Technological Capabilities at AT&T AT&T (Bell Telephone as it was) began with a much sharper focus in its technological specialisation than that at GE, concentrating its efforts on the telephone, and on related technologies. This is readily apparent from Table 5.8, and from a comparison of Table 5.8 with Table 5.7. In 1890–1919 the company was primarily focused on its origins in telecommunications (an RTA value of 4.50), together with the closely allied fields of other electrical communication systems (3.68) and image and sound equipment (3.12). Secondarily, it had established a related base in metal product technologies (an RTA of 1.22), no doubt given the need to work on the detailed development of telephone receivers, transmitters, cables and the like. Between them, these four fields of technological endeavour were the only ones out of the 23 areas of activity under consideration in which AT&T held RTA values
147
0.61 0.48 0.70 1.57 0.53 2.53 2.21 0.70 2.48 0.59 0.64 0.15 2.89 0.75
0.34 0.41 1.22 0.28 0.37 4.50 3.68 0.35 3.12 0.42 0.73 0.15 0.00 0.31
Chemicals and pharmaceuticals Metallurgical processes Miscellaneous metal prod. Electrical lamp manuf. equip. Other mach. and ind. equip. Telecommunications Other electrical communication systems Special radio systems Image and sound equipment Illumination devices Electrical devices and systems Other general elect. equip. Semiconductors Office equipment, computers, and other data processing
1920– 24
1890– 1919
1.02 1.77 0.56 0.70 0.24 0.00 1.40
0.85 1.08 0.69 0.47 0.75 2.32 1.28
1925– 29
0.93 1.43 0.35 0.71 0.35 0.39 1.07
0.87 1.54 1.50 0.18 1.04 2.26 1.09
1930– 34
0.47 1.50 0.39 0.91 0.34 0.62 1.12
0.87 1.38 1.82 0.82 1.02 1.95 1.83
1935– 39
0.48 1.05 0.58 0.73 0.40 0.51 0.98
0.73 1.80 1.22 0.36 1.40 2.15 2.04
1940– 47
Evolution of patterns of technological specialisation at AT&T, 1890 to 1995
Technological sectors
Table 5.8
0.58 1.59 0.44 0.74 0.26 0.45 0.95
0.71 1.27 1.21 0.47 0.90 2.65 1.94
Cumul. 1890– 1947
0.20 0.86 0.08 0.99 0.95 1.19 0.99
0.82 0.86 0.30 0.16 0.42 2.77 1.12
1991– 95
0.63 1.00 0.39 0.92 0.45 1.04 1.13
0.72 1.20 0.85 0.37 0.95 2.77 1.48
Cumul. 1947– 95
148
as for Table 5.3.
0.18 0.12 0.00 0.17 0.51 0.00 0.71 0.32 0.65
Motor vehicles and engines Other transport equipment Textiles, clothing and leather Rubber and plastic products Building materials Coal and petroleum products Photographic equipment Other instruments and controls Other manuf. and non-ind.
Source:
1890– 1919
(continued)
Technological sectors
Table 5.8
0.43 0.15 2.89 0.38 0.51 0.00 0.00 0.61 1.09
1920– 24 0.53 0.07 0.00 1.33 0.43 0.00 1.67 0.58 1.34
1925– 29 0.09 0.32 0.64 2.11 0.89 0.32 1.23 0.87 1.00
1930– 34 0.12 0.15 1.41 1.09 0.99 0.00 0.82 0.77 1.27
1935– 39 0.00 0.29 0.00 1.93 1.13 0.00 0.00 0.69 1.14
1940– 47 0.21 0.14 0.76 1.41 0.80 0.12 0.83 0.65 1.10
Cumul. 1890– 1947
0.00 0.00 0.00 0.69 0.55 0.00 0.00 1.18 0.26
1991– 95
0.17 0.20 0.80 1.19 0.73 0.28 0.63 0.86 0.43
Cumul. 1947– 95
Path dependence and diversification
149
greater than 1 in 1890–1919. In contrast, GE had an RTA above unity in no less than 16 out of 23 fields over the same period (see Table 5.7). From the 1920s onwards, and especially after the formation of Bell Laboratories in 1925, AT&T was engaged in a substantial technological diversification largely in support of (rather than away from) its continuing core interests in telecommunications (Reich, 1985). This theme is again well illustrated in Table 5.8. In 1925–29, AT&T had added to the fields in which it held an RTA value above 1 (among others) the three areas of metallurgical processes (1.08), rubber and plastic products (1.33) and office equipment and other data processing (1.40). These three new fields of strength are likely to have been related to respectively the firm’s development in the 1920s of metallic contacts for telephone switching apparatus and the properties of magnetic materials, enamel and phenol-fibre insulation, and telephone and telegraph transmission parameters (Reich, 1985). In 1930–34 an RTA above one was further attained in other machinery and industrial equipment (then 1.04), representing again a recognition of the role of vertical systems integration in the direction of diversification in the electrical equipment industry. By 1940–47 all these new fields just referred to continued to hold an RTA greater than unity in AT&T – metallurgical processes at 1.80, machinery and industrial equipment at 1.40 and rubber and plastic products at 1.93 – but with the partial exception of office equipment and data processing at 0.98. Needless to say, the established advantage was retained through to 1940–47 in telecommunications (2.15), other electrical communication systems (2.04), image and sound equipment (1.05) and metal products (1.22). In other words, over a combined period of well over 50 years AT&T preserved its core technological competence, but effectively built around it. As has been commented upon already, AT&T was also involved as a major contributor to the development of the inter-war radio industry. However, while it had moved substantively into special radio systems (an RTA of 1.02 in 1925–29) and various instrument technologies (0.87 in 1930–34), these never became its relative strength compared with its major rivals (and most notably compared with RCA, of course). Even in image and sound equipment there was a significant relative decline in the 1940s (from 1.50 in 1935–39 to 1.05 in 1940–47), following the advent of commercial television in the USA and the new focus of experimentation that this provided (Abramson, 1995). In essence, AT&T’s technological profile persisted also into the post-war period. In the years 1947–95 considered together, AT&T retained its indicator of positive corporate technological specialisation in telecommunications (2.77), other electrical communication systems (1.48), image and sound equipment (1.00), as well as in metallurgical processes (1.20) and
150
Path dependence in technical change
rubber and plastic products (1.19), but had lost its position in metal products (0.85). Key strengths persist all the way through to the latest period of 1991–95 in its core fields of telecommunications (2.77) and other electrical communication systems (1.12). However, by 1991–95 the company had witnessed some decline of its relative capabilities in image and sound equipment (0.86) and metallurgical processes (0.86), and a clear fall in metal products (0.30) and rubber and plastic products (0.69). Replacing these areas, AT&T now has advantages in instruments (1.18) and semiconductors (1.19), for which the antecedents had been laid in the inter-war years, but on which a relative focus of development attention emerged only recently. Yet despite some shift in the overall composition of its technological trajectory, AT&T’s continuing strong concentration on its telecommunications origins (2.77) is evident.
5
DIVERSIFICATION REVISITED AND SOME CONCLUSIONS
The evidence seems to confirm that all the large firms considered here followed specific and path-dependent corporate technological trajectories, in that the distinctive characteristics of their early years exercised an influence on the composition and breadth of their subsequent technological capabilities, and the direction in which they evolved. Despite the fact that these large companies broadened their areas of research, until at least the postwar period they remained specialised in those fields which had been their original stronghold. However, Du Pont and GE seem to have experienced a more radical departure from their original core technology than did IG Farben (and later Bayer) and AT&T. Nevertheless, even the extension of Bayer’s and AT&T’s research activities also led to a gradual diversification into some other related areas of strength. It is noticeable that the intensification of research in certain key areas appears to have spun-off allied innovations in other fields. This underlines the importance of interrelatedness of technology, whereby a major technological breakthrough tends to generate further innovations in connected fields, a feature of particular significance for firms in the science-based industries. It also helps to show the usefulness of patent statistics as a means of tracing the historical paths followed by large firms when considered across their entire distribution, as important patents are unlikely to be isolated while unimportant ones may be. While the patterns described here are entirely consistent with the qualitative evidence of business histories, they add some precision by facilitating clearer comparisons between the positions and paths of firms operating in similar industries.
151
Path dependence and diversification
Having commented on the extent of corporate technological diversification and how it evolved historically in each of our individual company descriptions of paths of specialisation, it is time to summarise these trends through an examination of our more formal indicator of diversification, DIV, the reciprocal of the coefficient of each company’s RTA distribution across fields (as explained above). The values of this indicator are set out in Table 5.9. The first row of the table, concerning 1890–1919, affirms what has already become clear from the discussion of Tables 5.5 to 5.8. That is, corporate technological diversification was much more pronounced from the outset in general electrical systems (as represented by GE, with a DIV value of 1.64 in 1890–1919) than it was in chemicals (1.04 in what became the IG group combined, and a still more concentrated 0.71 in Du Pont), but telecommunications was more like the latter than the former (with a DIV as low as 0.61 in AT&T for the equivalent period). By the end of the inter-war period both Du Pont and AT&T had largely caught up with the span of diversified technological development at GE. Thus, in 1935–39 Du Pont’s DIV value had risen as high as 2.00, and that of AT&T to 1.71, as against a value of 2.01 in the case of GE. IG Farben Table 5.9 Evolution of corporate technological diversification, 1890 to 1995 Period
IG Farben
Bayer
Du Pont
GE
AT&T
1890–1919 1920–24 1925–29 1930–34 1935–39 1940–59 1960–64 1965–68 1969–72 1973–77 1978–82 1983–86 1987–90 1991–95
1.04 0.57 1.23 1.27 1.10 n/a n/a n/a n/a n/a n/a n/a n/a n/a
n/a n/a n/a n/a n/a 0.79 0.83 0.91 0.91 0.95 1.34 1.34 1.50 1.31
0.71 1.06 1.00 1.14 2.00 2.68 1.90 2.66 2.63 1.98 1.58 1.50 1.46 1.17
1.64 1.35 1.32 2.03 2.01 1.87 1.42 1.68 1.87 1.41 1.31 1.36 1.20 1.28
0.66 1.07 1.33 1.59 1.71 1.65 1.37 1.09 1.16 1.15 1.19 1.32 0.89 0.98
1890–95
0.89
1.15
2.56
2.04
1.52
Note: n/a not applicable. Source: As for Table 5.3.
152
Path dependence in technical change
was the firm out of step in this respect, in that while it had been gradually diversifying through to 1930–34, at which stage it remained more diversified than Du Pont (a DIV value of 1.27 as against 1.14 for Du Pont), in the later 1930s what appeared to be its natural commercial path became distorted by the need to be attuned to the military objectives of the new German government. The fall in IG’s DIV value in 1935–39 (to 1.10) was particularly associated with its further move into oil-related chemicals, exactly at the when Du Pont was diversifying heavily into a range of new chemical processes following (and associated with) its successful transition into synthetic resins and fibres. Bringing the story closer to the present day, there tends now to be less difference between large firms in the scope of their technological diversification, compared to the inter-company variety of diversification that was often observed in the past. By 1991–95 the DIV measure seems to have converged among firms to a range of around 1.0 through to 1.3. Compared to the long-term historical average value of DIV (for 1890–1995 as a whole) this represents a rise for Bayer (to 1.31 in 1991–95, as opposed to 1.15 in its own post-war history, or even 0.89 for the IG group in earlier times), a fall for AT&T (from 1.52 as a long run average through to 1.32 in 1983–86, and then 0.98 in 1991–95, although the sharpness of this recent structural shift reflects the break-up of AT&T and the greater focusing of its remaining business), but a significant decline for Du Pont (1.17 in 1991–95 v. 2.56 in 1890–1995) and GE (1.28 in 1991–95 v. 2.04 in 1890–1995). Looking across companies and over time the general trend that is observed is of a steady initial increase in technological diversification historically, followed by a renewed concentration in more recent times. The three of our four companies with the greatest continuity of historical identity (Du Pont, GE and AT&T) describe this pattern most clearly, in that the DIV values with which they began in 1890–1919, and those with which they finished in 1991–95, were both well below their respective long-term averages for DIV in 1890–1995 as a whole. If one allows for the specificities of the experience of IG Farben in its later years (its contribution to the German war effort), and for the smaller size of Bayer which was only part of the original group and the recovery of the German chemical industry in the early post-war period, one can also see an historical trend towards diversification, from a DIV value of 1.04 in 1890–1919 through to 1.27 in 1930–34 and then to 1.34 in Bayer in 1978–82. Although Bayer’s DIV value did not decline after 1978–82, it has seen no further sustained increase since that time (standing at 1.31 in 1991–95). While in the early post-war years there was some continuation of the inter-war diversification trend (into, for example, photographic chemistry) in both Du Pont and Bayer, from around 1970 Du Pont has refocused its
Path dependence and diversification
153
technological efforts (with a fall in DIV from 2.63 in 1969–72 to 1.17 in 1991–95). In comparison the reversal of the diversifying trend began earlier in the post-war period in the electrical equipment industry, with some moderate refocusing of technological efforts in GE until around 1970 (the DIV indicator fell from 2.01 in 1935–39 to 1.87 in 1969–72), and in AT&T until the mid-1980s (a drop from 1.71 in 1935–39 to 1.32 in 1983–86, having been as low as 1.16 in 1969–72). Since then there has been a clearer refocusing upon a more closely related set of technological activities in GE from around 1970 (from 1.87 in 1969–72 to 1.28 in 1991–95) and in AT&T following its break-up (from 1.32 in 1983–86 to 0.98 in 1991–95). Some explanation can be offered for the apparent switch in the longterm direction of corporate technological diversification in at least this group of the largest firms (in terms of their patent volume), away from pro-diversifying change and towards an increasing focus of effort. In the first phase of the growth of large industrial companies, from around the end of the nineteenth century through to the Second World War, product diversification and technological diversification were much more closely connected to one another, through attempts to realise the joint economies of scale and scope (as documented in depth by Chandler, 1990). In the second phase of such growth since 1945, and especially since around 1970, corporate technological diversification has acquired a new motive apart from the simple support of product or market diversification. That is, in more recent times the primary motive has become the potential rewards from rising technological interrelatedness between formerly largely separate and discrete branches of innovative activity. These are new and more dynamic economies of scope, associated with continuous knowledge spillovers between allied fields of learning, and with the creation of new and more complex technological combinations. While for many smaller companies this shift of motives has meant a new impulse towards greater technological diversification than in the past to incorporate what have become the most closely related areas to their own core business, in some giant firms the greater potential inner benefits of interrelatedness has meant, instead, a refocusing of efforts around that combination of their established areas which have become most closely related (Cantwell and Santangelo, 2000). Thus, the drivers of corporate technological diversification have shifted from the coverage of related products and markets in the first phase of large company growth, to the relatedness to be found in innovative activity itself, and in the construction of new technological combinations in the second phase in the growth of large firms. So technological diversification associated with a steady movement outwards into new markets and technologically related products has been gradually replaced by often more focused combinations of technological activity
154
Path dependence in technical change
to capture the fruits of interrelatedness in the competence creation process itself. Meanwhile, path dependency has prevailed throughout the last hundred years or so of the corporate technological trajectories of these large firms, but it is a path dependency accompanied with the gradual drift that is associated with most stochastic processes. Bayer has continued with IG Farben’s traditions in dyestuffs, photographic chemistry and fibres to the present day, but Du Pont has eventually moved away from its historic origins in explosives. AT&T continues to hold its primary position in electrical communication systems, but GE has eventually moved away from illumination devices and lamp manufacture. Taken together these large company technological histories show how corporate technological trajectories have the typical property of path dependency with some continual drift, and not the strong kind of ‘lock-in’ configuration of the QWERTY kind.
NOTES 1.
2.
3. 4.
The author is grateful for the support of the UK Economic and Social Research Council, who funded the project on long-term technological change in the largest US and European firms on which this chapter draws, and to Pilar Barrera, who worked with him on that project. He is also grateful for the help of Jane Myers and Jim Hirabayashi at the US Patent and Trademark Office. The CV measure has often been used as well in the analysis of business concentration across firms within an industry, as opposed to concentration or dispersion across sectors within a firm (see Hart and Prais, 1956). It is worth noticing that alternative measures could be used (for example, the Herfindhal index) but that for a given number of firms or sectors (N), there is a strict relationship between the Herfindahl index (H) and the coefficient of variation (CV) (Hart, 1971). The relationship is H (CV2 1)/N. There is evidence that during the interwar period there was a close overlap between oil and petrochemical research in the chemical industry (Freeman, 1982). ‘Nylon became far and away the biggest money-maker in the history of the Du Pont Company’ (Hounshell and Smith, 1988: 273).
REFERENCES Abramson, A. (1995), Zworykin: Pioneer of Television, Urbana, IL: University of Illinois Press. Aitken, H.G.J. (1985), The Continuous Wave: Technology and American Radio, 1900–1932, Princeton, NJ, Princeton University Press. Archibugi, D. (1992), ‘Patenting as an indicator of technological innovation: a review’, Science and Public Policy, 19, 357–68. Arthur, W.B. (1994), Increasing Returns and Path Dependence in the Economy, Ann Arbor, MI, University of Michigan Press. Basberg, B.L. (1987), ‘Patents and the measurement of technological change: a survey of the literature’, Research Policy, 16, 131–41.
Path dependence and diversification
155
Beaton, K. (1957), Enterprise in Oil: A History of Shell in the United States, New York: Appleton-Century-Crofts. Beer, J.J. (1959), The Emergence of the German Dye Industry, Urbana, IL: University of Illinois Press. Bound, J., Cummins, C., Griliches, Z., Hall, B.H. and Jaffe, A.B. (1984), ‘Who does R & D and who patents?’, in Z. Griliches (ed.), R & D, Patents and Productivity, Chicago, IL, University of Chicago Press. Cantwell, J.A. (1989), Technological Innovation and Multinational Corporations, Oxford, Basil Blackwell. Cantwell, J.A. (1993), ‘Corporate technological specialisation in international industries’, in M.C. Casson and J. Creedy (eds), Industrial Concentration and Economic Inequality: Essays in Honour of Peter Hart, Aldershot, Edward Elgar. Cantwell, J.A. (2000), ‘Technological lock-in of large firms since the inter-war period’, European Review of Economic History, 4, 147–74. Cantwell, J.A. and Andersen, H.B. (1996), ‘A statistical analysis of corporate technological leadership historically’, Economics of Innovation and New Technology, 4, 211–34. Cantwell, J.A. and Barrera, M.P. (1998), ‘The localisation of corporate technological trajectories in the interwar cartels: cooperative learning versus an exchange of knowledge’, Economics of Innovation and New Technology, 6, 257–90. Cantwell, J.A. and Fai, F.M. (1999), ‘Firms as the source of innovation and growth: the evolution of technological competence’, Journal of Evolutionary Economics, 9, 331–66. Cantwell, J.A. and Santangelo, G.D. (2000), ‘Capitalism, profits and innovation in the new techno-economic paradigm’, Journal of Evolutionary Economics, 10, 131–57. Chandler, A.D. (1990), Scale and Scope: The Dynamics of Industrial Capitalism, Cambridge, MA, Harvard University Press. Coleman, D.C. (1969), Courtaulds: An Economic and Social History (Volume II: Rayon), Oxford and New York: Oxford University Press. David, P.A. (1985), ‘Clio and the economics of QWERTY’, American Economic Review, 75, 332–7. David, P.A. (1991), ‘The hero and the herd in technological history: reflections on Thomas Edison and the battle of the systems’, in P. Higonnet, D.S. Landes and H. Rosovsky (eds), Favorites of Fortune: Technology, Growth and Economic Development Since the Industrial Revolution, Cambridge, MA, Harvard University Press. David, P.A. (1993), ‘Historical economics in the long run: some implications of path dependence’, in G.D. Snooks (ed.), Historical Analysis in Economics, London and New York, Routledge. David, P.A. (1994), ‘Why are institutions the “carriers of history”? Path dependence and the evolution of conventions, organisations and institutions’, Structural Change and Economic Dynamics, 4, 205–20. David, P.A. (2001), ‘Path dependence, its critics and the quest for “historical economics” ’, in P. Garrouste and S. Ioannides (eds), Evolution and Path Dependence in Economic Ideas: Past and Present, Cheltenham, Edward Elgar. Dornseifer, B. (1989), ‘Research, innovation and corporate structure: Du Pont and IG Farben in comparative perspective’, Harvard University Graduate School of Business Administration Working Paper, Spring.
156
Path dependence in technical change
Dornseifer, B. (1995), ‘Strategy, technological capability and innovation: German enterprises in comparative perspective’, in F. Caron, P. Erker and W. Fischer (eds), Innovations in the European Economy Between the Wars, Berlin: de Gruyter. Freeman, C. (1982), The Economics of Industrial Innovation, London, Frances Pinter. Griliches, Z. (1990), ‘Patent statistics as economic indicators: a survey’, Journal of Economic Literature, 28, 1661–707. Haber, L.F. (1971), The Chemical Industry: 1900–1930, Oxford, Oxford University Press. Hart, P.E. (1971), ‘Entropy and other measures of concentration’, Journal of the Royal Statistical Society, Series A, 134, 73–85. Hart, P.E. and Prais, S.J. (1956), ‘The analysis of business concentration: a statistical approach’, Journal of the Royal Statistical Society, Series A, 119, 150–91. Hounshell, D.A. (1995), ‘Strategies of growth and innovation in the decentralized Du Pont company 1920–1950’, in F. Caron, P. Erker and W. Fischer (eds), Innovations in the European Economy Between the Wars, Berlin, de Gruyter. Hounshell, D.A. and Smith, J.K. (1988), Science and Corporate Strategy: Du Pont R&D, 1902–1980, Cambridge, Cambridge University Press. Hughes, T.P. (1983), Networks of Power: Electrification in Western Society, 1880–1930, Baltimore MD, Johns Hopkins University Press. Hughes, T.P. (1989), American Genesis: A Century of Invention and Technological Enthusiasm, New York, Viking. Jones, R. and Marriot, O. (1971), Anatomy of a Merger: A History of GEC, AEI and English Electric, London: Cape. Maclaurin, W.R. (1949), Invention and Innovation in the Radio Industry, London and New York, Macmillan. Nelson, R.R. and Winter, S.G. (1982), An Evolutionary Theory of Economic Change, Cambridge, MA, Harvard University Press. Nobel, D.F. (1979), America by Design: Science, Technology and the Rise of Corporate Capitalism, New York, Alfred A. Knopf. Patel, P. and Pavitt, K.L.R. (1997), ‘The technological competencies of the world’s largest firms: complex and path-dependent, but not much variety’, Research Policy, 26, 141–56. Patel, P. and Pavitt, K.L.R. (1998), ‘The wide (and increasing) spread of technological competencies in the world’s largest firms: a challenge to conventional wisdom’, in A.D. Chandler, P. Hagström and Ö. Sölvell (eds), The Dynamic Firm: The Role of Technology, Strategy, Organisation, and Regions, Oxford and New York, Oxford University Press. Pavitt, K.L.R. (1985), ‘Patent statistics as indicators of innovative activities: possibilities and problems’, Scientometrics, 7 (1–2), 77–99. Pavitt, K.L.R. (1988), ‘Uses and abuses of patent statistics’, in A. van Raan (ed.), Handbook of Quantitative Studies of Science Policy, Amsterdam, North Holland. Pavitt, K.L.R. and Soete, L.L.G. (1980), ‘Innovative activities and export shares: some comparisons between industries and countries’, in K.L.R. Pavitt (ed.), Technical Innovation and British Performance, London: Macmillan, 38–66. Plumpe, G. (1990), Die IG Farbenindustrie AG. Wirtschaft, Technik und Politik 1904–1945, Berlin, Duncker and Humblot. Plumpe, G. (1995), ‘Innovation and the structure of the IG Farben’, in F. Caron, P. Erker and W. Fischer (eds), Innovations in the European Economy Between the Wars, Berlin, de Gruyter.
Path dependence and diversification
157
Reich, L.S. (1977), ‘Research, patents and the struggle to control radio: a study of big business and the uses of industrial research’, Business History Review, 51, 208–35. Reich, L.S. (1985), The Making of American Industrial Research: Science and Business at GE and Bell, 1876–1926, Cambridge and New York, Cambridge University Press. Rosenberg, N. (1982), Inside the Black Box: Technology and Economics, Cambridge and New York, Cambridge University Press. Scherer, F. (1983), ‘The propensity to patent’, International Journal of Industrial Organisation, 1, 107–28. Scherer, F.M., Herzstein, S.E., Dreyfoos, A.W., Whitney, W.G., Bachmann, O.J., Pesek, C.P., Scott, C.J., Kelly, T.G. and Galvin, J.J. (1959), Patents and the Corporation: A Report on Industrial Technology Under Changing Public Policy, Boston, MA, J.J. Galvin (privately published). Schmookler, J. (1966), Invention and Economic Growth, Cambridge, MA, Harvard University Press. Soete, L.L.G. (1987), ‘The impact of technological innovation on international trade patterns: the evidence reconsidered’, Research Policy, 16, 101–30. Stocking, G.W and Watkins M.W. (1946), Cartels in Action, New York: Twenty Century Fund. Wilkins, M. (1974), The Maturing of Multinational Enterprise: American Business Abroad from 1914 to 1970, Cambridge, MA: Harvard University Press. Wilkins, M. (1989), The History of Foreign Investment in the United States to 1914, Cambridge, MA: Harvard University Press. Winter, S.G. (1988), ‘On Coase, competence and the corporation’, Journal of Law, Economics and Organization, 4, 163–80.
6. Is the world flat or round? Mapping changes in the taste for art G.M. Peter Swann1 1
INTRODUCTION Public taste, I believe, as far as it is the encourager and supporter of art, has been the same in all ages; a fitful and vacillating current of vague impression, perpetually liable to change, subject to epidemic desires, and agitated by infectious passion, the slave of fashion, and the fool of fancy. (Ruskin 1843 [1996], vol. 3: 617–18) The main lesson imparted by the test of time is the fickleness of taste whose meanderings defy prediction. (Baumol 1986: 14)
This chapter is a preliminary attempt to map the changing tastes for works of art as manifested in the prices of paintings sold at auction. There are two main goals in this work: first, to describe a space in which we can represent the work of different artists; and second, to describe how ‘cultivated taste’ moves around that space. In pursuing the first goal, moreover, we have to confront a further quandary: what is the appropriate shape of the space suitable for this purpose? Is the appropriate world ‘round’ or ‘flat’? This is an empirical question, which can be assessed by reference to measures of ‘goodness of fit’. But there is also an interesting theoretical interpretation. We shall argue that the answer depends on whether one takes a historicist or a modernist perspective on the development of art. When we turn to the second goal, the quote from Ruskin illustrates why our task will be a difficult one: it is not easy to map a ‘fitful and vacillating current’. Even more, we shall argue that movement in taste around the space of painters is a path-dependent process. That provides the link between this chapter and the work of Paul David. One of the many areas in which his work has been very influential is in the economics of path dependence – see for example David (1985; 1986; 1987; 1988; 1992; 1993; 1994; 1997), David and Foray (1994), David et al. (1998). Antonelli (1997) introduces a special issue of the International Journal of Industrial Organisation on path dependence, inspired by Paul David’s work. 158
Mapping changes in the taste for art
159
In a very loose sense, path dependence is the idea that ‘history matters’ when we try to understand how markets evolve today. But Paul David has provided us with much more precise definitions of path dependence, and we shall turn to this later (section 4). The path dependence we observe in this case is a little different from that observed in much of his work. For here, as art historians and theorists have recognised (see section 3), there is both positive and negative feedback: positive feedback from the elite to the aspirants but negative feedback from aspirants to the elite. The rest of the chapter is in eight sections. Section 2 briefly describes the economic models of taste change, which guide what follows – though these are discussed at greater length in the companion chapter by Cowan in this book (Chapter 7). Section 3 looks at a few central themes in art theory and history that illuminate our study. Section 4 provides a more precise definition of path dependence. Section 5 compares two leading techniques for mapping products in a common space: the characteristics approach and multidimensional scaling (MDS). Section 6 then examines (in the light of section 3) whether an appropriate space for mapping should be flat or round: precisely, should we work with a (two-dimensional) plane or the surface of a (three-dimensional) sphere? Section 7 states precise results for mapping from price correlations to points on a plane or on the surface of a sphere. Section 8 presents some preliminary calculations, which illustrate in a rough way how tastes have changed over a 100-year period. Section 9 concludes.
2
SOCIAL THEORIES AND ECONOMIC MODELS OF TASTE
This chapter aims only to map changes in taste, and not to explain why they come about. The companion chapter by Robin Cowan (Chapter 7) offers a model of changing tastes towards works of art. However, to help motivate the chapter, here is a brief sketch of one possible process leading to changes in taste. For a long time, we have known that the character of demand for prestige goods, including works of art, is rather different from the elementary neoclassical picture of demand. Veblen (1899) described how the newly wealthy indulged in conspicuous consumption – that is, the visible consumption of things that other people do not have. So long as these items of conspicuous consumption are not owned by others from whom one wishes to distinguish oneself (Bourdieu, 1984), then they serve their purpose well. But when others who aspire to share the consumption activities of the elite start to catch up, then it is time for the elite to move on to other forms of conspicuous consumption.
160
Path dependence in technical change
Cowan et al. (1997; 2004) have developed a model of demand driven by the conflicting desires of association, distinction and aspiration. The consumer may seek to associate in consumption with some groups, may seek distinction in consumption from some other groups, and all the while is aspiring to share some consumption activities with an elite. In this model, demand exhibits waves: to begin with, a particular good may be in demand from the elite, but then demand shifts downmarket, and the elite desert the good. Under some circumstances, these waves can repeat themselves. This model generates rich and complex patterns in the consumption of particular goods. The companion chapter by Cowan (Chapter 7) shows how models of this sort may be applied to understanding changing tastes towards works of art. In this last study, cycles in relative popularity are a natural outcome when the products are arrayed around a circle. Although some other social scientists believe that economics has relatively little to say about such historical patterns in demand, the opposite is true. There is indeed a large economics literature on association, distinction and aspiration in demand. We do not attempt to review that literature here, but have done so elsewhere (Swann, 1999; 2002). See also McPherson (1987) for a very useful overview. Some of the most important recent contributions are by Becker (1996), Bianchi (1998), Dosi et al. (1999) and Frank (1985). Moreover, following an influential paper by Baumol (1986), an interesting and substantial literature has grown up recently on the economics of the arts and culture, including Cozzi (1998), essays in De Marchi and Goodwin (1999), Frey (1997; 2000), Frey and Pommerehne (1989), essays in Ginsburgh and Menger (1996), Grampp (1989), McCain (1981) and Throsby (2001). Nor does this chapter try to describe how painters have reacted to changes in taste. Elsewhere, we have studied how the nature of consumer demand for distinction goods might influence the product design strategies of producers (Swann, 2001a). It is certainly true that some of the painters in our sample painted primarily for the market at the time, and as a result some of their works do not appeal today. By contrast, Van Gogh only sold one painting during his lifetime, but his popularity has grown monotonically since his death.
3
SOME THEMES IN ART HISTORY AND THEORY
Ruskin’s remark, quoted at the start of the chapter, highlights the volatility of taste. In view of that, it would not make much sense to assume that taste can be approximated by a constant! But Ruskin was talking about taste
Mapping changes in the taste for art
161
over all forms of art. Individual episodes are less chaotic but show curious cycles of popularity. As the art dealer and historian, Maurice Rheims (1959 [1961]) shows us, the fluctuating fortunes of art in different times can be quite entertaining. But he notes that it can also be cutting, when he says (Rheims, 1959 [1961]: 133): ‘Fashion is a sorcerer’s charm or talisman changing the masterpieces of today into the laughing-stock of tomorrow.’ Why do these changes in taste come about? It is unlikely that there could ever be one explanation that accounts for all cases, but some examples are suggestive. Take the history of the Pre-Raphaelite group of painters. Bell (1984) discusses the fall in popularity of the Pre-Raphaelites in the 1920s. Part of their error was to abuse their position as the establishment, and fail to understand innovations elsewhere (Bell, 1984: 14). In turn, the ‘Bloomsbury Group’ (notably Bell’s father, Clive Bell, 1927) were influential in turning the tide against the Pre-Raphaelites. As Barnes (1998: 113) puts it: ‘The Bloomsbury set laughed at the Pre-Raphaelites. To them it seemed that “everything of importance in the second half of the 19th century had happened in France” .’ In addition, Bell (1984: 16) recognised that part of the reason for their fall was that the wrong patrons had bought their work: ‘From the very first these painters found their market among those whom contemporaries would have considered an ignorant and philistine clientele, the “self-made” men and manufacturers of the North.’ In that respect, Bell takes a view very similar to that in the economic models of distinction and aspiration described in the last section. He is even closer when he goes on to describe how reproductions of Pre-Raphaelite paintings became commonplace on middle class, aspirational, walls (Bell, 1984: 16): We all want to exhibit a cultivated taste, we all want to be enlisted in the cultural elite and of course in so doing we deprive the elite of its elitist character; that which had been distinguished becomes in the truest sense vulgar and the public is ready for something else; it is thus I would suggest that the wheel of fashion is made to revolve.
While the Pre-Raphaelites became unpopular from the 1920s, and this can be seen in prices paid at auction (see below), the wheel of fashion came round again in the latter part of the twentieth century. Barnes (1998: 115) notes that attitudes started to change in the 1960s, and 1970s, and that the Tate Gallery exhibition on Pre-Raphaelites in 1984 was one of the most popular ever mounted by the gallery. Art theorists also stress how innovators in art can dispel the unattractive associations of the current establishment. Gombrich summarises what he
162
Path dependence in technical change
calls (with a little irony) the ‘treasured legend of the modern movement’ (Gombrich, 1963a: 145): What is called loosely ‘modern art’ sprang indeed from a protest against the lie in the soul, a revulsion from false values. When new classes of patrons acquired unexpected wealth and were bent on ostentation, cheap vulgarity stifled our cities and choked our drawing rooms. Sentimental trash was taken for Great Art. This sickened the heart of the true artists who went on their lonely and perilous way in the face of public neglect and derision.
Others went further and argued that true modernity was a negation of all that is past – not just the current establishment. Marcus (1998: 7) argues as follows: ‘To be modern, design did not just have to be new, it had to be free of any reference to the decorative styles of the past.’ Some, by contrast, recognised that negation of the present establishment could lead to a rediscovery of past styles. So, for example, while Malraux (1954) describes a concept of perpetual revolution in art, of the artist as a ‘defiant animal’ and hence a theory of changing styles of (and tastes for) of art, he nevertheless recognises a potential connection with the distant past. Gombrich summarises Malraux’s position very succinctly (Gombrich, 1963b: 83): ‘Modern art came into being as a protest against the commercial pseudoart of prettiness. It is this element of negation that establishes its kinship with the religious art of the past . . .’. Gombrich (1963b: 83) recognises, however, the fundamental paradox that today’s revolutionaries in art become tomorrow’s establishment: ‘True, modern art will not be able to “outlive its victory intact”. As an act of defiance it will wither away when it becomes dominant.’ Historicism, the practice of borrowing from the more distant past (even that untainted by recent associations) was an anathema to true modernists. Marcus (1998: 7) argues: Ever since the middle of the previous century, reformers had condemned design’s dependence on historicism (and its handmaiden, ornamentalism), and the progress of modern design could be measured by the extent to which the mining of historic styles was supplanted by the creation of new, anonymous, and universal forms, forms that looked to the future instead of the past.
Nevertheless, it was still widely accepted, especially in the nineteenth century, and as McDermott (1992: 120–21) notes, an obsession for the past lead to the publication of a large number of design source books at that time. The history of what was popular and unpopular at particular times in the past shapes what is popular and unpopular today. Using the term, ‘path
Mapping changes in the taste for art
163
dependence’ in an informal way, we can say that the history of taste is pathdependent. Some of the manifestations of path dependence that have drawn greatest attention in the literature occur when there is positive feedback only, leading perhaps to ‘lock in’. Here, in contrast, there is a mix of positive and negative feedback.
4
PATH DEPENDENCE
In some parts of the literature, the term ‘path-dependent’ has been used in a rather casual fashion. David (1997: 13) helps us to be much more precise. Path dependence is a dynamic property of allocative processes. It can be defined with reference to either: (a) the relationship between process dynamics and outcome(s) to which it converges; or (b) the limiting probability distribution of the stochastic process under consideration. To understand the true meaning of path dependence, it is easiest perhaps to start by defining path independence. A path-independent process is one whose dynamics guarantee that it will converge to a unique, globally stable equilibrium, or, in the case of stochastic systems, where the outcome has an invariant stationary asymptotic probability distribution. Stochastic systems with this latter property are ergodic. That means they are able to shake off the influence of their past. David (1997:13) offers a ‘negative definition’ of path dependence: ‘Processes that are non-ergodic, and thus unable to shake free of their history, are said to yield path dependent outcomes.’ Working on from this, he can provide a ‘positive definition’ (David, 1997: 14): ‘A path dependent stochastic process is one whose asymptotic distribution evolves as a consequence (function) of the process’ own history.’ In short, there are three key points about path dependence (David, 1997: 18–19): ● ● ●
Path dependence is a property of stochastic dynamic systems. It is natural to interpret a path-dependent process as a contingent branching process. The definition of path dependence is independent of any issues of economic efficiency or inefficiency.
From this point of view, there seems little doubt that the evolution of tastes and prices in the market for art is truly a path-dependent process – even if not all the models that could be invoked to analyse these are strictly path-dependent.
164
5
Path dependence in technical change
CHARACTERISTICS AND MULTIDIMENSIONAL SCALING
We can now return to the primary aim of the chapter: to map the changing tastes for works of art as manifested in the prices of paintings sold at auction. Whenever economists seek to represent different products in a common space, the natural starting point is the characteristics approach, developed in the modern literature by Gorman (1980), Ironmonger (1972) and Lancaster (1971). But what characteristics are needed to give an adequate representation of objects so subtle as works of art? Here Bacharach’s (1990) work on commodities and language is helpful. We can judge the necessary characteristics from the things people say about these objects. But in the case of works of art, people have said a great deal. For example, Wildenstein’s (1996) catalogue of paintings by Monet runs to four volumes and over 1500 pages. It would seem to be a difficult task indeed to capture all this in a set of characteristics scores! Nevertheless, both economists and art historians have tried to capture the essence of good art in a few characteristics. As De Marchi (1999: 6) notes, Adam Smith argued that with objects of ‘art’, we derive pleasure from four characteristics: ● ● ● ●
Form Colour Rarity Ingenuity of design and manufacture.
Equally, the nineteenth-century art and social critic, John Ruskin, said that there were four essential characteristics of great art (Ruskin 1856 [1996], vol. 5: 48–63): ● ● ● ●
Choice of Noble Subject Love of Beauty Sincerity Invention (Imagination).
And, amongst classical writers on this theme, perhaps the greatest advances were made by De Piles, who again identified four characteristics of art (De Piles, 1708, here quoted from De Marchi, 1999: 8–10): ● ●
Design Colouring
Mapping changes in the taste for art ● ●
165
Composition Expression.
De Piles, indeed, attempted to rank the work of 56 different painters using his own estimates of their scores on these four characteristics. And in the modern literature, the hedonic technique has been applied to works of art (Chanel et al., 1996). However, in this chapter, we shall take a different approach. We seek to identify an implicit taste space by reference to observed trends in the prices of works of art. The basic logic of our approach is this. If the prices of works by two painters are closely correlated over time, then we assume that these painters are located close together in the taste space. If, on the contrary, the prices of works by two painters are negatively correlated, then we assume that these painters are located far apart in the taste space. A variety of techniques exist to enable us to map from an n*n matrix of correlations between objects into a two-dimensional (or higher dimensional) representation of those objects. These include principal components, factor analysis and multidimensional scaling. Such methods give a ready way of illustrating the similarities and differences between different entities, whether they are nations, companies, products or, even, consumers. For a wide range of empirical data-sets in economics, and other business studies, two components capture a large share of the total variance. However, these statistical techniques are often used in a rather ad hoc way, which does not describe the precise microeconomics of how correlations in price map into proximity in taste space. That mapping is set out in detail below. Accordingly, in what follows, we assume that prices are sufficient statistics for describing the evolution of tastes in the art market. In reality, they are not. The art historian has much to tell us about influences which change tastes, that are not captured in prices. Even stronger, this chapter assumes in effect that a matrix of correlation coefficients between the prices of different artists’ work is a sufficient set of statistics. This is somewhat stronger, because it does not make use of the more detailed trends in popularity of different artists. Are these assumptions too strong? In the second edition of The Principles of Economics, Marshall provided an interesting insight into why prices might not give a wholly accurate measure of taste (here quoted from White, 1999: 79):2 And therefore the price at which such a thing is sold will depend very much on whether rich persons with a fancy for that particular thing happen to be present at its sale. If not, it will probably be bought by dealers who reckon on being able to sell it at a profit; and the variations in the price for which the same picture sells
166
Path dependence in technical change
at successive auctions, great as they are, would be greater still if it were not for the steadying influence of professional and semi-professional purchasers. The ‘equilibrium price’ for such sales is very much a matter of accident.
Indeed, since our data source (Reitlinger, 1961; 1970) simply records the auction prices of works (and also some prices relating to private sales), we do not know how many potential buyers were ‘in the market’ for a particular work at any time. However, Ruskin believed that price was a reasonable measure of the artist’s rank (Ruskin 1843 [1996], vol. 3: 617–18): ‘Of course a thousand modifying circumstances interfere with the action of the general rule; but, taking one case with another, we shall very constantly find the price which the picture commands in the market a pretty fair standard of the artist’s rank of intellect.’ The technique developed below refers to the ‘height of taste’. We mean by this that location on our space of tastes and artists where demand is strongest. We shall assume throughout that at any time the artist whose work comes closest to the ‘height of taste’ will ‘enjoy’ the highest prices. This is an important, if strong, assumption in what follows.
6
FLAT OR ROUND? MODERNISM OR HISTORICISM?
Before starting to construct our map of painters, there is one further issue that needs attention. Should the space be flat or round? It is conventional in multidimensional scaling and principal components to use a linear or planar representation of the data. But there is an implicit but rarely explored assumption behind this procedure, which could be described as the ‘transitivity of distance’. This is an inherent property of planar representations but would not be found, for example, in a spherical projection.3 But it is possible that goods exhibiting cycles in popularity may be better represented by circular or spherical projections. Ultimately, this is an empirical question. And in what follows we shall create both planar and spherical representations. However, the ideas from art history and art theory summarised in section 3 also give us an important lead here. To achieve the modernist goal that design has to be different from any designs of the past, it is necessary to locate painters in an unbounded plane. If there are any bounds to this space, it is impossible for design to be innovative and continually to evade any reference to the past. Sooner or later the artist will be forced back into a part of the space that has been occupied by earlier artists.
Mapping changes in the taste for art
167
On the other hand, if one accepts that a degree of historicism is an inevitability, then a bounded space will suffice. In that case, we may wish to distinguish between two types of historicism: ● ●
Distant past historicism: it is legitimate to use the styles of the distant past, but not those of the recent past. Recent past historicism: it is legitimate to use the styles of the recent past, but not those of the distant past.
Assume that the movement of the ‘height of taste’ around the space of artists is continuous and smooth. If artists and tastes are represented on a real line, then when the ‘height of taste’ hits the boundary of the space, tastes must move in the reverse direction, and this will entail recent past historicism. On the other hand, if artists and tastes are represented on a circle, then the ‘height of taste’ can continue to cycle in one direction (say clockwise) without reversal: this will entail distant past historicism. When we move from lines and circles up to planes and spheres, some more subtle possibilities emerge. But if distant past historicism is the more common than recent past historicism (and the references cited in section 3 suggest it is), then a spherical projection may be more useful for the purposes of this chapter.
7
TECHNICAL DETAILS
This section presents techniques for mapping from a correlation matrix representing the similarities and differences between products to a planar spherical representation of the positions of those products. The reason we choose to represent products on the surface of a sphere and not on the perimeter of a circle is simply so that two degrees of freedom (or two components) are obtained, rather than one. A result of considerable power and great simplicity is derived: for two products located on the surface of a sphere, the correlation between their prices is equal to the cosine of the angle between them – as measured from the centre of the sphere. A variety of standard data reduction methods (principal components, multidimensional scaling) take a matrix of distances (or similarities) between entities and project these entities onto a plane. Points close together imply that the entities are similar, while points far apart imply that the entities are dissimilar. As indicated above, however, there are some disadvantages from projecting onto a plane. In particular, if the ‘height of taste’ moves around in a continuous fashion, it implies a particular and perhaps restricted pattern of
168
Path dependence in technical change
fashion cycles. By contrast, there are some attractions in locating points on the surface of a sphere for in that case a rather different pattern of cycles is available. In this section we show how to map from a matrix of correlations into planar and spherical projections. But in each case, we start with the easier case of a line and a circle – because these help us to grasp the intuition of what is happening. Line Assume that each artist is located at some point along a line (0,xmax). The ‘height of taste’ is defined as x*, and the strength of demand at any other point (xi) along the line is defined as: S(xi, x*) 1 |x* xi |
(6.1)
As Figure 6.1 shows, the signal is at a peak when x* xi, and drops away on either side. We can use this simple model to compute the correlation between demand prices for artists located at different points along the line. To see this, consider Figure 6.2. Swann (2001b) shows a very simple and convenient result. Under some assumptions, we can define the correlation between these demand prices as follows: "
12
[(x1 (x2 x1)] (xmax x2) 2(x2 x1)
1 xmax xmax
(6.2)
Demand 1
0
0 Figure 6.1
x1 Signal strength along a line
xmax
Mapping changes in the taste for art
Demand 1
D1
D2
x1
x2
169
0
0 Figure 6.2
xmax
Correlation in demand prices
The basic intuition of this result is as follows. We can split Figure 6.2 into three areas. From 0 to x1, the correlation between demand prices is 1; from x1 to x2 that correlation is 1; and from x2 to xmax that correlation is again 1. In short, there is a natural mapping from positions to correlations: when x1 and x2 are located close together, the correlation between their demand prices is close to 1; when they are located far apart, at either end of the painter spectrum, then, the correlation between their demand prices approaches 1. Plane Extending the previous result to a plane is messy rather than difficult. To make it as simple as possible, it is helpful to use a grid measure of distance (Figure 6.3) rather than an Euclidean measure. Thus, the distance between any two points (x1,y1) and (x2,y2) is defined as: L |x1 x2| |y1 y2|
(6.3)
And as above, we assume that when the ‘height of taste’ is (x*,y*), then the demand at point (x1,y1) is given by: S 1 L 1 |x1 x*| |y1 y*|
(6.4)
170
Path dependence in technical change
ymax
y2
y1
0 Figure 6.3
x1
x2
xmax
A grid measure of distance
Swann (2001b) shows that the correlation between demand for 1 and demand for 2 can be calculated by the following procedure. Split the plane into nine zones as shown in Figure 6.4. (Note that this can be done wherever the two points lie. If y1 y2 then the central horizontal band W–C–E disappears; if y2 0 then the lower horizontal band disappears; and so on). Swann (2001b) shows that under some conditions, the overall correlation between 1 and 2 over the entire plane can be computed as: corr12
9
a(i)"12(i)
(6.5)
1
where a(i) is the area of zone i and "12 is the correlation between demand for 1 and demand for 2 in zone i. These zone areas and zone correlations are given in Table 6.1.: Once again, if artists are clustered together in a particular part of the plane, then there is a strong positive correlation in their demand prices. By contrast, if they are located at opposite corners of the plane, then this correlation will be strong and negative.
171
Mapping changes in the taste for art
ymax NW
N
NE
W
C
E
SW
S
SE
y2
y1
0 Figure 6.4 Table 6.1
x1
x2
xmax
Nine zones to describe correlation in demand Zone areas and zone correlations
Zone i
Area a (i)
Correlation12 "12(i)
NW N NE W C E SW S SE
(ymax y2)·x1 (ymax y2)·(x2 x1) (ymax y2)·(xmax x2) (y2 y1)·x1 (y2 y1)·(x2 x1) (y2 y1)·(xmax x2) y1·x1 y1·(x2 x1) y1·(xmax x2)
1 [(ymax y2) (x2 x1)]/[(ymax y2) (x2 x1)] 1 [x1 (y2 y1)]/[x1 (y2 y1)] 1 [(xmax x2) (y2 y1)]/[(xmax x2) (y2 y1)] 1 [y1 (x2 x1)]/[y1 (x2 x1)] 1
Circle While superficially it may look harder to project points onto a circle than a line (and on the surface of a sphere than on a plane), in fact it is in some respects easier. We shall see that a very simple result obtains: the correlation between the demand for and price of work by two different artists is given by the angle between those two artists as located on the circle and viewed from the centre of the circle. Consider Figure 6.5. Suppose that a particular artist is located at point x1 on the circle, and that at a particular time and date, the most popular artist or ‘height of taste’ is at point x*. As in the
172
Path dependence in technical change
x* .
1
cos(#)
x1 #
0
1 Figure 6.5
Signal strength around a circle
case of the line or plane, we need a function relating demand or demand price for x1 to x*. Here we adopt a slightly different convention to that used in the case of the line or plane. We liken the question to that of computing the brightness of daylight at different places on the globe. Suppose that demand is a signal transmitted from the dotted line at the top of Figure 6.5 (which is tangent to the circle at x* and which passes through 1 on the vertical axis). The intensity of demand felt at any other point on the circle depends on the vertical distance from the transmitter line to that point on the circle. From the diagram, it is readily apparent that this vertical distance depends on the angle between x1 and x*. As drawn, the perpendicular distance from transmitter line to x1 is 1 cos(#). Hence, if we measure the strength of the signal at x1 on a scale from 1 to 1, then that strength of signal is given by cos(#). This is reasonable. When the angle between x1 and x* is small, so that our chosen artist is close to the ‘height of taste’, then demand is very strong: as # → 0, cos(#) → 1. By contrast, when the angle # is large, suggesting our chosen artist is far from the ‘height of taste’, then demand is small. Now, using this framework, we can obtain a remarkably simple and powerful result about the correlation between demand prices for different artists. Take two artists, 1 and 2, located at points x1 and x2 on the circumference of the circle. Suppose that the angle between each artist and
Mapping changes in the taste for art
173
the vertical (as drawn in Figure 6.5) is denoted by #1 and #2 respectively. Suppose also that the angle between the location of the ‘height of taste’ at any date and the vertical is given by #*. (Note that in the diagram, #* 0.) Then, the strength of demand for the work of painters 1 and 2 is simply defined by cos(#1 #*) and cos(#2 #*), respectively. From this we can obtain the following expression for the covariance of demand prices for 1 and 2. Assume that the ‘height of taste’ might in the fullness of time occur above any point on the circle, and with equal probability. This means that we can assume that the density of #* is constant over the range $ to $. We obtain the covariance by integrating from $ to $ as shown: cov12
1 $ cos (#1 #*)cos (#2 #*)d#* 2$ $
1 $ 1 $ cos (# #*)d#* · 1 cos (#2 #*)d#* 2$ $ 2$ $
(6.6)
In fact, this expression can be simplified considerably by using two results. First, when k is an integer (positive or negative) and because sin (a k$)
sin(a k$): $
1
cos (a k#*)d#* k [sin(a k$) sin (a k$)] 0 $
(6.7)
Second, Ayers (1987: 143, eqn 9) shows that: 1 cos (x) cos (y) [cos (x y) cos (x y)] 2
(6.8)
Applying the first result to equation (6) we see that the second row of that expression is simply zero. Moreover, using the second result, we see that the first line of equation (6.6) simplifies to: 1 $ 1 $ cos (# # )d#* 1 2 cos (#1 #2 2#*)d#* 4$ $ 4$ $ cos(#1 #2) (6.9)
2
cov12
Now, we can use equation (6.9) to derive the variance of demand for 1 (or 2): var1
cos(#1 #1) 1
2 2
(6.10)
174
Path dependence in technical change
From which it is clear that the simple correlation between the demand prices of 1 and 2 is given by: "12
cos(#1 #2)2
cos(#1 #2) √(12) · (12)
(6.11)
In short, the formula describing the demand intensity for one artist when another is the ‘height of taste’ also defines the correlation between demand prices for the work of different artists. This simple but powerful result plays a central role in what follows. Sphere An equivalent result obtains in the case of the sphere, but the proof is rather more cumbersome. It is also helpful to use a slightly different notation. Suppose that we follow cartographical principles and identify any point on a sphere by its latitude (a) and longitude (n). Start by assuming that the ‘height of taste’ is located at the ‘North Pole’, that is, where latitude 90%N (or 90%, or $/2). In this case the result from the previous section carries over in a straightforward way. To see this, examine Figure 6.6. As in the case of the circle, a radiating plane is tangent to the sphere at the point (n0,a0), where a0 90%, which defines the ‘height of taste’. The strength of demand signal at another point is defined by the vertical distance between the radiating plane and that other point. Now, as
(n0, a0)
(n0, a1)
(n1, a1) S(n0, a1) = cos(a0 a1)
Figure 6.6
Signal strength on a sphere
Mapping changes in the taste for art
175
is clear from the diagram, when the ‘height of taste’ is at the North Pole, then the strength of demand anywhere else depends only on the latitude at that other point. The diagram shows two points (n0,a1) and (n1,a1) on the same latitude a1, and it is readily apparent that the demand strength is the same at both: cos(a0 a1). This convenient result does not apply when the ‘height of taste’ is not at the North Pole. Then a more complex formula applies. This asymmetry does not imply any imperfection in the sphere, rather it results from a fundamental asymmetry in the treatment of longitude and latitude in cartography: latitude is defines between – 90% and 90% while longitude is defined between 180% and 180%. Figure 6.7 shows the more general case. Here the ‘height of taste’ is at (n0,a0) and we wish to compute the strength of demand at another point (n1,a1). It is easiest to do this in two stages. First, define another point (n0,a1) which is located on the same longitude as the ‘height of taste’ and the same latitude as the other artist. Compute the strength of the demand signal at (n0,a1) by the perpendicular distance
R
(n0, a0)
(n0, a1)
Figure 6.7
(n1, a1)
Signal strength on a sphere
S(n0, a1)
176
Path dependence in technical change
between the ‘radiating plane’ tangent to the sphere at (n0,a0) and this intermediate point (n0,a1). The strength of demand at (n0,a1) is given by S(n0,a1). Then, second, compute the perpendicular distance between this intermediate point and the final point (n1,a1). This second distance is defined by R in the diagram. The strength of the signal at (n1,a1) can then be computed as S(n0,a1) R. The first stage is easy. Because (n0,a0) and (n0,a1) are on the same longitude, then we can use the result of the previous section to compute the strength of signal at (n0,a1). It is simply given by cos(a0 a1). The second stage is a bit harder. First, compute the horizontal distance from (n0,a1) and (n1,a1) along the horizontal (and dotted) latitude line. This can be calculated as follows. The horizontal girth of the sphere at latitude a1 is given by 2cos(a1). At the equator (a1 0), the horizontal girth is at its maximum (equal to the diameter of the sphere, which is 2). Away from the equator, the girth is smaller. The horizontal distance between (n0,a1) and (n1,a1) however, is only a proportion of this girth, equal to: 2cos(a1) ·
[1 cos (n0 n1)]
cos (a1) · [1 cos (n0 n1)] 2
(6.12)
Moreover, this horizontal distance between (n0,a1) and (n1,a1) overstates the distance when measured perpendicular to the ‘radiating plane’ (R). To obtain R, we need to multiply the horizontal distance by cos(a0). Hence the strength of demand at (n1,a1) when the ‘height of taste’ is at (n0,a0) is defined by: S(n1, a1) cos(a0 a1) cos(a0) cos(a1) · [1 cos(n0 n1)]
(6.13)
By putting a0 90%, we can see that this formula handles the special case where the ‘height of taste’ is at the North Pole, and where the strength of demand is simply cos(a0 a1). Once again, it can be shown that this formula for the strength of demand can also be used to compute the correlation between demand prices of two different artists. The proof is very cumbersome but the basic idea is as follows. Once again, define two different artists 1 and 2 by the longitude/ latitude coordinates: (n1,a1) and (n2,a2). Again, assume that the ‘height of taste’ could be found with equal probability above any point on the surface of the sphere. Then we can show (Swann, 2001b) that the correlation between the demand prices of 1 and 2 is given by: corr12 cos (a1 a2) cos (a1) cos(a2) · [1 cos (n1 n2)]
(6.14)
Mapping changes in the taste for art
177
Finally, we can – as a special case of the sphere – locate points on a hemisphere. This is done simply by limiting the longitude to the range { $/2, $/2}.
8
SOME ILLUSTRATIVE RESULTS
The main data for this study come from Reitlinger’s (1961; 1970) studies of the auction prices of works of art. In addition, we have generated an approximate price deflator using data from Mitchell (1980) and Mitchell and Deane (1971). In particular, we have focused on the auction prices of oil paintings by some of the major artists of the seventeenth, eighteenth, nineteenth and twentieth centuries. We should add, of course, that some of the artists commanding high prices in the past have fallen right out of favour in the twentieth century. Indeed, that is one of the phenomena that this study has set out to explore. Accordingly, the criterion for inclusion is that the artist commanded relatively high prices (in real terms) at some point of the evolution of the art market between 1800–1970, even if his/her work is not highly priced now. Guerzoni (1995) has discussed some of the shortcomings of these data. The raw data are not in an ideal form for econometric analysis. This is no reflection on the accuracy of the data. Indeed, since these are for the most part, auction prices, where the agreed sums are recorded, then the data are very accurate. A few prices are estimated by Reitlinger, but this is not a serious source of error. Rather, the problems with these data reflect two main factors. The first is the non-homogeneity of works of art. Clearly, different works are unlikely to be of equal merit. Some are small paintings, others are large canvasses. While art historians have documented these paintings in great details, we did not (as discussed above) think it practical to attempt to turn these qualitative descriptions into a list of characteristics or to construct a hedonic analysis of art prices. At most, we have tried to reduce the degree of variance in art prices by restricting our attention to oil paintings, neglecting most watercolours or prints, and neglecting the smaller works or studies. Also, where two or three paintings were sold as a lot we have attributed the total price paid between the constituent parts of the bundle. Second, while the volume of art traded increased markedly in the postwar period, and in particular during the 1960s, this is still a fairly thin market. Amongst the great old masters, indeed, few works of significance came up for auction in recent years. As a result, it has not been practical to include a number of old masters in our sample – there is just too little data,
178
Path dependence in technical change
and what sales there are do not represent their great works. Moreover, sales are infrequent and unevenly spaced. While some econometricians might wish to delve into the peculiar timeseries properties of these data, we have chosen instead to perform a very simple analysis of these data, in three steps. First, we have constructed an average price for each artist in each year as follows. We take a 20-year moving average covering all the items sold in the last 20 years. Naturally, this smoothes the raw data considerably. Second, we have deflated these prices by a general price index derived from Mitchell (1980) and Mitchell and Deane (1971). A little splicing was required to make the index continuous back to 1840, and while it is far from perfect, it is adequate for our immediate purposes. Some have argued that it would be even more interesting to deflate by an art price index, so that one can look simply at relativities within the art market, and abstract from the secular upwards trend in art prices during (for example) the 1960s. Eventually, we may be able to do that, but it would require a more comprehensive set of prices that we have analysed so far. Third, we have computed a simple correlation coefficient between the price series for each painter. In practice, it is unlikely that a simple correlation coefficient is a sufficient statistic for all the analysis we might want to do here. For example, some artists have shown price cycles with relatively low periodicity, while others have exhibited perhaps one major cycle in 100 years or more. Figure 6.8 illustrates this for three artists, chosen more or less at random, but who do happen to show some very different timeseries properties. However, for the preliminary analysis in this chapter, we shall work just with these correlation coefficients using data on the prices of works by 20 artists (listed in Appendix 1) for up to 130 years. Econometric Methods With the correlation matrices computed as described in the previous section, we have applied the methods of Section 7 to create some preliminary artist maps. Of course, as noted above, this is by no means a new procedure since there is a very well-established tradition of using principal components (or multidimensional scaling) to construct such maps. However, the analysis of Section 7 helps to bring out more precisely the way in which product locations in a characteristics space map into demand price correlations. And it is only with that precision that we can hope to discriminate between planar and spherical projections as a means of representing these data. The basic procedure is straightforward. For any set of coordinates for all the artists in our sample, we can compute the matrix of implied
179
1880
1900
1920
1940
1960
Author’s calculations based on data in Reitlinger (1961; 1970), Mitchell and Deane (1971) and Mitchell (1980).
1860
Figure 6.8 Price of oil paintings – 1970 prices, 20-year moving average
Source:
10 1840
100
1000
10 000
100 000
£
1980
Alma-Tadema
Canaletto
Cézanne
180
Path dependence in technical change
correlations between them. We can then compare that to the actual and compute a matrix of residuals. We then add up the absolute values of these residuals across each element in the correlation matrix to obtain a sum of absolute residuals. The aim of our optimisation algorithm, then, is to choose the set of coordinates which minimises this sum of absolute residuals. The minimand is of course a highly non-linear function of the coordinates. But we can apply the Lasdon–Waren GRG2 (Generalised Reduced Gradient) nonlinear optimisation routine to solve this (Lasdon and Waren, 1978; 1981; Lasdon et al., 1978). The problem can be seen as an exercise in mathematical programming: we have to minimise the sum of absolute residuals with upper and lower limits on all the coordinates. For the planar representation, we assume that all artist coordinates are between 0 and 1 on both axes. For the spherical representation, the latitude must be between $/2 and $/2 while the longitude must be between $ and $. Results The sum of absolute residuals for the 400 elements of the correlation matrix comes to 23.5 for the planar representation and 17.4 for the spherical representation. This could suggest that the spherical representation is slightly preferable, but we are cautious about making such a claim. The results described here are preliminary and we cannot be certain that we have reached the global optimum in each case. These statistics correspond to a mean absolute error of 4 percentage points in the spherical representation and 6 percentage points in the planar representation. These figures are quite acceptable. Figures 6.9 and 6.10 show, respectively, the coordinates obtained for the planar and spherical representations. They are actually strikingly similar. In Figure 6.9 we have superimposed a dotted circle, within the plane, and it is striking how most artists cluster around this circle. Indeed, this suggests that were we to constrain the artists to lie on a circle and not on a sphere then that constraint would not do too much damage to the data. By contrast, it is clearly inappropriate to constrain the artists to a (onedimensional) line. We can also use these charts to interpret the main trends in the prices of these artists during the period 1840–1970. In the nineteenth century and at the start of the twentieth century, the artists (Landseer, Collins, Meissonier, Alma-Tadema) in the top left-hand part of these charts were at the ‘height of taste’. During the early and middle part of the twentieth century, the ‘height of taste’ was moving in an anticlockwise fashion, with a strong revival in the prices of Nattier and Hals.
181
Mapping changes in the taste for art 1
Landseer
Collins
Meissonier Alma-Tadema
Boucher
0.5 Claude
Canaletto
Constable Nattier
Monet Sisley Pissarro van Gogh Cézanne Manet Degas
Rembrandt
Hals
0
Figure 6.9
Renoir 0.5
0
Bonnard
1
Planar map of 20 artists
The first of the impressionists in the bottom right (Renoir) was also the first to enjoy a rapid growth in prices in the early twentieth century, while the prices of artists further to the right took off later (van Gogh and Bonnard). In short, the ‘height of taste’ has, over 100 years or so, completed rather more than half an anticlockwise circuit from top left to bottom right. Where next? The results here probably do not bring out the full advantage of the spherical projection over the planar. Indeed, in the best solution obtained to date, no painter is located on the back of the sphere: the front hemisphere fits everybody in comfort. This work is continuing, and when we complete computations for a much larger number of artists, it is likely that the algorithm will need to spread out the artists over a wider area, and some will locate on the back of the sphere.
182
Path dependence in technical change $/2 Collins Landseer Meissonier Alma-Tadema
Boucher $/2
$/2
0 Claude Constable Canaletto
Nattier
Rembrandt Hals
Monet Manet Degas Renoir
Bonnard
Sisley van Gogh Cézanne Pissarro
$/2
Figure 6.10
9
Spherical map of 20 artists
CONCLUSIONS
The aim of this work is to construct a map of artists and to illustrate how tastes evolve within that map. We have seen that there is an interesting similarity between some of the economic models of evolving tastes and what art historians and art theorists have written about the evolution of art. Moreover, we have seen that the evolution of tastes is clearly a pathdependent process, in the sense of the term defined by Paul David. Rather than constructing a characteristics space for works of art, the chapter constructs an implicit space of painters derived from the correlations between prices of different painters’ work. We discuss whether a planar representation or a spherical representation would be preferable. We suggest that the modernist ideal, that new art be divorced from any reference to the past, requires an unbounded planar representation. We also distinguish between two types of historicism: distant past historicism and recent past
Mapping changes in the taste for art
183
historicism. The former fits more comfortably with a spherical representation while the latter fits best with a planar representation. The main theoretical results of the chapter are to show exact methods for deriving planar and spherical representations of artists from a correlation matrix of prices. The spherical representation offers a result of striking power and similarity. For two artists located on a circle, the correlation between the prices of their work is shown to equal the cosine of the angle between them’ and an analogous result applies to the spherical case. These methods are applied to a representative sample of data on the prices of work by 20 artists, taken from the period 1840–1970. The maps generated show these painters arrayed over three-quarters of a hemisphere, and the ‘height of taste’ has, over 100 years or so, moved top left to bottom right. The results presented here are incomplete, but the techniques described in this chapter can help us to understand the path-dependent nature of tastes in art, and the associated waves in popularity of different artists.
APPENDIX 1 LIST OF ARTISTS ANALYSED IN THIS STUDY Alma-Tadema, Lawrence: 1836–1913 Bonnard, Pierre: 1867–1947 Boucher, François: 1704–70 Canaletto, Antonio: 1697–1768 Cézanne, Paul: 1839–1906 Claude Le Lorrain (or Claude Gellée): 1600–82 Collins, William: 1788–1847 Constable, John: 1776–1837 Degas, Edgar: 1839–1917 Hals, Frans: 1584–1666 Landseer, Edward: 1802–73 Manet, Edouard: 1832–83 Meissonier, Ernest: 1815–91 Monet, Claude: 1840–1926 Nattier, Jean-Marc: 1685–1766 Pissarro, Camille: 1831–1903 Rembrandt van Ryn: 1606–69 Renoir, Pierre Auguste: 1841–1919 Sisley, Alfred: 1840–99 Van Gogh, Vincent: 1853–90
184
Path dependence in technical change
NOTES 1. Nottingham University Business School, Jubilee Campus, Wollaton Road, Nottingham, NG8 1BB, UK. Email:
[email protected]. I am grateful to participants in the conference on New Frontiers in the Economics of Innovation and New Technology, held in honour of Paul David, Turin, May 2000, at a seminar at London Business School, and at the International Schumpeter Society Conference in Manchester, June 2000, and also to Catherine Beaudry, John Cantwell, Robin Cowan, Stuart MacDonald and Jenny Swann for helpful discussion about this chapter, but none of these are responsible for any errors and idiosyncrasies. Funding from the Social Sciences and Humanities Research Council of Canada (with R.A. Cowan) is gratefully acknowledged. 2. As White (1999) notes, this paragraph was abbreviated in later editions. 3. To explain this, consider the travel analogy. Suppose we take a train journey from London to Edinburgh via York. Since York is north-west of London and Edinburgh is north-west of York then the distance from London to Edinburgh exceeds the distance from London to York. This transitivity works on a local scale, but it does not work on a global scale. Thus, for example, to fly from London to Tokyo is a journey of 9 600 km to the east, and to fly from Tokyo to New York is a further journey of 10 800 km to the east. But this does not imply that the distance from London to New York is greater than the distance from London to Tokyo. Indeed, at 5600 km that is the smallest distance of the three. So on this global sphere, distance rankings are not transitive.
REFERENCES Antonelli, C. (1997), ‘The economics of path dependence in industrial organisation’, International Journal of Industrial Organisation, 15, 643–52. Ayers, F. (1987), Calculus, 2nd edition, New York: Schaum/McGraw-Hill. Bacharach, M. (1990), ‘Commodities, language and desire’, Journal of Philosophy, 87 (7), 346–68. Barnes, R. (1998), The Pre-Raphaelites and Their World, London, Tate Gallery. Baumol, W.J. (1986), ‘Unnatural value: or art investment as floating crap game’, American Economic Review, 76 (2), 10–14. Becker, G.S. (1996), Accounting for Tastes, Cambridge, MA, Harvard University Press. Bell, C. (1927), Landmarks in Nineteenth-Century Painting, London, Chatto and Windus. Bell, Q. (1984), ‘The Pre-Raphaelites and their critics’, in L. Parris (ed.), PreRaphaelite Papers, London: Tate Gallery, pp. 11–22. Bianchi, M. (ed.) (1998), The Active Consumer: Novelty and Surprise in Consumer Choice, London and New York: Routledge. Bourdieu, P. (1984), Distinction: A Social Critique of the Judgement of Taste, London, Routledge and Kegan Paul. Chanel, O., L-A. Gerard-Varet and V. Ginsburgh (1996), ‘The relevance of hedonic price indices: the case of paintings’, Journal of Cultural Economics, 20, 1–24. Cowan, R.A., W. Cowan and G.M.P. Swann (1997), ‘A model of demand with interaction between consumers’, International Journal of Industrial Organisation, 15, 711–32. Cowan, R.A., W. Cowan and G.M.P. Swann (2004), ‘Waves in consumption with interdependence between consumers’, Canadian Journal of Economics, 37 (1).
Mapping changes in the taste for art
185
Cozzi, G. (1998), ‘Culture as a bubble’, Journal of Political Economy, 106, 376–94. David, P.A. (1985), ‘CLIO and the economics of QWERTY’, American Economic Review, 75, 332–36. David, P.A. (1986), ‘Understanding the economics of QWERTY: the necessity of history’, in W.N. Parker (ed.), Economic History and the Modern Economist, London, Basil Blackwell. David, P.A. (1987), ‘Some new standards for the economics of standardization in the information age’, in P. Dasgupta and P.L. Stoneman (eds), The Economics of Technology Policy, London: Cambridge University Press. David, P.A. (1988), Path-Dependence: Putting the Past into the Future of Economics, Technical Report 533, Institute for Mathematical Studies in the Social Sciences, Stanford University. David, P.A. (1992), ‘Heros, herds and hysteresis in technological history: Thomas Edison and the “Battle of the systems” reconsidered’, Industrial and Corporate Change, 1 (1), 129–80. David, P.A. (1993), ‘Path dependence and predictability in dynamic systems with local network externalities: a paradigm for historical economics’, in D. Foray and C. Freeman (eds), Technology and The Wealth of Nations: The Dynamics of Constructed Advantage, London, Pinter. David, P.A. (1994), ‘Why are institutions the “carriers of history”? Path-dependence and the evolution of conventions, organizations and institutions’, Structural Change and Economic Dynamics, 5 (2), 205–20. David, P.A. (1997), Path Dependence and the Quest for Historical Economics. One More Chorus of the Ballad of QWERTY, Discussion Papers in Economic and Social History, Number 20, University of Oxford. David, P.A. and D. Foray (1994), ‘Percolation structures, Markov random fields and the economics of EDI standards diffusion’, in G. Pogorel (ed.), Global Telecommunications Strategies and Technological Change, Amsterdam, Elsevier Science. David, P.A., D. Foray and J.-M. Dalle (1998), ‘Marshallian externalities and the emergence and spatial stability of technological enclaves’, Economics of Innovation and New Technology, 6 (2/3), 147–82. De Marchi, N. (1999), ‘Introduction’, in N. De Marchi and C.D.W. Goodwin (eds), Economic Engagements with Art, Durham, NC, and London, Duke University Press, pp. 1–30. De Marchi, N. and C.D.W. Goodwin (eds) (1999), Economic Engagements with Art, Durham, NC, and London: Duke University Press. De Piles, R. (1708), Cours de peinture par principes, reprinted as English edition (1743), The Principles of Painting, London. Dosi, G., R. Aversi, G. Fagiolo, M. Meacci, and C. Olivetti (1999), ‘Demand dynamics with socially evolving preferences’, in S.C. Dow and P.E. Earl (eds), Economic Organisation and Economic Knowledge: Essays in Honour of Brian Loasby, Cheltenham, UK and Lyme, USA: Edward Elgar. Frank, R.H. (1985), Choosing the Right Pond: Human Behaviour and the Quest for Status, New York, Oxford University Press. Frey, B.S. (1997), ‘Art markets and economics: introduction’, Journal of Cultural Economics, 21, 165–73. Frey, B.S. (2000), Art and Economics, Heidelberg: Springer-Verlag. Frey, B.S. and W.W. Pommerehne (1989), Muses and Markets: Explorations in the Economics of the Arts, Oxford, Basil Blackwell.
186
Path dependence in technical change
Ginsburgh, V.A. and P.-M. Menger (eds) (1996), Economics of the Arts: Selected Essays, Amsterdam, North-Holland. Gombrich, E.H. (1963a), ‘The vogue of abstract art’, in E.H. Gombrich (ed.), Meditations on a Hobby Morse and Other Essays on the Theory of Art, London and New York, Phaidon Press. Gombrich, E.H. (1963b), ‘André Malraux and the crisis of expressionism’, in E.H. Gombrich (ed.), Meditations on a Hobby Morse and Other Essays on the Theory of Art, London and New York, Phaidon Press. Gorman, W.M. (1980), ‘A possible procedure for analysing quality differentials in the egg market’, Review of Economic Studies, 47 (5), 843–56. Grampp, W.D. (1989), Pricing the Priceless: Art, Artists and Economics, New York, Basic Books. Guerzoni, G. (1995), ‘Reflections on historical series of art prices: Reitlinger’s data revisited’, Journal of Cultural Economics, 19, 251–60. Ironmonger, D.S. (1972), New Commodities and Consumer Behaviour, Cambridge, Cambridge University Press. Lancaster, K.J. (1971), Consumer Demand: A New Approach, New York, Columbia University Press. Lasdon, L.S. and A.D. Waren (1978), ‘Generalized reduced gradient software for linearly and nonlinearly constrained problems’, in H.J. Greenberg (ed.), Design and Implementation of Optimization Software, Leiden, the Netherlands, Sitjhoff and Noordhoff. Lasdon, L.S. and A.D. Waren (1981), ‘GRG2 – an all FORTRAN general purpose nonlinear optimizer’, ACM SIGMAP Bulletin, 30, 10–11. Lasdon, L.S., A.D. Waren, A. Jain and M. Ratner (1978), ‘Design and testing of a generalized reduced gradient code for nonlinear programming’, ACM Transactions on Mathematical Software, 4 (1), 34–50. Malraux, A. (1954), The Voices of Silence, London, Secker and Warburg. Marcus, G.H. (1998), Design in the Fifties, Munich and New York, Prestel-Verlag. McCain, R.A. (1981), ‘Cultivation of taste, catastrophe theory, and the demand for works of art’, American Economic Review, 71, 332–4. McDermott, C. (1992), Essential Design, London, Bloomsbury. McPherson, M.S. (1987), ‘Changes in tastes’, in J. Eatwell, M. Milgate and P. Newman (eds), The New Palgrave: A Dictionary of Economics, London, Macmillan, pp. 401–03. Mitchell, B.R. (1980), European Historical Statistics 1750–1975, second revised edition, London: Macmillan. Mitchell, B.R. with P. Deane (1971), Abstract of British Historical Statistics, Department of Applied Economics Monograph No. 17, Cambridge, Cambridge University Press. Reitlinger, G. (1961), The Economics of Taste: The Rise and Fall of Picture Prices, 1760–1960, London, Barrie and Rockliff. Reitlinger, G. (1970), The Economics of Taste, Volume III: The Art Market in the 1960s, London, Barrie and Jenkins. Rheims, M. (1959), La Vie Etrange des Objets, published in English translation by D. Pryce-Jones (1961), Art on the Market, London, Weidenfeld and Nicholson. Ruskin, J. (1843), Modern Painters: Volume I, reprinted in E.T. Cook and A. Wedderburn (1996), The Works of John Ruskin, Library edition on CD-ROM, Cambridge, Cambridge University Press.
Mapping changes in the taste for art
187
Ruskin, J. (1856), Modern Painters: Volume III, reprinted in E.T. Cook and A. Wedderburn (1996), The Works of John Ruskin, Library edition on CD-ROM, Cambridge, Cambridge University Press. Swann, G.M.P. (1999), ‘Marshall’s consumer as an innovator’, in S. Dow and P. Earl (eds), Economic Organisation and Economic Knowledge: Essays in Honour of Brian Loasby, Cheltenham, UK and Lyme, USA, Edward Elgar. Swann, G.M.P. (2001a), ‘The demand for distinction and the evolution of the prestige car’, Journal of Evolutionary Economics, 11 (1), 59–75. Swann, G.M.P. (2001b), ‘From correlations to implicit spaces: some results for planes and spheres’, unpublished paper, Manchester Business School. Swann, G.M.P. (2002), ‘There’s more to the economics of consumption than utility maximization’, in A. McMeekin, K. Green, M. Tomlinson and V. Walsh (eds), Innovation by Demand: Interdisciplinary Approaches to the Study of Demand and its Role in Innovation, Manchester, Manchester University Press. Throsby, D. (2001), Economics and Culture, Cambridge, Cambridge University Press. Veblen, T. (1899), The Theory of the Leisure Class: An Economic Study of Institutions, New York: Macmillan. White, M.V. (1999), ‘Obscure objects of desire? Nineteenth-century British economists and the price(s) of “rare art” ’, in N. De Marchi and C.D.W. Goodwin (eds), Economic Engagements with Art, Durham, NC, and London, Duke University Press, pp. 57–84. Wildenstein, D. (1996), Monet Volumes I–IV, Köln: Taschen Verlag.
7. Waves and cycles: explorations in the pure theory of price for fine art* Robin Cowan 1
INTRODUCTION
The popularity of painters rises and falls. The Impressionists were scorned in the mid-1800s, became the mode a few decades later, largely disappeared from view early in the twentieth century and stormed back to popularity at the end of that century. Their history illustrates two interesting phenomena. First is the rise and fall in popularity of artists. Artists become the mode; their prices rise, and galleries, critics and the public alike praise them as expressing the spirit of the times. But later they fade from view and are replaced by others. Second is that artists often come and go as a group. While there may be the ultimate Impressionist (some say that Manet was ‘the finest painter of all the Impressionists’ [Janson and Janson, 1997: 722), as his popularity rises and falls so does that of other ‘similar’ artists. Whatever is driving the popularity of Manet, drives also the popularity of Monet and Renoir. This grouping together of painters acts as a type of conformity effect in art fashion – if Van Gogh is popular, artists whose work is similar to his in the appropriate respects will also be popular, whereas painters whose work is different from his will tend to be unpopular. If one can imagine painters located in a space, with their popularity represented as a distribution over that space, then the modal painter will be surrounded by similarly popular painters. The mode, though, moves through the space over time. There is another striking feature of the market for fine art. This is the presence of ‘avant-garde’ consumers. These are the fashion-setters who are unwilling to have yesterday’s heroes on their walls. This need not sound quite so snobbish. One explanation for the behaviour of the avant-garde is that their search for new painters is driven by their need to express new ideas and concerns. Currently popular art will express today’s (or yesterday’s) concerns and ideas, but will in all likelihood not express tomorrow’s. Avant-garde consumers perform the valuable role of finding the modes of expression for emerging concerns. While the conformity effect of painting 188
Explorations in the pure theory of price for fine art
189
‘schools’ creates inertia in fashion, the presence of avant-garde consumers creates motion. These consumers will create popularity for painters situated far from the modal painter in painting space. These two features of the art market suggest the possibility of waves in popularity. Groups of painters become popular together, though one may stand out, but their popularity fades as avant-garde consumers look for new painters. The choices of these fashion leaders are thereafter taken up by less forward-looking consumers, which contributes to a shift in overall demand within the market. This is the phenomenon at the centre of this chapter. Our concern here is to model it in a way that sheds light on the patterns that can emerge from this interplay of conformity and distinction. The world of fine art is not the only one in which waves of fashion come and go. While the consumers of textbook theory have preferences independent of each other, this seems restrictive when considering the wide variety of both behaviour and type within the population of agents referred to as consumers. And indeed, a more general view, namely, that there are externalities in consumption or that the utility of an agent will depend in part on what other particular agents are consuming (whether there are physical spillovers such as pollution or not), is not new in economics. It stretches back to Smith, who claimed that the ‘the chief enjoyment of riches consists in the parade of riches’. Recently, the idea has been taken up by Becker (1996) and Akerlof (1997) each in his own way. One key aspect of externality in consumption, and the one emphasised by sociologists, is distinction.1 The idea here is that individuals gain utility from what they perceive to be their relative status in some hierarchy, and that one way to express, or even to change that status is through consumption. Consumption of some things will raise our relative status, and others will lower it. Frank (1985) has described some of the economic effects of the desire for distinction. Desire for distinction may be a powerful motivator but, nonetheless, for most agents it is important to function within a peer group – after all, from whom would we like to distinguish ourselves but our ‘former peers’ – so it is important that some activities create conformity with at least some part of the population around us. Thus within any utility function we would expect to see two (possibly occasionally conflicting) forms of non-independence: desire for distinction and desire for conformity. The effects of conformity and distinction have been addressed in a general way by Cowan et al. (1997). That model forms the background for much of the work presented in this chapter. Curiously, given the occasion for which this chapter was written, the model developed here shows no path dependence. It is a dynamic model, but the dynamics are deterministic, and the final, very long-run outcome is
190
Path dependence in technical change
predictable from the start. One place it does pick up ideas from Paul David (ignoring the general idea that some things are better than others) is in the notion of bandwagons, heroes and herds. Bandwagons of taste form, leaving yesterday’s hero a mere member, or at best icon, of today’s herd. How this comes about is through a form of standardisation – ideas and their expression become standardised. A new idea or new form of expressing something important emerges, to the scorn of the majority of the population. It is picked up by those who can see the coming thing, and pushed by them, they perhaps acting as translators, or gateways, between the new unusual idea and the old standard. Slowly, the herd come to see its value, it is absorbed (somewhat twisted perhaps to fit) by the mass of the population and, before you know it, everyone is doing it. But what of the innovator? Is he or she doomed to a future of anonymity? Not if he or she is able to continue producing new, unusual ideas to address emerging issues. If history is anything to go by, and especially if it matters, this is not something that need concern us on this occasion. We turn now to a very stylised model of the art world, in which prices are determined through the utility that paintings (or, more properly, the works of painters) provide the consumer. This utility is driven in part by externalities and, if snobbery is an unappealing trait on which to build a model, externalities can be seen as an expression of the notion that some art forms or genres are better fitted to the concerns of the day. Avant-garde consumers are looking at tomorrow’s ideas, and they pull the market with them in their search for new means of expression or needs to express new ideas. From these simple externalities we can derive rich dynamic patterns in which prices rise and fall as painters come into and go out of fashion.
2
A SIMPLE MODEL OF ART PRICE DYNAMICS
In this model we examine demand for painters’ outputs. That is, we consider a painter to be a brand name, associated with which there is a fixed supply of the commodity. We acknowledge the fact that painters produce nothing after they have died, but assume also that none of their works disappears. We consider the median consumer as representative of the market demand. While this is acknowledged to be problematic in some cases (see Kirman, 1992) this problem is alleviated somewhat in the heuristic wherein we treat the consumer’s demand as distributed over painters. This can be seen as capturing the notion that the median consumer represents a population of heterogeneous consumers whose demand will be distributed. The model is developed in the usual way as a consumer maximisation problem. The assumption of a fixed supply of every good immediately produces the
Explorations in the pure theory of price for fine art
191
equilibrium quantities consumed. What is of interest, however, rather than quantity is price. Prices change to equilibrate the market for each painter. This is not the trivial problem it seems, due to the presence of externalities in consumption – the popularity of a painter, indicated by a high price for his works, will affect the utility gained from consumption of other like, and unlike, painters. In addition, price adjustment is temporal in the sense that, because not every painting is auctioned every period (neither in the model nor in the actual art market), price adjustments are necessarily partial. There are two types of goods – paintings and other goods. Other goods are aggregated into the good Z, whereas paintings remain branded by painter. Define A as the set of painters: A {a [1, N]}. The consumer’s utility function is additively separable, written as U(X1,X2,. . ., Z)
&AUa(Xa) Z where Z is the aggregate bundle of other goods, and Xa represents consumption of the paintings of artist a. Normalise by setting the price of Z to 1. Utility maximisation under a fixed budget yields a first-order condition: dUdXa
dUb dXa p(a). bA
where p(a) is the price of the work of artist a.2 Separating a from the other artists: p(a) dUa dXa
dUb dXa.
(7.1)
ba
The first term is simply the marginal utility of consuming paintings by painter a; the sum represents the effects of consumption of painter b on the utility gained from consuming painter a. Assume now that d2Ua/dXbdXc
0∀c{a, b}. If the number of painters is large, this permits us to approximate equation (7.1) as p(a) g(a)
F(b, a)db,
(7.2)
bA
where g(a) dUa/dXa|X. We can suppress the Xa argument since by assumption there is a fixed number of paintings per painter, X, and the market clears through price adjustments. The term g(a) can be seen as representing the inherent value of artist a, regardless of his current (un)popularity. Some artists are just simply better than others. The integral contains externality effects: consumption of other painters affects the marginal utility of the consumption of a. We decompose F(a, b) into a product: 1/f(b a)p(b). The first element represents the strength of the externality as determined by the distance between the two painters in question – made up of the conformity and avant-garde effects; the second
192
Path dependence in technical change
is the standing price of the second painter, which is a measure of the extent to which he is consumed, or how popular he is (recalling that supply is fixed). To ease the presentation that follows we do a small violence to the notation and will treat b as indicating ‘the painter at distance b a from a’. This allows us to write f(.) as a function of b alone: f(b). Equation (7.2) describes the equilibrium price vector for the set of artists A.3 The nature of the art market, however, makes the disequilibrium process extremely important. Many out of equilibrium trades take place, in part because for any painter, only a small proportion of his work is traded in any period. This is the nature of the art auction market. To capture this effect we assume that the work of every artist is trade at the rate . That is, speaking in discrete time, each period, '100 per cent of every artist’s works come up for auction. The price of paintings actually traded is set by equation (7.2). Consider now the ‘standing’ price for work by a particular artist, a. Exposition is more transparent in discrete time: p(a, t 1) (1 )p(a, t) g(a)
f(b)p(b)db,
(7.3)
bA
or, p(a, t 1) p(a,t) p(a, t) g (a)
f(b)p(b)db.
(7.4)
bA
Writing now in continuous time, dp(a)
g(a) p(a, t) dt
f (b)p(b)db.
bA
This dynamic structure has been studied in other contexts (see for example Cowan et al., 1997). A solution is generated by doing a Laplace transform on the time variable, and a Fourier transform on the a variable (see the appendix in Cowan et al., 1997). Drawing on that work we can state three propositions: Proposition 1 the limit:
If the system is convergent, the steady state is described by lim P (k,t)
t→
G(k) F(k)
e ikap(a, t) da, F(k)
where we use the definitions P(k,t) e ikbf(b)db and G(k) e ikag(a)da. This proposition states that the ‘natural prices’, namely, those that would prevail if behaviour were based solely on the inherent worth of the artist
Explorations in the pure theory of price for fine art
193
and not on any considerations of externalities, are, in equilibrium, ‘distorted’ by any externalities that exist. We can retrieve the dependence of p on a by the following transforma 1(2$)eikaP(k, t) dk. In the absence of specified functions, tion: p(a, t) the form of p(a, t) is less easy to interpret than P(k, t). Consider an arbitrary painter a. P(k, t) measures the extent to which a painter of distance $/k from painter a affects the price (or equivalently the popularity) of painter a. Thus large values of k are associated with effects of nearby painters, small values of k are associated with effects of distant painters. The next two propositions concern the dynamics of this market, and whether waves of popularity will be observed. Proposition 2 If f(b) is an even function ( f( b) f(b)) then the dynamics are strictly diffusive. That is, the initial state decays and the final equilibrium state builds up exponentially. Proposition 3 If f(b) is an odd function ( f( b) f(b)) then the dynamic behaviour is captured by travelling waves. These two propositions follow from basic properties of Fourier transformations. For a discussion of these results see Cowan et al. (1997). Remark 1 Any well-defined function can be expressed uniquely as the sum of even and odd functions.4 When f(b) contains both even and odd elements non-trivially, price dynamics will be a sum of the dynamics described in propositions 2 and 3. The quantitative features will depend on the detailed forms of the even and odd parts of f(b).
3
PURE FAD
There is almost certainly a strong ‘fad’ component to taste in art. That is, the externalities represented by the sum in the utility function (equation [7.1]) are a major source of utility from art consumption. We can consider two extreme cases. In the first case, art per se has no inherent value to the consumer. Formally, dUa /dXa g(a) 0. In the second case, there is inherent value to the consumption of art, but no painter is better or worse than any other as a direct source of utility: for all a, g(a) c where c is a constant. Proposition 4 If g(a) 0 then if the system is convergent the limiting price is p(a) 0 for all a. From Proposition 1, the limiting price function is described by the limit of its transform limt→P(k, t) G(k) ( F(k)). If g(a) 0 then G(k) 0
194
Path dependence in technical change
by definition of the transform. Thus the inverse transform of P(k, t) is also equal to 0: limt→ p(a, t) 0 for all a. Proposition 5
If g(a) c then if the system is convergent lim p(a, t)
t→
c . f (b)db
If g(a) c then G(k) 2c$(k) where (k) is the Dirac delta function. Substituting into the limit from Proposition 1: lim P(k, t) 2c$(k) ( F(k)).
t→
Since (k) 0 (k 0, this reduces to lim P(k, t)
t→
2c$ 1 . 2$ F(0)
e ikbf (b)db so F(0) f (b)db, and substitution By definition, F(k) yields the propostion.
Corollary If g(a) c and (4c) F(b)db then there is a positive price p* such that p(a) p* (a A. This can be shown be checking consistency. Substitute p* for p(a): p*
c . F(b)p*db
1p*
p* F(b)db . c
Define q 1/p*. Then
qc q 1 F(b)db 0, which has real roots (and therefore at least one positive root), if and only if /(4c) F(b)db. Propositions 4 and 5 give insight into the stability of demand over long periods. The fact that prices have not gone to zero, even for painters long since dead, suggests that there is some inherent utility to be gained from consuming art. On the other hand, the fact that there is variation in the prices fetched by different painters implies one (or both) of two things:
Explorations in the pure theory of price for fine art
195
either not all painters are equal in the eyes of the median consumer, or there are strong ‘fad’ elements in the consumption of art works, and the market continues to exhibit out-of-equilibrium behaviour. We have not analysed the case in which some painters are ‘better’ than others – analytical results are extremely difficult to generate. This is a situation we explore below, however, by simulating the model.
4
CONVERGENCE
One concern that always exists with dynamic systems is whether or not they coverge. Steady-state results given above were conditional on the system converging. In this section we explore the conditions for convergence. From the solution in the appendix in Cowan et al. (1997), in terms of the conjugate variables k and z the dynamics of the system are defined by P(k,t)
G(a) G(a) ( F(k))P(k,0) ( F(k))t e . F(k) F(k)
(Ibid.A4)
Thus convergence turns on the final term: e ( F(k)). If, for some k, F(k) 0 then the system is divergent. Recall that F(k) is the transform of e ikbf (b)db the externalities in consumption, defined as F(k) where f(b) is the sum of the conformity and avant-garde effects. Convergence is determined in a critical way by the functional forms of these externalities. Assume that the conformity effect is even and the avant-garde effect is odd. We can then write the total, net externality effect as f(b) fc(b) fa(b) where fc( b) fc(b), and fa( b) fa(b). In passing we can note that from Remark 1 the dynamics in this case will exhibit both waves in prices and a secular trend toward the final price distribution. Fourier transforms are e ikbf (b)db e ikbf (b)db. linear, so F(k) Because fc(b) is even and c a 2 cos(bk) f (b)db 2i sin f (b)db. fa(b) is odd, this simplifies to F(k) c a To illustrate, suppose that the two externality effects are each a member of families of functions: fc(b) C1 f1(b) and fa(b) C2 f2(b). In this case the transforms become C1F1(k) and C2 F2(k). Substituting the functional family description of fc and fa we get
F(k) C1
2 cos(bk) f1(b)db C2
2i sin f2(b)db.
The Dirichlet conditions (which have been assumed to hold) ensure that the integrals are finite. Thus C1 and C2 are scaling parameters that will, jointly with , determine the sign of F(k). The stronger are the externalities, the faster the system diverges; similarly the smaller is , that is, the smaller
196
Path dependence in technical change
the proportion of works that come onto the market each period, the faster the divergence. If externalities are weak enough, or enough works come on to the market each period, then the system converges.
5
SIMULATION OF THE MODEL
To illustrate some of the results, showing both dynamics and long-run properties, we simulate the model of art prices developed above. An initial problem is the space in which painters exist. In a Lancasterian world, paintings obviously exist in a very high dimensional space – they have many types of qualities that can vary from one painter to another. To what extent the dimension of the space can be reduced without doing extreme violence is as yet unknown. Peter Swann explores this issue in his contribution to these proceedings. He finds that it is indeed possible to reduce the dimension dramatically. Exactly how far this can be done, and the nature of the space itself (linear, periodic, spherical or toroidal and so on), is something he is currently exploring I believe. Because these details, which are essentially empirical matters, are still being explored, we can use this opportunity to implement the model in a variety of spaces. We create a world of 600 painters, each at a fixed location in space. Proximity in this space indicates that painters (or more properly their outputs) are ‘like’ each other in the eyes of their beholders. Consumers gain utility from consuming artworks, and there are externalities in that consumption. Painters, and schools of painters become fashionable, that is, a consumer will gain utility from owning the works of a painter who is like other painters who are widely appreciated. This is the conformity effect, though it could be described as the effect of having concerns and ideas in common with other consumers. This we model as an even function. On the other hand, there is an avant-garde effect. There are some consumers who are ahead of the pack in terms of ideas and concerns, and are thus looking for ‘new’ painters to express those concerns. In general, this effect is to seek ‘unfashionable’ parts of the space in which painters reside. There is, though, a difference between yesterday and tomorrow, so this is modelled as an odd function. It is this effect that gives direction (rather than just change) to art prices. 5.1
One Dimension
The analytic results developed above were in the context of a one dimensional linear space. They apply directly to a (one-dimensional) periodic space, making allowances for the fact that in this space if a wave travels
197
Explorations in the pure theory of price for fine art
forever in one direction, it is actually travelling around a circle. One caveat is that the externality functions must go to zero (in distance) at a distance less than half the circumference of the circle. This seems entirely reasonable in this case. For completeness we explore both linear and periodic spaces. 5.1.1 Linear space Six hundred painters are arrayed along the whole line. Every period 1 per cent of each painters paintings comes up for auction. Prices of those paintings are set as in equation (7.2). Two types of externalities exist: the conformity effect is modelled as f1(b) c/|b|. The avant-garde effect is modelled as f2(b) sgn (b)a(1 1/|b|). The dynamic pattern can be seen in Figure 7.1. This figure should be read as a relief map. Darker colours indicate higher prices, and thus higher popularity. What can be observed here is a main wave of popularity. Initially the painter located at 550 on the horizontal
300
250
Time
200
150
100
50
0 0 Figure 7.1
100
200
300 Painters
400
Prices of painters over time: linear space
500
600
198
Path dependence in technical change
axis is the painter of the day. His popularity fades, though, and popularity moves to painters to his left. This is driven by the avant-garde effect. What is noteworthy, however, is that painters ‘reappear’. There is a second and third wave – an echo if you will, in which prices rise again, to be followed by a decline. What is interesting here is that this occurs without the introduction of new painters, which is clearly one source of new waves in the actual world of art and artists. Here the simple dynamics are alone enough to produce a main wave and subsequent resurgence of formerly popular painters. 5.1.2 Periodic space The interpretation of the dimensions of ‘painter-space’ is unclear. Fine art has a variety of attributes; precisely which of them should appear as major axes is unclear. Further, some of these properties seem naturally modelled as extending indefinitely, while others seem naturally periodic (consider a colour wheel for example). In this section we implement the same simulation but in a periodic space. Painters are arrayed around a circle rather than on a line. This is consistent with some of the empirical results of Peter Swann. The periodicity of the space permits a painter to reappear if there are waves in consumption. A wave travels in one direction, and eventually returns to its original position, travelling round and round. Again, cycles emerge without the introduction of new painters. Figure 7.2 shows a typical pattern, using the same representation as Figure 7.1. Darker grey indicates higher popularity (and prices). Notice that with these parameters prices are secularly increasing, as the waves shown by the diagonal patches get darker and darker as time passes. 5.1.3 Inherently good painters In the previous illustrations of the model, no painter was any better, inherently, than any other. Thus the dynamics were driven purely by externalities or fashion effects. It may be, though, that some painters or schools are inherently better than others in providing consumers with higher utility regardless of fashion. The analytic results included this aspect; here we illustrate it. Figure 7.3 shows the same dynamics as Figure 7.2 (a periodic space with secularly increasing prices) but with the painter located at position 100 having inherent value. His inherent value spreads a short distance to include those nearby above and below.5 As can be seen in the figure, the gradual darkening of the graph as time passes indicates general increases in prices. Repeated waves occur – the diagonal stripes running northwest/south-east. We can see though, that relative to Figure 7.2, those waves are distorted by the price of painter 100. His inherent, persistent value creates a vacuum of low prices to his right, caused by the avant-garde
199
Explorations in the pure theory of price for fine art
1600 1400 1200
Time
1000 800 600 400 200 0 0 Figure 7.2
100
200
300 Painters
400
500
600
Prices of painters over time: periodic space
shunning him and those like him. (Reading horizontally from left to right just to the right of 100 the graph becomes lighter ‘than it should be’ relative to the overriding wave pattern). This effect gains in extent in that as time passes more and more painters are affected. This can be seen by the changing shape of the wave – the stripes are not of uniform width, and have odd patterns emerging between the waves. Similar effects exist in the one-dimensional linear space. 5.2
Two Dimensions
Reducing the dimension to one is a dramatic simplification of the space in which fine art must exist. Nonetheless, it does lend insights into cycles in pricing. A higher dimension space adds realism, but raises an in principle problem: how does the avant-garde effect operate? In a single dimension
200
Path dependence in technical change
1600 1400 1200
Time
1000 800 600 400 200 0 0
100
200
300 Painters
400
500
600
Figure 7.3 Prices of painters over time: periodic space with one inherently valuable painter this was relatively simple, in that the temporal aspect of avant-garde-ism had to be equated with a spatial aspect, and this determined the direction of motion. In higher dimensions, however, at any time the avant-garde effect operates, it could be pulling the art world in more than one direction, depending on what the avant-garde consumer (or artist) is attempting to express. To assume at the outset that this motion always occurs in the same direction (the avant-garde effect favours painters to the right, as in the one-dimensional case, for example) will simply reproduce the effects of the one-dimensional model, the other dimensions not interacting very much with the motion. A more interesting, and probably more realistic, notion is that from time to time one point, currently ‘ignored’ in the space becomes very popular. The avant-garde effect would be to pull popularity from the current mode towards and beyond this newly popular artist. This implies
Explorations in the pure theory of price for fine art
201
that popularity would follow a more random path around the space, but tending to move from ‘old’ to ‘new’ ideas. Ignoring the possibility that new painters enter, thereby disturbing exiting market structures and dynamics, the same dynamic patterns observed in the single-dimension cases are observed in two dimensions, whether the space is a plane or a torus. There are waves of popularity as painters’ prices rise, fall, and then rise again as the fashion is pulled by the avant-garde away from the currently popular painters. Using the same framework it is possible to introduce entirely new fashions, by creating a small island of high prices in a part of the space that is currently out of fashion. Doing this in an ad hoc way indicates that the motion of prices is as expected. The island of high prices creates immediate conformity and distinction effects, and changes the direction in which art fashion was moving. It is difficult to perform this experiment in a systematic way (and it is equally difficult to present the results graphically and concisely), so we leave that discussion at this point as part of the agenda for future work. 5.3
Planes and Circles
One of the fascinating results of Peter Swann’s contribution to these proceedings is that even when painters are constrained to exist in a two- or three-dimensional space, neither the planar nor the spherical projection is full. In fact, in both cases the painters are (statistically) located around a circle. This suggests that the one-dimensional periodic space may be a good representation in which to analyse the dynamics of prices or fashion in the fine art world. Swann’s analytic results show that in this representation correlations of prices are proportional to the cosine of the angle between the painters. Thus a test of the suitability of modelling a high dimensional space as a circle is whether the dynamics of the model can be parametrised in such a way that the price correlations satisfy the cosine relationship. Figure 7.4 is a scatter plot of price correlations versus the cosine of the angle between painters on the circle. In this implementation there are 600 painters, avant-garde and conformity effects that increase with distance, but are truncated at a distance of 25.6 As is clear in Figure 7.4, this parametrisation comes very close to satisfying the cosine relationship exactly. It departs from this relationship when the cosine is near 1, that is, between painters who are located on opposite sides of the circle. There are three explanatory factors: (1) here the externality effects, which are key to determining price relationships, are weakest; (2) in general the model places no bounds on prices, but for reasons of realism prices have been bounded below by zero, which will interfere with the inter-painter price relationships, and will ‘distort’ this relationship most for distant painters; and (3) under
202
Path dependence in technical change
1
Price correlation between painters
0.8 0.6 0.4 0.2 0 0.2 0.4 0.6 0.8 1 0.8 0.6 0.4 0.2 0 0.2 0.4 0.6 Cosine between painters on the circle
0.8
1
Figure 7.4 Correlations between prices versus cosine of the angle between painters the parameters used in this implementation there is a secular increase in prices which is a source of positive correlation among prices of all painters, which pulls the correlation ‘up’ most where the pattern created by externalities is weakest.
6
DISCUSSION
The rise and fall of painters’ popularity is striking over a period of several hundred years. This is a phenomenon that has been felt intuitively by many observers of modern culture, but is also evident from the hard economic measure, namely, prices. It is a common place that the art world goes through fads and fashions, so a tempting explanation for the rise and fall of
Explorations in the pure theory of price for fine art
203
a painter is something like sunspots. But to explain the rise and fall of groups of painters implies that there are qualitative similarities between the members of any group. This suggests that something more stable is going on (unless one simply invokes sunspots that operate on groups rather than on individuals, but this begs the question why members of a group fetch different prices from each other). An economist is tempted to point to stable preferences at least in explaining the grouping phenomenon. But stable preferences can be used for a more general explanation, provided one is willing to abandon inter-agent independence in preferences. That has been the approach here. Stable preference in a stable world are enough to generate waves in which painters emerge, disappear and emerge again. And a very simple model has produced a rich set of dynamics which can be treated analytically and numerically to help understand the dynamics of the art market. One of the observations made by Swann regarding the price time series is that there appear to be cycles of different period. This can in principle be reproduced in this simple model, by introducing more complex externality functions. Every function used here was monotonic. But a non-monotonic function, with more than one maximum or minimum will produce strong reactions at wherever there is an optimum. If there are optima at several distances, this implies that waves of different frequencies will form. The stability of the world of the model, in the sense that no new painters are suddenly appearing to upset existing dynamic patterns, is both a strength and a weakness. It shows the power of externalities in creating interesting dynamics, but it departs somewhat from reality. Introducing new painters in a non ad hoc way remains a research challenge.
NOTES *
This chapter was written for the conference New Frontiers in the Economics of Innovation and New Technology, held in honour of Paul David at the Accademia delle Scienze, Torino, 20–21 May 2000. The chapter has benefited from the comments of the participants in that conference and especially from ongoing discussions with William Cowan and Peter Swann. Funding from the Social Sciences and Humanities Research Council of Canada is gratefully acknowledged. 1. See most notably Bourdieu (1984). 2. Because we are interested in the evolution of prices of all the painters, rather than using the common pa notation we treat price as a function of the painter, p(a). 3. The price on the left-hand side of equation (7.2) is the ‘standing price’, that is, the average prices most recently paid for the entire oeuvre of an artist. This introduces some history into the notion of popularity, which seems appropriate in this context. The market for fine art, at least for the painters that survive the test of time is indeed faddish to some degree, but nonetheless it has a strong sense of value and history. History is gradually eliminated (and replaced) however, as the entire work of an artist comes to auction over time, and must face competition from other artists and in other eras.
204
Path dependence in technical change
4. Formally, any function satisfying the Dirichlet conditions can be so written. The Dirichlet conditions are that the function is square integrable, single-valued, piece-wise continuous and bounded above and below. 5. Specifically there is a scaled normal distribution with variance 10, and mean 100 describing inherent value. 6. Because the circle is translation invariant, many pairs of painters are separated by the same angle – painters 0 and 10 are separated by the same angle as 1 and 11, and so on. The data in the figure are not averages for an angle but rather are the correlations between each pair of painters plotted against the cosine of the angle between them.
BIBLIOGRAPHY Akerlof, G. (1997), ‘Social distance and social decisions’, Econometrica, 65 (5), 1005–27. Becker, G.S. (1996), Accounting for Tastes, Cambridge, MA: Harvard University Press. Bourdieu, P. (1984), Distinction: A Social Critique of the Judgement of Taste, London, Routledge and Kegan Paul. Cowan, Robin, William Cowan and Peter Swann, (1997), ‘A model of demand with interaction among consumers’, International Journal of Industrial Organisation, 15, 711–32. David, P.A. (1985), ‘CLIO and the economics of QWERTY’, American Economic Review, 75, 332–36. David, P.A. (1992), ‘Heroes, herds and hysteresis in technological history’, Industrial and Corporate Change, 1 (1), 129–80. David, P.A. (1997), ‘The economics of path dependence in industrial organization’, International Journal of Industrial Organisation, 15, 643–52. Frank, R. (1985), Choosing the Right Pond: Human Behavior and the Quest for Status, New York, Oxford University Press. Janson, H.W. and Anthony F. Janson, (1997), History of Art, 5th edition, New York, Prentice-Hall. Kirman, A. (1992), ‘Whom or what does the representative individual represent?’, Journal of Economic Perspectives, 6 (2), 117–36. Smith, A. (1937), The Wealth of Nations, New York, Random House. Swann, G.M.P. (2000), ‘Is the world flat or round? Mapping changes in the taste for art’, MERIT-Infonomics Research Memorandum series #2001–009.
PART III
The Economics of Knowledge
8. Learning in the knowledge-based economy: the future as viewed from the past W. Edward Steinmueller The pervasiveness of economies of scale opens up the prospect that past market configurations, which neoclassical theory tempts one to interpret as globally stable equilibria, were in reality unstable positions away from which the system moved when disturbed by shifts in demand. The added presence of ‘learning’ effects in production (and the implied suggestion that they may have also been present in consumption, in the form of habituation or endogenous taste reformation) introduces a degree of irreversibility in the ensuing market adjustments of relative costs and prices. As a result of this previous economic configurations become irrevocably lost, and in trying to work backwards by entertaining counterfactual variations on the present, one cannot hope to exhibit the workings of historical process. [. . .] Under such conditions, market divergences between ultimate outcomes may flow from seemingly negligible differences in remote beginnings. There is no reason to suppose that dynamic processes are ergodic, in the sense of ultimately shaking free of hysteresis effects and converging from dispersed initial positions towards a pre-determined steady state. To understand the process of modern economic growth and technological development in such an untidy world necessarily calls for the study of history. For, change itself ceases to be mere locomotion. Economic growth takes on an essentially historical character, and the shape of the future may be presumed to bear the heavy impress of the past. (David, 1975: 15–16)
1
INTRODUCTION
Since this early contribution of Paul A. David, many economists, from a variety of perspectives and using a variety of methodologies, have come to share the viewpoint that economic change involves processes of organisational and individual learning. This chapter examines the relationship between individual and organisational learning as it bears upon the accumulation of knowledge (the result of learning) from a historical perspective and the recent opportunities afforded by information and communication
207
208
The economics of knowledge
technologies for changes in the ‘technology of learning’ (the means for reproducing and exchanging knowledge). 1.1
A Closer Look at Learning
Marking a distinction between individual and organisational learning is a necessary step in explaining one of the most persistent and troubling of economic problems, the uneven and uncertain transfer of technological capability, what is commonly called ‘technology transfer’. The two fundamental premises underlying technology transfer, that the acquisition of capability is related to the acquisition of knowledge and that knowledge is acquired through learning processes, seem simple enough. The problem is that, between any two individuals, learning processes are likely to produce different knowledge outcomes and hence different capabilities. These differences are magnified by distance (cultural, social, physical) between the individuals and compounded when a group of receiving individuals attempts to rearticulate the knowledge in an organisation. Within an organisation, knowledge can only achieve and retain value through processes of exchange that result in the construction of common understandings which, in turn, provide the basis for consistent and predictable actions.1 The result is that there is rarely a clear map or plan available for specifying how individual learning processes can be translated into effective or usable organisational capabilities. In other words, in an organisation, individual knowledge is unlikely to ‘self-assemble’ into organisational capability. Instead, the construction of organisational capabilities from individual knowledge is likely to require processes of iteration and interaction that involve constructing a ‘common ground’ between the individuals engaged in purposive activities. This requirement of constructing a ‘common ground’ of knowledge within organisations is one of the reasons why the knowledge of organisations is not the same as that of individuals. Taking account of this requirement,2 along with the costs of transactions and principal-agent incentive problems, offers a basic toolkit for explaining the division of labour within and between organisations and a simple or first-order explanation for the difficulty of the technology transfer problem. The recognition that knowledge must be articulated within an organisation to become an effective capability leads to a more specific theory of what economists refer to as economies of learning.3 In the usual economic treatment of learning economies, the practice is to take instrumental variables reflecting the accumulation of experience, such as the accumulated scale of output or the passage of time, as an explanatory variable for cost reduction. If, instead, we focus on the ‘technology of learning,’ the means by which
Learning in the knowledge-based economy
209
organisations ‘construct’ knowledge by building a common ground of understanding based on individual learning processes, a new set of questions about the knowledge acquisition and exchange processes emerge. First, it is possible to ask whether, in a competitive selection environment, the knowledge of two different organisations may differ in significant ways. In other words, can different ‘stocks’ of knowledge have the same competitive fitness? The answer to this question clearly depends upon a technological context. The modern-day ‘low-tech’ cement or brick industries seem to offer less opportunity for distinct accumulations of knowledge than the ‘high-tech’ pharmaceutical, integrated circuit or software industries. It is a common belief that these differences are the results of ‘technological opportunity’; that is, that greater opportunities for product design and discovery are available in high technology industries. It is also true, however, that these differences arise from the character of demand.4 The demand for variety in cement or brick design is far more limited than the demand for variety in ‘high-technology’ industry products. In essence, knowledge accumulation is more valuable in ‘high-technology’ industries because of the interest of customers in differentiated product offerings. If tomorrow, everyone were to wake up with a determination to have dwellings and workplaces that were truly unique, the relative value of knowledge accumulation in the brick and cement industries would increase and, with some delay, so would the pace of innovation in these industries. For the purposes of this chapter, what is significant is that changes in knowledge accumulation initiated by the new characteristics of demand would likely lead to greater diversity in what is known by cement and brick producers. Second, the nature of competitive advantage in an industry is pertinent. For variety in knowledge accumulation between firms to emerge, it is necessary that the sources of advantage be pluralistic. Advantages in one area must be offset by weaknesses in others. For example, if a single parameter determines the technological trajectory of a particular industry, the uneven results of knowledge accumulation (learning) would produce leaders and laggards. Leading and lagging firms could only coexist to the extent that the technological advantage of the leader(s) was not fully translated into market advantage due to the possibility of non-technological advantages such as marketing capability or market location. Similarly, if the industry’s technological trajectory is characterised by many technological parameters, there may be many coexisting ‘leaders’, none of which has a clear technological advantage over the others. In this case, limitations to the ‘scope’ of technological mastery will limit the dominance of individual firms.5 These first two issues suggest that diversity in ‘what’ is known by firms is a consequence of the characteristics of demand, of technological
210
The economics of knowledge
opportunity, and of specific sources of technological advantage that shape the accumulation of knowledge. Learning provides a basis for the emergence of diversity and, perhaps, disparity in firm capabilities. Where technological disparities exist and some firms are at a technological disadvantage, these firms must find offsetting and non-technological advantages in order to remain in the market. 1.2
The Circulation of Knowledge within the Organisation
The third observation to be made, and the focus of this chapter, is on the relation between the sources of diversity (or disparity) of firm capabilities and the processes by which knowledge is held and circulated within and between firms.6 The possibilities for knowledge circulation may be highly constrained. That is, knowledge may only be accumulated ‘locally’ in close proximity to the work processes and within very tightly bound social groupings of individual workers. Alternatively, the knowledge underlying a firm’s capabilities may circulate widely and with few constraints, providing each of the employees of the firm with access to the knowledge underlying the firm’s capabilities. These may be taken as the two extreme points on an axis along which it is possible, in principle, to locate firms.7 Even if knowledge is highly localised in the sense of not easily circulating within or between organisations, it is not necessarily true that what is known by various organisations is fundamentally different. In principle, faced with similar environmental conditions and opportunities, individuals and organisations may independently accumulate similar types of knowledge. This is especially likely to occur, despite the proclivities of humans to be inventive, when the technological and demand conditions in an industry punish rather than reward creativity and invention. Moreover, if knowledge exchange between organisations occurs through processes of imitation, labour mobility or feedback from capital suppliers and customers, the knowledge of entire industries may follow convergent paths. Individual firms either learn or die and, if they survive, they eventually learn the same things. The above observations suggest two possible research agendas. The first is to characterise industries according the sources of their knowledge and the processes by which knowledge is exchanged and reproduced.8 The second is to examine how changes in market or technological conditions alter the processes of knowledge accumulation and exchange, and expose underlying differences in capabilities among firms. When such changes have similar effects on all firms in an industry it is likely that each of the firms is operating with similar knowledge. If the effects of environmental and technological change have an uneven or irregular effect on individual firms, it
Learning in the knowledge-based economy
211
is likely that firms are operating with different knowledge. Thus, market or technological changes may help to reveal differences in knowledge ‘holdings’ by firms within an industry. The next section examines market and technological changes in learning (and knowledge holding) within organisations from an historical perspective. This discussion sets the stage for the second half of this chapter, which explores some of the structural changes and transformations accompanying the application of information and communication technologies to the learning process within firms and industries.
2
A HISTORICAL PERSPECTIVE ON KNOWLEDGE AND LEARNING
There is no immediately obvious reason to believe that contemporary economic activities are inherently more knowledge-intensive than those of the past. Despite the high tide of rhetoric concerning ‘knowledge-based’ economies and industries, manufacturing and service activities have always involved knowledge inputs from skilled workers, accumulated productionrelated know-how, the design of products and services, and the configuration of production processes to achieve well-defined levels of quality and reliability. Indeed, the embodiment of this knowledge in individual workers in ways that were unarticulated outside a local context is likely to have provided more scope for variety and differentiation in nineteenth-century workplaces than in contemporary ones. Given a time machine, it would be tempting to transport and to strand (at least briefly) a few modern knowledge management experts in a steel, textile or grain mill of 1870. The survivors of this exercise would likely conclude that the processes they observed did encompass bodies of knowledge, albeit ones that would be largely unfamiliar even to those with some knowledge of the modern equivalents of these factories. The accumulation of knowledge is, in fact, situated within historical circumstance and it does not immediately follow that modern ‘stocks’ of knowledge are larger or more complex than those accumulated at previous points in history. A close interrogation of these visitors to the past would, however, reveal that the structures of knowledge that they saw being employed, involved a much greater degree of personal expertise of typical individuals directly engaged in production activities. The creation and sustenance of hierarchies in nineteenth-century industry often reflected the accumulation of experience over a lifetime, with individual workers closely guarding the maintenance and reproduction of this knowledge. Correspondingly, the transportation into the future of industrially knowledgeable individuals from 1870 to the present would provoke a
212
The economics of knowledge
symmetrical observation. They would observe that modern production operations were comparatively ‘de-skilled’ in their absence of individuals with specific knowledge of the tools and procedures being employed on the factory floor. This change in the locus of knowledge ‘holding’, rather than in the extent of employed knowledge is the fundamental distinction between the role of knowledge in contemporary industry and that of the late nineteenth century. In the hierarchy of the modern enterprise, production engineering management and operations are activities conducted behind the scenes and, for the most part, off the factory floor. While this mode of organisation is sometimes criticised as discarding valuable information, the trend to relocate these activities largely continues, as does the growth in recorded productivity in manufacturing of all industrialised countries employing this type of organisation.9 There must, therefore, be a reason for the learning activities that are responsible for the accumulation of knowledge having been ‘de-localised’ from direct involvement in factory operations. Examining these reasons provides a deeper understanding of the learning processes at the level of the organisation and at the level of the industry. 2.1 The De-localisation of Knowledge Accumulation in Nineteenth-Century American Manufacturing There appear to be two fundamental reasons for the de-localisation of knowledge in late nineteenth-century American manufacturing.10 The first is the progressive articulation of knowledge required to produce a largescale producer goods industry. The expertise required to produce state of the art producer goods was, by the late nineteenth century, already moving outside the factory.11 Producing complex machines such as textile looms or reapers follows the admonition of Smith, Young and Stigler that the division of labour is limited by the size of the market.12 Widespread markets for producer goods assured that the design and manufacture of these machines was increasingly remote from the site of production. Factories without machinists who would be capable, if necessary, of rebuilding the factory’s machinery became increasingly common and, with the end of these individuals’ employment, a vast body of specialised knowledge disappeared. The new site of knowledge accumulation became the capital goods producer and a new form of industrial co-ordination, the producer–user relationship, came to the fore. This change is important for the ensuing discussion. The second reason for the de-localisation of knowledge in production was the rise of ‘systemic’ approaches to the problems of increasing productivity and standardisation of industrial output. Disentangling the
Learning in the knowledge-based economy
213
origins and spread of these ideas has become a monumental quest and a proving ground for historians of technology. One reason for this is that ‘systemic’ approaches are tightly bound to understanding the origins of ‘modernism’ and for a critical appraisal of its consequences.13 For our purposes, however, the main point is established by Rosenberg’s (1976) history of the emergence of generic and multi-purpose machines to replace the specialised tools developed within specific industries (for example, locomotive, bicycles and sewing machines). These new machines could produce mechanical parts for a wide variety of industries and thus could be produced on a larger scale than the industry-specific machine tools. These developments produced a degree of interdependence between mechanical manufacturing industries, creating the possibility for knowledge spillovers as machinists and mechanics practising in one industry could master skills that were applicable for employment in another industry. The argument that these two forces, scale and interdependency, foster standardisation and the need for control is a particularly attractive hypothesis in explaining the increasing interest in mechanisation that emerged in the latter half of the nineteenth century. Several complementary hypotheses support this explanation. For example, Beniger (1986) links the origins of systematisation with the control requirements for safely maintaining the speed of railroads, a critical parameter in reaching higher levels of scale economies in a transportation network.14 Similarly, Hughes (1989) has argued that solving the ‘reverse salient’ problems introduced by new technologies required the development of a more pervasive ‘systems’ view of technology in a variety of sectors, most notably in the network industries. Hounshell (1984) contends that the idea of interchangeable parts was adopted in view of the impetus it would provide to scale manufacture, rather than that increasing scale led to the rationalisation represented by interchangeable parts. Systematisation produced a need for organisational structures within modern enterprises that could record, analyse and plan the extension and development of production. This obligated enterprises to restructure the ‘holding’ of knowledge from the shop floor or even from its individual custodians within engineering departments to the managerial and research functions of the enterprise. Only by achieving a global overview of production was it possible to strive for systemic improvements in the organisation and conduct of production, to apply ‘scientific methods’ based upon ordered trial, or quantitative analysis, to the operations of business. The ‘de-localisation’ of knowledge from the shop floor and its relocation into the managerial and research functions of the organisation provided the foundation for the mass production of Henry Ford and perhaps the most pivotal moment in the history of industrial learning processes. Ford’s vision
214
The economics of knowledge
of the future of mass production was the creation of standardised goods that all could afford. As Edward A. Filene argued, Fordism offered America economic salvation from the inter-war years growth problems, in supporting mass consumption by allowing an ‘increasing standardisation of products, and an increasing mechanisation of the process of production’.15 The shift in knowledge ‘holding’ in the USA from 1870 until the mid1920s involved supplanting learning and knowledge at level of the factory or shop floor with knowledge managed at the level of the organisation. The transformation from a ‘craft-based’ to a mass production system obliterated this localisation of knowledge and replaced it with large and complex production systems with many interdependent components and subprocesses. These systems served as the new locus of knowledge accumulation. The ‘canvas’ upon which learning processes were inscribed had grown. 2.2 An Archetypical Case of the Transformation in Knowledge Holding and Circulation Providing systematic evidence concerning the rate and direction of the transformation of knowledge holding and circulation in nineteenthcentury manufacturing is beyond the scope of this chapter. It is possible, however, to provide an archetypical narrative of this transformation in one organisation, the McCormick Reaper Works. The following account examines the struggle between Cyrus Hall and his brother Leander McCormick over the operation of the McCormick Reaper Works following its reconstruction after the Chicago fire of 1871 and until the year of Cyrus Hall’s death 1884.16 During these years, the struggle, often acrimonious, between the two brothers concerned the adoption of what Hounshell calls the American system of manufacture, a system of standardised parts based on the use of jigs and fixtures and a combination of general purpose and specialised machines. The machines were produced by the emerging New England machine tool industry. The context of this dispute was a continuing disagreement between the brothers about the scale of production of the mechanised reaper for which the company had become famous. Leander held the view that there was a significant danger of overproduction in any particular model year of the company’s reaper, a view that Cyrus Hall shared for many years. Leander’s view, however, was strengthened by the craft methods employed in the McCormick factories that were under his direction. These methods involved many parts being individually constructed by skilled workmen following patterns derived from the current year’s model design, but without the techniques necessary to achieve consistent reproduction of parts and, therefore, the need for additional ‘fitting’ operations
Learning in the knowledge-based economy
215
in the production process. In economic terms, the variable costs of production remained high despite the existence of a relatively ‘standardised’ model design. Weeks before the Chicago fire of that year destroyed their north-side factory, Cyrus Hall McCormick decided and announced in 1871 that he thought the company should expand production. To achieve this aim following the disaster he dispatched a series of production experts to the newly constructed factory in Chicago’s south-west side. Hounshell contends that in equipping this factory, the nature of Leander’s orders with New England machine companies clearly revealed his inability to exploit the opportunities offered by specialised machinery.17 After a protracted struggle with his brother and following the advice of the manufacturing experts he had sent to their factory, Cyrus Hall reached a similar conclusion. He first bought out his brother’s share in the business and then, with the concurrence of the Board of Directors of the recently incorporated business, dismissed Leander as factory superintendent in favour of Lewis Wilkinson, a veteran of several New England manufacturing companies. Working with the young Cyrus McCormick Jr, Wilkinson began the process of creating a truly standardised McCormick reaper in the 1880s, a path Cyrus Jr continued after Wilkinson left the company in the following year. The resulting changes between 1880 and 1884 were, however, not viewed neutrally by the craftworkers who had devoted their lives to the production methods. In 1885, a strike united all the major unions for the first time, a development that one author has concluded was a ‘prelude to Haymarket’.18 There are three lessons from this case. First, the issues of market size and growth are relevant to the transformation. Cyrus Hall McCormick’s decision to expand production was taken in the expectation that the market would absorb the resulting expansion. If this expectation had not been fulfilled, the McCormick Reaper Works would be a subject for the local history of Chicago, and its successor company, International Harvester, would not have been born. Second, to transform the scale of production it was necessary to replace one set of skills with another, a transformation that required considerable planning and tooling investment. In the case of the McCormick Reaper Works, the shortcomings of craft-based production made these investments worthwhile, although this need not always be the case. Third, the process of destroying one set of skills and replacing it with another had consequences for the livelihood of individual workers and was collectively resisted. The consequence of the change, however, was that the new production system, under the direction of Cyrus McCormick Jr was able to increase its output fivefold by 1902. The new system involved a different learning process from the old, one based upon the knowledge
216
The economics of knowledge
required for altering the entire production system to accommodate changes in model designs. 2.3
The Middle Distance: Why Wasn’t Fordism Enough?
The contention that the specific forms in which knowledge is organised within the organisation creates a differentiated advantage that may not persist with changes in market conditions is well illustrated by the Ford Motor Company. During the first quarter of the twentieth century, the logic of a mass production system based upon standardised parts was widely deployed. The principles enunciated by Ford of ‘power, accuracy, economy, system, continuity and speed’ became the guiding principles for mass production.19 Ford and his production engineers devised and perfected this logic in the production of the Model T, a single basic model for the automobile. The Model T was produced, with some setbacks, in ever greater numbers and with incremental changes from 1908 through 1927 when the ‘last’ of the 15 million Model Ts was driven by Ford to its resting place in Dearborn, Michigan. The difficulties that Ford and his company had in making the changeover to the Model A, and the costs to the company whose market share had slipped to less than 15 per cent in 1927, have become a case study followed by several generations of business school students.20 The central lesson is that General Motors achieved its stronger position through a combination of organisational, marketing and production innovations. These innovations created ‘flexible mass production’, the ability to re-tool model annual changes in the automobile and to promote these changes to buyers as an improvement over the standard represented by Ford’s Model T. Thus, as Hounshell concludes, ‘Ford had driven the strategy of mass production to its ultimate form thereby into a cul-de-sac’.21 This example of a fundamental change in the nature of the learning process is no less valid because the Ford Motor Company was ultimately able to make the transition to the new paradigm, for in doing so the original visionary path of Henry Ford, of mass production, had to be abandoned. This path, admired and emulated throughout the socialist world, supported the industrialisation of the Soviet Union and China, as well as many smaller countries, even as it was displaced in the USA and other industrialised countries. Few illustrations of changes in learning processes could be more dramatic than the contest between the Ford Motor Company and General Motors for the future architecture of industrial production. Along the branch pioneered by Ford is the indefinite expansion of standardised mass production, achieving ever-larger scales of production but imposing upon
Learning in the knowledge-based economy
217
the buyer a very limited choice. The branch that Alfred Sloan created produces technologies for creating variety and diversity in industrial output, the shifting of skills and employment towards stimulation of new consumer demands and the creation of new fashions in artefacts. In following this second path, the USA was able to discover the new technological frontier represented by the application of distributed information technology. This possibility might never have emerged had Fordism succeeded in producing what Filene claimed was a ‘way out’ of the American economic dilemmas of the inter-war years.
3
APPROACHING TODAY: KNOWLEDGE AND THE COMPUTER REVOLUTION
Early commercial computer manufacturers in the USA such as IBM and Burroughs attempted to produce general-purpose computers that could be reconfigured as information-processing needs shifted. This choice did not reflect technological necessity. Alternative means to build computational office machinery were being demonstrated by the British company, Lyons Confectionary Limited, which had created a special-purpose computer designed around many of the most common business processes in accountancy and inventory control.22 It is tempting to conclude that computer manufacturers had learned the lessons of the past century of American manufacturing in producing a flexibly configurable and mass-produced system. It is, however, more likely that the initial commercial designs were a response to the uncertainties of commercial computer applications. Although mass production had resulted in the centralisation of many business operations, those amenable to computerisation might be either smaller or larger than anticipated by computer manufacturers. A generalpurpose strategy not only simplified the design of the machine, it also assured that software would become the principal means for ‘customising’ the machine to particular applications. To achieve this aim, it was necessary for computer manufacturers to collaborate with users in the design of software systems. Users also exchanged a certain amount of software with each other through user groups sponsored and organised by the computer manufacturers, a harbinger of the contemporary ‘open source’ software movement. Computer manufacturers abandoned this collaboration in stages.23 During the first stage, customers were provided with ‘general purpose’ languages that could be used to create customised or user-owned applications, and computer manufacturers retreated from the direct implementation of specific software applications. A second stage of computer manufacturers’
218
The economics of knowledge
retreat, beginning in the late 1960s, involved further retrenching of software offerings and the blossoming of an independent software vendor market.24 It is difficult to conclude whether these developments would have followed a similar path had Ford’s vision of mass production prevailed over Sloan’s. Until the late 1960s, centralised computation was a necessity because the costs of equipment and ancillary support were simply too high to be distributed throughout the company. Had Ford’s vision prevailed, however, the longer planning horizon it entails might have encouraged a greater degree of hardware and software integration. By the mid-1960s, the progress in electronic components (another example of a producer goods industry that emerged from the continuing process of outsourcing) had provided the means for a new generation of ‘minicomputer’ companies. Digital Electronics Corporation and Data General promoted a decentralised model of computation, a model allowing the computer to either be integrated as a dedicated part of the production process or to ‘stand alone’ as a data processing engine for use as needed. In the framework of this chapter, the minicomputer provided the means for ‘localising’ the generation of knowledge at the site of application and provided a means for escaping from the development and tasking priorities established for the central computer facility.25 The trajectory of decentralising information processing applications continued with the development of the personal computer, which, instead of being an incremental and specialised innovation, created a profound alteration in the paradigm of information processing. As David has emphasised in his examination of the parallels between electrification of US industry and the growth of distributed computational power, the transformational feature of the distributed technology is its capacity to be utilised at the ‘site’ of application.26 For the personal computer, this ‘site’ was the desktop of the ordinary business professional as well as the desks of numerous clerks, administrators and secretaries. Electrification involved the transformation of the physical plant of manufacturing operations, which no longer needed to rely on the economies of the central drive and vertical plant architectures to best utilise central drives. In the case of the personal computer, the ability to distribute documents (initially by the manual exchange of the floppy disks and soon thereafter through local area networks) offered numerous opportunities to restructure the work flow within organisations. These changes reduced vertical information-processing layers in management. What in the past needed to be centrally collected and analysed can now be built into the local information processing tools and these can, in turn, be used to construct integrated systems in which horizontal as well as vertical communication paths in the organisation can be exploited in the same way as
Learning in the knowledge-based economy
219
fractional horsepower electrical motors can be distributed in mechanical operations. A new integrated and information-laden system in manufacturing and services is now characteristic of modern establishments. In one respect these developments are a direct application of the decentralised management and rapid changeover principles of Alfred Sloan. In another, however, they offer potential for more fundamental change. This potential is the consequence of the social issues involved in employing the new technologies. As in the past, there is resistance to the new technologies.27 It appears in obvious forms such as non-cooperation with the routines established for data entry or professed incapacity to master the new tools. One means for addressing these problems is to create a more transparent information environment in which the individual users of this technology receive direct and useful feedback for the performance of their jobs and in which these same users are encouraged to suggest new tools or improvements in existing ones. Some have suggested that implementing such capacities would support the ‘learning organisation’, an improved process of knowledge accumulation and distribution that would augment the designs originally built into such systems. Whether such optimistic assessments of the deconstruction of managerial control and authority will, in fact, be either widely accepted or constitute a competitive advantage for those organisations adopting such changes remains an unresolved but intriguing issue which is explored further in the next section (4). The central feature of these developments is that they involve the use of an information processing system to support local learning. The ultimate goal is similar to that pursued within Sloan’s model, facilitating innovation and product variety while simultaneously intensifying the use of fixed tangible and intangible capital. The means for doing this, however, involve new learning processes, both in the central design of such systems and their local application. The remarkable feature of decentralised information processing systems is the fluidity that they introduce for reconfiguring the loci of learning within the organisation. As information technology use intensifies, organisations are free to evolve in patterns that are peculiar to the composition of their employees and local circumstances. In this sense, contemporary learning opportunities involve re-energising the ‘local’ layer that has been displaced by prior management practices involving larger systemic planning and implementation of information systems. It is possible that these new processes may deliver a more participatory and, therefore, in certain respects, more democratic work environment. On the other hand, it is possible that these new processes will be structured and managed using new forms of hierarchy, ones that are perhaps more shallow in layers but that are no less autocratic in defining what procedures individual workers follow in their everyday work lives. This new environment
220
The economics of knowledge
offers a new contest between centralised and decentralised modalities of learning and knowledge accumulation. In summary, the intensification in the use of computer technology has been accompanied by several shifts in the locus of knowledge generation, from a highly centralised and collaborative activity between computer manufacturers and their customers to develop organisation-wide solutions to information processing problems, towards a more ‘local’ accumulation of knowledge and capability that is, nonetheless, co-ordinated through the use of an ‘information processing infrastructure’. The history of learning processes involving production planning and information processing systems in the twentieth century has been mixed. On the one hand, the creation of vast integrated systems managed by very large companies have substituted a new organisational-based learning for the localised and idiosyncratic learning of the shop floor and individual routines. On the other hand, when these larger organisational structures prove unable to exploit new opportunities, it appears that the articulation of the producer goods industries supports entry and growth of new industrial players.28 What is new, and in particular in need of further analysis, is how the further development of information and communication technologies is likely to influence the processes emerging in the ‘learning organisation’ or variants based on this idea that more strongly emphasise managerial control and authority. The ‘relocation’ of learning activities, in fact, the potential for the wholesale dispersal of these activities throughout the organisation, suggests a further profound set of changes in how organisations operate. This change provides a rematch of the contest, in a different context, between centralised and decentralised knowledge accumulation. By reflecting on the similarities with and differences from previous experience, it is possible to hazard a few predictions about the outcomes of this contest in the future. This is the principal theme of the next section.
4
FORWARD INTO THE FUTURE: THE NATURE OF THE KNOWLEDGE-BASED ECONOMY
Mass production supports the convergence of all productive processes that employ it towards an ‘information model’ in which the characteristics usually associated with ‘information goods’, their high initial costs of production and the relatively lower costs of subsequent copies, become the model for all mass produced products. There are, of course, significant differences in the scale between the ‘first copy’ costs of a General Motors automobile and those of a multimedia product, and the variable costs of
Learning in the knowledge-based economy
221
tangible artefacts such as automobiles are much higher than the reproduction costs for information goods.29 Nonetheless, many of the same principles may be employed in product planning such as the ‘reuse’ of design and development efforts, the planning of product ‘versions’ that incorporate upgrades and updates to the initial design and, perhaps most significantly from an organisational perspective, the granting of substantial autonomy to the product manager. 4.1
The Production of Variety
In terms of economic analysis, a principal decision of the firm is whether the relative investment costs of producing more varieties, that is, reinvesting in the fixed costs to produce a new variety, have a greater return than extending the marketing and production of existing varieties. The ‘changeover’ decision is also of direct relevance for issues of learning. A particular path of development may offer greater ‘spillover’ from the production of one variety to the production of another, reducing the costs of introducing new varieties and producing sub-additive costs (economies of scope) in producing more varieties. Technological change can be seen as opening up new opportunities for learning, some of which will lead to substitute products, eroding the share of existing varieties and perhaps extinguishing further development along previously developed paths. Learning influences not only the supplier’s development path, but also the nature of demand (for example, brand or product loyalty or product type familiarisation), and may either sustain or detract from the continuing production of a particular set of varieties.30 A complementary analysis can be derived from employing the traditional learning curve. If cost reduction is accumulated through repeating the same productive process, the relevant economic issue is when to make a change to a new learning curve. The same sorts of elaboration are possible additions to the basic learning curve model. Again, a principal issue is the timing and nature of changeover processes that alter the variety of products under production.31 These same basic principles can be extended beyond manufacturing to a number of services, although not to all. When services incorporate reproducible elements that are designed once and then replicated in whole or in a significant part of the whole, the ‘information model’ of fixed costs is applicable. Similarly, when services involve significant learning processes, the same basic model can be applied to services as to manufactured goods. The distinguishing feature of many services, however, is that they are almost entirely driven by variable costs. For example, the costs of a surgical operation involve commitment of the surgical team’s time and skill on
222
The economics of knowledge
a one-to-one basis with the patient. Although the costs of the operating theatre will be subject to economies of throughput planning and the artefacts utilised are individually subject to the information model, a major share of the costs is for the professional services rendered to a particular patient.32 Services of this nature are of considerable interest, but are not dealt with in this chapter. In the ‘information model’ of production, organisations innovate by creating generic products that are capable of reaping economies of scale by reusing the fixed costs of their development. These economies are inherent in the relatively lower fixed costs of creating the ‘first copy’ of the product compared to the costs of creating succeeding copies. What makes the world not only more ‘untidy’, but also more interesting, is the enormous variety of possibilities for reuse and recombination of knowledge. The resource allocation decisions governing the number and scope of this reuse and recombination are of principal importance in determining the rate and direction of economic growth in the ‘knowledge-based’ economy. This raises questions about how to best organise the accumulation of knowledge within and between organisations. As the model of flexible production has developed, capabilities to reconfigure and tailor products to specific markets have improved dramatically. In industrialised nations, the two most desirable strategies are (1) the creation of ‘hit’ products in which the full extent of mass production and consumption can be engaged or (2) the development of portfolios of specialised products that dominate their market ‘niches.’ Given the uncertainties of finding ‘hit’ products, a strategy of risk reduction is to employ the second strategy as a means of searching for potential ‘hits’. A less attractive third strategy is to produce ‘commodity’ products in which price competition reduces profits towards competitive levels, the ‘normal’ rate of profit. This strategy may still be viable because it supports the construction of a manufacturing base that can be used to introduce imitations of ‘hit’ or specialised products. If the preceding account is an accurate characterisation, we would expect to observe a continuing growth of product variety. Despite the continuing shortcomings of national income statistics in measuring variety (a legacy of the historical concern with recording the performance of mass production), evidence can be produced about the extent of this increase. A significant compound annual growth rate of varieties of consumer product categories has been experienced in the USA over the past two decades (see Table 8.1). Projecting the number of varieties of consumer packaged goods available for another two decades at these rates would yield over 140 000 different consumer goods on offer in the year 2018. An individual devoting one minute on average to considering each of them would devote over
223
Learning in the knowledge-based economy
Table 8.1
Indicators of increasing variety in the US economy
Product category
Number
Year
Number
Year
CAGR
Consumer packaged goods Vehicle models Vehicle styles Breakfast cereals National soft drink brands New book titles Mouthwashes and dental flosses Levi’s jean styles
4 414
1980
24 965
1998
10.1 %
140 654 160 20
1978 1978 1978 1978
260 1 212 340 87
1998 1998 1998 1998
3.5 % 3.5 % 4.3 % 8.5 %
40 530 27
1978 1978
77 446 130
1998 1998
3.7 % 9.1 %
41
1978
70
1998
3.0 %
Note: The number of consumer packaged goods is estimated from the shelf keeper unit code registrations, part of the uniform product code standard for standardised point of sale scanners. Sources of other figures are available in the source. In order to compute the CAGR (compound annual growth rate) the original sources’ approximation of the years of observation ‘late 70s’ and ‘late 90s’ have been fixed at 1978 and 1998 respectively. Source: Federal Reserve Bank of Dallas (1998).
2000 hours to the task of becoming a ‘fully informed’ consumer compared to the 400 hours now required. The problems of competitive analysis and strategic formulation will, of course, also expand. If these trends are to continue, the technologies available for conceiving, designing and implementing new product and service varieties must continue to improve. Just as the consumer faces information overload in considering the range of choice available, the companies producing these goods and services are likely to experience an overload in their design and execution capabilities using the centralised planning and design structure of Sloan’s flexible manufacturing model. Further decentralisation in which the term ‘product managers’ achieves a growing independence of action and initiative seems a likely consequence of these developments. 4.2
Organising Knowledge to Produce Variety
Implementing greater decentralisation within the organisation requires the development of intra-organisational interfaces between working groups that minimise the time required to set up and execute the complex interactions needed to conceive, design, and implement new products and services. Doing this requires organisational capabilities that are ‘modular’, they can
224
The economics of knowledge
be reconfigured and recombined to adapt to the growing variety and complexity of products and services. Information and communication technologies may aid in this process, but there are many uncertainties about whether the ‘hype’ and promises of ‘work group’ and ‘collaborative’ software are really meeting the needs of organisations for this flexibility and modularity.33 At the same time, however, negative ‘systemic’ effects, such as negative spillovers from one group’s decisions on the performance of others, are likely to persist. Limiting these effects requires managerial control. There are two issues that prominently appear in organisations’ attempts to grapple with the issue of new forms of organisation associated with variety and the ‘information model’ of product and service innovation and production. The first is the extent to which the ‘centre’ has to intervene in order to achieve coherent and cohesive outcomes among the decentralised ‘work groups’ or ‘production teams’ that are responsible for creating and producing products and services. Can the centre govern by adopting ‘procedural’ rules for mediating conflicting priorities or contests over common resources or is it necessary to engage in a more comprehensive planning process?34 The second issue is whether new ‘institutional standards’ (that is, norms, rules and practices) can be devised that reduce the need for such intervention and permit the internal organisation of companies and the relation of companies with each other to exhibit ‘emergent order’, much as markets achieve co-ordination between supply and demand without an explicit co-ordinator or auctioneer. The first of these issues reflects the culmination of outsourcing and the devolving of operational decision-making to lower levels within the organisation (that is, the elimination of middle management layers that previously processed information and attempted to steer the organisation towards strategic goals). Management provides not only ‘procedural authority’ but also the resources and priorities granted to these groups and a set of rules for governing the interactions among them. The disadvantage of decentralisation, however, is that potential synergies in the accumulation and reuse of knowledge may be ignored, reducing the organisation to a loose confederacy of smaller-sized enterprises operating within a common financial framework. These potential problems indicate that further co-ordination may have to be devised to capture the benefits available from operating within a single organisational framework. One means of doing this is to devise incentives and institutional standards that are effective in reducing central management ‘overheads’ (in terms of managerial time and therefore, ultimately, cost). How these incentives and institutional standards can be devised and who will devise them are areas of active experimentation
Learning in the knowledge-based economy
225
within contemporary organisations. A vast array of research opportunities exist for identifying their emergence in specific industries and thereby aiding in the process of their diffusion and recombination within other industries.35 Those who are sceptical about the possibilities for achieving this sort of decentralisation emphasise the situated nature of knowledge and individual cognition in order to deny the transferability of understandings reached by one social group within an organisation to others.36 This sceptical position, however, seems to deny that anything has happened since craftworkers were the principal repositories of knowledge in manufacturing establishments. There can be little doubt that the processes of labour resistance and solidarity have, on occasion, prevented the complete appropriation of local knowledge through the redesign and centralisation of production processes and designs. As was noted earlier, however, wherever this issue has proved to be a significant barrier, the nature of the production system has been changed to make skills and knowledge obsolete. This is done by a centralised authority devising a set of procedures and routines and the induction and indoctrination of an (often new) industrial labour force through a relatively short period of training, which reflects the modest extent of local knowledge required to perform the job.37 The question remains, however, whether standards for localised decisionmaking and knowledge accumulation can be re-established without re-creating the problems of opportunism, co-ordination failure and idiosyncrasies that historically motivated the usurpation of decentralised control. This amounts to assessing whether Sloan’s idea of divisional responsibility can be further extended to permit work groups to negotiate (either internally or with a ‘lean’ and process-oriented centralised management) the incentives and institutional standards that will govern their work. In some cases this process will not work and central authority will make direct interventions in the content of incentives and standards. These interventions will forgo the potential advantages of worker participation in the design and improvement of their roles, including the more detailed worker knowledge of context and situation. These developments are not just a source of interesting research questions; they reflect two alternative ‘paths’ for the predominant mode for the organisation of work in modern enterprises. They will influence the conditions of working life for the foreseeable future. There are many possible points of observation within the organisation and between organisations for observing the changing locus of learning within the organisation and how these changes stimulate the development of new incentives and institutional standards. They include human resource policies regarding recruitment, induction and training, the design of business processes and
226
The economics of knowledge
structures, and the configuration of information and communication technologies. This chapter focuses only on the last of these areas because of its direct and relatively straightforward linkages to institutional standardisation issues. 4.3
Information and Communication Technologies and Knowledge
The design and execution of information and communication technology systems is a central issue in the development and implementation of incentives and institutional standards for meeting the challenges of complexity and variety. The following discussion briefly considers three aspects of these technologies that are likely to inform future developments. First, we will examine the use of interpersonal communication within the organisation and between organisations to enhance the exchange of knowledge necessary for forming institutional standards. Second, we examine the use of information technologies to model the operations of the company with the aim of identifying systemic dependencies and then resolving them either by the decisions of central authority or by introducing incentives to limit systemic failures and the propagation of negative externalities. The third aspect of the use of information and communication technologies is in augmenting collective memory and, ideally, accelerating processes of collective learning and adaptation. The role of interpersonal communications Information and communication technologies provide a means to augment interpersonal communication within organisations. The use of email for the exchange of information allows fluid interpersonal communication required for learning and the negotiation of incentives and institutional standards, while reducing the overhead costs of scheduling and conducting meetings. There is evidence, however, that electronic communication does require ‘real world’ interaction to build interpersonal trust and to resolve non-routine transactions.38 It is, however, unclear whether reliance on ‘real world’ interpersonal interactions is a persistent or transitory feature in the use of electronic communications. Many current users have relied upon other types of communication for most of their careers, and succeeding generations with more experience with the medium may exhibit different behaviour. Correspondingly, the tools for facilitating more complex types of communication, such as exchanges about work or negotiations in progress, have only recently begun to be used on a broad scale.39 Some period of learning may be required before they can be considered as replacements for existing methods for negotiating standards about artefacts or procedures and a
Learning in the knowledge-based economy
227
continuing critical appraisal of these methods is needed to predict their eventual consequences. Evidence from Europe’s Telematics Application Programme, a Europe-wide research, development and demonstration programme, suggests that while some ‘new ways of working’ including email were used by a majority of participants, a minority of the participants utilised more advanced information and communication technologies such as collaborative software or videoconferencing.40 The relative immaturity in the use of these technologies suggests limits to their applicability both to the exercise of procedural authority by centralised authority and to their efficacy in negotiating incentives and institutional norms within decentralised work groups. Recent experience in the operation of open source software communities offers examples of ‘leading edge’ experimentation with new mechanisms for implementing collaborative work.41 Successful development of a widely used microcomputer operating system (Linux) and the software that dominates the application of ‘serving’ World Wide Web pages on the Internet (Apache) are two recent examples of successful developments undertaken by worldwide collaboration among software developers, many of whom have only ‘met’ through virtual communications.42 Since the working processes involved in the development of computer software involve many institutional standards issues (and several incentive issues as well), the open source software development is a likely source of evidence about the possibility of more broadly employing electronic communication to support shared development of artefacts involving the use of incentives and institutional standards within companies. For example, open-source software developers must develop a number of tools for the smooth interchange of ‘work in progress’ (the source code), which they are collectively developing. One of the features of these tools is internal documentation of revisions that allows smooth ‘backtracking’ to earlier versions. This capability, in combination with modular code and standardised interfaces between modules, allows the recombination of modules across versions of the system under development. To co-ordinate these activities, open-source software developers also employ real-time communication channels such as ‘chat’ lines (concurrent short message exchanges) and a hierarchy of procedural authorities (at the level of the main project as well as sub-projects) to achieve closure on versions of the software as it is developed. All these elements involve learning and exchange processes that are likely to be reproduced within other successful virtual development efforts. Interpersonal communications about artefact and procedure standards are now able to benefit from the very rapid increase in abilities to document current practice. The Internet technologies for creating web pages as
228
The economics of knowledge
applied to intranet communication networks substantially reduce the costs of intra-company ‘publications’, the internal dissemination of company information designed to facilitate learning activities. These technologies also have the potential to accelerate the timeliness of information publication and to hasten the revision and updating processes. As in any publication process, the costs of producing the first copy of information remain significant and it is possible that organisations can over-invest in the production of information that does not provide significant value. Nonetheless, these technologies do appear to offer potential reductions in the distribution costs and increases in the accessibility of information about artefact and procedure standards. Electronic communication can further augment this process by providing feedback about the value and correctness of such publications. Organisational incentives that support the currency and accuracy of publications are necessary for effective results and provide a means for preventing over-investment by individual enthusiasts.43 Effective means for organising information about artefact and procedure standards will also be required to decentralise institutional standardsmaking processes and devise effective incentives. In medium-sized or larger organisations it is relatively easy to create information structures whose complexity creates a bottleneck to effective use. The development of the skills and procedures for organising this information is likely to become a relatively high priority. The development and application of new metrics for assessing the complexity of information structures such as web sites is urgently needed.44 Once created, however, the information structure of a particular organisation is likely to be subject to ‘lock in’ as users invest in learning how to navigate it to find the information they need. Modelling business processes The value of creating a ‘virtual model’, an electronic representation of the processes and transactions of the organisation, is the ability to define organisational dependencies, overlaps and redundancies that can be fruitfully resolved through decentralised initiative or centralised intervention. Integrating the analysis and execution of transactions is particularly important because it avoids a separate and often difficult to justify data entry process. Thus, for example, inventory control systems that rely upon a separate data entry procedure are far less reliably maintained than ones where placing an item in inventory or removing an item requires a ‘logging’ procedure. Moreover, the requirement for data logging provides the impetus needed to implement complementary technologies such as scanner codes, along with the readers and writers of these codes. In retail establishments, for example, reorders are often automatically generated by
Learning in the knowledge-based economy
229
the integrated processing of received shipments and ‘point of sale’ terminal information. The recording of individual transactions is the basis for an organisation’s ‘data warehouse’ and the growth of these warehouses is providing the impetus for fundamental change in the architecture of enterprise management software. The goal of software systems devised to exploit data warehouses is to provide a means for the ex post imposition of an analytical structure on data records that may either be tightly or loosely linked to one another. Thus, enterprise software may be used to model and analyse the processes that originally generated the stored data. They may also be used, however, for entirely different purposes. For example, in retail applications data on weather patterns or television schedules may be correlated with the purchasing patterns of customers. The data from factories where many of the processes are automatically measured and recorded may be analysed for previously unrecognised interdependencies and the propagation of local machine breakdowns on the pattern of work flow and error rates. The ability of such systems to support many different users may allow substantial decentralisation in problem identification, problem solution and design activities. The particularly intriguing feature of these systems is their capacity for the dynamic restructuring of interrelationships among the data that have been warehoused. It is important to emphasise that these tools are not costless or automatic; their application requires substantial investment and skill. In principle, these tools can support the formation of institutional standards and the specification of incentives that will improve organisational performances. In the first instance, such systems lower the incremental costs of organisational analysis; ultimately, they may encourage creative experimentation, learning and critical thought. Organisational memory and learning The issues surrounding organisational memory and learning are particularly important for assessing the value of localised knowledge accumulation. The activities of engineering consulting companies, architects, law firms and management consultants involve a constantly shifting pattern of novel problem-solving activities and the replication of established knowledge. Many other organisations, including those that have high numbers of ‘similar’ transactions, could also benefit from improved recovery of previous solutions to related problems. On the one hand, localised problemsolving is stimulated by the immediacy and tangibility of the problem to be solved. On the other hand, localised problem-solving runs the risk of providing idiosyncratic solutions that ignore potentially valuable organisational experience. Smaller organisations, particularly those with relatively stable employment, can develop effective social referral networks so
230
The economics of knowledge
that the same individual or group consistently deals with problems of a similar type.45 Larger organisations, including those involved in the generation of greater variety, will face greater problems in creating social referral networks. They face three challenges. The first is the challenge of identifying the salient features of a particular problem that make it ‘like’ some other problem the organisation has experienced. If the challenge of ‘likeness’ can be met, the second problem is identifying the relevant information source. At best, this will involve the identification of a current employee of the organisation who can service the referral (assuming that proper intraorganisational incentives are in place to do so). Perhaps as likely, part or all of the organisational memory regarding the particular problem has been disassembled through departures or transfers so that it is not possible to find the appropriate individual. These three challenges, ‘likeness’, identifying relevant referrals and recovering information with no active caretaker are among the most common ‘memory’ problems that medium- and largersized organisations face. The ways that these types of problems are resolved is central to the prospects of decentralisation. 4.4
Knowledge Management: Control versus Empowerment
Eigentlich weiss man nur wenn man wenig weiss; mit dem Wissen wachst des Zweifel. (We know accurately only when we know little; with knowledge doubt increases.) (Goethe, Spruche in Prosa [Proverbs in Prose])
Each of the above instances of the use of information and communication technologies in supporting decentralised formation of institutional standards and incentives intersects with the field of studies that has come to be known as ‘knowledge management’. The growing recognition of the intangible asset value of knowledge as a productive input and, in some cases, as an output of the organisation suggests the need for a management strategy to maximise this asset’s value. The controversy and struggles arising from knowledge management recapitulate and surmise about the contest between centralised and decentralised strategies with regard to learning and knowledge in modern organisations. Much of this struggle turns on an understanding of the meaning of the term ‘knowledge’. Throughout this chapter, knowledge has been depicted as deeply intertwined with learning and transactional or interactive experience. The alternative view is that knowledge can be meaningfully described as a ‘stock’ that can be ‘held’ in some repository such as a data warehouse or in the documentation produced by the organisation about its organisation and operations. In the first view, knowledge is transitory and provisional, while,
Learning in the knowledge-based economy
231
in the second view, knowledge is an equilibrium state reflecting the attainment of a ‘best’ solution or understanding. The contention is that the growing complexity of modern organisations as well as the variety of their outputs and processes make the first view the most relevant. A version of the second view, involves developing and employing a ‘best’ set of understanding and solutions for the circumstances, for ‘coping’ and for using this knowledge to impose routines and standard operating procedures on the workforce. Which view of knowledge prevails has implications for the systems used to gather, store and process information in the organisation as well as for the applications of these information capacities to solve problems (knowledge creation) and enable learning (knowledge exchange or reproduction). Within the community of scholars engaged with ‘knowledge management’ there are clearly adherents to both viewpoints. As each of the viewpoints is also allied to a position on the issues of decentralisation it would be comforting to believe that competitive selection would test the respective value of the viewpoints in practice. Unfortunately, there are real possibilities that the ‘best available’ (also called ‘best practice’) approach will prevail, despite the merits of decentralised options, as the ‘best practice’ approach when used as a means for indoctrination serves preferences for power and dominance. It is also the case that organisational solutions involving hierarchical control and dominance are the most familiar. Decentralised solutions need to be examined critically and comparatively. For example, what organisations choose to remember and what they forget should, in principle, be an investment decision. This immediately conjures up images of painful reporting procedures in which individuals are required to maintain records of their problem-solving activities categorised and classified in ways the organisation’s ‘knowledge managers’ believe will eventually facilitate the retrieval of relevant information and thereby a return on investment. In contemplation of this prospect, it is important to confront this strategy with a decentralised solution based upon incentives and institutional standards. There are three requirements for such a decentralised knowledge management system. First, there must be rewards (positive incentives) for ‘publication’ (disclosure) of knowledge as information that might be used by others. Second, these rewards must create greater value for the first disclosure of the relevant information (as copying the available information is an opportunistic possibility). Third, the reward system must be funded in a way that does not require the information user to bear the costs of the reward. Thus, this system cannot employ the common practice of establishing ‘internal markets’ for consulting and other services within the organisation in which the user must bear the costs (and uncertainties) of utilising information produced by others. Such systems are often underutilised because
232
The economics of knowledge
of the difficulties of assessing the information to be provided in advance. By separating rewards from utilisation the organisation encourages the diffusion of information and is, in effect, investing in the distribution of information within the organisation.46 A single institutional standard is a necessary complement to these features of the incentive system. A norm of ‘recognition’ in using others ideas must be established so that it is possible to attribute credit in a decentralised way. Failing this, it is necessary to monitor the use of information within the organisation to ascertain credit. This incentive system and institutional standard is, of course, virtually identical to the one that Dasgupta and David (1994) propose lies at the root of the social system known as ‘science’, one of the most powerful systems devised by humans for generating and distributing knowledge. By providing incentives for disclosure and identifying ‘priority’ of ‘discovery’ (or contribution to the common knowledge), the system encourages others to consider the current state of knowledge within the organisation and to hasten to add their own contribution lest someone else receive recognition and the subsequent reward.47 Severing the use of knowledge from a requirement to ‘pay’ for its use generates further incentives to look for a solution rather than engage in the risks and incur the costs of trying to create a solution (that is, to reinvent the wheel). The level of reward can be adjusted to discourage ‘over-investment’ by individuals in disclosure, since the value of only a few uses may not justify the costs of submission. There are, of course, some problems with this scheme as there are with the social system of science that suggests it. The opportunity costs of the individual’s time in making disclosures is a further investment in the scheme by the company and the collective opportunity costs plus the costs of the rewards may exceed the benefits since, although rewards may be adjusted from time to time, they need to be set arbitrarily. The parallel problem in the case of science is the problem of setting the overall level of funding for scientific endeavour (as, in science, it is the proposed investigation rather than the outputs that govern compensation48). There are also practical problems in packaging knowledge for disclosure in ways that prevent its ‘leakage’ to rivals and in governing rewards for the creation of ‘incremental improvements’ which may be just that or opportunistic attempts to share the rewards.49 The purpose of this stylised example is to illustrate that it is possible to view issues of knowledge management involving organisational memory and learning as a problem of creating the corrective incentive structure and institutional standards for knowledge disclosure. It is certainly true that memory represents an important asset and learning a major investment in modern organisations. It does not follow, however, that these assets and
Learning in the knowledge-based economy
233
investments can be readily committed to inventory or distributed within the organisation, as are other corporate assets. Instead, it is necessary to bring producers and users of different types of knowledge in the organisation together through a system of procedural authority (in this example represented by the mechanisms of establishing and operating a ‘disclosure’ process). Aligning the incentives of producers and users becomes the principal incentive design problem, one for which the social institutions of science provide an intriguing solution.
5
CONCLUSION
This chapter began with an examination of the relocation of learning and knowledge accumulation within the organisation with the aim of illuminating how these processes have changed over time and may be subject to further transformation. The principal examples utilised in the first half of the chapter stressed the historical processes by which localised learning and knowledge accumulation was displaced by a process of centralisation needed to support the standard and flexible mass production of Ford and Sloan respectively. In the second half of this chapter, Sloan’s system of flexible mass production is portrayed as suffering from increasing stress in managing the processes of variety creation and problems shared by non-manufacturing organisations such as services as well. The prospects for avoiding diseconomies of scale in the management of variety are chosen as the principal focus of analysis along with the specific contributions of information and communication technologies to the standards regarding procedures and artefacts within the organisation. The information and communication technology examples highlight the considerable variability in the existing systems for managing the decentralisation that is a concomitant of increasing variety. In general these systems cannot be characterised as either mature or well integrated for the purposes of managing decentralised learning or accumulation of knowledge, or avoiding diseconomies as variety continues to expand. There remains considerable opportunity that has not yet been fully exploited in such systems for the purposes of standardising artefacts and procedures within the organisation as well as devising appropriate incentive systems that would address the problems of achieving greater product and service variety. Thus, it would be premature to conclude that the Sloan model of centralised design authority (at least at a divisional level) is faced with a viable and currently available substitute. At the same time, it would appear that the ‘learning organisation’ represented by a decentralisation of
234
The economics of knowledge
the processes of making intra-organisational standards has considerable further potential than has yet been exploited. This potential is not only relevant for further improvements in productivity and economic growth, it also offers fundamentally more democratic and participatory processes in the workplace. Whether the new ‘learning organisations’ created by these processes will overcome tendencies towards the assertion of dominance and power that would indoctrinate workers with ‘best practice’ and lock down the potential of the new technologies remains to be seen. In their infancy, these organisations will be vulnerable to market pressure and the undeveloped state of the information and communication technologies necessary to support them.
NOTES 1. 2.
3. 4. 5. 6. 7. 8. 9.
10. 11. 12. 13. 14. 15. 16. 17. 18. 19.
A prominent contemporary theory for understanding this process is provided by Lave and Wenger (1991). Such accounting is not usually made in economic theory where it is more straightforward simply to argue that the firm understands its own technological capabilities and may achieve an understanding of any other ‘technology’ (productive capability) that it deems useful in the pursuit of profit. Coming to such understanding may represent a fixed ‘getting started’ cost worth noting, but the details of the process and its corollaries are not further developed. Arrow (1962) and David (1975). Schmookler (1966). One type of market structure consistent with the existence of multiple ‘niche’ players. This approach is a generalisation of some of the issues raised in David and Foray (1995). This is a good example of a theoretical abstraction that would be very difficult to measure. The argument that follows is only based on the premise that such rankings exist, not that they can be empirically ascertained given the large number of co-determinants involved. Among the most influential efforts to undertake this is Pavitt (1984). It is often argued that production engineering and operations are more integrated in Japanese factories than in their Western counterparts and that this is a major source of the high rates of Japanese productivity improvement. Two poles defining the spectrum of views on this subject are (Dore, 1973) who reflects on the social consequences of this form of organisation and (Schonberger, 1982) who uncritically accepts its inherent superiority. In addition, ancillary sociological issues related to the ‘professionalisation’ of certain classes of employment may have created further impetus to this trend. Rosenberg (1976). Smith (1937), Young (1928) and Stigler (1951). For example, Mumford’s (1934) monumental effort at identifying the origins and critiquing the consequences of industrialism. Mumford (1934), Giedion (1948), Hounshell (1984) and Beniger (1986). Filene (1925), as quoted in Hounshell (1984: 305). The following draws heavily upon Hounshell’s (1984) account of these events. Hounshell (1984: 174). Ozanne (1967). Henry Ford, ‘Mass production’, as quoted by Hounshell (1984: 217).
Learning in the knowledge-based economy 20. 21. 22. 23. 24.
25.
26. 27. 28. 29.
30. 31. 32. 33. 34. 35. 36. 37. 38.
39. 40. 41.
235
Among the texts often used are Nevins and Hill (1957), Abernathy (1978) and Chandler Jr (1962). Hounshell (1984: 267). Caminer et al. (1996). See Steinmueller (1996) for a more complete account of these developments. IBM retained a significant position in software production. What distinguished its later activities in the software field from the earlier period was the lack of sustained commitment to producing software in any particular category of application. IBM’s software activities in the 1980s and 1990s focused on what has come to be called ‘enterprise computing’, large integrated systems assembled from more generic sub-systems and components. IBM’s participation in any particular sector was, in this period, based upon the demand for these services in the use of their computer systems rather than the effort to pre-specify bundled packages of hardware and software for particular applications. See Steinmueller (1996) for further details of this history. The emergence of minicomputers was made possible by shortcomings in the ‘divisibility’ of computing power that, in turn, would have created yet another alternative path for the development of computation. One of the ironies of historical development is that time-sharing related ideas are re-emerging in the use of the Internet. These new forms range from ‘net computers’ (personal computers whose system and application software is regularly upgraded through network access) to the distributed control of logistic and production systems that use Internet technology and centralised computers. David (1991). See Zuboff (1988) and Mansell and Silverstone (1996). See Pavitt and Steinmueller (2000) for a complementary analysis reaching a similar conclusion. A significant qualification to this point of view is the recognition that the relative prices of tangible inputs may change over time, selecting against particular sets of product variety in ways that were not initially anticipated. Although this is theoretically an important issue for all tangible goods, it is often only of practical significance for relatively simple products that have many close substitutes and therefore relatively small profit margins. Other demand side influences including saturation, fashion, habituation or novelty preference may also involve learning elements. See Gulledge Jr and Womer (1986) for an extended analysis along these lines. For example, the margin between the re-sterilisation of instruments and their disposal is continuously closing. See Steinmueller (2000) for an overview of these issues. See Cowan et al. (2000). Hertog and Huizenga (2000). For example, see Ancori et al. (2000). Those who believe that the local knowledge of production workers is of particular significance need to explain how this belief can be consistent with relatively high turnover in the labour force. Granovetter (1985) provides a general theory regarding the tradeoffs between interpersonal trust and market exchanges. The issue of routine and non-routine transactions is examined empirically in Hart and Estrin (1991). Kraut et al. (1998) examine the issue of intra-organisational networks as well as providing a thorough conceptual review of these issues. Steinmueller (2000). ASSENT (1998). Open-source software involves the publication of the source code that is used to compile executable computer programs. Typically, source code is the proprietary intellectual property of the software author as its analysis would allow the creation of functionally identical software by competitors. In the open-source model, alternative business models must be developed to fund the costs of software development and current developments
236
42. 43.
44. 45. 46. 47.
48.
49.
The economics of knowledge have relied upon voluntary contributions of programmers. See Mateos-Garcia and Steinmueller (2003). Jeong (1999). The design of such incentives is a difficult practical problem since it is desirable to encourage learning and experimentation with innovative information resources as well as the adaptation of existing organisational information resources to the electronic media. See Steinmueller (1992) for a pre-World Wide Web examination of these issues. It is, of course, important to ask whether consistently good answers are found for such problems. See David and Foray (1995) for the implications of this argument for science and technology investment from a social viewpoint. In practice, the efforts required to assign priority are best left primarily to the community of users since it is undesirable to exclude incremental improvements to existing solutions. Instead, the role of ‘procedural authority’ (management in this case) should be confined to linking suggestions that appear to be related so that first disclosure is rewarded even if later disclosures are more heavily utilised. This practice is necessary given the often lengthy delay in making use of science and the difficulty of ‘tracing’ these applications. In the example here, the practicality and instrumentality of results suggest a shorter delay while the proximity of production and use indicate better ‘tracing’. The latter will ‘of course’ depend on how well the institutional standard (or norm) of recognition operates. The parallel issues in science concern the release of interim research results that might allow rivals to make a larger claim of discovery or generalisation and the governance system of peer review which helps define the significance of claimed advances in knowledge.
REFERENCES Abernathy, W.J. (1978), The Productivity Dilemma: Roadblock to Innovation in the Automobile Industry, Baltimore, MD, Johns Hopkins University Press. Ancori, A., A. Bureth and P. Cohendet (2000), ‘The economics of knowledge: the debate about codification and tacit knowledge’, Industrial and Corporate Change, 9 (2), 255–88. Arrow, K.J. (1962), ‘The economic implications of learning by doing’, Review of Economic Studies, 29 (3), 155–73. ASSENT (1998), Assessment of the Telematics Application Programme (ASSENT), Assessment of the Results of the Projects, Telematics Application Programme, Deliverable 9.01, Brussels, European Commission. Beniger, J.A. (1986), The Control Revolution. Cambridge, MA, Harvard University Press. Caminer, D., J. Aris, P. Hermon and F. Land (1996), User-Driven Innovation: The World’s First Business Computer, London, McGraw-Hill. Chandler Jr, A.D. (1962), Strategy and Structure: Chapters in the History of the Industrial Enterprise. Cambridge, MA: MIT Press. Cowan, R., P.A. David and D. Foray (2000), ‘The explicit economics of knowledge codification and tacitness’, Industrial and Corporate Change, 9 (2), 211–54. Dasgupta, P. and P.A. David (1994), ‘Toward a new economics of science’. Research Policy, 97 (387): 487–521. David, P.A. (1975), ‘Learning by doing and tariff protection: a reconsideration of the case of the ante-bellum US cotton textile industry’, in P.A. David, Technical
Learning in the knowledge-based economy
237
Choice, Innovation and Economic Growth, Cambridge, Cambridge University Press, pp. 95–173. David, P.A. (1991), ‘Computer and dynamo: the modern productivity paradox in a not-too-distant mirror’, in OECD (eds), Technology and Productivity: The Challenge for Economic Policy, Paris: OECD. David, P.A, and D. Foray (1995), ‘Accessing and expanding the science and technology knowledge Base’, STI Review, 16: 13–68. Dore, R. (1973). British Factory Japanese Factory. Berkeley, CA, University of California Press. Federal Reserve Bank of Dallas (1998), The Right Stuff: America’s Move to Mass Customization, The 1998 Annual Report, Dallas, TX: Federal Reserve Bank of Dallas. Filene, E.A. (1925), The Way Out: A Forecast of Coming Changes in American Business and Industry, New York, Doubleday. Giedion, S. (1948). Mechanization Takes Command: A Contribution to Anonymous History, New York, W.W. Norton. Granovetter, M. (1985), ‘Economic action and social structure: the problem of embededness’, American Journal of Sociology, 91 (3), 481–510. Gulledge Jr, T.R, and N.K. Womer (1986), The Economics of Made-to-Order Production Theory with Applications Related to the Airframe Industry, in M. Beckmann and W. Krelle (series eds), Lecture Notes in Economics and Mathematical Systems 261, Berlin: Springer-Verlag. Hart, P. and D. Estrin (1991), ‘Inter-organizational networks, computer integration, and shifts in interdepdence: the case of the semiconductor industry’, ACM Transactions on Information Systems, 9 (4), 370–98. Hertog, J.F.D. and E. Huizenga (2000), The Knowledge Enterprise, Implementing Intelligent Business Strategies, London, Imperial College Press. Hounshell, D.A. (1984), From the American System to Mass Production, 1800–1932, Baltimore, MD, Johns Hopkins University Press. Hughes, T.P. (1989). American Genesis, New York: Viking. Jeong, B.S. (1999), ‘Analysis of the Linux system, a new entrant in the operating system market: technological innovations and business models, SPRU – Science and Technology Policy Research’, unpublished SPRU MSc dissertation, University of Sussex, Brighton. Kraut, R., C. Steinfield, A. Chan, B. Butler and A. Hoag (1998), ‘Coordination and virtualization: the role of electronic networks and personal relationships’, Journal of Computer Mediated Communications, 3 (4), http://www.ascusc.org./ jcmc/ vol3/issue4/kraut.htm Lave, J. and E. Wenger (1991), Situated Learning: Legitimate Peripheral Participation, Cambridge, Cambridge University Press. Mansell, R. and R. Silverstone (eds) (1996), Communication by Design: The Politics of Information and Communication Technologies, Oxford, Oxford University Press. Mateos-Garcia, J. and W.E. Steinmueller (2003), The Open Source Way of Working: A New Paradigm for the Division of Labour in Software Development?, Falmer, East Sussex, SPRU Science and Technology Policy Research, http://siepr. stanford.edu/programs/OpenSoftware_David/NSFOSF_Publications.html Mumford, L. (1934), Technics and Civilization, New York: Harcourt, Brace. Nevins, A. and F.E. Hill (1957), Ford: Expansion and Challenge, 1915–1933, New York, Charles Scribner’s Sons.
238
The economics of knowledge
Ozanne, R. (1967), A Century of Labor – Management Relations at McCormick and International Harverster, Madison, WI, University of Wisconsin Press. Pavitt, K. (1984), ‘Sectoral patterns of technical change: towards a taxonomy and a theory’, Research Policy, 13, 343–73. Pavitt, K, and W.E. Steinmueller (2000), ‘Technology, Strategy and the Information Society’, in A. Pettigrew, H. Thomas and R. Whittington (eds), Handbook of Strategy and Management, London, Sage Publications. Rosenberg, N. (1976), ‘Technological change in the machine tool industry, 1840–1910’, in N. Rosenberg (ed.), Perspectives on Technology, Cambridge, Cambridge University Press, pp. 9–31. Schmookler, J. (1966), Invention and Economic Growth, Cambridge, MA: Harvard University Press. Schonberger, R.J. (1982), Japanese Manufacturing Techniques: Nine Hidden Lessons in Simplicity, New York, Free Press. Smith, A. (1937), The Wealth of Nations, 5th edition (first published 1789), New York: Modern Library. Steinmueller, W.E. (1992), ‘The economics of production and distribution of userspecific information via digital networks’, in C. Antonelli (ed.), The Economics of Information Networks, Amsterdam: North Holland. Steinmueller, W.E. (1996), ‘The US Software Industry: An Analysis and Interpretative History’, in D.C. Mowery (ed.), The International Software Industry, New York, Oxford University Press. Steinmueller, W.E. (2000), ‘Will new information and communication technologies improve the “codification” of knowledge?’, Industrial and Corporate Change, 9 (2), 361–76. Stigler, G. (1951), ‘The division of labor is limited by the extent of the market’, Journal of Political Economy, 59 (3): 89–90. Young, A. (1928), ‘Increasing returns and economic progress’, Economic Journal, 38 (152): 527–40. Zuboff, S. (1988), In the Age of the Smart Machine, New York, Basic Books.
9. The economics of open technology: collective organisation and individual claims in the ‘fabrique lyonnaise’ during the old regime Dominique Foray and Liliane Hilaire Perez INTRODUCTION1 What we call ‘knowledge openness’ is a system in which the principles of rapid disclosure of new knowledge are predominant, and in which a number of procedures facilitate and reinforce the circulation not only of knowledge such as that which is codified in instructional guides and documentation, but also of tacit knowledge and research tools. It is not pure chance that in this context new knowledge is codified and carefully systematised in order to facilitate its transmission and discussion. But particular attention is also paid to the reproduction of knowledge, that is, to learning. It is not because knowledge flows freely – in the form of manuals and codified instructions – that it is necessarily reproduced from one place to the next. It is also necessary to create and maintain relationships between ‘masters and apprentices’, either in the context of work communities or in formal processes of teaching practical knowledge. The significance of knowledge openness is particularly important for knowledge, which is an input for further cognitive works. In this case the principle of openness allows external users of that knowledge to reproduce it for investigation, modification and improvement. Systems of knowledge openness relate to public (or semi-public) spaces in which knowledge circulates. Such spaces can include areas in which exclusive property rights cannot be granted, either constitutionally (as in the case of open science) or within the framework of organisations specially designed for the purpose (research networks where partners share their knowledge) and markets whose modi operandi are conducive to efficient knowledge dissemination. In such circumstances, a fundamental economic issue is the design of private incentives (to give credit to the knowledge producer) without creating exclusivity rights. 239
240
The economics of knowledge
The economic analysis of ‘knowledge openness’ as a system has been developed extensively in the field of scientific research owing to the seminal works of Dasgupta and David (1994), David (1998; 1999) and David et al. (1999). The approach of the ‘new economics of science’ develops two important arguments for theoretical analysis as well as policy implication in the field of the economics of knowledge. First, knowledge openness and sharing behaviours do not only express some kind of ethics or moral attitude (although ethical conviction certainly plays a role). Knowledge openness is viewed, above all, as a mechanism generating economic efficiency that people in certain circumstances are willing to implement and maintain in order to be players in a positive sum game. In fact, knowledge openness that entails rapid and complete distribution, facilitates co-ordination between agents, reduces risks of duplication between research projects and functions as a sort of ‘quality assurance’ in so far as disclosed results can be reproduced and verified by other members of the community. They are thus peer evaluated. Both static efficiency and dynamic efficiency are, therefore, expected to be enhanced: (1) ‘the wheel is not re-invented’ and each ‘great’ invention will benefit from a strong collective focus on it; (2) propagating knowledge within a heterogeneous population of researchers and entrepreneurs increases the probability of later discoveries and inventions and decreases the risk that this knowledge will fall into the hands of agents incapable of exploiting its potential (David and Foray, 1995).2 Second, knowledge openness does not mean the absence of individual incentives. There is a need for individual rewards, which are compatible with the complete disclosure norm. In the case of open science, a remarkable mechanism comes into play consisting of the granting of moral property rights which are not concretised in exclusivity rights. These two features apply in the world of open science as well as in local systems of open technology, such as the particular case of the ‘fabrique lyonnaise’ to which this chapter is devoted.
1
THE ECONOMICS OF OPEN SCIENCE
Private markets, even when equipped with a system of intellectual property rights, are ill-suited to the production and exploitation of certain forms of knowledge. There is, thus, a need for some other economic institutions that can be relied upon to create and exploit knowledge in an efficient manner. One main institutional arrangement consists in financing knowledge production from public (or private) funds while at the same time identifying mechanisms aimed at providing forms of self-discipline, evaluation and
The economics of open technology
241
competition within the beneficiary community. In return for aid received, the beneficiary is expected not so much to pursue objectives set by the financier, but rather to relinquish exclusive rights on knowledge produced. In concrete terms, society is responsible for covering the costs of resources needed to produce knowledge. This means, however, that anything produced is the property of society as a whole and cannot be privately controlled. Knowledge is often disclosed through scientific publications, and since anything published can no longer be patented, it definitively becomes public knowledge (in the US system the grace-period mechanism allows patenting in the year following publication). Rapid communication and sharing of knowledge are the norm, facilitating the creation of co-operation networks. Knowledge openness characterises, therefore, research undertaken in public institutions such as universities where in most cases exclusive rights cannot be granted on knowledge and where salaries and equipment are paid from public funds. In many countries public funding of a large part of this system is facilitated by the close ties that exist between research and higher education. As Arrow (1962) points out, the fact that research and teaching activities are two sides of the same profession is a ‘lucky accident’ since it ensures that researchers are remunerated not on the basis of what they find (their income in that case would be highly irregular and only the best would survive) but on that of regular teaching. It is because this public system produces both knowledge and human capital that it easily harnesses a large proportion of public resources.3 1.1 A First Look at a Great Problem: Voluntary Spillovers and Private Incentives Yet there is still a piece missing in this system. How can people be encouraged to be efficient and effective researchers if their work is immediately disclosed, without any possibility of private appropriation, and their salaries guaranteed? An ingenious mechanism comes into play here, consisting of the granting of moral property rights that are not concretised in exclusivity rights (in other words, they are compatible with the complete disclosure norm). It is the priority rule which identifies the author of the discovery as soon as he or she publishes and which thus determines the constitution of ‘reputation capital’, a decisive element when it comes to obtaining grants. ‘The norm of openness is incentive-compatible with a collegiate reputational reward system based upon accepted claims to priority’ (David, 1998, p. 17). The priority rule creates contexts of races (or tournaments), while ensuring that results are disclosed. It is a remarkable device since it allows for the creation of private assets, a form of intellectual property, resulting
242
The economics of knowledge
from the very act of forgoing exclusive ownership of the knowledge concerned. Here the need to be identified and recognised as ‘the one who discovered’ forces people to release new knowledge quickly and completely. In this sense the priority rule is a highly effective device that offers non-market incentives to the production of public goods (Callon and Foray, 1997; Dasgupta and David, 1994). Maximising knowledge externalities is the raison d’être of such systems (for instance, of an open science system). This is based on a set of consistent institutions: weak intellectual property protection; funding largely from government or private foundations; and a reward system (based on priority) compatible with the fast and broad dissemination of knowledge. Moreover, management of externalities, namely, the organisation of access to and integration of knowledge, is accomplished through norms and institutions. For example, it is usual for researchers to write and share ‘surveys’ aimed at making available to the rest of the community the state of the art of a particular domain. Nothing like that exists in the private property system.4 Of course, the ideal world of openness described here does not exclude the possibility of bending or departing from the rules. On the contrary, the tournament contexts created by the priority rule, as well as the size of related rewards, tend to encourage bad conduct. The notion of ‘open science’ is therefore based on an ideal that is never achieved (in other words, there will always be many cases of various degrees of retention). In Dasgupta and David (1994) it is argued that the norms are prescriptive, and that beliefs that are instilled in scientists as part of the ‘culture of science’ have an effect on their behaviour – making it easier to form co-operative networks where it is in their mutual interest (and that of society at large) to organise research co-operatively. 1.2
Modelling Exercise
These ‘good properties’ have recently been modelled by David (1998), who shows how the disclosure norm positively influences the cognitive performance of the system under consideration. David models stochastic interactions in a group of rational researchers individually engaged in a continuous process of experimental observation, information exchange and revision of choices in relation to locally constituted majorities. This modelling is then used to link micro-behaviours (being open, being closed) and macroperformances. Simulations suggest that the social norm of openness, which influences micro-behaviours, favours free entry into knowledge networks and, in so doing, prevents researchers from closing in on themselves too quickly and excluding different opinions. David shows that a system situated
The economics of open technology
243
beyond the critical openness threshold ensures confrontation of ideas and provides a mechanism that guarantees the production of consensus and preserves the diversity of opinions. The capacity to produce scientific statements collectively while preserving a degree of diversity of opinions and arguments is thus an important feature in an open research network, and standards of disclosure and openness appear to be decisive in the cognitive performances of the network. The advantage of such an approach is that it produces formal results, derived from the mathematical theory of percolation, on the basis of which more political reflection can be envisaged: ●
●
2
The size of the network is important. The smaller the network, the greater the risk of it rapidly becoming trapped in one of those ‘absorbing states’, namely, in a situation of complete agreement of all agents, from which it is difficult to collectively withdraw. The network can tolerate certain shortcomings and divergence from the openness norm. In other words, the same cognitive performance is guaranteed as long as the network is above a certain critical threshold. Co-operative behaviour can emerge and be maintained without everyone complying perfectly with the openness standard.
THE ‘GRANDE FABRIQUE LYONNAISE’: KNOWLEDGE OPENNESS OUTSIDE THE SCIENTIFIC FIELD
We have discussed ‘open science’ because it is probably the organisation of science that is closest to this standard of openness. Yet in the past there have been numerous cases of ‘open technology’, albeit limited in time and space. Historically, most situations of openness were linked to a specific territory: Lyons in the case of the circulation of techniques and inventions relating to the silk industry (Hilaire Perez, 2000), Lancashire in the case of collective invention in the metallurgical industry (Allen, 1983), the Clyde area in the case of collective invention in shipbuilding (Schwerin, 2000), the Cornish mining district in the case of collective invention related to pumping engine technology (Nuvolari, 2004). More recent cases are those of mature industries (von Hippel, 1988) as well as emerging activities such as virtual reality (Swann, 1999) or software.5 2.1
A Preliminary View of the Process of Collective Invention in Lyons6
Lyons was the second largest French town, with 143 000 inhabitants (1789) and 25 per cent were working in the silk industry (35 000). This huge sector
244
The economics of knowledge
was fostering an important internal and foreign trade for luxury silk cloth. The dominance of the French silk industry was based upon changing fabric patterns according to taste and fashion, and upon research either to invent new stitches or to set up (to ‘read’) more easily the drawings on the looms in order to make the rich cloth as quickly as possible (setting up the pattern on the loom could take 25 days). Some inventions aimed to programme the patterns on the loom and to select warp threads (like JeanPhilippe Falcon’s), others were intended to change quickly parts of the weft thread to reduce their number (Philippe de Lasalle’s movable ‘sample’) or to ease the pulling of the weft linked to the threads (Jacques Vaucanson’s hooks). Two factors were critical to induce intensive inventive activities. First, the cost of draw-girls (assistants) was a growing burden for guild families. By mid-century, these girls, who came from nearby provinces, were also very scarce. Second, the speed and synchronisation of the work became the core of inventions at the end of the century as the taste moved from brochés (heavy silk cloth with complicated patterns in gold and silver threads and many shuttles) to façonnés (lighter cloth which could be manufactured with smaller number of shuttles). This product, especially small façonnés, was the basis for successful research in suppressing the pulling of ropes. Jacquard loom (rewarded in 1804), which combined Falcon’s programme and Vaucanson’s hooks, was intended for façonnés. Nearly all Lyonnais’ inventions, which were addressed to the commerce office in Paris from 1700 to 1789, were related to the silk industry (181 in 265, for Lyon), and more precisely to weaving (116) (generally new devices of looms either for brochés or façonnés) and they occurred mostly after 1730. Lyonnais’ artisans also represented a high proportion of inventors applying to the government: there were 170 inventors from Lyons, in a total of 875 inventors (from all crafts) addressing the office of commerce. Inventor members of the Grande Fabrique were 73 and only 12 of them were large merchants. 2.2
Institutions Promoting Knowledge Openness
This innovative context was sustained by local institutions, traditionally involved in the management of innovation, since the sixteenth century, by the means of local monopolies granted in ordonnances consulaires and financial rewards. In the eighteenth century, few monopolies were granted, and there was a reward fund officially established, the Caisse du droit des étoffes étrangères, created in 1711 (from a tax upon foreign silk) and intended to promote industry since 1725.
The economics of open technology
245
This particular mechanism was designed to reward inventors who agreed to disclose their knowledge and actively to participate in the diffusion of that knowledge (teaching). The setting up of a reward fund, the process of examining inventions and the system of financial bonuses awarded to those who agreed not only to disclose but also to teach their knowledge were institutional mechanisms which made the system very effective. The system of bonuses shows how well the conditions for an efficient reproduction of knowledge, once created, were understood. From 1752, the Intendant was at the head of the Caisse and the procedure involved the business community, the local council and the academy of Lyons. The management of the Caisse was based on a contradictory proof procedure, contrived for getting more information about the invention considered, so as to reduce uncertainties and secure the public investment. This procedure was unique in France since it actually institutionalised the plurality of judgements as a method of governance: there was a double process of examination running in parallel; one involving the Intendant and a member of the academy and another involving the local council and the guild inspectors. This double procedure of examination resulted in stimulating exchanges between various kinds of competences, and compelled stakeholders to negotiate the rewards as they often reached contradictory conclusion and to mobilise their own networks. The bonus system was fostering contacts between guild inspectors and artisans as there were many visits in the workshops to quantify the spread of the new looms. Let us take the example of Michel Berthet, who, inspired by Falcon, invented a loom for easing the work of the draw-girls (an essential matter in the Lyonnaise silk industry): in 1760, the Intendant de la Michodière agreed with the academician de Goiffon to grant him £1000: £600 immediately and the rest of the sum if he taught the other artisans how to use the new loom and if four of his looms were put in other workshops. In 1765, Berthet made a new technical improvement and the Intendant proposed £1500 in exchange for the secret and for setting up some of these improved looms in town. The Intendant compelled Berthet to deposit a model and a description at the Fabrique’s office. The grants were not only rewarding the presumed economic utility of inventions; they reflected the efforts of the inventor for sharing his knowledge within the whole community. Thus, secrecy was actively opposed. There were few monopolies for invention in Lyon: nine affairs ended with a patent, concerning seven inventors. And seven of these patents were granted before 1750. The Lyonnais elites preferred to invest in innovation, to make inventions a common wealth, and this was not just a fancy ideal, as the rewards were often bonuses based upon the spreading of the inventions within the town.7 Each inventor was encouraged to be a dynamic actor collaborating for the innovation diffusion and
246
The economics of knowledge
the official credit (financial and symbolic) for the invention involved the choices and decisions of the potential adopters.
3
OPEN SCIENCE AND OPEN TECHNOLOGY
The historical analysis of open technology – in the particular case of the ‘fabrique lyonnaise’ – allows us to draw a parallel with the economics of open science. 3.1
The Basic Ingredient of Knowledge Openness
Ethos In both systems some kind of collective ethos is present, generating a sort of ‘natural’ inclination of inventors to diffuse their knowledge. In the case of Lyons, such an ethos can be seen both at the policy/administrative level and at the individual inventor level. At the policy level, the municipality, following the ancien régime tradition, kept on rewarding inventions to put them into the public domain. Much hope was placed in the free circulation of knowledge; even a slight improvement could bring about huge effects because all trades were viewed as naturally interdependent. Exclusive rights were rejected in favour of grants and bonuses for spreading knowledge and teaching. Inventors were rewarded for the practicality of their inventions and the valuation of inventions involved complex negotiations on applicability between officials and users. Liberalism, in the administrators’ eyes, meant growing exchanges between autonomous agents as a means of reinforcing social cohesion. Thus, collaboration and collective inventions were strong, because very ancient patrimonial policy was reaffirmed by new ideals and practices. This agreement between artisans, merchants and elites was essential: new techniques should not bring tensions nor disorders but, on the contrary, they should cement the social cohesion through knowledge sharing and collective emulation. At the individual level, some inventors were emblematic of this natural inclination to reveal knowledge freely. The best example is Philippe de Lasalle’s career path (1723–1804). De Lasalle was very famous in the eighteenth century, in France and abroad, and he was largely rewarded by the Grande Fabrique and the city of Lyon (£122 000). The Lyonnais elite cherished him but he devoted his effort to the progress of the whole community. Enlightened administrators like Trudaine’s son and Turgot, and writers like Voltaire, were friends of his. He belonged to the republic of arts and letters as well as to the economic world. What he did and what he thought derived from general ideals and principles he was eager to realise.
The economics of open technology
247
He began by learning drawing from local painters and became a draughtsman and a merchant. He was rewarded from 1758 by a pension for excelling in halftones for floral patterns. He also imitated tiger fur in silk cloth and he innovated by printing silk cloth like calicos. Soon after, in 1760, he was asked to teach drawing in the Fabrique and his pension was enlarged. Ten years later, his inventions for accelerating the changing of patterns on the looms (reversible loom and movable ‘sample’) increased his pension and he gained a bonus for helping to spread use of his looms. After creating machine tools to assure better diffusion of his looms, he was granted £6000 in 1778. According to the administrators and to de Lasalle himself, artistic creativity, technical invention and transmitting knowledge were closely connected. Collaborating and imitating were the main principles everywhere and the only ways to progress. Art and invention rested on a cumulative process, methods, rules, devices, lines and colours to be learnt side by side with the master, teacher, contriver or nature itself. De Lasalle had created a garden in the South of France where he sent his best students to train in drawing flowers. For him, there was no genius without copying: You are not unaware that art is learned through emulation and great examples. Work and my observations of the works of those who have distinguished themselves in the career that I follow have shaped my talents. Even more ardour to warrant the protection that you grant them can afford them one day that celebrity which offers models to imitate and stimulates other geniuses to outdo it. Thus, amongst us, as soon as a striking piece has left the hand of a skilled artist, it is lifted up to be seen by all rivals seeking the means to acquire it, and often provides, by its character, either the season’s fashion or the example of a beautiful subject. When in 1756 I treated a tiger skin worked with a touch of art on a golden background, one witnessed budding in each workshop tasteful drawings representing diverse furs. The same happened on other occasions when I introduced landscapes, birds and people.8
De Lasalle would not condemn the theft of patterns or inventions; his aim was the circulation of knowledge and the progress of qualifications which could result. He was even pleased when his printed silk cloth was copied and his workers seduced by rivals. All means were good if diffusion were at stake: teaching, imitating, stealing and, not least of all, deeds and free offers. Several times, de Lasalle gave away inventions and taught about his new device without asking for anything in return. In 1760 he was offered a £200 bonus for each student he taught, but he refused and preferred to offer all his knowledge freely: ‘it appears . . . that he gives up the gratification of 200 pounds for each of his six students and, moreover, that he teaches them everything learned from years of experience’ (note of the Lyons’council, 1760). How such an ethos appears and becomes forceful is
248
The economics of knowledge
a broad question, addressed for instance by Hilaire Perez (2000) in the case of the fabrique lyonnaise. Collective belief of being part of a positive sum game Similar collective belief, in both open science and open technology, of being part of a positive sum game plays a key role as well. A common knowledge that open technology is a positive sum game was particularly effective and ‘had force’ in the case of Lyons since the city was engaged in international competition with London and the inventors knew full well that the prosperity of the local system to which they belonged directly influenced their own individual prosperity. Individual rewards Both collective ethics and common knowledge about the efficiency of open technology are not enough to sustain a system based on the free dissemination of knowledge. There is a need for some kind of mechanism aiming at rewarding inventors without granting exclusive rights. We have presented above the particular mechanisms that were designed in Lyons to reward inventors who agreed to disclose their knowledge and actively to participate in the diffusion of that knowledge. 3.2
System’s Efficiency
The efficiency of systems of open technology is similar to the efficiency of open science: both are a way to increase the performance of a system of invention by making the existing stock of knowledge more socially useful, through improved transfer, transformation and access to the existing innovations. In Lyons a good example is the diffusion of the Jacquard loom. Massive diffusion of new technologies For the nineteenth century, A. Cottereau (1997) documented the massive diffusion of the Jacquard loom in Lyons (20 000 existed at the mid-century) and he compared this success to London where ‘sweated’ labour conditions, specialisation and private strategies obstructed its dissemination. Cottereau (1997) explains that the London and Lyons silk manufacturing were based on a similar amount of machines: there were 12 000 looms in London in 1815 and 14 500 in Lyons. Though London could compete with Lyons between 1790 and 1810, because the French revolutionary crisis disrupted production for a while, the London silk industry began to decline as French enterprises used Jacquard’s loom to make a revival of sophisticated and varied silk production. In London, only 5000 looms could be found in 1853; in Lyon, there were 30 000 (and 30 000 more in rural areas
The economics of open technology
249
outside the town). Before the First World War, French production exceeded that of England and most of it was exported, while England was importing substantial amounts of silk fabrics for home consumption. In London, Jacquard did not spread and generally speaking, there were not many inventions in the London silk industry (Cottereau even speaks of the ‘backwardness of all British handlooms’). As for the Jacquard loom, its introduction gave rise to a frantic race between important manufacturers; one, G. Wilson, succeeded, was allowed to keep the secret of the machine, took a patent in 1821 and did not sell the invention or build new looms. Cottereau does not mention any use of licences, though they existed in the cotton industry which was the model referred to by silk manufacturers. Cumulative knowledge Jacquard’s invention matched the needs of the Lyonnaise silk industry. The new loom immediately spreads and this creates a mental mobilisation and a collective focus resulting in several useful improvements. Other loombuilders made hundreds of Jacquard’s loom, compared to the first inventor who built only 57. These networks were the basis for the pattern of innovation in Lyon. Inventive artisans, either weavers or not, were quickly informed of new devices; they watched working new looms, listened to weavers, talked with other artisans, worked on rewarded looms and contrived improvements to them. The open system generated huge cumulative effect. For example, already in 1759, Berthet had presented an improvement of Falcon’s 1742 loom. In 1765, he said he had improved the new Falcon’s loom he had just acquired. And the Jacquard’s invention itself was often called ‘Vaucanson–Falcon–Breton’ to show how important was knowledge recombination in the production of this major invention. Jacquard really stood on the ‘shoulders of giants’. There were, in Lyons, many other examples of cumulative progress involving successive improvements of a new loom. Moreover, inventors like Falcon and de Lasalle kept improving their own devices. One invention was never definitive but always evolving, and these improvements were encouraged by the local council, which, for instance, blamed Jacquard’s disinterest in amending his own loom. Technical standards and intergenerational compatibility Another positive effect of knowledge openness was the establishment of technical standards. The historian Cottereau (1997) found an essay written in 1863 describing the networks of newly invented looms in Lyons: ‘The most convincing proof that these successive inventions were borrowed from one another is that a Jacquard card in use today may be applied both to Vaucanson’s planchette with needles and to Falcon’s, and the match is so good that Falcon’s initial matrix must have fixed dimensions’ (Cottereau,
250
The economics of knowledge
1997, p. 143). According to Cottereau, the effects ‘were comparable to what could easily have been the case today if computer systems had been standardised from the start and made cumulatively compatible as they progressed’ (ibid., p. 142), even if contrived by several different firms. Then, collaboration and open technology in Lyon was highly efficient for the spreading of inventions, for sharing technical innovative culture and for helping research in craftwork. This model had fostered a professional elite of indefatigable researchers, skilled inventors and artisans, ‘artists’, who often devoted more time to research than to their own business. Fame, excellence and performance were these inventors’ aims. But what was the boundary with self-strategies? Although Cottereau describes an equilibrium in the Lyon industry, conflicts and private interests were very harsh. In a paradoxical way, collective innovation did usher in a disruption of community ethos; it did foster a burst of opportunism, and, most of all, claims for priority and posterity amongst inventors. 3.3
The Metaphor of the Jacquard Flight
In both cases – open science and open technology – the reward system introduces competition and increases the risk of disputes. Then the force of ethics as well as the effectiveness of the common knowledge about the efficiency of the system come to the fore to mitigate individual misconducts and frustrations. Indeed, the ‘collective fabrique’ appears very fragile, and somewhat vulnerable to individual claims, frustrations and hopes. Jacquard agreed initially to give up his rights to patents and ‘left the fruit of his art to the community’ (Cottereau, 1997, p. 151). The invention became the property of the town and quickly spread. But later on, Jacquard started to complain that the Lyonnais administrators had not treated him well enough, considering the importance (‘the social return’) of his invention. A conflict arose between the great inventor and the local council, which compelled him to stay in Lyons, fearing that he would sell the invention to competitors. In 1814, Jacquard left Lyons to go to Paris where he wanted to patent his invention. The police of Lyons were urged to take him back and to check if he had transmitted his invention to rivals! The history of Jacquard’s flight is a good metaphor to capture all situations where the coexistence of different incentive systems makes fragile those not based on private property.9 There was a degree of fragility in the Lyons system of knowledge openness, especially when areas close by (Paris, in this case) offered inventors the possibility of obtaining a patent. This last point is particularly important. Apart from the beauty of systems of collective invention and the fine economic performance such systems can
The economics of open technology
251
produce, the individual incentive dimension remains decisive and calls for substantial institutional mechanisms to give credit to inventors without granting them exclusivity. This is the kind of mechanism Dasgupta and David have explored in the case of open science and which remains uncertain in the case of open technology, although the case of the ‘fabrique lyonnaise’ provides some ideas about how such credit and reward incentives might need to be structured in order to support an ‘open technology’ environment. The prize system is an efficient mechanism because it creates incentives while keeping the knowledge in the public domain. However, the amount of the reward should be equal to the social surplus afforded by the invention. As this case shows, the ex ante prediction of the social value of the invention is not a trivial condition.
4
CONCLUSION
In this chapter we tried to show that there was a golden age in the regulation of the production and distribution of silk loom technology in Lyon which resembled the open science structure but made specific accommodation with the practical needs of inventors to be rewarded in excess of what they might earn by retaining their knowledge as a secret. The result was effective diffusion and economic growth for the whole local system of innovation. This system then began to break down in the nineteenth century as merchant pressures of various sorts increased and as vertical class conflict (between masters and merchants) became more important. This story also raises the issue of the type of knowledge which is relevant to the kind of reward system that has been described here.10 This was because of the ability to implement the knowledge investment in superior looms that inventors were rewarded. Thus, la fabrique lyonnaise can be qualified as a collective enterprise in which the rewards were distributed according to contribution to the collective benefit. This would suggest that there was a distinct bias in the types of knowledge that could be relevant to the reward system, and another possible source for decay. When it becomes possible to make inventions through knowledge rather than the practice of the mechanical arts, it is no longer possible for the community to retain exclusivity over the machinery. Indeed, the machinery now becomes subject to an innovation process that lies outside the direct and co-evolving experiences of machine construction and use. If true, then it would follow that a peculiar condition of the ‘open technology’ is the necessity for the division of labour to be limited so that the externalities of knowledge production (invention) can be captured as a local externality. Once the possibility of broader externalities comes into play the dynamic interaction collapses and
252
The economics of knowledge
we are reduced to the model of commodified knowledge which is what the patent system offers. We hope that this study opens new research avenues – historical as well as analytical – about these classes of mechanisms which allowed credit to be given to individual inventors while supporting strongly the disclosure as well as the reproduction of knowledge; those mechanisms which strictly govern the solidity and stability of open systems, as Paul David so clearly demonstrated in the case of science and academic research.
NOTES 1.
2.
3. 4. 5.
6.
7.
The present version of this chapter has benefited from comments received on an earlier draft, especially those from Cristiano Antonelli, Paul A. David, Alfonso Gambardella, Bronwyn Hall, Jacques Mairesse and Peter Swann at the conference ‘In honour of Paul A. David’ (Turin, May 2000). We are grateful to all the aforementioned and we are particularly grateful to Ed Steinmueller who made numerous suggestions to clarify the exposition and greatly help us to improve the final draft. In economic terms, since the marginal cost of use of knowledge is nil, maximum efficiency in its use implies that there is no restriction to access and that the price of use is equal to 0. Knowledge should be a ‘free’ good; that is, the condition for optimum use of a non-rival good. Note that the union between research and teaching is not always maintained, which is what determines the partition of the public research system between universities and national (or regional) laboratories. This argument comes from an oral comment by Iain Cockburn made at an OECD/CERI meeting (Paris, July 2001). As defined in a recent research project (Cassier and Foray, 1999; Foray and Steinmueller, 2002) the type of open knowledge we are dealing with in this chapter is different from the collusive and explicit forms of collective invention (such as high technology consortiums) which require explicit co-ordination mechanisms as well as the formalisation of agreements on both the distribution of tasks and the attribution of results. Moreover collusive forms delimit semi-private areas for the circulation and pooling of knowledge, which may in some cases be less open than informal networks we are studying here. The main difference between these two types of collective enterprise deals with the mode of production of knowledge. In the cases studied here, trading or sharing concerns knowledge that is already available. The participants do not participate in a co-ordinated research project; they trade or share existing technical data. This is an incremental process based on the dissemination and reuse of knowledge available within a group of firms. In the case of collusive and explicit forms of collective invention, the actors engage in operations of knowledge production. Our analysis (extensively presented in Hilaire Perez, 1994; 2000) is based upon archival sources that were not fully exploited in the previous studies on the fabrique lyonnaise (Cottereau, 1997; Poni, 1998): the letters and reports relating to eighteenth-century Lyonnais inventors’ claims for grants and privileges were adjudicated both in Lyons and in Paris by the Bureau du Commerce. For instance, in 1760, Ringuet presented a new loom for brochés which imitated paintings and embroidery; he was granted a £300 (livres tournois) bonus for the 10 first looms set up, £200 for the next 10 looms and £100 for the 100 next ones during 10 years. He was very successful: as early as 1760, he had set up the first 10 looms; in 1762, the next 10 and even 17 more; in 1763, 47 others and, in 1764, 85. Thus Ringuet had even passed the
The economics of open technology
8. 9. 10.
253
quota (169 instead of 120, and in less than 10 years). He was paid for all the looms, even the ones which were not planned in the grant (£19 900 instead of £15 000). Translation by Liz Carey-Libbrecht. Quoted in Hilaire Perez (2000), p. 76. Thanks to Bronwyn Hall for her help in deciphering this case as a very powerful metaphor for the economics of knowledge production and diffusion. We are grateful to Ed Steinmueller who so well captured this issue and shared it with us.
REFERENCES Allen, R. (1983), ‘Collective invention’, Journal of Economic Behavior and Organization, 4, 1–24. Arrow, K.J. (1962), ‘Economic welfare and the allocation of resources for inventions’, in R.R. Nelson (ed.), The Rate and Direction of Inventive Activity: Economic and Social Factors, Princeton, NJ, Princeton University Press. Callon, M. and Foray, D. (1997), ‘Nouvelle économie de la science ou socioéconomie de la recherche scientifique?’, Revue d’Economie Industrielle, 79, 13–35. Cassier, M. and Foray, D. (1999), ‘The sharing of knowledge in collective, spontaneous or collusive forms of invention’, Colline WP 02, IMRI, University Paris Dauphine. Cottereau, A. (1997), ‘The fate of collective manufactures in the industrial world: the silk industries of Lyons and London, 1800–1850’, in C.F. Sabel and J. Zeitlin (eds), World of Possibilities: Flexibility and Mass Production in Westerns Industrialization, Cambridge, Cambridge University Press. Dasgupta, P. and David, P.A. (1994), ‘Toward a new economics of science’, Research Policy, 23 (5), 487–521. David, P.A. (1998), ‘Common agency contracting and the ‘emergence of “open science” institutions’, The American Economic Review, 88 (2), 15–21. David, P.A. (1999), ‘Patronage, reputation and common agency contracting in the scientific revolution: from keeping “nature’s secret to the institutionalization of open science” ’, All Souls College, Oxford, December. David, P.A. and Foray, D. (1995), ‘Accessing and expanding the science and technology knowledge base’, STI Review, 16, 13–68. David, P.A., Foray, D, and Steinmueller, W.E. (1999), ‘The research network and the new economics of science: from metaphors to organizational behavior’, in A. Gambardella and F. Malerba (eds), The Organization of Inventive Activity in Europe, Cambridge: Cambridge University Press. Foray, D. and Steinmueller, W.E. (2002), ‘On the economics of R&D and technological collaborations: insights from the project Colline’, Economics of Innovation and New Technology, 12 (1), 77–97. Hilaire Perez, L. (1994), ‘Inventions et inventeurs en France et en Angleterre au XVIIIè siècle’, Doctorat de l’Université de Paris I, Atelier National de Reproduction des Thèses, Lille III. Hilaire Perez, L. (2000), L’invention technique au siècle des lumières, Paris, Albin Michel. Nuvolari, A. (2004), ‘Collective invention during the British industrial revolution: the case of the Cornish pumping engine’, Eindhoven Centre for Innovation Studies.
254
The economics of knowledge
Poni, C. (1998), ‘Mode et innovation: les stratégies des marchands en soie de Lyon, XVIII%’, Revue d’Histoire Moderne et Contemporaine, 45–3, 25–58. Schwerin, J. (2000), ‘The dynamics of sectoral change: innovation and growth in Clyde shipbuilding, 1850–1900’, 8th International J.A. Schumpeter Society Conference, Manchester, June. Swann, P. (1999), ‘Collective invention in virtual reality’, Colline WP 05, IMRI, University Paris Dauphine. Von Hippel, E. (1988), ‘Trading trade secrets’, Technology Review, February/ March, 58–64.
10. Measurement and explanation of the intensity of co-publication in scientific research: an analysis at the laboratory level* Jacques Mairesse and Laure Turner 1
INTRODUCTION
Since the scientific research system has become an essential sector in our modern knowledge-based economies, an important new research field has opened up. The challenge is to illuminate the role of scientific institutions in the production, diffusion and transfer of knowledge and that of science in economic development and social welfare. The ‘new economics of science’ therefore is interested in a variety of issues concerning the functioning of scientific institutions, the labor market, training and careers of scientists, their productiveness, the allocation of public funds to basic research, the design of intellectual property rights, and so on. It thus contributes to the understanding of the organization of science and of ways it can be improved (Dasgupta and David, 1994; Gibbons et al., 1994; Diamond, 1996; Stephan, 1996; Callon and Foray, 1997; Shi, 2001; Foray, 2004). The analysis of co-publications between scientists presented in this contribution is in keeping with the main focus of the economics of science on knowledge production, and is part of a broader study of the determinants of scientific research productivity. We believe that membership in a dynamic and productive laboratory favours collaboration between researchers and improves their own individual productivity, and that it may be part of a process of cumulative advantage by which these researchers enhance their productivity and reputation.1 Given the substantial increase in the proportion of co-authored articles, it also seems that the relevant units of knowledge production tend to be more and more specific networks of researchers, whether they belong or not to the same institutions and/or countries (Gibbons et al., 1994).
255
256
The economics of knowledge
In the economics of science, until recently, the literature on the interactions that favour knowledge production and diffusion primarily concerns geographic externalities. Authors have mainly studied such externalities within industries or from universities and other public research institutions to firms and industries, relying on the analysis of patent data.2 Our work moves upstream to study knowledge externalities within the scientific research system using co-publication data. We wish to look beyond the observation of the spatial dimensions of research activity to investigate the determinants of the occurrence and intensity of collaborative relations between researchers. Audretsch and Stephan (1996) have done similar work but, again, concerning the relations between public research and industry. Based on data on academic scientists collaborating with US biotechnology firms, they show that such collaboration between firms and researchers is more likely when the researchers have a good academic reputation, when they belong to a geographically extensive network, and when they are involved in practice in the transfer of knowledge towards the firm (as participants in the creation of the firm or as members of the Scientific Advisory Board). Regional and local characteristics also seem to influence the strength of relations between scientists and firms. In the sociology of science and in bibliometry, a number of studies have already highlighted some of the factors facilitating collaboration within academic research (see Beaver and Rosen, 1978 and 1979, and Katz, 1994, for a summary presentation). They include, above all, researchers’ reputation and visibility, the need to access or to share the use of specific research instruments and facilities, the increasing specialisation in science and geographic proximity. Two types of analyses can be found, however, in this literature, depending on their explicit or implicit conception of a network (Shrum and Mullins, 1988). In one line of analysis the actors in networks are identified through their interrelations, being mainly differentiated by their different positions in the structural configuration of their networks (for example, whether they occupy a central position or not), not by their individual characteristics such as age, gender or skills.3 By contrast, the second line of research takes into explicit account the status, capacities and strategies of actors, and it is these individual characteristics that mainly determine the position of agents in networks and the nature of interactions between them.4 Yet it would be desirable to be able to include in the same analysis structural and individual elements as determinants of network interactions, and particularly for collaboration in research. Knowledge production and diffusion are based on the interactions of multiple agents and institutions with diverse interests: scientists in public and private laboratories, firms, financiers, public authorities, and so on (Callon, 1999). Investigating the existence and intensity of collaboration between researchers in relation to
The intensity of co-publication in scientific research
257
their specific characteristics should afford insight into the various mechanisms at play. In this chapter we present the first results of such an attempt. We propose an intensity measure of collaboration between researchers, which has an intuitive interpretation and can be simply aggregated to the laboratory level or higher levels of aggregation. Our unit of analysis in this contribution is the laboratory and the group of laboratories at the geographic level of a town (which to be short we will call ‘town’). Our purpose is to explain measured differences of intensity of collaboration as revealed by co-publications by various potential determinants: precisely the geographic distance between laboratories, their thematic specialisation, their size, their productivity in terms of average number of publications per researcher, their quality in terms of average citation impact factor per publication, and their international openness.5 In particular, to what extent does the geographic distance between researchers and their laboratories strongly impede, or not, their scientific collaboration? Our approach is basically descriptive. We measure the intensity of copublications among the researchers of the French Centre National de la Recherche Scientifique (CNRS) in the field of condensed matter physics, during the six-year period 1992–97.6 We first estimate the intensity of copublication among these researchers, within their laboratories and between them, and also within and between the towns in which these laboratories are located. Next we consider by means of simple correlations the possible influence of geographical distance and other determinants on the occurrence and intensity of co-publication. We then try to better assess the specific impacts of these different factors by estimating their relative weight in a regression analysis. The chapter is organised as follows. In section 2, we give necessary information on the scope of our study, the construction of our sample, and some descriptive characteristics of co-publication. In section 3, we define our measure of intensity of collaboration, giving a detailed example of its computation. In section 4, we present our correlation and regression results, and comment on what they tell us of the respective importance of the various determinants of co-publication we have been able to consider. We briefly conclude in section 5.
258
2
The economics of knowledge
SCOPE OF THE STUDY AND GENERAL CHARACTERISTICS OF CO-PUBLICATION
2.1 Scope of the Study: The Collaboration between CNRS Researchers in Condensed Matter Physics In this chapter we study the determinants of collaboration among a group of 493 physicists belonging to the condensed matter section at the CNRS, over the six-year period 1992–97. This sample consists of practically all the CNRS physicists in this field who were born between 1936 and 1960 and were still working at the CNRS in 1997.7 Condensed matter physics investigates, at various scales (atom, molecules, colloids, particles or cells), all states of matter from liquids to solids in which molecules are relatively close to each other. It is based on a heritage of traditions, both experimental (crystallography, diffusion of neutrons and electrons, magnetic resonance imagery, microscopy, and so on) and theoretical (solid state physics). It has recently developed a closer relation with industry, contributing to the development of materials used in electronics, plastics, food or cosmetic gels, and so forth. We chose condensed matter physics for three main reasons. First, the characteristics of this field are particularly well suited to our study: it is a domain of basic research, which is clearly defined and where the journals with a sound reputation are easily identifiable. Second, condensed matter is a fast-growing field, honoured by the Nobel Prize for Physics awarded to Pierre-Gilles de Gennes in 1991, and currently accounting for close to half of all French research in physics. Third, there is relatively little mobility among CNRS researchers outside of the field to other fields of research in CNRS, or out of CNRS towards academia or industry. The sample of 493 physicists studied here represents a majority of all CNRS researchers in the field. The CNRS and higher education institutions are the only public research institutions in this field in France. In 1996, there were a total of 654 condensed matter physicists in CNRS, as against 1475 in universities and ‘Grandes Ecoles’ (Barré et al., 1999). The fact that our study is limited to researchers belonging to the same institution, the CNRS, comes as an advantage. It implies a strong organisational proximity between the researchers, characterised by the sharing of common knowledge and implicit or explicit rules of organisation that favour interaction and co-ordination (Rallet and Torre, 2000; Foray, 2004). Because they all belong to the same scientific community within the same institution, they work in a context directly conducive to co-operation that does not involve prior agreement on rules of behaviour. The existence of such strong organisational proximity thus makes it possible to isolate more clearly the effects on collaboration of geographic distance proper and other factors.
The intensity of co-publication in scientific research
259
The indicator of collaboration that we use in this study is co-publication. It seems to be a reliable indicator of collaboration without being an exhaustive measurement, in so far as collaboration can have results other than publications. Our database has been compiled on the basis of all the publications drawn from the Science Citation Index (SCI), for 518 CNRS condensed matter physicists over the period 1992–97, of whom 493 published at least one co-authored article during this six years.8 Of the remaining 25 physicists, 21 published no articles in this period, and the other four published only a total of five non-co-authored articles. Collaboration appears to be the main mode of publication for the 493 researchers. Only 132 of them also wrote articles without co-authors over the period (for a total of 252 articles) and, from the total corpus of 7784 articles they wrote over the period, 7532 (97 per cent!) are co-authored. In order to improve the measurement of the intensity of collaboration in our analysis, we thought it appropriate to weight co-authored articles in proportion to the number of pairs of co-authors they involve. In other words, we simply chose to study the network of collaboration ‘link by link’, that is, by pairs of co-authors. In practice, this means that an article appears in the database we constructed as many times as the number of different pairs of its CNRS co-authors.9 We also chose to centre our study at the level of the laboratory, and even at the more aggregate level of groups of laboratories in the same towns or localities (‘towns’). We thus consider networks of collaboration between laboratories and towns rather than directly between individual researchers. When two researchers belonging to different laboratories (towns) collaborate, we consider that these laboratories (towns) collaborate, and on this basis we can measure the intensity of collaboration between laboratories (towns). When two researchers belonging to the same laboratory (town) collaborate, we also simply consider it as a case of collaboration ‘within’ this laboratory (town), and likewise we compute the intensity of collaboration within laboratories (within towns). We can also similarly compute intensity of collaboration between laboratories-within towns. Carrying out our study at the aggregate level of laboratories and towns simplifies the analysis and makes the use of our measure of collaboration intensity perhaps more convincing, since networks of collaboration are, of course, much denser at these levels than at the individual researcher level. But, as we shall see, it also has the advantage that it allows for a direct characterisation of the influence on collaboration of working in the same laboratory or town, and thus of the importance of spatial proximity and easy face-toface relations.
260
2.2
The economics of knowledge
Two Configurations of Co-publication
The co-authors of the articles of our group of 493 CNRS researchers, whom we will simply call ‘CNRS researchers’ from now on, can be (these) CNRS researchers themselves, or other researchers, mainly belonging to universities or other institutions, either French or foreign, whom we will call ‘external researchers’. In our analysis, we are led to distinguish between two configurations of co-publications, depending on whether a publication involves at least two CNRS researchers and possibly other researchers (CNRS or external), or whether it concerns at most one CNRS researcher and one or more external researchers. An important reason for this distinction is a practical one. We not only preferred a priori to focus our analysis on the collaboration among CNRS researchers in the same field, but also we could not extend it in practice to the external researchers in this field. The CNRS researchers were the only ones for whom we could have access to the name, location and some characteristics of their laboratories, in addition to their individual characteristics (age, gender, seniority, and so on).10 This was not possible for the external researchers since we could not even retrieve the name and location of their laboratories with sufficient reliability from the SCI.11 We were thus left with much more limited information for them than for the CNRS researchers and their laboratories. Our group of 493 CNRS researchers generally co-publish both with the other CNRS researchers in the group and with external researchers. As indicated in Figure 10.1, 38 of them are collaborating only with CNRS researchers (never with external researchers), and 69 are collaborating only with external researchers (not with the other CNRS researchers), and thus 386 ( 493 38 69) are collaborating in both ways. The first configuration of co-publication (involving at least two CNRS researchers and possibly other researchers) corresponds to ‘Group 1’ with a total of 1823 articles ( 1741 82), while the second (involving only one CNRS researcher with external researchers) corresponds to ‘Group 2’ with a total of 5709 articles ( 5012 697). Group 1 thus concerns 424 of our CNRS researchers (
493 69), while Group 2 concerns 455 of them ( 493 38). Table 10.1 shows the two-way distribution of the number of articles in Group 1 and Group 2 with respect to the number of their CNRS authors and that of their external authors (see also the two related distributions shown in Figure 10.2). We observe immediately that, for both Group 1 and Group 2 articles, collaboration generally involves several ‘external’ researchers (82 articles are written by CNRS researchers only!). We can also note that most articles of Group 1, which have at least two CNRS coauthors, do not involve a third (or more) CNRS co-author (1498 out of 1823). Thus, the average number of authors per article for Group 1 is 5.9,
261
Figure 10.1
Collaboration only with ‘others’ and never with CNRS researchers *** 697 articles
69 researchers
Collaboration in both modes: At least 2 CNRS and ‘others’ (Group 1) or at most 1 CNRS and ‘others’ (Group 2) *** 6753 articles (1741 for Group 1 and 5012 for Group 2)
386 researchers
Choosing the sample
Collaboration only with CNRS researchers and never with ‘others’ *** 82 articles
38 researchers
• 7532 articles co-authored • 252 articles alone
493 CNRS researchers
262
196 31 2 1 726 956
64 15 3 0
0
82
1411
1087
268 45 9 2
324
2
1489
1114
300 60 12 3
375
3
1236
976
218 31 8 3
260
4
917
708
172 33 4 0
209
5
568
441
106 15 6 0
127
6
322
241
61 15 4 1
81
7
184
128
47 6 3 0
56
8
367
288
66 7 4 3
161
9
5709 (455**) 7532 (493)
1498 257 55 13
1823 (424*)
Total
Notes: ( ) The three numbers in parentheses below the numbers of articles are the numbers of CNRS researchers co-authoring these articles. * Including 38 CNRS researchers who never published with external researchers and who account for 82 publications of Group 1 articles. ** Including 69 researchers who never published with other CNRS researchers and who account for 697 publications of Group 2 articles. 132 CNRS researchers who also published co-authored papers have written 252 articles alone (not included in the first group or second group of articles).
230
82
Group 1 (at least 2 CNRS co-authors) Of which: 2 CNRS co-authors 3 CNRS co-authors 4 CNRS co-authors 5 or more CNRS co-authors Group 2 (only 1 CNRS co-author) Total (Group 1 and 2)
1
0
Number of articles in Group 1 and Group 2 by number of CNRS and external co-authors
Number of ‘external’ co-authors:
Table 10.1
The intensity of co-publication in scientific research
263
of which 2.2 are CNRS researchers and 3.7 external researchers, and for Group 2 it is 4.9 (that is, 1 CNRS researcher and 3.9 external researchers). 2.3
The Selected Sample of Co-publications and Some Characteristics
Four main reasons determine our choice of limiting our analysis to the first configuration of co-publications and to the Group 1 sample of articles. The first reason, which we have already stressed, is analytical. By studying copublication between couples of CNRS researchers, we control for institutional and organisational proximity resulting in ‘common knowledge’ of rules and practices and strongly favouring collaboration. Organisational proximity and geographical proximity usually being confounded, this has the great advantage of allowing us to unravel clearly the impact of the latter on collaboration. The second reason, which we also mentioned, is simply that we cannot identify precisely enough the laboratories of the ‘external’ researchers, and thus cannot locate them or characterise them, as we can do for the laboratories of the CNRS researchers. But there is a third important reason of an empirical nature for focusing our investigation on the collaboration between CNRS researchers. The occurrence of co-publication between a CNRS and an external researcher is extremely low, while it is much higher, as we would expect, between couples of CNRS researchers. The 1823 articles in Group 1, written by 424 CNRS authors and about 3500 external co-authors, actually involve only 880 different couples of CNRS researchers out of the 89 676, ( 423 ' 424/2) number of potential couples (that is one out of 100). The 5709 articles in Group 2, written by 455 CNRS authors with close to 10 000 external co-authors, involve by contrast as much as 17 500 couples of a CNRS researcher with an external researcher, out of the 4 550 000 potential couples (that is only four out of 1000). Thus, the average number of articles per effective couple of co-authors is 2.1 in Group 1 and only about 0.3 in Group 2. Likewise the probability (frequency) of a CNRS author having another CNRS co-author (in Group 1) is much higher than having an external co-author (in Group 2): 0.021 as against 0.001. A last consideration arises from the fact that some characteristics of copublication in Group 1 and Group 2 are nonetheless close enough. This suggests that, hopefully, a number of the results we find in the analysis of co-publication between CNRS researchers might not be too different from those we would have obtained if we had been able to extend the analysis to the co-publication with non-CNRS researchers. This is clear for the three following characteristics that we can compute for the sample of 6753 articles published by the 386 CNRS researchers involved in both types of
264
Frequency of articles in %
The economics of knowledge 20 18 16 14 12 10 8 6 4 2 0
Group 1
0
1
2
3
Group 2
4 5 6 7 8 9 10 11 12 13 14 15 Number of external co-authors
Figure 10.2 Frequency of the number of articles written by CNRS researchers with external co-authors in Groups 1 and 2 of articles publication (see Figure 10.1). The first of these is the frequency distribution of the number of articles per number of external co-authors. As shown in Figure 10.2 the probability (frequency) that an article is coauthored by a given number of external researchers is nearly the same in the two groups of articles. The second very close characteristic concerns the degree of concentration of the number of articles published in the two groups of articles by the CNRS researchers. As shown in Figure 10.3 the concentration curves practically coincide in both cases, with nearly 40 per cent of the articles being co-authored by 10 per cent of the most productive CNRS researchers, and about 80 per cent by the more productive half of them. Yet, as can be seen in Figure 10.4, the distribution of the number of articles written per CNRS researcher (our third characteristic) differs somewhat for the two groups of articles. During the six-year period 1992–97, the cumulative probability that a CNRS researcher publishes less than six articles in Group 1 is 50 per cent, while it is about 35 per cent in Group 2. Likewise, during this period, a CNRS researcher published an average of 9.9 articles in Group 1, as against 13 in Group 2.12 2.4
Other Restrictions on the Selected Sample
In practice, in order to avoid having laboratories and towns with too few CNRS researchers we thought it better to put two further restrictions on our
265
Cumulative frequency of the number of articles in %
The intensity of co-publication in scientific research 100 90 80 70 60 50 40 30 20 Group 1
10
Group 2
0 0
10
20 30 40 50 60 70 80 90 Cumulative frequency of the number of CNRS researchers in %
100
Cumulative frequency of articles in %
Figure 10.3 Concentration curves of the number of articles written by CNRS researchers in Groups 1 and 2 of articles 100 90 80 70 60 50 40 30 20 10 0
Group 1 0
5
Group 2
10 15 20 25 30 35 40 45 50 Number of articles of CNRS researchers
55
60
Figure 10.4 Distribution of the number of articles written by CNRS researchers in Groups 1 and 2 of articles sample. We imposed that laboratories in our sample had at least five CNRS researchers, and towns had at least nine CNRS researchers. Our final sample thus consists of 470 CNRS researchers in condensed matter physics (out of the initial group of 493), located in 34 laboratories and 17 towns. Likewise,
266
The economics of knowledge
in our analysis, we thought it better to avoid characterising collaboration between two laboratories, or collaboration between two towns, on the basis of too few co-publications between their CNRS researchers. We thus defined collaboration between a couple of laboratories as involving more than four co-publications over the six-year study period, and between a couple of towns as involving more than six co-publications. These two types of restrictions had the consequence of limiting also the number of Group 1 articles (with at least two CNRS co-authors), on which our analysis concentrates, to 1634 articles (out of 1741). To summarise, our investigation is thus mainly based on a sample of 470 CNRS condensed matter physicists (located in 17 towns and 34 laboratories) and a sample of 1634 articles they have co-published over the period 1992–97.
3
MEASUREMENT OF INTENSITY OF COLLABORATION
The behaviours of agents in networks is determined by ‘intrinsic’ individual characteristics such as age, gender, skills, motivations and objectives, and by more ‘structural’ variables such as the density of their networks, their more or less central or peripheral situation, geographic distance, and so on. As a result, the form and functioning of networks differ. If the actors were not differentiated and if they collaborated with all the others with equal probability, we would expect to observe a uniform structure of relations between all the individuals. We take this extreme case of ‘homogeneity’ as a reference. At the aggregate level of entities such as the laboratories and groups of laboratories (towns) on which we centre our analysis, the case of homogeneity corresponds to a configuration in which the frequency of collaboration of agents, the CNRS researchers, is the same, irrespective of the entities to which they belong, their geographic localisation and other characteristics. Our simple measure of (relative) intensity of collaboration between two entities is simply based on the comparison between the real network as portrayed by the data and the network that would be observed in the hypothetical case of homogeneity. We generally define this measure in sub-section 3.1, and comment on its aggregation properties and on the weighting issues in sub-sections 3.2 and 3.3. In sub-section 3.4 we then provide a detailed example of its calculation. 3.1
Definition
In this sub-section we assume for simplicity that collaboration always involves at the most two (CNRS) researchers (this assumption is discussed
The intensity of co-publication in scientific research
267
in the next sub-section). The overall or ‘complete’ network of collaboration studied has a finite number of entities (laboratories or towns) consisting in total of N researchers who can form C collaboration pairs, or couples, where by definition C N(N 1)/2, the total number of possible pairs. Let n be the total number of articles produced in collaboration between the N researchers, then p, the frequency of the number of co-publications per pair in the complete network, is the ratio between the total number of articles n, and the number of possible pairs C, that is, p n/C. Using similar notations at the level of the network’s entities, consider now two entities X and Y, where NX and NY are the numbers of researchers working in them respectively. The numbers of possible pairs of researchers within X and within Y are respectively CX NX(NX 1)/2 and CY
NY (NY 1)/2, and the number of possible pairs that can be formed between researchers from X and Y is CXY NXNy. If the total numbers of articles written jointly within X and Y are respectively nX and nY, and the total number of articles written in common by researchers in X and Y is nXY, the frequencies of collaboration within the entity X and Y, noted as pX and pY, are the corresponding ratios between the total number of articles nX and nY written together by researchers from X or Y, and the number of possible pairs CX and CY of researchers in entity X and Y, that is pX nX/CX and pY nY/CY. Similarly the frequency of collaboration pXY between the two entities X and Y is the ratio between the total number of articles nXY written in common by researchers in X with researchers in Y, and the number of possible pairs CXY of researchers from the two entities, that is pXY
nXY /CXY. The intensity of collaboration relates the frequencies obtained at the entities’ level to the frequency p obtained for the complete network. We thus respectively define the two intra- or within-intensity and the inter- or between-intensity as: iX
n C n C pX nX CX p p
iY Y Y Y iXY XY XY XY p nC p nC p nC
Note that in what follows we will be using indifferently the expression intraor within-intensity, and inter- or between-intensity. In the reference case of homogeneity of the network we have pX pY
pXY for all X and Y, and consequently we can see that pX pY pXY p, or in terms of the intensity measure: iX iY iXY 1. In the case of homogeneity, the frequencies of collaboration intra- and inter-entities are all equal to the overall frequency p for the network, and the intra- and interintensities of collaboration are all equal to unity. Otherwise, in the case of a real network, as the one we are considering, various factors influence
268
The economics of knowledge
intensities of collaboration; we can expect them, of course, to be very different from unity, which can be viewed as an average benchmark value. Note that another way of looking at our measure of intensity of collaboration of an entity is to interpret it as its contribution of co-authored articles nX to the total number n of co-authored articles in the network, normalised by its size relative to that of the complete network measured in terms of possible pairs of co-authors, that is, iX pX /p (nX/n)/(CX/C). Note also that the structure of intensity of a network of E entities can be represented by means of a symmetrical matrix E by E with positive or zero coefficients where the diagonal terms are equal to the intra-entity intensities and the off-diagonal terms are equal to the inter-entity intensities.13 Appendix 1 gives this matrix for the 17 towns in our sample. 3.2
Aggregation Properties
The (relative) intensity of collaboration as defined above has the advantage of being easy to aggregate at different levels of analysis. In order to see this, suppose that V is a town with two laboratories, X and Y. The total number of co-authored articles written in V is the sum of co-authored articles by researchers from X and Y separately, and from X and Y jointly. Likewise, the number of possible pairs of researchers in V is the sum of the possible pairs of researchers in X and in Y separately, and between X and Y. We thus can write: C C C nV n nY nXY n n n
X
X ' X Y ' Y XY ' XY CV CX CY CXY CX CV CY CV CXY CV or in terms of frequencies and intensities of collaboration: pV wX pX wYpY wXYpXY or iV wX iX wY iY wXY ixy where CX C wY Y CV CV
wX
wXY
CXY CV
with wX wY wXY 1 This formula can easily be extended to groups of more than two laboratories. Aggregating over the entire network, we have
wIiI wIJiIJ 1 I
I,JI
The intensity of co-publication in scientific research
269
with
wI wIJiIJ 1 I
3.3
I,JI
Remark on the Weighting
Until now, we have considered for simplicity that the articles were coauthored by two (CNRS) researchers. In reality, they can also be written by threesomes or foursomes of (CNRS) researchers, and so on. But, as already indicated (in sub-section 2.1), we thought it appropriate to study the network of collaboration ‘link by link’, that is, by couples or pairs of coauthors. In practice, this means that an article is repeated in our database (and thus counted) as many times as there are pairs of different (CNRS) coauthors. For example, for an article published by three CNRS researchers, one belonging to a laboratory X and the two others to a laboratory Y, we count three co-publications – two between X and Y and one within Y.14 Note that, if we follow this procedure, the aggregation formula (as we simply write it in the previous sub-section) applies more generally in the case where there are more than two (CNRS) co-authors for an article. Note also that in practice in our case, since only 20 per cent of the articles in Group 1 are co-authored by more than two CNRS researchers, the choice of the weighting assumption should not make an important difference. 3.4
Practical Calculation: An Example
Let us take the concrete example of the town of Marseille to describe in detail the calculation of our measure of the intensity of collaboration, using the information displayed in Table 10.2, which also gives the results of this calculation for the other towns. Marseille (as indicated in column 1) is a town with 18 CNRS researchers (among the 470). These researchers are involved in 34 co-publications among themselves (column 3) and in 18 copublications with the CNRS researchers from two other towns (column 4), ten of them with Grenoble and eight with Strasbourg.15 The number of possible couples of researchers working in Marseille is 18 '17/2, or 153. The frequency of collaborations per couple of researchers in Marseille is therefore 34/153 or 0.22. Given that the numbers of researchers in Grenoble and Strasbourg are 105 and 14, the number of possible couples of researchers linking Marseille and Grenoble and Marseille and Strasbourg are respectively 1890 ( 105'18) and 252 ( 14'18). The corresponding frequencies of collaborations per couple are therefore 0.0053 ( 10/1890) and 0.0317 ( 8/252).
270
Bagneux Gif sur Yvette Grenoble Marseille Meudon Montpellier Orléans Orsay Palaiseau Paris Poitiers Saint Martin d’Hères Strasbourg Talence Toulouse
Towns
1 1 6 1 1 3 1 3 2 6 1 2 1 1 2
14 9 29
Number of laboratories per town (*)
9 16 105 18 9 20 10 66 18 86 11 31
Number of CNRS researchers (*)
2 0 4
6 3 12 2 2 7 0 9 4 7 0 5
Number of partner towns (**)
72 8 88
51 11 666 34 27 47 7 174 15 249 31 161
Number of articles ‘within’
20 0 63
171 40 449 18 19 83 0 192 45 148 0 193
Number of articles ‘between’
35.2 9.9 9.6
63.0 4.1 5.4 9.9 33.3 11.0 6.9 3.6 4.4 3.0 25.1 15.4
Intensity within-town
0.1 0.0 0.4
1.3 0.2 0.7 0.1 0.1 0.3 0.0 0.2 0.2 0.3 0.0 0.3
Intensity between-town (average computed on all other 16 towns)
Table 10.2 Descriptive statistics and within- and between-town intensity of co-publication at the town level#
0.9 0.0 1.6
3.5 0.9 0.9 0.8 0.5 0.8 0.0 0.4 0.9 0.6 0.0 0.9
Intensity between-town (average computed on partner towns only)
271
10 9 470 27.6
1 1 34 2.0
3 2 68 4.0
39 35 1715a —
31 58 765b —
38.5 43.2 — 18.9
0.6 0.2 — 0.3
3.0 1.4 — 1.0
Notes: # The overall frequency of co-publications for the sample of 470 CNRS researchers is p 0.0225. * Towns with less than nine CNRS researchers and laboratories than less than five CNRS researchers are not considered. ** Partner towns are defined as having more than six articles co-published by their CNRS researchers over the six-year period, 1992–97 (that is, at least an average of one co-publication per year). a Each article is weighted by the number of pairs of authors that contribute to its publication, otherwise the number of articles would be 1222. b Each article is weighted by the number of pairs of authors that contribute to its publication, otherwise the number of articles would be 412.
Villeneuve d’Ascq Villeurbanne Total Mean
272
The economics of knowledge
In order to compute the intensities of collaboration, we have also to calculate p, the overall frequency of collaboration per couple of researchers for the complete set of the 17 towns. It is the ratio between the total weighted number of articles, 2480 ( 1715 765), and the number of possible couples that can be formed by the 470 CNRS researchers, that is, 110 215 pairs ( 470'469/2). We thus have p 0.0225. This overall frequency p is also the intra- (or within-) and inter- (or between-) frequency of collaboration that would have been obtained for Marseille and all the other towns in the hypothetical case of homogeneity. In fact, the intrafrequency for Marseille is much higher (0.22) than this reference value, the inter-frequency of collaboration with Grenoble much lower (0.0053), and that with Strasbourg relatively closer (0.0317). Finally, the intra-town intensity for Marseille is of 0.22/p or 9.88 (column 5). Likewise, the Marseille–Grenoble and Marseille–Strasbourg inter-intensities are 0.24 and 1.41 respectively, yielding a mean inter-intensity of collaboration of Marseille with all the other 16 towns of (0.24 1.41)/16 or 0.1 (column 6), and a mean inter-intensity of Marseille with only its two effective partners of (0.24 1.41)/2 or 0.82 (column 7).
4
RESULTS: THE IMPORTANCE OF GEOGRAPHICAL PROXIMITY AND QUALITY OF SCIENTIFIC ENVIRONMENT
We look first at the estimated intensities of co-publication between the CNRS researchers at the town level (Table 10.2 and Appendices 1 and 2). Next, we consider in detail the statistical evidence on the potential determinants of co-publication we have been able to measure, which is mainly provided by simple correlations computed both at the town and laboratory levels (Tables 10.3 to 10.5 and Appendix 3). Finally, we assess the robustness of these results by examining the multivariate regressions of the occurrence and intensity of co-publication on these various determinants (Table 10.6). 4.1
Intensity of Co-publication at the Town Level
The estimated inter-town intensities of co-publication among all the different couples of towns, as we can see from the matrix of co-publication intensity in Appendix 1 and from the graph of the co-publication network in Appendix 2 (and also from their averages by towns computed in Table 10.2), are extremely dispersed. Of the 136 possible couples of towns, only 34 are effectively collaborating.16 Grenoble, Orsay and Paris are the main
273
4 149 7 264 38 325 8 136 4 149 9 000 4 600 26 664 8 136 33 024 5 049 13 609
6 384 4 149 12 789 4 600 4 149 194 176 —
9 16 105 18 9 20 10 66 18 86 11 31
14 9 29 10 9 470 27.6
91 36 406 45 36 13 127 —
36 120 5 460 153 36 190 45 2 145 153 3 655 55 465 248 193 379 184 139 7 056 —
328 246 1 870 235 99 365 63 922 274 985 88 438 449 0 410 305 188 — 279
344 171 421 361 208 548 0 334 297 291 0 418 6.45 11.07 5.30 3.12 4.55 — 6.13
3.02 3.93 10.19 11.17 5.97 5.63 8.24 4.97 4.32 6.72 5.17 4.44 17.71 21.44 13.07 18.40 15.44 — 15.71
36.44 15.38 17.81 13.06 11.00 18.25 6.30 13.97 15.22 11.45 8.00 14.13
3.69 3.94 2.71 4.13 3.02 — 3.44
3.68 3.07 3.39 2.84 2.63 3.47 3.54 3.69 4.77 3.75 2.34 3.78
Number of Number of Number of Stock of Mean Mean distance Mean Mean scientists possible possible publications geographic of productivity quality of couples couples between 1992 distance to specialisation publications ‘between’ ‘within’ and 1997 partners
Descriptive statistics for the main determinants of co-publication at the town level
Bagneux Gif sur Yvette Grenoble Marseille Meudon Montpellier Orléans Orsay Palaiseau Paris Poitiers Saint Martin d’Hères Strasbourg Talence Toulouse Villeneuve d’Ascq Villeurbanne Total Mean
Town
Table 10.3
0.16 0.43 0.24 0.28 0.23 — 0.29
0.12 0.14 0.50 0.44 0.20 0.21 0.32 0.25 0.33 0.48 0.26 0.27
Mean proportion of articles co-authored with foreigners
274
The economics of knowledge
Table 10.4 Correlations at the town level with the occurrence and intensity of co-publication Intensity within-town (N 17) Geographic distance Distance in specialisation Overall profile Physics-chemistry General physics Solid-state physics Applied physics Materials science Crystallography Other Size Number of researchers NI Maximum (NI, NJ) Minimum (NI, NJ) Average (NI NJ )/2 Publications 1992–97 SI Maximum (SI, SJ) Minimum (SI, SJ) Average (SI SJ)/2 Number of couples of researchers CIJ NI * NJ Productivity PI Maximum (PI, PJ) Minimum (PI, PJ) Average (PI PJ)/2 Quality of publications QI Maximum (QI, QJ) Minimum (QI, QJ) Average (QI QJ)/2
Occurrence between-town (N 136)
Intensity between-town (N 34)
—
0.09
0.16
—
0.02 0.18** 0.20*** 0.14 0.06 0.02 0.16* 0.00
0.21 0.14 0.20 0.24 0.08 0.06 0.17 0.16
0.48* —
— 0.52*** 0.42*** 0.56***
— 0.41** 0.31* 0.45***
0.29 —
— 0.52*** 0.54*** 0.60*** 0.49***
— 0.29* 0.21 0.33* 0.32*
0.39
0.62*** —
0.11 —
— 0.11 0.26*** 0.19***
— 0.67*** 0.47** 0.69***
— 0.03 0.23*** 0.16*
— 0.09 0.03 0.08
275
The intensity of co-publication in scientific research International openness peI Maximum (peI, peJ) Minimum (peI, peJ) Average (peI peJ)/2
0.37 — — —
— 0.14* 0.34*** 0.26***
— 0.26 0.12 0.22
Note: The stars ***, ** and * indicate that the correlations are statistically significant at a confidence level of 1 per cent, 5 per cent and 10 per cent, respectively.
nodes in the network of collaboration, being respectively linked to twelve, nine and seven other towns, whereas Poitiers, Orléans and Talence appear to be isolated.17 Among all effectively linked couples of towns (or partner towns), the intensity of co-publication ranges from a lowest value of 0.24 for Grenoble–Marseille and Grenoble–Meudon to a highest value of 7.40 for Bagneux–Villeuneuve d’Ascq. As could be expected (and checked by looking at the number of CNRS researchers in our sample per town, given in column 2 of Table 10.2), the towns with the largest number of CNRS researchers are those which tend to have more links with other towns but also lower inter-town intensity estimates. The estimated intra-town intensities of co-publication (given in column 6 of Table 10.2) are much higher than the inter-town intensities, with very few exceptions. These are always greater than one, and on average equal to 18.9, as compared with an average inter-intensity of 0.3 when computed over all couples of towns and of 1.0 when computed only over the partner towns. This strongly points to a major influence of geographical proximity on the intensity of collaboration. Note also that intra-town intensity tends to be high in towns with few partners, like Meudon, Poitiers, Strasbourg, Villeneuve-d’Ascq or Villeurbanne, compared with towns with many partners like Grenoble, Orsay and Paris, which have among the lowest intratown intensities (5.4, 3.6 and 3.0 respectively). This result could mainly be explained by the larger size of the CNRS research community in Grenoble, Orsay and Paris, and the fact that these towns host several laboratories, both characteristics entailing numerous potential links among which relatively many do not occur. As a matter of fact, the co-publication intensities estimated at the laboratory level for Grenoble, Orsay and Paris are much higher, being on average equal to 19.4, 33.4 and 34.5 respectively, and quite comparable to that of the laboratories of the other towns (see column 3 in Table 10.5).
276
The economics of knowledge
Table 10.5 Town averages of within- and between-town intensity of co-publication at the laboratory level Town
Bagneux Gif sur Yvette Grenoble Marseille Meudon Montpellier Orléans Orsay Palaiseau Paris Poitiers Saint Martin d’Hères Strasbourg Talence Toulouse Villeneuve d’Ascq Villeurbanne Mean* Mean**
Number of Intensity of Intensity of Intensity of laboratories collaboration collaboration collaboration per town withinbetweenbetweenlaboratory laboratorylaboratorywithin-town* between-town* 1 1 6 1 1 3 1 3 2 6 1 2 1 1 2 1 1 2 2
58.4 9.2 19.4 11.6 30.9 28.1 6.4 33.4 11.9 34.5 23.2 37.1 38.0 15.7 25.7 98.9 40.0 30.7 30.7
— — 2.7 — — 0.0 — 1.4 0.0 0.2 — 0.0 — — 1.5 — — 0.8 3.2
2.0 0.4 0.5 0.1 0.1 0.3 0.0 0.2 0.2 0.2 0.0 0.4 0.1 0.0 0.2 0.5 0.4 0.3 4.0
Notes: * Mean computed on all the laboratories. ** Mean computed on partner laboratories only.
4.2 Determinants of the Occurrence and Intensity of Co-publication: Correlation Evidence We consider six a priori influential determinants of collaboration at the laboratory or town levels, which we have been able to approximately measure or proxy: geographic distance, specialisation, size, productivity, quality of publications and international openness. Apart from geographic distance and distance in specialisation which are directly defined for a couple of laboratories or towns, it is somewhat problematic to adopt a priori a single measure for our four other variables, such as their average over the two laboratories or towns concerned, say, for example, (SI SJ)/2 where S is a measure of size of the laboratories or towns I and J. Thus, in addition to
The intensity of co-publication in scientific research
277
the average, we used the maximum and minimum values, say SI and SJ for the couple of laboratories or towns I and J. Note that for size we also have three different possibilities: the number of CNRS researchers, say NI for laboratory or town I; the number of possible pairs of CNRS co-authors, say (NI 'NJ) for the couple (I, J) of laboratories or towns – or (NI 'NI 1)/2 within the laboratory or town I – and the number of publications over the six-year period, say SI for laboratory or town I. We will examine all these variables in turn. Their means at the town level are given in Table 10.3; and their correlations with the binary indicator of occurrence of co-publication and with our measures of both intra- and inter-intensity, also at the town level, are displayed in Table 10.4. These correlations are also recorded in Appendix 3 at the laboratory level, both overall (for the 34 laboratories) and within towns (for the seven towns out of 17 which have more than one laboratory). The statistical evidence is quite consistent at both the town and laboratory levels, with the notable exception of the correlations of the occurrence of co-publication with specialisation at the laboratory level within towns. It is also, in general, qualitatively comparable for the occurrence and intensity of co-publication, the one major exception being the size variable positively correlated with occurrence and negatively with intensity. 4.2.1 Geographic distance The average distance of a town from its partners can vary widely (see Table 10.3). At the two extremes, Montpellier collaborates with seven other towns, situated at an average distance of 550 km (kilometres), while Gif sur Yvette is related to three towns much closer, at an average distance of 170 km, two of these being also located in the Parisian region. Four of the five towns situated less than 300 km on average from their partners are in the Parisian region (besides Paris intramuros, Bagneux, Gif sur Yvette, Meudon, Orsay and Palaiseau). The geographic distance apparently plays a negligible role, or only a slightly negative one, in the occurrence of collaboration, as well as on its intensity. This is shown by the correlations computed at the town level but also at the laboratory level. The relevant two correlations at the town level ( 0.09 and 0.16) are negative but both statistically non-significant, and the two at the laboratory level ( 0.09 and 0.06) are also negative, with only the first moderately significant (at a 5 per cent confidence level). In fact, as we have already noted in comparing the values of the intraand inter-town intensities of co-publication presented in Table 10.2, proximity has a major influence on collaboration. This is confirmed by the comparison of the intra- and inter-laboratory intensities shown in Table 10.5. But more interestingly, this can also be qualified, since by comparing the inter-laboratory–intra-town intensity to the inter-laboratory–inter-town
278
The economics of knowledge
intensity we distinguish between what we shall call immediate proximity and local proximity. Immediate proximity, which is that of researchers working in the same laboratory, usually located in a common building, favours frequent face-to-face interactions and can be expected to induce and facilitate collaboration. Local proximity, which is that of researchers working in different laboratories but still relatively close, as when in the same town, can be expected to be less conducive to collaboration than immediate proximity. Nonetheless, local proximity should be more favourable to collaboration than when researchers are working in really distant laboratories, as when located in different towns. This is, indeed, very clearly what we find. For example, for Grenoble and its six laboratories, the average intra-laboratory intensity is 19.4, which is about seven times higher than the inter-laboratory–intra-town intensity (2.7), itself about 5 times higher than the average inter-laboratory–inter-town intensity (0.5). On average, for all 17 towns the pattern is the same, the three average intensities being 30.7, 0.8 and 0.3 respectively. One can thus distinguish three scales of geographic distance, which influence collaboration very differently. Immediate proximity has a considerable impact on collaboration and local proximity is also favourable but much less so. By contrast, beyond proximity, geographic distance strongly hinders collaboration, but per se and only slightly, if at all, in proportion to real distance (say in kilometres). Such findings are well corroborated by prior studies on knowledge flows between public laboratories and industrial firms, which show that proximity allows face-to-face interactions and easy exchanges of tacit knowledge between actors, inducing them to build a common understanding rather than referring only to a ‘common text’ (see, for example, Zucker et al., 1998a and 1998b, and Leamer and Storper, 2000). New communication technologies have certainly contributed to the ‘death of distance’, by helping researchers to collaborate in research much more easily and faster: however, they have not done away with the crucial importance of proximity. 4.2.2 Specialisation Proximity in specialisation, not only geographic proximity, should also strongly influence collaboration in a field as diverse and large as condensed matter physics. The network of co-publication presented in Appendix 2 is by itself suggestive of such an influence. Orsay and Grenoble, which appear to be two central nodes of the network, are indeed the location of the two French storage rings, which are very large facilities used by physicists of condensed matter.18 We have tried to take into account the specialisation of laboratories (or towns), although there is no easy and good way to do so. We have defined a
The intensity of co-publication in scientific research
279
profile of specialisation of a laboratory (or town), based on the classification by the main theoretical and/or experimental ‘sub-domains’ of the journals in which the CNRS researchers of this laboratory (or town) are publishing. Such classification is difficult but seems carried out relatively well by the Science Citation Index (SCI). We found that the most frequently listed subdomains of the journals in which the articles of the CNRS condensed matter physicists are published were physics-chemistry, general physics, solid-state physics, applied physics, materials science and crystallography. We then characterised an ‘overall specialisation profile’ of each laboratory (or town) by the [7, 1] column vector defined by the proportions of publications of its CNRS researchers in these six main sub-domains and the group of other sub-domains. We also considered the seven ‘specific specialisation profiles’ corresponding to the seven [2, 1] column vectors defined by the proportions of publications in each of the seven sub-domains and all the six others. Next, to measure the distance in specialisation between all couples of laboratories (or towns) we adopt the chi-squared distance between their specialisation profiles. To facilitate interpretation we also normalised this measure in such a way that the average distance between any one laboratory (or town) and all others will at most be equal to one if the laboratories (or towns) had specialisation profiles which were not statistically different at the 1 per cent confidence level.19 The average specialisation distances (in terms of the overall profiles) between towns, given in Table 10.3, show that we are in fact far from this hypothetical situation. Specialisation profiles are quite diverse, and the specialisation distances as we measure them vary widely (the lowest town average distance being about 3 for Bagneux and Villeneuve d’Ascq and the highest one about 11 for Marseille and Talence). With the major exception of the puzzling positive and significant between-laboratory-within town correlations with the occurrence of copublication (and some minor ones concerning particularly the specialisation indicator in crystallography and again the occurrence of co-publication), all other correlations of our overall and specific specialisation distance measures with the occurrence and intensity of co-publication are consistently negative (see Table 10.4 and Appendix 3). The correlations with the existence of co-publication, however, are mostly small and not significant, while the correlations with the intensity tend to be sizeable and statistically significant. The two between-laboratory-between-town and between-laboratorywithin-town correlations of intensity of co-publication and overall specialisation distance are, for example, as large as 0.3 and 0.4 respectively. Note also that the puzzling exception of the between-laboratory-withintown correlations with the occurrence of co-publication may largely reflect the correlations of our measure of specialisation with the other determinants of co-publication, since the corresponding coefficient in our
280
The economics of knowledge
regression analysis is not statistically significant (see sub-section 4.3 and Table 10.6). On the whole, our evidence thus tends to show that proximity in specialisation favours strongly the intensity of collaboration, but much more weakly so, or not, its existence. This latter result is not what we expected. It is clear, however, that our attempt here at measuring specialisation is crude, and that much remains to be done to better characterise what it is and to assess its impact on collaboration. 4.2.3 Size We considered three different measures of size. The first relies on the number of CNRS researchers in the laboratories (or towns) concerned, and the second on the total number (or stock) of their publications over the six years 1992–97.20 The third, which is particularly well suited to our definition of intensity of co-publication, is the number of possible couples of CNRS researchers for each couple of laboratories (or towns). For the first two measures, as already explained, we experimented with the average, the maximum and minimum of the number of researchers and of the stock of their publications for each couple of laboratories (or towns). Not surprisingly, these three types of size measures are overall quite consistent, as shown by the descriptive statistics in Table 10.3. Grenoble comes first of all towns, with 22 per cent of the total number of CNRS researchers involved, 20 per cent of the total number of possible couples of co-authors among them, and 27 per cent of their total publications. Paris comes second and Orsay third (with about 18 per cent and 14 per cent respectively of the number of researchers and of all possible couples of co-authors, and for each of them roughly 14 per cent of all their publications). A priori one would expect that the size of laboratory (or town) would impact positively on the chance of collaboration, but not its intensity, since by construction our measure of intensity already takes into account such a size effect. One would even think it likely that the larger the laboratories (or towns) involved in collaboration, the smaller its intensity. This is indeed what we see clearly when looking at the correlations in Table 10.4 and in Appendix 3 for the different indicators of size we used. The correlations of nearly all of them, both at the town level and the laboratory level (withinand between-town), are thus very significantly positive and substantial (ranging from 0.2 to 0.6) with the occurrence of co-publications, while very significantly negative in a comparable range (from 0.2 to 0.6) with its intensity. Note that it is also the case that the correlations of the intensity of co-publication of the researchers within their own laboratories (or towns) with the size of their laboratories (or towns) tend to be significantly negative.
The intensity of co-publication in scientific research
281
4.2.4 Productivity The productivity of the laboratories (or towns) is simply measured as the stock of publications of their CNRS researchers in the period 1992–97 per researcher (that is, as the ratio of our two first measures of size just defined in sub-section 4.2.3).21 As can be seen in Table 10.3, productivity varies widely from one town to another, from a minimum of 6.3 articles per researcher over six years for Orleans to a maximum of 36.4 for Bagneux, the overall mean being 15.7 articles per researcher (that is, 2.6 articles per year). In contrast with size, it seemed a priori likely that both the correlations of productivity with the occurrence of co-publication and its intensity should be positive. This is definitely what we find. Nearly all these correlations, including the two of within-town and within-laboratory intensities with productivity, are very significantly positive and of a sizeable order of magnitude, from 0.2 up to 0.7 (see Table 10.4 and Appendix 3). 4.2.5 Quality of publications Our measure of the quality of publications of laboratories (or towns) is consistent with that of their productivity. It is the average impact factor (or impact score) per publication of their stock of publications over the period 1992–97. Precisely, it is the weighted mean of the impact factors of the journals in which these publications have appeared (the weights being the numbers of publications in the different journals). The impact factors of the journals are provided by the SCI; they are defined and computed as the average number of citations per article received by the articles published in the journals over a period of two and five years. We used here the two-year impact factors, but using the five-year impact factors did not make a difference in our results. Our measure of the quality of publications of a laboratory (or town) is thus an estimate of the expected number of citations that the publications of its researchers will on average receive over two years. In Table 10.3, we see that this average number overall is 3.4 citations per article over two years, and that it can differ by a factor of 2 at the town level, being lowest for Poitiers, with a citation rate of 2.3, and highest for Palaiseau, with a citation rate of 4.8. Although we expected that the quality of publications, like productivity, would be positively correlated with both the occurrence and intensity of collaboration, the evidence (recorded in Table 10.4 and Appendix 3) is mixed. There are no statistically significant correlations with the intensity of co-publication at the town or laboratory levels. We find a significantly positive correlation with the occurrence of co-publication only when we use as our quality indicator the minimum value for the couples of laboratories or towns involved (0.23 at the town level and 0.06 at the laboratory level). This suggests, interestingly but tentatively, that what matters in establishing
282
The economics of knowledge
a collaboration is a minimum quality requirement on the two partners involved. To confirm such a proposition would, of course, need a more detailed analysis and should be performed at the researcher level, not only at the aggregate level of the laboratory. Note, however, that the betweenlaboratory-within-town correlations with the occurrence of co-publication are all very significantly negative, raising a similar puzzle as the one we have with our specialisation indicators. 4.2.6 International openness As we indicated in sub-section 2.2, we cannot precisely locate the laboratories of the very many ‘external’ researchers who are co-publishing with our sample of 470 CNRS physicists. However, it is possible to identify the foreign (non-French) addresses among all those listed in the SCI electronic records for all their articles (both in Group 1 and Group 2). In spite of the imprecision of such information, we can thus build an indicator of the international openness of the laboratory as the proportion of articles of their CNRS researchers (over the six years, 1992–97) involving a least one foreign co-author. In Table 10.3, we see that this proportion is overall about 30 per cent, and that it is the highest, about 50 per cent, for Grenoble and Paris, and the lowest, about 15 per cent, for Bagneux, Gif sur Yvette and Strasbourg. Our a priori thought was that international openness would also go together with greater occurrence and intensity of collaboration between the CNRS researchers themselves and their laboratories (or towns). This is what we observe, although the evidence is not strong. Many of the correlations given in Table 10.4 and in Appendix 3 are not significantly different from zero, but those that are significant are all positive. 4.3
Regression Confirmatory Evidence
To assess the robustness of the evidence provided by our analysis of simple correlations, we did a number of regressions of both the occurrence and the intensity of co-publication on the six a priori influential variables we have been able to consider. All of these mainly told the same story, confirming our observations, which were already strongly supported by the correlation evidence.22 We present in Tables 10.6 and 10.7 the regressions we did at the laboratory level, which include all six variables measured in the simplest way (that is, for specialisation in terms of overall profile, for size as the average of the numbers of researchers in all couples of laboratories, and similarly for productivity, quality of publications and international openness as the averages of the corresponding indicators in all couples of laboratories). The geographic distance between laboratories does not influence the intensity of collaboration and has only a small, significantly negative
The intensity of co-publication in scientific research
283
Table 10.6 Regression results at the laboratory level on the occurrence of co-publication Variables
Geographic distance
Occurrence of co-publication between laboratories Within-town (N 39)
Between-town (N 522)
—
0.008** (0.004) 0.009* (0.005)
Distance in specialisation
0.02 (0.04)
Size
0.04*** (0.01)
0.015*** (0.001)
Productivity
0.03* (0.02)
0.013*** (0.003)
Quality of publications
0.18*** (0.02)
International openness
0.08 (1.5)
Adjusted R2
0.374
0.017 (0.029) 0.17 (0.22) 0.107
Notes: The standard errors of the estimated coefficients are given in parentheses. The stars ***, ** and * indicate that they are statistically significant at a confidence level of 1 per cent, 5 per cent and 10 per cent, respectively.
impact on its occurrence – an increase of 100 km in the distance between two laboratories corresponding to a decrease in the frequency of co-publication of less than 1 per cent (0.8 per cent). As expected, the distance in specialisation has a negative impact on the intensity of co-publication, which seems sizeable although statistically not very significant. An increase of one standard deviation in the distance of specialisation, as we characterise it, will thus imply a fall of nearly 30 per cent in the intensity of co-publication between-laboratory-between-town. Laboratory size has a very significant and large impact on collaboration: positive in its occurrence, while negative in its intensity. A 10 per cent increase of the average size of each laboratory will thus entail an increase of the frequency of collaboration within-town and between-town of about 15 per cent and 25 per cent respectively, while it will correspond to a decline of the intensity of co-publication within-laboratory of nearly 10 per cent, and between-laboratory-between-town of about 10 per cent. Laboratory productivity has positive and mostly significant effects on both the
284
The economics of knowledge
Table 10.7 Regression results at the laboratory level on the intensity of co-publication Variables
Intensity within-laboratory (N 34)
Intensity of co-publication between-laboratory Within-town (N 15)
Between-town (N 41)
Geographic distance
—
—
0.02 (0.19)
Distance in specialisation
—
0.78* (0.43)
0.41* (0.23)
1.51*** (0.50)
0.31 (0.34)
0.24** (0.09)
Productivity
3.31*** (1.09)
0.32 (0.19)
0.19 (0.12)
Quality of publications
6.88 (5.41)
0.83 (2.42)
0.48 (1.82)
Size
International openness Adjusted R2
40.1 (58.5) 0.312
33.4 (29.8) 0.162
0.93 (12.1) 0.369
Notes: The standard errors of the estimated coefficients are given in parentheses. The stars ***, ** and * indicate that they are statistically significant at a confidence level of 1 per cent, 5 per cent and 10 per cent, respectively.
occurrence and intensity of collaboration, which are of the same or even larger orders of magnitude than the size effects. A 10 per cent increase of productivity will thus involve an increase of 15 per cent and 30 per cent in the frequency of collaboration within-town and between-town respectively, and will result in a rise of about 20 per cent in the intensity of co-publication within-laboratory. The quality of laboratory publications does not seem to have a significant impact on collaboration, except one which is negative, contrary to our a priori expectation, on the frequency of co-publication betweenlaboratory-within town (confirming the puzzling simple correlations we already noted). Clearly the international openness of laboratories, at least in the way we can proxy for it, has also apparently no significant influence on collaboration.
The intensity of co-publication in scientific research
5
285
CONCLUSION
In order to study networks of collaboration between researchers, we proposed a simple measure of the intensity of collaboration, which can be intuitively interpreted in terms of relative probability and easily aggregated at the laboratory level. We first used this measure to characterise the relations of collaboration between the scientists of the French Centre National de la Recherche Scientifique (CNRS) in the field of condensed matter physics, as defined in terms of co-publication during the six-year period, 1992–97. We then used it to investigate the importance of various factors of collaboration: mainly the geographical distance between laboratories, but also their specialisation and size, their productivity, the quality of their publications and their international openness. We find that the average intensity of co-publication of researchers within laboratories is about 40 times higher than the average intensity between laboratories if they are located in the same town, and about 100 times higher than the intensity between laboratories if they are not located in the same town. Yet, geographical distance does not have a significant impact, or has a very weak one, on the existence and intensity of co-publication between laboratories located in different towns. There are basically three scales of geographic distance. Immediate proximity, which allows easy face-to-face interactions, has a considerable impact on collaboration, while local proximity is also relatively favourable but much less so. Geographic distance per se, that is, beyond proximity, remains by contrast a strong obstacle to collaboration, but only slightly, if at all, in proportion to real distance. Although our measure of specialisation between laboratories remains crude, we find that proximity in specialisation has also a large positive influence on the intensity collaboration. The size of laboratories and their productivity in terms of number publications per researcher appear to be influential determinants of collaboration, having both a positive impact on the occurrence of co-publication, but a negative impact for size and a positive one for productivity on the intensity of co-publication. Contrary to our expectations, we do not really observe significant effects on collaboration of the average quality of publications and of international openness of laboratories. However, this may be due, at least in part, to the fact that these two indicators, as we have been able to construct them, are at best imperfect proxies. In future work, it will thus be necessary to improve the measurement of the potential determinants of collaboration that we have been able to consider, as well as to extend the list of these determinants. Clearly it will also be important to broaden the scope of our study, which remains mainly illustrative. In particular, although we think it is appropriate and interesting to
286
The economics of knowledge
analyse collaboration at the level of the laboratory, as we did here, it will be both enlightening and challenging to carry out this analysis together with an investigation at the individual researcher level. This will undoubtedly lead to an assessment of the role of ‘star scientists’ in the scientific performance of their own laboratories and in the formation and development of networks of collaboration. By focusing on the co-publication between researchers working in the same institutional setting, that of the French CNRS, we have been able to control for organisational proximity. Comparing similar studies in different research environments could be instructive by itself. But, of course, trying more generally to integrate institutional and organisational characteristics in the analysis and to understand how they can enhance or hinder collaboration should be a central objective in the research agenda – one that will keep up with the high standards of Paul David.
287
Villeurbanne
Villeneuve d’Ascq
Toulouse
Talence
Strasbourg
Saint Martin d’Hères
Poitiers
Paris
Palaiseau
Orsay
Orléans
Montpellier
Meudon
Marseille
Grenoble
Gif sur Yvette
Bagneux
Matrix of intensities of co-publication betweenn CNRS condensed matter physicists at the town level
62.96
4.07
5.42
0.53
3.53
9.88
0.24 33.33
0.24
11.00
0.49
1.73
6.91 3.61
0.81
0.44
0.38
0.37
4.36
0.52
0.99
0.28
1.70
3.03
0.29
0.13
0.80
0.13
2.57
25.05 15.39
0.38
0.33
0.79
1.84
0.36
35.16
1.41
9.88 9.63
0.35
0.38
0.15
5.62
38.52
1.29
0.36
7.4
43.21
0.37
2.50
Bagneux Gif sur Grenoble Marseille Meudon Mont- Orléans Orsay Palaiseau Paris Poitiers Saint Stras- Talence Toulouse Ville- VilleurYvette pellier Martin bourg neuve banne d’Hères d’Ascq
Appendix 1
APPENDICES
288
Appendix 2
Toulouse
Orléans
Network of co-publications at the town level
LEGEND: The intensity of co-publication between two towns is indicated by a line: • thick grey if more than 1 • thick black if between 0.4 and 1 • thin black if less than 0.4.
Talence
Poitiers
Paris
Bagneux
Montpellier
Palaiseau
Orsay
Gif sur Yvette
Meudon
Villeneuve d’Ascq
Marseille
Villeurbanne
Saint Martin d’Hères
Grenoble
Strasbourg
The intensity of co-publication in scientific research
289
Appendix 3a Correlations at the laboratory level with the occurrence of co-publication Occurrence of co-publications between-laboratorybetween-town (N 522)
within-town (N 39)
Geographic distance
0.09**
Distance in specialisation Overall profile Physics-chemistry General physics Solid-state physics Applied physics Materials science Crystallography Other
0.02 0.05 0.09*** 0.04 0.01 0.05* 0.06** 0.03
0.40*** 0.30*** 0.35*** 0.24** 0.05 0.41*** 0.10 0.11
— 0.21*** 0.22*** 0.24***
— 0.45*** 0.57*** 0.57***
— 0.23*** 0.35*** 0.31*** 0.27***
— 0.58*** 0.63*** 0.63*** 0.61***
— 0.19*** 0.30*** 0.26***
— 0.35*** 0.40*** 0.39***
Size Number of researchers NI Maximum (NI, NJ) Minimum (NI, NJ) Average (NI NJ )/2 Stock of publications between 1992 and 1997 SI Maximum (SI, SJ) Minimum (SI, SJ) Average (SI SJ)/2 Number of couples of researchers CIJ NI * NJ Productivity PI Maximum (PI, PJ) Minimum (PI, PJ) Average (PI PJ)/2 Quality of publications QI Maximum (QI, QJ) Minimum (QI, QJ) Average (QI QJ)/2
— 0.03 0.06** 0.02
—
— 0.46*** 0.30*** 0.44***
290
The economics of knowledge
Appendix 3a (continued) Occurrence of co-publications between-laboratorybetween-town (N 522) International openness peI Maximum (peI, peJ) Minimum (peI, peJ) Average (peI peJ)/2
— 0.02 0.03 0.03
within-town (N 39)
— 0.07 0.06 0.01
Note: The stars ***, ** and * indicate that the correlations are statistically significant at a confidence level of 1 per cent, 5 per cent and 10 per cent, respectively.
Appendix 3b Correlations at the laboratory level with the intensity of co-publication Intensity within-laboratory (N 34)
Intensity of co-publication between-laboratory between towns (N 41)
Geographic distance
—
Distance in specialisation Overall profile Physics-chemistry General physics Solid-state physics Applied physics Materials science Crystallography Other Size Number of researchers NI Maximum (NI, NJ) Minimum (NI, NJ) Average (NI NJ )/2
0.42**
0.06
within towns (N 15) —
0.32*** 0.18 0.29*** 0.25** 0.05 0.21* 0.26** 0.11
0.42** 0.10 0.30 0.05 0.26 0.35* 0.16 0.26
— 0.52*** 0.51*** 0.60***
— 0.34* 0.02 0.27
291
The intensity of co-publication in scientific research Stock of publications between 1992 and 1997 SI 0.06 Maximum (SI, SJ) — Minimum (SI, SJ) Average (SI SJ)/2 Number of couples of researchers CIJ NI * NJ 0.34** Productivity PI Maximum (PI, PJ) Minimum (PI, PJ) Average (PI PJ)/2 Quality of publications QI Maximum (QI, QJ) Minimum (QI, QJ) Average (QI QJ)/2 International openness peI Maximum (peI, peJ) Minimum (peI, peJ) Average (peI peJ)/2
0.36** —
0.20 —
0.14 — — —
— 0.08 0.14 0.12 0.49*** — 0.51*** 0.42*** 0.54*** — 0.03 0.12 0.05 — 0.34*** 0.17 0.29***
— 0.18 0.13 0.17 0.23 — 0.33* 0.36** 0.36** — 0.25 0.11 0.01 — 0.005 0.35* 0.22
Note: The stars ***, ** and * indicate that the correlations are statistically significant at a confidence level of 1 per cent, 5 per cent and 10 per cent, respectively.
292
The economics of knowledge
NOTES *
1. 2.
3.
4.
5.
6.
7.
We wish to thank Michèle Crance, Serge Bauin and the members of the Unité des Indicateurs de la Politique Scientifique (UNIPS, CNRS), for their help in compiling the database and their advice. We have also benefited from remarks by Ajay Agrawal, MarieLaure Allain, Céline Allard-Prigent, Anne Crubelier, Dominique Foray and Claudine Hermann, and are particularly grateful, for detailed comments, to Bronwyn Hall and Edward Steinmueller. For a simulation analysis of this process in the institutional context of the USA, see David (1994), and for a first attempt of an econometric analysis on the same data as the one used in the present work, see Turner and Mairesse (2002). Three of these studies can be mentioned here. Jaffe (1989) shows that there is in the USA a close relationship at the state level between the number of patents and the importance of university research, which he interprets as evidence of geographic externalities. Jaffe et al. (1993) investigate the localisation of knowledge externalities using patent citation data. The authors show that citing and cited patents belong to the same geographic region with a very high probability. Jaffe and Trajtenberg (1998), also on the basis of patent citation data, study the localisation of flows of knowledge at an international scale. They find that patents cite much more frequently patents whose inventors live in the same country than patents whose inventors live in different countries. For example, by adopting a definition of a network as a ‘clique’ in the sense of the theory of graphs (that is, a set of points which are connected or such that the intensity of their interconnections exceeds a certain threshold), Blau (1973) makes the following observations for a group of 411 physicists. Members of large networks are often young, work in new and innovative specialties, have a teaching post and are relatively well known; by contrast, members of small networks are older, work in established specialties, in prestigious university departments and are involved in administration. These findings seem to reflect the existence of a cycle in research careers, leading the most productive scientists to be also part of the administrative elite. The analysis by Cole and Cole (1973) on stratification in science is typical of this approach. The authors classify physicists in terms of different criteria such as age, prestige within university departments, productivity and scientific awards. They then measure the impact of these characteristics on the researchers’ ranking in terms of scientific reputation and visibility. They finally use their results in an attempt to assess the existence and the extent of discrimination possibly arising from differences of race, gender and religion. In future work, a possibility will be to carry on this research at the level of individual researchers as well. In addition to what we can already observe at the laboratory level, this would allow the analysis of the role of productive and well-known ‘star’ scientists in shaping research networks. See for example the work by Crane (1969 and 1972), Crawford (1971) and Zucker et al. (1998a and 1998b). The Centre National de la Recherche Scientifique (CNRS) is the main French public organisation for basic research. With 25 000 employees (11 000 researchers and 14 000 engineers, technicians and administrative staff) and over 1200 research and service units (laboratories) throughout the country, the CNRS covers all fields of science. Directly administered by the ministry responsible for research, which is also usually responsible for higher education, the CNRS has very close links with academic research, researchers from the CNRS and from universities often working in the same laboratories. These criteria are mainly based on two practical considerations: researchers had to be ‘not too young’ so that we had a history of their publications (the youngest researchers born in 1960 had already been publishing for a few years in 1992, when they were 32 years old), and 1997 was the year for which we could know precisely the laboratories in which the researchers were working, when we first started compiling our database. Note that we have been able to follow up female researchers who happened to change names during the study period because they married.
The intensity of co-publication in scientific research 8.
293
The Science Citation Index (SCI) is produced by the Institute for Scientific Information (ISI). It encompasses all the (‘hard’) scientific disciplines and is constructed on the base of a compilation of over 3200 of the most cited international periodicals. The quality of the data is remarkable and, in particular, the coverage of scientific publications by CNRS units is very satisfactory (UNIPS, 1999). Ninety-five per cent of the scientific articles written by the CNRS researchers are in English and these are fully covered by the SCI. 9. Another solution would be to count each article only once, by simply weighting them by the inverse of the number of pairs of authors concerned. This point is discussed in subsection 3.3. It seems that the main results of our analysis would have been qualitatively unchanged. 10. This information on the individual researchers and their laboratories was provided to us by the Unité des Indicateurs de la Politique Scientifique (UNIPS) of CNRS. 11. There is no strict rule for correspondence between authors and addresses in the SCI, since the number of authors recorded for a scientific article often differs from the number of addresses listed for them. It is possible that several co-authors have the same address, in which case their address may be listed only once. Or when the collaboration involves different laboratories, the correspondence between authors and addresses is not always clear. Another possibility is that of multiple affiliations, with one co-author mentioning his or her affiliation to two or more laboratories, which may result again in a problem of attribution. 12. Note that for Group 1 articles this is a weighted average in which each article is counted as many times as the number of pairs of CNRS co-authors. The simple average for Group 1 is 4.5 ( 1741/386). See sub-sections 2.1 and 3.3 and note 9. 13. Note that this matrix is similar to the adjacency matrix used in the graph theory. The coefficients of the adjacency matrix are equal to 1 when there is a link between the entities corresponding to rows and those corresponding to columns; otherwise it is 0. The adjacency matrix thus characterises only the occurrence of collaboration between entities but not their intensities. 14. It is possible, of course, to proceed otherwise; that is, in this example we could have counted the article for one article giving rise to ‘two-thirds’ of a co-publication between X and Y and ‘one-third’ within Y. But, as we said, since we are interested in the analysis of collaboration relations, we deemed it better to consider that the more co-authors, the greater the weight of an article. 15. Marseille in fact has also relations with Poitiers, Gif sur Yvette, Orsay, Toulouse and Villeurbanne, but these are not taken into account because they all involve less than six co-publications (see sub-section 2.4). 16. As explained in sub-section 2.4, we only estimated the intensity of co-publication between any two towns (or two laboratories) when the actual number of co-publications between the CNRS researchers in these two towns (or two laboratories) was not too small, that is, less than six (or four) over the six-year period, and set it to zero otherwise. 17. The average number of links per town with other towns is only four. Although we do not know of such a result in another study with which we could compare this estimate, it may seem somewhat on the low side for towns with at least nine CNRS researchers in our sample (and given our adoption of a rather small threshold of at least six copublications over six years for the definition of an actual link between two towns). 18. Storage rings have become of great importance throughout the world. They are used to curve or oscillate the trajectory of light-charged particles (electrons or positrons) that emit ‘synchrotron radiations’. They thus constitute an extraordinary source of radiations of varying wavelengths, especially X-rays. The European ring of the ESRF (European Synchrotron Radiation Facility) is situated at Grenoble and employs as many as 500 people on a permanent basis. France has two other rings situated at Orsay at the LURE. About 30 outside laboratories collaborate on a permanent basis with the LURE, as do 20 industrial partners, in the field of physics but also chemistry, biology and environmental science, micro-production, lithography and astrophysics. The LURE rings
294
19.
The economics of knowledge should soon be replaced by the ‘SOLEIL’ ring, which will constitute a source of ‘super’ synchrotron radiation (several thousand times brighter). The chi-squared distance between the specialisation profile column vectors 1 and 2 of two laboratories 1 and 2 is thus defined as: ^ 2 ^ 2 ( 1i ( 2i i) i) 2
n1 n , 2 ^ ^ i i i where 1i and 1i denote the coefficients of vectors 1 and 2 with i varying from 1 to 7 (or 1 to 2) for the overall (specific) profiles, and where n1 and n2 are the numbers of ^ publications of the two laboratories, and is the specialisation profile column vector of the two laboratories taken together (or weighted average profile). This chi-squared distance (as shown in Table 10.3) is normalised by dividing it by the 99 percentile value of the chi-squared statistic of 6 degrees of liberty for the test of equality of the overall specialisation profile vectors (of dimension 7) for any two given laboratories or towns. Note that the stock of publications not only includes the co-publications of Group 1 of our final sample of 470 CNRS researchers, but also their co-publications of Group 2 and their (few) publications alone. Note also that each publication is counted only once, irrespective of the number of co-authors. Our measure of productivity thus corresponds to all the publications of our 470 CNRS researchers in the period 1992–97 (see previous note). Note that estimating a generalised Tobit regression model of both the occurrence and intensity of co-publication also provides practically the same picture as the two corresponding separate linear regressions (the estimated correlation between the probit occurrence equation and the linear intensity equation being not statistically different from zero).
20.
21. 22.
BIBLIOGRAPHY Audretsch, D. and Stephan, P. (1996), ‘Company-scientist locational links: the case of biotechnology’, American Economic Review, 86 (3), 641–52. Barré, R., Crance M. and Sigogneau, A. (1999), ‘La recherche scientifique française: situation démographique’, Etudes et dossiers de l’OST, 1. Blau, J. (1973), ‘Patterns of communication among theoretical high energy physicists’, Sociometry, 37, 391–406. Beaver, D. and Rosen, R. (1978), ‘Studies in scientific collaboration. Part 1. The professional origins of scientific co-authorship’, Scientometrics, 1, 65–84. Beaver, D. and Rosen, R. (1979), ‘Studies in scientific collaboration. Part 2. Scientific co-authorship, research productivity and visibility in the French scientific elite’, Scientometrics, 1, 133–49. Callon, M. (1999), ‘Le réseau comme forme émergente et comme modalité de coordination: le cas des interactions stratégiques entre firmes industrielles et laboratoires académiques’, in M. Callon et al., Réseau et coordination, Paris: Economica. Callon, M. and Foray, D. (1997), ‘Introduction: nouvelle économie de la science ou socio-économie de la recherche scientifique?’, Revue d’Economie Industrielle, 79, 13–37. Cole, J. and Cole, S. (1973), Social Stratification in Science, Chicago IL: University of Chicago Press. Crane, D. (1969), ‘Social structure in a group of scientists: a test of the invisible college hypothesis’, American Sociological Review, 34, 335–52.
The intensity of co-publication in scientific research
295
Crane, D. (1972), Invisible Colleges: Diffusion of Knowledge in Scientific Communities, Chicago IL: University of Chicago Press. Crawford, S. (1971), ‘Informal communication among scientists in sleep research’, Journal of the American Society for Information Science, 22, 301–10. Dasgupta, P. and David, P.A. (1994), ‘Toward a new economics of science’, Research Policy, 23 (5), 487–521. David, P. (1994), ‘A science simulation for studying the US and similar institutional setups (SCISIMUS)’, mimeo. Diamond, A. (1996), ‘The economics of science’, Knowledge and Policy, 9 (2–3), 6–49. Foray, D. (2004), The Economics of Knowledge, Cambridge, MA: MIT Press. Gibbons, M., Limoges, C., Novotny, H., Schwartzman, S., Scott, P. and Trow, M. (1994), The New Production of Knowledge, London: Sage. Jaffe, A. (1989), ‘Real effects of academic research’, American Economic Review, 79 (5), 957–70. Jaffe, A. and Trajtenberg, M. (1998), ‘International knowledge flows: evidence from patent citations’, NBER Working Paper 6507. Jaffe, A., Trajtenberg, M. and Henderson, R. (1993), ‘Geographic localization of knowledge spillovers as evidenced from patent citations’, Quarterly Journal of Economics, 108 (3), 557–98. Katz, J.S. (1994), ‘Geographical proximity and scientific collaboration’, Scientometrics, 31 (1), 31–43. Leamer, E. and Storper, M. (2000), ‘The economics of geography at the Internet age’, mimeo, University of California at Los Angeles and Management and Public Policy, November. Rallet, A. and Torre, A. (2000), ‘Is geographical proximity necessary in the innovation networks in the era of global economy?’, mimeo, Université Paris Dauphine et Institut National de la Recherche Agronomique. Shi, Y. (2001), The Economics of Scientific Knowledge, Cheltenham, UK: Edward Elgar. Shrum, W. and Mullins, N. (1988), ‘Network analysis in the study of science and technology’, in Handbook of Quantitative Studies of Science and Technology, ed. A.F.J. Van Raan, Amsterdam: North Holland. Stephan, P. (1996), ‘The economics of science’, Journal of Economic Literature, 34, 1199–235. Turner, L. and Mairesse, J. (2002), ‘Explaining individual productivity differences in public research: how important are non-individual determinants? An econometric analysis of French physicists (1986–1997)’, Working Paper – Cahiers de la MSE 2002-66, Université Paris I – Panthéon-Sorbonne. Unité des Indicateurs de la Politique Scientifique (UNIPS) (1999), ‘Les publications des laboratoires du CNRS et leur impact’, March. Zucker, L., Darby, M. and Armstrong, J. (1998a), ‘Intellectual human capital and the birth of U.S. biotechnology enterprises’, American Economic Review, 88, 290–306. Zucker, L., Darby, M. and Brewer, M. (1998b), ‘Geographically localized knowledge: spillovers or markets?’, Economic Inquiry, 36, 65–86.
11. Epistemic communities and communities of practice in the knowledge-based firm1 Patrick Cohendet and Ash Amin INTRODUCTION As Fransman (1994) has argued, the traditional theoretical approaches to the firm – transaction costs theory in particular – considers the firm as a ‘processor of information’. For these traditional approaches, the behaviour of the firm can be understood as an optimal reaction to external environmental signals detected by the firm. The focus is thus on the process of allocation of resources needed to cope with this adaptation. In this perspective, the establishment of incentive schemes to avoid informational asymmetries is the main driving force for firm governance mechanisms. The competence-based approach, in contrast, relies on a different point of view: the firm is conceived as ‘a processor of knowledge’, that is, as a locus of construction, selection, usage and development of knowledge. This vision strongly differs from the above ‘information-based’ theories of the firm. Considering the firm as a processor of knowledge leads to the recognition that cognitive and related processes are essential, and that routines play a major role in keeping the internal coherence of the organisation. In this perspective, the governance of the firm is no more focused on the resolution of informational asymmetries but on the co-ordination of distributed pieces of knowledge and distributed learning processes. The focus of theory now falls on the process of creation of knowledge resources, as evidenced through the work of, among others, Cyert and March (1963), Cohen et al. (1972), Cohen (1991), Loasby (1976; 1983), Eliasson (1990), Dosi and Marengo (1994), and Marengo (1994; 1996). The existence of two alternative visions of the firm (processor of information versus processor of knowledge) raises some fundamental theoretical questions. As Langlois and Foss (1996) have pointed out, we are confronted with the choice between, on the one hand, a contractual approach based on transaction costs, where ‘firms and other institutions 296
Epistemic communities and communities of practice
297
are alternative bundles of contracts understood as mechanisms for creating and realigning incentives’, and, on the other hand, ‘a qualitative co-ordination, that is helping co-operating parties to align not their incentives but their knowledge and expectations’ (pp. 10–11). But, does the competence approach represent a complementary or competing approach to traditional theories of the firm, in particular the dominant transaction based approach? Which of the two approaches is the best equipped to explain what Casson (2000) considers to be the main issues addressed by modern theories of the firm, namely: (1) the boundary of the firm; (2) the internal organisation of the firm; (3) the formation growth and diversification of the firm, and (4) the role of the entrepreneur? The aim of this contribution is to shed light on some aspects of this theoretical debate, by addressing two issues. First, we argue that the two approaches to the firm are more complementary than substitutes. We assume that firms simultaneously manage competences and transactions, but they do so according to a specific order of priorities. They rank their activities according to their intensity of knowledge. Within the domain of core competences, the governance mechanisms are specifically devoted to knowledge co-ordination. Then, as we move away from the core, firms tend to allocate resources and adapt to the environment in accordance with governance mechanisms that are well analysed by the transaction cost approach. Second, we argue that the growing need to co-ordinate knowledge and achieve coherence within the firm (to avoid an excessive tension between different governance mechanisms) requires a reconsideration of the actual process of production and circulation of knowledge. We emphasise the role played by two specific communities, epistemic communities and communities of practice, in the formation and evolution of the routines of the firm.
1 1.1
COMPETENCES AND TRANSACTIONS RECONSIDERED The Firm as a Processor of Information
Traditional theories of the firm, especially the transaction-based approach and the principal/agent theory, share some common essential features. For them, the behaviour of the firm can be understood as an optimal reaction to information from the external environment. The firm as a rational processor of information implies that the same signals will through time give rise to the same pattern of action, provided that the technical conditions (as expressed by the production function) remain unchanged. The neoclassical theory of the firm, in particular principal/agent theory, have
298
The economics of knowledge
basically reduced co-ordination principles to a bundle of bilateral contracts which are meant to achieve co-ordination through appropriate incentive schemes to align self-interested action with common organisational goals. An important ommission of this approach is any serious treatment of the production process as a collective activity. The transaction cost approach, even if conceptually different and with a specific focus on the boundaries of the firm, comes to a similar fundamental conclusion: the firm could be seen as a nexus of contracts. Its very reason for existence is to correct market failures, when the operation of market mechanisms in terms of information processing is too costly. Transaction cost theory agrees with the principal/agent vision that information is imperfect and that the existence of potential asymmetries of information authorises unproductive rent-seeking behaviour. The firm is thus conceived as an institutional mechanism offering a governance structure to solve the problem of misaligned incentives related to imperfect information. It should be emphasised that, in such a vision, the unit of analysis is the transaction, with the productive activities which a firm undertakes out of focus. The transactional approach is by definition a defensive one in so far that it reacts to an environment characterised by imperfect information and agents who are opportunistic. Thus, it is in a sense a ‘theory of frictions’ to quote Favereau (1989). The focus on adaptation to imperfect information signals from the environment does not imply that the contractual approaches are unable to incorporate some aspects of cognition and learning among economic agents. The transaction cost approach assumes action based on bounded rationality, which is to admit the existence of cognitive constraints upon individuals, and to analyse key learning processes as learning by doing. However, the scope of analysis is extremely narrow, since the cognitive capabilities of agents are taken as given. Agents do not change their representation of the world through time, they do not differ in their perception of the environment, and they do not pay attention to the definition and evolution of common sets of rules, codes and languages within the organisation. To some extent, it can also be said that the traditional approach tackles the problem of knowledge. But here, too, the analysis is restricted to a very limited conception of knowledge: that knowledge is a mere stock resulting from the accumulation or loss of information considered as a flux. This is a narrow conception of information that does not acknowledge the cognitive and non-cognitive mechanisms involved in the production of knowledge. 1.2
The Firm as a Processor of Knowledge
A substantial volume of research, from different disciplines (economic history, industrial organisation, sociology of organisation, evolutionary
Epistemic communities and communities of practice
299
theory, management science, and so on) has seriously questioned the contractual vision of the firm (see Penrose, 1959; Richardson, 1960; 1972; Chandler, 1962; 1992; Nelson and Winter, 1982; March and Simon, 1993). The criticisms converge towards a common view that the concept of competence is a leading variable for explaining the organisation of firms, as well as their diversity and persistence (Dosi and Marengo, 1994). The concept of competence, which relies on that of routines and rules, centres on a view of the firm as a social institution, the main characteristic of which is to know (well) how to do certain things. Competences are coherent sets of capabilities used in an efficient way. Some of the competences are strategic (‘core-competences’ according to Teece, 1988) and constitute the main sources of the competitiveness of a firm (‘what a firm does well and better than the others’). They are the product of a selection process both internal and external to the firm. How these competences are constructed, combined and managed is critical for understanding the boundaries of the firm as well as the co-ordination and incentive structure of the firm. According to this alternative approach, the firm is conceived as an institution where competences are actively and consciously built, shaped, maintained and protected. This is a cumulative and strategic process that relies intensively on the management of knowledge. This has important consequences in particular for co-ordination by the firm. First because knowledge of complex production processes is necessarily distributed (cf. Hayek, 1937) and cannot be fully grasped and controlled by a single individual, a primary role of organisation becomes that of co-ordinating this dispersed knowledge. Second, co-ordination in this case generally involves the creation of commonly shared bodies of knowledge: sets of facts, notions, ‘models of the world’, rules, procedures which are – at least partly – known to all the members of the organisation involved in a given interaction. In a sense this kind of co-ordination is a precondition for the co-ordination of actions that are examined by the literature which implicitly assumes that all these mechanisms for the co-ordination of dispersed knowledge are already in place. It is most unlikely that incentive mechanisms alone could be sufficient to promote this kind of co-ordination. But, and perhaps even more important, the focus on knowledge issues highlights the question of how such knowledge is generated, maintained, replicated and modified (and possibly also lost) – that is, the question of learning and its nature. As repeatedly argued (see, for example, Nelson and Winter, 1982; Dosi and Egidi, 1991), innovative activities involve a kind of learning that is quite different from Bayesian probability updating and regression estimation. It requires agents to build new representations of the environment, which remains largely unknown, and to develop new skills which enable them both to explore and exploit this world of ever-expanding opportunities.
300
The economics of knowledge
Such representations are embedded in the routines which characterise the organisation. 1.3 Competences and Transactions Viewed as Complementary Mechanisms In our view the core statement of the competence theory is that the firm must be seen in primis as a processor of knowledge, not just as a mere information processing device. But in this perspective, the relationships between competences and transactions may be viewed as complementary. To be more specific: 1.
The firm will first focus its limited attention on its core competences. Within this set of core competences, the firm functions as a knowledge processor that gives full priority to the creation of resources. Such a focus signifies that the activities that belong to the ‘core’ of the firm are not considered as tradable in the market: they are ‘disconnected’ from any ‘make or buy’ trade-off as suggested by transaction cost theory. However, the scope of the set of core competences is limited; focusing on core competences is by definition very costly. This requires specific sunk costs, forging and managing co-operation with institutions that have complementary forms of knowledge, accessing and absorbing the most recent scientific results related to the core competences, and so on. For a given firm, in terms of the exchange of knowledge, this zone is characterised by ‘partners’ or ‘quasi-integrated’ suppliers, who produce high-value components or systems that are highly strategic. These could be wholly owned suppliers or partly owned suppliers in which the firm holds an equity stake and typically transfers personnel to work on a part-time or full-time basis. These suppliers participate in long-term strategic plans, capital investments and capacity planning, and personnel transfers. The formal duration of the typical contract is long-term, and most contracts are renewed automatically. The suppliers also tend to participate in building the knowledge base of the firm, and benefit from the absorptive capacities accumulated by the firm. But it is also important for the firm to enhance the absorbing capacities of the suppliers themselves. The firm provides assistance to suppliers not only in the areas of quality, cost reduction, factory layout and inventory management, but also in terms of increasing technological competences and research facilities. What is essentially transferred in this zone are creative ideas through multiple functional interfaces (manufacturing to manufacturing, engineering to engineering, and so on). This requires permanent capabilities benchmarking within the
Epistemic communities and communities of practice
2.
301
group of partners, a substantial investment in inter-firm knowledge shared routines, and regular activities of socialisation. In contrast the relationship with competitors in this zone is highly unstable and conflicting, and leads generally to acquisitions or mergers. In decreasing order of attention, next comes the domain of non-core competences. This is the domain of activities, which the firm ‘knows’ well to do, but does not necessarily invest much in the systematic production of new knowledge to place it at the leading edge of competition. To make the knowledge effective, the firm mostly has to function within networks and diverse types of alliances where it can access the complementary forms of knowledge required to makes its own knowledge valuable. In terms of transfer of technology what is at stake in this zone is the mutual exchange of complementary forms of knowledge. Networks offer precisely such an opportunity. What differentiates a given economic agent from another is its specific body of tacit knowledge. Through networks, agents can organise an efficient circulation of codified knowledge through a structure that renders compatible different segments of agent-specific tacit knowledge. Agents agree to specialise in a given area of tacit knowledge, because they are confident that the other agents will increase their specialisation in complementary forms. This reduces the risks of overspecialisation, but relies centrally on building mutual trust and reciprocity in the production of knowledge. Taking into account the degree of trust raises an important issue, which has to do with the choice between specialisation and co-operation in the production of knowledge. As argued by Zuscovitch (1998: 256): Trust is a tacit agreement in which rather than systematically seeking out the best opportunity at every instant, each agent takes a longer perspective to the transactions, as long as his traditional partner does not go beyond some mutually accepted norm. Sharing the risks of specialization is an aspect of co-operation that manifests an important trust mechanism in network functioning. Specialization is a risky business. One may sacrifice the ‘horizontal’ ability to satisfy various demands in order to gain ‘vertical’ efficiency in an effort to increase profitability. Any specializing firm accepts this risk, network or not. A risk-sharing mechanism is essential because, while aggregate profits for participating firms may indeed be superior to the situation where firms are less specialized, the distribution of profits may be very hazardous. To make specialization worthwhile, the dichotomous (win-lose) individual outcome must be smoothed somehow by a cooperative principle of risk sharing.
Trust is relevant for the reliability of other specialised producers of complementary knowledge.2 In such a perspective, one should not worry too much about excessive uncontrolled spillovers and risks of
302
3.
The economics of knowledge
excessive imitation, precisely because of significant transaction costs. Imitating is very costly, and loose co-operation in informal networks allowing a certain control of the diffusion of spillovers between agents, can be an efficient form of collaboration. One of the key determinants of innovative networks is the constant trade-off by agents between, on the one hand, delimiting property rights and, on the other hand, termination of rights of access to complementary forms of knowledge. Finally, away from the focus of attention on competences, one finds the peripheral activities. Once the set of activities that belong to the core and non-core competences has been chosen, the other activities that do not belong to the domain of competences are managed under traditional methods which may rely on the transaction cost approach. These activities are necessary to support the domain of competences and they generally correspond to the larger number of activities and employment positions in the firm. These activities do not require by definition a strong commitment in terms of knowledge. The firm just needs to ‘be informed’ of the best practices of external firms and organisations that can offer equivalent support services and, if it appears that these activities are too costly to be run within the firm compared to market mechanisms (according to transaction costs criteria), they will be outsourced. This is a zone of ‘quasi-market’ relations, where the degree of supplier–buyer interdependence is generally low. Products are standardised, and require few interactions with other inputs. Contracts are at arm’s length, the duration of which depends on the classical transactional parameters. For a given firm, in terms of supplier management practices, this zone requires minimal assistance to suppliers, accompanied by single functional interfaces (sales to purchasing, for instance), and the practice of price benchmarking. In terms of technology transfer what is at stake in this zone is the exchange of an artefact, rather than innovative ideas or new tacit knowledge.
Such a ranking of activities is suggested by Langlois and Foss (1996), when they argue that beyond core competences, firms will rank their activities according to an index of growing distance. As we move away from the core, we enter a domain which is more and more regulated by the classical need to process information. A rendering along the lines of the modern economics of organisation may be: as firms move increasingly away from their core businesses, they confront increasing adverse selection and moral hazard, since management becomes increasingly unable efficiently to monitor employees or to evaluate their human capital. Agency costs rise correspondingly, producing the net profitability disadvantage associated with further integration (for a similar story, see Aghion and Tirole, 1995).
303
Epistemic communities and communities of practice Intensity of focus of knowledge
Partnering Zone 1
Network Zone 2
Market Zone 3
Distance from the core
Source: Amesse and Cohendet (2000).
Figure 11.1
Ranking of activities of the firm from core competences
Figure 11.1 illustrates the above ranking of knowledge activities within the firm. The first zone (Zone 1) is the core itself. The second zone (Zone 2) is one where the firm holds significant pieces of knowledge, but needs to access complementary forms of knowledge held by other firms to be able to develop and use the knowledge efficiently. This zone is characterised by ‘networks’. The third zone (Zone 3) is the peripheral zone, where the firm does not hold any specific advantage in terms of knowledge. 1.4 Economic Consequences of the Ranking of Activities from the Core to the Periphery The consequences of this ‘lexicographic’ choice first focusing on competences, then managing the periphery, are significant. Let us emphasise two of them: 1.
In terms of dynamic evolution: the above representation is essentially static. It corresponds to the actual ranking of activities within the firm at a given moment of time. However, the dynamic functioning of activities could be interpreted along the lines of the evolutionary theory of the firm (Cohendet et al., 2000). Through the combined mechanisms of selection and variation in the body of existing routines, lies always
304
2.
1.5
The economics of knowledge
the possibility of transforming a set of secondary routines situated in the periphery into a new competence. Naturally, the reverse mechanism is also possible. For instance, routines that belong at a given moment in time to the core domain could be ‘declassified’ to the competence or the peripheral domain over time if they happened not to be successful. It must be emphasised that in the selection process that operates on routines, in addition to the classical external competitive environment, the attention of the firm operates as an internal element of selection. In terms of governance structures: the ranking of activities emphasises the need for the firm to define at least two distinct governance priorities; first, a structure to manage competences in order to align dispersed knowledge and expectations, and second, a structure conceived along the transaction costs criteria to manage the periphery. In a recent article, Nonaka and Konno (1998) underline the importance of dominant learning practices for managing the core of the organisation (concept of ‘ba’). Within this ‘core’ structure, some contractual schemes may naturally be implemented (for example, stock options, or specific rewards for inventors within the organisation), but these are not essential when compared with the priority given to the stimulation of collective learning processes. In the second structure of governance, classical contractual schemes are dominant to ensure the information processing that is central to the functioning of the periphery.3 Summary
We have argued thus far that a competence perspective on the firm – as opposed to a contract or transaction-based perspective – opens up considerably the scope for exploring how firms learn and adapt in complex and changing business environments. The acquisition and renewal of knowledge – tacit and formal – is crucial for survival, it cannot be taken as pregiven, and it occurs at a variety of levels through a variety of means. For such reasons, the firm has to be seen as much more than an allocating mechanism. It is primarily a generator of resources, defined as distinctive knowledge and organisational routines, locked in core and non-core competences. We have also argued that the cognitive set-up of firms – more specifically their organisational rationality – is crucial for framing expectations and outcomes. Thus, for example, substantive rationality, based upon the principle of rulefollowing behaviour, is efficient for decision-making in stable and simpleresponse situations (for example, mass production for planned markets), but inappropriate for a continually changing environment. In contrast, a procedural rationality favours learning through continual adjustment as
Epistemic communities and communities of practice
305
agents modify their behaviour to external circumstances, but is ill-equipped for strategic action in the context of radical change. In practice, individual firms tend to draw on these rationalities selectively depending on the nature of the task in hand, but routines often settle around one dominant cognitive set-up, which is precisely why firms vary in their learning potential. We have also claimed that a key governance challenge for firms is to reconcile organisational arrangements for transactional efficiency with those for sustaining learning. This is much the same as the exploration versus exploitation trade off signalled by James March (1991) to mark the dilemma faced by firms in balancing the short-term exploitation of existing competence and the long-term exploration of new competence. Thus in the ‘learning domain’ organisational and management practices must facilitate the creation and circulation of knowledge as well as strengthen decision-making. The challenge is to build trust, long-term commitments and knowledge externalities, to encourage experimentalism, variety and creative friction, and to facilitate the conversion of knowledge (between tacit and explicit, between individual to collective, between local and global). All this tends to privilege de-centred management, distributed capability and, to a degree, organisational ‘excess’ (Nohria and Ghoshal, 1997). In contrast, the transactional domain of routine activities (for example, securing supply, achieving scale economies, make–buy trade-offs) demands the efficient allocation of resources, largely through substantive or procedural responses to the environment. Here, as ever, the governance choice is between hierarchy and market, based on the cost-efficiency of transactions largely of a contractual nature.
2
THE COHERENCE OF THE FIRM AS A PROCESSOR OF KNOWLEDGE: ORGANISATIONAL LEARNING BETWEEN EPISTEMIC COMMUNITIES AND COMMUNITIES OF PRACTICE
The perspective that we have set out, based on matching two distinct operational domains and governance imperatives, might imply that the firm is a neatly divided entity. The aim of this section is to avoid this risk by framing the firm as a set of overlapping learning practices, involving recursive interchange between the domains and solutions emerging from the practices themselves. Ongoing learning in ‘epistemic communities’ and ‘communities of practice’ within firms bridges different types of rationality in separate domains. Consequently, in practice, the resolution of the governance problem of reconciling creation and allocation of resources is
306
The economics of knowledge
less a matter of formally integrating two separate models of governance, than a question of how meaningful links are generated in daily practice across the distributed mechanisms of governance. 2.1
Bridging Epistemic Communities and Communities of Practice
In the traditional vision of the firm, the distinction between epistemic communities and communities of practice is based on a linear representation of the process of transformation of knowledge. This process is viewed as evolving from separate departments in charge of producing new (deliberate) knowledge or handling and distributing information to the other departments that assimilate and use this new knowledge to improve their current activities. These latter departments can to some extent produce some new knowledge from their routine activities, but this would be a nondeliberate form of production of knowledge that emerges as a by-product of learning by using or learning by doing. We will argue that the differentiation between deliberate and non-deliberate forms of knowledge production is becoming strongly blurred. In a knowledge-based context the essence of the coherence of the firm precisely relies on the ways these two types of communities deliberately interact and organise simultaneously the production and circulation of knowledge. But, first the main characteristics of each community need to be detailed: 1.
2.
Epistemic communities are involved in the deliberate production of knowledge. They comprise ‘agents who work on a mutually recognized subset of knowledge issues, and who at the very least accept some commonly procedural authority as essential to the success of their collective building activities’ (Cowan et al., 2000: 234). The existence of procedural authority aids in the resolution of potential disputes and provides a reference point for achieving ‘closure’ in various stages of the codification process. In these communities the knowledge base from which the agents work is generally highly codified, but ‘paradoxically, its existence and contents are matters left tacit among the group unless some dispute or memory problem arises’ (Cowan et al., 2000: 234). What characterises the knowledge activities within these communities is that they are deliberately focused on the production of new knowledge, with a priori little reference to the different contexts in which the new knowledge produced will be used. Communities of practice involve learning in doing, or the nondeliberate production of knowledge. Wenger (1998) defines a community of practice as one marked by three dimensions, which take shape through repeated interaction, rather than rule or design. The first is
Epistemic communities and communities of practice
307
mutual engagement among participants, involving negotiating diversity, doing things together, mutual relationships and community maintenance. The second is joint enterprise, involving the negotiation of diversity among members, the formation of a local code of practice and a regime of mutual accountability. The third dimension is a shared repertoire that draws on stories, artefacts, discourses, concepts, historical events, discourses, reflecting a history of mutual engagement, ambiguity of meaning through new metaphors, and dynamic co-ordination through the latter ‘tools’ of engagement. Thus a community of practice – drawing on the subconscious, interaction, participation and reified knowledge to act, interpret, innovate and communicate – acts as ‘a locally negotiated regime of competence’ (ibid.: 137), as ‘shared histories of learning’ (ibid.: 86).4 What characterises these communities is that the production of knowledge is a by-product of the practice. In practice, increasingly the frontier between the two communities is becoming blurred. As Lundvall (2000) has pointed out, the emergence of new forms of learning such as ‘experimental learning’ makes the differentiation between ‘on-line’ and ‘off-line’ learning activities less and less relevant. The importance of experimental learning has been first emphasised by Paul David in a contribution with Warren Sanderson (David and Sanderson, 1996). For Lundvall (OECD, 2000: 25) experimental learning may start ‘on-line’, that is to say, during the process of producing a good, but consists in deliberately experimenting during the production process: By doing so, one creates new options and variety. This form of learning is based on a strategy whereby experimentation allows for collecting data, on the basis of which the best strategy for future activities is chosen. For example, a professor can undertake pedagogical experiments; the craftsman can seek new solutions to a problem even during the fabrication process. The possibility of moving this type of learning in many activities represents an important transition in the historical emergence of the knowledge-based economy. In effect, as long as an activity remains fundamentally based on learning processes that are routine adaptation procedures and leave no room for programming experiments during economic activity, there remains a strong dichotomy between those who deliberately produce knowledge and those who use and exploit it. When an activity moves to higher forms of learning, and where the individual can programme experiments and obtain results, the production of knowledge becomes much more collectively distributed . . . With the emergence of experimental learning, the feedback and reciprocal links that tie ‘on-line’ learning processes and in house R&D together – and whereby a potential creative activity effectively contributes to the production of knowledge – become crucial.
308
The economics of knowledge
This implies a complete reconsideration of the separation between epistemic communities and communities of practice. The management of the collectively distributed knowledge within the organisation that brings together epistemic communities and communities of practice is one of the cornerstones of the coherence of the firm in a knowledge-based context. The development of different modes of interaction between the two types of communities (that rely on particular processes of codification of knowledge) becomes critical. Devices are needed to overcome the problem of knowledge fit or integration across boundaries (for example, longer-term ‘languaging’ devices, informal crossovers and deployment of ‘boundary spanning’ individuals such as brokers and intermediaries). Nooteboom (1999a) has emphasised the role of third party ‘go-betweens’ as vital brokers of innovation who help to sediment trust, maintain unique secrets, resolve conflicts, reveal mutual advantage and introduce innovation without destabilising established competences within each firm. Learning, then, is a fine-grained and grounded process (Gibbons et al., 1994) that straddles across the deliberate and the routine, across the epistemic community and the community of practice. It involves trial and many errors, chance discoveries, and mistakes, in a context of firms operating as ‘experimental learning machines’ (Eliasson, 1994) in uncertain circumstances. Their daily hazard is to ‘act prematurely on a very incomplete information base’ (Eliasson, 1994: 184). This is not to say, of course, that all is left to chance as we shall see below, but it does mean that distributed or varied learning, is neither guaranteed nor that easily ‘arranged’ (Metcalfe, 1998). In many regards, the daily practice of learning across the boundaries of the firm helps adaptation and continuity. And here, the ‘anthropological’ literature on firms, stressing the generation of knowledge through practice and social interaction, is particularly insightful. For example, in their seminal article on communities of practice, John Seely Brown and Paul Duguid (1996) argue that learning and innovation only too often are situated practices in the everyday humdrum of interaction with one’s peers and with the environment. They explain: Alternative worldviews, then, do not lie in the laboratory or strategic planning office alone, condemning everyone else in the organisation to a unitary culture. Alternatives are inevitably distributed throughout all the different communities that make up the organisation. For it is the organisation’s communities, at all levels, who are in contact with the environment and involved in interpretative sense making, congruence finding, and adapting. It is from any site of such interactions that new insights can be co-produced. If an organisational core overlooks or curtails the enacting in its midst by ignoring or disrupting its
Epistemic communities and communities of practice
309
communities-of-practice, it threatens its survival in two ways. It will not only threaten to destroy the very working and learning practices by which it, knowingly or unknowingly, survives. It will also cut itself off from a major source of potential innovation that inevitably arises in the course of the working and learning. (Ibid.: 76)
It is clear that, for Brown and Duguid, organisational innovation is related to tolerance for diversity (as most evolutionary economists would agree), but the important point here is the emphasis placed on the development of alternatives (that is, exploration) within autonomous communities of practice within the organisation. Once again, they explain: Within an organisation perceived as a collective of communities, not simply of individuals, in which enacting experiments are legitimate, separate community perspectives can be amplified by inter-changes among communities. Out of this friction of competing ideas can come the sort of improvisational sparks necessary for igniting organisational innovation. Thus large organisations, reflectively structured, are perhaps well positioned to be highly innovative and to deal with discontinuities. If their internal communities have a reasonable degree of autonomy and independence from the dominant worldview, large organisations might actually accelerate innovation. (Ibid.: 77–8)
In this account of innovation, the impression we have is that groups within organisations act, at one and the same time, recursively (or reflectively) and procedurally. The variety and redundancy necessary for organisational learning is situated within each part of an organisation or firm. Every organisation is made up of many communities of practice in which learning is not a matter of conscious design or recognisable rationalities and cognitive frames, but a matter of new meanings and emergent structures arising out of common enterprise, experience and sociability – learning in doing.5 Consistent with the argument thus far that learning is distributed, composite and realised in communities of practice, might also be the suggestion, increasingly peddled in the organisational literature, that the firm is an autopoetic cognitive system, a site for developing knowledge in a selfreferential manner, usually through internal communication and ‘languaging’ procedures at different scales of organisation (von Krogh and Roos, 1995; Magalhães, 1998). Indeed, Salvatore Vicari and Gabriele Toniolo (1998) argued that the firm should be seen as a cognitive system which enacts and makes sense of the environment only from its own individual point of view. As such, it makes no sense to theorise learning as a linear or procedural reaction to an environment ‘out there’, since, at least according to Vicari and Toniolo, the latter is known only through the firm’s cognitive schemata (scripts, maps, data and stories), which define the structure of its
310
The economics of knowledge
knowledge. The firm has knowledge only of itself but, equally, it needs to produce knowledge in order to survive. For Vicari and Toniolo, a firm does this via enactment, that is, by constructing maps and images of reality based on raw data that are monitored and interpreted, expectations about customer and competitor behaviour, time and motion studies, and so on. The performance of these enactments, rather than ‘in here’ anticipation of ‘out there’ signals, is the basis upon which firms know and act. Such an interpretation of knowledge formation in the firm considerably opens up the scope for seeing learning and innovation as a fragile, experimental and uncertain process. So much so, that Vicari and Toniolo (1998) claim that innovation and learning is made possible via the ‘production of errors’ (ibid.: 211) – errors which could imperil the firm if the gap between the enacted knowledge scripts and say customer preferences and expectations remains wide. They distinguish between chance errors, generated by alterations in the external environment (for example, new competitors or technical changes) or transformations in the cognitive structure (for example, the discovery of unusual behaviours or product shifts) and intentional errors, produced through intentional search for disturbing events (for example, customer surveys or firing management) or experimentation (for example, via acquisitions or the launch of new product lines). 2.2
Governance Implications
The management implication of stressing the role of error could be twofold: either that there is only so much that can be done in order to overcome the ever-present gap between autopoietic knowledge and market preferences; or, as argued by Vicari and Toniolo, that firms should foster learning through error production, for example, by developing procedures to detect and amplify market signals (for example, benchmarking to detect the size of error, or acknowledgement of error throughout the firm). Perhaps both implications are valid, especially if we take seriously the claim that learning occurs in communities of practice which combine procedural and recursive knowledge, exploration and exploitation. The broader implication of the emphasis on learning in communities of practice is that the ‘management by design’ of learning is not feasible. Instead, as already indicated earlier, the imperative is to maximise the potential for de-centred and distributed learning. This includes the removal of organisational barriers designed for time- and resource-efficiency in the field of innovation, or, put differently, root-and-branch evaluation of the varied tasks of the organisation followed by the application of transactional rules of governance only to a restricted set of functions and corporate priorities. It requires, above all, the identification and full acknowledgement of
Epistemic communities and communities of practice
311
the communities of practice in play within and across the boundaries of the firm. It also requires analysis of the learning processes within these communities, and actions designed to facilitate new learning within them as an ongoing process – not dispositions from the centre. Part of this recognition of de-centred learning involves mobilising the ‘soft’ infrastructure for learning – both procedural and recursive – and recognising its integral relationship with the formal infrastructure for learning (from courses and texts to artefacts and technologies). We have already had a glimpse of some of the elements. Slack, memory and forgetting are clearly of crucial importance: slack, facilitated by the retention of skills and capabilities in excess of the minimum necessary for immediate use, helps ‘innovative projects to be pursued because it buffers organisations from the uncertain success of these projects, fostering a culture of experimentation’ (Nohria and Ghoshal, 1997: 52); memory, facilitated by remembering, inter-generation mixture, and stories among long-serving employees, mobilises the fruits of experience, including knowledge of past trials and errors; forgetting, stimulated by employee rotation and mobility, new training and new routines, helps to weed out practices that are not suited for changing circumstances. Another aspect of the soft infrastructure, so clearly evident from the work of Wenger and Brown and Duguid is the practice itself of community. This does not necessarily mean consensus or trust and loyalty (if the latter two are meant to imply lack of conflict and dissonance). Instead, it refers to the daily practices of a group bound together by common purpose and expertise (for example, insurance claims processors, middle mangers, R&D workers). Learning and innovation is the product of shared expertise, social discourse, sociability, argument, disagreement, negotiation skills, and so on. It is the realisation of potential through the practices of joint interaction that is important. Of course, not all communities of practice are learning communities, leaving much to be done by management to stimulate ‘learning in doing’: decisional autonomy, allocation of complex tasks, away days and group reflections, opportunities for socialising, group work, opportunities for dispute resolution, and encouragement of talk, exchange of ideas and ‘manageable’ disagreement. A third aspect of soft learning is dissonance and experimentation. Creative communities are those which are able to mobilise difference, variety and counter-argument. Thus, to accompany exploitation of current opportunities through procedural efficiency, they need to regularise exploration, as a means of generating new routines, goals and possibilities. Exploration, or learning to learn, is routinely sacrificed in most organisational cultures largely for the threat it poses to current repertoire and competitive advantage. It therefore requires active pursuit, perhaps through the
312
The economics of knowledge
encouragement of ideas within groups, competition for ideas between groups, use of brokers and intermediaries to manage external exposure and adoption of new routines, encouragement of scenario-building and experimental games, and so on. Such a pursuit cannot be restricted to individual communities of practice, but has to involve the entire organisation, for as Richard Nelson (1994: 238) notes: ‘devising and learning to use effectively a significantly new organisational form involves much the same kind of uncertainty, experimental groping, and learning by making mistakes . . . that makes technological invention and innovation’. What remains for the central management in a de-centred organisational environment, other than actions such as those above to support the ongoing practices of distributed learning? Perhaps the central imperative is to hold the network of distributed competences in place. As Nohria and Ghoshal (1997: 87) explain in the context of the multinational corporation which relies on its subsidiaries for the creation, adoption and distribution of innovations: ‘the real leverage lies in creating a shared context and common purpose and in enhancing the communication densities within and across the organisation’s internal and external boundaries’. It is interesting to note that at the level of the organisation as a whole, here too, it is the soft infrastructure for learning that is increasingly stressed. What seems to be important is to find ways of binding the different subsidiaries, nodes in a network of alliances, or communities of practice, into a common enterprise, and perhaps in ways which go beyond the traditional mechanisms of propriety rights and the employment contract. Nohria and Ghoshal seem to want to emphasise the role of socialisation (for example, via corporate encounters, conferences and recreational clubs) and mechanisms that lead to normative integration (for example, membership incentives such as access to health care or travel concessions, company rituals, slogans, and mission statements, inculcation of corporate or brand quality standards, and so on). No doubt there are many more integration mechanisms to ensure that the parts do somehow belong to the same whole. Nohria and Ghoshal also stress the architecture of communication as a central management concern in a system of de-centred learning. Such communication, however, is no longer a simple matter of information flow within and beyond a firm, as it is in the classical contract-based model designed to minimise transaction costs and other frictions impeding information processing. Now it is a matter of ensuring that there is effective communication between self-governing but interdependent units. It is crucially a matter of relational or cognitive proximity (Nooteboom, 1999b), implying linguistic and semantic equivalence, shared tacit knowledge, rapid flow and processing of information, trust or other conventions of
Epistemic communities and communities of practice
313
negotiation. Indeed, perhaps the better word is interaction, with its active promotion and management a key management priority. The use of state-of-the-art communications technologies to facilitate real-time and two-way interchange across the boundaries of the firm is, of course, one aspect of the architecture of interaction, but it is not the only aspect. Von Krogh et al. (2000) for instance have argued the centrality of ‘knowledge enablers’ as boundary managers in learning-based networks. For example, they stress the role of interaction through dialogue with customers based on shared tacit knowledge as a way of constructing markets, they argue the importance of entering into the domain of suppliers or users of technology through personnel exchanges, they emphasise the need to develop organisational conversations through block conferences, personnel rotation, and exchange of business plans among decentralised units, and they recommend knowledge managers as a new professional cadre dedicated, among other things, to managing relationships across internal and external boundaries. Governance of the architecture for dissipated learning, thus, is the central management challenge in the domain of resource mobilisation, alongside, of course, as argued earlier, managing the domain of resource allocation. On this point, to return to the original paradox of reconciling exploration and exploitation, the major conflict may well lie in the culture clash between procedural and recursive thinking within whatever is designated as ‘management’ in an organisation. At the level of communities of practice, we suggested that everyday ongoings tend to generate a hybrid culture that is capable of both experimentation and routine response. At the level of designated corporate or net-wide management, groups tend to be specialised in either the preservation of routine or the search for novelty. How to de-centre and reintegrate the centre? One of the main issues to solve the tension between centralisation and decentralisation in the organisational learning process is the building of a common knowledge, specific to the firm, that integrates the decentralised bodies of knowledge held by members of the firm. As has been said, firms require both centralisation and decentralisation to operate successfully in a changing environment. Decentralisation in the acquisition of knowledge is a source of diversity, experimentation and ‘ultimately’ of learning. But, eventually, knowledge has to be made available for exploitation by the entire organisation. When agents differ with regard to their representations of the environment and their cognitive capabilities, a body of common (or collective) knowledge must exist, within the organisation, which guarantees the coherence of the various learning processes (Crémer, 1990). This is a prerequisite for an efficient management of the competences. In order to cope with changing environments, the process of generation and modification of
314
The economics of knowledge
such a body of common knowledge, although fed by decentralised learning processes, has to undergo some forms of centralisation, even if it is basically maintained by decentralised learning processes (Cohendet and Llerena, 1991).
CONCLUSION This contribution has stressed that the production and circulation of collective organisational knowledge is a key determinant of the capability of the organisation to innovate. It considers the ‘cognitive architecture’ of knowledge within the firm (the way knowledge is produced, stored, exchanged, transmitted and retrieved) to strongly influence the process of organisational learning and, in turn, the creation process. Two main features of the cognitive architecture of the firm have been emphasised: first the ‘ranking of activities’ by the firm in terms of the level of attention required in a knowledge context; second, the nature of interactions between epistemic communities and communities of practice within the firm. Of course, many other types of community participate in the production and circulation of knowledge within the firm. For instance, project teams have always been considered as key organisational devices to bring together different forms of heterogeneous functional knowledge held by agents, in order to develop new knowledge. Our purpose has not been to cover all knowledge actor-networks. However, we are assuming that in the knowledge-based economy, of growing importance will be the mutual interactions between communities of practice and epistemic communities, as a tool for both extensive learning and governance cohesion. This perspective on knowledge communities opens a new avenue of research in the competence-based theory of the firm. In this contribution we have concentrated on governance mechanisms. It certainly would be of major interest to investigate, within such a perspective, individual incentive mechanisms, the relationships between managers and stakeholders, the role of the entrepreneur and many other key features of the learning and governance environment of the firm.
NOTES 1. This contribution has been particularly inspired by ideas brought to the TIPIK project by Paul David on epistemic communities and the collective building of knowledge. It draws on our joint work, in particular, Amin and Cohendet (2000). 2. The institutionalisation of incentives for validation (peer refereeing , for instance) in epistemic communities may vary widely. The choice for an agent to specialise in one domain
Epistemic communities and communities of practice
315
of knowledge (and to bear the sunk costs) in co-operation with others agents that accept specialising in complementary types of knowledge is an important line of research to understand the management of knowledge by organisations. 3. The possibility of a dual structure of governance within the same firm raises the problem of internal coherence which could also impact on the internal process of technological transfer. The conflicting tensions within a firm that is at the same time trying to organise and manage its information flow and its knowledge base have been analysed by Marengo (1994), through a discussion of the limits of the well-known multi-divisional form (‘M-form’) and functional form (‘U-form’). He notes that these traditional forms are conceived to solve information problems (information overload by managers, in particular), but that they are not appropriate for creating and transferring knowledge. He argues, for example, that the U-form centralises competences in inter-functional co-ordination and decentralises instead to functional departments competences in many strategic issues concerning products and diversification. With the growing multiplicity of products the functional structure does not seem that of information overload, but that of mismatch between competences and tasks. Chief executives are unable to do their job effectively, not because they are burdened by an excess of information, but rather because the organizational structure does not enable them to develop the necessary competences. Chief executives should respond to environmental changes, but when such changes push towards product diversification, many of the competences that are necessary to promote and manage diversity remain, in the U-form, at the level of functional departments. In the same vein, he also argues that the M-form, by preventing cross-fertilisation of innovative ideas between separate departments, does not favour the internal transfer of knowledge. 4. Learning, for Wenger, occurs through ongoing practice and draws on social energy and power generated through interaction in joint enterprises with some history. He identifies three infrastructures of learning – corresponding to the three dimensions of a community of practice – which potentially have enough novelty, perturbation and emergence in them to sustain incremental and discontinuous learning, as well as procedural adaptation and goal monitoring. These are infrastructures which draw upon a staggeringly broad range of facilities, tools, practices and conventions (lest we are inclined to think that learning in action is a simple process). One infrastructure is engagement, composed of mutuality (supported by such routines as joint tasks and interactive spaces), competence (supported by training, encouragement of initiative and judgement) and continuity (supported by reified memory locked in data, documents, files as well as participatory memory unlocked by storytelling and inter-generation encounters). Another is alignment, composed of convergence (facilitated by common focus, shared values and leadership), co-ordination (helped by such devices such as standards, information transmission, feedback, division of labour, and deadlines) and arbitration (facilitated by rules, policies and conflict resolution techniques). The third infrastructure is imagination, composed of orientation (helped by visualisation tools, examples, explanations, codes and organisational charts), reflection (supported by retreats, time-off, conversations and pattern analysis) and exploration (facilitated by scenario-building, prototypes, play, simulations, experimentation). These are embedded infrastructures of learning built into the routines and daily practices of members, and the facilities put into place through experience or management decision. They are features of all the communities of practice that are to be found within and across organisations. Wenger’s example of insurance claim processors should not be taken to mean that learning of the sort he describes does not apply to top management, strategists and scientists. Indeed, one of the remarkable early insights of applications of actor network theory in the literature on sociology of science, was to show that incremental and radical learning in the R&D laboratory is no different from the processes described by Wenger, locked as it is in routines, conversations, artefacts, things, memory, stories and so on. It has been argued recently that the resource-based perspective, with its emphasis on the centrality of knowledge creation for competitive advantage, ‘does not provide an explanation about how resources such as organisational knowledge develop over time’ (Probst et al., 1998: 243). Thus, the sophistication with which competitive advantage is
316
The economics of knowledge
explained at any point in time seems not matched in diachronic analysis of learning trajectories. Probst et al. provide a partial corrective by mobilising metaphors from evolutionary economics (for example, imitation, selection, variety, replication, and so on). For example, they relate changes in organisational knowledge to the cumulative effects of individual and group learning: when such learning becomes routinised and synthesised through agent interaction and shared beliefs over time, the organisational knowledge set can be said to be transformed. Thus, replication at one level becomes a precondition for innovation at another level. 5. The above perspective properly contextualises also the processes that feed into radical innovations forced by dramatic events. A revealing example is provided by Edwin Hutchins (1996) in his study of how a ship’s navigation team arrived at a new stable procedure when, upon entering a harbour, a large ship suffered an engineering breakdown that disabled a vital piece of navigational equipment. Following a chaotic and unsuccessful search for a solution from reflection involving thought experiments to computational and textual alternatives, the team developed an answer through acting. As local task were found for individuals distributed across the ship, the ensuing sequence of actions and conversations, drawing on experience and experimentation, led to the construction of a solution based on trial and testing. On this occasion, a solution was found on time, and there is every possibility that the ship could have got into very serious trouble. The point, however, is that other alternatives such as learning by design, once a snap solution failed to emerge, were not viable.
BIBLIOGRAPHY Aghion, P. and Tirole, J. (1995), ‘Some implications of growth for organizational form and ownership structure’, European Economic Journal, 39 (3), 440–55. Alchian, A. (1951), ‘Uncertainty, evolution and economic theory’, Journal of Political Economy, 57, 211–21. Alchian, A. and Demsetz, H. (1972), ‘Production, information costs and economic organization’, American Economic Review, 62, 777–95. Amesse, F. and Cohendet P (2000), ‘Technology transfer revisited from the perspective of the knowledge-based economy’, paper presented at the conference on Economics and Management of Knowledge, HEC Montreal, 5–6 May. Amin, A. and Cohendet, P. (1999), ‘Learning and adaptation in decentralized business networks’, Society and Space, 17, 87–104. Amin, A. and Cohendet, P. (2000), ‘Organisational learning and governance through embedded practices’, Journal of Management and Governance, 4, 93–116. Andersen, E.S. (1998), ‘The evolution of the organisation of industry’, Aalborg University, DRUID Working Paper 98–13. Aoki, M. (1986), ‘Horizontal vs. vertical information structure of the firm’, American Economic Review, 76, 971–83. Aoki, M. (1988), Information, Incentives and Bargaining in the Japanese Economy, Cambridge, Cambridge University Press. Aoki, M. (1990a), ‘Toward an economic model of the Japanese firm’, Journal of Economic Literature, 28, 1–27. Aoki, M. (1990b), ‘The participatory generation of information rents and the theory of the firm’, in Aoki, M., Gustafsson, B. and Williamson, O.E. (eds), The Firm as a Nexus of Treaties, London, Sage, pp. 26–52. Aoki, M., Gustafsson, B. and Williamson, O.E. (eds) (1990), The Firm as a Nexus of Treaties, London, Sage.
Epistemic communities and communities of practice
317
Arrow, K.J. (1974a), The Limits of Organization, New York and London, W.W. Norton. Arrow, K.J. (1985), ‘Informational structure of the firm’, American Economic Review, 75, 303–7. Avadikyan, A., Cohendet, P. and Llerena, P. (1995), ‘Coherence, diversity of assets: towards an evolutionary approach’, Revue International de Systémique, 7 (5), 505–31. Brown, J.S. and Duguid, P. (1996), ‘Organizational learning and communities-ofpractice’, in Cohen, M. and Sproull, L. (eds) Organizational Learning, London, Sage. Casson, M. (2000), ‘An entrepreneurial theory of the firm’, in Foss, N. and Mahnke, V. (eds), Competence, Governance, and Entrepreneurship, Oxford, Oxford University Press, pp. 116–46. Chandler, A.D. Jr (1962), Strategy and Structure: Chapters in the History of Industrial Enterprise, Cambridge, MA, MIT Press. Chandler, A.D. Jr (1992), ‘Corporate strategy, structure and control methods in the United States during the 20th century’, Industrial and Corporate Change, 1 (2), 263–84. Coase, R.M. (1937), ‘The nature of the firm’, Economica, 4, 386–405. Coase, R.M. (1988), ‘Lectures on “The nature of the firm” 50 years after’, Journal of Law, Economics and Organization, 4, 3–47. Cohen, M.D. (1984), ‘Conflict and complexity: goal diversity and organizational search effectiveness’, American Political Science Review, 78, 435–51. Cohen, M.D. (1987), ‘Adaptation of organizational routines’, mimeo, Ann Arbor, University of Michigan, Institute of Public Policy Studies. Cohen, M.D. (1991), ‘Individual learning and organizational routine: emerging connections’, Organization Science, 2, 135–9. Cohen, M.D. and Bacdayan, P. (1991), ‘Organizational routines are stored in procedural memory: evidence from a laboratory study’, mimeo, Ann Arbor, University of Michigan, Institute of Public Policy Studies. Cohen, M.D., Burkhart, R., Dosi, G., Egidi, M., Marengo, L, Warglien, M. and Winter, S. (1996), ‘Routines and other recurring action patterns of organizations: contemporary research issues’, Industrial and Corporate Change, 5 (3), 653–98. Cohen, M.D., March, J.G. and Olsen, J.P. (1972), ‘A garbage can model of organizational choice’, Administrative Sciences Quarterly, 17, 1–25. Cohendet, P. (1996), ‘Information, connaissances et théorie de la firme évolutionniste’, séminaire Economie de l’Information, Commissariat au Plan, Paris, 16 April. Cohendet, P. and Llerena, P. (1991), ‘Integration and learning processes’, in Bourgine, P. and Walliser, B. (eds), Economics and Artificial Intelligence, Oxford, Pergamon Press. Cohendet, P., Llerena, P. and Marengo, M. (2000), ‘Is there a pilot in the evolutionary firm?’, in Foss, N. and Mahnke, V. (eds), Competence, Governance, and Entrepreneurship, Oxford, Oxford University Press, pp. 95–116. Coriat, B. and Dosi, G. (1995), ‘Learning how to govern and learning how to solve problems: on the coevolution of competences, conflicts and organizational routines’, IIASA Working Paper No. 95–06 Laxenburg (Austria). Coriat, B. and Weinstein, O. (1995), Les nouvelles theories de l’entreprise, Livre de Poche, Collection Référence, Paris: La Découverte.
318
The economics of knowledge
Coriat, B. and Weinstein, O. (1996), ‘Sur la théorie évolutionniste de la firme’, Colloque ‘L’évolutionnisme, fondements, perspectives et réalisations’, Sorbonne, Paris, 19–20 September. Cowan, R., David, P. and Foray, D. (2000), ‘The explicit economics of knowledge codification and tacitness’, Industrial and Corporate Change, 9 (2), 212–53. Crémer, J. (1990), ‘Common knowledge and the co-ordination of economic activities’, in Aoki, M., Gustafsson, B. and Williamson, O.E. (eds), The Firm as a Nexus of Treaties, London, Sage, pp. 53–75. Crémer, J. (1993), ‘Corporate culture and shared knowledge’, Industrial and Corporate Change, 2 (3), 351–86. Crémer, J. and Riordan, M.H. (1987), ‘On governing multilateral relationships with bilateral contracts’, RAND Journal of Economics, 18, 436–51. Cyert, R.M. and March, J.G. (1963), A Behavioural Theory of the Firm, Englewood Cliffs, NJ, Prentice-Hall. David, P. (1994), ‘Why are institutions the “carriers of history”? Path dependence and the evolution of conventions, organizations and institutions’, Economic Dynamic and Structural Change, 5 (2), 205–20. David, P. and Foray, D. (1996), ‘Accessing and expanding the science and technology knowledge base’, STI Review, 16, 87–116. David, P. and Sanderson, W. (1996), ‘Making use of treacherous advice: cognitive learning, Bayesian adaptation; and the tenacity of unreliable knowledge’, MERIT working paper, 2/96–004, University of Limburg. Dosi, G. (1988), ‘Sources, procedures and microeconomic effects of innovation’, Journal of Economic Literature, 26, 1120–71. Dosi, G. and Egidi, M. (1991), ‘Substantive and procedural uncertainty. An exploration of economic behaviours in complex and changing environments’, Journal of Evolutionary Economics, 1, 145–68. Dosi, G. and Marengo, L. (1994), ‘Toward a theory of organizational competences’, in England, R.W. (ed.), Evolutionary Concepts in Contemporary Economics, Ann Arbor, MI, University of Michigan Press. Dosi, G., Freeman, C., Nelson, R., Silverberg, G. and Soete, L. (eds) (1988), Technical Change and Economic Theory, London, Pinter. Dosi, G., Teece, D.J. and Winter, S.G. (1991), ‘Toward of theory of corporate coherence’, in Dosi, G., Giannetti, R. and Toninelli, P.A. (eds), Technology and the Enterprise in a Historical Perspective, Oxford, Oxford University Press. Egidi, M. (1994), ‘Routines, hierarchies of problems, procedural behaviour: some evidence from experiments’, IIASA Working Paper No. 94–58, Laxenburg (Austria). Egidi, M. and Marengo, L. (1995), ‘Division of labor and social coordination modes: a simple simulation model’, in Gilbert, G.N. and Conte, R. (eds), Artificial Societies: The Computer Simulation of Social Life, London, UCL Press. Egidi, M. and Ricotilli, M. (1997), ‘Co-ordination and specialisation’, Working Paper 1997–03, University of Trento. Eliasson, G. (1990), ‘The firm as a competent team’, Journal of Economic Behaviour and Organization, 13, 275–98. Eliasson, G. (1994), ‘The theory of the firm and the theory of economic growth’, in Magnusson, L. (ed.), Evolutionary and Meo-Schumpeterian Approaches to Economics, Dordrecht, Kluwer. Farma, E. (1980), ‘Agency problems and the theory of the firm’, Journal of Political Economy, 88, 288–307.
Epistemic communities and communities of practice
319
Favereau, O. (1989), ‘Organisations et marchés’, Revue Française d’Économie, 4, 273–328. Favereau, O. (1993), ‘Suggestions pour reconstruire la théorie du salaire sur une théorie des règles’, WP Laetix, University of Paris X, April. Favereau, O. (1995), ‘Apprentissage collectif et coordination par les règles: application à la théorie des salaires’, in Lazaric, N. and Monnier, J.M. (eds), Coordination économique et apprentissage des firmes, Paris, Economica, pp. 23–38. Fontana, W. and Buss, L.W. (1994), ‘What would be conserved “if the tape were played twice”?’, Proceedings of the National Academy of Science USA, 91, 757–61. Foss, N.J. (1993), ‘Theories of the firm: contractual and competence perspectives’, Journal of Evolutionary Economics, 3, 127–44. Foss, N.J. (1996), ‘The “alternative” theories of Knight and Coase, and the modern theory of the firm’, Journal of the History of Economic Thought, 18 (Spring), 1–20. Fransman, M. (1994), ‘Information, knowledge, vision and theories of the firm’, Industrial and Corporate Change, 3 (3), 1–45. Geanakoplos, J. and Milgrom, P. (1991), ‘A theory of hierarchies based on limited managerial attention’, Journal of the Japanese and International Economies, 5, 205–25. Gibbons, M., Limoges, C., Novotny, H., Schwartzman, S., Scott, P. and Trow, M. (1994), The New Production of Knowledge, Beverly Hills, CA, Sage. Hammond, T.H. (1990), ‘The agenda of the firm: structure and incentive in institutional design’, conference on Fundamental Issues in Strategy: A Research Agenda for the 1990s, Napa, CA, December. Hart, O. (1988), ‘Incomplete contracts and the theory of the firm’, Journal of Law, Economics and Organization, 4, 119–39. Hayek, F.A. von (1937), ‘Economics and knowledge’, Economica, 4, 33–54. Hutchins, E. (1996), ‘Organizing work by adaptation’, in Cohen, M. and L. Sproull (eds), Organizational Learning, London, Sage. Itoh, H. (1987), ‘Information processing capacities of the firm’, Journal of the Japanese and International Economies, 1, 299–326. Jensen, M.C. and Meckling, W.H. (1976), ‘Theory of the firm: managerial behaviour, agency costs and ownership structure’, Journal of Financial Economics, 3, 305–60. Kauffman, S.A. (1993), The Origins of Order. Self-Organization and Selection in Evolution, New York and Oxford, Oxford University Press. Klepper, S. (1997), ‘Industry life cycles’, Industrial and Corporate Change, 6, 145–81. Knight, F.H. (1921), Risk, Uncertainty and Profit, Boston, MA, Houghton Mifflin. Kreps, D. (1992), ‘Corporate culture and economic theory’, in Alt, J. and Shepsle, K. (eds), Positive Perspectives on Political Economy, Cambridge, Cambridge University Press. Langlois, R. and Foss, N. (1996), ‘Capabilities and governance: the rebirth of production in the theory of economic organisation’, working paper, University of Connecticut/Copenhagen Business School. Levinthal, D. (1997), ‘Adaptation on rugged landscapes’, Management Science, 43, 934–50. Levinthal, D.A. and March, J.G. (1981), ‘A model of adaptive organizational search’, Journal of Economic Behaviour and Organization, 2, 307–33. Levitt, B. and March, J.G. (1988), ‘Organizational learning’, Annual Review of Sociology, 14, 319–40.
320
The economics of knowledge
Loasby, B.J. (1976), Choice, Complexity and Ignorance, Cambridge, Cambridge University Press. Loasby, B.J. (1983), ‘Knowledge, learning and the enterprise’, in Wiseman, J. (ed.), Beyond Positive Economics? London, Macmillan, pp. 104–21. Loasby, B.J. (1991), Equilibrium and Evolution: An Exploration of Connecting Principles in Economics, Manchester, Manchester University Press. Lounamaa, P.H. and March J.G. (1987), ‘Adaptive coordination of a learning team’, Management Science, 33, 107–23. Lundvall, B.A (2000), ‘Unlocking knowledge in society’, in OECD report Knowledge in the Learning Economy, OECD. Machlup, F. (1983), Knowledge, its Creation, Distribution, and Economic Significance, Princeton, NJ, Princeton University Press. Magalhães, R. (1998), ‘Organizational knowledge and learning’, in von Krogh, G., Roos, J. and Kleine, D. (eds), Knowing in Firms, London, Sage. March, J.G. (1991), ‘Exploration and exploitation in organizational learning’, Organization Science, 10 (1), 71–87. March, J.G. and Simon, H. (1993), ‘Organizations revisited’, Industrial and Corporate Change, 2 (3), 299–316. Marengo, L. (1994), ‘Knowledge distribution and coordination in organizations: on some social aspects of the exploration vs. exploitation trade-off’, Revue Internationale de Systémique, 7, 553–71. Marengo, L. (1996), ‘Structure, competence and learning in an adaptive model of the firm’, in Dosi, G. and Malerba, F. (eds), Organisation and Strategy in the Evolution of the Enterprise, London, Macmillan, pp. 124–54. Marengo, L. (1998), ‘Interdependencies and division of labour in problem-solving technologies’, presented at the conference, Competences, Governance and Entrepreneurship, DRUID, Bornholm, June. Marschak, J. and Radner, R. (1972), Economic Theory of Teams, New Haven, CT, Yale University Press. McGuire, C.B. and Radner, R. (eds) (1986), Decision and Organization, 2nd edition, Minneapolis, MN, University of Minnesota Press. Metcalfe, S. (1998), ‘Innovation as a policy problem: new perspectives and old on the division of labour in the innovation process’, mimeo, Centre for Research on Innovation, University of Manchester. Milgrom, P. and Roberts, J. (1988), ‘Economic theories of the firm: past, present, future’, Canadian Journal of Economics, 21, 444–58. Nelson, R. (1994), ‘The role of firm difference in an evolutionary theory of technical advance’, in Magnusson, L. (ed.), Evolutionary and Neo-Schumpeterian Approaches to Economics, Dordrecht, Kluwer. Nelson, R.R. and Winter, S.G. (1982), An Evolutionary Theory of Economic Change, Cambridge, MA, Harvard University Press. Nohria, N. and Ghoshal, S. (1997), The Differential Network: Organizing Multinational Corporations for Value Creation, San Francisco, CA, Jossey-Bass. Nonaka, I. (1994), ‘A dynamic theory of organizational knowledge creation’, Organization Science, 5 (1), 14–37. Nonaka, I. and Konno, N. (1998), ‘The concept of Ba: building for knowledge creation’, California Management Review, 40 (3), 40–54. Nonaka, I. and Takeuchi, H. (1995), The Knowledge Creating Company, New York, Oxford University Press. Nooteboom, B. (1999a), Inter-Firm Alliances, London, Routledge.
Epistemic communities and communities of practice
321
Nooteboom, B. (1999b), ‘The triangle: roles of the go-between’, mimeo, Faculty of Management and Organization, Groningen University. OECD (2000), Knowledge Management in the Learning Economy: Education and Skills, Paris: Organisation for Economic Cooperation and Development. Pavitt, K. (1990a), ‘What we know about the strategic management of technology?’, California Management Review, 32, 17–26. Pavitt, K. (1990b), ‘Some foundations for a theory of the large innovating firm’, mimeo, Brighton, SPRU, University of Sussex. Pavitt, K. (1997), ‘Technologies, products and organization in the innovative firm: what Adam Smith tells us and Joseph Schumpeter doesn’t’, mimeo, SPRU, September, presented at the conference, Competences, Governance and Entrepreneurship, DRUID, Bornhom, 9–11 June 1998. Penrose, E. (1959), The Theory of the Growth of the Firm, Oxford, Oxford University Press. Postrel, S. and Rumelt, R.P. (1992), ‘Incentives, routines and self-command’, Industrial and Corporate Change, 1 (3), 397–425. Prahalad, C.K. and Hamel, G. (1990), ‘The core competence of the corporation’, Harvard Business Review, 68 (May–June), 79–91. Probst, G., Büchel, B. and Raub, S. (1998), ‘Knowledge as a strategic resource’, in von Krogh, G., Roos, J. and Kleine, D. (eds), Knowing in Firms, London, Sage. Radner, R. (1986), ‘The internal economy of large firms’, Economic Journal, 96 (supplement), 1–22. Reynaud, B. (1996), ‘Types of rules, interpretation and collective dynamics: reflections on the introduction of a salary rule in a maintenance workshop’, Industrial and Corporate Change, 5 (3), 699–723. Richardson, G.B. (1960), Information and Investment, Oxford: Oxford University Press. Richardson, G.B. (1972), ‘The organization of industry’, Economic Journal, Sept., 883–96. Schelling, T. (1978), Micromotives and Macrobehavior, New York: W.W. Norton. Simon, H.A. (1957), Models of Man, New York, Wiley. Simon, H.A. (1976), Administrative Behaviour, 3rd edition, New York, Free Press. Simon, H.A. (1979), ‘Rational decision making in business organizations’, American Economic Review, 69, 493–513. Teece, D.J. (1988), ‘Technological change and the nature of the firm’, in Dosi, G., Freeman, C., Nelson, R., Silverberg, G. and Soete, L. (eds), Technical Change and Economic Theory, London, Pinter, pp. 256–81. Teece, D.J. and Pisano, G. (1994), ‘The dynamic capabilities of firms: an introduction’, Industrial and Corporate Change, 3, 537–56. Teece, D.J., Rumelt, R.P., Dosi, G. and Winter, S.G. (1994), ‘Understanding corporate coherence: theory and evidence’, Journal of Economic Behaviour and Organization, 23, 1–30. Tell, F. (1997), ‘Knowledge and justification – exploring the knowledge based firm’, thesis, Department of Management and Economics, University of Linköping. Vicari, S. and Toniolo, G. (1998), ‘Errors and learning in organizations’, in von Krogh, G., Roos, J. and Kleine, D. (eds), Knowing in Firms, London, Sage. Von Krogh, G., Ichijo, K. and Nonaka, I. (2000), Enabling Knowledge Creation, Oxford, Oxford University Press. Von Krogh, G. and Roos, J. (1995), Organizational Epistemology, Basingstoke, Macmillan.
322
The economics of knowledge
Wenger, E. (1998), Communities of Practice: Learning, Meaning, and Identity, Cambridge, Cambridge University Press. Williamson, O.E. (1967), ‘Hierarchical control and optimum firm size’, Journal of Political Economy, 75, 123–38. Williamson, O.E. (1975), Markets and Hierarchies: Analysis and Antitrust Implications, New York, Free Press. Williamson, O.E. (1985), The Economic Institutions of Capitalism: Firms, Markets, Relational Contracting, London, Macmillan. Williamson, O.E. (1993), ‘Transaction costs economics and organisation theory’, Industrial and Corporate Change, 2 (2), 107–56. Winter, S.G. (1982), ‘An essay on the theory of production’, in Hymans, H. (ed.), Economics and the World around It, Ann Arbor, MI, University of Michigan Press, pp. 55–93. Zuscovitch, E. (1998), ‘Networks specialization and trust’, in Cohendet, P., Llerena, P., Stahn, H. and Umbhauer, G. (eds), The Economics of Networks, Berlin, Springer Verlag, pp. 243–64.
12. Markets for technology: ‘panda’s thumbs’, ‘calypso policies’ and other institutional considerations Ashish Arora, Andrea Fosfuri and Alfonso Gambardella 1
INTRODUCTION
We have argued elsewhere that markets for technology have become important (see Arora and Gambardella, 1994; 1998; Arora et al., 2001b; 2001c; Arora and Fosfuri, 2003). Our research in this field is related to our intellectual legacy from Paul David. It was Paul who originally suggested to us that the ‘technology of technical change’ (an expression that Ashish and Alfonso heard from him when they were graduate students, and that Andrea heard from them) was an important phenomenon to understand. Moreover, Paul’s work on the nature of knowledge directed some of our work, especially when he emphasised that differences in the nature of knowledge can give rise to different opportunities of exchanging knowledge or technologies among parties that do not belong to the same organisation (for example, David, 1993a). But while one may think that our intellectual debt to Paul David is largely associated with the study of the nature and the role of technology, his emphasis on institutions and norms, and their role in affecting economic growth and technology, probably influenced our way of thinking even more deeply. In this chapter, we therefore touch upon several topics about norms and institutions at large that Paul addressed in several ways in his work. Specifically, we discuss how standards, intellectual property rights, institutions and the related social, cultural or political norms shape the rise and development of markets for technology. We draw inspiration from Paul’s work on the scientific community as an institution (for example, Dasgupta and David, 1987, 1994; David, 1993a; 1998; 2003) and his concerns about the threats to the norms of open science. As we shall discuss in this chapter, while markets for technology may on the one hand encourage greater diffusion of knowledge and technology, 323
324
The economics of knowledge
the implied ‘privatisation’ of knowledge has serious side effects for open science. To set this discussion within the specific context of the markets for technology, a natural starting point is to note that markets do not arise or function in a vacuum. They need a supporting infrastructure. This is evident in the current context of on-line markets and the fortunes of firms that provide the technical infrastructure for such markets. But no less important is the institutional and policy infrastructure. This includes not just the formal laws and policies that govern such markets, but also norms and ‘rules of the game’ that determine the transaction costs of participating in them. But this also suggests that rather than an ‘exhortative’ approach to policy analysis, whose objective is a list of recommendations for policy changes, one can more usefully rely on a more ‘institutionalist’ approach, which as we shall see, appears to be particularly suited for the context that we set forth in this chapter. As Douglass North (for example, 1990), among others, put it, such an approach focuses on understanding how adequate institutions for supporting economic growth are created, how they function, and the role of governments in creating, supporting and shaping these institutions. Several studies in the new institutionalist approach, especially those developed in a historical context, deal specifically with the genesis and development of new markets, and their effects on economic growth (for example, Rosenberg and Birdzell, 1986; Alston et al., 1996). In this chapter, we argue that the support that policy can provide to the growth of markets for technology is relatively more important when they have to be created, compared with supporting their functioning when they are in place.1 As those creating the new electronic virtual market places well understand, market creation may require explicit transfers between market participants that enter at different times. Put differently, creating markets often implies large externalities, and public policy can play an important role in subsidising those creating the positive externalities. For instance, the difficulties in establishing standards hinder the development of new markets. These standards need not be merely technical standards, but they can also pertain to contracts or the forms in which commercial information is gathered and recorded, and these are typically the domain of legal and policy infrastructures. Further, these standards need not be set by governments or other public bodies. As Rosenberg and Birdzell (1986) show, critical institutions for the expansion of markets and commerce in the fifteenth and sixteenth centuries in Europe, such as a bill of exchange, insurance or double entry bookkeeping, were produced independently by the economic agents. Even so, public
Markets for technology
325
policy can play an important role in encouraging compliance. For instance, as Bordo et al. (1999) describe, the flow of foreign capital to the USA and the development of an equity market were greatly facilitated by the adoption, by US firms, of English accounting practices and the regular reporting of financial results. State legislatures, particularly those in influential states such as New York, played an important role in the widespread adoption and diffusion of these accounting standards. In this chapter, we discuss the role of standards in section 2.2. Specialised technology suppliers can play an important role in the context of markets for technology. Some societies tend to provide greater opportunities for new firm formation and for the extent to which firms, particularly new firms, can enter new market niches. Specialised technology suppliers are a special case, albeit a very important special case, of the latter category. Needless to say, as we discuss in section 2.3, public policy can facilitate or discourage new technology-based startups and risk-taking activities more generally. To say this is not necessarily to advocate such policies. Whether such policies are appropriate and necessary is itself something that needs to be examined more closely on its own merits. The most obvious relevant institution for technology markets is intellectual property rights. As Paul David noted in several papers (for example, David, 1993a; 1993b), intellectual property rights are a social institution. Although much of the discussion has focused on their role in providing incentives for innovation, a market for technology perspective focuses attention on their role in facilitating transactions in technology. Simply put, intellectual property rights are almost a precondition for a market for technology to exist. That said, there are some subtleties about how these rights are defined and interpreted that have important implications for the functioning of these markets. For example, as noted by Heller and Eisenberg (1998), and as we discuss in section 3, fragmentation of intellectual property rights may hinder the efficient working of markets for technology. Excessive fragmentation (or excessive overlap) of intellectual property rights may prevent the development of innovations when complementary property rights are owned by independent agents and no one is capable of collecting or co-ordinating all the rights to develop the innovation. A related consequence of the increasing strength of intellectual property rights that we observe, for instance, in the USA today is an increase in litigation costs. Thus, mechanisms like stronger intellectual property rights, which, as we argued elsewhere (for example, Arora, 1995; Arora et al., 2001c), may help the formation of markets for technology by reducing some of their transaction costs, may create other transaction costs.2
326
The economics of knowledge
An institutional perspective also points to an important interaction between the growth of markets for technology and other institutions in society, most notably universities. Academic research has contributed significantly to the growth of new scientific and technological knowledge. Further, universities have played a crucial role in codifying and standardising the language and terminology used to communicate scientific findings and, more generally, in creating and sustaining scientific communities. A natural result has been the development of conceptual categories that are general and universal in scope, an important requirement for an effective market for technology. These activities of the research university system form an important ‘public good’ upon which all participants in a market for technology draw and for which they make no direct recompense. These public good creating activities of universities are greatly enhanced by norms of disclosure and collegiality that appear to have arisen in response to specific historical factors unrelated to their present function but which, as Dasgupta and David (1994) have noted, appear to be critical to the role research universities play in modern economies. But, as we discuss in section 4, the greater ‘privatisation’ of knowledge and the ability to directly value knowledge that markets for technology make possible often clash with these academic norms. In times of weakening public support, this may seriously attenuate these norms. In turn, this presents a serious challenge for public policy broadly construed. Finally, in a globalising world, markets for technology are also likely to be global. But the globalisation of the markets for technology may limit the scope and effect of national policies. It also means that a country may benefit from the rise of markets for technology elsewhere, even though little effort was made domestically to create them. The direct implications of this argument for policy are twofold. First, in the European context, policies for encouraging markets for technology are best considered at the level of the European Union rather than the individual nation-state. Second, particularly in developing countries, science and technology policies should be mindful of whether markets for technology exist and, if they do, their efficiency and, working. In turn, this argues for sector- or industry-specific policies, rather than a ‘one size fits all’ policy. Where markets for technology do exist, the policy questions may well narrow down to how best to take advantage of the ongoing growth in the worldwide technology trade. We discuss these issues in section 5 of this chapter. Section 6 concludes the chapter.
Markets for technology
2 2.1
327
TRANSACTION COST-REDUCING MECHANISMS IN THE MARKETS FOR TECHNOLOGY Institutions and the Rise of New Markets
Whenever new markets arise, and as long as they develop and grow, there will be some private incentive to create the institutions that support these markets. Thus, as Rosenberg and Birdzell (1986) note, the expansion of commerce, particularly international commerce, in the fifteenth and sixteenth centuries was accompanied by several institutional innovations. Some of these were made by the economic agents themselves (for example, insurance and bills of exchange), while others required supra-individual policy interventions (for example, taxation). In addition, there is a question of timing. Institutions are less ‘plastic’ than technology or industry structure. As a result, new technologies diffuse rapidly compared to the development of complementary organisational innovations (for example, David, 1990). Since the formation of complementary institutions takes time, these institutions are more likely to follow rather than anticipate the new markets. With these broad remarks in mind, the institutional settings that are required for the markets for technology to arise are in many respects similar to those that are required for any new market to arise. There are striking similarities between the institutional innovations that gave rise to the growth of new commercial markets in sixteenth-century commerce and those needed for the development of technology markets today. Just as the growth of sixteenth-century commerce required well-defined property rights, so also markets for technology require better defined intellectual property rights (further discussed in the next section). Similarly, markets require standards. In the special case of the markets for technology, technological standards reduce the market risks of new technological developments by ensuring the compatibility of specific technological components with existing technological architectures and complementary technologies. In other words, provided that their devices are consistent with a commonly known architecture, they only face the technological risk that their device may not function, that it might not be costeffective or that it might be inferior to competitors. Finally, new markets, and their expansion, are typically associated with greater economic experimentation and risk. Thus, we also enquire into the institutions that support economic or technological experimentation.
328
2.2
The economics of knowledge
Standards
The absence of standards can significantly increase transaction costs. The importance of standards for technology markets or the exchange of knowledge should not be confined to ‘hard’ technological standards. ‘Soft’ standards can also be crucial for enabling independent parties, often located far away from one another, to exchange knowledge or information. A striking example is the scientific community. The scientific community has developed for many centuries a common language, and a common set of norms and customs. We discuss the importance of such norms and customs for the diffusion of knowledge in section 5. Here we want to highlight a related effect. As David (1991; 1993a) and Dasgupta and David (1987; 1994) point out, the current norms and customs of the scientific community require that scientific discoveries should be reproducible by peers, and the results be cast in a ‘language’ that is commonly understood by the scientific community. The common language of science, in its various fields and disciplines, has acted as a natural standard for research. The growing use of mathematical and computer modelling has furthered standardisation. In many ways, the scientific system is a prototypical case of widespread effective exchange of knowledge among individuals or groups who have never had any direct contact with one another, but who exchange ‘codified’ knowledge through journal articles and build upon the work of others.3 Accordingly, one would expect that a greater diffusion of standardised R&D techniques, such as standardised software and simulation tools (for example, for testing products or for computerised product design) can promote methodological standards, and therefore create opportunities for exchange among independent parties (Arora and Gambardella, 1994). Of course, ‘hard’ technological standards are also critical for the growth of markets for technology. The argument is not new (for example, see David and Greenstein, 1990, for a survey), and Langlois and Robertson (1992), among others, have shown how the development of an ‘open architecture’, based on well-defined architectural interfaces and standards, was critical for giving rise to widespread innovation activities and experimentation with personal computer (PC) components by many independent suppliers. Established architectural interfaces meant that independent innovators did not face the risk of creating incompatible innovations, but only the risk of failing to develop and commercialise the innovation. Technological standards have become common in many markets for technology. One example is the development of a component-management software technology, called CORBA, produced by a US-based non-profit consortium – Component Management Group (CMG) – with 20 000
Markets for technology
329
members (individuals and companies) in the USA, Canada, Italy, Japan and India. CORBA provides standard interfaces for designing software components that can be plugged and played into the systems without interfering with the basic structure of the architecture, along with other standard setting operations and procedures. Moreover, the rise of such standard-setting institutions is complementary with the creation of open electronics markets for software components. Web-based markets exist today to help match buyers and sellers, thereby reducing search and related transaction costs. Standards for components and their reuse further reduce the transaction costs, and therefore enhance the value of such markets. Linden and Somaya (2003) have shown that institutional developments in the semiconductor industry have helped create standards that facilitate licensing and cross-licensing of design modules (see also Hall and Ziedonis, 2001). Linden and Somaya note that a critical event in the 1980s was the establishment of the silicon-based CMOS technology as the dominant design in semiconductor process technology. In the 1990s, when fabless design firms arose, they could take advantage of this standard processing technology by focusing on designing integrated circuits for users and rely upon merchant foundries for manufacturing. The existence of a standard process like CMOS reduced transaction costs and was critical for an independent fabless design industry to arise. Put differently, the existence of a standard manufacturing process implied that it was easier to de-link product design from process requirements and, therefore, to separate the skills, knowledge and activities that were needed to design the chips and those that were required to fabricate them. The separation of the knowledge and related domains for designing the product could be achieved to a greater extent by ensuring the compatibility of the interfaces between the design and the process. During the 1990s, the semiconductor industry gave rise to two major standard setting alliances. The first, called the Virtual Socket Interface Alliance (VSIA), was established in 1996 by 35 founding members, which included Electronic Design Automation (EDA) software firms, fabless semiconductor design companies and electronics companies. The goal of VSIA was to define and establish open compatibility standards (‘virtual sockets’) in semiconductor design. Virtual Socket Interface Alliance both releases the specifications, and it actively tries to encourage their use by the participating firms. Reusable Application-Specific Intellectual Property Developers (RAPID) aims at improving access to information about design modules. Thus, for instance, RAPID developed a standard catalogue for featuring commercially available design modules on the Internet. Similarly, the Virtual Component Exchange (VCX), was created in 1998 by the Scottish economic development agency, ‘Scottish Enterprise’, and a few
330
The economics of knowledge
major players from VSIA. VCX is addressing business and legal issues related to trade in design modules by developing standard contracts, monitoring systems, a matchmaking service and customised arbitration services (see Linden and Somaya, 2003). There are several noteworthy aspects of the foregoing. First, in the case of semiconductors, it is interesting that coalitions for establishing standards have not been confined to the setting of purely technological standards. Virtual Component Exchange, in particular, is addressing standardisation in areas such as contracts for design modules. Thus, standards related to the efficient functioning of various aspects of the market for technology, including norms about legal settings, have been central to the formation of such institutions. Second, some of these coalitions are private initiatives, with no evident government intervention. This suggests that direct policy intervention is not strictly necessary for addressing or solving co-ordination problems involved in standard setting. However, public agencies like the ‘Scottish Enterprise’ (for VCX) can help catalyse such initiatives. Thus, indirect policy interventions, possibly through such agencies, may be important to encourage the industry to coalesce in order to promote standards. At the same time, as noted earlier, VCX deals with the creation of standard contracts. This suggests that such decentralised institutions may even embrace areas that once might have been thought the preserve of government alone. Clearly, government intervention may be critical when private initiatives are in conflict with each other, or absent. But rather than directly creating such standards, it is probably more effective if governments encouraged ‘private’ coalitions for standard setting, which would be closer to the actual needs and information of the specific industry producers and stakeholders, and therefore more likely to develop valid and competent solutions. Governments may still have important roles to play, and in this respect there may be different remarks to be made for governments with different policy traditions, like the USA and Europe. For example, in the USA, where antitrust is more stringent, antitrust considerations could be traded off against the need to promote effective standards and prevent fragmentation of intellectual property rights, as for instance in the case of patent pools described later in this chapter. This also requires that such alliances are monitored closely to verify, for example, whether industry prices increase after the alliance. By contrast, in Europe, where inter-firm alliances have attracted less antitrust scrutiny, greater attention should be paid to these coalitions to make sure that they actually act as institutions for the creation of standards rather than as means for price-fixing. At the same time, the substantial barriers to information exchange, mobility and interaction among companies
Markets for technology
331
and other institutions located in different European countries suggest that the European Commission should take a more active role in favouring the formation of such coalitions. 2.3
Financial and Other Institutions
Specialised technology suppliers play a key role in the market for technology. Although firms operating in standard product markets can, and do, supply technology, there are some obvious ways in which specialised technology suppliers are important. For one, unlike technology producers that are also suppliers in the final markets, technology suppliers that specialise only in the production of the technology do not face the inherent conflict of having to compete with their customers. On the one hand, this encourages them to supply their ‘best’ technology, because of the natural market incentives to provide the highest possible value to the market, as opposed to second-rate technologies to avoid that competitors catch up with them. On the other hand, the buyers themselves feel more comfortable when buying technologies from firms that have no vested interests in the final markets. Both factors reduce problems of asymmetric information and other transaction costs or sources of opportunism. A related reason is that the independent technology suppliers are not locked into an installed base of products and need not fear cannibalising their existing market by developing new technology. Finally, their corporate culture is likely to be more flexible and open to communication, and their management less likely to be distracted by the needs of manufacturing and distribution. For such specialised firms, many of which are startups and small, there are many other types of barriers and transaction costs to contend with. Obtaining finance is a particularly important one. Financial institutions can play a critical role in fostering or hindering the markets for technology. A general discussion of the role of financial institutions and policy is beyond the scope of the present discussion. Instead, we will focus upon policy initiatives for promoting risk-taking, to encourage specialised technology suppliers and startups. Venture capital, initial public offerings and ‘new’ financial and equity markets have grown in parallel with the rise of new business opportunities and innovation in high-tech industries. Institutional innovations like venture capital have proved to be extremely flexible. They have adapted in various ways to the actual needs and conditions for supporting new business activities. For example, many analysts have noted that not only do venture capitalists provide finance and managerial support, but, most important of all, they provide their startups with connections to their broader networks of people and resources (see, for example, Gomperts,
332
The economics of knowledge
1999). This networking capability has been critical for maximising the exploitation of the external economies that exist in areas like California or elsewhere (see also Bresnahan and Gambardella, 2004). To a large extent, these institutions are themselves the result of private responses to profitmaking opportunities. In successful cases, policy has mainly played the role of creating the general ‘ambience’, rather than direct intervention. Even with venture capital, technology-based startups face a high degree of risk. Thus, in many countries institutions for supporting technological risk have taken the form of direct policy interventions as well. For example, many European governments, often with the financial and political support of the European Union, have invested sizeable public resources in the creation of science and technology parks, especially in less developed regions. The stated objective is to provide the physical and business infrastructure needed to nurture startup firms, particularly, R&D-intensive firms. These initiatives have had mixed success. Rather than enter into a discussion of the pros and cons of such initiatives, a more interesting approach is to note that other, more indirect ways, of reducing risks for technology-based new business exist. Compared with direct measures, indirect measures have the advantage that the benefits from the measure accrue primarily to those who have attained some independent achievement, rather than indiscriminately to everybody. A report by the European Commission (ETAN, 1999) suggests three areas for institutional developments that would encourage risk-taking behaviour – changes in fiscal policy, the creation of security interests in intellectual property rights and changes in insolvency laws. As far as the fiscal incentives are concerned, R&D and innovation tax credit can be useful institutional innovations to favour startup companies and the markets for technology. These objectives, however, have to be taken explicitly into account in designing the schemes. For example, the US Research and Experimentation Tax Credit issued in the early 1980s was amended in 1993 to properly extend the underlying incentives to smaller firms. Small firms, new startups or, more generally, firms that could not claim a basis of R&D expenditures in the previous three years on which to compute the incremental R&D tax credit, were assigned a fixed percentage increase of 3 per cent for the first five taxable years beginning after 1993. Another problem with startups is that they may not be able to enjoy the benefits of the R&D credit because of having no taxable income in the year. In order to make an unused R&D credit a valuable asset, the 1993 Amendment established that firms could carry back the credit three years and carry it forward up to 15 years. In so doing, the credit becomes a hidden asset that can be unlocked in the future when the company becomes profitable or is sold. Venture capitalists and lenders understand the importance
Markets for technology
333
of these hidden assets and may grant more favourable terms if they know a credit exists and can be deployed in the near future. There is a more general perspective to this point. A well-known problem with technological activities is how to account for the value of these activities – and of the technology-based companies more generally – given that such a value is often made up of intangible rather than tangible assets. The valuation of intangible assets, and specifically the valuation of technology, is particularly relevant in cases where the firm lacks downstream assets to commercialise the technology. This is a complex issue and well beyond the scope of the present discussion. The point is that accounting practices and norms, decided through a complicated interaction between private and public bodies, can affect the fortunes of technology-based firms, particularly startups, in important ways. Current accounting practices and norms, derived as they are from times when measuring tangible and material assets was their crucial task, will have to be modified in order for technology markets to flourish. This point is not new, and other writers have emphasised this as well (Lev, 2000; Hand and Lev, 2003). There has also been some interesting research using firm-level R&D and patent and patent citation data to develop implicit measures of the technological assets of firms (Deng et al., 2003). What is less well understood is the role that technology markets themselves can play in improving the accounting for intangible technological assets. A market for technology improves the accuracy of any valuation attempt. It does so in the most obvious way, by providing an objective measure of the value, if the asset has been traded in the past or if similar assets have been traded. Needless to say, technology is highly differentiated, and its ‘price’ is likely to reflect factors idiosyncratic to the buyer and the seller. Thus, any monetary measure is likely to be imperfect. That said, such problems are not unique to the measurement of the value of technology. Further, when investing in R&D, firms are implicitly making such measurements, as do investors when they value the firms in capital markets. Markets for technology allow for the possibility of valuing the contribution of technology separately from the value of other valuable assets the firm may possess. In turn, such valuation may enable firms to specialise in the developing technology without necessarily having to acquire downstream capabilities. This problem is also closely related to attempts in the USA and Europe to remove legal obstacles to the creation of security interests in intellectual property rights, as discussed in the ETAN report (1999: 46). The point is that once lenders, investors or the entrepreneurs themselves can meaningfully assess the value of these assets, the assets can be used in a variety of ways, including as collateral to obtain financing. When such assets can be
334
The economics of knowledge
‘securitised’ and traded in a market, this is likely to further encourage the growth of firms specialising in developing technology.4 The third area identified by the ETAN report (1999) is that of the insolvency laws. In fact, the introduction of limited liability in the sixteenth century is a major institution for limiting the risk of the entrepreneur who sets up a new business. However, as the report notes, the limited liability can be in many practical cases more apparent than real. For example, on many occasions, the life savings and dwelling house of the entrepreneurs have been used as collateral for company debts. The issue is at least as serious in the case of technological business where companies may take even higher risks. This could limit, for instance, their ability to invest in expensive equipment to carry out research experiments, computerised product designs and the like. The ETAN report argues that one ought to encourage institutional changes that would further reduce liability for failure in technologically risky industries. The report also argues that in the USA the Federal Bankruptcy Code Chapter 11 is more favourable to the setting up of new companies for marketing technological innovations than existing laws in most European Union member countries. This is clearly a quite complicated area for intervention, as one has to properly balance the need for encouraging risky business against the need to discourage excessive or imprudent risks. But this only increases the importance of carefully crafted policies. As with any other market, intermediating institutions for reducing information search costs will improve the working of technology markets. However, the reduction of information search costs in the market for technology may often be carried out by agents who would either specialise in acquiring and diffusing the information, or who would do so jointly with other supporting activities for these markets. For example, patent attorneys and patent agents did play such a role in the development of the market for technology in the USA during the nineteenth century and early twentieth century (see Lamoreaux and Sokoloff, 1996; 1999). In addition to providing patent counselling and related services, they also helped match demand and supply. A similar role is played today by several independent firms and technology traders on the Internet (see for instance BTG, 1998). Public institutions can step in where private initiative is lacking. Thus, for instance, the European Union has created its own Internet information providing service, CORDIS, which collects information about potential technologies for licensing, as well as requests for technological partnerships and the like. Similarly, the German government, along with the Länder and private investors, has created the Steinbeis Foundation – a large network of German firms, research institutions and academic professors, whose task is to co-ordinate and match technology demands with the supply of technologies and related competencies.
Markets for technology
3
3.1
335
INTELLECTUAL PROPERTY RIGHTS AND THE MARKET FOR TECHNOLOGY IN A NON-COASIAN WORLD Fragmentation of Intellectual Property Rights and Related Issues
It is difficult to think that a market, or at least what economists mean by employing this word, could ever function properly without property rights on the object of the transaction. Markets for technology are no exception. Indeed, intellectual property rights are to the markets for technology what property rights are to the markets for cars or for PCs. They recognise enforceable claims to some benefits or use to the technology. Intellectual property rights consist of patent rights, copyrights, design rights, trademark rights, trade secret rights and a few other special property rights in contemporary law. Our focus here is on patents. Patents are the most important type of property right for technology, although as we have repeatedly pointed out, technology is much more than what is covered by patents. However, selling a technology without an enforceable patent would be difficult, although not impossible. The moment one has to disclose a piece of information in order to sell it, one is running the risk of being cheated. Potential buyers and other third parties can therefore appropriate these ideas and knowledge without having to pay for them. Anticipating this, potential sellers may be reluctant to show the object of the transaction and let it be evaluated by the prospective buyers. The latter will not pay money for something whose value they cannot appraise up front. The net result is that such transactions may not take place at all.5 Therefore, patents, or, more generally, well-defined enforceable intellectual property rights, are not only critical to protect the incentives for innovation but are also the supporting institutions for the existence and functioning of markets for technology. Arora and Merges (2004) use the incomplete contracting approach (Grossman and Hart, 1986; Hart and Moore, 1990) to argue that welldefined enforceable patents reduce transaction costs, and thereby help increase transactions in technology. Efficient contracting for technology enhances the opportunity to profit from innovation through licensing. Patents can be used to structure technology transfer contracts, thereby playing an important role in determining the efficiency of knowledge flows. Arora and Merges (2004) also find support for the argument put forward in Arora and Gambardella (1994) that patents are likely to have a greater value for small firms and independent technology suppliers as compared to large established corporations. Whereas the latter have several means to protect their innovations – for instance, through their
336
The economics of knowledge
extensive downstream manufacturing and commercialisation assets – the former can only appropriate the rents to their innovation by leveraging the protection that patents provide.6 The role of patents in facilitating transactions in technology has largely been ignored in formal economic analysis. The focus there has been on the trade-off between the ex ante incentives to innovate and the ex post advantages of innovation diffusion, with the limitation to ex post diffusion being the price that society has to pay in order to encourage market-based innovation ex ante. The major policy question related to the optimum length (and, later, length and breadth) of the temporary monopoly to be granted (see, for instance, Gilbert and Shapiro, 1990; Klemperer, 1990). One feature common to both views of patents – as inducement for innovation and as the basis for markets for technology – is that they treat patents as covering inventions with well-defined and clearly de-limited scope of application. In fields such as chemicals, biology, materials and electronics, the growth in our understanding of the underlying physical phenomena makes it possible to represent the invention succinctly and effectively, through the use of abstract generalisation that a scientific approach makes possible. However, the very same growth in scientific understanding, and the growing power and use of the abstraction that this understanding makes possible, also makes it possible to relate knowledge created in a specific context to a much broader array of applications. This growth in generality, which has spurred the growth of many sciencebased startups, especially in biotechnology, has also created several challenges for the patent system and for the role of intellectual property rights. For the first of the challenges, consider the example raised by patenting of parts of the human genome. This is a controversial and emotive issue but our focus here is somewhat different. Understanding the structure of a gene provides information about the proteins it codes for. If one also understands the role of the protein in the context of some disease, then understanding the structure of the gene provides an opportunity to try to prevent or cure the disease. A patent on the gene would therefore allow the patentholder to share from the economic rents created by this therapy. However, these rents would have to be shared with the firm that uncovers the role of the protein coded for by the gene as well as with the firm that uses that knowledge to develop a cure, to test the cure in clinical trials and to manufacture, market and distribute it. This raises the question of how the different contributors should be rewarded. One might expect that the relative bargaining power of the parties involved would determine the rewards. In principle, the situation is not very different from a landowner bargaining with a real estate developer
Markets for technology
337
who will put up a shopping mall on the land. In simple terms, one would expect higher transaction costs as the various parties to the negotiation try to get the best deal for themselves. The major difference is that the knowledge of the structure of the gene (and the working of the protein, for that matter) is a non-rival good in that it may be applied in other contexts without reducing the economic value derived from its application in the first context. In other words, there is a strong ‘public goods’ character.7 Indeed, applying the knowledge about the structure of the gene to cure one disease in no way reduces the value of applying the same knowledge to cure other diseases. In this sense, knowledge is non-rival. As is evident upon further thought, the key here is that the knowledge has multiple potential applications, so that users do not compete. When knowledge is non-rival, protecting that knowledge through patents creates potential inefficiencies. For instance, in the case of ex ante contracting, a number of different potential users may have to get together to invest in creating the knowledge. Such contracts are problematic because users will differ in the value they place upon the enterprise and, consequently, are likely to under-report their value. Similar problems are likely with ex post contracting, with different users being charged different prices. Moreover, the closer a patent comes to covering knowledge that amounts to a basic understanding of the physical phenomena involved, the broader the likely sweep of the patent and the further in time its applications. In other words, the implicit one-to-one relationship between a patent and an innovation that many economic models assume, although analytically convenient, has obscured the point that in cumulative or systemic technologies, a commercialisable innovation may require many different pieces of knowledge, some of which may be patented and owned by people with conflicting interests.8 In turn, an agent holding a patent on an important component may cause severe ‘hold-up’ problems, retarding the development of the technology (see also Scotchmer, 1991, and Green and Scotchmer, 1995, for further discussion.) In a similar vein, Merges and Nelson (1990; 1994) argue that broad patents increase the likelihood that an innovator would try to control future innovations based upon its own innovation, thereby slowing down the pace of technological progress. However, the essential problem is not caused by patents, but by factors (such as negotiation costs) that prevent agents from entering into contracts for the use of patents. In a Coasian world with no transaction costs, agents will bargain for a Pareto superior solution given any initial distribution of property rights over the fragments. More realistically, the required collection of property rights, although socially efficient, might not occur because of transaction costs and ‘hold-up’ problems. An agent holding a patent on an important fragment (‘blocking patent’) may use his or her patent as a
338
The economics of knowledge
‘hold-up’ right in an attempt to extract as much of the value of his or her innovation as possible. Thus the issue at stake is the impact that strengthening and expansion of patent rights – which is what is happening today, particularly in the USA – would have on transaction costs. An especially problematic case is when the property rights are defined around very narrow fragments of knowledge and owned by separate entities. In this case, each patent-holder has the right to exclude the others from the use of his or her piece of knowledge. In other words, when several pieces of intellectual property rights have to be combined, the transaction costs implied could be so high as to prevent some otherwise productive combinations. This problem has been studied in a broader context as the ‘anti-commons’ problem (Heller, 1998; Heller and Eisenberg, 1998). To fix ideas, suppose that the development of a new technology involves the use of N fragments invented and patented by separate firms. In addition, the technology innovator has to pay ex ante a fixed cost, I, which might be thought of as expenditures in R&D. In order to assemble the new technology either the innovator has to buy licences on the fragments or alternatively he or she has to invent around them. The cost of inventing around depends, among other things, on the strength on intellectual property rights defined around the single fragments.9 By definition, a ‘blocking patent’ implies that such cost is extremely large. Let us assume that in the process of collection of the rights, the parties agree to sign a licensing contract which stipulates an up-front fee negotiated through bilateral bargaining. There are two straightforward results that one can derive. First, the higher the cost of ‘inventing-around’ the fragments, the weaker the bargaining power of the innovator in the licensing negotiations to collect the rights for the use of the different fragments. This is simply because the innovator’s outside option – ‘inventing around’ the patents on the fragments – becomes less attractive. Second, the larger the number of fragments, the higher the number of contracts that should be signed to guarantee the use of the innovation. If transaction costs are increasing with the number of transactions, a larger N is likely to increase the total transaction costs to assemble the fragments. A more interesting and less straightforward result emerges when one considers opportunistic behaviour of firms holding ‘blocking patents’ on the fragments. Indeed, the further the innovator goes in the collection of the rights for use of the fragments, the more resources are subject to an irreversible commitment and, therefore, the weaker is his or her bargaining power in future licensing negotiations for the collection of the remaining fragments. This implies that in subsequent negotiations he or she will have
Markets for technology
339
little chance to recoup all costs sunk up to that moment, that is, the fixed investment, I, and the fees paid for the rights on the fragments already bought. Furthermore, firms might also try to delay selling their blocking claims in order to hold out for more ‘quasi-rents’ that become available. Put differently, since the last firm that is going to negotiate its ‘blocking patent’ has the strongest bargaining power vis-à-vis the innovator and can capture the largest amount of rents, all firms have incentives to be the last. This is likely to introduce some further inefficiencies and delays in assembling the technology. Obviously, all these ‘hold-up’ problems can only be exacerbated when the number of ‘blocking patents’ on separate fragments increases.10 So far we have analysed a scenario with basically no uncertainty. Technologies, by definition, are characterised by a high level of uncertainty. This can only worsen the picture. First, it can sometimes be very difficult to know a priori N. In other words it is a hard task to determine which domains of technology have a legitimate bearing on the commercial product and who all the relevant intellectual property rights holders are. Ensuring access to all potentially blocking rights can therefore become extremely cumbersome. Second, when the market value of the innovation is uncertain firms might agree to sign royalty-based payment for the use of the fragments. In principle, these offer the advantage to the innovator of delaying the payments till the moment profits from the innovation start to materialise and to the patent-holders they confer the chance at larger pay-offs from sales of downstream products rather than certain, but smaller, up-front fees. However, the presence of such royalty-stacking negotiated on individual basis might imply that the total amount of royalties per unit of output is inefficiently high both from a private and social point of view.11 Third, as Langlois (2002) pointed out, in environments marked by Knightian uncertainty, transaction costs include not only the problems of hold up, bargaining and imperfect contracts. Rather, they include the problem that Langlois calls dynamic transaction costs. For instance, Langlois argues that Henry Ford’s consolidation of all production steps in a vertically integrated company was critical to the successful introduction of the Model T in the 1920. Had the various stages of production remained under separate ownership, Ford would have had difficulty experimenting with new techniques, machines and parts, all of which had to fit with each other. In other words, till the overall architecture of the product, in this case, the Model T, was settled, the costs of co-ordinating the actions of independent parts suppliers and machine-makers would have been very high. Similarly, when an innovation is based on combining the intellectual property of several independent agents, the costs of persuading them to ‘rent’ or part with their property for a particular application will be high,
340
The economics of knowledge
because each person’s pay-off will be contingent on all relevant parties coming on board. The net results might be that the parties involved are unable to reach an agreement. One example of what might occur when several companies hold patents on different components is provided by the early development of the radio (Merges and Nelson, 1990). The Marconi Wireless and Telegraph Company, AT&T, General Electric and Westinghouse all held important patent positions in the early stages of the development of the industry. The ensuing fragmentation of property rights is said to have caused serious delays in the pace of technological innovation. For instance, the basic patent on the diode was granted to Marconi, while the patent on the triode vacuum tube was assigned to AT&T. Marconi’s patent was needed for using the triode technology, yet neither party would license the other and, as a consequence, no one used the revolutionary triode for some time. Software, semiconductors and computers are other good examples of industries where the nature of innovation is systemic and cumulative, and where the intellectual property is very fragmented. In these industries the opportunities for hold up are enormous. Indeed, as reported in Grindley and Teece (1997) and Hall and Ziedonis (2001), this has led industry actors to sign cross-licensing agreements covering whole portfolios of patents related to an entire technical field (including both existing and future patents). These concerns have been echoed by industry participants as well. Cecil Quillen, former Senior Vice President and General Counsel of the Eastman Kodak Company, claims that since the early 1980s, the legal costs of intellectual property protection has risen dramatically to the point of substantially raising the cost of innovation itself. Michael Rostoker, former head of LSI Logic, a semiconductor manufacturer, has also suggested that, due to stronger patent protection, firms holding old technology have been in a position to command licensing fees from a current generation of innovators even while the original patent-holders have long ceased advancing the state of the art, leading to a stacking of licensing fees that impede the development of new generations of chips (Hadley, 1998). Similar situations arose in the early stages of development of the automobile and aircraft industry, and in the chemical process technology industry. Biomedical research may provide another possible example. A particular concern raised by Heller and Eisenberg (1998) and the National Research Council (1997, ch. 5) was the prospect that, by potentially increasing the number of patent rights corresponding to a single gene, patents on expressed sequence tags would proliferate the number of claimants to prospective drugs and increase the likelihood of bargaining breakdowns. Although plausible, the available evidence, limited as it is, suggests that anticipated problems have not yet materialised. For instance, based on
Markets for technology
341
about 70 interviews, Walsh et al. (2003b) report that although there has been an increase in the 1990s in patents on the inputs to drug discovery, drug discovery has not been substantially impeded by these changes. They report some evidence of delays associated with negotiating access to patented research tools. Further, there are instances where patents over targets limit access, and where access to foundational discoveries can be restricted. There are also cases where research is redirected to areas with more intellectual property freedom. Still, the vast majority of respondents say that there are no cases where valuable research projects were stopped due to intellectual property (IP) problems. One does not observe as much breakdown or even restricted access to research tools as one might expect because firms have been able to develop ‘working solutions’ including licensing, inventing around and legal challenges. Importantly however, institutional responses, particularly new PTO guidelines, active intervention by NIH and some shift in courts’ views towards research tool patents, appear to have further reduced the threat of breakdown and access restrictions.12 Interestingly enough, the institutional response had important private components as well (Walsh et al., 2003b). For example, public databases (for example, GenBank, or the Blueprint Worldwide Inc. venture to create a public ‘proteomics’ database ) and quasi-public databases (such as the Merck Gene Index and the SNPs Consortium) have been created, with substantial public, private and foundation support. Merck has sponsored an $8 million programme to create 150 patent-free transgenic mice to be made available to the research community at cost, without patent or use restrictions.13 These initiatives represent a partial return to the time before the genomics revolution, when publicly funded university researchers produced a body of publicly available knowledge that was then used by pharmaceutical firms to help guide their search for drug candidates.14 Scientific journals have also pushed for access to research materials, and biology journals require that authors deposit sequences in public databases such as GenBank or Protein Data Bank (Walsh and Bayma, 1996). Similarly, when Celera published their human genome map findings, Science’s editors were able to gain for academics largely unrestricted access to Celera’s proprietary database. 3.2
Policy Responses
This leads to some straightforward implications for patent offices as well. The main purpose of the Patent Office is to issue patents that the legislation permits or deems desirable. The Patent Office evaluates whether the claims are enabled under the terms of the statute or the patent has the
342
The economics of knowledge
statutory novelty in light of the prior art. The role of the Patent Office is to assure that only well-specified applications receive a patent. The better the definition of the claims, the less uncertainty about their scope and validity. This translates into lower transaction costs for technology trade and hence better functioning of markets for technology. Moreover, owing to the growing importance of patents, the social cost of ‘bad’ patents has increased, along with the number of patents themselves, as Kortum and Lerner (1999) have documented. This argues for more resources to be made available to patent offices for examining patents and, in particular, for searching for prior art.15 One may conjecture that larger patent offices and stronger incentives to patent might increase patent disputes and litigation, creating deadweight loss. The spurt in litigation activity that we have witnessed in recent years is not a consequence of a greater number of granted patents alone. Indeed, it also points to the likelihood that patent offices, particularly in the USA, have issued poorly defined patents, with overly broad scope or of dubious ‘non-obviousness’ and novelty over prior art. Often, broad and imprecise patents are issued because patent offices are under-funded, and the patent examiners inadequately trained and lacking the necessary capabilities to search for the prior art. In software, for instance, the US Patent Office has issued what are widely seen as overly broad patents, in large measure because the examiners rely very heavily upon previous patent applications to discover prior art. Since software patents are relatively new (copyrights having been the typical way of protecting software until recently), the result is bad and socially harmful patents, which nonetheless carry with them the presumption of validity. By the same token, patent offices should pay more attention to patenting requirements. Specifically, in the USA, the patentee is required to ‘reduce to practice’ the invention, show that it ‘possesses’ the invention, demonstrate the best known way the invention is to be used or ‘enabled’ and show the usefulness or ‘utility’. In the late 1990s, there was a perception that these requirements were not been enforced very seriously, at least in certain well-known cases. For instance, some early patents on gene fragments (ESTs) were issued without any clear knowledge of what proteins the gene fragment was coded for and what functions the proteins performed. In principle, these fragments could prove to be useful in a broad spectrum of applications, as yet unknown. If granted, the patent-holder might be able to demand a large share of the rents from any such applications or even block such applications, without having contributed to their discovery. Thus, the fear was that such patents could perversely block genomic research and, particularly, the commercialisation of such research.
Markets for technology
343
As indicated, these fears have, for the most part, not been realised. It appears that the institutional response to the concerns about the possibility of intellectual property rights impeding research and its commercialisation has been very important. In January 2001 the US Patent Office issued guidelines clarifying that a clear and specific utility would have to be indicated in applications for ESTs. Similarly, the US Patent Office announced that it would undertake a more careful examination of prior art for business method patents involving the Internet (www.uspto.gov, posted April 2000). In addition, in a series of judgments, the Court of Appeals of the Federal Circuit (CAFC) in the US (the so called ‘patent court’) narrowed the scope of many patents and limited the ability of patent-holders on upstream discoveries to block downstream development.16 In addition, there is room for policy-oriented interventions to facilitate the functioning of markets for technology even under very fragmented intellectual property. One possibility to modify the traditional stance of antitrust authorities on patent-pooling agreements. A patent-pooling agreement involves typically two or more companies with similar or overlapping patents. Rather than pursuing interference proceedings, or engaging in long and costly litigation to determine issues such as patent validity or infringement, they put their collective efforts to more productive use. For example, they may form a separate entity to which they assign or license their patents. The entity collects money for the service or product and pays out a royalty to each of the patent owners, according to the terms of the agreement. A similar argument applies to cross-licensing agreements where firms agree to license each other the use of their respective fragments.17 Traditionally, antitrust authorities in the USA have aggressively scrutinised patent pools and cross-licensing agreements, because such kind of agreements were sometimes used for restricting entry, controlling prices and market shares. However, recently the antitrust stance appears to have changed somewhat, favouring the emergence of market-based responses to the problem of excessive fragmentation of intellectual property rights. For instance, Grindley and Teece (1997) attribute the extensive use of crosslicensing agreements in electronics and semiconductors, where innovations are typically based on hundreds of different existing patents, to the large transaction costs required to bundle together patent portfolios. Recently, the Department of Justice has given the green light to a group of nine companies and one university to create a pool of patents that are essential to the MPEG-2 video standard. Another patent pool involving 11 patentholders has been agreed for the IEEE 1394 bus, a popular solution for transferring audio and video data.18 In the chemical process industry, technology-sharing agreements have a long history and were established to alleviate the transaction costs involved
344
The economics of knowledge
in market relationships. The case of the chemical process industry is interesting for another reason as well. The specialised engineering firms (SEFs) mentioned earlier, which supplied chemical process engineering and design, have sometimes acted as technology integrators which have helped in getting around the ‘hold-up’ problem of fragmented property rights. Thus, another potential benefit of specialised technology suppliers is that they can act as technology integrators limiting the hold-up problem created by the fragmentation of intellectual property rights. A final set of, more controversial, policies whose merits remain under debate is the extension of ‘eminent domain’ (that is, the legal doctrine that allows the government to take over private property for public purpose) to intellectual property. In principle, the threat that the government may step in and buy out a patent-holder at a ‘fair’ price can be a powerful deterrent to the sort of opportunism that underlies the fragmentation problem. But governments may not be the best agencies to take over a technology where public good considerations might be quite indirect. Determining the price for the patent is an important challenge. Kremer (1998) suggests using an auction as mechanism to determine the private value of patents. The government would use this price to buy out the patents and place them in the public domain. Alternatively, the law may simply allow for ‘efficient breach’ – that is, let people ‘infringe’ the patent and leave the courts decide about a ‘fair’ royalty. The latter is very similar in spirit to the compulsory licensing provisions and provisions that require the patent to be ‘worked’. Both these provisions have been present in many countries, especially in the past, and require courts to intervene more aggressively than is probably desirable.
4
THE PRIVATISATION OF KNOWLEDGE: CRUMBLING ACADEMIC NORMS?
The growth of markets for technology and the concomitant strengthening of intellectual property rights raises another fundamental challenge, this time at an institutional level. As basic scientific knowledge, such as the structure of genes, becomes eligible for patenting, universities are both pulled and pushed into entering the market for technology. In the USA, the decline of the cold war reduced government funding for research and, hence, an important source of revenues for research universities. David (1997) describes the trend in policy discourse towards a more instrumentalist view of university research, moving from national defence to national competitiveness, to wealth creation in the most recent stage. These trends have contributed to an increasing pressure upon universities and
Markets for technology
345
university researchers to demonstrate the economic value of their work. Concomitantly, universities have increasingly resorted to patenting and copyrighting their research. As Mowery et al. (2001) suggest, this is partly (but not only) a consequence of the passage of the Bayh–Dole Act in the USA in 1980, which allowed universities and government laboratories to claim patents on federally funded research. Other factors have encouraged the trend towards greater privatisation of university-based research. A general expansion of patentable subject matter and a more patent-friendly legal environment (with the establishment of the so called ‘patent court’, the Court of Appeals of the Federal Circuit) in 1981 meant that the patents that universities could obtain would be more valuable.19 The result was a great expansion in university patenting, as well as the establishment of university licensing offices and encouragement of ‘spin offs’ to commercialise university-based research. In short, universities are beginning to resemble, albeit only partially, both a source of technology as well as incubators for developing independent suppliers of technology. The number of patents issued to the US universities increased significantly between 1980 and the late 1990s. The share of patents assigned to US universities grew from less than 1 per cent of all the patents assigned to US inventors in 1975 to almost 4 per cent in 1997 (Mowery et al., 2001). Since US patents have grown rapidly during this period as well, it implies that university patents have grown faster still. Moreover, Mowery et al. report that while the patents per $1 billion R&D spending (in constant terms) declined from 780 in 1975 to 429 in 1990, university patenting has shown an increase in this ratio from 57 to 96 over the same period. These increases reflect more systematic attempts by universities to assert rights over inventions, including attempts by university licensing offices to elicit the disclosure of such inventions. Alongside this, US universities stepped up their efforts to license their patents. Mowery et al. report that licensing revenues of US universities (those that are members of the Association of University Technology Managers) have increased, in real terms, from $222 million in 1991 to $698 million in 1997. This is a notable increase, which is confirmed by their case studies of three leading US universities – Columbia, the UC system and Stanford. These trends raise three concerns. First, to some there seems to be something wrong with the notion of publicly subsidising research, whose results will later be monopolised. Put differently, even if one accepts that the temporary patent-based monopoly is a necessary evil to provide incentives for investments in research, then surely research that is publicly funded does not require the lure of patents. Patents are both unfair and inefficient. This line of reasoning, however, ignores the other role that patents can play,
346
The economics of knowledge
namely, in encouraging the efficient transfer of knowledge between the inventors and the commercialisers. It has been argued that without such patenting, much of the research now being commercialised would lie fallow and unused. In so far as this is important, one benefit of university patenting is that university researchers can effectively benefit from the invention by licensing the technology and the know-how, instead of attempting to commercialise the innovation themselves. If the latter were to happen, the university would lose the services of, potentially, very valuable researchers and teachers.20 Patents affect the commercialisation of university technology through another route as well. They provide incentives, not for invention, but for development. Indeed, at the heart of the Bayh–Dole Act was the belief that without some measure of exclusivity, companies would not invest in developing university research. University licensing also seem to rely heavily on exclusive licensing contracts. For instance, Mowery et al. (2001) report that for the period 1986–90, the fraction of licensed disclosures that were through exclusive contracts was 58.8 per cent, 59.1 per cent and 90.3 per cent for Stanford, Columbia and University of California, respectively. They seem to worry about the rather heavy use of exclusive contracting, arguing that nonexclusive licensing would balance the need for exclusivity with the public interest of broad dissemination of knowledge. The heavy use of exclusive licences should raise preoccupations, particularly if they cover technology with broad applicability. The second concern relates to the nature of university research. For instance, might the push to earn revenues from research increase emphasis on more applied research at the cost of basic, or fundamental, research?21 Paradoxically enough, the analysis of patent citations provides one source of empirical evidence. Henderson et al. (1998) found that the rise of university patenting after 1980 was associated with a decline in the ‘importance’ and ‘generality’ of university patents. Since patents relating to more fundamental and broad-ranging discoveries were likely to score higher on the proxies used to measure importance and generality, this evidence seems to support the concern about a move towards more short-term and applied research at universities. However, a more recent study (Mowery and Ziedonis, 2002), which controls more carefully for the changing composition of universities that patent, reaches a different conclusion. They analyse citations to patents held by Stanford and the University of California and find that, relative to a control group of similar patents by industry, there has been no decline in the importance and generality of patents from these universities. Using data on all university patents, they find that the decline in importance and generality is largely as a result of the increased share of university patents
Markets for technology
347
due to universities that are new to patenting. Thus Mowery and Ziedonis conclude that the evidence does not support a major shift in the content or culture of university research. As before, we lack systematic evidence to come to a definite conclusion on this issue. However, a shift towards more applied research and a neglect of more basic and fundamental research is certainly a possibility, and universities and governments have to be watchful. The third, and perhaps most important, concern raised by the growing involvement of universities and university faculty and researchers in patenting and commercialising their discoveries is the impact on academic norms, and the consequences for the growth of new knowledge. Dasgupta and David (1994) and David (1993a) have highlighted both the features that distinguish the production and dissemination of university research from that carried out in firms and also provided an economic framework for understanding these distinguishing features. They argue that the difference is not so much in the methods of enquiry or the nature of the knowledge obtained. Rather, it is the nature of the goals accepted as legitimate within the two communities of researchers, the norms regarding disclosure of knowledge and the reward systems that are held to be the distinguishing features. Roughly speaking, university research is undertaken with the intent of disclosure and the rewards include the approval and respect of a broad invisible college of peers. Inevitably, these differences are associated with differences in the types of questions tackled and the methods for representing and communicating the results of the research. As Walter Vincenti (1990) put it, scientists are interested in understanding how things are, whereas engineers focus on how they ought to be.22 These differences have evolved along with the corresponding institutions in response to some specific features of research as an economic activity. For instance, David (1991; 1993b) argues that an inability on the part of European princes and noblemen in the Middle Ages to monitor the quality and effort of the scientists they patronised provided an impetus for open disclosure of research findings. With open disclosure and peer review, the merits of the research findings and, hence, the quality of the researcher, would be easy to establish. As noted earlier, this also required consensus on the methodology and terminology. The open and rapid dissemination of research findings and the associated academic norms of scientific conduct, especially peer review, academic freedom and an apprenticeship-type relationship between research students and professors, have become part and parcel of what we think of as university research. These norms are sustained by the public subsidy for university research, allocated principally through a peer-review based mechanism. In essence, the community of
348
The economics of knowledge
researchers decides upon how funds should be allocated, subject to some general guidelines and constraints. Perhaps even more important are the norms of cooperation and collegiality. As Dasgupta and David (1994) point out, the high importance attached to priority creates a tension between complying with the norm of full disclosure and the individual urge to be the first to publish.23 Since the solution of scientific problems typically requires research into several subproblems, full disclosure yields a better outcome for the community as a whole in the long run. However, individual researchers have an incentive to free-ride by learning from others but not co-operating in turn. Although co-operative behaviour can be sustained in repeated games by threat of exclusion, Dasgupta and David (1994) argue that scientific norms greatly increase the likelihood that networks of co-operative information sharing will arise because of an increased trust. More generally, these norms are critical for the formation and sustenance of these communities. In turn, these scientific communities act as an agent of society at large, punishing those that violate co-operation (by withholding findings), reviewing the validity of results as well as training new researchers, and providing some degree of verification of the quality of the researchers themselves. That funding for scientific research is allocated by the community itself is complementary to these other functions and reinforces the ability of the scientific community to enforce scientific norms. Privatisation of knowledge weakens these norms by reducing the ability of the scientific community to sanction violators and by increasing the rewards to violators. To hark back to Dasgupta and David’s model of the situation as a repeated prisoner’s dilemma game, privatisation of knowledge increases the pay-off to withholding co-operation when secrecy followed by patenting of the results will yield large monetary pay-offs. There are indirect effects as well. Focusing universities on earning revenues through research leads to a dilution of the role of the larger scientific community in the allocation of funds for scientific research, weakening the power of the community and, hence, also weakening the hold of the norms of disclosure and collegiality. The foregoing is certainly closer to the Platonic ideal of university research rather than an accurate description of existing reality. The point is that these norms are valuable and useful, and need to be reinforced rather than weakened. Moreover, although resilient, norms are easier to destroy than build – once a sufficiently large fraction of the research community moves away from them, it will be hard to sustain them anywhere. Indeed, the growing commercial applicability of scientific research in biotechnology has been accompanied by growing anecdotal evidence of the violation of these norms by scientists, including the withholding of
Markets for technology
349
important information, delays in disclosure, refusal to co-operate with other researchers, as well as tales of abuse of graduate students and postdoctoral students (for example, Kenney, 1986). As noted earlier, the available evidence, limited as it is, does not support any significant shift in norms associated with patenting. Mowery et al. (2001) conclude that the bulk of the increased patenting appears to have been associated with the increased importance in software and biomedical research, which, though more amenable to patenting, is no less basic or fundamental than other research at universities. Based on their case study of patenting at Columbia University, they also conclude that the impacts on norms are in any case likely to be limited to a few departments most heavily involved in patenting: electrical engineering, computers science and the medical school. Even so, it is likely that in so far as the privatisation of knowledge adversely affects universities, it will be by diluting and degrading the norms of openness and collegiality. Moreover, the ‘few’ departments that are most involved in patenting are likely to be quite important ones for the production of ‘open’ academic research and public domain knowledge that is critical for socio-economic growth. In this respect, the fact that two quite important fields such as software and biomedical research are more involved in patenting and may loose their reliance on open academic settings, can be by itself a source of concern. A broader question is whether good research can flourish anywhere without such norms. This is a question of institutional design that we raise but to which we do not know the answer. Although the academic model of research has been very successful in the last 100–150 years, to our knowledge, alternatives to that model have not been tried. Indeed, there is evidence that successful research-oriented firms have tended to adopt the academic mode of organising at least some of their research.24 In other words, the prevailing Western university model has implicitly been assumed to be the only sensible way of organising research, especially basic or fundamental research. With the growing privatisation of knowledge, this assumption may well be tested in the future.
5
‘GLOBAL’ MARKETS FOR TECHNOLOGIES AND NATIONAL POLICIES
Along with other markets, markets for technology are becoming global. In some ways, this is only to be expected, given the smaller ‘transport’ costs and the greater appreciation, by even otherwise protectionist governments, of the benefits of technology. Rapid advances in communications, with
350
The economics of knowledge
the Internet being only the most recent, has only hastened the process of globalisation. Markets for technology are far more likely to arise in large and technologically and economically advanced regions, than in developing countries. The latter, therefore, need not focus on developing such markets. Instead, they can focus on developing institutions that will enable their firms to participate more effectively in these markets. The example of the Western European chemical industry in the years after the Second World War is a case in point. Prior to the war, the European chemical industry was technologically far ahead of the USA industry. The disruption due to the war and the rise of the petrochemical industry, and the associated process technologies, in the USA, ought to have provided the US chemical industry with a decisive advantage over its European rivals whose expertise lay in coal-based processes. Yet, in a period of a few years, the German, British and French chemical industries had largely switched over to using petroleum and natural gas as basic inputs. The availability of US-developed refining and chemical engineering expertise made this switch possible. Further, the specialised engineering producers, the SEFs, played an important role in integrating and supplying technology to European customers. In the 1960s, the SEFs played a similar role in Japan. Japanese industrial policy, which tended to restrict access to the Japanese market for foreign firm, was far more receptive to foreign technology imports. Indeed, the policy focus in this context was in creating the ability to absorb and adapt foreign technology (Arora and Gambardella, 1998.) The point is simple and well known: global markets tend to circumscribe the role of policy in being able to improve market outcomes. For smaller countries like the individual European countries or the less developed countries, the impact of their own policies, if they are not co-ordinated with those of other countries, is likely to be small. For example, policies by smaller countries to develop standards or other types of supporting institutions are unlikely to induce the development of technology markets on substantial scale. Similarly, strengthening or weakening intellectual property rights will probably have little effect on the global market for technology, although this may affect the extent to which technology flows into their country or technology trades takes place within it. Policies for encouraging, co-ordinating or controlling the markets for technology will be most effective when they are developed by large countries (for example, the USA), or by sets of countries (for example, the European Union). Such policies require co-ordination among countries, and this requires super-national interventions in international policy settings. But it is precisely at this super-national level that policy decisions are
Markets for technology
351
harder to take because of the many conflicting interests involved, and the lack of strong enforcement mechanisms. And this is why policies developed by a large homogeneous country like the USA (for example, in intellectual property rights, in the development of standards) can have a strong impact on the world development of markets for technology, as we are observing today with the effects and the debate that has been raised worldwide by the US attitude towards stronger patents. Likewise, the European Union can play a significant role, especially if it can harmonise the policies of the individual member states, and avoid that individual member states adopt different rules and standards. For most other countries, the key policy question may be how to take advantage of the worldwide growth in technology trade. This will require encouraging the effective use of existing technologies, rather than the creation of new ones. Also, policies aimed at monitoring international technological developments increase in importance, as do institutions for enhancing the efficiency of contracts and reducing search costs. In this view, countries may increase the emphasis on the ability to identify and select technology, and develop complementary capabilities. In sectors where markets for technology develop, and technology can be traded more effectively, countries or regions should specialise according to comparative advantages. This does not imply that countries should cease to invest in research and development. Rather, it implies that they should be more selective in terms of which sectors they focus on, and more selective in terms of the types of activities they focus on, at least in the short to medium term. It is well known that R&D and technology production is quite concentrated worldwide. The rich countries, and in particular the USA and Western Europe, have a head start in terms of basic research, developing ‘generic’ technologies like semiconductors and genetics. Their advantage lies not only in first-mover advantage, but also the broader industrial base over which to apply these findings. These advantages are less salient when technologies and products need to be adapted for local uses and needs. If one accepts that companies or industries located ‘near’ users have an advantage in communicating with their markets and in acquiring the relevant information for adapting the technologies, this suggests that firms in other parts of the world could seize this niche. Thus, even if the production of more basic technologies is concentrated regionally, if markets for technology work well, other regions can get access to these basic technologies and exploit their proximity to users, or their comparative advantage in developing complementary technologies (for example, the growing development of software components for leading software producers worldwide by Indian firms today – see Arora et al., 2001a).
352
The economics of knowledge
These recommendations are not new and in some quarters, are viewed as a prescription for perpetual technological ‘backwardness’, and some countries may resist such an international division of labour in technology production and adaptation. The reasons may range from national pride to the willingness to control strategic technologies. Thus, some form of the ‘not-invented-here’ syndrome, at the level of countries, is likely to operate. Whether justified or not, it is important to know that markets for technology, where they exist, increase the opportunity cost of such an attitude. Simply put, if others have already paid the fixed cost of developing technology, and competition among sellers implies that the price of the technology is related to the marginal cost of technology transfer, a strategy of developing the technology in-house and incurring the fixed cost all over again must provide some additional benefits over mere ownership of the technology. There is little point in national policies aimed at ‘reinventing the wheel’, except where such reinvention is a part of the process of building ‘absorptive capacity’ or as a part of a longrun strategy to develop international technological leadership. Second, in a dynamic setting, the international division of labour, with implied specialisation in technology production and adaptation, means that countries that specialise in the latter need not give up the possibility of becoming technology producers, at least in some well-defined areas. For example, by starting with a policy of developing technologies that are complementary to those developed by some leading areas or regions, the local firms and industries may gradually learn about the basic technology as well, and they could possibly escalate up to becoming the producers of some key technologies. The Indian software industry, for instance, started as a lowend supplier of software components to the major software companies especially in the USA, and this has proven to be an overly successful strategy by many standards (employment growth, exports, and so on, see Arora et al., 2001a). A similar argument can be made for the Irish software companies, which seem to have improved their ability to become producers of new software products in some niches of this market (Arora, et al., 2004). In short, in a dynamic setting, the pattern of specialisation is not immutable. With proper technology policies, the advantages of specialisation in lower-end technological activities (adaptation) could even become the springboard for a move up the value chain. Learning through systematic interactions with the markets or the technology producers of more advanced countries may be critical for this process to take place. Indeed, some countries, like Russia and Israel, and to a lesser extent, India, have a relatively well-developed scientific and engineering infrastructure. However, they lack the market size and the complementary technological and economic infrastructure that could best exploit their
Markets for technology
353
scientific and engineering infrastructure. In this respect, they are similar to specialised technology suppliers. A well-developed and globalised market for technology will enable firms from these countries to derive more value from their investments in science and engineering, by supplying technology to others that can develop and commercialise it more effectively. Here, too, one may encounter opposition from those who would see this as ‘giving away the store’. Once again, our objective is not to act as advocates, for the appropriate policy will depend on the specifics of the situation, but to highlight the option that markets for technology would create.
6
CONCLUSIONS
Markets do not arise simply because the benefits of having them outweigh the costs. They also require institutions that support them. Further, markets develop over time, along with these complementary institutions. As Paul David has long taught us, this development has to be understood as a historical process, with the pace and form of the development conditioned by starting conditions and chance. Further, the rise of a new market affects other markets and other existing social and economic institutions. Their development raises new challenges for policy-makers but creates new policy options as well. So also with markets for technology. Policy can play a relatively more important role for encouraging these markets when they arise rather than after they start functioning. Further, policies that encourage ‘decentralized’ institutional innovations, along the lines of the new ‘market-enhancing view’ of policy are likely to be more successful than direct policy attempts to create such institutions (Aoki et al., 1996). In this chapter we have highlighted some of the major policy challenges posed by the development of markets for technology. Intellectual property rights are a sine qua non for the development of markets for technology. But given the nature of knowledge, property rights in knowledge, such as patents, can create problems. In some cases, they can retard the development and commercialisation of innovations, as for instance when such a use requires combining the intellectual property rights controlled by a number of independent agents. How serious this problem is in practice is uncertain and further research in this area would be very valuable. The privatisation of knowledge can also undermine an important institution of modern capitalism, namely, the research university, by weakening academic norms of open disclosure and collegiality. Weakening public support for academic research exacerbates the problem by forcing universities to look to generate additional resources by patenting and licensing their research findings. As we note, empirical research on this topic is just
354
The economics of knowledge
beginning and the available evidence suggests that the situation is not irrevocable. However, by their nature, norms are easier to destroy than to create and it seems sensible to try to modify only very slowly a system that appears to have worked well as a way of organising basic research. With markets becoming global, the exercise of national policy has to be more circumscribed. Especially for ‘smaller’ countries like the individual European countries, or the less developed countries, markets for technology imply a focus on how best to benefit from the growth of these markets. We suggest that this would mean becoming more open to outside technology and re-examining arguments for investments based on national pride. It would mean participating in an international division of labour, by increasing the emphasis on using technology and building complementary capabilities, possibly at the cost of investing in basic research. In a dynamic setting, the learning potential that is embedded in a division of labour with more advanced technology producers can create the opportunities for later specialisation in some of the more basic technology areas.
NOTES 1.
2.
3.
4. 5.
6.
Aoki et al. (1996: 6) make a similar point. They argue that technological and other complementarities (for example, of institutions, or expectations) are more likely to occur at the early stages of development of new markets or technologies. Hence, government intervention and the need for co-ordination are more useful earlier than later when such markets function, or the technologies are more mature, and there is greater competition (and substitutability) among them. Heller and Eisenberg’s concerns were specifically directed to the growth in genomic patenting, and more specifically, on patenting of Expressed Sequence Tags (ESTs). The available evidence suggests that a variety of factors, including institutional responses by the Patent and Trademark Office (PTO), the courts and by the National Institute Health (NIH), as well as private responses by firms, have largely avoided the problem. Further details are contained in Walsh et al. (2003a; 2003b). Clearly, conferences, seminars, visiting positions in other schools or departments show that the exchange of papers is not enough, and many scholars have pointed to the importance of tacit communication. Even so, one striking example of what a common language can do is open-source software, whereby software programs are developed by developers located far away from each other and communicating principally over the Internet and by exchanging the software code itself. In this respect, another area for further study is the effective prohibition on lawyers undertaking patent infringement cases on a ‘contingency’ basis. This is the same argument put forward in Arora (1995) to explain the problems in transferring technology based on tacit, unpatented knowledge. Anton and Yao (1994) develop a clever model which shows that in principle, a technology holder can sell the technology without any intellectual property protection, in effect, by threatening to put it in the public domain (and destroying its value) if the buyer reneges on it. Anton and Yao (2002) show that a seller can partially reveal the technology to signal its value, mitigating the asymmetric information problem. In some cases, policies designed in the naive hope of encouraging small inventors have encouraged the abuse of the patent system. In the USA, for instance, there have been
Markets for technology
7.
8.
9. 10.
11.
12.
13.
14.
15.
16.
355
well-known cases where patents filed in the 1950s ultimately were issued more than 20 years later. In the mean time, the patentee could legally amend the application so that it covered inventions made well after the filing date. Since patents in the USA are published only upon issue, many established firms have been surprised by such patents (sometimes referred to as ‘submarine’ patents because they are not visible for long periods after they are filed). The move towards patent harmonisation, which will require publication of patent applications after a certain period, will be helpful in this respect. The point is not that information can be reproduced at low cost or that information is non-rival in the sense that one person’s knowing something does not preclude another from possessing the same information. A familiar counter-example is as follows. If only one person knows what is going to happen to the price of a stock, he or she is likely to benefit greatly. But if all (or sufficiently many) were to have the same information, none is going to benefit. Thus information can be rival in use, although in the physical sense, it is non-rival. As Paul David (1993a) noted, knowledge is different from the prototypical public goods such as lighthouses and airport beacons. One important point of differentiation is that the acquisition of knowledge is cumulative and interactive: knowledge itself is an important input into the production of knowledge. Other factors such as standards may also raise the cost of inventing around. A related consequence is that non-manufacturing firms that hold patents on key components are likely to bargain more aggressively for licensing fees. The strategies of firms that have significant market shares in the downstream markets (in which the technology is applicable) are more complex. However, they are likely to co-operate, particularly if there is a stable group of such firms. Interestingly enough, the ownership of mutually blocking patents can actually support licensing in this context, since each party will have the ability to block commercial development by the other. Think for instance of two patent holders fixing separately their royalty rates for selling their patents to a unique licensee vis-à-vis the case in which the two patents are pooled and a single royalty is set up. This is a distortion similar to the one generated by the double marginalisation in a chain of monopolies. For instance, as Walsh et al. (2003b) note, NIH negotiated with DuPont to provide more favourable terms for transgenic mice for NIH and NIH-sponsored researchers, to relax restrictions on publication and sharing of animals and eliminate reach-through provisions. The NIH has also begun a ‘mouse initiative’ to sequence the mouse genome and create transgenic mice. One of the conditions of funding is that grantees forgo patenting on this research. Then NIH also pushed for broader access to stem cells, as well as for a simplified, one-page Material Transfer Agreement (MTA) without reach-through claims or publication restrictions. The firms in the SNPs Consortium include Bayer, Bristol-Meyers Squibb, Glaxo Wellcome, Hoechst Marion Roussel, Monstanto, Novartis, Pfizer, Roche, SmithKline Beecham and Zaneca. Each firm contributed $3 million, and Wellcome Trust added another $14 million to the effort. One could speculate that the shift of Human Genome Science (HGS) away from the database business and toward the drug development business may be a response to both the higher returns available to drug companies and the lower returns available to genomics companies that are competing with increasingly developed public databases. Merges (1999) shows that the US Patent Office has about $3000 to spend on each patent application. Further research is needed to assess whether this is the optimal amount to spend. Any such assessment should take into account the impact of intellectual property rights on the functioning and development of markets for technology. Cockburn et al. (2002) find that the CAFC went from upholding the plaintiff in about 60 per cent of the cases to finding for the plaintiff in only 40 per cent of the cases in recent years. Similarly, in University of California v. Eli Lilly and Co, the court ruled against the University of California’s argument that its patent on insulin, based on work on rats, covered Lilly’s human-based bio-engineered insulin production process. Similarly, it
356
17. 18.
19. 20.
21.
22.
23.
24.
The economics of knowledge ruled against the University of Rochester in its attempt to enforce its COX-2 patent against Searle. More nuanced institutional arrangements are also possible. For copyrights, organisations such as ESCAP that hold the copyrights of individual song writers and singers and collect fixed royalty payments for their use on behalf of the artists, have worked well. However, in other circumstances firms failed to reach a satisfactory agreement for pooling together the patents. This is the case, for instance, for DVDs, where agreements to cross-license the rights or to pool them together have not been reached for quite a few years. Needless to say, ours is a hindsight view, not a reconstruction of how university leaders saw the situation at the time. We should also note that there may be many benefits for university researchers who leave the university to start ‘spin-off ’ firms. These benefits may take many forms, including providing the researchers with better information on financially (and economically) promising areas of research, and in providing teachers with better information on the types of skills and competencies students need. However, for the most part, university spin-offs are celebrated as evidence of the university’s contribution to the national and regional economy, ignoring the potentially much greater contributions in terms of training and other types of technology transfer, such as faculty consulting with industry. In this context, one must note that American universities have historically been very responsive to industry needs. Collaborative research relationships between university and industry in a broad range of fields have been a distinctive hallmark of the American university system (Rosenberg and Nelson, 1994). Rosenberg (1992) in particular, has convincingly argued the critical role that American universities have played in supporting innovation, often by helping in the solution of very practical, and sometimes, scientifically mundane problems. Dasgupta and David (1987; 1994) distinguish between what they call the realm of Science and the realm of Technology, associating the first with open, university-type research and the latter with research in firms. It is tempting to interpret this as implying that researchers in firms never participate in open research, or that university research is never applied nor with immediate practical utility. This interpretation is incorrect. Rosenberg and Nelson (1994) have argued, for instance, that, in the USA at least, university researchers have also performed a variety of important applied activities, such as simple chemical assays, or development of instruments like instruments for ascertaining the fat content of milk at the University of Wisconsin. One reason is that the results of one project feed into the next. Full disclosure of the results of one project would put all researchers on the same footing in terms of being the first to complete the next. By contrast, by only imperfectly disclosing the research finding, a researcher completing a stage ahead of others would get a head start on completing the next stage as well. See Gambardella (1995) for a discussion of this point in the context of the pharmaceutical industry.
REFERENCES Alston, L., Eggertsson, T. and North, D. (1996) Empirical Studies in Institutional Change, Cambridge University Press, Cambridge. Anton, James J. and Dennis A. Yao (1994) ‘Expropriation and inventions: appropriable rents in the absence of property rights’, American Economic Review, 84, 190–209. Anton, James J. and Dennis A. Yao (2002) ‘The sale of ideas: disclosure, property rights, and incomplete contracts’, Review of Economic Studies, 69, 513–31.
Markets for technology
357
Aoki, M., Murdock, K. and Okuno-Fujiwara, M. (1996) ‘Beyond the East Asian miracle: introducing the market-enhancing view’, in Aoki, M., Kim, H.-K. and Okuno-Fujiwara, M. (eds), The Role of Government in East Asian Economic Development, Oxford University Press, Oxford. Arora, A. (1995) ‘Licensing tacit knowledge: intellectual property right and the market for know-how’, Economics of Innovation and New Technologyi, 4, 41–59. Arora, A. and Fosfuri, A. (2003) ‘Licensing the market for technology’, Journal of Economic Behavior and Organisation, 52, 277–95. Arora, A. and Gambardella, A. (1994) ‘The changing technology of technological change: general and abstract knowledge and the division of innovative labour’, Research Policy, 23, 523–32. Arora, A. and Gambardella, A. (1998) ‘Evolution of industry structure in the chemical industry’, in Arora, A., Landau, R. and Rosenberg, N. (eds), Chemicals and Long-Term Economic Growth, John Wiley, New York. Arora, A. and Merges, R. (2004) ‘Intellectual property rights, firm boundaries and R&D inputs’, Industrial and Corporate Change, 13 (3), 451–75. Arora, A., Arunachalam, V., Asundi, J. and Fernandes, R. (2001a) ‘The Indian software service industry’, Research Policy, 30, 1267–87. Arora, A., Fosfuri, A. and Gambardella, A. (2001b) ‘Specialised technology suppliers, international spillovers and investment: evidence from the chemical industry’, Journal of Development Economics, 65, 31–54. Arora, A., Fosfuri, A. and Gambardella, A. (2001c) Markets for Technology: The Economics of Innovation and Corporate Strategy, MIT Press, Cambridge, MA. Arora, A., Gambardella, A. and Torrisi, S. (2004) ‘In the footsteps of the Silicon Valley? Indian and Irish software in the international division of labor’, in Bresnahan, T. and Gambardella, A. (eds), Building High Tech Clusters: Silicon Valley and Beyond, Cambridge University Press, Cambridge. Bordo, M., Eichengreen, B. and Irwin, D. (1999) ‘Is globalization today really different from globalization a hundred years ago?’, NBER Working Paper 7195, NBER, Cambridge, MA. Bresnahan, T. and Gambardella, A. (eds) (2004) Building High Tech Clusters: Silicon Valley and Beyond, Cambridge University Press, Cambridge. British Technology Group (BTG) (1998) ‘IPR market benchmark study’, www.btgplc.com Cockburn, I., Kortum, S. and Stern, S. (2002) ‘Are all patent examiners equal? The impact of examiner characteristics’ NBER Working Paper 8980, www.nber.org Dasgupta, P. and David, P. (1987) ‘Information disclosure and the economics of science and technology’, in Feiwel, G. (ed.), Arrow and the Ascent of Economic Theory, New York University Press, New York. Dasgupta, P. and David, P. (1994) ‘Towards a new economics of science’, Research Policy, 23, 487–521. David, P. (1990) ‘The dynamo and the computer’, American Economic Review Papers and Proceedings, 80 (2), 355–61. David, P. (1991) ‘Reputation and agency in the historical emergence of the institutions of “open science,” ’ CEPR Working Paper N.261, CEPR, Stanford University, Stanford, CA. David, P. (1993a) ‘Knowledge, property, and the system dynamics of technological change’, Proceedings of the World Bank Annual Conference on Development Economics 1992, World Bank, Washington, DC.
358
The economics of knowledge
David, P. (1993b) ‘Intellectual property institutions and the panda’s thumb: patents, copyrights, and trade secrets in economic theory and history’, in Wallerstein, M., Mogee, M. and Schoen, R. (eds), Global Dimensions of Intellectual Property Rights in Science and Technology, National Academy Press, Washington, DC. David, P. (1997) ‘From market magic to calypso policy: a review of Terence Kealey’s The Economic Laws of Scientific Research’, Research Policy, 26, 229–55. David, P. (1998) ‘Common agency contracting and the emergence of “open science” institutions’, American Economic Review, 88 (2), 15–21. David, P. (2003) ‘The economic logic of “open science,” ’ SIEPR Working Paper 02–30, SIEPR, Stanford University, Stanford, CA; forthcoming in The Role of the Public Domain in Scientific and Technical Data and Information, National Academy Press, Washington, DC. David, P. and Greenstein, S. (1990) ‘The economics of compatibility standards: an introduction to recent research’, Economics of Innovation and New Technology, 1, 3–41. Deng, Z., Lev, B. and Narin, F. (2003) ‘Science and technology as predictors of stock performance’, in Hand, J. and Lev, B. (eds), Intangible Assets: Values, Measures and Risks, Oxford University Press, Oxford. European Technology Assessment Network (ETAN) (1999) ‘Strategic dimensions of Intellectual property rights in the context of science and technology policy’, ETAN Expert Working Paper, European Commission, Directorate General Science and Technology, Brussels, June. Gambardella, A. (1995) Science and Innovation: The US Pharmaceutical Industry in the 1980s, Cambridge University Press, Cambridge, MA. Gilbert, R. and Shapiro, C. (1990) ‘Optimal patent length and breadth’, Rand Journal of Economics, 21, 106–12. Gompers, P. (1999) ‘Resource allocation, incentives and control: the importance of venture capital in financing entrepreneurial firms’, in Acs, Z., Carlsson, B. and Karlsson, C. (eds), Entrepreneurship, Small and Medium Sized Enterprises and the Macroeconomy, Cambridge University Press, Cambridge. Green, J. and Scotchmer, S. (1995) ‘On the Division of Profits in Sequential Innovations’, Rand Journal of Economics, 26 (1), 20–33. Grindley, P.C. and Teece, D.J. (1997) ‘Licensing and cross-licensing in semiconductors and electronics’, California Management Review, 39 (2), 8–41. Grossman, S. and Hart, O. (1986) ‘The costs and benefits of ownership: a theory of vertical and lateral integration’, Journal of Political Economy, 94, 691–719. Hadley, W. (1998) ‘Rapporteur’s report’, conference on intellectual property rights and competition policy, Stanford Law School, Stanford, CA, April. Hall, B.H. and Ziedonis, R. (2001) ‘The patent paradox revisited: determinants of patenting in the US semiconductor industry, 1980–1994’, Rand Journal of Economics, 32 (1), 101–28. Hand, J. and Lev, B. (eds) (2003) Intangible Assets: Values, Measures and Risks, Oxford University Press, Oxford. Hart, O. and Moore, J. (1990) ‘Property rights and the nature of the firm’, Journal of Political Economy, 98, 1119–58. Heller, M. (1998) ‘The tragedy of the anticommons: property in the transition from marx to markets’, Harvard Law Review, 111, 621–88. Heller, M. and Eisenberg, R. (1998) ‘Can patents deter innovation? The anticommons in biomedical research’, Science, 280, 698–701.
Markets for technology
359
Henderson, R., Jaffe, A. and Trajtenberg, M. (1998) ‘Universities as a source of commercial technology: a detailed analysis of university patenting, 1965–1988’, Review of Economics and Statistics, 80, 119–27. Kenney, M. (1986) Biotechnology: The University–Industrial Complex, Yale University Press, New Haven, Ct. Klemperer, P. (1990) ‘How broad should the scope of patent protection be?’, Rand Journal of Economics, 21, 113–30. Kremer, M. (1998) ‘Patent buyouts: a mechanism for encouraging innovation’, Quarterly Journal of Economics’, 113, 1137–67. Kortum, S. and Lerner, J. (1999) ‘What is behind the recent surge in patenting’, Research Policy, 28, 1–22. Lamoreaux, N. and Sokoloff, K. (1996) ‘Long term change in the organisation of inventive activity’, Proceedings of the National Academy of Science USA, 93, 12686–92. Lamoreaux, N. and Sokoloff, K. (1999) ‘Inventors, firms, and the market for technology: US manufacturing in the late nineteenth and early twentieth centuries’, in Lamoreaux, N., Raff, D. and Temin, P. (eds), Learning by Firms, Organisations, and Nations, University of Chicago Press, Chicago, IL. Langlois, R. (2002) ‘Modularity in technology, organisation, and society’, Journal of Economic Behavior and Organisation, 49, 19–37. Langlois, R. and Robertson, P. (1992) ‘Networks and innovation in a modular system: lessons from the microcomputer and stereo component industries’, Research Policy, 21, 297–313. Lerner, J. (1995) ‘Patenting in the shadow of competitors’, Journal of Law and Economics, 38, 563–95. Lev, B. (2000) Intangibles: Management, Measurement and Reporting, Brookings Institutions, Washington, DC. Linden, G. and Somaya, D. (2003) ‘System-on-a-chip integration in the semiconductor industry: industry structure and firm strategies’, Industrial and Corporate Change, 12, 545–76. Merges, R. (1999) ‘As many as six impossible patents before breakfast: property rights for business concepts and patent system reform’, Berkeley Technology Law Journal, 14, 577–615. Merges, R. and Nelson, R. (1990) ‘On the complex economics of patent scope’, Columbia Law Review, 90 (4), 839–916. Merges, R. and Nelson, R. (1994) ‘On limiting or encouraging rivalry in technical progress: the effect of patent scope decisions’, Journal of Economic Behavior and Organisation, 25, 1–24. Mowery, D., Nelson, R., Sampat, B. and Ziedonis, A. (2001) ‘The growth of patenting and licensing by US universities: an assessment of the effects of the BayhDole Act of 1980’, Research Policy, 30, 99–119. Mowery, D. and Ziedonis, A. (2002) ‘Academic patent quality and quantity before and after the Bayh–Dole Act in the United States’, Research Policy, 31, 399–418. National Research Council (1997) Intellectual Property Rights and Research Tools in Molecular Biology, National Academy of Sciences, Washington, DC. North, D. (1990) Institutions, Institutional Change and Economic Performance, Cambridge University Press, Cambridge. Rosenberg, N. (1992) ‘Scientific instrumentation and university research’, Research Policy, 21, 381–90.
360
The economics of knowledge
Rosenberg, N. and Birdzell, L. (1986) How the West Grew Rich, Basic Books, New York. Rosenberg, N. and Nelson, R. (1994) ‘American university and technical advance in industry’, Research Policy, 23, 323–48. Scotchmer, S. (1991) ‘Standing on the shoulders of giants: cumulative research and the patent law’, Journal of Economic Perspectives, 5 (1), 29–41. Vincenti, W.G. (1990) What Engineers Know and How They Know It, Johns Hopkins University Press, Baltimore, MD. Walsh, J., Arora, A. and Cohen, W. (2003a) ‘Working through the patent problem’, Science, 299 (5609), 1021. Walsh, J., Arora, A. and Cohen, W. (2003b) ‘Research tool licensing and patenting and biomedical innovation’, in Cohen, W. and Merril, S. (eds), Patents in the Knowledge Based Economy, NAS Press, Washington, DC. Walsh, J.P. and Bayma, T. (1996) ‘Computer networks and scientific work’, Social Studies of Science, 26, 661–703.
13. The key characteristics of sectoral knowledge bases: an international comparison* Stefano Brusoni and Aldo Geuna1 1
INTRODUCTION
This chapter builds upon and extends existing studies of scientific and technological specialisation by proposing a unifying theoretical framework in which to compare sectoral knowledge bases across countries. In conducting this comparison, we elaborate upon the large body of literature that analyses national systems of innovation (NSI) (Lundvall, 1992; Nelson, 1993). An NSI is defined as ‘being comprised of those elements of social organisation and behaviour, and the relationships among them, that are either located within or rooted inside the borders of a national state, and that interact in the production, diffusion and use of new, and economically useful knowledge’ (David and Foray, 1995, p. 14). The concept of NSI gained wide popularity that goes beyond the boundaries of the academic community as it became (often unwillingly) entangled with ‘techno-nationalistic’ positions that have animated the industrial policy debate throughout the 1980s and 1990s. As stressed by David and Foray (1995), such positions are based upon two related (and nowadays widely held) assumptions. First, technical capabilities lie at the core of a country’s international competitiveness. Second, the development of such capabilities is influenced by issues of national localisation and can be managed via proper government action. Recent research has challenged the relevance of the national dimension. In particular, it stresses that firms and researchers are entwined in thick networks of international relationships that cut across national boundaries. National systems of innovation come under increasing strain, as the research and development (R&D) activities of large firms are progressively internationalised. Such internationalisation is caused by emerging imbalances between what a country’s science base has to offer and the knowledge requirements of innovative processes. However, despite their undeniable 361
362
The economics of knowledge
increase, R&D linkages have not developed on a global scale, but rather they involve mainly US, European Union (EU) and, to a lesser extent, Japanese firms (Patel and Pavitt, 2000). In this situation of internationalised rather than globalised R&D activities, it is very important to understand why specific countries lie at the core of such international networks. Standard explanations refer to a number of factors considered to be key determinants of ‘national competitiveness’ (Porter, 1990). Following a well-established tradition (Fagerberg et al., 1999), this chapter acknowledges that a country’s specialisation pattern in specific scientific and technological fields plays a key role: firms establish R&D facilities for which they perceive they have the relevant capabilities. However, most studies that empirically explore specialisation patterns at country level focus on a rather narrowly defined concept of specialisation. The emphasis falls squarely on the fields in which countries and/or firms patent. Classic specialisation studies focus on the cumulative evolution of countries’ technological capabilities and, in most cases, scientific specialisation is not analysed. The stability of specialisation patterns over time (what we will term ‘knowledge persistence’) is well established; however, persistence and cumulativeness are not the only dimensions relevant to a study of knowledge bases. It is well known that design and development activities capture a relevant share of the R&D funded by companies (Rosenberg, 1994). A country’s knowledge base may have a strong science base but lack the engineering capabilities to embody scientific results in profitable products. Or it can have strong development capabilities that are not supported by robust basic scientific knowledge. Different typologies of knowledge are complementary and interrelated. A strong presence in each typology of research induces an easier multi-directional flow of knowledge that can facilitate the production of successful innovation. Micro-level innovation studies strongly support this view, for example, Pisano (1997). Therefore, what type of research is carried out in each field (for example, basic versus engineering-oriented research) becomes a key issue. The chief aim of this chapter is to develop a framework in which to analyse knowledge specialisation both over time and across research typology. We put forward this framework as a way to approach questions related to industry decisions to source knowledge internationally. In particular, we want to link these decisions to specific characteristics of the sectoral knowledge base that is drawn on. The chapter identifies and operationalises, at sectoral level, the relevant dimensions that make the comparison of the knowledge bases of different countries a meaningful exercise. Particular attention is devoted not only to examining whether each country’s specialisation is stable over time (knowledge persistence),
The key characteristics of sectoral knowledge bases
363
but also to whether specialisation by field is similar across different typologies of research (knowledge integration). The operationalisation of these two dimensions is based upon the design of a comprehensive data set of peer-reviewed papers that was obtained by combining the standard Institute for Scientific Information (ISI) classification by science field with the Computer Horizons Inc. (CHI) classification by type of research (that is, Applied Technology and Engineering, Applied Research and Basic Research). The result is an original data set encompassing some 630 000 papers in 11 different sub-fields of chemistry and pharmacology published between 1989 and 1996. The limitations of peer-reviewed publications as an indicator of the knowledge bases is discussed. This dataset will allow for a quantitative analysis of the characteristics and evolution of the specialisation profile of the four largest European countries (the UK, Germany, France and Italy), the EU as a whole, the USA and Japan. This data-set is analysed in combination with the Policies, Appropriation and Competitiveness in Europe (PACE) survey (Arundel et al., 1995). The results of the PACE questionnaire pinpoint the pharmaceutical industry as being a highly internationalised industry. The PACE survey data show that not only do EU R&D managers in the pharmaceutical sector value the results of public research, but also that they rely upon international research much more than those in the chemical sector and in other manufacturing industries. Also, PACE stresses that the pharmaceutical industry relies more on North American research than on EU research. The questions that demand explanation are why do EU pharmaceutical firms rely to such a great extent on North American research? What makes it attractive to EU firms? In attempting to answer these questions, we discuss some evidence related to the existence of a ‘European paradox’ in the case of traditional pharmaceuticals. To do this, we compare sectoral knowledge bases across countries by developing a grid designed along the two dimensions identified above: integration and persistence. The chapter is organised as follows. Section 2 discusses the concept of knowledge persistence and integration. Section 3 presents an empirical exploration of the concepts developed in section 2 in the case of the pharmaceuticals and chemicals knowledge bases. Finally, section 4 offers concluding observations and raises a few policy issues.
2
TOWARDS A THEORETICAL FRAMEWORK OF KNOWLEDGE SPECIALISATION
Although the recent literature has devoted increasing attention to analysis of the economics of science and its implication for the innovation process
364
The economics of knowledge
(Dasgupta and David, 1994; Mansfield, 1991; Narin et al., 1997), the analysis of national science and technology specialisation profiles has remained, so far, largely independent. Despite token acknowledgement of the complexity and intricacy of the relationships between the science and the technology domains, specialisation studies tend to focus either on science or on technology. The former traditionally rely on bibliometric indicators; the latter on patent studies. The former are dominated by sociologists of science; the latter by economists who study technical change. The rhetoric of the linear model still determines the intellectual division of labour in this area of research. This chapter represents a first step towards the redefinition of such a division of labour. This is achieved by complementing the analysis of knowledge specialisation over time with an analysis of knowledge specialisation across type of research. First, we briefly review a few classic specialisation studies developed in the historical, sociological and economic literature to stress the cumulative and path-dependent process of knowledge production and accumulation. The concept of knowledge persistence (that is, of specialisation over time) is based upon these notions. Second, micro-level analysis of technical change will inspire the introduction of the concept of knowledge integration (that is, specialisation across type of research). In this respect, and somewhat paradoxically, this chapter is a first attempt to develop the ‘macro-foundations of innovation management studies’. We will argue that the combined study of specialisation, in terms of both specialisation over time and specialisation across types of research, enables analysts to start addressing issues such as why different industries utilise national versus international sources of knowledge to different extents. 2.1
Specialisation Patterns over Time: Knowledge Persistence
Research in the history of science has stressed the cumulative and social aspects of scientific endeavour. Historians have provided a number of accurate case histories that reveal how the accumulation of results over time influences the rate and direction of the discovery process. For instance, Conant and Nash (1964) describe the process of accumulation of quantitative results in physics that led to Lavoisier’s revolution in modern chemistry. Such a process did not entail the substitution of inaccurate explanations with more accurate ones; rather, it involved the re-conceptualisation of existing findings to deliver a new, more general, explanation. In addition, it is particularly interesting that scientific advancement is often focused on a common frontier. The evidence for this is the incidence of multiple discoveries that Merton characterised as endemic rather than isolated features of science (Merton, 1965). The cumulative development of science has also
The key characteristics of sectoral knowledge bases
365
been studied following the seminal work of Price (1963). Price sketches a macro ‘growth of knowledge’ approach that highlights the acceleration of scientific publication that accompanied the growth of the scientific community. This approach is probably more congenial to economists who can advance a number of established theoretical propositions to explain Price’s empirical results. First, the increasing size of the scientific community would enable increasing division of labour and generate network externalities so that ‘increasing returns’ in scientific endeavour would be activated. Second, the growth of the scientific community stimulates the race for priority in discovery. This would create a powerful incentive to publish more prolifically in order to share some of the credit for ‘discovery.’ Scientific advance would then occur in smaller steps with greater overlap and duplication. Third, as the scientific community grows it becomes more difficult to assess individual contributions which in turn provides an incentive to produce more publications in order to make claims about ‘productivity’. As the three mechanisms are not mutually exclusive, cumulativeness is the most likely outcome.2 On the basis of the above-mentioned literature, studies in the fields of bibliometrics and the sociology of science have analysed the scientific base of individual countries in terms of publications share (Braun et al., 1995). However, the analysis of absolute shares does not allow for meaningful cross-country comparisons. Only recently has the methodology used to analyse technological specialisation (based upon relative specialisation indicators) been applied to the publication output of countries in an attempt to develop a comparative analysis of scientific specialisation patterns (European Commission, 1997; Geuna, 2001; Godin, 1994; OST, 1998; Pianta and Archibugi, 1991). The works of Soete (1981), Pavitt (1989) and Cantwell (1989) provide the building blocks for the analysis of stability of technological specialisation patterns at the country level. Following these studies a large body of literature has been devoted to the study of technology and trade specialisation. The analysis of country-level technological specialisation patterns is nowadays a methodology commonly used to study the relationship between innovation and performance in terms of international trade and/or growth. In a nutshell, as technical change is a cumulative process that generates clusters of innovations, it is not indifferent to which technological areas countries are specialised in (Meliciani, 2001). Different technical fields are characterised by different degrees of innovative opportunities and appropriability conditions (Carlsson, 1997; Malerba and Orsenigo, 1997). Furthermore, the learning processes that underpin technical change tend to be localised and cumulative (Pavitt, 1992): it is easier to learn in the proximity of what one already knows, so to speak. Therefore, if one is specialised in the ‘wrong’(that
366
The economics of knowledge
is, low opportunity) technical or scientific fields, one should not expect to be able to refocus one’s own specialisation pattern in the short term. Trade and growth indicators will reflect such ‘bad’ specialisation. Scholars of technical change have therefore devoted much effort to matching technological specialisation indicators and countries’ growth indicators (Fagerberg et al., 1999). Although there is some consensus about the importance of the knowledge base (or science base) of a country in the process of economic growth, the empirical and theoretical analyses have focused almost entirely on technology (especially patents) and, generally, do not attempt to provide measurement of the scientific base of the country. The work of Archibugi and Pianta in the early 1990s (see, for example, Archibugi and Pianta, 1992) is a rare example of the combination of patent studies and bibliometric analysis to examine national specialisation in EU countries. Expanding upon the bodies of literature discussed above, we define knowledge persistence as the stability of the knowledge specialisation pattern over time. 2.2
Specialisation across Research Typologies: Knowledge Integration
Persistence and cumulativeness are not the only dimensions relevant to a study of the knowledge bases of firms or countries. Micro-level studies of technical change have highlighted how the integration of different types of research plays a crucial role in the process of innovation. Integration issues have been studied at length in the innovation management literature. Pavitt (1998) stresses that the key role played by modern firms is to map an increasing range of relevant disciplines into products. Integration efforts at firm level have been thoroughly discussed by a number of authors. Granstrand et al. (1997) studied the distributed capabilities that enable firms to monitor and integrate technologies. Iansiti (1998) analysed integration issues in the mainframe industry. Prencipe (1997) studied similar problems in the aero-engine industry. Engineering disciplines are commonly stressed as being powerful, although often overlooked, enablers of such integration. They provide the problem-solving techniques to handle complex problems by decomposing them into simpler sub-tasks, which can be solved and then integrated back into a consistent whole. For instance, Patel and Pavitt (1994) studied the pervasiveness of mechanical engineering skills across a variety of sectors. Landau and Rosenberg (1992) analysed chemical engineering as the key engine of growth in the modern chemical industry. Vincenti (1990) stressed the key role played by engineers and engineering sciences in solving the problems and finding the explanations that led to the birth of the aircraft industry. Pisano (1996; 1997) studied in detail a sample of pharmaceutical development projects in order to conclude that success is related to the capability to carry out,
The key characteristics of sectoral knowledge bases
367
in a co-ordinated and timely manner, a number of activities that go well beyond the traditional boundaries of the R&D laboratory. The development of economically viable routes to produce drugs on an industrial scale is fraught with complex engineering issues, particularly where new untested routes are being explored. Nevertheless, more aggregated studies continue to have a focus on indicators that do not make it possible to analyse whether the country possesses a strong knowledge base that spans basic, applied and development research activities in any specific sector. A country’s sectoral knowledge base may have a strong science base, but lack the engineering capabilities to embody scientific results within profitable products, or strong development capabilities, but a not sufficiently robust base of scientific knowledge. In either case, firms may need to access those capabilities that are lacking, from where they exist, for example, from another country. This view is not based on a simple linear model that sees basic research as the source of the whole knowledge that is then transformed into technology. On the contrary, what we want to stress here is that the various typologies of knowledge are complementary and interrelated. A strong base in each typology of research induces an easier multi-directional flow of knowledge that can facilitate the production of successful innovation. Such intuition is consistent with the theoretical framework developed by David and Foray (1995: 40), when they argue that ‘an efficient system of distribution and access to knowledge is a sine qua non condition for increasing the amount of innovative opportunities’. Consistent with the results of micro-level studies of technical change, we argue that the successful exploitation of such combinations requires the existence of capabilities spanning a range of disciplines that go beyond the traditional boundaries of scientific endeavour. Knowledge bases that are too narrowly focused around core scientific disciplines (with no competencies in the related, but different, engineering sciences) may fail to close the feedback loop between the science and the technology domains. Such failure would seriously hamper the ‘ “distribution power” of the system’ (David and Foray, 1995: 46). In other words, in order to close the feedback loop between the science and the technology domains, countries (as well as firms) need to maintain distributed (rather than narrowly focused) competencies at sectoral level. As a national bias seems to exist in terms of the effectiveness of the linkages between business practitioners and academic research (Arundel and Geuna, 2004; Malo and Geuna, 2000; Narin et al., 1997), it is likely that such a bias exists also with respect to the linkages between the scientific and engineering communities. Thus, particular attention should be devoted not only to examining whether each country’s sectoral specialisation is stable over time (knowledge persistence), but also to
368
The economics of knowledge
Knowledge integration Low
High
High
Country B
Country A
Low
Country D
Country C
Knowledge persistence
Figure 13.1
Matrix of knowledge specialisation
whether each country’s sectoral specialisation cuts across different types of research (knowledge integration). A sectoral knowledge base with high knowledge integration would have similar specialisation by field across different typologies of research. To conclude, what we propose is a simple analytical framework capable of combining the analysis of specialisation profiles over time with the specialisation across type of research. This framework, built upon the notions of knowledge persistence and knowledge integration, should shed light on the ‘morphological’ characteristics of different countries’ knowledge bases in certain sectors, and thus help to explain firms’ international knowledge sourcing decisions. Figure 13.1 summarises the above discussion. In what follows, we will argue that this typology can be usefully deployed to study a number of issues related to the characteristics and evolution of countries’ sectoral specialisation profiles, as well as firms’ decisions about where to source useful knowledge and capabilities. With respect to any specific sector, a country can be positioned in one of the four quadrants of the matrix of knowledge specialisation (Figure 13.1). Country A, in the top right-hand quadrant is characterised by a persistent pattern of scientific specialisation and high level of knowledge integration. In the fields where it is positively specialised, Country A has developed capabilities in basic, applied and engineering research. Country B (top lefthand quadrant) would be persistently specialised in one or more fields, but its capabilities would be focused on, say, basic research only. Country C (bottom right-hand quadrant) would be characterised by integrated, although somewhat erratic, scientific and technological skills. Finally, Country D would be both erratic and unfocused in terms of research types: the fields of positive specialisation would change frequently and would be different in different types of research.
The key characteristics of sectoral knowledge bases
3
369
AN EMPIRICAL EXPLORATION OF KNOWLEDGE PERSISTENCE AND KNOWLEDGE INTEGRATION
In this section, we argue that by introducing the notion of knowledge integration alongside the more traditional one of persistence we can actually quantitatively explore problems that are only partially answered by more traditional specialisation studies. Specifically, in this chapter we focus on two main issues. First, we are interested in understanding the differences between patterns of internationalisation of knowledge sourcing activities pursued by different industries. Or, in other words, why firms in different industries appear to rely more on foreign innovation systems. Traditional explanations of this type of behaviour stress that firms go abroad whenever they (think they) can access ‘better’ capabilities relevant to their innovative and manufacturing efforts (Cantwell, 1995). Implicitly, these explanations assume that firms go abroad when their home knowledge base is not specialised in the ‘right’ fields. Second, the notion that a European paradox exists has gained wide support in the public policy arena (European Commission, 1996). According to this position, in some sectors EU firms would be very good at developing new ideas, but would tend to fail to exploit them commercially. Something would be ‘missing’ from the EU system of innovation (or its national components) that would leave EU firms at a disadvantage to their US counterparts. While the anecdotal evidence is abundant, rigorous empirical studies to prove (or disprove) the existence of such a problem are scant. Tijssen and van Wijk (1999) provide one of the few systematic efforts to solve this difficulty using robust empirical data in the specific case of the ICT sector. To contribute to this debate, we operationalise our framework in the case of the international pharmaceutical industry, using the chemical industry as a yardstick. The pharmaceutical industry is an interesting case for our purposes for a number of reasons. First, it relies heavily on basic, highly codified research at the forefront of human knowledge; thus, the scientific and technological knowledge base contributes to the development of this industry in a crucial way. Second, the pharmaceutical industry appears to be one of the most internationalised manufacturing sectors, not only in terms of product markets, but also, specifically, in terms of the knowledge sourcing strategies pursued by the major players (Patel and Pavitt, 2000). Third, and consistent with the previous point, the results of the PACE survey (Arundel et al., 1995)3 show that ‘general and specialised knowledge’ produced by public research institutes is particularly valuable to pharmaceutical firms (much more than to other manufacturing sectors), and that these firms consider scientific publications to be the key channel
370
The economics of knowledge
to internationally access this knowledge. Thus, publications can be used as a proxy for the measurement of the characteristics of persistence and integration of the knowledge base.4 Chemicals are used as a benchmark for establishing the divergence of pharmaceuticals. This benchmark is appropriate because the chemical industry, overall, behaves similarly to the other industrial sectors (see also, Geuna, 2001). If all sectors are similar, why choose the chemical industry as a benchmark, and not some other? The principal reason is that we wish to make inter-country and inter-sectoral comparisons simultaneously. Countries differ in the nature and extent of their development of specific industries. Since these differences are very difficult to capture, it is useful to choose industries that share a common knowledge base as the point of reference, but that rely on knowledge generated outside their home countries to different extents. Therefore, differences in sectoral behaviour may be related to the country-specific characteristics of the foreign NSI. Needless to say, the knowledge bases of the chemical and pharmaceutical industries differ greatly. The key difference is the increasing reliance of pharmaceuticals on biology and biotechnology, rather than chemistry (Gambardella, 1995; Orsenigo, 1989; Pisano, 1997). The chemical sector seems not to have seriously explored the potential of biotechnologies, although recent developments in combinatorial chemistry and biology provide evidence of the possibility for convergence (Malo and Geuna, 2000). By leaving aside the biotechnological knowledge base, the portion of the pharmaceuticals knowledge base that relies on the more traditional chemical processes can be analysed. This knowledge base is fairly similar to the knowledge base relied on by the chemical industry. 3.1 International Knowledge Sourcing Activities and the European Paradox The results of the PACE survey reveal the sources of the public research activities most useful to EU R&D managers and the specific channels used to find out about such research activities. Studying the frequency with which EU firms source knowledge from different regions (with respect to each method for learning about public research) reveals the geographic origin of the most useful research activities (Table 13.1). In the chemical industry, respondents obtained the results of research conducted by public research institutes or universities from the whole range of sources at levels similar to those for all the other industrial sectors. However, a few particularities are worth mentioning. Chemical firms obtain information from conferences in their own country and those in other European countries with the same frequency (about 88 per cent of
371 16.50 33.20 43.80 36.50 24.90
34.00 76.00 73.20 45.70 29.70
Authors’ elaboration of PACE data.
82.70 81.00 71.90 32.70
91.70 87.00 89.10 71.00
Publications and technical reports Conferences and meetings Informal personal contacts Hiring trained scientists and engineers Temporary exchanges of personnel Contract research done by the institute or university Joint research projects Affiliated firms Joint ventures
Source:
EU
19.70 33.80 16.30
11.10 20.30
77.50 66.20 59.40 19.80
NA
All industrial
Method of knowledge Knowledge Home acquisition source
Sector
5.10 11.10 12.20
4.80 4.00
54.80 36.70 31.90 4.40
Jap
72.90 39.60 28.10
28.10 79.20
90.60 87.50 89.60 75.00
Home
42.70 45.80 28.10
15.60 39.60
82.30 87.50 68.80 34.40
EU
28.10 50.00 20.80
14.60 30.20
84.40 67.70 66.70 33.30
NA
Chemicals
4.20 24.00 24.00
4.20 8.30
61.50 40.60 36.50 10.40
Jap
80.50 48.80 24.40
48.80 82.90
95.10 87.80 92.70 85.40
Home
48.80 51.20 17.10
31.70 58.50
92.70 92.70 87.80 61.00
EU
48.80 56.10 24.40
24.40 56.10
95.10 90.20 90.20 46.30
NA
Pharmaceutical Jap
14.60 26.80 22.00
4.90 17.10
68.30 53.70 58.50 19.50
Table 13.1 Sources of useful public research activities by geographic area and sector: percentage of respondents that reported obtaining results from the listed sources
372
The economics of knowledge
respondents). In the case of publications, informal contacts and hiring, respondents from the chemical industry attribute equal weights to other European countries and North America (exhibiting a lower EUlocalisation effect than all industrial sectors combined). In the pharmaceutical industry, the home country localisation effect tends to vanish. About 90 per cent of respondents obtained information through publications, informal contacts and conferences within their own country, other European countries and North America with only a small percentage receiving information from Japan. In seeking to source public research results EU R&D managers approach the North American science system, the EU and domestic sources with similar frequency. In particular, North American papers are used with the same frequency as home-country publications (95 per cent) and more frequently than papers from other EU countries (92 per cent). The behaviour of the pharmaceutical industry is of particular importance, if only because it is widely considered as one of the main areas of strength for the EU (Sharp et al., 1997). However, despite past successful performance, the EU industry seems to be rather pessimistic about the future (Sharp et al., 1997). Rising levels of R&D, decreasing profit margins and the struggle to refocus research efforts toward biotechnologies have been undermining the competitive position of the industry. United States pharmaceutical companies are often considered to be way ahead of their EU competitors, particularly with respect to the adoption of biotechnologies. It is often argued that the comparative success of the US industry is related to its capability to effectively transform the results of basic research into blockbuster drugs, rather than its ability to generate such results per se. In this respect, a European paradox is commonly evoked to emphasise the gap between the seemingly good performance of EU basic research and the relatively bad commercial performance of EU firms (when compared with their US competitors): European firms are not particularly good at transforming brilliant ideas into successful products. While most observers now agree on this last point, explanations of the reasons for this are in short supply. Generally speaking, the firms themselves are often blamed for not applying sensible management practices that would enable them to fully exploit the wealth of insights provided by the EU system of innovation and its national components. British firms are too short-sighted; German firms are too slow in making decisions and pursuing new research routes; French firms are sheltered from the pressures of the global economy by a complacent state-managed health system, and so on. While appealing, such propositions fail to consider the complementary possibility that there is something systematically different between the EU
The key characteristics of sectoral knowledge bases
373
and the US knowledge bases that enables US firms to be more competitive and induces EU firms to look to the US scientific knowledge base to source new knowledge. In what follows, we operationalise our framework to interpret the results of the PACE survey and assess the anecdotal evidence of the European paradox. 3.2 Mapping and Measuring Countries’ Sectoral Knowledge Bases We examined the publication profiles of different countries in the fields of chemistry and pharmacology. Following Geuna (2001) we used the Science Citation Index (SCI) database of the Institute for Scientific Information to analyse the publication output of the four largest European countries, the EU, Japan and the USA in the period 1989–96 (see Geuna, 2001, for a description of the data). Eleven scientific fields relevant to the chemical and pharmaceutical industries are identified.5 Each publication in these fields is classified in a typology of research using the CHI journal classification: Applied technology, Engineering and technological sciences, Applied research, and Basic research. Godin (1994), who studied a sample of large innovating firms in order to analyse the complementarities between science and technology, proposed a similar approach. He developed a database of publications that were then divided into four groups in a spectrum that varied from very applied to basic (‘untargeted’) research. Unlike this chapter, his work focused on firm-level activities, rather than sectors. To develop a comparative analysis of the knowledge base in chemistry and pharmacology, the relative specialisation of a country was studied. The symmetric Relative Specialisation Index (RSI) (see Appendix I for methodological issues concerning the RSI) is calculated on the basis of data from the SCI database for six countries and the EU, 11 scientific fields and three research areas between 1989 and 1996 (Balassa, 1965; Soete, 1981). The statistical results are used to operationalise the theoretical framework of knowledge specialisation for the pharmaceutical and chemical industries. 3.2.1 Knowledge persistence (stability of specialisation patterns over time) In the eight-year period under examination the specialisation of the EU and the six countries considered has changed, in some cases quite substantially. To verify the stability (or lack thereof) of overall specialisation patterns we examined how all 11 specialisation indices had changed over time. Following the work of Pavitt (1989), we calculated the Pearson correlation coefficient for each country at the start and at the end of the period considered. Positive and significant coefficients would hint at the cumulative
374
The economics of knowledge
and path-dependent nature of knowledge accumulation processes. We discovered that the knowledge specialisation in France, Germany, the UK, Japan and the USA is positively correlated in the two periods, while no significant correlation was found for Italy and the EU. Furthermore, in order to analyse the path of specialisation or despecialisation of a country, we regressed the symmetric specialisation index in 1996 on the 1989 value, country by country. Such a methodology was originally proposed by Cantwell (1989) and consists of a simple countryby-country regression at two different points in time. The dynamic path, therefore, cannot be studied. Also, nothing can be said about the determinants of the initial pattern of specialisation. Despite these limitations, this methodology has been widely used in specialisation studies. Its main advantage is its simplicity. If the coefficient is equal to 1, then the country specialisation pattern has remained unchanged over the period. If 1 then the country is increasing its positive specialisation in fields where it was already specialised. If 0 1, the country has decreased its non-specialisation in those fields where it was negatively specialised at the beginning of the period (or decreased its positive specialisation where it was positively specialised). In all cases, variations in specialisation occur in a cumulative way, as 0. In the case that is not significantly different from zero, the hypothesis that changes in specialisation are either not cumulative or are random cannot be excluded. If is negative we are witnessing a process of reversion in the specialisation. The case where 1 is often referred to as -specialisation (Dalum et al., 1998). Cantwell (1989: 31–2) argues that 1 is not a necessary condition for increasing specialisation. Therefore, we have also analysed the so-called s-specialisation (Dalum et al., 1998). The dispersion of a given distribution does not change if R; if R the specialisation increases (s-specialisation) and if R the specialisation decreases (s-specialisation). For each country we ran regressions on all fields, basic research only, applied research only, and applied technology and engineering research only. For the most general regressions, we found that Germany was the country with the most stable specialisation pattern ( .96, R .724). Italy and EU do not have significant coefficients. All the other countries de-specialised cumulatively (that is, 0 1) in terms of both and s-specialisation, with the UK ( .843, R .897) being the most cumulative, followed by the USA ( .779, R 852), Japan ( .659, R 926) and France ( .42, R 799). All coefficients are significant at the 1 per cent level (2 per cent for Germany). In terms of basic research, only the USA, Japan and the UK have coefficients with a significance level higher than 5 per cent – respectively 1.04 (1 per cent), 0.95 (2 per cent) and
The key characteristics of sectoral knowledge bases
375
0.44 (4 per cent). The USA, with both 1 and /R 1, increased specialisation in sectors where it was already specialised, and became less specialised where initially specialisation was low. Japan, with both 1 and /R 1, showed a high stability in its specialisation patterns. In particular, the USA deepened its specialisation in fields related to the pharmaceutical industry: medical chemistry (C6) and pharmacology (C10). The four largest European countries saw an increase in the dispersion of their basic research specialisation. European Union countries, especially Germany and France, show a tendency to remain more focused on traditional chemistry fields. 3.2.2 Knowledge integration (specialisation across research typology) Table 13.2 presents a summary of the relative specialisation of the EU and the six countries under consideration over the entire eight-year period. It lists, by type of research, the chemical fields in which each country exhibited positive specialisation (top two outside square brackets). The first observation that can be made is that there is some degree of overlap between the positive specialisation in applied research and that in basic research, while the area of applied technology and engineering tends to differ from the other two areas of research. If knowledge integration is defined as the presence of positive specialisation in the same scientific fields in the three typologies of research, it can be stated that the USA has a much Table 13.2
EU15 France Germany Italy UK USA Japan
Fields of positive specialisation by type of research Applied technology Applied research and engineering
Basic research
C10 C3 [C8] C8 C4 [C6 C10] C3 C4 [C11] C10 C6 [C8] C3 C10 C1 C2 [C9 C6 C10] C4 C11
C2 C5[C4 C8 C6 C7] C7 C4 [C8] C5 C4 [C8] C2 C6 [ C7 C5 C4] C6 C4[C5 C2] C6 C10 [C2 C8 C3 C7] C1 C10 [C7]
C4 C7 [C10 C8 C2 C1] C7 C4[C8 C6 C9] C1 C5[C4 C9 C8] C7 C6 [C10] C4 C7[C10 C3 C6] C6 C10 [C3 C1] C10 C9
Notes: Top two positive specialisation fields out of brackets. C1 General chemistry, C2 Analytical chemistry, C3 Applied chemistry, C4 Crystallography, C5 Inorganic and nuclear chemistry, C6 Medical chemistry, C7 Organic chemistry, C8 Physical chemistry, C9 Polymer science, C10 Pharmacology and pharmacy, C11 Chemical engineering. Source: Authors’ elaboration of ISI and CHI data.
376
The economics of knowledge
higher degree of knowledge integration than the EU. Indeed, the USA has a positive specialisation in medical chemistry (C6) and pharmacology and pharmacy (C10) in all research typologies. Among the four largest EU countries, France has positive specialisation in crystallography (C4), organic chemistry (C7) and physical chemistry (C8) in all three research typologies. Similarly, Germany is positively specialised in all typologies in crystallography (C4) and inorganic chemistry (C5). Finally, Italy is consistently positively specialised in medical chemistry (C6) and organic chemistry (C7).6 A simple indicator of integration is calculated by dividing the fields in which a country is positively specialised in all research typologies by the total number of fields in which a country is positively specialised. So, for each country we have: INT
no. of fields of positive specialisation in all types of research no. of fields of positive specialisation
This indicator varies from 0 to 1. It is 0 when the country considered does not exhibit any overlap between the three types of research. It is 1 when the country considered is fully integrated across all types of research in all the fields in which it exhibits positive specialisation. The USA is positively specialised in medical chemistry and pharmacology and pharmacy in all typologies of research, and positively specialised in a total of eight fields. Its indicator of integration is 2/8 0.25. France, Italy and Germany are less dispersed – that is, more integrated – than the USA. France scores 3/6 0.5 (where the fields of full integration are crystallography, physical chemistry and inorganic chemistry out of a total of six fields of positive specialisation). Italy and Germany are integrated in two fields (medical chemistry and organic chemistry versus crystallography and inorganic chemistry) out of a total of seven fields of positive specialisation, giving an integration coefficient of 2/7 0.29. Japan and the UK are not integrated at all and do not exhibit any overlap across the different typologies of research. 3.2.3 A taxonomy of knowledge specialisation By combining the results on persistence and integration, it is possible to map the science and engineering bases of different countries in a two-dimensional space that summarises the results sketched above. We have mapped the indicator of integration on the horizontal axis and the indicator of persistence on the vertical axis. For persistence, we have used the results of the regressions for all chemistry and pharmacology fields, first and last year, country by country. We have also used the coefficients for Italy and the EU although, as stated above, they are not significant. As all coefficients except
377
The key characteristics of sectoral knowledge bases
Knowledge integration Below average (< .22)
Above .5
*USA (0.25)
*Germany (0.29)
* Japan (0.00)
Knowledge persistence
* EU15 (0.22) Below .5
Figure 13.2
* UK (0.00)
Above average (> .22)
* Italy * France (0.29) (0.5)
Integration and persistence: a rough-and-ready map
those for Germany (whose equals 1) are 0 1 (-de-specialisation), we set 0.5 as the threshold. For integration, we have used the simple indicator sketched above. The threshold between high and low integration is given by the arithmetic average of the indicator (0.22). Figure 13.2 reports the result of such a combination. It is fairly apparent that the USA and Germany combine high levels of both integration and persistence. France, despite a high level of integration, exhibits low persistence over time. Neither Japan nor the UK shows any integration, but the pattern of specialisation in the UK is more stable. Italy and the EU are somewhere in between. The EU as a whole is characterised by both average integration and low persistence (this latter coefficient was not significant in the regression). Italy appears to be relatively integrated, but exhibits low persistence (Italy’s coefficient for persistence is not significant). It is worth combining the results of this taxonomy with the analysis of the specific fields of specialisation listed in Table 13.2. Despite the high persistence and integration exhibited by both the USA and Germany, their specialisation profiles appear to be very different. In particular, Germany’s specialisation revolves around traditional chemistry fields, such as crystallography (C4) and inorganic chemistry (C5). The USA is specialised in those fields more directly related to pharmaceuticals: medical chemistry (C6) and pharmacology and pharmacy (C10). The other EU countries studied also are more specialised in ‘chemistry for chemicals’, rather than pharmaceuticals. Furthermore, it is evident from the regressions we ran by
378
The economics of knowledge
type of research that the EU countries’ specialisation in medical chemistry and pharmacology decreases as we move away from development type research towards applied and then basic research. Such results are consistent with other studies of specialisation that rely on traditional methodologies. So, for Germany, specialisation in ‘traditional’ chemistry (that is, inorganic and organic) is confirmed by Sternberg (2000, p. 98) who also highlights the German disadvantage in medical sciences. The Office of Science and Technology (OST, 1998) confirms both the integration of the German pattern of specialisation and its focus on chemistry. Furthermore, the UK seems to be more specialised in medical research than France and Germany. The OST (1998) also confirms the strong EU specialisation in chemistry and its relative disadvantage (in terms of publications) in biology (basic research). These different specialisation profiles hint at a possible explanation for the results of the PACE questionnaire. The PACE survey revealed that public research carried out in North America was valued and used extensively (even more than public research carried out in other European countries) by the largest EU R&D firms in the pharmaceutical sector. The PACE questionnaire does not allow speculation about why this happens, though. We argue that the reliance of EU firms on the North American knowledge base is consistent with the fact that the USA exhibits a persistent as well as an integrated specialisation pattern in medical chemistry and pharmacy and pharmacology. The results for the chemical industry confirm this. European Union chemical firms do not use US-generated research to the same extent as pharmaceutical firms. Their home-country knowledge base is relatively more specialised in a persistent and integrated manner in those fields that are particularly relevant to the innovative efforts of the chemical industry. Thus, they rely heavily on the public research of their own country or other European countries. Particular attention should be devoted to specialisation by type of research in EU countries. It was noted above that they are positively specialised in either medical chemistry or pharmacology at the level of applied and development research. However, those two fields do not show up as areas of positive specialisation in basic research (Table 13.2). Also, the results of the regression by type of research clearly show that only the USA and Japan are increasing their specialisation in basic research. No clear pattern is discernible for EU countries except for the UK, which is -despecialising. Therefore, these data do not allow us to talk about a ‘European paradox’, according to which EU firms would not be capable of exploiting an efficient basic research system because of lack of ‘development’ capabilities. Our data seem to point to the fact that these types of capabilities do exist. What is missing is the basic research bit, with the result
The key characteristics of sectoral knowledge bases
379
that EU pharmaceutical firms have to source research results from the USA. The pattern of sourcing is consistently different when chemical firms are considered, as their home-country knowledge bases seem more capable of providing basic research capabilities. Despite the limitations of the data and the simplicity of this analysis, the location of different countries along the grid defined by the measures of persistence and integration matches with a few things we know about the institutional structure of each country, and also raises some interesting questions. For instance, the results concerning the 15 countries of the EU as a whole are hardly surprising. An EU-wide system of innovation is still in the process of formation. National industry and science and technology (S&T) policies still heavily influence country-level specialisation patterns, preventing them from converging toward a homogeneous whole.
4
CONCLUSIONS
The evolution of country-level sectoral specialisation has been conceptualised by the discussion on knowledge persistence and knowledge integration. Persistence is related to the evolution of specialisation over time. It hints at the cumulative, path-dependent nature of the learning processes. Integration is related to the evolution of specialisation across different typologies of research. It suggests the complex, non-linear interdependencies that link the scientific and technological domains. The interaction of the concepts of knowledge persistence and knowledge integration provide a head start for the development of a robust conceptual framework in which to compare countries’ sectoral knowledge bases. It is quite significant that the conceptualisation proposed in terms of persistence is consistent with the results of micro-studies of technical change that pinpoint learning processes as cumulative and path dependent (David, 1985). Also, innovation studies hint at the key role played by distributed (rather than narrowly focused) capabilities in enabling technical change (Granstrand et al., 1997). This is captured by the concept of knowledge integration. This chapter represents a first attempt to operationalise this framework on the basis of a statistical analysis of a huge, original and custom-built data-set that describes the scientific and engineering knowledge base in chemistry and pharmacology in the four largest European countries, the EU as a whole, the USA and Japan during the period 1989 to 1996. Analysis of the relationships between core positive and negative specialisation, and of the typology of research (Applied technology and engineering, Applied research and Basic research) has shown that the countries considered have different degrees of knowledge integration and knowledge
380
The economics of knowledge
persistence. Specifically, the USA and Germany exhibit the highest coefficients of persistence and integration. However, the USA is more heavily specialised in fields related to pharmaceuticals (that is, medical chemistry and pharmacology and pharmacy) than Germany and the other EU countries, which appear to be more specialised around traditional chemistry. These results are consistent with the views expressed by the EU R&D managers that responded to the PACE questionnaire. They stressed that public research developed in North America was particularly useful to their innovative efforts in pharmaceuticals. In contrast, domestic and EU localisation effects prevail in the case of the chemical industry. As for the policy implications, the empirical results presented (although preliminary) allow us to make two main observations. First, our data-set does not identify any ‘European paradox’ in pharmacology. European Union countries exhibit capabilities in terms of applied and engineering research, but not in basic research. Instead, the USA only increases its -specialisation in basic research in pharmacology and medical chemistry. No clear pattern is discernible for EU countries, with the exception of the UK, which is -despecialising. Such lack of basic research capabilities may well explain the frequency with which EU R&D managers in the pharmaceutical industry approach the US knowledge base. For chemicals, the pattern of sourcing is different. As their home-country knowledge bases seem more capable of providing a more integrated pattern of research capabilities, EU chemical firms rely chiefly on their home-country knowledge base and then approach that of the EU. At least for pharmacology and medical chemistry we found no evidence of paradoxes. Second, our approach hints at the possibility that government can actually influence the rate of technical change by fostering the development of an ‘integrated’ specialisation profile. Empirically, one can identify the NSI that firms consider to be more helpful to their innovative activities (for example, the USA for pharmaceuticals), analyse it in terms of integration and then target the type of research that is lacking in the home country. We may call this the ‘policy for integration’ option. In fact, despite the enormous resources devoted by policy-makers to the exploration of emerging technologies, ‘picking a winner’ remains a rather hazardous activity. The greatest successes of recent years are the unintended consequences of policies aimed at fostering other paths of research – for example, biotechnology being the unintended offspring of US cancer research programmes and the beneficiary of military research for the bio-war (Martin et al., 1990). Which specific scientific field will be responsible for the next revolution continues to be difficult to predict. We argue that our approach would not allow governments to pick the winners, but would allow them to support the development of an integrated knowledge base once a new path has emerged.
The key characteristics of sectoral knowledge bases
381
The limitations of this work open up a number of challenging questions for future research. First, publications are a very good way to trace the scientific knowledge base of a country, but are less successful as far as engineering research is concerned. Merging traditional data-sets of patent activities with our data-set of publications would provide a better picture of the interaction between scientific and engineering specialisation. Second, the concepts of knowledge integration and persistence can also be applied to the study of firms’ knowledge bases to further confirm the consistency between micro- and macro-level dynamics. It is important to expand such analysis to test the existence of a correlation or causation between knowledge integration and knowledge persistence in certain fields on the one side, and technological and economic performances of firms and countries, on the other.7 The qualitative indications provided by the PACE questionnaire are but a first step. Third, this analysis should be extended to a sample of ‘small countries’. These may be much less integrated and persistent than large countries as they may find it more convenient (or just more feasible) to exploit the advantages of flexibility by specialising narrowly in terms of fields and/or types of research and then switching when new research trajectories emerge. Finally, on a more theoretical note, PACE reveals that firms can source knowledge not available in their home country by looking abroad. However, there are costs attached to such a choice. Traditionally, costs are related to the geographic distance between source and user. This chapter hints at the possibility that there might be costs attached also to the relative position in the ‘knowledge spectrum’, so that the farther from a typology of research the more expensive it will be to develop knowledge exchange.
APPENDIX 1: METHODOLOGY The symmetric Relative Specialisation Index (RSI) is given by the ratio between the share of the given scientific field in the publication of the given country and the share of the given scientific field in the world total of publications (activity index – AI) minus one, divided by AI plus one. It may take values in the range [ 1, 1]. It indicates whether a country has a higherthan-average activity in a scientific field (RSI1) or a lower-than-average activity (RSI 1).
AI
⁄ ⁄ pij
j
pij
i
pij
ij
pij
(13.1)
382
The economics of knowledge
where p number of publications, i 1 . . . n number of scientific fields
11 and j 1 . . . m number of countries 7. RSI
AI 1 AI 1
(13.2)
As the denominator of AI is the share of the given scientific field in the world total of publications, the number and choice of the countries in the comparative analysis does not influence the robustness of this indicator.
NOTES * 1.
2. 3.
4.
5.
6.
A version of this chapter has appeared in Research Policy, 32 (2003). Comments and suggestions made by Anthony Arundel, Dominique Foray, Ben Martin, Richard Nelson, Ammon Salter, Ed Steinmueller and Robert Tijssen are gratefully acknowledged. The participants in the New Frontiers in the Economics of Innovation and New Technology: A Conference in Honour of Paul A. David and the Sixth International Conference on Science and Technology Indicators conferences provided insightful comments. We are grateful to Anthony Arundel for the PACE data, and to Diana Hicks for the CHI classification. The financial support of the Commission of the European Communities, TSER project ‘From Science to Products’ and IST project ‘NewKind – New Indicators for the Knowledge-based Economy, no. IST 1999–20782’ is acknowledged. We are indebted to Ed Steinmueller for the development of the discussion on the cumulativeness of science. The PACE questionnaire surveyed the largest R&D performing industrial firms in 1993 in 12 of the EU countries. The responses are from 414 large manufacturing firms across nine EU countries (Belgium, Denmark, Germany, Ireland, Italy, Luxembourg, the Netherlands, Spain and the UK). Before proceeding, an important caveat related to the use of publications as a descriptor of a knowledge base needs to be discussed. We fully acknowledge that by adopting peer-reviewed publications as a descriptor of a country’s knowledge base we limit our analysis to the most codified (and codifiable) bits of this base. This limitation is determined by obvious data constraints (tacit knowledge is rather difficult to capture ‘alive’), and also by the responses to the PACE questionnaire, which pinpoint scientific papers as a key mechanism to locate relevant sources of knowledge. Hicks (1995) thoroughly discusses the role of scientific papers as signals of information about the presence of valuable ‘hidden’ tacit skills. This chapter considers publications as elements of a signalling system whose morphological characteristics reveal something of the deeper structure of a country’s sectoral knowledge base. Publications would be a sort of observable ‘sufficient statistics’ of the underlying unobservables. Finally, as publications represent a preliminary and incomplete proxy of the knowledge base, more inclusive indicators or combinations of indicators should be developed in the future to operationalise the interpretative framework. C1: general chemistry, C2: analytical chemistry, C3: applied chemistry, C4: crystallography, C5: inorganic and nuclear chemistry, C6: medical chemistry, C7: organic chemistry, C8: physical chemistry, C9: polymer science, C10: pharmacology and pharmacy and C11: chemical engineering. A problem emerged with respect to inorganic chemistry (C5) and organic chemistry (C7). For these fields, no publications are recorded in applied technology and engineering. Thus, for these two fields, we considered as integrated those countries that exhibits positive specialisation in applied research and basic research only.
The key characteristics of sectoral knowledge bases 7.
383
A first attempt in this direction has been made within the project NewKind. All results, deliverables and data sources are available at http://www.researchineurope.org/ newkind/index.htm (accessed on 21 March 2005).
REFERENCES Archibugi, D. and M. Pianta (1992), The Technological Specialisation of Advanced Countries, Dordrecht, Kluwer Academic. Arundel, A. and A. Geuna (2001), ‘Proximity and the use of public science by innovative European firms’, Economics of Innovation and New Technology, 13 (6), 559–80. Arundel, A., G. van de Paal and L. Soete (1995), Innovation Strategies of Europe’s Largest Industrial Firms: Results of the PACE Survey: 23, Brussels, European Innovation Monitoring System, European Commission. Balassa, B. (1965), ‘Trade liberalization and “revealed” comparative advantage’, The Manchester School of Economic and Social Studies, 32, 99–123. Braun, T., W. Glänzel and H. Gruup (1995), ‘The scientometric weight of 50 nations in 27 science areas. 1989–1993. Part I All fields combined, mathematics, engineering, chemistry and physics’, Scientometrics, 33, 263–93. Cantwell, J.A. (1989), Technological Innovation and Multinational Corporations, Oxford, Basil Blackwell. Cantwell, J.A. (1995), ‘Multinational corporations and innovatory activities: towards a new evolutionary approach’, in J. Molero (ed.), Technological Innovation, Multinational Corporations and New International Competitiveness, Chur, Harwood, pp. 21–57. Carlsson, B. (ed.) (1997), Technological Systems and Industrial Dynamics, Boston, MA, Kluwer Academic. Conant, J.B. and L.K. Nash (1964), Harvard Case Histories in Experimental Science, vol. 1, Cambridge, MA, Harvard University Press. Dalum, B., K. Laursen and G. Villumsen (1998), ‘Structural change in OECD export specialisation patterns: de-specialisation and “stickiness” ’, International Review of Applied Economics, 12, 447–67. Dasgupta, P. and P.A. David (1994), ‘Towards a new economics of science’, Research Policy, 23, 487–521. David, P.A. (1985), ‘Clio and the economics of QUERTY’, American Economic Review, 75, 332–37. David, P.A. and D. Foray (1995), ‘Accessing and expanding the science and technology knowledge base’, STI Review, 16, 13–68. European Commission (1996), Green Paper on Innovation, Brussels: European Commission. European Commission (1997), Second European Report on S & T Indicators 1997, Brussels, European Commission. Fagerberg, J., P. Guerrieri and B. Verspagen (eds) (1999), The Economic Challenge for Europe, Cheltenham, Edward Elgar. Gambardella, A. (1995), Science and Innovation: The US Pharmaceutical Industry During the 1980s, Cambridge, Cambridge University Press. Geuna, A. (2001) ‘The evolution of specialisation: public research in the chemical and pharmaceutical industries’, Research Evaluation, 10, 67–79.
384
The economics of knowledge
Godin, B. (1994), ‘The relationship between science and technology’, unpublished DPhil thesis, University of Sussex, Brighton. Granstrand, O., P. Patel and K. Pavitt (1997), ‘Multi-technology corporations: why they have “distributed” rather than “distinctive core” competencies’, California Management Review, 39, 8–25. Hicks, D. (1995), ‘Published papers, tacit competencies and corporate management of public/private character of knowledge’, Industrial and Corporate Change, 4, 401–24. Iansiti, M. (1998), Technology Integration: Making Critical Choices in a Dynamic World, Boston, MA, Harvard Business School Press. Landau, R. and N. Rosenberg (1992), ‘Successful commercialization in the chemical process industries’, in N. Rosenberg, R. Landau, and D.C. Mowery (eds), Technology and the Wealth of Nations, Stanford, CA, Stanford University Press. Lundvall, B.A. (ed.) (1992), National Systems of Innovation. London: Pinter. Malerba, F. and L. Orsenigo (1997), ‘Technological regimes and sectoral pattern of innovative activities’, Industrial and Corporate Change, 6, 83–118. Malo, S. and A. Geuna (2000), ‘Science-technology linkages in an emerging research platform: the case of combinatorial chemistry and biology’, Scientometrics, 47, 303–21. Mansfield, E. (1991), ‘Academic research and industrial innovation’, Research Policy, 20, 1–12. Martin, B.R., J. Irvine and P.A. Isard (1990), ‘Trends in UK government expenditure on academic and related research: a comparison with the Federal Republic of Germany, France, Japan, the Netherlands and United States’, Science and Public Policy, 17, 3–13. Meliciani, V. (2001), Technology, Trade and Growth in OECD Countries. Does Specialisation Matter? London, Routledge. Merton, R.K. (1965), On the Shoulders of Giants – a Shandean Postscript, New York, Free Press. Narin, F., K.S. Hamilton and D. Olivastro (1997), ‘The increasing linkage between U.S. technology and public science’, Research Policy, 26, 317–30. Nelson, R.R. (ed.) (1993), National Systems of Innovation: A Comparative Study, Oxford, Oxford University Press. Office of Science and Technology (OST) (1998), Science & Technologie Indicateurs. Paris: Economica. Orsenigo, L. (1989), The Emergence of Biotechnology: Institutions and Markets in Industrial Innovation, London, Pinter. Patel P. and K. Pavitt (1994), ‘Uneven (and divergent) technological development amongst advanced countries: evidence and a framework of explanation’, Industrial and Corporate Change, 13, 759–87. Patel, P. and K. Pavitt (2000), ‘National systems of innovation under strain: the internationalisation of corporate R&D’, in R. Barrell, G. Mason and M. O’Mahoney (eds), Productivity, Innovation and Economic Performance, Cambridge, Cambridge University Press. Pavitt, K. (1989), ‘International patterns of technological accumulation’, in N. Hood and J.E. Vahlne (eds), Strategies in Global Competition, London, Croom Helm. Pavitt, K. (1992), ‘The key characteristics of the large innovating firm’, in G. Dosi, M. Giannetti and L. Toninelli (eds), Technology and Enterprise in Historical Perspective, Oxford, Clarendon Press.
The key characteristics of sectoral knowledge bases
385
Pavitt, K. (1998), ‘Technologies, products and organisation in the innovating firm: what Adam Smith tells us that Schumpeter doesn’t’, Industrial and Corporate Change, 7 (3), 433–52. Pianta, M. and D. Archibugi (1991), ‘Specialisation and size of scientific activities: a bibliometric analysis of advanced countries’, Scientometrics, 22, 341–58. Pisano, G.P. (1996), ‘Learning-before-doing in the development of new process technology’, Research Policy, 25, pp. 1079–1119. Pisano, G.P. (1997), The Development Factory: Unlocking the Potential of Process Innovation, Boston, MA, Harvard Business School Press. Porter, M.E. (1990), The Comparative Advantage of Nations, New York, Free Press. Prencipe, A. (1997), ‘Technological competencies and a product’s evolutionary dynamics: a case study from the aero-engine industry’, Research Policy, 25, 1261–76. Price, D. de S. (1963), Little Science, Big Science, New York, Columbia University Press. Rosenberg, N. (1994), Exploring the Black Box: Technology, Economics and History, Cambridge, Cambridge University Press. Sharp, M., P. Patel and K. Pavitt (1997), Europe’s Pharmaceutical Industry: An Innovation Profile, Brussels, report prepared for DG XIII–D-4. Soete, L.L.G. (1981), ‘A general test of the technological gap trade theory’, Weltwirtschaftliches Archiv, 117, 638–66. Sternberg, R. (2000), ‘University–industry relationships in Germany and their regional consequences’, in Z.J. Acs (eds), Regional Innovation, Knowledge and Global Change, London, Pinter. Tijssen, R.J.W. and E. van Wijk (1999), ‘In search of the European paradox: an international comparison of Europe’s scientific performance and knowledge flows in information and communication technologies research’, Research Policy, 28, 519–43. Vincenti, W.G. (1990), What Engineers Know and How They Know: Analytical Studies from Aeronautical History, Baltimore, MD, and London, Johns Hopkins University Press.
PART IV
The Diffusion of New Technologies
14. Uncovering general purpose technologies with patent data1 Bronwyn H. Hall and Manuel Trajtenberg 1
INTRODUCTION
In ‘The computer and the dynamo,’ Paul David (1991) makes a persuasive case for considering the process by which the electric dynamo spread throughout the economy during the turn of the twentieth century and the process by which the use of information technology (specifically, computing technology) is currently being spread throughout different industries as similar manifestations of the diffusion of ‘general purpose technologies,’ a term introduced into the economics literature by Bresnahan and Trajtenberg (1995). All these authors, as well as Helpman and Trajtenberg (1998a; 1998b), emphasise the singular contribution to economic growth made by this type of technology, because of its ability to transform the means and methods of production in a wide variety of industries. At the same time and using historical data, David (1990; 1991), Rosenberg (1976), and others have argued that the diffusion of these technologies throughout the economy may take decades rather than years because of co-ordination problems and the need for complementary investments (both tangible and intangible) in using industries. For this reason it may take some time for the benefits of the technologies to be manifest in economic growth. On the theoretical side, Bresnahan and Trajtenberg (1995) have studied the non-optimality of innovation and diffusion when a decentralised market system is called upon to try to solve the co-ordination problem between technology-innovating and technology-using industries. However, there has been relatively little empirical and econometric work that incorporates the insights of these various authors to analyse specific technologies. Our modest goal in this chapter is to see what might be learned about the existence and technological development of general purpose technologies (GPTs) through the examination of patent data, including the citations made to other patents. Such measures would be useful both to help identify GPTs in their early stages of development and also as proxies for the 389
390
The diffusion of new technologies
various rates of technical change called for in a fully developed growth model such as that in Helpman and Trajtenberg (1998b). In doing this exploration we are also motivated by the observation that not all technologies or, indeed, R&D dollars are equal, but that economists too often ignore that fact, primarily because of data limitations. As has been pointed out by others before us patenting measures have the potential to allow more detailed analysis of the ‘direction’ as well as the ‘rate’ of technical change.2 Although such an exploration might be made using data from a variety of countries, our focus here is on the use of US patent data, where the citations have a well-defined meaning and also where they have been computerised since 1977, enabling us to work with them relatively easily. Given the importance of the USA as a locus of technical change in the late twentieth century, we do not feel that this limitation to US patenting activity is a serious drawback for a preliminary investigation of this kind. We begin with the definition (description) of GPTs offered by Helpman and Trajtenberg (1998a): 1.
2.
3.
They are extremely pervasive and used in many sectors of the economy. Historical examples are the steam engine and the electric dynamo (the engine of electrification). Contemporary examples are the semiconductor and perhaps the Internet. Because they are pervasive and therefore important, they are subject to continuous technical advance after they are first introduced, with sustained performance improvements. Effective use of these technologies requires complementary investment in the using sectors; at the same time, the GPT enhances the productivity of R&D in the downstream sector. It is these points that are emphasised by David.
Using this definition, the contribution of the effort described here is to define measures using patents and citations that quantify the insights of David and Trajtenberg and their co-authors. Our study is subject to a variety of limitations, however. First, it is based on patent data, which provides imperfect coverage of innovative activity, as not all innovations are patented or patentable.3 Second, it relies heavily on the US Patent and Trademark Office (USPTO) classification system for technology, treating each three-digit patent class as roughly comparable to the ‘size’ of a technology. An examination of the classes suggests that this is unlikely to be strictly true (for example, the chemistry of inorganic compounds is a single class, whereas there are multiple optics classes). Making use of the sub-classes to refine the class measures would be a formidable
Uncovering general purpose technologies with patent data
391
task, because sub-classes are spawned within the three-digit class ad libitum and may descend either from the main class or from another sub-class. Thus some sub-classes are more ‘important’ than others, but this fact has to be uncovered by a tedious search of the text on the USPTO website. Rather than attempting to construct our own classification system in this way, we chose to look at measures based on the primary International Patent Classification (IPC) class of the patent.4 Finally, because patent citation data is only available in computerised form in 1975, and because severe truncation due to the application-grant lag and the citation lag sets in by around 1995 our period of study is necessarily fairly short and emphasises the 1980s and 1990s. However, truncation in the later years means that we are unable (yet) to explore fully the implications of changes in information processing technology during the very recent past. Because of time and resource constraints, we have focused the large part of our analysis on an extremely small subset of nearly 3 million patents available to us, the 780 most highly cited patents that were granted between 1967 and 1999. There is considerable evidence that the value or importance distribution of patents is highly skewed, with most patents being unimportant and a few being highly valuable.5 We expect that one reason for this finding is that true GPT patents are concentrated among the highly cited patents, so the current endeavour is centred on those patents which represent the extreme tail of a very skew distribution. In order to understand how patent data might help us identify GPTs and explore their development and diffusion, it is necessary first to understand something more about patent citations. This is the subject of the next section. We then discuss the GPT-related measures we have constructed from the patent data, and show how our sample of highly cited patents differs in various dimensions from the population as a whole.
2
PATENT CITATIONS6
A key data item in the patent document is ‘References Cited – US Patent Documents’ (hereafter we refer to these just as ‘citations’). Patent citations serve an important legal function, since they delimit the scope of the property rights awarded by the patent. Thus, if patent B cites patent A, it implies that patent A represents a piece of previously existing knowledge upon which patent B builds, and over which B cannot have a claim. The applicant has a legal duty to disclose any knowledge of the ‘prior art’, but the decision regarding which patents to cite ultimately rests with the patent examiner, who is supposed to be an expert in the area and hence to be able to identify relevant prior art that the applicant misses or conceals.
392
The diffusion of new technologies
The presumption is thus that citations are informative of links between patented innovations. First, citations made may constitute a ‘paper trail’ for spillovers, that is, the fact that patent B cites patent A may be indicative of knowledge flowing from A to B; second, citations received may be telling of the ‘importance’ of the cited patent.7 The following quote provides support for the latter presumption: the examiner searches the . . . patent file. His purpose is to identify any prior disclosures of technology . . . which might be similar to the claimed invention and limit the scope of patent protection . . . or which, generally, reveal the state of the technology to which the invention is directed. If such documents are found . . . they are ‘cited’ . . . if a single document is cited in numerous patents, the technology revealed in that document is apparently involved in many developmental efforts. Thus, the number of times a patent document is cited may be a measure of its technological significance. (OTAF, 1976: 167)
The aspect of citations that is important for the present effort is that they provide a record of the link between the present invention and previous inventions. Thus they can tell us both the extent to which a particular line of technology is being developed (if they are made to patents in the same technology area) and whether a particular invention is used in wide variety of applications (if they are made to patents in different technology areas). In principle, given that we know which firm owns the relevant patents, it is possible to ask these question both using the technology field, which is a classification made by the USPTO,8 and using the industry in which the patent falls, as indicated by the firm to which it is assigned.
3
MEASURES OF GPTS
The definition of GPTs paraphrased in the introduction suggests that the following characteristics (observations) apply to the patents associated with GPT innovations: (1) they will have many citations from outside their particular technology area or perhaps from industries outside the one in which the patented invention was made; (2) they will have many citations within their technology area, and the citations will indicate a pattern of cumulative innovation, or trace out a technology trajectory; (3) more speculatively, citing technologies will be subject to a burst of innovative activity as complementary goods are developed; and (4) given the length of time it takes for a GPT to pervade the economy, citation lags for patents in this area may be longer than average. In this section we report on the construction of a number of proxies for these characteristics. We use these proxies to identify patents that are in the extreme tail of the distribution of
Uncovering general purpose technologies with patent data
393
patent characteristics, in an effort to identify some candidate GPTs. Not surprisingly, we find that looking at a single characteristics may be misleading, so in the later sections of the chapter we use a more multivariate approach to refine the analysis. It is well known that the distribution of patent values and patent citations is very skewed with almost half of all patents receiving zero or one cite and less than 0.1 per cent receiving more than 100 cites (see Hall et al. 2005, for evidence on both points). Observations (1) and (2) above also suggest that GPT patents are likely to be highly cited. Therefore, we began our investigation by focusing on highly cited patents. We identified these patents by requiring that the number of citations the patent received be greater than three times the number received by the patent in the 99th percentile of the distribution. The results of this selection process are shown in Table 14.1. It yielded 780 patents granted between 1967 and 1999 that were ultimately granted, together with the name and type of their assignee (owner), the three-digit patent classification, and similar information on all the patents issued between 1975 and 2002 that cited this patent, for a total of 159 822 citations. Table 14.1 also makes it clear how skewed the citation distribution is: our sample of 780 patents is about one out of 3700 patents, whereas the 160 000 citations are one out of 100 citations (there are 16 million citations in all). Thus our patents are 37 times more likely to be cited than predicted by the average probability. 3.1
Generality
Observation (1) above suggests the use of a measure that is similar to the Trajtenberg, Jaffe and Henderson ‘generality’ measure, which is defined in the following way: Generality Gi 1
ni
s
2 ij
j
where sij denotes the percentage of citations received by patent i that belong to patent class j, out of ni patent classes (note that the sum is the Herfindahl concentration index). Thus, if a patent is cited by subsequent patents that belong to a wide range of fields the measure will be high, whereas if most citations are concentrated in a few fields it will be low (close to zero).9 Observation (2) suggests that even if generality is relatively high, the absolute number of citations should also be high, implying that there may still be a large number of citations in the patent’s own technology class. It also suggests that ‘second-generation’ citations be examined. We implement this using two variables, the average number of citations to the citing patents and the average generality of the citing patent.
394
The diffusion of new technologies
Table 14.1
Selecting the sample of highly cited patents
Grant year
Cutoff no. citations
1967 1968 1969 1970 1971 1972 1973 1974 1975 1976 1977 1978 1979 1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999
78 84 87 93 99 108 111 117 126 132 135 138 141 144 147 150 156 159 159 168 180 177 177 177 171 174 174 171 156 141 114 90 63
All years
Highly cited patents
All patents
Highly cited share (%)
Number
Median cites
Number
Median cites
20 15 28 19 22 30 25 24 19 22 17 16 7 26 17 19 20 22 25 21 29 29 27 27 38 41 33 26 21 13 28 33 21
103.0 131.0 115.0 105.0 165.0 136.0 132.0 136.5 152.0 156.5 182.0 179.0 190.0 204.5 177.0 232.0 224.5 188.5 199.0 204.0 213.0 252.0 258.0 215.0 204.5 218.0 231.0 214.5 178.0 171.0 136.0 103.0 69.0
65 652 59 104 67 559 64 429 78 317 74 810 74 143 76 278 73 690 72 015 66 883 67 862 50 177 63 371 67 373 59 462 58 435 69 338 73 824 72 977 85 522 80 345 98 567 93 290 99 789 100 760 100 980 104 317 104 091 112 832 115 337 151 745 153 486
2 2 3 3 3 4 4 5 5 5 5 5 5 5 5 5 5 5 5 6 6 6 5 5 5 5 4 4 4 3 3 2 1
0.030 0.025 0.041 0.029 0.028 0.040 0.034 0.031 0.026 0.031 0.025 0.024 0.014 0.041 0.025 0.032 0.034 0.032 0.034 0.029 0.034 0.036 0.027 0.029 0.038 0.041 0.033 0.025 0.02 0.012 0.024 0.022 0.014
780
183.0
2 756 760
3
0.028
Note: *Patents with zero cites 1975–2002 are excluded.
Uncovering general purpose technologies with patent data
395
In actual measurement, the preceding two predictions interact in ways which make our task a bit more complex. Like patents, citation counts are a discrete random variable bounded below by zero. This means that fewer citations in total imply that fewer classes will be observed to have citations than should be observed were the total number of citations larger. That is, ni is biased downward by the fact that fractional citations are not observed, and generality will tend to be lower when there are fewer citations. This is quite visible in the graph of average generality over time shown in Figure 14.1, where we show two different versions of generality, one based on US patent classes, and another based on the International Patent Classification, as assigned to these patents by the USPTO. Note that the average of either generality measures begins to decline fairly steeply in 1993–95, at the same time as our measure of average citations per patent turns down sharply due to the effects of lag truncation (see Figure 14.2). In this case, this is a spurious rather than real decline in generality, owing to the fact that our patent grant data ends in 2002, and therefore our application-dated data ends around 1999, so that patents in the years after about 1994 have had less chance to receive citations.10 Using a simple binomial model of the probability of observing a citation in a given cell, Hall (2002) shows that an unbiased estimate of the generality of the ith patent can be computed using the following correction: Gi
Ni G Ni 1 i
where Ni is the number of citations observed. Note that this measure is not defined when Ni 2 and will be fairly noisy when Ni is small. We have used this bias correction for the first three generality measures described below. The US patent classification system has grown over time in ways that make it not ideal for the purpose we have in mind here. Generality measures essentially assume that all categories are equidistant from each other if they are to be compared, but this is not the case for the US patent class system. Therefore, we explore the use of generality measures based on five different classification systems: 1 2 3 4
5
US patent class (approximately 400 cells). Hall–Jaffe–Trajtenberg technology subcategories (36 cells). Main International Patent Class (approximately 1200 cells). Industry classification based on Silverman’s IPC-SIC concordance (Silverman, 2002) for industry of manufacture, aggregated to the Hall– Vopel (Hall and Vopel, 1997) level (37 cells). Industry classification based on Silverman’s IPC-SIC concordance for industry of use, aggregated to Hall–Vopel level (37 cells).
396
1979
1981
1983
1993
1995
1997
Generality (IPC class)
Average measures of generality and originality
Generality (US class)
Cites per patent (as of 1996)
Cites per patent (as of 2002)
0.00 1999
1.00
0.05
1985 1987 1989 1991 Patent application year
2.00
0.10
0.00 1977
3.00
0.15
4.00
0.20 Cites/pat (1996)
5.00
0.25
6.00
7.00
0.35
0.30
8.00
0.40 Cites/pat (2002)
9.00
0.45 Generality
10.00
Figure 14.1
Index
0.50
Cites/patent
397
0.0
0.1
0.2
0.3
0.4
0.5
0.6
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 Citation lag
Figure 14.2 Citation lag distribution (1976–94). Trajtenberg, Jaffe and Henderson Methodology
Relative citation probability
398
The diffusion of new technologies
The rationales for these various choices are the following. First, in addition to using the US patent classification system (measure 1), we also constructed generality based on the more equal groupings of technologies constructed by Hall et al. (2002) from the patent classes (measure 2) and from the International Patent Classification system main four-digit classes (measure 3), which is more detailed that the US patent classification system. Second, it could be argued that a GPT is not likely to manifest itself as a series of citations by patents in different technology classes, but rather as citations by firms in different industries. To consider this possibility, we would like to base a measure on the shares of citations that come from firms in different industries at the roughly two and one-half digit level. That is, our fourth measure is a Herfindahl for patent citation dispersion across industries rather than across technologies. Based on the discussion of GPT diffusion to using industries in the introduction, an industry-based measure would seem to be intrinsically preferred for this exercise. There are basically two ways to construct such a measure: the first uses the industry of ownership of the patents based on identification of the patent assignees and the second determines the industry for each patent class/sub-class from some type of industry-patent class concordance. The first approach is difficult to implement in practice, given the number of patents that are unassigned and the number of assignees that are not identifiable, either because they are small firms or because they are foreign firms for which we do not yet have a match to other data sources.11 Therefore we used the SIC-technology class concordances of Silverman (2002) to assign these patents and their citations to industries of manufacture and of use. Then we collapsed the distribution of citations by SIC codes into a 37-element vector of industries using the SIC-industry correspondence given in Appendix Table A1, and used this vector to construct the generality measures 4 and 5. The computation of these measures was able to make use of all the patents rather than just those held by US industry.12 One major drawback of using the Silverman concordance ought to be mentioned, especially in the light of our subsequent findings: it is based on assignments to industry of manufacture and use made by Canadian patent examiners between 1990 and 1993. This means that it will do a poor job on patents in technologies related to the growth of the Internet and software, because there were unlikely to be many of these in the Canadian patent system prior to 1994. Figures 14.3 and 14.4 show the distribution of two of the computed generality indices for the highly cited patents, Figure 14.3 the US patent class index and Figure 14.4 the index based on Silverman’s industry of manufacture map. As the figures show, these indices range from zero to one and the measure based on the industry of manufacture has a somewhat
Uncovering general purpose technologies with patent data
399
60
Frequency
40
20
0 0
Figure 14.3
0.5 Generality index
1
Distribution of US patent class-based generality
Frequency
60
40
20
0 0
0.5 Generality based on sic of mfg
Figure 14.4
Distribution of industry of manufacture-based generality
1
400
The diffusion of new technologies
different distribution from that based on the US classification system. In Appendix Table A3, we show the correlation matrix for all five generality measures for our highly cited sample. Although they are generally fairly highly correlated, the industry of use measure (5) is not very correlated with the US class-based measures (1 and 3), and the industry of manufacture (4) not very correlated with the US class measure (1). Table 14.2 shows the 20 highly cited patents which also have the highest generality, where generality is measured by each of the five measures. In general, the most general patents are those in chemicals, especially when we consider the industry of use. Looking at the industry of manufacture, those in other technologies seem to be the most general. However, looking by US class, we can see the drawback of this generality measure: there are a number of chemical classes that are all essentially the same large class (the series 532–570), whereas in the case of some of the physics-based classes, there is only a single class. This fact will tend to bias the index towards generality in the chemicals case; however, the fact that the IPC classification which does not have this structure produces a similar result is somewhat reassuring. 3.2
Patenting Growth
Observation (3) suggested that we look at patent classes with rapid growth in patenting. Using the entire patent database aggregated to patent class, we constructed three sub-periods (1975–83, 1984–92 and 1993–99) and computed the average growth within class for each of the periods.13 The results are shown in Table 14.3. As might be expected, in all three periods, the patent classes with rapid growth are dominated by the information and data processing classes (395 and the 700 series), with the addition of the new multicellular biotechnology class 800 in the latter two periods. Highly cited patents are slightly more common in rapidly growing classes, although only a few of these classes have significant numbers of highly cited patents and the difference may not be very significant. It does appear that the patent classes that are growing rapidly include technologies that have more of the character of what we think of as GPTs, but that although highly cited patents are two to three times more likely to be found in rapidly growing classes (as we might expect if citations tend to come from the same class), they do not seem to be disproportionate in these classes. Another way of looking at the growth in patenting following the introduction of a GPT is to look at the growth of the patent classes that cite such a technology. The hypothesis is that innovations which build on a GPT-like innovation will themselves spawn many new innovations. Table 14.4 shows the patent classes for the top 20 patents in terms of the growth of their citing patent classes, both for the highly cited patents and for all patents, excluding
401
521 523 524 528
15
21
19
455
21 340 342
510
366 430
156
106 118 540 556 568
12
14
442
Communications Communications: Electrical Communications: Directive Radio Wave Systems and Devices (e.g., Radar, Radio Na) Telecommunications
Textiles: Web or Sheet Containing Structurally De Coatings Compositions: Coating or Plastic Coating Apparatus Organic Compounds – Part of the Class 532–570 Series Synthetic Resins or Natural Rubbers – Part of the Class 520 Series Miscellaneous chemicals Adhesive Bonding and Miscellaneous Chemicals Agitating Radiation Imagery Chemistry: Process, Composition, or Product Thereof Cleaning Compositions for Solid Surface TOTAL CHEMICALS
US patent class Class description
11
HJT subcategory
Table 14.2 Number of top 20 highly cited patents
0
0 0 0
0 6
0 0
0 0
3
0 0 0 0 3
US class
1
2 0 1
0 9
0 0
0 0
5
1 0 0 0 3
0
1 1 0
0 6
0 0
0 0
1
4 1 0 1 0
US IPC subcategory
0
0 0 0
1 6
0 1
2 0
1
1 1 1 0 1
0
0 0 0
1 10
1 2
6 2
1
3 0 0 0 0
Industry of Industry of manufacture use
Generality measure
402
128 604
174 257
32
41 46 49
264 359 49
138 53 248
51 54 59
67 68 69
348 386
345
Pipes and Tubular Conduits Package Making Supports TOTAL OTHER
Plastic and Nonmetallic Article Shaping Optics: Systems (including Communications) Movable or Removable Closures TOTAL MECHANICAL
Electricity: Conductors and Insulators Active Solid-State Devices (e.g., Transistors) Miscellaneous electrical Television Television Signal Processing for Dynamic Recording or Reproducing TOTAL ELECTRICAL
Surgery Surgery TOTAL DRUGS AND MEDICAL INSTRUMENTS
Selective Visual Display Systems TOTAL COMPUTING
US patent class Class description
(continued)
23
HJT subcategory
Table 14.2
0 0 1 1
0 0 1 1
3 0 1 4
2
6 3 0 0 3
1 1 0 0 0
2 0 2
0 2
1 0 2 3
4 0 0 4
2
1 1 0 0 0
2 2 4
0 1
US IPC subcategory
1 1 4 3 1
1 2 3
1 1
US class
1 1 6 8
2 0 0 2
1
1 0 0 0 0
2 0 2
0 0
0 0 3 3
1 2 0 3
1
1 0 0 0 0
1 2 3
0 0
Industry of Industry of manufacture use
Generality measure
403
0
0 0
12
7
0
0
0 0 0
19 148
90 311
152
523
1 061
742
255 37 42
7 548 146 045
Number
0.25 0.10
0.00 0.00 0.00
0.00
0.00
1.34
7.89
0.00 0.00
0.00
Share (%)
Highly cited patents 1993–99
US patent classes with rapid growth
335
Number
Patents in 1999
Table 14.3
12.1 12.0 11.7
13.0
13.1
15.6
21.6
22.2 22.1
22.2
(%)
Annual growth 1992–99
Total for selected classes All classes
Data Processing: Design and Analysis of Circuit or Semiconductor Mask Interactive Video Distribution Systems Data Processing: Software Development, Installation, or Management Data Processing: Structural Design, Modeling, Simulation, and Emulation Multicellular Living Organisms and Unmodified Parts Thereof and Related Processes Data Processing: Database and File Management, Data Structures, or Document Processing Semiconductor Device Manufacturing: Process Amusement Devices: Games Foundation Garments Chemistry: Fischer–Torpsch Processes; or Purification or Recovery of Products Thereof
Class description
463 450 518
438
707
800
703
725 717
716
Class
404
4
0 4
0 0
16 259
264
18 221
32 27
1 670 106 626
4
300
4
0 0
204 91
273
0
Number
0.96 0.24
0.00 0.00
0.00 1.81
1.52
1.47
1.33
0.00 0.00
0.00
Share (%)
Highly cited patents 1984–92
(continued)
240
Number
Patents in 1992
Table 14.3
14.2 14.1
15.9 15.7
16.1
17.2
17.6
21.8 18.2
22.2
(%)
Annual growth 1983–92 (%)
Total for selected classes All classes
Superconductor Technology: Apparatus, Material, Process Data Processing: Artificial Intelligence Multicellular Living Organisms and Unmodified Parts Thereof and Related Processes Electrical Computers and Digital Processing Systems: Multiple Computer or Process Data Processing: Database and File Management, Data Structures, or Document Processing Information Processing System Organization Textiles: Cloth Finishing Electrical Computers and Digital Processing Systems: Support Roll or Roller Railway Switches and Signals
Class description
492 246
26 713
395
707
709
706 800
505
Class
405
1
1
0
3
0
0
0
6 322
75
126
223
54
135
35
257
1 081 63 383
0.56 0.51
0.00
0.00
0.00
5.56
0.00
0.79
1.33
2.63
0.00 0.00
Share (%)
10.5
10.7
10.9
11.6
11.3
11.6
12.1
13.0
18.8 13.7
(%)
Annual growth 1975–83 (%)
Total for selected classes All classes
Information Processing System Organisation Data Processing: Speech Signal Processing, Linguistics, Language Translation, and Electrical Computers and Digital Processing Systems: Support Electrical Computers and Digital Processing Systems: Processing Architectures and Data Processing: Vehicles, Navigation, and Relative Location Data Processing: Generic Control Systems or Specific Applications Data Processing: Financial, Business Practice, Management, or Cost/Price Determination Data Processing: Measuring, Calibrating, or Testing Electrical Computers and Digital Processing Systems: Multiple Computer or Process Error Detection/Correction and Fault Detection/Recovery
Class description
Patent classes with fewer than 10 patents at the end of each period have been omitted from the table.
1
38
Note:
0 0
Number
Highly cited patents 1 975–83
42 96
Number
Patents in 1983
714
709
702
705
700
701
712
713
395 704
Class
406
TOTAL CHEMICALS Communications Multiplex Communications Telephonic Communications Computer Hardware and Software Cryptography Information Processing System Organisation Data Processing: Speech Signal Processing, Linguistics, Language Data Processing: Financial, Business Practice, Management, or Data Processing: Database and File Management, Data Structure Electrical Computers and Digital Processing Systems: Multiple Electrical Computers and Digital Processing Systems: Processing Electrical Computers and Digital Processing Systems: Support Data Processing: Software Development, Installation, or Management Selective Visual Display Systems TOTAL COMPUTING Chemistry: Molecular Biology and Microbiology TOTAL DRUGS AND MEDICAL INSTRUMENTS Miscellaneous electrical Television Semiconductor technology: apparatus, etc TOTAL ELECTRICAL TOTAL MECHANICAL TOTAL MISCELLANEOUS
Class description
2 0 0
2
3 18 1 1
3
1 2 1 3
0 4 3 1 11 1
Highly cited patents
1 1 0 0
0
1 1 19
1
16
18
0
All patents excl. highly cited
Note: *Classes for the top 20 patents in each category are shown. The average growth rate of the classes of the highly cited patents is above 28 per cent per annum; those for all patents above 52 per cent per annum.
348 505
33
49
435
23
370 379
US patent class
Patent classes whose cited classes have high growth rates*
380 395 704 705 707 709 712 713 717 345
22
21
HJT subcategory
Table 14.4
407
Uncovering general purpose technologies with patent data
those that are highly cited. The message is clear: using this measure, almost all the patent classes identified are in computing and communications technology, and most are in data processing technologies more narrowly defined. 3.3
Citation Lags
Finally, observation (4) suggested that the average citation lag to GPT patents might be longer. Of course, citations to patent data for a fixed time period such as ours (1967–99) are always subject to truncation. For this reason, we look at mean citation lags that are large relative to the average citation lag for patents applied for in the same year. Table 14.5 shows that the 20 highly cited patents with long lags (greater than 70 per cent of the average citation lag) are typically in older Table 14.5
Patent classes with highly cited patents that have long cite lags
HJT subcategory
US patent class
24
365
32 39
604, 606 623
51 264 425 54
359
61 68
47
69
53 206 383 428
Class description
Number of patents
TOTAL CHEMICALS
0
Static Information Storage and Retrieval TOTAL COMPUTING
1 1
Surgery and Med. Instruments Prosthesis (i.e., Artificial Body Members), Parts Thereof, or A TOTAL DRUGS AND MEDICAL INSTRUMENTS
3 1
TOTAL ELECTRICAL
0
Mat. Proc. and Handling Plastic and Nonmetallic Article Shaping or Treating: Processes Plastic Article or Earthenware Shaping or Treating: Apparatus Optics TOTAL MECHANICAL
2 1
Agriculture, Husbandry, Food Receptacles Package Making Special Receptacle or Package Flexible Bags Stock Material or Miscellaneous Articles TOTAL MISCELLANEOUS
4
1 2 4 3 7 1 3 3 1 11
408
The diffusion of new technologies
technologies. It is noteworthy that there are none in the chemicals or electrical industries and only five in computing and drugs, most of which are to surgical innovations. The only highly cited computing patent with long citation lags is a patent on an aspect of computer architecture taken out by Siemens in 1976; this patent has a mean citation lag of 23 years and is noteworthy because it is has essentially no citations until after it expired in 1994. It now has over 200. In general, given the fact that long lags by themselves often simply identify older and slower-moving technologies such as packaging, we will want to use this indicator in combination with our other indicators when looking for GPTs. 3.4
Summary
The GPT measures we have identified (the five generality indexes, the generality of citing patents, within class growth in patenting, growth in citing patent classes and the average citation lag) are promising, but clearly give contradictory messages when taken separately. The goal is to combine them in a reasonable way to give an indication of the types of evidence GPTs leave in the patent statistics. We explore solutions to this problem in section 5, but first we summarise the relationship between them and the probability that a patent is highly cited.
4
HIGHLY CITED PATENTS
Table 14.6 shows that the highly cited patents differ in almost all respects from the population of all patents, and also from a 4 per cent sample of patents with at least one cite that we will use later as a control sample. They take longer to be issued, they have about twice as many claims, they are more likely to have a US origin, and more likely to be assigned to a US corporation, more likely to have multiple assignees and have higher citation lags on average.14 They also have higher generality, no matter how generality is measured, and are in patent classes that are growing faster than average. Although the patents that cite them are more likely to be cited themselves, they have only slightly higher generality than citing patents in general. More than half of the highly cited patents are in two of our six main technology classes: computing hardware and software, and drugs and medical instruments. Of course, these are indeed the technology classes where we expect to find modern-day GPTs. In Appendix Table A2, we broke this down, in order to identify the important technologies more precisely. Highly cited patents are more than twice as likely to be found in computer
409
Uncovering general purpose technologies with patent data
Table 14.6
US patents granted 1967–99
Statistic
All patents
4 % sample of patents (1 cite)
Highly cited patents
Number of patents Year applied for Year granted Average grant lag (years) Number of claims
2 768 011 1983.1 1984.6 1.5 12.1
100 634 1982.8 1984.8 2.0 12.5
780 1981.6 1983.9 2.3 23.6
Number of forward citations (to 2002) Average citation lag Average class growth (5 years) Share US origin Share assigned to US corporations Share multiple assignees
6.72
8.85
204.71
8.73 NA
9.98 2.98%
13.48 7.28%
59.90% 46.50%
61.20% 47.70%
88.30% 75.90%
Generality 1 (US class) Generality 2 (IPC) Generality 3 (US subcategory) Generality 4 (SIC of mfg-IPC) Generality 5 (SIC of use-IPC) Average cites to citing patents Total cites to citing patents Average growth of citing patent classes* Average generality of citing patents Broad technology classes Chemicals Computing Drugs and medical Electrical Mechanical Other Type of assignee US corporation Non-US corporation US individual Non-US individual US government Non-US government
0.50%
0.60%
1.00%
0.3417 0.3548 0.2711
0.5261 0.5484 0.4167
0.6416 0.5716 0.4569
NA
NA
0.5856
NA
NA
0.6444
4.02
4.7
12.92
46.5 NA
55.7 3.57%
2663.9 7.69%
0.3094
0.3487
0.3887
20.80% 10.20% 7.30% 17.10% 23.00% 21.60%
19.20% 11.40% 7.00% 17.80% 22.90% 21.80%
18.00% 23.90% 32.60% 9.70% 6.10% 9.70%
1 247 030 885 533 378 394 135 756 43 048 9 845
47 975 31 798 14 347 4 645 1 499 370
596 75 98 9 6 1
410
The diffusion of new technologies
hardware and software, computer peripherals, surgery and medical instruments, genetic technologies, miscellaneous drugs and semiconductors. Table 14.7 shows a series of probit estimations for the probability that a patent with at least one cite will be highly cited, in order to provide a multivariate summary of the data in Table 14.6.15 Table 14.7 shows the derivative of the probability with respect to the independent variable that are implied by the coefficient estimates. In the case of dummy variables, it shows the change in probability when the variable changes from zero to one. Because the probability of being one of the highly cited patents in the sample is very small (0.77 per cent), the values in Table 14.7 are small. Taking the grant lag as an example, the interpretation is that an additional year between application and grant increases the probability of being highly cited by 0.06 per cent, or from 0.77 per cent to 0.83 per cent at the mean. Being a patent in the drugs and medical category increases the probability by 2.9 per cent, which is a very large change at the mean probability. Table 14.7 confirms the univariate differences between highly cited and all patents. In addition, this table shows that variations over time in the probability of high citation do not greatly affect the coefficients (compare column 4 with column 2). The only generality measure that enters significantly and positively in this regression is that based on the US class; the others were all insignificant (IPC, technology sub-category and industry of use) or slightly negative (industry of manufacture). Also note that highly cited patents are far more likely to be cited by patents that are themselves cited by patents in many technology classes, once we control for the other differences between highly cited and other patents.
5
IDENTIFYING GPT PATENTS
It is not obvious how to combine these measures to choose a sample of GPT patents. In this first investigation of the topic, we have chosen simply to look for patents that are outliers in several of the categories, on the grounds that such patents are likely to give us an idea of the technologies that have given birth to the most subsequent inventive activity in the largest number of technological areas. Accordingly, we began with the 780 highly cited patents and then we chose a set of patents that fell in the top 20 per cent of these patents according to generality, citing patent generality and the subsequent five-year growth of the patent’s class. We performed this exercise for each of the five generality measures in turn. Table 14. 8 shows the result: 20 patents out of the 780 were selected, many by several of the different criteria. Selection by each of the five generality measures is indicated by the presence of the measure in the table. Of these
411
0.124% 0.400% 0.062% 0.397% 0.210%
dp/dx*** 0.011% 0.063% 0.015% 0.047% 0.046%
Std. err
Cited patent characteristics
0.006% 0.025% 0.008% 0.025% 0.026% 0.032% 0.005% 0.044%
0.226% 0.068% 0.331%
Std. err
0.062% 0.033% 0.057% 0.146% 0.147%
dp/dx***
Cited and citing patent char.
0.470%
0.096% 0.078%
0.076% 0.039% 0.066% 0.163% 0.166%
dp/dx***
0.050%
0.063% 0.005%
0.007% 0.029% 0.009% 0.028% 0.029%
Std. err
Cited and citing patent char.
Probit regression for highly cited patents (101 414 observations; 780 highly cited)*
Number of claims/10 D (claims missing) Average grant lag (years) Dummy for US origin Dummy for US corporation Generality 1 (US class) Generality 5 (SIC of use) Average citation lag (relative to year average) Average generality of citing patents
Variable
Table 14.7
0.334%
0.066%
0.202%
0.055% 0.508% 0.054% 0.137% 0.132%
dp/dx***
0.044%
0.005%
0.030%
0.006% 0.452% 0.008% 0.023% 0.024%
Std. err
Including year dummies
412 No 0.130 3980.69
No
0.052% 0.147% 0.291% 0.048% 0.029%
Std. err
0.222 3556.96
0.201% 1.007% 2.432% 0.118% 0.077%
dp/dx***
Cited and citing patent char.
0.060% 0.158% 0.292% 0.052% 0.033%
Std. err
0.217 3583.02
No
0.239% 1.005% 2.454% 0.119% 0.085%
dp/dx***
Cited and citing patent char.
0.049% 0.138% 0.275% 0.044% 0.027%
Std. err
0.229 3528.15
Yes
0.191% 0.904% 2.222% 0.106% 0.069%
dp/dx***
Including year dummies
Notes: Coefficient estimates in italics are not significant at the 1 per cent level. * The sample of non-highly-cited patents is a 10 per cent sample of all patents that have two or more citations. The average probability of being highly cited in the sample is 0.7 per cent. ** The excluded class is other technologies. *** Estimated derivative of probability with respect to independent variable. For dummy variables ( ), the discrete change in probability from 0 to 1.
Scaled R-squared Log likelihood
Year dummies
0.085% 0.162% 0.318% 0.072% 0.057%
Dummies for technology classes** Chemicals 0.308% Computing 1.085% Drugs and medical 2.882% Electrical 0.079% Mechanical 0.157%
Cited patent characteristics Std. err
(continued)
dp/dx***
Variable
Table 14.7
413
186
377
4558413
125
3842194
4528643
181
3636956
178
129
3624019
3956615
Number of Cites
Patent number
1983
1983
1974
1971
1970
1970
Application year
CA, US
TX, US
CA, US
NJ, US
DE, US
IL, US
Inventor state, country
Xerox
FPDC (Freeny patent)
IBM
RCA Corporation
Ethicon, Inc.
Nalco Chemical Compay
Assignee
Process for Rapidily Dissolving Water-soluble Polymers Polyactide sutures (absorbable) Information records and recording playback system therefore (video disc) Transaction execution system with secure data storage and communications System for reproducing information in material objects in a point of sale location Software version management system
Patent description
0.826
0.880
0.801
0.846
By US class
0.797
0.843
0.841
0.907
By IPC
0.696
0.730
0.696
0.659
By tech subcategory
Table 14.8 Highly cited patents with high generality, class growth and citing patent generality
0.856
By ind. of manufacture
Generality
0.830
0.825
0.798
By industry of use
414
183
286
181
4916441
4953080
195
4783695
4885717
200
4672658
180
186
4575621
4821220
Number of Cites
1988
1988
1986
1986
1986
1986
1984
Application year
(continued)
Patent number
Table 14.8
CA, US
CO, US
OR, US
WA, US
NY, US
NJ, US
PA, US
Inventor state, country
HewlettPackard Co
Clinicom Inc
Tektronix
Tektronix
General Electric Co
Corpra Research Inc AT&T
Assignee
Portable electronic transaction device and system therefor Spread spectrum wireless PBX Multichip integrated circuit packaging configuration and method System for animating program operation and displaying timebased relationships System for graphically representing operation of objectoriented programs Portable handheld terminal Object management facility for maintaining data in a computer system
Patent description
0.794
0.912
0.816
0.796
0.824
By US class
0.824
0.844
0.804
By IPC
0.827
0.674
0.714
By tech subcategory
By ind of manufacture
Generality By industry of use
415
178
173
5307456
255
5347632
5132992
217
5093914
200
224
5155847
5119475
210
5133075
1992
1991
1991
1989
1989
1988
1988
CA, US
NY, US
TX, US
NJ, US
IL, US
ON, CA
CA, US
Sony Electronics Inc
Schlumberger Technology Corp unassigned
Prodigy Services Co
AT&T
Minicom Data Corp
HewlettPackard Co
Method of monitoring changes in attribute values of object in an object-oriented database Method and apparatus for updating software at remote locations Method of controlling the execution of object-oriented programs Reception system for an interactive computer network and method of operation Object-oriented framework for menu definition Audio and video transmission and receiving system (compression) Integrated multimedia production and authoring system 0.856
0.810
0.796
0.832
0.870
0.812
0.682
0.764
416
4558413
4575621 4672658 4783695
4821220
4885717
0.615 0.521 0.609
0.570
0.581
3956615
0.658
0.547
3636956 3842194
0.573 0.546
4528643
3624019
0.566
0.616
Patent number
9.1
8.8
10.4 9.9 9.9
13.1
14.6
18.2
21.5 8.6
17.3
Mean cite lag
(continued)
Citing patents generality
Table 14.8
0.6
0.3
1.2 1.4 1.4
3.5
5.0
4.4
6.1 6.0
1.9
Mean cite lag relative to average
22
22
59 21 46
22
22
22
32 24
15
Sub-category
395
395
235 455 257
707
705
705
606 369
523
US class
19.9
19.9
13.3 13.2 20.1
18.9
14.7
20.7
13.0 12.9
15.2
Growth of class ( %)
Synthetic Resins or Natural Rubbers – Part of the Class 520 Series Surgery Dynamic Information Storage or Retrieval Data Processing: Financial, Business Practice, Management, or Cost Data Processing: Financial, Business Practice, Management, or Cost Data Processing: Database and File Management, Data Structures Registers Telecommunications Active Solid-State Devices (e.g., Transistors, SolidState Diodes) Information Processing System Organisation Information Processing System Organisation
Class description
417
5347632
5119475
5132992
5307456
0.566
0.707
0.539
5155847
0.541
0.542
5133075
0.557
5093914
4953080
0.593
0.593
4916441
0.539
5.9
6.6
6.5
6.5
6.6
8.3
6.5
8.0
9.2
21 23
0.1 0.1
23
22
0.9
0.0
22
0.7
22
22
1.3
0.5
22
23
0.2
1.4
345
375
345
709
395
709
707
707
345
23.7
18.5
26.5
24.1
18.6
30.2
18.6
18.6
15.1
Selective Visual Display Systems Data Processing: Database and File Management, Data Structures Data Processing: Database and File Management, Data Structures Electrical Computers and Digital Processing Systems: Multiple Computers Information Processing System Organisation Electrical Computers and Digital Processing Systems: Multiple Computers Selective Visual Display Systems Pulse or Digital Communications Selective Visual Display Systems
418
The diffusion of new technologies
patents, all but two were in technologies related to information and communication technology (ICT). The remaining two are the oldest (applied for in 1970) and cover a process that is useful in the making of paper, and in sewage and waste treatment, and in absorbable sutures for surgery. All but one of the patents cover US inventions, five from California, three from New Jersey and the remainder from a number of other states. The sole exception comes from a Toronto-based company. All but one of the patents were assigned to corporations at the time they were taken out; the exception was a patent for a method of compressing audio and video data for transmission. The ICT-related patents cover a range of technologies: integrated circuit manufacturing, handheld computers, spread spectrum technology, and so forth. What is noteworthy is the number of patents that relate to Internet transactions (e-commerce) and software development, especially objectoriented programming. Some of the e-commerce patents greatly precede the actual use of the technology. For example, the celebrated Freeny patent (US4528643, shown in Figure 14.5) was applied for in 1983 and issued in United States Patent Freeny, Jr.
4,528,643 July 9, 1985
System for reproducing information in material objects at a point of sale location Abstract The present invention contemplates a system for reproducing information in material objects at a point of sale location wherein the information to be reproduced is provided at the point of sale location from a location remote with respect to the point of sale location, an owner authorization code is provided to the point of sale location in response to receiving a request code from the point of sale location requesting to reproduce predetermined information in a material object, and the predetermined information is reproduced in a material object at the point of sale location in response to receiving the owner authorization code. Inventors: Freeny, Jr.; Charles C. (Fort Worth, TX) Assignee: FPDC, Inc. (Oklahoma City, OK) Appl.No.: 456730 Filed: January 10, 1983 http://www.e-data.com/e-freeny.htm Figure 14.5
The Freeny patent (4528643)
Uncovering general purpose technologies with patent data
419
1985, but has been successfully asserted against such corporations as Microsoft and Apple almost to the present day.16 The fact that the original use for which this patent was contemplated is unlikely to have been Internet-based e-commerce reminds us to be cautious in our interpretation of the results in Table 14.8: we do not argue that the patents we identify are necessarily the source of the GPT itself, but we do suggest that by identifying them via the subsequent growth and generality in their citing patents, we are observing the symptoms of the diffusion and development of a general purpose technology. Looking at the actual ICT patents in Table 14.8 (rather than at the classes in which they have been placed) yields the following summary: seven are related to object-oriented and windows-based software, four to Internet commerce and communication, three to audio-video applications, two to handheld computing, and one each to telecommunications and semiconductor manufacturing. Thus the specific technologies identified as being both general and spawning rapid patenting growth are those related to the effective use of the computer, especially for interacting and transacting over distance. That is, they are not computing hardware patents per se, but patents on the technologies that allow a network of computers to operate together effectively and to interact with the users of those computers. This seems to us to characterise the GPT of the 1980s and 1990s, and we would therefore declare our prospecting exercise a qualified success.17
6
CONCLUSIONS
Many empirical papers close with interpretive cautions and calls for further research. This chapter is no exception, but the caution and the call are stronger than usual. For reasons of limited time and computing power, we have not been able to explore the validity and use of the measures we have constructed as much as we would like and we encourage further work in this area. In particular, all the generality measures suffer from the fact that they treat technologies that are closely related but not in the same class in the same way that they treat very distant technologies. This inevitably means that generality may be overestimated in some cases and underestimated in others. One suggestion for future research would be to construct a weighted generality measure, where the weights are inversely related to the overall probability that one class cites another class. A second area of concern has to do with changes in the strategic uses of patents during the two decades we have studied. These changes are not unrelated to the growth in importance of ICT technologies but they may also have had an distorting impact on some of the measure we have used.
420
The diffusion of new technologies
In particular, as Hall and Ziedonis (2001) have shown, one reason for rapid growth in semiconductor patenting after the mid-1980s is a conscious decision on the part of some major firms to build up their patent portfolios in order to fend off litigation and negotiate cross-licensing agreements. This type of strategy has spread throughout the industry and the consequences for patenting by firms such as IBM, Lucent and Hewlett-Packard has been confirmed by Bessen and Hunt (2004) and Hall (2005). The implication is that citations to earlier patents in the ICT sector may be growing rapidly partly because of a strategic shift as well as because the underlying technology is growing in importance and diffusing throughout the economy. Sorting this out from our data will require more attention to the time series behaviour of the indicators, improved generality measures and more detailed investigation of the firms involved. In the interim, this chapter has demonstrated the potential validity of patent-based measures of GPTs and we hope it will spur further investigation into the use of patent data in this way.
APPENDIX Table 14.A1
SIC-industry correspondence for generality indices
Hall–Vopel quasi 2-digit industry
SIC codes (1987)
01 02 03 04 05 06 07 08 09 10 11 12 13 14 15
20xx, 21xx 22xx, 23xx, 31xx, 3021, 3961, 3965 24xx 25xx 26xx 27xx 28xx, excl. 283x, 284x 13xx, 29xx 30xx, excl. 3021 32xx 33xx 34xx 35xx, excl. 357x, 358x, 3524 357x 358x, 3596, 360x, 361x, 362x, 363x, 364x, 3677, 369x, excl. 3690, 3695 3651, 3652, 366x, 367x, excl. 3677, 3678; 3690, 3695, 381x, 382x, excl. 3827 372x, 373x, 374x, 376x–379x, excl. 3790, 3792, 3799 371x, excl. 3714; 375x, 3790, 3792, 3799
Food and tobacco Textiles, apparel and footwear Lumber and wood products Furniture Paper and paper products Printing and publishing Chemical products Petroleum refining and prods Plastics and rubber prods Stone, clay and glass Primary metal products Fabricated metal products Machinery and engines Computers and comp. equip. Electrical machinery
16 Electronic inst. and comm. eq. 17 Transportation equipment 18 Motor vehicles
421
Uncovering general purpose technologies with patent data
Table 14.A1
(continued)
Hall-Vopel quasi 2-digit industry
SIC codes (1987)
19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37
3827, 384x, 386x 283x, 3851 387x, 39xx, excl. 3961, 3965 284x 3714 737x 48xx 50xx 73xx, excl. 737x 01xx–09xx 10xx, 11xx, 12xx, 14xx 15xx–19xx 40xx–47xx 49xx 51xx-59xx 60xx–69xx 80xx 87xx 70xx–99xx and not 73xx, 80xx, 87xx
Optical and medical instruments Pharmaceuticals Misc. manufacturing Soap and toiletries Auto parts Computing software Telecommunications Wholesale trade Business services Agriculture Mining Construction Transportation services Utilities Trade Fire, Insurance, Real Estate Health services Engineering services Other services
Table 14.A2
Breakdown by technology sub-category
Sub-category
All patents Number
Ratio
Share ( %)
Number
Share ( %)
0.9 1.5 0.5 4.2 3.5 10.2 20.8
8 7 4 13 40 69 141
1.0 0.9 0.5 1.7 5.1 8.8 18.0
1.17 0.58 1.04 0.39 1.45 0.86 0.86
118 316 90 326
4.3 3.3
51 93
6.5 11.8
1.52 3.63
24 147 49 963 282 752
0.9 1.8 10.2
28 16 188
3.6 2.0 23.9
4.09 1.13 2.34
Agriculture, Food, Textiles 24 134 Coating 42 235 Gas 13 614 Organic Compounds 116 334 Resins 96 948 Miscellaneous 282 717 Chemical technologies 575 982 total Communications Computer Hardware and Software Computer Peripherials Information Storage Computer hardware and software total
Highly cited patents
422
Table 14.A2
The diffusion of new technologies
(continued)
Sub-category
All patents Number
Highly cited patents
Ratio
Share (%)
Number
Share (%)
83 410 69 344
3.0 2.5
35 164
4.5 20.9
1.48 8.34
31 794 16 312 200 860
1.1 0.6 7.3
24 33 256
3.1 4.2 32.6
2.66 7.13 4.49
Electrical Devices Electrical Lighting Measuring and Testing Nuclear and X-rays Power Systems Semiconductor Devices Miscellaneous Electrical technologies total
92 508 44 738 80 315 40 746 97 739 51 950 66 440 474 436
3.3 1.6 2.9 1.5 3.5 1.9 2.4 17.1
1 6 2 4 4 38 21 76
0.1 0.8 0.3 0.5 0.5 4.8 2.7 9.7
0.04 0.47 0.09 0.35 0.14 2.58 1.11 0.56
Mat. Proc. and Handling Metal Working Motors and Engines Parts Optics Transportation Miscellaneous Mechanical technologies total
155 200
5.6
16
2.0
0.36
88 661 102 504
3.2 3.7
11 1
1.4 0.1
0.44 0.03
62 832 82 854 143 849 635 900
2.3 3.0 5.2 23.0
4 0 16 48
0.5 0.0 2.0 6.1
0.22 0.00 0.39 0.27
59 793
2.2
19
2.4
1.12
28 095 50 477 40 857
1.0 1.8 1.5
0 0 0
0.0 0.0 0.0
0.00 0.00 0.00
57 362
2.1
0
0.0
0.00
38 146 25 198 58 616
1.4 0.9 2.1
0 3 32
0.0 0.4 4.1
0.00 0.42 1.93
Drugs Surgery and Med. Instruments Genetics Miscellaneous Drugs and med. instruments total
Agriculture, Husbandry, Food Amusement Devices Apparel and Textile Earth Working and Wells Furniture, House Fixtures Heating Pipes and Joints Receptacles
Uncovering general purpose technologies with patent data
423
(continued)
Table 14.A2 Sub-category
All patents Number
Highly cited patents
Share (%)
Number
Ratio
Share (%)
Miscellaneous
239 537
8.7
22
2.8
0.32
Other technologies total
598 081
21.6
76
9.7
0.45
2 768 011
100.0
785
100.0
1.00
All technologies total
Table 14.A3
Correlation matrix for generality indices (N 780) 1 US class 2 IPC
Generality 1 (US class) Generality 2 (IPC) Generality 3 (US subcategory) Generality 4 (industry of manufacture) Generality 5 (industry of use)
3 US sub-category
4 Industry 5 Industry of of manufacture use
1.000 0.555
1.000
0.621
0.523
1.000
0.238
0.590
0.599
1.000
0.143
0.627
0.389
0.632
1.000
NOTES 1.
2. 3.
A preliminary version of this chapter was presented at the conference ‘New Frontiers in the Economics of Innovation and New Technology,’ held in honour of Paul A. David at the Accademia delle Scienze, Torino, 20–21 May, 2000. We are grateful to participants in that conference, especially Paul David, John Cantwell, Giovanni Dosi, Ove Granstrand and Ed Steinmueller, for comments on the earlier draft. The first author thanks the Centre for Business Research, Judge Institute of Management, University of Cambridge for hospitality while this version was being written. Griliches (1990); Pavitt (1988). In the US environment, this statement is increasingly less true, although the converse, that not all patent subject matter is innovative, may be becoming more true.
424 4. 5.
6. 7. 8.
9.
10.
11.
12.
13. 14.
15.
16. 17.
The diffusion of new technologies As a general rule, the USPTO does not classify patents individually into IPC classes, but relies on a map based on US classes and sub-classes to determine them. This is not ideal, but does mean that they incorporate some subclass information. For recent evidence on this point, see Harhoff et al. (1999) for the results of a survey of patent owners, and Hall et al. (2005) for results showing that the market value–citation relationship is highly non-linear, with firms owning highly cited patents subject to very large premia, as well as a graph showing the frequency distribution of patent citations. This description of the meaning of patent citations is drawn from Hall et al. (2002). See Jaffe et al. (2000) for evidence from a survey of inventors on the role of citations in both senses. The USPTO has developed over the years a highly elaborate classification system for the technologies to which the patented inventions belong, consisting of about 400 main (three-digit) patent classes, and over 120 000 patent sub-classes. This system is being updated continuously, reflecting the rapid changes in the technologies themselves. Trajtenberg, Jaffe and Hall have developed a higher-level classification, by which the 400 classes are aggregated into 36 two-digit technological sub-categories, and these in turn are further aggregated into six main categories: Chemical, Computers and Communications (C&C), Drugs and Medical (D&M), Electrical and Electronics (E&E), Mechanical, and Others. Note that generality is not defined if a patent receives no citations, and is zero by construction when a patent receives only one. We have omitted such patents in the tables and graphs shown in this chapter. They comprise about one-quarter of all patents in our sample. A typical citation lag distribution is shown in Figure 14.2. This curve was estimated from the observed data using the methodology of Trajtenberg et al. (1997). See Appendix D of Hall et al. (2002) for details. From the graph it appears that over half the citations ever made are made in the first six years since the cited patent’s application date. Slightly fewer than half the patents granted between 1967 and 1999 are assigned to US corporations that we can identify (see Hall et al., 2002). However, many of these are in multiple industries so the primary industry assignment may not be relevant for the particular patent or citation that we are using. The actual industry classification we use was developed by Hall and Vopel (1997) from an earlier classification used by Levin and Reiss (1984). It is based on four-digit SICs aggregated up to a level that is coarse enough to include most, but not all, whole firms in the US manufacturing sector. We augmented this industry list with ten industries for the non-manufacturing sector. See Appendix Table A1 for details. We have omitted patent classes with fewer than 10 patents at the end of each period. Unlike the case of generality measures, the mean citation lag is linear in the citation counts and therefore not a biased estimate, conditional on the total number of citations. It is, however, truncated at the end of the period, but this truncation affects both highly cited and non-highly cited patents equally. The sample used is the 780 highly cited patents plus the four per cent random sample of patents with at least one cite shown in Table 14.2. The fact that we use a random sample rather than a population affects the constant term in this probit regression, so we do not report it. The other coefficient estimates will not be affected by this procedure, although the interpretation of the changes in probability will depend on the average probability in the sample used. This patent is currently owned by E-Data Corporation and was agressively asserted by that company in the US beginning in 1996 (http://www.prpnet.net/7604.html). Note that the industry of manufacture and industry of use measures do not identify the software and Internet patents as GPTs, for reasons discussed earlier: they have been obtained using a concordance that did not really admit these as patentable areas.
Uncovering general purpose technologies with patent data
425
BIBLIOGRAPHY Abramovitz, Moses and Paul A. David (1996), ‘Technological change, intangible investments and growth in the knowledge-based economy: the US historical perspective’, in D. Foray and B.A. Lundvall, Employment and Growth in the Knowledge-based Economy, Paris, OECD. Bessen, James and Robert M. Hunt, (2004), ‘An empirical look at software patents’, Philadelphia: Federal Reserve Bank Working Paper No. 03–17/R. Bresnahan, Timothy J. and Manuel Trajtenberg (1995), ‘General purpose technologies: “Engines of growth”?’, Journal of Econometrics, 95, 83–108. David, Paul A. (1990), ‘The dynamo and the computer: an historical perspective on the modern productivity paradox’, American Economic Review, 80, 355–61. David, Paul A. (1991), ‘Computer and dynamo: the modern productivity paradox in a not-too-distant mirror’, in OECD (ed.), Technology and Productivity, Paris: OECD. Also Stanford University, CEPR Discussion Paper No. 172 (1989). David, Paul A. and Gavin Wright (2003), ‘General purpose technologies and productivity surges: historical reflections on the future of the ICT revolution’, in P. David and M. Thomas (eds), The Economic Future in Historical Perspective, Oxford: Oxford University Press for the British Academy. Griliches, Zvi (1990), ‘Patent statistics as economic indicators: a survey’, Journal of Economic Literature, 28, 1661–707. Hall, Bronwyn H. (2005), ‘Exploring the patent explosion’, Journal of Technology Transfer, 30, 35–48. Hall, Bronwyn H. (2002), ‘A note on the bias in the Herfindahl based on count data’, in A. Jaffe and M. Trajtenberg (eds), Patents, Citations, and Innovation, Cambridge, MA, MIT Press. Hall, Bronwyn H. and Katrin Vopel (1997), ‘Market value, market share, and innovation’, NBER, the University of California at Berkeley, and the University of Mannheim. Industry classification available at http://emlab.berkeley.edu/users/ bhhall/index.html Hall, Bronwyn H. and Rosemarie Ham Ziedonis (2001), ‘The determinants of patenting in the US semiconductor industry, 1980–94’, Rand Journal of Economics, 32, 101–28. Hall, Bronwyn H., Adam Jaffe and Manuel Trajtenberg (2002), ‘The NBER patent citations data file: lessons, insights and methodological tools’, in A. Jaffe and M. Trajtenberg (eds), Patents, Citations and Innovations, Cambridge, MA, MIT Press. Also NBER Working Paper No. 8498 (October 2001). Hall, Bronwyn H., Adam Jaffe and Manuel Trajtenberg (2005), ‘Market value and patent citations’, Rand Journal of Economics, 36 (1), 16–38. Also NBER Working Paper No. 7741 (June 2000). Harhoff, Dietmar, Francis Narin, F.M. Scherer and Katrin Vopel (1999), ‘Citation frequency and the value of patented inventions’, Review of Economics and Statistics, 81 (3), 511–15. Helpman, Elhanan and Manuel Trajtenberg (1998a), ‘Diffusion of general purpose technologies’, in E. Helpman (ed.), General Purpose Technologies and Economic Growth, Cambridge: MIT Press, pp. 85–119. Helpman, Elhanan and Manuel Trajtenberg (1998b), ‘A time to sow and a time to reap: growth based on general purpose technologies’, in E. Helpman (ed.), General Purpose Technologies and Economic Growth, Cambridge, MIT Press, Chapter 3.
426
The diffusion of new technologies
Jaffe, Adam, Manuel Trajtenberg and Michael Fogarty (2000), ‘Knowledge spillovers and patent citations: evidence from a survey of inventors’, American Economic Review, Papers and Proceedings (May), 215–18. Levin, Richard C. and Peter C. Reiss (1984), ‘Tests of a Schumpeterian model of R&D and market structure’, in Z. Griliches (ed.), R&D, Patents and Productivity, Chicago, University of Chicago Press for the National Bureau of Economic research, pp. 175–208. Office of Technology Assessment and Forecast (OTAF) (1976), US Department of Commerce, Patent and Trademark Office, Sixth Report, Washington, DC, US Government Printing Office. Pavitt, Keith (1988), ‘Uses and abuses of patent statistics’, in A.F.J. van Raan (ed.), Handbook of Quantitative Studies of Science and Technology, Amsterdam: Elsevier Science, pp. 509–36. Rosenberg, Nathan (1976), ‘Technological change in the machine tool industry, 1840–1910’, in N. Rosenberg (ed.), Perspectives in Technology, Cambridge, Cambridge University Press, pp. 9–31. Sakakibara, Mariko and Lee G. Branstetter (2001), ‘Do stronger patents induce more innovation? Evidence from the 1988 Japanese patent law reforms’, Rand Journal of Economics, 32 (1), 77–100. Silverman, Brian (2002), ‘International Patent Classification – US SIC concordance’, http://www.rotman.utoronto.ca/~silverman/ipcsic/ipcsicfiles.ZIP Trajtenberg, Manuel, Adam Jaffe and Rebecca Henderson (1997), ‘University versus corporate patents: a window on the basicness of invention’, Economics of Innovation and New Technology, 5 (1), 19–50.
15. Equilibrium, epidemic and catastrophe: diffusion of innovations with network effects Luís M.B. Cabral* 1
INTRODUCTION
It seems that important papers are characterised by long publication lags. Maskin’s famous mechanism design theorem and Holmstrom’s seminal paper on managerial concerns each took about 20 years to get published.1 Prominent among the list of famous works that remained unpublished for a long time is Paul David’s ‘Contribution to the theory of diffusion’ (David, 1969). In that paper, David develops an equilibrium model of new technology adoption and shows how S-shaped diffusion paths reflect heterogeneity among adopters. In this chapter, I too focus on the issue of diffusion of innovations, specifically innovations subject to network effects. Like David and others, I start from an equilibrium model of adopter heterogeneity. However, I will argue that, in the presence of strong network effects, the nature of the adoption process is quite different from what was previously characterised. In particular, I show that network effects imply discontinuous adoption paths – mathematically speaking, a catastrophe. In a previous paper (Cabral, 1990), I noted how network effects may lead to discontinuous adoption paths. This chapter goes beyond Cabral (1990) in two ways. First, I provide a more precise set of conditions under which a catastrophe takes place (section 3). Second, I suggest a possible test to distinguish between alternative theories of new technology adoption (section 5).2 S-shaped diffusion paths, one of the most robust empirical regularities found in the literature, are consistent with a number of theories. I consider two types: (1) equilibrium diffusion theories based on adopter heterogeneity, and (2) epidemic theories based on some form of imperfect information and/or word-of-mouth effects. I start from a model of the first type and add network effects to it. I then compare it to a model of the second type, also allowing for the possibility of network effects. 427
428
2
The diffusion of new technologies
AN EQUILIBRIUM MODEL OF NEW TECHNOLOGY ADOPTION
Consider a new technology available from some time t0. The cost of adopting such innovation is ct per period. That is, upon adoption, a flow cost ct must be paid. I assume that ct is decreasing, which reflects gradual, post-invention, technological development, as well as increased competition in supply. I am particularly interested in the case when the benefit from an innovation can be measured by its use. For example, the benefit from having a telephone is proportional to the use that is made of such telephone (or, if we also want to consider the ‘stand-alone’ benefit from owning a telephone, then total benefit is a linear function of use). Formally, Assumption 1
Each adopter’s benefit is proportional to use.
I am also interested in the case when the innovation is subject to network effects, that is, the case when adoption benefits are increasing in the number of adopters. Specifically, suppose that each potential user derives a benefit from communicating with a set of other users. Such benefit can only be gained if the other users are also hooked up to the network, that is, if the other users have adopted the innovation as well. Suppose, moreover, that the event of being part of the list of desirable links is independent of the user’s type. Then the use of (and benefit from) the innovation is a linear function of the number of users. Assumption 2 Each adopter’s willingness to pay is a linear function of the number of users. I assume that potential adopters are different from each other. Specifically, each potential adopter is characterised by a parameter ) that measures its willingness to pay for the innovation. Specifically, I assume that Assumption 3 Each adopter’s use of the innovation is proportional to the adopter’s type . Assumption 4
N( ,).
The above assumptions imply that, upon adoption (which I assume is irreversible), an adopter of type receives a benefit flow given by ut ((1 *) *nt),
(15.1)
Diffusion of innovations with network effects
429
where nt is the measure of adopters at time t. The parameter * measures the importance of network effects. In the extreme when * 0, benefit is simply given by (stand-alone utility), that is, independent of network size. In the opposite extreme, when * 1, stand-alone utility is zero and benefit is proportional to network size. It is straightforward to show that, if type finds it optimal to adopt before time t, then the same is true for type +. It follows that, in equilibrium, the set of adopters at time t is given by all types with greater than some critical value. Let + +(t) be such critical value. The equilibrium conditions are then ct (1 *)+ *+Nt nt
f ()d,
+
where f is the density of . The first equation guarantees that the marginal adopter (type +) is just indifferent between adopting and not adopting at time t: the left-hand side is the (flow) cost of adoption, whereas the righthand side is the benefit from adoption. The second equation is a ‘closure’ condition: it implies that the network size is the measure of all adopters of type greater than the marginal type. The above equations can be combined to yield the following equilibrium condition: ct 1 *F(+) , (15.2) + where F is the cumulative distribution function of . For each time t, and the corresponding value of ct, (15.2) can be solved for + . Each value of + in turn corresponds to a value of n. Therefore, (15.2) induces an equilibrium correspondence E(t) giving the possible equilibrium values nt for each t.3 Although the graph E(t) is a continuous and smooth manifold (Cabral, 1990), the equilibrium correspondence can, in principle, be multi-valued. In fact, in the presence of network effects this would not be a surprising feature.
3
NETWORK EXTERNALITIES AND CATASTROPHIC ADOPTION PATHS
If network effects are non-existent or mild, then (15.2) induces a singlevalued equilibrium correspondence E(t) and a continuous equilibrium adoption path (EAP). This is illustrated in Figure 15.1. The left-hand side
430
The diffusion of new technologies
nt
LHS RHS
1
LHS
E(t )
RHS1 RHS2 RHS3 +
t
Figure 15.1 Continuous adoption path with ‘mild’ network externalities ( 5, 1, * .25) nt
LHS RHS
n** E(t ) LHS RHS1 RHS2 RHS3 +
n* t+
t
Figure 15.2 Strong network externalities and catastrophe adoption path ( 5, 1, * .5) depicts the two sides in (15.2). As can be seen, for each value of t (and ct), there exits a unique solution to the equation. This implies that the equilibrium correspondence E(t) is single valued and there is a unique EAP, namely nt E(t). Consider now the case when network externalities are significant. This case is illustrated in Figure 15.2. The left-hand side of the figure shows that, for values of t slightly greater than the one corresponding to RHS2, several solutions exist to equation (15.2) (in this figure, equation (15.2) corresponds to LHS RHSi). This results in an equilibrium correspondence E(t) that is multi-valued for an interval of values of t. Such equilibrium correspondence is depicted in the right-hand side of the figure. A multi-valued equilibrium correspondence means that there are multiple possible EAPs, in fact a continuum of them.
Diffusion of innovations with network effects
431
Despite this multiplicity of equilibria, it can be readily seen that every EAP is discontinuous at least for some t. In fact, the most reasonable EAP consists of following the lower branch of E(t) up to time t + and then jumping from nt n* to nt n**. In the jargon of topology, the point (t, n*) is a catastrophe point: although the equilibrium correspondence is continuous, a small increase in t implies a discontinuous change in the value of nt. Figures 15.1 and 15.2 suggest that catastrophes are more likely when network externalities are stronger. My first result formalises this intuition: Proposition 1 For given and , a catastrophe point exists if and only if ***. Conversely, for given and *0, a catastrophe point exists if and only if *. A formal proof may be found in the Appendix. In terms of actual behaviour, one would normally not expect to observe a discontinuous EAP like the one suggested above. The shift from n* to n** would very likely occur over a period of time. In fact, if one assumes that potential adopters make their decisions at time t based on the installed base at time t (that is, there is an observation lag ); then it can be shown that, for small , the adoption path follows closely E(t) up to time t + and then moves gradually towards the upper portion of E(t) along a concave path. The result of this process is an S-shaped adoption path.4
4
ALTERNATIVE THEORIES
As I mentioned before, the economics literature has produced a large number of theoretical explanations consistent with the stylised fact of an S-shaped adoption path. Any claim for the worth of a new explanatory theory has to be confronted with competing claims. At the risk of over-generalising, we may classify the different theories into two different categories.5 First, we have the equilibrium diffusion theories based on adopter heterogeneity. These theories are similar to the model I presented above (or vice versa), except for the inclusion of network effects. In words, these theories explain diffusion as a result of adopter heterogeneity. Specifically, an S-shaped adoption path results from the shape of the cumulative distribution function of the adopters’ type. In particular, the steep portion of the adoption path corresponds to a high density of adopters around the relevant valuation parameter.
432
The diffusion of new technologies
The epidemic theories, which are based on some form of imperfect information, provide an alternative explanation for S-shaped diffusion paths. In its simplest form, the epidemic theory assumes that potential adopters become aware of the existence of the innovation by word of mouth. Wordof-mouth dynamics are known to have the dynamics of medical epidemics, where the rate of change is proportional to the product of the number of infected and not-infected agents. This results in an S-shaped path very similar to the empirically observed paths.6 To summarize: an S-shaped adoption path does not require a catastrophe. In fact, it does not even require that there are network externalities at all or that adopters are heterogeneous, so long as there is imperfect information of some sort. Therefore, the simple observation of the aggregate rate of adoption is not sufficient to validate any theory in particular. In the next section, I focus on empirical implications that separate the different theories.
5
TESTING BETWEEN THEORIES
As I have argued in the previous sections, there exist many theories that are consistent with an S-shaped adoption path. Different theories must then be distinguished by observables other than the diffusion path. My second result implies one such test for the case when the intensity of use can be easily measured. As I mentioned before, one natural motivation for the utility function (15.1) is the distinction between stand-alone and network-related benefit. Network benefit is proportional to total use, which in turn is proportional to nt. Based on this observation, different theories have different implications with respect to the time path of average use, at, given by
at
+ (1 * *nt)gt()d() , + gt()d
where gt() is the density of types who adopt by time t. Specifically, we have the following results: Proposition 2 Under equilibrium adoption with heterogeneous adopters, * 0 implies that at is decreasing for all t; * 1 implies that at is increasing for low t and decreasing for high t. Proposition 3 Under epidemic diffusion, * 0 implies that at is constant for all t; * 1 implies that at is increasing for all t. Table 15.1 summarises Propositions 2 and 3.
433
Diffusion of innovations with network effects
Table 15.1
Summary of Propositions 2 and 3
Theory
No net effects
→ →
→
Heterogeneous adopters Epidemic
Net effects
→
→
Installed base (IB)
Pages per machine (PPM)
6m
12000
IB
3m
3000
PPM
0
Year 1970
1980
1990
Source: Farrell and Shapiro (1992); see also Economides and Himmelberg (1995).
Figure 15.3
6
Fax machines in the US: installed base and intensity of use
AN EXAMPLE: FAX MACHINES
In order to test the applicability of my theoretical results, I consider data on the diffusion of fax machines in the USA. Figure 15.3 plots the value of the installed base of fax machines as well as the average use per machine (pages per machine), from the mid-1960s to 1990. The data seem roughly consistent with the theory of diffusion with heterogeneous adopters and strong network externalities. First, around 1987 there was a sharp increase in the installed base, which suggests a catastrophe point in the diffusion of fax machines. Moreover, the time path of usage per machine seems consistent with the prediction of Proposition 2 for the case * 1: the value of at is initially increasing and then decreasing.
434
The diffusion of new technologies
Several qualifications are in order, however. First, the time series in Figure 15.3 is a bit too short to uncover a clear pattern in the evolution of at. I am currently working on trying to extend this series, in the hope of finding stronger results. One problem with extending the series to the 1990s is that serious consideration must be given to the emergence of email as an alternative to fax (including the emergence of electronic faxing). Second, it should be noted that the value of pages per minute (PPM) is not necessarily the best measure of at. In fact, it is not uncommon for fax machines to be shared among several users. For this reason, using PPM as a measure of average use implicitly amounts to assuming that the number of users per machine remained constant throughout the sample period. However, anecdotal evidence suggests that, as the price of fax machines dropped over time, so did the number of users per machine. Figure 15.3 is therefore consistent with the epidemic-theory-cum-network-effects story. In other words, the time path of PPM is consistent with an ever increasing path of at.
7
CONCLUDING REMARKS
Referring to the plethora of theories of S-shaped adoption paths, David (1969: II/13) argues that It would be possible to find some pair of specifications [of F() and ct] that would give rise to a diffusion path of the appropriate shape. Hence, meaningful efforts to distinguish between and verify alternative models of diffusion ought to involve some attempt at direct empirical validation of the component specifications, including the postulated characteristics of the distribution f(X).
The analysis in this chapter suggests that the problem is deeper than that: not only there are multiple functional forms consistent with a given time path; there are multiple theories that would produce identical outcomes. On the positive side, Propositions 2 and 3 suggest that, even with aggregate data only, there are ways of distinguishing between competing theories. To conclude, I should acknowledge that the model in this chapter is based on a somewhat narrow class of innovations, namely communication technologies, where benefits are derived from actual links between potential users. However, as in much of the networks literature, results from the direct network effects model can be extended to the case of indirect effects as well. The crucial point is that Assumptions 1 and 2 (or a variation thereof) hold.
435
Diffusion of innovations with network effects
APPENDIX Proof of Proposition 1 In a continuous equilibrium path (no catastrophe), all values 0 will correspond to the marginal adopter at some time t. The condition determining the marginal adopter is 1 *F(+)
ct +
(15.3)
A necessary and sufficient condition for the EAP to be continuous is that the RHS cut the LHS from above for every t, that is,
|
| | |
, , ct (1 *F()) , , ,
or simply * f ()
ct , 2
for all 0. Substiuting (15.3) for ct and simplifying, we get *(F() f()) 1.
(15.4)
The next step is to show that F f is greater than 1 for some value of .7 The derivative of F f with respect to is
, (F() f ()) f () f () 2 . , It follows that, for sufficiently large, , (F() f ()) 0. , Since, in addition lim (F() f ()) 1,
→
it follows that there exists a such that F() f() 1. It follows that if * is sufficiently large, then the condition (15.4) is violated for some . The second part of the proposition is quite straightforward. For , (15.4) reduces to *
1 1 1. 2 √2$
436
The diffusion of new technologies
Clearly, for given and *0 this condition is violated for sufficiently close to zero. Proof of Proposition 2 use is given by at
Under heterogeneous adopter diffusion, per capita ((1 *) *(1 F(+)))f ()d + f ()d +
Given that is normally distributed, we have at ((1 *) *(1 F(+)))2
f(+) . 1 F(+)
If * 0, then at 2fi(+)/ (1 - F(+)), which is increasing in . Since + is decreasing in t, it follows that at is decreasing in t. If * 1, then at 2f(+). Since is decreasing in t, the value of at follows the value of f(+t): increasing for low values of t, decreasing for high values of t. Proof of Proposition 3 Under epidemic diffusion, the population of adopters at time t is a representative sample from the population of potential adopters. Moreover, type ’s use at time t is given by (1*) *nt. Together, these facts imply that at (1*) * nt. If * 0, then at , which is constant over time. If * 1, then at nt, which is increasing over time.
NOTES *
1. 2.
3.
My first work in this area dates back to a second-year student paper at Stanford (in 1986) that eventually led to Cabral (1990). I am grateful to Paul David, Brian Arthur, and many others who encouraged me on this line of research. Regretfully, I alone remain responsible for all the shortcomings of this and previous related papers. Both appeared in a recent issue of the Review of Economic Studies. Cowan (2005), also included in these proceedings, considers a model of cycles in art appreciation and prices. Although I do not consider the possibility of cycles (I assume that technology adoption is irreversible), our models share the prediction that, over time, consumers will shift between equilibria. Specifically, E(t) is obtained by solving 1 *(1 nt)
where G() is the inverse of F().
ct , G(1 nt)
Diffusion of innovations with network effects 4. 5. 6.
7.
437
Cf Cabral (1990). Notice I do not call this an equilibrium adoption path since, for a period of time, the system is in disequilibrium, gradually moving from a low-adoption to a high-adoption static equilibrium. See Geroski (1999) for a recent survey. Jensen (1982) proposes an interesting variant of the epidemic theory based on imperfect information about the value of the innovation. Specifically, he assumes that adopters differ with respect to their prior beliefs that the innovation is profitable (other than this, he assumes adopters are identical). He shows that, starting from a uniform distribution of prior beliefs, an S-shaped equilibrium adoption path is obtained. One side-result of the above condition is that, for given and , the EAP is continuous if * is sufficiently small.
REFERENCES Cabral, Luís (1990), ‘On the adoption of innovations with “network” externalities’, Mathematical Social Sciences, 19, 299–308. Cowan, Robin (2005), ‘Waves and cycles: explorations in the pure theory of price for fine art’, chapter 7 in this volume. David, Paul A. (1969), ‘A contribution to the theory of diffusion’, Research Center in Economic Growth Memorandum No. 71, Stanford University. Economides, Nicholas and Charles Himmelberg (1995), ‘Critical mass and network size with application to the US FAX market’, NYU Stern School of Business Working Paper No. EC-95-11, August. Farrell, Joseph and Carl Shapiro (1992), ‘Standard setting in high-definition television’, Brookings Papers on Economic Activity (Microeconomics), 1–77. Geroski, Paul (2000), ‘Models of technological diffusion’, Research Policy, 29, 603–25. Jensen, Richard (1982), ‘Adoption and diffusion of an innovation of uncertain profitability’, Journal of Economic Theory, 27, 182–93.
16. Technological diffusion under uncertainty: a real options model applied to the comparative international diffusion of robot technology Paul Stoneman and Otto Toivanen INTRODUCTION The purpose of this Festschrift is to celebrate the contribution of Paul David to various areas of Economics and especially the Economics of Innovation. It is Paul’s work on technological diffusion that has had the most impact upon our work and especially Paul’s seminal contribution, David (1969), that introduced into the literature the probit-type model of diffusion. Prior to this work, diffusion was largely explained as the result of an information-spreading process, whereas, here, for the first time, differences across firms in the returns that are to be realised from the use of new technology was argued to be the central element of the diffusion process. This not only emphasised the importance of user heterogeneity in the process, but also, for the first time, provided the basis by which the standard tools of rational choice, the key elements of economic analysis, could be brought to bear upon the analysis of diffusion. The actual model presented in David (1969) is in many ways very basic. The principle argument is that firms (farms) make the decision whether to adopt a new technology or not on the basis of the cost of adoption relative to the expected profit gain, the profit gain differing across firms (according to firm size). Changes in costs or benefits over time drive the diffusion process. Later modifications in the literature have extended the insight by modifying the adoption condition to introduce an arbitrage as well as a profitability criterion (not only is it profitable to adopt today but is it not more profitable to wait until a later date), added expectations (on technology and prices) to the adoption criterion and also considered the competitive environment in which firms operate (for a literature review see 438
Technological diffusion under uncertainty
439
Stoneman, 2001). There is a further topic, however, that David did not address to any degree, and nor has much of the later literature, and that is the impact of risk and uncertainty on the diffusion process. In this chapter the diffusion of new technology in an uncertain environment is considered via a probit-type model of the David type. Investment in general and the adoption of new technology in particular is a process inherently characterised by risk and uncertainty on both the cost and demand side. As an investment, the potential pay-off to the use of new technology is uncertain because future market revenues and production costs will not generally be known with certainty, while at the same time the cost of adopting a new technology, for example, the price to be paid for acquiring new capital goods at different points of (future) time, is also unlikely to be known with certainty. It may even be the case that new technology investments are inherently more uncertain than investment in general, for example, (1) new technology may well require additional investments in human capital whereas investments in existing technology may not and (2) new technology investments may have a higher degree of irreversibility for there may only be very limited second-hand markets for such technologies. Although the modelling of technological diffusion has made considerable advances in the last 10 years (Stoneman, 2001) uncertainty has played little part in this. The dominant treatment of uncertainty in the diffusion literature goes back to Mansfield’s (1968) seminal but flawed (see Stoneman, 1983) contribution, with a more suitable but dated means variance treatment in Stoneman (1981), and a further set of papers by Jensen (for example, Jensen, 1982; 1983). On the other hand, the investment literature has made good use of models that place uncertainty and irreversibility as the centrepiece of the analysis (for a recent example see Caballero and Engel, 1999) although the empirical literature has placed little emphasis upon analysing the effects of uncertainty directly (but see Driver and Moreton, 1991; Pindyck and Solimano, 1993; Leahy and Whited, 1996). It is thus somewhat surprising that to date the diffusion literature has not followed this lead. In continuing research, for which this chapter should be considered as a progress report rather than a final outcome, we attempt to fill this gap by building a model of technology diffusion that is based upon an optimising firm (project holder) facing an irreversible investment in a new technology in a world of uncertainty and then aggregating up to get an economy-wide level expression for the diffusion of the technology. The model is then applied to an inter-country data-set upon the diffusion of robot technologies.1 Our approach is largely based upon the real option approaches which were developed and have become dominant in the investment literature (see, for example, Dixit and Pindyck, 1994). To the best of our knowledge, real option
440
The diffusion of new technologies
methods have rarely been used to analyse diffusion phenomena and, in particular, we know of no empirical work into diffusion based upon a real options approach. The approach allows a relatively direct way of incorporating well-defined measures of uncertainty into the estimating equation. The exercise we undertake has five main potential pay-offs. Initially, of course, it will provide insight in to the impact of uncertainty on the diffusion process, how diffusion may be modelled under uncertainty and the utility of real option approaches for this purpose. Secondly, by exploring the real options approach we are able to provide a new and more rigorous theoretical foundation for some widely used empirical models of technology diffusion and show that these can be thought of as the aggregate implications of individual optimising behaviour where an ad hoc structure has been placed upon the adoption hazard. Third, although there is some literature that compares diffusion across countries, that literature is quite sparse and our comparative international analysis also makes a relevant contribution in this area. Fourth, there has been a long-running argument in macroeconomic policy discussion as to whether macroeconomic stability encourages investment (see, for example, Caballero, 1991; Driver and Morton, 1991; Pindyck and Solimano, 1993), and the work reported upon here provides further evidence for this policy debate in that a by-product of the model that we construct gives some insight into whether the rate of technological change in an economy as measured by the speed of adoption of a particular technology is related to several indicators of macroeconomic stability. Fifth, on the basis of the empirical results we are able to conduct policy experiments to explore the effects of changing the volatility of the environment in which firms make investment decisions. In the next section we provide an introduction to robot technology. In the third section we develop the real options model of technology choice and then from that model construct an aggregate diffusion curve for an economy. In the fourth section we discuss the data used to test the model. In the fifth section we discuss econometric issues. In the sixth section we present the results of the estimation and discuss their implications. The work that we report upon in this paper is still ongoing, thus to a considerable extent the results that we present as this stage should be considered preliminary. We then use the estimated parameters in the sixth section to conduct the policy experiments in the seventh section. The eighth section contains conclusions.
ROBOT TECHNOLOGY The International Federation of Robotics (IFR) has created an international standard (ISO TR 8373) for robot technology with robots customarily
Technological diffusion under uncertainty
441
classified into standard and advanced robots and by the application area. International statistics2 on the use of industrial robots are compiled by the United Nations (UN) and the IFR, however, for the period 1981–93 only aggregate data is available. We use this data to measure the inter-country stock of robots. These statistics cover 28 countries, 16 of which were included in our actual sample,3 although due to gaps in the data our panel is unbalanced, yielding a total of 161 observations. The 16 countries are listed together with summary statistics in Tables 16.2a and 16.2b (see below); the tables also report the observation period for each country. It is estimated (the statistics are not water-tight) that there were some 610 000 robots in use worldwide at the end of 1993. The rate of growth of this stock has been fast, ranging from the 16–23 per cent per annum. recorded at the turn of the decade to latest levels of 6–8 per cent per annum. The average yearly change in the robot stock varies between almost 30 000 in Japan to 41 in Norway. As evidence that robot adoption is significant, but different from aggregate manufacturing investment, in 1993 robot investment amounted to 12 per cent of total machine tool investments in the USA, to 11 per cent in both Germany and the UK, and to 6 per cent in France. Japan is by far the largest user of robots whether measured by absolute (some 60 per cent of world stock) or relative numbers (in 1993 Japan had 264 advanced robots per 10 000 employees in manufacturing when the country with the second highest density (Singapore) had 61). Robots are used in several industries, and perform a variety of tasks. Worldwide, the traditional ‘vehicle’ for diffusion of robots has been the transport equipment industry (especially the motor vehicle industry), but lately, for example in Japan, the electrical machinery industry has adopted more robots. The major application areas are welding, machining and assembly, with the leading application area varying over countries. Although it would certainly be beneficial to have more detailed country-level data on the composition of the robot stock, and its use, such country-level idiosyncrasies are to a great extent constant over time, and can be captured by country-level controls in the econometric model. Intuition suggests that investments into a new technology such as robots may be more volatile than aggregate manufacturing investment. The reason is that as the degree of technological uncertainty is greater (leading to a greater variance of future revenue streams), such investments are more sensitive to changes in other variables (such as prices and interest rates) that affect investment decisions. To check whether this is the case, we compared manufacturing investment volatility to that of robot diffusion volatility in the OECD countries of our sample (thereby excluding Singapore, Switzerland and Taiwan, for which it proved difficult to obtain a comparable investment series at this point). To be able to compare two different
442
The diffusion of new technologies
Table 16.1 Volatility of robot adoption in comparison to manufacturing investment Country
Ratio of robot diffusion coefficient of variation to manufacturing investment coefficient of variation
Australia Austria Denmark Finland France Germany Italy Japan Norway Spain Sweden UK USA
4.15677 5.43715 3.340727 1.91579 5.724699 2.729275 4.548866 1.73117 2.944321 2.780212 1.256557 1.422169 5.592314
forms of investment that are measured differently (manufacturing investment in monetary terms, robots in units), we employed the coefficient of variation. Comparing for each of the 13 countries the two series over the robot observation period (see Table 16.1) we found that for all countries, the coefficient of variation of robot diffusion was larger than that of aggregate manufacturing investment (which includes robots, thereby biasing these figures upwards). The mean ratio of robot coefficient of variation to that of manufacturing investment was 3.60, with a minimum of 1.26 (Sweden) and a maximum of 5.74 (France). We read this as evidence that robot diffusion is indeed more volatile than aggregate manufacturing investment, as discussed in the Introduction.
THE MODELLING FRAMEWORK In this section we develop a theoretical model of technology adoption under uncertainty that may be applied to our data upon robot diffusion. The initial unit of analysis is the project (as opposed to the more usual unit of the firm) where a project corresponds to investment in one unit of robot technology. As our data does not contain observations upon individual projects, we aggregate up from the project level to the country level, the model predicting the number of robots installed in each country at time t which corresponds to the data available.
Technological diffusion under uncertainty
443
The Adoption Timing Decision We employ a real options approach adapted from a model proposed by Dixit and Pindyck (1994: 207–11). Defining Pjt as the cost of a unit of robot technology in country j in time t, assumed the same for all projects i, and Rijt as the annual gross profit increase generated by project i in country j in time t, we assume that both Pjt and Rijt are uncertain but exhibit geometric Brownian motion such that, dRijt /Rijt Rij dt Rijdz
(16.1)
dPjt /Pjt Pj dt Pjdz
(16.2)
and that
where dz is the increment of a standard Wiener Process. Note that the s measure the expected rates of growth or drift of the variables over time while the s show the volatility or uncertainty attached to the variable. It is through the terms that in this chapter we explore the effects of uncertainty on the diffusion process. We assume that Pjt and Rijt are independent covariates with zero covariance. We also assume that Pjt and Rijt are independent of the (existing) number of robots in use and actual dates of adoption. In terms of the diffusion literature, this is the same as assuming that there are no stock and order effects in the diffusion process (see Karshenas and Stoneman, 1993). The resulting model thus falls within the class of probit or rank models of diffusion introduced by Paul David and in which different rates of adoption across countries will reflect the different characteristics of those countries, the characteristics being exogenous to the diffusion process. Defining rjt as the riskless real interest rate in country j in time t, a project i, which has not previously been undertaken, will be undertaken (started) in time t if Rijt /Pjt "ijt ij (rjt Rij)/(ij 1)
(16.3)
where "ijt is the threshold ratio of profit gains relative to the cost of acquisition above which project i will be undertaken in time t and below which it will not (this being allowed to be time dependent for generality), and ij is the larger root of the quadratic equation 2 , 2 ) ( 1) ( ) ( r ) 0 0.5(Rij Pj ij ij Rij Pj ij Pj jt
(16.4)
444
The diffusion of new technologies
enabling us to write (16.3) more generally as (16.5). 2 , 2, , ) Rijt /Pjt "ijt F(rjt, Rij Pj Rij Pj
(16.5)
Following Dixit and Pindyck we assume that rjt Rij 0 and rjt Pj 0 which implies that ij 1, and then using (16.3) and (16.4) we may deduce that F2 0, F3 0, F5 0 but the signs of F1 and F4 are indeterminate. Thus, 2 , 2 ) will lead to increases in the threshold increases in uncertainty (Rij Pj value "ijt and an increase in the drift rate of increase of robot prices (Pj) will reduce the threshold, however, the impact of an increase in the interest rate and the drift (rate of increase) of profit gains (Rij) are indeterminate. Basically, although the direct impact of increases in the latter two parameters on (rjt Rij) is clear, their impact on ij /(ij 1) is of the opposite sign. If there is no uncertainty so that Rij Pj 0 then (16.3) collapses to Rijt /Pjt "ijt rjt Pj
(16.6)
which is the standard intertemporal arbitrage condition for adoption of a new technology in time t (see Ireland and Stoneman, 1986). In this case the impact of changes in rjt and Rij can be clearly signed as positive and zero respectively. 2 and/or 2 increase, then the The main result thus generated is that if Rij Pj threshold value for the ratio of profit gains relative to the cost of acquisition, "ijt, will also increase. This does not necessarily imply however that increases in uncertainty will lead to a slower diffusion process. To illustrate the point we consider the probability that a randomly chosen project i will be undertaken in time t given the adoption rule (16.3) and the paths of Rijt and Pjt as given by (16.1) and (16.2). The basis of the argument is that given the geometric Brownian motion assumed for P and R, the probability in time t that the realised or observed ratio of R to P will exceed any given 2 and 2 and in such a way that value will also be dependent upon on Rij Pj one cannot predict a priori whether the probability that (16.5) will be met 2 and 2 . will be increasing or decreasing in Rij Pj From Dixit and Pindyck (1994: 71–2) we know that if a variable x follows geometric Brownian motion with drift at the rate and volatility then x at any time will be lognormally distributed with mean x0exp(t) and variance x02exp(2t)[exp(2t 1] where x0 is the value of x at time zero. Aitchinson and Brown (1957) show that the ratio of two lognormally distributed variables is also lognormally distributed with a mean equal to the difference of the two means and a variance equal to the sum of the two variances. Given (16.1) and (16.2) above, we may then state that Rijt /Pjt in time t will be lognormally distributed with mean ijt (R0 /P0)exp[(Rij Pj)t]
Technological diffusion under uncertainty
445
2 2 [exp( 2 2 ) 1]. Clearly, the greater is 2 and/or and variance ijt ijt Rij Pj Rij 2 Pj the larger will be the variance of Rijt /Pjt in time t. Using L to represent the lognormal distribution function, we may then state that the probability that a project i will be initiated in time t is given by the probability that Rijt /Pjt "ijt, written as Pr{Rijt /Pjt "ijt} where 2) Pr{Rijt /Pjt "ijt} 1 L("ijt | ijt, ijt
(16.7)
From the properties of the lognormal we may immediately state, ceteres paribus, that Pr{Rijt /Pjt "ijt} is decreasing in "ijt, thus (1) the lower is the threshold return to adoption the greater is the probability of adoption, (2) the greater is ijtthe greater is the probability of adoption, but (3) the 2 depends upon the value of " and cannot be impact of changes in ijt ijt signed a priori. 2 and 2 are The implications for the date of adoption of changes in Rij Pj thus unclear in that an increase in these volatility parameters leads to an increase in "ijt which reduces the probability of meeting the adoption criterion in time t, but it also leads to an increase in the variance of Rijt /Pjt which may impact positively or negatively upon the probability of adoption in time t. One cannot predict a priori the overall effect. One cannot, thus, make any predictions of the impact of increases in uncertainty upon the date of adoption of a new technology. In addition, although for a given "ijt one may predict that the greater is ijt the greater is the probability of adoption, this does not imply that increases in (Rij Pj) will lead to increased adoption. This is because an increase in (Rij Pj) will change both "ijt and 2 as well as and we are unable to predict the overall impact of such ijt ijt changes. Given that the modelling framework cannot predict a priori whether increases in volatility (or drift) of the cost and returns of adopting new technology will have a positive or negative impact upon the date of adoption, there is little advantage in attempting to empirically explore the relationship between diffusion and uncertainty in terms of the impact of uncertainty on a diffusion curve expressed as a function of time. In the absence of any clear predictions from the model on the impact of uncertainty upon timing, one can never use the empirical results to reject hypotheses relating timing to uncertainty, in that positive, negative or no relationships are all equally possible. However, we do have one clear prediction from the theory – that uncertainty as measured by the volatility of the costs and returns impacts positively upon the critical adoption threshold ("ijt). It is around this prediction that the rest of this chapter is constructed. Essentially, we concentrate upon adoption condition (16.3), that is, Rijt /Pjt "ijt, but rather than replacing
446
The diffusion of new technologies
Rijt /Pjt by a stochastic function of time and looking at the probability that at any time t a project will be started, we instead explore whether the diffusion patterns observed are consistent with the prediction of (16.3) that a project i will be adopted at the first point in time at which the realised ratio of Rijt /Pjt is greater than or equal to "ijt. In essence, therefore, we estimate a diffusion curve not as function of time but as a function of realised Rijt /Pjt. In doing this we model "ijt inter alia as a positive function of volatility and jointly test this hypothesis. It should be noticed, however, that even if we cannot reject that the threshold is positively related to volatility, this will not necessarily imply that uncertainty slows diffusion when diffusion is considered solely as a function of time. We would instead only be able to say that, given the realised path for Rijt /Pjt, diffusion has been slowed by uncertainty. The basic model of firm behaviour we therefore have in mind is that for a project i the costs and returns follow geometric Brownian motion as detailed above. The decision-maker knows this and is aware of size of the drift and volatility indicators. From this he/she is able to calculate a threshold value of the ratio of returns to cost. As time proceeds a realised time path for Rijt /Pjt is mapped out. Eventually (or not, for some projects) Rijt /Pjtwill equal or exceed the critical value (for the first time) and the decision-maker will, upon that criterion being met, commit to the project. In pursuing this basic framework the obvious issue arises as to how the decision-maker obtains estimates of drift and volatility. It would, of course, be possible to allow some sort of Bayesian updating process of assumed priors but this is not really consistent with the modelling detailed above. In our empirical explorations we have explored the properties of the costs of and (determinants of) returns to robot technologies in each of the countries in our sample and are unable to reject the hypothesis that these costs and returns follow geometric Brownian motion processes with constant drift and volatility. Given the constancy of drift and volatility, very few realised observations would be required by a decision-maker to get a true estimate of the drift and volatility or, on average, decision-makers would generate true estimates at an early date. We thus assume that decisionmakers from the earliest days of the diffusion process have such estimates and these estimates are those that are apparent when estimated from data upon the diffusion process from its start date to the end date of our sample. The Hazard Rate Equation (16.5) above is the condition to be met if project i is to be initiated in country j in time t. This condition may be written as (16.8) Rijt /Pjt "ijt 0.
(16.8)
Technological diffusion under uncertainty
447
To move to a model that is appropriate to the macro data available to us, and as we are unable to measure Rijt directly, we assume that Rijt may be written as Rijt exp (aij &kbkXkjt)
(16.9)
where the term &kbkXkjt is a weighted sum of a number k of macroeconomic variables that are the determinants of the return to robots on average across projects, and the aij term picks up the heterogeneity of returns to different projects arising from various real-life factors that, although they may be known with certainty to individual decision-makers, cannot be incorporated into the model. This may include such factors as firm-specific risk, expectations and characteristics. If the Xkjt terms all follow absolute Brownian motion (with drift and volatility Xkj and Xkj), as we can empirically show that the variables that we include in Xkjt do, then Rijt will follow 2 will also be weighted sums geometric Brownian motion so that Rij and Rij 2 over k of Xkj and Xkj and independent of i. "ijt will thus also be independent of i and may be written as (16.10) 2 , 2, & b "ijt F(rjt, &kbk2Xkj Pj k k Xkj, Pj).
(16.10)
Substituting from (9) into (8) and taking logs then yields the adoption condition that &kbkXkjt lnPjt ln"jt aij 0
(16.11)
Following Karshenas and Stoneman (1993) for each country j, assuming that the distribution of ai, V(ai), remains invariant across the projects over time and is distributed independently of &kbkXkjt lnPjt ln"jt, one may write that the probability of a project i, that has not been started previously, being undertaken in the small time interval {t, t dt} that is, the hazard rate, is given by (16.12) hijt Prob{&kbkXkjt lnPjt ln"jt aij 0}
V(&kbkXkjt lnPjt ln"jt)
(16.12)
which is independent of i, and thus may be written as hjt. The Aggregate Diffusion Curve Define Sjt as the number of robots installed or projects being undertaken in country j in time t, S*jt as the number of robots that would be installed if
448
The diffusion of new technologies
all projects in country j for which robot use were feasible in a technological as opposed to an economic sense were being pursued and j as the rate at which past robot investments disappear from the stock on account, perhaps, of the churning of firms (a consequence of our assumption that R follows a geometric Brownian motion is that it is bounded below to be non-negative and a robot will be operated forever and thus physical obsolescence is ruled out). Then immediately we may write (16.13) Sjt (1 j)Sjt 1 hjt{S*jt Sjt 1}
(16.13)
where, as defined above, hjt is the hazard rate, that is, essentially the probability that any project in country j not started by time t will be initiated at time t. Two issues with respect to this specification merit discussion. First, usually in diffusion models the attrition effect is implicitly allowed to be zero, probably on the basis of an unspecified assumption that either such effects do not exist or that any such effects are immediately counteracted through greater gross investment in robots. Here, instead, we rigorously test as to whether this ‘depreciation term’ is non-zero. Secondly, S*jt plays a rather different role than is often assumed in diffusion analysis. It is often assumed in such models that as t tends to infinity that Sjt tends to S*jt. Here, however, this is not necessarily so. Here the diffusion process will only continue as long as hjt is positive and hjt may tend to zero well before Sjt S*jt. Equation (16.13) is a standard logistic diffusion curve (see Stoneman, 2001) with the rate of diffusion (setting j equal to zero) given by the hazard rate, hjt, as determined by (16.12). A useful variant on (16.13) can be obtained using the approximation that (x y)/y logx log y which yields the Gompertz diffusion curve (16.14) lnSjt lnSjt 1 j hjt{lnS*jt lnSjt 1}.
(16.14)
It is the Gompertz curve upon which we concentrate below.4 One may note that diffusion curves such as (16.13) and (16.14) are usually based upon information spreading or epidemic arguments (see Stoneman, 2001) but in this case that is not so. When they are so based, the lagged stock usually appears on the on the right-hand side as a proxy for current stock. In the current model the lagged stock appears in its own right. Detailed Model Specification There are three remaining issues that need to be settled before the model can be operationalised. The first is the modelling of S*jt. We take two
Technological diffusion under uncertainty
449
different approaches to this. The simple approach is to assume, given that robots are generally only used in the manufacturing sector, that S*jt is proportional to manufacturing output, INDjt, in country j such that S*jt INDjt .
(16.15)
However, it is very likely that, for example, the optimal robot stock varies over countries in a manner that is country specific, time-invariant and unobservable to us.5 This suggests the use of a random effects specification such as S*jt INDjte-j
(16.16)
where, E[-j] 0, var(-j) 2-2 and is a measure of the variance share of -j. Such an approach, however, implies that in the estimating equation the country specific time-invariant error term is multiplied by hjt. This creates a random coefficients – type (equi-correlated) model that cannot be estimated with standard methods. It is thus necessary to resort to a simulation estimator (see below) to estimate this model. The second issue to address is to list the macro variables that are included as determinants of Rijt. We initially assume that the potential profit gain from installing robots is determined by real gross domestic product, GDP, and the share of investment in GDP. The logic is that demand conditions as measured by GDP will be a main determinant of profit gains from adoption, whereas I/GDP will reflect the general investment climate and, thus, also the profitability of robot adoption. Both variables are expected to carry positive coefficients. In addition, because the rate of inflation appears in a number of the investment and uncertainty studies referred to above as a key measure of uncertainty or stability, for the purposes of empirical exploration we have further included the growth rate of the price level of manufactured products, i, as a proxy for the rate of inflation, into the empirical model (although not the actual price level due to difficulties in cross-country comparison). Finally, we need to be specific as to the functional form of the hazard function. Given our limited sample size, we decided in favour of parameterised models and have assumed that the hazard function is exponential (we test the assumption of an exponential functional form against that of a Weibull hazard). We also assume that the baseline hazard takes an exponential functional form (to guarantee non-negative hazards) and is linear in its arguments. Additionally, to simplify the model we assume that F() in (16.10), which specifies the threshold level for investment, is linear in its arguments.
450
The diffusion of new technologies
Bringing together all the parts of the model with these assumptions leads to the estimating equation (16.17) into which the more complex version of the expression for S*jt is incorporated.
ln
Sjt
Sjit 1
d exp(0 1GDPjt 11GDPj 12GDPj 2IGDPjt 21IGDPj 22IGDPj
3lnPjt 31Pj 32Pj 4rjt 5i) (lnINDjt -j lnSjt 1) (1 )jt
(16.17)
The analysis above suggests that d 0, 1 0, 12 0, 2 0, 22 0, 3 0, 31 0, 32 0, and 0 with other parameters unsigned.
DATA SOURCES In the Appendix we discuss data sources in more detail. Here we provide an overview. The sources and nature of the international data upon the robot stock was discussed in the second section above. Robot price data were not available on an individual country basis, but we located such data for Germany (and Italy). The results that we present are based on the German data. The prices were converted into real US dollars using the respective consumer price index and the yearly (average) exchange rates as reported in the International Monetary Fund’s (IMF’s) International Financial Statistics Yearbook (IFS). The data on macro variables (see Tables 16.2a and 16.2b) comes mainly (but not solely, see the Appendix) from Penn World Tables Mark 6 (Summers and Heston, 1991). The relevant variables are GDP, the investment share of GDP, the real interest rate (taken from IFS statistics) and the rate of inflation measured using the Penn World Tables’ index on prices in manufacturing. Tables 16.2a and 16.2b reveal that there is considerable variation in the macro variables over countries. The mean of GDP growth varies between 11.72 per cent (Taiwan) and 4.4 per cent (Norway). The mean of the real interest rate varies between 0.69 per cent (Switzerland) and 7.45 per cent (Taiwan) whereas the lowest mean inflation rate (note that this is the price level of manufactures, not the consumer price index) is found in USA ( 2.37 per cent) and the highest in Spain (5.3 per cent). There are also considerable inter-country differences in the volatility measures, but the different measures do not necessarily move in parallel. The structure of the model requires for estimation purposes not only estimates of variables in levels but also estimates of the js and js for
451
271.7382 41.1980 98.6655 21.5569 74.6674 14.6161 67.5622 13.2520 948.4007 86.0395 933.3297 228.9880 729.1065 158.3957 1 739.0154 468.7260 61.6330 7.7665 28.9984 10.4600
Australia 1985–91 Austria 1982–92 Denmark 1982–92 Finland 1982–92 France 1988–92 Germany 1982–92 Italy 1982–92 Japan 1982–92 Norway 1983–92 Singapore 1982–92
0.0594 0.0310 0.0644 0.0163 0.0633 0.0191 0.0535 0.0506 0.0622 0.0191 0.0685 0.0196 0.0629 0.0141 0.0787 0.0166 0.0461 0.0391 0.1064 0.0460 0.0226 0.0258 0.0291 0.0319 0.0457 0.0595
0.0821 0.0801 0.0843 0.1080 0.0802 0.1284
0.0242
0.0749 0.0467
0.0279
0.0822
0.0785
0.0347
GDP growth s.d.
0.0812
GDP growth
GDP (millions of 1985 US dollars)
Country
GDP growth percentage
Country-level descriptive statistics
Table 16.2a
25.6375 2.2772 24.7909 1.7335 21.2909 2.0364 30.5182 3.4713 26.8200 1.0208 23.9727 0.8855 24.1364 0.6772 34.4273 3.3142 27.7000 3.7205 35.8000 3.6927
Investment share of GDP (I/GDP) percentage
0.0121
0.0811
0.0965
0.0114
0.0655
0.0083
0.0591
0.0522
0.0077
0.0155
0.0549
0.0976
0.0127 0.0021
0.0914
0.0096
0.0802 0.0641
0.0064
0.0183 0.0757 0.0035 0.0540 0.0015 0.0843 0.0315 0.0959 0.0040 0.0512 0.0011 0.0332 0.0029 0.0341 0.0094 0.0443 0.0247 0.1011 0.0049 0.0808
I/GDP growth s.d.
0.0073
I/GDP growth
Percentage change in I/GDP
0.0124 0.0899 0.0211 0.1334 0.0196 0.1302 0.0072 0.1235 0.0084 0.0742 0.0216 0.1283 0.0177 0.1044 0.0158 0.1292 0.0110 0.1029 0.0042 0.0540
Inflation
452
GDP (millions of 1985 US dollars)
387.2430 87.2258 128.5969 24.3775 117.1019 23.1568 130.6258 44.1117 742.3100 153.1283 4 250.0572 1 370.0140
Country
Spain 1984–92 Sweden 1982–92 Switzerland 1983–92 Taiwan 1982–90 UK 1982–92 USA 1982–92
0.0737 0.0242 0.0545 0.0210 0.0606 0.0190 0.1172 0.0302 0.0605 0.0227 0.0577 0.0224
GDP growth percentage
0.0242 0.0349 0.0252 0.0251
0.1306 0.0699 0.0749 0.0016
0.0273
0.0705 0.0750
0.0262
GDP growth s.d.
0.0922
GDP growth
25.3778 3.5643 21.4909 2.1741 30.9800 2.6599 22.5667 2.1575 17.8727 1.6298 22.0833 1.8004
Investment share of GDP (I/GDP) percentage 0.0240 0.0607 0.0045 0.0771 0.0064 0.0542 0.0205 0.0947 0.0189 0.0642 0.0038 0.0706
Percentage change in I/GDP
0.0771 0.0638 0.0042
0.0023 0.0016
0.0955 0.0027
0.0214
0.0800
0.0833
0.0036 0.0035
0.0546
I/GDP growth s.d.
0.0117
I/GDP growth
0.0531 0.0948 0.0027 0.1129 0.0226 0.1322 0.0135 0.0831 0.0136 0.1082 0.0237 0.0162
Inflation
Notes: First column gives the country name and period used in estimation (which is due to differencing one year less than the observation period). In most cases, robot data limits the observation period. Each column gives the country level mean and s.d. of the respective variable. Note that, for example, the GDP growth percentage and GDP growth differ because the former is calculated over the robot stock observation period only, the latter over a longer period (defined in the Appendix). For the drift (growth) and volatility variables, no s.ds are reported as they are by definition zero.
(continued)
Table 16.2a
453
1 229.3750 402.7757 667.6364 566.1875 301.4545 193.6958 494.2727 342.9624 8 380.2000 2 071.8759 17 533.1818 12 165.3533 7 573.3636 5 451.8190 167 296.6364 111 333.3099 422.4000 128.8825 807.4545 798.0270
Australia 1985–91 Austria 1982–92 Denmark 1982–92 Finland 1982–92 France 1988–92 Germany 1982–92 Italy 1982–92 Japan 1982–92 Norway 1983–92 Singapore 1982–92
154.2500 50.1960 150.0909 101.2496 48.4545 29.7098 92.3636 37.4039 1 289.0000 180.4896 3 371.8182 1 782.6613 1 513.3636 685.0799 29 859.8182 14 306.0971 42.6000 19.0916 189.5455 238.4023
0.1506 0.0683 0.3091 0.0808 0.2216 0.1095 0.3093 0.1797 0.1811 0.0638 0.2582 0.0856 0.3307 0.1996 0.2556 0.1040 0.1345 0.0970 0.5487 0.4845
Percentage change
Robot stock (No. of robots)
Country
Change in robot stock
Country-level descriptive statistics
Table 16.2b
98.9510 31.0942 90.4676 29.8562 90.4676 29.8562 90.4676 29.8562 114.1126 29.4699 90.4676 29.8562 90.4676 29.8562 90.4676 29.8562 93.0191 30.1806 90.4676 29.8562
Robot price RP (1985 US dollars)
0.2144 0.2016 0.2194 0.2238 0.2150 0.3636
0.0810 0.0516 0.0962 0.0333 0.0817 0.0457
0.2102
0.0760 0.2182
0.2007
0.0497
0.0756
0.2139
RP s.d.
0.0832
RP drift
6.3238 1.3662 4.6629 0.9528 7.0989 1.4206 6.1285 3.0889 6.0377 0.4904 3.8156 1.1037 4.6662 1.8523 3.7269 0.7464 5.5871 1.4845 5.7492 5.6178
Real interest rate (percentage points)
59.6261 2.1126 13.9798 4.4845 11.4156 1.0931 13.8356 1.1285 176.1589 3.8855 282.7428 30.7243 2.5276 0.2000 472.7758 65.7128 9.1056 1.2137 7.0426 2.9696
Industrial production (millions of 1985 US dollars)
454
Robot stock (No. of robots)
1 642.1111 1 006.8790 2 781.2727 1 106.0298 868.5000 695.1109 462.4444 435.5362 4 301.5455 2 150.9057 2 5 849.8333 1 4 659.5393
Country
Spain 1984–92 Sweden 1982–92 Switzerland 1983–92 Taiwan 1982–90 UK 1982–92 USA 1982–92
342.2222 217.3671 311.3636 89.7700 197.7000 132.3976 143.5556 111.3217 625.9091 127.8870 3 468.6667 1 890.8775
Change in robot stock
0.2326 0.0287 0.1270 0.0318 0.3335 0.1330 0.7961 0.8907 0.2151 0.1345 0.1787 0.1376
Percentage change
96.0272 30.3798 90.4676 29.8562 93.0191 30.1806 88.0445 32.7886 90.4676 29.8562 91.6561 28.7630
Robot price RP (1985 US dollars)
0.1924 0.2668 0.2045 0.2447 0.0127
0.0422 0.0576 0.0686 0.0035
0.2199
0.0877 0.0524
0.2308
RP s.d.
0.0902
RP drift
6.4693 2.7346 5.8330 3.9539 0.6964 1.0631 7.4458 1.8341 4.6691 1.4663 4.8861 1.6758
Real interest rate (percentage points)
85.6590 5.5779 26.0212 1.6903 28.3199 2.8359 53.4771 10.2842 146.1418 9.3619 800.3874 221.5590
Industrial production (millions of 1985 US dollars)
Notes: First column gives the country name and period used in estimation (which is due to differencing one year less than the observation period). In most cases, robot data limits the observation period. Each column gives the country level mean and s.d. of the respective variable. Note that, for example, the GDP growth percentage and GDP growth differ because the former is calculated over the robot stock observation period only, the latter over a longer period (defined in the Appendix). For the drift (growth) and volatility variables, no s.ds are reported as they are by definition zero.
(continued)
Table 16.2b
Technological diffusion under uncertainty
455
several variables, in particular, GDP, the investment share of GDP (I/GDP), and robot prices. The measures of growth rates and volatility for each of these variables we get by assuming that potential adopters know the true growth rates and volatilities over identical periods in each country (the length of the period is 1960–92, but modest changes in this make little difference), and apply these in their decision-making. Then, relying on the assumption that each of the country-specific time series is a random walk, we estimate the country-specific growth rate and the volatility using maximum likelihood (we have tested the null hypothesis of a random walk, and are unable to reject it for any of the series. See the Appendix).
ECONOMETRIC ISSUES In estimating (16.17) there are several relevant econometric issues. In particular we note the following. Spurious versus True State Dependence It is well known that initial conditions (we do not observe robot diffusion from T 0) may be correlated with time-invariant unobservables (Heckman, 1981). We control for this initially by allowing the lagged stock of robots to be correlated with -j. Note, however, that we do observe country-wise robot stocks ‘almost’ from the beginning (apart from in France): this is likely to reduce the problem considerably. As can be seen from Table 16.3, for 12 out of 16 countries, the first observed stock accounts for less than 13 per cent of final observed stock and for some, substantially less. For four countries, the first observed stock is over 20 per cent of the final observed stock. These countries are Australia (29.97 per cent, 1st observation year 1984), France (40.44 per cent, 1987), Norway (26.04 per cent, 1982), and Sweden 24.73 per cent, 1982). Together, these countries account for 4 per cent of sample robot stock at the end of 1992. In future work, it is our intention to use the correction suggested by Heckman, and regress the initial (first observed) stock of robots on observables, but we have not done so here. Endogeneity of Price Terms Although it may be plausible to think that individual firms take prices as given when contemplating the adoption of robots, it is less plausible to assume that the price of robots is exogenous at the country level. This
456
Table 16.3 Country
Australia Austria Denmark Finland France Germany Italy Japan Norway Singapore Spain Sweden Switzerland Taiwan UK USA
The diffusion of new technologies
Country-level initial stocks and final stocks of robots Initial stock of robots
Final stock of robots
Ratio of initial to final stock of robots
1st year of observation/no. of observations
528 57 51 35 4 376 2 300 450 21 000 150 5 433 1 125 73 1 713 6 000
1 762 1 708 584 1 051 10 821 39 390 17 097 3 49 458 576 2 090 3 513 4 550 2 050 2 217 7 598 47 000
0.2997 0.0334 0.0 873 0.0333 0.4044 0.0584 0.0263 0.0601 0.2604 0.0024 0.1233 0.2473 0.0356 0.0005 0.0938 0.1277
1984/7 1981/11 1981/11 1981/11 1987/5 1981/11 1981/11 1981/11 1982/10 1981/11 1983/9 1981/11 1982/10 1981/9 1981/11 1981/11
would seem to be the case at least for the countries with higher adoption rates (Japan being the prime example). We will test for the exogeneity of robot prices in future work but have not done so here. Aggregation Our current framework implicitly assumes homogenous firms up to the iid. error term in the hazard function. In future work we plan to exploit the knowledge that (1) large firms usually adopt earlier (for example, Rose and Joskow, 1990; Karshenas and Stoneman, 1993) and (2) the firm size distribution is (close to) log-normal, and thus use a simulation estimator that allows us to aggregate over unobserved firm size differences, where the mass of the distribution is given by industrial output, but have not done so here. This will affect both our hazard rate, and our estimate of S*jt. Estimation Methods We use a generalised method of moments estimator (GMM) and a method of simulated moments estimator (MSM) respectively, for models with and without a country-specific time-invariant error term. Both
457
Technological diffusion under uncertainty
estimators minimise #jt in equation (16.18), thereby matching the predictions as closely as possible with observations (ln(Sjt /Sjt 1)), where (16.18) is derived from (16.17) above.
#jt ln
Sjt
Sjit 1
d exp(0 1GDPjt 11GDPj 12GDPj 2IGDPjt 21IGDPj 22IGDPj 3lnPjt 31Pj 32Pj 4rjt 5i) (lnINDjt -j lnSjt 1) (1 )jt
(16.18) A no equi-correlation assumption is identical to imposing 1 on the data. Currently we use S 30, where S is the number of simulations, for the MSM estimator unless otherwise stated. Instruments Both estimators require instruments. Berndt et al. (1974) show that optimal instruments are of the form A*(xjt) GoD(xjt)+. 1
(16.19)
where is a vector of parameters, xjt is a data vector, D(xjt) E[,#jt /, |xjt]
(16.20)
. E[#jt#+jt |xjt].
(16.21)
and
These, however, suffer from reliance on functional form assumptions: if there is functional form misspecification, the resulting estimator is consistent but inefficient (Newey, 1990). Newey (1990) has proposed instruments that are asymptotically optimal even under functional form misspecification. Currently we use (untransformed) explanatory variables as instruments. Generated Regressors Our drift and volatility variables are generated, leading to biased standard errors. In future work we intend to correct these as suggested by Pagan (1984) but have not done so here.
458
The diffusion of new technologies
RESULTS Our estimation results are presented in Table 16.4. We experimented with estimating models where the constant term (the depreciation rate times minus one) was constrained to be non-positive. These generally failed to converge, and the constraint was always binding. We therefore impose the constraint of a zero depreciation rate in what follows. In column (1) we present the GMM estimates where no equi-correlation is assumed. In column (2) we then present the MSM estimates that allow for countryspecific unobserved component in the frictionless stock of robots. From the GMM estimates, the (natural log of) S*jt is given by 0.62 times manufacturing output. Turning then to the determinants of the hazard rate, we find first of all a large negative constant term (point estimate 1.856 with s.e. 1.184) that together with other estimated coefficients guarantees that the estimated (baseline) hazard is less than unity for all observations. Of the level variables, the investment share of GDP (I/GDP) (we here omit country and time subscripts) and GDP obtain positive coefficients as expected, but of these only the first one is significant. The coefficient of the real interest rate carries a negative sign, but is insignificant. The log of robot prices carries the correct, negative, sign and is significant at the 5 per cent level. Of the drift (growth) variables, the growth in the investment share and the growth of robot prices (incorrectly) obtain negative coefficients, but neither is significant. The growth of GDP carries a precisely measured positive coefficient. All four volatility variables (inflation, the volatility of the growth of the investment share of GDP, of GDP and of robot prices, respectively) carry negative coefficients in line with the predictions of theory. However, of these, only one (volatility of GDP growth) is significant at even the 10 per cent level. This would suggest that uncertainty does not play a major role in the diffusion of robots. Turning to column (2) one immediately notices that our estimate of , the variance share of the country-specific time-invariant error term, is statistically insignificantly different from unity but significantly different from zero. This implies that all the variance of the error term is captured by the time-invariant, country-specific part. A high variance share is not a surprise, given the large differences in stocks of robots across countries. Also, the zero variance share of the linear additive iid error carries the interpretation that the depreciation rate is uniform, constant, and zero. Note also that we find no sign of spurious state dependence being a problem, as the estimated correlation between lagged robot stock and the time-invariant error term is very small ( 0.000023) and very imprecisely measured. Comparing the point estimates to those in column (2) it is clear that the coefficients on the drift and volatility variables are most affected, together
459
Technological diffusion under uncertainty
Table 16.4
Estimation results
Variable/parameter Baseline hazard Constant/0 GDP/1 GDP/11 GDP/12 I/GDP/2 IGDP/21 IGDP/22 lnP/3 P/31 P/32 r/4 i/5 Frictionless robot stock lnIND/ Variance share of the country specific time-invariant error term/ Correlation coefficient between the country specific error term and lagged robot stock/ nobs R2 min. dist. s.e.
(1) GMM
(2) MSM
1.855681 1.188489 0.000051 0.000101 30.822349*** 8.895551 24.749952* 14.490138 0.050468*** 0.017385 19.602290 12.600056 6.496848 4.589849 0.551984** 0.234123 3.953244 4.371913 3.121895 3.678035 0.011807 0.020639 0.817569 0.540648
6.004694*** 0.657220 0.003182*** 0.000248 109.904215*** 6.509909 59.492837*** 8.306346 0.106185*** 0.012490 97.052057*** 7.446970 44.250598*** 3.948424 0.944563*** 0.144453 0.003910 0.004187 20.161940*** 2.646841 0.008959 0.006719 1.219140*** 0.368441
0.623337*** 0.021108 —
94.977696*** 9.993399 1.000001*** 0.000022
—
161 0.6755 0.4791 0.1789
0.000023 0.000051 161 — 31.2525 0.1725
Notes: S 30 (S no. of simulation draws). *** sign at the 1 per cent level, ** sign at the 5 per cent level, * sign at the 10 per cent level.
460
The diffusion of new technologies
with the constant which is now estimated to be 6.00 (p-value .00). We now find that of the levels variables, I/GDP and robot prices carry significant coefficients of the right sign (I/GDP positive, robot price negative), and that of GDP a coefficient of the wrong (negative) sign. The growth (drift) of I/GDP carries a negative and significant coefficient. That of robot prices (contrary to expectations) obtains a negative coefficient but this is imprecisely measured. The coefficient of the drift of GDP carries a positive sign and is highly significant. Most importantly, all the volatility variables’ coefficients carry coefficients of the expected sign (negative) and are precisely measured. Finally, another large change resulting from allowing equi-correlation is that the coefficient of industrial output obtains a very large (and very precisely measured) coefficient. To check the robustness of the results in column (2), we conducted the following tests. First, we estimated the model (and that in column (1)) assuming a Weibull hazard which is a generalisation of the exponential (Weibull generating the exponential when the Weibull parameter p 1). Our estimate of the Weibull parameter was p 1.2732 (s.e. 1.9769) and we therefore cannot reject the null hypothesis of an exponential hazard function. Secondly, we estimated the model in column (2) allowing for a nonpositive constant (non-negative depreciation rate). Our point estimate of the depreciation rate was .0000 (s.e. .01673), and we thus consider it justified to impose a zero depreciation rate.
POLICY EXPERIMENTS To quantify and illustrate the effects of policy on investment decisions, we conducted several policy experiments at country level. The central question, of course, is whether a government policy that yields a more stable (less volatile) environment will lead to substantially faster adoption (this assertion being based on the assumption that [robot] diffusion is welfare enhancing). It should be noted, however, that the exercises we undertake are only partial. We allow that changes in volatility impact upon the threshold rate of return above which adoption will take place and thus the hazard rate. We have not allowed for the second order effect which means that as volatility changes the time paths of Rijt and Pjt will also be changing. It is possible that these second order changes that we do not consider could outweigh the effects we do consider. We report the results of our experiments below in Table 16.5. The first column gives the country in question and column (1) reports the estimated hazard rates of adoption (using the preferred equi-correlated specification) using the country-specific means of the explanatory variables,
461
Notes:
0.0047 0.0006 0.0006 0.0752 0.0000 0.0001 0.0003 0.0000 0.0533 0.0021 0.0000 0.0002 0.0006 0.0053 0.0000 0.0000 0.0089 0.0220
(1)
Predicted hazard
1.1158 1.1766 1.1720 1.1625 1.0947 1.1693 1.1358 1.1705 1.1336 1.0680 1.1225 1.1476 1.1749 1.1066 1.1409 1.0199 1.1320 0.0437
One s.d. decrease (2)
0.8962 0.8499 0.8532 0.8602 0.9135 0.8552 0.8805 0.8543 0.8821 0.9363 0.8909 0.8714 0.8511 0.9037 0.8765 0.9805 0.8847 0.0358
One s.d. increase (3)
Inflation
Policy experiments GDP volatility
I/GDP growth
I/GDP volatility
Robot price volatility
0.4096 0.4051 0.4393 0.4220 0.4057 0.4149 0.3959 0.3052 0.4144 0.2438 0.3630 0.4607 0.4384 0.2380 0.4636 0.4390 0.3912 0.0698
2.4417 2.4684 2.2766 2.3697 2.4646 2.4105 2.5261 3.2761 2.4133 4.1020 2.7545 2.1705 2.2812 4.2025 2.1568 2.2777 2.6620 0.6390
1.4241 1.4667 1.4569 1.3744 1.4479 1.5027 1.4536 1.5975 1.3155 1.8836 1.5507 1.3832 1.4344 2.0085 1.4330 1.4095 1.5089 0.1845
0.7022 0.6818 0.6864 0.7276 0.6906 0.6655 0.6880 0.6260 0.7602 0.5309 0.6449 0.7230 0.6971 0.4979 0.6978 0.7095 0.6706 0.0688
0.8377 1.0343 0.9854 0.7364 1.0393 0.9893 0.9718 1.0960 0.7870 0.9537 1.2623 1.0444 1.0639 0.8196 1.2011 0.9642 0.9866 0.1420
1.1938 0.9669 1.0148 1.3580 0.9621 1.0108 1.0290 0.9124 1.2707 1.0486 0.7922 0.9575 0.9399 1.2200 0.8326 1.0371 1.0342 0.1550
1.4241 1.4667 1.4569 1.3744 1.4479 1.5027 1.4536 1.5975 1.3155 1.8836 1.5507 1.3832 1.4344 2.0085 1.4330 1.4095 1.5089 0.1845
0.7022 0.6818 0.6864 0.7276 0.6906 0.6655 0.6880 0.6260 0.7602 0.5309 0.6449 0.7230 0.6971 0.4979 0.6978 0.7095 0.6706 0.0688
1.5391 1.4986 1.5277 1.5526 1.5408 1.5016 1.5563 1.5701 1.5426 2.0814 1.5926 1.5578 1.4740 1.7124 1.5102 1.6379 1.5872 0.1437
0.6497 0.6673 0.6546 0.6441 0.6490 0.6660 0.6426 0.6369 0.6483 0.4804 0.6279 0.6419 0.6784 0.5840 0.6622 0.6106 0.6340 0.0468
10 % 10 % 10 % 10 % 10 % 10 % 10 % 10 % 10 % 10 % decrease increase decrease increase decrease increase decrease increase decrease increase (4) (5) (6) (7) (8) (9) (10) (11) (12) (13)
GDP growth
The growth variables refer to those used in estimation.
Australia Austria Denmark Finland France Germany Italy Japan Norway Singapore Spain Sweden Switzerland Taiwan UK USA Average s.d.
Country
Table 16.5
462
The diffusion of new technologies
calculated over the estimation period. As can be seen, the predicted hazard rates vary substantially between countries (even those that are zero at four digits – UK, USA, Japan and France – are strictly positive at eight digits). Surprisingly, Japan’s predicted hazard rate is small: this reflects at least partially the fact that Japan’s high stock of robots is captured by a high estimate of S*jt. Columns (2) to (13) report the ratio of the predicted hazard after the experiment to the predicted hazard before the experiment (that is, the hazard given in column (1)). The first experiment was to increase and decrease inflation by one standard deviation from its country mean. Column (2) shows that decreasing inflation by one standard deviation increases the hazard rate of robot adoption by between 2 and almost 18 per cent, and on average by 13 per cent. Increasing inflation by the same amount decreases the average hazard rate by over 11 per cent. For all other variables, the experiment we conducted was a 10 per cent increase or decrease;6 the reason being that as these measures are constant for each country, no standard deviations are available. We find that decreasing GDP growth by 10 per cent leads to a 60 per cent decrease in the hazard rate; mirroring this, an equivalent increase more than doubles the hazard rate in each country, with the highest increases being more than fourfold (Singapore and Taiwan). Changes in the volatility of GDP growth had pronounced effects, too, with an over 50 per cent increase following a 10 per cent decrease in GDP growth volatility. Increasing volatility by the same amount results in an average hazard rate that is just two-thirds of the predicted. Looking then at the effects of growth of the investment share of GDP, we find the effects are moderate. A 10 per cent decrease (increase) only slightly lowers (raises) the hazard rate. The effects of similar relative changes in the volatility of this growth rate are much more pronounced, leading to a 50 per cent increase and a one-third decrease respectively. Finally, we looked at the effects of changes in the volatility of the stochastic process determining robot prices. These are of the same order of magnitude as changes in other measures of volatility.
CONCLUSIONS In this chapter a real options based model of the diffusion of a new process technology has been developed and applied to international data upon the diffusion of robots. There are limits on the predictive power of this model, but it indicates that uncertainty will impact negatively upon the threshold value of the ratio or returns to costs above which adoption will take place. Using the volatility of several macroeconomic indicators
Technological diffusion under uncertainty
463
as measures of uncertainty, the preliminary empirical analysis reported here confirms this prediction. In addition, it is found that there are significant country-specific effects, that variables in their levels generally impact as expected but that the results with the respect to the drift (or growth) of relevant variables are weaker. There is still considerable further work to be undertaken upon the estimation, but the results as presented here are encouraging and (1) support the use of real options based methods, (2) generally confirm the hypothesis that uncertainty deters investment in new technology, and (3) support the view that, to some degree at least, different rates of diffusion in different countries reflect the uncertainties in their macro environments. Preliminary policy experiments suggest, within their stated limitations, that modest changes in GDP growth and all our measures of volatility have pronounced effects on the hazard rate of adoption. As the effects of such changes cumulate rapidly over time, our results suggest – under the hypothesis that more rapid diffusion of new technologies is welfare enhancing – that managing the stability (volatility) of the (macro) environment in which firms make investment decisions is of paramount importance.
APPENDIX Data Sources and Variable Definitions Robot and robot price data: World industrial robots 1994, United Nations. For robot prices, we use the German prices on the unit value of robot production. Prices for 1979 and 1978 were calculated using the results of a regression of prices on a constant, years and squared years. The price series was deflated using the consumer price index reported in the IFS statistics (the only price series available for all countries). We use prices in constant 1985 US dollars for all countries in the regressions. To calculate drift and volatility for robot prices we use the whole series. GDP, from Penn World Tables Mark 5.6. Defined as GDP real GDP per capita (series name CGDP) times population in 000s (POP)/1 000 000. To calculate drift and volatility for GDP, we use 1960–92 data, as this was the longest series available to all countries. Manufacturing output. Manufacturing in 1986 defined as Man. index of manufacturing times GDP (defined as above) times the manufacturing share of GDP. For other years, the level of manufacturing is calculated from the 1986 (1992 for Taiwan, see below) figure using the index of manufacturing.
464
The diffusion of new technologies
The index is from International Financial Statistics (for Taiwan from Financial Statistics, Taiwan District Republic of China: these are designed to conform to the IFS statistics) and GDP is derived from Penn World Tables as described above. The share of manufacturing as a percentage of GDP in 1986 is from the UN Statistical Yearbook 1995 for all other countries but Italy and Taiwan (not available for these two). For Italy, the manufacturing share of GDP was calculated as the ratio between figures ‘industria in senso stretto’ (total industry output) and ‘totale’ in Table ‘Tavola 10 – Produzione al costo dei fattori – Valori a prezzi 1995’ (‘Table 10 – Production at factor cost – 1995 values’). The table can be found on the web page of the National Institute of Statistics of Italy (http://www.istat.it/) as file TAVOLE-PRODUZIONE.XLS. The manufacturing share of GDP was calculated for 1986. For Taiwan, the source is the file VIGNOF4D.XLS to be found on the web page of the Directorate General of Budget, Accounting and Statistics of Taiwan (http://www.dgbasey.gov.tw/). The file contains Table H-2 Structure of Domestic Production, in which the manufacturing share of GDP is reported for 1992–95. We used the 1992 figure. Investment share of GDP: Penn World Tables (CI). Drift and volatility calculations as with GDP. Price level in manufacturing, Penn World Tables (PI). Inflation in country i in year t defined as inflit ln(PIit) ln(PIit–1). Real interest rate. Defined as the difference between the nominal interest rate and inflation, both calculated from indexes from the IFS. Inflation is calculated from the consumer price index as reported in IFS (it is the only price index available for all sample countries in IFS). We use the money market rate from IFS statistics as the nominal interest rate. Testing the Assumption of a Random Walk Our theoretical model assumes that the relevant (country-level) time-series are random walks (with drift). In the estimations, we use as explanatory variables ML estimates of the drift (growth rate) and standard error of GDP, investment share of GDP, and of robot prices. To test that our timeseries are random walks, we regressed the difference in the series on a constant, time trend, and lagged level using data from the period 1960–92 (we also estimated the model without the time trend, with similar results). We then calculated a Dickey-Fuller test. Table 16.A1 summarises our findings, presenting the D-F test value (p-value) of the GDP and I/GDP estimations for each country separately (32 degrees of freedom), and ‘finally’ the robot price tests.
465
Technological diffusion under uncertainty
Table 16.A1
Dickey-Fuller tests
Country Australia Austria Denmark Finland France Germany Italy Japan Norway Singapore Spain Sweden Switzerland Taiwan UK USA
GDP 1.3640 0.0234 1.1187 2.0427 1.0780 1.5006 0.9978 1.1567 1.9845 2.8244 0.8829 2.0859 0.5560 4.6409 1.8332 1.3868
I/GDP (1.0000) (1.0000) (1.0000) (1.0000) (1.0000) (1.0000) (1.0000) (1.0000) (1.0000) (1.0000) (1.0000) (1.0000) (1.0000) (1.0000) (1.0000) (1.0000)
4.4769 2.3260 2.5019 2.8554 2.2507 2.9139 3.0360 2.0718 1.9929 1.6285 1.8924 2.9694 2.2222 1.3666 2.5871 4.3405
(1.0000) (1.0000) (1.0000) (1.0000) (1.0000) (1.0000) (1.0000) (1.0000) (1.0000) (1.0000) (1.0000) (1.0000) (1.0000) (1.0000) (1.0000) (1.0000)
NOTES 1. In an earlier paper, Toivanen et al. (1999), we have also looked at the comparative international diffusion paths of robot technology. That paper contains no formal modelling of the impact of uncertainty upon the diffusion process, instead some rather ad hoc theorising was relied upon. Here, with the explicit theoretical model we are able to more precisely define appropriate measures and functional forms and also more reliably interpret the results. In addition, we have undertaken considerably more estimation experiments than we had completed at the time the previous paper was published. 2. Our discussion on robots and robot statistics relies heavily on World Industrial Robots 1994 (Geneva: United Nations). 3. Twelve countries were excluded for a variety of reasons, for example, for Russia and Hungary there was no reliable data on required macrovariables (and even the robot data is suspicious) and the Benelux countries (Netherlands, Belgium and Luxembourg) were summed together in the robot statistics. 4. The choice of the Gompertz as opposed to logistic formulation is to some degree conditioned by the success of this formulation in our earlier work, Toivanen et al. (1999), but experiments also indicate that empirically the Gompertz model is to be preferred to the logistic model. 5. Neither of these approaches take any note of changes in the nature of robot technology over time. In principle this could be catered for by the introduction of trend terms or by the definition of robot price Pjt as quality adjusted. In this chapter, however, neither route is explored. 6. Halving the changes to one-half of a standard deviation and 5 per cent respectively led to effects that were, to a rough approximation, on average half of those reported here.
466
The diffusion of new technologies
REFERENCES Aitchinson, J. and J.A.C. Brown (1957), The Lognormal Distribution, Cambridge, Cambridge University Press. Berndt, E.R., B.H. Hall, R.E. Hall and J.A. Hausman (1974), ‘Estimation and inference in nonlinear structural models’, Analysis of Economic and Social Measurement, 3, 653–66. Caballero, R-J. (1991), ‘On the sign of the investment–uncertainty relationship’, American Economic Review, 81, 279–88. Caballero, R-J. and E.M.R.A. Engel (1999), ‘Explaining investment dynamics in US manufacturing: a generalized (S,s) approach’, Econometrica, 67, (4), 783–86. David, P.A. (1969), A Contribution to the Theory of Diffusion, Stanford Centre for Research in Economic Growth, Memorandum No. 71, Stanford University. Dixit, A. and R. Pindyck (1994), Investment Under Uncertainty, Princeton, Princeton University Press. Driver, C. and D. Moreton (1991), ‘The influence of uncertainty on UK manufacturing investment’, Economic Journal, 101, 1452–9. Heckman, J. (1981), ‘The incidental parameters problem and the problem of initial conditions in estimating a discrete time-discrete data stochastic process’, in C.F. Manski and D. McFadden (eds), Structural Analysis of Discrete Data with Econometric Applications, Cambridge, MA: MIT Press. Ireland, N. and P. Stoneman (1986), ‘Technological diffusion, expectations and welfare’, Oxford Economic Papers, 38, 283–304. Jensen, R. (1982), ‘Adoption and diffusion of an innovation of uncertain profitability’, Journal of Economic Theory, 27 (1), 182–93. Jensen, R. (1983), ‘Innovation adoption and diffusion when there are competing innovations’, Journal of Economic Theory, 29 (1), 161–71. Karshenas, K. and P. Stoneman (1993), ‘Rank, stock, order and epidemic effects in the diffusion of new process technology’, Rand Journal of Economics, 24 (4), 503–28. Leahy, J. and T. Whited (1996), ‘The Effect of uncertainty on investment: some stylized facts’, Journal of Money, Credit and Banking, 28, 64–83. Mansfield, E., (1968), Industrial Research and Technological Innovation, New York, Norton. Newey, W.K. (1990), ‘Efficient instrumental variables estimation of nonlinear models’, Econometrica, 58 (4), 809–37. Pagan, A.R. (1984), ‘Econometric issues in the analysis of regressions with generated regressors’, International Economic Review, 25 (1), 221–47. Pindyck R.S. and A. Solimano (1993), ‘Economic instability and aggregate investment’, NBER Macroeconomics Annual, Cambridge, MA: MIT Press, pp. 259–303. Rose, N. and P. Joskow (1990), ‘The diffusion of new technologies: evidence from the electricity utility industry’, Rand Journal of Economics, 21, 354–73. Stoneman, P. (1981), ‘Intra-firm diffusion, Bayesian learning and profitability’, Economic Journal, 91 (362), 375–88. Stoneman, P. (1983), The Economic Analysis of Technological Change, Oxford: Oxford University Press.
Technological diffusion under uncertainty
467
Stoneman, P. (2001), The Economics of Technological Diffusion, Oxford, Blackwell. Summers, R. and A. Heston (1991), ‘The Penn World Table (Mark 5): an expanded set of international comparisons, 1950–1988’, Quarterly Journal of Economics, 106 (2), 327–68. Toivanen, O., P. Stoneman and P. Diederen (1999), ‘Uncertainty, macroeconomic volatility and investment in new technology’, in C. Driver (ed.), Investment, Growth and Employment: Perspectives for Policy, London, Routledge; pp. 136–60.
PART V
Postscript
17. An appreciation of Paul David’s work Dominique Foray* Paul’s work is like an extraordinary theatrical performance. It is full of heroes, including Galileo1 and Edison,2 you will meet primitive tribes – the Pomo Indians and the !Kung Bushmen of the Kalihari,3 as well as nymphs4 and strange little monsters (our ‘less reliable American cousins’5). It is a body of work where amazing machines are driven by a frantic will to survive – QWERTY,6 the reaper, and the robot.7 There is sex (the estimation of ‘objective probabilities of conception from isolated coitus on various days of the menstrual cycle’8), blood (or at least weapons like the boomerang9 or near-weapons like nuclear power10), and weird, almost imaginary objects such as the marine chronometer11 or the so called ‘N times 384 Kbps standard’.12 You can visit the 1900 Paris Exposition,13 as well as the Crystal Palace Great Exhibition of 1851,14 and you can explore the secret Venice of 1332.15 Apart from the panda16 and a few horses, there are not really many animals, but there are all kinds of networks, made up of strange and disturbing people, like the secret society of snow-shovellers,17 or the mysterious Hungarian sect of Zipernowsky, Blathy and Deri.18 It is a body of work with lots of drama and all sorts of accidents: nuclear ones,19 a fire in Baltimore,20 even a train derailment.21 And, of course, the millennium bug.22 It is thrilling, fertile, evocative and exuberant, with frenzied battles giving way to galleries of portraits. It makes one dizzy to contemplate it all. Is that the primitive tribe that uses the digital boomerang to hunt with? Did the Grand Duke of Tuscany save us from the year 2000 bug? But there is a key that allows us to understand the meaning of this great spectacle, to put all the characters in the right place and to grasp what happens to the machines and the institutions and why disasters happen. I found this key when I read Robert Musil’s great work.23 Musil writes: ‘The course of history is shaped by the action of myriad little causes which all operate in an unpredictable way.’ Musil labels this notion anti-hero and petit bourgeois, while remarking that the philosophy of the history of great
471
472
Postscript
causes, with its fine intellectual pathos is really only heroic in appearance, because it does not take the facts at face value. Indeed, says Musil, sometimes a very small movement at the right time is all that is needed to change substantially the end result. The philosophy of small causes, of sensitivity of effects to the very tiniest differences in initial conditions, and the huge contrast possible in the end between minute causes and enormous effects, introduce vagueness, doubt and a basic ambiguity into human activity. Musil defends this view of history, which he firmly shares with Paul. He says it both respects the facts, which other theories do not, and leaves us free to take action, and perhaps even be left with a better outcome at the end of the day. Musil concludes: people say a taste for this way of thinking betrays a crudely mechanical mind and a cynical Philistine attitude. I’d like to say that it contains enormous optimism. Because if we do not have to rely any longer on the likes of a spindly scarecrow in a field deciding our destiny, but are just covered with a lot of little weights tangled up like pendants, it’s up to us to tip the balance.24
Our steps are influenced by a tangle of weights, whose balance we may be able to shift. Abundance of resources is up to us!25 And knowledge openness as well!26 This is a key to Paul’s work. ‘It’s up to us to tip the balance’ (ibid.). Thank you, Paul, for giving us an economics of optimism!
NOTES * 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19.
I am grateful to Bronwyn and Ed for editorial assistance. ‘Patronage, reputation and common agency contracting in the scientific revolution’. ‘The hero and the herd’. ‘Information technology, social communication and the wealth and diversity of nations’. ‘From market magic to Calypso science policy’. ‘European nuclear power and their less reliable American cousins’. ‘Clio and the economics of QWERTY’. ‘The reaper and the robot’. ‘Making use of treacherous advice’. ‘Digital technology boomerang: new IPR protections’. ‘Performance-based measures of nuclear reactor standardization’. ‘Keeping your bearings on public R&D policy’. ‘The ISDN bandwagon is coming’. ‘Computer and dynamo’. ‘The landscape and the machine’. ‘The evolution of intellectual property institutions’. ‘Intellectual property institutions and the panda’s thumb’. ‘Path-dependence and predictability’. ‘The economics of gateway technologies and network evolution’. ‘Learning from disasters’.
An appreciation of Paul David’s work 20. 21. 22. 23. 24. 25. 26.
473
‘New standards for the economics of standardization’. ‘Transport innovations and economic growth’. ‘The millennium bug meets the economics of QWERTY’. Musil’s words come from R. Musil, ‘Der Man ohne Eigenschaften’ (French translation: ‘L’homme sans qualité) and from J. Bouveresse, ‘L’homme probable: R. Musil, le hasard, la moyenne et l’escargot de l’histoire’. In L’Europe désemparée, Essais, Conferences Critiques, Aphorisme et Réflexions, l’Europe désemparée, Paris: Seuil, 1984, p. 142 (our translation). ‘Resource abundance and American leadership’. ‘Knowledge, property, and the system dynamics of technological change’.
Index Abramovitz, M. 10, 13 Abramson, A. 149 adoption of new technology 16, 17, 434, 438 adoption decisions 24–5, 29, 30 catastrophe adoption path 430, 431, 435 continuous adoption path 430 epidemic adoption theory 432, 433, 436 equilibrium adoption theory 428–9, 431, 432, 433, 436 fax machines 433, 434 increasing returns to adoption 23–4, 34 network effects 428, 429, 430, 431 uncertainty 439 see also competing technologies; real options model of technology adoption; robots Aghion, P. 302 Aitchinson, J. 444 Aitken, H. 141 Akerlof, G. 189 Allen, R. 243 Alma-Tadema, L. 179, 180, 181, 182, 183 Alston, L. 324 American Telephone and Telegraph (AT&T) 120, 122, 127, 141 share of US patenting 141, 142, 143 technological specialisation path dependency 154 profile of 46–8, 149, 150, 151, 152, 153 Andersen, H. 128 Antonelli, C. 158 Aoki, M. 353 Archibugi, D. 128, 365, 366 Arora, A. 80 Arrow, K. 13, 55, 241 art
inter-painter price relationships 201, 202 oil paintings prices of 179 price measures of demand 165, 166, 167–77, 190–93, 195–6 ‘fad component’ 193–5 inherently good painters 198–9, 200 study data 177, 178 study methodology 177, 178, 180 tastes avant-garde effect 188–9, 196, 197, 198, 199, 200–201 characteristics approach 164–5 conformity effect 188, 196, 197, 201 path dependence 162, 163, 182, 183 volatility of 160, 161, 162 trends in popularity of artists 180–82, 183, 188, 189, 197, 198, 201, 202–3 see also taste Arthur, W. 24, 25, 26, 30, 31, 36, 39, 60, 118 Arundel, A. 363, 367, 369 AT&T, see American Telephone and Telegraph Atkinson, A. 58 Audretsch, D. 256 Ayers, F. 173 Bacharach, M. 164 Balassa, B. 373 Banerjee, A. 24 Barnes, R. 161 Barré, R. 258 Barrera, M. 127, 137 Basberg, B. 128 Baumol, W. 158, 160 Bayer 120 475
476
Index
share of US patenting 131 technological specialisation path dependency 154 profile of 133, 134, 135, 136, 137, 150, 151, 152 see also IG Farben Bayma, T. 341 Beaton, K. 122 Beaver, D. 256 Becker, G. 160, 189 Beer, J. 119, 129, 136 Bell, A. 141 Bell, C. 161 Bell, Q. 161 Beniger, J. 213 Berndt, E. 457 Bernheim, B. 24 Berthet, M. 245, 249 Bessen, J. 420 Bianchi, M. 160 Bienz-Tadmor, B. 78 Birdzell, L. 324, 327 Bonnard, P. 181, 182, 183 Bordo, M. 325 Boucher, F. 181, 182, 183 Bound, J. 126 Bourdieu, P. 159 Boyer, H. 77 Braun, T. 365 Bresnahan, T. 332, 389 Brock, W. 24 Brown, J. 308, 309, 311, 444 Brynjolfsson, E. 24 Burstall, M. 85 Caballero, R.-J. 439, 440 Callon, M. 242, 255, 256 Canaletto, A. 179, 181, 182, 183 Cantwell, J. 365, 369, 374 Carlsson, B. 365 Casson, M. 297 Centre National de la Recherche Scientifique (CNRS) study of collaboration between researchers 258–91 Cézanne, P. 179, 181, 182, 183 Chandler, A. 71, 121, 122, 130, 153, 299 Chanel, O. 165 chemical industry 119–20
geographic origin of research activities 370, 371, 372 knowledge specialisation 373–5, 376, 377, 378, 379, 380 shares of US patenting 131 technological specialisation 151 see also Bayer; Du Pont; IG Farben Chien, R. 84 Church, J. 24, 25 Claude Gellée 181, 182, 183 CNRS, see Centre National de la Recherche Scientifique co-publication in scientific research 255 determinants of 255, 256 future research on 285, 286 see also Centre National de la Recherche Scientifique Coase, R. 7 Cockburn, I. 77, 79, 81, 82 Cohen, M. 296 Cohendet, P. 303 Coleman, D. 122 Collins, W. 180, 181, 182, 183 communities of practice interaction with epistemic communities 314 learning in 306–7, 309, 310, 311 competing technologies 29 models of 27–9, 30–34, 35, 36–9 proof of proposition 1 40–41 proof of proposition 2 41–3 proof of proposition 4 43–6 rate of convergence to technological monopoly/market sharing 35–9 sequence of historical events 38 time required 38, 39 relative impact of increasing returns and degree of heterogeneity 38, 39 competitive advantage in industry 209 composition effects 59, 60, 61 computer manufacturers 217 software systems 217, 218 see also information and communication technology Conant, J. 364 condensed matter physics 258 conspicuous consumption 159 Constable, J. 181, 182, 183 Cottereau, A. 248, 249, 250
Index Cowan, R. 24, 25, 30, 31, 160, 306 Cozzi, G. 160 creativity 62, 63, 64 Crémer, J. 313 cross-licensing agreements 340, 343 Cusumano, M. 24, 29 Cyert, R. 296 Dalum, B. 374 Dasgupta, P. 7, 232, 240, 242, 251, 255, 323, 326, 328, 347, 348, 364 data warehouses 229 David, P. 5–18, 23, 25, 51–2, 56–8, 60, 62, 70, 118–20, 141, 143, 158, 163, 182, 190, 207, 218, 232, 240–42, 251–2, 255, 286, 307, 323, 325–8, 344, 347–8, 353, 361, 364, 367, 379, 389–90, 427, 438–9, 443, 471–2 de-centred and distributed learning 310, 311, 312 communication 312–13 de Gennes, P.-G. 258 de Lasalle, P. 244, 246–7, 249 de-localisation of knowledge 212–14 De Marchi, N. 160, 164 De Piles, R. 164, 165 Deane, P. 177, 178, 179 Debreu, G. 13 decentralised system of knowledge management 231–2 decision-making structures 64 Degas, E. 181, 182, 183 Deng, Z. 333 Diamond, A. 255 diffusion of innovations and new technology 16, 17 fax machines 433, 434 see also adoption of new technology; real options model of technology adoption; robots dissonance 311 distant past historicism 167 distributed information technology 217, 218 Dixit, A. 439, 443, 444 Dornseifer, B. 130 Dosi, G. 160, 296, 299 Driver, C. 439, 440 Du Pont 120, 122, 127 share of US patenting 131
477
technological development 129–30 technological specialisation path dependency 154 profile of 137–40, 150, 151, 152 Duguid, P. 308, 309, 311 Durlauf, S. 24 dynamic efficiency of economic systems, conditions for 61–6 , 67 Eckhardt, S. 82 economic theory 3, 4 economics of science 255, 256 Economides, N. 24, 433 economies of learning 208 Edison, T. 141, 143 Egidi, M. 299 Eisenberg, R. 325, 338, 340 electrical equipment industry 120 development of 141 shares of US patenting 141–3 technological specialisation 151 vertically integrated systems 141, 149 see also American Telephone and Telegraph; General Electric Eliasson, G. 296, 308 Engel, E. 439 enterprise management software 229 epistemic communities interactions with communities of practice 314 production of knowledge 306 ergodic processes 52, 53 Ernst and Young 81 ETAN 332, 333, 334 European Commission 365, 369 European Technology Assessment Network (ETAN) 332, 333, 334 experimental learning 307, 308, 311–12 Fagerberg, J. 366 Fai, F. 118, 119 Falcon, J.-P. 244, 249 Farrell, J. 24, 34, 433 Favereau, O. 298 fax machines 433, 434 Federal Reserve Bank of Dallas 223 Filene, E. 214, 217 firms building of a common knowledge specific to the firm 313–14
478
Index
core competences 300 exchange of knowledge through networks 301, 302 governance 304, 305 knowledge formation in the firm 309, 310 management of collectively distributed knowledge within the organisation 308, 309 non-core competences 301 peripheral activities 302 ranking of activities 302, 303, 304, 314 declassifying routines 303–4 structure of governance 304 theories of 296, 297 competence 300, 304, 305 principal/agent theory 297–8 processor of information, as 297–8 processor of knowledge, as 298–9 transaction cost 298, 300 flexible production 216, 217, 222; see also product variety Foray, D. 158, 255, 258, 361, 367 Ford, H. 213, 216, 339 Ford Motor Company 216 forgetting 311 Foss, N. 296, 302 Frank, R. 160, 189 Fransman, M. 296 Freeman, C. 133 Freeny Jr., C. 418 Frey, B. 160 Frost, R. 8 Galambos, L. 71 Gambardella, A. 71, 77, 79, 80, 81, 370 Gandal, N. 24, 25 General Electric 120, 122, 127, 141 share of US patenting 142, 143 technological specialisation path dependency 146, 154 profile of 143–6, 150, 151, 152, 153 General Motors 216 general purpose technologies (GPTs) analysis of data 390, 391 generality measurement 393, 395–400, 419, 420, 421, 423
identifying GPT patents 410, 413–17, 418, 419 definition of 390 ICT-related patents 418, 419 patent characteristics 392 see also patent citations; patents geographic origins of research activities 370, 371, 372 Ghoshal, S. 305, 311, 312 Gibbons, M. 255, 308 Gilbert, R. 336 Ginsburgh, V. 160 globalisation as cause of technological change 65 Godin, B. 365, 373 Gogh, V. van 160, 181, 182, 183, 188 Gombrich, E. 161, 162 Gomperts, P. 331 Goodwin, C. 160 Gorman, W. 164 Gould, S. 57 GPTs, see general purpose technologies Grabowski, H. 78, 84 Grampp, W. 160 Granstrand, O. 366, 379 Green, J. 337 Greenstein, S. 328 Griliches, Z. 16, 123, 127, 128 Grindley, P. 340, 343 Grossman, S. 335 growth 66 conditions for 61–6 Guerzoni, G. 177 Haber, L. 119, 129 Hadley, W. 340 Hall, B. 329, 340 Hals, F. 180, 181, 182, 183 Hand, J. 333 Hart, O. 335 Hayek, F. von 299 Heckman, J. 455 Heller, M. 325, 338, 340 Helpman, E. 389, 390 Henderson, R. 71, 77, 79, 81, 82, 85, 346, 393, 397 Heston, A. 450 higher education, see university-based research Hill, B. 41, 43
Index Himmelberg, C. 433 Hounshell, D. 122, 126, 130, 137, 140, 213, 214, 215, 216 Hughes, T. 141, 213 Hunt, R. 420 Iansiti, M. 366 ICT, see information and communication technology IG Farben 120, 122, 127, 133 share of US patenting 131 technological development 129, 130 technological specialisation path dependency 154 profile of 133, 134, 135, 136, 137, 150, 151, 152 see also Bayer IMF 450 incentives and institutional standards 224, 225, 226, 249, 328, 329–30 increasing returns to adoption 23–4, 25, 26, 34, 39 individual knowledge 208 individual learning 208 information and communication technology (ICT) 11 construction of integrated systems 218–19 decentralising information processing 218, 219 distributed technology 217, 218 incentives and institutional standards 226 interpersonal communication 226–8 minicomputers 218 modelling business processes 228–9 patents 418, 419 supporting local learning 219, 220 see also computer manufacturers; software information search costs 334 innovation economics 3, 4, 5, 6 innovative capacity 4, 5 intellectual property 12, 335 security interests in 333–4 see also patents intensive use of knowledge 8 International Monetary Fund (IMF) 450
interpersonal communication exchange of knowledge 226–8 interrelatedness of technology 150 Ireland, N. 444 Ironmonger, D. 164 irreversibility 59 Islas, J. 24 Jacquard, J.-M. 248, 249, 250 Jaffe, A. 393, 395, 397 Janson, A. 188 Janson, H. 188 Jensen, R. 439 Jones, R. 122 Joskow, P. 456 Kahneman, D. 55 Karshenas, K. 443, 447, 456 Katz, J. 256 Katz, M. 23, 24, 25, 27, 29, 31 Kemerer, C. 24 Kenney, M. 349 Kirman, A. 190 Klemperer, P. 336 Klevorick, A. 83 knowledge 6 circulation 210 codification 10 de-localisation of 12–14 individual 208 intensive use of 8 management 230, 231, 232–3 decentralised system of 231–2 meaning of 230–31 organisational 208 value of 209 production of 8 public domain 7, 12 financing of knowledge production 240–41 public–private interactions 12 role of, in industry 211, 212 tacit 10, 11 transfers of 12 see also knowledge commons; knowledge integration; knowledge openness; knowledge persistence; knowledge specialisation
479
480
Index
knowledge commons 8, 9 localised 58, 60 knowledge integration 363, 366, 367, 368 chemical and pharmaceutical industries 375–6, 377, 378, 379, 380 future research 381 policies for 380 knowledge openness 239, 240, 241, 242 collective ethos 246, 247, 248 efficiency of 248–50 establishment of technical standards 249 reward system 248, 250, 251 see also open science; open source software; open technology knowledge persistence 362, 364, 365, 366 chemical and pharmaceutical industries 373–5, 377, 378, 379, 380 future research 381 see also knowledge integration knowledge specialisation RSI index 381–2 specialisation profiles of chemical and pharmaceutical industries 377–8, 380 see also knowledge integration; knowledge persistence; technological specialisation Konno, N. 304 Kortum, S. 342 Koski, H. 37 Kremer, M. 344 Krugman, P. 26 Lamoreaux, N. 71, 334 Lancaster, K. 164 Landau, R. 366 Landseer, E. 180, 181, 182, 183 Lane, D. 24 Langlois, R. 296, 302, 328, 339 Lasdon, L. 180 Leahy, J. 439 Leamer, E. 278 learning 308 by doing 9, 10
de-centred and distributed 310, 311, 312, 313 economies of 208 experimental 307, 308, 311–12 governance for 313 individual 208 organisational 208 ‘technology of 208–9 through error production 310 see also communities of practice Leibig, J. 82 Lerner, J. 342 Lev, B. 333 Liebowitz, S. 24, 29 Linden, G. 329, 330 Llerena, P. 314 Loasby, B. 296 localised introduction of new technologies 58 factors affecting 65 localised knowledge commons 58, 60 localised problem-solving 229 lock-in effects 14, 15, 61, 119 Lundvall, B. 307, 361 Lyons silk industry diffusion of new technology 248–50 invention in 243–4 sharing of knowledge 244–8, 250, 251 Maclaurin, W. 143 Madison, J. 16 Magalhães, R. 309 Malerba, F. 365 Malo, S. 367, 370 Malraux, A. 162 Manet, E. 181, 182, 183, 188 Mansfield, E. 364, 439 March, J. 296, 299, 305 Marcus, G. 162 Marengo, L. 296, 299 Margolis, S. 24, 29 market sharing 23, 26, 34; see also competing technologies markets for technology financial institutions, role of 331 global market 349–53 information search costs 334 institutional settings 327
Index intellectual property rights 335 security interests in 333–4 limitation of liability 334 patent offices, role of 341–3 patent-pooling agreements 343 patents 335, 336, 337–8 ‘efficient breach’ 344 extension of ‘eminent domain’ 344 fragmentation 338, 339, 340, 343, 344 legal costs 340 research and development tax credits 332–3 standards 328, 329–30 startup firms, government support for 332 technology suppliers, role of 331, 344 university research 344 valuation of technology 333 venture capitalists, role of 331–2 Marriot, O. 122 Martin, B. 380 mass production system 213, 214, 216, 217 information goods 220–21 Matraves, C. 74 Maxwell, R. 82 McCain, R. 160 McCormick, C. 214, 215 McCormick, L. 214, 215 McCormick Reaper Works 214, 215 production system 214–16 McDermott, C. 162 McPherson, M. 160 Meissonier, E. 180, 181, 182, 183 Meliciani, V. 365 memory 230, 311 Menger, P.-M. 160 Merges, R. 78, 335, 337, 340 Merton, R. 364 Metcalfe, J. 65 Metcalfe, S. 308 minicomputers 218 Mitchell, B. 177, 178, 179 Monet, C. 164, 181, 182, 183, 188 Moore, J. 335 moral property rights 241, 242 Moreton, D. 439, 440 Mowery, D. 345, 346, 347, 349
481
Mullins, N. 256 Musil, R. 471, 472 Narin, F. 364, 367 Nash, L. 364 national competitiveness, scientific and technological specialisation, role of 362 National Research Council 340 national systems of innovation (NSI) 361 Nattier, J.-M. 180, 181, 182, 183 Nelson, R. 78, 120, 136, 299, 312, 337, 340, 361 network effects 15, 24, 29, 428, 429, 430, 431 Newey, W. 457 Nijkamp, P. 37 Nobel, D. 143 Nohria, N. 305, 311, 312 non-ergodic processes 52, 53; see also past dependence; path dependence Nonaka, I. 304 Nooteboom, B. 308, 312 North, D. 324 NSI 361 Nuvolari, A. 243 OECD 307 Office of Science and Technology (OST) 365 Office of Technology Assessment and Forecast (OTAF) 392 oil firms, shares of US patenting 132 open science 7–8; see also knowledge openness open source software 227 open technology 243, 251; see also knowledge openness; Lyons silk industry Oren, S. 24 Organisation for Economic Cooperation and Development (OECD) 307 organisational capability 208 organisational knowledge 208 value of 209 organisational learning 208 organisational memory 230 Orsenigo, L. 365, 370
482
Index
OST 365 OTAF 392 Pagan, A. 457 Pareto, V. 13, 14 past dependence 52, 53, 54, 56 role of internal factors 57 Patel, P. 128, 362, 366, 369 patent citations 391–2 citation lags 407, 408, 409 highly cited patents 393, 394 characteristics of 408, 409, 410 generality measures 400, 401–2 probability of 410, 411–12 technology sub-categories of 421, 422, 423 see also general purpose technologies; patents patents characteristics of 409 cross-licensing agreements 340, 343 growth of patent classes 400, 403–5, 406, 407, 409 legal costs 340 pharmaceutical industry 78, 83, 87 technology sub-categories of 421, 422, 423 university-based research 345, 346 see also general purpose technologies; markets for technology; patent citations path dependence 51, 52, 53, 54, 56, 60–61, 66, 67, 118 characteristics of 52, 54 definitions of 163 external factors, role of 58, 59, 60 feedbacks 54, 55, 56, 119 internal factors, role of 58, 59, 60, 61 irreversibility 54, 55 local externalities 54, 55, 56, 58 lock-in 61 sequence of steps 54 strength of 119 theory of 13–15 path independence 163 Pavitt, K. 123, 128, 362, 365, 366, 369, 373 Peltzman, S. 84 Penrose, E. 299 personal computers
construction of integrated systems 218–19 decentralising information processing 218, 219 supporting local learning 219, 220 pharmaceutical industry 70, 71 biotechnology development of 81 impact of 77, 78, 79 collaborative research 80, 81 commercialisation of penicillin 72 competition 75 development of 71–2, 113 geographic origin of research activities 371, 372 health-care systems, structure of 84 innovation 85–6 economic benefits from 73, 74, 75 forms of 75 imitator firms 93 innovative firms 93 levels of 73, 74 knowledge specialisation 373–5, 376, 377, 378, 379, 380 levels of concentration 75–6, 86 model of new drug development 87–92, 93, 114 extension of time of patent protection, effect of 108, 113 firms’ activity in different therapeutic categories 106 imitative products, number of 101, 102, 103, 104 increase in number of firms, effect of 108, 113 increase in stringency of approval procedures, effect of 108, 113 innovative products, number of 101, 102, 103, 104 innovative products, share of 105 market concentration 94, 95, 96, 108 number of firms in each therapeutic area 100 number of innovative and imitative products in each therapeutic area 108, 109, 110, 111, 112 number of products in therapeutic area 99
Index number of therapeutic areas discovered 98 performance index 107 surviving firms 97 new firm entrants 75, 78 patents 78, 83, 87 price regulation 84, 85 product approval 83–4, 87 publicly funded research 76, 82–3 random screening 71, 73, 74, 75, 76 source of first-mover advantage 75 rate of technological change 81 rational drug design 76, 77 research approach 70, 72, 73 transforming research into successful products 372 university research 82 university spin-offs 77 vertical integration 81 Pharmaceutical Manufacturers Association 78 Pianta, M. 365, 366 Pindyck, R. 439, 440, 443, 444 Pisano, G. 71, 362, 366, 370 Pissarro, C. 181, 182, 183 Plumpe, G. 122, 126, 127, 129 Pommerehne, W. 160 Porter, M. 362 Prencipe, A. 366 Price, D. de S. 365 Price, R. 16 producer–user relationship in industry 212 product diversification, link to technological diversification 153 product selection decisions 25, 29, 30, 37 product variety 221, 222, 223 decentralisation, need for 223, 224, 225 see also flexible production productivity of scientific research, determinants of 255 prospect theory 55 public and quasi-public databases 341 Quillen, C. 340 Rallet, A. 258 Rauch, J. 26
483
real options model of technology adoption 439–40, 442–50 data sources 450, 463–4 descriptive statistics 451–4 methodology 455–7, 463, 464, 465 see also robots recent past historicism 167 Reich, L. 120, 122, 126, 127, 141, 143, 146, 149 Reitlinger, G. 166, 177, 179 Rembrandt van Ryn 181, 182, 183 Renoir, P. 181, 182, 183, 188 reputation capital 241 research and development (R&D) tax credits 332–3 Rheims, M. 161 Richardson, G. 299 Robertson, P. 328 robots 440–41 adoption of 441 determinants of 458, 459, 460 government policy changes, effect of 460–62 numbers 453, 454, 456 uncertainty, impact of 462–3 application areas 441 investment in 441 volatility of 441, 442 prices 450, 453, 454, 455, 456, 458, 460, 463 see also real options model of technology adoption Rohlfs, J. 24 Roos, J. 309 Rose, N. 456 Rosen, R. 256 Rosenberg, N. 24, 136, 213, 324, 327, 362, 366, 389 Rostoker, M. 340 Ruskin, J. 158, 160, 164, 166 Ruttan, V. 55 Saloner, G. 24, 34 Sanderson, W. 9, 307 Santangelo, G. 153 Saviotti, P. 24 Scherer, F. 123, 126 Schmookler, J. 121, 123, 126 Schwartzman, D. 71 Schwerin, J. 243
484
Index
Science Citation Index (SCI) 279, 281, 282 Scotchmer, S. 337 sectoral knowledge bases 367, 368; see also knowledge specialisation security interests in intellectual property rights 333–4 self-sustaining process of growth and innovation 62, 63 Sewell, J. 71 Shapiro, C. 23, 24, 25, 27, 29, 31, 336, 433 Sharp, M. 372 Shi, Y. 255 Shrum, W. 256 Silverman, B. 395, 398 Simon, H. 299 Sisley, A. 181, 182, 183 slack 311 Sloan, A. 217, 219 Smith, A. 164, 189, 212 Smith, J. 122, 126, 130, 137, 140 Smith, S. 24 social referral networks 229–30 Soete, L. 123, 365, 373 software enterprise management 229 open source 227 systems 217, 218 Sokoloff, K. 334 Solimano, A. 439, 440 Somaya, D. 329, 330 startup firms 332 Stephan, P. 255, 256 Sternberg, R. 378 Stigler, G. 212 Stiglitz, J. 58 Stocking, G. 122 Storper, M. 278 Sturchio, J. 71 Summers, R. 450 Sutton, J. 74 Swann, G. 198, 201, 203 Swanson, R. 77 systematisation 213 tacit knowledge 10, 11 taste aspiration 160 association 160
bandwagons of 190 conformity 189 distinction 159, 160, 189 path dependence of 162, 163, 182 price as a measure of 165, 166 volatility of 160, 161–2 see also art technological commons 57 technological disparities between firms 210 technological diversification interrelatedness of technological activities 153–4 link to product diversification 153 motives for 153 see also technological specialisation technological knowledge 4, 5 collective activity, as 56–7 technological monopolies 23 different monopolies in different markets 39 increasing returns to adoption 25, 26, 39 see also competing technologies ‘technological opportunity’ 209 technological specialisation 362 research study data 121–2 measure of specialisation 122–8 see also American Telephone and Telegraph; Bayer; Du Pont; General Electric; IG Farben; knowledge specialisation; technological diversification technology adoption, see adoption of new technology ‘technology of learning’ 208–9 technology suppliers 331, 344 technology transfer 208 Teece, D. 299, 340, 343 Thomas, L. 84, 85 Throsby, D. 160 Tijessen, R. 369 Tirole, J. 302 Toniolo, G. 309, 310 Torre, A. 258 Trickett, A. 24 Tversky, A. 55
Index university-based research 82, 241, 344, 345, 346, 347 commercialisation 345 impact on academic norms 347–9 licensing 345, 346 patents 345, 346 university spin-offs 77 valuation of technology 333 van Gogh, V. 160, 181, 182, 183, 188 van Wijk, E. 369 Vaucanson, J. 244, 249 Veblen, T. 159 Venables, A. 26 venture capitalists 331–2 Vernon, J. 78, 84 Vicari, S. 309, 310 Vincenti, W. 347, 366 von Hippell, E. 243
von Krogh, G. 309, 313 Vopel, K. 395, 420, 421 Walras, M.-E.L. 13, 14 Walsh, J. 341 Waren, A. 180 Watkins, M. 122 Wenger, E. 306,311 White, M. 165 Whited, T. 439 Wilkins, M. 122 Wilkinson, L. 215 Wilson, G. 249 Winter, S. 26, 118, 120, 136, 299 Witt, U. 55 Young, A. 212 Ziedonis, R. 329, 340, 346, 347, 420 Zucker, L. 278 Zuscovitch, E. 301
485
