Flexibility and Stability in the Innovating Economy

FLEXIBILITY AND STABILITY IN THE INNOVATING ECONOMY This page intentionally left blank Flexibility and Stability in...

Author: Maureen McKelvey | Magnus Holmén

10 downloads 1072 Views 2MB Size Report

This content was uploaded by our users and we assume good faith they have the permission to share this book. If you own the copyright to this book and it is wrongfully on our website, we offer a simple DMCA procedure to remove your content from our site. Start by pressing the button below!

Report copyright / DMCA form

DOWNLOAD PDF

FLEXIBILITY AND STABILITY IN THE INNOVATING ECONOMY

This page intentionally left blank

Flexibility and Stability in the Innovating Economy Edited by MAUREEN McKELVEY and MAGNUS HOLME´N

1

3

Great Clarendon Street, Oxford OX2 6DP Oxford University Press is a department of the University of Oxford. It furthers the University’s objective of excellence in research, scholarship, and education by publishing worldwide in Oxford New York Auckland Cape Town Dar es Salaam Hong Kong Karachi Kuala Lumpur Madrid Melbourne Mexico City Nairobi New Delhi Shanghai Taipei Toronto With ofﬁces in Argentina Austria Brazil Chile Czech Republic France Greece Guatemala Hungary Italy Japan Poland Portugal Singapore South Korea Switzerland Thailand Turkey Ukraine Vietnam Oxford is a registered trade mark of Oxford University Press in the UK and in certain other countries Published in the United States by Oxford University Press Inc., New York ß Oxford University Press 2006 The moral rights of the author have been asserted Database right Oxford University Press (maker) First published 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, without the prior permission in writing of Oxford University Press, or as expressly permitted by law, or under terms agreed with the appropriate reprographics rights organization. Enquiries concerning reproduction outside the scope of the above should be sent to the Rights Department, Oxford University Press, at the address above You must not circulate this book in any other binding or cover and you must impose the same condition on any acquirer British Library Cataloguing in Publication Data Data available Library of Congress Cataloging in Publication Data Data available Typeset by SPI Publisher Services, Pondicherry, India Printed in Great Britain on acid-free paper by Biddles King’s Lynn, Norfolk ISBN 0-19-929047–4

978-0-19-929047–5

1 3 5 7 9 10 8 6 4 2

Contents

List of Figures List of Tables List of Boxes Contributors Preface

1 Introduction

x xi xii xiii xvii

1

Maureen McKelvey and Magnus Holme´n 1.1 Introduction 1.2 Why this book? 1.3 Overview of chapters 1.3.1 Chapters in Theme 1: experimenting and inertia 1.3.2 Chapters in Theme 2: evolution and adaptation of structure 1.3.3 Chapters in Theme 3: innovating and technological transformation 1.4 Beyond this book

1 3 11 12 14 16 18

THEME 1: EXPERIMENTING AND INERTIA

25

2 The New Craft Skills of Engineering: The Impact of Innovation Technology on Engineering Practice

27

Mark Dodgson, David M. Gann, and Ammon Salter 2.1 Introduction 2.2 Changing nature of engineering practice and knowledge 2.3 Case studies 2.3.1 Arup 2.3.2 Ricardo Engineering 2.4 Discussion and conclusions

27 28 32 33 38 42

v

Contents

3 Innovative Opportunities and Dependencies: Illustrations from Mobile Communications

48

Magnus Holme´n, Mats Magnusson, and Maureen McKelvey 3.1 3.2 3.3 3.4

Introduction Innovative opportunities Innovative opportunities in 3G and i-mode Dependencies in innovative opportunities 3.4.1 Perceived economic value 3.4.2 Perceived ability to mobilize resources 3.4.3 Perceived appropriability 3.5 Conclusions

4 The Great Experiment: Public–Private Partnerships and Innovation in Design, Production, and Operation of Capital Goods in the UK

48 49 53 58 59 61 63 66

73

Andrew Davies and Ammon Salter 4.1 Introduction 4.2 Empirical and theoretical background 4.3 Innovation in capital goods and repositioning for PPPs 4.3.1 The capital goods value stream 4.3.2 Repositioning in the value stream 4.3.3 Moving from unique to repeatable solutions 4.3.4 Challenges of providing PPPs 4.3.5 Impact of PPP on government departments and agencies 4.4 Discussion and conclusions

73 75 81 83 86 88 89 90 91

THEME 2: EVOLUTION AND ADAPTATION OF STRUCTURE

97

5 Complexity, Evolution, and the Structure of Demand

99

John Foster and Jason Potts 5.1 5.2 5.3 5.4 5.5 5.6 5.7 5.8 5.9 5.10

vi

Introduction Consumption networks not production functions The economy is a complex rule-system The growth of demand and the growth of economies Correlated preferences Analysis of economic networks Orders of complexity Micro–meso–macro The complexity of consumption and demand Evolution and aggregate demand

99 101 103 104 106 107 109 111 114 117

Contents

6 Self-Transformation, Self-Organization, and Evolutionary Adaptation in the Economic Process

121

J. Stan Metcalfe and Ronnie Ramlogan 6.1 6.2 6.3 6.4 6.5 6.6 6.7

Introduction Some evidence for structural adaptation Accounting for evolutionary adaptation The logistic principle Adaptation, complexity, and the problem of knowledge The correlation of knowing Conclusions

121 128 135 140 144 148 152

7 Changing Boundaries of Firms in the Evolution of the Computer Industry: Towards a History-Friendly Model

157

Franco Malerba, Richard Nelson, Luigi Orsenigo, and Sidney Winter 7.1 Introduction 7.2 The conceptual background 7.3 A brief discussion of the semiconductor and computer industries 7.4 Some theoretical statements on the changing vertical boundaries of ﬁrms 7.4.1 First prediction–vertical integration 7.4.2 Second prediction–vertical integration 7.4.3 Third prediction–vertical integration 7.4.4 Fourth prediction–disintegration 7.4.5 Fifth prediction–disintegration 7.4.6 Conclusions from the theoretical statements 7.5 The model 7.5.1 Computers 7.5.2 The market for components 7.5.3 Firms’ behaviour and technical progress 7.5.4 Demand for computers 7.5.5 Vertical integration and specialization 7.5.6 The working of the model with two technological discontinuities in components 7.6. The simulations 7.6.1 The benchmark case: vertical integration 7.7. Conclusions

157 158 160 163 163 164 164 164 165 166 167 167 167 168 170 171 173 175 175 193

vii

Contents THEME 3: INNOVATING AND TECHNOLOGICAL TRANSFORMATION

199

8 The Effects of Technological Change on the Boundaries of Existing Firms

201

Paul L. Robertson and Gianmario Verona 8.1 Introduction 8.2 Technological stability and the boundaries of the ﬁrm 8.2.1 Core and distinctive competences 8.2.2 Transaction costs 8.2.3 Behavioural factors 8.2.4 Summary 8.3 Technological change and the boundaries of the ﬁrm 8.3.1 Dynamic capabilities 8.3.2 Dynamic transaction cost 8.3.3 Modularity and ﬁrm boundaries 8.4 Conclusions

9 Transitions, Transformations, and Reproduction: Dynamics in Socio-Technical Systems

201 205 206 207 208 209 209 209 212 214 217

227

Frank W. Geels and Rene´ Kemp 9.1 Introduction 9.2 Multilevel perspective and types of change 9.2.1 Reproduction 9.2.2 Transformation 9.2.3 Transition 9.3 Case studies 9.3.1 The hygienic transition from cesspools to integrated sewer systems in the Netherlands (1870–1930) 9.3.2 The transformation of Dutch waste management (1960–2000) 9.4 Conclusions and policy implications 9.4.1 Steering and management 9.4.2 Transition management in the Netherlands

viii

227 229 234 235 235 236 236 243 249 251 252

Contents

10 Analysing Flexibility and Stability in Co-evolutionary Processes

257

Magnus Holme´n and Maureen McKelvey 10.1 Introduction 10.2 Transformation as involving novelty, destruction, or renewal? 10.2.1 The character of change 10.2.2 Doing and interpreting empirical research 10.3 Co-evolutionary processes in the innovating economy 10.4 Discussion Index

257 259 260 262 266 274 283

ix

List of Figures

4.1 The industry value stream in capital goods

84

4.2 The system–service innovation cycle

85

6.1 (a) Correlation of employment shares

131

6.1 (b) Correlation of output shares

131

6.1 (c) Correlation between employment and output

131

6.2 Adjusted Herﬁndahl index, 1958–96

132

6.3 Employment share weighted labour productivity

133

6.4 Distribution of percentage productivity change, 1958–96

133

6.5 (a) The Fabricant relationship

135

6.5 (b) Fabricant coefﬁcients for manufacturing

135

6.6 Logistic process

143

6.7 Logistic share proﬁle

144

7.1 (a–f) Benchmark run

177

7.2 (a–f) Specialization: large external market

180

7.3 (a–f) Vertical integration, no uncertainty in demand

182

7.4 (a–f) Vertical integration, high in-house capabilities, cumulativeness of technological advances

184

7.5 (a–f) Changing boundaries of ﬁrms

187

7.6 (a–f) Specialization: high role of technological uncertainty

189

7.7 (a–f) Integration decision depends only on the maturity of the technology 191 7.8 (a–f) The disappearance of the component industry

194

8.1 Activities within the boundaries of a ﬁrm

210

9.1 Three interrelated analytic dimensions

230

9.2 A dynamic multilevel perspective on system innovations

233

9.3 Reduction of landﬁll

246

9.4 Transformation of waste management

247

9.5 Arrows of inﬂuence in the Dutch waste management transformation (1970–2000)

248

x

List of Tables

1.1 An initial conceptualization of ﬂexibility and stability

5

1.2 Conceptualizing ﬂexibility and stability in the three themes

10

3.1 Examples of ﬁrms that abandoned 3G markets around year 2001

56

8.1 Summary of the expected outcomes of the propositions

218

9.1 Different mechanisms in change processes

234

9.2 Rejected plans for sewer systems

241

10.1 Outlining questions about novelty, destruction, and renewal at different levels of analysis

275

10.2 Concepts relevant for type of change occurring in economic systems

276

xi

List of Boxes

4.1 The case of Alstom Transport

88

7.1 Computers

167

7.2 The market for components

168

7.3 Technological progress

169

7.4 Demand for computers

170

7.5 Vertical integration and specialization

172

8.1 The impact of ICT on ﬁrm boundaries

218

xii

Contributors

Andrew Davies (SPRU) is Principal Research Fellow at the Innovation Studies Centre, Tanaka Business School, Imperial College London. He has published widely on the management of innovation in complex products and systems. Mark Dodgson (Prof.) is Director of the Technology and Innovation Management Centre at the University of Queensland Business School, Australia. He has published extensively on innovation management and policy. John Foster (Prof.) is Head of the School of Economics and Acting Dean at the University of Queensland, Australia. He does research in the use of complex system theory in economics; the application of self-organization theory to statistical/ econometric modelling in the presence of structural transition; and the role of entrepeneurship and innovation in economic growth. David Gann (Prof.) holds the Chair in Technology and Innovation Management at Imperial College, London. He is also Head of Entrepreneurship, Innovation and Technology Management in the Tanaka Business School. He has worked extensively with government on innovation policies and advises a number of leading ﬁrms on their innovation strategies. He has published widely in the literature on innovation management focusing speciﬁcally on ﬁrm-level strategies and the development and growth of project-based ﬁrms. Frank Geels (Dr) is a post-doctoral researcher in the Department Technology Management at Eindhoven University of Technology, the Netherlands. His main research topics are technological transitions and system innovations, studying combinations of disciplines such as the sociology of technology, innovation studies and evolutionary economics with historical case studies and possible future case studies. Magnus Holme´n (Dr) is Research Associate of the National Graduate School of Management at the Australian National University, Australia (www.ngsm.anu. edu.au). He is an expert on innovation system studies and on techno-economic changes in ICT. Rene´ Kemp (Dr) is Senior Research Fellow at the Maastricht Economic Research Institute on Innovation and Technology at Maastricht University, the Netherlands, and he is Senior Advisor of TNO-STB (www.meritbbs.unimaas.nl/rkemp). He has written extensively about environmental issues as a policy challenge. Maureen McKelvey (Prof.) is Director of the R&D, Innovation and Dynamics Centre at Chalmers University of Technology, Sweden (www.chalmers.se/tme).

xiii

Contributors Her research addresses innovation systems as relevant for industrial and policy studies, with a particular interest in biotechnology and open source software. Mats Magnusson (Dr) is Senior Lecturer at the Department of Innovation Engineering and Management in the School of Technology Management and Economics at Chalmers University of Technology, Sweden (www.chalmers.se/tme). His main research interests are innovation strategy and organization, entrepreneurship, and organizational learning. Franco Malerba (Prof.) of Bocconi University, Italy, is Director of CESPRI research centre (www.cespri.it). His work on industrial dynamics is well recognized, including work while serving as President of the International Schumpeter Society and while coordinating two large Eu projects: European Sectoral Systems of Innovation and Knowledge-Based Entrepreneurship in Networks. Stan Metcalfe (Prof.) is Executive Director of the ESRC Centre for Research on Innovation and Competition (CRIC) at the University of Manchester, UK (www. cric.ac.uk). He is a leading ﬁgure in evolutionary economics and in the innovation communities, with CRIC being an active centre for policy debates, academic workshops, and visiting researchers. He also recently chaired a DTI working group on innovation. Richard Nelson (Prof.) is the George Blumenthal Professor of International and Public Affairs at Columbia University. His research has been largely focused on the process of long run economic change, with particular emphasis on technological advance and on the evolution of economic institutions. His recent work has looked at the varied roles of government in modern mixed economies. Lnigi Orsenigo (Prof.) of Brescia University, Italy, is Deputy Director of CESPRI (www.cespri.it). He is a leading expert on transformations in biotechnology and pharmaceutical industries as well as in evolutionary economics. Jason Polts (Dr) is a lecturer in the School of Economics at the University of Queensland, Australia (www.uq.edu.au/economics/staff/polts.htm). He does research in the ﬁeld of foundations of evolutionary economic analysis and complex systems theory. Ronnie Ramlogan (Dr) is Research Fellow at the ESRC Centre for Research on Innovation and Competition (CRIC) at Manchester University, UK (www.cric. ac.uk). He has been working on problems of complexity and innovation in medicine. He holds a PhD in development economics from Manchester University. Paul Robertson (Prof.) is a professor at the Graduate School of Management at Grifﬁth University (www.grifﬁth.edu.au). His principal research interests are in technological strategy and in technology and economic development. Ammon Salter (Dr) of Imperial College, London, is Research Fellow at the Innovation Studies Centre at the Tanaka Business School. He has published widely on the

xiv

Contributors management of innovation in project-based industries and sources and determinants of innovative performance. Gianmario Verona is Associate Professor of Management at Bocconi University, Italy (www.unibocconi.it). He is Associate Director of the PhD in Business Administration and Management, and Senior Lecturer at the SDA Bocconi, the Graduate School of Management at Bocconi University. He has published several articles on innovation and technology in international journals. Sidney Winter (Prof.) is the Deloitte and Touche Professor of Management at Wharton School in the University of Pennyslvania, and is Co-Director of the Reginald H. Jones Centre for Management Policy, Strategy and Organization. He does research within areas of ﬁrm capabilities technological change, and competitive advantage (www.wharton.edu/faculty/winters.html).

xv

This page intentionally left blank

Preface

A book project is always a journey, taking you to both expected and unexpected places. This book should allow us to address evolutionary processes and complex transformation in the economy, with a particular emphasis on ﬂexibility and stability. We would particularly like to thank the Ruben Rausing Foundation in Sweden, which made this creative journey possible. This book has combined an intellectual and global journey by the editors and many authors. Usually, much of the journey associated with a book is an intellectual one, because the interactions among people with different opinions—despite agreement on some core issues—will challenge each author to explore new ways of conceptualizing and explaining the world around us. This has deﬁnitely been the case with this book, exploring as it does relatively uncharted territory for the communities of innovation studies, industrial dynamics, and evolutionary economics. And yet this book has also been associated with journeys around the globe, in that we have had book meetings in Sweden, Australia, and the UK over a two-year period. So, we shall remember this book as a particularly memorable journey, combining interesting debates and disagreements with wine and dining across the globe. The editorial idea was to start with authors and end up with a book. Thereby, the editors had the task of approaching the authors, asking for abstracts, and starting the discussion of book themes amongst the whole group of authors. All authors have been involved in commenting upon each others’ chapters and in articulating the core book themes. The criticisms have at times been tough—but the end result is what we believe to be a set of novel and challenging scholarship. Authors were approached and initial abstracts submitted and selected. A small project meeting was held at Chalmers University of Technology, Sweden, in May 2003, to initiate work on the book. The draft book chapters were discussed at a three-day workshop held in Brisbane, University of Queensland (UQ), Australia, in March 2004. In addition to the book discussions, we held an open ‘International Workshop on Innovation Research and Policy’ at the UQ Business School, Australia. The revised chapters were discussed at a ﬁnal meeting held in November 2004 at Imperial College, London, UK. We would argue that this series of workshops in Sweden, Australia, and the UK have been critical for ‘making’ ﬂexibility and stability core research themes for this book.

xvii

Preface We would like to thank all the authors—and Keith Smith—for their contributions to developing and realizing this project, including commenting upon all chapters. We have presented versions of the editors’ chapters at various conferences, especially DRUID and the International J. A. Schumpeter society. Sven Lindmark commented upon Chapters 1 and 3. In addition, we would like to thank colleagues at our local research environments at, respectively, Department of Technology Management & Economics and RIDE research center at Chalmers University of Technology and at Australian National University (ANU). Particular thanks to Mark Dodgson and John Foster for helping to organize the two conferences at UQ and to David Gann and Ammon Salter for organizing the workshop at Imperial College. The Institute for Management of Innovation and Technology (IMIT) has done a great job in administering the grant from the Ruben Rausing Foundation, with particular thanks to Birgitta Andersson, Bengt Karlsson, and Christer Karlsson. Mattias Jonsson and Eva Burford helped in the preparation of the ﬁnal book manuscript. The usual caveats apply. Finally, we would like to acknowledge sources of ﬁnancing for this speciﬁc book project. Maureen McKelvey would like to especially thank the Ruben Rausing Foundation. Their grant has led to a number of other papers and book chapters. And yet, this book is the most visible outcome of a series of interesting debates on what ﬂexibility and stability mean in the modern economy. Maureen would also like to acknowledge support from two strategic projects that she leads at Chalmers University of Technology: IT and Biotechnology and Economics of the Innovation. These projects have been vital in building and sustaining a group of researchers, interested in many related questions about the nature of innovation and industrial dynamics in the modern economy.

xviii

1 Introduction Maureen McKelvey and Magnus Holme´n

1.1 Introduction Flexibility and Stability in the Innovating Economy is a book that addresses the nature of industrial dynamics, structural change, and transformation in our time. The aim is to understand, conceptualize, and explain what changes and what does not in the economy, and how individual actors and ‘systems’ relate to one another through differences in perceptions and actions. Thus, the central focus of this book is on the question of evolutionary processes and complex transformation in the economy, with a particular emphasis on the nature of ﬂexibility and stability. One common starting point for all chapters in this book is that innovation and entrepreneurship continue to disrupt the economy, thereby sometimes fundamentally changing activities and moving the economy in new directions. At other times, resistance and inertia may be more prevalent. This can lead to ‘tensions’ across the system, where tensions can arise between actors, elements, and processes that tend to exhibit ﬂexibility and those that tend to exhibit stability, in different parts of the economic system. Such tensions spring, for example, from differential rates of change, from the variable abilities of actors to respond to systemic changes, and from the existence of both turbulence and inertia at different levels within the same system. Conceptualizing actors—whether individuals, ﬁrms, or other organizations—as capable of innovating and learning has important implications for understanding economic transformation over time. This can be illustrated through two opposing sets of assumptions. In one set, researchers may simply assume that existing organizational forms survive and that known information provides all the relevant signals about the economic system. In this case actors only have to adjust to these signals, and various actors likely interpret the signals in similar ways. Within the mainstream of economics, the analytical framework rests on given environmental conditions to which ﬁrms adapt in optimal ways; changes in external conditions generate reactions by ﬁrms,

1

Introduction and consequently new positions of ﬁrms in the competitive landscape. This is a common way of thinking about industrial dynamics and the economy as a system—and in this case, the actors just need to modify behaviour to adjust to those signals, and thereby the system as a whole makes a transition to a new phase. This book makes a set of rather different assumptions that imply that economic transformation involves fundamental change and is driven by actors, a view that has implications for how we need to understand and explain networks and systems. We assume that knowledge and information are separate, that the value and validity of much information is uncertain and difﬁcult to interpret, and that information does not evenly distribute among actors but is only localized in certain places, industries, or actors. In this case, no one actor can be assumed to have all the correct and relevant ‘signals’. Moreover, it is impossible for any decision-maker to have full information and to assess all possible outcomes. Change is a historical process, ﬁlled with uncertainty. The present and future must therefore be seen as the result of history as a process unfolding, with different possible branching points and trajectories. The theoretical assumptions made in this book have a number of implications for understanding ﬂexibility and stability in the innovating and transforming economy. Under these conditions, various actors will make ‘mistakes’ and use diverse assessments of the future when trying to interpret and react to signals. As knowledge may be distributed among different actors, they may decide to team up to pool resources, to network, or to make alliances for speciﬁc goals. Therefore, what actors perceive and the problems they select, what they do, how they learn, and how well they can use and transform that knowledge to solve later problems or take advantage of later opportunities are crucial. This requires a new way of problematizing such endogenous, innovation-driven processes. It also implies that diverse actors will make choices in the present and future that will also fundamentally change the rate, direction, and outcomes, or future trajectories, of the economic system. Finally, the economic system as a whole will transform internally through aspects such as self-organization so that economic transformation is irreversible. Thereby the concept of economic transformation is deﬁned here in a speciﬁc way, in relation to the ongoing processes of industrial dynamics, structural change, and transformation. Economic transformation as a concept used in this book refers to a non-reversible process, encompassing quantitative and qualitative changes in components and connections, driven by opportunities and innovations. Such economic transformation may well be driven by processes of complexity and self-organization as well as processes of actors acting, adapting to contexts. Moreover, the concept of transformation, as used here, may result from very different processes, including ones driven by very large and discontinuous changes as well as ones driven by very small changes,

2

Introduction which follow upon an existing trajectory. We would argue that such qualitative transformation is a normal part of the market economy—not a special event.1 Hence the authors in this book share a common approach to the economy as an inherently dynamic and complex system, consisting of diverse, changing, and interacting components and activities. We will argue that industrial dynamics primarily consists of endogenous, innovation-driven transformation of the economy, which in turn incorporates aspects of business, technological, public policy, and organizational processes. Framing the book this way stresses the necessity of viewing the economy as a sequence of processes and temporally dependent events occurring in irreversible time. But in doing so, this book also stresses that diverse actors are deciding, doing, and acting in ways that affect the outcome in evolutionary processes and complex transformation of the economy. Section 1.2 asks the question, ‘Why this book?’ in order to place this book in its intellectual framework and specify our book themes, which are further developed throughout the nine subsequent chapters. Section 1.3 provides an overview of each of these nine chapters. A short paragraph introduces the title, authors, and main ideas of each chapter. These chapters are grouped into the three themes, each of which plays upon and reﬂects the book title in some way. Section 1.4 engages the reader to consider issues ‘Beyond this book’. The reason for this section is that although Flexibility and Stability in the Innovating Economy addresses issues related to industrial dynamics, structural change, and transformation, which are issues of interest to many readers, much more research may be developed around the speciﬁc issue of ﬂexibility and stability in the innovating and transforming economy.

1.2 Why this book? As stated earlier, this book focuses on the question of evolutionary processes and complex transformation in the economy, with a particular emphasis on the nature of ﬂexibility and stability. This book contrasts with, but also contributes to, the vast literature on industrial dynamics, evolutionary economics, systems of innovation, management of innovation and technology, history of science and technology, economic history, and economic transformation. This section explains where our thinking started in this book project, in order to address these issues, as well as how the development of ideas led to three themes and ten chapters by 19 researchers. The key concepts addressed within this edited volume are ‘ﬂexibility’ and ‘stability’ set in relation to the innovating and transforming economy. From the start of this project, these concepts focused our attention on three issues. First this book project started from the idea of ‘change and not change’ and of differential change in economic transformation. These concepts signify that

3

Introduction change de facto occurs at different rates, in different directions, on different levels, and in different dimensions of the economy. Our communities in innovation studies, industrial dynamics, and evolutionary economics tend to stress the ‘changing’ aspects rather than continuity (Constant 2002). Still, ‘change’ can only be understood in contrast to something that is not changing. Hence, when we started writing this book one issue that was highlighted is how and why different endogenous dimensions of an economy change and how other dimensions remain stable over some period of time. Second, we started from the premise that aspects of ﬂexibility and stability would require a conceptualization that could lead to debates about how and why diverse actors drive systems, as well as how and why systems constrain and enable actors, albeit perhaps on different timescales. This implied that the book should bring together scholars working on related topics but still different issues, and in quite different ways. This also implied that relationships between different levels of analysis would be particularly important, such as whether and why different actors such as ﬁrms and organizations also constitute a systemic level, such as sectors or nations. Third, this book project also wanted to raise the question of whether the nature of ﬂexibility and stability is in some ways related to debates about the subjective choice of actors relative to an objective, heterogeneous albeit structured reality. The entrepreneurial ﬁrm may, for example, experience change as rapid and destructive because many ﬁrms are being started, going bankrupt, diversifying while at the industry level, market concentration remains constant over a long period, such that change appears slow and conservative. Hence to the extent that authors wished to address the issue of subjective versus objective ‘reality’, we started with the idea that the perspective likely depended on differing interpretations of economic transformation, in terms of aspects such as the rate, direction, level, and dimension of change. These three issues formed the starting point for why this book project examines ﬂexibility and stability in the innovating and transforming economy, and all three are classical themes within, say, sociology, organizational science, and evolutionary economics. Yet relative to these rich debates, little research has addressed the conceptual and analytical foundation upon which we can analyse ﬂexibility and stability in economic transformation. Using the concepts of ‘ﬂexibility and stability’ and of ‘actor level and system level’, Table 1.1 provides an overview of relevant and sometimes overlapping concepts that are useful to analyse evolutionary processes and complex transformation in the economy. Table 1.1 illustrates many relevant concepts for such research. These notions can be related, where e.g. ‘learning’ at the ﬂexibility/actor level can be related to ‘routine’ at the stability/actor level and to ‘responsiveness’ at the ﬂexibility/system level. Table 1.1 clearly shows that linking ﬂexibility and stability to the level of analysis opens up new vistas of analysis. Depending on the research questions and the theories developed,

4

Introduction Table 1.1. An initial conceptualization of ﬂexibility and stability Flexibility

Stability

Actor level (e.g. a ﬁrm, university, organization, government).

Learning Adaptability New connection

Routine Resistance Regeneration of competencies and resources Rigidity

System level (e.g. national economy, network, industry, sectoral, and regional innovation systems).

Responsiveness Adaptation Turbulence Phase shift Transition

Rigidity Inertia Instituted practices Trajectory

these types of concepts may be useful to highlight particular puzzles. One set may be useful to identify aspects such as the rate and direction of change at different levels and in different dimensions of the economy. But then the question becomes, how and why are they related? Another set may be useful to explain why it is that what at one level appears to be turbulent and persistent change, appears, at another level, to exhibit consistency over time. Similar concepts may be relevant within several ‘boxes’ or across several theories. Hence, we mean that the conceptualization of ﬂexibility and stability as a research topic opens up new vistas of research, in order to ask questions such as: . How much, and why do the actors have ‘ﬂexibility’ systemic dynamics?

in relation to

. How might different dimensions and characteristics of the system help drive ‘ﬂexibility’ in responses by actors? . How might elements which seem to indicate ‘stability’ at one level still require much diversity, learning, and turbulence among the particular set of actors? Taking up ﬂexibility and stability in the innovating and transforming economy as a research topic has meant that not only do we need to address many new questions but also we may need to reinterpret existing questions through new insights. And each question represents ideas—and opens up room for new ideas—which can be debated and further developed beyond this book. Indeed, once one begins thinking in these terms, it is clear that existing literature abounds with many different sets of concepts that help us capture and highlight aspects of such processes, but neglect other aspects. This implies that different and sometimes almost contradictory concepts are used in the literature. For example, the idea of routines has been a recurring theme within evolutionary economics—and yet much of this literature and of entrepreneur-

5

Introduction ship literature stresses the opposite concept, namely the need for quick adaptability and new connections. Similarly, literature about national institutional frameworks often stresses aspects of inertia and trajectories in explaining the differential performance of national economies—and yet much of this literature is also concerned with points of turbulence and the need to respond to new regimes and trends, such as the globalization of capitalism. So, through this book, we wished to ﬁnd ways to discuss and debate across narrow academic specializations in order to address the larger issues. To carry on such a debate, we simply wish to start by reminding ourselves and the reader that even though this book broadly adheres to the modern Schumpeterian, or evolutionary economics tradition, there is a historical legacy for understanding industrial dynamics, structural change, and economic transformation. Arguably, the conceptualization of ﬂexibility and stability in industrial dynamics and economic transformation needs to be understood in terms of the writings of many classic as well as more modern scholars. Much of the modern understanding of economic transformation draws upon and extends the insights of scholars such as Joseph Schumpeter, Adam Smith, Carl Menger, Karl Marx, Alfred Marshall, Thorstein Veblen, Max Weber, and Allyn Young.2 For example, this understanding may draw upon Adam Smith’s and Allyn Young’s views that the progression of the division of labour and specialization is both dependent upon and shapes the size of the market in a non-tautological way. Or, Karl Marx’s view that basic science and technological change are harnessed to the productive core of capitalism; Joseph Schumpeter’s view that economic transformation is endogenously driven by innovations and entrepreneurship; Carl Menger’s view that the subjective nature of needs and wants and complementarities and causal interdependencies among economic goods and knowledge drive the evolution of sectors; and Thorstein Veblen’s view that economic change is non-teleological and that habits of thought are at the heart of any economic activity. Hence, a long historical legacy can be developed for many modern ideas of relevance to ﬂexibility and stability. While encompassing many scholars, modern industrial dynamics can still be described in part as the legacy of Joseph Schumpeter (Hanusch 1999). Schumpeter’s key insight was that change is endogenous to economic systems: it is not imposed from without, but rather is generated within, and it is that insight that is followed here. Schumpeter argued that fundamental change in existing activities as well as the introduction of entirely novel activities would keep providing the ‘fuel’ to the capitalist engine. Through innovations and creative destruction, the existing and old activities are modiﬁed but keep existing—as well as helping to give rise to new ones. This can be claimed to be a forward looking view of the economy where heterogeneous expectations and actions among a range of different organizations and individuals keep shifting and reorienting economic activities.

6

Introduction Nelson (1996: 87) argues that this is ‘Schumpeter’s most consistent and elaborated argument about innovation and economic transformation, that it fundamentally involves disequilibrium and that standard equilibrium theory in economics cannot cope with it and its economic consequences’. In other words, innovation and entrepreneurship continue to disrupt the economy, thereby fundamentally changing activities and moving the economy into new directions. This part of the Schumpeterian legacy can be best captured by a quotation. In Capitalism, Socialism and Democracy, Schumpeter ([1947](1975): 82–3) states: Capitalism, then, is by nature a form or method of economic change and not only never is but never can be stationary. And this evolutionary character of the capitalist process is not merely due to the fact that economic life goes on in a social and natural environment which changes and by its change alters the data of economic action; this fact is important and these changes (wars, revolutions and so on) often condition industrial change, but they are not its prime movers. Nor is this evolutionary character due to a quasi-automatic increase in population and capital or to the vagaries of monetary systems of which exactly the same thing holds true. The fundamental impulse that sets and keeps the capitalist engine in motion comes from the new consumers’ goods, the new methods of production or transportation, the new markets, the new forms of industrial organization that capitalist enterprise creates.

In this quote, Schumpeter provides a broad view of the importance of internal processes of change within many aspects of the economy. Although Schumpeter stresses that different types of innovations drive economic change, this book interprets this view to mean that diverse actors are driving such innovation processes. Schumpeter emphasized two broad types of change—a more radical and a more incremental type. These have been tested in the sense of hypotheses about what degree and type of change tends to matter the most for long-term economic growth. On the one hand, there is likely to be radical, disruptive change that rests on new knowledge bases and the rise of new industries. This dimension of change is stressed by Schumpeter himself in his major work, Business Cycles. More recent innovation scholars like Freeman and Perez (1988) stress that major shifts in the ‘techno-economic paradigm’ spark productivity increases across the economy where new technologies are accompanied by institutional shifts. This line of argumentation can be found within other parts of economics, where for example Kuznets (1954, 1959) stresses that economic growth is driven by the emergence of new industries. On the other hand, the second type of change is incremental in that the economy exhibits continuous small changes in products and processes. These are incremental shifts that, over time, create major changes in outputs, methods, skills, and productivity

7

Introduction levels. This dimension of change relates to the argumentation about how division of labour and specialization drives economic growth (Young 1928) and it does so by stressing the long-term impacts of a series of small changes. In terms of ﬂexibility and stability, the debates about radical and incremental innovations can be linked to level of analysis, as well as aspects of rate, direction, and outcomes of innovation processes. Another important writer for understanding ﬂexibility and stability was Menger ([1871](1976). A major contribution of Menger was that he stressed the importance of individuals and subjectivity through the focus on (subjective) needs and wants. From this perspective, Menger and his successors went on to outline how knowledge and interdependences among heterogeneous capital goods shape or determine industrial organization.3 Flexibility comes about in that different actors may choose different uses of capital goods, but at the same time, the choice set is constrained because of the nature of the individual capital goods as well as its overarching structure, providing some stability to the economic system. In terms of ﬂexibility and stability, this perspective opens up debates about ‘subjective’ choice versus ‘objective’ systemic conditions and how an existing structure may affect problem or opportunity identiﬁcation and selection. Thus, in somewhat different ways, this book draws upon a rich historical legacy. The classical economists focused on key issues underlying transformation, such as how increases in productivity come about, the sources of economic returns, the reasons and returns on investment into technology and R&D, the importance of organizational design, and the like. Such issues have remained important in modern scholarship, represented in many articles and books and thereby remain key for analysing evolutionary processes and complex transformation in the economy. And yet there has previously been too little attention on ﬂexibility and stability within such processes. This is problematic in the view of the editors in that we argue that the issue of ﬂexibility and stability are at the core of evolution and transformation, and as such need to be studied more systematically. However, given the relatively little direct attention, it is far beyond the scope of this book to ‘solve’ this issue. Therefore, this book and each chapter in this book are set to problematize issues related to ﬂexibility and stability in an ever-evolving economy, rather than to solve a known and well-deﬁned problem. This book provides a collection of new scholarship that can debate these issues. In the process of writing and debating each chapter and the book as a whole, we have sharpened our core common assumptions as well as points of differences. Our broadly Schumpeterian approach to understanding the economy provided some shared intellectual commons for this endeavour—including an understanding that business, technological, public policy, and organizational changes are somehow linked in driving endogenous innovation-driven

8

Introduction transformation. Beyond that, all nine of the subsequent chapters problematize issues related to Flexibility and Stability in the Innovating Economy, not only in terms of their own contribution, but also in relation to whether they mainly focus on ‘ﬂexibility or stability’ or on ‘actor and component level or system level’. Speciﬁc concepts may appear at different levels, and authors may deﬁne similar concepts in somewhat different ways. Clearly, each subsequent chapter will therefore make its own unique contribution. The processes underlying ﬂexibility and stability are difﬁcult to catch because they are complex, each contribution can only focus on a narrow aspect, as authors use their chosen set of theories, providing their unique interpretation of a research question. Yet these chapters also relate to each other by taking their own position within debates of relevance to the book as a whole. The chapters have therefore been grouped into three themes which reﬂect the book title Flexibility and Stability in the Innovating Economy: Theme 1: Experimenting and inertia Theme 2: Evolution and adaptation of structure Theme 3: Innovating and technological transformation Each of these three themes thus reﬂects and plays upon the issue of ﬂexibility and stability in the innovating and transforming economy. The three themes are the result of discussions, writings, and debates among the authors, as interpreted by the editors. As such, these three themes are a better, if clearly still imperfect, map to addressing evolutionary processes and complex transformation in the economy, with a particular emphasis on the nature of ﬂexibility and stability. Theme 1, ‘Experimenting and inertia’, addresses what happens, as new organizational forms and new knowledge are developed in existing and new organizations. In this case, ﬂexibility and stability are interpreted as processes occurring primarily at the actor level, with dependences and interactions to the system level. The three chapters in this theme focus on questions related to how and why actors try out new things, from within existing engineering practices, existing views of business opportunities, and existing assumptions underlying contracting and provision of public goods. On the one hand, ﬁrms and other actors are engaged in search activities to do new things—such as develop and beneﬁt from knowledge and innovations. On the other hand, the actors are working, starting from old ways of doing things but face unclear choice sets about the future, for example in terms of economic environment, technological opportunities, and competitive conditions. To survive under such conditions they have to experiment. At the same time, they draw upon past experience and competencies to solve new problems thereby leading to inertia in some dimensions. Theme 1 thus reﬂects research questions that are directly related to what happens over

9

Introduction time in the economy, as actors perceive, think, decide, act, and inﬂuence others through business relationships. Theme 2, ‘Evolution and adaptation of structure’, has three chapters addressing issues about how to understand the overall evolving economy, where evolution and adaptation of structure are endogenous and can be seen as historical processes. In this case, ﬂexibility and stability are primarily interpreted at the system level while still analysing interactions to the actor level, or to components of the system. These three chapters present a view on the economy where co-evolution, adaptation, and the making of connections between elements matter because they help shape and reorient search, innovation, and production. They all start from an explicitly evolutionary economics perspective, albeit with different interpretations of what these theories mean. Hence, each chapter provides a somewhat uniquely different theoretical perspective on what structure and order mean and especially how and why they might change—or not—over time. Theme 2 is thus closely linked to Theme 1 through their common interest in the observation that the economy can be simultaneously characterized as actors’ diversity and systemic order. Theme 3, ‘Innovating and technological transformation’, has three chapters explicitly addressing innovating and technological transformation, in particular the degree and type of change occurring over time. Indeed, they have a shared interest in the role of technological change as related to different sets of processes in industrial dynamics, socio-technological systems, and co-evolutionary processes. In doing so, all three chapters consider the relationship and interactions among actor and system levels as processes involving feedback over time. They thereby set technological transformation in relation to inertia and evolution. Theme 3 focuses on a particular kind of change and each chapter presents perspectives on how and why the diversity of actor behaviour are linked to broader systemic characteristics. Table 1.2 places the six concepts of ‘experimenting’, ‘inertia’, ‘evolution’, ‘adaptation’, ‘innovating’, and ‘technological transformation’ in relation to the two dimensions of actor/sytem level and ﬂexibility/stability. Thus, it uses the same dimensions as Table 1.1, placing only the concepts of Themes 1, 2, and 3.

Table 1.2. Conceptualizing ﬂexibility and stability in the three themes Flexibility

Stability

Actor level

T1: Experimenting T2: Evolution T3: Innovating

T1: Inertia

System level

T2: Adaptation

T1: Inertia T2: Evolution T3: Technological Transfosrmation

10

Introduction In this way, Table 1.2, albeit still imperfectly, relates concepts to the central ideas and three themes explored within this book. Indeed, Table 1.2 suggests that the exploration of these themes in this book will answer certain questions but also open up new vistas and thereby new questions. For, by doing this simple exercise, Table 1.2 also outlines how concepts may stretch across either of the two dimensions in that some concepts go across several ‘boxes’. The reason is that many concepts bring several dimensions into play. While some concepts may ﬁt neatly within one box, other concepts can be categorized in several boxes because they focus on linkages, tensions, and relationships across dimensions. In summary, this book problematizes the issues identiﬁed earlier through nine chapters grouped in the three themes. In doing so, this book takes one of the challenges within evolutionary economics and innovation literature seriously, namely how and why to combine levels of analysis related to actor and system, or what has been traditionally called individual and structure, in order to explain both unique trajectories of development and general characteristics of the transformation of the economic system. Because of the complexity of the overarching issues, there is no single level of analysis or approach that will provide a correct and exhaustive answer.

1.3 Overview of chapters This section provides a short summary of each chapter, including author, titles, and abstract, and the chapters are organized within the three themes. Thereby, this section should help direct the reader who is interested in particular issues and authors to the relevant chapters—as well as giving more detailed insights into the book as a whole. As mentioned earlier, researchers writing subsequent chapters all make their unique contribution by focusing on his/her more speciﬁc research question, and they also work within rather speciﬁc academic communities. Theories used in this book therefore draw from a range speciﬁc disciplines and ﬁelds, especially economics of innovation, industrial dynamics, evolutionary economics, business history, organization studies, technology and innovation management, science and technology policy, and entrepreneurship. Thus to address speciﬁc questions subsequent chapters in this book will deal with different theories, different levels of the economic system, and different aspects or variables of the economic system to bridge a few of these problems. Nonetheless, the chapters taken together provide a means to explain and stress why analysing Flexibility and Stability in the Innovating Economy requires putting together the many diverse pieces of our common conceptual and empirical puzzle into an interesting mosaic.

11

Introduction

1.3.1

Chapters in Theme 1: experimenting and inertia

Chapters 2, 3, and 4 in Theme 1 discuss ﬂexibility and stability, primarily at the actor level. Chapter 2, ‘The new craft skills of engineering: the impact of innovation technology on engineering practice’ by Mark Dodgson, David Gann, and Ammon Salter, explores the impact of the use of innovation technologies such as simulation, modelling, and rapid prototyping on engineering practice for a business context. These innovation technologies help redeﬁne the role of engineers in the innovation process, creating a new division of innovative labour both with organizations and across organizations. In doing so, this chapter explores the boundaries of experimentation and inertia within particular domains of problem-solving to create new opportunities and value. The difference between engineering knowledge and scientiﬁc knowledge retains a central position in the main theories of innovation. And yet there is a growing body of literature focusing on the ‘distinctiveness’ of engineering problem-solving, and the nature of engineering knowledge is itself coming under increasing examination. Williams (2003) argues that engineering is witnessing two major patterns of change, one involving a deepening of knowledge in specialist domains, and the other a broadening set of interests and pursuits outside normal engineering practice. She calls these two trends the ‘expansive disintegration’ of engineering. This chapter focuses on innovation technologies in relation to experimenting and inertia. Using the experiences of two leading engineering organizations, Arup and Ricardo, this chapter explores how the use of the new technologies can help reshape engineering problem-solving and, in turn, engineering knowledge. The chapter suggests that the use of the new technologies is creating opportunities for some engineers to create new value by allowing users and clients to play with different alternatives. However, the creation of this new value will require new forms of engineering problemsolving and knowledge, often far removed from the traditional demarcations of traditional engineering practice. Chapter 3, ‘Innovative opportunities and dependencies: illustrations from mobile communications’, by Magnus Holme´n, Mats Magnusson, and Maureen McKelvey, addresses the role played by innovative opportunities as one force driving economic transformation through dependencies. As such, this chapter explores the extent of experimentation and inertia when actors identify, act upon, and realize innovative opportunities. Innovative opportunities refer to a set of different elements within the process whereby actors identify, act upon, and realize new combinations of resources and market needs—to try to beneﬁt from their future economic potential. The proposed concept draws upon earlier proposed opportunity

12

Introduction concepts: technological, entrepreneurial, and productive, but provides a more comprehensive tool for capturing innovation as it includes elements of value perception, resource mobilization, and value appropriation. In this chapter, innovative opportunities are used as a concept for capturing dependencies between key processes in innovation, as well as understanding the systemic effects present in complex innovation activities characterized by uncertainty. The dependencies are illustrated by case studies of the early stages of development of technology and business platforms in the telecom area, more specifically 3G and i-mode. The empirical observations demonstrate dependencies among different actors which result in ‘systemic’ effects, and these take the form of dependencies across different types of opportunities (technological, entrepreneurial, and productive) as well as dependencies across different elements of these processes (value perception, resource mobilization, and value appropriation). Such interdependencies are argued be one force helping drive ﬂexibility and stability in economic transformation, as drawn out in the conclusions of this chapter. In Chapter 4, ‘The great experiment: public–private partnerships and innovation in design, production, and operation of capital goods in the UK’, by Andrew Davies and Ammon Salter, public–private partnerships (PPPs) are becoming the key mechanism for the design, production, and operation of capital goods in the public sector in the UK and elsewhere. There has been little independent empirical research on the impact of new arrangements on the innovation process in capital goods. Thus, this chapter explores the boundaries of innovating, learning, and experimenting within particular types of organization forms of contracting for capital goods. This type of contracting is becoming increasingly important to the economy, because long-term service contracts between the public and private sectors or PPPs are increasingly common. They have become a much used ‘tool’ for governments looking for opportunities to lower risk and shift responsibility for the production and operation of ﬁxed capital goods to the private sector. Moreover, they represent a Schumpeterian transformation in the capital goods sector that reshapes the division of labour between the public and private sector and alters the potential for innovation in the design, production, and operation of capital goods. The chapter focuses on the central motivations behind the movement towards PPPs and explores the impact of these new contractual arrangements on innovation in the capital goods sector. The greatest impact of the new arrangements appears to have been the emergence of a PPP industry and, along with this, a new division of labour among private providers and public procurers of capital goods. In addition to its traditional role as designer and builder of systems, the private sector has taken on greater downstream responsibility for operation and maintenance of systems previously handled in-house by public sector organizations.

13

Introduction

1.3.2

Chapters in Theme 2: evolution and adaptation of structure

Chapters 5, 6, and 7 in Theme 2 mainly analyse ﬂexibility at the actor and system levels. Chapter 5, ‘Complexity, evolution, and the structure of demand’, by John Foster and Jason Potts, argues that evolutionary economics should be founded upon complex systems theory rather than upon neo-Darwinian analogies concerning natural selection which have been focused especially on supply-side considerations and competition among ﬁrms and technologies. Such analogies are useful to understand the resolution of variety, but they do not help us understand how variety is generated and diffused. Hence, the interpretation of evolution and adaptation of structure presented in this chapter stresses complex systems theory. Generally, an economic system is a complex geometry of connections between rules and elements. Thus, economic evolution involves change in the connective structure of an economic system towards more ordered complexity. Network connections exist between consumers, between producers, and between producers and consumers. When dealt with in an explicit and systematic way, the ﬂexibility and stability of economic systems can be studied in a more coherent way than hitherto. The authors argue that conceptions such as production and consumption functions should be replaced by network representations in which the preferences or, more correctly, the aspirations of consumers are fundamental and, as such, the primary drivers of economic growth. Technological innovation is viewed as a process that is intermediate between these aspirational networks and the organizational networks in which goods and services are produced. Consumer knowledge becomes at least as important as producer knowledge in determining how economic value is generated and it becomes clear that the stability afforded by connective systems of rules is essential for economic ﬂexibility to exist but that too many rules result in inert and structurally unstable states. In contrast, too few rules result in a more stable state, but at a low level of ordered complexity. The chapter explores economic evolution from this perspective using random and scale-free network representations of complex systems. Chapter 6, ‘Self-transformation, self-organization, and evolutionary adaptation in the economic process’, by Stan Metcalfe and Ronnie Ramlogan, focuses on the connection between economic adaptation and economic development and growth. Thus, this chapter also explicitly draws upon evolutionary economic theory and complexity theory to address the speciﬁc issue of adaptation. The authors argue that economic adaptation is inseparable from the growth of knowledge and that market-based dynamics give adaptation a form that is central to change in modern capitalism. Reallocation of resources is the con-

14

Introduction sequence of different beliefs, the imagination that the economic world can be organized differently. This is the primary reason why economies evolve and adapt, in that they are instituted variation, selection, and developmental processes. Understanding at multiple levels is the necessary condition for an economy to operate, which in turn stimulates the growth of new knowledge and thus the emergence of new understandings. This is a system, therefore, in which neither the component parts, which consist of knowledgeable individuals, nor their patterns of intercommunication, the social relations, are given. The point about the complex knowledge system is that it is evolving in parts and connections, as such it is ‘restless’ in that it is always becoming something else in an unpredictable way (Metcalfe 1998). Within this conceptualization of the innovating and transforming economy as restless capitalism, the effects of adaptation to change may be positive or negative. Chapter 7 is ‘Changing boundaries of ﬁrms in the evolution of the computer industry: towards a history-friendly model’, by Franco Malerba, Richard Nelson, Luigi Orsenigo, and Sidney Winter. It analyses the changing boundaries of ﬁrms in terms of vertical integration and disintegration (specialization) in dynamic and uncertain technological and market environments. Like the previous chapters in this theme, this chapter also explicitly draws upon evolutionary economic theory, but here with a focus on modelling industrial dynamics as co-evolutionary processes. In particular, this chapter addresses the question of stability and change in ﬁrms’ decisions to ‘make or buy’ in contexts characterized by periods of technological revolutions punctuating periods of relative technological stability and smooth technical progress. The analysis is centred upon the coupled dynamics of competencies, market structure and on the co-evolution of the upstream and downstream industries. The chapter is inspired by the case of the computer and semiconductor industries and proposes the building blocks of a model in the ‘history-friendly’ style, showing how alternative dynamics of demand and technical change might generate profoundly different patterns of evolution in the two industries. After a brief recount of the long-term evolution of the two industries, as it emerges from the historical and economic literature, the model examines why certain computer producers (e.g. IBM) were vertically integrated into semiconductors early on in the evolution of the industry and why later on they disintegrated, becoming specialized system producers acquiring part or all of their needs for semiconductors from the market, i.e. from specialized producers. The main argument proposed in the chapter concerns the role of co-evolution in the upstream and downstream industries in explaining the changing boundaries of ﬁrms.

15

Introduction

1.3.3

Chapters in Theme 3: innovating and technological transformation

Chapters 8, 9, and 10 in Theme 3 link ﬂexibility at the actor level with stability at the system level. Chapter 8, ‘The effects of technological change on the boundaries of existing ﬁrms’, by Paul L. Robertson and Gianmario Verona, explores the reciprocal relationship between technological innovation and the boundaries of the ﬁrm, and thereby complements Chapter 7 with an interest in industrial dynamics. Innovation and technological change are part of the core of discussing ﬁrm boundaries, not least because the authors argue that innovation is highly unlikely to lead to uniform changes in the relationship between ﬁrms. While in some cases, technological change and increasing transaction costs might lead to convergence (vertical integration) among ﬁrms in certain industries, other types of technological change and decreasing transaction costs (brought on by a diffusion of knowledge) could lead to divergent patterns (vertical differentiation or disintegration), even in the same industries. Far from being random, however, the nature and degree of these changes may be predicted through an examination of the economic and technological contexts in which particular ﬁrms operate. This chapter explores the paths that ﬁrm boundaries might be expected to follow under a variety of circumstances. The basic argument is that, while one might follow Langlois (2003) in believing that there will be a tendency for large, vertically integrated enterprises to become less common in the technological environment that seems to unfold at present, other tendencies will also be afoot, leading to a spectrum of different types of inter- and intraﬁrm relationships. This includes, in some cases, the creation of new giant enterprises of the sort whose evolution in the early twentieth century was described by Chandler (1962, 1977) as the imposition of a ‘visible hand’ to reduce uncertainties in market-based relationships. Chapter 9, ‘Transitions, transformations, and reproduction: dynamics in socio-technical systems’, by Frank W. Geels and Rene´ Kemp, explores innovating and technological transformation in terms of competing socio-technical systems that evolve over long time periods. This chapter examines changes in functional systems from both a theoretical and empirical perspective. It offers a typology of system changes and two brief case studies based on multilevel analysis. The case studies are the hygienic transition from cesspools to integrated sewer systems (1870–1930) and the transformation in waste management (1960–2000) in the Netherlands. Three types of processes through which systems may change are described: reproduction, transformation, and transition. In the case of reproduction, there is no fundamental change at the meso level (the orientation of dominant actors, regime rules, and key technology or knowledge base), which instead is

16

Introduction the case when stabilizing factors dominate. In the case of transitions and transformations in socio-technical systems there are meso-level changes, created either by problems or new opportunities afforded by changes in technology and changes in the overall landscape. Transitions and transformations can only occur when developments at multiple levels amplify each other. This is worked out in a scheme of dynamic interactions between actors, systems rules, and social networks. The novelty of the scheme is that a regime perspective is combined with an actor perspective. Thus, the two empirical cases are used to explore the conceptual framework. The empirical cases illustrate that transitions and transformations are nonpredictable processes with emergent systemic properties, in which niches, surprises, and crises play an important role. Even the outcomes of goaloriented transitions can only be understood through events and crossover effects from various processes at different levels. Whereas diversity is desirable from a lock-in point of view, diversity may also create uncertainty and delay transition and transformation processes. Chapter 10, ‘Analysing ﬂexibility and stability in co-evolutionary processes’, by Magnus Holme´n and Maureen McKelvey, poses the question of how we can analyse conceptually and empirically whether or not certain types of change have occurred, as well as by pointing out research areas to further address how, why, and in what dimensions such transformation has occurred. Our motivation for writing this chapter is simple: to encourage research on the interlocking importance of business, technological, public policy, and organizational change over time. This chapter focuses on considerations related to systematically linking theoretical arguments with empirical material. This chapter ﬁrst addresses how the amount or degree of ‘change’ can be conceptualized, as compared to the amount or degree of ‘not change’. The proposal is that we can do so by distinguishing between the relative amounts of ‘old’ and ‘new’ that is found—within some deﬁned, speciﬁc characteristic of the economic system that is observed empirically. Three concepts are therefore introduced here—namely novelty, destruction, and renewal. The chapter then explains and outlines six points about how ﬂexibility and stability can be understood from a paradigmatic perspective on the innovating and transforming economy. The reason for pursuing this line of (paradigmatic) argumentation is twofold. One is that there are many researchers from different disciplines who are coming into the study of innovation and technological change in relation to industrial dynamics, structural change, and transformation. Thus there is a need to provide a brief introduction that links our concepts into a way of understanding economic transformation. The second is to use these six points, in relation to the discussion of ﬂexibility and stability, in order to sketch out some interesting research questions.

17

Introduction

1.4

Beyond this book

This concluding section presents a few reﬂections to help stimulate readers’ further development of the ideas outlined earlier. The comments in this section are meant to stimulate readers to think about ‘Beyond this book’ while reading subsequent chapters. To begin with, we would like to stress that each of the individual chapters in this book opens up room for new ideas and ways of approaching this topic. Many different concepts, theoretical explanations, and methods to conduct research can be found there, symbolizing the heterogeneity of ideas still waiting to be explored. The reason this book involves researchers from different ﬁelds is not simply because there are many ways of attacking the issues of dynamics. The editors’ view is that for a deeper understanding of tensions and dynamics, we need progress in understanding how and why to connect the incentives and behaviour of the individual actors with system characteristics and system dynamics. Such progress will depend on communication among researchers who are specialized in different areas. Moreover, the three themes—‘Experimenting and inertia’, ‘Evolution and adaptation of structure’, and ‘Innovating and technological transformation’— suggest areas of research in relation to analysing ﬂexibility and stability. The following discussion of research within each theme is done in a fairly abstract way to point to a few of the issues raised by the preceding chapters in this book. Theme 1: Given that economic change fundamentally relies upon endogenous processes, diverse actors are developing new market and technological knowledge, identifying and exploiting opportunities, and revising the organization of business relationships. From this perspective, therefore, economic transformation is likely to be driven by the development of new formalized and fundamental knowledge—as well as the development of knowledge based on organizational learning and experience-based judgements. Innovating and learning are key because rather than assuming there is one ‘optimal’ solution, the shared assumption across the chapters is that the outcome is the result of historical processes. Hence, we should address what happens when actors are experimenting with different organizational forms. Such experimenting may be driven by both ﬁrm strategy and adaptation to environmental conditions, and their choices will affect the trajectory of further economic development. Yet the boundaries of innovating, learning, and experimenting may be affected by systemic conditions and not all actors strive for change, some resist change. Thus, experimenting and inertia may be linked together when understanding what diverse actors are doing during economic transformation. Under certain circumstances, actors will work within the existing organization and knowledge as appropriate and applied to the task at hand. And yet, in

18

Introduction other circumstances, new and different types of business, technological, public policy, and organizational processes will need to be developed in order to solve the business and economic problems at hand. This implies that research should further explore the tensions that will arise between ‘old’ and ‘new’ knowledge and organizations as experimentation, strategies, learning, and institutional context inﬂuence which speciﬁc types of knowledge and organizational forms are tried out in any given case. Theme 2: This theme addresses how certain economic phenomena and inherent properties of the system may be channelling economic transformation through connections in complex systems, adaptation of structure and co-evolutionary processes. The chapters in this book primarily interpret ﬂexibility and stability at the system level, with interactions to the actor level. They thereby address some core issues where more research could be done. For one thing, the chapters in this theme all start from an explicitly evolutionary economics perspective, albeit with different interpretations of what these theories mean. This implies that more research is necessary to further develop this theoretical paradigm, which is core to the idea of an innovating and transforming economy. In relation to ﬂexibility and stability, particularly important issues are about demand as a fuel for the creation and production of goods and services, as well as the role of adaptation and of networks in driving economic transformation. A particularly fruitful avenue may be to try to seriously link our understanding of consumers and users to an understanding of economic transformation. Theme 3: This theme raises questions of ﬂexibility and stability in relation to how and why economic change can be channelled into developing along particular trajectories. The chapters here sketch different ways to proceed on such topics as related to different sets of processes in industrial dynamics, socio-technological systems, and co-evolutionary processes. For example, there are questions related to concepts and explanations for how and why certain ‘transitions’, ‘transformations’, and ‘trajectories’ occur as well as how and why stability may occur. It appears that such processes may be explained either in terms of systemic characteristics and emergent properties, or as arising from interactions and interdependencies of individual actors.4 The research may particularly wish to explore ﬂexibility and stability, in relation to the dependencies between the decisions and actions of actors and systemic conditions. The idea of long-term trajectories could further be explored in relation to ﬁrms but also in relation to the particular role of government and public policies. In short, each individual chapter and each of the three themes highlights insights related to explaining and exploring ﬂexibility and stability in evolutionary processes and complex transformation of the economy. On the one hand, these processes of industrial dynamics, structural change, and transformation are the result of individuals, ﬁrms, and organizations actively

19

Introduction shaping, learning about, predicting under uncertainty, and reacting to their context(s). On the other hand, these broader processes and systems also exhibit their own dynamics, and thereby constrain and enable these actors. This book does not solve this challenge, because instead of trying to provide the ‘correct’ interpretation, the collection of scholarship, which constitutes this book, answers only some questions and opens up a variety of new debates. In our opinion, one of the key issues that requires further exploration is how and why aspects of cognition and learning inﬂuence broader systemic transformation. Hence, one area for future research would be to explore the interactions between individual agents and systems, especially what triggers individual actors to alter their behaviour, whether through competencies, cognition, or learning. This could be used, for example, to explain differential interest in starting businesses and entrepreneurship in different times and places. Speciﬁcally, how do individual actors recognize an innovation, a chance to do something new, an opportunity? According to Loasby (2001) the triggering mechanism for an individual (actor) identifying an opportunity or a problem can be understood in what he refers to as Pound’s principle.5 The principle states that problems (or opportunities) are identiﬁed by the difference between some existing situation and some desired situation. The argument here is that these differences go back to the distinction between a perception of what something could be like or how something could be done and a perception of something factual.6 Hence, perception and cognition is an important aspect of the ability for ﬁrms and agents of public policy to react to environmental changes—but they also act in advance of expected future states. This leads to the development of new sets of questions. For example, under these conditions, what does experimentation and learning actually mean? And how and why are new connections made, or new feedback mechanisms developed to interpret the environment? A related question is: How is cognition economized? The answer to this question helps explain the extent to which knowledge functions as a public good, as well as the division of innovative labour, between different actors and systems (Loasby 1999). Loasby refers to this as ‘bounded cognition’ and puts it into sharp contrast with Simon’s ‘bounded rationality’, which is a rationalistic information processing view. Finally, it would be useful to develop the epistemological foundations of the evolving and innovating economy further, because doing so is necessary to address some of the challenges facing researchers and decision-makers. Differences in epistemological foundations matter when deciding how to ‘do’ social science and also to develop useful knowledge for practitioners. Many of the debates in subsequent chapters address—directly or indirectly—how much (and which parts) of empirical and historical accounts can be used to develop predictive and explanatory models. For example, one aspect of this is how to continue the tradition of confronting conceptual and theoretical approaches

20

Introduction with empirical understanding, known as ‘appreciative theorizing’ (Nelson and Winter 1982). Theoretical constructs do not take precedence over other ways of perceiving and intermediating with the world, but neither are empirical cases and illustrations considered the only way to approach ‘truth’ about the economy. Hence, we must debate how science is ‘done’, especially the epistemological foundations of what we ‘know’ and how we know it in social sciences. Indeed, all of social science—and most of medicine, engineering, and large parts of natural science—face the same problem. That is, how does one go about sorting out the speciﬁc, context-dependent outcomes from more generalizable laws and predictions of behaviour? Our community is faced with particular challenges in explaining ﬂexibility and stability in economic transformation, given the insight that history matters in such a fundamental sense that no two states of the system are the same, nor may the dynamics be explained through simple mechanisms. On the one hand, this book stresses that very history-speciﬁc and unique aspects—such as phenomena, events, people, and organizations—are inherently part of what may, on the other hand, be the result of ‘explanatory variables’ conceptualized in terms of systematic outcomes and trends. This implies that theoretical explanations are not simply ‘proving’ hypotheses to make predictions but they are instead ‘exploring’ how and why serendipity, human choice, and speciﬁc conditions combine with more ‘structural’ causal factors. In summary, we believe that debating issues about Flexibility and Stability in the Innovating Economy lifts up many themes that could be of relevance to readers. Given that this issue represents one of the front lines of theory development in the ﬁeld, there are many points of contention among similarly minded researchers and hence, many debates are ongoing in subsequent chapters. For this reason, this book provides an interdisciplinary analysis, with a collection of scholarship that represents diverse perspectives on the core issues. In doing so, this book is primarily conceptual and theoretical, with relevant empirical material to illustrate the arguments and to make the case. In different ways subsequent chapters argue that understanding these fundamental issues are critical for academic debate, public policy, and ﬁrm strategies—as well as for stimulating further research.

Notes 1. See Nelson 1996; Metcalfe 1998. 2. There are many references for each of the classic writers, but some of the important references include Schumpeter 1947; Smith [1776] (1991); Menger [1871] (1976); Marx (1890); Marshall 1890; Weber 1961; Veblen 1898; and Young 1928. 3. Such as Hayek (1937, 1945), Lachmann ([1956] 1978).

21

Introduction 4. In this case, the notion of ‘systemic characteristics’ could refer to understanding the linkages between system changes (i.e. changes of the system as a whole) or to the ongoing dynamics (or lack thereof) between actors and components within that system, or to the interaction among the two previous ideas. 5. This was ﬁrst outlined by Pounds (1969) in an article ‘ . . . noticed by hardly anyone’ (Loasby 1976: 96). 6. This raises the question of the source of differing perceptions, and Loasby (1976) lists four categories from which such differences with the perception of the present situation can be found: a) historical, b) external to the ﬁrm, c) planning which should have led to an anticipated or intended outcome, and d) imaginative, a notion of what might be.

References Chandler, A. D. (1962). Strategy and Structure: Chapters in the History of Industrial Enterprise. Cambridge, MA: MIT Press. Chandler, A. D., Jr. (1977). The Visible Hand: The Managerial Revolution in American Business. Cambridge, MA: The Belknap Press. Constant, E. W. II (2002). ‘Why Evolution is a Theory About Stability: Constraint, Causation, and Ecology in Technological Change’, Research Policy, 31(8–9): 1241–56. Freeman, C. and Perez, C. (1988). ‘Structural Crises of Adjustment, Business Cycles and Investment Behaviour’, in Dosi et al. (eds.), Technical Change and Economic Theory. London: Pinter. pp. 38–66. Hayek, F. A. (1937). Economics and Knowledge. Economica, New Series, 4: 33–54. Hayek, F. A. (1945). ‘The Use of Knowledge in Society’, American Economic Review, 35(4): 519–30. Hanusch, H. (ed.) (1999). The Legacy of Joseph A. Schumpeter. Two-volume reference collection. Cheltenham, UK: Edward Elgar. Kuznets, S. (1954). Economic Change. London: Heinemann. Kuznets, S. (1959). Six Lectures on Economic Growth. Glencoe, IL: Free Press. Lachmann, L. M. [1956] (1978). Capital and Its Structure, 2nd edn. Kansas City, KS: Sheed Andrews and McMeel. Langlois, R. N. (2003). ‘The Vanishing Hand: The Changing Dynamics of Industrial Capitalism’, Industrial and Corporate Change, 12(2): 351–85. Loasby, B. J. (1976). Choice, Complexity and Ignorance: An Enquiry into Economic Theory and the Practice of Decision Making. London: Cambridge University Press. —— (1999). Knowledge, Institutions and Evolution in Economics. London: Routledge. —— (2001). ‘Cognition, Imagination and Institutions in Demand Creation’, Journal of Evolutionary Economics, 11: 7–21. Marshall, A. (1890). Principles of Economics: An Introductory Text. London: Macmillan. Marx, K. (1890). Das Kapital. Menger, C. [1871] (1976). Principles of Economics. Translated by J. Dingwall and B. F. Hoselitz. New York: New York University Press. Metcalfe, J. S. (1998). Evolutionary Economics and Creative Destruction. London: Routledge.

22

Introduction Nelson, R. (1996). The Sources of Economic Growth. Cambridge, MA: Harvard University Press. Nelson, R. R. and Winter, S. G. (1982). An Evolutionary Theory of Economic Change. Cambridge, MA: The Belknap Press. Pounds, W. F. (1969). ‘The Process of Problem Finding’, Industrial Management Review, 11: 1–19. Schumpeter, J. A. (1947). Capitalism, Socialism and Democracy, 2nd edn. New York: Harper and Brothers. Smith, A. [1776] (1991). The Wealth of Nations. Amherst, NY: Prometheus Books. (Original title ‘An Inquiry into the Nature and Causes of the Wealth of Nations’.) Veblen, T. (1898). ‘Why is Economics Not an Evolutionary Science?’, The Quarterly Journal of Economics, 12. Weber, M. (1961). General Economic History. New York: First Collier Books. Young, A. (1928). ‘Increasing Returns and Economic Progress’, The Economic Journal, 38(152):527–42 Williams, R. (2003). Retooling: A Historian Confronts Technological Change. Cambridge, MA: MIT Press.

23

This page intentionally left blank

THEME 1 EXPERIMENTING AND INERTIA

This page intentionally left blank

2 The New Craft Skills of Engineering: The Impact of Innovation Technology on Engineering Practice Mark Dodgson, David M. Gann, and Ammon Salter

2.1 Introduction This chapter explores the impact of the use of innovation technologies1 on engineering practices. It seeks to contribute a better understanding of how the practices of experimentation are changing within engineering due to the use of new technology. There is a wide body of literature focusing on the ‘distinctiveness’ of engineering problem-solving, and the nature of engineering knowledge is itself coming under increasing examination. Williams (2003) argues that engineering as a ﬁeld is witnessing two major patterns of change, one involving a deepening of knowledge in specialist domains, and the other a broadening set of interests and pursuits outside normal engineering practice. She calls these two trends the ‘expansive disintegration’ of engineering. One source of the disintegration of engineering practice is the use of new technologies, including simulation, modelling, and rapid prototyping, which enable new forms of problem-solving and reshape the division of innovative labour. These technologies are helping to redeﬁne the role of engineers in the innovation process, creating new practices that span different networks and knowledge domains within and across organizations. They also help reshape the knowledge base of engineers, requiring them to collaborate with more nonengineers and to deal with social, political, and economic challenges as they go about developing new technical solutions to problems. In this sense, the use of new tools may involve a signiﬁcant change in the way experimentations are conducted in the economy. They could create a more ‘democratic’ and ‘open’ method for ﬁnding new solutions, products, and processes, by integrating a wider range of organizations, including suppliers and users, in the process of experimentation (Chesbrough 2003; Von Hippel 2005).

27

New Craft Skills of Engineering The chapter is based on the analysis of the experiences of two leading engineering companies, Arup and Ricardo Engineering. In particular, the chapter explores how the use of the new technologies in the design engineering process has helped reshape practices inside these organizations. The chapter suggests that the use of these new technologies is creating opportunities for engineers to create new value by allowing users and clients to play with different alternatives. They also enable engineers to work more closely with partners outside their organization. This process of transformation also creates new tensions between experimentation and change on the one hand, and the need to sustain and enhance the core knowledge and problem-solving routines of engineering on the other. These organizations need to balance ﬂexibility in work practices with stability in the core knowledge of technical decision-making. It is the combination of stability in the knowledge base and ﬂexibility in practices that is the core of the new creative challenges facing engineering. The chapter is organized into three sections. Section 2.2 examines some of the literature on engineering knowledge and explores the impact of the use of new technologies on engineering practice. Section 2.3 reviews experiences of the two different engineering organizations. Section 2.4 discusses the emergence of new craft skills in engineering that require a degree of continuity with the past in combination with a willingness to change to take advantage of the opportunities created by the new technologies.

2.2

Changing nature of engineering practice and knowledge

Engineering design is a central part of the innovation process and it plays a key role in shaping the way new products and processes are developed. It involves extensive experimentation to ﬁnd new technical solutions that can meet functional needs of potential users and be readily and cost-effectively manufactured (Thomke 2003). Many of these solutions involve ﬁnding new combinations of existing components, materials, or technology. To ﬁnd these new combinations, engineers often need to invest considerable time, effort, and resources in learning about different technologies and exploring how these can be integrated in new and productive ways (Hargadon and Sutton 1997; Brusoni et al. 2001). The experimentation process itself often generates a wide variety of potential solutions. Many of these potential solutions may meet the functional needs of the user, but each solution may require a very different development path before it can be successfully implemented. Indeed, engineers often have to make difﬁcult choices among a number of potentially appealing solutions, weighing different objectives and goals, and making trade-offs between different aspects of performance and cost. In making these choices, engineers

28

New Craft Skills of Engineering often work with potential users, suppliers, other technical functions such as production, operations, and other organizations (Von Hippel 2001, 2005). In this way, engineers are continuously engaged in a process of experimentation and selection. Through such a process, they are continuously balancing different options and experimenting with a number of potential futures. In the search for these new solutions, researchers have long highlighted the distinctive properties of engineering problem-solving (Constant 1980; Pavitt 1987; Vincenti 1990; Simon 1996; Perlow 1999). Simon, for example, distinguishes between engineering, which he says is concerned with ‘synthesis’ (often synthesis of artiﬁcial objects), and science, which is concerned with ‘analysis’ (Simon 1996: 4). Whereas scientiﬁc problem-solving primarily concerns the development of theories and predictions about the physical world, technological problem-solving normally starts out with a functional requirement and seeks to use scientiﬁc principles and technological components and systems to help achieve desired goals. In so doing, engineers work differently from scientists. They work closely with a variety of actors in the innovation process, such as with the eventual users of innovations, to help understand the function requirements and goals of the task they are undertaking. As a consequence of this approach, engineering knowledge is often described as practical, practice-based, sometimes tacit, output-oriented, and recursive (Pavitt 1987, 1998; Nightingale 1998). Engineering knowledge draws from scientiﬁc principles, but relies heavily on ‘rules of thumb’, ‘informed guesses’, ‘routines’, and ‘norms’ of problemsolving that are built-up through engineering education and experience with real world problems (Vincenti 1990). Engineering knowledge grows incrementally and ‘traditional’ knowledge, such as the Second Law of Thermodynamics, remains an essential element in problem-solving routines and methods. Indeed, engineering knowledge relies on a set of ‘heuristics’ about problemsolving that are framed by the technological trajectories the engineers are working within (Dosi 1982, 1988). They are said to be less interested in developing theory or general knowledge than scientists and more interested in meeting functional needs through the creation of new technologies (Dasgupta and David 1994). Constant (2000) refers to development of engineering knowledge as the development of recursive practice, slow and steady accretion of knowledge about how things work and how they ﬁt together. Engineers learn to use practical, iterative approaches in their education, which often combines a problem-oriented approach with theory about physical properties. Engineering itself remains a highly regulated activity, governed by professional organizations, practices, codes, and regulations (Becher 1999). This is not to suggest that all engineering involves routine or humdrum activities or the simple application of problem-solving tools to new problems. As Vincenti (1990) highlighted, engineering work varies from procedures in ‘normal design’, to experimentation, and the development of ‘radical design’ solutions.

29

New Craft Skills of Engineering To solve problems, engineers rely on a social and informal culture, networks of connections, and patterns of communication (Allen 1977). Project work is sometimes carried out in diverse, multidisciplinary teams that span a number of different organizations. Team members communicate by sketching and telling stories about related engineering work in order to develop solutions (Ferguson 1992; Henderson 1999; Salter and Gann 2003). Studies have shown that engineers need time for face-to-face interaction and time to work alone (Perlow 1999). This balance between the need for interaction and solitary work creates a deep and structural tension in engineering that is often reﬂected in heavy workloads, late nights, and cycles of heroics and overwork (Perlow 1999). Simon (1996) argued that for engineers to solve complex problems they need to break them down into modules or small parts. By decomposing tasks engineers are able to focus their work on an area of manageable complexity and apply their techniques and capabilities to a narrow range of problems while leaving other problems to other groups. In this way, engineers are able to divide tasks across many different subteams and units and increase both the efﬁciency and effectiveness of problem-solving (Pahl and Beitz 1996). Engineers also need deep knowledge of different technologies and components and how they ﬁt together. Innovation often involves new combinations of existing technologies applied to a new problem (Hargadon 2003). Some engineers work as systems integrators, choosing components and technologies, specifying the interfaces between different systems, and combining new components with different vintages of technology (Baldwin and Clark 1997, 2000; Brusoni et al. 2001). Engineering design and problem-solving involves more than simply choosing from a variety of technological options. Technological choices are shaped by the social and economic context in which they are made such that engineering principles usually form only part of the solution. They have to be integrated within a set of ideas from a range of diverse disciplines including economics, management, sociology, and political science. Engineering has, furthermore, become so ubiquitous that engineers are involved in producing and maintaining artefacts in almost every facet of society and the economy. As Williams suggests, on the one hand, engineering is becoming more specialized with many new subdisciplines emerging (such as ﬁre, acoustics, and lighting engineering), and on the other, it is becoming more generalized (Williams 2003). Simulation and modelling have always played a central role in engineering problem-solving. For engineers, models provide a mechanism for learning about artefacts before and after they have been built. Models enable engineers to examine different options and weigh the choices of structural elements, materials, and components against one another. It is commonly accepted that

30

New Craft Skills of Engineering the adoption of technology has changed the way engineers work (Schrage 2000; D’Adderio 2001, 2004; Thomke 2003). More engineering is now done on computer rather than on paper, although most engineers are still heavy users of paper. Physical prototypes are expensive, time consuming to create, and often unreliable, whereas many digital tools are relatively cheap and easily available. They can allow the engineer to do more than they could in the past. They provide the basis for digital models that assist in abstracting physical phenomena, allowing engineers to experiment, simulate, and play with different options. This is leading to a new culture of prototyping in which traditional practices of design are being opened to more concurrent diagnostic enquiry (Schrage 2000; D’Adderio 2004). Added value in design and development processes, however, comes from creative and schematic work where designers produce radical solutions that are beyond the calculations that are embedded in routinized software programmes. They do this through ‘conversations’ and the use of ‘visual cues’, interacting with one another and their clients and suppliers (Schrage 2000; Whyte 2002). This schematic work involves seeing the interfaces between components and integrating inputs from different specialists. One of the most interesting developments appears to be occurring in the manipulation of digital symbols and models by highly skilled creative craftspeople often working in small teams: what McCullogh (1996) calls ‘abstracting craft’. Rather than replacing traditional design skills, the new digital technologies can complement them in novel and evolving ways. Despite the evidence of the use of new technologies in engineering problemsolving, there have been relatively few studies on the impact of these technologies on engineering organizations. Given the widespread use of new innovation technologies, it is necessary to return to Vincenti’s fundamental questions (1990) of ‘what engineers do, what they know and how they know it’, and explore how the answers to these questions are changing. Our focus is on the role played by a speciﬁc category of technology— innovation technology (IvT)—on engineering practices. IvT includes a wide range of new digital technologies that are being applied to the innovation process. It includes modelling and simulation, virtual reality, data mining, artiﬁcial intelligence, and rapid prototyping. It also encompasses a number of new technologies, like high throughput chemistry and high throughput screening applied to drug discovery. Many of these technologies are becoming ubiquitous. It is possible to think of these tools as the new ‘capital goods’ of the innovation process. There is evidence of extensive use in a wide range of sectors, from pharmaceuticals to the mining and construction industries, and across many different types of organization. They are a new category of technology that function alongside information and communications technology and operations and manufacturing technology (Dodgson et al. 2005).

31

New Craft Skills of Engineering The use of innovation technologies has increased greatly in recent years among leading innovative organizations (Schrage 2000; Thomke 2001, 2003; Dodgson et al. 2002; Tuomi 2002; D’Adderio 2004). Its adoption is forcing many organizations to reconsider the ways they conduct and manage innovative activities. Relatively little has been written about the impact of these tools on the innovation process. Most of the discussion in the literature on innovation has focused on the impact of use of new electronic tools on the nature of knowledge in a limited number of technological and organizational environments (see Steinmueller 2000). In particular, it has focused on the potential for the ‘codiﬁcation’ of tacit knowledge (David and Foray 1995; Cowan et al. 2000; Johnson et al. 2002). Instead, on this question, we explore how these new tools are being used in practice inside engineering organizations. Our past research suggests that the impact of innovation technology in design, research, and development enhances both the ‘code’ and the ‘craft’ required to innovate (Dodgson et al. 2004). In conducting this research, we are seeking to better understand what engineers ‘know’ and ‘do’ in the context of the increasing use of IvT. We are also keen to understand how the use of the new tools lead to new work practices and reshapes the division of labour inside engineering organizations.

2.3

Case studies2

In this section, we explore the experiences of the major engineering companies, Arup and Ricardo Engineering. Both favoured organizations have a long tradition of engineering excellence and have diversiﬁed into a variety of new businesses based on their core engineering skills. In the case of Arup, we review the ﬁrm’s approach to innovation and engineering, and examine the experiences of a new specialist group within the ﬁrm: ArupFire. In the case of Ricardo, we focus on the experiences of Ricardo Software. By exploring the experiences of these two different organizations, we are able to gain some insights into how traditional engineering skills are being reshaped by the use of IvT and how the use of these new technologies is creating new business diversiﬁcation opportunities. Both ArupFire and Ricardo Software represent examples of the ‘expansive disintegration’ of engineering that requires engineers to work in new ways across traditional organizational and professional demarcations. They require engineers to balance the need to retain and enhance their core engineering skills as well as to develop new ways of working and problem-solving. Both organizations are providers of ‘engineering services’ to a wide range of different industries. However, Arup mostly works with the construction industry, whereas Ricardo’s main clients are in the automotive sector. The engineering service sector itself is extremely diverse, containing a range of specialist

32

New Craft Skills of Engineering and general practices. Normally these organizations work on a series of projects for a diverse range of clients, providing specialist consultancy, advice, and input to help other ﬁrms develop new products, processes, and services. As such, they are often key gatekeepers of new technological knowledge and broker this knowledge across different projects and organizations, passing lessons from one project or client to another (Hargadon 2003). For example, in the construction sector, an organization such as Arup works with architects, construction management organizations, contractors, client organizations, developers, government departments, and private organizations. Each one of these actors in the industry may purchase services from Arup and in different projects a past client will become a future project partner. In contrast, Ricardo has several long-term contracts with a number of key clients, but it also works closely with a range of different organizations in the automotive sector, providing specialist inputs to ongoing projects inside these external organizations. The contrast between Arup and Ricardo is drawn to illustrate and contrast the experiences of two different engineering organizations. Our goal is to gain a better understanding of how a particular context of the organization shapes its problem-solving practices and the evolution of engineering knowledge. Both organizations were selected as leaders in their respective industries and they should be seen as exemplars of practice rather than representative of a larger population of organizations in the respective sectors. The goal of the case studies is to better understand the impact of the new tools of engineering problem-solving in the context of evolving business models.

2.3.1

Arup

Arup was founded in London in 1946 by Sir Ove Arup. It provides a range of design, engineering, and associated services and currently employs over 6500 members of staff in 71 ofﬁces in 50 countries. Arup is recognized for its concentration of technical and design knowledge. It was involved in some of the greatest building projects of the twentieth century, including the Sydney Opera House and the Pompidou Centre. In working on these and current projects, it has confronted many of the major structural and civil engineering challenges through developing highly specialized knowledge of mechanical, electrical, and electronic systems, earthquakes, ﬁre and smoke, acoustics, and environmental engineering. Arup works on several thousand projects simultaneously, providing specialist advice to a diverse client base. However, the growth in the ﬁrm has been almost entirely self-generated. The company started as a structural engineering ﬁrm and as it won new projects its capabilities expanded into a wide number of different areas. As the former Chairman of Arup, Bob Emmerson, states: ‘gifted people take us in unexpected directions’. It now contains over 50

33

New Craft Skills of Engineering specialist groups, spanning a range from environmental consultancy to acoustics. Arup sees its key advantage over competitors as the ability to combine a wide variety of specialist skills on projects. Reﬂecting the view of its founder, it aims to become a ‘total problem-solver,’ weaving together its diverse skills. Many of its competitors are small, specialized suppliers of design services with fewer competencies across a narrower range of ﬁelds. Another advantage claimed by Arup is its ability to recruit talented graduates. Within the ﬁrm it is believed that students are attracted to Arup because of its participation in complex and demanding projects and its ability to provide rich work experiences and training with opportunities for diversiﬁcation into new businesses. Arup employs 350 students annually. In some respects, Arup has found itself in a virtuous project-based cycle: the ﬁrm wins high proﬁle projects because of its reputation for problem-solving and highly skilled engineers are attracted to Arup because of its ability to win complex, exciting projects. As part of this virtuous project-based circle, Arup ‘gets problems that others don’t get’ suggested a senior manager in the ﬁrm. Arup has gained a reputation for ‘delivering difﬁcult projects’, and ‘creativity in problem-solving’, and for working with leaders in their ﬁelds. In part, the strength of Arup’s reputation is based on close, personal relationships among architects, clients, and Arup staff. These relationships have been built from working on past projects and also because of the high levels of recognition of Arup’s abilities. There are numerous problems for managing innovation caused by the way work is organized in the construction industry. Specialist independent ﬁrms join together in a team that usually disbands at the end of the project. The project-based nature of activities means organizations working in the sector often struggle to learn from project to project. Projects are often one-off and task oriented. Learning in one project is rarely fed back to the other areas of the organization as project teams operate semi-autonomously and outside the boundaries of the ﬁrm (see Gann and Salter 1998). Furthermore, design activities are often undertaken under severe time constraints and this lack of time acts as a barrier to innovation (Perlow 1999).

2.3.1.1

ARUP’S DIVERSIFICATION STRATEGY

As mentioned previously, the growth of Arup has been accompanied by an increasing number of specialist groups within the ﬁrm. These new groups developed inside the body of an existing team and as project leaders in these groups saw new market opportunities to develop a specialist service offering, they spun out a new team. Central management recognizes the de facto independence of these teams. Senior managers in Arup have adopted a ‘let a thousand ﬂowers bloom’ strategy to the management of these groups, and the

34

New Craft Skills of Engineering company is decentralized. They feel that attempts to impose central control on these groups might weaken their development and instead they argue that they should be left to get on with developing their own markets and skills. The management style adopted is similar to that found in many professional service ﬁrms employing highly creative people (McKenna and Maister 2002). The emergence of a new capability in Arup has often resulted from experience of working on particular projects where specialist and unique skills and technologies have been developed. These capabilities are honed and transferred to other projects as the market for them expands. IvT lies at the heart of many of the new specialist groups in Arup and much of its specialized simulation software has been developed in-house, or in close collaboration with software ﬁrms. In Section 2.3.1.2, we explore the experiences of one of these specialist groups and their use of simulation tools.

2.3.1.2

ARUPFIRE

ArupFire is a group of over 200 engineers with ofﬁces in 11 different locations in ﬁve countries. It specializes in ﬁre engineering services, including ﬁre modelling, escape and evacuation, ﬁre detection and alarm, ﬁre strategies, and smoke modelling. It was founded in 1998 with a staff of four specialist engineers in London. Originally, ﬁre engineering was located in Arup’s R&D department but, with the emergence of new market opportunities and in accordance with the company’s diversiﬁcation strategy, it was possible to set up a separate specialist division. The creation of ArupFire was led by Peter Bressington who developed a plan for an expansion of the ﬁre group. ArupFire expanded quickly. By 2000, there were around 20 engineers working in London, while ofﬁces were soon opened in Manchester, New York, Hong Kong, and several cities in Australia. ArupFire works closely with other parts of Arup, with close to 60 per cent of the work performed for projects within the company itself. The ratio of external and internal work undertaken by ArupFire is shaped by its own view of its role inside the ﬁrm. Peter Bressington believes that by bundling ﬁre engineering into other projects, Arup is able to increase the value of its service offering. It allows other areas of Arup engineering greater possibilities to ﬁnd more innovative solutions because they are able to draw on the expertise of Arup ﬁre engineers as their designs are emerging. The growth of ﬁre engineering in Arup is intimately connected to changes in the way ﬁre is regulated in the UK and elsewhere. In the UK in the late 1990s, there was a shift away from prescriptive code-based regulation towards performance-based regulation. In the old model, there was only limited opportunity for ﬁre engineering to develop as a service. Fire engineering was the responsibility of architects and designs were organized around the

35

New Craft Skills of Engineering requirements of the statutory codes governing egress and entry into the building. With the shift to performance-based regulation, however, it became possible to work outside the codes, allowing much more ﬂexibility in the designs available to architects, engineers, and their clients. Under the performance-based regulations, if the designers could prove that a design provided a high level of ﬁre safety then the design would be approved by local ﬁre control agencies. As Peter Bressington states, ‘We can’t make a prescriptive code ﬂame-proof—it’s not possible to design these extreme events out, so instead we look at speciﬁc scenarios. A better understanding of building performance, such as focusing on how people would be able to escape from lifts in an emergency, will have more bearing on building safety than simply changing codes.’ Essential to the process of gaining approval for a new building is convincing ﬁre prevention and control staff that the design would be safe not only for the occupiers of the building but also for emergency services personnel. Gaining approval for a design involves meetings between the designers and local planning and ﬁre authorities. To demonstrate the safety of the design to their clients and regulatory ofﬁcials, ArupFire engineers rely on multiple simulation and modelling tools. Bressington states that it is ‘using modelling as a pre-emptive design tool, rather than a way of analysing what happens after a ﬁre has already happened’. Using the simulation tools ‘makes the scenario more realistic and is a much clearer way of seeing and analysing ﬂow and the positioning of exits’.

2.3.1.3

THE ELEVATE MODEL

Much of ArupFire’s work focuses on ﬁre hazards in tall buildings with the goal of getting as many people out of a building on ﬁre as fast as possible. Egress in a tall building is extremely difﬁcult: stairwells can become jammed with both people leaving the building and emergency staff attempting to gain entry to the high ﬂoors of a ﬁre. Travelling downstairs from high levels can take a long time and be dangerous, and disabled people might be unable to climb down 40 ﬂights of stairs in an emergency. People at high levels often need to queue to gain access to the stairs. ArupFire engineers have come up with a radical solution to the problem of egress in tall buildings: they use elevators. By pressurizing the elevator shafts and blowing the smoke and ﬁre out of them, elevators can provide a means of escape. A key part of the development of the elevator evacuation strategy is the use of simulation models of the movement of people in tall buildings. One of these models is called Elevate. It shows the number of people on each ﬂoor and how they move down and out of the building over time. The model also shows where these people should assemble before evacuation. This simulation is essential to overcome the long-ingrained lessons of ‘in case of ﬁre, do not

36

New Craft Skills of Engineering use the elevator’ of building users, and of the spectacular failures of lifts to evacuate people successfully in some notorious ﬁres. After September 11, 2002, many of the residents of London’s tall buildings were concerned about safety and evacuation. ArupFire was approached by the facility managers of the HSBC Canary Wharf tower to examine its ﬁre strategy and to determine whether it was appropriate. Arup ﬁre engineers had been deeply involved in the development of the building and had developed its ﬁre strategy, and the client was now seeking assurance that their building was safe. Using the Elevate model, Arup engineers investigated the HSBC tower. They used their model to think about and play with different possibilities for the building and found that if people followed the traditional strategy of walking out of the building via the stairs it would take 22 minutes to remove everyone from the building. In contrast, if elevators were used, everyone could be evacuated from the building in 12 minutes. The added advantage of the elevator strategy is that it enables disabled people to escape from buildings at the same time, and makes employment of disabled people in high buildings less risky for the individual and employer. The use of elevators in ﬁre strategy has had a signiﬁcant impact on the design of new buildings. On the higher ﬂoors, workers are advised to congregate in the lobby in front of the elevators, so the ﬂoor plan of the buildings must accommodate a large number of people. The elevators must also arrive at the right ﬂoors for the evacuation from the building. The Elevate model allows Arup ﬁre engineers to play with different design options, exploiting relationships between lobby and stairwell size, speed and number of elevators, and the length of time people will wait in the lobby for elevators to evacuate them. Each of these possibilities is developed by engineers with experience of extreme events and deep knowledge of the materials and behaviour of people in extreme situations. The elevator strategy for ﬁre evacuation from tall buildings is not simply about changing technology: it requires considerable retraining for staff working in buildings. People need to be taught that elevators can be key to survival in case of ﬁre. It also required the ArupFire engineers to learn about the behaviour of individuals and groups during extreme events. The simulation and modelling of evacuation requires an in-depth understanding of the psychology of fear. It also requires knowledge of how the building is going to be used, i.e. what types of people will be inside it at the time of an extreme event. Fire engineers create a variety of different scenarios and situations and probe the implications of these scenarios for the safety of the people in the building. By using models, they visualize their scenarios and analyse the behaviour of people as they respond to particular episodes. To win support for these new designs, the engineers need to work with regulators and ﬁre control authorities. They must also convince their clients and users of the building that the structure will be safe in the case of an extreme event. The models also allow ﬁre

37

New Craft Skills of Engineering control authorities to assess the feasibility of the ﬁre strategy by visualizing the extreme events and exploring different evacuation scenarios. Undertaking such an engineering project often requires deep engineering knowledge about the type of materials used in the building, the movement of smoke and ﬁre, the behaviour of people in extreme situations, and the impact of the design of the building on smoke, ﬁre, and egress. It requires them to work closely with a range of engineering disciplines, including mechanical, electrical, structural, acoustic, and civil engineers. They also need to work closely with architects, clients, potential users, regulatory ofﬁcials, and ﬁre control authorities. Working in this environment also requires considerable interorganizational collaboration and providing specialist inputs into a variety of different project teams. The use of simulation tools helps support these networks of collaborations by enabling different communities to see different design options and to explore different scenarios with the ﬁre engineers.

2.3.2

Ricardo Engineering

Ricardo is an engineering technology company conducting research, design, and development for major automobile manufacturers. Based in the UK, it has expanded into the USA, Czech Republic, and Germany. With 1800 employees, it grew its sales from £91 million in 1999 to £137 million in 2003. It has highly specialized expertise in vehicle, engine, and transmission engineering and works closely with product development departments in leading automotive manufacturers. It also provides testing and calibration services and has developed a range of simulation tools in its own software division. 2.3.2.1

SOFTWARE FOR ENGINE DESIGN

The software engineering group in Ricardo grew in the late 1980s out of an inhouse requirement for specialized software analysis tools. There were no third party software products available at the time and it was necessary for Ricardo to develop its own analysis software to support engine design projects. The software evolved and enabled fundamental understanding of physical processes in order to make informed design changes. Ricardo’s customers demanded to use the software and the company repackaged it for them as a product. As this business grew, it acquired software capabilities from outside sources. In 1995, the software group was established in its own right as Ricardo Software, designing, developing, selling, and supporting its own software. Its products are used by a wide range of industries, including motor sport, automotive, motorcycle, truck, agriculture, locomotive, marine, and power generation. Most of the customers are large automobile manufacturers and their suppliers. There are two main simulation software areas: ﬂuid systems and mechanical systems.3

38

New Craft Skills of Engineering One of the drivers of the need for simulations was the increased complexity of the product to be designed. In the early 1980s, engines were largely mechanical and there were only a few elements that controlled their performance. Throughout the 1980s and 1990s, however, interactions between the engine and the rest of the vehicle increased signiﬁcantly, particularly through new electronic control systems. As the market for Ricardo software expanded, the ﬁrm became more dependent on external users for feedback and ideas for future developments to manage this new complexity. In 2003, Ricardo Software employed 55 staff and had three technical centres in Chicago, Shoreham-by-Sea, and Prague. Ricardo is developing products that integrate different software applications. Various simulation tools will allow designers and analysts to see how different systems within their vehicle interact with one another and to manage their complex integration, allowing designers to see how decisions in one area affect other parts of a system. This is a considerable technological challenge as in Ricardo’s business different parts of the system may be designed by organizations in various locations around the world; for example, the engine might be designed in country X, the transmission in country Y, and the vehicle in country Z. In addition to the geographic barriers, each of the different partners involved in the design process may be using different simulation software adding to the problems of systems integration and limiting opportunities for working collaboratively across organizational boundaries. Within the industry it is observed, furthermore, that as automakers are increasingly collaborating on new engine projects, this adds to the challenges of integration. Mistakes in transmission of data and information from one partner to another are common in engineering projects. Each organization and country has its own engineering approach and culture. Ideas about how much and what type of information should be shared with partners often differ from partner to partner. The potential for miscommunication rises considerably as the number of partners increases. The idea of this new generation of integration software for collaborative engineering is to help mitigate these types of dangers. Software for collaborative engineering requires the adoption of open systems. This allows organizations to ‘mix and match’ different software tools, including those developed internally. In-house software tools remain vitally important for many organizations as they usually contain methods, knowledge, and experience that have evolved over many years and are strongly embedded within the ﬁrm. New integration tools, then, must allow organizations to exchange information effortlessly between these different software environments. Many of Ricardo’s simulation tools have been built up slowly over several years and emerge out of internal research projects or efforts to satisfy external consultancy requirements. Lead users have occasionally asked the ﬁrm to

39

New Craft Skills of Engineering develop software for a particular task. In the case of most of the current simulation tools, their capabilities have been developed well beyond the original intention. Richard Johns, head of Ricardo Software, suggests that if the architecture of the software is well thought out then simulation tools can be highly ﬂexible, malleable, and robust, allowing future developers to include a range of new features and design elements. An example of robustness of design can be seen in the development of WAVE, its gas dynamics simulation programme. Originally designed as a simulation for engine performance, it has been extended into many areas such as the prediction of noise and integration with control, vehicle, and mechanical systems. Few of the initial designers of these simulations could have predicted the eventual uses for their programmes when they were initially written. But because of their ﬂexibility and robustness, they have made such a transformation possible. The movement to IvT has allowed designers and analysts to change the way they solve problems in Ricardo. With the earliest generation of simulation tools, individual components and elements would need to be analysed on a step-by-step basis, building up understanding of one system with relatively simple, often inexact, connections to other systems with which they would interact. Each model would run independently from the others. This was a slow and laborious process involving considerable effort in translating the impact of outputs from one part of the system into inputs for the next part of the system. According to Richard Johns, it made the process of managing system interactions a time consuming, error-prone, and approximate process requiring skilled operators and great care. The new generation of simulation tools helps resolve this problem by allowing designers and analysts to connect interacting processes and to co-simulate across different simulation tools. Simulation tools are used across all stages of engine design. There is no point, however, in conducting large-scale detailed simulation of the system until the design begins to take shape. Instead, the designers start by focusing on key features of the engine, such as the bore, stroke, number of cylinders, and their conﬁguration. In doing so, they use a range of conceptual tools and draw on past knowledge and experience. During this concept design phase the conﬁguration is established with basic geometric information about the engine and how different parts of the design relate to one another. At this point, the designers conduct ‘low cost’ and ‘fast concept’ analysis, exploring how the system might function and how its elements might interact. Elements of the design become more and more established over time, and more detailed simulation allows for reﬁnement of the concept design in the course of the design process. Designers in the early stages think about and explore different options and assess the performance of each alternative compared to design targets, then in later stages seek to optimize the design. The use of the tools is also reducing the time and effort expended in physical prototyping and testing. Once the design moves from digital to physical

40

New Craft Skills of Engineering testing the costs of changing the design increase exponentially. Ricardo would like to have conﬁdence that the design will satisfy functional requirements before moving onto physical prototyping and testing. This requires designers to have more and more faith in the simulation tools and their ability to predict the actual performance of the system. The use of the simulation tools alters both the speed and cost of the design process for engine design. Building physical prototypes and testing the design of a combustion system, such as the use of different pistons or cylinder heads, is both time consuming and expensive. Analysing them on a computer is relatively inexpensive with two advantages accruing, namely the tangible cost savings in product development and also a reduction in the time-tomarket that potentially provides a major commercial advantage. It is estimated that the number of physical models has declined dramatically in Ricardo over the past years. Regardless of the power of the simulation tools, however, the company appreciates there will still be a need to physically test model performance. The increasing use of new and more powerful simulations in auto-engine design has major implications for skills and roles of designers and analysts. Simulation work in engine design was traditionally undertaken by analysts, while designers focused on the conceptual, detailed dimensioning, and manufacturing. In this case, designers are generalists with knowledge about materials, engines features, and manufacturing processes. Analysts, in contrast, are highly specialized users of simulations able to correctly set up inputs and interpret outputs. Despite the increased complexity of the products, performing simulation has become easier as user interfaces have advanced. Thus, designers can become increasingly involved in analysis. Many designers, however, still lack the depth of knowledge in a particular area to be able to conduct analysis competently. According to Johns: It is now possible for designers to undertake quite sophisticated analysis and 9 times out of 10 they will do this without a problem. However, more dangerously, 1 time out of 10 they will analyse a design and get it wrong because they do not have the specialist knowledge and experience to be able to recognise mistakes in either input or output or limitations of the modelling that render the analysis plausible but wrong.

There is still plenty of room for mistakes in analysis and even experienced analysts can make mistakes from time to time, especially when the analysis is not routine. In theory, designers could work from design templates embedded in the tools and using these design templates they could quickly create functioning engines. By using such templates, however, the designers would limit the scope for innovation in the design of the engine. As Johns says, ‘templates are straitjackets. They are ﬁne if you want to do repetitive analysis but not if you want innovation.’ Once it is necessary to go beyond the template embedded in the programme then the skills and the experience of specialist analysts

41

New Craft Skills of Engineering come to the fore. It is only these skilled operators and interpreters of the inputs and outputs from the simulation who can deal with new features and elements in the design. The relationship between designers and analysts in engine design has become more intertwined as a result of the use of new simulations tools. Decisions by designers can have implications for a wide range of interacting systems. Analysts working with designers can explore and interpret these interactions, seeing how one change may spill over across the system. The tools force designers and analysts to see the engine as a system rather than a number of components. It forces them to collaborate with more people and to open up to new connections across the design process. Using IvT, Ricardo has changed the innovation process for engine design. IvT has created the potential for new forms of integration across organizations and across countries. New simulation suites allow ﬁrms to co-simulate across different simulation packages, enabling ﬁrms to adopt a ‘best-in-class’ approach to simulation.

2.4

Discussion and conclusions

The two cases of Arup and Ricardo demonstrate that the nature of engineering is changing as a result of the introduction of new technologies, business strategies that seek growth in new markets for specialized services, and new approaches to regulation. The combination of the use of technology and new strategies and regulatory approaches is recasting the role of the engineer and changing the division of labour among participants in technological problemsolving. Von Hippel (2005) has suggested that such changes could represent a ‘democratizing’ of the innovation process, enabling those who are working to develop new solutions to work more closely with potential users and customers. In this new environment, the boundaries between experts and nonexperts, user and producer, and the ﬁrm and its external environment are becoming more ﬂuid and porous (Chesbrough 2003). At the core of many of these changes is IvT. As IvT is a relatively novel technology, its consequences for the skills of engineers are still being enacted. It is possible to envisage a deskilling trajectory, similar to that predicted by the application of computers to Advanced Manufacturing Technology (Braverman 1974). While some ﬁrms may embark down that road, the result will be similar to those ﬁrms that unsuccessfully tried to use Advanced Manufacturing Technology to deskill. Productivity depends upon the merger of the new technological possibilities with the appreciation and understanding of the old principles, and ‘ways of doing things’: the established rules of thumb and norms. What is clear from our studies is that the technologies of modelling and simulation are being combined with more traditional forms of engineering

42

New Craft Skills of Engineering design and problem-solving in novel ways. Engineering skills are becoming more dispersed. At the same time, techniques for problem-solving are being routinized in computer models. These are changing the ways in which problems can be solved. But evidence from these case studies suggests that this is not leading to a simple process of automation and deskilling. Instead, deep craft knowledge is often required and this expertise can provide innovative solutions when aided by the new digital tools. The study illustrates the importance of craft knowledge and therefore limits attempts to routinize some aspects of engineering knowledge. Segmented approaches that attempt to automate all aspects of problem-solving and design activities will not realize the advantages to be achieved when IvT is combined with deep craft knowledge. These advantages are considerable. The eminent architect, Frank Gehry has commented upon the combination of new technology and processes that have enabled him to produce his innovative, malleable, plastic form of building design saying that This technology provides a way for me to get closer to the craft. In the past, there were many layers between my rough sketch and the ﬁnal building, and the feeling of the design could get lost before it reached the craftsman. It feels like I’ve been speaking a foreign language and now, all of a sudden, the craftsman understands me. In this case, the computer is not dehumanising, it’s an interpreter (Gerhy in Pacey 2003).

Further research is needed to explore how the use of these techniques is resulting in changes in engineering practice. IvT has the potential to provide the basis for a more open, transparent approach to engineering design and decision-making. It could also enable faster results with greater concurrency of engineering work on different parts of a problem simultaneously, allowing near real-time decision-making. Its capacity to assist in interdisciplinary decision-making by enabling people from different backgrounds to understand potential solutions opens a whole range of possibilities for a more informed, possibly a more democratic engineering praxis. There is a rich potential for new research to document and describe these new practices (Von Hippel 2005). There remain several barriers to such a transformation. In many countries in the Organization for Economic Cooperation and Development (OECD), especially in the UK, engineering remains dominated by strong professional demarcations. The professional institutions in the UK and elsewhere establish standards of knowledge required for an individual to be qualiﬁed to solve engineering problems. They have played a critical role in the development of engineering knowledge and practice (Becher 1999). Given the diffuse and elastic nature of problem-solving in many modern engineering projects, however, the notion of a distinct and specialist ‘professional engineer’ has come under strain. Projects may involve a range of specialists with little or no

43

New Craft Skills of Engineering professional training in engineering but who have the capabilities to design and develop new solutions. Indeed, the increasing shift to ‘user toolkits’ in innovation management means that products can be designed by users with little technical knowledge (Von Hippel 2001). The challenge for the engineer in this environment is to provide the opportunity for the lay users to create new combinations or innovations that are both safe and functional. In this respect, the engineer acts as a teacher or a guide to the lay user, helping to assist them as they attempt to achieve their design objectives. Given this problem-solving environment, a wider notion is required of what an engineer ‘is’ and ‘does’ that goes well beyond the conﬁnes of professional accreditation and membership and its formal set of core engineering knowledge. This poses considerable challenges for engineering education providers with their traditional demarcations of different kinds of engineering—mechanical, civil, electrical, and so on—and for representative professional bodies. The development of new practices in engineering may provide an opportunity for governments in the OECD to attract more young people into careers in science and engineering. In companies like Arup and Ricardo, engineering involves playing with different technical solutions and working closely with a wide range of users on practical problems. This environment might allow young people the opportunity to transfer their love of gameplaying and simulation into their professional lives. The US Army has successfully used young people’s interest in computer games in its recruitment of future soldiers and it may be that such strategies will be successful in generating interest in careers in science and engineering. Engineering will also be more attractive as a profession for young people as the role of engineers in constructing the future is redeﬁned and better articulated. Rather than identifying themselves as a particular kind of engineer, a new generation could see themselves more broadly as contributors to technological innovation and sustainable development. One of the big challenges that governments will face here is to make engineering education reﬂective of good practices that are increasing part of leading areas of engineering. This will mean a shift towards problem-based learning, multidisciplinary teams, and working with users in the education of engineers. In many ﬁelds of engineering education, such a shift in educational practices would mark a considerable transformation.

Notes 1. The term ‘innovation technology’ is taken from Dodgson et al. (2005) and refers to a range of technologies used in the innovation process to support scientiﬁc and engineering problem-solving. These technologies build upon the latest developments in information and communication technologies (ICT) and operation and manufac-

44

New Craft Skills of Engineering turing technologies (OMT) and include: simulation and models, rapid prototyping, artiﬁcial intelligence or expert systems, data mining, grid computing, and virtual reality. We are grateful to Paul Robertson, Maureen McKelvey, Keith Smith, Franco Malerba, Luigi Orsenigo, Andy Davies, Mats Holme´n, Stan Metcalfe, Jennifer Whyte, and Catelijne Coopmans for comments on previous versions of this chapter. 2. The two cases are based on an extensive range of interviews at Arup and Ricardo carried out over many years. Our research association with Ricardo began in 1991 and in Arup in 1996. The method used in the study is described in detail in Dodgson et al. 2005. The present chapter is based on 28 interviews and a month-long observation period at Arup, and an interview at Ricardo conducted in 2003. In the case of Arup, one of the research team spent three weeks of on-site participant observation. The Ricardo case draws from an extensive interview with Richard Johns, head of Ricardo Software, and his written comments on the interview transcript, and the case description of Ricardo. We are extremely grateful for the generosity of these two organizations to our research. 3. Fluid simulation packages include WAVE—a gas dynamics simulation programme enabling performance simulations to be carried out based on intake, combustion, and exhaust system design; and VECTIS—a three-dimensional computational ﬂuid dynamics programme for solving ﬂow equations governing conservation of mass, momentum, and energy. Ricardo has also developed a range of mechanical systems simulations, including ENGDYN—a simulation environment for analysing the dynamics of the engine and powertrain; VALDYN—for simulating the dynamics of the valvetrain; PISDYN and RINGPAK—simulation packages for design and analysis of piston/ringpack assemblies; and ORBIT—for detailed analysis of bearings.

References Allen, T. J. (1977). Managing the Flow of Technology. Cambridge, MA: MIT Press. Baldwin, C. and Clark, K. (1997). ‘Managing in an Age of Modularity’, Harvard Business Review (September–October): 84–93. —— and —— (2000). Design Rules, 1: The Power of Modularity. Cambridge, MA: MIT Press. Becher, T. (1999). Professional Practices: Commitment and Capability in a Changing Environment. New Brunswick, NJ: Transaction Publishers. Bravermann, H. (1974). Labor and Monopoly Capital: The Degradation of Work in the Twentieth Century. New York: Monthly Review Press. Brusoni, S., Prencipe A., and Pavitt, K. (2001). ‘Knowledge Specialization and the Boundaries of the Firm: Why Do Firms Know More Than They Make?’, Administrative Science Quarterly, 46(4): 597–621. Chesbrough, H. (2003). Open Innovation: The New Imperative for Creating and Proﬁting from Technology. Boston, MA: Harvard Business School Press. Constant, E. (1980). The Origins of the Turbojet Revolution. Baltimore, MD: Johns Hopkins University Press. —— (2000). ‘Recursive Practice and the Evolution of Technological Knowledge’, in Z. Ziman (ed.) Technological Innovation as an Evolutionary Process. Cambridge: Cambridge University Press, pp. 219–30.

45

New Craft Skills of Engineering Cowan, R., David, P., and Foray, D. (2000). ‘The Explicit Economics of Knowledge Codiﬁcation and Tacitness’, Industrial and Corporate Change, 9(2): 211–53. D’Adderio, L. (2001). ‘Crafting Virtual Prototype: How Firms Integrate Knowledge and Capabilities Across Organizational Boundaries’, Research Policy, 30(9): 1409–24. —— (2004). Inside the Virtual Product: How Organizations Create Knowledge through Software. London: Edward Elgar. Dasgupta, P. and David, P. (1994). ‘Towards a New Economics of Science’, Research Policy, 23(5): 487–522. David, P. and Foray, D. (1995). ‘Accessing and Expanding the Science and Technology Knowledge Base’, STI Review, 16. Dodgson, M. (1993). Technological Collaboration in Industry: Strategy Policy and Internationalization in Innovation. London: Routledge. ——, Gann, D. M., and Salter, A. (2002). ‘The Intensiﬁcation of Innovation’, International Journal of Innovation Management, 6(1): 53–84. ——, ——, and —— (2004). ‘Craft and Code: The Intensiﬁcation of Innovation and Management of Knowledge’, in K. Green, M. Miozzo, and P. Dewick (eds.) Knowledge and Technology: Implications for Firm Strategy and Industrial Change. Cheltenham, UK: Edward Elgar. ——, ——, and —— (2005). Think, Play, Do: Technology and Organization in the Emerging Innovation Process. Oxford: Oxford University Press. Dosi, G. (1982). ‘Technological Paradigms and Technological Trajectories: A Suggested Interpretation of the Determinants and Directions of Technical Change’, Research Policy, 11: 147–62. —— (1988). ‘Sources, Procedures and Microeconomic Effects of Innovation’, Journal of Economic Literature, 26: 1120–71. Ferguson, E. (1992). Engineering in the Mind’s Eye. Cambridge, MA: MIT Press. Gann, D. and Salter, A. (1998). ‘Learning and Innovation Management in Project-based, Service-enhanced Firms’, International Journal of Innovation Management, 2(4): 431–54. Hargadon, A. (2003). How Breakthroughs Happen: The Surprising Truth about How Companies Innovate. Cambridge, MA: Harvard Business School Press. —— and Sutton, R. (1997). ‘Technology Brokering and Innovation in a Product Development Firm’, Administrative Science Quarterly, 42(4): 716–49. Henderson, K. (1999). On Line and On Paper: Visual Representations, Visual Culture, and Computer Graphics in Design Engineering. Cambridge, MA: MIT Press. Johnson, B., Lorenz, E., and Lundvall, B-A. (2002). ‘Why all this Fuss About Codiﬁed and Tacit Knowledge?’, Industrial Corporate Change, 11(2): 245–62. McCullogh, M. (1996). Abstracting Craft: The Practiced Digital Hand. Cambridge, MA: MIT Press. McKenna, P. and Maister, D. (2002). First Amongst Equals. New York: Free Press. Nightingale, P. (1998). ‘A Cognitive Model of Innovation’, Research Policy, 27: 689–709. Pacey, S. (2003). ‘Interview with F. Gehry’, RIBA Journal (November): 78–9. Pahl, G. and Beitz, W. (1996). Engineering Design: A Systematic Approach, 2nd edn. (K. Wallace, trans.). London: Springer-Verlag. Pavitt, K. (1987). On the Nature of Technology. Inaugural Lecture, University of Sussex, SPRU.

46

New Craft Skills of Engineering Pavitt, K. (1998). ‘Technologies, Products and Organization in the Innovating Firm: What Adam Smith Tells us and Joseph Schumpeter Doesn’t’, Industrial and Corporate Change, 7: 433–52. Perlow, L. A. (1999). ‘The Time Frame: Towards a Sociology of Work Time’, Administrative Science Quarterly, 44: 57–81. Salter, A. and Gann, D. (2003). ‘The Sources of Ideas in Engineering Design’, Research Policy, 32: 1307–24. Schrage, M. (2000). Serious Play: How the World’s Best Companies Simulate to Innovate. Boston, MA: Harvard Business School Press. Simon, H. A. (1996). The Sciences of the Artiﬁcial. Cambridge, MA: MIT Press. Steinmueller, E. (2000). ‘Will New Information and Communication Technologies Improve the ‘‘Codiﬁcation’’ of Knowledge’, Industrial and Corporate Change, 9(2): 361–76. Thomke, S. (2001). ‘Enlightened Experimentation: The New Imperative for Innovation’, Harvard Business Review (February): 67–75. —— (2003). Experimentation Matters. Cambridge, MA: Harvard Business School Press. Tuomi, I. (2002). Networks of Innovation: Change and Meaning in the Age of the Internet. New York: Oxford University Press. Vincenti, W. (1990). What Engineers Know and How They Know It. Baltimore, MA: Johns Hopkins Press. Von Hippel, E. (2001). ‘User Toolkits for Innovation’, Journal of Product Innovation Management, 18: 247–57. —— (2005). Democratizing Innovation. Cambridge, MA: MIT Press. Whyte, J. (2002). Virtual Reality in the Built Environment. London: The Architectural Press. Williams, R. (2003). Retooling: A Historian Confronts Technological Change. Cambridge, MA: MIT Press.

47

3 Innovative Opportunities and Dependencies: Illustrations from Mobile Communications Magnus Holme´n, Mats Magnusson, and Maureen McKelvey

3.1

Introduction

This chapter explores the question of how and why actors innovate in terms of identifying, acting upon, and realizing new combinations of resources and identiﬁed market needs within a larger economic system.1 It aims to capture these processes by analysing innovative opportunities, which consist of three conceptual elements. The chapter is primarily focused on conceptual development in order to understand the dependencies within innovative opportunities. In addition to conceptual discussions, it also uses illustrative cases from the early development of technology and business platforms in mobile telecommunication to demonstrate dependencies. Hence the argument put forth is that the way in which speciﬁc actors can develop innovative opportunities related to a new technology are inﬂuenced by choices made by other actors, as analysed in relation to somewhat broader knowledge and market processes. In doing so, this chapter aims to contribute to an improved understanding of how economic transformation relates to the activities of individuals and organizational actors. The key starting point here, as outlined in Chapters 1 and 10 of this book, is that economic competition is driven by the ability of actors—especially different ﬁrms—to develop competencies and knowledge to innovate and to appropriate the returns from innovation (Schumpeter 1934). Our argument is that opportunities are a fundamental aspect of what actors perceive and do, and thus they are a core feature of economic transformation. Yet existing conceptualizations cover only some speciﬁc aspects, leaving other aspects relatively ill-deﬁned theoretically. This chapter contributes to the broad and growing literature that addresses opportunities in industrial dynamics and entrepreneurship.

48

Innovative Opportunities and Dependencies The concept of innovative opportunities includes a set of three conceptual elements to deﬁne substantial processes and to help analyse these processes empirically. As developed in other work (Holme´n et al. 2004, 2005), these elements include economic value, resources, and appropriation. This concept is a means to combine the literature on science, technology, and innovation especially around ‘search activities’ for new technologies with the literature on ‘entrepreneurship and venture creation’. This thereby places the innovating actor in relation to other actors and other processes within the economy. More speciﬁcally, the purpose of the chapter is to argue for, and demonstrate the existence of what we may refer to as dependencies within innovative opportunities. To discuss dependencies, the chapter draws on existing literature and uses illustrative cases from the early development of technology and business platforms in mobile telecommunication. The empirical cases demonstrate dependencies among different actors that result in ‘systemic’ effects, and these actor–system interactions take the form of dependencies across different types of opportunities as well as dependencies across different elements of these processes. The chapter concludes by examining how the concept of innovative opportunities can help us understand the shape of the trajectory of development, or overall transformation, of the economic system. Section 3.2 deﬁnes our concept, ‘innovative opportunities’, which draws upon earlier proposed opportunity concepts but is a separate deﬁnition and contribution. This discussion of opportunities is based on work done elsewhere, and stresses the interaction between actor and systems in innovation processes, drawing upon literature from evolutionary economics in a broad sense, entrepreneurship literature, and the knowledge- and resourcebased theories of the ﬁrm. Section 3.3 then uses this conceptualization to structure an overview of early development of technology and business platforms for mobile communication, focusing on 3G and i-mode. Section 3.4 articulates further what happens within the processes of the three conceptual elements of innovative opportunities by focusing on dependencies. It also goes back to the cases to reinterpret dependencies. Section 3.5 reﬂects upon our analysis of dependencies manifested in innovative opportunities—in order to understand the larger effect on transformation of the economic system.

3.2 Innovative opportunities This section presents our conceptualization of innovative opportunities, through deﬁnitions and elements, by drawing from the existing literature but also going beyond it.2 A main reason for proposing innovative opportunities

49

Innovative Opportunities and Dependencies is to better understand the innovation-driven economy, especially the incentives, perceptions, and possibilities of actors to innovate. This chapter uses a broad deﬁnition of innovation, as linked to business use. Following Edquist et al. (2001: 10): ‘Key distinctions . . . are between product and process innovations. With respect to product innovations, we distinguish between material goods and intangible services and, regarding process innovations, we distinguish between technological and organizational innovations.’ Innovation as a concept used in this chapter is speciﬁcally related to how technological knowledge is incorporated into hardware, services, infrastructures, consumer behaviour, and so on, as sets of innovations related to delivering goods and services purchased by consumers.3 In other words, the discussion is speciﬁcally concerned with innovations in terms of goods and services related to technologies and used for a business context—and not technologies per se. Technologies are considered from a business point of view, where it is assumed that ﬁrms use technologies to transform inputs into outputs, and as such technology may be useful to create and access value.4 A theoretical rationale for addressing these topics comes from our assumption that innovation as a process involves diverse actors who both compete and collaborate, and whose actions are coordinated through different types of dependencies. Given this starting assumption, it implies that the reader ﬁrst needs some understanding of how our concept of ‘innovative opportunities’ ﬁts in with the existing literature, which is to some extent divided into those stressing ‘micro’ processes and those stressing ‘macro’ processes. Hence, on the one hand, innovation processes are driven by single actors, so that human or organizational behaviour plays a fundamental role in why change occurs. There is a broad literature studying these processes. For example, some business strategists in the resource-based tradition tend to explain strategic change and innovation in terms of organizations’ core competencies (Prahalad and Hamel 1990; Leonard-Barton 1992). Human actors are crucial for the unfolding of innovation processes, and they may be active in various types of organizations. That human action is at the base of innovation is even more clearly reﬂected in the rapidly growing literature on entrepreneurship and new ventures where the individual’s skills and ambitions are considered to be of great importance. On the other hand, there are characteristics of change that may be considered to be systemic. Work by economic, business, and technology historians such as Rosenberg (1982), Hughes (1983), and David (2004) strongly suggest that science and technologies tend to be ‘systemic’ in the sense that changes are necessary within many interrelated technologies and social institutions. Hence, dependencies across variables are common in intertemporal co-evolutionary processes such that different variables or different processes are likely to be ongoing in parallel but also to exert mutual inﬂuence (Nelson 1995), a reasoning that is similar to co-evolution in that it also considers mutual

50

Innovative Opportunities and Dependencies inﬂuences across different variables and processes (see Chapter 10). This type of inﬂuence across the economic system may lead to particular trajectories of development during economic transformation. Research in innovation studies and evolutionary economics suggests that experimentation with different types of knowledge, industrial structures, and organizational and institutional forms interact and may direct speciﬁc national economies onto different trajectories of growth and development (Rosenberg and Birdzell 1986).5 Even though the literature tends to stress different levels, our ambition is to combine systemic and actor levels in discussing innovative opportunities in order to make further sense of the processes that correspond to the conceptual elements mentioned earlier and further explained later. The argument put forth here is that dependencies among different actors result in ‘systemic’ effects, and these take the form of dependencies across different types of opportunities as well as dependencies across different elements of these processes. These dependencies affect and shape different trajectories of growth and development in economic transformation. One type of dependency is across different types of opportunities. The literature discusses technological, entrepreneurial, and productive opportunities, as reviewed in Holme´n et al. (2004, 2005). One stream of literature, broadly deﬁned as evolutionary economics literature, discusses technological opportunities (Scherer 1965: 1121; Breschi et al. 2000; Palmberg 2004). It addresses how and why actors and their competencies co-evolve with institutional structures, knowledge, and industries at a broader system level. The second type of literature is the broadly deﬁned entrepreneurship and theory of the ﬁrm literature. This literature focuses on how and why individuals and ﬁrms (or, occasionally, non-ﬁrm organizations) perceive opportunities for new business activities and act upon them, thereby attempting to explain the generation and exploitation of different innovations (Kirzner 1997; Shane 2000). The third type of literature is the knowledgeand resource-based theories of the ﬁrm and the related concept of productive opportunities suggested by Penrose (1959) as a key factor for explaining the growth of individual ﬁrms (Kor and Mahoney 2000; Alvarez and Busenitz 2001). Hence a rich literature exists which addresses the question of opportunities in industrial dynamics and entrepreneurship, as further explored in Section 3.3. Based on Holme´n et al. (2004, 2005), we wish to start by arguing that the existing literature provides three useful concepts of opportunities— namely technological, entrepreneurial, and productive—but none of them is alone sufﬁcient to capture the complexity of the innovation processes in the economy. Hence, even though literature on each type of opportunity stresses aspects that are relevant to a more comprehensive understanding of how opportunities relate to economic transformation, none is sufﬁcient on its own.

51

Innovative Opportunities and Dependencies Our proposal is to instead develop a formulation of ‘innovative opportunities’, which incorporates aspects of the technological, entrepreneurial, and productive opportunities but which is a separate concept. Holme´n et al. (2004) propose that an innovative opportunity is ‘the perceived possibility to realize a potential economic value inherent in a new combination of resources and market needs, emerging from changes in the scientiﬁc or technological knowledge base, changes in customer preferences, or changes in the interrelationships between economic actors’. In other words, this concept comprises both ‘market pull’ and ‘science and technology push’, in terms of deﬁning aspects related to a potential market as well as aspects related to the scientiﬁc and technological knowledge needed to serve this speciﬁc market. These are also ﬁrm-driven process in making decisions about the development and allocation of resources. Holme´n et al. (2004) also propose, as further developed in Holme´n et al. (2005), three underlying conceptual elements to help deﬁne how the actors go about perceiving and realizing the innovative opportunity. These elements help us move from a deﬁnition to conceptual elements that can be analysed within speciﬁc empirical cases. Arguably, an innovative opportunity must comprise at least the following three conceptual elements: 1. A perceived economic value for someone. 2. A perceived possibility that the resources needed to realize the opportunity can be mobilized. 3. A perceived possibility that at least some part of the generated economic value can be appropriated by the actor pursuing the opportunity. Our argument is that all three conceptual elements must be present in order for an innovative opportunity to be realized in the sense deﬁned here, and this also implies that aspects of each element can be developed further theoretically and also tested empirically. Furthermore, we suggest that interlinkages across the mentioned constituent elements constitute another type of dependency that is of importance for understanding economic transformation. Our starting point for the presented deﬁnition of innovative opportunities and the three conceptual elements has been literature within innovation studies and evolutionary economics, as argued in Chapters 1 and 10 of this book. That is, in order to deﬁne opportunities conceptually, one must build upon a more nuanced understanding of the nature of innovation processes. This could be termed our ‘evolutionary paradigm’ of historical development, of search activities occurring under conditions of uncertainty. First, we wish to stress that the three elements of innovative opportunities should be understood as processes normally involving not only one ﬁrm but a number of different actors. Technological innovations often require series of

52

Innovative Opportunities and Dependencies improvements within goods, services, processes, and so on. Different actors may therefore develop different but complementary innovations. Second, we wish to stress that no one actor will necessarily succeed in dealing with all elements. Instead, in complex technologies it is likely that actors are active in developing different types of opportunities as well as being involved in identifying, acting upon, and realizing the three different conceptual elements. The reason for this is that given uncertainty about markets and technology, it is quite possible that different actors will have different perceptions of the feasibility and ‘value’ of possible ideas. Third, innovation processes are seen as multidimensional and are characterized by a non-negligible level of novelty, uncertainty, and ambiguity. This conceptualization stresses that individual actors have partial and incomplete knowledge and therefore often gain information and resources through systemic elements, such as through network relations with other actors, involvement in activities and processes, and so on. Finally, we wish to stress the importance of ‘subjective’ perception during conditions of uncertainty, as well as an ‘objective’ understanding of ongoing processes. Perception is necessary in relation to identifying, acting upon, and realizing these opportunities—if uncertainty and ambiguity are prevalent in these processes. This also helps explain differential behaviour of actors, which should be visible empirically as well.

3.3 Innovative opportunities in 3G and i-mode This section uses the concept of innovative opportunities and its three conceptual elements in order to interpret the early development of business and technology platforms in mobile communications. Two illustrative cases, namely the shift from existing mobile communication standards to 3G and i-mode, are analysed as processes where actors develop innovative opportunities.6 This empirical overview is also useful as background information for the more speciﬁc analysis of each element in the subsequent section. Our interpretation is that these two illustrative examples probably represent two extreme cases in terms of perceptions of innovative opportunities—idiosyncratic perceptions within 3G versus shared perceptions within i-mode. Thus, we conclude that the cases of 3G and i-mode areas are superﬁcially similar. Hence, on the one hand, they are paired with similar cases to help highlight differences, but also to show clear differences in how actors perceived economic value and means of appropriation as well as differences in how to organize the development of many innovations within these systems. On the other hand, the research design strategy is also that of contrasting extreme cases. The purposes of these empirical illustrations are to explore the

53

Innovative Opportunities and Dependencies boundaries of our three conceptual elements as well as to demonstrate whether dependencies exist or not. The empirical cases presented here have been extensively covered in the literature and we present no new empirical results, but merely put forward key observations with respect to innovative opportunities from the earlier studies.7 The illustrative cases provide interesting insights in relation to this view of innovative opportunities, as also explained in Holme´n et al. (2005).8 Mobile communication is here used in a sense primarily related to voice and data communication over radio waves. It can encompass different speciﬁc standards such as NMT, GSM, GPRS, and 3G. By early development, we mean a phase during which much investment is being made in developing technical standards, business frameworks, and in building the physical infrastructure (or network). We are particularly focused on the change in technical standards, where the new standard may be closer or further away from the existing standard (and infrastructure) seen in technical dimensions. A new standard is usually considered to either represent a break with previous means of communication, such as NMT as an alternative to ﬁxed phones, or else an improvement upon existing standards, such as the move from GSM to 3G. This early development is a phase that often requires considerable investment, and at a time when it is uncertain whether users and suppliers of other hardware, software, content providers, services, and so on, will use—and pay for—this infrastructure in the future. For example, Lindmark et al. (2004b: 8) stress the turbulent nature of the telecommunication industry around this period: ‘The boom of the late 1990s . . . was partly driven by unrealistic expectations of market growth and an overheated ﬁnancial market, a mature mobile market in need of renewed growth, and fuelled by an unfortunate 3G licensing regime. This led to overinvestment, particularly in 3G licenses, but also in network capacity and in acquisitions raising debt to levels that could not be sustained, leading to write-downs, divestments, bankruptcies and halted investments.’ Hence, this phase is a very particular one, involving much technical and market uncertainty within different dimensions of the possible future business. This outline has implications for understanding the types of opportunities involved (Holme´n et al. 2005). In terms of technological opportunities, the emergence and implementation of new scientiﬁc and technological knowledge itself can be considered to regenerate the pool of opportunities for mobile communication. In terms of entrepreneurial opportunities, the early development of technical standards for mobile communication has often opened up ‘economic space’ in the sense of access to capital, human capital, and other resources that enable new business ventures, or start-ups to enter the economy. In terms of productive opportunities, the early development of technical standards for mobile communication has also been related to

54

Innovative Opportunities and Dependencies intensiﬁcation of existing product lines and to market and technological diversiﬁcation of large ﬁrms. Mobile communication standards set the technical speciﬁcations of a large system, and are known to have network effects. The cases involve early phases, but in somewhat different geographical regions, namely cases of 3G, a predominantly European set of standards for mobile communications, and imode, the dominating Japanese operator’s (DoCoMo) mobile communication platform for services.9 Each of these two large systems in mobile communications consists of a range of somewhat unique, somewhat overlapping services, goods, and infrastructures. The time period covers roughly the period 1997– 2002, which only relates to the initial development on the markets. We only focus on this early phase and would simply like to point out that things have changed since. The ﬁrst case is 3G, or third generation, and it involves innovative opportunities that are of a more systemic nature than what was the case for GSM, the initially mainly European mobile communication standard that preceded 3G. Even so, 3G is not a completely different standard but instead was aimed to build on and augment the technical and market possibilities of GSM. Specifically, the focus of this new technical standard was intended to be packet switched data communication, allowing for any type of data instead of just voice together with limited data transfer abilities, such as the example of SMS. Technological, entrepreneurial, and productive opportunities were identiﬁed by different actors, as part of the innovative opportunities, as evidenced by high levels of R&D (public and private), start-ups in related ﬁelds, and intensiﬁcation and diversiﬁcation of large ﬁrms. First, in terms of the perception of economic value, a range of actors, including the mobile telecom infrastructure suppliers, the telecom operators, politicians, and perhaps customers perceived that the value of this new system would be tremendous. Empirical examples of these perceived future beneﬁts can be inferred from the multibillion-dollar auctions in the UK and Germany for acquiring a 3G license. However, some companies had difﬁculties in later capitalizing on their investment in purchasing of such licenses, as indicated by Table 3.1. Table 3.1 shows that huge investments had to be written off by individual ﬁrms, suggesting that while perceptions were that economic beneﬁts were high, this was difﬁcult to realize, at least for ﬁrms in particular periods and particular countries. It was difﬁcult to capture the returns, despite—or perhaps because of—the very high expectations initially, or at least hype, about high expectations. Financial modelling has demonstrated that in small markets, such as Sweden, made by operators were highly unlikely to be recouped within 15 years ¨ rkdahl and Bohlin 2002). These models were based on the assumption that (Bjo user behaviour was unlikely to be changed drastically compared to the user

55

Innovative Opportunities and Dependencies Table 3.1. Examples of ﬁrms that abandoned 3G markets around year 2001 Firm

Country of licences (full or partial ownership)

Money paid for licences (e)

Sonera MobilCom Telefonica Tele2

Germany, Norway, Italy, Spain Germany Austria, Germany, Switzerland, Italy Norway

4.3 billion 8.4 billion 6.1 billion 12.8 million

¨ rkdahl 2004: 9, Exhibit 2 Source: Bjo

acceptance and consumption of GSM, and the authors also felt this was highly unlikely to take place. The results in this study thereby seem to stand in sharp contrast to the expectations—or hype. Perhaps too much focus was put on the identiﬁcation of technological opportunities—but too little placed on the realization of the market aspects of these innovative opportunities. Thus, the perception of future economic potential—implicitly at least—should have been tied to the need to expand the user base and services. Despite this, one can argue that particularly in the earliest phase, the question of how to expand users and services of 3G was not addressed so explicitly. Issues like content providers and development of new hand-held terminals allowing for the use of new services most likely did not receive sufﬁcient attention in the early phases of 3G. Instead, for 3G, the mobilization of resources in the early phase seems to have been primarily mobilized around suppliers. Indeed, innovation in telecommunication has been primarily driven by the supply side, not least by the providers of mobile telecommunication infrastructure, instead of by the demand side. The problem with this was that the mobile databased services were supposed to complement or even substitute voice telephony as the big revenue source of telecom operators. Hence, this in fact implies a radical change in the way mobile hand-held terminals are used, and also how they are designed. Such changes to accommodate new users and new uses would be necessary areas of innovation in services and goods, and would be necessary to recoup the investments of the operators. Hence, our interpretation is that one problem here was that the foci of innovation was perhaps placed too much on the supply side in terms of infrastructure investments, while at least to some extent neglecting the demand side, which could for example have been stimulated in terms of encouraging content generation by diverse actors. Moreover, the competitive bidding situation probably encouraged rather idiosyncratic perceptions of each actor’s ability to innovate and capture returns. Hence, appropriation has also been an issue in the early phase. In particular, 3G suffered from the lack of clear appropriation models, which probably had a particularly adverse effect on all but the largest operators. Returns to investments in this type of mobile communication system probably requires

56

Innovative Opportunities and Dependencies agreed-upon ways of sharing the returns from services and goods, or in ‘bundles’ of goods and services used simultaneously. For example, should such payment be based on time spent on network, ﬂat rate, or on the size of the download? In the early phases, the European evidence suggests that operators were unwilling to create a revenue model that shared income between the operators and the content providers from the increased trafﬁc of new uses. This had broader effects on innovations because it consequently reduced incentives for content providers to develop new services. This then constrained the appropriation of returns and thereby the division of innovative labour among the different actors. This ﬁrst case thus represents an illustration of where innovative opportunities were identiﬁed and thought to ‘exist’ in different domains, but where the next step, to actually realize the opportunities and appropriate the returns, was not only difﬁcult but also affected the future trajectory of innovations necessary to use 3G. The second case is i-mode in Japan, which provides a striking contrast to 3G. To begin with, in contrast to 3G with many competing operators and actors, imode is a case where there was only one major actor that gave the system ‘credibility’, even though they were linked to a variety of other actors. That is, DoCoMo guaranteed the success and feasibility of the system in terms of market size and technology. This had implications for the perception of economic value, because their commitment signalled to other actors that there should be a high potential economic value of the system. This ﬁrm would exploit productive opportunities—but do so in such a way as to help realize innovative opportunities through innovations in a range of complementary technologies and services. Furthermore, much more focus was put on generating and diffusing a shared perception of the future system in Japan than was the case for 3G in Europe. In this way, a much more coordinated approach could develop between the different actors developing hardware, services, and infrastructure. This implied that different types of innovations could be developed and used together to deliver services demanded by users. This has an important implication during the early phase, because this generation of a common knowledge platform reduced uncertainty regarding the new system, and consequently also put limits on experimentation. Hence, the identiﬁcation, acting upon and realization of innovations related to the innovative opportunity were focused within a more narrow range, stressing complementary technologies and user services. This linked the shared perception of economic value to actors who mobilized resources in order to be able to innovate. Thus, at least in the initial phase, i-mode was able to align perceptions of innovative opportunities among the actors in such a way that innovation in many different areas of technology, services, and products worked together to increase the user base. Lindmark (2005: 55) argues that a number of complementary components were in place

57

Innovative Opportunities and Dependencies from the start, which helped to explain the success of i-mode. These components include: ‘(a) a packet-switched network, (b) a service gateway, (c) terminals supporting i-mode services (including micro-browsers and with standard screen size facilitating for content providers to develop content), (d) content development tools and support, (e) micro-payment system, (f) a simple and fairly html-compatible mark-up language, and (g) a portal’. These component technologies were developed, in turn, by both DoCoMo and external parties. One can thus argue that the speciﬁcation of the basic business model was instrumental for resource mobilization among a range of actors developing different components necessary to the functioning of a mobile communication standard. In terms of appropriation, a more centrally coordinated system was also developed very early on for i-mode. A case in point is the revenue sharing model that DoCoMo put in place, which gave a ﬁxed sum of the income (91 per cent) to the content providers (Lindmark et al. 2004a). From this business model of DoCoMo many content providers could then build up their own business services using the i-mode system. In closing this discussion of innovative opportunities, we wish to stress that our analysis is of the early development of these technology and business platforms. As such, one should not interpret 3G as a failure and i-mode as a success. Of course in both cases at times some components of the entire system have been lacking and there have been incompatibilities between the perceptions held by suppliers and customers, respectively. So the argument put forth here is not that one large dominant actor must necessarily ‘direct’ innovation. Instead, the argument is that these shared perceptions of the different conceptual elements of an innovative opportunity, among a diverse group of actors, helped stimulate innovations within different components and expand the user base. Hence, over time, it is likely that the different forms of ‘coordination’ may work more or less well to solve different challenges involved in identifying, acting upon, and beneﬁting from innovative opportunities in this type of complex technology, as represented by early development of business and technology platforms for mobile communication.

3.4

Dependencies in innovative opportunities

Given the earlier conceptual and empirical starting points, this section goes on to explore the three conceptual elements in more detail. Our interest in dependencies provides a way to explore what happens as actors identify, act upon, and realize innovative opportunities when their choices are related to systemic aspects. These cases demonstrate dependencies among different actors that result in ‘systemic’ effects, and these take the form of dependencies

58

Innovative Opportunities and Dependencies across different types of opportunities as well as dependencies across different elements of these processes. By analysing dependencies as both constraining and enabling behaviour, one can help explain why innovation is not only about change—but also involves some characteristics of ‘stability’ and ‘lock-in’ within a complex business, technological, public policy, and organizational context. Thus, to illustrate the impact of dependencies upon innovation outcomes, we ﬁrst brieﬂy discuss the conceptualization of the three elements, and then compare this view against that of the two historical cases within mobile communication.

3.4.1

Perceived economic value

The ﬁrst conceptual element of innovative opportunities is the perception of economic value. In other words, in order to use technology in a business context someone has to perceive that the knowledge, techniques, inputs, and so on may be used in such a way as to create and access economic value. The underlying conceptual issue here is how and why certain actors perceive a potential economic value at all, whether related to technological, entrepreneurial, productive, or innovative opportunities. This is a key issue because if innovations help drive dynamics in the economy, then some person and organization must think they will beneﬁt ﬁnancially from introducing modiﬁed or new products and activities. One view, related to technological opportunities literature, stresses that technological knowledge has to be used for business purposes. One way of thinking about the difference between ‘technical knowledge per se’ and the use of technologies in a business context is to consider the historically deﬁned difference between an invention and an innovation (Schumpeter 1934: 88–9). Of course, the modern conceptualization argues that no one object or piece of knowledge is exclusively an ‘invention’ or an ‘innovation’, but that instead an innovation draws from a variety of knowledge and industrial dimensions. This implies that actors must mix different ideas of what they are doing, which helps contribute to the non-linearity of innovation processes (Edquist and McKelvey 2000; Fagerberg et al. 2004). Division of innovative labour is an obvious part of this story. As pointed out by Sanz-Velasco and Magnusson (2004), complex entrepreneurial ventures can hardly be seen as linear processes composed of opportunity recognition followed by idea realization and launch, but rather as a process characterized by interaction and interpretation in which the venture imposes its ideas on a market in an intrusive manner. The other two types of dependencies to explore here are those across different actors that result in ‘systemic’ effects and across different elements of these processes. Under different conditions it is likely that diverse perceptions

59

Innovative Opportunities and Dependencies of economic value (of an innovation) may also lead to diverse outcomes of ﬂexibility or stability in the economic system. This argument can be explored through two extreme conditions. In the ﬁrst case, we can assume that all agents have extremely diverse perceptions of the future economic value from innovations. We would argue that a substantial number of idiosyncrasies in perception (or, what may also be called uncertainty in many dimensions) will probably increase the range of experimentation in this ﬁrst case. This holds even if there are shared perception up to a point, followed by diversity. In the second case, we can assume that all agents involved have some similar ideas about the ‘possible future’ economic value of an innovation or technology—and about how to innovate in order to appropriate that value. In this second case, we may assume that they address the underlying problem (or use the technologies) in similar ways such that a ‘dominant design’ or ‘platform’ emerges. The two empirical cases in Section 3.3 can help nuance this discussion by exploring these questions about the role of collective perceptions and coordination of a diverse set of innovative search activities. It should be noted that both 3G and i-mode were perceived to have a high economic value, but the two cases differed in the role of coordination. We have suggested that the success of i-mode is largely attributable to DoCoMo’s role as guarantor for the new technology, thus ensuring the market for content providers. However, the ﬁrst case of 3G also suggests that innovations may not be widely diffused— even if they are likely to have a high abstract economic value—if consumers are not able to identify beneﬁts, as seen in the case for 3G. First, this problem with the difference in perception and outcome between 3G and i-mode is telling in that a major difference in how i-mode was able to move towards a higher economic value resided in the difference in appropriation and in resource mobilizations. This illustrates the interdependencies of these dimensions. Second, within early development of technology and business platforms for mobile communication, it is clear that different types of innovations seem to require substantial matching between the development of different components, in terms of technologies, goods, and services within the large technological system. Third, this issue of coordination of possible trajectories during the development and use of scientiﬁc and engineering knowledge for this business context must also include some matching of the perceptions of supply and demand. Although we have assumed innovation from the side of hardware, service and infrastructure providers, one of the main questions that suppliers have had to address is how to ‘match’ their perception with what agents on the demand side are willing to buy, both today and in the future. Finally, for understanding the perception of innovative opportunities, these illustrative examples suggest that an innovative opportunity may not be realized at all, especially if a large system is involved, if it is not possible to attract enough developers to solve all the technical and business problems of

60

Innovative Opportunities and Dependencies bits and pieces of the system—in this case, innovations around a platform for mobile communication. It may be difﬁcult for each single innovation—or single innovator that is acting within a large system—to attract a critical mass of people to further develop and use it. Such systemic technology thus involves aspects such as ‘attraction to additional innovators’ or ‘network externalities’ that appear crucial here today and not just in history of business and technology, if the idea is to survive as a long-term business venture. It also suggests that for an innovative opportunity to be realized, there needs to be some shared interpretation (at some point) of the economic value of the new goods and service product or large technological system. This needs to be shared among actors on the supply and demand sides, and often by multiple and diverse actors on the supply side.

3.4.2

Perceived ability to mobilize resources

The second conceptual element of the process—after identifying possible economic value to an innovation—is that of identifying and acting to mobilize resources. As outlined brieﬂy earlier (and more fully developed in Holme´n et al. 2004, 2005), our view is that the ﬁrm’s ability to mobilize resources is a necessary element in innovation and entrepreneurship—but is linked not only to ﬁnance but also to knowledge ﬂows and to potential customers. These may be expressed either as productive or entrepreneurial opportunities. The most direct input in resource mobilization often comes from ﬁnancial ﬂows. For large companies it may be accessing internal capital or leveraging existing assets. For small and/or young companies, it may be ﬁnding investors who are willing to provide capital for a venture given a risk premium. In order to bring this risk premium down to a reasonable level, ambiguity and uncertainty concerning the feasibility of the business concept normally have to be reduced by showing proofs of market estimates, cost structures, and so on. Consequently, the management team and/or ﬁrm has to try to ‘prove’ that the innovation will attract customers, in order to mobilize resources. This is the reason why early users or lead customers may be very important in R&D investments in that they show that there is a genuine interest in the particular type of solution. Hence, resource acquisition and customer attraction are not two different parts of a linear process, but are instead two highly intertwined activities and it is far from evident even from the outset of ‘early phases’ which one will follow the other: investment following customers or vice versa. This clearly suggests that there are dependencies across elements. Another apparent dependency between the constituent elements is related to the mobilization of resources in terms of content providers and the other actors’ capability to clarify what the economic value of the content providers’ input would be.

61

Innovative Opportunities and Dependencies In relation to resource mobilization, an interesting attempt to conceptualize the entrepreneurial process is made by Bhave (1994). In particular, Bhave considers the build-up of resources to be one of the key subprocesses involved in entrepreneurial ventures, implicitly highlighting the need to mobilize resources that make it possible to realize a perceived opportunity. What is not included in Bhave’s article is that resource mobilization depends on a system of actors as argued in this chapter. We, following a different line of research, would argue that it is not enough that the single entrepreneur is capable of perceiving an opportunity for innovation, but it is just as important to convey this picture of a desired future state to other actors involved. Of course, there are many means of conveying such a view, including persuasion or communicating as well as relying on individual initiatives. In one such discussion, Witt (1998) has pointed out the need to consider the conception of a ﬁrm before we can talk about its realization—and hence mobilization of resources—in a sensible manner. Furthermore, Witt argues that the entrepreneur must be able to impose this conception on co-workers in order to transform the individual entrepreneurs’ ideas into organizational routines, which can constitute key capabilities of the ﬁrm. Hence, there are dependencies between opportunities, actors, and elements because the mobilization of resources may be considered in relation to different perceptions of risks and opportunities—by diverse actors. For example, it may be difﬁcult to mobilize the necessary resources if the demand side is too unclear; not knowing about future customers implies that this will require a high tolerance of risks from upper level managers or potential investors. Depending upon the uncertainty related to the innovative opportunity, the ﬁrms could make different choices about where to mobilize ﬁnancial resources. Different types of investors may be possible to mobilize, ranging from banks with very high tolerance for uncertainty, but limited cost for the capital, to venture capitalists accepting that a lot of investments fail, making up for this by getting a high return on capital in case of a successful venture. Apart from customers and investors, also suppliers and collaboration partners have to be mobilized.10 This applies to the ﬁrm in relation to the economic system. This reasoning can be extended to an even more general level in terms of resource mobilization. Not only does the entrepreneur (or manager) in small (or large) ﬁrms have to make the innovation, or business conception, communicable and convincing to an extent that it is possible to attract and coordinate co-workers, but it is also necessary to convince a range of critical actors, namely investors, customers, and also others who in some way invest into related innovation processes. Note, however, that regardless of which monetary resources that are mobilized in the short run, there is still the question of mobilizing resources for sustained innovation. It is well known since the seminal article of Dierickx and

62

Innovative Opportunities and Dependencies Cool (1989) that resources available in a marketplace cannot be a source of sustained competitive advantage. Hence, these generic resources must be utilized in such a way that they allow for the build-up of more unique, scarce, and difﬁcult-to-imitate resources and capabilities. First, only after such distinctive capabilities have been acquired can the ﬁrm establish a competitive advantage. As this process takes time, the need for opportunistic adaptation to capture other possible proﬁts during the venture development process may be of signiﬁcant importance, as suggested by Bhide´ (2000). The comparison between 3G and i-mode is interesting here, particularly in relation to the suggestion that the means of resource mobilization on behalf of the content providers differed greatly. In this case, resource mobilization went beyond the boundaries of the existing ﬁrms per se or of new ﬁrm start-up— that are issues relevant to the literature on business venture and entrepreneurship. Resource mobilization has also meant bringing in new actors—through start-ups, technological diversiﬁcation, and market diversiﬁcation—that could develop different parts of the hardware, services, and infrastructure. Hence, the two illustrative case studies also suggest interesting questions about how the mobilization of resources is related to the level of diversity seen at the system level. Basically, in areas where there is a higher level of experimentation, a comparatively large amount of innovative opportunities ought to be identiﬁed, something that in turn often may lead to the mobilization of more resources. However, the reversed causality is probably of even more importance, namely that the level of experimentation is conditioned by the perception and actual amount of resources available. This has been seen in periods of economic overheating in different areas such as nickel exploration in the late 1960s and mobile communication from the mid-1990s up to the early 2000s. As large amounts of venture capital have been available the number of identiﬁed innovative opportunities have radically increased, normally also leading to increased investing in less fruitful ideas.

3.4.3

Perceived appropriability

The ﬁnal element of innovative opportunities is that of acting upon the perception to appropriate value. This is related to how the actors interpret the intellectual property regimes and other strategies or means to appropriate the beneﬁts of R&D (Levin et al. 1987). That is, do—and how do—actors create and access value which was created through R&D or other innovative search activities? The above discussions of ‘economic value’ and ‘mobilization of resources’ implies that before moving to innovate, the actors must identify that someone will beneﬁt economically from the innovation and that the needed resources can be obtained. However, in order to take the step to engage in experimentation, we would argue that it is also necessary that the parties involved

63

Innovative Opportunities and Dependencies perceive that there is a fair possibility for them to appropriate a large enough share of the value that already exists (and thus a value that a new innovation is to exploit) or is expected to emerge. However, dependencies are also visible here because the actors are acting under conditions of uncertainty. Therefore, the strategies and perceptions will not always be correct, as is easily seen by observing the early days of the Internet industry. Here, the difﬁculties for providers of products and services to appropriate the value created has mainly beneﬁted the end-users who today have great functionality at a low cost. Or, one could say that productivity has increased (as a result of better products/services) but those investing in the innovation processes may have difﬁculties in appropriating and therefore reducing additional investment into search activities. As a result of these difﬁculties, the willingness of existing or new ﬁrms to get involved in certain product segments has most certainly dropped, not least among investors but also among people who could potentially be entrepreneurs. This implies that the uneven distribution (over time and actors) of the perceived ability to appropriate returns to search activities may lead to different systemic outcomes. It is possibly a particular problem to analyse further when one examines distributed agents with partial knowledge who are innovating in areas of technologies that are likely to be complex, interrelated, and uncertain. The early development of technology and business platforms for mobile communication is an example of this. This reasoning implies, in turn, that agents will have diverse perceptions of their ability to appropriate economic value—and that will very likely inﬂuence their investment into search activities and choices of speciﬁc technologies. Furthermore, this will affect future development and the probable use of technology for business contexts. Simply put, much more knowledge is generated and experimented with than is actually put into use. For example, all the technical (engineering) knowledge that is intended to be useful for a business context is not utilized, nor is it successful in the marketplace. In these processes, the actors must assume value—and in doing so experiment based on these apparently idiosyncratic perceptions. Indeed, the matching of ideas and feedback is essential here, in order to understand how and why actors identify, act upon, and realize different elements within an innovative opportunity. The entrepreneurship literature strongly suggests that entrepreneurs have a tendency to overestimate the value of the business idea—whether technical or otherwise—that they are proposing to the market (see e.g. Ardichvili et al. 2003). Similarly, Nelson (1996) and Eliasson (2000) have argued that the capitalist system(s) is not ‘efﬁcient’ in a traditional economic sense, but these systems instead tend to involve much trial and error and experimentation of different ideas, goods, ﬁrms, and so on which may in the future prove to have an economic value.

64

Innovative Opportunities and Dependencies This implies that the actors involved in innovations, especially ones for as of yet non-existent goods and service products, are acting under conditions of uncertainty about the innovations’ potential economic value as well as whether they might be able to beneﬁt from the returns. This speciﬁc problem has received substantial attention in the last few years, as entrepreneurs and academics have started to address the creation and appropriation of value explicitly in terms of business models (see e.g. Chesbrough and Rosenbloom 2002).11 While the concept of business models is relatively new, the issues of value creation and appropriation are of course not. However, there has been a tendency to focus primarily, or exclusively, on either value creation (e.g. in marketing theory) or appropriation (e.g. in studies of intellectual property rights), and this piece-meal approach misses the dual challenge in innovation as an activity that by necessity deals with both exploration and exploitation (March 1991). The comparison between 3G and i-mode showed that appropriation may be a key issue in understanding how and why actors do things that affect these processes related to perception of economic value as well as resource mobilization. This has already been discussed in relation to revenue-sharing models where as a consequence of the i-mode solution of appropriation there were investments made by the content providers and revenue streams early on for the content providers. In 3G, different models of appropriation have been tested out, often as related to other standards like GSM, WAP, and SMS, though no single dominant model of how to handle the appropriation of value provided by for instance new services has been singled out and used, Arguably, the lack of a clear model makes it more difﬁcult to make investment decisions and thereby some innovation processes are delayed or come to an abrupt end as they run out of funding. Dependencies among opportunities and actors are particularly evident in these two cases. Since the entrepreneurship literature shows that entrepreneurs have a tendency to overestimate the value of the business idea that they are proposing to the market, one should further investigate empirically whether the two cases differ in terms of the involvement and expectations of small entrepreneurial start-ups—or whether opportunities primarily were perceived and acted upon by established ﬁrms. Dependencies in various parts of the innovation process may affect the outcomes. For example, such an overestimation by the individual entrepreneur may, in turn, depend on all other actors’ prediction of future markets as well as the relative ‘ease’ of mobilizing resources, e.g. resulting from an excess supply of venture capital at different points in time. Finally, the complex nature of innovation in many types of technologies, services, infrastructure, and consumer behaviour in these cases also suggests that distributed agents with limited information have difﬁculties with the appropriability of the ‘investment’ into search activities. Their own revenues

65

Innovative Opportunities and Dependencies may be dependent upon innovations by others. Hence on a more general note, it is clear that appropriation strategies of the ﬁrm—whether these be in the form of intellectual property rights, ‘normal’ property rights, secrecy, integration into organizational competencies, and so on differ. In the comparison of the empirical cases, a major ﬁnding is of course that in many ways, many actors did not have a clear view of how to appropriate value. In fact the period was characterized by a continuous search for new business models in order to try to ensure that economic value could be captured by the ﬁrm. For example, it was sometimes not even clear which actor was a seller and which a buyer—or who should sell to whom—during this era. For example, if two companies develop software programmes for the network, would the equipment supplier develop and sell its software package to a smaller specialized company or vice versa? In these cases, the business models and strategies for appropriation of the main equipment providers and operators was important for overall development, because their choices in turn affected the possible strategies of other small and large ﬁrms to appropriate. In the 3G case, the large companies declared basically ‘business as usual’ and the question was more whether to charge per minute or per download or not at all. In the i-mode case, DoCoMo set up and coordinated a system where the various innovators could appropriate, through a decided-upon share of the total content fee, based on a percentage. This is a clear example of when dependencies between actors linked both the type of opportunity and the element within the innovative opportunity. Hence, the cases demonstrate that the models of appropriation by one central actor may inﬂuence the innovative activities of different actors. It is not only an individual choice. The actor will need to take into consideration institutional contexts, the innovative activities of other actors as well as the various organizational strategies. This in turn suggests that the economists’ model of the ‘appropriability problem’ becomes somewhat more complex given that the incentive structure may depend on both their own strategies and those of many others. It might even suggest that the expected value and strategies to obtain returns from R&D investment may differ, even for ﬁrms competing within the same industry.

3.5

Conclusions

This chapter has focused on innovative opportunities and the dependencies that arise during the process when actors innovate in terms of identifying, acting upon, and realizing new combinations of resources and identiﬁed market needs within a larger economic system. The question of dependencies has been developed in connection with empirical and theoretical work. The

66

Innovative Opportunities and Dependencies chapter develops the concept of innovative opportunities further in terms of perception of economic value, perception of mobilization of resources, and perception of appropriation. Perception has been linked to actors who identify, act upon, and realize different types of innovations. The illustrative case of early development of technology and business platforms in mobile communication has been used, especially 3G and i-mode, to illustrate innovative opportunities and dependencies. A few points can be drawn out of this discussion of relevance for understanding whether economic transformation tends to occur under conditions of relatively more ﬂexibility or relatively more stability. First, it is clear that dependencies exist in such a way as to link the actor to systemic effects. The cases discussed in Section 3.3 demonstrate dependencies among different actors that result in ‘systemic’ effects, and these take the form of dependencies across different types of opportunities as well as dependencies across different elements of these processes. These involve different actors exploiting different types of opportunities—but also different activities to identify, act upon, and realize the three elements of innovative opportunities. These are often parallel processes, which are temporally interdependent in that actors change their perceptions of one affected by ‘outcomes’ in the same domain or any of the other. There are also dependencies between different types of opportunities and between different elements within the innovative opportunity. The discussion of the 3G and i-mode cases illustrates these dependencies in that the perception of economic value, resource mobilization, and appropriation inﬂuenced the activities of actors elsewhere in the system. This implies that changing the system requires coordination. Second, the claim is that the earlier deﬁned types of dependencies tend to affect the direction and shape of the further trajectory of innovations. Changes in one or several of the processes associated with innovative opportunities tend to shape the other processes, or more precisely, the activities of actors engaged in other parts of the processes. This is linked to ‘subjective’ perceptions as well as ‘objective’ knowledge. That is, these changes in perception (presumably often but not always) come from some real-world feedback. Some new activities are initiated while existing activities may be terminated, signiﬁcantly altered, or they may keep occurring within a new context. Most importantly, different actors are ‘connected’ through the interdependencies in that perceptions change because of ‘outcomes’ over time. This type of linkage may only become clear over time, when the relationships between the elements of innovative opportunities change as organizations reconsider their strategies, relationships change, and so on. For example, some business models used by organizations may include perceptions both of economic value and of appropriation and as such they are actively a means to convince investors or complementary actors that they should invest in the company or the industry at large. Hence, in a fundamental way, the systemic aspects of

67

Innovative Opportunities and Dependencies economic transformation depend upon the unfolding of innovative opportunities over time in that complementary actors may alter their behaviour based on the dependencies. This implies that dependencies will likely inﬂuence whether innovative opportunities tend to lead to an outcome that is more ‘ﬂexible’ or more ‘inert’ and ‘resistant’ to the changes. Third, this suggests the need to reconsider division of innovative labour within the economy. Speciﬁcally, parts of this goes back to the Smithian division of labour, where different actors specialize in different types of activities and skills. Now, it is well known that economic progress critically depends upon the progression of the division of labour across industries (Young 1928). We argue that the dependencies of the three processes of innovative opportunities may constrain or reorient this division of labour. For example, even if there is a perception of a large future economic value, if the perception of the means of appropriation is unfavourable it may of course mean that individual actors may decide not to act upon this opportunity. More generally, this may also constrain the division of labour in that other complementary actors, either in terms of ﬁnance, consulting services, and so on do not ﬁnd investments in time or labour worthwhile to develop an industry. Some industries seem much more dependent upon a broad scope of innovations, as illustrated by the mobile telecommunication case, where there was a need to change both in terms of hardware, services, infrastructure, and consumer behaviour. This implies that ﬂexibility may sometimes come through the entrance of new actors and exit of old actors, and sometimes through the renewal of existing ﬁrms. Finally, these points suggest the need to further investigate the extent to which certain ways of organizing and beneﬁting from innovations are more ‘durable’ than others. If actors feel there is little uncertainty about where to make a proﬁt, then it is reasonable to assume that more actors will be aware of it, and act on it. This would increase competition and reduce the possibilities to make a proﬁt for any given agent. Moreover, since the simplest type of price differences making up arbitrage proﬁts are easily observable, they should not allow for any long-lasting proﬁts. Probably more long-lasting sources of proﬁts have to be less imitable. Unless there are appropriative speciﬁcities that protect these sources, such as legislation, this means that the sources for many emerging areas need to be more complex. The more ‘durable’ ways of beneﬁting from innovations may come in more complex technical domains. More complex innovative opportunities are for example the ones involving development of goods, services, and infrastructures, and the ones involving the parallel perceptions of actors on the supply side and demand side, respectively. The more complex innovative opportunities are likely to be more difﬁcult to perceive, especially since most agents will only appropriate a small part of the potential returns. The question then becomes how and where any individual ﬁrm can innovate—and to what

68

Innovative Opportunities and Dependencies extent they do so through other more virtual organizational forms such as networks. This implies that some ﬁrms—and some organizational structures— may be inherently more ﬂexible or more stable, depending upon the particular constellation of dependencies at any given point.

Notes 1. This work has been carried out at Chalmers University of Technology and at the Australian National University, in a grant administered by Chalmers and in the grant ‘Flexibility and Stability’, given by the Rueben Rausing Foundation and administered through IMIT. We thank all the authors in the book, Flexibility and Stability in the Innovating Economy, for comments on previous versions. Also, special thanks to Erik Bohlin and Sven Lindmark, for teaching us through your detailed studies of telecommunication. 2. The proposed concept of ‘innovation opportunities’ is found in more detail in Holme´n et al. (2004). 3. While it may be that ‘innovative opportunities’ could provide an argument that can be applied to many human endeavours, we wish to stress that this chapter will primarily consider innovations related to technological developments for a business context. 4. This characterization should not be confused with the view of technology as a production function. Albeit we have an admittedly ‘narrow’ focus in the broad deﬁnition of ‘innovations’, technological change is one type of change directly linked to industrial dynamics, productivity, and relevant skills among members of the labour market. 5. The aim of many economists could be seen, for example, to explain changes in the economy in terms of how the preferences, incentives, and institutional structure lead to speciﬁc choices for a representative agent and thereby move the system in a new direction. In contrast to the view presented in the text, there is a strong assumption of the ‘representative agent’ where all agents in a category will act as predicted. 6. This chapter is based on the premise that understanding the question of how and why ﬁrms develop different aspects of innovative opportunities must be explored both conceptually and through illustrative cases. This not only enables us to compare existing theoretical ideas with details from speciﬁc cases but also to demonstrate that paradigmatic statements are visible in the empirical material. Hence, in this chapter, the illustrative cases from early development of technology and business platforms in mobile communication, especially 3G and i-mode, are used to help analyse how actors identify, act upon, and realize new combinations of resources and market needs, in order to try to beneﬁt from and realize a perceived economic potential of new technology. Innovative opportunities can be examined empirically in terms of how, why, and whether or not the actors develop different types of innovations. 7. Given the fact that these existing case studies have been analysed within different frameworks, and also may not have covered the necessary types of data to appropriately cover innovation opportunities, we stress that the interpretation given here should be mainly seen as illustrative, and perhaps even speculative, about a possible interpretation of complex historical processes.

69

Innovative Opportunities and Dependencies 8. These statements are based on a variety of sources, ranging from trade industry data to scientiﬁc reports. In addition to the references in note 9, academic references include: Fransman 1995; Lemola and Palmgren 1998; McKelvey, Texier, and Alan 1998; Kraft 2000; McKelvey and Texier 2000; Holme´n 2001; Fransman 2002; Edquist 2003; Palmberg 2004. ¨ rkdahl 2004; Lindmark et al. 9. The cases are primarily based upon Lindmark 2003; Bjo 2004a, 2004b; Lindmark 2005. Additional references include Funk 2001; Ka¨rrberg and Marnung 2001; Lindmark 2002; Bohlin et al. 2004; McKelvey and Bohlin 2005a, 2005b. Note that 3G is mainly a set of technical standards whereas i-mode is a platform standard for services, but for clarity of argumentation, we simply call them two standards, or technology and business platforms. 10. Although not addressed here, one key point is of course to convince potential employees that they should contribute to the realization of the opportunity. 11. As mentioned by Chesbrough and Rosenbloom (2002), academics have been rather slow in picking up the business model discourse. However, even in practice, the term has only been explicitly used since the emergence of Internet-based business, as this demarcated a shift in the way ﬁrms could create and appropriate value.

References Alvarez, S. A. and Busenitz, L. W. (2001). ‘The Entrepreneurship of Resource-based Theory’, Journal of Management, 27: 755–75. Ardichvili, A., Cardozo, R., and Sourav, R. (2003). ‘A Theory of Entrepreneurial Opportunity Identiﬁcation and Development’, Journal of Business Venturing, 18(1): 105–23. Bhave, M. P. (1994). ‘A Process Model of Entrepreneurial Venture Creation’, Journal of Business Venturing, 9(3): 223–42. Bhide´, A. V. (2000). The Origin and Evolution of New Business. New York: Oxford University Press. ¨ rkdahl, J. (2004). ‘Imagination, Variation and Firm Behaviour: Observations From Bjo 3G’, Licentiate thesis. Gothenburg, Sweden: Department of Innovation Engineering and Management, Chalmers University of Technology. —— and Bohlin, E. (2002) ‘Financial Analysis of the Swedish 3G Licensees: Where are the Proﬁts?’, Info, 4: 10–16 ¨ rkdahl, J., Weber, A., Wingert, B., and Ballon, P. (2004). ‘The Bohlin, E., Lindmark, S., Bjo Future of Mobile Technologies in EU: Assessing 4G Developments’, IPTS Technical Report prepared for the European Commission – Joint Research Center, EUR 21192 EN, Seville; http://ﬁste.jrc.es/ Breschi, S., Malerba, F., and Orsenigo L. (2000). ‘Technological Regimes and Schumpeterian Patterns of Innovation’, The Economic Journal, 110: 388–410. Chesbrough, H. and Rosenbloom, R. (2002). ‘The Role of the Business Model in Capturing Value From Innovation: Evidence From Xerox Corporation’s Technology Spin-Off Companies’, Industrial and Corporate Change, 11(3): 529–55. David, P. (2004). ‘Understanding the Emergence of ‘‘Open Science’’ Institutions: Functionalist Economics in Historical Context’, Industrial and Corporate Change, 13(4):571–89. Dierickx I. and Cool K. (1989). ‘Asset Stock Accumulation and Sustainability of Competitive Advantage’, Management Science, 35: 1504–11.

70

Innovative Opportunities and Dependencies Edquist, C. (ed.) (2003). The Internet and Mobile Telecommunication Systems of Innovation: Developments in Equipment, Access and Content. Cheltenham, UK: Edward Elgar. Edquist, C. and McKelvey, M. (2000). Systems of Innovation: Growth, Competitiveness and Employment, a 2 vol. reference collection. Cheltenham: Edward Elgar. ——, Hommen, L., and McKelvey, M. (2001). Innovation and Employment: Process Versus Product Innovation. Cheltenham, UK: Edward Elgar. Eliasson, G. (2000). ‘Industrial Policy, Competence Blocs, and the Role of Science in Economic Development’, Journal of Evolutionary Economics, 10: 217–41. Fagerberg, J., Mowery, D., and Nelson, R. R. (2004). The Oxford Handbook of Innovation. Oxford: Oxford University Press. Fransman, M. (1995). Japan’s Computer and Communication Industry. Oxford: Oxford University Press. —— (2002). Telecoms in the Internet Age: From Boom to Bust to? Oxford: Oxford University Press. Funk, J. (2001) The Mobile Internet: How Japan Dialed Up and the West Disconnected. Hong Kong: ISI Publications. Holme´n, M. (2001). Emergence of Regional Actor Systems: Generic Technologies and the Search for Useful or Saleable Applications, PhD thesis. Gothenburg, Sweden: Department of Industrial Dynamics —— , Magnusson, M., and McKelvey, M. (2004). ‘What are Innovative Opportunities?’, Article submitted for review. Earlier version presented at the International Joseph A. Schumpeter Association conference, Milano. —— , —— , and —— (2005). ‘Innovative Opportunities: What Do We Learn from Mobile Communication?’, Chapter presented at DRUID conference, 10th Annual conference, June, Copenhagen, www.druid.dk Hughes, T. (1983). Networks of Power: Electriﬁcation in Western Society 1880–1930. Baltimore, MD: Johns Hopkins University Press. Ka¨rrberg, P. and Marnung, E. (2001) The Mobile Internet Industry in Tokyo: A Study of an ¨ ping University, Institute of Entertainment Cluster, Master thesis, No. 2001: 51, Linko Technology, Department of Management and Economics. Kirzner, I. (1997) ‘Entrepreneurial Discovery and the Competitive Market Process: An Austrian Approach’, Journal of Economic Literature, 35(1): 60–85. Kor, Y. Y. and Mahoney, J. T. (2000). ‘Penrose’s Resource-Based Approach: The Process and Product of Research Creativity’, Journal of Management Studies, 37(1):109–39. Kraft, J. (ed.) (2000). The Processes of Competition. Cheltenham, UK: Edward Elgar. Lemola, T. and Palmgren, C. (1998) ‘Case Study of Nokia’. Report within the EU project ‘ISE: Innovation Systems and Europe, Contract No. SOE1-CT95–1004, Brussels: European Union Commission. Leonard-Barton, D. (1992). ‘Core Capabilities and Core Rigidities: A Paradox in Managing New Product Development’, Strategic Management Journal, 13: 111–25. Lindmark, S. (2002). Evolution of Techno-Economic Systems: An Investigation of the History of Mobile Communications, PhD thesis. Department of Industrial Management and Economics, Chalmers University of Technology —— (2003). ‘The i-Mode Success Story: Towards a System Explanation’, Communications & Strategies, Issue 52, Special Issue ICT Market Dynamics and Development Models in Asia, 4th quarter.

71

Innovative Opportunities and Dependencies Lindmark, S. (2005). Coordinating the Early Commercialization of General Purpose Technologies: The Case of Mobile Data Communications. Special issue ‘Biotechnology and Telecommunications Innovation: Conditions and Processes for Emerging Technologies’, Innovation: Management, Policy & Practice, 5(1): 39–60. ——, Bohlin, E., and Andersson, E. (2004a). ‘Japan’s Mobile Internet Success Story: Facts, Myths and Implications’, info, 6(6). ——, Andersson,E.,Johansson,M.,andBohlin,E.(2004b).‘TelecomDynamics:HistoryandState of the Swedish Telecom Sector 1970–2003’, VINNOVA Analysis VA 2004: 04; www.vinnova.se. McKelvey, M. and Texier, F. (2000). ’Surviving Technological Discontinuities Through Evolutionary Systems of Innovation: Ericsson and Mobile Telecommunication’, in P. P. Saviotti and B. Norboom, (eds.), Technology and Knowledge: From the Firm to Innovation Systems. Cheltenham, UK: Edward Elgar ——, ——, and Alm, H. (1998). ‘The Dynamics of High Tech Industry: Swedish Firms Developing Mobile Telecommunication Systems’. Report submitted within the EU project ‘ISE: Innovation Systems and Europe, Contract No. SOE1-CT95–1004, Brussels: European union commission. —— and Bohlin E. (eds.) (2005a). ‘Introduction’, Special issue ‘Biotechnology and Telecommunications Innovation: Conditions and Processes for Emerging Technologies’, Innovation: Management, Policy & Practice, 5(1): 1–6. —— and —— (eds.) (2005b). ‘Conclusion’, Special issue ‘Biotechnology and Telecommunications Innovation: Conditions and Processes for Emerging Technologies’, Innovation: Management, Policy & Practice, 5(1): 96–103. March, J. (1991). ‘Exploration and Exploitation in Organizational Learning’, Organization Science, 2(1): 71–87. Nelson, R. (1995). ‘Co-evolution of Industry Structure, Technology and Supporting Institutions, and the Making of Competitive Advantage’, International Journal of the Economics of Business, 2(2): 171–84 —— (1996). The Sources of Economic Growth. Cambridge, MA: Harvard University Press. Palmberg, C. (2004). ‘The Sources of Innovation: Looking Beyond Technological Opportunities’, Economics of Innovation and New Technology, 13(2): 183–97. Penrose, E. (1959). The Theory of the Growth of the Firm. New York: John Wiley & Sons. Prahalad, C. K. and Hamel, G. (1990). ‘The Core Competence of the Corporation’, Harvard Business Review, (May–June): 79–91. Rosenberg, N. (1982). Inside the Black Box. Cambridge: Cambridge University Press. —— and Birdzell, L. E. (1986). How the West Grew Rich. New York: Basic Books. Sanz-Velasco, S. A. and Magnusson, M. G. (2004). ‘Opportunity Development in a Knowledge-intensive Venture: A Case Study’, Journal of Small Business and Entrepreneurship, 17(4): 277–92. Scherer, F. M. (1965). ‘Firm Size, Market Structure, Opportunity, and the Output of Patented Inventions’, American Economic Review, LV, (December), 1097–125. Schumpeter, J. A. (1934). The Theory of Economic Development: An Inquiry into Proﬁts, Capital, Credit, Interest, and the Business Cycle. Cambridge, MA: Harvard University Press. Shane, S. (2000). ‘Prior Knowledge and the Discovery of Entrepreneurial Opportunities’, Organization Science, 11(4): 448–69. Witt, U. (1998). ‘Imagination and Leadership: The Neglected Dimension of an Evolutionary Theory of the Firm’, Journal of Economic Behavior and Organization, 35: 161–77.

72

4 The Great Experiment: Public–Private Partnerships and Innovation in Design, Production, and Operation of Capital Goods in the UK Andrew Davies and Ammon Salter

4.1 Introduction Public–private partnerships (PPPs) are becoming the key mechanism for the design, production, and operation of capital goods in the public sector in the UK and elsewhere. PPP is, however, not a new phenomenon. Long-term service contracts between the public and private sectors have been a common tool for governments looking for opportunities to lower risk and shift responsibility for the production and operation of ﬁxed capital goods to the private sector. Yet the scale of use of PPP in the UK marks a departure from previous methods of government procurement. It represents a transformation in the way UK government procures, builds, and operates its ﬁxed capital assets. This ‘great experiment’ in public policy is changing the nature of the state and realigning its relationship with the private sector. PPP is part of a larger shift in the role and nature of the state in the UK and elsewhere, beginning in the early 1980s with the deregulation, liberalization, and privatization of state-controlled industries such as telecommunications, airlines, and electricity. It represents a retreat from the traditional role of the state across a broad range of sectors as both the provider of public services (such as health care and education) and the chief funder, procurer, and operator of the ﬁxed capital goods that underlie the delivery of these services. In the traditional welfare state, the private sector was responsible for design and construction of ﬁxed capital goods, but it played little direct role in their operations or ﬁnancing. Under PPP, the private sector takes on ﬁnancial debt associated with the cost of the design and construction of ﬁxed capital goods

73

The Great Experiment and the costs of maintaining these assets over a long period (usually 25–30 years). In return, private ﬁrms are paid a yearly service payment for the ﬁxed capital asset, usually tied to a highly formalized service-level agreement. In some areas, the government sets the guidelines or performance outcomes from the service and has little or no direct responsibility for the services themselves. In other areas, the government retains responsibly for the delivery of services such as health and education and the private sector provides auxiliary services to support these core activities. PPPs are highly controversial in the UK. Arguments in favour of PPP often stress the beneﬁts of these arrangements for easing the ﬁnancial, economic, and social burden imposed on the government for the development and operation of ﬁxed capital goods. It shifts ﬁnancial risk and debt to the private sector, freeing up scarce government resources to be invested in critical services. It may also help lower the costs of running and maintaining ﬁxed capital assets through performance improvements in auxiliary services or simply to lower the cost and amount of labour required to build, maintain, and operate ﬁxed capital assets. Despite the growth of PPPs, there has been little independent empirical research on the impact of these new arrangements on the innovation process in capital goods. As an innovation in public procurement policy, the adoption of PPP has created new opportunities and challenges for private sector involvement in public projects. However, there are no tried or tested solutions for PPP provision, which require ﬂexible and creative new approaches from the private sector. Firms continue to experiment with new strategies, to alter their positions in the industry value stream and to explore new ways of delivering PPP contracts. In order to gain a better understanding of these new arrangements, we examine the policy environment in the UK that supported the development of PPP. Although there are numerous studies that have explored whether PPPs offer ‘good value for money’ in comparison with projects procured under traditional contractual arrangements, little empirical research has assessed the ability of PPPs to act as a stimulus for innovation—to discover to what extent the new arrangements create a virtuous circle across the different stages of the design, production, and operation of capital goods. To improve our understanding of this issue, we explore the organizational arrangements that PPP providers have adopted in response to new government procurement strategies. In particular, we focus on potential for new forms of systems integration to be created through PPP and examine the implications of these changes for government and industry. As well as successfully designing and building complex products and systems, ﬁrms have to develop new capabilities to meet the challenges of operating, maintaining, and ﬁnancing the product during its life cycle. The chapter is organized as follows. Section 4.2 examines the empirical and policy background to the use of PPPs, reviewing UK government policy in this

74

The Great Experiment area since the mid-1990s. In doing so, we seek to explore the central propositions made about innovation in policy documents. We then examine some of the current literature on PPPs. In Section 4.3 we focus on the changing division of labour in capital goods industries, exploring the impact of this new division of labour for the innovation process. Section 4.4 presents the discussion and conclusions.

4.2 Empirical and theoretical background Private sector involvement in the design, construction, and operation of capital goods has an extensive history. For example, the Perrier brothers were given a licence to supply water to Paris for 15 years in 1782 (a contract that was later rescinded by the Revolutionary government) (Leiringer 2003). Large public projects in the nineteenth century, such as the Suez Canal and US railroads, were also designed, constructed, ﬁnanced, and operated by the private sector. However, Leiringer and Michaud have argued that the ﬁrst modern and signiﬁcant use of a PPP arrangement was by the Turkish government for the construction of nuclear power projects in the early 1980s (Leiringer 2003). In this case, the use of PPP was driven by strong ﬁnancial constraints imposed on the Turkish government by the International Monetary Fund (Michaud 2001). PPP offered a mechanism to transfer the risk and debt associated with these new capital projects to the private sector in return for a long-term service contract. The use of PPP as a central element in the procurement of design, production, and operation of public goods has grown tremendously in the late 1990s and early 2000s in the UK (Osborne 2000). PPPs now account for close to 15 per cent of all UK government capital procurement, and contracts with a cash value over £35.6 billion have been signed between the UK government and the private sector (HM Treasury 2003, 2004; Zitron 2004). Several factors have inﬂuenced the growth of PPP. First, when the Labour government was elected in May 1997, it had committed itself to the previous government’s spending plans for the ﬁrst two years of its administration and it also adopted a set of ‘golden rules’ for public ﬁnance (Allen 2001; Zitron 2004). At the same time, the government had set out a major programme of investment in public services, especially education and health care. The idea behind the expansion of PPP was partly to shift the burden of upfront costs for the creation of new ﬁxed capital assets from the public to the private sector, ensuring greater levels of investment in public assets than could be provided by the government operating alone (Akintoye et al. 2003; Zitron 2004). At the heart of this movement was realignment of the Labour government’s view of private ownership and the role of the state in the economy. The Blair government was deeply pragmatic about the role of the public and private sectors, and claimed

75

The Great Experiment to be unburdened by past arguments about the role of the state in the economy. Blair himself argued that state ownership of public goods does not matter—‘it is not matter of public or private, it is just what works’. To quote from the Chief Secretary to the Treasury: However, to create this new partnership approach we needed a fundamental shift of thinking, putting behind us the ideology and dogma of the past. In the modern world, governments are judged not by what they own, or on how much they spend, but on whether they deliver. In government, therefore, our focus in all that we do is on outcomes rather than inputs. In places of the debates between left and right, nationalisation versus privatisation, our goal is the modernisation of Britain (Smith in HM Treasury 2000).

A second factor in shaping the movement towards PPP was a belief that government itself had a poor record for delivery and operation of large capital projects. Many government-led projects had overrun and/or been signiﬁcantly over budget. For example, the Jubilee Line extension on the London Underground was over budget and was completed several years behind schedule. At the same time, the Docklands Light Rail system, funded by a combination of public and private funds, was on time and under budget. The failure of the Jubilee Line and other projects helped undermine the government’s conﬁdence in its ability to deliver large projects and to procure its own capital goods (HM Treasury 2000). A third factor in explaining the shift towards PPP was the movement towards market-based solutions for the operation and management of the public sector. Jessop called this movement the shift away from the Keynesian Welfare State (KWS) to the Schumpeterian Workfare State (SWS), representing a retreat of government from provision of social and public goods and reﬂecting a greater reliance on private management and market relations (Jessop 1994, 2002). In part, this shift is reﬂected in the widespread use of privatization and commercialization of publicly owned corporations in the 1980s and 1990s across the OECD as well as the contracting out of government services to external agencies. This movement had strong political support in the Thatcher and Reagan governments in the UK and the USA in the early 1980s, but was followed by many governments around the Organization for Economic Cooperation and Development (OECD) often regardless of their nominal political afﬁliation (Osborne 2000; Akintoye et al. 2003). Another source of inﬂuence in this movement towards the SWS was Osborne and Gaebler’s Reinventing Government that outlined a range of opportunities for the public sector to harness the capability of the private sector to improve service delivery and performance (Osborne and Gaebler 1993; Hood 2000). The UK is not alone in using PPP to support the development of public goods. The Netherlands also has a long tradition of PPP arrangements, and long-term facility management contracts are common throughout Europe. In

76

The Great Experiment the USA, large contractors such as Bechtel, have been involved in ‘design, build, and operate’ contracts with the Department of Defence and the Department of Energy (Morris 1994). Although PPP was ﬁrst developed in the USA in the early 1980s as an alternative way of delivering state government services (Osborne and Gaebler 1993: 17), it could be argued that in the UK the scale of PPP activity is greater than in other countries. In fact, the UK government has been attempting to export its brand of PPP to the rest of the world. The former Chief Secretary to the Treasury, Andrew Smith, argued that ‘it is an area of public policy where the UK leads the world’ and that PPPs have generated ‘huge international interest’ (Smith in HM Treasury 2000). What are PPPs? The ofﬁcial UK government deﬁnition includes a wide range of different types of partnerships between public and private sector. PPP may involve: 1. The introduction of private sector ownership in state-owned businesses. 2. The Private Finance Initiatives (PFI) where the public sector contracts to purchase services on a long-term basis. 3. Selling government services to ‘wider markets’ where private sector used to exploit the commercial potential of government assets (HM Treasury 2000: 10). In this paper, we will focus on PFI as an example of PPPs because they are the most commonly used type of PPP in the UK. It must be noted that the differences between PPP and PFI are rarely clear (Zitron 2004). In government publications, the terms are often used interchangeably (HM Treasury 2003). We are interested in public–private arrangements whereby the private sector designs, builds, and operates a public capital good over a long-term period for a yearly fee based on a service-level agreement. The policy rationale for PPP is based on efﬁciency and the innovative capabilities of the private sector over the public sector. The UK government argues that private sector ﬁrms are under greater competitive pressure and therefore they can develop innovative approaches to delivering public services and managing state-owned assets. PPPs seek to ‘harness the innovation and discipline of the private sector’ and enable government to ‘enhance the services it offers to its customers’. Overall, it argues that the private sector is ‘more skilled at . . . managing complex investment projects to time and budget’ (HM Treasury 2000: 10–20). Two different types of innovative beneﬁts are described by the UK government that arise from the use of PPP. First, PPPs are seen to create opportunities for innovation in the design, development, and operation of ﬁxed capital assets. In theory, they allow private actors to take on the whole life cycle of a capital good, creating a virtuous circle of learning between different stages of the capital goods innovation process. This beneﬁt could be called the ‘life cycle

77

The Great Experiment beneﬁt’. Second, PPPs call upon the ‘greater expertise’ of the private sector in the procurement and operation of capital goods, ensuring a high probability that the design and construction of the capital goods is ﬁnished on time and on budget. They are an attempt to introduce ‘private sector skills and expertise’ into the construction and operations of the ﬁxed capital assets of government (HM Treasury 2000). This beneﬁt could be called the ‘expertise beneﬁt’. At present, PPPs represent close to 14 per cent of the UK government’s capital procurement budget for 2004–05 (Zitron 2004). Hence the vast majority of government projects are still procured by non-PPP contracts. However, the growth of PPP has been dramatic in that since the late 1990s and early 2000s this growth has been severely constrained by lack of skills within local councils and government departments to manage the PPP procurement process (HM Treasury 2003). Once these skills are enhanced, it is envisioned that PPPs will become more widespread. PPPs have not been limited to a particular sector. PPPs have become common in prisons, schools, hospitals, transportation, and defense. Since the early 1990s, the UK government has signed 556 PPP projects with a total capital value of £35.6 billion (HM Treasury 2004). The largest PPP contracts were signed in the transportation sector. In total, there have been 37 projects in transport with a capital value of over £20.5 billion. This basket of projects includes the three largest contracts for the modernization and maintenance of the London Underground. The most active area for PPPs has been the health sector, with 117 contracts signed with a capital value of £3.2 billion between 1992 and 2003. Education is the second most active sector, with 96 signed projects with total capital value of £2.0 billion. There have also been 46 defense projects with a capital value of £2.5 billion. The number and economic value of PPPs increased dramatically in the late 1990s. In 1997, there were 61 PPP capital projects signed, with a total capital value of £2.3 billion. By 2003, there were only 17 projects but these had a capital value of £11.6 billion. The average size of a PPP project was £63.6 million (HM Treasury 2004). However, as PPPs have developed, the average project size has been increasing signiﬁcantly, indicating a shift to fewer and larger projects, such as the London Underground PPP. With the growth of PPPs, new research has attempted to assess the beneﬁts and costs of these arrangements (Glaister 1999; Osborne 2000; Allen 2001; IPPR 2001; Wakeford and Valentine 2001; Grimsey and Lewis 2002; Akintoye et al. 2003; Grimshaw et al. 2002). The UK’s National Audit Ofﬁce has been a leader in conducting this research, publishing a range of studies on different aspects of PPPs (NAO 2001, 2002, 2003a, b, c, d). These studies show that PPP projects have considerable beneﬁts to the design and construction process. For example, the NAO estimated that only 22 per cent of PPP projects were delivered late and 24 per cent delivered over budget in 2000. In comparison,

78

The Great Experiment in 1999, 73 per cent of traditionally procured projects were late and 70 per cent were delivered over budget (NAO 2003a). These stark differences in performance indicate substantial beneﬁts to the PPP mechanism of procurement, but the NAO is careful not to impute a direct effect of PPP on project performance. The NAO suggests that ‘it is not possible to judge whether these projects could have achieved these results using a different procurement route’ (NAO 2003a: 17). In a survey of construction managers and authorities working on the ﬁrst generation of PPP projects, the NAO found that use of PPPs has had little impact on the adoption of use of innovations in the delivery of construction projects. A considerable percentage of construction managers indicated that there was scope for innovation in PPP projects, but that their government clients tended to be risk adverse and made very conservative design choices (NAO 2001). Only a third of authorities surveyed indicated that the providers were demonstrating innovative solutions. In a signiﬁcant minority of projects, there was little or no innovation at all in both the design and construction and operation phases of the projects (NAO 2001: 26). However, this sample reﬂects a particular time period and makes no attempt to compare the innovativeness of PPPs with other projects. Parker and Hartley argue that in the case of the PPPs, it is almost impossible to measure the beneﬁts or costs of the projects in the short term, given the use of 30-year service contracts (Parker and Hartley 2003). In this respect, the beneﬁts of PPP arrangements are in the eye of the beholder. Hartley also suggests that there are signiﬁcant problems in writing long-term service supply contracts (Hartley 2002). Drawing on transaction costs economics (Williamson 1985), he argues that it is both very expensive and time consuming to fully specify these types of contracts and since it is difﬁcult to know what the future holds, many of these contracts will be not able to fully specify future requirements. As Hartley points out, 30 years can be a long time in the life of capital goods. The difﬁculty in specifying the contracts might lead ﬁrms to cut corners on those parts of the contract that are incompletely speciﬁed. This may have signiﬁcant implications for the performance of the system in the medium or long term (Hartley 2002). A recent study by the Major Project Association (MPA), based on interviews with leading actors in the PPP process, attempts to assess whether ‘PPPs are working’ (Zitron 2004). Overall, the report indicates that the PPPs are yielding beneﬁts for government, providers, and users. However, the report indicates that the evidential base for PPP remains very thin (Zitron 2004: 69). In terms of innovation, Zitron found that ﬁnancial pressures on PPP projects constrained opportunities for innovation. Funders and corporate decisionmakers ‘want tried, tested and therefore low risk, solutions’ (Zitron 2004: 47). One interview suggests that ‘high exposure to innovation and complex construction . . . could lead to non-investment-grade underlying ratings’

79

The Great Experiment (Zitron 2004: 47). The open nature of bidding between companies also means that many ﬁrms are unwilling to broadcast any new innovations to their competitors. Jeremy Wagge of the Babtie Group notes that ‘innovation, using new materials not tried and tested, is seen as very dangerous’ (Zitron 2004: 47). In sum, the MPA report suggests that ﬁnancial pressures, open bidding, and resistance by government authorities are helping to hinder the use of innovation approaches in PPPs. In support of the conclusions of the MPA study, an in-depth case study of four PPP construction projects in the UK and Sweden found that innovation was indeed possible on PPP projects but that adopting such innovations was no easy task (Leiringer 2003). In fact, the challenges of getting innovation adopted on PPP projects are similar to the challenges of getting innovations adopted on other types of projects procured by different means. Leiringer argues that there is no essential difference in PPP projects and other types of projects that might lead to more innovative outcomes. He states ‘it would be a rash conclusion to claim that PPP is a panacea for change within the construction industry’ (Leiringer 2003: 184). No evidence surfaced in the case studies that the long-term nature of the contract generated more innovation. In fact, in all of the projects, there was a tendency for providers to use ‘tried and tested’ solutions in order to lower their risk exposure. In this respect, PPPs are likely to ‘provide solutions that adhere to best practice and available knowledge and expertise rather than something new or unique’ (Leiringer 2003: 184). A difﬁculty faced in the NAO studies and other studies is small and selective samples and a strong selection bias in the research method. There are a number of pitfalls here. First, there is the potential of a sample bias because many projects follow the PPP procurement route as they may be less complex and therefore contracts are easier to specify. NAO studies in this respect may reﬂect a sample bias and comparing ‘apples with oranges’. Second, many of the studies are based on surveys or interviews of ﬁrms that have direct ﬁnancial interest in the industry. There are few external benchmarks. Third, there are no large-scale samples of the performance of PPPs versus traditional procurement or other forms of procurement. Given the fact that many factors might inﬂuence the success and failure of a project, it will be a difﬁcult task to disentangle the effect of the procurement process on the performance of the project. Several factors that authors highlight in shaping the success of a PPP project, such as trust among project members, are themselves extremely common across a range of different capital good projects (Morris 1994; NAO 2003c). To date, little research has focused on the impact of PPPs on the ﬁrms who actually design, build, and operate the capital goods. The use of PPPs has created a new industry of PPP providers. Many of these ﬁrms are traditional construction ﬁrms, but they also include engineering and design organizations, architects, ﬁnancial companies, service and outsourcing companies, and train manufacturers. Little is known about how these organizations have

80

The Great Experiment changed their activities as a result of the new procurement mechanism, yet it is any such changes within and among these ﬁrms that will enable them to experience the main beneﬁts of PPP for innovation. Accordingly, greater research is required on the impact of PPPs on the innovation process inside capital goods providers.

4.3 Innovation in capital goods and repositioning for PPPs It has long been recognized that the design, production, and operation of capital goods differs from traditional consumer goods manufacturing. In studies of complex products and systems (CoPS), authors have highlighted the high-cost, engineering-intensive nature of some capital goods (Miller et al. 1995; Davies 1996; Hobday 1998; Hobday and Rush 1999; Hobday et al. 2000; Davies et al. 2001; Acha et al. 2004). In this context, the term ‘complex’ is used to denote the high number of customized components, the breadth of the knowledge and skills required, and the extent of new knowledge involved in development and production. Given this product deﬁnition, the production process for CoPS can be viewed within the classic scheme provided by Woodward (1965) that describes and contrasts project, small batch, mass production, and continuous process production. Within this scheme CoPS are the high-cost, high-technology products made in projects and small batches. Typically, CoPS have very thin markets, often comprising a single client. Examples include mobile phone base stations, aircraft engines, software packages, chemical plants, intelligent buildings, ﬂight simulators, satellite systems, rail transit systems and so on. CoPS production often involves signiﬁcant client involvement in the design and speciﬁcation of the ﬁnal product. CoPS are produced by a wide variety of different types of organizations. Project-based organizations are common in construction and in the production of small batch or bespoke products (Woodward 1965). Many CoPS producers are vertically integrated. They typically concentrate on the hightechnology portion of CoPS, such as aero-engine manufacturers and train equipment manufacturers. They have extensive capabilities in design and production of capital goods but have, until recently, rarely taken on responsibility for operation of the capital goods themselves. In contrast, in construction, there are many ﬁrms which act as system assemblers, bringing together a range of components and systems and managing a wide range of specialist inputs from the different organizations which come together to complete a speciﬁc project. These ﬁrms have also been highly active in the PPP market and have shifted to downstream into operational services. There are a number of speciﬁc challenges faced by ﬁrms designing and producing CoPS. One central challenge is to integrate different subsystems

81

The Great Experiment and components. CoPS involve combining a vast array of different components, systems and technologies from a variety of industries. Many CoPS ﬁrms have found that in order to integrate, they need to retain a deep knowledge of a range of component technologies (Prencipe, Davies, and Hobday 2003; Prencipe 2000; Brusoni, Prencipe, and Pavitt 2001). A second problem relates to the integration of design, production, and operation of the product itself. CoPS ﬁrms often ﬁnd it extremely beneﬁcial to learn from the operation of their ﬁxed capital goods in order to reﬁne subsequent design and production. In traditional models of CoPS contractual relationships, ﬁnal capital goods were simply handed over to operators. In this approach, producers are denied opportunities to learn from the way their products are used. In some CoPS industries, such as construction, the lack of integration between different stages of design, production, and operation is stark (Gann 2000). Different organizations are involved in each stage of the process and there are few lessons passed from one stage to another. In this environment, there is a tendency to continuously ‘reinvent the wheel’ from one project to another. Each project begets a new set of actors, each of whom may have detailed technical knowledge in their specialist domains but who have little experience of working together. They need to ﬁnd common methods, procedures, and mechanisms to ensure successful communication across a variety of organizational boundaries. In some areas of CoPS, such as IT and telecommunications, ﬁrms have been able to develop ‘repeatable solutions’, passing lessons from one project to another (Davies and Brady 2000; Brady and Davies 2004). However, it appears that such ‘repeatability’ has not yet emerged in the construction sector (Gann 2000; Gann and Salter 2000). PPPs have the potential to help reshape managerial challenges faced by CoPS ﬁrms. PPPs change the division of labour in design, production, and operation of ﬁxed capital goods. They may allow CoPS ﬁrms to take on the whole life cycle of a capital good and therefore realize the ‘life cycle beneﬁt’ highlighted by the UK Government in its arguments in favour of PPPs. They also may allow these ﬁrms to ﬁnd new ways of integrating specialist knowledge from different partners in the project. PPPs shift the burden of operation upstream, back into the supply chain. To understand how suppliers of capital goods are occupying new positions in their industries in order to move into the provision of PPP, it is useful to identify the entire set of activities or steps involved in making, delivering and using a product to provide services to the ﬁnal consumer. The analysis of the ‘value chain’ in much of the business strategy literature (e.g. Wise and Baumgartner 1999; Slywotzky and Morrison 1998) is concerned with how an ‘individual ﬁrm’ can manage upstream and downstream activities to that ﬁrm’s advantage. To understand how ﬁrms change their position in the supply chain to provide PPP it is helpful to examine how value is added within an ‘industry’. A ﬁrm’s value chain for competing in an industry is embedded in a ‘larger stream of activities’ which Porter calls the value system (Porter 1990: 42). The concept

82

The Great Experiment of the ‘value stream’ is used in this chapter to refer to the activities entailed in creating and producing speciﬁc goods and services, ﬂowing from raw materials to the ﬁnal consumer (Womack and Jones 1996: 19).

4.3.1

The capital goods value stream

Until the mid-1990s a traditional capital goods industry—such as railways and telecommunications—typically consisted of two main vertically integrated stages: equipment manufacture and operations. In recent years, these sectors have developed more elaborate divisions of labour as an increasing number of ﬁrms specialize in performing an increasingly narrow range of activities in vertically disintegrated industries. In some sectors, such as construction and railways, PPPs have forced the suppliers to set up special purpose vehicles (SPVs). SPVs act as the key coordinating agency, taking on ﬁnancial, design, and operational responsibility for the ﬁxed capital good. These SPVs act as the system integrator, combining different specialist inputs and organizing the division of labour among the partners of the SPV. To account for this increase in the type and range of activities performed, previous research has identiﬁed four main value stream stages in a typical capital goods industry, as depicted in Figures 4.1 and 4.2 (Davies 2003). The outputs of one value-adding stage are the inputs of the next. Value accumulates at each stage to make up the overall value stream. Each of these stages in the value stream is progressively closer to the ﬁnal consumer, such as the railway passenger or mobile phone user. The stages include: . Manufacture: The ﬁrst stage is the process of taking raw materials and subassemblies and transforming them into physical components and subsystems that are manufactured to meet an overall system design. . Systems integration: The second stage adds value through the design and integration of physical components—product hardware, software, and embedded services—that have to work together as a whole in a ﬁnished product. Systems integrators are responsible for managing numerous inhouse or external contractors responsible for the design and manufacture of components that comprise a system. . Operational services: In the next stage, an operator or business user runs and maintains a system to provide services, such as a corporate telecom network, baggage handling, ﬂight simulation training, and train services. . Final service provision: In some industries, services are provided to the ﬁnal consumer through intermediary organizations called service providers. These ﬁrms buy in the system capacity they require from external operators and concentrate on brand, marketing, distribution, and customer care activities.

83

The Great Experiment

84

UPSTREAM (products)

DOWNSTREAM (services)

Manufacturing-services interface

Added value

Earlier stages

Raw materials, intermediate goods, primary product manufactures

M

SI

Manufacture (M) Systems Design and Integration (SI) produce Design, build, components and integrate subsystems products and systems

Figure 4.1 The industry value stream in capital goods

OS

Operational Services (OS) Maintain and operate products and systems

FSP

Final consumer

Final Service Provision (FSP) Buy in maintenance and operational capacity to provide services to final consumers

Consumption of service by final consumer (e.g. train passenger)

The Great Experiment

Improve design of current & future systems

SI

OS Feed back lesson learnt about inservice maintenance and operational activities

Figure 4.2 The system–service innovation cycle

The four stages can be illustrated by the example of the railway sector in the UK since the early 1990s. Train manufacturing is now dominated by a global oligopoly consisting of Alstom, Siemens, and Bombardier. Over the past decade these ﬁrms have been concentrating on becoming systems integrators and providers of a range of services, and outsourcing a growing proportion of their non-core manufacturing activities. Until 1993, the operational and service provision stages were controlled by a national monopoly. Privatization of British Rail in 1993 created a railway network composed of over 100 specialized operators and service providers, such as rolling stock leasing companies, training operating companies (e.g. Virgin Trains), technical consultancies (e.g. WS Atkins (1999) ), and maintenance companies (e.g. Serco). The line dividing the two segments into upstream and downstream stages corresponds to the traditional manufacturing–services distinction. These segments face different business problems, operate in different market environments, and require different organizations and capabilities. Upstream stages add value to the physical product through technology development, manufacture, and assembly. Downstream stages add value by performing intangible, service-based activities such as managing and maintaining system operations, customer care, advertising, billing, branding, marketing, and other service activities. Systems integration is the pivotal activity in the value stream linking manufacturing to the provision of services. Systems integrators are responsible for managing projects to design, integrate, and install systems that meet a particular customer’s speciﬁcations. They ensure that the value of the solution for the customer is greater than the sum of its parts. They remove the need for the customer to assemble or integrate the products and services that comprise a solution and take responsibility for negotiating with multiple

85

The Great Experiment suppliers of a solution’s component parts—hardware, software, and services. Because systems integrators have an in-depth knowledge of their customers’ operational needs as well as the products they have designed, they are best placed to provide services to monitor, operate, maintain, ﬁnance, and support a product. Rather than a simple linear step-by-step process, adding value involves a series of dynamic feedback loops and iterations between operations and earlier stages of systems integration and design (Hobday 1998: 694; Geyer and Davies 2000). Systems integrators ensure that manufacturers in an earlier stage of production are able to produce components as integrated packages that conform to an overall design. Through ‘learning by using’ (Rosenberg 1976, 1982; Prencipe 2003), operators and service providers can identify opportunities to improve system performance and feedback lessons learnt into the design of current and future product generations. Besides these main value-adding stages, ﬁnancial services and business consulting and other services support and underpin the creation of value by providing inputs at different stages up and down the stream.

4.3.2

Repositioning in the value stream

The changing activities performed by suppliers and customers in capital goods raises questions about the focus of a ﬁrm’s activities and its core capabilities. A ﬁrm’s capabilities can be identiﬁed in terms of the activities it performs in the value stream. When a ﬁrm changes position in the value stream—by moving upstream or downstream—it must develop new capabilities. The growing body of literature on capabilities is inspired by Penrose’s resource-based view (1959) of the ﬁrm—for a review of the literature see Teece and Pisano (1994) and Grant (2002). Since the mid-1990s, the traditional boundary between upstream suppliers and downstream customers continues to be redrawn. Buyers of capital goods are focusing on the provision of services to the ﬁnal consumer and outsourcing non-core activities. Intangible services such as reputation, brand, billing, and marketing are now regarded as more central to the competitive success of these customers than designing, building, or maintaining the systems on which their services depend. To meet the demand for outsourcing, suppliers are undertaking systems integration, operational, and ﬁnancing activities previously performed as part of their customers’ business. In full outsourcing solutions, this includes the transfer of assets and staff to supplier ﬁrms. Buyers of capital goods are entering into long-term partnerships with their suppliers to ensure that providers of solutions share the responsibility and risks of performing outsourced activities. For the outsourcing customer, a capital good no longer represents a ﬁxed cost incurred on an intermittent basis, but a variable cost paid for in regular

86

The Great Experiment instalments for the duration of a service-based contract. A supplier of capital goods can achieve efﬁciency gains by spreading the costs of providing solutions over a larger number of customers. These economies of scale beneﬁt customers that cannot achieve the same cost savings when designing and implementing one-off solutions solely for internal requirements. As Domberger (1998) points out, a supplier can achieve lower unit costs by strengthening its capabilities, because these capabilities represent a ﬁxed set-up cost, which the supplier does not have to reinvest in each time it serves a new customer (Domberger 1998: 78). The process of the customer outsourcing in capital goods has been accelerated by the liberalization and privatization of former state-controlled sectors, such as telecoms and railways, and the adoption of PPPs as the dominant form of public procurement. A variety of customers with different needs operate in more competitive markets. While experienced customers (e.g. incumbent mobile operators like Vodafone) often want to perform a broader range of activities in-house, less sophisticated customers with limited in-house capabilities (e.g. virtual network providers like Virgin Mobile) tend to rely on suppliers for complete solutions to their needs. Suppliers in the UK are also being encouraged to move downstream by increasing use of PPP. In an age when buyers of capital goods are outsourcing more and specializing in the provision of services to the ﬁnal consumer, there is an opportunity for manufacturers like Alstom or service-based consultancy ﬁrms like WS Atkins to integrate forwards in the value stream into the provision of services. In increasingly disintegrated value streams like the railway or telecom sectors, systems integrators are responsible for coordinating a network of upstream component suppliers and subcontractors, designing and integrating systems, and providing services that add value to the product. The adoption of PPP is part of a wider trend towards ‘integrated solutions’ provision that has led many ﬁrms to change their strategies, develop a broader range of service capabilities, and occupy new positions in the value stream (Davies 2003). All ﬁrms moving downstream into PPP have to provide services to operate, maintain, and ﬁnance a product such as a train through its life cycle. But a ﬁrm’s traditional position in the value stream—whether in manufacturing or services—shapes its path of development and the kinds of capabilities it must develop to manage the transition into PPP markets. Manufacturing ﬁrms like Alstom Transport are building on their systems integration capabilities and moving into the provision of train care services (see Box 4.1). Firms that are traditionally based in services like WS Atkins are specializing in being systems integrators of products supplied by best-in-class manufacturers and offering a range of services. The growth of PPP is one the factors encouraging ﬁrms—whether they started as manufacturers or service providers—to develop their capabilities as systems integrators—the capability to design and integrate internally or

87

The Great Experiment Box 4.1 THE CASE OF ALSTOM TRANSPORT Alstom illustrates the shift in strategies in response to the use of PPP in the UK railway industry. Within the diversiﬁed Alstom energy and transportation group, the Alstom Transport division is responsible for train and signalling system design and manufacture. Since the mid-1990s, Alstom has developed a strategy to evolve from a ‘seller of goods to a system and service provider’ (Owen 1997). In a move away from its traditional design and build activities, Alstom Transport, now provides its customers (e.g. UK train operating companies) with complete transport solutions for train availability during the life cycle of the system. The ﬁrst project of this kind, awarded in 1995, was a PFI contract to renew the train ﬂeet on London Underground’s Northern Line. Rather than specify the size of the total ﬂeet, the contract only required that 96 trains be available for service each day for the duration of a 20-year contract. To achieve the customer’s targets for ‘train availability’, Alstom built 106 trains and set up a maintenance organization to service them. Alstom’s managers responsible for maintenance and operational services were deeply involved in the front-end design of the rolling stock. As a result of their recommendations, the train designers made more than 250 modiﬁcations to create easy-to-maintain and easy-to-use trains. The ongoing challenge for Alstom is to improve the reliability of the trains in service in order to avoid penalty payments if it fails to meet the performance targets speciﬁed in its contract with the London Underground. In 1998, Alstom’s Service Business division was created as a result of a strategic review of Alstom’s global activities, which recognized the huge growth in the market for rolling stock services, such as maintenance, technical support, product upgrades, and renovation. The Systems Business division was set up to carry out major systems integration, turnkey and PFI contracts for complete bundles of train, signalling, and infrastructure.

externally developed components into a ﬁnished product (Prencipe et al. 2003). From this new position, ﬁrms are creating new forms of vertically integrated structures by moving forwards in the value stream to offer the broad base of services in operations, business consultancy, and ﬁnancing that are required to provide complete solutions to their customer’s needs.

4.3.3

Moving from unique to repeatable solutions

Performance in PPP depends on how quickly and successfully ﬁrms can move from unique to repeatable solutions (Davies and Brady 2000; Galbraith 2002). The challenge to the supplier is to create organizations that can package and deliver effective and efﬁcient PPP solutions to meet growing customer demand. Suppliers typically invest in the development of a solution with a lead customer so that it can be sold to many other similar customers. Solutions have to be repeatable so that a supplier can get a return on the upfront ﬁxed investment. In other words ‘If every solution is unique, the company cannot make much money on them’ (Galbraith 2002: 203).

88

The Great Experiment

4.3.4

Challenges of providing PPPs

PPPs have been enthusiastically adopted by UK ﬁrms and the use of PPPs has forced these companies to develop new organizational arrangements that provide the means to design, build, and operate these capital goods. In this section, we explore some of the effects of the adoption of PPPs in UK ﬁrms, reﬂecting on the ﬁrms’ experiences and the implications of these experiences of innovation in capital goods. One of the major transformations of the PPP process is cost and effort of bidding for new projects on the part of industrial ﬁrms. For large PPP projects, the bid costs can escalate quickly and become a serious barrier to entry for new ﬁrms. Indeed, WS Atkins, one of the UK leaders in PPPs states in its annual report that one of the attraction of PPPs for the company is that it creates barriers to entry through bid costs and track record requirements on providers (WS Atkins 2004). Given the costs of bidding, the UK government has attempted to manage the competitive process by limiting the number of bidders, usually to three or four bid teams, while still ensuring a degree of competition. The bidding process also requires the providers to develop an extensive plan for the design, development, construction, and operation of the capital good over its life cycle. In developing these bids, different ﬁrms from a range of industries usually band together in a special purpose vehicle (SPV). Each ﬁrm takes an ownership share in the SPV, relative to their contributions to the project. Each SPV acts as an independent organization, ﬁnancially separated from the main organizations. It is possible to think of these SPVs as corporate ‘spin-offs’ where the main organization maintains an equity share and provides staff to support the new venture. The SPV, however, only exists to fulﬁl the terms of the contract and therefore it can be called a project-based organization. However, given the size and length of PPP contracts, the SPVs can become large enterprises in their own right, developing operating routines and cultures often far removed from the original organizations that contributed to the SPV in the ﬁrst instance. The organizational implications of SPV for PPP provision have not yet been fully investigated. In particular, the temporary and multiﬁrm nature of the SPV raises questions about the integration of PPPs into long-term development of organizational capabilities. One of the key beneﬁts of the PPPs for design, construction, and operation of capital goods is that they can create opportunities for ‘economics of repetition’ within and among the organizations involved in the projects. However, given the extensive use of SPVs and often distant and weak links between the ﬁrm and its SPVs, it is difﬁcult to ﬁnd a tight coupling between PPP projects and the creation of bundled or repeated solutions. In fact, these SPVs may create a unique organizational arrangement that operates independently and free from internal routines. This limits the opportunity for PPP providers to develop the learning mechanisms associated

89

The Great Experiment with knowledge codiﬁcation and knowledge creation and capture. This might, in turn, limit the possibilities for the development of dynamic capabilities in organizations responsible for the design, construction, and operation of capital goods (Zollo and Winter 2002).

4.3.5

Impact of PPP on government departments and agencies

The impact of PPP on the procurement capabilities of the UK government is considerable but remains under investigation. The UK Treasury has been providing advice and support to different government departments to enable them to successfully launch more PPPs. As yet, the skill levels within the UK government are still well below what it believes are required to enable more PPP projects to be developed. The use of PPP, however, could mean a major shift in the role and responsibility of UK government agencies and their departments in the design, construction, and operation of capital goods. Traditionally, UK government departments and agencies have maintained internal departments to design and operate capital goods. For the most part, construction was left to the private sector. However, over the 1980s and 1990s, the capabilities to design and procure capital goods have been steadily declining or have been privatized. For example, numerous local authorities have transferred their design and engineering ofﬁces to the private sector. Also many large public agencies have closed down their internal R&D facilities and become more reliant on outside suppliers of specialized advice about the design of capital goods. A new consultancy industry, which provides advice to government about technical choices for capital goods, has emerged to help government agencies with the design and procurement of their capital goods. In many government departments, the capability to procure capital goods has become increasingly associated with the management of the procurement process rather than the design of the capital good itself. This ‘hollowing’ out of the UK government’s capability to design and operate capital goods could have signiﬁcant implications for its ability to procure capital goods in the future. Recent experience in the rail sector indicates some of the dangers of this approach. The creation of Network Rail in 2001 from the ashes of Railtrack was in part the result of a failure of Railtrack to manage its capital projects such as the modernizations of the West Coast Mainline and problems with the maintenance of tracks by external contractors. Network Rail decided in early 2003 to bring in in-house track maintenance to gain a tighter control over the rail system. It has also attempted to improve its ability to deliver and procure major projects by building up its internal capability in systems integration. The experiences of the rail sector may provide a salutary lesson for other government departments thinking of adopting similar arrangements. It also indicates that PPPs may erode the capacity of the UK government departments and agencies to procure capital goods. Given

90

The Great Experiment the central importance of the customer in the design of complex capital goods, the erosion of this capability could in turn undermine the potential for innovation in capital goods.

4.4 Discussion and conclusions The purpose of this chapter has been to examine the use of PPPs by the UK government and their effect on the innovation process in capital goods. We suggested that provision of long-term service contracts has the potential to reshape the design and operation of capital goods. Such contracts change the traditional division of labour between the private and public sector and between the producers and operators of capital goods. The contracts could allow for the integration of design, construction, and operation of capital goods inside a single organization and such integration could create a virtuous circle of development whereby lessons from operations are integrated into the design and production of future capital goods. Despite this potential, we have argued that little research has been done to assess whether such an opportunity has been achieved in practice. We began by reviewing research from secondary sources that indicated that PPPs have had a positive inﬂuence on many projects. The ﬁnancial beneﬁts of PPP are still open to debate, but PPP projects have generally been delivered on time and on budget. It remains, however, difﬁcult to understand the precise effect of PPP on project performance given all the other factors that might shape project performance. PPP providers have argued that the scope for innovation in PPPs has been limited by overly cautious government ofﬁcials. Efforts are being made to create more ﬂexibility in contracts and servicelevel agreements. The difﬁculty of specifying long-term contracts remains unresolved and could have serious implications for PPPs in the medium and long term. The greatest impact of the new arrangements appears to have been the emergence of a PPP industry and, along with this, a new division of labour among private providers and public procurers of capital goods. In addition to its traditional role as designer and builder of systems, the private sector has taken on greater downstream responsibility for operation and maintenance of systems previously handled in-house by public sector organizations, such as schools, the NHS and the Ministry of Defence. There can be substantial beneﬁts for innovation in capital goods from the new procurement arrangements if private sector ﬁrms can both successfully execute major public projects and provide efﬁcient and highly reliable operational services. Yet the jury is still out as to whether these changes will be long-lasting. For some equipment suppliers, PPPs have enabled expansion of their capabilities, forcing them to move downstream into operational services. The new arrangements have

91

The Great Experiment forced ﬁrms to build up their knowledge and skills of ﬁnancing projects, of working in large and diverse teams, and in the management of operational services. It has also helped them to create a virtuous circle between the different stages of design, production, and operation of capital goods. This new virtuous circle could allow capital goods producers considerable opportunities to gain lasting advantages for the design and production of capital goods, overcoming years of separation between users and producers of capital goods, and potentially providing more reliable and efﬁcient services. The transition to PPPs has been dramatic, often involving the strategic positioning of ﬁrms away from traditional design and engineering services towards facilities management and other operational services. Many organizations have had to reorient their approach to the market and its own capabilities. More research needs to be done to study the emergence of a ‘virtuous circle’ between different stages of design, production, and operation of capital goods in many areas of PPPs. There are some examples of PPPs where systems integration and operational activities are performed in-house by large vertically integrated suppliers. However, the ‘best in breed/class’ model of subcontracting adopted by many SPVs and widely followed in the UK construction sector means that there is a ‘moveable feast’ of project partners involved in different stages of the capital goods design, production, and operation process. In this case, there is less evidence of the emergence of a new model of innovation and there appear to be few beneﬁts from these SPV arrangements for capital goods innovation. In this respect, the use of PPPs has just shifted the separation of design, construction, and operation from the government to the private sector without creating higher levels of integration between these stages. One factor that might explain the difference between the two systems is that the use of PPPs represents an extension of the contracting mentality of the construction industry (Gann and Salter 2000). It may be that until this mentality is challenged there will be little opportunity to create a virtuous circle between design, production, and operation in capital goods. The growth of PPPs has not yet been accompanied by research on the impact of these arrangements for innovation. Long-term service contracts could have profound implications for the way capital goods are designed, produced, and run. More knowledge about how these arrangements work and how they shape the potential for changes and innovation in the development of capital goods and the services they render would be greatly beneﬁcial. Research needs to go beyond simply examining where PPP projects have been successfully delivered on time, to the required speciﬁcations and within budget to examine the operational performance of the PPP supplier in terms of reliability and efﬁciency. Such in-depth case studies of PPP suppliers would provide a greater understanding of how policy changes play out in the innovation process. Large-scale surveys of PPP practices could also be beneﬁcial in allowing researchers to examine the effects of the procurement mechanism on ﬁrm or

92

The Great Experiment project performance. Until such research is undertaken, the implications of the ‘great experiment’ of the UK government’s PPP policy will remain largely unresolved.

References Acha, V., Davies, A., Hobday, M., and Salter, A. (2004). ‘Exploring the Capital Goods Economy: Complex Product Systems in the UK’, Industrial and Corporate Change, 13: 505–29. (forthcoming). Akintoye, A., Beck, M., and Hardcastle C. (2003). Public-Private Partnerships: Managing Risks and Opportunities. London: Blackwell. Allen, G. (2001). The Private Finance Initiative, Library of House of Commons, Research Paper 01/177, 18 December. House of Commons Library, London. Brady, T. and Davies, A. (2004). ‘Building Project Capabilities: From Exploratory to Exploitative Learning’, Organization Studies, 25(9): 1601–21. Brusoni, S., Prencipe, A., and Pavitt, K. (2001). ‘Knowledge Specialization, Organizational Coupling, and the Boundaries of the Firm: Why do Firms Know More Than they Make?’, Administrative Science Quarterly, 46(4): 597–621. Davies, A. (1996). ‘Innovation in Large Technical Systems: The Case of Telecommunications’, Industrial and Corporate Change, 5(4): 1143–80. —— (2003). ‘Integrated Solutions: The Changing Business of Systems Integration’, in A. Prencipe, A. Davies, and M. Mobday (eds.) The Business of Systems Integration. Oxford: Oxford University Press. —— and —— (2000). ‘Organisational Capabilities and Learning in Complex Product Systems: Towards Repeatable Solutions’, Research Policy, 29(7–8): 931–53. ——, Brady, T., Tang, P. et al. (2001). Integrated Solutions: The New Economy Between Manufacturing and Services. Brighton, SPRU–Science and Technology Policy Research. Domberger, S. (1998). The Contracting Organization: A Strategic Guide to Outsourcing. Oxford: Oxford University Press. Galbraith, J. R. (2002). ‘Organizing to Deliver Solutions’, Organizational Dynamics, 31(2): 194–207. Gann, D. M. (2000). Building Innovation: Complex Constructs in a Changing World. London: Thomas Telford. —— and Salter, A. (2000). ‘Innovation in Project-based, Service-enhanced Firms: The Construction of Complex Products and Systems’, Research Policy, 29(7–8): 955–72. Geyer, A. and Davies, A. (2000). ‘Managing Project-System Interfaces: Case Studies of Railway Projects in Restructured UK and German Markets’, Research Policy, 29: 991–1013. Glaister, S. (1999). ‘Past Abuses and Future Uses of Private Finance and Public–Private Partnerships in Transport’, Public Money & Management (July–September): 29–36. Grant, R. M. (2002). Contemporary Strategic Analysis: Concepts, Techniques, Applications. Malden, MA: Blackwell. Grimsey, D. and Lewis, M. (2002). ‘Accounting for Public-private Partnerships’, Accountancy Forum, 26(3): 245–70. Grimshaw, D., Vincent, V., and Willmott, H. (2002). ‘Going Privately: Partnership and Outsourcing in UK Public Services’, Public Administration, 80(3): 475–502.

93

The Great Experiment Hartley, K. (2002). ‘The Economics of Military Outsourcing’, Public Procurement Law Review, 5: 287–97. HM Treasury (2000). Public-Private Partnerships: The Government’s Approach, with introduction by Andrew Smith, Chief Secretary to the Treasury. London: HMSO. —— (2003). PFI: Meeting the Investment Challenge. London: HMSO. —— (2004). PFI/PPP Projects Database. London: HMSO. Hobday, M. (1998). ‘Product Complexity, Innovation and Industrial Organisation’, Research Policy, 26: 689–710. —— and Rush, H. (1999). ‘Technology Management in Complex Product Systems (CoPS): Ten Questions Answered’, International Journal of Technology Management, 17(6): 618–38. ——, ——, and Tidd, J. (eds) (2000). ‘Innovation in Complex Products and Systems’, Research Policy, Special Issue, 29(7–8). Hood, C. (2000). The Art of the State: Culture, Rhetoric and Public Management. London: Oxford University Press. IPPR (Institute for Public Policy Research) (2001). Building Better Partnerships (June). London: IPPR. Jessop, B. (1994). ‘The Transformation to Post-Fordism and the Schumpeterian Workfare State’, in R. Barrows and B. Loader (eds.) Towards a Post-Fordist Welfare State. London: Routledge. —— (2002). The Future of the Capitalist State. Oxford: Polity Press. Leiringer, R. (2003). Technological Innovations in the Context of Public-private Partnerships, PhD thesis, KTH Industrial Economics and Management, Stockholm, Sweden. Michaud, P. (2001). Institutional Innovation and the Selection of Complex Engineering Projects: A Dual Relation, PhD thesis, University of Sussex. Miller, R., Hobday, M., Leroux-Demers, and Olleros, X. (1995). ‘Innovation in Complex Systems Industries: The Case of Flight Simulation’, Industrial and Corporate Change, 4(2): 363–400. Morris, P. (1994). The Management of Projects. London: London Thomas Telford. NAO (National Audit Ofﬁce) (2001). Managing the Relationship to Secure a Successful Partnership in PFI Projects, Report by the Comptroller and Auditor General, HC 375, 29 November. —— (2002). The Public–Private Partnership for National Air Trafﬁc Services Ltd. Report by the Comptroller and Auditor General, HC 1096, 24 July. —— (2003a). Managing the Relationship to Secure a Successful Partnership in PFI Projects, Report by the Comptroller and Auditor General, HC 375, 29 November. —— (2003b). National Savings and Investment Deal with Siemens Business Services, Four Years On, PPP in Practice Report by the Comptroller and Auditor General, HC 626, 8 May. —— (2003c). PFI: Construction Performance, Report by the Comptroller and Auditor General, HC 371, 5 February. —— (2003d). The Operational Performance of PFI Prisons, Report by the Comptroller and Auditor General, HC 700, 18 June. Osborne, D. and Gaebler, T. (1993). Reinventing Government: How the Entrepreneurial Spirit is Transforming the Public Sector. New York: Plume. Osborne, S. (ed.) (2000). Public–Private Partnerships: Theory and Practice in International Perspective. London: Routledge.

94

The Great Experiment Owen, D. (1997). ‘GEC Alstom in Career Discussions’, Financial Times, 19 November. Parker, D. and Hartley, K. (2003). ‘Transaction Costs, Relational Contracting and Public– Private Partnerships: A Case Study of UK Defence’, Journal of Purchasing and Supply Management, 9(3): 97–108. Penrose, E. (1959). The Theory of the Growth of the Firm (3rd edn., 1995). Oxford: Oxford University Press. Prencipe, A. (2000). ‘Breadth and Depth of Technological Capabilities in CoPS: The Case of The Aircraft Engine Control System’, Research Policy, 29(7–8): 895–911. —— (2003). ‘Corporate Strategy and Systems Integration Capabilities: Managing Networks in Complex Systems Industries’, The Business of Systems Integration. Oxford: Oxford University Press. ——, Davies, A., and Hobday, M. (eds.) (2003). The Business of Systems Integration. Oxford: Oxford University Press. Porter, M. E. (1990). The Competitive Advantage of Nations. London: Macmillan. Rosenberg, N. (1976). Perspectives on Technology. Cambridge: Cambridge University Press. —— (1982). ‘Learning by Using’, in N. Rosenberg (ed.) Inside the Black Box: Technology and Economics. Cambridge: Cambridge University Press, pp. 120–40. Slwotzky, A. and Morrison, D. J. (1998). The Proﬁt Zone: How Strategic Business Design Will Lead You to Tomorrow’s Proﬁts. Chichester, UK: John Wiley & Sons. Teece, D. and Pisano, G. (1994). ‘The Dynamic Capabilities of Firms: An Introduction’, Industrial and Corporate Change, 3(3): 537–56. Wakeford, J. and Valentine, J. (2001). ‘Learning Through Partnership: Private Finance and Management in the Delivery of Services for London’, Public Money & Management (October–December): 19–29. Williamson, O. (1985). The Economic Institutions of Capitalism. New York: Free Press. Wise, R. and Baumgartner, P. (1999). ‘Go Downstream: The New Proﬁt Imperative in Manufacturing’, Harvard Business Review (September–October): 133–41. Womack, J. P. and Jones, D. T. (1996). Lean Thinking: Banish Waste and Create Wealth in Your Corporation. New York: Simon & Schuster. Woodward, J. (1965). Industrial Organisations: Theory and Practice, 2nd edn. Oxford: Oxford University Press. WS Atkins (2004). Annual Review. Epsom, UK: WS Atkins. Available at www. atkinsglobal. com (accessed October 2005). Zitron, J. (2004). PFI and PPP Projects: Are They Working? A research study commissioned by the Major Project Association, 2nd edn. February, Major Project Association and Henley Management College, London. Zollo, M. and Winter, S. (2002). ‘Deliberate Learing and the Evolution of Dynamic Capabilities’, Organisational Science, 13(3): 339–51.

95

This page intentionally left blank

THEME 2 EVOLUTION AND ADAPTATION OF STRUCTURE

This page intentionally left blank

5 Complexity, Evolution, and the Structure of Demand John Foster and Jason Potts

5.1 Introduction Implicitly or otherwise, all talk of economic transformation is talk of a dynamical process. But a dynamical process can mean many things. So the question should be more speciﬁc: What kind of dynamical process is economic transformation? There are many ways of describing economic transformation. It may be a smooth and laminar process, for example, or jumpy and reticulated. It may be pushed along by drivers of growth, or pulled along by entrepreneurship and ﬁnance. It may or may not be pulled or pushed anywhere in particular. Its dynamics might be like a volume that expands, or more like a chemical reaction that unfolds. It may be a deterministic, stochastic, or evolutionary transformation process, or a mix. It may involve increasing differentiation, or perpetual revolution, or strange attraction. It may be ergodic or historical, about just the parts of the system or the whole. It may be about open or closed systems. A dynamical process can mean a great many things. For the most part, however, economic thinking about transformation has retained the dynamical systems categories of old, which tended to be mechanistic and closed in crucial aspects. This point is clearly evident in the almost complete supply-side bias in the theory of economic transformation, and on the predominance of analyses of closed system processes linking demand and supply. There are many ways of describing economic transformation. Yet despite the wealth of (largely unexplored) combinatorial possibilities implied by the many meanings of dynamical systems decomposition above, economic theorists have, more or less, self-organized themselves into just two overarching groups in answering questions concerning the identity of dynamical processes. The ﬁrst favours extensions of standard growth theory (e.g. ‘endogenous growth

99

Complexity, Evolution, Structure of Demand theory’), and views transformation as a process of expansion under constraint. The second favours neo-Darwinism (e.g. ‘neo-Schumpeterian economics’) and views the same process as guided by mechanisms of variation and selection. The work of Paul Romer in the ﬁrst case and Richard Nelson and Sidney Winter in the second exemplify these respective approaches. Clearly they are different; they emphasize different things. Yet these studies, and the many like them, are all inherently stories about production and the supply side. Most economists who have considered these things have usually found their way into one or other or these positions. But not us, however. We think that neither view yet has the measure of a general conception of economic transformation, and for two simple reasons—lack of a demand perspective and lack of an understanding of the complex systems under consideration. In this chapter we shall propose a new way of looking at economic transformation that is based in the ontology of evolutionary realism (Dopfer et al. 2004) and on the theory of complex economic systems (Potts 2000; Dopfer et al. 2004; Foster 2005). In our view, the story of aggregate demand is an evolutionary one about a complex system of consumption behaviour. The economic system is a highly complex rule-system on both the ‘supply’ and ‘demand’ sides—indeed the validity and usefulness of this kind of traditional analytical dichotomy must be questioned from a complex systems perspective. Economic transformation is the process of change in rules and their populations, and in the connective structure of both rules and populations. Production and consumption phenomena are an intertwined network that, if we are to get a clear sense of what economic transformation actually involves, needs to be understood as a complex network. And so we are critical of the seemingly default-setting (what David Hume called ‘dogmatic slumber’) of students and scholars of economic growth and transformation to overly concentrate on the supply side of the equation. Towards redress, we shall set out some rudiments of this general complex systems perspective in order to illuminate the deep relation and complex interconnection between supply and demand, or production and consumption, in the analysis of open system economic dynamics. Two ideas run through this chapter. First, production function-based analysis of economic dynamics has been overgeneralized, and, indeed, has crowdedout network and complex system analysis. We illustrate some network-based analytical tools and discuss their analytical relevance. Second, a network or meso-based perspective will rebalance the analytical relationship between production and consumption. Economic evolution, in this new view, which brings together Veblenian, Keynesian, Schumpeterian, Hayekian, and ‘Simonian’ economics, is as much about change in the connective structure of consumption systems as about change in production systems. Both stories are necessary for a proper analysis of economic evolution. Indeed, it is the meeting of these two stories that is the proper analysis of economic evolution.

100

Complexity, Evolution, Structure of Demand

5.2 Consumption networks not production functions Students of economic growth and transformation have perennially expressed a remarkably consistent preference for just one side of the Marshallian equation. They mostly prefer to think about production and supply, and rather less about consumption and demand. This is no recent phenomenon, for Schumpeter was like that too, as were the early Ricardians, Institutionalists, and Austrian economists. There have been very few exceptions. And so the study of transformation in economic systems has, over the years, developed as an almost exclusively supply-side story. This is curious, given that the very ﬁrst thing the young economist learns is that there are two sides to the micro story (supply and demand) and two sides to the macro story (income and expenditure). Yet somewhere along the way this tends to get forgotten. Why? A possible explanation is a philosophical bias towards materialism and actions. The supply side is about making things; it is hard and objective. The demand side, however, is about shopping: it is soft and subjective. The notion of transformation implies an energy conversion process, or work. Work requires an engine, and the engine (or, equivalently, drivers) of the economy are ‘clearly’ in production rather than consumption, which is, of course, the destruction of work. In this way, it is ‘natural’ to suppose that the proper and serious questions about growth and transformation are to be associated with the supply engine of the economy. (And that only sociologists, psychologists, marketers, or other intellectual-effeminates or refugees from hard-core economic theory will see ﬁt to consider the demand side pars toto.) To grow an economy is to increase the power of an engine, and that means thinking about production and supply, about efﬁciency. Demand, as Jean Baptiste Say taught, will just happen naturally and ﬂuidly. Supply is about putting order into a system, demand is about sucking it out. Pace Alfred Marshall, supply is ‘really’ analytically prior and the proper subject of interest; demand is what just passively follows. John Maynard Keynes thought this was nonsense, and indeed made it the basis of his general theory, but few subsequent Keynesians ever did. Who started this? The classical economists, of course. It was a way of thinking that was encapsulated in their very notions of value. They taught that the ultimate determinants of value are to be found in the combination and recombination of the basic factors of production, and so it has been ever since. The neoclassical revolution was revolutionary because, supposedly, it corrected this lopsided view of the origin and nature of value by showing how both supply and demand interact to create value. But looking over modern growth theory, you would not much know that. From Bob Solow to Bob Lucas to Charles Jones, and everywhere between, neoclassical growth theory has reverted to classical type. The theory of economic transformation,

101

Complexity, Evolution, Structure of Demand irrespective of whether we are referring to neoclassical, new classical, or even evolutionary variants, has, on the whole, maintained an overarching supply-side bias. Evolutionary models have a strong claim to superiority over neoclassical and new classical growth theory because they have further advanced our understanding of the resolution of variety and the nature of selection mechanisms. Still, the supply-side preference limits their analytical power. An overly narrow focus has made the full evolutionary story of economic transformation overly one-sided. Too much attention has been given to factor accumulation, as in growth models, or the resolution of technological variety, as in selection models, and insufﬁcient attention has been directed towards the forces that give rise to variety in the ﬁrst place. A full evolutionary story needs to have a conception of supply as the product of a network structure that ﬁts into the broader context of consumer demand networks. To conceptualize this, we need to begin with a view of the economic system as a complex structure of networks; both within and between producers and consumers. We will argue that conceptions of production or consumption functions should be replaced by network representations. In such models, the preferences of consumers—or, more correctly, their aspirations—are fundamental and, as such, the primary drivers of economic growth. In such models, technological, institutional, and market innovation processes are intermediate between the aspirational networks of consumers and the organizational networks that produce goods and services. Economic evolution is a story about how systems interact with other systems. When we make network connections between consumers, producers, and between producers and consumers explicit, both the ﬂexibility and stability of economic systems can then be analysed in a more coherent way than hitherto. Consumer knowledge becomes at least as important as producer knowledge in determining the generation of economic value. And it becomes clear that the stability afforded by connected systems of economic rules is essential for economic ﬂexibility to exist, which many rules can eventually result in inert and structurally unstable states and that too few rules can result in a kind of stability but at a low level of complexity. As a dynamical process, the economic system is badly misrepresented as a production function with constraints. Much better to think of it as a complex, self-organizing rule-system, and therefore to understand that this is a process on the demand (or consumption) side as much as on the supply (or production) side. The generation of micro value, a` la Marshall, is about the interaction of demand and supply, macro value, a` la Keynes, is an aggregation of the expenditures of demanders and suppliers, and what we can call ‘evolutionary value’ that views value as an outcome of dynamic interactions between such players within changing network structures. An economic system is, fundamentally, a complex system of complex systems.

102

Complexity, Evolution, Structure of Demand

5.3 The economy is a complex rule-system Rather than basing economic analysis on production functions for ﬁrms and preference functions for consumers, we gain a better understanding of economic structure and its dynamics by viewing both production and consumption as occurring in complex interacting networks. Value has both static and dynamic dimensions. Evolutionary theory is about how they interact (Straton 2004). Evolutionary economists have long sought to reject ﬁeld-based analytical techniques (see Potts 2000). Eschewing comparative static, demand and supply curves and closed-system analytical techniques alike, they have ultimately sought refuge in evolutionary analysis, which has generally been taken to mean neo-Darwinism (e.g. Hodgson 2002). This has been an important part of the story of economics growing-up. But it is not yet the whole story. Evolutionary economics should be founded upon high-order complex systems theory rather than on neo-Darwinian analogies about replicators and natural selection. Such analogies are useful to understand the process of variety resolution, but they do not help in understanding how variety is generated and diffused (Metcalfe et al. 2005). The weakness of competitive selection theory is that it does not provide any account of the network dimensions of variety or the network dynamics of selection. In neo-Darwinian biology, variety is a matter of random mutation with particular mutations making more favourable connections with the environment than others. Yet this is wholly unsatisfactory in evolutionary economic settings where variety begins as an act of imagination in a complex systems context. The existence of ‘variety’ presumes the availability of network connections. So, a ﬁrm that makes more successful connections with consumers will grow faster and its connective structure in production and distribution will tend to become dominant in that particular market. However, variety itself is the outcome of network dynamics that stretch well beyond the economic domain. Marketing expenditure, in prime instance, is fundamentally important to forge network connections both between producers and consumers and between consumers. The latter connection is vital because of the importance of imitation and group identiﬁcation. Thus, much TV advertising is not about the details of the product but rather about the supposed social status of a purchaser. In turn, this is related to the socio-psychological perceptions of those who design advertising campaigns. Consumer tastes and material aspirations are both diverse and conformist. Conformism arises because meso-rules exist, for example in ‘fashion’ (see Chai et al. forthcoming). We have a need to wear fashionable clothing but identical items of clothing tend to be avoided, cars with very similar technical characteristics vary in their styling, and so on. All suppliers, as well as the innovators of the products that they sell, depend upon the prior existence of network connections between products and

103

Complexity, Evolution, Structure of Demand consumers, and connections between consumers. Otherwise they have to manufacture connections themselves, and that is not always trivial. History is littered with innovations that did not make money because there were no connections with consumers and, often half a century later, an innovation was eventually adopted because the network structure of demand had evolved to a degree that allowed the relevant connections to be made. It is important to understand that the complexity of an economic system is constructed on the consumer side just as much as on the producer side, and sometimes, and often even, may lead it (see Earl and Potts 2000, 2004). Say’s Law was predicated on the notion that factor payments must turn into some form of demand. Keynes pointed out that a bout of pessimism by those purchasing durable goods can reduce demand signiﬁcantly, i.e. if the economy is perceived as a complex system of incomplete (and malleable) network connections with all the attendant uncertainties. The fact that producers supply goods and services is no guarantee that they will subsequently be demanded, and therefore reliance on price movements to equilibrate in any general sense is ultimately misconceived. It is undoubtedly true that goods falling out of fashion will fall in price, and that other novel fashion goods will attract high prices as new meso-rules develop. But it is equally clear that conventional price theory tells us little about this transformational process. Prices are not just determined by ﬁrms varying the supply of goods to the market, e.g. as in conventional perfectly competitive and monopoly pricing models. In our view of economic transformation as a dynamical process, it is the change in the structure of demand that is crucial. Keynes was interested in the macroeconomic consequences of this. In particular, he was concerned with cases where there are general reductions in demand because fears concerning the future diminished expenditure. This applies just as much to a family deferring the purchase of a new vehicle, because of fears that income is at risk due to an anticipated recession, as it does to a ﬁrm postponing an investment project for the same reason. Because such opinions concerning the future relate to uncertainty in a complex economic system, they tend to be interconnected through networks (see Potts 1999; Flieth and Foster 2002; Potts 2004).

5.4

The growth of demand and the growth of economies

The growth of economies depends on the expansion of consumption sets, either through our inherent desire for novelty or through the power of marketing in expanding the range of goods and services we aspire to consume. As economic growth proceeds, more and more of the connections that we forge with others become economic in nature, i.e. they are expressed through transactions and contracts across an ever-widening economic network. The

104

Complexity, Evolution, Structure of Demand result is an economy that is characterized by increases in organized complexity that facilitate greater consumption of materials, energy, and time use (in both producing and consuming). Thus, understanding economic growth involves both understanding the socio-psychological and cultural factors that lead to the formation and attainment of novel material aspirations and the continual shifting of preferences (Earl and Potts 2004), as well as oppositional tendencies towards inertia manifested in habitual and routinized behaviour. Economic growth involves the meeting of imagination with experimental possibility. Yet if we think of economics as the domain of reason and logic in our cognition, then much observed economic activity stems from the non-economic emotional or instinctual behaviours that we have inherited in our genes from our past. Many of the network connections that are formed because our emotional dispositions are ‘hard wired’ (i.e. adapted) to an environment that, in many and important ways, bears very little resemblance to the hunter-gather world that has made up most of human history (Rubin 2002; Potts 2003). This is why freedom and economic evolution co-evolve (see Dennett 2003). Today we go ‘bargain hunting’ and gathering tends to occur in supermarkets. Marketing professionals are acutely aware of these things, but economists have been reluctant to forsake the comparative safety of abstract logic and reason to study what Adam Smith and David Hume referred to as the ‘passions’ that they saw at the core of the economic system. While undeniably mathematically convenient, the abstract logic of idealized optimization is insufﬁcient for understanding how an economy works and changes. It is simply that computational processing by agents is only part of what is involved, and cannot, and should not, be assumed given or exogenous. Trouble awaits any bold simplifying assumptions at this juncture. Damasio (1999), for example, has conclusively demonstrated that people who can optimize but do not have an emotional feeling cannot actually function normally in the economic system. This, along with many other insights from neuroeconomics, experimental economics, and behavioural economics, points to a neglect of the nature of what Herbert Simon called bounded rationality and what Brian Loasby calls bounded cognition. Economic behaviour is a by-product of the connective networks that people forge. What is of crucial interest is how meso-rules have evolved that facilitate economic activity in the midst of the other rule-driven behaviours that human beings engage in. Thus, when we think of the economy and its components as complex adaptive systems, we are inevitably drawn into the study of human intentions and interactions. Neoclassical economists decided a long time ago that all this was intractable. It was better to summarize all this into well-deﬁned utility and production functions, and then to concentrate on logical outcomes. However, it is simply not possible to obtain a satisfactory representation of economic behaviour in

105

Complexity, Evolution, Structure of Demand historical time by adopting this stance. Also, as the network structures that are typical of complex economic systems are becoming better understood, it is clear that there can be analytical representations of economic behaviour that are superior, both in terms of insights and explanation, to those on offer in neoclassical economics. The emergence of economic novelty plays a central role in our evolutionary understanding of economic transformation. In much of the innovation and evolutionary economics literature the emergence and stabilization of novel products and processes is articulated as an exclusively supply-side problem. We argue here that growth cannot be understood without a similar articulation of the development of consumer behaviour in response to novelty. Consumers innovate too, and their changes in behaviour are just as central to growth and the sustainability of growth as are the innovative activities of ﬁrms. If we accept the importance of variety, learning, and selection on the supply side, we must accept it on the demand side as well.

5.5

Correlated preferences

Preferences can be idiosyncratic or related to basic needs but, in the modern economy, many are learnt, either through the efforts of marketers or by observing the preferences of others (contagion). In this way, the formation of preferences is socially embedded, for idiosyncratic preferences would never be sufﬁcient to sustain a market. Thus, markets depend on the emergence of correlated preferences; on the emergence of a degree of common understanding across consumers as to the functions of individual products. For the analysis of economic growth, the ‘correlated preference’ is the relevant micro-foundation, not individual utility functions. Correlated preferences are not to be confused with uniform preferences—such preferences will vary in their strength, but the higher the degree of correlation across the consuming population and, of course, they will not include all the population, the more effectively are preferences translated into demand. If preferences depend on consumer understanding they cannot be stationary. The stock of preferences subject to continual change is dynamic, with new ones being continually formed and existing ones growing, as they spread and intensify, sharing increases while the old are being abandoned. Correspondingly, preference shift means that products tend to follow ‘life cycles’. It follows that the product cycles we observe in the economy are also preference cycles. The spread of a correlated dynamics of preference sharing is determined by marketing efforts and socio-psychological processes that result in contagion, which means a growing correlation in understanding across consumers. If effective market arrangements exist, prices can mediate between preference expansion/contraction and the provision of supply changes. Prices do not

106

Complexity, Evolution, Structure of Demand initiate change, but, rather, act as coordinating devices as preferences and production costs change, i.e. as connections change. Economic evolution involves the evolution of consumption systems as well as production systems. Consumption systems evolve as connections form and change between the aspirations of consumers, the elements of consumer lifestyles and the consumption possibilities offered up by producers. The inevitable outcome of such an evolutionary process is a dynamical structure of correlated preferences (or a ‘meso-trajectory’, as we shall soon see). The demand-side economics of an evolutionary process of transformation are a long way indeed from the standard view of independent and stationary preferences possessed by rational agents

5.6 Analysis of economic networks An analytical hold on this perspective is achievable, not by trying to extend preference theory by ever higher order algebraic relaxations, but rather by rethinking the geometry of the space in which preferences are compiled and operated, and of the dynamics of this process. And for that we need to turn to network theory (e.g. Kirman 1997; Potts 2000, 2001) to conceive of an economic system as a complex structure of connected rules. In this way, an evolving economic system is amenable to network analysis (Dopfer et al. 2004). Network theory has come a remarkably long way in recent years, from a staid and esoteric branch of pure mathematics with only oblique application (e.g. chemical enumeration, packing problems) to one of the hottest new ﬁelds of applied mathematics in the study of complexity (e.g. Watts 2003). If the economic system is a complex rule-system, then network theory is the most natural foundation for analysis of economic structure and dynamics. With network theory, we can hope to formalize analysis of the dynamics of complex economic structure in general and of consumption in particular. A network, or graph, is a set of elements and a set of connections forming a system. Different classes of network are deﬁned by the density and distribution of the connections between elements. At the limit of integral density, where every element is directly connected to every other, a network collapses into an integral space known more generally as the real ﬁeld. Most micro- and macroeconomic theory implicitly assumes a real space. This makes for easy application of the calculus of variations, but for a rather more tortured path through analysis of structure and dynamics. When we abandon the calculus of variations we lose the intuitive logic of production functions, but we gain a ﬁnely grained insight into the distinct dynamical properties of differently connected systems. Network theory is the link between the connective structure of an economic system and the sorts of dynamics, or processes of transformation, which are possible.

107

Complexity, Evolution, Structure of Demand All networks can be mapped to a number of classes. The two simplest classes of analytical networks are a regular lattice and a random graph. In a regular lattice, each element has only local connections, with an equal density across the lattice. In a random graph, connections need not be local, but are randomly distributed between a given set of elements. Regular lattices have a lot of local structure, but very poor long-range connective properties (i.e. they have large diameter). Random graphs are the opposite. They have very little local structure, but excellent connective properties (i.e. small diameter). Long ¨ s showed that as connections are randomly ago, the mathematician Paul Erdo added, beyond a critical density, a phase transition occurs in the connected state of the system. Below the threshold, a network will consist of many separated islands; above the threshold, the network percolates into a giant connected cluster. When analysing the dynamics of Random Boolean Networks, Kauffman (1993) showed that this transition is associated with the complex regime of phase space, and is an evolutionary attractor. These general properties of complex networks as complex adaptive systems are discussed in Potts (2000, 2001). There is much that can be said about the structural and dynamical implications of the state of complexity for the structural and dynamical properties of an economic system, and we shall examine some of these soon with respect to the relation between consumption and demand systems and production and supply networks, but network theory has also moved on signiﬁcantly in the past few years, mostly due to the work of a scattering of non-linear mathematicians and solid-state physicists. In 1998, Watts and Strogatz published some analysis of what they called small-world networks. Small world networks are a hybrid of a regular lattice and a random graph, in that like a regular lattice, they have a lot of local structure, but unlike a lattice they also have a very small diameter. This enables them to combine in a single network properties usually only found in very different classes of network. Watts and Strogatz speculated, like Kauffman, that this property would make them evolutionary attractors, able to combine and balance forces of structure and change. In 2001, Barabasi presented some new thinking about the evolutionary growth of networks. He deﬁned a scale-free network as that which grows by preferential attachment of new elements to existing hubs. This results in a power law distribution of connectivity across the network. Barabasi (2002) argued that similar evolutionary properties accrue to such networks, such as resilience to attack and robustness to decay. What is interesting about network theory for the analysis of economic systems is that it connects speciﬁc structural properties of networks with speciﬁc dynamical properties. Only complex adaptive systems are capable of ongoing evolution. Network theory helps us understand the sorts of structural properties we would expect those dynamical systems to have. (And also show-

108

Complexity, Evolution, Structure of Demand ing why the assumption of an integral space can seriously mislead the analysis of open system dynamics.) However, the problem with the sorts of network theory that mathematicians, physicists, and computational biologists have been developing is that, from the perspective of economic analysis, it all tends to be rather ﬂat. For the most part, complexity is viewed as a global property of a system, which either is or is not present. But there is complexity and then there is complexity. Is the complexity of a chemical dissipative system the same sort of complexity as that found in a biological organism, and indeed, is this the same sort of complexity as found in an economic system? We think not.

5.7 Orders of complexity Foster (2005) has recently proposed a framework based on four fundamental orders of complex systems. Its foundations are drawn from self-organization and open system thermodynamics. The underlying idea is that an economic system is a class of dissipative structure, which, like all dissipative structures, builds and maintains a boundary between internal order and external chaos, and where complexity involves a component interaction between order and disorder, between core and periphery. This view is assembled by folding the self-organizational perspective on an economic system (e.g. Foster 1997) with the complex network theory perspective of an economic system (e.g. Potts 2000) with the theory of entropy, complexity, and emergence (e.g. Brooks and Wiley 1986; Schneider and Kay 1994; Holland 1998; Collier and Hooker 1999; Kauffman 2000). The outcome is a view of economic evolution as an open system process of high order complexity. Evolution is a process of replication and emergence. Component structures evolve from a process of specialization and integration that yields a coherence that contains the seeds of either system destruction or structural transition to a new state. Yet analytical models of a system often begin with the presumption that it is in a high state of order, or is capable of attaining such a state (equilibrium in the force ﬁeld not the thermodynamic sense). Analysis is then intended to reveal mechanisms (actual or designed) that can be subject to control. Its structure is so organized that its mechanisms can be represented in mathematical terms. Such a system does not, and cannot, evolve. Evolution can only occur when a structure has the capacity to both expand and unravel. While preserving topological form, variety, and disorder are features of its periphery. Complexity is a term that refers to the connective structure (or lack thereof) of a system that permits it to process energy and information, while variety refers to the connective potential of a system. Complex adaptive systems, which arise when we go above the physio-chemical

109

Complexity, Evolution, Structure of Demand level of structure, have been viewed as at the ‘edge of chaos’. Variation is balanced with selection. The system is maintaining a boundary while maintaining adaptive potential. The properties of order and chaos are melded into structural complexity which, at least for a time, exhibits coherence (see Kauffman 1993). This poses analytical problems. In the prime instance, the behaviour of such systems cannot be consistently represented as a control problem for the simple reason that system emergence and system failure are fundamentally important to the dynamics of the system. Emergence and failure are, classically, not control problems. And this is why evolutionary science is different. The maintenance of variety and novelty is essential for emergence and it can only be sustained if there is failure. All evolution stems from the existence of variety. Control deals with the design of mechanisms to avoid failure. But if you eliminate all failure, you eliminate evolution. Following Foster (2005) we can identify four orders of complexity: .

1st order complexity (the imposed energy case) arises in physio-chemical settings when energy is imposed on chemical elements. Non-adaptive structures or patterns facilitate the dissipation of energy as internal organization. This level of complexity can be modelled with conventional dynamical systems and non-linear dynamical techniques, including control theory.

.

2nd order complexity (the imposed knowledge, acquired energy case) involves responsiveness to information and, therefore, knowledge in some sense, which permits some control over energy ingestion and the impact of external forces. This is the level of complexity that is prevalent in the biological domain where physio-chemical component structuring is overlaid by selection mechanisms in the presence of genetic and behavioural variety. Knowledge is imposed by experience.

.

3rd order complexity (the acquired knowledge case) exists when a system does not interact directly with its environment as a data ﬁeld, but via internal models formed from the external world. Knowledge is no longer experience of reality but also ‘mental models’ that can determine aspects of reality. There is feedback from reality and feed forward to reality. This is the complexity we ﬁnd in the socio-economic domain, where imagination as well as selection is involved. The system can rearrange its environment advantageously.

.

4th order complexity (the interactive knowledge case) arises when the mental models themselves start to interact. My imaginings can still shape reality but knowledge that this is so leads to others imagining what my imaginings might be. As John Maynard Keynes perceptively observed, participants in ﬁnancial markets are more interested in the average level of sentiment in the market than the relation of prices to the ‘fundamentals’. This kind of complexity becomes important when the future is

110

Complexity, Evolution, Structure of Demand involved in terms of commitments, such as forward contracts and commitments that involve terminal dates in the future. As we proceed up through these orders of complexity, the role of connections and knowledge becomes increasingly apparent and important. The further up we go, the more that knowledge matters by a process whereby connections become layered until they become an environment in themselves. Fourth order complexity is really what is often referred to as the ‘knowledge’ or ‘new’ economy. A ﬁrm is an ordered system in terms of its organization and institutions. There is structured behaviour within a ﬁrm, but it is also a system at the edge of chaos in terms of its temporal behaviour. Businesses deal with forward commitments as promises that are turned into contracts if an institutional/legal structure exists in the environment. Standard models of the ﬁrm used to ignore complexity altogether. In the 1980s, with the development of the economics of imperfect information, rational expectations theory and game theory logic, third order complexity began to be explored. Fourth order complexity is still virtually ignored in the standard theory because it cannot be treated in a rational expectations or a game theoretic framework. But the essence of the complex nature of consumption and demand in an economic system is that it is fourth order complex. Just as ﬁnancial markets exhibit fourth order complexity in terms of interactive knowledge, so too do consumption systems as populations of rules begin to interact structurally. Economic evolution consists of both the process of emergence and regeneration of novelty, and the changing structure of network connections as this occurs. However, the main difﬁculty with this conception of economic evolution lies in its kaleidic nature, and it was this that prevented Austrian economists from offering analytical formalizations of such evolutionary processes in the presence of novelty generation (cf. Loasby 1999; Potts 1999). Although the work of von Hayek and others has emphasized the central role of rules in the economic system (e.g. Vanberg 1994), the Austrians did not have at their disposal the insights of modern network theory to ﬁnd their way through the complexity of the economic system (cf. Earl 2003). Without some kind of formal way of dealing with these rules, then the fourth order complexity that characterizes advanced economic systems cannot be approached from a scientiﬁc perspective. The behaviour of a complex system can be described in words in space and time, but no analytical generalizations are available than can be used for data-based, that is statistical, explanations.

5.8 Micro–meso–macro Dopfer et al. (2004) have recently proposed a new analytical framework for evolutionary economics composed of three analytical domains: micro,

111

Complexity, Evolution, Structure of Demand meso, and macro. The economic system is a micro and macro structure of meso-rule populations. Unlike the conventional micro–macro partition, micro is not the foundation of evolutionary analysis, rather meso is. And, we suggest, it is from this meso-centred framework that we ﬁnally get a clear view of the relative importance of consumption and demand in economic transformation. The meso unit is deﬁned as a generic rule and its population of actualizations. It is the fundamental building block of the economic system. An economic system, then, is deﬁned as a structure of meso-rules expressed in both micro and macro dimensions. At the micro level, we see these as systems of interacting rule actualizations, and at the macro level as structures of interacting rule-populations. An economic system is a highly complex system from the micro and macro perspectives, but when this view is centred on the meso, we see that it is also an open and evolving system. Economic evolution, by deﬁnition, involves change in the meso domain. Economic evolution occurs in a process described by what we call a mesotrajectory. A meso-trajectory consists of three phases: (a) the origination of a novel generic rule, (b) the adoption and adaptation of that rule through a population of carriers (agents), and (c) the retention and maintenance of the rule population. Over long periods of time, economic systems evolve by the ongoing process of replication of extant meso populations, and the micro and macro structure they entail, as well as by the emergence of new meso units and the effects they have on extant micro and macro structure. The micro–meso– macro framework allows us to see more clearly the full complexity of economic evolution. From this rather abstract perspective, then, the theme of stability and ﬂexibility in economic transformation seems to us to refer to the question of how new institutional level meso-rules emerge and diffuse. Again, these are Austrian themes, namely, the emergence of novelty, the emergence of institutions through ‘spontaneous order’ and the evolutionary character of the economy (see Witt 2003). Yet, Austrian economics has always had an ideological tinge, favouring individualism, universal markets, and minimalist government. The micro–meso–macro framework does not adopt such sharp delineations and lies most comfortably with the idiosyncratic neoAustrianism of George Shackle and, of course, Keynes’s vision of the economic system. Neither individualist (micro) nor collectivist (macro) perspectives are upheld, markets are limited in extent and government is an integral part in the economy (the extent of economically relevant complex structure cannot meaningfully be restricted to the private sector). In this chapter, however, our attention is upon consumers. Agents viewed as independent micro elements in conventional economics are here viewed as complex behavioural entities connected, usually in several dimensions, to other individuals through actualizations of meso-rules. Conventional economics diminishes the importance of consumer demand precisely because the consumer is pre-

112

Complexity, Evolution, Structure of Demand sumed to be an optimizer in a simplistic environment. The equilibrium state of the consumer, with all its attendant trade-offs, moves only when there are exogenous shocks. As such, the consumer is an entirely passive entity, not the active driver of economic evolution. (A critique forcefully made by Veblen as early as 1898.) In neo-Schumpeterian economics, entrepreneurs are thought of as organizational innovators, bringing together techniques, capital, and labour to form a productive organization. However, this is only a means to an end. The end is a novel product, which will match a new preference, or a cheaper existing product that opens up choice to more people. As a consequence, the growth of knowledge of supply amongst demanders and demand amongst suppliers market is vital to entrepreneurial success (Allen and Strathern 2004). Without new demands, new products cannot be sold. The entrepreneur must work with a model of market demand possibilities as much as with a model of supply possibilities Accumulated evidence suggests that in the traditional view of the unfolding of a market, it is assumed that products generally go through four distinct phases: (a) emergent (accelerating) growth, (b) stable growth, (c) saturated (decelerating) growth, and (d) bifurcation (involving either product abandonment by suppliers or the attainment of ‘core’ status absorption into a more stable core of correlated preferences). Frequently, the products with core status possess a high degree of complementarity with peripheral products and, thus, from this arises their more stable demand characteristics relationships. Over and above the intrinsic appeal of such a product, it is in demand because of a ‘derived preference’ relation to a range of other products. The signiﬁcance of the ‘four stages’ is that each stage has its own growth trajectory, and that these different velocities of change are extremely important to business decisions concerning the supply of a product and its ultimate survival in particular markets. What determines survival in all phases is not the level of proﬁts/losses but their rate of change. Emergent products will, initially, register losses—these will fall in the case of successful products, but not for the failures. Later on, in the bifurcation phase, proﬁts may still be positive, but falling, when abandonment occurs. In the stable growth phase, the rate of proﬁt tends to rise at a decelerating rate but the high level of proﬁts enjoyed can enable investments in capital goods that can lower costs. This cost-cutting strategy tends to intensify in the saturation phase. Proﬁts are, increasingly, due to rents in this phase rather than due to rising volume of production. As preferences move in favour of new products, productive organizations may seek to maintain proﬁts through both increased efﬁciency and protectionist strategies. Thus, how ﬁrms behave in relation to the maintenance of old, and the introduction of new, products depends upon the changing network structure of production and consumption. The complex system perspective can predict quite different outcomes to conventional neoclassical economics. For

113

Complexity, Evolution, Structure of Demand example, the best time to abandon a product is often when its proﬁtability is highest; excessive cost-cutting can damage the capacity of a ﬁrm to adapt its products and, thus, survive; the emergence of monopolist or oligopolistic ﬁrms is what competition should produce—competition is a process that destroys itself. In the micro–meso–macro framework, economic agents are both adapting to the changes bought by extant meso-trajectories, and, if they are entrepreneurial, starting their own meso-trajectories. These have consequences that can be analysed at both the micro and macro level. Each case, however, conforms to a generic process of emergent growth, stable growth, saturated growth, and ﬁnally bifurcation. Producers are much involved in this process, and what results is a network of supply. Consumers, similarly, are also much involved, and what results is a network of demand. Economic transformation is the story of how these two processes link up into a single process. This is well represented as a micro–meso–macro process of emergent complexity in the meso structure between micro agents and the macro structure of meso units. An economic system is a complex adaptive system of meso-rules, about which there is both micro structure and macro structure. In turn, the micro and macro structure is a dynamical function of the relative sizes of meso populations. This is the simplest possible representation of the necessary analytical structure to describe the process of economic transformation. The immediate implication of this view is to refocus analytical attention back to the origins of demand, and therefore of novelty.

5.9

The complexity of consumption and demand

Homo Sapiens Oeconomicus is a dissipative system, and, like all dissipative systems, a consumer. At base, consumption is an open-system thermodynamic process. But this process is only elucidated at high orders of complexity in an economic system. In Foster’s (2005) orders of complexity framework, second order complex states arise when animals essentially engage in two activities, seeking food and behaviour associated with reproduction. Knowledge is acquired concerning demand (what tastes good) and supply (best hunting and gathering sites and strategies). Because of the ‘Malthusian’ consequences of sexual reproductive activity, novelty is essential, i.e. novel hunting or gathering locations or experimentation with novel foods (see Potts 2003 on the role of nomadic journeys as generators of novelty). In the absence of a search for novelty, inertia, poverty, and death await. However, despite the longer term importance of novelty, in the shorter term, habit and routine facilitate the organization necessary to allow the search for novelty to occur, and in this sense are equally crucial. Hunter-gather societies

114

Complexity, Evolution, Structure of Demand operated in this way for most of human history until humans moved to third order complex states, whereby knowledge came to be imposed on the environment through the domestication of animals and then agriculture. This led to the formation of much more complex societies that could throughput more energy via larger populations (see Diamond 1997). Then, through the monetization of production and trade, urban communities and industries evolved with a vast proliferation of goods and services produced by specialists (Landes 1998). This was the eruption of trucking and bartering that Adam Smith was ﬁrst to comprehend as the outlines of an emergent evolving complex system (Loasby 1999). Provided that there were no natural disasters, the human capacity for generating novel solutions to economic problems kept human society ahead of a Malthusian state until the emergence of modern economies, where fourth order complexity became dominant. In this state, Malthusian pressure stopped—the genetically determined impulse to reproduce became dominated by desires to accumulate durables and to consume their services. Economic evolution became a more immediate and powerful force on the human species than biological evolution. Fourth order complex states involve long time horizons exempliﬁed in contracts and hedging arrangements. Once upon a time, long-term consumption was a matter concerning offspring; with economic evolution, it became something attainable within a lifetime. When we stand back and look at this process, what we see is that consumption always drove economic behaviour. Needs expanded because of the inherent search for novelty. And novelty, as we have noted, is a deep thermodynamic force in all behaviour. All innovation and other supply-side activities are undertaken with consumption in mind. What changed over time were the exhortations that individuals faced. In agrarian/feudal societies, the exhortation was to have as many children as possible to provide agricultural labour and soldiers. In modern economies, the exhortation is to consume expensive consumer durables funded by long-term loans. We ﬁnance our children, rather than using our children for ﬁnance, and this is a much better deal for all concerned. Clearly there have been important changes in the way that consumers behave. The micro–macro–meso framework is of assistance in helping us understand that process. Essentially, what drives consumption is meso-rules that are articulated in microeconomic settings. The feudal meso-rules, based on religion, family, community, and nationalism, gave way to new meso-rules based upon status, measured in terms of material afﬂuence. There were many novel ways that the feudal serf could serve and there are, equally, many kinds of goods and services that can convey material status. The prevalence of the latter meso-rules also had supply-side effects because entrepreneurs (who were effectively outlawed in feudal societies, as they were in many communist countries until recently) are largely driven by a quest for material afﬂuence. Thus it follows that in order

115

Complexity, Evolution, Structure of Demand to understand the process of economic development, i.e. the process whereby society becomes more complex economically, uses more energy per capita and consumes a greater range of market valued goods and services per capita, it is necessary to understand how meso units have come and gone. This also involves an understanding of the changes that have taken place in the network structure of the economy as it has moved from one of predominantly random networks to predominantly scale-free networks. It is here that the qualitative nature of the meso-rule merges with the quantitative dimension of the economic network, which is in terms of monetary valuations. The meso unit can, thus, be given a measurable dimension. As Austrian economists stress, constitutions, good common law procedures, and democratic principles all help allow this process to occur. Otherwise, most of our individual consumption decisions are just variations around mesorules. Instabilities occur in economic systems not because of our micro behaviour in a given meso-rule state, but rather because of large shifts in mesorules—and this is Keynes’ point, again. It is the durability of meso-rules in consumption that give the aggregate consumption function an econometric stability emphatically not shared by aggregate business investment functions. The latter reﬂects the considerable movement in meso-rules that exists because of the heavy role of conﬁdence factors and technological obsolescence. When Keynes advocated stabilization policy, it was precisely to encourage a meso-rule that would provide more conﬁdence amongst business investors and, thus, fewer recessions and a higher rate of growth (Foster 1987). In the mechanical representations of Keynesian policy in the textbooks, this dimension of the Keynesian strategy hardly rates a mention. All economic decision-making involving consumption streams in future time-periods supposes the formation of an aspiration which, when pursued, becomes an explicit goal. Meso-rules guide the formation of aspirations, and it is for this reason that so much is spent on marketing to inﬂuence aspirations. Of course, in a detailed sense, no two aspirations are the same and, thus, it does make sense to talk of different ‘tastes’. However, these are marginal to a primary main aspiration that is expressed in an abstract way. For example, it is true that there is a vast choice of bathroom ﬁttings for houses to meet the variety of tastes that people have, but, ultimately, this choice is derivative, of a larger aspiration to own a signiﬁcant consumer durable that is, more than likely, in excess of basic needs. It has been over a century since Thorstein Veblen coined the term ‘conspicuous consumption’ as one of the hallmarks of a modern capitalist system, and it would appear that the appetite for such consumption became insatiable in the twentieth century. Economic evolution involves the transformation of consumption systems as well as production systems. At one end of the economy, consumption is a driver of economic growth and, at the other end, is the accessibility of energy and materials that input

116

Complexity, Evolution, Structure of Demand into the productive process. Technological changes have enabled previously uneconomic primary reserves to be accessed and existing ones to be harvested at lower cost. The exploitation of energy and material resources has important long-term impacts, yet sustainability considerations have hardly impacted at all on consumption aspirations. From the perspective of meso-rules, this is interesting given that there have been examples in history of societal groups that have upheld strong meso-rules concerning sustainability and others that have not. The latter mostly died out. These could be viewed as largely due to ignorance concerning, for example, the salination effects of irrigation. However, the modern day problem seems to stem from a strong conﬁdence in technology to solve all problems, emphasizing an increasing reliance on scientiﬁc rationality at least since the early nineteenth century in Europe. This is consistent with our observations that economics, both orthodox and heterodox, has been over preoccupied with the supply side and the role of technological change. Thus, there appears to be a deeply held meso-rule in capitalist societies that relates to ‘scientiﬁc progress’ as a cure for all ills, eliminating all moral responsibility for the effects of excessive consumption. Human beings are dissipative structures and this is the essence of economic agency (Hass 1972). The complexity of an economic system is the complexity of 3rd and 4th order knowledge systems, and in which the economic problem is essentially that of tracking and then generating novelty, and of coordinating the results. Economic evolution is characterized by a process called a mesotrajectory that, from the demand side of the equation, looks a lot like a process of preference co-integration. The meso structure of an economic system is a statement about the structure of demand as much as it is about the structure of supply. And because it tends to move more slowly than the pace of embedded technological change, the structure of demand is ultimately the determinant of economic evolution and therefore of transformation of economic systems. From the meso perspective, the correlation of preferences is both the basic driving force of economic evolution and the ultimate cause of economic stationarity, depending upon the relative phase-structure of a meso-trajectory.

5.10

Evolution and aggregate demand

Evolutionary economics in general and technology studies in particular have tended to focus too much on supply-side considerations and competitive selection arguments. The consequence of this lopsided view is a poorly developed sense of the role of network and complexity analysis in understanding the process of economic transformation. We argued here that consumption is about individual meso-rules and their interactive structure between agents. We have also argued that demand is about the relative size of meso populations and their effect on the evolution of the macroeconomy. Aggregate ﬂows

117

Complexity, Evolution, Structure of Demand of expenditure or demand will follow paths shaped by meso trajectories that were initially forged by micro agents making new connections. Consumption, like production, is a complex system. Transformation means making new connections not just between production technologies and supply lines, but also between consumer aspirations and patterns of consumption. This network-based view helps us to see clearly the role of market institutions, among other things, in the process of economic transformation. We may even come to see this as a kind of ‘allocation of complexity’ problem in terms of the allocation of rules over carriers in evolving systems by the effect of the relative price of rule embedding in complex systems (Clark et al. 2004). Theoretical developments such as these, and others that may follow, will all help us better understand the evolutionary nature of aggregate demand. It is our view that the demand side of the economic value equation is more than just a passive absorbing medium, as often implicitly assumed and portrayed, but is instead a functionally signiﬁcant and scientiﬁcally interesting part of the open-system dynamical process of economic evolution. Surprisingly, perhaps, given the hoary arch-veracity of modern economic analysis, much empirical and theoretical work yet remains to be done at the interface of demand and supply. And particularly so in the dimensions of micro evolutionary consumer theory and macro evolutionary network or ‘aggregate’ demand theory.

References Allen, P. and Strathern, M. (2004). ‘Evolution, Emergence and Learning in Complex Systems’, Emergence, 5: 8–33. Barabasi, A. (2002). Linked: The New Science of Networks. New York: Perseus. Brooks, D. R. and Wiley, E. O. (1986). Evolution as Entropy. Chicago: University of Chicago Press. Chai, A., Earl, P., and Potts, J. (forthcoming). ‘Attention, Fashion and the Growth of Aggregate Demand’, in R. Franz (ed.), Renaissance in Behavioural Economics: Essays on Harvey Leibenstein. London: Routledge. Clark, J., Morrison, K., and Potts, J. (2004). ‘The Allocation of Complexity’, ACCS working paper, UQ. Collier, J. and Hooker, C. A. (1999). ‘Complexly Organised Dynamical Systems’, Open Systems and Information Dynamics, 6: 111–36. Damasio, A. (1999). The Feeling of What Happens: Body and Emotion in the Making of Consciousness. New York: Harcourt Brace. Dennett, D. (2003). Freedom Evolves. London: Penguin. Diamond, J. (1997). Guns, Germs and Steel: The Fate of Human Societies. New York: W.W. Norton. —— Foster, J., and Potts, J. (2004). ‘Micro–meso–macro’, Journal of Evolutionary Economics, 14: 263–79. —— and Potts, J. (2004). ‘Evolutionary Realism: A New Ontology for Economics’, Journal of Economic Methodology, 11: 195–212.

118

Complexity, Evolution, Structure of Demand Earl, P. E. (2003). ‘The Entrepreneur as a Constructor of Connections’, Advances in Austrian Economics, 6: 117–34. —— and Potts, J. (2000). ‘Latent Demand and the Browsing Shopper’, Managerial and Decision Economics, 21: 111–22. ——and—— (2004). ‘The Market for Preferences’, Cambridge Journal of Economics, 28: 619–33. Flieth, B. and Foster, J. (2002). ‘Interactive Expectations’, Journal of Evolutionary Economics, 12: 375–95. Foster, J. (1987). Evolutionary Macroeconomics. London: Unwin Hyman. —— (1997). ‘The Analytical Foundations of Evolutionary Economics: From Biological Analogy to Economic Self-organization’, Structural Change and Economic Dynamics, 8: 427–51. —— (2005). ‘From Simplistic to Complex Systems in Economics’, Cambridge Journal of Economics, 29. (forthcoming). Hass, H. (1972). The Human Animal. New York: Putman. Hodgson, G. (2002). ‘Darwinism in Economics: From Analogy to Ontology’, Journal of Evolutionary Economics, 12: 259–81. Holland, J. (1998). Emergence. New York: Addison-Wesley. Kauffman, S. (1993). The Origins of Order: Self-organization and Selection in Evolution. New York: Oxford University Press. —— (2000). Investigations. New York: Oxford University Press. Kirman, A. (1997). ‘The Economy as an Evolving Network’, Journal of Evolutionary Economics, 7: 339–53. Landes, D. (1998). The Wealth and Poverty of Nations. London: Abacus. Loasby, B. (1999). Knowledge, Institutions and Evolution in Economics. London: Routledge. Metcalfe, J. S. and Ramlogan, R. (2002). ‘The Limits to the Economy of Knowledge and of Knowledge of the Economy’, CRIC Discussion paper, University of Manchester. ——, ——, and Foster, J. (2005). ‘Adaptive Economic Growth’, Cambridge Journal of Economics, 29. (forthcoming). Potts, J. (1999). ‘Uncertainty, Complexity, and Imagination’, in P. Earl and S. Frowen (eds.), Economics as an Art of Thought: Essays in memory of G.L.S Shackle. Routledge: London, pp. 162–84. —— (2000). The New Evolutionary Microeconomics: Complexity, Competence and Adaptive Behaviour. Cheltenham, UK: Edward Elgar. —— (2001). ‘Knowledge and Markets’, Journal of Evolutionary Economics, 11: 413–31. —— (2003). ‘Toward an Evolutionary Theory of Homo Economicus: The Concept of Universal Nomadism’, in J. Laurent (ed.), Evolutionary Economics and Human Nature. Cheltenham, UK: Edward Elgar. —— (2004). ‘Liberty Bubbles’, Policy (Centre of Independent Studies), 20(3): 15–21. Rubin, P. (2002). Darwinian Politics: The Evolutionary Origins of Freedom. New Jersey: Rutgers University Press. Schneider, E. and Kay, J. (1994). ‘Life as a Manifestation of the Second Law of Thermodynamics’, Mathematical and Computer Modelling, 19: 25–48. Straton, A. (2004). ‘Evolutionary Value Theory’, (Ph.D. thesis, School of Economics, University of Queensland) Vanberg, V. (1994). Rules and Choice in Economics. London: Routledge.

119

Complexity, Evolution, Structure of Demand Veblen, T. (1898). ‘Why is Economics not an Evolutionary Science?’, Quarterly Journal of Economics, 12: 373–97. Watts, D. (2003). Six Degrees: The Science of a Connected Age. New York: W.W. Norton. —— and Strogatz, S. (1998). ‘Collective Dynamics of Small World Networks’, Nature, 393: 440–42. Witt, U. (2003). The Evolving Economy. Cheltenham, UK: Edward Elgar.

120

6 Self-transformation, Self-organization, and Evolutionary Adaptation in the Economic Process J. Stan Metcalfe and Ronnie Ramlogan Perhaps it is not too much to say that what we need is an evolutionary theory worthy of our best social theory, not a social theory trimmed to ﬁt a rapidly receding, overly simplistic, evolutionary theory. Depew and Weber (1996: 495)

6.1 Introduction The central theme of this chapter is the connection between economic adaptation and economic development and growth.1 In it we explore the idea that economies never grow without simultaneous development, and that development is an experimental, creative process. By development we mean structural change in the broad, with the relative growth and decline of activities and the addition and deletion of activities from the economic population. Developmental change is qualitative and quantitative, it requires the continual creation and reallocation of resources and over quite moderate periods of time the composition of an economy’s activity may change quite fundamentally and irreversibly. It is, as Schumpeter (1928) expressed, a process of ﬁnding new uses for resources; rather more grandly, it is a process of discovering new economic worlds. Adaptation implies ﬂexibility and a willingness to change to accept losses as well as welcome gains, it implies that forces in favour of the status quo are not absolute, and it is intimately connected with the idea of dispersed economic power. More fundamentally, we shall argue, it is inseparable from the human propensity to develop new beliefs and to ground those beliefs in reliable knowledge. Two kinds of adaptation may therefore be distinguished. The ﬁrst is the traditional case of adaptation in the context of given knowledge, as the result, say, of the opening of an industry to foreign trade or the imposition of excise taxes, the consequent changes in prices inducing appropriate changes in the patterns of production and consumption, market by market for a given set of commodities, services, and activities.

121

Evolutionary Adaptation in the Economic Process In the process labour is relocated between ﬁrms and industries and the physical capital stock is similarly adjusted. But this is not the principal kind of adaptation that we have in mind for the transformations that mark capitalism are inseparable from the growth of knowledge and involve much more than the changing relative importance of a given set of commodities and end uses. In this second kind of adaptation, innovation and the growth of knowledge are the principal transforming elements but only when placed in the context of market and other institutions that regulate the economic process of adaptation. Moreover, innovation and new knowledge presuppose novelty and it is the treatment of novelty that makes the connection between economic evolution, economic adaptation, and complexity. This takes us into deep waters for neither innovation nor its effects can be predicted with any accuracy, nor can the content and import of new knowledge be known in advance, so there is no possibility of a closed account of the economic process; development is open-ended, radically uncertain and thus impossible to foresee in its detailed consequences and it is the details that matter. However, this does not mean that the underpinning evolutionary processes are beyond detailed analysis and comprehension as long as we recognize that the only possible dynamics is a short-run dynamics that explores the immediate evolution of the system. In this account there are no stable, long-run attracting sets towards which the economy may converge, there are no equilibria, only degrees of order and structure of differing degrees of transience. As we shall see the fundamental reason for this claim, and the source of the link with complexity, is the endogeneity of the growth of knowledge. Only in a world of stationary knowledge is it possible to think of invariant attractors for a system, but a world of stationary knowledge is not a world of sentient, imaginative individuals, it is effectively a world populated by the dead. It is certainly not the world of capitalism, with its history of at least nine centuries of knowledge-led development (White 1962; Mokyr 1991, 2002; Landes 1998). Even the merest acquaintance with the facts demonstrates that the economic world of 2005 bears little resemblance to that of 1960 and even less to that of 1805 or even 1905 in that the entire pattern of production and resource allocation in terms of nature and location has changed as new products and methods of making them using new kinds of material and energy appeared, and old ones disappeared (Mokyr 1991, 2002; Landes 1998). Patterns of resource allocation become radically different over time, the activities and economic ways of life of consecutive generations bear little resemblance to each other, and patterns of consumption include practices and purchases that would be undreamt of by earlier generations. Adaptation is the constant theme, the one process that appears true of all periods of capitalism. Even in 1960 would many have imagined let alone believed that the desktop computer would be almost as ubiquitous as the television in the households of a modern economy, and that the typewriter industry would virtually disappear as a

122

Evolutionary Adaptation in the Economic Process consequence? Who would have foreseen that the ubiquitous ﬁlm camera would have its market radically cut by the application of digital computer technology, or that the telephone would contain within it a camera? Who in the 1930s would have foreseen the negative impact of the television on the cinema industry or would have imagined the effect of the refrigerator on patterns of household living? Few modern homes are lit by coal gas, not so in 1905; a virtually negligible proportion of the population today works directly on the land, not so in 1870; and very few make the trip from Europe to New York or Australasia by ocean liner, not so in 1960. The record, in this longterm perspective, appears to be one of radical discontinuity such that any comparison of a single economy over extended time is fraught with difﬁculty. Growth, the increase in some aggregate measure such as gross domestic product (GDP) per capita, never happens without development and the ongoing radical redevelopment of the economic structure so that economic change always appears to be uneven within and between countries. The scope of the adaptations involved is considerable, from changes in consumer spending patterns, through changes in the nature and location of the employment of labour, land, and produced means of production, to new methods of production and new commodities. In its broadest contours it is reﬂected in urbanization and the transfer of population out of agricultural pursuits, in the increasing use of inanimate energy in production activities of all kinds together with the development of new material bases for production, in the increasing importance of ‘service’ activities in total activity and the emergence of new international divisions of labour. In respect of these and other dimensions the characteristic feature is their connection with new ways of living, new patterns of organization, and new ways of knowing. It is on the latter that adaptation ultimately depends as cause and effect, and it is the epistemic dimension of adaptation that connects directly to complexity. In the presence of new knowledge existing patterns of order in markets and their institutional frames more generally become inappropriate for purpose and thus are subject to adaptive pressure. It is the precise response to this pressure that characterizes and shapes the development process in an economy. Slow adaptation means slow development and a slow revision of knowledge, perhaps the most compelling of the constraints that hold back an economy or industry or business. Adaptive response is facilitated by efﬁcient market processes but this is much more than the issue of ‘getting the prices right’. When markets allow use to be made of all the available information so that consumers and suppliers are fully informed of their respective offers and when the constellation of prices are not arbitrarily distorted by taxes, tariffs, or quantitative restrictions, then the process of economic evolution will be at its most efﬁcacious. But ensuring that prices reﬂect real resource costs in their totality is the lesser part of the adaptive problem. It is more fundamental that markets are open to change; that every position is open to challenge by

123

Evolutionary Adaptation in the Economic Process hitherto unspeciﬁed rivals, indeed this is the proper signiﬁcance of markets that we emphasize. Markets are social devices that not only facilitate the connection between production and exchange but also leave those connections open to radical transformation and generate the incentives to mount those challenges to the prevailing order. Bureaucracies and central plans can also make the production/exchange connections but markets are vastly superior at the process of radical transformation. What this points to is the creative aspect of the market system as a whole. Rivalry requires more than efﬁcient product markets in which consumers, whether ﬁnal or intermediate, are not locked into speciﬁc suppliers but can switch allegiance when other suppliers offer superior products or better prices. It also requires efﬁcient markets for access to labour and willingness of labour to be mobile between rival employment opportunities, and it similarly requires an efﬁcient capital market to facilitate business experimentation and the growth of businesses with real competitive advantages. None of these features of a market economy can be taken for granted. Inertia, habitual practice, the constraints of social ties, the costs of change more generally make it extremely unlikely that markets will be maximally efﬁcient in the conventional sense of requiring uniform prices for equivalent transactions. Moreover, the restless, uncomfortable features of capitalism give every producer an incentive to engage in practices that protect prior investments in capital and capability, which limit possibilities for entry, that seek politically sanctioned market privileges, and which slow down the adaptive response of consumers. Regulation is almost always needed to keep the system open, and this is quite different in scope and nature, for example, from regulation to limit the abuse of monopoly power or to keep the prices right. Moreover, markets do not come for free, the process of generating and disseminating the relevant information absorbs real resources. Where commodities are sufﬁciently homogeneous and sold in sufﬁcient quantity the market-forming functions can be undertaken by specialized intermediaries who in making the market may also hold stocks with the aim of stabilizing the more extreme ﬂuctuations in prices. The classical merchant trader and the modern supermarket fall into this category. More generally, however, in the case of markets for less clearly speciﬁed goods and services the market is made by suppliers who purchase information disseminating services to ‘market’ their products and who compete in an imperfect information market to attract customers. Thus the institution of the market is the principal device to generate ﬂexibility and to ensure that resources are adapted to new possibilities. What does this entail? In one of the few detailed attempts to deal with the economics of ﬂexibility, Killick (1995) also makes the useful distinction between responsive adjustment in conditions of given knowledge and innovative adjustment, and points out that the latter depends on such factors as degree of education, technological capabilities, and leadership, difﬁcult categories for an econo-

124

Evolutionary Adaptation in the Economic Process mist. Adjustment can occur at multiple levels, he distinguishes the individual, the organization, and the institution, but the important point is that adjustment at all levels requires incentives, commitments, and capabilities. Responsive adjustment, say, to a change in relative prices presupposes knowledge of alternative ways of consuming and producing, and ignorance of alternatives is a sure source of inertia. It is not the speed of adjustment that is the sole issue here but the potential range of adjustment, the horizon of change as it were. Interrelatedness of behaviours too adds to the scope for inertia, for the change in one element in behaviour may depend on adjustment to a package of other elements so raising the possibility of threshold responses and domains of inertia. Since all change requires some effort, some investment, then one would expect a generalized presence of ‘adjustment costs’ created by complementarities. The necessary investment in turn requires a capability for identifying new courses of action and for managing the change process. Existing knowledge can also be a constraint on change since knowledge too is costly to adjust and entails the abandonment of commitments and prior positions (Walker 2000). We dismiss at our peril the signiﬁcance of habit and routine as constraints on the adaptive process. For these reasons perfect adjustment is a facile concept. The reality of adaptation is that it is heavily constrained by the prevailing state of knowledge, neither organizations nor individuals readily abandon practices that hitherto have served their purpose well; it takes a signiﬁcant change in the environment to lead to adaptation. However, economies do adapt and markets facilitate adaptation and this discussion suggests that adaptive pressure often comes from outside-established practice. Signiﬁcant innovations, as Schumpeter ([1911] 1939) pointed out, are often associated with new entrant ﬁrms or more radically with the establishment of new activities and sectors. Cases in the managerial literature are many where established ﬁrms fail to read the signiﬁcance of the innovative products or processes produced by rivals and suffer the consequences (Fransman 1994; Utterback 1994; Christensen 1997). Thus what is intrinsic to market capitalism is that adaptation occurs at the level of the market as well as at the level of the ﬁrm, resource supplier, and consumer. The inability of established producers to perceive or respond to new investment opportunities leaves room for new entrants to innovate and for the market to guide the changes in resource allocation. An important dimension of this perspective is the emphasis on the restless nature of modern capitalist economies. They are organized, indeed self-organized, but they also self-transform as the loci of activity are redistributed within and between ﬁrms, economic sectors, and countries (Foster 1993). Change arises as the consequence of the operation of the system, it is not imposed from outside; capitalism is a system for promoting change from within, as Schumpeter ([1911] 1934, 1944) always emphasized. The idea of a capitalism that does not exhibit these properties is surely a contradiction in terms despite the frequent use in

125

Evolutionary Adaptation in the Economic Process economic theory of the benchmark stationary or static state.2 In fact, the economic process is an unceasing struggle with new activities emerging to compete with older, more established rivals, displacing them in the process and ultimately succumbing to competitive pressures themselves. The generation of novelty plays a central part in this dynamic process and this provides the important connections with innovation in capitalism and the fact that adapting to innovation requires decline in some activities as well as growth of others. Capitalist economies never expand in a uniform fashion in which the growth of all activities is equiproportional and time reversible. Furthermore, adaptation as self-transformation is not to be confused with adaptation as transition. The latter entails a move from initial state A to ﬁnal state B, with neither state inﬂuenced by the process of going between. Hence, transitions are reversible in principle and can take place in logical time. Transformations, by contrast, are open ended, what is being transformed is a set of possible states that are created in the process of transformation and take place in historic time. The process alters the end point and the beginning point because the process of movement itself generates new knowledge and hence renders the change irreversible. Economic transformations are a species of ‘far from equilibrium’ process and what keeps them ‘far from equilibrium’ is the particular set of knowledgegenerating and-application processes that deﬁne a modern economy. They constitute an instituted system for generating and adapting to innovations in technique, product, service, and organization that arise within the system. Thus, we suggest, complexity in economics rests on the interaction between processes for generating novelty and processes for adapting to that novelty. It is only when we ignore the knowledge dimensions of an economy that it is possible to treat economic systems in a non-complex, time-reversible fashion. The wider signiﬁcance of this is that the market system must be unstable if it is to evolve. That is to say the forces that generate order in markets, the typical role of the price mechanism, must coexist with but be outweighed by the disruptive, innovative forces transforming that order. Stable systems cannot evolve, they cannot be invaded by new patterns of activity, and they are closed to development. But this does not characterize capitalism, within its broad instituted frame the order is under continuous transformation as the opening paragraphs explained. Economic history is a history of instability induced by the invasion of new activities and the working out of the consequences of those multiple innovations. This is perhaps the most important aspect about capitalism viewed as a self-transforming system. The institutions of the market process generate sufﬁcient order to make it sensible to conceive of directional change, for chaotic systems do not evolve they drift, and the same durable institutions create the incentives to undermine order and generate uneven development. As in all evolutionary processes, nothing is constant absolutely, and what matters is the distribution of rates of change and particularly the fact that the institutional frame evolves far more slowly than the orders in terms of

126

Evolutionary Adaptation in the Economic Process resource allocation that it generates and transforms. Indeed, the disparity in degrees of transience is precisely the mechanism that generates structural change, the mechanism that Marshall captured in his famous periodization of changes in the economic order. Shackle (1965) expressed this very well. In commenting on the nature of Marshall’s periodization scheme and the idea of equilibrium, he writes, Equilibrium is a state of adjustment to circumstances, but it is ﬁction, Marshall’s own and declared ﬁction, for it is an adjustment that would be attained if the very endeavour to reach it did not reveal fresh possibilities, give fresh command of resources, and prepare the way for inevitable, natural, organic further change. (Shackle 1965: 36).

The point that Shackle is making is that order is not equilibrium, for each pattern of order contains within it the potential for its own destruction through the growth of knowledge and thus continual adaptation to the growth of knowledge. Self-organization through the market process generates self-transformation of those patterns of order. It is this sense of transformation from within that is so distinctive in modern innovation-led economies, it is exactly the point captured by Schumpeter in the notion of creative destruction (Schumpeter 1944; Witt 2002). In the remainder of this chapter we explore the idea that an evolutionary framework provides a productive way of understanding the adaptive, transformative properties of modern economies. What is it about the process of modern capitalism that makes it so revolutionary, perhaps too revolutionary for its own long-term good? Our suggested answer is that the dynamic of modern capitalism lies in the combinatorial growth of knowledge and investment opportunities combined with the instituted frameworks of the market economy, which taken together simultaneously stimulate and enable entrepreneurial activity. These elements are the building blocks for the evolutionary method, the framework of which is premised on three concepts: variation within a population, selection within and across populations, and the development of further variation within and between populations. Evolution is a three-stage process (Foster and Metcalfe 2001). The ﬁrst two concepts are not enough; no evolutionary explanation of economic transformation can dispense with an account of the processes that regenerate variety or fail to recognize that variety generation in biology will have little or no connection with variety generation in the economic domain. The development of individual entities is an integral and indispensable element in any causal theory or adaptation based on the evolution of individual characters jointly with the evolution of populations. Contrary to a prevailing viewpoint, development and selection are not competing, mutually exclusive explanations of adaptation rather they are the two different levels of explanation needed for any adaptive theory to be credible

127

Evolutionary Adaptation in the Economic Process (Walsh 2003).3 The necessity of this dualism follows, of course, from the fact that selection winnows variation and, unless that variety is replenished, evolution comes to a full stop. That we cannot ignore the processes by which selective characteristics are acquired by individuals opens up evolutionary explanation to the central importance of innovation processes and the signiﬁcance of emergent novelty. Consequently, in the following we will treat the question of adaptation from four perspectives. First, we explore the population method and its companion the replicator dynamic the building blocks of a theory of adaptation. This also allows us to elaborate upon the Fisher/Price dynamics of population evolution. Second, we apply this accounting scheme to a ‘one ﬁnger piano exercise’, the relation between the ‘logistic principle’, and the dynamics of population adaptation. Third and ﬁnally, we make the connection between adaptation of populations and complexity thinking. However, before we undertake these three tasks it will be useful to consider some evidence on the ubiquity of adaptation in the economic process.

6.2

Some evidence for structural adaptation

It should not be necessary to belabour the evidence in favour of the ongoing structural transformation of economies as they develop and grow, the support for this most important of stylized economic facts is conclusive (Pasinetti 1981; Freeman and Louca 2001; Cornwall and Cornwall 2002). At whatever level we focus we can trace the pattern of structural change of the emergence of new activities and the disappearance of obsolete ones. Indeed, to an earlier generation of growth economists nothing was more natural than to point to the changing composition of economic activity both qualitatively and quantitatively with the passage of time. Depending on the level of aggregation we choose we ﬁnd a very different pattern of transience, and in general, the more we aggregate the slower the process appears and the more we hide the evidence for economic evolution. Thus Colin Clark’s (1940) distinction between primary, secondary, and tertiary activities provides one high level classiﬁcation within which we can observe the structural shifts between agriculture, industry, and service activities over long developmental sequences. These shifts map onto the changing urban–rural balance so that it has become natural to deﬁne developed economies as ones in which the majority of the population live in towns and cities and is engaged in non-agricultural pursuits. When we turn to the ﬁner classiﬁcation of industries, recognizing that the list also evolves over time, we ﬁnd evidence of more rapid change. Fabricant (1940) summarizes the issue with clarity based on intensive empirical investigation into the US economy.

128

Evolutionary Adaptation in the Economic Process When we turn from the averages and concentrate upon the movements of manufacturing production in individual industries, we ﬁnd sharp differences in the secular rates of change in the physical output of these industries. In every period, some decline, some forge ahead, and only a few industries follow the general trend of manufacturing output. These disparate rates of growth affect, and are affected by changes in the structure of industry, in technical processes, in the kind of goods produced, and in the distribution of employment. (Fabricant, italics added, 1940: 9)

No better statement could be found of the themes of this chapter but it is a statement that just as readily could have ﬂowed from the pen of Schumpeter ([1911] 1934, 1944) or Kuznets (1954) or Burns (1934), economists who deeply understood the transformative nature of capitalism. The crucial insight is that systems transform through the emergence of differential growth rates for different activities and that these growth rate differences are interdependent and mutually determining. What is to be explained in the theory of transformation and adaptation is not why growth rates might be the same but why they are invariably different. For it is upon the diversity of growth experience that any understanding of economic evolution is premised. It is when we look within industries and their submarkets that we ﬁnd the most persuasive evidence for the self-transformation of the economy. Largely under the pressure of innovation, ﬁrms experience wide variations in rates of growth and decline combined with substantial rates of entry and exit. Utterback (1994), for example, charts the life cycle evolution of several industries, including the automobile industry in which an initial burst of entry is followed by a ‘shakeout’ as the industry consolidates around a small number of dominant ﬁrms. Christensen (1997) provides further examples of this competitive dynamic in which adaptation is forced by what he terms ‘disruptive technologies’ reﬂecting a set of concerns that occupy a central place in much of the literature on innovation and technology management (Georghiou et al. 1984; Foster 1986). Thus Christensen charts the entry of new ﬁrms exploiting a different innovation trajectory (based on internal combustion rather than steam power) in the mechanical digger industry and explains how the incumbents were unable to adapt effectively to this challenge and declined and disappeared. The story of the effect of the computer on the typewriter industry (Utterback 1994), of the semiconductor on the thermionic valve industry (Braun and McDonald 1978) are equally well-known examples of economic adaptation through self-transformation. These are essentially evolutionary stories of the competitive process in which one set of business activities is replaced by a different set in an innovation-induced differential growth dynamic. At this micro level the divergence in rates of change are more obvious as are the qualitative variations in the nature of the products, their end uses, and their methods of production.

129

Evolutionary Adaptation in the Economic Process To provide some different evidence on the empirical importance of growth rate diversity and structural change we use an National Bureau of Economic Research (NBER)–Centre for Economic Studies (CES) Manufacturing Productivity data-set covering 459 four-digit Standard Industrial Classiﬁcation (SIC) industries over the period from 1958 to 1996.4 Relative to most macro datasets, this provides highly disaggregated information although from another perspective, that of within market competition, it can be argued that it is still highly aggregated relative to the level of individual products, markets, and ﬁrms where evolution makes its clearer mark. While this data-set can still be used to identify some features of structural transformation, much remains hidden. This data allow us to compute the shares of each industry in total employment and total output (measured by deﬂated shipments) together with the levels and rates of change of labour productivity. If structure is changing the ﬁrst place to look is in terms of the patterns of the employment and output shares and the changes they evince over time. In semi-stationary, proportional growth these shares will be constant, which is only possible if all industries grow at the same rate and if all rates of productivity growth are the same. This is a hypothesis that we can reject with conﬁdence. In the absence of any structural change the employment share structure in the base year will exactly predict the employment share structure in all subsequent years, and similarly for the output shares. Figures 6.1a and 6.1b show the consequences of using the output and employment shares in the base period to predict the corresponding shares in the years up to 1996. Each graph shows the correlation coefﬁcient between the shares in year t and the shares in the base year for employment and output, respectively. With proportional growth these correlation coefﬁcients would remain constant at unity but as we see they decline virtually monotonically becoming progressively weaker as time passes. Also shown in Figure 6.1c are the results of a different test carried out, namely the correlation between employment and output shares over time and as shown, this also weakens but less dramatically a reﬂection of the divergent rates of labour productivity growth between the industries. An alternative way of measuring the rate of structural change is to compute the variation over time for the Herﬁndahl indices of employment and output, each index measuring the relevant ‘average’ share in the population. In proportional growth these indices would be constant. Figure 6.2 shows the variation over time in the Herﬁndahl index for employment shares, H ¼ e2j . This index measures the average employment share at each date. From Figure 6.2 we see persistent evidence of structural change in the economy’s employment pattern. The rate of change of the Herﬁndahl is readily seen to be proportional to the covariance between employment shares and employment growth rates,

130

1.0

1.0

(base year = 58)

Correlation zt zt+n

Correlation et et+n

(base year = 58)

0.9

0.8

0.8

0.6

0.7

(a)

58 60 62 64 66 68 70 72 74 76 78 80 82 84 86 88 90 92 94 96

58 60 62 64 66 68 70 72 74 76 78 80 82 84 86 88 90 92 94 96 59 61 63 65 67 69 71 73 75 77 79 81 83 85 87 89 91 93 95

Correlation et zt et+n zt+n

0.8

(b)

59 61 63 65 67 69 71 73 75 77 79 81 83 85 87 89 91 93 95

(base year = 58)

0.7

0.6

0.5 58 60 62 64 66 68 70 72 74 76 78 80 82 84 86 88 90 92 94 96

(c)

59 61 63 65 67 69 71 73 75 77 79 81 83 85 87 89 91 93 95

131

Figure 6.1 (a) Correlation of employment shares; (b) Correlation of output shares; (c) Correlation between employment and output shares

Evolutionary Adaptation in the Economic Process

0.4 0.6

Evolutionary Adaptation in the Economic Process

0.0048 0.0046 0.0044

H

0.0042 0.0040 0.0038 0.0036 0.0034 58 60 62 64 66 68 70 72 74 76 78 80 82 84 86 88 90 92 94 96 59 61 63 65 67 69 71 73 75 77 79 81 83 85 87 89 91 93 95 Year Figure 6.2 Adjusted Herﬁndahl index, 1958–96

dH ¼ 2ej nj n ej ¼ 2Ce ðe nÞ dt where nj is the growth rate of industry employment and n ¼ ej nj is the aggregate employment growth rate across all the industries. Consequently, the Herﬁndahl index is rising or falling as employment shares are positively or negatively correlated with the growth rates of employment in each sector. The rather dramatic fall and rise of the index provide clear evidence against the hypothesis of proportional growth.5 If structural transformation is pervasive and ongoing it can only be because the forces shaping the evolution of demand and the development of technology in the various industries are operating unevenly. Any theory of structural transformation must be capable of connecting together these ‘demand and supply’ forces to show how the evolution of individual industries is connected to the evolution of economy as a whole. The same data-set can be used to explore the rate of transformation within the US manufacturing sector in terms of the pattern of productivity growth that we know already must be divergent given the evolution of the output and employment shares. In Figure 6.3, we show the trend in the growth of labour productivity for US manufacturing as a whole. This series we have constructed by

132

Evolutionary Adaptation in the Economic Process 200 180

Productivity

160 140 120 100 80 60 58 60 62 64 66 68 70 72 74 76 78 80 82 84 86 88 90 92 94 96 59 61 63 65 67 69 71 73 75 77 79 81 83 85 87 89 91 93 95 Year Figure 6.3 Employment share weighted labour productivity

weighting the industry productivity growth rates by the corresponding shares in output. There is undoubted progress on average but around this average there is a wide dispersion of rates of productivity growth in the individual industries, as shown in the frequency distribution, Figure 6.4. Taking all the industries together, sixty-ﬁve have negative productivity changes over the period while the

70 60

Frequency

50 40 30 20 10 0

−1 0

1

2

3

4

5

6

7 8 9 10 11 12 13 14 15 16 17 18 19 20 Productivity growth

Figure 6.4 Distribution of percentage productivity change, 1958–96

133

Evolutionary Adaptation in the Economic Process mean percentage increase over the period for all the industries is 22.8 per cent. The highest percentage increase in productivity over the period is 2809 per cent in the computing sector. Fabricant further explored the relation between productivity growth and output growth, in essence an assessment of Adam Smith’s central theme that the division of labour is limited by the extent of the market. The central idea is that the rate of productivity growth for some activity, the outcome of processes of innovation, should be positively related to the rate of growth of its output and thus its market, the outcome of the process of self-transformation. To test this we compare the rates of growth of employment (n) with the corresponding rates of output growth (g) in a regression of the form n ¼ Æ þ g. The constant Æ measures the rate of productivity growth that is independent of the growth of output while the coefﬁcient measures the direct effect of output growth upon productivity growth. A check on the Fabricant relation for all the industries in this data-set is provided in Figure 6.5a where we show the regression line across the set of industries. The estimated equation is n ¼ 1:55 þ 0:57 g (0:37)(0:7)R2 ¼ 0:67 from which we infer that on average a 10 per cent increase in aggregate output is associated with a 4.3 per cent increase in labour productivity. In Figure 6.5b we show the scatter plot of the individual regression coefﬁcients, one for each industry to demonstrate a considerable diversity of outcome for the employment, output growth and productivity growth, relations. Figure 6.5b gives the scatter plot of the estimates for Æ and in those 419 sectors where the estimates are signiﬁcant at a 5 per cent conﬁdence level.6 With one exception all the coefﬁcients are less than unity conﬁrming the presence of a positive relation between output growth and productivity growth. To a much lesser degree the Æ coefﬁcients are positive since there are a substantial proportion of the industries where there has been measured technical regress. It will be apparent that there is considerable diversity in the technical progress functions at industry level but no apparent correlation between the estimated values of Æ and for these progress functions. The Fabricant Law stands up remarkably well as a robust empirical descriptor of the relation between technical progress, investment, and the growth of the market. It is not our purpose to explore here the origins of the differences in the productivity output growth relations summarized in Figure 6.5b, for that would be a major undertaking. Rather we turn our attention to how the evidence on structural change and growth rate diversity can be used as the building block for a theory of self-transformation.

134

Evolutionary Adaptation in the Economic Process

Employment growth

5

−0 n = −1.55 + 0.57g (0.37) (0.7) R 2 = 0.67

−5

2

(a)

4

−10

6 Output growth

8

10

1.2 1.0

β

0.8 0.6 0.4 0.2 0.0 −3 (b)

−1

1

a

3

5

7

Figure 6.5 (a) The Fabricant relationship; (b) Fabricant coefﬁcients for manufacturing

6.3 Accounting for evolutionary adaptation All of the examples provided in the previous section are cases of structural change within some suitably deﬁned population within an economy, and it is this property that opens up an evolutionary approach to economic adaptation. In economic terms the population adaptations reﬂect the ongoing redistribution of resources and consumer demand across the relevant markets. It is within markets that the causes of differential growth are found and it is market processes that establish the order, which is necessary before purposeful change can be conceived and implemented. Thus evolutionary adaptation in economic populations is essentially a ‘market’ story but it is a dynamic story one that reﬂects the creative, restless nature of a knowledge-based capitalism. Ever since R. A. Fisher (1930), the population method has been recognized as the central frame for any

135

Evolutionary Adaptation in the Economic Process evolutionary theory (Mayr 1982) for it is in the context of populations that we deﬁne adaptation and embed the notions of variation and selection. It is important from the beginning to inject some structure and clarity into the notion of adaptation. We need to distinguish the process of adaptation from the state of adaptedness or maximal ﬁtness for purpose between individual and environment, and we also need to distinguish adaptation of individual entities from adaptation of populations. In a state of adaptedness there is no scope for further evolution in the existing environment, given the existing properties of the population members. Adaptation as process, however, does not require either that that process has terminal states or that these states are stable in an appropriate dynamic sense. Adaptation is nothing more than directional change in populations imparted by variation and selection processes to better ﬁt the constituent entities into their environment. Adaptive processes can be open-ended; indeed, we shall argue that the lesson from complexity in the economic sphere is that they are precisely so because of the presence of development processes that create innovation and new knowledge from within. Adaptationism, in the sense we mean it, is a short-run account of evolution from which accumulates longterm change. Thus adaptation of populations is a process of changing the structure of those populations in such a way that the relative importance of their constituent entities changes as a result of environmental, selection ‘forces’. Evolutionary adaptation in one guise is perfectly consistent with individuals that are invariant in their properties; it is a type of process that explains change in structure but not change in components.7 But this would be a too limited framework to deal with economic change. We must also add changes in the individual entities through innovation, entry and exit from the population, and indeed their combination or dissipation through merger or divestiture if we are to have a complete framework for the study of adaptation. More crucially, we cannot use the population method unless we have reason to believe that there is prior variation in the population and the only causal ground for this belief is that there is differential adaptation, whether active or passive, at the individual level. Innovation processes in general are the developmental processes that lead us to the conclusion that population-level adaptation is premised upon adaptation at the individual level. Thus, the population method is naturally statistical in that it deals with the relative frequency of the activities in the population from which perspective it is natural to use the change in the moments, cumulants, or other statistics to characterize population-level change (Horan 1995). This population method is also incompatible with the idea of the representative agent, if by that we mean the uniform agent, and it is also incompatible with any macroeconomic explanation of growth and development. The aggregate economy perspective prevents the identiﬁcation of transformations for these are necessarily relative movements that are aggregated out of the picture when we focus on GDP. It loses sight of the fact that growth in some activities necessarily involves the decline of other activities, and it necessarily misses the

136

Evolutionary Adaptation in the Economic Process creation of new activities and the elimination of old ones, all the painful adjustments that characterize restless capitalism. This does not mean that we cannot produce meaningful aggregate statistics rather that any explanation of those statistics is not to be had by treating those statistics as ‘real’ entities. Indeed, because evolutionary explanations are naturally statistical, they embody the idea that change is premised on variety, that outliers matter greatly in explaining the course of adaptation. Thus growth is not generated at the macro level, and the aggregate growth rate is a statistical construction in relation to which no industry might grow at the average rate for all industries. The notions of semi-stationary growth or proportional dynamics have their uses but understanding innovation-led development is not one of them. Consequently, the system evolves and with it the relations between averages and their components, the induced changes in structure continually redeﬁne the relations between activity and development. We now provide an outline of an accounting framework to capture the possible processes of population change in an economy. A population may be deﬁned at any level but to ﬁx ideas it is useful to focus on the concept of an industry, deﬁned as a group of ﬁrms producing similar products and drawing on the same markets for resources and customers. The accounting has no explanatory value; of itself it is no more than a ﬁling system to place different notions of population change in the proper relation one to another. However, it is a basis for incorporating any explanatory theory of economic adaptation, all of which must ﬁt with the constraints of this accounting. Let us commence with an abstract description of a general evolutionary process located at the level of an ‘industry’ and its associated ‘markets’. A population consists of a set of ﬁrms, each one associated with its set of distinguishing characteristics. The individual characteristics of the ﬁrms may or may not have economic selective value and, in general what is evaluated by the environment is the bundle of characteristics that deﬁne each member of this population. Moreover, the values of each individual characteristic may be constrained by values of other characteristics such that the overall bundle is a reﬂection of prevailing trade-offs and constraints within the entity. This population method is a remarkably general tool of analysis in that it provides an exhaustive way to account for all the changes that occur in this population over a time interval. Four processes exhaust the possibilities of population change: . Pure replication of the continuing (surviving) ﬁrms in the industry either in terms of changes in the number of ﬁrms or, as here, in terms of changes in the scale of activity of each ﬁrm. . The entry (birth) of additional ﬁrms in that population, each one deﬁned by the characteristics in question and making a contribution to the scale of activity.

137

Evolutionary Adaptation in the Economic Process .

The exit (death) of ﬁrms in the population, taking with them a fraction of the output of the population.

.

Innovations (mutations) in the characteristics possessed by the continuing ﬁrms so that they vary individually, to which we can add innovations in organizations that combine together different ﬁrms into a new whole or the converse their fragmentation into separate businesses.

By partitioning the population of entities into these categories we can perform a complete analysis of the change in the population between the two dates. A pure replication process would focus exclusively on the changes in respect of the continuing ﬁrms and, indeed this is a standard method of much evolutionary analysis. It is what we shall focus on in our discussion of the logistic process below. In economic terms this is not entirely satisfactory for it loses sight of extremely important processes in relation to the birth and death of ﬁrms or their recombination. Innovation too at the level of activities is an essential element in economic evolution for it corresponds to a change in the characteristics of the entities and thus a change in the distribution of selective advantage in the population. We deﬁne development processes to include innovation in continuing ﬁrms, entry of new ﬁrms (often premised on innovation), exit of established ﬁrms, and the recombination or dissolution of ﬁrms. Adaptation is then a combination of replication and development processes. Since adaptation is a reﬂection of differential growth our accounting frame must provide a way of linking structural change to growth rate diversity. We can make this a little more precise as follows.8 Consider an industry consisting of a population of ﬁrms and measure the relative importance of each ﬁrm by its proportional contribution to total output. Deﬁne a surviving ﬁrm as one that neither enters nor exits in the relevant time interval. Construct output shares such that ci is the share of a surviving ﬁrm in the output of the surviving ﬁrms and let si be its corresponding share in total output. Let e and n respectively be the instantaneous rates of exit and entry in the population measured in terms of their proportional contributions to total output gained and lost by the respective ﬁrms. Let g be the instantaneous growth rate of total output and let gi be the rate of growth of the output of the surviving ﬁrms’ net of entry and exit. Then it can be shown that9 g ¼ gc þ n e

(1)

linking the two output growth rates to the rates of entry and exit. It follows from the deﬁnition of the output shares that dsi ¼ si (t)(gi g) ¼ si (t)(gi gc n þ e) dt and that

138

(2)

Evolutionary Adaptation in the Economic Process dci ¼ ci (t)(gi gc ) ¼ ci (t)(gi g þ n e) dt

(3)

In these two equations the growth rate averages are deﬁned by g ¼ si (t)gi (t) and gc ¼ ci (t)gi (t), respectively. Equations (2) and (3) are the primitive replicator dynamic relations in the population for they tie together the diversity of growth rates with the pattern of structural change. They deﬁne a distance from mean dynamic for the population, in which its adaptation is governed by the distribution of the growth rates around the population average. As an accounting frame these relations have no causal signiﬁcance. Rather they are the frames in which to generate an economic explanation of the causal chain that leads to the distribution of the growth rates. Structural change is caused by the interaction between those causal factors that lead to growth at the level of the individual ﬁrms and at the level of the selection environment. If we change the causal characteristics of any ﬁrm or the selective characteristics of the environment, we generate a different distribution of growth rates and a different pattern of adaptation according to relations (1) to (3). Selection, entry, exit, and recombination exhaust the structural components of population change but for a complete description of the dynamics of adaptation we need also to add the effects of changes in the individual growth rates. If, to illustrate, we focus on the average growth rate in the population of surviving ﬁrms, its evolution over time has two components as follows. Since gc ¼ ci gi , the rate of change of the average growth rate of the survivors becomes X dgi X dgi dg X dci gi þ ¼ Vc (g) þ ci ci ¼ dt dt dt dt

(4)

Where, Vc (g) is the ci weighted variance of the growth rates across the population. The ﬁrst term is that due to selection and the second is that due to innovation. This form of the selection effect is known as Fisher’s fundamental theorem, after its originator, the distinguished biologist and statistician R. A. Fisher (1930). If the growth rates are interpreted as ﬁtness values, then selection has the effect of increasing average ﬁtness in the population. However, its signiﬁcance lies in it being a very special case of a much wider principle, Fisher’s principle (Metcalfe 1998), namely that the statistical variability within the population accounts for the rate and direction of evolutionary change—the variation-cum-selection view of development.10 In fact equation (4) is a variant of a well-known result in evolutionary theory called the Fisher/Price equation (Price 1970; Metcalfe 1998; Frank 1998; Gintis 2002; Andersen 2004), which decomposes the change in average value into additive effects due to selection and innovation. As Andersen (2004) suggests, the Price equation ‘eats its own

139

Evolutionary Adaptation in the Economic Process tail’, an attribute of considerable signiﬁcance in the analysis of multilevel evolutionary processes. By this is meant the fact that each gi can be treated as a sum across a population so that the equation (4) can be applied recursively to sets of nested populations at multiple levels. It means that we can decompose any population change into change between subpopulations and change within subpopulations in an identical fashion so that at each level of aggregation we can reﬂect the same forces of adaptation and development. So far we have developed an accounting scheme for any evolutionary process within some arbitrary population, a hypothetical industry. To repeat, it is entirely neutral as to the explanation of the growth rates, innovation rates, and entry, exit, and recombination rates in any population, providing the framework into which more substantive theories can be located, compared, and tested. What gives the scheme its content in any case is the particular explanation that causally links the characteristics of the entities to the differential growth of their scales of activity. These schemes may be deterministic, stochastic, or combinations of both in their causal structure but they will all fall within the accounting net for population change in which differential growth gives rise to adaptation. Of course what gives this accounting an economic signiﬁcance is that these processes of innovation and development take place within markets and the wider institutions in which markets are regulated. Market processes shape exit and entry, recombination and differential growth in general so that what matters in a theory of adaptation is the market process that drives these different phenomena.

6.4

The logistic principle

In the empirical study of industry adaptation referred to earlier it is frequently found that a logistic curve provides an adequate summary of the time evolution of a population. This is so, for example, in the literatures that deal with the diffusion of innovation, processes of technological substitution, and the evolutionary, life cycle dynamics of industries. That the logistic should apply to so many empirical domains is of interest in itself and it reﬂects a deeper theme, namely that the logistic is a population phenomenon premised on the differential growth of the relevant members of a population. Indeed, we shall now show that the logistic process is a deep signature of any evolutionary process within a population that is governed by a variation and a selection dynamic. It is the connection with the evolution of populations that helps explain the empirical ubiquity of the logistic curve precisely because all processes of structural change and adaptation are population-based phenomena.11 However, we will also show that a logistic process need not generate the familiar ‘S’ shaped logistic curve expressed as a function of time, indeed this general logistic principle may be, and generally will be, associated with distinctly non-logistic time proﬁles, for

140

Evolutionary Adaptation in the Economic Process the changing relative importance of the entities in a population. The logistic process outlined here has, therefore, a degree of generality that the logistic curve does not possess as a signature of adaptation. To take one example, Marchetti and Nakicenovic (1979) in a well-known study of the evolution of populations of rival energy technologies recognized that in the substitution process the logistic phase of growth and saturation over time is normally followed by a phase of decline such that a logistic trend only captures the ﬁrst part of the evolutionary process.12 That growth may be followed by decline is unexceptional but that both aspects of population evolution could be captured in the same general logistic process is perhaps worthy of further investigation. We now show that this logistic principle depends on the distance from mean dynamic and its close relatives, the replicator processes, that we have alluded to already in our accounting for evolution. To ﬁx ideas, consider a population of distinct entities, we might imagine that they are rival technologies. The measure of the scale of activation of each technology at date t is x(t).13 By the ﬁtness of each technology we mean the exponential growth rate of its scale of activation. The relative importance of each member of the population is deﬁned by its population share si (t) ¼ xi (t)=

X

xi (t)

(5)

and, to prevent the discussion becoming too complex, we ignore the possibility of new technologies entering the population or of existing ones leaving it.14 We have already seen how these relative measures of population structure are the central focus of any variation selection approach to evolution and we shall now show how their evolution obeys a logistic lawlike property. It follows as a matter of the deﬁnition of s(t) and g(t) that the dynamic process of selection for each entity in the population will obey the following replicator relation dsi ¼ si (t)[gi (t) gs (t)] dt

(6)

P P with the mean ﬁtness value deﬁned by gs (t) ¼ si (t)gi (t); si (t) ¼ 1. Equation (6) is the familiar equation of a replicator dynamic process and it exactly embodies the distance from mean dynamic principle alluded to earlier. Whatever may determine the evolution of the individual growth rates, equation (6) is a complete description of the dynamics of the population and it holds exactly whatever the nature of the population. If the individual ﬁtness values gi (t) are treated as independently determined constants, we have a pure sorting process, although, as long as the individual growth rates differ, the Fisher/Price theorem indicates that the average value gs (t) will not be constant.15 If the individual growth rates are interdependent

141

Evolutionary Adaptation in the Economic Process and simultaneously determined, we have a selection process proper within this population. On integrating equation (6) for each of the members of the population we have ð t si (t) ¼ si (0) exp [gi (t) gs (t)]dt 0

The important point about this expression is that each population share evolves according to the history of its ﬁtness relative to average ﬁtness in the population as a whole. This evolutionary dynamic is precisely a distance from mean dynamic. However, this way of describing the process of population adaptation can be expressed differently to expose the logistic principle contained within it. Take each si (t) then we can write for that technology the relation 0

gs (t) ¼ si gi þ (1 si )gsi P 0 0 and where the quantity gsi ¼ sj gj is a weighted average of ﬁtness values j6¼i across all the remaining technologies in the population, and it is different for each entity.16 Then we can rewrite equation (6) as dsi 0 ¼ si (1 si )[gi gsi ] dt

(7)

In this expression si (t) is a logistic function of the distance function 0 (gi gsi ) ¼ Gi (t) and we can see this by integrating equation (7) to give si (t) ¼

1 ¼ Li (t) 1 þ Ai exp Di (t)

i (0) with Ai ¼ 1s si (0) determined as an initial condition. The integral function

Di (t) ¼

ðt 0

0

[gi (t) gsi (t)]dt,

we call the ‘Downie’ function, after the economist Jack Downie who explored this dynamic process, the transfer process he called it, in his pioneering exposition of competitive industry dynamics (Downie 1958). Thus the logistic process does not give the conventional logistic curve over time but rather a logistic expression that captures the distributed nature of evolutionary dynamics in a population. This relation applies to all the entities in the population and it follows that each of them must evolve along their own speciﬁc logistic curve as a function

142

Evolutionary Adaptation in the Economic Process 1•

s2(t )

•

1 2

g1 > g2

•

0

s1(0) Di (t )

Figure 6.6 Logistic process

of their individual ‘Downie’ functions as shown in Figure 6.6. The curves have an upper asymptote of unity and two such curves are shown relative to initial date t ¼ 0 remembering that the inﬂection point for each logistic curve always occurs when s(t) ¼ 0:5. Now as long as Gi (t) > 0 that entity is ﬁtter than ‘average’ and si (t) increases over time and the relative importance of that entity is increasing in the population. But clearly this cannot be so for all the technologies, some must necessarily be less ﬁt than ‘average’ and be declining in relative importance. Thus when Gi (t) < 0 the technology’s share moves down the logistic, and this will be 0 so when gi (t) has fallen below the population value of gs (t)—in Figure 6.7 this ** * occurs at time t , while at time t the value of D(t ) ¼ 0, and at this date the value of the relative share in the population has returned to its initial value. However, the important lesson from this exercise is that the logistic process does not in general generate a logistic curve measured against time simpliciter. In empirical terms this means that we cannot interpret the absence of the simple logistic trend curve as evidence of the non-operation of the logistic process. What is the condition for the logistic principle to support a logistic curve in the traditional sense? It is that we can express the Downie function in the linear form Di (t) ¼ Æi t. Clearly, this can only be true if there are no more than two competing members of the population and if the difference in their respective growth rates is a constant. This would be achieved not only in a sorting process with given growth rates but also in a selection process in which the growth rates are independent.17 Thus while the logistic principle is general, the instantiation of it in a simple logistic curve is very special. Figure 6.7 shows the more general relation of a population share plotted over time, with growth followed by decline and intermediated by momentary ‘saturation’. In part, this time proﬁle helps explain why other non-logistic ‘S’ curves such as the Gompertz of log normal ﬁnd frequent empirical representation, and it also helps explain the decline phase of an entity in the population which has often been observed in technological substitution studies.18

143

Evolutionary Adaptation in the Economic Process Of course this is no more than an accounting for evolutionary adaptation. The dynamic relations must hold for every population when we deﬁne the notions of population share and growth rate in the way that we have. It only becomes the basis for a refutable theory of evolutionary change when we impose a particular theory of why the individual growth rates differ and vary over time.

6.5

Adaptation, complexity, and the problem of knowledge

Let us set these formal arguments aside and turn to broader and deeper issues in relation to the adaptation of economies. The problem of economic adaptation and its relation with the growth of knowledge provides a natural focus for connecting together the replication and developmental stands of the evolutionary process. Recent debates, particularly following the publication of Kauffman’s The Origins of Order (1993), have brought the issue of the relation between self-organization, complexity, and population selection to the forefront of discussion (Depew and Weber 1996; Richardson 2001). We do not follow this discussion here but restrict the argument to the claim that complexity arises only in the course of the growth of knowledge and is intimately connected to the emergence of novelty in terms of new entities in a population or new selective characteristics for the existing entities. In terms of the seven ‘niches’ proposed by Depew and Weber for linking selection and complexity, we are ﬁrmly in the camp that sees self-organization and selftransformation as a uniﬁed process; complexity and selection are in this view inseparable and the link between them is the growth of knowledge.

si (t)

si (o) •

•

t **

t*

Figure 6.7 Logistic share proﬁle

At t , Gi (t) ¼ 0; at t , Di (t) ¼ 0:

144

t

Evolutionary Adaptation in the Economic Process The core of the argument is that the adjectives complex and complicated have quite different meanings when applied to situations that are difﬁcult to understand in their entirety. Many systems, the economy included, are very complicated, involving multiple connections between multiple components within partially overlapping sets of boundaries. An input–output model of interlocking industries or a computable general equilibrium model would ﬁt this description perfectly. However, such a system would not contain any features not already contained in its founding description. Although its dynamics may be technically complicated, every possible state is predictable in principle from knowledge of the antecedent states. In this sense time passes but nothing new happens, there are sequences of states but there is no history. By contrast, a complex system generates states that are not completely reducible to antecedent states, it creates genuine novelties from within and change from within depends on the emergence of new knowledge. A non-complex world would be a world of invariant beliefs, a world of the stationary or proportionally expanding state in which irrespective of its scale the structure of the economy is invariant over time. Moreover, it would be invariant over time because the scope for adaptation through selection is exhausted. This is a world without development in which there are no reasons to change the nature of economic activity, for if beliefs are stationary we can only conclude that knowledge is stationary too. A complex world, therefore, is not only self-organizing but also self-transforming, and it is this latter characteristic that forms the signature of economic complexity. It can never be a world in equilibrium, as the word equilibrium is normally intended, for in equilibrium all change is only explicable by extraneous forces not included in the determinants of the postulated equilibrium state. In short, when economies are out of equilibrium they stay out of equilibrium (Robinson 1974). Nevertheless, they always exhibit order and without that order would not be capable of coherent self-transformation. Hence, we argue that complexity is a problem in dynamics that transcends the complicated nature of the economic world, while recognizing that all economies are complicated in the sense that they are multidimensional and beyond the comprehension of a single mind. Perhaps, one of the most surprising aspects of capitalism from the eighteenth century onwards is how relatively few individuals have led the process of self-transformation, the crucial changes in beliefs are not initiated in a widespread fashion, a fact which points to the central importance of entrepreneurial activity and economic leadership to the complexity perspective (Witt 2000). Complexity analysis cannot avoid recognizing the role of socially situated individuals in the economic process, and the only purpose of introducing individuals is to recognize their different beliefs and states of knowing and how those differences are a basis for structural change in economic populations. Complexity is an antidote to more simple-minded versions of methodological individualism, or the extreme reductionism that is expressed in terms of the

145

Evolutionary Adaptation in the Economic Process methodological atomism, so prevalent in modern economics. Individualism remains a central tenet in any adaptive evolutionary account of an economy but it is a heavily nuanced version of individualism, one that places the individual and what the individual knows in systemic population contexts. A serious treatment of complexity therefore requires an understanding of the process that generate the components of an economy/society and the process that generate their interconnections, so that at a deeper level we can understand how complexity is itself created and renewed by economic and social processes. It is these processes of organizing complexity that lead us to a co-evolutionary account of the ongoing development of knowledge and the development of the economy. We believe that an evolutionary adaptive model of development, in which the emergence of macro structure arises from micro diversity and which recognizes explicitly the coordinating properties of markets, can explain best the dynamic features of modern capitalism, in relation to novelty, innovation, and structural change (Potts and Dopfer 2004). Self-organization brings with it self-transformation so that order and change are inseparable. Therefore, it seems that one of the most important insights that complexity analysis bring to the study of social phenomena is that complex economic systems must be understood as adaptive, evolving processes, in which adaptation of the economy is made possible by the adaptation of knowledge. The crucial point here is that complex systems are evolving systems in which the components, ultimately the individual actors, change their behavioural properties because they become more knowledgeable. A selftransforming system is renewed on a continuous basis; all complex, developmental processes have this property of self-discovery. An understanding of complexity, therefore, allows for a dynamic characterization of the economy, not one that is deterministic, predictable, and mechanistic but one that is process-dependent, organic, and endogenous (Arthur 1994). Built into this view is another deep evolutionary issue, namely that the components of a evolving system are different, they are ‘individuals’ in the proper sense of their distinctive properties, among which what they know as individuals is of paramount importance. Evolutionary change is impossible without variation and variation captures the element of individuality located within populations of other individuals. Few scholars would deny that economic and social evolution are contingent on the continued growth of knowledge but precisely who is said to know more when we link the growth of knowledge to the growth of the economy? Our answer is that only individuals can possess knowledge, that all changes in knowledge require changes in the state of individual minds, and that the correlation of knowledge across minds underpins the shared understandings necessary for coordinated economic and social action. Knowledge is always private and unarticulated; knowledge always implies a ‘knower’. Consequently, the total of what is known can at best only mean

146

Evolutionary Adaptation in the Economic Process the union of the vectors of the knowledge held by every individual within the relevant domain. In this aspect knowledge is a consequence of life; it is the dominant characteristic of human living. What is articulated in public is information, data open to sensory perception, and information is not knowledge, it is a representation of knowledge transmitted and received via the sensory apparatus. It is information that lies in the public domain and it is the transmission of information that is essential to the correlation of knowledge across individual minds. Thus the distinctions codiﬁed and tacit only apply to information not to knowledge. Yet, information can only be an imperfect representation of individual knowledge, depending on the respective capacities to encode and decode the implicit messages. There is always a tacit background to everything that we seek to express and interpret; as Polanyi (1966) insisted, we always know more than we can communicate through the senses. Strictly speaking, knowledge itself is never codiﬁed, it is only its imperfect representation as information that can so be digniﬁed. Codiﬁcation can be in terms of sounds or symbols but the dominant kind of codiﬁcation that has engaged modern scholars is the written, storable symbolic codiﬁcation associated with language and mathematics. Moreover, what is codiﬁed is, in part, an economic decision dependent on the scale at which the information is intended to be used (Cowan et al. 2000; Johnson et al. 2002). How does this relate to an adaptive, complexity view of the economy? Under capitalism much new knowledge results from the conduct of the market process as suppliers and customers interact and learn what to produce and from whom to buy. To this extent economically valuable knowledge is a product of coordination and can be expected to accumulate differently in different coordination systems.19 Consequently, it is not possible to ‘hold knowledge constant’ while allowing human economic activity to occur and to construct our analysis on this premise hides more than it illuminates about the dynamic nature of capitalism. It follows that every position of temporary economic order creates within it the conditions to change that order, and this is especially true of knowledge accumulated in the pursuit of innovation. No clearer concept in this domain can be found than that of the division of labour being contingent on the scale of the market. If selection accounts for differential growth of different activities, the differential extension of the market, then it opens up from within prospects for the further division of labour to different degrees and thus reshapes the forces impinging on selection (Young 1928; Richardson 1975). The consequence is that processes of economic selection guide and constrain the growth of knowledge just as much as the constraints on the growth of knowledge shape the process of selection. A selection theory that fails to account for these feedbacks will only capture a fraction of the evolutionary account of economic life.20 In the modern economy, of course, this is

147

Evolutionary Adaptation in the Economic Process reinforced by the practice of allocating a non-trivial portion of the economy’s resources to the acquisition of knowledge and the dissemination of information, together with its embodiment in the population via processes of education and training. One of the major steps in the evolution of capitalism, a step that makes the transition to complexity more apparent, has been the deliberate allocation of resources to research and development activities conducted independently from day-to-day experience of production, distribution, and consumption (Mowery and Rosenberg 1998). This was a key institutional development that enabled society to allocate its resources to the purposeful exploration of the unknown, which has had a profound effect on the conduct of economic life for it builds into the system a further powerful source of novelties. With the passage of time, an increasingly reﬁned division of labour has emerged in the production and application of laboratory-based knowledge, the reign of the ‘philosophers and men of speculation’. New ﬁelds of knowledge have emerged together with a complex skein of systemic interrelation between laboratories in ﬁrms, their customers and suppliers, universities, and other public and private research organizations.

6.6

The correlation of knowing

Since knowledge is always private and can only emerge as novelty in individual minds this poses a severe problem, ‘How is it that we can act in common in pursuit of human objectives if we do not understand in common?’ Quite crucially for the following argument, private states of mind are not accessible by any other individual and this carries with it an implication of great importance, an implication that is a natural, unavoidable consequence of the limited mental capacity of all individuals.21 Namely, that those processes, by which we come to know as individuals, are augmented greatly by instituted social processes that permit exchanges of information, representations of knowledge communicated between individuals such that they can lay claim to common understanding or social knowledge. Moreover, these processes of organizing the growth of knowledge are themselves the products of human agency. Through sense experiences of the knowledge of others, that is to say through the communication of information, our individual states of knowledge become interdependent. In this way, private knowledge of any individual is correlated, but not identical, with the private knowledge of others. The correlation of knowledge is the necessary condition for the extensive spread of knowledge and thus for the effectiveness of any activities that require individuals to act in concert. It is also essential for the growth of knowledge in the intensive sense in which any individual comes to know more. Thus, we are suggesting that what is known is private and that what is understood socially is necessarily systemic and emergent and, like all knowledge, continually evolving. This leaves open the ques-

148

Evolutionary Adaptation in the Economic Process tion of whether knowledge within individual minds is also systemic and evolving in an adaptive sense but this is a step we need not take here. The point is that social knowledge or understanding in common as it is better expressed is an order that self-organizes and self-transforms, and does so via variation, selection, and developmental processes. This extended reliance upon the testimony of others is one of the key factors in comprehending understanding as a complex system predicated on the knowledge of individuals and indeed in comprehending the nature of capitalism as a knowledge-based system. In this way, Adam Smith reaches the core of the matter identifying capitalism as a self-exciting system that has evolved and continues to evolve new methods of inventing novelties. It is this that makes possible an edge of modernity as a persistent feature of capitalism, a system in which fundamental changes are always in train, fundamental changes that arise from within the system. What Smith does not develop is how this growth of knowledge is to be coordinated, how individual knowledge is to be shared in the wider social context for this will determine how individual knowledge will grow. What is the instituted process that achieves for knowledge activities what markets achieve for conventional productive activities? If information ﬂow is to convey personal knowledge with sufﬁcient accuracy to achieve commonality of understanding, there must be vehicles of communication and common standards of communication, language, or other forms of symbolic representation, and agreed standards for the justiﬁcation of that which can be said to be known. Moreover, there must also be shared interpretive frames and theoretical schema to judge the content of information; otherwise private knowledge cannot develop into collective understanding. We must have shared grounds to agree what the facts are, and what facts are relevant to the issue in hand. As Nelson puts it, there must be ‘social technologies’ to make testimony possible (Nelson and Sampat 2001). In this regard, institutions and information technologies matter in three fundamental ways in relation to the connection between knowledge and understanding. First, they constitute the means to store and communicate information in general and the means to support particular patterns of interaction, ‘who talks to whom with what frequency and with what authority’, in a society. This is the question of language, commensuration, and symbolic representation in general. Different patterns of interconnection imply different distributed patterns of understanding and thus different paths for the adaptive evolution of knowledge. Nor are the patterns of interconnection given, they are to be explained as a process of institutionalization; whom we communicate with naturally changes as our individual knowledge develops, the networks and modes of intercommunication are themselves emergent properties of our kinds of knowledge-based economy. Second, institutions embody the rules, the standards of socially agreed belief, that are the means to accumulate justiﬁably true knowledge in relation to science, technology, as well as organization and social

149

Evolutionary Adaptation in the Economic Process discourse. It is the institution of understanding in common that makes economic and social life possible while simultaneously constituting a powerful engine for the differential growth of personal knowledge. North (1990) is correct in arguing that institutional rules constrain behaviour by facilitating the growth of common understanding. However, these same rules are to a degree enabling and facilitating in that the spread of understanding opens up opportunities for the further growth of private knowledge. Third, and equally crucial, has been the invention and adoption of instituted standards or norms to distinguish reliable from less reliable knowledge. The process of establishing error, of identifying mistakes through criteria for falsiﬁcation and rejection have provided the critical edge that combats the problem of super fecundity, the problem of being unable to distinguish which of the too numerous rival courses of action to follow (Mokyr 2002). Fundamental changes in the nature of the processes of knowledge generation have greatly enhanced the basis for social testimony to shape what is understood in common. Indeed, perhaps the distinguishing feature of a traditional society (non-complex?) is that testimony is limited almost exclusively to direct, immediate social contact, whereas in advanced knowledge-based societies, communication is much more impersonal, at a distance; and, we may note, that impersonality undoubtedly helps soften the bounds of tradition. Consequently, whether we realize it or not, our daily lives in modern capitalism rely upon the information provided by many individuals whom we never meet let alone know of. This is the impact of the long sequence of novel technologies that begins with printing and trade in books and newspapers, and leads on via the telegraph and telephone to wireless communication and the Internet. Exchanges of information are thereby quicker, denser, and broader so changing fundamentally the generation of social testimony. In turn, this changes the information stimuli and their distribution across individuals, thus reshaping the emergence of understanding and the development of private knowledge. This is why understanding is the product of a complex, information network system, in which every new thought is a novelty opening up potential new dimensions of economic and social behaviour. Modern science is one canonical example of this process, with the academic journal and conference providing the instituted vehicles to transmit information and correlate understanding, but the point applies far more generally, particularly to economic entrepreneurship. So is the modern economy, in which the stimuli to innovation are broadly and deeply distributed and lead us directly to the foundations of capitalism as a restless system, a system in a continuous state of self-induced adaptation. To summarize, human interaction generates directly and indirectly a ﬂow of information between individuals who, at best, treat that information as a representation of the knowledge of others, and interpret it through confrontation and association with their own sense of knowing. In the process, the disjunctures that arise are a powerful stimulus to new thoughts, new beliefs, and new know-

150

Evolutionary Adaptation in the Economic Process ledge. Information ﬂow may change the knowledge states of the recipients but there can be no expectation that the change of knowledge will be complete, that it will be identical for all recipients, or that it will be unanimously accepted. Indeed, it is the fact that information is not interpreted in uniform fashion, that there is disagreement, which is essential to the continued growth of knowledge and thus understanding. Understanding cannot be completely correlated or nothing new will ever be known. Oddly enough, in science, invention and entrepreneurship, the big ‘prizes’ go to those who have disagreed most fundamentally with prevailing patterns of understanding whether they realized it at the time or not. A world, in which all private knowledge was identical, would be a world in which the problem of knowledge had ceased to exist, and the individual had ceased to exist also; such a world would not be complex. This is why the growth of knowledge depends on individual differences and why it is a variation–selection process. In regard to disagreements, it is important to recognize that they may have no grounding in the current state of understanding. The ability to imagine and conjecture in advance of a supporting base of evidence is the human characteristic par excellence that most directly leads to the generation of novelty, innovation, and thus complexity. As with individual knowledge, shared understanding is an open system, it is emergent, it is adaptive, it can grow combinatorially fast, at least in local domains. If it has no rest points or stable, invariant attractors then neither will an economy. It can undergo subtle changes as information percolates across networks of relationships or it can undergo sweeping changes that take understanding into entirely new dimensions. To this degree understanding is unstable; populations of beliefs can be invaded by new entrants and some existing beliefs abandoned. Yet the kinds of understanding we develop and share are not unconstrained. The path of understanding is we suggest chreodic and it is channelled primarily by the particular nature of the social, economic, and institutional context. The conventions as to which information is made public and in relation to who could communicate with whom about what are deeply important for the growth of knowledge. Institutions, conventions, social standards, and the system shapers can suppress the generation of understanding just as they can enable it. We can ﬁnd here the unpredictability of knowledge accumulation, the uneven nature of knowledge accumulation, and the corollary, the unpredictability of the surface forms of capitalism in terms of what is produced and consumed. But this is only to say that all orders are distributed kinds of organization whether as markets or as organizations simpliciter. Hence the restless nature of ﬁrms and whole economies: capitalism can never be at rest because understanding and knowledge are never at rest and never can be given the rules of the game; ‘equilibrium of knowledge’ is an oxymoron. These considerations tell us a great deal about the unique properties of capitalism as an adaptive, complex, knowledge-based system. It is a system for generating business experiments based on the accumulation of scientiﬁc,

151

Evolutionary Adaptation in the Economic Process technical, organizational, and market knowledge. Business conjectures create sequences of new problems to be solved, give meaning to entrepreneurship, and give to the ﬁrm as organization the unique role of combining the multiple knowledge elements that are needed to innovate successfully. In being a problem-solving system, any organizational form such as a ﬁrm or a market is necessarily a problem-generating system. Markets and ﬁrms, the twin forms of organization, provide the context in which economic knowledge is generated and understandings correlated. Of course, it is important to emphasize that the growth of understanding cannot be random without further processes to focus that randomness to good effect. Random systems do not evolve, they drift. We make rapid economic progress precisely because the underlying processes of variation are guided that they are chreodic, in the sense discussed earlier, and explore only limited regions of the space of possibilities. Nor can the growth of understanding be entirely deterministic for it involves choice, judgement, and creativity in the sense of the expansion of thought and action into new dimensions and spaces. For this reason alone genuine uncertainty is an unavoidable element in the complexity picture for only when the space of possibilities is closed can probability judgement and its formal calculus be entertained (Shackle 1958). One ﬁnal aspect of this scheme is worth brief mention, the connection with enterprise and ‘the’ entrepreneur. If economic order depended on the correlation of knowledge and only the processes of correlation, it is clear that evolution would come to a halt and with it economic adaptation. The paradox is that knowledge has to be de-correlated if the new understandings essential to economic adaptation are to take place. The agency by which this is achieved is what we normally term entrepreneurship. Here we ﬁnd a new perspective on the Schumpeterian entrepreneur, whose chief role becomes the de-correlation of private knowledge, the sowing of doubt where previously there was understanding in common. Hence, the emphasis on novelty, on challenging existing practices and understandings that is typical of the adaptive Schumpeterian model and, indeed, typical of the Kuhnian notion of the paradigm-breaking scientist.22

6.7

Conclusions

The emergent theme of this chapter is that adaptation is inseparable from the growth of knowledge and that market-based dynamics give adaptation a form that is central to change in modern capitalism. What we observe as the reallocation of resources is the consequence of different beliefs, the imagination that the economic world can be organized differently. This is the primary reason why economies evolve and adapt, they are instituted variation, selection, and developmental processes. In the emergence of new conjectures the

152

Evolutionary Adaptation in the Economic Process growth of knowledge is of central importance. We have drawn a distinction between private knowledge and social understanding, and suggested that shared understanding is an emergent, self-transforming, complex system predicated on processes and institutions for generating social testimony, that is to say for correlating private knowledge. Understanding at multiple levels is the necessary condition for an economy to operate and the operation of the economy stimulates the growth of new knowledge and thus the emergence of new understandings. This is a system, therefore, in which neither the component parts, the knowledgeable individuals, nor their patterns of intercommunication, the social relations, are given. The point about the complex knowledge system is that it is evolving in parts and connections, it is always becoming something else but not in a way that anyone can predict. That the knowledge system is complex and the fact that any economy is knowledge-based suggests that the degree of complexity of the former largely shapes that of the latter.

Notes 1. We are grateful to Maureen McKelvey, Keith Smith, and the other contributors to this volume for their comments and suggestions, most of which we have been able to take into account. Sins of omission and commission remain the author’s alone. 2. An economy expanding in all respects at a uniform rate is in effect a stationary state, a state deﬁned by proportional dynamics. It is, of course, a ﬁction useful to a degree but severely limited in its relevance to growth and development as it is experienced. 3. For further elaboration on the history of the evolutionary developmental biology ‘evo-devo’ controversy see Love (2003). 4. This data-set is published on the NBER website (www.nber.org/data). 5. The measure of the Herﬁndahl index is sensitive to the level of aggregation and the number of sectors included at each level. We have scaled its value to lie between zero and unity. Unscaled, its minimum value is 1/459 which equals 0.002179. 6. Because of the statistical problems that arise in directly regressing productivity growth on output growth all our results are based on the regression of employment growth on output growth from which the productivity relation is inferred. 7. R. A. Fisher drew the not entirely appropriate analogy with statistical thermodynamics. On this see Walsh (2003). 8. To keep the discussion as clear as possible we abstract from the effects of recombination and dissolution of ﬁrms. 9. The proof of these relations can be found in Metcalfe (2004b). 10. In fact, there is a deeper interpretation of the selection effect in the Fisher/Price accounting. It is that the rate of change of the nth cumulant of the distribution of any characteristic is proportional to the magnitude of the (nþ1)th cumulant. We call this the cumulant theorem in Metcalfe (1998). 11. It is perhaps not surprising to ﬁnd that the logistic is one of the principle devices developed by population ecologists. See Lotka (1956) and Kingsland (1985) and also Geroski (2003) for further discussion.

153

Evolutionary Adaptation in the Economic Process 12. For useful references to the technology substitution literature, see Fisher and Pry (1971), Kwasnicki and Kwasnicki (1996), and Mahajan and Petersen (1985). On the use of the logistic in relation to economic development see in particular, Nelson (1968) and Pack and Nelson (1999). 13. In a biological model x(t)may represent the number of individuals said to be of the same kind. 14. The necessary generalizations follow from (1) to (3) above. 15. As explained above the evolution of this average obeys the Price equation and its particular instantiation in Fisher’s principle. 0 16. Thus Sj (1 Si ) ¼ Sj . 17. The Fisher Pry model (1971) is precisely a binary substitution model leading to the simple logistic curve. 18. See references cited in Note 12 above. 19. It is this fact, which links evolutionary explanation with some Austrian approaches to economic evolution as a discovery process, a matter we treat further below. 20. The idea that development constraints limit selection processes thus seems to miss the point, cf. Gould, 2002:1025 et seq. 21. We cannot pursue here the related philosophical issues in relation to epistemic dependence that we hold beliefs dependent not on our own understanding but through a trust in the understanding held by others. This is of course a central theme in the quite different economic writings of Adam Smith and Hayek, encompassing a remarkable divergence of approach that relies on the same principle. See Harding (1988) for elaboration of the underlying issues. 22. This theme is explored further in Metcalfe (2004a).

References Andersen, E. S. (2004). ‘Population Thinking, Price’s Equation and the Analysis of Economic Evolution’, Evolutionary and Institutional Economics Review, 1: 127–47. Arthur, W. B. (1994). ‘Inductive Reasoning and Bounded Rationality’, American Economic Review (Papers and Proceedings), 84(2): 406–11. Braun, E. and McDonald, S. (1978). Revolution in Miniature. Cambridge: Cambridge University Press. Burns, A. F. (1934). Production Trends in the United States since 1870. Boston, MA: NBER. Christensen, C. (1997). The Innovator’s Dilemma. Boston, MA: Harvard University Press. Clark, C. (1940). Conditions of Economic Progress. London: Macmillan. Cornwall, J. and Cornwall, W. (2002). ‘A Demand and Supply Analysis of Productivity Growth’, Structural Change and Economic Dynamics, 13: 203–30. Cowan, R., David, P. A., and Foray, D. (2000). ‘The Economics of Codiﬁcation and the Diffusion of Knowledge’, Industrial and Corporate Change, 9: 211–53. Depew, D. J. and Weber, B. H. (1996). Darwinism Evolving: Systems Dynamics and the Genealogy of Natural Selection. Boston, MA: MIT Press. Downie, J. (1958). The Competitive Process. London: Duckworth. Fabricant, S. (1940). The Output of Manufacturing Industries: 1899–1937. New York: NBER. Fisher, J. and Pry, R. (1971). ‘A Simple Substitution Model of Technological Change’, Technological Forecasting and Social Change, 3: 75–88.

154

Evolutionary Adaptation in the Economic Process Fisher, R. A. (1930). The Genetical Theory of Natural Selection. Oxford: Oxford University Press. Foster, J. (1993). ‘Economics and the Self-organization Approach: Alfred Marshall Revisited’, Economic Journal, 103: 975–91. —— and Metcalfe, J. S. (2001). Frontiers of Evolutionary Economic. Cheltenham, UK: Edward Elgar. Foster, R. (1986). Innovation: The Attacker’s Advantage. London: Macmillan. Frank, S. A. (1998). Foundations of Social Evolution. Princeton, NJ: Princeton University Press. Fransman, M. (1994). ‘Information, Knowledge, Vision and Theories of the Firm’, Industrial and Corporate Change, 3: 713–58. Freeman, C. and Louca, F. (2001). As Time Goes By. Oxford: Oxford University Press. Georghiou, L., Metcalfe, J. S., Evans, J., Ray, T., and Gibbons, M. (1984). Post Innovation Performance. London: Macmillan. Geroski, P. A. (2003). The Evolution of New Markets. Oxford: Oxford University Press. Gintis, H. (2002). Game Theory Evolving. Princeton, NJ: Princeton University Press. Gould, S. J. (2002). The Structure of Evolutionary Theory. Cambridge, MA: The Belknap Press. Harding, J. (1985). ‘Epistemic Dependence’, Journal of Philosophy, 82: 335–49. Horan, B. (1995). ‘The Statistical Character of Evolutionary Theory’, Philosophy of Science, 61: 76–95. Johnson, B., Lorenz, E., and Lundvall, B. A. (2002). ‘Why all this Fuss about Codiﬁed and Tacit Knowledge?’ Industrial and Corporate Change, 11: 245–62. Kauffman, S. (1993). The Origins Of Order: Self-organisation and Selection in Evolution. Oxford: Oxford University Press. Killick, T. (1995). The Flexible Economy. London: Routledge. Kingsland, S. (1985). Modelling Nature: Episodes in the History of Population Ecology. Chicago, IL: Chicago University Press. Kuznets, S. (1954). Economic Change. London: Heinemann. Kwasnicki, W. and Kwasnicki, H. (1996). ‘Long-term Diffusion Factors of Technological Development: An Evolutionary Model and Case Study’, Technological Forecasting and Social Change, 52: 31–57. Landes, D. (1998). The Wealth and Poverty of Nations. New York: Little Brown & Co. Lotka, A. J. (1956). Elements of Mathematical Biology. New York: Dover Books. Love, A. C. (2003). ‘Evolutionary Morphology, Innovation, and the Synthesis of Evolutionary and Developmental Biology’, Biology and Philosophy, 18: 309–45. Mahajan, V. and Peterson, R. (1985). Models for Innovation Diffusion. London: Sage. Marchetti, C. and Nakicenovic, N. (1979). ‘The Dynamics of Energy Systems and the Logistic Substitution Model’, Research Report RR-79–13. Laxenburg, Austria: IIASA. Mayr, E. (1982). The Growth of Biological Thought: Diversity, Evolution and Inheritance. Cambridge, MA: The Belknap Press. Metcalfe, J. S. (1998). Evolutionary Economics and Creative Destruction. London: Routledge. —— (2004a). ‘The Entrepreneur and the Style of Modern Economics’, Journal of Evolutionary Economics, 8(2): 157–75. —— (2004b). ‘Accounting for Evolution: An Assessment of the Population Method’, Jena Papers in Evolutionary Economics, MPI, Jena. Mokyr, J. (1991). The Lever of Riches. Oxford: Oxford University Press.

155

Evolutionary Adaptation in the Economic Process Mokyr, J. (2002). Gifts of Athena: Historical Origins of the Knowledge Economy. Princeton, NJ: Princeton University Press. Mowery, D. C. and Rosenberg, N. (1998). Paths of Innovation: Technological Change in 20th Century America. Cambridge: Cambridge University Press. Nelson, R. R. (1968). ‘A ‘‘Diffusion’’ Model of International Productivity Differences in Manufacturing’, American Economic Review, 58: 1219–48. —— and Sampat, B. N. (2001). ‘Making Sense of Institutions as a Factor Shaping Economic Performance’, Journal of Economic Behaviour and Organization, 44: 31–54. North, D. (1990). Institutions, Institutional Change and Economic Performance. Cambridge: Cambridge University Press. Pack, H. and Nelson, R. R. (1999). ‘The Asian Miracle and Modern Growth Theory’, Economic Journal, 109: 416–36. Pasinetti, L. L. (1981). Structural Change and Economic Growth. Cambridge: Cambridge University Press. Polanyi, M. (1966). The Tacit Dimension. London: Routledge & Kegan Paul. Potts, J. and Dopfer, K. (2004). ‘Evolutionary Foundations of Economics (A Cyborg Manifesto)’, in J. Foster and S. Metcalfe (eds.) Organisations, Technology and Complex Adaptive Systems. Cheltenham, UK: Edward Elgar. Price, G. R. (1970). ‘Selection and Covariance’, Nature, 227: 520–1. Richardson, G. B. (1975). ‘Adam Smith on Competition and Increasing Returns’, in A. S. Skinner and T. Wilson (eds.) Essays on Adam Smith. Oxford: Oxford University Press. Richardson, R. C. (2001). ‘Complexity, Self-Organization and Selection’, Biology and Philosophy, 16: 655–83. Robinson, J. V. (1974). ‘History Versus Equilibrium’, Thames Papers in Political Economy. London: Thames Polytechnic. Schumpeter, J. A. ([1911] 1934). The Theory of Economic Development. Oxford: Oxford University Press. —— (1928). ‘The Instability of Capitalism’, Economic Journal, 28: 361–86. —— (1939). Business Cycles (2 vols). New York: McGraw-Hill. —— (1944). Capitalism, Socialism and Democracy. London: George Allen & Unwin. Shackle, G. L. S. (1958). Decision, Order and Time in Human Affairs. Cambridge: Cambridge University Press. —— (1965). A Scheme of Economic Theory. Cambridge: Cambridge University Press. Utterback, J. (1994). Mastering the Dynamics of Innovation. Boston, MA: Harvard Business School Press. Walker, W. (2000). ‘Entrapment in Large Technical Systems: Institutional Commitment and Power Relations’, Research Policy, 29: 833–46. Walsh, D. M. (2003). ‘Fit and Diversity: Explaining Adaptive Evolution’, Philosophy of Science, 70: 280–301. White, L., Jr. (1962). Medieval Technology and Social Change. Oxford: Oxford University Press. Witt, U. (2000). ‘Evolutionary Economics: An Interpretative Survey’, in K. Dopfer (ed.) Evolutionary Economics: Program and Scope. Boston, MA: Kluwer. —— (2002). ‘How Evolutionary Is Schumpeter’s Theory of Economic Development’, Industry and Innovation, 9: 7–22. Young, A. (1928). ‘Increasing Returns and Economic Progress’, Economic Journal, 28: 527–42.

156

7 Changing Boundaries of Firms in the Evolution of the Computer Industry: Towards a History-Friendly Model Franco Malerba, Richard Nelson, Luigi Orsenigo, and Sidney Winter

7.1 Introduction The aim of this chapter is to address ﬂexibility and stability in the economy by focusing on the micro level, in particular at the changing boundaries of ﬁrms—in terms of vertical integration and ‘disintegration’ (specialization)— in dynamic and uncertain technological and market environments. We address the issue of the ﬁrms’ decisions to ‘make or buy’ in contexts characterized by periods of technological revolutions punctuating periods of relative technological stability and smooth technical progress. Our aim is to examine when ﬁrms show stable structures and when they show change and ﬂexibility in their organization. The analysis of vertical integration and specialization constitutes a key issue in economics ever since Adam Smith’s statement that the division of labour is limited by the extent of the market (Smith 1776). Since then, various theories in microeconomics and industrial organization have been suggested to explain the conditions at which a ﬁrm would decide to resort to hierarchical coordination or to markets for the supply of components, ranging from issues related to bilateral monopoly, to transaction costs and information asymmetries, to uncertainties of supply, and so on. In this chapter, we propose an analysis of vertical integration and specialization centred on the coupled dynamics of competences, market structure, and the co-evolution of the upstream and downstream industries. The chapter refers to the case of a speciﬁc industry (computers) and proposes the building blocks of a model in the ‘history-friendly’ style (Malerba et al. 1999). We describe why certain computer producers (e.g. IBM) were vertically

157

Evolution of the Computer Industry integrated into semiconductors and software early on in the evolution of the industry and why later on they disintegrated, becoming specialized system producers acquiring part or all of their needs for semiconductors and software from the market, i.e. from specialized semiconductor and software producers. The main argument proposed in the chapter, however, concerns the role of coevolution in the upstream and downstream industries in explaining the changing boundaries of ﬁrms. Focusing our attention on the case of computers and semiconductors, we brieﬂy recount the appreciative story of these processes during the long-term evolution of the two industries, as it emerges from the historical and economic literature. On these bases, we present the sketch of a formal model, which should in principle be able to reproduce the history, to serve as a tool for checking the consistency of the verbal argument and as a basis for further theoretical analysis, using counterfactuals and ‘thought experiments’. Here, we present only the essential logic of the model and some preliminary simulations aiming at verifying the transparency of the formal model itself. The model clearly shares the distinctive characteristics of the evolutionary approach. Agents are characterized by ‘bounded rationality’, i.e. they do not completely understand the causal structure of the environment in which they are set. Moreover, they are unable to elaborate exceedingly complex expectations about the future. Rather, ﬁrms’ actions are assumed to be driven by routines and rules that introduce inertia in their behaviour. Agents, however, can learn and are able to improve their performance along some relevant dimensions, in particular technology.

7.2

The conceptual background

The currently leading theories of vertical integration and specialization are mainly based on some version of the transaction costs approach. They focus their attention on the market failures that may emerge in the exchange of goods and services under particular conditions and propose the view that hierarchical coordination is to be considered as a substitute for market transactions in those cases. As it is well known, transaction costs are likely to be present in cases of asset speciﬁcity, asymmetric information, and unclear deﬁnition of property rights. Under these conditions, limitations on effective writing and/or enforcement of contracts leave room for opportunistic behaviour, often leading to suboptimal solutions. The analysis of these issues may be framed in different theoretical settings: for instance, either bounded rationality (as in Williamson 1975) or full rationality (as in contract theory) may be assumed. The language of strategic interactions is increasingly used in this context. In all cases, though, the transaction costs approach has a distinct static ﬂavour and—quite

158

Evolution of the Computer Industry obviously—considers transactions as the main unit and exchange as the primary object of analysis. Technologies, the properties of the goods and the characteristics of the agents are taken as given and the processes of vertical integration/specialization are commonly examined as a choice at a given moment of time. Choices about integration—and the relative efﬁciency of markets versus hierarchies—are viewed as determined by a calculation that weighs incentives advantages of markets against the governance advantages of hierarchical organization. In the tradition of industrial economics, one ﬁnds an older and quite different perspective. Here, the characteristics of production and dynamic processes are central to the analysis. The Smith–Babbage–Stigler explanations focus on dynamic process of specialization, fuelled by static and dynamic economies of scale (Smith 1776; Stigler 1951). Thus, the growth of the market is identiﬁed as a major reason for vertical disintegration. With the growth of the upstream and downstream industries, division of labour may take place because static and dynamic (the Babbage effect) economies of scale induce efﬁcient specialization (Arora and Bokhari 2000). In this chapter we build on this second perspective. Yet, we do not emphasize economies of scale (static and dynamic) as such. Rather, we focus on the evolution of ﬁrms’ competences as a main inﬂuence on vertical integration or specialization. Thus our perspective is intrinsically dynamic. More speciﬁcally, our approach stresses the coupled dynamics of the upstream and downstream industries, in terms of the characteristics of the relevant technologies, ﬁrms’ competences, and market structures. Thus we claim that processes of vertical integration and specialization can be fruitfully analysed as the result of a coevolutionary process (Nelson 1994). In the following discussion, we refer to an upstream component industry (that produces semiconductors or software) and to a downstream system industry that produces a ﬁnal product (computers), whose characteristics and quality are deﬁned jointly by those components and other constituent parts. Our argument, however, is more general and may refer in principle to a variety of products, in which components and systems are present. A central theme in our analysis is that processes of vertical integration and specialization are likely to be deeply affected by differences in ﬁrms’ capabilities and by the nature of the knowledge base. Following Nelson and Winter (1982), we stress that ﬁrms are to be conceptualized as repositories and integrators of fragments of tacit and speciﬁc productive knowledge, which is partly embedded in routines. Through learning, ﬁrms accumulate over time capabilities in speciﬁc technological, productive, and market domains (Teece and Pisano 1994). Such competences take time to be developed and are then typically sticky and local (Winter 1987). Thus, they adapt only slowly to changes in technology and demand (Christensen and Rosenbloom 1997). The growth and dynamics of competences—as well as the processes of

159

Evolution of the Computer Industry coordination and integration of such capabilities—in each one of the two industries considered here inﬂuences the evolution of the other sector and shapes the dynamics of vertical integration and specialization (Langlois 1992; Langlois and Robertson 1995, Jacobides and Winter 2005). Thus, in our model the boundaries of ﬁrms are inﬂuenced jointly by the evolution of capabilities, complementarities, and market structure in the two sectors. Moreover, these factors should not be examined in isolation, because they interact and change at different paces over the evolution of industries. Hence, the forces determining vertical integration and specialization may change over time. The time dimension is particularly important. Competence-based accounts of the boundaries of ﬁrms take a long run, historical perspective. Competences take time to be developed and the previous history of the processes of construction of those capabilities often is important in determining what ﬁrms can and cannot do (better or satisfactorily enough). For example, if a ﬁrm decides to discontinue the development and production of certain components, it might ﬁnd it difﬁcult to resume such activities later on and in any case time and efforts are required. Thus, decisions are neither entirely ﬂexible nor symmetric as time goes by.

7.3

A brief discussion of the semiconductor and computer industries

During the evolution of the computer industry, the pattern of vertical integration, disintegration, and specialization of computer producers has been changing over time. In particular, in various periods of the history of the industry, computer producers have been integrated into semiconductors (and into software as well). In the following pages we will sketch a highly stylized history of the extent of vertical integration of computer producers into semiconductors. At the very beginning of the industry, most computer producers were not integrated. As Kricks (1995) discusses at length, the ﬁrst computer producers mainly purchased the very early receiving tubes components (1950–8) from the open market. However, with the introduction of transistors and the related beginning of the semiconductor industry in the second half of the 1950s and the early 1960s, there emerged two main behaviour patterns. The largest ﬁrms, such as IBM, RCA, and GE, were totally or at least partially vertically integrated. In particular, IBM was partially integrated and had fully automated manufacturing plants of its own; it also maintained close relationships with one of the leading merchant semiconductor suppliers—Texas Instruments. The smaller ﬁrms purchased components on the market. With the introduction of integrated circuits (1964–70) IBM became fully vertically integrated into semiconductors, ﬁrst with a hybrid integrated circuit

160

Evolution of the Computer Industry technology, solid logic technology (SLT), and then with monolithic ones. Other producers also partially integrated into integrated circuits. There are three basic reasons for vertical integration into integrated circuits. First, integrated circuits embedded system elements and thus required close coordination between the system and the component producer in the design and development of both components and systems. Second, semiconductor designs became more and more ‘strategic’ and key for system development, and therefore their design, development, and production was kept in-house for fears of leakage of strategic information. Third, the rapid growth of the mainframe market and later on of the minicomputer market (1960s and 1970s) generated fears of shortages of various key semiconductor components among some of the largest computer producers. With the full development of the semiconductor industry (1970s, 1980s, and 1990s) and the introduction of microprocessors, very large-scale integrated circuits, and RAM and ROM memory devices, those computer producers that were vertically integrated—including IBM—exited more or less completely from large-scale production of semiconductor components. Disintegration took place because the new demand for semiconductors coming from personal computer (PC) producers had grown greatly; in response, a variety of highly advanced components were introduced by several merchant microelectronics ﬁrms. A key ﬁrm—Intel—emerged as the industry leader for microprocessors, thus determining a de facto standard in the semiconductor industry to which computer producers, out of necessity, complied. A different story can be told for vertical integration by computer producers into software. At the beginning of the industry (1950s) computer hardware and software were highly integrated in mainframe systems for technological and strategic reasons. Few mainframe machines were produced at the beginning, and hardware was usually delivered with tailored software. This situation lasted until December 1969. At that time IBM decided to ‘unbundle’—to price software separately from hardware. The motives were several. American antitrust authorities were closely monitoring IBM for possible anticompetitive behaviour. Also, it was becoming more and more expensive for a large mainframe company to develop tailored software programs for speciﬁc small customers buying similar hardware machines. IBM’s decision opened the way to the independent production of software by software specialists. This production, however, remained small during the 1970s, even with the advent of minicomputers (which required custom software for speciﬁc uses). During the 1980s and 1990s, with the enormous growth of the PC industry, a very large demand for system and application software emerged from a wide variety of computer producers (Mowery 1996). This demand was met by a growing population of small software specialists that had previously done custom software development or that were new entrants. The rapid growth

161

Evolution of the Computer Industry of the software industry, and IBM’s decision to outsource the operating system for its PCs, led to the emergence of Microsoft as the industry leader and of Microsoft DOS as a key industry standard. In addition, the large and heterogeneous demand for software applications based on the Microsoft standards opened the way for specialized development of high ﬁxed-cost software programs. With this rapid expansion of potential applications, the entry and growth of software specialists and the variety of initiatives and products was very high. This led several vertically integrated mainframe manufacturers to disintegrate and specialize in hardware computer production. A similar story holds—with different speed and intensity—for operating system software, package application software, and custom software. While at the beginning of the computer industry, disintegration took place mainly for custom software to small users, later on with the development of a mass market for PCs and the diffusion of standards and open system architectures, the exit of system producers from software development occurred also for package software and operating system software. Application software became a segmented market, with limited economies of scope among different applications, each requiring a different type of knowledge regarding the speciﬁc uses governing development targets. As mentioned earlier, and similarly to semiconductors, a technological and market leader—Microsoft—emerged in the software industry. This led to a de facto standard in operating systems that further encouraged the exit of computer producers from in-house development of system software. In sum, in computers and semiconductors the evolution of the upstream industry is likely to be profoundly inﬂuenced by the evolution of the downstream sector—and vice versa (e.g. Bresnahan and Malerba 1999; Langlois and Robertson 1995). For example, had a large dominant ﬁrm in mainframes (IBM) not been present in the 1960s when integrated circuits were introduced, vertical integration would have been probably less relevant. On the contrary, had a large dominant ﬁrm (IBM) entered quite soon the PC market, it might have vertically integrated into microprocessors, thus reducing the possibility of rapid growth of a capable component producer like Intel. Conversely, one might speculate that in the absence of Intel, IBM might not have successfully entered the PC market. In all these cases, not only the evolution of the computer industry would have been rather different, but also the evolution of the semiconductor sector. The model that follows tries to capture in a compact and simple way the main mechanisms that might have led to waves of vertical integration and specialization in the computer industry. Although the formal model aims at reproducing the stylized history of the computer and of semiconductors on the basis of the underlying appreciative model, here we simply present the basic logical structure of the formal model and some preliminary exercises meant to explore how the model works.

162

Evolution of the Computer Industry

7.4 Some theoretical statements on the changing vertical boundaries of ﬁrms A full understanding of the process of vertical integration and disintegration has to take into account the presence and the differences in the level of ﬁrms’ capabilities in an industry and the nature of the relevant knowledge base. Following Nelson and Winter (1982), we stress that ﬁrms are best interpreted as repositories of tacit and speciﬁc knowledge, which is often embedded in routines. Through learning, ﬁrms accumulate over time capabilities in speciﬁc technological, productive, and demand domains (Teece and Pisano 1994). These capabilities take time to develop, and are then sticky and adapt slowly to changes in technology and demand (Christensen and Rosenbloom 1997). The coordination, integration, growth, and dynamics of capabilities in related industries create processes of vertical integration and specialization (Langlois 1992; Langlois and Robertson 1995; Jacobides and Winter 2005). Here are some theoretical predictions on vertical integration and specialization based on the interplay between capabilities, ﬁrm size, and market dimensions.

7.4.1

First prediction—vertical integration because of lack of external capabilities in components

The lack of already developed external capabilities or a clear superiority of internal competences with respect to external ones is a ﬁrst reason for vertical integration (Langlois and Robertson 1995). We have therefore a ﬁrst prediction. Firms will vertically integrate in the development and production of upstream components if capabilities for the development and production of components are not available on the market. Similarly, even in the presence of external capabilities, ﬁrms will remain vertically integrated if internal capabilities are superior to external ones.

In the ﬁrst case, in a highly uncertain environment the lack of available external competences in component development production pushes ﬁrms to develop internal capabilities, because it would take too much time, cost, and effort to persuade external suppliers or distributors with no speciﬁc competences in an activity to learn and build speciﬁc competence in that activity. In the second case, superior internal capabilities with respect to external ones would push ﬁrms to remain vertically integrated into components. Note that in either case our proposition is not a mere tautology: downstream ﬁrms could alternatively respond by trying to stimulate a market response, offering higher prices and advertising their needs more widely.

163

Evolution of the Computer Industry

7.4.2

Second prediction—vertical integration because of the advantages of coordination and integration capabilities

If external capabilities and internal capabilities are at a similar level, a second reason for vertical integration relates the dynamic governance costs (Langlois and Roberston (1995)). Here, building again on Nelson and Winter (1982), we stress that a fundamental aspect of ﬁrms’ activities consists in the integration of different pieces of knowledge. Particularly in the case of systemic innovations, coordination and integration capabilities become important in responding effectively to complementarities. If new components and subsystems have to be inserted into new systems, dynamic interdependencies and feedbacks may take place between components and systems, requiring continuous changes and adjustments in both (Langlois and Robertson 1995). Thus, vertical integration in the development and production of components and systems may be necessary. This is particularly likely to be the case whenever the knowledge underpinning ﬁrms’ activities has a strong tacit component. This is a second reason for vertical integration related to capabilities and dynamic governance costs. In the case of systemic innovations (Teece 1986), ﬁrms may decide to vertically integrate because ‘the cost of persuading, negotiating with, coordinating among, and teaching outside suppliers in the face of economic change or innovation’ may be too high (Langlois and Robertson 1995).

7.4.3

Third prediction—vertical integration because of a large one-ﬁrm demand for components directed to many Small component ﬁrms with similar capabilities in a relatively stable environment

Moderate rates of technological change or competence-enhancing technological change in the upstream activity are likely to induce vertical integration in case a system ﬁrm is very large and suppliers are too small and cannot grow quickly enough. In this case, a large downstream ﬁrm is likely to have the resources to (at least) keep the pace of technical change, while suppliers cannot progress fast enough or produce at sufﬁciently high scale. If in the upstream industry the overall technological change is not extremely rapid, innovation is incremental along certain trajectories and several small component ﬁrms with similar capabilities are present, a large downstream ﬁrm may decide to vertically integrate into components because of the need to have a large secure supply of components.

7.4.4

Fourth Prediction—Disintegration Because of the Wide Variety of Component Firms with Different Capabilities in a Highly Turbulent Environment

In case of a fast rate of technological change, high uncertainty, and a wide distribution of capabilities in the upstream industry, relevant innovations may

164

Evolution of the Computer Industry come from any quarter. Therefore, a downstream ﬁrm may decide not to vertically integrate because it does not want to be locked unto an inferior technology. This is particularly true in the ﬁrst stages of an industry life cycle (often following major technological discontinuities) and/or in cases where in the upstream industry barriers to entry are low, e.g. in ‘entrepreneurial regimes’ (Winter 1984). In addition, as we will see later, to the extent that economies of specialization exist (because of large markets, cumulative technical change, learning-by-doing, economies of scale, and so on), the upstream producers will quickly accumulate increasingly superior technological capabilities that will be more efﬁcient. This will further strengthen the tendency towards vertical disintegration. This reason for vertical disintegration derives from what we term the ‘variety effect’ in capabilities, visions, and strategies (Cohen and Malerba 2001). If in the upstream industry the overall rate of technological change is extremely rapid, innovation is highly uncertain and several capable ﬁrms are active at the technological frontier, a downstream ﬁrm may decide not to vertically integrate into upstream development and production (variety effect).

7.4.5

Fifth prediction—disintegration because of the emergence and establishment of an upstream highly capable monopolist

Finally, there is the case of the emergence of a capable component monopolist. Over time during the evolution of an industry a technological and market leader may emerge as a result of competition and selection in the upstream industry, with the possibility of the emergence also of a dominant design (Klepper 1996). This technological and market leader becomes extremely advanced in terms of competences and research and production facilities is the standard setter in the industry. Therefore, a downstream ﬁrm may decide not to vertically integrate (and to purchase instead components from the market), because it will never be able to match the R&D efforts and the pace of technological advance of the supplier. The emergence of a capable monopolist is likely to be linked to the existence of various forms of static and above all dynamic increasing returns: typically, cumulative R&D and innovation and large marketing expenditures generating brand loyalty, but also more conventional variables like large markets and static economies of scale. These factors—as noticed earlier—constitute a powerful engine of specialization. Thus, these variables induce the development of a strong upstream industry (and hence vertical disintegration) and the emergence of high concentration among the specialized component producers. If in the upstream industry a very advanced technological and production leader has emerged, a downstream ﬁrm may decide not to vertically integrate.

165

Evolution of the Computer Industry

7.4.6

Conclusions from the theoretical statements

These determinants should not be examined in isolation. Rather, more than one may be present in a speciﬁc situation, thus creating trade-offs between specialization and integration. Moreover, the determinants of vertical integration and specialization may change during the evolution of the upstream and downstream industries. Thus in different moments of the evolution of any industry we may observe different reasons for vertical integration and disintegration. The time dimension is particularly important. In fact while transaction costs–based explanations have a short-term dimension, capability-based explanations have a long-term view. In particular, it takes time to develop competences. Therefore, if a ﬁrm decides to exit from the development and production of components, and later on decides to re-enter through internal development and production (and not through the acquisition of an already existing external supplier), it takes time and a lot of effort to do that. More important, given the relevance of dynamic increasing returns, the evolution of the patterns of vertical integration/disintegration is likely to be characterized by strong elements of irreversibility and path dependency. A fortiori, this is likely to be the case when one considers that the patterns of vertical integration or specialization are determined by the co-evolution of two industries. The evolution of the upstream industry is likely to be profoundly inﬂuenced by the evolution of the downstream sector. In the example discussed earlier, had a large dominant ﬁrm in mainframes (IBM) not been present in the 1960s when integrated circuits were introduced, vertical integration would have been less relevant in the industry. On the contrary, had a large dominant ﬁrm (IBM) entered quite soon into PCs, it probably would have vertically integrated into microprocessors, thus reducing the possibility of a rapid growth of a capable monopolist such as Intel. In all these cases, not only the evolution of the computer industry but also the evolution of the semiconductor industry would probably have been rather different. Finally, it is worth considering that in this approach, the decision to vertically integrate cannot be treated symmetrically to the decision to vertically disintegrate. If anything, the decision to disintegrate presupposes the existence of suppliers, whose performance can be observed and compared to that of the downstream producer. The opposite does not apply, though: a ﬁrm contemplating vertical integration has no available comparison before starting production. In sum, vertical integration and disintegration are characterized by path dependency, irreversibilities, and asymmetries and result from the co-evolution of the upstream and downstream industries. The model that follows tries to capture in a compact and simple way the main mechanisms that might have led to waves of vertical integration and specialization in the computer industry. Although the formal model aims at

166

Evolution of the Computer Industry reproducing the stylized history of the computer and of semiconductors on the basis of the underlying appreciative model, here we simply present the basic logical structure of the formal model and some preliminary exercises meant to explore how the model works.

7.5 The model 7.5.1

Computers

The essence of the story that the model tries to reproduce can be summarized as follows. At the beginning of the simulation, a number of ﬁrms start producing and selling computers. Computers can be thought of as a point in the space of characteristics. Speciﬁcally, a computer is characterized by two attributes, cheapness and performance. As a consequence of ﬁrms’ R&D investment, the characteristics of computers of a given type improve over time. The position of a particular computer design at any time deﬁnes its ‘merit of design’ (Mod) or quality. In turn, computers result from the combination of two main inputs, systems and components (see Box 7.1).

7.5.2

The market for components

Systems are designed in-house by system producers, while components may be also produced in-house or bought by specialized producers of semiconductors.

Box 7.1 COMPUTERS The level of the merit of design (Mod) is given by a constant elasticity of substitution (CES) function:

ModiCOMPUTER ,t

æ æ 1æ COMP SYS ¼ A Modi ,t þð1 Þ Modi ,t

(1)

with A > 1, 0 < < 1 and æ > 1. The elasticity of substitution is ¼ 1=( 1 þ æ). In the CES functions the weight attributed to the Mod of components () is always higher than the weight on the Mod of systems. In the model there are two broadly different types of computers, mainframes and personal computers (PCs), with mainframes having a high ratio of performance to cheapness and PCs having a high ratio of cheapness to performance. Similarly, the elasticity of substitution, , in PCs is higher than in mainframes. Different computers are produced by different companies. In other words, a computer company produces only either mainframe or PCs: diversiﬁcation is not contemplated in this model.

167

Evolution of the Computer Industry Suppliers are chosen by computer ﬁrms on the basis of a ranking of the Mod of the components produced by each supplier. Given uncertainty and imperfect information, this choice is stochastic. Semiconductor suppliers sell their product to an external market also, e.g. the military or other industries. At the beginning, computer producers rely on semiconductor ﬁrms for their supply of components. The latter are designed on the basis of the available technology, e.g. transistors (see Box 7.2).

7.5.3

Firms’ behaviour and technical progress

At the beginning of our story, ﬁrms start with a given Mod and they start to sell, make proﬁts, and invest in R&D spending. The price of a computer is Box 7.2 THE MARKET FOR COMPONENTS The demand for components, faced by component-specialized ﬁrms, comes from two sources: 1. demand for components from users different from computer ﬁrms (e.g. consumer electronics, the military, the automobile industry). This demand is not modelled explicitly and the size of the external market is exogenous. External demand plays a critical role in the model, since it allows component producers to survive and grow in the early stages of development of a new technology and to improve the quality of their components. 2. demand for components from computer ﬁrms that have decided to outsource component production (specialized ﬁrms). When a computer ﬁrm decides to outsource components production it starts to scan the market for potential suppliers. A specialized computer producer will ‘sign’ a contract with a component producer, who is selected by using a probability function that considers the technical quality of the components: the higher the quality of the component, the higher the probability of signing a contract with a computer producer. Formally: Æ LCOMP ¼ ModiCOMP i ,t ,t (2) LCOMP i ,t Pr COMP ¼ P COMP i ,t Li , t where L,i,tCOMP is the propensity of component producer i to be selected and Pri,t is the probability of a supplier to be selected. A component ﬁrm that signs up a contract sells a number of components, which is related to the proportion to which components and systems combine in order to build a computer. After signing the contract the computer ﬁrm is tied to the component supplier for a certain number of periods. When this period expires, a new supplier might be selected, using the same procedure, if the ﬁrm still decides to buy component on the open market.

168

Evolution of the Computer Industry given by adding a mark-up on unit costs, which in turn are derived by the Mod of the computer: thus, unit production costs decline over time as a result of ﬁrms’ innovative activities and consequent improvements of the Mod.1 The price of components charged by component suppliers is determined symmetrically by adding a ﬁxed mark-up to unit production costs. Technical progress in components and systems—and hence in computers—is the result of the R&D activities of ﬁrms, both as computer and semiconductor producers are concerned. R&D expenditures are calculated following a simple rule of thumb, i.e. a certain fraction of proﬁts is invested in R&D in each period. By investing more in R&D, ﬁrms buy themselves higher probabilities to increase their Mod: technological progress is a stochastic process (see Box 7.3). Box 7.3 TECHNOLOGICAL PROGRESS In this model technical progress is modelled using the ‘double draw scheme’ used in Nelson and Winter (1982) for both systems and components. There are two draw schemes, similar in term of mechanisms but different in terms of mean and variance for systems and components. Integrated ﬁrms have two progress functions: one for components and one for systems. Specialized computer ﬁrms and component ﬁrms have only one progress function: for systems and for components, respectively. According to this scheme, in each period ﬁrms draw the value of their merit of design (Mod) from a normal distribution. The number of draws that any one ﬁrm can take is set proportional to its R&D spending; the parameter of proportionality is called drawcost. The algorithm that gives the number of draws is the following: NumberOfDrawsi ,t ¼

RdExpenditurei ,t drawCost

(3)

In each period, the values of the Mod obtained through the ﬁrms’ draws are compared with the current Mod, and the higher among these values is kept. Thus, more draws increase the likelihood to get a higher Mod for both systems and components. The extent to which technical progress is possible for each ﬁrm, given their R&D investment, depends in turn on the level of public knowledge (PK) (e.g. published academic research, technical information available in specialized journals) and the value of the Mod achieved by the ﬁrm in the previous period: in other words, technological change is only partly cumulative at the ﬁrm level and opportunities of innovating are ﬁrm-speciﬁc.* The mean of the normal distribution from which the values of the Mod of system or component are taken is a linear combination of the Mod at time t1 of ﬁrm i and of the level of PK at time t. i ,t ¼ h Modi ,t1 þ (1 h) PKtK

(4)

Integrated producers enjoy some coordination advantages as compared to specialized producers, because they can produce components tailored to their system. As a consequence, the productivity of their R&D efforts on components is enhanced. This effect can

169

Evolution of the Computer Industry Box 7.3 (Continued ) be simply expressed as if their component R&D expenditures were augmented by a certain factor. * Public knowledge (PK) is speciﬁc to each basic component technology (i.e. transistors, integrated circuits, and microprocessors) and it grows exogenously over time. When a new technology is introduced, its corresponding level of public knowledge is lower than that reached by current technology, but then it grows faster and at a certain time it overtakes the PK of the older technology. The rate of growth of PK starts to slow down as time goes by. An integrated computer ﬁrm decides to adopt the new technology when the mean of its own distribution becomes inferior to the PK of the new technology.

7.5.4

Demand for computers

Customers of computers are characterized by their preferences about the two attributes that deﬁne a computer design—performance and cheapness. There are two customer groups: (a) one consisting of ‘big ﬁrms’ who are especially interested in performance, and care less about cheapness; and (b) the other of ‘small users’ who are especially concerned about cheapness, and who value performance less than do big ﬁrms. These differences in preferences show up in terms of how performance and cheapness ‘trade off’ in terms of customer evaluation of merit. Each customer group consists of a large number of heterogeneous subgroups. Within a particular subgroup, customers buy computers valuing its ‘merit’ compared with other products. In addition, however, markets are characterized by frictions of various sorts, including imperfect information and sheer inertia in consumers’ behaviour, brand-loyalty (or lock-in) effects as well as sensitiveness to ﬁrms’ marketing policies. These factors are captured in a compact form by the share of computer brands in that overall submarkets at time t1: the larger the share of the market that a product already holds, the greater the likelihood that a customer will consider that product. Moreover, there is a stochastic element in consumers’ choices between different computers (see Box 7.4). Box 7.4 DEMAND FOR COMPUTERS The ‘value’ that customers attribute to any speciﬁc computer design, Mi,t, of a computer designed by company i at time t is deﬁned as: ð1a1Þ Mi ,t ¼ Alfa Cheapnessia1 ,t Performancei ,t

170

(5)

Evolution of the Computer Industry The probability calculated for ﬁrm i is applied to each submarket, i.e. every submarket m selects the computer ﬁrm i with a probability Pri,t. Formally, the propensity Li of computer i to be sold to a group of customers at time t is deﬁned as: Li ,t ¼ MiÆ,t 1 þ Si ,t1

(6)

with S as the market share of ﬁrm i The probability Pri,t of the computer i to be sold to a group of customers at time t is given by: Li , t Pr i ,t ¼ P Li ,t

(7)

i

From this ﬁrms’ computer i—if selected—the submarket buys a number of computers equal to Mi,t.

7.5.5

Vertical integration and specialization

As a consequence of the cumulative and stochastic nature of technological change on the one hand and of the lock-in effects in demand, both computer and semiconductor ﬁrms undergo a process of selection and the industry becomes increasingly concentrated. Both computer ﬁrms and component suppliers exit the market when their market share falls under a certain minimum threshold. Semiconductor producers also exit if they do not sell anything to computer ﬁrms for a number of consecutive periods. Computer producers may decide to vertically integrate into semiconductors, if they think that they can design and produce components that are comparable in quality to those offered by specialist suppliers. In turn, this is more likely to be the case if computer producers are larger enough as compared to extant suppliers, so that they can fund a much larger ﬂow of R&D expenditures. Moreover, the decision to vertically integrate depends probabilistically on the age of the component technology. In the early stages of development of the new technology, when specialized semiconductor producers are likely to control the new technical developments, technical change is fast and comes from every quarter, given the risk of getting stuck in an inferior trajectory a computer producer is not likely to vertically integrate. Rather, computer ﬁrms would wait and see how the new technology develops. Instead, if the technology for designing and producing components is settled along relatively well-deﬁned and established trajectories, the probability that new superior generations of components may be frequently invented by component suppliers is lower. Also,

171

Evolution of the Computer Industry fears of supply shortages may induce vertical integration. Again, this is likely to be the case if semiconductor ﬁrms are small because the external market for transistor is not too big and/or no dominant ﬁrm has emerged (see Box 7.5). Box 7.5 VERTICAL INTEGRATION AND SPECIALIZATION The probability of integration for each computer ﬁrm may be deﬁned, in a compact way, as follows. Let

b1

b2 Qi ,t AgeOfTechK ;1 Zi ,t ¼ min COMP w biggestQt

(8)

where AgeOfTechK (K ¼TR,IC,MP) ¼ t (starting time of technology K) and w is a parameter. Q refers to ﬁrm size. Then Pr ob(Integrate)i ,t ¼

B Zi ,t 1 þ Zi ,t

(9)

If the probability of integration is bigger than a number drawn from a uniform distribution (0–1), integration occurs. The probability of specialization for each ﬁrm is a probabilistic function of the proportionate difference between the merit of design (Mod) of the components produced internally and the Mod of the best component available on the market. The probability of specialization for each ﬁrm is deﬁned as follows. Xi ,t ¼ max

max ModCOMP ModCOMP t i ,t ModCOMP i ,t

! ,0

(10)

where max ModCOMP is the higher component Mod available on the market. Then: Pr ob(Specialize)i ,t ¼

A Xi ,t 1 þ Xi ,t

(11)

As before, if Prob(Specialize) is bigger than a number randomly drawn by a uniform distribution (0–1), specialization will occur. A specialized computer ﬁrm may also decide to change its supplier, if a better computer producer has emerged in the market. The procedure for changing supplier follows the same rule for the specialization process. That is to say, every n periods after the last decision to specialize or the last change of supplier, a specialized ﬁrm checks if a better supplier than the current one exists. If this is the case, a new supplier is chosen using the rating mechanism described in the discussion of the demand module.

172

Evolution of the Computer Industry

7.5.6

The working of the model with two technological discontinuities in components

Computer ﬁrms start developing systems with transistors. After some time, a new component technology, e.g. integrated circuits, becomes available. This technological discontinuity allows for the entry of new semiconductor ﬁrms. As they invest in R&D and the new technology improves, they will gradually become more efﬁcient than competitors belonging to the older generation, eventually displacing them. Computer ﬁrms may face, in these circumstances, pressures towards vertical disintegration. As mentioned earlier, an integrated computer ﬁrm may decide to specialize for motives that are not entirely symmetrical to the decision of integration. The existence of a large, competent supplier, who is able to design and produce high-quality components, to sustain high R&D expenditures and hence high rates of technical advance, is likely to induce specialization. In case of vertical integration, the company compares the Mod of the components produced in-house to the best components available on the marketplace. By observing the behaviour of suppliers and comparing it to its own performance, the computer producer is also able to have an idea of the degree of ferment that characterizes innovative activities in the design of components. Thus, a computer producer will decide to buy— instead of make—components as a probabilistic function of the proportionate difference between the Mod of the components produced internally and the quality of the best component available on the market. However, computer ﬁrms are able to adopt integrated circuits technology very rapidly and, unless the new specialized semiconductor producers are large enough to be able to innovate very quickly (e.g. they have a large external market), large dominant computer ﬁrms will end up producing in-house their own components. Three main factors push towards integration. First, the largest of the computer ﬁrms by now are big enough to spread their R&D costs for components across at least as large a total production as it can be achieved by any of the specialized component suppliers (again, absent a large external market that allows semiconductor ﬁrms to have very big sales) and can catch up and overtake quickly in terms of the Mod of their integrated circuits. Second, there are some signiﬁcant gains from being able to design both the system and the components, so that they complement each other. Third, in the history of the computer, one of the main reasons that led IBM to vertically integrate into the production of integrated circuits was the risk of supply shortages. Producers of integrated circuits were too small to guarantee that they would have been able to satisfy IBM demand in each period. Thus, in the model it is assumed that the production of components might be constrained by the productive capacity of semiconductor ﬁrms: speciﬁcally, it is assumed that capacity constraints manifest themselves in an erosion of the ﬁrm’s component Mod, reﬂecting, for example, a decrease in the quality of the

173

Evolution of the Computer Industry component, due for example, to delivery delays. Lower component quality feeds back on sales, proﬁts, and R&D expenditures. After capacity constraints have been experienced, ﬁrms gradually adjust their capacity drawing from their proﬁts (and therefore decreasing correspondingly their R&D). The relative strength of these conﬂicting forces determines whether integration or specialization will prevail. Note, however, that if the largest computer ﬁrm chooses vertical integration (as IBM actually did in the integrated-circuits period), the ability of the specialized integrated-circuit producers to achieve economies of large-scale production is in turn reduced. As a result of these dynamics, the mainframe computer industry is likely to become highly concentrated, as well as integrated, while no very large dominant ﬁrm emerges in the component supply business. After some time, a second discontinuity in component technology (microprocessors) allows for the entry of new generation of component ﬁrms, which again undergo a process of selection so that a new component leader emerges. The previous story applies again in this case. However, some important differences characterize this second technological transition. On the one hand, the mainframe industry has become strongly concentrated and it is dominated by a very large monopolist. On the other hand, microprocessors make it possible to design and start selling a new—previously unattainable—type of computers, i.e. PCs. PCs differ from mainframes—in the language of this model—in two important respects. The weight of components with respect to systems is much higher in determining the Mod of the computer as compared to mainframes. Moreover, PCs are much cheaper than mainframes. Thus, a whole new class of customers, who attribute much more value to cheapness than to performance, starts buying the new type of computer: the PC market opens up and grows rapidly. PC producers, however, are relatively small as compared to the microprocessor suppliers that can sell also to a large external market. Thus, quite soon large specialized microprocessor supplier begins to emerge before any of the new producers of PCs becomes very large. Moreover, by assumption, the advantages of tailoring microprocessor design to systems design are less for PCs than they had been for mainframes and lock-in effects on the demand side are less important in the case of PCs as compared to mainframes. As a result, PC computer ﬁrms remain specialists in systems. And a dominant microprocessor supply ﬁrm emerges. The rise of strong and large microprocessor ﬁrms, selling their wares on a large but not concentrated PC market, soon makes it costly and risky for mainframe producers to continue to design and produce their own components—now microprocessors. This leads to vertical disintegration in the mainframe industry also. In sum, the dynamics of vertical integration and specialization is driven in this story—and in the model—by the evolution of competences in the design

174

Evolution of the Computer Industry of components in computer and semiconductor ﬁrms, respectively; in turn, such evolution depends on the ability to invest in R&D and hence on the size of the external market for semiconductor suppliers demand and on the strength of the lock-in effects in the computer market. The dynamics of technological capabilities and demand conditions determine also the dynamics of market structures in both industries.

7.6 The simulations In order to test the hypotheses discussed earlier we have run a set of simulations. Rather than trying to replicate the stylized story discussed earlier, we explore some extreme cases, in order to explore the logic and the functioning of the model itself. Fifteen computer ﬁrms and twelve component ﬁrms start at period 0. A new component technology (integrated circuits) is introduced at time 30 and the new components with the new technology are available on the market at time 40. At time 100, another completely new component technology (microprocessors) is introduced and the new components with the new technology become available on the market at time 110. All computer ﬁrms start specialized. Firms reinvest all their proﬁts. Mark-up is equal to 10 per cent. A simulation covers 250 periods: a period could be considered a quarter. We discuss ﬁrst a benchmark case, in which vertical integration emerges quite soon and persist over time. We then discuss factors that either reinforce vertical integration or that move the system industry from integration to specialization. We then address two other issues such as the continuous changing of the boundaries of ﬁrms and the eventual disappearance of the component industry.

7.6.1

The benchmark case: vertical integration due to lock-ins in system demand and cumulativeness of technical advance

In the benchmark run we ﬁrst examine a situation in which there are lock-ins in system demand and cumulativeness of technical advance. In the model we obtained that by increasing the value of the bandwagon effect in computers and by reducing the role of public knowledge in the advancement of the Mod in computers. As a consequence, the computer industry becomes quickly concentrated and integrated and remains so despite technological discontinuities. In particular, the run shows that in the computer industry ﬁrms start specialized and due to the lock-in in demand and cumulativeness of technical progress, they undergo a process of selection. The component industry becomes also concentrated. Concentration emerges already with the initial

175

Evolution of the Computer Industry technology. The discontinuity introduced by the new component technology (integrated circuits) allows for the entry of new ﬁrms, which undergo a new selection process, and a new component leader emerges. With the second discontinuity in component technology (microprocessors) a new generation of component ﬁrms enters the market, and again competition generates the emergence of a new component leader. As Figure 7.1 shows, in this benchmark run the growth of computer ﬁrms due to the lock-in effects in demand and cumulativeness of technical change leads to high concentration in the computer market (see Herﬁndhal index) and also to an initial drive towards integration as the component technology matures and the size of computer ﬁrms increases. The ﬁrst technological discontinuity in components reduces the number of integrated ﬁrms because the uncertainty of the new component technology is very high. However, with the maturity of the second generation of semiconductor components and the increase in size of the computers leader, all the remaining computer ﬁrms (very few in numbers) become vertically integrated again. The high rate of technical progress due to its large size (and therefore high R&D expenditures) allows the integrated ﬁrm to remain vertically integrated in spite of the second discontinuity in component technology (see the ratio integrated ﬁrms/total number of ﬁrms in Figure 7.1). In sum, in the benchmark case vertical integration is driven by demand conditions (lock-ins) and by cumulativeness of advance in systems technologies: these factors create a large and technologically progressive system leader that remains integrated.

7.6.1.1

CASE 1: FROM INTEGRATION TO SPECIALIZATION IN COMPUTERS: THE ROLE OF EXTERNAL DEMAND FOR COMPONENTS AND LACK OF IN-HOUSE COMPETENCES BY SYSTEM FIRMS IN COMPONENT TECHNOLOGIES

The choice of vertical integration by system ﬁrms changes drastically if component producers can grow in size and competences. In the model this could be the result of two factors. First, the presence of a large external market (in addition to the demand coming from computer producers) allows for the emergence of large and capable component producers. Think for example to the case of mainframes for the military market for American integrated circuits producers or to the PC market for semiconductor producers. In addition, this pattern is reinforced if computer producers do no have advanced capabilities to replicate in-house the components produced by their suppliers. In the run, this is done by introducing a very large external market and a low share of the Mod of the last component purchased by a system producer that a newly integrated ﬁrm is able to obtain.

176

Evolution of the Computer Industry Number of firms 25 System F 20

Component F

15 10 5 0 1

22 43 64 85 106 127 148 169 190 211 232

(a) Herfindahl index 1.0 0,9 0,8 0,7 0,6 0,5 0,4 0,3 0,2 0,1 0 1

16 31 46 61 76 91 106 121 136 151 166 181 196 211 226 241

(b) Number of integrated firms 5.0 4.5 4.0 3.5 3.0 2.5 2.0 1.5 1.0 0.5 0 1

16 31 46 61 76 91 106 121 136 151 166 181 196 211 226 241

(c) Figure 7.1 (a–f) Benchmark run

177

Evolution of the Computer Industry Ratio: number of integrated firms/total number of firms 1.0 0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0 1

16 31 46 61 76 91 106 121 136 151 166 181 196 211 226 241

(d) Share of specialized leader and of integrated leader 1.0 0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0

Specialized Integrated

1

18

35

52

69

86 103 120 137 154 171 188 205 222 239

(e) Total share of specialized firms and of integrated firms 1.0 0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0

Specialized Integrated

1

21 41 61 81 101 121 141 161 181 201 221 241

(f) Figure 7.1 (Continued)

178

Evolution of the Computer Industry This situation is very ‘Smithian’: the size of the external component market affects the specialization of computer producers. As Figure 7.2 shows, computer producers remain specialized and the computer industry remains with low concentration for the whole simulation (with just a small tendency towards integration in the last periods, when microprocessor technology becomes very old). The reason is that the existence of strong and competent suppliers reduces the differences among computer ﬁrms who have access to the best component technology, so that the whole computer industry remains competitive. 7.6.1.2

CASE 2: FACTORS AFFECTING INTEGRATION: LACK OF UNCERTAINTY IN DEMAND FOR COMPUTERS

Integration may be driven by a very high growth of computer producers due to the lack of uncertainty in the demand for computers: the better the computers offered on the market, the more computers are sold by a ﬁrm. In the model, this is obtained by making the demand for computers deterministic. As Figure 7.3 shows, no uncertainty in demand for computers allows some computer ﬁrms to grow and the computer market to become concentrated. The large size of system producers drives the integration. 7.6.1.3

CASE 3: FACTORS AFFECTING INTEGRATION: HIGH IN-HOUSE CAPABILITIES OF DEVELOPING ADVANCED COMPONENTS AND HIGH CUMULATIVENESS OF TECHNICAL ADVANCE

Another factor affecting integration could be related to capabilities and cumulativeness, even without the presence of bandwagon effects. Therefore, we explored a case in which, in addition to high cumulativeness of technological advance in computers (as in the benchmark case), computer ﬁrms have the capabilities of developing advanced components in-house. In the model we reduced the size of the external market and we increased the share of the Mod of the last component purchased by a system producer that a newly integrated ﬁrm is able to obtain. As Figure 7.4 shows, computer ﬁrms become integrated and the industry remains quite competitive. 7.6.1.4

CASE 4: CHANGING BOUNDARIES OF FIRMS (IN TERMS OF VERTICAL INTEGRATION AND SPECIALIZATION): LOW CUMULATIVENESS, LOW IN-HOUSE CAPABILITIES TO DEVELOPING ADVANCED COMPONENTS BY COMPUTER PRODUCERS AND LIMITED EXTERNAL MARKETS

What happens on the contrary if computer ﬁrms have limited cumulativeness of technological advance and limited in-house capabilities in component

179

Evolution of the Computer Industry Number of firms 25 System F 20

Component F

15 10 5 0 1

22 43 64 85 106 127 148 169 190 211 232

(a) Herfindahl index 1.0 0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0 1 16 31 46 61 76 91 106 121 136 151 166 181 196 211 226 241 (b) Number of integrated firms 2.0 1.8 1.6 1.4 1.2 1.0 0.8 0.6 0.4 0.2 0 1 16 31 46 61 76 91 106 121 136 151 166 181 196 211 226 241 (c) Figure 7.2 (a–f) Specialization: large external market

180

Evolution of the Computer Industry Ratio: number of integrated firms/total number of firms 1.0 0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0 1

16 31 46 61 76 91 106 121 136 151 166 181 196 211 226 241

(d) Share of specialized leader and of integrated leader 1.0 Specialized 0.8

Integrated

0.6 0.4 0.2 0 1

21 41 61 81 101 121 141 161 181 201 221 241

(e) Total share of specialized firms and integrated firms 1.0 0.8 0.6

Specialized

0.4

Integrated

0.2 0 (f) 1

21 41 61 81 101 121 141 161 181 201 221 241

Figure 7.2 (Continued)

181

Evolution of the Computer Industry Number of firms 18 16 14 12 10

System F Component F

8 6 4 2 0 1

22 43 64 85 106 127 148 169 190 211 232

(a) Herfindahl index 1.0 0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0 1 16 31 46 61 76 91 106 121 136 151 166 181 196 211 226 241 (b) Number of integrated firms 2.5 2.0 1.5 1.0 0.5 0 (c)

1

16 31 46 61 76 91 106 121 136 151 166 181 196 211 226 241

Figure 7.3 (a–f) Vertical integration, no uncertainty in demand

182

Evolution of the Computer Industry Ratio: number of integrated firms/total number of firms 1.0 0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0 1

16 31 46 61 76 91 106 121 136 151 166 181 196 211 226 241

(d) Share of specialized leader and of integrated leader 1.0 0.8 Specialized

0.6

Integrated

0.4 0.2 0 1

21 41 61 81 101 121 141 161 181 201 221 241

(e) Total share of specialized firms and of integrated firms 1.0 0.8 0.6

Specialized

0.4

Integrated

0.2 0 (f)

1

21 41 61 81 101 121 141 161 181 201 221 241

Figure 7.3 (Continued)

183

Evolution of the Computer Industry Number of firms 25 System F 20

Component F

15 10 5 0 1

22 43 64 85 106 127 148 169 190 211 232

(a) Herfindahl index of system market 1.0 0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0 1

16 31 46 61 76 91 106 121 136 151 166 181 196 211 226 241

(b) Number of integrated firms 7 6 5 4 3 2 1 0 1 (c)

16 31 46 61 76 91 106 121 136 151 166 181 196 211 226 241

Figure 7.4 (a–f) Vertical integration, high in-house capabilities, cumulativeness of technological advances

184

Evolution of the Computer Industry Ratio: number of integrated firms/total number of firms 1.0 0.8 0.6 0.4 0.2 0 1

20 39 58 77 96 115 134 153 172 191 210 229 248

(d) Share of specialized leader and of integrated leader 1.0 Specialized 0.8

Integrated

0.6 0.4 0.2 0 1

21 41 61 81 101 121 141 161 181 201 221 241

(e) Total share of specialized firms and of integrated firms 1.0 0.8 0.6

Specialized

0.4

Integrated

0.2 0 1

21 41 61 81 101 121 141 161 181 201 221 241

(f) Figure 7.4 (Continued)

design, in the presence also of limited external markets? In other words, what happens if the major factors affecting integration and specialization discussed in Cases 1 and 3 are absent? In the model, the size of the external market for components has been greatly reduced, while the role of public knowledge in the advancement of the Mod in computers has been increased together with the share of the Mod

185

Evolution of the Computer Industry of the last component purchased by a system producer that a newly integrated ﬁrm is able to obtain has been lowered. As Figure 7.5 shows system ﬁrms are continuously changing boundaries in terms of specialization and integration. Integration is driven essentially by the maturity of each component technology, while specialization is brought in by each component discontinuities. After discontinuities, the choice of vertical integration compared to specialization is due to the balance between the Mod of component producers (leading to disintegration) and the maturity of component technology and the size of computer producers (leading to integration). In general, the limited cumulativeness of technological advance in systems creates a very competitive computer market. 7.6.1.5

CASE 5: FROM INTEGRATION TO SPECIALIZATION: THE ROLE OF BIG TECHNOLOGICAL DISCONTINUITIES IN COMPONENTS

Even in the presence of a major drive towards integration such as high cumulativeness and high capabilities of developing advanced components by the computer industry (as in Case 3 earlier), specialization by system producers may take place because of major technological discontinuities. In our model, we have done that by having the merit of microprocessor design introduced by the new microprocessor component producers very high compared with the previous integrated circuit technology. As Figure 7.6 shows, the technological discontinuity is so strong and the new components produced by the new entrants are so superior to the components produced in-house by the system producers that vertically integrated ﬁrms have to move to specialization. Computer producers will remain specialized until the new technology matures and uncertainty about the new technology diminishes. 7.6.1.6

CASE 6: CHANGING BOUNDARIES OF FIRMS: INTEGRATION DECISION DEPENDS ONLY ON THE MATURITY OF THE TECHNOLOGY

So far, in our model the choice to integrate by computer ﬁrms was a function of the maturity of component technology and of their large relative size compared to component producers, while their decisions to disintegrate was a function of the relative merit of in-house components as compared to the merit of the components produced by suppliers. What would happen if the integration strategy depends only on the maturity of the component technology (i.e. if in the early stages of component technology computer ﬁrms may want to be specialized because of the young age of the new component technology, while they may want to integrate when technology matures)? Even in a scenario like the one discussed in Case 1 (which would lead to specialization), ﬁrms’ boundaries are changing continuously, as Figure 7.7

186

Evolution of the Computer Industry

Number of firms 25 System F 20

Component F

15 10 5 0 1 22 43 64 85 106 127 148 169 190 211 232 (a) Herfindahl index 1.0 0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0 1

16 31 46 61 76 91 106 121 136 151 166 181 196 211 226 241

(b) Number of integrated firms 8 7 6 5 4 3 2 1 0 1

16 31 46

61 76 91 106 121 136 151 166 181 196 211 226 241

(c) Figure 7.5 (a–f) Changing boundaries of the ﬁrms

187

Evolution of the Computer Industry Ratio: number of integrated firms/total number of firms 1.0 0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0 1

16 31 46 61 76 91 106 121 136 151 166 181 196 211 226 241

(d) Share leader of specialized leader and of integrated leader 1.0 Specialized

0.8

Integrated 0.6 0.4 0.2 0 1

21 41 61 81 101 121 141 161 181 201 221 241

(e) Total share of specialized firms 1.0 0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0 (f)

Specialized Integrated

1

21 41 61 81 101 121 141 161 181 201 221 241

Figure 7.5 (Continued)

188

Evolution of the Computer Industry Number of firms 25 System F 20

Component F

15 10 5 0 (a)

1 22 43 64 85 106 127 148 169 190 211 232 Herfindahl index

1.0 0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0 1

16 31 46 61 76 91 106 121 136 151 166 181 196 211 226 241

(b) Number of integrated firms 9 8 7 6 5 4 3 2 1 0 1 16 31 46 (c)

61 76 91 106 121 136 151 166 181 196 211 226 241

Figure 7.6 (a–f) Specialization: high role of technological uncertainty

189

Evolution of the Computer Industry Ratio: number of integrated firms/total number of firms 1.0 0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0 1

16 31 46 61 76 91 106 121 136 151 166 181 196 211 226 241

(d) Share of specialized leader and integrated leader 0.18 0.16 0.14 0.12 0.1 0.08 0.06 0.04 0.02 0

Specialized Integrated 1

21 41 61 81 101 121 141 161 181 201 221 241

(e) Total share of specialized and integrated firms 1.0 0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0

Specialized Integrated

1

21 41 61 81 101 121 141 161 181 201 221 241

(f) Figure 7.6 (Continued)

190

Evolution of the Computer Industry Number of firms 25 System F 20

Component F

15 10 5 0 1

22 43 64 85 106 127 148 169 190 211 232

(a) Herfindahl index of system market 1.0 0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0 (b)

1

16 31 46 61 76 91 106 121 136 151 166 181 196 211 226 241 Number of integrated firms

10 9 8 7 6 5 4 3 2 1 0 1

16 31 46 61 76 91 106 121 136 151 166 181 196 211 226 241

(c) Figure 7.7 (a–f) Integration decision depends only on the maturity of the technology

191

Evolution of the Computer Industry Ratio: number of integrated firms/total number of firms 1.0 0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0 1

16 31 46 61 76 91 106 121 136 151 166 181 196 211 226 241

(d) Share of specialized leader and of integrated leader 1.0 Specialized 0.8

Integrated

0.6 0.4 0.2 0 1

21 41 61 81 101 121 141 161 181 201 221 241

(e) Total share of specialized firms and integrated firms 1.0 Specialized 0.8

Integrated

0.6 0.4 0.2 0 (f)

1

21 41 61 81 101 121 141 161 181 201 221 241

Figure 7.7 (Continued)

192

Evolution of the Computer Industry shows. Initially ﬁrms integrate as a consequence of the increasing maturity of the semiconductor technology, but when a new technology comes they specialize. Because integration decisions is driven only by the maturity of the technology after the second discontinuity has taken place, over time ﬁrms will have a major incentive to integrate, disregarding their size compared to the one of the largest component suppliers. This proves wrong decisions for the small computer producers, which are then forced to disintegrate again. However, again, the maturity of the component technology drives them to integration, followed again by specialization. Thus, the model generates continuous shifts between integration driven by maturity and specialization driven by the higher Mod of component suppliers. 7.6.1.7

CASE 7: AN EXTREME CASE: VERTICAL INTEGRATION BRINGS TO THE DISAPPEARANCE OF THE COMPONENT INDUSTRY

Finally, as an extreme case we have examined the case in which the dynamics of vertical integration leads to the disappearance of the component industry. In the model this is the case when the component industry sells mainly to the computer ﬁrms and the external market is very small while the system industry has high capabilities of producing in-house the components. As Figure 7.8 shows, because of the capability of developing components in-house, the computer industry becomes vertically integrated quite soon. Once the industry becomes integrated, the component producers face a very small demand and they exit the market. The two technological discontinuities in the component industry do allow for entry of new suppliers. However, these suppliers are not able to survive, given the limited external markets and the absence of demand coming from the fully integrated system producers. In sum, in this case vertical integration affects the actual growth and the very existence of the component industry.

7.7 Conclusions This chapter has examined ﬂexibility and stability in the organizational boundaries of ﬁrms. It has focused on stability as well as change in vertical integration and specialization and has related them to a set of key factors: capabilities, maturity of technology, size of ﬁrms, and size of markets. While the relative size of system and component producers, the age of component technology, and the different Mod of in-house and external components are explicitly expressed in the model as major factors leading to specialization or vertical integration, other factors affect the boundaries of ﬁrms by acting on ﬁrms’ relative size, components’ relative Mod, and age of technology.

193

Evolution of the Computer Industry Number of firms 25 System F 20

Component F

15 10 5 0 (a)

1

22 43 64 85 106 127 148 169 190 211 232 Herfindahl index of system market

1.0 0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0 (b)

1

16 31 46 61 76 91 106 121 136 151 166 181 196 211 226 241 Number of integrated firms

8 7 6 5 4 3 2 1 0 1 16 31 46 (c)

61 76 91 106 121 136 151 166 181 196 211 226 241

Figure 7.8 (a–f) The disappearance of the component industry

194

Evolution of the Computer Industry Ratio: number of integrated firms/total number of firms 1.0 0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0 1 16 31 46 61 76 91 106 121 136 151 166 181 196 211 226 241 (d) Share of specialized leader and of integrated leader 1.0 Specialized 0.8

Integrated

0.6 0.4 0.2 0 (e)

1

21 41 61 81 101 121 141 161 181 201 221 241 Total share of specialized firms and of integrated firms

1.0 Specialized

0.8

Integrated 0.6 0.4 0.2 0 (f) 1

21 41 61 81 101 121 141 161 181 201 221 241

Figure 7.8 (Continued)

195

Evolution of the Computer Industry One is the capability of producing components in-house, which affects the size of ﬁrms (through proﬁts and growth) and the relative merit of internally developed components with respect to externally purchased components. Relatedly, the cumulativeness at the ﬁrm level of technological advance in systems affects the innovativeness, proﬁtability, growth, and therefore size of ﬁrms. The second is the level of a technological discontinuity, which affects the difference between the merits of internally produced components and the merit of the new components introduced with the discontinuity. The third is the size of the market, both in terms of external market for components and in terms of derived demand of components from systems producers. The presence of a large external market affects the growth of capable and large components producers, which then affects the specialization of system producers. For similar reason, the presence of a large number of vertically integrated ﬁrms reduces the derived demand for components from the computer industry, and reinforces the trend for vertical integration by reducing the possibility for growth of component producers. Therefore, as mentioned in the introduction, the level and distribution of competences, the structure of markets, and the size and type of demand come to play a very important role in the explanation of vertical integration and specialization. And equally important are their coupled dynamics and the coevolution of the upstream and downstream industries. Exogenous technological discontinuities, coupled with the endogenous innovative activities of ﬁrms set in motion changes and transformation in the distribution of ﬁrms’ sizes and market structure, which in turn feed back in innovativeness, market structure, and ﬁrms’ boundaries. Periods of stability are punctuated by sharp discontinuities and structural transformation. In sum, the ability of ﬁrms to react promptly and ﬂexibly to changing market and technological conditions is a crucial variable in determining industry dynamics. But ﬂexibility here does not simply imply adaptive capabilities, but also the ability to endogenously induce structural change, through innovation and growth.

Note 1. While the production costs of integrated computer producers are a function of the achieved Mod, the production costs of specialized producers are instead determined as the costs of the system plus the cost of buying the components on the marketplace, i.e. the price charged by the particular supplier from which the computer company is buying. In the model, we assume that an integrated and a specialized ﬁrm having the same computer Mod have also the same production costs for a computer. For a given component Mod, the cost of internally produced components is equal to the cost of

196

Evolution of the Computer Industry the externally produced components. The additional costs that would be associated to the mark-up charged by component suppliers and ‘saved’ by an integrated ﬁrm are invested in R&D and treated as a cost.

References Arora, A. and Bokhari F. (2000). Vertical Integration and Dynamics and Industry Evolution. Carnegie Mellon University Mimeo. Bresnahan, T. F. and Malerba, F. (1999). ‘Industrial Dynamics and the Evolution of Firms’ and Nations’ Competitive Capabilities in the World Computer Industry’, in D. Mowery and R. Nelson (eds.) The Sources of Industrial Leadership. Cambridge: Cambridge University Press. Christensen, C. M. and Rosenbloom, R. S. (1994). ‘Technological Discontinuities, Organizational Capabilities, and Strategic Commitments’, Industrial and Corporate Change, 3: 655–85. Cohen, W. and Malerba, F. (2001). ‘Is the Tendency to Variation a Chief Source of Progress?’ Industrial and Corporate Change, 3. Jacobides, M. and Winter, S. G. (2005). ‘The Coevolution of Capabilities and Transaction Costs: Explaining the Institutional Structure of Production’, Strategic Management Journal. Klepper, S. (1996). ‘Entry, Exit and Innovation over the Product Life Cycle’, American Economic Review, 86(3): 562–82. Kricks, G. (1995). ‘Vertical Integration in the Mainframe Computer Industry: a Transaction Cost Interpretation’, Journal of Economic Behavior and Organization, 26: 75–91. Langlois, R. N.(1992). ‘Transaction Cost Economics in Real Time’, Industrial and Corporate Change, 1(1): 99–127. —— and Robertson, P. L. (1995). Firms, Markets and Economic Change: A Dynamic Theory of Business Institutions. London: Routledge. Malerba, F., Nelson, R., Orsenigo, L., and Winter, S. (1999). ‘History-Friendly Models of Industry Evolution: The Computer Industry’, Industrial and Corporate Change, 1: 3–41. Mowery, D. (ed.) (1996). The International Computer Software. A Comparative Study of Industry Evolution and Structure. Oxford: Oxford University Press. Nelson, R.(1994). ‘The Co-evolution of Technology, Industrial Structure, and Supporting Institutions’, Industrial and Corporate Change, 3: 47–63. —— and Winter, S. (1982). An Evolutionary Theory of Economic Change. Cambridge, MA: Harvard University Press. Smith, A. (1776). The Wealth of Nations. Stigler, G. (1951). ‘The Division of Labour Is Limited by the Extent of the Market’, Journal of Political Economy, 59: 185–93. Teece, D. J. (1986). ‘Proﬁting from Technological Innovation’, Research Policy, 15: 285–306. —— and Pisano, G. P. (1994). ‘The Dynamic Capabilities of Firms: An Introduction’, Industrial and Corporate Change, 3: 537–56. ——, ——, and Shuen, A. (1992). Dynamic Capabilities and Strategic Management. Working Paper, Berkeley University.

197

Evolution of the Computer Industry Williamson, O. (1975). Markets and Hierarchies. New York: Free Press. Winter, S. (1984). ‘Schumpeterian Competition in Alternative Technological Regimes’, Journal of Economic Behavior and Organization. —— (1987). ‘Knowledge and Competence as Strategic Assets’, in D. J. Teece (ed.) The Competitive Challenge. Cambridge, MA: Ballinger, pp. 159–84.

198

THEME 3 INNOVATING AND TECHNOLOGICAL TRANSFORMATION

This page intentionally left blank

8 The Effects of Technological Change on the Boundaries of Existing Firms Paul L. Robertson and Gianmario Verona

8.1 Introduction Transformation is a vital process at the level of the economy as a whole, but it is often accomplished through ﬂexibility and stability at the level of individual ﬁrms. In fact, the words ﬂexibility and stability neatly characterize the technological evolution of the 90 per cent or more of modern economies that operate in established industries and use well-known technologies, but are nevertheless regularly affected by changes of greater or lesser degrees of importance (Robertson et al. 2003). Except in the rare cases in which technological underpinnings are rendered totally obsolete, the primary requirement for ﬁrms in established industries is to learn how to blend old and new technologies, that is to ﬁnd efﬁcient ways of ﬁtting new technologies into their existing rosters of products, processes, and routines. Moreover, technological change within existing ﬁrms may need to be compatible not only with existing technologies but also with other activities that affect the ability of ﬁrms to implement changes. These extend beyond conventional organizational factors associated with ‘change management’ to include areas such as sales, marketing, and procurement in which personnel need to understand the nature of technological improvements and their importance both for themselves and for their customers and suppliers. Dealing with change is clearly a dynamic process in which some (aspects of) old technologies become obsolete while new technologies may simultaneously gain in importance. It is also an irregular and contingent process that is not totally susceptible to managerial control. As new technologies often emerge from outside a given industry (and, even more so, from outside a given ﬁrm), assimilating new technologies can be a reactive process in which ﬁrms seek to gain or retain competitive advantage by encouraging ﬂexibility in order to permit orderly evolution within their more or less stable stock of

201

Effects of Technological Change resources and capabilities. Thus they need to develop ‘receptive capacity’ (Robertson et al. 2003). Receptive capacity includes control over physical and ﬁnancial resources as well as the intellectual elements that Cohen and Levinthal (1990) have termed ‘absorptive capacity’. Acquisition of knowledge is clearly a necessary condition for technological updating but it must be accompanied by other types of resources to ensure the successful implementation of change. In this chapter, we argue that technological change may have diverse effects on how ﬁrms organize their operations, and in particular on whether they internalize speciﬁc activities or choose to outsource them. We proceed by investigating some ways in which ﬁrms alter their boundaries to adjust to both the acquisition and the generation of new knowledge and the maturing or even obsolescence of old knowledge. In some cases, the challenges posed by change may be best accommodated by outsourcing activities that can be supplied most efﬁciently through market processes. In others, ﬁrms may need to add to their range of in-house activities if they are to manage change successfully. We argue that the pattern of change is highly variable in practice because it depends not only on the receptive capacity (access to knowledge and to physical and ﬁnancial resources) of the ﬁrm itself but also to distribution of these resources throughout the economy. To take a simple example, many shipbuilders have not felt a need to develop or produce their own power plants because independent suppliers have undertaken research, development, and production more efﬁciently than the shipbuilders can. On the other hand, the overall conceptualization and design of new types of vessels is a major source of competitive advantage and has therefore been retained by the shipbuilders. This division of labour is not solely the result of an uneven distribution of knowledge:1 as ships and their power plants are co-specialized, both shipbuilders and engine manufacturers must have substantial amounts of absorptive capacity concerning the other’s activities. Furthermore, as many shipbuilders and engine manufacturers are large ﬁrms with impressive technological knowledge held in-house, they may well have the other resources needed to integrate forward or backward if they felt it worth their while. Why, therefore, do the two types of ﬁrms specialize? The answer is that, given the market conditions that they face, the opportunity cost of integration is excessive (Langlois and Robertson 1995) and modularity of functions offers the cheapest joint solution for development and production of ships and their engines. There are several, but not necessarily mutually exclusive, explanations for why ﬁrms may choose to either internalize or outsource. Chandler (1977, 1990) presents vertical integration as a defensive move to compensate for market failure that might reduce the throughput of ﬁrms with high ﬁxed investments. Williamson (1985) goes further by singling out asset speciﬁcity and opportunistic behaviour as the major (although not the only) causes of market failures that drive ﬁrms to internalize activities in order to protect

202

Effects of Technological Change themselves. A third type of explanation has been offered by Langlois and Robertson (1995) who argue that a major cause of internalization has been impacted technological knowledge. In contrast to Williamson’s general thrust, however, the transaction costs that they emphasize are ‘dynamic’ (Langlois 1992) in that they can be expected to diminish as uncertainty is reduced and the rate of change moderates. For this reason one might expect that, in the long run within a single industry, the level of vertical integration would decrease in step with reductions in those transaction costs related to maturing technology. Langlois and Robertson (1995) illustrate their thesis with a study of vertical integration in the US automobile industry between 1900 and 1940, concluding that patterns of development in the industry were highly consistent with their story. Recent research has reinforced the message of Langlois and Robertson that ﬁrms might be expected to become less vertically integrated as they and their industries mature.2 Writers on modularity, for instance, have emphasized that the need for ﬁrms to learn about new technologies might be restricted if they could segment their activities through the use of standardized interfaces to allow technologies to develop virtually independently in each segment. These may then be assembled (integrated) on a ‘plug and play’ basis on the assumption that optimizing the performance of each segment will lead to optimization for the system as a whole (Sanchez and Mahoney 1996; Baldwin and Clark 1997, 2000). Similarly, new communications technologies may reduce transaction costs and therefore make markets increasingly efﬁcient. As a result, ﬁrms might spin off activities and choose instead to buy from specialist suppliers in competitive markets with less risk of opportunism than in the past.3 Nevertheless, it is not altogether clear that technological change will necessarily reduce the attractions of vertical integration in innovative situations or diminish the importance of vertically integrated ﬁrms. For example, a study of globalization in the motor vehicle industry (Robertson 2003b) found that recent patterns of spin-offs are not exactly as might have been anticipated on the basis of the ideas of Langlois and Robertson (1995). While there has been a great increase in the range of components that are being outsourced by the major automobile producers, vertical integration has retained its importance because, over the years, the auto ﬁrms have embraced new technologies dependent on proprietary knowledge as they have simultaneously been outsourcing older technologies for which knowledge has become widespread. Likewise, evidence from the pharmaceutical industry (e.g. Yeoh and Roth 1999; Cockburn et al. 2000; Thomke and Kuemmerle 2002) highlights the relevance of size without neglecting the emergence of more complex, network-based relationships in the market. While the passage from the traditional and stochastic random-screening approach to drug discovery to the more recent rational drug design (Henderson 1994) and IT-based methodologies (Arora et al. 2001) has reduced the importance of ﬁrm size

203

Effects of Technological Change per se, it seems to have not reduced the relevance of vertical integration in the management of the relational web that ﬁrms use to generate competitive advantage. To make sense of this contrasting empirical evidence, our aim in this chapter is to explain why the a priori effects of technological change on ﬁrm boundaries are indeterminate, that is that introducing new technologies into an existing technological framework may lead to increased internalization of activities, increased outsourcing, or both depending on the context in which the change occurs and depending on the nature of change—whether it is incremental or radical. Our presentation rests on two subsidiary pole hypotheses that we test to prove the broader hypothesis of indeterminism. They are: 1. that technological change will necessarily lead ﬁrms to undertake greater internalization (or less outsourcing) of activities; and 2. that technological change will necessarily lead ﬁrms to undertake more outsourcing (reduced internalization) of activities. By showing that neither of these hypotheses holds, the necessity of indeterminacy is demonstrated. In the course of testing the hypotheses, we also develop a model to explain why, as a result of technological change, ﬁrms might simultaneously outsource some activities that they had formerly preferred to produce in-house while increasing their vertical integration in other areas that had previously been beyond their boundaries. Our argument is that dynamic transaction costs may continue to be important even in longestablished industries because the types of technologies that are of importance are continually changing, forcing ﬁrms to enter into new and unfamiliar areas in order to maintain their competitiveness; and that, at least in their pure forms, new organizational forms such as modularity are frequently impractical during design, development, and production because they do not provide for as much integration in learning as may be achieved when there is vertical integration.4 Because our argument is meant to show the importance of context and contingency for the effects of technological change on ﬁrm boundaries, we have drawn eclectically from a wide range of underlying arguments. Although the initial assumptions that underpin the various arguments vary considerably, we believe that they can be used side by side because, at the point on which we focus, all can be translated meaningfully into terms that are consistent with basic premises of evolutionary economics. Therefore, they all provide evidence that we can use to test our two hypotheses. Our approach is ﬁrst to deﬁne the determinants of the boundaries of the ﬁrm under stable technological conditions and then to investigate systematically how those boundaries might alter in the face of technological change in the ﬁrm’s internal and external environments. In Section 8.2, we establish a

204

Effects of Technological Change sort of baseline by examining the activities that will generally be part of the internal activities of a typical ﬁrm in a stable technological environment. We analyse these on the basis of three sets of factors: (a) distinctive and core capabilities; (b) behavioural issues that may lead to indivisibilities; and (c) transaction costs. Section 8.3 then investigates the effects of technological change on the boundaries of a ﬁrm that has hitherto been operating in a stable environment. The discussion encompasses the effects of changing technologies on dynamic transaction costs, and the relationship between technological change and modularity. Our conclusion is that, depending on circumstances, technological change will have diverse and possibly conﬂicting outcomes for the boundaries of the ﬁrm.

8.2 Technological stability and the boundaries of the ﬁrm Technological change does not feature prominently in most analyses of the boundaries of the ﬁrm. This argument may seem puzzling in that, in neoclassical economics, the boundaries of the ﬁrm are technologically determined by the relevant production function (Blair and Kaserman 1983; Williamson 1985), but the approach taken both by the neoclassicists and later writers generally ignores the effects of subsequent major changes. Thus, the dynamics in Chandler’s story (1962, 1977, 1990) are based on particular characteristics common to many technologies of the ﬁrst half of the twentieth century—the presence of substantial economies of scale and capital-intensive production methods—but he generally treats technology per se as a given and pays little attention to the effects of technological change on ﬁrm boundaries because he believes that later incremental changes did not undermine the importance of these characteristics. Similarly, in his discussion of hierarchies, Williamson (1985) is primarily concerned with impacted knowledge, the manifestations of power relationships, and their effects on transaction costs. While he acknowledges that some of the relevant factors may have technological bases, he is adamant that technology is not the principal inﬂuence on ﬁrm boundaries (Williamson 1985: esp. ch. 4). Even the theory of dynamic transaction costs (Langlois 1992; Langlois and Robertson 1995) is more concerned with the diffusion of existing knowledge (including incremental changes) within a particular generation of technological knowledge than with the way in which the later introduction of a new, more highly disruptive, generation of technological knowledge might alter the boundaries of the ﬁrm. As a matter of analytical convenience, we begin with a discussion of these stable scenarios, in which technological change is incremental and occurs only within what may be considered the ‘life cycle’ of a particular technology.5 Our intention is to show how inclusive a ﬁrm’s boundaries might be expected to be at a particular point in time.

205

Effects of Technological Change

8.2.1

Core and distinctive competences

The concept of the ﬁrm as a bundle of resources goes back to the pioneering work by Penrose (1959) and has been widely considered as a key driver of the processes that Chandler analysed as contributing to the evolution of the American ﬁrm (1962, 1977, 1990). In recent years, owing to its consistency with attempts in strategic management to explain the roots of competitive advantage (Cockburn, Henderson, and Stern 2000), the resource-based perspective (RBP) (Pandian and Robertson 2003) has received a tremendous boost. A key contribution to the RBP is the development of the concepts of core competences and distinctive competences.6 When ﬁrms are conceived as a bundle of assets, conceptual and empirical models show that they ought to internalize competences that are valuable, rare, and inimitable (e.g. Lippman and Rumelt 1982; Barney 1986; Dierickx and Cool 1989; Reed and DeFillippi 1990; Peteraf 1993). These properties have helped clarify that the most critical ingredients of the resource endowment are generally neither tangible, such as physical and ﬁnancial assets, nor intangible, such as human capital and reputation, resource categories already identiﬁed in the traditional literature on strategic management (Grant 1991; Amit and Schoemaker 1993). These types of resources tend, in fact, to be tradable in markets (Barney 1986), and few of them can be productive in isolation (Grant 1991). Instead, rents are generated primarily by intellectual and knowledge-based capabilities (Verona 1999), which, because they accumulate over time, are strictly idiosyncratic (Dierickx and Cool 1989). While many ﬁrms tend to have few distinctive competences (Barney 1991), this does not mean that they have tight limits on the range of resources they possess or that they tend to reduce this range. Distinctive competences are the basis of the process of ﬁrm diversiﬁcation in related products and markets (Rumelt 1974; Prahalad and Hamel 1990; Mahoney and Pandian 1992), and therefore may lead to increases in the number of additional (or ‘ancillary’) activities that a ﬁrm undertakes as it enters related markets. (Furthermore, no empirical study so far has shown that an enhanced priority given to resources in the strategic agenda of ﬁrms leads to a reduction in the internalized activities devoted to serving existing markets.) Finally, competences present a path-dependent trajectory that is cumulative (Nelson and Winter 1982; Dosi and Malerba 1996). This means that it is sometimes hard to get rid of speciﬁc competences even when they become less distinctive (Leonard-Barton 1992), and most important to our argument, that the cumulative nature of learning helps companies ﬁnd fruitful new ways to deploy resources over time (Teece et al. 1997). Distinctive competences may ﬁrst of all be described in relation to a ﬁrm’s technology.7 They tend to be embodied in core products (Prahalad and Hamel 1990), which represent the interface between the ﬁrm’s technical and distinctive knowledge and the products it delivers to the market. More broadly, competences can be described in relation to a ﬁrm’s value chain (Grant 1996)

206

Effects of Technological Change or to its system of activities (Porter 1996; Ghemawat and Pisano 1999), and can therefore be discerned in all the functions a ﬁrm manages. The nature of distinctive capabilities, then, includes both technological functions (e.g. R&D and operations management) and commercial functions (e.g. marketing and sales), as well as the ability to integrate the two (through organizational and strategic processes). However, their speciﬁc relevance varies across different industries (Henderson and Mitchell 1997). In the case, for instance, of the pharmaceutical industry, the presence of a large sales force and superior ability in managing R&D processes are considered important competences for generating competitive advantage (Yeoh and Roth 1999). How do distinctive competences affect organizational boundaries? While the empirical evidence on this issue is ambiguous, it is clear that due to the complexity of resource-based management, managers often ﬁnd it strategically desirable to internalize competences that are: (a) key to the present creation of industry value (e.g. Amit and Schoemaker 1993); (b) consistent with the strategy of related diversiﬁcation they want to put into practice (e.g. Rumelt 1974); and/ or (c) useful to build new avenues of value creation for the ﬁrm’s competitive future (e.g. Prahalad and Hamel 1990). Therefore, because a ﬁrm needs to protect the integrity of its core competences, and in particular of its distinctive competences, these will be included within the ﬁrm’s boundaries.

8.2.2

Transaction costs

Williamson (1985: 1) contends that markets may potentially be employed whenever there is a transaction: ‘A transaction occurs when a good or service is transferred across a technologically separable interface. One stage of activity terminates and another begins’. As a result, the boundaries of a ﬁrm (the extent to which it internalizes activities that might otherwise be conducted through markets) are, other things being equal, a function of the relative costs of conducting transactions internally and externally. To Williamson, common— perhaps nearly universal—attempts by suppliers and distributors to act ‘with guile’ undermine the efﬁciency of markets and drive up transaction costs to the point at which it is cheaper for a ﬁrm to undertake many activities itself that might otherwise be purchased through markets. If transaction costs are relatively low, a ﬁrm will outsource activities, but in the presence of signiﬁcant asset speciﬁcity, opportunism, or other sources of transaction costs, a ﬁrm will tend to undertake activities itself as a cheaper alternative to employing markets. Williamson’s analysis is not dependent on technology per se. Indeed, he devotes considerable space to explaining why he believes that technology is not an important factor in determining the extent of vertical integration (although he concedes that the types of technology employed can certainly affect the costs of conducting transactions as well as production costs, the other major variable in the make-buy decision) (Williamson 1985). By con-

207

Effects of Technological Change trast, in his discussion of dynamic transaction costs, Langlois (1992) makes technology a central factor in determining the boundaries of the ﬁrm. Langlois avoids the ‘technological determinism’ of the neoclassicists that Williamson ﬁnds unconvincing, and instead constructs his argument along circumstantial lines. Following Morris Silver (1984), Langlois (1992) and Langlois and Robertson (1995) note that innovative ﬁrms may well be forced to undertake activities they would rather outsource. The primary reason for this is the difﬁculty that ﬁrms may have in conveying unfamiliar knowledge to others because the concepts are new and therefore hard for others to comprehend. Alternatively, potential suppliers may be unwilling to supply customized inputs because of scepticism as to the viability of the innovation.8 As time passes, however, knowledge spreads, increasing the potential pool of technologically qualiﬁed external suppliers and reducing scepticism. Whereas virtually all factors of production may be candidates for outsourcing in Williamson’s scheme,9 provided that the cost structure is appropriate; Langlois and Robertson, in common with supporters of the RBP, contend that some activities should remain part of a permanent core within the boundaries of the ﬁrm. The activities that they nominate as candidates for (often involuntary) insourcing in the initial phases of an innovation, followed by outsourcing as knowledge spreads, are instead ancillary to the main activities of the ﬁrm. As innovative technologies age, ﬁrms will increasingly outsource ancillary aspects of their activities. Furthermore, because the diffusion of knowledge increases the pool of potential suppliers, later entrants will be less highly vertically integrated than early entrants.

8.2.3

Behavioural factors

As Williamson notes, transaction costs are only one of several possible factors that underlie the structure of complex organizations. In fact, although he does not raise the issue, some of the most important activities within a ﬁrm may not involve ‘technologically separable interfaces’. These include learning by individuals within the ﬁrm through interaction with other employees as well as collective learning, as through the development of routines (Nelson and Winter 1982). The product of this localized learning has two main characteristics (Langlois and Robertson 1995). First, it is often synergistic, that is, of greater value than the combined learning of the individuals would be when operating in isolation. Furthermore, we contend that learning is context speciﬁc in two senses. For one thing, it is ‘situated’ (Lave and Wenger 1991) in that time and place may inﬂuence the intellectual and practical inputs into thought processes.10 More importantly for our purposes, the learning that occurs may depend on the membership of the group—people will ‘spark off’ others in different ways. Persons A and B will often interact differently, and develop different knowledge, than persons A and C would. Therefore, if C purchases

208

Effects of Technological Change the services of A, who until now has worked with B, this will not guarantee that A will be able to produce equivalent results in a new environment with C. As a result, the ways in which tangible, intangible, and intellectual resources are developed and used can be expected to vary from ﬁrm to ﬁrm. The essential core of each ﬁrm (what differentiates it from other ﬁrms and gives it intellectual cohesion) will be the ‘idiosyncratic synergy’11 (Langlois and Robertson 1995) that comes from the interaction of the speciﬁc group of people involved, especially in the management of innovation (i.e. R&D and marketing) and among the upper ranks of management. Similarly, the boundaries of the ﬁrm will encompass the activities that comprise the routines that give the ﬁrm its speciﬁc behavioural characteristics.

8.2.4

Summary

Figure 8.1a depicts the boundaries of a typical ﬁrm under stable conditions. E and F represent the core and distinctive capabilities of the ﬁrm, while D encompasses the idiosyncratic synergy and routines that give the ﬁrm its individual learning and behavioural characteristics. A–C are ancillary capabilities that may have been incorporated either because of high, relatively stable transaction costs in Williamson’s sense or high dynamic transaction costs as described by Langlois. But since dynamic transaction costs can be expected to lessen as knowledge spreads, in a technologically stable (but not static) environment the outcome illustrated in Figure 8.1b can be expected to emerge eventually. Here, the number of ancillary activities included within the boundaries of the ﬁrm has decreased owing to the diffusion of technological knowledge that makes outsourcing more feasible than in the early stages of an innovative product.

8.3 Technological change and the boundaries of the ﬁrm In practice, technologies and industries are not stable. Change may arise continually from diverse sources, affecting individual ﬁrms in major or (more frequently) minor ways. Wholesale destruction of a ﬁrm’s competences (Tushman and Anderson 1986) can occur, but it is uncommon. Instead, at any particular point in time, a ﬁrm is likely to use a mix of technological vintages, ranging from old and established to new and innovative knowledge, equipment, and procedures. Moreover, some new technologies may originate within an industry, but others are imported from different industries and require adjustments for use in new environments.

8.3.1

Dynamic capabilities

As highlighted by empirical research on technology–environment relations at the population level (e.g. Tushman and Anderson 1986; Anderson and

209

Effects of Technological Change

(a)

(b) A B C D E F

(c) C D E F

Distinctive/core activities Other activities

C M N D E G H Activities involving idiosyncratic synergy and routines

Figure 8.1 Activities within the boundaries of a ﬁrm

Tushman 1990; Tushman and Rosenkopf 1992), unstable technological change may impact incrementally on technological knowledge and actually enhance the current technological competences in the industry. The change may not, in fact, affect the technological core of the industry, but be primarily related to modular and architectural competences (Henderson and Clark 1990) and to problems related to the internal allocation process of ﬁrms that tend to favour large customers (Christensen and Bower 1996) or other environmental actors. Thus, even in the case of unstable technological change, ﬁrms may continue to outsource part of their ancillary activities without affecting their ability to gain competitive advantage with their core technologies. Yet in order to appropriate the rents accruing to the innovation, they must be able to deal with the complementary knowledge required not only to invent, but also to develop and launch the innovation in the market. Firms, then, must ﬁnd adequate ways to access and use complementary assets and capabilities (Teece 1986). For instance, small biotech ﬁrms could not appropriate the rents from the design and development of new drugs in the late 1980s because they lacked the commercial competences of big pharmaceutical corporations (Gambardella 1995). While part of these complementary assets may be possessed by partners and may be coordinated through market contracts (Teece 1986), it is important to directly manage part of them. For these reasons we suggest the following propositions: Proposition 1a: When unstable technological change is competence enhancing, the ﬁrm will persist in internalizing its core activities. Proposition 1b: When unstable technological change is competence enhancing, the ﬁrm will persist in outsourcing technologically related ancillary activities. Proposition 1c: When unstable technological change is competence enhancing, the ﬁrm will tend to internalize complementary knowledge to improve its ability to deliver products to the market and appropriate the consequent rents.

210

Effects of Technological Change Afuah (2001) has made a similar argument based on transaction costs and the knowledge-based theory of the ﬁrm. He contends that fortune favours the prepared in the sense that a ﬁrm’s ability to beneﬁt from a technological change that replaces one set of capabilities with another depends on the extent to which, at the time of the change, the ﬁrm is committed both to the old technology (the value of whose underlying competences is diminished by the change) and to the competences that underlie the replacement technology (whose value is increased by the change). His argument, which is related to the absorptive capacity arguments of Cohen and Levinthal (1990), is summarized by his third set of hypotheses (Afuah 2001: 1217): Proposition 2a: When unstable technological change is competence destroying (to suppliers of a major component), incumbent ﬁrms that had been vertically integrated into the old technology but are not vertically integrated into the new technology will have the worst performance of all groups of incumbents. Proposition 2b: When unstable technological change is competence destroying (to suppliers of a major component), incumbent ﬁrms that had not been vertically integrated into the old technology but are vertically integrated into the new technology will have the best performance of all the groups of incumbents. These two sets of propositions make it clear that the level of vertical integration as such does not necessarily either increase or decrease as a result of technological change. While there may be more outsourcing of activities related to core technological knowledge (Proposition 1b), this may be counterbalanced by an increase in vertical integration to internalize complementary capabilities that allow the ﬁrm to appropriate the rents (Proposition 1c). Similarly, ﬁrms that have integrated vertically into a technology that is becoming obsolete may lose proﬁtability, but ﬁrms that have integrated vertically into a technology that is becoming more signiﬁcant stand to gain from their prior investments (Propositions 2a and 2b). If, however, there is a substantial change in the technologies of core products (i.e. the change from hydraulic to digital control systems described in Brusoni et al. 2001), the boundaries of the ﬁrm could expand again, through the addition of new distinctive/core activities (G and H in Figure 8.1c). Whenever unstable technological change is competence destroying (Tushman and Anderson 1986) or disruptive (Christensen and Bower 1996), ﬁrms must pay speciﬁc attention to the quality of their technical capabilities and may be required to internalize new core competences. To analyse how ﬁrms can equip themselves to deal with both competenceenhancing and competence-destroying change, Teece et al. (1997) have developed the concept of dynamic capabilities to highlight the skills that ﬁrms need to cope with abrupt changes in markets and technologies. Dynamic capabilities are ‘the processes that use resources—speciﬁcally, the processes to integrate,

211

Effects of Technological Change reconﬁgure, gain, and release resources—to match and even create market change’ (Eisenhardt and Martin 2000: 1107). They represent the subset of capabilities that allow ﬁrms to generate new products and processes and respond to changing market circumstances (Teece et al. 1997). The importance of dynamic capabilities lies in barriers to innovation, also called core rigidities. This refers to the fact that the people, structures, managerial systems, and values that constitute the ﬁrm’s capabilities to develop and integrate knowledge can also create inertia that prevents adaptation of the innovation system (Nelson and Winter 1982; Leonard-Barton 1992; Tushman and O’Really 1997). Through the use of dynamic capabilities, ﬁrms can reconﬁgure their organizational capacity to adapt to changes in markets and technologies. This process of reconﬁguration represents a key dynamic capability of the ﬁrm and may induce a ﬁrm to internalize part of the new ancillary activities. For this reason, we posit that: Proposition 3: When unstable technological change is competence destroying, the boundaries of the ﬁrm will widen in order to include new technological core competences. Patent data (Granstrand et al. 1997; Patel and Pavitt 1998) illustrate how a ﬁrm’s activities change as its technological needs evolve in line with Propositions 1a, 1b, and 3. In their study of the US patenting behaviour of the world’s 440 largest ﬁrms, Granstrand et al. (1997) demonstrate that most of these ﬁrms maintain considerable technical ability outside their core areas. In part, this is due to a desire to maintain absorptive capacity (Cohen and Levinthal 1990), but it also reﬂects changing demands as some technologies became obsolete while others gained prominence. Granstrand et al. (1997) discuss how the range of technologies in which patenting activity occurred at Rolls-Royce and Ericsson changed as their products evolved, with additions of new technologies outweighing exits from existing ﬁelds. For example, at Ericsson, cellular phones and telecommunications cables became more ‘multitechnology’ as the products developed, leading to an extension of the ﬁrm’s technology base. ‘The new technological competences that were required outnumbered the old ones that were made obsolete; and as a result of this process, ‘‘competence enhancement’’ dominated over ‘‘competence destruction’’ just as in the case of Rolls-Royce’ (Grandstrand et al. 1997: 14).12

8.3.2

Dynamic transaction costs

Technological change will not necessarily increase or diminish the factors such as asset speciﬁcity and opportunism that Williamson (1985) has emphasized. Vertical integration may still be necessary because a ﬁrm cannot locate suppliers who are willing to take the risk of acquiring specialized dies or other equipment to aid in the production of an untried good or service, or because

212

Effects of Technological Change the ﬁrm fears that its suppliers will act with guile. As in other cases, however, there is reason to expect that Williamsonian transaction costs will be less for producers of mature goods in which a new technology is replacing an established one, than for producers of truly innovative products. For example, providing components for products with developed markets should be less risky than providing them for products for which there is, as yet, little veriﬁable demand. In addition, established ﬁrms may have a cache of experience unavailable to new ﬁrms that makes them better able to recognize opportunistic behaviour on the part of potential suppliers and, therefore, to protect themselves through shrewd negotiations. Proposition 4a: Technological change may make ﬁrms vulnerable to increased transaction costs and therefore lead to the internalization of new components and processes. Proposition 4b: Established ﬁrms producing mature products will be less vulnerable than new ﬁrms producing innovative products to transaction costs arising from asset speciﬁcity and opportunism; therefore, technological change affecting established ﬁrms producing mature products will lead to lower degrees of subsequent internalization than it will for new ﬁrms making innovative products. The principles of dynamic transaction costs are important in determining how the patterns outlined earlier operate in practice. Although the deﬁnition of ‘competence enhancement’ used by Granstrand et al. (1997) differs from that of Tushman and Anderson (1986), both deﬁnitions are valuable from the standpoint of an individual ﬁrm. In some cases, the value of existing competences will increase as a result of technological change (Tushman and Anderson 1986), but in others, ﬁrms will be forced by change to add new competences—to broaden their range of knowledge and skills—without necessarily affecting the value of their existing competences at all (Granstrand et al. 1997). While, strictly speaking, the latter process does not enhance existing competences, it is necessary to equip the ﬁrm for operating in new conditions. However, the new competences acquired may not be (often will not be) distinctive or even core competences. Instead, as a result of technological changes, a ﬁrm will continually replicate the patterns described in our earlier discussion of dynamic transaction costs. As is shown in Figure 8.1c, because some technologies age while at the same time new technologies are being added to the repertoire of knowledge that a ﬁrm draws upon, it may be necessary to intensify the resources devoted to some new activities (M, N, G, and H) while cutting back on the resources needed for others (A, B, and F), which have either been discarded as obsolete or have become so commonplace that they may be easily outsourced. As a result, the boundaries of a ﬁrm may simultaneously expand and contract. These developments are affected to a substantial degree by both traditional (‘Williamsonian’) transaction costs and

213

Effects of Technological Change by dynamic transaction costs. Over lengthy periods of time, this can be expected to happen repeatedly when the technologies used in a product alter. Furthermore, in the case of complex products with many components whose technologies change at varying rates, it may be expected that ancillary activities associated with some components will be in the process of being shed at the same time as ancillary activities associated with other components are being added to the ﬁrm’s repertoire of internalized activities: Proposition 5: Because technological change is a repeated process, boundary changes associated with dynamic transaction costs will continue to arise even for ﬁrms that produce mature products. Proposition 6: Because of varying rates of technological change among components, a ﬁrm may outsource ancillary activities needed to produce some mature components while simultaneously adding ancillary activities needed to produce other components that are technologically innovative. The extent to which a ﬁrm needs to invest in new activities as technology changes, however, varies depending on circumstances in external markets. In the case of mature products, for which supplier networks are highly developed, it is possible that independent ﬁrms may already exist to produce new components as technologies change. In fact, these independent ﬁrms may themselves develop new component technologies, thus permitting ﬁrms to get by with investments in absorptive capacity to recognize and implement important changes arising in the marketplace without needing to invest in production capabilities and with only modest investments in development activities. For example, Robert Bosch, a large German producer of electronic components for the automotive industry, has been a leader in developing as well as producing new technologies including fuel injection equipment and body electronics (for locking systems etc.). This has permitted many automobile manufacturers to restrict their activities to integrating Bosch equipment into their products without having to develop a full range of development and production capabilities themselves. Proposition 7a: Firms operating in mature industries will need to possess competences to ensure their absorptive capacity. Proposition 7b: Because they already possess strong supplier networks, ﬁrms operating in mature industries are less likely than ﬁrms in less well-developed industries to need to develop a wide range of competences when there is a signiﬁcant technological change affecting a major component.

8.3.3

Modularity and ﬁrm boundaries

In recent years, increased product and design modularity have been suggested as means of reducing the need for vertical integration in ﬁrms. When a

214

Effects of Technological Change product is a modular set of components, with standardized interfaces to ensure connectability, both producers and customers can choose among a variety of component suppliers because all standardized variations will be acceptable within the product’s architecture (Langlois and Robertson 1992). Modularity leads to potential reductions in cost because it makes it easier for several users (not all of whom are necessarily in the same industry) to share a common input and therefore gives greater access to economies of scale in development and production (Langlois and Robertson 1995). Proposition 8: Technological change within a context of modular components is less likely to lead to internalization of component production than when there is no modularity. Since the middle of the 1990s, however, modularity has acquired broader meaning as a result of being applied to activities such as design, as well as to artefacts. Writers including Sanchez and Mahoney (1996) and Baldwin and Clark (1997, 2000) have extended modularity to include both processes and organizational forms. Modularity can ease the product development process by permitting individual components to be designed independently as long as there are standardized interfaces which have been speciﬁed in advance. This makes it possible to introduce improved versions of individual components without having to alter the system as a whole, and allows for the maximization of performance for each component, leading (it is assumed) to optimal performance for the artefact as a whole. In addition, modularization can reduce the need for specialists from different areas to interact, a timeconsuming process that can lead to confusion, and it accelerates movement along the learning curve for each component/segment. Schilling and Steensma (2001) also argue that the use of modular design practices can speed up the product development process by allowing components to be used interchangeably and permitting segments to be developed in parallel rather than sequentially.13 Proposition 9: Firms using modular design practices to reduce the time and cost required for product development will have greater opportunities and incentives for the outsourcing of design and development activities. Schilling and Steensma (2001) contend that, because of its beneﬁts, modularization of design practices will be more common in industries with high rates of technological change than in industries with more stable technological structures. Nevertheless, the use of standardized interfaces is not always possible. This is especially true when an improvement in one component renders it incompatible with other components that were used in the past. When this occurs, it may be necessary to redesign large parts of the product to take advantages of the improvements that ﬂow from the improved component.

215

Effects of Technological Change Proposition 10: The introduction of a systemic technological change that destroys existing standardized interfaces may lead to reduced modularity and increased internalization of design functions. Sanchez and Mahoney (1996) and Baldwin and Clark (1997, 2000) have also argued that the concept of modularity should be applied to organizational forms. This is, in effect, a rebadging of the principles traditionally associated with the division of labour. In relation to product development and design, these authors claim that organizational forms that divide workers according to their tasks make it possible for workers to specialize and therefore to think more deeply and productively about the tasks to which they are assigned. They are also encouraged to remain up to date with their disciplinary specialisms and do not need to waste time coordinating their thoughts and their output with those of engineers and scientists in other ﬁelds because standardized interfaces make it possible for each specialism to perform its work independently. These are all principles familiar to readers of the ﬁrst three chapters of The Wealth of Nations (Smith [1776] (1937)), and they are undoubtedly correct under certain circumstances, leading to the outcomes already suggested in Propositions 8 and 9. Nevertheless, there are some drawbacks to applying modularity to organizational forms, largely because the concept tends to treat people as if they are inanimate objects and therefore ignores some of the dangers that may arise from using modularized organizations as well as the beneﬁts that may derive from closer integration (Langlois and Robertson 2003). In any project, differentiation must be accompanied by integration. In modern modular theory, integration is accomplished through the use of standardized interfaces. For better or worse, however, human thought processes cannot reasonably be characterized as being equipped with interfaces of this type. Not only do people differ from each other, but (as noted earlier) they also interact differently with different people. As a consequence, by limiting the ways in which developers can work together, the ‘design rules’ (Baldwin and Clark 2000) that modularity imposes on research may restrict how people may think and the outcomes they are likely to reach. This is true not just in the formal sense that the technical limitations imposed by standardized interfaces must be observed, but also because of limitations on thought processes as such. The best outcomes (however measured) may, therefore, be ruled out when modular design rules are applied to organizations and to interpersonal relations. Under appropriate conditions, the beneﬁts claimed for multidisciplinary research teams and close cooperation (Stalk and Hout 1990; Smith and Reinertsen 1998) will outweigh the beneﬁts foregone when modularity and specialization are curtailed. Proposition 11: When modular forms of organization for new product development or R & D place undesirable restrictions on conceptualization, ﬁrms will

216

Effects of Technological Change internalize a wider range of design, development, and research functions than when modular forms do not restrict conceptualization. Even when there is organizational modularity through the outsourcing of functions, systems integration may be needed to coordinate activities (Prencipe 1997; Brusoni et al. 2001; Brusoni 2003). This is true not only of new product development but also of complex projects involving a range of contractors and subcontractors, each of whom is charged with fulﬁlling a particular function. In the words of Brusoni et al. (2001: 617), ‘ . . . there is no one-to-one mapping between product architecture and organizational structure, as put forward by Sanchez and Mahoney (1996: 83).’ Among other things, there may be no standardized interfaces in complex projects that are designed to meet the precise needs of their users.14 Few chemical plants may be alike because of different commercial needs and also because evolving technology dictates that technical relationships tend to change constantly. This increases the need, which exists in any case in complex projects, for skilled integration to ensure that the work is accomplished on budget, on time, and according to speciﬁcation. After a careful examination of the issues involved, Brusoni (2003: 16) concludes that the role of systems integrator is best met by the customers ‘who know more than they actually do in-house’.15 Proposition 12: Even when they outsource activities, ﬁrms need to maintain enough knowledge in-house to supervise their contractors efﬁciently.

8.4 Conclusions As is summarized in Table 8.1, our analysis illustrates the ways in which technological change can promote both ﬂexibility and stability in ﬁrm structures. Our conclusions are consistent with the effects of current changes in a speciﬁc setting: information and communications technology (ICT) on ﬁrm boundaries (Box 8.1). Therefore, the shape of the ‘representative ﬁrm of the future’ remains to be seen. In fact, there may well be a proliferation of types of ﬁrm structure. Although the vertically integrated structure suggested by Chandler and Williamson is unlikely to disappear altogether, there is no reason to believe that it will not be altered signiﬁcantly under emerging conditions. The effects of technological change on new ﬁrms and on existing vertically integrated ﬁrms cannot be speciﬁed without examining a wide range of factors including: (a) the market structures for components; (b) the technological vintages of the various components used in a production process; (c) the extent to which economies of scale are present; (d) whether a process may plausibly be

217

Effects of Technological Change Table 8.1. Summary of the expected outcomes of the propositions Proposition number Proposition 1a Proposition 1b Proposition 1c Proposition 2a Proposition 2b Proposition 3 Proposition 4a Proposition 4b Proposition 5 Proposition 6 Proposition 7a Proposition 7b Proposition 8 Proposition 9 Proposition 10 Proposition 11 Proposition 12

Likely to lead to outsourcing

Likely to lead to internalization X

X X X

X X X X X

X X

X X X

X X X X X X

Box 8.1 THE IMPACT OF ICT ON FIRM BOUNDARIES Information and communication technologies (ICT) provides ﬁrms in any industry with the possibility of open, cost-effective, and ubiquitous networks (Afuha 2003). These attributes make ICT a new, global medium of unprecedented reach, substantially reducing the major constraints of geography and distance (Cairncross 1997). Modern ﬁrms can gain efﬁcient access to knowledge and competence providers throughout the world. Furthermore, ICT allows ﬁrms to go beyond traditional trade-off between richness and reach because such technologies are interactive in nature (Evans and Wurster 1999). In the classic world of tangible and physical objects, communicating and absorbing rich information required close physical proximity or personal interactions. Thus working with a large audience entailed compromises in the quality of the information ﬂows. Now, Internet-based virtual environments make it possible to interact with a large number of partners with fewer compromises in the richness of the interactions. ICT also increases the speed and the persistence of engagement with a ﬁrm’s external actors (Verona, Prandelli, and Sawhney forthcoming), extending capacity to tap into external knowledge sources. Even innovation, which since Schumpeter (1942) has traditionally been conceived as primarily an intra-organizational activity, now results increasingly from interaction among organizations and an increased division of inventive labour (Arora and Gambardella 1994). The growing ability of ﬁrms to rely on external sources thanks to ICT for it has been one of the primary reasons that many scholars have predicted increased decentralization and a tendency towards greater outsourcing. This decentralization is, among other things, the result of improved distribution of knowledge and reduced costs of using markets (Langlois 2003). Also, as suggested by Afuah (2003), ICT has contributed to a substantial reduction in component costs, asset speciﬁcity, information asymmetry, and opportunism. Hence it can be argued that the overall effect of greater ICT usage will be a general reduction in the range of activities that a typical ﬁrm undertakes in-house, allowing it to specialize on

218

Effects of Technological Change particular activities in a value chain and deliver greater value to the end market. This vertical disintegration can be seen today in several industries that, due to the more sophisticated connections with the external world that improved ICT permits, are opting for increased degrees of specialization. For instance, in the increasingly globalized automotive industry, several large ﬁrms (notably General Motors and Ford) have spun off many of their components activities into separate ﬁrms and now make extensive use of ICT to open up access to independent suppliers throughout the world. Nevertheless, the increasing ability of ﬁrms to share activities with others and the greater degree of specialization that has affected many industries in recent years has not led to an unambiguous contraction in the boundaries of ‘typical’ or ‘representative’ ﬁrms. As was argued in an earlier paper (Robertson 2003a), while new ICT can improve the efﬁciency of markets and reduce transaction costs by improving information ﬂows, this does not necessarily bring about a major shift towards outsourcing because many other complementary functions still have to be performed internally. These include not only functions for which ICT is of little relevance, but also others involving ‘transactions’ in anyone’s vocabulary. For example, customers can easily ﬁnd out what prices are being asked when potential suppliers post price lists on the Internet. But, as Hall (2001) points out, posted prices are often meant to be a starting point for negotiation and not as the amount that the ﬁrm expects to receive when a sale is concluded. Therefore, many buyers of intermediate goods cannot realistically reduce their purchasing to the equivalent of an Amazon.com ‘one-click’ procedure in which they are able to conclude transactions quickly but must accept the posted price. Instead, ﬁrms need to negotiate prices (especially, if they are large buyers who might have substantial leverage in dealing with suppliers), as well as to monitor quality, check on reliability of delivery, etc. Moreover, as transaction costs are just one aspect of the total cost equation (Williamson 1985; Langlois and Robertson 1995), reductions based on ICT improvements can be expected to have only a limited effect. Thus a breakthrough in ICT could have a variety of outcomes, but which one will prevail in any given case is an empirical matter. As is also suggested in Afuah (2003), there is no theoretical basis for determinism.

modularized; (e) the strength of Williamsonian transaction costs; and (f) ﬁrm strategy. Furthermore, as none of the papers written so far has examined the effects of technological change on the scope of activities undertaken by a ﬁrm, the complexity of the larger issue is almost certainly underestimated here. It seems clear, however, that the effects of changes will be both ﬁrm speciﬁc and path dependent. But none of this implies that the outcomes, diverse as they may be, are random. All of the propositions presented here describe rational ways of dealing with technological change in speciﬁc environments. As both internal and external environments will vary across ﬁrms and industries, it is reasonable to expect that organizational responses will also vary as ﬁrms seek out efﬁcient responses to the situations in which they are placed.

219

Effects of Technological Change

Notes 1. Although this may play a role in some cases. 2. Other authors, such as Burgelman in his work on Intel (1991), have found a similar tendency on the part of innovative manufacturers to become highly involved in generating their own process technology, much as Henry Ford was. Brusoni, et al. (2001) discovered that, in the jet engine industry, the way in which engine manufacturers dealt with control functions corresponded broadly to the predictions of Langlois and Robertson (1995). As hydraulic controls matured, engine manufacturers tended increasingly to outsource the function. When digital controls replaced hydraulic ones, however, the engine manufacturers increased their internalization of functions associated with the new control technology. 3. This and a number of related issues are canvassed in Lamoreaux et al. (2003), Langlois (2003), and Robertson (2003a). 4. As posited by Afuah (2003), while information and communication technologies are extremely effective in favouring the exchange of codiﬁed knowledge, they are less powerful when it comes to tacit knowledge—which we argue is a key resource in the ideation and development of new knowledge and technology. 5. Another way of expressing this is to say that, within a ﬁrm or industry, change does not require a move to a new experience curve (Robertson and Langlois 1994). 6. A problem with the RBP is related to terminology. As stated by Foss (1997: 348): ‘the contributions that may be seen as constituting the resource-based perspective (RBP) are undeniably quite heterogeneous in terms, for example, of the style and of the discipline from which they draw.’ Here, we will broadly use the term competences to also describe intangible resources, capabilities, and organizational assets. 7. Following Prahalad and Hamel (1990: 82): ‘Core competences are the collective learning in the organization, especially how to coordinate diverse production skills and integrate multiple streams of technologies . . . If core competence is about harmonizing streams of technology, it is also about the organization of work and the delivery of value . . . Core competence is communication, involvement, and deep commitment to working across organizational boundaries. It involves many levels of people and all functions.’ 8. The latter is closely related to Williamson’s notion of asset speciﬁcity. 9. In other words, in his model there are no core or distinctive capabilities—everything may be purchased. 10. This sense was captured nicely by Alfred Marshall when he wrote of concepts being ‘in the air’ (Marshall 1961). 11. This is similar to what Schilling (2000) subsequently termed ‘synergistic speciﬁcity’. 12. This dominance in competence enhancement (by which Granstrand et al. (1997) seem to mean the addition of new competences rather than increased value of existing ones) over competence destruction may be in part a reﬂection of the extent to which a transition to old technologies has been accomplished. When the transition is incomplete, a ﬁrm can be expected to maintain skills in the old area but may drop them later when the new technology has gained complete dominance.

220

Effects of Technological Change 13. Modularity can be used equally well for the development of non-tangible items such as computer programs, where segments of code can be written independently and easily assembled because the interfaces between the segments are known in advance. This is carried even further in the case of object-oriented programming, in which small programs written for semi-generic purposes are joined together into larger and more complicated programs through the use of standardized interfaces, thereby reducing the need to write entire programs from scratch. 14. In the account of the development of the IBM System 360 presented by Langlois (1997), IBM switched from modularity to increased vertical integration precisely to avoid changes by components suppliers that would conﬂict with IBM’s own systemic design vision. 15. A similar form of control is the monitoring of performance through ‘taper integration’. In this case, ﬁrms maintain operations to produce a small proportion of a given input, the supply of most of which is outsourced. In this way, ﬁrms learn about technical details involved in production and gain leverage in negotiations because they are better able to assess the price structures of their suppliers.

References Afuah, A. (2001). ‘Dynamic Boundaries of the Firm: Are Firms Better Off Being Vertically Integrated in the Face of a Technological Change?’ Academy of Management Journal, 44(6): 1211–28. —— (2003). ‘Redeﬁning Firm Boundaries in the Face of the Internet: Are Firms Really Shrinking?’ Academy of Management Review, 28(1): 34–53. Amit, R. and Schoemaker, P. (1993). ‘Strategic Assets and Organizational Rent’, Strategic Management Journal, 14: 33–46. Anderson, P. and Tushman, M. L. (1990). ‘Technological Discontinuities and Dominant Design: A Cyclical Model of Technological Change’, Administrative Science Quarterly, 35: 604–33. 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. —— , Fosfuri, A., and Gambardella, A. (2001). Markets for Technology. The Economics of Innovation and Corporate Strategy. Cambridge, MA: MIT Press. Baldwin, C. Y. and Clark, K. B. (1997). ‘Managing in the Age of Modularity’, Harvard Business Review, (September/October), 75(5): 84–93. —— and—— (2000). Design Rules, 1, The Power of Modularity. Cambridge, MA: Harvard University Press. Barney, J. B. (1986). ‘Strategic Factor Markets: Expectations, Luck, and Business Strategy’, Management Science, 32: 1231–41. —— (1991). ‘Firm Resources and Sustained Competitive Advantage’, Journal of Management, 17: 99–120. Blair, R. D. and Kaserman, D. L. (1983). Law and Economics of Vertical Integration and Control. New York: Academic Press.

221

Effects of Technological Change Brusoni, S. (2003). ‘Authority in the Age of Modularity’. Working paper, Science Policy Research Unit, University of Sussex. Brusoni, S., Prencipe, A., and Pavitt, K. (2001). ‘Knowledge Specialization, Organizational Coupling, and the Boundaries of the Firm: Why Do Firms Know More than They Make?’ Administrative Science Quarterly, 46(4): 597–621. Burgelman, R. A. (1991). ‘Intraorganizational Ecology of Strategy Making and Organizational Adaptation: Theory and Field Research’, Organization Science 2: 239–62. Cairncross, F. (1997). The Death of Distance: How the Communication Revolution will Change Our Lives. Boston, MA: Harvard Business School Press. Chandler, A. D., Jr. (1962). Strategy and Structure: Chapters in the History of Industrial Enterprise. Cambridge, MA: MIT Press. —— (1977). The Visible Hand: The Managerial Revolution in American Business. Cambridge, MA: The Belknap Press. —— (1990). Scale and Scope: The Dynamics of Industrial Capitalism. Cambridge, MA: The Belknap Press. Christensen, C. M. and Bower, J. L. (1996). ‘Customer Power, Strategic Investment, and Failure of Leading Firms’, Strategic Management Journal, 18: 509–33. Cockburn, I. M., Henderson, R. M., and Stern, S. (2000). ‘Untangling the Origins of Competitive Advantage’, Strategic Management Journal, 21:1123–45. Cohen, W. M. and Levinthal, D. A. (1990). ‘Absorptive Capacity: A New Perspective on Learning and Innovation’, Administrative Science Quarterly, 35: 128–52. Dierickx, I. and Cool, K. (1989). ‘Asset Stock Accumulation and Sustainability of Competitive Advantage’, Management Science, 35: 1504–11. Dosi, G. and Malerba, F. (eds.) (1996). Organization and Strategy in the Evolution of the Enterprise. London: Macmillan. Eisenhardt, K. M. and Martin, J. A. (2000). ‘Dynamic Capabilities: What Are They?’ Strategic Management Journal, 21: 1105–23. Evans, P. and Wurster, T. (1999). Blown to Bits. How the New Economics of Information Transforms Strategy. Boston, MA: Harvard Business School Press. Foss, N. J. (1997). ‘Resources and Strategy: A Brief Overview of Themes and Contributions’, in N. J. Foss (ed.) Resources, Firms, and Strategies. A Reader in the Resource-Based Perspective. New York: Oxford University Press, pp. 3–18. Gambardella, A. (1995). Science and Innovation. Cambridge: Cambridge University Press. Ghemawat, P. and Pisano, G. (1999). ‘Building and Sustaining Success’, in P. Ghemawat (ed.), Strategy and the Business Landscape. Text and Cases. Reading, MA: Addison-Wesley, pp. 111–34. Grandstand, O., Patel, P., and Pavitt, K. (1997). ‘Multi-Technology Corporations: Why They Have ‘‘Distributed’’ Rather Than ‘‘Distinctive Core’’ Competences’, California Management Review, 39(4): 8–25. Grant, R. (1991). ‘The Resource-Based Theory of Competitive Advantage’, California Management Review, 33(3): 114–34. —— (1996). ‘Prospering in Dynamically Competitive Environments: Organizational Capability as Knowledge Creation’, Organization Science, 7: 375–87. Hall, R. E. (2001). Digital Dealing: How e-Markets are Transforming the Economy. New York: Norton.

222

Effects of Technological Change Henderson, R. (1994). ‘The Evolution of Integrative Capability: Innovation in Cardiovascular Drug Discovery’, Industrial and Corporate Change 3(3): 607–30. —— and Clark, K. (1990). ‘Architectural Innovation: The Reconﬁguration of Existing Product Technologies and the Failure of Established Firms’, Administrative Science Quarterly, 35: 9–30. —— and Mitchell, W. (1997). ‘Introduction to the Special Issue’, Strategic Management Journal, 18(summer special issue): 119–42. Lamoreaux, N., Raff, D. M. G., and Temin, P. (2003). ‘Beyond Markets and Hierarchies: Toward a New Synthesis of American Business History’, American Historical Review, 108(2): 404–33. Langlois, R. N. (1992). ‘Transaction-Cost Economics in Real Time’, Industrial and Corporate Change, 1(1): 99–127. —— (1997). ‘Cognition and Capabilities: Opportunities Seized and Missed in the History of the Computer Industry’, in R. Garud, P. R. Nayyar, and Z. B. Shapira (eds.) Technological Innovation: Oversights and Foresights. Cambridge, UK: Cambridge University Press. —— (2003). ‘The Vanishing Hand: The Changing Dynamics of Industrial Capitalism’, Industrial and Corporate Change, 12(2): 351–85. —— and Robertson, P. L. (1992). ‘Networks and Innovation in a Modular System: Lessons from the Microcomputer and Stereo Components Industries’, Research Policy, 21: 297–313. —— and —— (1995). Firms, Markets and Economic Change: A Dynamic Theory of Business Institutions. London: Routledge. —— and —— (2003). ‘Commentary’, in R. Garud, A. Kumaraswamy, and R. N. Langlois (eds.) Managing in the Modular Age: Architectures, Networks, and Organizations. Oxford: Blackwell, pp. 101–13. Lave, J. and Wenger, E. (1991). Situated Learning: Legitimate Peripheral Participation. Cambridge: Cambridge University Press. Leonard-Burton, D. (1992). ‘Core Capabilities and Core Regidities’, Strategic Management Journal, 13: 111–25. Lippman, S. A. and Rumelt, R. P. (1982). ‘Uncertain Imitability: An Analysis of Interﬁrm Differences in Efﬁciency under Competition’, The Bell Journal of Economics, 13: 418–38. Mahoney, J. T. and Pandian, J. R. (1992). ‘The Resource-Based View Within the Conversation of Strategic Management’, Strategic Management Journal, 13: 363–80. Marshall, A. (1961). Principles of Economics, 9th edn. London: Macmillan. Nelson, R. R. and Winter, S. G. (1982). An Evolutionary Theory of Economic Change. Cambridge, MA: Harvard University Press. Pandian, J. R. and Robertson, P. L. (2003). ‘Introduction’, Managerial and Decision Economics, 24(4): 241–43. Patel, P. and Pavitt, K. (1998). ‘The Wide (and Increasing). Spread of Technological Competences in the World’s Largest Firms: A Challenge to Conventional Wisdom’, ¨ . So ¨ m and O ¨ lvell (eds.) The Dynamic Firm: The Role in A. D. Chandler Jr., P. Hagstro of Technology, Strategy, Organization, and Regions. Oxford: Oxford University Press, pp. 192–213. Penrose, E. T. (1959). The Theory of the Growth of the Firm. Oxford: Blackwell.

223

Effects of Technological Change Peteraf, M. A. (1993). ‘The Cornerstones of Competitive Advantage: A Resource-Based View within the Conversation of Strategic Management’, Strategic Management Journal, 14: 179–91. Prahalad, C. K. and Hamel, G. (1990). ‘The Core Competence of the Corporation’, Harvard Business Review, 68(3): 79–91. Prencipe, A. (1997). ‘Technological Compenencies and Product’s Evolutionary Dynamics: A Case Study from the Aero-Engine Industry’, Research Policy, 25: 1261–76. Porter, M. E. (1996). ‘What is Strategy?’ Harvard Business Review, 74 (November–December): 61–78. Reed, R. and DeFillippi, R. J. (1990). ‘Causal Ambiguity, Barriers to Imitation, and Sustainable Competitive Advantage’, Academy of Management Review, 15: 88–102. Robertson, P. L. (2003a). ‘The Future of Management: Does Business Have Anything to Tell Us?’ Australian Economic History Review, 43(1): 1–21. —— (2003b). ‘Strategic Interaction and Knowledge Distribution in the Location of Supply Chain Activities in the Global Automotive Industry’. Unpublished research paper, School of Management, Marketing and Employment Relations, University of Wollongong. —— and Langlois, R. N. (1994). ‘Institutions, Inertia, and Changing Industrial Leadership’, Industrial and Corporate Change, 3(2): 359–78. —— , Pol, E., and Carroll, P. (2003). ‘Receptive Capacity of Established Industries as a Limiting Factor in the Economy’s Rate of Innovation’, Industry and Innovation, 10(4): 457–74. Rumelt, R. P. (1974). Strategy, Structure, and Economic Performance. Boston, MA: Harvard Business School Press. Sanchez, R. and Mahoney, J. T. (1996). ‘Modularity, Flexibility, and Knowledge Management in Product and Organizational Design’, Strategic Management Journal, 17: 63–76. Schilling, M. A. and Steensma, H. K. (2001). ‘The Use of Modular Organizational Forms: An Industry-Level Analysis’, Academy of Management Journal, 44(6): 1149–68. Schumpeter, J. A. (1942). Capitalism, Socialism and Democracy. New York: Harper and Brothers. Shilling, M. A. (2000). ‘Towards a General Systems Theory and its Application to interﬁrm product Modularity’, Academy of Management Review, 25: 312–34. Silver, M. (1984). Enterprise and the Scope of the Firm. London: Martin Robertson. Smith, A. [1776] (1937). An Inquiry into the Nature and Causes of the Wealth of Nations, Edwin Cannan (ed.). New York: Modern Library. Smith, P. G. and Reinertsen, D. G. (1998). Developing Products in Half the Time: New Rules, New Tools. New York: Van Nostrand Reinhold. Stalk, G., Jr. and Haut, T. M. (1990). Competing Against Time: How Time-Based Competition Is Reshaping Global Markets. New York: Free Press. Teece, D. J. (1986). ‘Firm Boundaries, Technological Innovation, and Strategic Management’, in L.G. Thomas III (ed.) The Economics of Strategic Planning. Lexington, MA: Lexington Books, pp. 187–99. —— , Pisano, G., and Shuen, A. (1997). ‘Dynamic Capabilities and Strategic Management’, Strategic Management Journal, 18: 509–33. Thomke, S. and Kuemmerle, W. (2002). ‘Asset Accumulation, Interdependence and Technological Change: Evidence from Pharmaceutical Drug Discovery’, Strategic Management Journal, 23: 619–35.

224

Effects of Technological Change Tushman, M. L. and Anderson, P. (1986). ‘Technological Discontinuities and Organizational Environment’, Administrative Science Quarterly, 31(3): 439–65. —— and O’Really, C.A. III (1997). Winning Through Innovation: A Practical Guide to Leading Organizational Change and Renewal. Boston, MA: Harvard Business School Press. —— and Rosenkopf, L. (1992). ‘On the Organizational Determinants of Technological Change: Towards a Sociology of Technological Evolution’, in B. Staw and L. Cummings (eds.) Research in Organization Behavior, Vol. 14. Greenwich, CT: JAI Press, pp. 311–47. Verona, G. (1999). ‘A Resource-Based View of Product Development’, Academy of Management Review, 24(1): 132–42. —— , Prandelli, E., and Sawhney, M. (forthcoming). ‘Innovation and Virtual Environments: Towards Virtual Knowledge Brokers’, Organization Studies., 27(3). Yeoh, P. L. and Roth, K. (1999). ‘An Empirical Analysis of Sustained Advantage in the U.S. Pharmaceutical Industry: Impact of Firm Resources and Capabilities’, Strategic Management Journal, 20(7): 637–53. Williamson, O. E. (1985). The Economic Institutions of Capitalism, New York: Free Press.

225

This page intentionally left blank

9 Transitions, Transformations, and Reproduction: Dynamics in Socio-Technical Systems Frank W. Geels and Rene´ Kemp

9.1 Introduction This chapter is about stability and change at the systemic level and the role of public policy. It addresses some of the challenges set out in the introduction to this book (Chapter 1 by McKelvey and Holme´n), in particular the need for a theoretical framework that understands change as caused on the one hand by systemic processes and on the other by agents and their decisions. The perspective proposed in this chapter builds upon some of the ideas described in the book introduction. One idea is that external changes in the wider environment are important, as implied by the reference to Schumpeter (1947: 82) in Chapter 1: And this evolutionary character of the capitalist process is not merely due to the fact that economic life goes on in a social and natural environment which changes and by its change alters the data of economic action; this fact is important and these changes (wars, revolutions, and so on) often condition industrial change, but they are not its prime movers.

But this quote also highlights that developments in the exogenous environment do not mechanically cause system changes. Instead, system change is enacted by various types of actors, such as ﬁrms, policymakers, researchers, and consumers. This means that change is endogenous to the system, as well as situated in and inﬂuenced by a wider environment. We agree with the editors’ view that ‘we need progress in understanding how and why to connect the incentives and behaviour of the individual actors with system dynamics’. To make such progress, this chapter presents a multilevel perspective (MLP) on system change and reﬁnes it by distinguishing three kinds of change

227

Transitions, Transformations, Reproduction processes: reproduction, transformation, and transition. Our understanding of transformation and transition differs in some respects from the editors’ view expressed in Chapter 1. They deﬁne (economic) transformation as a ‘non-reversible process, referring to quantitative and qualitative changes in components and connections’ that encompasses ‘both processes driven by very large and discontinuous changes as well as processes driven by very small changes, which follow upon an existing trajectory’. While their term ‘transformation’ basically means change, we think it is useful to distinguish different types of change, varying in scope and underlying mechanisms. In this chapter ‘reproduction’ refers to incremental change along existing trajectories. ‘Transformation’ refers to a change in the direction of trajectories, related to a change in rules that guide innovative action (e.g. search heuristics, guiding princples, goals). And ‘transition’ refers to a change to a completely different trajectory, a discontinuous shift to a new system. So while the editors deﬁne transitions as ‘movement between two states of the system’, we see it as a shift to a new socio-technical system, with new technologies and new social groups. In Section 9.2 we will further articulate the mechanisms behind these three change processes. What do we mean by ‘systems’ in this chapter? There are many kinds of systems in the literature. For instance, the large technical systems (LTS) approach looks at big infrastructural systems, such as electricity systems, railroad networks, telephone systems (Hughes 1983, 1987; Mayntz and Hughes 1988). LTS researchers have developed the seamless web approach, highlighting that LTS achieve functionalities through the alignment of many heterogeneous elements, e.g. physical artefacts, organizations, natural resources, scientiﬁc elements, and legislative artefacts. Another approach is formed by sectoral systems of innovation (Breschi and Malerba 1997; Malerba 2002) which are deﬁned as ‘a system (group) of ﬁrms active in developing and making a sector’s products and in generating and utilizing a sector’s technologies’ (Breschi and Malerba 1997: 131). Likewise, the technological systems approach (Carlsson and Stankiewicz 1991; Carlsson 1997) looks at ‘networks of agents interacting in a speciﬁc technology area under a particular institutional infrastructure to generate, diffuse, and utilize technology’ (Carlsson and Stankiewicz 1991: 111). The last two approaches basically use a social network approach, widening the attention from ﬁrms to networks of actors involved in technological innovation, focusing attention on interaction, knowledge ﬂows, network dynamics, and co-evolution. But both approaches say more about the functioning of systems than about their change. In a recent review of sectoral systems of innovation Malerba (2002: 259) noted that one of the key questions that need to be explored in depth is: ‘How do new sectoral systems emerge, and what is the link with the previous sectoral system?’ This question is addressed in this chapter.

228

Transitions, Transformations, Reproduction Building on the seamless web approach, we understand systems at the sectoral level as socio-technical systems (Geels 2004). Societal functions such as transport, communication, and waste management are fulﬁlled by a cluster of elements, involving technology, science, regulation, user practices, markets, cultural meaning, infrastructure, production, and supply networks. This cluster of elements forms a socio-technical system. Building on the sectoral systems of innovation approach, we highlight the role of social groups. The elements of socio-technical systems are created, maintained, and reﬁned by multiple social groups, both supply-side actors (ﬁrms, research institutes, universities, policymakers) and demand-side actors (users, special-interest groups, media). The structure of the chapter is as follows. In Section 9.2 we describe a basic MLP on change in socio-technical systems. While this perspective was initially developed to understand transitions, it can also explain reproduction and transformation. Section 9.3 presents two historical case studies, one about transition and the other about transformation. The case studies are the hygienic transition from cesspools to integrated sewer systems (1870–1930) and the transformation in waste management (1960–2000) in the Netherlands. In Section 9.4 we will draw conclusions and address the role of policymakers in system change.

9.2 Multilevel perspective and types of change: reproduction, transformation, and transition MLP was originally developed to understand transitions and regime shifts (Schot et al. 1994; Rip and Kemp 1998; Kemp et al. 1998; Kemp et al. 2001; Geels 2002, 2005). We will ﬁrst introduce this perspective and then reﬁne it by describing the mechanisms of reproduction and transformation. The basic ontology behind the MLP stems from the sociology of technology, where three interrelated dimensions are important: (a) socio-technical systems, the tangible elements needed to fulﬁl societal functions; (b) social groups who maintain and reﬁne the elements of socio-technical systems; and (c) rules (understood as regimes) that guide and orient activities of social groups (see Figure 9.1). Actors in social groups do not act autonomously, but in the context of social structures and formal, normative, and cognitive rules. Rules form a coordinating context that guides and orients action. On the other hand, rules are reinforced and changed through action and enactment. Rules do not exist individually, but are linked together in semi-coherent sets of rules called regimes. Nelson and Winter (1982) coined the term ‘technological regimes’, which referred to the cognitive routines (e.g. search heuristics) that are shared in a community of engineers and guide engineers in their R&D activities.

229

Transitions, Transformations, Reproduction Socio-technical regime (rules) Socio-technical systems

Actors and social groups Figure 9.1 Three interrelated analytic dimensions (Geels 2004: 903)

Through their coordinative effects, technological regimes create relative stability at the sectoral level and lead to incremental innovation along technical trajectories. Rip and Kemp (1998: 340) have widened the deﬁnition of technological regimes to the sociological category of ‘rules’: A technological regime is the rule-set or grammar embedded in a complex of engineering practices, production process technologies, product characteristics, skills, and procedures, ways of handling relevant artefacts and persons, ways of deﬁning problems; all of them embedded in institutions and infrastructures.

This deﬁnition indicates that rules are not only embedded in the minds of engineers, but also in elements of the socio-technical system. The regime rules are also wider than search heuristics and may include problem agendas, guiding principles, rules of thumb, standards, government regulations, sense of identity, and role expectations. The regime concept has been further widened to ‘socio-technical regimes’ which include additional social groups besides engineers, for instance, scientists, users, policymakers, and special-interest groups (Geels 2004). These social groups interact and form networks with mutual dependencies, resulting in the alignment of activities. This intergroup coordination is represented with the concept of socio-technical regimes. The socio-technical regime forms the meso-level in the MLP of niche–regime– landscape. Through providing orientation and coordination to the activities of relevant social groups, the rules of socio-technical regimes account for the stability of existing socio-technical systems. Cognitive routines make engineers and designers look in particular directions and not in others (Dosi 1982; Nelson and Winter 1982). This can make them ‘blind’ to developments outside their focus. Core capabilities can thus turn into core rigidities (Leonard-Barton 1995). Established systems may be stabilized by legally binding contracts (Walker 2000). Systems are also stabilized because they are embedded in society. People adapt their lifestyles to them, favourable institutional arrangements and formal regulations are created, and accompanying infrastructures are set up. The alignment between these heterogeneous elements

230

Transitions, Transformations, Reproduction leads to technological momentum (Hughes 1994). Existing systems are further stabilized by social relationships. Actors and organizations are embedded in interdependent networks, which represent a kind of ‘organizational capital’, and create stability through mutual role expectations. Organizational commitments and vested interests of existing organizations also contribute to the stability of existing socio-technical systems (Hughes 1987: 76–7). Furthermore, industries may create special organizations to lobby on their behalf, e.g. industry associations or branch organizations (Unruh 2000). The material aspects of socio-technical systems also contribute to stability, because of sunk investments. Once artefacts and material networks are in place, they are not easily abandoned and acquire a logic of their own (Walker 2000). So, for many reasons, existing socio-technical systems are characterized by stability. This stability is not inertia, but dynamic stability, meaning that innovation still occurs but is of an incremental nature. Because of the stabilizing mechanisms, it is difﬁcult to create radical innovations within socio-technical systems. So how do radical innovations emerge? Scholars in sociology of technology and evolutionary economics have highlighted the importance of niches as the locus of radical innovations (Kemp 1994; Levinthal 1998; Schot 1998). Niches act as ‘incubation rooms’ for radical novelties, shielding them from mainstream market selection. New technologies need such protection, because ‘most inventions are relatively crude and inefﬁcient at the date when they are ﬁrst recognized as constituting a new invention. They are, of necessity, badly adapted to many of the ultimate uses to which they will eventually be put’ (Rosenberg 1976: 195). Niches may have the form of small market niches, where selection criteria are different from the existing regime. Or they may have the form of technological niches, where resources are provided by public subsidies or private strategic investments. The literature on strategic niche management distinguishes three crucial nicheinternal processes (Schot et al. 1994; Kemp et al. 1998; Kemp et al. 2001; Hoogma et al. 2002). First, technical learning processes (R&D) are crucial to improve performance. But learning also occurs with regard to user preferences, regulation, symbolic meaning, infrastructure, and production systems. Through the alignment of these learning processes a new seamless web can be build up. The second process is the building of social networks and ‘sociotechnical constituencies’ (Molina 1995) that support the new innovation and invest in its further development. The third process is the articulation of visions and expectations to provide orientation towards the future and give direction to learning processes. The macro level is formed by the socio-technical landscape, which refers to aspects of the exogenous environment. The content of the socio-technical landscape is heterogeneous and may include aspects such as economic growth, broad political coalitions, cultural and normative values, environmental problems, resource scarcities. But the socio-technical landscape also includes

231

Transitions, Transformations, Reproduction the large-scale material context of society, e.g. the material and spatial arrangements of cities, factories, highways, and electricity infrastructures (Rip and Kemp 1998). Landscapes are beyond the direct inﬂuence of actors, and cannot be changed at will. Changes at the landscape level usually take place slowly, in the order of decades. The importance of this macro level was also highlighted in Schumpeter’s quote in Section 9.1. The relationship between the three conceptual levels can be understood as a nested hierarchy. The key point of the MLP is that transitions come about through the interplay between processes at different levels in different phases. In the ﬁrst phase, radical innovations emerge in niches, often outside or on the fringe of the existing regime. There are no stable rules (e.g. dominant design), and actors improvise and engage in experiments to work out the best design and ﬁnd out what users want. The networks that carry and support the innovation are small and precarious. The innovations do not (yet) form a threat to the existing regime. In the second phase, the new innovation is used in small market niches, which provide resources for technical development and specialization. The new technology develops a technical trajectory of its own and rules begin to stabilize (e.g. a dominant design). But the innovation still forms no major threat to the regime, because it is used in specialized market niches. New technologies may remain stuck in these niches for a long time (decades), when they face a mismatch with the existing regime and landscape. The third phase is characterized by wider breakthrough of the new technology and competition with the established regime. On the one hand, this depends on internal drivers in the niche. Improvements in the price/performance ratio are an important driver, as well as support from powerful social groups. On the other hand, external circumstances at regime and landscape levels are crucial, creating windows of opportunity for novelties in niches. Social, cultural, or economic changes at the landscape level may put pressure on the regime. Or the existing regime may be plagued by increasing internal problems that cannot be solved by incremental improvements. Such tensions create opportunities for the breakthrough of innovations. MLP emphasizes that both internal niche dynamics and external developments at regime and landscape level are important for wider breakthrough and diffusion (see Figure 9.2). As the new technology enters mainstream markets it enters a competitive relationship with the established socio-technical regime. In the fourth phase this leads to replacement, something that is accompanied by wider socio-technical changes. Replacement often happens in a gradual fashion, because the creation of a new socio-technical regime takes time. The new system may eventually inﬂuence wider landscape developments. In the MLP there is no simple ‘cause’ or driver in transitions. There are downward inﬂuences from the landscape level, and upward inﬂuences from new technologies that capture a niche and ﬁnd support by product

232

Transitions, Transformations, Reproduction Landscape developments

Markets, user preferences Sociotechnical regime

Landscape developments put pressure on regime, which opens up, creating windows of opportunity for novelties.

New socio-technical regime influences landscape.

Science Policy

Culture

Technology Socio-technical regime is ‘dynamically stable’. On different dimensions there are ongoing processes.

New technology breaks through, taking advantage of ‘windows of opportunity’. Adjustments occur in socio-technical regime.

Elements are gradually linked together, and stabilize around a dominant design. Internal momentum increases.

Technological niches

Learning processes with novelties on multiple dimension. Different elements are gradually linked together.

Time

Figure 9.2 A dynamic multilevel perspective on system innovations (Geels 2002: 1263)

constituencies, who may be able to build political cloud. For a transition to occur dynamics at different levels should come together and reinforce each other. MLP in Figure 9.2 works from the ‘outside in’, describing and mapping the entire long-term process. With its emphasis on the alignment of processes at different levels it may seem mechanical and linear. But processes in sociotechnical regimes and systems are the outcome of perceptions and (inter)actions of actors and social groups. These social groups have their own vested interests, problem perceptions, strategies, and resources (money, knowledge, and contacts). This implies that dynamics in socio-technical systems involve interactions between many social groups. The ‘outside in’ approach is thus complemented with an ‘inside out’ approach. Because the linkages between processes at different levels are made by actors in their cognitions and activities, the dynamics are not mechanical, but socially constructed. As social groups try to navigate a transition, they engage in commercial transactions, political negotiations, power struggles, coalition building, controversies, and debates. The dynamics of transitions are not linear, because perceptions and

233

Transitions, Transformations, Reproduction strategies of actors change over time. Transitions are complex processes that cannot be overseen or steered from one viewpoint. They are emergent outcomes of interactions between social groups with myopic views and differing interests. Transitions are contested, because groups with vested interests have something to lose, while others hope to gain certain advantages. So the overall dynamic is a crooked and non-linear process. The conceptual perspective enables a systematic distinction between three kinds of change processes: reproduction, transformation, and transition. Table 9.1 summarizes the differences between these change processes in terms of underlying mechanisms.

9.2.1

Reproduction

In this change process there are only dynamics at the regime level, not at the landscape and niche level. The existing socio-technical system and regime form a stable context for (inter)action of social groups. Existing rules are reproduced by the incumbent actors, and elements in the socio-technical system are reﬁned. The orientation of dominant actors, key technology, and knowledge base do not change fundamentally. There is incremental and cumulative change along trajectories. This is the normal situation at the regime level. As indicated earlier, there are many reasons why existing regimes and systems are stable (e.g. sunk investments, role expectations in networks, standards, contracts, cognitive routines). This is dynamic stability, meaning that incremental innovations still occur. Incremental innovations in stable regimes are important, because, over time, they can accumulate and result in major performance improvements. A large portion of the total growth in productivity takes the form of a slow and often invisible accretion of individually small improvements in innovations. ( . . . .) Such modiﬁcations are achieved by unspectacular design and engineering activities, but

Table 9.1. Different mechanisms in change processes Reproduction Levels involved Regime dynamics

Role of actors

234

Transformation

Transition

. Pressure from landscape . Pressure from landscape . Adaptation and reorientation . Increasing problems in regime, in regime and attempts at reorientation . New innovation in niches that eventually break through

Incumbent regime . Pressure from outsiders actors . Incumbent regime actors respond through reorienting innovative trajectories

. Pressure from outsiders . Incumbent actors fail to solve regime problems . Outsiders develop new innovations

Transitions, Transformations, Reproduction they constitute the substance of much productivity improvement and increased consumer well-being in industrial economies. (Rosenberg 1982: 62)

The reproduction process is represented in Figure 9.2 if we leave out the niche and landscape dynamics, and extrapolate the stable trajectories represented with the straight ongoing lines in the regime. In this pattern, there can still be interesting actor dynamics such strategic games in markets, power struggles, and strategic coalitions. But these take place in a stable system and regime context.

9.2.2

Transformation

In this change process there are interacting dynamics at the regime and landscape level, but litte inﬂuence from niches. The basic mechanism is that changes at the landscape level create pressure on the regime, leading to reorientation of the direction of innovative activities. This happens through a change in the regime rules that coordinate actions of regime actors, e.g. changes in technical problem agendas, visions, goals and guiding principles, relative costs and incentive structures, regulations and perceptions of opportunities. The adjustment and reorientation to external landscape pressure does not happen in a mechanical fashion, but through negotiations, power struggles, and shifting coalitions of actors. Because incumbent regime actors initially tend to downplay the need for transformation, a change in the social network is often important to start a transformation process. Outsiders, public and regulatory pressure, or the entry of new actors may help challenge previously held assumptions and place new issues on the problem agenda (Van de Poel 2000). Such outsiders may express concerns over negative externalities of the existing system, and demand responses from regime actors (Van de Poel 2003). But in the transformation process these outsiders do not develop competing technologies to replace the existing system. So, the survival of incumbent regime actors is not threatened, and they are the ones to enact the redirection of the development trajectory of the existing system. In the transformation process a new system may grow out of the old one, through cumulative adjustments in a new direction.

9.2.3

Transition

A transition refers to a shift from one socio-technical system to another. It is not about the reorientation of an existing trajectory, but a shift to a new trajectory. An example is the transition from a transport system based on horse-drawn carriages to a transport system based on automobiles. This transition involved changes in the socio-technical system (e.g. technologies, knowledge base, infrastructure, regulations, user practices, cultural prefer-

235

Transitions, Transformations, Reproduction ences), social groups, and regime rules. In a transition process there are interactions between dynamics at landscape, regime, and niche level. Landscape developments create pressure on the regime, leading to major problems. Regime actors react with adjustments in the system (as in the transformation process), but they are not able to solve the problems. This creates a window of opportunity for new innovations, developed in niches and carried by a new network of social groups. If a new innovation breaks through and replaces the existing system, this will be accompanied by ‘creative destruction’ and the downfall of (some) incumbent actors. Once a transition has taken place, a new period of dynamic stability and reproduction sets in.

9.3

Case studies

To illustrate the usefulness of the MLP and the three change processes, this section presents two case studies. Because we assume that reproduction is the normal dynamic in existing regimes, we have chosen a transition case and a transformation case. Both the cases are about the Netherlands, because of the practical reason of easy data collection. The ﬁrst case is the hygienic transition from cesspools to integrated sewer systems (1870–1930), and the second is the transformation of Dutch waste management (1960–2000). Both the cases are often seen as goal-oriented: there were pressing problems (bad hygiene and disease, pollution), causing a search for solutions, followed by implementation. The case studies are chosen to challenge this rationalistic and functionalistic view. They will show that dynamics were more complex and that there was a lot of contestation and struggle going on. The case studies are also chosen, because there are important roles for policymakers and government, a topic we will come back to in Section 9.4. The case studies will show that different kinds of multilevel interactions result in different change processes. The ﬁrst case study will show the crucial importance of exernal landscape changes to create pressure on the regime and opportunities for niches. The second case study is more about interactions between regime and landscape and will show in more detail the role of rule changes in the transformation process.

9.3.1 9.3.1.1

The hygienic transition from cesspools to integrated sewer systems in the Netherlands (1870–1930) PROBLEM ARTICULATION (1840–70)

For most of the nineteenth century, urine, faeces, and other domestic wastes were the largest waste streams in Dutch cities (Van Zon 1986). People

236

Transitions, Transformations, Reproduction commonly relieved their bowels in public space, dumping urine and excrements on streets and in canals. Middle and upper classes had personal privies in-house, where excrements were collected in cesspools that were emptied only a few times a year. Local public authorities issued regulations and prohibitions against waste dumping on public roads and in canals, but with little effect, because of a lack of policing (Houwaart 1993b: 82). The presence of organic waste on streets and in canals created stench problems, especially in summers and in dense urban areas (Corbin 1994). Many people were concerned, because it was widely thought that bad smells (so-called miasmas) caused diseases. In the 1840s and 1850s new groups of Dutch doctors emerged, the so-called hygienists, who began to investigate the relationship between bad hygiene and infectious disease using quantitative data and medical statistics. Building on insights from the ‘sanitarian movement’ in France and England (Houwaart 1991), they collected data about the distribution of diseases and death rates over different cities and neighbourhoods. They then tried to correlate these differences to hygienic variables, such as waste heaps, insufﬁcient street cleaning, canals with still water and decaying organic material (Houwaart 1993a: 38). They presented their ﬁndings in tables and maps, showing clear relationships between hygiene and disease. The hygienists were thus crucial in the articulation of a problem that was subsequently discussed in city councils, newspapers, and public associations. But there was uncertainty about underlying causal mechanisms that linked hygiene and disease. Debates about the origin of diseases continued until late in the nineteenth century. Hygienist doctors increasingly saw drinking water polluted with human excrements as important cause of infectious diseases (in particular cholera). But many doctors in the established Dutch medical community hung on to the traditional miasma-based theory, and developed additional hypotheses, for instance how quality and level of groundwater affected soil conditions that, in turn, affected the blossoming of cholera germs that were then spread through miasmas (Houwaart 1991: 132). Although doubts remained about underlying causal mechanisms, it was clear by the 1850s that there were relationships between bad hygiene and infectious disease. As cities grew in size, waste disposal problems increased. Heaps of waste accumulated in canals. This blocked water circulation and hindered the supply of fresh water and the removal of waste (Van Zon 1986: 34). It also caused stench and fears about miasmas. Between 1850 and 1880 the faeces problem was hotly debated in city councils and newspapers. In other countries visions about encompassing and radical solutions for urban hygiene were developed in the 1840s and 1850s by engineers and hygienists, ¨ nchen, Villerme´ and Parente.g. Virchow in Berlin, Von Pettenkofer in Mu Duchaˆtelet in Paris, Liernur in Amsterdam, Shattuck in Boston, and Chadwick in London (De Swaan 1989: 141). Foreign cities began building integrated

237

Transitions, Transformations, Reproduction sewer systems to deal with waste problems, ﬁrst in the German city Hamburg in 1843. London commissioned Joseph Bazalgette in 1852 to plan and design a sewer system. Actual work began in 1859, and the ‘main drainage’ was completed in 1865. In Paris design of an underground sewer system began in 1863, as part of Haussman’s reforms. By 1871 an integrated sewer system of 560 km length had been created (Reid 1991). In the USA the ﬁrst sewer systems were built in Brooklyn in 1855, Chicago in 1856, Jersey City in 1859 (Tarr 1988:166). And in Germany sewer systems were built in Berlin (1873), ¨ nchen (1880) (Van den Akker 1952: Breslau (1875), Karlsruhe (1877) and Mu 220). Dutch cities did not follow the examples from abroad. There was limited interest in alternative niches for waste management, because of conditions at landscape and regime level. The Dutch political cultural was highly liberal with a very small state that hardly interfered in society. The Municipal Law (1851) gave local authorities much autonomy, and made them responsible for public works, public health, and hygiene. To keep taxes low for their voters, city governments had low involvement in public life. Voters included a small part of the population, because the census was limited. In 1850 around 10.7 per cent of Dutch men above 23 years could vote for Parliament, and 18 per cent for local city councils (Verdoorn 1965: 340). So, in terms of political accountability, there was little incentive for public authorities to care much about the health of poor people without vote. Health was seen as an individual responsibility, not as a responsibly of public authorities (Houwaart 1993b: 82). In the waste disposal regime, city governments remained inactive about accumulating waste problems, despite public concerns about hygiene and disease. They mainly implemented incremental changes within the existing regime, such as measures to improve water circulation in canals, so that waste would be ﬂushed away (Van Zon 1986: 37). Canals were dredged more frequently to maintain sufﬁcient depth, and sometimes steam engines were implemented to pump in more fresh water. Another option was to ﬁll up canals that stank worst and had little water circulation (Buiter 2005). At the niche level, there was a small reform network, consisting of hygienist doctors and some city engineers, which lobbied for sewer systems. Stimulated by foreign examples, many Dutch cities set up local commissions in the 1850s and 1860s to investigate the option of sewer systems. These commissions produced an endless number of reports, which were discussed in city councils. But concrete sewer designs were rejected by city governments, because of technical and ﬁnancial uncertainties related to characteristics of the Dutch landscape. One characteristic was that much of the soil was wet and sponge-like. Because the soil was unstable, it could subside, leading to tensions and fractures in ﬁxed sewer pipes (Daru and Van Zon 1987: 58).

238

Transitions, Transformations, Reproduction So there was uncertainty if the Dutch soil was suitable for sewers. Another problem was the ﬂatness of the country. Ideally sewers functioned on a gravity-ﬂow basis, where natural slopes gave water sufﬁcient speed to prevent sedimentation and clogging. Because many places in the Netherlands had insufﬁcient slope, pumps should be added to the system to stimulate water ﬂows. This would increase costs, making them less attractive to city governments. So decisions were postponed, and cities stuck to incremental solutions.

9.3.1.2

EXPERIMENTATION AND COEXISTENCE OF MULTIPLE NICHES (1870–90)

In the 1870s and 1880s waste problems grew worse, because of wider landscape developments. As industrialization gathered speed, poor farmers moved to the cities, leading to rapid urbanization (Kossman 1976). Many people lived in crowded, damp, and unhygienic conditions. While the middle classes continued to use the cesspool for waste disposal, the expanding working classes had no sanitary facilities (Bolderman 1992: 44). Hence waste dumping on streets and in canals continued, resulting in exacerbating stench problems as urban populations expanded. Human excrements and water pollution rose on the problem agenda, after the cholera epidemic during 1866–7. In 1868 a National Drinking Water Commission concluded that drinking water in the Netherlands was highly polluted with organic compounds from human excrements. The Commission recommended that public authorities should play a more active role with regard to drinking water and waste disposal (Van Zon 1986). The Commission’s report was widely discussed, but did not lead to action. City governments continued to rely on traditional solutions, such as the improvement of water circulation through dredging and using steam engines to pump fresh water in canals. Nevertheless, city governments facilitated some degree of experimentation with new technologies in niches. This minor, but important change in attitude was related to several macro developments. The economic situation in the Netherlands began to improve somewhat in the 1870s and 1880s, so there was more tax money available. Another landscape change was the emergence of the ‘social issue’ on the political agenda. Labour unions became more political, particularly the socialist labour union, the SDB, created in 1880. The SDB drew attention to social issues such as poverty, socio-economic inequality, and class struggle. Public health was increasingly seen as part of the social issue, and city governments wanted to show that they took the issue seriously. Another reason was expansion of the social network that wanted hygienic reform. Hygienist doctors were increasingly joined by engineers, a new group that enjoyed much societal esteem. Public opinion and discussions in newspapers also increased pressure on city governments

239

Transitions, Transformations, Reproduction to be more active about waste problems. The coalescence of these developments created more willingness to experiment with alternative waste-removal options. One alternative solution was the barrel-collection system: people deposited their excrements in barrels, which were collected several times a week (Van Zon 1986: 79). Full barrels were transported to a central collection place, where their contents were processed into compost, and then sold to farmers as fertilizer. The barrels were cleaned and then reused. Many cities became interested in this niche, because it looked like a win-win solution, dealing both with the faeces problem and earning money. Most engineers opposed the barrel-collection system, because of its imprecision and leakages during the process. But hygienists and agricultural experts praised the system, because excrements fulﬁlled a useful function as fertilizer. Many cities implemented the barrel system in the 1870s to some extent (e.g. Groningen, Leeuwarden, Rotterdam, and Amsterdam). Barrel collection was a viable niche. Amsterdam and Rotterdam used it until the 1910s, and small cities like Alkmaar used it until after the Second World War (Vis 1996: 110). But the niche was not always viable, because the price of human excrements varied between locations, depending on soil conditions, needs of farmers, and availability of other types of fertilizer (e.g. guano). The second niche was the pneumatic Liernur system. This system consisted of a toilet, a funnel, and underground pipes that connected the house to main pipes that ended in a collection reservoir. A steam pump was used to create a vacuum and collect faeces in the reservoir. Excrements were collected daily, processed and sold as fertilizer. A beneﬁt of the system was its cleanliness. Faeces were collected without spilling, and it did not involve labourers carrying dirty barrels. Hygienists favoured this system, because it was clean and produced faeces for agriculture. But the system was relatively complex and expensive. City governments feared that high construction costs could not be recovered from excrements sales. Experiments were done on a small scale in Leiden (1871), Amsterdam (1872), and Dordrecht (1873). The experiments were moderately successful. In 1879 Amsterdam’s city council decided to expand the Liernur system at the expense of implementing an integrated sewer system. In some neighbourhoods the system was used until 1916 (Buiter 2005). The third niche was formed by sewer systems. Individual sewer pipes had already been created in the 1850s and 1860s, to facilitate rainwater run-off when canals had been ﬁlled up. Dutch municipal engineers made plans to link these individual pipes in an integrated sewer system. Many commissions were installed to investigate integrated sewer systems, reports were written, and discussions were held. But sewer systems were not implemented in practice. Table 9.2 gives an impression of the plans that were developed and rejected. The unwillingness of city governments was due to the liberal political climate, and high costs for underground infrastructures. For city governments,

240

Transitions, Transformations, Reproduction Table 9.2. Rejected plans for sewer systems Year

City

Engineer

1858 1863 1870 1870 1872 1872 1876, 1878 1897 1902

Rotterdam Arnhem Amsterdam Tilburg The Hague Arnhem The Hague Amsterdam Amsterdam

Scholten Van Gendt Van Niftrik Havelaar Van der Waayen Pieterszen Henket Reinders Lambrechtsen van Ritthem Van Hasselt

Source: Daru and Van Zon 1987: 59

ﬁnancial considerations were more important than solving hygienic problems (Daru and Van Zon 1987: 55). A third reason was that the barrel system and the Liernur system emerged, just when many cities began taking sewers seriously. The coexistence of multiple niches created much uncertainty, resulting in an ‘information chaos’ (Van den Noort 1990: 48). Competing claims were made about the advantages and disadvantages of different niches. Most engineers opposed the barrel-collection system, because of its imprecision and leakages during the process. Hygienists, on the other hand, praised the system, because excrements fulﬁlled a useful function as fertilizer. But hygienists liked the pneumatic system even better, because it was clean in operation and made good use of faeces. Agricultural experts were also in favor of separate faeces collection (either barrel collection or pneumatic), because of its economic use in agriculture. Engineers were mainly in favour of sewer systems, which removed faeces efﬁciently. But many hygienists and agricultural experts opposed sewer systems, because of the loss of fertilizer value from excrements. So there were many conﬂicting claims from different kinds of experts. Uncertainty was further increased, because many local factors inﬂuenced the technical and economic feasibility of the different niches, e.g. geo-hydrological conditions, soil conditions, city size, population density, vicinity, and needs of agriculture (Daru 1985). Even when cities implemented the same waste-removal option, their assessments and evaluations could differ substantially. Variety also existed within cities. In large cities such as Rotterdam and Amsterdam multiple waste-collection options coexisted. Given the uncertainties, most cities opted for the cheapest option, the barrel system. So the sewer niche, which was seriously discussed in the 1850s and 1860s, was placed on the back burner in the 1870s and 1880s, as the Liernur and barrel-collection niche received much attention. 9.3.1.3

TRANSITION TO SEWER SYSTEMS (1890–1930)

In the 1890s and early twentieth century several important macro changes took place that affected the choice between waste-removal options. One

241

Transitions, Transformations, Reproduction important development was the change in the role perception of public authorities from liberal refrainment to more active intervention. It came to be seen as legitimate that city governments would intervene to improve urban life for all residents. This change was related to cultural and political concern about the condition of working classes. More vocal labour unions and the socialist movement forced the social issue on the political agenda. Awareness of the social issue was also stimulated by several parliamentary inquiries (e.g. in 1887 into working conditions in factories). And it was stimulated by further democratization. In 1887 the census was lowered and the right to vote widened. In 1896 the attributive right to vote was installed. And in 1917 the general right to vote was extended to all men, and in 1919 to all women. The widening of the right to vote formed an incentive for public authorities to pay more attention to living conditions of all people. More active public involvement was also enabled by strong economic developments between 1890 and 1914, as industrialization ﬁnally took off (Kossman 1976: 314). Economic growth led to higher tax incomes, providing ﬁnancial means for public interventions. Population growth and urbanization continued to exacerbate problems and stench in the waste-removal regime. Together with the increased sensitivity about the social issue and a new role perception of public authorities, this resulted in a widespread feeling that something had to be done to improve hygienic conditions. Better ﬁnancial conditions enabled local authorities to implement more expensive solutions such as sewer systems. But the turn towards sewers was also strongly inﬂuenced by developments in other regimes, such as agriculture, water supply, and housing. These regime developments strongly inﬂuenced the relative competitiveness of the three alternative waste-removal options. One such development was the diffusion of piped water. By 1900, around 40 per cent of the Dutch population was connected to piped water (De Swaan 1989). The diffusion of piped water systems was accompanied by the breakthrough of water closets. As a consequence waste streams had higher water content, reducing the fertilizer value and economic feasibility of the Liernur system. And water closets stimulated sewer systems, because ﬂushing became easier. An inﬂuential development in the agricultural regime was the emergence of artiﬁcial fertilizer factories in the 1890s that produced cheap fertilizer based on phosphates or sulphates. As a result, farmer’s demand for human excrements decreased and faeces prices dropped, having a negative effect on the Liernur system and barrel system. In the housing regime, continued urbanization was a driver for cities to build new neighbourhoods. These new neighbourhoods provided good locations for the construction of underground infrastructures, since pipes could be laid cheaply before houses were build. As a result of these niche, regime, and landscape interactions, the Liernur system disappeared, the barrel-collection system was gradually phased out,

242

Transitions, Transformations, Reproduction and sewer systems became more popular. Sewer systems were ﬁrst implemented in large cities, because construction was expensive and required a certain threshold to be worthwhile, e.g. the Hague in 1893 and Amsterdam in 1913. The transition to sewer systems in small cities was much slower. Financial means were often lacking, but also hygienic problems were less pressing, since pollution was more diluted. In 1938, 47 per cent of all municipalities had sewers (Van den Akker 1952: 242). Especially in small cities and on the countryside, cesspools and barrel-collection systems remained in use until after the Second World War. We conclude that the case study has a good match with the multilevel perspective, in the sense of important interactions between levels. From the mid-nineteenth century onwards, the waste-disposal regime, based on cesspools and waste dumping in canals, encountered increasing problems with the expanding quantities of waste and excrements. The problems were made worse by landscape developments such as urbanization, and new medical insights that correlated disease with bad hygiene. As the regime ran into problems, three niches emerged. But these niches could not gather enough momentum as long as the principal actor (city governments) stuck with incremental solutions in the existing regime. Only the barrel-collection niche was implemented to some extent in the 1870s and 1880s, because it was cheap and promised to make money. Sewer systems eventually broke through in the 1890s as changes in other regimes (piped water supply, agriculture, and housing) and landscape changes on economic, social, and political dimensions created more favourable conditions. The case study showed that a transition process comes about through the interaction of dynamics at all three levels. This is a difference with the transformation process, which includes only regime and landscape interactions. The case study also showed that outside actors are important to get problems on the agenda and support niche innovations (hygienist doctors, engineers, and local health councils). The transition also saw major changes in the network of social groups. There was some ‘creative destruction’ with the disappearance of existing groups (actors in the collection, processing, and distribution of faeces). New groups appeared (municipal sewage organizations), and some existing groups took on new roles (more active public authorities). These major changes in social networks form another difference with the transformation process, where social changes are smaller (see next case study).

9.3.2

The transformation of Dutch waste management (1960–2000)1

In the last 40 years, the Netherlands witnessed a transformation in waste management: from uncontrolled landﬁlling (waste dumping) towards a differentiated waste-handling system of recycling, incineration with energy

243

Transitions, Transformations, Reproduction reuse, and controlled landﬁlling. It is unclear whether this transformation has ended. Changes at the European level (the disappearance of waste borders) may lead to further change (even backwards), which is why we talk about transformation and not about a transition. In some ways the transformation meant a return to the old practice of recycling. Before 150 years, recycling was a common practice in the Netherlands: glass, metals, old fabrics, and certain types of organic waste were being collected by individual traders (Loorbach 2003). At the end of the nineteenth century, such activities became less economical and more and more private entrepreneurs stopped collecting waste. There no longer is the ‘schillenboer’ with his horse collecting shells of vegetables. Waste collection became a public task handled by municipalities. Most of the waste (including rising quantities of chemical waste) was being landﬁlled; a small part was reused or incinerated in newly built incinerators. In 1912, the ﬁrst incineration plant was opened in Rotterdam, while Amsterdam and Leiden followed in 1918 and 1914, respectively. In Den Haag, in 1918 a small incineration plant was opened that even generated electricity on a small scale. The incinerators were built in urbanized areas lacking landﬁll sites in the vicinity. Waste was also used for ﬁlling swamps and ditches (‘slotenrijden’), to generate new land for settlements. No track was kept of the types of waste being disposed. The Netherlands basically had an uncontrolled waste management subsystem, in which waste was disposed off with few environmental considerations. The principal issue was to get rid of waste. In the 1970s, waste and unsustainable waste management practices were receiving increasing attention: concerns were raised about how waste was being managed; there were growing problems with creating new landﬁll sites (because of local resistance); and the 1972 Report to the club of Rome followed by the oil crisis in 1973 put attention to scarcity of materials.2 Waste disposal was increasingly seen as a problem. Special legislation for waste was being made and responsibilities were given to provinces. With the introduction of the Hazardous Waste Act (1976) and the Waste Act (1977), the Dutch provinces received the planning and coordinating task, while the implementation to a large degree remained with (cooperating) local authorities (collection and disposal). The reason for this change in responsibilities was to put an end to the (uncontrolled) dumping on landﬁlls and to beneﬁt from economies of scale for incineration. Provincial borders were closed for waste transports and the operators were given the exclusive right and obligation to collect waste in a certain region. Operators were guaranteed necessary supply (processing certainty) and transporters had a guaranteed demand. The activities were organized as municipal service, controlled by local politicians formally in control, responsible for funding (from www.aoo.nl).

244

Transitions, Transformations, Reproduction Central to policy thinking was the ‘waste hierarchy’ proposed in the parliamentary motion of Ad Lansink in 1979. The waste management hierarchy went from prevention, through reuse (of products), recycling (of materials), and incineration (with energy production) to landﬁlling as the last option. The motion became law in 1986 and was an important cognitive institution (Parto et al. 2003). From the late 1970s on, waste was increasingly seen as ‘a waste of resources’ in polity. Business also started investigating ways to reduce waste as part of its environmental policy. To reduce the volumes of waste for disposal the Dutch government opted for a differentiated waste-stream approach in which certain types of waste (notably paper and glass) were singled out for recycling. The initial reluctance to adopt the separate waste system came from the municipal waste-collecting services that had to change their practices. Other actors such as non-governmental organizations (NGOs) and private businesses performed new activities like the collection of paper and glass. The systematic collection of the bulk of recyclable waste and organic materials would only become institutionalized in the 1990s (Parto et al. 2003: 7). Despite these intentions for upgrading waste practices, many activities in the area of waste management still suffered from their small-scale nature and from inadequate environmental protection. For example, up until the 1990s, soil protection measures were absent in virtually all landﬁlls and ﬂue gas scrubbing in waste incineration facilities was inadequate (from www.aoo.nl). There was considerable political and community resistance to the construction of new landﬁlls and incineration plants, with the resistance reached a high peak in the 1980s, following the discovery of leaking landﬁlls (Vogelmeerpolder) and contaminated land (Lekkerkerk and Griftpark). Waste scandals were a frequent news item in the 1980s. The two most important ones were: Lekkerkerk in which it was discovered in 1980 that new houses had been built on soil containing chemical waste that had been landﬁlled, and Lickebaert, where in 1989 dioxins (coming from incinerators of AVR and AKZO) were discovered in the milk of grazing cows. Five waste incinerators were closed because of dioxin emissions and at least one plan for a new landﬁll (Does in Leiden) was abandoned because of opposition. Whilst capacity was decreasing, waste volumes kept growing, leading to capacity problems. In 1991, as a result of lack of regular waste management capacity, it even became necessary to store waste in push barges. At the end of the 1980s, the Dutch waste management system was in a state ¨ rdinatie Commissie of crisis. The system was reviewed by the Landelijke Coo Afvalbeleid (Commissie Welschen) in 1989 who concluded that ‘the current organization is fragmented, dispersed, and small scale’. It argued for the creation of a nationally oriented organization for disposal, to manage overall waste volumes and keep disposal costs under control. Not only for incineration, but also for organizing waste management from cradle to grave (chain

245

Transitions, Transformations, Reproduction management), four waste regions (encompasses several provinces) were envisaged, each with 3–4 million inhabitants (from www.aoo.nl). This advice led to the appointment of the Waste Management Council (AOO), through the cooperative agreement for waste disposal VROM/IPO/VNG (1990). The AOO would play an important role in the modernization of the waste system. From the beginning there were problems with the four waste-regions system. Municipalities wanted to sign contracts with waste companies in other regions, and because of capacity problems waste had to go to other regions for incineration. In 1996, at the advice of the Commission Epema, it was decided to centralize the responsibility for waste control to the national level. The legal basis for the centralization is the last amendment of the Environmental Management Act that came into force in May 2002. Especially, efﬁciency considerations lied behind this decision. The centralization was very much wanted by new private collecting and transport companies who wanted to operate nationally. In the 1977–2000 period, the number of landﬁll sites fell from 450 in 1977 to 34 in 2000 (Figure 9.3) thanks to the differentiated waste-handling approach and targeted policies (e.g. the packaging covenants), the ban of thirty-two waste stream for landﬁlling, and steadily increasing costs for landﬁlling, creating an incentive to move up on the waste ladder. The amount of waste being landﬁlled fell 14 million tonnes in 1990 to 5 million tonnes in 2002, a reduction of 9 million tonnes. Today all landﬁlls have advanced systems of soil protection and systems of methane extraction. In the same period, the capacity of incineration increased gradually, from 2.2 million tonnes in 1980 to 4.9 million tonnes in 2000. Between 1995 and 2000, incineration

600

2,

26

5

0 24 2,

400

2,

300

48

0

200

2,

Number of sites

10

0

500

0 1975

1980

1985

1990 Area landfill in ha

Figure 9.3 Reduction of landﬁll (Source: AOO)

246

0 18 1,

1,

47

0

100

1995

2000

Transitions, Transformations, Reproduction capacity increased with 2 million tonnes. Recycling increased between 1985 and 2000 from 23.5 million tonnes to 45.3 million tonnes, almost a doubling (Figure 9.4). The transformation to a system of recycling and increased incineration with controlled landﬁlling as a last resort option is often viewed as the result of policy. Such a view, although not wrong in itself, overlooks that policy itself was the result of various changes: the growing volumes of waste, the waste scandals in the 1980s and early 1990s, and changes in beliefs (such as the belief that waste is ‘a waste of resources’ and the belief that landﬁlling should be done in a hygienic manner and only be used as a last resort option) in a period in which environment was very much on the mind of people. The waste scandals helped close down old incinerators and build better ones. Various waste acts laid the basis for policy and the AOO, created in 1990, brought together the three layers of government (local, provincial, and central) to work together in a policy network with no clear legal status under an independent chairman. The AOO played an important role in the transformation process. Negotiations between different layers of government and with private waste companies took place within the AOO with the actors agreeing on the general direction of creating a modern and efﬁcient system of waste management with less waste being landﬁlled. The environmental movement, while being ofﬁcially opposed to incineration, were not creating too much trouble because they understood that high costs of advanced systems of incineration necessitated a high tax for landﬁlling for burnable waste,3 which encouraged waste prevention and recycling. The waste companies were happy with the greater scale at which they could operate. The reorganization of the sector with big companies from North America such as Waste Management Inc. and BFI taking over small companies was

70 60

Landfill Incineration

Mton

Discharge Recycling

50 40 30 20 10 0 1985

1990

1995

2000

Figure 9.4 Transformation of waste management (Source: AOO)

247

Transitions, Transformations, Reproduction seen by AOO as a blessing. The big companies were committed to full compliance and had a strong incentive to respect the law. A simple causality analysis disclosed that there was no single driver but several drivers inﬂuencing each other (Figure 9.5). Packaging policies were targeted as a waste stream to reduce the amount of waste. Waste scandals allowed policymakers to modernize the waste management subsystem. Within the waste regime, the rule system and roles of the different actors changed. Policy was thus endogenous, a response to immediate issues. To deal with problems of capacity, a new network organization (AOO) was created, which served an important coordinating function. The AOO is viewed by some (such as Geelhoed in a speech at the AOO lustrum conference in 2001) an example of the ‘poldermodel’ of consensus-based politics but the organization itself sees itself more as a change agent and mediator (interview with Daemen and Huisman from AOO, 7 September 2004). We ﬁnd that the case study had a good match with the MLP. Changes at the landscape level (environmental consciousness and changed beliefs about waste) created pressure on the waste management regime. But this did not automatically lead to transformation. Pressure from outsiders was important, in particular public outrage over waste scandals and subsequent stricter policies. In Section 9.2 we argued that transformations come about through changes in the regime rules that guide activities of regime actors. Such rules changes were crucial in the case study. For instance, in the 1970s greater environmental awareness led to a new problem deﬁnition: using landﬁll sites for waste disposal was increasingly seen as problematic, because of scarcity of space, stench, and pollution. The waste issue rose rapidly on the problem agenda in the 1980s, as waste scandals led to a public outcry. New regulations were introduced, e.g. the Hazardous Waste Act (1976) and Waste Act (1977),

Growing volumes of waste

Opposition to landfills Waste scandals

Packaging policies AOO Waste capacity problems

Stricter regulations Waste hierarchy

Environmental awareness

Better incineration and landfilling

Waste minimization and recycling

Rising costs waste

Figure 9.5 Arrows of inﬂuence in the Dutch waste management transformation (1970–2000)

248

Transitions, Transformations, Reproduction the Parliamentary motion of Lansink (1979) and several laws and plans in the 1990s (NMP1 and NMP2, EVOA, LAP), and the Environmental Management Act (2002). Stricter regulations changed the selection environment and the organizational framework of responsibilities. Other examples of rule changes are the development of new concepts, visions, and guiding principles (e.g. the waste hierarchy, seeing waste as resource, ‘cradle to grave’ chain management of material streams). Incentive structures were changed through a gradual increase in landﬁll tax. These rule changes led to a marked transformation in waste management in the direction of less landﬁll (a two-thirds reduction between 1990 and 2002), much more recycling (almost a doubling between 1985 and 2000), and more incineration (which more than doubled between 1980 and 2000). The case study not only illustrated that rule changes drive transformations, but also showed that changes in social networks were limited. It were incumbent regime actors (municipal waste-collection agencies, municipalities, provinces, national policymakers, the public) that redirected developments, although their responsibilities and roles changed. The only new actors were private waste collecting and transport companies and the AOO. These limited network changes are a difference with the transition process to sewers. Another difference is that niches did not play an important role in the transformation process. The transformation process depended more on the adaptive capacity of regime actors than on radical technological innovations such as the sewer system in the ﬁrst case study on waste.

9.4 Conclusions and policy implications In Section 9.2 we distinguished three kinds of change processes: reproduction, transformation, and transition. The differences depended on the kinds of interactions in the MLP, and kinds of social groups that took the initiative in the change process. In reproduction processes the system gets improved but maintains its basic structure. The system is dynamically stable, implying that innovation is mainly incremental along trajectories. Dynamics are mainly at the regime level and driven by regime actors who reproduce existing rules as they move along trajectories. In transformation processes development trajectories are redirected through changes in regime rules. Interactions between landscape and regime level are important, as well as pressure from outsiders. But the enactment of transformation is done by regime actors. In transition processes there is a shift to a new socio-technical system and development trajectory. Transitions come about through interactions between all three levels, and the main drive comes from outside actors that develop radically new innovations. Incumbent regime actors may disappear in transitions, giving way to new social groups and networks.

249

Transitions, Transformations, Reproduction Interactions between multiple levels did indeed occur in both case studies, demonstrating the general usefulness of the MLP. The case studies also replicated the differences between transformation and transition processes. It was shown that different multilevel interactions lead to different change processes, and that the role of social groups was different in transition and transformation. In sum, we conclude that the MLP was a useful framework to organize both case studies, and explain the differences between transition and transformation processes. The case studies challenged the existing view that they were coordinated, planned, and goal-oriented processes. The case studies showed that dynamics were more complex, that there were multiple groups involved with different interests and views, leading to contestation and struggle. In the midnineteenth century it took a lot of effort (e.g. by hygienic doctors) to get the problems of hygiene and disease on the agenda. Once these problems were accepted, hygienic doctors, agricultural experts, engineers, and city governments had different views and visions about the best solution. There was contestation and struggle instead of a central plan and widely accepted vision. Furthermore, the eventual implementation of sewer systems depended as much on wider changes at regime and landscape level as on actor strategies. The second case study also showed that none of the actors involved could oversee or control the entire transformation process. Problem agendas, goals, guiding principles (e.g. the waste ladder), and strategies evolved during the process as actors gained experience and responded to wider ongoing developments (growing waste volumes and increased environmental awareness). Furthermore, sudden events such as waste scandals and public outrage created unexpected windows of opportunity that accelerated the transformation. So both transitions and transformations are non-linear and endogenous change processes, emerging from the interactions between social groups. Both cases showed, however, that these endogenous processes are situated in and inﬂuenced by external landscape developments. Neither a regime perspective nor an actor perspective is sufﬁcient for understanding transitions and transformations; they must be combined with each other, as we have done in this chapter. In so doing, we have addressed the challenge mentioned in the introduction of the book, namely the development of a theoretical framework that understands change as caused by both systemic processes and agents and their decisions. Through our typology of socio-technical change processes (reproduction, transition, and transformation) and discussion of underlying mechanisms we offered a theoretical contribution to the issue of societal change. As to the theme of ﬂexibility, we ﬁnd that diversity allows for better selection, but we also want to highlight a tension. While the creation of multiple niches may be good to increase variety, it may also lead to uncertainty, postponement of substantial choice, and delay in transitions. That is what happened in the ﬁrst

250

Transitions, Transformations, Reproduction case study in the 1870s and 1880s. The emergence of two new niches (Liernur and barrel collection) slowed down the niche of sewer systems by creating uncertainty and dispersion of resources. So, while niches are important in transitions, too many niches can lead to dispersion and delay.

9.4.1

Steering and management

There is much policy interest in recent years in system changes, especially in the context of environmental sustainability issues. Through transitions to new transport and energy systems, major jumps in environmental efﬁciency may be achieved. But can transitions or major transformations be managed? From the conceptual perspective and the case studies it may be clear that steering of transitions and transformations is not easy. System changes are non-predictable processes with emergent systemic properties, in which surprises and crises play an important role. As a result, there are limited possibilities for managing transitions and transformations. Power to achieve change is distributed, knowledge is distributed, and there is usually great deal of uncertainty about solutions. For a transition to occur many things have to come together and become aligned. Sometimes deliberative processes of coordination are used for this. An example of a goal-oriented (teleological) and coordinated transition was the shift from coal to gas in the Netherlands in the 1960s (Correlje´ and Verbong 2004). A semi-public organization called the Gasunie was set up for the exploitation and distribution of gas. Shell and Exxon were given 25 per cent ownership stakes into the Gasunie, and users were persuaded through public campaigns to have their furnaces converted and shift to natural gas for heating. It was possible to coordinate this because there were beneﬁts for almost all actors involved and because in that period the state had a strong societal mandate to plan and coordinate developments. This situation is rather exceptional. Historical studies show that few transitions were planned and goal oriented (Geels 2002, 2005). Most transitions, it seems, have emergent properties and cannot be oriented from the start towards particular outcomes as policymakers cannot oversee and control the process. Transitions rather occur through decentralized choices of mutually adapting agents. At the same time both case studies showed that policy actions were important to facilitate system change. This raises the question: What do the case studies learn us about the management of transitions? A ﬁrst lesson, demonstrated in the second case study, is that it is useful to have a more or less shared long-term orientation that serves as the basis for coordination. The ‘waste hierarchy’ provided a general guiding principle to policy thinking about the direction of transformation. Without such a strategic sense of direction, policy can only react to immediate problems (act in a ‘ﬁre brigade’ fashion of putting out ﬁres). A second lesson, demonstrated in both case studies, is that active and substantial steering from policy actors is not likely until problems

251

Transitions, Transformations, Reproduction have become so pressing that solutions can no longer be postponed. In the hygienic transition public authorities ﬁrst implemented incremental wasteremoval solutions (more ﬂushing of canals) or new solutions that were cheap (barrel collection). So they initially preferred to implement non-radical options (mining possibilities for improvement within the existing system). Only when problems became acute and the barrel system was no longer economic (when farmers started using artiﬁcial fertilizers) did city governments switch to sewer systems. Also the second case showed the importance of acute problems (opposition to landﬁlls and incinerators because of waste scandals of contaminated soil and dioxin emissions) for creating new institutions and for doing things that are helpful for longer-term change. Problems provide opportunities for transitions and transformations, however to utilize such opportunities policy has to be forward-looking, and not just reactive. A third lesson, demonstrated in the ﬁrst case study, is that a change in the role perception and ‘sense of self’ of public authorities can be important. The decision to implement sewer systems on a large scale depended strongly on the change from a liberal role perception to a more intervention-oriented view. But such a change in role perception does not come about at will. Instead, it usually depends on wider societal, cultural, and political changes, such as those occurring in the Netherlands in the 1890s.

9.4.2

Transition management in the Netherlands

It is interesting to note that the Dutch national government is committed to achieve change in functional systems as part of sustainability policy. The following systems are viewed unsustainable and in need of change: the energy system based on fossil fuels, the system of motorized mobility characterized by congestion and pollution, and the system of intensive farming with frequent disease problems requiring occasional mass destruction of animals to maintain public trust. Policies towards system change are pursued under the name of transition management. The national authorities are opting for a ﬂexible, forward-looking approach through the use of strategic experiments and the consideration of systemic effects. Long-term visions are guiding societal experiments and to avoid lock in to suboptimal solutions various paths are explored simultaneously. This makes sense given the uncertainty about which option is best. In doing so Dutch authorities rely on the wisdom of variation and selection processes rather than the ‘intelligence’ of planners. A mechanism of self-correction based on policy learning and social learning is part of transition management. Whereas other countries are engaged in managing transitions in an implicit way, the Netherlands does so in an explicit way. The commitment to transitions not allows only for cooperation between ministries but also helps to make political choices that are needed for bringing production and consumption closer to sustainability. Within the transition

252

Transitions, Transformations, Reproduction management perspective, diversity and ﬂexibility are key elements but there is also an element of stability.4 It is too early to tell whether these policies will be successful. The model of transition management marries incrementalism with indicative planning in the form of long-term goals, combining the advantages of both approaches. It is an interesting model for socio-technical development, which is based on the MLP described in this chapter.

Notes 1. This part is based on the project ‘Institutional change in the transition of waste management in the Netherlands’ for the NWO research programme ‘Milieu en Economie’ and draws on contributions of Derk Loorbach and Saeed Parto (in particular Daemen 2003; Loorbach 2003; and Parto et al. 2003). 2. For Ad Lansink, the inventor of the waste management hierarchy (which became widely known as the ‘ladder of Lansink’), there was a direct link between raw materials and energy. As he said: ‘The Club of Rome report really established this link for me because it talked about a shortage of not only raw materials but also of energy. I felt that waste was potential raw material for energy generation.’ (interview with Lansink, 17 February 2004). 3. In 2002 the landﬁlling tax for burnable waste amounted to e79 per ton (62 per cent of the price to be paid). 4. The authors of the present chapter were involved in the development of the model of transition management described in Rotmans et al. 2000 and Rotmans et al. 2001. A description of Dutch policies in the ﬁeld of energy can be found in Kemp and Loorbach (2005) who also offered a discussion of the reasons and interactions that led to the adoption of the model of transition management in 2000.

References Afval Overleg Orgaan (2002). ‘Het poldermodel op de afvalhoop? De rol van overleg in het toekomstige afvalbeleid?’. AOO 2002–03. Bolderman, R. (ed.) (1992). Maatschappelijke aspecten van riolering: De geschiedenis van riolen [‘Societal Aspects of Sewers: The History of Sewers’], Woerden: Vereniging van Producenten van Betonleidingssystemen. Breschi, S. and Malerba, F. (1997). ‘Sectoral Innovation Systems: Technological Regimes, Schumpeterian Dynamics, and Spatial Boundaries’, in C. Edquist (ed.) Systems of Innovation: Technologies, Institutions and Organizations. London: Pinter, pp. 130–156. Buiter, H. (2005). De straat. De wisselwerking tussen techniek, gebruik en betekenissen van Nederlandse stadstraten, 1860–1940 [‘The Street: Interaction Between Technology, Use and Meaning of Dutch City Streets 1860–1940’], PhD thesis, Technische Universiteit Eindhoven. Carlsson, B. (ed). (1997). Technological Systems and Industrial Dynamics. Boston: Kluwer. Carlsson, B. and Stankiewicz, R. (1991). ‘On the Nature, Function and Composition of Technological Systems’. Journal of Evolutionary Economics, 1: 93–118.

253

Transitions, Transformations, Reproduction Corbin, A. (1994). The Foul and The Fragrant: Odour and the Social Imagination. London: Picador. Correlje´, A. and Verbong, G. P. J. (2004). ‘The Transition from Coal to Gas: Radical Change of the Dutch Gas System’, in B. Elzen, F. W. Geels, and K. Green (eds.) System Innovation and the Transition to Sustainability: Theory, Evidence and Policy. Cheltenham, UK: Edward Elgar, pp. 114–34. Daemen, J. (2003). Waste Management Planning in the Netherlands. Paper presentation at waste management council meeting on 5–6 November. Daru, M. (1985). De Kwestie der Faecalie¨n [‘The Matter of Faeces’], PhD thesis, Erasmus University Rotterdam, Netherlands. —— and Van Zon, H. (1987). ‘Afval en milieu in de negentiende eeuw’ [‘Waste and Environment in the Nineteenth Century’], De Ingenieur, 99(5): 55–61. De Swaan, A. (1989). Zorg en de Staat: Welzijn, Onderwijs en Gezondheidzorg in Europa en de Verenigde Staten in de Nieuwe Tijd [‘Care and State: Welfare, Education and Health Care in Europe and the United States in the New Time’]. Bert Bakker, Amsterdam. Dosi, G. (1982). ‘Technological Paradigms and Technological Trajectories: A Suggested Interpretation of the Determinants and Directions of Technical Change’, Research Policy, 6: 147–62. Geels, F. W. (2002). ‘Technological Transitions as Evolutionary Reconﬁguration Processes: A Multi-level Perspective and a Case-Study’, Research Policy, 31(8–9): 1257– 74. —— (2004) ‘From Sectoral Systems of Innovation to Socio-Technical Systems: Insights about Dynamics and Change from Sociology and Institutional Theory’, Research Policy, 33(6–7): 897–920. —— (2005). Technological Transitions and System Innovations: A Co-evolutionary and SocioTechnical Analysis. Cheltenham, UK: Edward Elgar. Hoogma, R., Kemp, R., Schot, J., and Truffer, B. (2002). Experimenting for Sustainable Transport. London: Spon Press. Houwaart, E. (1991). De Hygie¨nisten: Artsen, staat en volksgezondheid in Nederland, 1840–1890 [‘The Hygienists: Doctors, State and Public Health in the Netherlands, 1840–1890’], PhD thesis, Groningen: Historische Uitgeverij Groningen. —— (1993a). ‘Medische Statistiek’ [‘Medical Statistics’], in H. W. Lintsen (ed.) Geschiedenis van de techniek in de negentiende eeuw, Zutphen, deel 3: 19–45. —— (1993b). Professionalisering en staatsvorming’ [Professionalisation and State Formation’], in H. W. Lintsen (ed.) Geschiedenis van de techniek in de negentiende eeuw, Zutphen, deel 3: 81–92. Hughes, T. P. (1983). Networks of Power: Electriﬁcation in Western Society, 1880–1930. Baltimore, CMD: Johns Hopkins University Press. —— (1987). ‘The Evolution of Large Technological Systems’, in W. E. Bijker, T. P. Hughes, and T. Pinch (eds.) The Social Construction of Technological Systems: New Directions in the Sociology and History of Technology. Cambridge, MA: MIT Press, pp. 51–82. —— (1994). ‘Technological Momentum’, in M. R. Smith and L. Marx (eds.) Does Technology Drive History? The Dilemma of Technological Determinism. Cambridge, MA: MIT Press, pp. 101–13. Kemp, R. (1994). ‘Technology and the Transition to Environmental Sustainability. The Problem of Technological Regime Shifts’, Futures 26(10): 1023–46.

254

Transitions, Transformations, Reproduction —— and Loorbach, D. (2005). ‘Dutch Policies to Manage the Transition to Sustainable ¨ kologische Energy’, in F. Beckenbach, U. Hampicke, C. Leipert et al. (eds.) Jahrbuch O ¨ Okonomik: Innovationen und Transformation. Band 4. Metropolis, Marburg, pp. 123–150. ——, Schot, J., and Hoogma, R. (1998). ‘Regime Shifts to Sustainability Through Processes of Niche Formation: The Approach of Strategic Niche Management’, Technology Analysis and Strategic Management, 10: 175–196. ——, Rip, A., and Schot, J. (2001). ‘Constructing Transition Paths Through the Management of Niches’, in R. Garud and P. Karnoe (eds.) Path Dependence and Creation. Mahwah, NJ: Lawrence Erlbaum, pp. 269–99. Kossman, E. H. (1976). De Lage Landen 1780–1914. Amsterdam: Netherlands (also published in 1978 as The Low Countries 1780–1914. Oxford: Clarendon Press). Leonard-Barton, D. (1995). Wellsprings of Knowledge: Building and Sustaining the Sources of Innovation. Boston, MA: Harvard Business School Press. Levinthal, D. A. (1998). ‘The Slow Pace of Rapid Technological Change: Gradualism and Punctuation in Technological Change’, Industrial and Corporate Change, 7(2): 217–47. Loorbach, D. (2003). A Short History of Waste in the Netherlands. Mimeo, Maastricht. Malerba, F. (2002). ‘Sectoral Systems of Innovation’, Research Policy, 31(2): 247–64. Mayntz, R. and Hughes, T. P. (eds.) (1988). The Development of Large Technical Systems. Frankfurt: Campus Verlag; and Boulner: Westview Press. Molina, A. (1995). ‘Sociotechnical Constituencies as Processes of Alignment: The Rise of a Large-Scale European Information Technology Initiative’, Technology in Society, 17: 385–412. Nelson, R. R. and Winter, S. G. (1982). An Evolutionary Theory of Economic Change. Cambridge, MA: Belknap Press. Reid, D. (1991). Paris Sewers and Sewermen: Realities and Representations. Cambridge, MA: Harvard University Press. Rip, A. and Kemp, R. (1998). ‘Technological Change’, in S. Rayner and E. L. Malone (eds.) Human Choice and Climate Change. Columbus, OH: Battelle Press, 2: 327–99. Rosenberg, N. (1976). ‘Factors Affecting the Diffusion of Technology’, in N. Rosenberg (ed.) Perspectives on Technology. Cambridge MA: Cambridge University Press, pp. 189–210. —— (1982). Inside the Black Box: Technology and Economics. Cambridge, MA: Cambridge University Press. Rotmans, J., Kemp, R., van Asselt, M. et al. (2000). Transities & Transitiemanagement. De casus van een emissiearme energievoorziening. Eindrapport van studie ‘Transities en Transitiemanagement’, ten behoeve van NMP-4, Oktober 2000, ICIS & MERIT, Maastricht. ——, Kemp, R., and van Asselt, M. (2001). ‘More Evolution than Revolution. Transition Management in Public Policy’. Foresight 3(1): 15–31. Parto, S., Loorbach, D., and Kemp, R. (2003). Institutional Change During Transitions: The Case of the Dutch Waste Management Sector Paper to be presented at the IHDP Meeting October 16–18, Montre´al, Canada. Schot, J. W. (1998). ‘The Usefulness of Evolutionary Models for Explaining Innovation. The Case of the Netherlands in the Nineteenth Century’, History of Technology, 14: 173–200. ——, Hoogma, R., and Elzen, B. (1994). ‘Strategies for Shifting Technological Systems. The Case of the Automobile System’, Futures, 26: 1060–76.

255

Transitions, Transformations, Reproduction Schot, J. W. ([1947] 1975). Capitalism, Socialism and Democracy, 2nd edn. New York: Harper. Tarr, J. A. (1988). ‘Sewerage and the Development of the Networked City in the United States, 1850–1930’, in J. A. Tarr and G. Dupuy (eds.) Technology and the Rise of the Networked City in Europe and America. Philadelphia, PA: Temple University Press, pp. 159–85. Unruh, G. C. (2000). ‘Understanding Carbon Lock-in’, Energy Policy, 28: 817–30. Van den Akker, J. (1952). Rioleringen, Deel I: Het ontwerpen en berekenen van een riolennet [‘Sewers, Part 1: The Design and Calculation of a Sewer Network’]. Leiden. A.W. Slijthoff’s Uitgeversmaatschappij Van den Noort, J. (1990). Pion of Pionier: Rotterdam-Gemeentelijke bedrijvigheid in de Negentiende Eeuw (‘Pawn or Pioneer: Rotterdam’s Urban Activities in the Nineteenth Century’), PhD thesis, University Leiden (in Dutch). Rotterdam: Stichting PK. Van de Poel, I. (2000). ‘On the Role of Outsiders in Technical Development’, Technology Analysis & Strategic Management, 12(3): 383–97. —— (2003). ‘The Transformation of Technological Regimes’, Research Policy, 32: 49–68. Van Zon, H. (1986). Een zeer onfrisse geschiedenis: Studies over niet-industrie¨le vervuiling in Nederland, 1850–1920 [‘A Very Dirty History: Studies on Non-industrial Waste in the Netherlands, 1850–1920’], PhD thesis, Groningen: Rijksuniversiteit Groningen. Verdoorn, J. A. (1965). Volksgezondheid en sociale ontwikkeling: Beschouwingen over het gezondheidswezen te Amsterdam in de 19e eeuw [‘Public Health and Social Development: Consideratons about Health Care in Amsterdam in the 19th Century’]. Utrecht, Antwerpen: Aula. Vis, G. N. M. (1996). Van ‘vulliscuyl’ tot huisvuilcentrale: Vuilnis en afval en hun verwerking in Almaar en omgeving van de middeleeuwen tot heden [‘From Waste Dumping to Processing of Household Waste: Rubbish and Waste Collection in Alkmaar and Environments from the Middle Ages to the Present’]. Hilversum, Verloren. Walker, W. (2000). ‘Entrapment in Large Technology Systems: Institutional Commitments and Power Relations’, Research Policy, 29: 833–46.

256

10 Analysing Flexibility and Stability in Co-evolutionary Processes Magnus Holme´n and Maureen McKelvey

10.1

Introduction

The analysis of Flexibility and Stability in the Innovating Economy has been pursued in many different ways in this book. This chapter continues the discussion started in Chapter 1 of sketching out research ideas ‘Beyond this Book’ about co-evolutionary processes.1 It does so by considering one topic, namely how we can analyse conceptually and empirically whether or not certain types of change have occurred, as well as by pointing out research areas to further address how, why, and in what dimensions such transformation has occurred. Our motivation for writing this chapter is simple; to encourage research on the interlocking importance of business, technological, public policy, and organizational change over time. We ﬁrmly believe that as a community, we need better concepts and techniques in order to explore further these evolutionary processes of complex transformation in the economy, with a particular emphasis on the nature of ﬂexibility and stability. This chapter focuses on considerations related to systematically linking theoretical arguments with empirical material. Chapter 1 outlined how the book chapters explore three themes, each of which reﬂect and play upon the issue of ﬂexibility and stability in the innovating and transforming economy. They are: Theme 1: Experimenting and Inertia Theme 2: Evolution and Adaptation of Structure Theme 3: Innovating and Technological Transformation These three themes relate ﬂexibility and stability to the two levels of actor and system. Based on the argumentation found in Chapter 1 and the subsequent eight chapters, this chapter starts from the assumption that ﬂexibility and

257

Analysing Flexibility and Stability stability are observable characteristics of economic transformation. Because of the inherent complexity of such processes, there is not one best way (or framework) to analyse them. Likely, these characteristics may differ such that economic systems can be analysed in terms of differential rates of change, differential directions of change, and likely differential outcomes (trajectories). Hence, there are two main motivations for continuing this analysis here, as set in the context of this book and community of scholars represented through the book authors. First, there exists a vast literature within the evolutionary paradigm, which has gone beyond simple, linear causality. The question is, how to go further in analysing economic transformation and still retain an aim of social science, which is not simply a fragmented, postmodern story. For example, many of us make the argument—and empirically demonstrate—that innovation processes are difﬁcult, complex, involve many actors, and so forth, such as found in the innovation system approach, and that such innovation processes are core to economic transformation. Moreover, many researchers in the management and economics of innovation disagree with very simple models, of the type ‘basic science (or small ﬁrms) are the prime driver of commercial innovations’. However, a major challenge remains. That challenge is to ﬁnd conceptual and theoretical models that help us further explain how and why the learning of diverse actors inﬂuence the transformation of the economic system and vice versa. Do processes co-evolve in a complex way? And what effect do they have on other dimensions, including the environment (ﬁtness landscape)? Second, the question remains of how to ‘do’ research and to show and explain causality. Our purpose in stimulating a debate on doing and interpreting empirical research is that in the spirit of appreciative theorizing, as introduced by Nelson and Winter (1982), it is necessary to systematically go back and forth between theoretical arguments and empirical material. Finding ways to study issues related to ﬂexibility and stability more systematically would make it easier to design new research as well as to analyse the theoretical hypotheses, key variables, empirical indicators, and research design of different studies. For example, a key question is explaining how and why two similar ﬁrms—each with their particular combination of competencies, creativity and learning—may likely respond in different ways to the same environmental pressure. Using similar concepts, empirical studies can be compared and contrasted despite differing in many ways, such as casestudy methodology versus large quantitative studies. In this discussion, the explanation of causality refers to rather complex chains of events, and multiple levels of causality, a broader theme that is explored in many chapters in this book. Due to these motivations, this chapter is structured as follows. Section 10.2 addresses how the amount or degree of ‘change’ be conceptualized, as com-

258

Analysing Flexibility and Stability pared to the amount or degree of ‘not change’. This section proposes that we can do so by distinguishing between the relative amounts of ‘old’ and ‘new’ that is found—within some deﬁned, speciﬁc characteristic of the economic system that is observed empirically. Three concepts are therefore introduced here—namely novelty, destruction, and renewal. Novelty implies that something new appears whereas destruction implies that something old disappears. Renewal implies a combination of the other two concepts, such that the modiﬁcation through renewal implies that some new appears, some old disappears but most things do not change (much). Section 10.3 explains and outlines six points about how ﬂexibility and stability can be understood from a paradigmatic perspective on the innovating and transforming economy. The reason for pursuing this line of (paradigmatic) argumentation is twofold. One is that there are many researchers from different disciplines who are coming into the study of innovation and technological change, in relation to industrial dynamics, structural change, and transformation. Thus there is a need to provide a brief introduction that links our concepts into a way of understanding economic transformation. The second is to use these six points, in relation to the discussion of ﬂexibility and stability, in order to sketch out some of the interesting research questions. In other words, this focuses on raising questions about how and why to translate such differential processes and outcomes into concepts and empirical data of use for research. Section 10.4 draws out the conclusions.

10.2

Transformation as involving novelty, destruction, or renewal?

This section addresses the issue of how to conceptualize aspects of change and not change, which are crucial aspects in order to sensibly discuss the innovating and transforming economy. It does so ﬁrst by characterizing and ‘operationalizing’ different amount or degree of change, or what we can call character of change and second by brieﬂy illustrating such changes in relation to the choices about the research design that are necessary to pursue empirical work. In other words, we propose some simple answers to the complex and fundamental question; how can we capture and contrast whether something has changed or not? These ideas of types of change and not change are related to the vast and varied literature that relate to co-evolutionary processes. Co-evolutionary processes are here understood as a way of studying different aspects of industrial dynamics where the emphasis is on processes, learning in new choice situations, complex causality, and historical trajectories. This chapter takes the co-evolutionary paradigm as a starting point but with the intention of probing deeper in how various forms of change across a range of dimensions can be analysed.

259

Analysing Flexibility and Stability Literature that broadly adhere to a co-evolutionary argument often stresses the interlocking importance of business, technological, public policy, and organizational change over time. Bodies of relevant literature include evolutionary economic theory, innovation system approaches, as well as economic history, innovation management, history of science, and history of technology.2 Many of them provide rich descriptions of correlated processes occurring in parallel, over time in such a way as to provide positive and negative feedbacks to certain directions of economic transformation. Speciﬁcally, many authors stress that co-evolutionary processes is a useful way to conceptualize the processes. The main idea behind co-evolution is that change in one dimension may trigger change in other dimensions, as exempliﬁed in (Nelson 1995; Murmann 2003) as well as in (McKelvey 1996) for explaining commercialization of biotechnology.3

10.2.1

The character of change

This subsection addresses one question raised here, namely how and why certain changes in, say, business, technological, public policy and organizational dimensions can be analysed as having greater degrees of novelty whereas others involve greater degrees of resistance (or inertia) to change. This focus on character of change is combined with the ideas discussed in Chapter 1 and in Section 10.3 about the links between learning and transformation. The reason for these links between character of change and learning are as follow. Chapter 1 stressed that economic change involves learning and innovating processes, where actors inﬂuence the trajectories of economic systems, and vice versa. Under these conditions, various actors will make ‘mistakes’ and use diverse assessments of the future when trying to interpret and react to signals. What actors do, how they learn, and how well they can use and transform that knowledge to solve later problems and take advantage of later opportunities are crucial for economic transformation. This speciﬁc interpretation of the innovating and transforming economy has implications for analysing ﬂexibility and stability. The type of ﬂexibility and stability that characterizes a particular economy may be interpreted as somewhat different phenomena, depending on whether actors directly inﬂuence outcomes and the system transforms endogenously or on whether actors mainly respond in a mechanical way to external stimuli. These concepts are linked to the discussion on learning and whether change is driven endogenously or exogenously to the economy, because they are close to what Schumpeter might have called creative innovation and imitative innovation. Schumpeter (1942) distinguished adaptive and creative responses, based on the degree of novelty involved in the act of creation. Adaptive innovations, in these terms, meant imitating something that already existed,

260

Analysing Flexibility and Stability whereas creative innovations meant novelty, where novelty could be built upon novelty and new combinations. In other words, we can call the ﬁrst case endogenously driven change, which may lay closer to creative innovation and the second case reactive change, which is perhaps closer to adaptive innovation.4 Whereas much existing literature on the innovating and transforming economy suggests that many dimensions are changing at once, the questions as deﬁned in the context of this book are more speciﬁcally related to ﬂexibility and stability. This question of whether a particular case of economic transformation, say of the Argentinian economy during the twentieth century, is more endogenously driven or more reactive can only be analysed in relation to some understanding of the differential rates of change, differential directions of change, and likely differential outcomes (trajectories). To better understand the issue of if, where, how, and why change has occurred, and whether these are generated endogenously or in reaction to external stimuli, we draw upon the ideas spelt out in Chapter 1 in this book of reconsidering the Schumpeterian idea of innovation being not only completely new things but also being combinations of new and old things. Concepts like ‘new’, ‘old’, and ‘combinations’ help pinpoint the degree of change involved in innovation or transformation. For the purpose of this book, we therefore wish to further explore how one can analyse the degree of novelty, as compared to existing elements. Using three simple concepts the core idea can be deﬁned as follows: Novelty refers to the appearance, or creation, of something that is new, for example the entrance of new actors, activities, knowledge, and resources. Destruction refers to the termination of something that already exists, for example the exit of existing actors, activities, knowledge, and resources. Finally, renewal refers to something that combines novelty and destruction, for example some existing actors, activities, knowledge, and resources disappear whereas others continue existing and also become modiﬁed in some form through the introduction of novelty. In this case, we are only concerned with the degree of relative newness, and we propose these concepts as three simple ‘answers’ to the complex and fundamental question of how change and not change can be analysed. One reason for proposing these three concepts here is that they can be operationalized in ways to analyse relevant questions about transformation and stability in co-evolutionary processes. They do not allow us to explain why such processes occur. But we argue that they do help us identify, conceive, and structure such analysis in ways to facilitate more nuanced debates. Questions about the degree of novelty, and the degree of change relative to the existing, are ones that run through much of the literature concerned with innovation (Freeman 1995). On the one hand, there are scholars such as

261

Analysing Flexibility and Stability Freeman and Soete (1997) who argue that radical technological discontinuities, found at certain periods, involve not only a renewal of productive opportunities in certain industries but also a renewal and restructuring of institutions and political relationships. On the other hand, there are scholars such as Pavitt (1996, 1998) who consistently argued that the slow build of actor competencies lead to small steps along a continuum of change. A similar debate is carried out within the management literature, in terms of the effects of new technology on industrial dynamics and ﬁrm dynamics, as exempliﬁed by debates about ‘competence-destroying versus competence-enhancing innovations for an industry’.5 Hence, in contrast to the previous discussion of Schumpeter, parts of the modern debate focuses on the effects of different types of technological change on the economy, in relation to businesses and the building up of competencies of ﬁrms, regions, and nations. Answers to questions about the degree of novelty and degree of change involved in a speciﬁc empirical study usually depend upon the operationalization through the research design. Whether new or old is stressed often depends upon the perspective, including time period studied, types and level of effects studied, and unit of analysis. In fact, the argument could be made that many different types of novelties, if analysed carefully, could be seen as exhibiting some degree of ‘newness’ and some degree of ‘oldness’, which affect the economic impacts of transformation.6 If that is true, then for example, the dichotomy between ‘radical’ and ‘incremental’ types of technological and economic change is likely a false one, or at least a red herring. Quite simply, the differences in perception of radical and incremental will likely depend upon the perspective. This implies that one needs to understand more about the type of change involved—including combinations of radical and new innovations as well as more systematically develop empirical research designs in our community, in order to analyse within which dimensions that change occurs or not.

10.2.2

Doing and interpreting empirical research

This section therefore turns our attention to doing and interpreting empirical research in order to develop research designs, which can provide a way to discuss the issues mentioned earlier, such as radical versus incremental. We wish to illustrate here how the application of the ideas outlined in Section 10.2.1 raises a number of methodologically related questions, which must be answered in order to design a new empirical study, or alternatively, to test whether speciﬁc conceptual and theoretical frameworks can help explain empirical phenomena. This section thus links the preceding discussions to the set of choices necessary to operationalize concepts and collect data in order to answer a research question(s). Obviously, the way the study is designed, as well as the way the questions are posed, will have implications

262

Analysing Flexibility and Stability for analysing the results. Our aim is that the following illustration of choices helps highlight why theories, frameworks, and research questions regarding ﬂexibility and stability in particular need quite a nuanced approach to comparing and contrasting data. Developing a research design is a matter of making choices, in order to better match the data and the theory around the increasingly more speciﬁc, research question(s). To illustrate the types of choices necessary to design interesting studies of such phenomena, an example of the choices necessary to make is given, when the researcher intends to design and analyse an empirical study of an innovating ﬁrm, in order to answer a particular theoretical question.7 Having chosen a ﬁrm or industry, the researcher wants to know in which dimensions the ﬁrm ‘exit’ or had access to ‘new’ or ‘old’ activities, knowledge, and resources in order to innovate in technological domains. For example, the question might be ‘which types, and why, of knowledge can be reused within the ﬁrm and lead to better ﬁrm performance, including higher ﬁrm growth?’ Obviously, to answer such a question, there are many types of issues that need to be resolved, including deﬁnitions of concepts. One relevant deﬁnition here may be to state that old knowledge, which is used in a new technical domain or product application may be considered to be knowledge reuse. This may be analysed or measured through whether the same individuals (skilled labour) worked on several projects or at the ﬁrm level, by indicators of market and technological diversiﬁcation. Another set of choices is how to operationalize the idea of ‘innovation’. After all, what does ‘innovating’ mean in this case? Many deﬁnitions and indicators are possible. Based on existing research praxis such as the Community Innovation Surveys (CIS), the researcher decides that innovation in the ﬁrm includes something new with economic value, and may be differentiated in terms of categories such as goods, services, and organization. Innovation may therefore be deﬁned as products, whether goods or services or as process innovations, including technological, and organizational routines. In this case, we assume that the researcher chooses to particularly focus on events leading to product innovations, especially when an existing ﬁrm wants to launch a new bundle of products and services on the market. The researcher then has to determine how to characterize the market, involving new sets of difﬁcult choices. Deﬁning one market may be difﬁcult, given that the concept of product markets often refers to single products, whereas the ﬁrm may rather see their market in terms of the bundle of physical products and services demanded by the customer. The researcher observes a particular product innovation in the ﬁrm and decides to further study it. They observe that the ﬁrm wishes to develop this product and sell it to what to them is a new market, involving new customers, while if the innovation were analysed for the industry at large, then one can observe that the product and market already exist. Another way of putting this is that this ﬁrm is an imitator to other ﬁrms in the

263

Analysing Flexibility and Stability sector, at least in terms of this innovation. Their innovation would not be considered particularly new to the world—but it is new to the ﬁrm. The researcher can continue along this path and use ﬁner grain distinctions related to novelty, destruction, and renewal in order to understand dimensions of innovating in the ﬁrm. Such a characterization is developed when the researcher ﬁnds that in order to launch a new product, the innovating ﬁrm needed to produce the product by using quite a speciﬁc combination of activities, knowledge, and resources. During the study, the researcher ﬁnds out that the ﬁrm primarily innovates through in-house R&D and combinations of resources, although also drawing upon a few but important external networks. To produce the product (whether ﬁnal or intermediate), the ﬁrm uses production technology and service delivery routines, which draw upon an old process technology in the industry as well as existing organizational routines in the ﬁrm. In order to deliver the product, the ﬁrm also has to develop to some extent a new process technology as well as new service delivery routines. The amalgamation of the old and the new across the process technology and service delivery constitute combinations that are new both to the ﬁrm and the industry. Over time, the ﬁrm has to continue to make choices about how much existing ﬁrm routines and competencies are a hindrance—or an aid—to their ultimate goal of innovating and making proﬁts. Then, it turns out that the ﬁrm needs to develop a new business model, which allows them to develop and use this bundle of services and goods based on technological innovations in order to create economic value. At the same time, they need to terminate some in-house routines that have been successfully used many times in their old business model. The ﬁrm expects to appropriate the returns to their innovative activities, including both adaptive and creative innovations, but their success in doing so can only be evaluated ex post facto. Predictions can be made, however, as to who might capture this economic value. The combination of new and old process technology and service delivery routines along with a new product may create economic value for the ﬁrm, and by substitution, it may also create value for the product market and industry. This illustration uses the notions of novelty (new), destruction (old and exit), and renewal (new, old, improvements) in order to discuss and draw out implications of the choices necessary in order to do and interpret empirical research. These three simple notions may thus prove very useful in order to more systematically carry out empirical work on the range of dimensions involved in industrial dynamics, structural change, and economic transformation. Drawing upon this discussion, ﬁve implications are discussed below. The ﬁrst implication for operationalization is how to ‘translate’ the ideas of novelty, destruction, and renewal into more nuanced indicators of the relative

264

Analysing Flexibility and Stability degree of newness. In dictionary deﬁnitions and in much of the literature, novelty is primarily viewed as something entirely novel. Thereby the phenomena that one should observe empirically through indicators and other data and to explain theoretically is the introduction of something ‘new’. Given this interpretation, novelty should consist of actors, activities, knowledge, and resources that did not exist prior to the point of analysis. In contrast, destruction implies a distinct termination of actors, activities, knowledge, and resources that had existed prior to the point of analysis. Thereby, the phenomena that one should observe and explain are the ‘exit’ of something existing, at some point in time. Finally, renewal would logically be a combination involving both novelty and destruction, whereby the relative degrees of ‘new’ ‘old’, and ‘existing’ in activities, knowledge and resources should be studied. The second implication is that studies over time may demonstrate how novelty, destruction, and renewal across a variety of dimensions may be related. On the one hand, novelty is a concept that can be used to describe a start-up ﬁrm or scientiﬁc breakthrough whereas destruction can be applied to discuss the radical technological shift from horse and buggy to the automobile. On the other hand, however, in each case, a more nuanced empirical study would likely suggest that transformation comes about through renewal, due to a combination of ‘new’ and ‘old’ in the development and use of an innovation. The start-up ﬁrm, for example, can likely be analysed in terms of the mobility of skilled individuals, who brought skills and experience into the new organization. Similarly, the scientiﬁc breakthrough relies upon much existing knowledge, skills, and equipment—even if the breakthrough in itself represents a high degree of novelty.8 This is clearly in line with the literature on co-evolutionary processes. The third implication is that, the idea of relative change across dimensions over time is likely linked to broader aspects such as substitution and division of labour amongst industries. Over time, the relative degree of new and old may change, due to substitution effects. The automobile, for example, substituted for and replaced many modes of horse-drawn transportation that were formerly common, but horses and buggies were also used in parallel long after the initial introduction of the car. Technologies continue to ‘compete’ with each other, due to aspects such as consumer behaviour and productive efﬁciencies. The point here is that it is useful to apply ideas about ‘new’, ‘exit’ and ‘old’ in order to specify the time period in which dimensions have changed—or not changed. How long before the innovation or technology ‘took over’ from an earlier competitor? How much was new during the different phases? Time periods matter because it is well known that in any given empirical case, innovations usually do not immediately substitute for existing goods and service products and existing organizational and technological processes. Instead, different varieties and different substitute ways of serving the same

265

Analysing Flexibility and Stability consumer need likely coexist over a longer period, such that the relative degree of novelty, destruction, and renewal looks quite different, depending on whether the analysts chooses longer or shorter time periods. The fourth implication is that this way of operationalizing novelty, destruction, and renewal helps demonstrate why an empirical study of innovation can be useful for understanding market dynamics, by placing the innovating ﬁrm, in relation to the industrial dynamics of competition and demand. For example, one way to develop it further is to identify the relative value created through having a new ﬁrm start up, as opposed to an existing ﬁrm diversifying into new areas. The ﬁfth implication is that such an operationalization of the concepts could be applied to compare the industrial dynamics within a new and ‘hot’ industry like biotech and a ‘traditional’ industry like paper and pulp. By understanding which aspects of an industry are new, which are old, and which are combinations, we may better understand how, when, and why any particular set of economic actors, activities, knowledge, and resources are used for innovating and transforming the economy as opposed to when they are used for preserving existing structures, organizations, and distribution of power.

10.3

Co-evolutionary processes in the innovating economy

Section 10.2 spelt out in some detail how different types of changes can be conceptualized and also suggested a means towards operationalization into studies relevant to understanding ﬂexibility and stability. While this discussion was explicit, the important issue of what dimensions that should be included in analyses of economic transformation was not dealt with. To deal with this issue to some extent, this section discusses how to understand changes through six points about how ﬂexibility and stability can be understood from a paradigmatic perspective on the innovating and transforming economy. Each point, in turn, raises research questions that draw on the discussion in Section 10.2 of novelty, destruction, and renewal when doing and interpreting empirical research. Hence, to make these arguments, this section thus outlines six points that we feel are necessary elements to include in analysing co-evolutionary processes, including both the dimension of change as well as whether change is more endogenously-driven or more imitative. That is, this section considers novelty, destruction, and renewal as observable characteristics of co-evolutionary processes in the innovating economy.9 These six points form the core of our theoretical view of this type of co-evolutionary processes. The ﬁrst point is that the economy consists of actors who affect the innovating and transforming economy, where those actors must be seen as both

266

Analysing Flexibility and Stability active and diverse in how they perceive things, and in what they know, have, and do. The reason this point matters is not only because actors drive multiple exploration and exploitation processes but also because diversity helps drive economic transformation. In other words, focusing on the actor highlights that we need to understand who is acting, to promote (or stop) business, technological, public policy, and organizational changes as well as why they are promoting—or hindering—the innovating, learning, and experimenting inherent in technological transformation. One particularly important category is the ﬁrm. The ﬁrm is not seen as some kind of calculating machine that simply adapts to signals and changes in its environment. Firms are here seen to be actors, which only respond to signals in the environment but also have diverse and sometimes unique competences, resources, and strategies to try to shape their environments. In other words, ﬁrms are seen as economic actors, which respond to both internal and external pressures for change and not to change. This way of thinking about ﬁrms is related both to modern management of technology and innovation literature as well as evolutionary economics and industrial organization literature (see Chapters 2, 3, 7, and 8). Indeed, many ﬁrms actively shape their environment—even if many fail to do so—and it also implies that change is integral to the operations of ﬁrms. To put it more bluntly, given the nature of capitalist competition, ﬁrms are actors that can only remain inert at the risk of extinction, and so they must initiate change as a condition of their existence. And yet, in many circumstances, ﬁrms are well served by being stable, in that existing routines and activities may at times be the basis of continuing ﬁrm success. In the management literature, different types of competencies that accrue over time are crucial for theoretical explanations of the past differential performance of ﬁrms—and for making strategic recommendations about ways to compete in the future.10 So, on the one hand, experimentation and inertia are inherent and endogenous to these processes—but on the other hand, the ﬁrms may well resist the pressure to react in certain directions, to respond and shape environmental conditions. Another actor that shapes the possible trajectories of innovations and thereby economic development are government and public policy actors. Policymakers make actions and decisions, which can affect the societal creation of technological opportunities as well as the broader framework conditions for the economy. This would imply that the proper role of policy is on stimulating and driving creativity and experimentation rather than from any efﬁciency-oriented argument.11 Our view on government and public policy also represents them as actor in the economic system, in that they can proactively inﬂuence, or constrain the wider context for innovation (see Chapters 4 and 9). By stating that actors in the public policy arena are important, however, this does not imply their role is necessarily a positive one. Indeed,

267

Analysing Flexibility and Stability while they have traditionally been important, there are debates about whether government actors are losing or gaining power, as companies go international and as economic processes increasingly occur outside their domains of expertise and of political mandates. Finally, we would like to stress that a long list of other organizations are also active and relevant players to explaining processes of industrial dynamics. The list could be made very long and include such organizations as universities, venture capital, consulting engineering ﬁrms, consumer advocacy groups, NGOs, service operators, and so forth. In fact, the list of societal actors which engage directly in knowledge production or exploitation is an empirical question, and as such an open question. The ﬁrst point is thus about the necessity of considering active and diverse actors as both enabling and sometimes constraining the capability of an economy to engage in innovating, learning, and experimenting. In particular, this perspective stresses two areas for further research. One area is why the ﬁrms experiment with the development of knowledge, opportunities, and organizational forms, to survive—and succeed—in a changing environment. Another area is how much novelty, destruction, and renewal of organizational forms takes place, when seen through the lenses of particular technologies, time periods, and type of response. Questions include: .

How and why do ﬁrms, universities, and governments sometimes substitute for each other, in terms of developing new scientiﬁc and technical knowledge for business contexts?

.

Within the history of particular actors, what causes points of radical destruction like bankruptcy as well as processes of incremental renewal like gradual renewal of product markets?

.

Under what conditions do different types of organization drive such transformatory economic change through internal processes and under what conditions are they more reactive to external stimuli?

.

‘How much’ new and old are involved when actors choose to reorient their strategies? Do the degrees of novelty, destruction, and renewal differ, when they decide to do so through endogenously driven processes, as compared to when, they ﬁnd themselves ‘forced’ to do so to survive?

The second point stresses that uncertainty is integral to innovating, learning, and experimenting. Uncertainty as opposed to risk was deﬁned by Frank Knight (1921), and this is a concept relevant for modern literature, as exempliﬁed in the discussions by Loasby (1999). The philosopher of science Campbell (1987) explains that uncertainty can be understood if the further development of knowledge is viewed as blind generation. Blind generation means that although actors generate alternatives to ﬁt their interpretations of

268

Analysing Flexibility and Stability environmental conditions, they cannot know in advance which alternative(s) will turn out to be well adapted. As such, actors generate novelty in the dark, or with only an approximate understanding of selection pressures. This type of thinking has been developed in various ways. For example, apart from the distinction between Knightian uncertainty and risk, one could make a distinction between ontological and semantic uncertainty (Lane and Maxﬁeld 2005). Ontological uncertainty is where the future is unknowable, for example because the outcome of the activities of one actor is dependent upon the future outcome of the activities of other actors, where these actors themselves do now know what they will be able to come up with. Semantic uncertainty is where there is an interpretative ambiguity so that interpretation may differ. Uncertainty is particularly important for economic transformation in relation to the value of search, in the sense that further development of knowledge affects further choices (Rosenberg 1963, 1976). A key feature of uncertainty is that it is not constant, because over time, the domain of existing uncertainty is often reduced, through learning and routinization about internal activities, knowledge and resources, and learning and institutionalization about external environment processes. Nonetheless, the activities of other actors, as well as possibilities for differences in interpretation may introduce new dimensions of uncertainty. We may also express co-evolution as a learning process within different types of actors as well as one of ignorance and forgetting. The latter statement does not mean that actors become ‘senile’ (even if this may very well happen); indeed learning is cumulative and knowledge tends to be accumulated. However, as transformation proceeds, some old lessons are or need to be forgotten or reinterpreted. Despite an apparent ‘mandate’ to learn, we must remember that actors must often decide whether or not to invest in innovating processes, which are often costly in terms of consuming resource and time. Still, time periods remain important to understanding transformation because new domains of uncertainty will arise over time, due to inherent uncertainties about technological and market futures of innovations. This view of uncertainty and learning being linked to each other has implications for understanding how actors inﬂuence the dynamics of systems.12 Thus, due to such inherent uncertainty about future paths (or about the future results of current choices), information is not ‘correct’, instead actors must interpret and act upon partial information, and both the information and the competence to interpret it are unevenly distributed among different actors. This second point helps explain why the ﬁrst point is necessary to our theoretical understanding of actors—deﬁned within an analysed population—needing to be analysed as both active and diverse in the innovating and transforming economy.13 Actors often make very different choices about whether and why to change or not change, as well as whether to innovate in an adaptive or creative way. The more theoretical explanation for how, why,

269

Analysing Flexibility and Stability and whether the actors differ may be analysed through different lenses, such as perceptions, experimentation, adaptation, adoption, and co-evolutionary processes (see Chapters 3, 4, 6, and 7). One aspect of this is how the search for information to solve problems in such a way as to reduce uncertainty affects public and private choices. Certain types of knowledge are considered ‘sticky’ to particular competencies of actors and not easily transferred whereas other types are more of a public goods character, may be transferred and may more easily affect productivity in many ﬁrms and industries. This has implications for both the distribution of knowledge in society and the reduction of uncertainty in market and technical domains. Any given actor’s ability to capture and interpret information will be partially dependent upon the actor’s own competencies as well as partially dependent upon the type of information necessary for innovating, learning, and experimenting to occur within that speciﬁc technological domain. Here, there are questions about whether and when learning that helps reduce uncertainty primarily builds upon new, old or combinations of activities, knowledge, and resources. .

Why, when, and to what extent does a new ﬁrm build upon competencies developed in existing organizations but transferred to the new one through employees or imitation of routines?

.

Given that industries may shrink or grow in size, under what conditions are the activities, knowledge, and resources moved to other productive uses as opposed to simply becoming inactive and not applied to productive uses?

.

To what extent does ﬂexibility in terms of labour markets stimulate learning as opposed to stiﬂe learning, given that it may shift the cost of learning from society to individuals?

The third point is that an innovating economy involves both individual and societal creation of opportunities and knowledge. This point explicitly relates back to the previous two, in that there is a historical political economy legacy of seeing capitalism as endogenously generating new opportunities through transformation in that certain type of actors may ‘drive’ processes, whereby economic processes must be understood as sociological, political, and psychological as well as economic. One aspect is that knowledge—whether developed through private R&D or by public ﬁnancing—often has impacts on productivity and on the business opportunities identiﬁed by other actors in society. In other words, there may be positive spillovers to other actors than those who developed the technological knowledge, which has been an important topic in economics. Hence, scientiﬁc and technical knowledge has long been seen to have characteristics

270

Analysing Flexibility and Stability of ‘public goods’ and ‘externalities’ in economics. While agreeing with many of these propositions, the evolutionary economics and innovation debates have also stressed that each actor must invest resources to access, interpret, and use knowledge for a business context. One assumption in the literature is that knowledge is a scarce resource and as a consequence needs to be economized, given the assumption that it is difﬁcult to create (and diffuse) economically relevant knowledge within society. In other words, it may be difﬁcult for economies to endogenously create an ability to regenerate opportunities. Thereby, there are debates about why and how aspects such as business, technological, public policy, and organizational change opens up new spaces for the societal creation of opportunities—but also deﬁnes clear limits to novelty and renewal (see Chapters 3, 5, 7, and 8). If innovations are endogenously driven, then one should be able to identify speciﬁc interdependencies between technological change with business, public policy, and organizational change, when new technological opportunities arise over time.14 This may lead to a sequence of events where an innovator is followed by a rapid entrance of imitators, in what Schumpeter vividly referred to as ‘swarming’. This behaviour should in turn lead to further investment into R&D and other activities which are investment into knowledge creation, thereby replenishing the pool of technological opportunities over time. Various mechanisms may help create new technological opportunities, such as internal ﬁrm R&D, learning by using, advances in scientiﬁc understanding, and technological advance from public research or from other industries. If innovations are imitative, then one should be able to identify speciﬁc institutional conditions and/or technological traditions that continuously shape (but do not fully determine) an expected or likely response, which is shared among many actors within a population15. Speciﬁcally, certain innovations function as pointers or guideposts to further innovations. These are often imitative, and they likely result in patterns that we recognize as regularity along a trajectory of search activities, which has been denoted as innovation avenues or natural paradigms. This may differ in different sectors. Concepts such as technological (learning) regimes have been further developed to address how patterns of innovation systematically differ between sectors, as implied by the nature of the technology.16 One way to explore this aspect is that these search activities are not random, but instead appear to follow directions or avenues of search within a broader (possible) search space. This third point raises questions about the extent to which new technological opportunities and knowledge largely involve renewal of existing or novel activities, knowledge, and resources. . Why and how are some economic systems better at generating endogenous change in opportunities whereas others are more reactive?

271

Analysing Flexibility and Stability .

How are novel trajectories to exploit new domains of technological opportunities generated?

.

And to what extent are some technological trajectories within particular industries combinations of previously existing, but fragmented activities?

.

Do new trajectories arise due to competing interpretations of business and technological opportunities as introduced by new actors?

.

And to what extent are actors’ ability to do the same—or change—to some extent conditioned by broader environmental conditions or by network relationships with other types of actors?

The fourth point concerns our view that what we may loosely refer to ‘an innovation system level of analysis’ is useful in order to understand the constraints and opportunities for actors. Approaches such as innovation systems and large technological systems are concepts that refer to the combination of components such as ‘knowledge’ infrastructure, organizations, institutions, and public policy.17 One of the most persistent themes in modern innovation studies is the idea that innovation by enterprises cannot be understood purely in terms of independent decision-making at the level of the enterprise. This literature stresses that economic processes are also shaped by institutional contexts and intensive interactions among organizations18 (see Chapters 4 and 9). This means that the level of analysis shifts from, say, individual innovations and ‘heroic’ inventors or entrepreneurs, to patterns or sequences of innovations and the ongoing changes in specialization, division, and coordination of innovative labour across a range of different (types of) actors. This type of analysis suggests that innovation systems at large may change only slowly, due not only to the organizations and institutions but especially to decisions of actors. As argued earlier, distributed actors within an innovation system only have partial and incomplete knowledge, and many diverse actors contribute to the future development of technological knowledge and innovation. This may affect trajectories—and thereby the relative degree of novelty, destruction, or renewal that occurs within an innovation system over time. The type of questions these points about the importance of open-ended transformation include: .

How and why do different national economies and different sectors differ in their ability to organize explorative and exploitative search activities?

.

To what extent do new innovation systems depend upon the entrance of new organizations and technologies and the exit of old ones, as opposed to continual reconﬁguration into new combinations?

.

What are the organizational forms of interaction inside the innovation system, by which business-relevant knowledge is ‘renewed’?

272

Analysing Flexibility and Stability . What are the effects of different organizational forms on the degree and type of radically new or incremental learning that occurs? The ﬁfth point about ﬂexibility and stability that we wish to raise here concerns the general nature of systems and how systems change. From an abstract point of view, systems consist of a range of elements, connection between elements, rules for connections, and connections among some of the rules (see Chapter 5). Changes in the systems take place through a new connection to or disconnection of some element(s), which may have taken place through alteration in the rule-structure. A fundamental feature of any economic system is that there are many elements and that not all are connected, nor can all elements in the economic system of the real world (or realistic model) ever be connected.19 That is, metaphorically we can talk of open economic systems, while mathematically we denote them as non-integral. A second phenomenon about ‘system dynamics’ is that economic systems depend upon human beings, and hence one could choose either a subjective or an objective approach to analysis. On the one hand, human beings act and interact in time and therefore a model of human behaviour needs to (explicitly or implicitly) conceptualize activities that run in parallel or in sequences. Any economic system has a subjective side, consisting of perceptions, ideas, and so on of different actors, whether they be individuals, ﬁrms, consumers, and so on. However, on the other hand, this book takes the stance that ontologically any economic system also has some ‘factual’ characteristics, regardless of perceptions. The two characterizations of subjective versus objective are the two sides of the same coin—but the relative degree of novelty, destruction, or renewal that is observed may differ, depending on whether the analysis focuses on the individual actor’s perception of change or on more systematically observed factual events.20 This leads to questions about whether and how the overall system coevolves with the internal organizations and competencies. . Do different national political economic systems have different degrees of ‘ﬂexibility’ and ‘inertia’, depending on the underlying elements that determine system dynamics? . To what extent does a speciﬁc cultural interpretation of the market rely upon existing cultures, as opposed to new forms of globalized culture? The sixth point returns to the deﬁnition of economic transformation, which was presented in Chapter 1. We argued that transformation here means something speciﬁc, in relation to the ongoing processes of industrial dynamics, structural change, and transformation. Economic transformation as a concept used here refers to a non-reversible process, referring to quantitative and qualitative changes in components and connections, and often driven by opportunities afforded by innovations, deﬁned in a very broad sense. This

273

Analysing Flexibility and Stability book has argued conceptually and shown empirically that such qualitative transformation is a normal part of the market economy—not a special event.21 Questions to be further addressed here are how to conceptualize what drives transformation as well as what types of change are involved. Such economic transformation may well be driven by processes of complexity and selforganization as well as processes of adaptation and co-evolution. There are also follow-on questions to our statement that the concept of transformation, as used here, refers to both processes driven by very large and discontinuous changes as well as processes driven by very small changes, which follow upon an existing trajectory (see Chapters 5, 6, 7, and 9). .

Which, and why, are different processes more important in different cases? Under what conditions are one or the other more common?

.

What are the implications for actors’ ability to learn, innovate, and experiment, if an innovation system is changing more or less rapidly?

.

When are actors more imitative, more reactive, or even resist transformation through pockets of inertia?

In summary, the above discussion has deﬁned six points that we believe are key to interpreting what goes on within co-evolutionary processes of economic transformation. The resulting research questions pose fairly abstract questions, but they illustrate the need to consider degree of change in terms of novelty, destruction, and renewal as well as whether the change is endogenously driven in the sense of actors directly inﬂuencing outcomes and the system transforms in this interaction between actor and system or whether the change is more reactive in the sense of actors mainly respond in a mechanical way to external stimuli. Given that our communities in innovation studies, industrial dynamics, and evolutionary economics tend to emphasize the change in co-evolutionary processes, we wish to stress the need to analyse both endogenously driven and imitative innovation as well as resistance to change. If we do so, then we will come further towards explaining the differential rate and directions of change among different sets of actors, activities, knowledge, and resources in the economy.

10.4

Discussion

The aim of this chapter has been to systematically link conceptual and theoretical arguments with empirical material in order to conceive and structure analyses of these evolutionary processes of complex transformation in the economy, with a particular emphasis on ﬂexibility and stability. We wish to

274

Analysing Flexibility and Stability conclude the chapter by returning to the issues of research questions and how to interpret endogenously driven change as opposed to reactive change. Section 10.2 addresses the question of how the amount or degree of ‘change’ be conceptualized, as compared to the amount or degree of ‘not change’. Section 10.2 proposes that we can do so by distinguishing between novelty, destruction, and renewal. These concepts require a set of choices in order to be operationalized into empirical work, where the most fruitful path there is the relative amounts of ‘old’ and ‘new’ that is found—within some deﬁned, speciﬁc characteristic that is observed empirically. Section 10.3 explains and outlines six points about how ﬂexibility and stability can be understood from a paradigmatic perspective on the innovating and transforming economy. When combined with the preceding discussion of degree of change, this leads to the suggestion of interesting areas for future research. One area of discussion relates to the research questions highlighted here. The issue is how the proposed research questions relate to choices about the operationalization of novelty, destruction, and renewal across different levels of analysis. Table 10.1 uses the concepts of novelty, destruction, and renewal along with the actor and system levels, in order to outline sets of some relevant questions. Table 10.1 thus illustrates different types of questions, relevant for additional research. Other questions are raised in Section 10.3. Here, the purpose is to differentiate sets of questions by type of processes and by level of analysis. On the one hand, the researcher could ask questions about how and why particular actors innovated—or disappeared—and on the other, questions about the effects on a sector or national economy.

Table 10.1. Outlining questions about novelty, destruction, and renewal at different levels of analysis

Actor level

Novelty

Destruction

Why and how do people and organizations introduce novelty? What activities, knowledge, and resources do they draw upon to develop new technologies?

How can actors exit What synergies and conﬂicts arise when ‘old’ the system? and ‘new’ organizations Why do actors exit attempt to form new the system? combinations to survive What happens to activities, knowledge, and resources, or expand? when existing organizations exit a business area?

System level How many new entities Does creative destruction appear in a period, within occur in a way beneﬁcial an industry in a national to economic growth? economy? Are, and if so how, resources freed-up for other parts What effect does technological novelty have on the rate of the process, when the and direction of economic old exits? development?

Renewal

What is the relative rate of renewal in particular sectors in different economies? Why do they differ (or not)?

275

Analysing Flexibility and Stability Indeed, we would even go on to argue that Table 10.1 and Section 10.3 outline some of the questions that we ought to be asking to better understand ﬂexibility and stability. They could be useful for designing new empirical studies but also for comparing and contrasting the results of existing studies in terms of explanations for why the economy does—or does not—demonstrate innovative and transformative qualities. Another topic of discussion is how to interpret endogenously driven change as opposed to reactive change. The discussion in this chapter provides a range of relevant concepts for further interpreting ﬂexibility and stability, ranging from whether that change involves fundamental transformation, change in response to external signals (stimuli), a small degree of change but mostly the same as before, or essentially no change. Table 10.2 uses the concepts discussed in relation to on the one hand, creative innovation and endogenously driven change, and on the other hand, to imitative innovation and reactive change. These help us reinterpret the initial discussion of concepts relevant to ﬂexibility and stability, as found in Tables 1.1 and 1.2. Table 10.2 below is similar to these tables, but rather than ‘ﬂexibility and stability’, this table uses concepts ranging from ‘fundamental change in the system, where learning occurs’ to ‘no change’. Two levels are also considered here including, where the focus is on actors and on system dynamics. Table 10.2 thus uses a broader range of concepts, in order to capture whether the characteristics observed as ﬂexibility and stability tend to come through change introduced fundamental transformation or change that occurred in response to external stimuli. In connection with the two levels of actor and system, Table 10.2 thus provides a reinterpretation of concepts relevant for analysing ﬂexibility and stability, in relation to industrial dynamics, structural change, and transformation. Our contribution in this chapter is a modest one, related to our proposal of how to study these evolutionary processes of complex processes in the economy, with a particular focus on ﬂexibility and stability. The small shift in perspective proposed here may facilitate for other researchers, who are interTable 10.2. Concepts relevant for type of change occurring in economic systems Fundamental transformation, Change, in respond to Mostly the same, generated endogenously environmental changes but some change No change Actor level

Learning Innovating Flexibility

System level Phase shift Adaptation Technological transformation

276

Flexibility Experimenting Adoption New connection

Resistance Experimenting

Resistance

Responsiveness Turbulence

Stability Evolution

Inertia Rigidity

Analysing Flexibility and Stability ested in debating and operationalizing ﬂexibility and stability as system characteristics of the evolving economy. Reason for carrying out this simple exercise include ﬁnding ways to debate what co-evolutionary processes, industrial dynamics, structural change, and transformation may mean, if we focus on the dimensions, degree, and types of change involved. In relation to that, we perceive a need to ﬁnd ways to stimulate more communication and debate among researchers, not least to the newer generations, about how to ‘do research’. In general, we would argue that our communities still face the fundamental issue of how and why to do research in such a way as to analyse economic growth as the result of endogenous industrial dynamics, which involves complexity, multiple processes (dimensions), and multiple-levels of causality.

Notes 1. To do this, the chapter uses the notion of ‘co-evolutionary processes’ in order to help explain why it is that different processes occur in the economy in parallel over time, and that types of changes in some dimensions affect other processes of industrial dynamics, structural change, and transformation. (See Nelson 1995; McKelvey 1996; Lewin and Volberda 1999; Lewin et al. 1999; Dosi 2000; Murmann 2003). This chapter sees that co-evolution as a useful ‘lens’ for analysis, by combining an understanding of the degrees of change involved with an explicit discussion of the dimensions and character of change. 2. De Solla Price (1963); Rosenberg (1976); de Solla Price (1984); Dosi et al. (eds.) (1988) Witt (1992); Freeman (1995); Edquist (1997); Metcalfe (1998). 3. Indeed, a stronger claim is that co-evolution may be necessary for economic transformation or transition, using very speciﬁc deﬁnitions. Murmann (2003: 22) deﬁnes it as follows: ‘Two evolving populations coevolve if and only if they both have a signiﬁcant causal impact on each other’s ability to persist. Such causal inﬂuence can proceed through two avenues: (1) by altering the selection criteria or (2) by changing the replicative capacity of individuals in the population without necessarily altering the selection criteria. Kauffman (1993) uses the idea of coupled ﬁtness landscapes to express this conception of coevolution. In coevolution a` la Kauffman, one partner deforms the ﬁtness landscape of the second partner and vice versa. As a result, a coevolutionary relationship between entities can increase the average ﬁtness of both populations, decrease the average ﬁtness of both, or have a negative or positive impact on the average ﬁtness of one but not the other.’ This is a speciﬁc view of co-evolution, including aspects of causality, where the objective is to ﬁnd speciﬁc inﬂuences across different populations and environments. Drawing upon (Nitecki 1983), Murmann goes on to specify cross-ﬂows to establish evidence of reciprocal inﬂuence between two co-evolving partners, as opposed to sequential adaptations from different causes or simultaneous adaptation to the same environment.

277

Analysing Flexibility and Stability 4. As a community, we are perhaps more interested in the ﬁrst type of endogenously driven change, although also aware of the importance of reactive change. 5. Tushman and Anderson 1986; Teece 1986; Christensen 1997; Christensen and Overdorf 2000. 6. Landes 1998; Mokyr 2002. 7. Of course, the same holds true in that the research design matters in that the way questions are posed can lead to very different recommendations to the ﬁrm. For example, if the question is about ﬁrm strategy, then one could focus only on innovations as novelty. However, such a study is likely leading to recommendations to innovate but even a cursory glance of successful companies show that this may not say much, and sometimes such a recommendation may be erroneous. However, by studying novelty, destruction, and renewal in relation to business products and processes, this may help the ﬁrm understand whether and why they should ditch all existing products or else whether to slowly transform their core area into a new product bundle of goods and service. 8. Indeed, most empirical cases probably allow the careful analyst to observe characteristics of activities, knowledge, and resources relevant for innovation that suggest that some dimensions involve novelty whereas other dimensions involve destruction—and continuation of the old. 9. Co-evolution is here deﬁned in terms of dynamic processes, where actors and components affect each other as well as the economic system as a whole. As explained earlier, we do not impose a strict view on the need to inﬂuence mutual trajectories including survival (a) by altering the selection criteria or (b) by changing the regenerative or replicative capacity of individuals. 10. See Abernathy and Utterback 1978; Utterback and Abernathy 1975; Bourgeois 1985; Dodgson 2000; Grant 2003; Tidd et al. 2001. 11. Metcalfe 1994; Moreau 2004. 12. This discussion of uncertainty and learning can also be developed in other directions. For example, it may help us understand that competition in industrial market economies is driven by innovations in three ways. First, in the face of consumer demand for products, ﬁrms compete in terms of the design, quality, and performance characteristics of products; that is, in terms of a set of technically determined product attributes. Second, ﬁrms compete with process technologies which shape both the technical forms of the product, and the cost structure of the ﬁrm. In general, such technological and market competition is a dynamic, evolving phenomenon: patterns of demand evolve and change, and this together with the innovative activities of competitors means that ﬁrms must innovate on a continuing basis. Third, they face uncertainty about markets. Hence, the structure of demand, processes of consumption and the perception of consumers are here considered active in shaping the economic system. A means way to conceptualize this is to view consumption as a process that alters the nature or even terminates the existence of different ‘solutions’ (product or services). Thus, changes in demand and consumption may provide new opportunities for innovation or decreasing the appropriability of ‘old’ innovations. The future trajectories of development will depend not only on public policy and ﬁrms but also on demand and new connections driven by consumers. A special thanks to Keith Smith on this issue.

278

Analysing Flexibility and Stability 13. McKelvey 2004. 14. Nelson and Winter 1982; Malerba and Orsenigo 1993; Klevorick et al. 1995; Breschi et al. 2000; Palmberg 2003. 15. Sahal 1981; Dosi 1984; Nelson and Winter 1982; David 1985; Dosi et al. 1988. 16. Malerba and Orsenigo 1993; Breschi et al. 2000; Palmberg 2003. 17. Constant 1980; Hughes 1987; Hughes 1992. Within the innovation system literature, various approaches stress somewhat different explanatory variables and ‘levels’ in order to address the importance of institutional context. The national system stresses the national knowledge structure and industrial activities whereas the sectoral system focuses on production and innovation in relation to industrial dynamics. In contrast, the regional system deﬁnes activities as increasing in relation to regional boundaries; and the technological system in relation to the underlying technological knowledge and knowledge ﬂows. See Lundvall 1992 and Nelson 1993 for national, Malerba 2002 and Malerba 2004 for sectoral; Braczyk et al. 1998 for regional; Carlsson et al. 2002 for technological systems. See comparisons of approaches in Edquist (ed.) 1997 and in Edquist and McKelvey (eds.) 2000. 18. See Hodgson 1989; Nelson and Sampat 2001; Fagerberg et al. 2004 19. Arthur 1989; Kauffman 1993; Ayres 1994; Kauffman 1995; Potts 2000. 20. Points ﬁve and six are in many ways related to a larger discussion beyond this book, where transformation may be analysed using a variety of tools available for understanding system dynamics. Of course, if the whole economic system (or innovation system) is analysed, then this requires an understanding of what the system might be and how it might change. While the concept of ‘systems’ has been developed in many different branches of research such as system theory, system dynamics, and control engineering, this book is particularly interested in complex, self-organizing and open economic systems. 21. See Saviotti and Metcalfe 1991; Metcalfe 1998.

References Abernathy, W. J. and Utterback, J. (1978). ‘Patterns of Innovation in Technology’, Technology Review, 80(7): 40–7. Arthur, B. (1989). ‘Competing Technologies, Increasing Returns and Lock-in by Historical Events’, The Economic Journal, 99: 116–31. Ayres, R. U. (1994). ‘Toward a Non-linear Dynamics of Technological Progress’, Journal of Economic Behavior and Organization, 24: 35–69. Bourgeois, L. J. III (1985). ‘Strategic Goals, Perceived Uncertainty, and Economic Performance in Volatile Environments’, Academy of Management Journal, 28(3): 548–73. Braczyk, H-J., Cooke, P., and Heidenreich, M. (eds.) (1998). Regional Innovation Systems— The Role of Governances in a Globalized World. London: UCL. Breschi, S., Malerba, F., and Orsenigo, L. (2000). ‘Technological Regimes and Schumpeterian Patterns of Innovation’, The Economic Journal, 110: 388–410. Campbell, D. (1987). ‘Blind Variation and Selective Retention in Creative Thought as in Other Knowledge Processes’, in G. Radnitzky and W. W. Bartley III (eds.) Evolutionary Epistemology, Rationality and the Sociology of Knowledge. La Salle, Il: Open Court.

279

Analysing Flexibility and Stability Carlsson, B., Jacobsson, S., Holme´n, M., and Rickne, A. (2002). ‘Innovation Systems: Analytical and Methodological Issues’, Research Policy, 31(2): 233–45. Christensen, C. (1997). The Innovator’s Dilemma: When New Technologies Cause Great Firms to Fail. Boston, MA: Harvard Business School Press. Christensen, C. and Overdorf, M. (2000). ‘Meeting the Challenge of Disruptive Change’, Harvard Business Review, 78 (March–April): 66–76. Constant, E. W. II (1980). The Origins of the Turbojet Revolution. Baltimore, MD: Johns Hopkins University Press. David, P. (1985). Clio and the Economics of QWERTY, American Economic Review, 75: 332–37. de Solla Price, D. J. (1963). Little Science, Big Science. New York: Columbia University Press. de Solla Price, D. J. (1984). ‘The Science–Technology Relationship, the Craft of Experimental Science, and Policy for the Improvement of High Technology Innovation’, Research Policy, 13: 3–20. Dodgson, M. (2000). The Management of Technological Innovation: An International and Strategic Approach. Oxford: Oxford University Press. Dosi, G. (1984). Technical Change and Industrial Transformation. New York: Macmillan. Dosi, G. (2000). Innovation, Organization and Economic Dynamics, Selected Essays. Cheltenham, UK: Edward Elgar. Dosi, G., Freeman, C., Nelson, R., Silverberg, G., and Soete, L., (eds.) (1988). Technical Change and Economic Theory. London: Pinter and New York: Columbia University Press. Edquist, C. (ed.) (1997). Systems of Innovation: Technologies, Institutions and Organizations. London: Pinter. Edquist, C. and McKelvey, M. (eds.) (2000). Systems of Innovation, A Two volume reference collection. Cheltenham, UK: Edward Elgar. Fagerberg, J., Mowery, D., and Nelson, R. (2004). The Handbook of Innovation. Oxford: Oxford University Press. Freeman, C. (1995). ‘The ‘‘National System of Innovation’’ in Historical Perspective’, Journal of Economics, 19: 5–24. Freeman, C. and Soete, L. (1997). The Economics of Industrial Innovation. London: Pinter. Grant, R. M. (2003). Contemporary Strategy Analysis, 4th edn. Cambridge, MA: Blackwell. Hodgson, G. M. (1989). Institutional Rigidities and Economic Growth, Cambridge Journal of Economics, 13(1): 79–101. Hughes, T. (1987). ‘The Evolution of Large Technological Systems’, in W. Bijker, T. Hughes, and T. Pinch (eds.), The Social Construction of Technological Systems: New Directions in the Sociology and History of Technology. Cambridge, MA: MIT Press, pp. 51–82. Hughes, T. (1992). ‘The Dynamics of Technological Change: Salients, Critical Problems, and Industrial Revolutions’, in G. Dosi, R. Giannetti, and P. A. Toninelli (eds.) Technology and Enterprise in a Historical Perspective. Oxford: Oxford University Press. Kauffman, S. (1993). The Origins of Order. Self-Organization and Selection in Evolution. Oxford: Oxford University Press. Kauffman, S. (1995). At Home in the Universe: The Search for Laws of Self-Organization and Complexity. New York: Oxford University Press.

280

Analysing Flexibility and Stability Klevorick, A. K., Levin, R. C., Nelson, R. R., and Winter, S. G. (1995). ‘On the Sources and Signiﬁcance of Interindustry Differences in Technological Opportunities’, Research Policy, 24: 185–205. Knight, F. (1921). Risk, Uncertainty and Proﬁt. Boston: Houghton Mifﬂin Co. Landes, D. (1998). The Wealth and Poverty of Nations Why Some Are So Rich and Some So Poor. New York: W. W. Norton. Lane, D. A. and Maxﬁeld, R. R. (2005). ‘Ontological Uncertainty and Innovation’, Journal of Evolutionary Economics, 15: 3–50. Lewin, A. Y. and Volberda, H. W. (1999). ‘Prolegomena on Coevolution. A Framework for Research on Strategy and New Organizational Forms’, Organizational Science, 10(5): 519–34. Lewin, A. Y., Long, C. P., and Carroll, T. N. (1999). ‘The Coevolution of New Organizational Forms’, Organizational Science 10(5): 535–50. Loasby, B. (1999). Knowledge, Institutions and Evolution in Economics. London: Routledge. ˚ . (ed.) (1992). National Systems of Innovation: Towards a Theory of Innovation Lundvall, B-A and Interactive Learning. London: Pinter. Malerba, F. (2002). ‘Sectoral Systems of Innovation and Production’, Research Policy, 31(2): 247–64. Malerba, F. (2004). Sectoral Systems of Innovation. Cambridge: Cambridge University Press. Malerba, F. and Orsenigo, L. (1993). ‘Technological Regimes and Firm Behaviour’, Industrial and Corporate Change, 2(1): 45–74. McKelvey, M. (1996). Evolutionary Innovations: The Business of Biotechnology. Oxford: Oxford University Press. McKelvey, M. (2004). ‘How and Why Dynamic Selection Regimes Affect the Firm’s Innovative Search Activities’, Innovation Management, Policy, Practise, 6(1): 3–24. Metcalfe, S. (1994). ‘Foundations of Technology Policy—Equilibrium and Evolutionary Perspectives’, in P. Stoneman, P. Dasgupta, and R. Nelson (eds.) Handbook in the Economics of Innovation. London: Blackwell. Metcalfe, S. (1998). Evolutionary Economics and Creative Destruction. London: Routledge. Mokyr, J. (2002). The Gifts of Athena Historical Origins of the Knowledge Economy. Princeton, NJ: Princeton University Press. Moreau, F. (2004). ‘The Role of the State in Evolutionary Economics’, Cambridge Journal of Economics, 28: 847–74. Murmann, J. P. (2003). Knowledge and Competitive Advantage. The Coevolution of Firms, Technology and National Institutions. Cambridge: Cambridge University Press. Nelson, R. R. (1991). Why Do Firms Differ and How Does It Matter? Strategic Management Journal, 12(1): 61–74. Nelson, R. R. (ed.) (1993). National Innovation Systems. A Comparative Analysis. Oxford: Oxford University Press. Nelson, R. R. (1995). Co-evolution of Industry Structure, Technology and Supporting Institutions, and the Making of Competitive Advantage, International Journal of the Economics of Business, 2(2): 171–84. Nelson, R. R. (1996). The Sources of Economic Growth. Cambridge, MA: Harvard University Press. Nelson, R. R. and Winter, S. G. (1982). An Evolutionary Theory of Economic Change. Cambridge, MA: Harvard University Press.

281

Analysing Flexibility and Stability Nelson, R. R. and Sampat, B. N. (2001). Making Sense of Institutions as a Factor Shaping Economic Performance, Journal of Economic Behaviour and Organization, 44: 31–54. Nitecki, M. H. (1983). Coevolution. Chicago, IL: University of Chicago Press. Palmberg, C. (2003). Turning opportunities into innovation. PhD thesis, Royal Institute of Technology, Stockholm. Pavitt, K. (1996). ‘National Policies for Technical Change Where Are the Increasing Returns for Economic Research’, Proceedings of the National Academy of Sciences USA, 93: 12693–12, November, Colloquium Paper. Pavitt, K. (1998). ‘Technologies, Products and Organization in the Innovating Firm: What Adam Smith Tells us and Joseph Schumpeter Doesn’t’, Industrial and Corporate Change, 7(3): 433–52. Potts, J. (2000). The New Evolutionary Microeconomics. Complexity, Competence and Adaptive Behaviour. Cheltenham, UK: Edward Elgar. Rosenberg, N. (1963). ‘Technological Change in the Machine Tool Industry 1840–1910’. Journal of Economic History, 23(4): 414–43. Rosenberg, N. (1976). Perspectives on Technology. Cambridge: Cambridge University Press. Rosenberg, N. (1982). Inside the Black Box. Cambridge: Cambridge University Press. Sahal, D. (1981). Patterns of Technological Innovation. Reading, MA: Addison-Wesley. Saviotti, P. P. and Metcalfe, J. S. (eds.) (1991). Evolutionary Theories of Economic and Technological Change. Reading, UK: Harwood Academic. Schumpeter, J. A. (1934). The Theory of Economic Development. Cambridge, MA: Harvard University Press. Schumpeter, J. A. (1942). Capitalism, Socialism and Democracy. New York: Harper and Brothers. Schumpeter, J. A. (1954). History of Economic Analysis. New York: Oxford University Press. Teece, D. J. (1986). ‘Proﬁting from Technological Innovation: Implications for Integration, Collaboration, and Public Policy’, Research Policy, 15: 285–305. Tidd, J., Bessant, J., and Pavitt, K (2001). ‘Managing Innovation’, 2nd edn. New York: John Wiley & Sons. Tushman, M. L. and Anderson, P. (1986). ‘Technological Discontinuities and Organizational Environment’, Administrative Science Quarterly, 31(3): 439–65. Utterback, J. and Abernathy, W. J. (1975). ‘A Dynamic Model of Process and Product Innovation’, Omega, 33(4): 639–56. Witt, U. (ed.) (1992). Explaining Process and Change: Approaches to Evolutionary Economics. Ann Arbor, MI: University of Michigan Press.

282

Index

Note: The index includes all authors cited in the text. 3G (third generation), and innovative opportunities 55–7 appropriation 56–7, 65, 66 economic value 55–6, 60 resources 56, 63 absorptive capacity 202 Acha, V 81 actors: and co-evolutionary processes 266–8 and conceptualizing ﬂexibility/stability 4–5, 10–11, 260 and innovation 50 and systems 4 adaptation 19 and accounting for evolutionary adaptation 135–40 and capitalism 125–6 and complexity 14, 123, 144–5 and constraints on 125 and development 121, 122–3, 127–8 and differential growth rates 130–4, 138–9 and economic growth 121 and economic transformation 10 and evidence for structural 128–34 and evolutionary model of development 146 and implications of 121 and innovation 260 and knowledge 14–15, 146 correlation of 148–52 given 121–2 growth in 144 new 122, 123 and logistic principle 140–4 and markets 123–5, 140

and novelty 126 as process 136 and selection 127–8, 145 as self-transformation 125–6, 129 as transition 126 advertising 103 Afuah, A 211 Akintoye, A 75, 76, 78 Allen, G 75, 78 Allen, P 113 Allen, T J 30 Alstom 85, 87, 88 Alvarez, S A 51 Amit, R 206, 207 Andersen, E S 139–40 Anderson, P 209–10, 211, 213 appreciative theorizing 258 appropriation, and innovative opportunities 49, 52, 56–7, 63–6 Ardichvili, A 64 Arora, A 159, 203 Arthur, W B 146 Arup 12, 28 and ArupFire 32 development of 35 impact of regulatory environment 35–6 modelling ﬁre evacuation 36–8 and characteristics and advantages of 34 and foundation of 33 and growth of 33–4 and innovation technologies 35 and management style in 34–5 Arup, Ove 33 automation, and engineering 42, 43

283

Index automotive industry: and design process 40, 41 and simulation tools for 38–42 Babtie Group 80 Baldwin, C 30, 203, 215, 216 Barabasi, A 108 Barney, J B 206 Baumgartner, M 82 Bazalgette, Joseph 238 Becher, T 29, 43 Bechtel 77 Beitz, W 30 Bhave, M P 62 Bhide´, A V 63 ¨ rkdahl, J 55 Bjo Blair, R D 205 blind generation, and uncertainty 268–9 Bohlin, E 55 Bokhari, F 159 Bolderman, R 239 Bombardier 85 boundaries of ﬁrms 15 and change in 157 and co-evolution 158, 159 and computer industry model: computer characteristics 167 demand for computers 170 market for components 167–8 simulations 175–93 technical progress 168–9 technological discontinuities 173–5 vertical integration/specialization 171–2 and ﬁrm competences 159–60, 163, 166, 206–7, 209–12 and industry maturity 203 and technological change 202, 217–19 dynamic capabilities 209–12 dynamic transaction costs 212–14 indeterminate effects of 204 modularity 214–17 simultaneous outsourcing/vertical integration 204 and technological innovation 16 in technologically stable conditions 205 behavioural factors 208–9

284

core/distinctive competences 206–7 transaction costs 207–8 and temporal aspect 160, 166 and theoretical determinants of 163–7 and upstream/downstream industries 159 and vertical integration/specialization 193–6 explanations of 202–3 industrial economics perspective 159 theoretical determinants of 163–7 transaction cost perspective 158–9, 203, 207–8 bounded cognition 20, 105 bounded rationality 20, 105, 158 Bower, J L 210, 211 Brady, T 82, 88 Braun, E 129 Braverman, H 42 Breschi, S 51, 228 Bresnahan, T F 162 Bressington, Peter 35, 36 Brooks, D R 109 Brusoni, S 28, 30, 82, 211, 217 Buiter, H 238, 240 Burns, A F 129 Busenitz, L W 51 business models 65 Campbell, D 268 capital goods: and capital goods value stream 82–6 acquiring integration capabilities 86–8 systems integration 83–6 and complex products and systems (CoPS): challenges in producing 81–2 nature of 81 production process 81 and nature of 81 and public-private partnerships 13, 73–5 challenges of 89–90 repeatable solutions 88 and systems integration 83–6 see also public-private partnerships

Index capitalism: and adaptation 125 and dynamic nature of 127 and economic change 7, 125–6 and evolutionary model of development 146 and innovation 126 as knowledge-based system 149, 151–2 and knowledge-led development 122 as self-transforming system 126 Carlsson, B 228 Chai, A 103 Chandler, A D 16, 202, 205, 206 change 3–4 and actor/system levels of analysis 4–5, 10–11, 275–6 and character of 260–2 adaptive innovation 260 creative innovation 261 degree of novelty 261–2 destruction 261, 264–6 empirical research on 262–6 and learning 260 novelty 261, 264–6 renewal 261, 264–6 and co-evolutionary processes 259–60 actors’ role 266–8 and conceptualization of 17 and economic transformation 6–7 and endogenous/exogenous factors 227 and occurrence of 257, 275–6 and perceptions of 4 and types of 7–8 see also transformation Chesbrough, H 27, 42, 65 Christensen, C 125, 129, 159, 163, 210, 211 Clark, C 128 Clark, J 118 Clark, K 30, 203, 210, 215, 216 Cockburn, I M 203, 206 co-evolution 15, 19, 257 and boundaries of ﬁrms 158, 159 and change 259–60 and innovating economy 266–74 actors’ role 266–8

creation of knowledge and opportunity 270–2 economic transformation 273–4 innovation systems 272–3 nature of systems 273 system dynamics 273 uncertainty 268–70 and innovation process 50–1 cognition 20 Cohen, W 165, 202, 211, 212 Collier, J 109 communication, and understanding in common 149–51 Community Innovation Surveys (CIS) 263 competences, and boundaries of ﬁrms 159–60, 163, 166, 206–7, 209–12 competition, and drivers of 48 competitive advantage, and resources 63 complex products and systems (CoPS): and challenges in producing 81–2 and nature of 81 and production process 81 see also capital goods complex systems theory: and analysis of economic networks 107–9 and complexity of consumption and demand 114–17 and economic transformation 100 and evolutionary economics 14, 103 and orders of complexity 109–11 and product life cycles 113–14 complex transformation 1, 3 complexity: and adaptation 123, 144–5 and downstream system industry 159 and dynamic nature of economy 146 and economic development 122 and knowledge growth 144 and nature of 145 and novelty 144, 145 and role of individuals 145 computer industry 15, 157–8 and semiconductor industry: disintegration/specialization 161 vertical integration 160–1

285

Index computer industry (Contd ) and software: disintegration/specialization 161–2 vertical integration 161 and specialization/disintegration 157–8 and upstream component industry 159 and vertical integration 157–8 and vertical integration/specialization model: computer characteristics 167 demand for computers 170 market for components 167–8 simulations 175–93 technical progress 168–9 technological discontinuities 173–5 vertical integration/specialization 171–2 conformism, and consumers 103 conspicuous consumption 116 Constant, E 29 construction sector 33 consumers: and correlated preferences 106–7 and formation of preferences 106 and micro-meso-macro framework 112–14 and network connections 103–4 consumption: and complexity of 114–17, 118 and economic growth 104–6 and meso-rules 115–17 consumption networks, and economic transformation 102 consumption systems, and evolution of 107 contagion, and preference formation 106 Cool, K 62–3, 206 coordination, and innovative opportunities 60 Corbin, A 237 Cornwall, J 128 Cornwall, W 128 correlated preferences 106–7 Correlje´, A 251 Cowan, R 32, 147 craft knowledge, and engineering 43 creative destruction 127

286

D’Adderio, L 31, 32 Damasio, A 105 Daru, M 238, 241 Dasgupta, P 29 David, P 29, 32, 50 Davies, A 81, 82, 83, 86, 87, 88 De Swaan, A 237, 242 DeFillippi, R J 206 demand: and complexity of 114–17 and correlated preferences 106–7 and economic growth 104–6 and economic transformation 100, 117–18 and formation of preferences 106 and micro-meso-macro framework 112–14 and structure of 104 Dennett, D 105 dependencies: and innovation 50 and innovative opportunities 12–13, 48, 49, 51, 58–9, 66 division of innovative labour 68 impact on ﬂexibility/stability 67–8 mobilization of resources 61–3 perceived appropriability 63–6 perceptions of economic value 59–61 and systemic effects 51, 67 Depew, D J 121, 144 deskilling, and engineering 42, 43 destruction: and operationalizing idea of 275 and transformation 259, 261, 264–6 development: and adaptation 121, 122–3 and economic growth 123 and nature of 121 and open-ended nature of 122 Diamond, J 115 Dierickx, I 62–3, 206 discontinuity, and development 122–3 disintegration and specialization 157, 193–6 and computer industry 158 semiconductors 161 software 161–2

Index and computer industry model: computer characteristics 167 demand for computers 170 market for components 167–8 technical progress 168–9 technological discontinuities 173–5 vertical integration/specialization 171–2 and explanations of 203 and ﬁrm competences 159–60, 163, 166 and industrial economics perspective 159 and industry maturity 203 and technological change 202, 204 dynamic capabilities 209–12 dynamic transaction costs 212–14 modularity 214–17 in technologically stable conditions: behavioural factors 208–9 core/distinctive competences 206–7 transaction costs 207–8 and theoretical determinants of 164–7 and transaction costs 158–9, 203, 207–8 division of labour, and innovation 68 Docklands Light Rail system 76 DoCoMo 57, 58, 60 and i-mode 55 Dodgson, M 31, 32 Domberger, S 87 Dopfer, K 100, 107, 111, 146 Dosi, G 29, 206, 230 Downie, J 142 Downie function 142, 143 dynamic capabilities, and boundaries of ﬁrms 211–12 dynamic transaction costs, and boundaries of ﬁrms 212–14 Earl, P E 104, 105, 111 economic growth: and correlated preferences 106–7 and demand growth 104–6 and development 123 and differential growth rates 130–4, 138–9 economic networks, and analysis of 107–9

economic theory, and shortcomings of 105–6 economic transformation, see transformation economic value, see value, economic Edquist, C 50, 59 Eisenhardt, K M 212 Eliasson, G 64 Emmerson, Bob 33 engineering: and changing nature of 42, 43–4 recruitment and education 44 and collaborative 39 and craft knowledge 43 and design process 40, 41 and diversity of engineering services sector 32–3 and expansive disintegration 12, 27, 32 and experimentation 27, 28–9 and innovation process 28 and innovation technologies 12, 27, 28, 31–2 impact of 32, 40, 42–3 and nature of engineering knowledge 29 and problem-solving 29–30, 43 development of simulation tools 38– 41 modelling 30–1 modelling ﬁre evacuation 36–8 and professional accreditation 43 see also Arup; Ricardo Engineering entrepreneurship: and appropriation of value 64–5 and de-correlation of knowledge 152 and demand 113 and disruption 1, 7 and innovative opportunities 51 and resource mobilization 61–2 equilibrium: and economic transformation 126–7 and nature of 145 ¨ s, Paul 108 Erdo evolutionary economics 5, 6, 100 and adaptation, accounting for 135–40 and capitalism 127–8 and complex systems theory 14, 103

287

Index evolutionary economics (Contd ) and economic transformation 10 and logistic principle 140–4 and micro-meso-macro framework 111–14 and neo-Darwinism 103 and orders of complexity 109–11 and technological opportunities 51 evolutionary processes 1, 3, 19 evolutionary realism, and economic transformation 100 expansive disintegration, and engineering 12, 27, 32 experimentation 9, 18–19, 51 and engineering 27, 28–9 and innovation opportunities 12 and innovation technologies 12 and resource availability 63 and transition to sewer systems (Netherlands) 239–41 Fabricant, S 128–9, 134 Fagerberg, J 59 Ferguson, E 30 ﬁrms: and complexity 111 and design of research on 263–4 as economic actors 267 and innovative opportunities 51 and technological change 201–2 receptive capacity 202 see also boundaries of ﬁrms Fisher, R A 135, 139 Fisher/Price equation 139 Fisher’s fundamental theorem 139 ﬂexibility 1 and actor/system levels of analysis 4–5, 10–11, 260 and change 3–4 and experimentation 9 and markets 124 and research questions raised by 5 Flieth, B 104 Foray, D 32 Foster, J 100, 104, 109, 110, 114, 116, 125, 127

288

Foster, R 129 Frank, S A 139 Fransman, M 125 Freeman, C 7, 128, 261 Gaebler, T 76, 77 Galbraith, J R 88 Gambardella, A 210 Gann, D 30, 34, 82, 92 Geels, F W 229, 230, 251 Gehry, Frank 43 Georghiou, L 129 Geyer, A 86 Ghemawat, P 207 Gintis, H 139 Glaister, S 78 government, and economic development 267–8 Granstrand, O 212, 213 Grant, R M 86, 206 Grimsey, D 78 Grimshaw, D 78 growth theory: and economic transformation 99–100 and supply-side bias 101–2 Hamel, G 50, 206, 207 Hargadon, A 28, 30, 33 Hartley, K 79 Hass, H 117 Haut, T M 216 Henderson, K 30 Henderson, R 203, 206, 207, 210 Herﬁndahl indices 130–2 Hobday, M 81, 82, 86 Hodgson, G 103 Holland, J 109 Holme´n, M 48, 51, 52, 54, 61 Hood, C 76 Hoogma, R 231 Hooker, C A 109 Horan, B 136 Houwaart, E 237, 238 Hughes, T 50, 228, 231 Hume, David 100, 105

Index IBM: and specialization/disintegration 161–2 and vertical integration 160–1 idiosyncratic synergy, and learning within ﬁrms 208–9 i-mode, and innovative opportunities 57–8 appropriation 58, 65, 66 economic value 57, 60 resources 57–8, 63 individualism 145–6 individuals: and knowledge 146–7 and role in self-transformation 145 industrial dynamics 1, 2 and economic transformation 2, 3 and Schumpeter 6–7 and types of change 7–8 industrial economics, and vertical integration/specialization 159 inertia 1, 9, 18–19 and ﬁrm behaviour 158 and innovation opportunities 12 and innovation technologies 12 information: and comparison with knowledge 147 and economic transformation 2 and ﬂows of 149–51 information and communications technology (ICT), and ﬁrm boundaries 217–18 innovating economy, and co-evolutionary processes 266–74 actors’ role 266–8 creation of knowledge and opportunity 270–2 economic transformation 273–4 innovation systems 272–3 nature of systems 273 system dynamics 273 uncertainty 268–70 innovation: and actors’ role 50 and adaptive 260 and capitalism 126 and complexity of process 258

and creative 261 and ‘democratizing’ of process 42 and design of research on 262–6 and disruption 1, 7 and economic change 7 and economic transformation 10, 258 and ﬁrm boundaries 16 and incremental 7–8 and network connections 104 and niches 231 and novelty 261 and operationalizing idea of 263, 264–6 and process of 50 and public-private partnerships 13 and radical 7 and sectoral systems of 228 and systemic nature of 50–1 and transition process 232–4 niches 235–6 and types of 50 innovation systems, and innovating economy 272–3 innovation technologies 44 n1 and Arup 35 and engineering 12, 27, 28 and engineering practice 31–2, 42–3 and impact of 32 and Ricardo Software 40, 42 innovative opportunities 12–13, 48 and appropriation 49, 52, 56–7, 58, 63–6 and concept of 49–53 and coordination 60 and deﬁnition of 52 and dependencies 12–13, 48, 49, 51, 58– 9, 66 division of innovative labour 68 impact on ﬂexibility/stability 67–8 mobilization of resources 61–3 perceived appropriability 63–6 perceptions of economic value 59–61 systemic effects 67 and economic value 49, 52, 55–6, 57, 59–61 and mobile communications 53–5 3G (third generation) 55–7 appropriation 56–7, 58, 65, 66

289

Index innovative opportunities (Contd ) economic value 55–6, 57, 60 i-mode 57–8 resources 56, 57–8, 63 and multi-actor involvement 52–3 and perceptions of 53 and resources 49, 52, 56, 57–8, 61–3 and systemic factors 53 instability, and markets 126 institutional frameworks, and economic performance 6 Intel 161 interdependencies 13 Internet, and appropriation of value 64

and innovating economy 270–2 and markets 147 and nature of engineering 29 and new 122, 123 and orders of complexity 111 and research and development 148 and selection 147 and uncertainty 269–70 and understanding in common 149–51 Kor, Y Y 51 Kossman, E H 239, 242 Kricks, G 160 Kuemmerle, W 203 Kuznets, S 7, 129

Jacobides, M 160, 163 Jessop, B 76 Johns, Richard 40, 41 Johnson, B 32, 147 Jones, Charles 101 Jones, D T 83 Jubilee Line extension 76

Landes, D 115, 122 Lane, D A 269 Langlois, R N 16, 160, 162, 163, 164, 202, 203, 205, 208, 209, 215, 216 Lansink, Ad 245, 249 large technical systems (LTS) 228, 272 Lave, J 208 learning 20 and change 260 within ﬁrms 208 and uncertainty 268–70 Leiringer, R 75, 80 Leonard-Barton, D 50, 206, 212, 230 Levinthal, D A 202, 211, 212, 231 Lewis, M 78 Lindmark, S 54, 57, 58 Lippman, S A 206 Loasby, B J 20, 105, 111, 115, 268 logistic principle, and adaptation 140–4 Loorbach, D 244 Louca, F 128 Lucas, Bob 101

Kaserman, D L 205 Kauffman, S 108, 109, 110, 144 Kay, J 109 Kemp, R 229, 230, 231, 232 Keynes, John Maynard 101, 104, 110, 116 Keynesian Welfare State (KSW) 76 Killick, T 124 Kirman, A 107 Kirzner, I 51 Klepper, S 165 Knight, Frank 268 knowledge: and adaptation 14–15, 122, 123, 146 and blind generation 268–9 and capitalism 149, 151–2 and comparison with information 147 and correlation of 148–52 and de-correlation of 152 and economic transformation 2, 18 and given 121–2 and growth of 122, 144 and individual nature of 146–7 and information ﬂows 149–51

290

McCullogh, M 31 McDonald, S 129 McKelvey, M 59, 260 McKenna, P 35 Magnusson, M G 59 Mahoney, J T 51, 203, 206, 215, 216, 217 Maister, D 35

Index Major Project Association (MPA), and public-private partnerships 79–80 Malerba, F 157, 162, 165, 206, 228 March, J 65 Marchetti, C 141 marketing 103 markets: and adaptation 123–5, 140 and creative aspects of 124 and ﬂexibility 124 and knowledge 147 and need for instability 126 and regulation 124 Marshall, Alfred 6, 101, 127 Martin, J A 212 Marx, Karl 6 Maxﬁeld, R R 269 Mayntz, R 228 Mayr, E 136 Menger, Carl 6, 8 Metcalfe, J S 103, 127, 139 methodology, and research design 262–6 Michaud, P 75 Microsoft, and emergence of 162 Miller, R 81 Mitchell, W 207 mobile communications: and innovative opportunities 53–5 3G (third generation) 55–7 appropriation 56–7, 58, 65, 66 economic value 55–6, 57, 60 i-mode 57–8 resources 56, 57–8, 63 and technical standards 54–5 modelling: and engineering problem-solving 30–1 and evacuation of tall buildings 36–8 modernity, and capitalism 149 modularity, and boundaries of ﬁrms 214–17 Mokyr, J 122, 150 Molina, A 231 Morris, P 77, 80 Morrison, D 82 Mowery, D C 148, 161

multilevel perspective (MLP) and system change 227–8, 229–34, 249–51 and niches and radical innovation 231, 235–6, 250–1 and reproduction 234–5, 249 and rules 229–30 changes in 235, 248–9 and socio-technical landscape 231–2, 235 and socio-technical regimes 230 and stability 230–1 and technological regimes 229–30 and transformation 235, 249 and transformation of Dutch waste management (1960–2000) 243–9 impact of rule changes 248–9 and transition 235–6, 249 management of 251–3 process of 232–4 and transition to sewer systems (Netherlands) 236 experimentation and multiple niches (1870–90) 239–41 problem articulation (1840–70) 236–9 sewer systems (1890–1930) 241–3 Murmann, J P 260 Nakicenovic, N 141 National Audit Ofﬁce (UK), and publicprivate partnerships 78–9 Nelson, R R 21, 50, 64, 100, 149, 159, 163, 164, 169, 206, 208, 212, 229, 230, 258, 260 neoclassical economics, and shortcomings of 105–6 neo-Darwinism 100, 103 Netherlands: and management of transitions 251–3 and public-private partnerships 76 and transformation of waste management (1960–2000) 243–9 impact of rule changes 248–9 and transition to integrated sewer systems 236 experimentation and multiple niches (1870–90) 239–41

291

Index Netherlands: (Contd) problem articulation (1840–70) 236–9 sewer systems (1890–1930) 241–3 Network Rail 90 network theory, and analysis of economic networks 107–9 networks: and consumers 103–4 and economic system 102 and economic transformation 2 see also social networks New Labour, and public-private partnerships 75–6 niches: and radical innovation 231, 250–1 and transition 235–6 and transition to sewer systems (Netherlands) 239–41 Nightingale, P 29 North, D 150 novelty 122 and adaptation 126 and complexity 144, 145 and degree of 261 and demand growth 104–6 and economic evolution 111 and innovation 261 and operationalizing idea of 275 and transformation 114–15, 259, 261, 264–6 opportunity: and identiﬁcation of 20 and innovating economy 270–2 see also innovative opportunities O’Really, C A 212 Organization for Economic Cooperation and Development (OECD) 43, 76 organizational forms, and economic transformation 1–2 Osborne, D 76, 77 Osborne, S 75, 76, 78 Owen, D 88 Pacey, S 43 Pahl, G 30

292

Palmberg, C 51 Pandian, J R 206 Parker, D 79 Parto, S 245 Pasinetti, L L 128 Patel, P 212 Pavitt, K 29, 82, 212, 262 Penrose, E 51, 86, 206 perception 20 and innovative opportunities 53, 67 appropriation 63–6 economic value 59–61 resources 61–3 and uncertainty 53 Perez, C 7 Perlow, L A 29, 30, 34 Peteraf, M A 206 Pisano, G 86, 159, 163, 207 Polanyi, M 147 population method, and accounting for evolutionary adaptation 135–40 Porter, M E 82, 207 Potts, J 100, 103, 104, 105, 107, 108, 109, 111, 114, 146 Pound’s principle 20 Prahalad, C K 50, 206, 207 preferences: and correlated preferences 106–7 and formation of 106 Prencipe, A 82, 86, 88, 217 Price, G R 139 prices: as coordinating device 106–7 and determination of 104 Private Finance Initiatives (PFI) 77 see also public-private partnerships problem-solving, and engineering 29–30, 43 modelling 30–1, 36–8 simulation tools 38–42 process innovation 50 product innovation 50 product life cycles 106, 113–14 productive opportunities 51 public health, see waste management (Netherlands)

Index public policy: and economic development 267–8 and management of transitions 251–2 Netherlands 252–3 public-private partnerships 13, 73 and assessment of 78–80, 91–3 difﬁculties in analysing 80 and capital goods industry: acquiring integration capabilities 86–8 repeatable solutions 88 systems integration 83–6 value stream 82–6 and changing nature of the state 73–4 and complex products and systems (CoPS): challenges in producing 81–2 nature of 81 production process 81 and controversy over 74 and deﬁnition of 77 and expertise beneﬁt of 78 and growth of 75, 78 and historical precedents 75 and impact of 91–3 on ﬁrms 80–1, 89–90 on government 90–1 and innovation 74, 77, 79–80 and life-cycle beneﬁts of 77–8 and market-based solutions 76 and New Labour’s attitude 75–6 and poor performance of governmentled projects 76 and rationale for 77 and value of contracts 78 and widespread adoption of 76–7 railway industry 85, 88, 90 receptive capacity, and technological change 202 Reed, R 206 regimes 229–30 and reproduction 234–5 regulation: and impact on ArupFire 35–6 and markets 124 Reid, D 238

Reinertsen, D G 216 renewal: and operationalizing idea of 275 and transformation 259, 261, 264–6 repeatable solutions, and public-private partnerships 88 reproduction, and system change 16–17, 228, 234–5, 249 research, empirical, and character of change 262–6 research and development 148 resistance 1 resource-based perspective (RBP) 206, 208 resources: and changing patterns of allocation 122 and experimentation 63 and innovative opportunities 49, 52 mobile communications 56 mobilization of resources 61–3 Ricardo Engineering 12, 28 and characteristics of 38 and Ricardo Software 32 development of 38 innovation technologies 40, 42 simulation tools for automotive industry 38–42 Richardson, G B 144, 147 Rip, A 229, 230, 232 Robertson, P L 160, 162, 163, 164, 201, 202, 203, 205, 206, 208, 209, 215, 216 Robinson, J V 145 Romer, Paul 100 Rosenberg, N 50, 86, 148, 231, 234–5, 269 Rosenbloom, R 65, 159, 163 Rosenkopf, L 210 Roth, K 203, 207 routines 5–6 Rubin, P 105 rules: and changes in 235, 248–9 and technological regimes 229–30 Rumelt, R P 206, 207 Rush, H 81 Salter, A 30, 34, 82, 92 Sampat, B N 149 Sanchez, R 203, 215, 216, 217

293

Index Sanz-Velasco, S A 59 Say, Jean Baptiste 101 Say’s Law 104 Scherer, F M 51 Schilling, M A 215 Schneider, E 109 Schoemaker, P 206, 207 Schot, J W 229, 231 Schrage, M 31, 32 Schumpeter, Joseph 6–7, 48, 59, 101, 121, 125, 127, 129, 227, 260 Schumpeterian Workfare State (SWS) 76 sectoral systems of innovation 228 selection 103 and adaptation 127–8, 145 and knowledge 147 self-organization, and economic transformation 2 self-transformation: and adaptation 125–6, 129 and role of individuals 145 semiconductor industry 15, 158 and downstream system industry 159 and specialization 161 and upstream component industry 159 and vertical integration 160–1 see also computer industry Shackle, George 112, 127, 152 Shane, S 51 Siemens 85 Silver, M 208 Simon, H A 29, 30 simulation: and automotive industry 38–42 and modelling ﬁre evacuation 36–8 Slywotzky, A 82 Smith, Adam 6, 105, 115, 134, 149, 157, 159, 216 Smith, Andrew 76, 77 Smith, P G 216 social groups, and system change 229, 230, 233, 249, 250 social networks: and systems change 228 and transformation 235 socio-technical systems 229

294

and changes in 16–17 and management of transitions 251–2 Netherlands 252–3 and multilevel perspective on change in 229–34, 249–51 reproduction 234–5, 249 social groups 229, 230, 233, 249, 250 transformation 235, 249 transition 232–4, 235–6, 249 and niches and radical innovation 231, 250–1 transition process 235–6 and rules 229–30 changes in 235, 248–9 and socio-technical landscape 231–2, 235 and socio-technical regimes 230 and stability of 230–1 and technological regimes 229–30 and transformation of Dutch waste management (1960–2000) 243–9 impact of rule changes 248–9 and transition to sewer systems (Netherlands) 236 experimentation and multiple niches (1870–90) 239–41 problem articulation (1840–70) 236–9 sewer systems (1890–1930) 241–3 Soete, L 262 software industry: and specialization 161–2 and vertical integration 161 see also computer industry Solow, Bob 101 special purpose vehicles (SPVs), and capital goods industry 83, 89 specialization, see disintegration and specialization stability 1 and actor/system levels of analysis 4–5, 10–11, 260 and change 3–4 and inertia 9 and research questions raised by 5 Stalk, G 216 Stankiewicz, R 228

Index Steensma, H K 215 Steinmuller, E 32 Stern, S 206 Stigler, G 159 Strathern, M 113 Strogatz, S 108 structural change 1 and accounting for evolutionary adaptation 135–40 and differential growth rates 130–4, 138–9 and economic transformation 2 and evidence for 128–34 and logistic principle 140–4 Sutton, R 28 swarming 271 systems: and actors 4 and conceptualizing ﬂexibility/stability 4–5, 10–11 and economic transformation 2 and innovating economy 273 and innovation 50–1 and large technical systems (LTS) 228 and management of transitions 251–2 Netherlands 252–3 and orders of complexity 109–11 and seamless web approach to 228, 229 and sectoral systems of innovation 228 and technological systems approach 228 see also multilevel perspective (MLP) and system change; socio-technical systems systems integration: and capital goods industry 83–6 and repositioning in value stream 86–8 Tarr, J A 238 technical standards, and mobile communications 54–5 technological change: and ﬁrm boundaries 201–2, 217–19 continued relevance of vertical integration 203–4 dynamic capabilities 209–12

dynamic transaction costs 212–14 indeterminate effects on 204 modularity 214–17 simultaneous outsourcing/vertical integration 204 and receptive capacity of ﬁrms 202 and transformation 10 technological regimes 229–30 and reproduction 234–5 technological systems approach 228 Teece, D 86, 159, 163, 206, 210, 212 Texas Instruments 160 third generation, see 3G (third generation) Thomke, S 28, 32, 203 transaction costs, and vertical integration/ specialization 158–9, 203, 207–8, 212–14 transformation 1, 2, 106–7 and assumptions about 1–2 and co-evolutionary processes: actors’ role 266–8 creation of knowledge and opportunity 270–2 economic transformation 273–4 innovation systems 272–3 nature of systems 273 system dynamics 273 uncertainty 268–70 and complexity of consumption and demand 114–17 and concept of 2–3 and consumption networks 102 and demand perspective 100, 117–18 structure of demand 104 and differential growth rates 130–4, 138–9 as dynamical process 99 and evolutionary adaptation, accounting for 135–40 and evolutionary models 102 as ‘far from equilibrium’ process 126–7 and growth theory 99–100 and innovating economy 273–4 and innovation 10, 258 and learning 260 and logistic principle 140–4

295

Index transformation (Contd ) and micro-meso-macro framework 111–14 and neo-Darwinian explanations 100, 103 and open-ended nature of 126 and Schumpeter 6–7 and structural adaptation, evidence for 128–34 and supply-side bias 101–2, 117 and system change 17, 228, 235, 249 and types of change 7–8 and waste management (Netherlands 1960–2000) 243–9 impact of rule changes 248–9 see also change transition: and adaptation 126 and management of 251–2 Netherlands 251–3 and sewer systems (Netherlands) 236 experimentation and multiple niches (1870–90) 239–41 problem articulation (1840–70) 236–9 sewer systems (1890–1930) 241–3 and system change 17, 228, 235–6, 249 niches 235–6 process of 232–4 Tuomi, I 32 Turkey, and public-private partnerships 75 Tushman, M L 209–10, 211, 212, 213 uncertainty: and demand 104 and innovating economy 268–70 and perception 53 understanding, and knowledge 149–51 United States, and public-private partnerships 76–7 Unruh, G C 231 Utterback, J 125, 129 Valentine, J 78 value, economic: and appropriation of 64–5 and creation of 65

296

and innovative opportunities 49, 52 mobile communications 55–6, 57, 60 perceptions of economic value 59–61 value stream, and capital goods industry 82–6 repositioning in 86–8 Van de Poel, I 235 Van den Akker, J 238, 243 Van den Noort, J 241 Van Zon, H 236, 237, 238, 239, 240, 241 Vanberg, V 111 variety, and network connections 103 Veblen, Thorstein 6, 113, 116 Verbong, G P J 251 Verdoorn, J A 238 Verona, G 206 vertical integration 157, 193–6 and computer industry 157–8 semiconductors 160–1 software 161 and computer industry model: computer characteristics 167 demand for computers 170 market for components 167–8 simulations 175–93 technical progress 168–9 technological discontinuities 173–5 vertical integration/specialization 171–2 and continued relevance of 203–4 and explanations of 202–3 and ﬁrm competences 159–60, 163, 166, 207 and industrial economics perspective 159 and industry maturity 203 and technological change 202, 204, 217–18 dynamic capabilities 209–12 dynamic transaction costs 212–14 modularity 214–17 in technologically stable conditions: behavioural factors 208–9 core/distinctive competences 206–7 transaction costs 207–8

Index and theoretical determinants of 163–4, 166–7 and transaction costs 158–9, 203, 207–8 Vincenti, W 29, 31 Vis, G N M 240 Von Hippel, E 27, 29, 42, 43, 44 Wagge, Jeremy 80 Wakeford, J 78 Walker, W 125, 230 Walsh, D M 127–8 Waste Management Council 246, 247–8 waste management (Netherlands): and transformation of (1960–2000) 243–9 impact of rule changes 248–9 and transition to sewer systems 236 experimentation and multiple niches (1870–90) 239–41 problem articulation (1840–70) 236–9 sewer systems (1890–1930) 241–3 Watts, D 107, 108

Weber, B H 121, 144 Weber, Max 6 Wenger, E 208 White, L 122 Whyte, J 31 Wiley, E O 109 Williams, R 12, 27, 30 Williamson, O 79, 158, 202, 205, 207, 212 Winter, S G 21, 90, 100, 159, 160, 163, 164, 165, 169, 206, 208, 212, 229, 230, 258 Wise, R 82 Witt, U 62, 112, 127, 145 Womack, J P 83 Woodward, J 81 WS Atkins 87, 89 Yeoh, P L 203, 207 Young, Allyn 6, 68, 147 Zitron, J 75, 77, 78, 79–80 Zollo, M 90

297