RECOVERING FROM SUCCESS
This page intentionally left blank
RECOVERING FROM SUCCESS Innovation and Technology Manage...
242 downloads
2016 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
RECOVERING FROM SUCCESS
This page intentionally left blank
RECOVERING FROM SUCCESS Innovation and Technology Management in Japan Edited by D. HUGH WHITTAKER Doshisha University
ROBERT E. COLE Doshisha University/UC Berkeley
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 authors 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 Recovering form success: innovation and technology management in Japan/edited by D. Hugh Whittaker, Robert E. Cole. p. cm. Includes bibliographical references and index. 1. Technological innovations–Japan–Management. 2. Industrial management–Japan. I. Whittaker, D. H. (D. Hugh) II. Cole, Robert E. HD70. J3R435 2006 338. 0640952–dc22 2006007846 Typeset by SPI Publisher Services, Pondicherry, India Printed in Great Britain on acid-free paper by Biddles Ltd., King’s Lynn, Norfolk ISBN 0–19–929731–2 978–0–19–929731–3 ISBN 0–19–929732–0 (Pbk.) 978–0–19–929732–0 (Pbk.) 1 3 5 7 9 10 8 6 4 2
Contents
List of figures and tables List of contributors Acknowledgements 1
Introduction Robert E. Cole and D. Hugh Whittaker
Part 1: 2
Industries, technologies, and value chains
The telecommunication industry: A turnaround in Japan’s global presence Robert E. Cole
3
Modular production’s impact on Japan’s electronics industry Timothy J. Sturgeon
4
Technology management and competitiveness in the Japanese semiconductor industry Takashi Yunogami
5
Global value chains in the pharmaceutical industry Jocelyn Probert
6
Software’s hidden challenges Robert E. Cole
Part 2: 7
MOT in and between enterprises
The open innovation model: Implications for innovation in Japan Henry W. Chesbrough
vii ix xi 1
29
31 47
70 87 105
127
129
8
Managing creativity and control of knowledge workers Clair Brown
145
9
Rethinking innovation Eiichi Yamaguchi
166
v
Contents 10
‘Microbursts’ of knowledge and creative work in Japan Philippe Byosiere
184
11
Hitachi’s nascent ‘new production(ist)’ system D. H. Whittaker
199
12
Interfirm networks and the management of technology and innovation in Japan James R. Lincoln
Part 3:
Transforming Japan’s innovation system
215
235
13
Innovation policy for Japan in a new era Tateo Arimoto
237
14
Security and techno-systems: A comparative analysis Yuzo Murayama
255
15
Technology management training in Japan: Government initiatives and their effects Atsushi Kaneko, Yoshi-fumi Nakata and Muneaki Yokoyama
271
Electronic government in Japan: Towards harmony between technology solutions and administrative systems Toshiro Kita
286
16
17
Conclusions and reflections: Emergent models D. Hugh Whittaker and Robert E. Cole
Index
vi
298 321
List of figures and tables
Figures 1.1 2.1 4.1 4.2 4.3 4.4 4.5 4.6 4.7 5.1 6.1 6.2 6.3 9.1 9.2 9.3 9.4 9.5 10.1 10.2 11.1 11.2 13.1 13.2 13.3 13.4 13.5
Evolution of the semiconductor industry Japan’s trade balance in telecommunications equipment Changing DRAM share by country Japanese computer shipments Three phases of wafer process technology Superior etching technology Technology and quality Quality and yield Two evaluation axes of technology Value chain fragmentation options International comparison of rate of introduction of IT-related applications Proportion of Japanese and US companies that found their IT investment effective Measures by Japanese and US companies to optimize effectiveness of IT investment Paradigm disruptive innovation and performance disruptive innovation Innovation process for the blue light emitting diode The number of published papers vs. the number of researchers with PhDs in 1999 Trends in academic papers by year Academic papers and company value Decline in R&D efficiency in manufacturing industry Degree of satisfaction with treatment by researchers ‘Inspire A’ businesses (Stage 2) Monozukuri, MOT, and corporate management (Stage 3) Projected population decrease in Japan Projected decrease of researchers Convergence of disciplines The pipeline and the tree: a new framework for training and career development in the natural sciences Universities, society, and S&T in the 21st century
12 32 71 73 75 75 77 79 81 88 113 115 121 169 173 177 178 179 190 192 207 210 240 241 242 244 246
vii
List of figures and tables 13.6 16.1 16.2 16.3 17.1
Innovation system for Japan after catch-up Configuration of the Juki-net Configuration of the Juki-card ‘C2G2B’ service model Dual innovation system
250 288 289 294 304
Tables 1.1 ICT equipment exports, 1990–2000 1.2 LCD firm’s average profitability by national site, Q1 2004 3.1 Net income (loss) of the ten largest Japanese electronics firms, 1996–2004 3.2 Examples of recent and planned electronics factory investments in Japan 3.3 Examples of recent restructuring in the Japanese electronics industry 5.1 Changes in the domestic landscape, 1992–2002: An increasing Western presence 5.2 Selected consolidation moves in the Japanese pharmaceutical market 6.1 Ways of building information systems 8.1 Time spent working in a team and independently 8.2 Training by current employer 8.3 Importance of sources of technical information 8.4 Channels of technical information acquisition from other semiconductor companies 8.5 Sources of knowledge to solve a specific technical problem 11.1 Corporate senior staff (2001–03) (Stage1) 12.1 Denso’s share of Toyota’s total inputs of selected electronic parts by year 15.1 Management skills for MOT 15.2 Types and characteristics of MOT programmes in the US and Europe 15.3 Types and characteristics of MOT programmes in Japan 17.1 Open and closed innovation orientations
viii
4 7 54 57 59 96 97 112 151 151 153 155 156 203 226 275 277 283 304
List of contributors
Tateo Arimoto is Executive Research Fellow at the Economic and Social Research Institute, Cabinet Office, Japanese Government, and Visiting Professor at the Tokyo University of Science. Clair Brown is Professor of Economics and the head of the Center for Work, Technology, and Society at the Institute of Industrial Relations, University of California, Berkeley, as well as Omron Fellow and Visiting Professor of Graduate School of Policy and Management at Doshisha University, Kyoto. Philippe Byosiere is Professor at Doshisha Business School and Faculty Research Fellow at the Institute for Technology, Enterprise and Competitiveness (ITEC), Doshisha University, Kyoto. Henry Chesbrough is Lecturer and Executive Director at the Center for Open Innovation, Institute of Management, Innovation & Organization and Management of Technology Program at the University of California, Berkeley. Robert E. Cole is Omron Distinguished Professor, Doshisha Business School, Kyoto, Professor Emeritus, Haas School of Business and Department of Sociology, as well as Co-Director of Management of Technology Program, Haas School of Business and College of Engineering at the University of California, Berkeley. Atsushi Kaneko is Chief Consultant of the MOT Project Office and General Manager of the Learning Innovation Department at Mitsubishi Research Institute, Inc, Tokyo. Toshiro Kita is Professor at Doshisha Business School and Faculty Research Fellow at the Institute for Technology, Enterprise and Competitiveness (ITEC), Doshisha University, Kyoto. James Lincoln is Warren E. and Carol Spieker Professor at the Haas School of Business, University of California, Berkeley, and Omron Fellow and Visiting Professor of the Graduate School of Policy and Management at Doshisha University, Kyoto. Yuzo Murayama is Professor at Doshisha Business School and Faculty Research Fellow at the Institute for Technology, Enterprise and Competitiveness (ITEC), Doshisha University, Kyoto.
ix
List of contributors Yoshi-fumi Nakata is Professor at Doshisha Business School and Graduate School of Policy and Management, and 21st Century Centre of Excellence Programme Leader at the Institute for Technology, Enterprise and Competitiveness (ITEC), Doshisha University, Kyoto. Jocelyn Probert is Lecturer at Birmingham Business School, and Visiting Research Fellow at the Institute for Technology, Enterprise and Competitiveness (ITEC), Doshisha University, Kyoto. Timothy Sturgeon is Senior Research Affiliate at the Industrial Performance Centre, MIT, and Visiting Research Fellow at the Institute for Technology, Enterprise and Competitiveness (ITEC), Doshisha University, Kyoto. D. Hugh Whittaker is Professor at Doshisha Business School and Director of the Institute for Technology, Enterprise and Competitiveness (ITEC), Doshisha University, Kyoto. Eiichi Yamaguchi is Professor at Doshisha Business School, Deputy Director of the Institute for Technology, Enterprise and Competitiveness (ITEC), Doshisha University, Kyoto, and Executive Vice President of Powdec KK. Muneaki Yokoyama is Consultant of the MOT Project Office at the Learning Innovation Department at Mitsubishi Research Institute, Inc., Tokyo. Takashi Yunogami is Research Fellow of the Institute for Technology, Enterprise and Competitiveness (ITEC), Doshisha University, Kyoto, and Visiting Professor at Nagaoka University of Technology, Niigata.
x
Acknowledgements
This book grew out of our experiences in setting up a new business school, a new research institute, and developing executive education programmes in technology management (MOT) at Doshisha University, Kyoto. On the one hand, these experiences gave us insights into the changing institutional framework of innovation in Japan, especially the role of universities, and the challenges currently facing Japanese technology companies. One the other hand, they forced us to question our assumptions about business education, the goals and nature of technology management, and a host of related issues such as labour mobility and the encouragement of spin-offs, which were to some extent influenced by practices and debates in the US and Europe. Consequently we sought contributors who were involved in the emerging MOT and innovation framework and/or who held perspectives which might shed light on it. We are most deeply indebted to the participants of MOT seminars, especially TBI (Technological Business Innovation) at Keihanna, who provided a ‘reality check’ for our ideas and contributed ideas of their own. Ronald Dore did likewise, and we appreciate his support, as well as that of ITEC COE Project Leader Yoshi-fumi Nakata. We are grateful, too, to Miles Dodd, Tom Cole, and the staff of ITEC, especially Makiko Kawasaki, for their help in preparing the manuscript. Ministry of Education, Culture, Sports, Science and Technology (MEXT) financial support through the 21st Century Centre of Excellence Programme is gratefully acknowledged, as is support from the New Energy and Industrial Technology Development Organization (NEDO), and from the Omron Corporation benefaction, which helped to launch many of the initiatives at Doshisha University and funded Robert Cole’s Chair. Finally, we would especially like to thank our colleague, Philippe Byosiere, who came up with the original idea to have Doshisha faculty do a book on management of technology and helped us at various points to accomplish the project.
xi
This page intentionally left blank
1 Introduction Robert E. Cole and D. Hugh Whittaker
Japan is a prolific importer and creator of acronyms. Around 2002 a new acronym spread throughout industry and the business media: MOT–management of technology–was heralded as a cure to Japan’s ‘lost decade’ tribulations and was seen as a means of rehabilitating the Japanese economy to face the 21st century. That the meaning of MOT was vague only served to heighten interest and debate. Some observers reacted with skepticism, dismissing MOT as another here today–gone tomorrow fad, or as a US import which would not work in Japan. Others, however, felt that it gave expression and conceptual coherence to emergent changes which had taken place during the ‘lost decade’, and that its transformative potential should not be under-estimated. We have chosen to edit this book which is broadly about MOT because a) it offers an excellent prism or vantage point to observe challenges to Japan’s previously successful innovation model and how that model is undergoing change; b) even fads can have lasting and unpredicted consequences, as happened in the quality movement in the US in the 1980s and early 1990s; and c) MOT has been used as a lever to try to engineer changes to Japan’s innovation system, and these attempts and the vision behind them are noteworthy. This book is not, however, a systematic exposition of MOT practices in Japan. ‘Innovation’ in the title signals our broader interest in the Japanese macro-level innovation system, as well as corporate level innovation and technology management. We see the hitherto successful Japanese model of innovation facing new challenges from the mid-1990s. The first challenge is that other countries began to emulate aspects of the successful Japanese model and were able to shrink the productivity and quality gap. The second is that new competitive models undermined some of Japan’s advantages in a number of industries, notably information and communication technology (ICT). These two challenges, in turn, exposed previously hidden weaknesses in the Japanese innovation system and corporate level technology management. Thus, there are the two major sets of (external) challenges which we highlight through our MOT prism; emulation and new business models.1 We ask how the
1
Recovering from success challenges were perceived, the responses, and the significance of the responses. Are Japan’s technology champions destined to adopt the US model(s), either partially or substantially, or will they plot a new path, or indeed decline without effectively addressing the challenges? Is the national innovation system being reshaped, with higher education incorporated, for instance, as has been the case in the US? Are the levers of industrial policy changing, and if so, how? In terms of sectoral coverage, we are necessarily selective. We focus on the industries most directly challenged by the resurgent US. We do not cover, on the one hand, the highly successful automobile industry or emerging strengths like functional materials or, on the other hand, industries in the sheltered sector, which do not face open, global competition. Our focus is rather on the ICT sector defined broadly to include electronics as well as pharmaceuticals. This selective approach may annoy some readers, but a systematic exposition of MOT in a broad range of industries would mean a fundamentally different book. Instead, we offer pithy ‘snapshot’ chapters which illuminate aspects of the challenge and response–or would-be response–theme we outlined above. In doing so, we seek to move beyond the ‘lost decade,’ which has been the preoccupation of much writing about Japan in recent years, to consider the potential for corporate and economic renewal in the early 21st century. There are recent signs of renewal, but how far will they take Japan? Can one expect another role reversal with the US, or has the global competitive environment changed too much relative to Japan’s capabilities? By way of clarification, technology management is a comprehensive approach to how firms can capture value from technology, involving not just technological capabilities, but business and organizational factors as well. In the broadest sense of MOT, firms are supported or supplemented in their attempts to capture value by a range of other actors and institutions, including universities, research institutes, government agencies, and venture capital firms. These all fall within the scope of interest of this book. Different combinations of technological capabilities, business models, and organizational architectures and skill sets are required for success in different industries. The remainder of the Introduction examines the emulation and new business model challenges, and the MOT boom, or ‘fad,’ as a response. Having set the scene, we then provide a brief synopsis of the chapters which follow. Our assessment is left to the Conclusions.
The challenges The ‘Japanese model’ which so dazzled the world in the 1980s fell into disrepute in the 1990s. The model was constructed on a succession of manufacturing industries which became fiercely competitive on the world stage and
2
Introduction demonstrated rising levels of technological sophistication as well as innovative practices, which combined efficiency and continuous improvement in cost competitiveness and quality.2 This productionist model was based on intense exchange and circulation of design information throughout the firm along with decentralized decision making at the workshop level (Aoki 1988; Fujimoto 2004). Innovations in human resource management and subcontracting relations, not to mention patient capital, loaned further support to the productionist system. Scholars fiercely debated the significance of government policy, especially industrial policy, but it was generally considered constructive though exaggerated in some quarters. Odagiri and Goto (1993) analysed the key factors of the productionist system that brought great success to the Japanese in the 1980s. They stressed the importance of Japanese managers investing heavily in large R&D efforts, into making products efficiently, and incorporating the results swiftly and smoothly into manufacturing and marketing. They saw four factors as lying at the heart of Japan’s innovation system: 1 2 3 4
Bias to growth maximization (willingness to invest in technology) Familiarity of management with research, production, and marketing Close R&D–production–sales links Smooth transfer of new processes and products into production (Odagiri and Goto 1993: 109).
All four factors are in a narrow sense about the management of technology, though they are supported by a range of other factors, such as those mentioned above. The argument for the importance of Japanese management’s familiarity with research, production, and marketing was based on the fact that the largest proportion of company directors (who were full-time) came from production and technology departments. This was in contrast to the situation in the US and the UK where many directors are part-time outsiders, with a significant proportion of directors having an accounting and finance background. The backgrounds of Japanese managers and company directors were seen as enabling them to be particularly effective in managing technology (including being better able to evaluate outcomes from R&D and to having more favourable views of what R&D can contribute). However, Odagiri and Goto recognized education as a contributor to Japan’s system of innovation, primarily in terms of the education system’s ability to provide firms with well educated individuals who had generalized technical skills rather than in terms of its contribution to research. The university share of national R&D expenditure had declined, from 20 percent in 1978 to just 11.6 percent in 1990, and within this figure the proportion spent on basic R&D declined from 57.3 percent to 52.9 percent (Lee 1997).3 R&D expenditure per researcher at universities was less than half that in companies (Odagiri and
3
Recovering from success Goto 1993: 111). Such under-funding of Japanese educational R&D limited its contribution to economic growth. A depiction of the US innovation system in the same volume provides a stark contrast. Mowery and Rosenberg (1993) characterize core elements of the postwar US system as publicly funded (federal) R&D, representing two-thirds of basic research expenditure; growth of academic research; commercialization of technology through new firms; growing research collaboration between industry and universities; increasing exploitation of external sources of technology; and strengthening of enforcement of intellectual property. We shall explore later the extent to which Japanese leaders are seeking to adopt some elements of this model. Despite–or in some cases because of–its formidable strengths, Japan’s productionist system ran into trouble in the 1990s. The recession of the early 1990s could be ascribed at first to fallout from the bursting of the country’s massive asset bubble in 1990, and subsequently to woes in the financial sector. By the second half of the decade, however, problems were emerging in hitherto competitive manufacturing industries. Most immediately there were falling profits, declining market share, and in some major cases, especially among the large electronic makers, lots of red ink. From top position in 1989–93, Japan began to slide rapidly down IMD’s World Competitiveness Survey rankings. This was not due to a lack of R&D investment. On many key R&D indicators, such as per capita R&D expenditure, industrial R&D costs, and number of R&D personnel and patents issued, Japan was still ranked at or near the top. Yet many companies were having increasing difficulties in capturing the benefits of this investment. Table 1.1, for instance, shows the weak Japanese growth in key ICT sectors in 1995–2000 compared to those of the EU and the US. Reasons behind the difficulties were varied, some of them nation, industry, and firm specific.4 Many of them will be treated in subsequent chapters; here we focus on two of the largest drivers–emulation of Japanese best practices of Table 1.1 ICT equipment exports, 1990–2000 (value in millions of current US dollars and growth in percentages) Exports
1990
1995
2000
Average Annual Growth (2000/95)
Computer equipment
US Japan EU
23,005 18,584 40,119
34,476 29,521 66,460
54,685 27,558 94,131
9.7 1.4 7.2
Communication equipment
US Japan EU
4,063 5,614 9,541
10,933 6,904 26,440
20,680 8,106 69,179
13.6 3.3 21.2
Electronic component
US Japan EU
13,826 14,678 16,330
27,668 43,270 36,393
70,001 50,348 55,972
20.4 3.1 9.0
Source : OECD, ITS database, January 2002 (cited in Nezu 2002: 12)
4
Introduction the 1980s by Japan’s foreign competitors in the 1990s, and the transformation of global competition from being among large vertically integrated firms in many ICT industries to competition among networks of specialist firms. Examination of why the large Japanese producers had so much difficulty responding to these challenges leads to a discussion of adjustment difficulties and ‘large firm malaise’. Failure to respond strongly to these challenges in the context of falling profits, declining market share with still high R&D investment in technology driven industries led to increasing self-doubt among Japanese corporate leaders and provided fertile ground for the emerging interest in MOT.
The emulation challenge Since the mid- to late 1980s, there has been widespread diffusion of critical elements in the Japanese design/manufacturing model sketched above. Some of the key elements that have diffused include practices making for shorter product development cycles; the importance of managing the handoff from design to manufacturing; management of suppliers; standardized work processes; continuous improvement; and an emphasis on quality. Asian tigers, Korea and Taiwan, have largely mastered these approaches in key sectors and, significantly for the future, Chinese firms in the PRC are in the process of mastering them; their progress has been strikingly faster than many Japanese firms anticipated. Not only have many of Japan’s new competitors emulated the Japanese model, but sometimes they have dramatically improved on it in ways that the Japanese have found difficult to follow (e.g., the Taiwanese development of the foundry model in semiconductors coupled with ‘fabless’ semiconductor designers around the world). Japanese capital spending was low for much of the 1990s. Without strong investment in new equipment and factories, many companies didn’t go far beyond their 1980s capabilities. The average age of manufacturing equipment in 2003 stood at 12 years (9.3 in 1991), compared to 7.9 for the United States (Nikkei Weekly, 15 Sept. 2003), and even lower in key growth industries in countries like Korea that have invested heavily in capital equipment. Korean manufacturers like Samsung have been particularly aggressive. It is not only Japan’s new East Asian competitors that have mastered or improved on the Japanese 1980s success formulas, but also its traditional competitors from the US and Europe have made giant strides in improving the quality of their products, optimizing supply chains (not just supplier relationships), reducing time to market, capitalizing on their skill in global standard setting (e.g., telecommunications), and so on. Nowhere is the progress of US firms more evident than in the global market for PCs. Dell, in particular, pioneered the build-to-order model for computers. The top three global market shares for PCs in 2003, totaling 39 percent, were held by the
5
Recovering from success American firms Dell, Hewlett-Packard, and IBM, with NEC standing at 3 percent and Fujitsu-Siemens at 4 percent (Nikkei Weekly, 2 Aug. 2004). Moreover, contract manufacturers (EMS firms) like Flextronics and Solectron have emerged which, by virtue of their specialization in manufacturing, also have been able to master the model. A key sector in which the Japanese lead in quality and R&D cycle times has been eliminated is in semiconductors (Leachman and Hodges 1996; Macher et al. 1998: 113–18; Leachman 2002). Intel’s ‘copy exactly’ strategy has been very effective for a firm with a limited product line focused in shrinking Wintel architecture (Microsoft software coupled with Intel chips) at regular time intervals. In the case of the foundries (contract fabrication plants) with a greater product mix a different solution was required. The Taiwanese creative solution was to integrate R&D into manufacturing. Thus, R&D tends to be done concurrently with manufacturing. To do so, the foundries invested heavily in upgrading their manufacturing engineering human capital. Korea’s Samsung was also effective in reducing R&D cycle times and co-locating design and manufacturing facilities. In both the semiconductor and computer cases, the improvement of the competitive position of Western competitors rested very much on the embrace of modular production enabling Western firms to take advantage of low cost Asian suppliers. One sector where US producers have made only modest incursions on Japanese performance advantages is the automotive sector. US firms reduced the quality differential in the 1980s, and the average performance differences between US and Japanese producers had narrowed on many measures by the mid-1990s. Since the late 1990s however the gap seems to be increasing once again between Toyota’s performance (e.g., product development lead time) and its American competitors (Fujimoto 1999: 220–1).5 As we shall see, however, the automobile industry is characterized by product architectures and organizational practices that distinguish it from many of the ICT sector firms. Moreover, not all Japanese auto manufacturers are doing so well. Japanese ICT firms have found it increasingly difficult to protect their intellectual property from other East Asian competitors. To take advantage of lower costs, they have somewhat reluctantly located many of their plants in China and other Asian locations. That has, however, made it easier for competitors to access the renowned Japanese process technology through hiring away skilled workers from these facilities. In key technologies such as LCD panels, moreover, the Koreans and the Taiwanese have been able to purchase the same advanced production machinery from Japanese and other equipment vendors. While the Japanese long-term employment system still makes it difficult for East Asian competitors to poach key technical personnel, some of those laid off in the late 1990s restructuring and retirees, including key managers, have found new employment with Korean, Taiwanese, and Chinese firms.
6
Introduction In the 2003 Nikkei Survey of global shares of 23 product categories (mostly IT products), Japanese firms maintained top ranking in several categories but continued to lose ground across a broad range of products. The Samsung Electronics Group was placed in the top five in seven of the 23 categories, with its share climbing in six of the seven. It took top rank in DRAMs and the flash memory market, the latter a technology first commercialized by Toshiba. In LCD panels (TFT LCD devices), LG Philips LCD (South Korea) and Samsung solidified their position as first and second ranked, while Sharp continued to loose market share (Nikkei Weekly, 2 Aug. 2004). In 2004, Samsung and LG jumped into first and second place for plasma panels, while Fujitsu-Hitachi Plasma Display plummeted to fourth (Nikkei shinbun, 19 July 2005). The Koreans have established a strong global position in cell phones, whereas the Japanese, despite world class technology, are limited mostly to the domestic market. Strong business models trumped strong technology. Even when the Japanese producers haven’t lost market share, they have often had to cut prices to meet other East Asian competitors’ prices so that profit margins have fallen substantially. In the case of memory chips, PCs, cell phones, digital cameras, and LCD panels, they have suffered losses in market share and have had to cut prices to survive. In these IT markets such reduction in profit margins makes it difficult to raise funds for the next product generation investment cycle. Nor are these short-term fluctuations. Japanese firms led by Sharp created the LCD industry. As late as 1997, there were 11 Japanese firms competing in one or more segments of this industry. Aggressive investment by Taiwanese and Korean manufacturers has forced most of them either to leave the industry or merge with domestic competitors. Japanese market share for the total LCD market fell from 62 percent in 1999 to 24 percent at the end of 2002.6 Table 1.2 shows the kinds of profit pressures being applied by Taiwanese and Koreans in the total LCD market. The great disparity in profitability has made it difficult
Table 1.2 LCD firm’s average profitability by national site, Q1 2004 (in millions of US dollars)
Revenue Profit Profitability
Japan: Sharp, Toshiba, Hitachi
Korea: Samsung, LG-Philips
Taiwan: AUO, CMO, CPT (TFT), GDI, Hannstar
2,668 143 5.4%
3,835 1,311 34.2%
3,261 845 25.9%
Note : Compared with Japan, South Korea is estimated to have 5% less R&D cost, 5% less salary cost, 10% less depreciation cost and the two major producers LG-Philips and Samsung pay respectively 7% and 30% as their effective tax rate compared to roughly 41% for large Japanese companies (Exchange rates calculated at $1¼1165.4 won, $1¼108 yen, $1¼33.965 NT$). Source : This table was prepared by Hirohisa Kawamoto, Nara Institute of Science and Technology based largely on data prepared by Masahiko Ishino, Senior Analyst, Mitsubishi Securities.
7
Recovering from success for Japanese firms to match the capital investment levels of their Asian competitors thereby jeopardizing their future in these technologies.
Rise of the specialist firm and the modularization challenge A second major development in the competitive environment has been the transformation of many ICT/electronics sectors from large vertically integrated firms with broad product lines to a network of specialist firms closely coordinating their value added activities. There are two major drivers for the rise of the specialist firms. One is the shift from integral architecture and closed proprietary standards to modular architectures with often open standards. The second is increasing recognition by managers that they can make breaks in the value added production chain that allow them to outsource parts of production. Most notable is the transformation of the computer and semiconductor industries and major parts of the telecommunication industry. These industries in varying degrees shifted, or are in the process of shifting, from either integral and/or closed technology standards to modular and/or open technology standards. Fujimoto Takahiro argues that Japan’s global competitive strength lies in industries like automotive with integral architecture and closed proprietary systems. Integral architecture requires exceedingly close coordination among product designers and external or internal production units, something at which the Japanese excel. In contrast, US firms tend to be more competitive in industries with modular architecture and open systems (Fujimoto et al. 2001; Fujimoto 2002: 23). The latter more often involve alliance strategies and systems integration at which US firms seem to excel. A brief introduction to integral and modular architectures is in order to grasp its competitive implications. Integral architecture has a set of common requirements that all components must share. Integral architecture allows for optimization, for example, of compactness and functionality in ways that make very attractive, but often expensive, products (such as Japanese mobile phones). Integral architectures can be said to appeal to the engineer’s desire for creativity. Modular architecture is based on separate modules operating as part of a system. Modules are units whose structural elements are powerfully connected among themselves and relatively weakly connected to elements in other units but nevertheless these units work well together in a larger system. Modularity provides flexibility by enabling firms to tailor a variety of products and service variations to better fit consumer needs (enabling mass customization). Modularization provides efficiency and speed. Components are interchangeable, providing plug and play compatibility. This in turn creates ease of use and variety for consumers. The components are also individually upgradeable, and thus innovation is less constrained by other parts of the system. Above all, the
8
Introduction component’s interface with the overall system is standardized; this provides the very premise for modularity (Baldwin and Clark 2000). These are valuable assets from both selected producers’ and consumers’ points of view and, as such, their competitive significance is profound. Japanese manufacturers’ weakness in modularized design (coupled with open systems) and production has put them at increasing competitive disadvantage in a number of key industries. There are some sectors where Japanese producers have embraced modular design and production with closed proprietary systems and thrived, such as selected machine tool manufacturers. However, while individual firms may thrive using this model, it does limit the number of complementary innovations produced by suppliers, since the competitive field is open only to selected suppliers. Modularity and the development of standardized interfaces are often coupled with the development of ‘open standards’ architecture. Open standards enable open systems by ensuring the spread of standardized design specifications to all possible industry participants, thereby lowering barriers to entry and increasing competition. One needs to distinguish between modularization in design, modularization in manufacturing, and modularization in use. Much of the scholarly literature is about modularization in design, and the assumption is that modularization in design precedes modularization in the manufacturing process. This is indeed the way the process evolved in the computer industry. In some industries, however–automotive is one example–there is a great deal of ongoing modularization in manufacturing, especially in Europe, but not as much in design. This means that even with an integral architecture, firms find ways to make a logical cut in the production chain where they can outsource a given system, such as an auto seating system or cockpit design. Typically this occurs in areas where design standards have been formalized.
The computer industry The Japanese high point in computer sales in worldwide markets in the early 1980s was associated with their overtaking of IBM in its mainframe business. Yet at the very moment this was occurring, fundamental changes were occurring in American mainframe architecture. Until the mid-1960s, mainframe computer design was based on an integral/closed computer architecture. Change came with the release of the hugely popular IBM System/360 in 1964. While still based on a closed proprietary design, the IBM/360 was the first truly modular design architecture. New subindustries emerged (e.g., computer storage devices, computer peripheral devices, and computer programming services) which did not make the whole computer system, but rather made parts or provided services–modules–for the larger computer system. In the era of the IBM System/360 series, they were ‘plug compatible’ with IBM
9
Recovering from success equipment. As modular designs spread throughout the industry, first in mainframes and then in portable computers, ‘the industry changed from being a virtual monopoly to a very large modular cluster’ where economic power was a function of control over critical modules (Baldwin and Clark 2000: 6–18). Japanese computer producers failed to embrace the modularization of the industry and for a long time continued on their own path (as did many US firms that are no longer with us, like Burroughs and, later, DEC). This meant a forgoing of the benefits of specialization and the network alliances and geographically distributed production chains that it made possible. Dell Computer exemplified an embrace of modularity. It developed deep logistic capabilities to coordinate its various partners, enabling optimization of its direct sales and build-to-order business model. The efficiency and cost competitiveness of this model, based on modularly designed and manufactured products, has yet to be matched by Japanese firms.
The semiconductor industry The semiconductor industry provides the foundational building blocks for the computer industry as well as other electronics-based industries, not to mention its strong impact on a wide variety of other industries such as automotive, retail, and finance. Japanese manufactuers rose to worldwide prominence in the mid-1980s by capturing larger and larger shares of the memory device market–DRAMs. The very strength of the Japanese in low margin DRAMs pushed many of their US competitors into logic devices, ‘mixed signal’, and other digital signal processes that allowed them to exploit their strengths in product innovation.7 These new emphases, and the aforementioned ability of American firms to emulate the success of their Japanese competitors in quality and process management, provided the basis for the strong American retaking of worldwide semiconductor market sales at the same time that Japanese manufacturers were succumbing to new competition in DRAMs from the Koreans, Taiwanese, and US-based Micron Technology. To understand developments in this industry, one needs to consider broader developments that have shaped the recent evolution of the electronics industry. First, in the US starting in the early 1990s, partly in response to Japanese competitive pressures, American manufacturing firms started to outsource a variety of those functions they decided were not part of their central mission and which could be done more efficiently and to a high quality elsewhere. This movement had its intellectual rationalization in the framework of core competencies developed by Prahalad and Hamel (1990). Firms were encouraged to identify and develop their core competencies and then to consider outsourcing those activities that didn’t fall into the inner core. The movement spread and contributed to the rise of many new specialist firms in a variety of industries.
10
Introduction The seminal development that contributed to the breakup of the organizational structure of the semiconductor industry (the Integrated Device Makers or IDMs), however, was the rise of the foundry model. ‘Fabless’ semiconductor firms, mostly located in the US and specializing in design, linked with Asian foundries (Taiwanese producers dominate this group) specializing in manufacturing. These foundries at the end of 2002 accounted for some 16 percent of all chips produced, and they were the fastest growing segment of the industry. Fabless semiconductor firms are concentrated in computer and telecommunications and are able to offer more innovative designs (as long as they meet the specifications of the Asian foundries) with shorter delivery times than the extant merchant semiconductor firms (Macher et al. 1998: 119–20). This development is linked to the source of much of American innovation being concentrated in start-up companies. Those start-up companies whose expertise was chip design were unable to afford the huge costs associated with building manufacturing capabilities. Thus, the rise of the foundries created the basis for a marriage that led to an alternative organizational architecture for the industry.8 More generally, the 1990s witnessed a growing reliance of firms in the US semiconductor industry on collaborative strategies among networks of specialist firms. These networks are both vertical–linking suppliers and OEMs–and horizontal–linking OEMs with one another–and have stretched across national borders. The spread of creative alliance and partnership models allows specialized firms to compete with integrated producers. They create economic networks that offer customers the virtues of specialization combined with the offerings of variety (Kogut 2000). Figure 1.1 depicts the growing specialization of the worldwide semiconductor industry.9 Up to the mid-1980s, the IDMs were the dominant organizational form for chip makers. The 1990s saw the rise of the fabless/foundry model just discussed. By the late 1990s new firms specializing in contract assembly, packaging, and testing had entered. We are also witnessing further vertical disintegration, with system design separating from chip design (Qualcomm is moving in this direction) and the rise of specialized intellectual property firms (such as Rambus). It has now become more common to license reusable design components (Intellectual Property (IP) blocks); specialized software firms producing automated chip design (Synopsis) have emerged. Corporate leaders in the Japanese semiconductor industry were slow to act upon these developments. The Japanese semiconductor operations were typically divisions or subsidiaries of large diversified electrical equipment firms. Compared with nimble, specialized firms, especially start-ups, they were slower to make decisions. Only with huge mounting losses did they exit DRAMs in the very late 1990s and early 2000s, leaving one fledgling national champion, Elpida, a joint venture between Hitachi and NEC. Until very recently they had very broad product lines, with each producer more or less
11
Recovering from success Specialized chip and process suppliers license and re-use design components (IP blocks)
Integrated chip company Modified chip company
Chip design
Chip design
Chip design
Chip design
System design
Intellectual property
Chip design
System design Chip design
Design services Fabrication
Assembly and testing Before 1986
Fabrication
Fabrication Foundry
Foundry
Assembly Contract and assembly testing and testing
Assembly and testing
1990s
After 2000
Assembly and testing
Figure 1.1 Evolution of the semiconductor industry Source : Shafter 2000: 174, 176
mimicking the product mix of its domestic competitors. Their failure to appreciate the ‘select and concentrate’ approach of their American and European competitors ultimately proved very costly. Prominent in the computer and semiconductor industries, as well as other electronics industries, has been the growing importance of contract manufacturers. These specialist producers have altered the basis of competition by drawing on economies derived from shared production, and by forcing firms to rethink their strategies for differentiation.10 Japanese firms have been only modest users of this model. This is understandable since they saw their core strength as manufacturing. As noted above, Fujimoto Takahiro argues that the long-term competitive strengths of Japanese firms have been in integral architecture and the concomitant dense circulation and communication of design relevant information delivered to the customer in the form of sophisticated products. It is this movement of information that is necessary for the critical coordination between design and other functions. Modular design architecture and manufacturing undercuts and fundamentally devalues this strength. Modular systems increase the ability to innovate while reducing costs of technical (though not necessarily other forms of) coordination and integration. Technical integration is handled by the presence of standardized interfaces. Minimizing the need for coordination and integration also reduces the need for the traditional,
12
Introduction highly personalized ‘relational contracting’ with its heavy emphasis on trust, which has been a forte of Japanese manufacturers (Dore 1987). Japanese productionist firms, in the ICT sector, have been increasingly swimming in a modular ocean. At the same time this is not a unidirectional trend. New generations of technology may create a shift back to integral product architectures, catching nonadapting firms in a ‘modularity trap’ (Chesbrough and Kusunoki 2001), one in which organization architecture designed for modularity becomes a liability. The dynamics of technologies, products, and industries vary greatly so that it can be exceedingly difficult to know if and when the pendulum may be swinging in the opposite direction, and how to calibrate organizational practices to these developments.11
Adjustment difficulties We have documented key challenges posed for Japan by two major changes in the competitive environment: the emulation of Japan’s 1980s success formula by old and new competitors, and the transformation of several key sectors, especially in ICT, into networks of specialist producers facilitated by modular technologies and open standards. While we have treated these two factors for the most part as independent forces, they interact in powerful ways (modularization in design and production led to the very effective network of alliances between Japan’s competitors: the US and Asian upstart producers). We turn now to some of the reasons for the slow response to these forces by Japanese producers. A well established axiom of organizational analysis sees the failure of once successful firms as rooted in their past success. Dosi (1982) argues that through successive organizational and technological choices, a firm’s trajectory leads it to become more accomplished in a given set of capabilities, while at the same time declining in those capabilities that would allow it to pursue different directions. Those persons that led the firm to the top develop over time, with reinforcement, a confidence in the soundness of their success formula and a blind spot for alternative approaches. Their expertise is based on previously successful strategies and work routines and they promote employees and favour strategies that will keep these tasks central. When hit with a competitive threat the magnitude and character of which are not well understood, they initially fall back on familiar solutions (Starbuck and Milliken 1988: 53; Burgelman and Grove 1996; Fligstein 1996: 667). For Japanese leaders, the success formula was low cost high quality precision hardware achieved through continuous process improvement in a framework of dense communication of design information across organizational units. It was a success formula, however, that over time was increasingly characterized by slowed decision making processes, caution, and a reduced tolerance for risk.
13
Recovering from success Inagami and Whittaker (2005: 141–2) call this ‘large firm malaise’ in their detailing of of Hitachi practices in the 1990s. Analyses of ‘large firm malaise’ place emphasis on the decision maker. Japanese firms are well known for their practice of late screening and late promotion. Kato’s study in the early 1990s found that the average age at time of appointment to CEO was 49 for the American and 56 for the Japanese firms. The Nikkei Weekly reported a trend toward selecting younger leaders for Japanese companies for the first half of 2005. Yet the average age for those becoming President (or CEO) in this period was still a high (by American standards) 56.3 years old, about the same as Kato had found some 13 years earlier. Kato also found that the length of service before promotion to chief executive was notably higher in Japanese firms; one in four American top executives had 25 or more years of firm experience before being promoted compared to one out of two in Japan. The larger the firm, moreover, the more firm-specific human capital was required for promotion (Kato 1993: 109–25). Such differences were held up by Kato and others as a reason for Japanese success in the 1980s. When managers finally arrived in positions of leadership, they had a strong stake in the firm, an intimate knowledge of firm culture and practices, and a long-term perspective. When a firm is faced with discontinuous change and new business models, however, organizational memory can become a barrier to innovation rather than an asset. Hitachi’s problems in the rapidly evolving ICT sector in the 1990s, for example, may have been intensified because many top managers were drawn from the heavy electric operations, or within ICT from mainframe operations, that were part of Hitachi’s past success. A strong institutional memory should be seen as a contingent asset. Finally, a rapid adaptation to modularization trends seems to have posed the risk of embracing a new unproven methodology at the expense of the tried and the true. It would have required a more rapid reduction of Japan’s domestic manufacturing labour force, which most large firms believed they could not do, especially in the early and mid-1990s (Inagami and Whittaker 2005: 148–51). In summary, there were ample problems in the late 1990s to which a solution packaged as a better way to manage technology and capture its value would have strong appeal. MOT was just such a solution.
The MOT boom Fads and the birth of MOT Fads are not to be dismissed out of hand. They can have lasting–and sometimes unanticipated–consequences. This is well illustrated by the US response to the Japanese quality challenge in the 1980s. The initial response to the
14
Introduction Japanese quality challenge was, predictably, denial. As Schein (1992: 298–302) argues, ‘unfreezing’ of a system requires not just a mass of disconfirming data, but anxiety or guilt on the one hand, and a belief that the problem can be tackled on the other. Crises typically accelerate the transition from denial through to a new readiness to learn and implement changes. There was a sufficient sense of crisis in the 1980s US to trigger this transition, and ‘In the course of responding to the Japanese quality challenge, a social movement developed, filled with zealots, nonbelievers, inspirational leaders, opportunists, and institution builders’ (Cole 1999: 231). Cole notes: Many scholars treat individual managers as atomistic agents, weighing all the evidence and then making rational decisions on behalf of their firms to improve or not to improve in the context of their specific environment. Yet, with the Japanese cast in the role of challengers in the early 1980s, the American incumbents banded together to mobilize resources and defend their markets. Managers developed networks to reduce knowledge barriers and to further adoption. Organizational action developed in response to a collective challenge, and these collectivities acted with common purpose, solidarity, and continuity to promote change in industry practices, often meeting strong resistance by those with vested interest in existing arrangements. Individual employees often saw quality as a kind of crusade and indeed were called quality zealots by their detractors. Thus, the very vocabulary suggests a social movement. (Cole 1999: 13–14)
Media interest came and went. Many companies abandoned quality circles. As we have seen, however, some of the key lessons were learned, and by the 1990s more effective quality practices became institutionalized and the quality gap across a range of industries had significantly closed. US manufacturers in the 1980s also sought to learn from the Japanese about the benefits of building partnerships with key suppliers. Imitation was imperfect here, too, especially in the automotive industry, but it led to innovation. Supplier management ideas and practices mutated, combined with the new information technologies, and evolved into supply chain management. The focus was now not just on optimizing bilateral ties between Original Equipment Manufacturers (OEMs) and suppliers, but on optimizing the whole supply chain. Supply chain management, in turn, has become an object of study and learning for Japanese manufacturers. By focusing on improving business processes, the Japanese-inspired quality movement in the US also became the basis for other new developments, including new fads like business process re-engineering. Again, aided by information technology and modular product architecture, these developments (albeit with problems in their implementation) boosted the revival of US manufacturing which, along with new challenges from Asian competitors, contributed greatly to the late 1990s crisis in Japan. Ironically, MOT in the US can hardly be called a fad, or movement with a public face. In their distress, US industrialists called on American business
15
Recovering from success schools and engineering colleges to make their training and research more relevant to the needs of business faced with a rapidly changing competitive landscape. Some of this pressure on universities was public, such as the open letter on TQM (total quality management) from the chairmen of American Express, Ford, IBM, Motorola, P&G, and Xerox published in the Nov./Dec. 1991 issue of the Harvard Business Review (Robinson et al. 1991: 94–5). Much of it, however, was privately expressed through industry advisory boards to these professional schools, and through the selection of university programmes that corporations chose to support financially. At the same time, many researchers were trying themselves to come to terms with the changing competitive environment, globalization, and the growing role of technology as a competitive factor (e.g. Teece 1998). A combination of these factors led to the emergence of Management of Technology–MOT–in the late 1980s–early 1990s.12 There was little government leadership and no coherent model of MOT, covering all phases of the production and enterprise life cycle. Instead there were a range of ad hoc innovations in industry and academia spread sometimes, as in the case of quality, through consultancies and industry-based associations, but at other times they came from leading firms such as IBM being imitated by a host of followers. A variety of programmes came into being in academia, sometimes going under the name MOT, but just as often using terms like innovation management, technology management, networking management, and entrepreneurship. These terms are scholarly descriptions of the subject matter rather than being part of the working language of those in the corporate world. The Technology and Innovation Management Division of the Academy of Management was formed in 1987 to bring together scholars interested in innovation, research, and development, and the management of technology-based organizations. The Division had over 1100 members in 2005, making it one of the larger divisions within the Academy of Management. Its research domain encompasses the strategic management of technology and innovation processes.13
MOT hits Japan Fast forward a decade to Japan. In line with the Odagira and Goto analysis, we see that Japanese companies have long practiced the management of technology. Indeed, it has been one of their strengths. It has been buttressed by the training, consultation and promotional activities of numerous industrial organizations. When MOT courses began to emerge in the US, moreover, Japanese engineers and R&D managers attended, and invited leading US MOT academics to Japan to share their knowledge. The immediate impact, however,
16
Introduction was local and limited. Japanese manufacturing companies were recording profits into the 1990s, and were able to attribute difficulties to a host of environmental factors. Except for a minority who sensed a change in the ground rules of competition, faith in their own productionist models was largely intact. The situation changed around 1998 when financial crisis was compounded by mounting losses in manufacturing, which triggered restructuring and even factory closures. This was enough to spark a questioning of previously successful practices, and a more urgent interest in the causes of the US resurgence, not just by individual managers but collectively. The wave of corporate governance reform hit Japan simultaneously, and firms were pressed to manage for profitability, with greater attention to shareholder interests. In 1999 the Japan Productivity Centre for Socio-Economic Development (JPC-SED) began work to establish the Japan Research Centre for Technology and Innovation Management (TIM-Japan). The Centre was launched in 2001, along with an MOT course in JPC-SED’s Management Academy. At the Kansai Science City, too, the Kansai Research Institute launched the Technological Business Innovation (TBI) executive seminar. Such seminars marked the emergence of MOT as a collective management issue in Japan, in 2001. More significantly, it was emerging as a policy issue. MOT was cited in a number of governmental reports in 2002, from the Council on Economic and Fiscal Policy to the Intellectual Property Strategy Council, and even the Central Education Deliberation Council. The Ministry of Economy, Trade and Industry (METI) had been contemplating MOT for some time, in fact. It was moved to action, in part, because of the seeming paralysis of many Japanese industry leaders in the late 1990s. In 1999, it started a feasibility study and then accepted bids, finally choosing Mitsubishi Research Institute (MRI) as its contractor to create an MOT Consortium in 2002. MRI invited bids from universities to develop MOT courses, which would create the basis of an MBA programme in all but name. Thirteen universities were selected to develop 16 courses. MRI and METI hoped–optimistically–that the courses, when placed online for use by other would-be MOT educators, would leverage its seed money into a nationwide campaign to bring management and entrepreneurship to technologists and engineers. It pointed to the large numbers of engineers graduating from MOT courses in the US, and their virtual absence in Japan. Later in the year METI secured a significant increase in funds (roughly ¥3 billion) from the Supplementary Budget, and seed funded 48 MOT education projects in universities, research institutes, and private companies. Companies took note and began to sound out universities to develop MOT courses for their engineers. Universities, in turn, sought to buttress their MOT credentials in order to access such funds and to make closer links with the corporate sector.
17
Recovering from success All this ignited media interest. Articles on technology management (gijutsu keiei) in the Nikkei group newspapers (Nikkei shinbun, in passing, was a vociferous advocate of restructuring for profitability and shareholder interests in the late 1990s) rose from around ten a year prior to 2002 to 34 in the second half of the year. In the first half of 2003, articles topped 100, a rate sustained throughout the following two years. Articles referring to the acronym MOT similarly began to rise sharply in 2002 to around 100 in the first half of 2003, and continued to rise, more modestly, over the next two years.14 Business journals began to run special articles on MOT from 2003 and academic books with MOT in the title began to appear shortly afterwards (some no doubt were hastily re-titled to exploit the upsurge in interest on the subject). New journals on the subject such as Nikkei Biztech (2003) and Technology Management Journal (2004) were also started.
University-industry (U-I) relations Adding to the foment were attempts to strengthen university-industry (U-I) relations, and changes to universities themselves. There is nothing new about the former. As far back as 1977 the Council of Science and Technology called for strengthening of basic research through U-I (and public research lab–P) collaboration, and programmes such as ERATO and U-I were created to do this in the early 1980s.15 It can be argued, moreover, that U-I collaboration did strengthen significantly over the next two decades. Commonly used indicators such as university held patents, licence income and university-based entrepreneurship highlight continued weaknesses, but different indicators such as joint publications suggest a more nuanced picture.16 These relations, however, were often informal, and constrained by legal and bureaucratic obstacles, as well as the under-funding of universities we noted earlier. As Japanese companies sought to leverage basic research in universities, they frequently turned overseas. Research contracts awarded by Japanese companies to domestic and foreign universities were broadly comparable in 1989 (¥38.4 billion and ¥43.0 billion respectively), but by 2000 domestic contracts had grown to ¥67.5 billion, while those for universities abroad had grown to ¥157.0 billion.17 Inspired by the US model (depicted above by Mowery and Rosenberg) and specifically the 1980 Bayh-Dole Act, the Japanese government responded by passing laws almost every year from 1998 to 2002 to encourage technology transfer from universities; to make it easier to trade intellectual property rights derived from publicly funded research; to promote use of university technology licensing offices (TLOs–25 were recognized by late 2001, 32 by 2004); and to encourage collaboration through tax measures. Then in late 2001, key governmental and nongovernmental bodies jointly launched a series of Regional U-I-P Summits, and in 2002 the massive
18
Introduction Conference for the Promotion of U-I-P Collaboration in Kyoto, which attracted over 4000 participants. These events have since been held annually. The launch of the regional summits, moreover, coincided with the first phase (2001–05) of METI’s Industrial Cluster Plan, for which 19 projects were chosen, and MEXT’s Knowledge Clusters scheme (2002), for which 18 projects were chosen. There were compelling reasons on the universities’ side to engage in these activities. In 2001 public sector research institutes and national hospitals became independent administrative institutions, after the UK agency model. This model, with some modifications, was applied to national universities in 2004. One objective was cost cutting. Another was to encourage–force– universities to become more active in new business creation through science and technology resources. Whatever the balance, and irrespective of the controversy about university autonomy (see Tabata 2005), U-I relations were set to assume a growing institutional, strategic importance in 2001–02.
The life cycle and appeal of MOT While neither MOT nor U-I collaboration were new, the specific circumstances of 2001–02, following the upheavals of the late 1990s, gave them a higher profile and potency than might otherwise have been the case. A new ingredient was enthusiastic, institutional participation by universities. The enthusiasm with which the business media took up MOT, with which universities raised the MOT flag to declare their intent to interact with business, and the rebranding of existing seminars, publications, and initiatives with the MOT label evoked criticism and in some cases derision. It was, critics claimed, a) a fad, with a shelf life of a couple of years at most; b) a superficial phenomenon of repackaging; c) a mish-mash of concepts jumbled together; d) another ‘three letter’ (acronym) import, further symbolizing Japan’s loss of confidence; and/ or e) a ‘black ship’ import tempting Japan’s manufacturers to try to emulate foreign practices and forgo their own strengths. Proponents, on the other hand, were attracted to MOT for a variety of reasons: a) it would make top managers situate technology more centrally in corporate strategies; b) it would encourage researchers and engineers to incorporate business model considerations in their technology and product development decisions; c) it would provide insights into the inroads made by US and Asian competitors, and how to counteract them; d) it would free up canalized thinking, allowing for new competitive concepts and approaches to be incorporated; e) it would accelerate participation by, and change in, inertia-bound universities; and/or f) it would encourage coordinated microand macro-level renewal strategies which built on Japanese strengths. In one sense the critics were right. The shelf life of media interest was about two years. As the economy recovered, Nikkei references to gijutsu kanri and
19
Recovering from success MOT halved in the first half of 2005 from the previous half year. That does not mean, however, that the movement has run its course. Non-degree and inhouse MOT programmes run by a variety of organizations continue to experience strong market demand.
The monozukuri boom Finally, in closing this section, we would like to consider another vogue term which is not an imported acronym but claims an indigenous heritage (arguably as an invented tradition). ‘Monozukuri’ means the making of things. We have used the adjective ‘productionist’ to convey its value orientation.18 Monozukuri has few detractors in Japan; it evokes the sense that it is the backbone of the Japanese economy, and should be in the future as well. The concept captured public attention in the depths of Japan’s despair to reaffirm the core strengths of Japanese manufacturing firms.19 In 1999 the Basic Law for the Promotion of Monozukuri Base Technology was enacted to shore up the position of manufacturing in the economy. Notable in this law was the recognition that interministerial action was required to address manufacturing problems; three ministries–METI, MEXT and the Ministry of Health, Labour and Welfare–produce the Monozukuri White Paper, published annually since 2000. Media citations of monozukuri began to rise in the second half of 1998, and have continued to rise ever since.20 While Fujimoto (2004) has presented a sophisticated version of what monozukuri means, subsequent popularizers have been less than careful. The popularized versions may allow monozukuri to be used as a justification of established ways of doing things that include many outmoded management practices. By encouraging firms to focus on their traditional strengths, it may divert them from adjusting to the new competitive environment. It is conceivable, conversely, that MOT concepts can be diffused through monozukuri since it is indeed the case that, for many firms, success will entail building on their traditional strengths. Significantly, references to monozukuri in regional Nikkei shinbun editions are twice as high as in the Nikkei sangyo shinbun (Nikkei Industrial Newspaper), but the ratio is reversed for gijutsu keiei and MOT. Interest in MOT appears to be concentrated in larger industrial firms, whereas monozukuri additionally reaches down into the myriad of smaller local manufacturers as well as reaching a general audience. The MOT boom appears to be accelerating the process of unfreezing, and to be opening up possibilities for new business models and strategies. If past Japanese experience is anything to go by, cognitive restructuring will take on a dialectical quality. In the same way that scientific management was introduced enthusiastically but applied selectively and creatively following friction with Japanese ‘traditions’ (especially in employment relations, see Tsutsui
20
Introduction 1998), MOT may be used to introduce tension and explore new possibilities, even as it is ultimately indigenized.
Synopsis The chapters have been organized into three parts. Those in Part 1–Industries, technologies, and value chains–explore challenges to Japan’s innovation and technology management practices at an industry level. Three of the chapters look at the ICT/electronics sector. First, Cole shows how former strengths became debilitating weaknesses in the telecom industry. NTT was the ‘locomotive’ which powered the industry with its ‘family’ of equipment suppliers in the 1980s. In the 1990s, however, it placed its bets on ‘competency enhancing’ ATM and ISDN, and fell victim to the ‘disruptive’ technology of TCP/IP. Strong relational ties with its suppliers led them as well to be weak competitors in the emergent Internet networking equipment industry, and a lack of new entrants forestalled alternatives from emerging within Japan. Cole also points to Japanese problems in international standard setting, betting on the established committee process and failing to engage in the fluid IETF process. Cole’s analysis finds strong echoes in Yunogami’s treatment of the semiconductor industry in Chapter 4. This is not surprising, perhaps, because many of the protagonists are the same. The quality regime ultimately deriving from NTT (and the Japanese government as customer) was vital in the ascent of DRAMs in the 1980s, as in telecom, but it created a path-dependent approach to innovation which proved disastrous when the main use of DRAMs switched from mainframes to personal computers. As market share plummeted, Japanese semiconductor engineers continued to take solace in the fact that they were ‘not beaten in technology’, an attitude which Yunogami argues further condemned them to failure. Chapter 3 takes up modular production/manufacturing (as well as emulation) which, Sturgeon notes, poses a direct challenge to a host of the most cherished strategies of Japanese electronics firms, including employment protection. Contradictory pressures to respond to the modularization challenge, on the one hand, and maintain cherished strategies on the other, lead to ‘simultaneously shedding and protecting jobs, getting out of old business lines and adding new ones, opening their sourcing networks and investing in new in-house component plants, expanding some facilities and shrinking or closing others’. Overall, Sturgeon finds adoption of aspects of modular production in low-end manufacturing, but much less for advanced products and technologies, creating a ‘mixed model.’ In Chapter 5, Probert looks at value chains in the Japanese pharmaceutical industry, which has until recently been relatively isolated from global consolidation, value chain modularization, and indeed global markets. One reason
21
Recovering from success has been the institutional and regulatory environment–including, again, governmental emphasis on safety over effectiveness–but corporate strategies have also played a part, including (until recently) a strong preference for organic growth and employment security. She also notes the limited contribution from university-based research activity, especially in biotechnology. But this industry is now in the throes of rapid change. Probert offers the prospect that leading pharmaceutical companies may well respond to their challenges while retaining cherished traditions through global niche strategies. In the final chapter in Part 1, Cole looks at another key ICT industry– packaged software–of which Japan is a huge net importer. In what goes to the heart of the modularization issue, Cole notes that the large electronics firms were reluctant to unbundle hardware and software, and insisted on maintaining proprietary systems; indeed, free software was (is) frequently used to boost hardware sales. Spinning out software divisions as group companies led to the maintenance of this orientation, rather than the creation of a thriving set of independent software firms focused on innovation. In what he calls the ‘curse of genba shugi’, Cole argues that the cherished shopfloor focus of Japanese manufacturing firms, despite its many positive contributions to productivity and quality, creates enormous pressures for customization which add heavy costs and foreclose benefits of standardization and corporatewide optimization that would come from the use of packaged software. In Part 2 we move to a micro-level view of innovation and technology management, within firms and between them. Chesbrough explores the implications of the transition from closed to open innovation over the past two decades which as Probert showed in the bio-pharma case, Japanese companies have been relatively late to address. Chesbrough gives examples of the fate of similar US companies. On the one hand the wide dispersion of critical knowledge for innovation is ignored at great risk, on the other alternative paths to market for internal ideas may be ignored at great loss. Chesbrough explores strategies for dealing with ‘false negatives’–which at first look unpromising but later turn out to be valuable–including the time-honoured Japanese practice of spin-offs. Brown looks at the HRM (human resource management) and knowledge system dimensions of open innovation in Chapter 8. In line with Chesbrough, she highlights the rising importance of externally derived knowledge. She establishes a link between the internal–external orientation of HRM systems and the type of external knowledge accessed. Internally/externally oriented HRM systems are associated with public/private external knowledge respectively. The latter tends to be closer to the cutting edge. Once again, this analysis points to difficulties Japanese companies face in participating in global knowledge networks. However, it also points to tradeoffs, between support for individual creativity on the one hand, and team work and control on the other.
22
Introduction Yamaguchi (Chapter 9) amplifies a number of the preceding themes and provides new insights in his case study of the blue LED invention. Distinguishing between ‘paradigm-disruptive’ innovation and ‘performance-disruptive’ innovation, he argues that large firms face difficult problems with both, but especially the former (because of their ‘competency-enhancing’ innovation bias, noted earlier, as well as growing bureaucracy and, more recently, restructuring). Using the blue LED case, he argues that there is a greater chance of top managers following a hunch, or creating a ‘field of resonance’ with their scientists in smaller companies. The implication of Yamaguchi’s paper is a call for hastening the shift to a more open innovation system in Japan, with a greater role for start-ups and networks of innovative small firms. Homing in even further on the individual, Byosiere explores the challenge of unleashing individual creativity in large corporations through examining the environment and incentives of R&D researchers. He finds considerable tension in the relationship between the individual researcher and the firm, especially over special rewards given to successful research results and research expenses available to researchers. Deeper causes, however, flow from the fracturing of the sense of community brought on by corporate restructurings. Finally, Byosiere notes that researchers see risk avoidance by their superiors as the major obstacle to doing creative work, the result of the shift by Japanese firms from being aggressive challengers in the postwar period through the 1980s, to becoming defenders of existing assets. In Chapter 11, Whittaker provides a case study of Hitachi’s response to its first postwar loss in 1998, in the face of growing competitive challenges we have outlined, and ‘large firm malaise’. It began with HRM reforms and an attempt to reshape employment relations, as well as organization and governance reforms. From 2001 MOT concepts were progressively introduced, through what he calls exploratory, strategic and systemic stages. These attempted to combine technologies with new business models, to develop areas of technology focus, and to develop management systems to more effectively mobilize and integrate group and external resources. Whittaker speculates as to whether these constitute a ‘new production(ist) system’. Keiretsu are important for understanding the openness–or otherwise–of Japan’s innovation system and MOT, corporate specialization, and boundary management. In the final chapter of Part 2, Lincoln draws on a series of studies which portray keiretsu ties as both a cause and a consequence of innovation. They foster organizational learning and innovation, and they grow through strategies which allow spin-offs to commercialize core company R&D. Innovation strategies can cause the ties to strengthen (as in the case of Matsushita), or weaken or become strained (Toyota-Denso). A quantitative study finds that keiretsu ties remained influential for non-R&D activities in the 1990s, while R&D relations became more strategic and less influenced by keiretsu, especially ‘legacy relations’ (see also Lincoln and Gerlach 2004).
23
Recovering from success Part 3 moves to the macro level, and policy attempts to reshape Japan’s innovation system. Prominent among these has been moves to create an environment conducive to basic research and breakthrough innovations. Arimoto, a policymaker himself, reminds us that this is not simply a matter of strengthening university research and forging closer links between universities and industry, but requires universities to address a broad range of skill needs, particularly in professional fields like law, management and intellectual property matters, which individual companies can no longer address on their own. He notes that social issues must also be addressed, including greater efforts being made to promote public understanding of and support for science and technology. This requires directing S&T and innovation policies towards the needs of the many rather than the few. Ultimately, Arimoto argues, the necessary values and creativity have to reside in individuals, a point with which Byosiere would surely agree. Murayama (Chapter 14) extends this discussion into the area of security, which has been thrust onto the policy and innovation system debate agenda in the wake of 9/11. He first compares key features of the US and Japanese national innovation systems, and the respective roles of defence. Defence has played a relatively minor role in Japan’s innovation system, but like civilian industry in the 1990s, Japan’s defence industry failed to adapt to a dramatically changed (post Cold War) environment in the 1990s. In the new century, however, anzen-anshin (safety and security/peace of mind) has the potential to influence both. The trajectory of influence is still fluid, but Murayama’s preference is clearly for a civilian-oriented security techno-system. Kaneko, Nakata, and Yokoyama return to skills, focusing specifically on MOT. (The significance of the growing policy emphasis on skills is discussed in the Conclusion.) They outline five problem areas for MOT which must be addressed with new skill sets. Identifying these is one thing; resolving them is another. After reviewing MOT education in the US, Europe, and elsewhere in Asia, the authors discuss its provision in Japan, identifying problems in curriculum design (hastily assembled, objectives unclear), teaching methods (too much one-way transmission), and teaching resources (too few competent teachers). On the corporate side, too, there is much reliance on tacit knowledge. A common challenge is to find ways to make the key precepts for MOT explicit, so that they can be debated, developed, and diffused. The final chapter in Part 3 looks at government and policy from a different perspective: e-Government has been an important if overlooked part of the e-Japan strategy, and central to this is Juki-net. Kita analyses the debacle of its introduction, which was marked by initial confrontation with anti Juki-net campaigners concerned about privacy and information security, and subsequently between administrative agencies and residents, where passive resistance virtually consigned the Juki-card to oblivion. Kita proposes a ‘customer-oriented’ solution to the impasse, which he considers symptomatic of
24
Introduction the whole e-Japan programme. In his analysis, the policymakers are as much in need of MOT education as the engineers and managers who still believe in the linear model of innovation. We end the book with the Conclusion in which we attempt to tease out the implications of the individual chapters for the future of innovation and MOT in Japan.
Notes 1. There are other, related challenges, of course, such as escalating foreign direct investment in the 1980s, which accelerated in the 90s as Japan moved more production to China and raised the spectre of ‘hollowing out’. 2. Not all manufacturing industries were competitive though, and service industries visibly lagged in productivity growth (see Porter et al. 2000). 3. In national research institutes R&D spending also declined relatively, from 27.5 percent in 1978 to 21.7 percent in 1990, and within this, basic research from 18.5 percent to 14.2 percent (Lee 1997). 4. Macro factors include the appreciation of the yen, from ¥258/$1 in the first quarter of 1985 to ¥84/$1 in the second quarter of 1995. 5. We are indebted to Prof. Fujimoto for conversations with regard to more recent developments, 19 July 2004. 6. Asahi shinbun (2002). ‘Japan’s LCD Muscle Shrinks’, http://www.asahi.com/english/ business/K2002122800357.html 7. This subsection draws heavily on Macher et al. 1998: 107–36). 8. Conversations with Robert Leachman, UC Berkeley, College of Engineering, proved especially helpful in this formulation. 9. Notwithstanding, there are still semiconductor areas such as leading edge DRAM design where close coordination with production is critical, and in these cases there are pressures for integration rather than specialization (Macher et al. 2002: 6). 10. See Sturgeon 2002, as well as in this volume for an extended treatment of contract manufacturing. 11. See also Fine 1999; Kusunoki 2004. 12. At UC Berkeley, a committee set up by the Haas School of Business and the College of Engineering in 1988 drafted a Long Range Plan for the Joint Program in the Management of Technology, which was launched in 1989. MIT began MOT education in 1985, and started the Leaders for Industry Program in 1988. 13. In the UK, which also faced the brunt of competition from Japan in the 1980s, there were numerous initiatives encompassing elements of MOT, but they did not coalesce into what we would call MOT until the mid-1990s (see Gregory 1995). 14. In many cases MOT and gijutsu keiei appear in the same article. Figures are from the Nikkei Telecom 21 database, accessed 8 December 2004, and 3 August 2005. A similar if less pronounced trend can be found in references in the Asahi shinbun’s Asahi DNA database. References to an alternative expression for technology management–gijutsu kanri–showed no marked trend over the period. This term is less
25
Recovering from success
15.
16.
17. 18. 19. 20.
associated with the imported concept of MOT, and commonly refers to technology management in a less strategic sense. Lee 1997. ERATO stands for Exploratory Research for Advanced Technology. It was under the Science and Technology Agency, while U-I were under the Education Ministry. In 2001 the STA was merged into the Ministry of Education, now the Ministry of Education, Culture, Sports, Science and Technology (MEXT). Pechter and Kakinuma (1999) compare joint publications in Japan and the US and find as much if not more collaborative activity on this count in Japan. Branscomb, Kodama and Florida (eds.) (1999) provide a good survey on U-I relations in Japan and the US until the late 1990s. In 2003 the figures were ¥83.4 billion and ¥198.5 billion respectively (Somucho 2003). Japan’s capitalism has also been depicted as ‘production-oriented’ capitalism (see Dore 1987: 13). For a critical view of the monozukuri vogue see Tsai 2005. Again, this is according to the Nikkei Telecom 21 data base. Citations until the first half of 1998 were under 50, by the second half of 2004 they exceeded 300, well over twice the figures for gijutsu keiei and MOT.
References Aoki, M. (1988). Information Incentives and Bargaining in the Japanese Economy, Cambridge: Cambridge University Press. Baldwin, C. and K. Clark (2000). Design Rules, Cambridge, MA: MIT Press. Branscomb, L., F. Kodama and R. Florida (1999). Industrializing Knowledge: University– industry linkages in Japan and the United States, Cambridge, MA: MIT Press. Burgelman, R. and A. Grove (1996). ‘Strategic Dissonance’, California Management Review, 38(2): 8–28. Chesbrough, H. and K. Kusunoki (2001). ‘The Modularity Trap: Innovation, technology phase shifts and the resulting limits of virtual organizations’, in I. Nonaka and D. Teece (eds.) Managing Industrial Knowledge: Creation, transfer and utilization, London: Sage. Cole, R. (1999). Managing Quality Fads: How American business learned to play the quality game, New York: Oxford University Press. Dore, R. (1987). Taking Japan Seriously, Stanford: Stanford University Press. Dosi, G. (1982). ‘Technological Paradigms and Technological Trajectories’, Research Policy, 11: 147–62. Fine, C. (1999). Clockspeed: Winning industry control in the age of temporary advantage, New York: Basic Books. Fligstein, N. (1996). ‘A Political–Cultural Approach to Market Institutions’, American Sociological Review, 61 (August): 656–73. Fujimoto, T. (1999). The Evolution of a Manufacturing System at Toyota, New York: Oxford University Press. —— (2002). ‘Architecture, Capability, and Competitiveness of Firms and Industries’, Working Paper, University of Tokyo. —— (2004). Nihon no monozukuri tetsugaku (The Philosophy of Japanese Production), Tokyo: Nihon keizai shinbunsha.
26
Introduction —— , A. Takeishi and Y. Aoshima (2001). Bijinesu akitekuchia, seihin, soshiki, purosesu no senryakuteki sekkei nijumon (A Primer on Business Architecture, Strategic Design of Products, Organizations, and Processes), Tokyo: Nihon keizai shinbunsha. Gregory, M. (1995). ‘Technology Management: A process approach’, Proceedings of the Institute of Mechanical Engineers, 209: 347–56. IMD (2004). The World Competitiveness Yearbook, Lausanne. Inagami, T. and D. H. Whittaker (2005). The New Community Firm: Employment, governance and management reform in Japan, Cambridge: Cambridge University Press. Kato, T. (1993). ‘Internal Labor Markets for Managers and the Speed of Promotion in the U.S. and Japan’, Japan Institute of Labour, An International Comparison of Professionals and Managers, JIL Report Series No.2, Tokyo: 1–245. Kogut, B. (2000). ‘The Network As Knowledge: Generative rules and the emergence of structure’, Strategic Management Journal, 21: 405–25. Kusunoki, K. (2004). Synthesizing Modular and Integral Knowledge, in H. Takeuchi and I. Nonaka (eds.) Hitotsubashi in Knowledge Management, Singapore: Wiley. Leachman, R. (2002). ‘Competitive Semiconductor Manufacturing: Final report on findings from benchmarking eight-inch., sub 350nm wafer fabrication lines’, CSM-52, Engineering Systems Research Center, University of California, Berkeley, CA. —— and D. Hodges (1996). ‘Benchmarking Semiconductor Manufacturing’, IEEE Transactions on Semiconductor Manufacturing, 9(2): 158–69. Lee, K. (1997). ‘Managing Basic Research in Japan: Towards a Japanese system of breakthrough innovation?’ PhD dissertation, Judge Institute of Management Studies, Cambridge University, Cambridge. Lincoln, J. and M. Gerlach (2004). Japan’s Network Economy: Structure, persistence and change, Cambridge: Cambridge University Press. Macher, J., D. Mowery, and D. Hodges (1998). ‘Reversal of Fortune? The recovery of the U.S. industry’, California Management Review, 41(1): 107–36. —— , D. Mowery, and T. Simcoe (2002). ‘eBusiness and the Semiconductor Industry Value Chain: Implications for vertical specialization and integrated semiconductor manufacturers’, Industry & Innovation, forthcoming. Ministry of Economy, Trade and Industry (METI) (2001). Trends in Japan Industrial R&D Activities–Principal Indicators and Survey Data, Tokyo: METI Technology Research and Information Office. —— (2003). Trends in Japan Industrial R&D Activities–Principal Indicators and Survey Data, METI, Technology Research and Information Office. Mowery, D. and N. Rosenberg (1993). ‘The U.S. National Innovation System’, in R. Nelson (ed.) National Innovation Systems, New York: Oxford University Press. Nezu, R. (2002). ‘Perspective and Strategies for Japanese Industry’, presented at the Prospects for Core Industries in Japan and Germany Conference, Japanese-German Center, Berlin, Fujitsu Research Institute and German Institute for Economic Research, Nov. 28–9. Nikkei Weekly (2005). ‘Companies Tap Young Leaders’, Nikkei Weekly, 43(2) (13 June): 1,7. Odagiri, H. and A. Goto (1993). ‘The Japanese System of Innovation: Past, present and future’, in R. Nelson (ed.) National Innovation Systems, New York: Oxford University Press.
27
Recovering from success Pechter, K. and S. Kakinuma (1999). ‘Co-authorship Linkages Between University Research and Japanese Industry’, in L. Branscomb et al. (eds.) Industrializing Knowledge, Cambridge, MA: MIT Press. Porter, M., H. Takeuchi, and M. Sakakibara (2000). Can Japan Compete? Basingstoke: Macmillan. Prahalad, C. K. and G. Hamel (1990). ‘The Core Competence of the Corporation’, Harvard Business Review, 3: 79–91. Robinson, J. et al. (1991). ‘An Open Letter: TQM on the campus’, Harvard Business Review, 69 (Nov./Dec.): 94–5. Schein, E. (1992). Organizational Culture and Leadership, 2nd edn, San Francisco: Jossey Bass. Shafter, S. T. (2000). ‘The Chipping News’, Red Herring, 30 October: 174, 176. Somucho (ed.) (2003). ‘Heisei 15 nendo kagaku gijutsu kenkyu chosa hokoku’ (Research Survey Report on Science and Technology), Tokyo. Starbuck, W. and F. J. Milliken (1988). ‘Executives Perceptual Filers: What they notice and how they make sense’, in D.C. Hambrick (ed.) The Executive Effect: Concepts and Methods for Studying Top Managers, Greenwich, CN: JAI Press. Sturgeon, T. (2002). ‘Modular Production Networks. A new american model of industrial organization’, Industrial and Corporate Change, 11(3): 451–96. Tabata, H. (2005). ‘Reform of Japan’s National Universities’, Social Science Japan Journal, 8(1): 91–102. Teece, D. (1998). ‘Capturing Value from Technology Assets: The new economy, markets for know how, and intangible assets’, in R. Cole (ed.) California Management Review, 40: 55–79. Tsai, M. H. (2005). ‘The Myth of Monozukuri: Manufactured manufacturing ideology’, ITEC Working Papers Series, Doshisha University. Tsutsui, W. (1998). Manufacturing Ideology: Scientific management in twentieth century Japan, Princeton: Princeton University Press.
28
Part I Industries, technologies, and value chains
This page intentionally left blank
2 The telecommunication industry: A turnaround in Japan’s global presence Robert E. Cole
Japanese telecommunication firms entered the 1990s as major players in the global market. Despite the explosion in global demand for telecommunications products, they witnessed a strong decline in global competitiveness over the subsequent decade. We explore the factors accounting for this shift, focusing on a key emergent sector: the Internet network equipment industry. The global explosion of demand for telecom products during the 1990s can be seen in the rise of carrier telecommunication revenues in the OECD from roughly US $400 billion in 1992 to US $800 billion in 2001 (Matsuo 2003: 20). For a closer look, we can examine exports of communications equipment; this is a broader category than just telecommunication equipment (telecommunication equipment accounts for roughly 21 percent of total communication exports). Exports of communication equipment from OECD countries increased from US $49 billion in 1991 to US $165 billion in 2001. Communication equipment export growth was 12 percent per annum in the EU, 11 percent in the United States and zero in Japan. In 1991, Japan accounted for 27 percent of total OECD communication equipment exports. By 2001, Japan’s share had fallen to 8 percent (OECD 2003: 228–30). In short, Japanese firms did not share much in the increased trade in communication equipment. If we focus only on the telecommunication equipment balance of trade for Japan from 1990–2001 as shown in Figure 2.1, we see first that exports rose from roughly US $11.7 billion in 1990 to US $13.2 billion in 2001. Set against this very modest increase (in view of the growth of worldwide demand), we see a dramatic growth of telecommunication equipment imports from roughly US $2 billion in 1990 to almost US $10 billion in 2001. This growth in imports more than cancels out the rise in exports, so that Japan’s overall telecommunication trade balance shows a sharp decline from almost US $10 billion in 1990 to US $4 billion in 2001.
31
Industries, technologies, and value chains 18 16
Billion US$
14 12 10 8
Total exports Total imports Trade balance
6 4
2001
2000
1999
1998
1997
1996
1995
1994
1993
1992
1991
0
1990
2
Figure 2.1 Japan’s trade balance in telecommunications equipment Source: OECD Telecommunications database, created by Kenji Kushida, 2004
A key component of this sector is the Internet network equipment industry, a segment of the industry that did not exist much before 1990. The market for this sector was estimated to be US $39 billion in 2000 (Semilof 2000). American firms, especially Cisco, dominate the global market. Cisco, for example, is estimated in 2002 to have 80 percent of the US $6.6 billion market for routers, 69 percent of the US $10.4 billion market for switches, 48 percent of the US $2.1 billion market for IP telephones, and 30 percent of the market for network security equipment. Moreover, its closest competitors are typically North American firms (Yamazaki 2003: 50). Japanese firms, with some modest exceptions, are notable by their absence. In response to the developments described above, the then Director-General of Commerce and Information Policy Bureau of the Ministry of Economy, Trade, and Industry (METI) met with top officials at NEC, Hitachi, and Fujitsu in December 2002 to urge these companies to integrate their telecommunications equipment businesses. The most promising opportunity was to create a ‘merger among equals’ between NEC and Hitachi but the effort failed. Such government interventions are typically reserved for troubled industries perceived to be in crisis. This action to consolidate the domestic industry into a few strong corporations reflected not simply the troubled state of the worldwide telecommunications industry but was also spurred by ‘a sense of crisis about declining Japanese telecom equipment technology’ (Nihon Keizai Shinbun 2003: 10). The cost of R&D to compete in telecoms has risen so much that Japanese government officials believe that mergers must take place for Japanese firms to be competitive in the future.
32
The telecommunication industry How do we account for these dramatic shifts in fortune? The causal factors are not consistent across industry sectors. The factors which accounted for the dramatic decline in Japan’s global share of second generation handset sales are not the same as the ones that explain Japan’s failure to participate in the explosive growth of the network equipment industry (see Funk 2002 for an exposition of the mobile phone case). There are, however, some overlapping factors such as the strategic failure of Japanese firms to effectively play the global standards game. Keeping these issues in mind, we will focus on the network equipment industry both because of its position as being critical for future IT developments and because there is relatively little literature on this subject.
NTT’s miscalculations and their consequences We begin with a change in the competitive environment that has had profound reverberations for the development of the Information and Communication Technologies (ICT) sector worldwide. This was the worldwide movement toward controlled deregulation of the telecommunications sector. First were the various regulatory reforms around the world that led to continued liberalization of the sector. Second, various technological innovations of the 1990s (including fibre optics, high capacity, and high speed hard disk drives and digital subscriber lines) expanded the volume and capacity of communications. Third, and most importantly, has been the convergence of the telecommunications and information technology sectors, especially in the mid-1990s with the emergence of the World Wide Web and the browser; this ‘linked the existing capital stock of computers and communications systems in an open network that significantly increased their utility’ (OECD 2003: 56–7). The US took the lead in these developments. In the 1970s, the arguments for deregulation of regulated industries gathered strength as the ideology of competition received renewed emphasis (Temin 1987). It was in this environment that the breakup of the Bell System and the creation of the ‘Baby Bells’ as the cornerstone of a 1982 anti-trust settlement took place. Along with AT&T, seven new ‘regional operating companies’ were created. The divestiture ushered in a new era characterized by a more open competitive environment throughout the communications sector. This environment, when combined with the creation of the World Wide Web and wireless radio-based telephony, allowed for the entry of thousands of new competitors in the communications sector. Messerschmitt (2000: 212) estimates that venture capital played more of a role in networked computing than in perhaps any other industry other than bio-tech. Until 2003–04, Japan was a consistent outlier when it came to the costs of accessing the Internet and the pricing of leased lines. Despite early public
33
Industries, technologies, and value chains discussion of the coming importance of the convergence between communications and computers, Japan lagged in adopting and applying these new capabilities. Much of that lag can be traced to the failure to deal creatively with the NTT telephone monopoly in Japan. Powerful institutional forces and vested interests contributed to both slow the process of deregulation and shape it in ways that preserved much of NTT’s structure and pricing power (Vogel 1997; Tilton 2003). It is striking that one of the major groups adamantly opposing the NTT breakup was the traditional ‘family’ of NTT equipment suppliers: NEC, Hitachi, Fujitsu and Oki Electric. Yet, it is these companies that, long-term, would be the most likely major beneficiaries of such a breakup insofar as the changes could be expected to accelerate the creation of an ICT infrastructure and associated products and services. It would appear that the equipment suppliers focused primarily on the disruptions that a breakup might create for their up to then guaranteed NTT markets. This suggests contingency for the view that relationship contracting (networks of preferential stable, obligated trading relationships among upstream and downstream customers and suppliers), so prevalent in Japan, has strong positive benefits (Dore 1987: 173–91). Ronald Dore’s analysis of what he called relational contracting stressed the criticality of long-term trust-based relationships, risk sharing among partners in good times and bad, and the efficiency benefits that flow from such cooperation. When innovation and exploitation of new opportunities, however, are the objective (not ever-improving efficiency), and the environment is characterized by discontinuous change and strong first mover advantages, then seeking out new partners and employees with ready made capabilities may be required to successfully take advantage of new opportunities. The NTT suppliers ignored the long-term benefits they might receive from reorienting product lines to compete in worldwide markets in favour of the short-term benefits associated with continuing to cooperate with NTT. It is instructive to note that a recent analysis of Chinese success in global telecom markets showed that the Chinese have learned from Japanese mistakes, to wit, that Japanese telecom equipment companies tied their fortunes too closely to Japanese carriers, ‘which developed technological standards that they expected the rest of the world to adopt. When that didn’t happen, the Japanese vendors became captives of their home market’ (Rhoads and Hutzler 2004). NTT was Japan’s largest employer throughout the 1980s and into the 1990s (291,000 employees in 1989), the world’s most valuable company until the early 1990s, the centre of Japanese telecom R&D activity, and an engine of national economic growth. Not surprisingly, privatization and breakup had major implications for other institutional actors including the unions, NTT equipment suppliers, the ministries, and politicians. The unions were particularly active and effective in opposing the breakup (Tilton 2003: 3). Finally, as
34
The telecommunication industry NTT was seen as the core of Japan’s R&D capabilities in telecom, a breakup was seen as a threat to these capabilities. The breakup of AT&T and resultant dissipation and splintering of R&D capabilities was held up as an example to be avoided at all costs. Predictably, the combined external forces for inertia were large and strong. There were also powerful internal institutional rigidities that slowed NTT’s support for the emergence of networking technologies in the form of new products and services. NTT researchers didn’t see the potential of TCP (Transmission Control Protocol)/IP (Internet Protocol)–layers four and three of the layered network reference model) and the Ethernet (layers two and below of the layered network), in part because of their ingrained focus on the need for high reliability systems for provision of domestic universal service. This, after all, had been their long-term mandate. We see, however, that under certain conditions the vaunted quality of Japanese firms can be the enemy of innovation. The Internet, based on packet technology, was a ‘best effort’ network that did not initially match the traditional quality benchmarks (in the network area these are known as Quality of Service (QoS) benchmarks) provided in universal voice service. In particular, it was in the beginning quite deficient in minimizing delays and in providing sufficient bandwidth guarantees as well as in insuring reliability (correctness of data transfer). As a best effort network, the Internet does not guarantee delivery of specific messages and involves re-transmission of dropped packets. In the early 2000s some 3 percent of all packets sent daily were dropped.1 Moreover, each packet is delayed by variable and unknown amounts and the bandwidth available to each connection is unpredictable. The traditional ‘five nines’ (99.999 percent) reliability target of telephone companies was simply not a design requirement for the Internet architects. All this was anathema to the QoS culture of NTT. The many low QoS and reliability features exhibited by the early Internet are a common feature of disruptive technologies (Christensen 1997). Engineers, however, are often able to incrementally add new features and improve reliability as one after another of the technical problems of disruptive technologies gets resolved. High reliability organizations, like traditional telephone companies in particular, have a great deal of difficulty in understanding and responding positively to disruptive technologies with these trajectories because they initially challenge existing value propositions. (See Yunogami in this volume for the consequences of strong emphasis on reliability in the semiconductor industry when applications and markets changed.) In the late 1990s there were still senior NTT executives who didn’t understand TCP/IP. Moreover, most NTT researchers, well into the mid-1990s, still by and large preferred Asynchronous Transfer Mode (ATM) technology as their mainstream approach to networking and viewed TCP/IP as an interesting option. Indeed, the IP router was only one of many possible pieces of
35
Industries, technologies, and value chains equipment for building the data network. Many NTT researchers preferred ‘IP over ATM’ as the ideal solution from the point of view of providing high QoS. It was not a matter, however, of choosing between two new promising technologies. NTT began research on ATM switching in the mid-1980s. ATM was a ‘competency-enhancingtechnology’(Anderson andTushman1997: 48–50) in that it was a natural extension of the existing public telephone network relying on circuit switching. In circuit switching, distance and duration of connection determine the cost of communication service (Yamashita 2004: 1). Such competency enhancing technologies are typically easier to incorporate into incumbent organizations and thus they are more attracted to them. NTT predictably wanted to continue to extract high levels of profit from their existing fixed line investments. Moreover, ATM was consistent with the high reliability culture of NTT. ATM is connection oriented, meaning that all host to host communications requests are provided a connection (fixed route) through the network. There are two alternative approaches to transmission. The first is to devise complex mechanisms in the core to reduce error; this is the path followed by ATM in the effort to deliver guaranteed QoS. The second is to have terminals located around the network and to engage in re-transmission when error occurs; this is the path followed by Internet architects. The development of ATM was designed to improve NTT’s existing digital switches. Originally designed for voice, NTT officials were convinced ATM could be made to be multipurpose. They believed in the early 1990s that ATM was the ultimate solution mixing voice and data traffic over fibre. In 1991 Fujitsu became the first company worldwide to offer an ATM switching system that enabled high speed, two way transmission and routing of voice, video, and data simultaneously. In the early 1990s, NTT anticipated replacing the existing current narrow band digital network with the large capacity broadband ATM trunkline network somewhere around the year 2000 (Fransman 1995: 86, 116, 123). NTT officials were confident that they were leaders en route to building the new information highway. They pursued this dead end trajectory with the strong support and urging of the Ministry of Post and Telecommunications, buttressed by tax incentives and public money. Rather than wait and let equipment suppliers take the lead in developing ATM switches, NTT took the initiative and led its suppliers in developing this new technology. This also involved NTT taking the lead in developing the software required for the broadband ISDN services that they expected to be deployed over this network. In this way NTT thought it would insure that it accumulated and controlled the key competencies required for operating, maintaining, and modifying the switch software necessary for providing new services. This was in keeping with its long-standing view that it was their job to take the lead in advance of the equipment vendors in developing complex new technologies (Fransman 1995: 115–16, 119). Correspondingly, this led equipment vendors to
36
The telecommunication industry take a passive view and wait for NTT to take the lead when confronting complex technologies. As we shall see shortly, this passivity had disastrous consequences for Japan’s emergent network infrastructure industry. In 1995 NTT was experimenting with vBNS, a network built on the commercial ATM lines. It was built with a speed of 155 Mb/s and was expanded to 622 Mb/s in two years. Those collaborating with NTT aimed for a speed of 2.4 Gb/s. This group concluded at the end of 1997 that it was not feasible to achieve 2.4 Gb/s and that it would be more effective to exclude ATM from the network and utilize IP directly over SONET [Synchronous Optical Networking]. SONET is a layer two network technology for communication over optical fibre. It is only at this point that NTT executives began to realize that ATM was not the ultimate end to end solution (Oie et al. 2001: 184–5). In 1997–98 they finally realized that everything needed to be changed. This recognition was delayed by the strong internal political commitment to ATM. NTT had incurred a substantial investment in ATM switches and had developed rated products. Its executives didn’t want to admit failure even after key engineers in their Basic Research Lab concluded that ATM could not provide the speed possible with TCP/IP over SONET and the Ethernet. In particular the Network Service Systems group, which made telephone switching systems for ATM, continued to push ATM and lobbied NTT executives to continue supporting it throughout much of the 1990s. It was not until the late 1990s that NTT finally stopped their research on telephone switching units based on ATM. In late 1999 the mobile phone market was growing rapidly and the number of subscribers to DoCoMo’s i-mode was exploding. NEC, accustomed to following NTT’s lead, shifted resources including personnel who had worked on ATM over to second generation phones. They thought it would have potential for export markets to China and Europe, a hope that would prove unfounded. Of course, NTT engineers were not ignorant of the emergent Internet technology. A small informal group promoting Internet concepts emerged in NTT in the early 1980s.2 Dr Shigeki Goto, a research group leader (kacho) at the NTT Research Laboratories, arranged to send a Dr Okuno to Stanford University. With Ken Murakami as the lead person, the Japanese team finally succeeded in 1988 in connecting the NTT Laboratories Computer Network to the CSNet (Computer Science Net) and ARPANET through CSNET in the US. Even this was done informally because at the time NTT was forbidden by the government to engage in overseas activities. The Internet group operated initially as a ‘skunkworks’ (an informal group flying under the radar of the formal organization). A key step in the process of formal recognition of this group came in 1992 when Shigeki Goto became a department head (bucho). This enabled him to start several Internet projects with the official support of the Director of NTT Software laboratories. The initially small pro-Internet group promoted TCP/IP and the Ethernet as desirable solutions and gradually key researchers were won over. The turning
37
Industries, technologies, and value chains point for them to accept TCP/IP was the arrival of Mosaic in the period 1993–95. Nevertheless, as we saw above, top NTT executives continued to resist these networking solutions until the late 1990s.3 With this background in mind it is not surprising that the Japanese electronics industry has lagged in the introduction of cutting edge products and services associated with networking. ISDN requires digital switches, and many of their researchers were kept busy with the very consuming task of developing all sorts of equipment to allow the telephone lines to handle data. By contrast, TCP/IP requires routers, network interface cards, and sophisticated software such as network configuration management software. TCP/IP protocols are mostly implemented in software, running at both the router’s and the user’s computers. When one examines the global market for these products one is struck by the almost complete absence of Japanese vendors (International Data Corporation 2000). Without a strong domestic market in the networking products associated with TCP/IP and the Ethernet Japanese electronic firms were unable to build up scale economies that could serve as a platform for competing in international markets. In a Soumusho-commissioned survey of Japanese and foreign information communications researchers, respondents were surveyed on the superiority of Japan, the US and Europe in specific information and communication technologies. Respondents saw great Japanese strength in intelligent home appliances, mobile terminals, and optical communications. They ranked the Japanese weakest, however, in software, the Internet, content production support, computer systems, and security. In these areas, they acknowledged overwhelming American superiority (Soumusho 2003: 10).
Other (non)players In summary, the major Japanese electronic firms were accustomed to relying on NTT to set future technology directions in the field of communications. With NTT being slow to grasp the significance of TCP/IP and the Ethernet, the major electronic firms lagged in the development of ICT products and services. We can see this being played out in the behaviour of NEC and Fujitsu. In the early 1990s, Masao Hibino, President and CEO of NEC Magnus Communications, was General Manager of Modem Development at NEC and stationed in Silicon Valley. He thought TCP/IP and the Ethernet were important developments and sent information to NEC offices in Tokyo to that effect. They responded, however, that ‘TCP-IP’ wasn’t real communication because it was ‘connection-less’. In short, without a dedicated connection, they believed that there was no real communication. At this time he says, ‘NEC people thought ATM delivering ISDN services was the final solution to broadband. Everyone in Japan thought so and we worked with ITU-T
38
The telecommunication industry (International Telecommunication Union–Telecommunication Standardization Sector) to get each standard approved for ATM’. NEC also had mid-career Industrial Associates at UC Berkeley and these individuals were aware of the evolving Internet technology. Even when these individuals realized the significance of the Internet, however, they were reluctant to take the kind of risks that would be entailed in strongly championing what was seen at their firm as a nonmainstream technology. By the time NTT and the electronic companies began to focus strongly on the Internet, Cisco had already disseminated its products in the market and it was hard for NEC to differentiate its products and find a niche market. At the same time, the hardware and especially the software had become complex. Cisco had proprietary IOS intellectual property based on TCP/IP protocol. To simply copy Cisco products would have led to legal action from Cisco, so in 1997 NEC made the decision to distribute Cisco routers, hubs, and switches. The problem, however, was not simply Cisco’s intellectual property rights per se; Japanese vendors could develop their own router codes to get around that. The challenge was posed by the operator’s (enterprise’s) deployment and familiarity with Cisco’s router command interface (CLI). Customers wanted to insure that any new hardware and software had interoperability and compatibility with Cisco products and operations. Cisco also protects its competitive position from new entries by virtue of the extensive versions of its software that accommodate different legacy systems. No other competitor can match that connectivity.4 Seeking to keep pace with the evolving technology, NEC and Hitachi announced plans in December 2003 to jointly develop next generation routers designed for high speed Internet connections to telecom service providers. The total development costs over a three year period were estimated to be about ¥20 billion (US $180 million dollars) with half the cost to be subsidized by METI (Nihon Keizai Shimbun 2003: 3). It is remarkable (or perhaps unremarkable) that despite strong foreign and domestic criticism of METI’s old style ‘industrial targeting’, it continues to orchestrate and invest in such downstream product development activities. The situation at Fujitsu was somewhat similar to NECs, but with some variance. In the mid- and late 1990s, Fujitsu had two groups that were relevant for adoption of Internet related technologies: the Communications Systems Group that focused on sales to the telecom sector and carriers like NTT, and the Computer and Information Processing Group concentrated on sales to enterprises. The Communications Systems Group, like NEC, was accustomed to following the direction set by its lead customer, NTT, and thus saw the future as one dominated by ATM delivering ISDN services. As a consequence, it was not open to the Internet’s potential. By contrast the Computer and Information Processing Group, focused on sales to enterprises, was more open to Internet technologies. The problem here
39
Industries, technologies, and value chains was that investment decisions were made for them by top leaders higher up in the corporation who, at the time, did not appreciate the Internet’s potential and ignored the needs of emergent Internet businesses. So Fujitsu ended up investing heavily in carrier routers but not enterprise routers. In so doing, they abandoned the enterprise global market to Cisco.5 Fujitsu continued to manufacture a relatively full line of data communication products such as switches and routers and simple equipment such as repeaters and hubs. Most products, however, are designed for the domestic market, though there are some modest sales of routers and switches to SE Asia, and it does have a leading position in the global market for SONET, including an estimated 28 percent of the American market (Takemoto 2004: B3). In early 2004, Fujitsu announced it would stop developing routers by itself and would distribute routers from outside companies. Fujitsu said it would seek to rebuild its telecom business by strengthening the development of lowprice servers with router functions, distributing products of other manufacturers and exploring co-development with other large electronic companies (Nihon Keizai Shimbun 2004: 13). Cisco moved with incredible speed to build market share, using its elevated stock price to acquire start-ups that had the desired sets of skills and product lines. Originally they had made only a small part of the router but with the build up of their capabilities through mergers and acquisitions, they could market the whole router. As a result they secured tremendous first mover advantages as discussed above. None of the large vertically integrated Japanese electronic companies were capable of that kind of speed. The big US players like HP, IBM, and Lucent were somewhat faster to recognize the significance of the Internet than the large Japanese electronic companies, but they too were not fast enough to respond effectively to Cisco. The pace at which NTT came to recognize the significance of TCP/IP undoubtedly would have accelerated if the conditions had been created for the entry of new firms committed to innovation; they would have put pressure on NTT and would have provided the opportunities for acquisition that were available to Cisco. Even if NTT had speeded up its timetable for recognizing the importance of TCP/IP, however, what NTT did or did not do would not have mattered as much if the institutional field for networking had been augmented by new firm entries. Such firms would not have been constrained by the traditional mission of ‘five nines’ reliability, or the commitment to building on existing competencies by using the public telephone network, or by NTT’s past political commitments to ATM. Nationally, the conditions favourable for market entry by new venture start-ups, however, had not been created. To be sure it can be argued that from the time information about the foundations of the Internet first came to the attention of NTT researchers to the point in the late 1990s when TCP/IP and the Ethernet became accepted as the de facto standard for networking, NTT made–from a historical
40
The telecommunication industry perspective–a rather quick adaptation to a discontinuous technology. Shigeki Goto makes this argument and he certainly is correct.6 From the perspective of coping effectively with the rapidly changing competitive environment, however, one can make an equally strong argument that the process of adaptation was painfully slow.
Setting standards Of particular relevance is NTT’s decision to have the entire ATM network go through the ITU (International Telecommunications Union) standardization process prior to implementation; this consumed a great deal of time. ITU is the international organization within the United Nations’ System where governments and private sector companies coordinate global telecom networks and services. The Europeans, who traditionally liked to standardize around one technology, have been particularly active in ITU. The ITU has been known for its very slow standardization process whereby protocol suites are discussed at face to face meetings and then put out to review by mail. Each country’s representatives can propose the ultimate solution that they would like and this leads to a long iterative process of negotiation and further discussion before a common standard is finally selected. Moreover, to satisfy as many of these constituencies as possible, the solutions have tended to involve complex protocol solutions that are difficult to implement. NTT had long been committed to working with ITU and it was only natural that they would continue to do so with ATM. If researchers/engineers working on international standards wanted to be promoted at NTT it was fully expected that they would work with ITU. A quite different approach, however, was taken by the IETF (Internet Engineering Task Force) in which Americans have been the most active members. This organization is made up of volunteers from the international community. The IETF had its beginnings in 1986 at a meeting in San Diego attended by 21 individuals and has evolved into the principal body engaged in the development of new Internet standard specifications. It holds three meetings a year but the primary goal at these meetings is not to set policy or to agree on standards but to reinvigorate the working groups where the vast majority of IETF work is done. Each working group has a charter and a chair whose job it is to keep the discussion on track. Much of the work effort is done between meetings relying heavily on online communication; emergent policies are by rough consensus. Proposed specifications are repeatedly discussed and their merits debated in open meetings and especially in public electronic mailing lists and made available via worldwide online directories. One joins a working group by subscribing to the mailing list for a given group. The face to face working group meetings are much less important than the need to gain consensus on
41
Industries, technologies, and value chains the working group mailing list. Every IETF standard is published as an RFC (Request for Comments) and every RFC starts out as an Internet Draft. Internet Drafts can be submitted by an individual or as a product of a working group. After an Internet draft has been sufficiently discussed, and if there is rough consensus that it would be a useful standard, it is presented to the Internet Engineering Steering Group (IESG) for consideration. The IESG, after soliciting still further comments, has the authority to approve the draft as a Proposed Standard. Before this happens, however, it is ‘highly desirable’ that there be independent interoperable implementations of each part of the standard. The more the number of successful independent interoperable implementations, the more substantial the operational experience, and the more the candidate specification is used in increasingly demanding environments, the more likely the draft will be accepted and eventually become a Draft or Internet Standard. Most of the standards in common use are Proposed or Draft Standards that have never moved forward to the status of ‘full’ Internet Standards. Often this is because people found more important projects to work on or the specification is not considered all that important. This suggests a much less bureaucratic organization than the ITU. Moreover, many of the protocols that are standardized at the IETF have been relatively simple compared to ATM, and required fewer people to work on them. The simplicity of the Internet architecture and the low expectations of performance (relative to ATM) make this possible.7 Moreover, standards were approved even more swiftly in the early days of the IETF when the numbers participating were still smaller. Over time, and with the growing interest of commercial organizations, it inevitably is becoming more structured in its procedures. As a result of these working routines, IETF has been much more responsive to real time market forces in its development of new Internet standard specifications than was ITU in its development of ATM standards. As we have seen, the Internet community’s support for a given standard, ceteris paribus, is stronger for technology solutions that have been successfully deployed. Letting the market place decide the winners, while not without its problems, tends to be a faster process. This case is also instructive for researchers who commonly distinguish between de jure standards created by committees and de facto standards created by markets (Besen and Farrell 1994; Funk 2002: 1). The IETF process shows that a committee-based approach facilitated by online communication can be a powerful force that is quite in line with market forces (see Shapiro and Varian 1999). Finally, Japanese researchers played only a minor role in setting IETF’s Internet standards in the critical early years (1986–96). While the number of Japanese attending IETF meetings has grown in recent years, they participate mostly as observers rather than as active members. That is telling in terms of their continuing status as followers not leaders. Failure to master the standard setting process can leave ICT firms at an acute competitive disadvantage. This is because firms and nations typically seek to use the global standard setting
42
The telecommunication industry process to enhance their own competitive positions by seeking to incorporate into the global standard their solutions or solutions that are compatible with their capabilities.
Conclusion The decades of the 1980s and 1990s saw a huge opportunity missed with Japan’s failure to deregulate telecom and break up NTT’s monopoly power. As a result, NTT was able to keep prices high (therefore limiting the spread of the Internet) and to push its own proprietary technologies. This delay severely retarded the development of ICT infrastructure and that, in turn, led to a retarded development of ICT products and services. Institutional rigidity and ill-conceived decisions regarding standard setting have clearly slowed the growth of Japan’s ICT sector. The absence of a hospitable environment for new ventures and the constraints imposed by relationship contracting on key electronic firms by their ties to an NTT committed to ATM technologies slowed the private sector’s embrace of the Internet and related networking technologies. As a result, US firms reaped huge first mover advantages in the global network equipment industry. With respect to Internet related patents, the US holds a commanding lead over Japan with 4.6 times as many patents (Soumusho 2003: 41). To be sure, the advantages held by Japan’s competitors are by no means unassailable. There is growing deregulation in the Japanese telecom market with real competition being provided by up-starts like Softbank. Japanese firms need to learn to participate more effectively in the international standard setting process, however. The weakness they have demonstrated in the past in setting global standards, whether it be a function of poor English capabilities, or a naive belief that the best technology will always win, or their insularity, or their own engineering arrogance, or a failure to join the shift from a traditional committee-based approach to the more dynamic IETF model, or some combination of all of these factors, has put them in a disadvantaged position. There are larger forces on the horizon that contain the seeds of a Japanese revival. While US firms are superior to Japanese firms in the conventional PC-centred Internet and related technologies of content, production, and security, many experts predict that the future will be one of ubiquitous networks centred on mobile communications technology. It is here that Japanese firms have significant advantages. Their wireless technology is quite sophisticated. Mobile communications technology requires terminal technology for overcoming restrictions of receiving devices and of terminals; it also requires optical technology for overcoming performance problems. These are areas in which the Japanese are quite strong (Soumusho 2004: 10–11). Wireless networks and ubiquitous computing using cell phones provide Japanese firms an opportunity to break Cisco’s near monopoly on the network equipment business.
43
Industries, technologies, and value chains There are many US researchers, however, who believe that the future lies in wireless sensor nets providing the key platform, not the mobile phone. They envision a smorgasbord of seamlessly interfaced wireless technologies such as WiMax (long-range wireless networking), WiFi (short-range networking), and self-configuring self-powered mote networks enabling a range of wireless applications. Should the cell phone be less central to the emergent solutions then Japanese potential advantages will be correspondingly reduced. In the networking area, although their global market share is very low, Japanese producers have maintained a high level of technology through servicing domestic firms so that should an opportunity occur, they are in a position to capitalize on it. Japanese firms and government officials are exploring possible alliances with Chinese and Korean firms to challenge Cisco’s dominant position, and were they to unite on setting technology standards, their international clout would be formidable. We have attributed Japan’s problems to a complex set of factors around the management of technology ranging from NTT management making the wrong technology bets on ATM and ISDN, institutional rigidities within NTT, damaging approaches to standards, to a failure to move more strongly to restructure the slow moving monopolistic NTT. There are those, however, who attribute Japan’s decline in telecommunication global markets to the weakening of NTT. To those proponents we can ask the following question. Would Japanese consumers have experienced the recent rapid spread of DSL and rapidly falling prices if NTT had been stronger? The answer is clearly, no. Indeed, the consequences of an even more rapid restructuring of NTT accompanied by the strong encouragement of new entries and the rapid introduction of new technology might have led to even more positive outcomes for the Japanese telecom sector and consumers. Such initiatives would have produced wide benefits for Japan’s whole economy and society. AT&T’s standing in the US now is but a shell of its former formidable organizational body. Who can deny, however, that the US, despite the excesses of the 1990s and resultant over-capacity in fibre networks, has emerged much stronger in global telecommunications as a result of the AT&T breakup. Such is the power of ‘creative destruction’ as envisioned by Joseph Schumpeter. It is one of the features of industrial evolution that market forces sometimes may lead to better management of technology than governments’ efforts to protect incumbents and micro-manage business decisions.
Notes 1. Lecture by John Chuang, 10 Sept. 2001, UC Berkeley. John Chuang is Professor of Information Management Systems at UC Berkeley.
44
The telecommunication industry 2. I am indebted above all for this account of Internet development at NTT to Dr Shigeki Goto of the School of Science and Engineering, Waseda University and formerly of NTT, and also to Ken Murakami, Senior Research Scientist at NTT laboratories. 3. This section is based on an interview with Prof. Shigeki Goto, Waseda University, Tokyo, on 11 June 2003. 4. This section draws heavily on an interview with Masao Hibino, President and CEO of Magnus Communications Ltd, 23 Oct. 2003. 5. I am indebted to Haruki Koretomo, Chief Scientist, Network Systems Group of Fujitsu for this account, 22 Oct. 2003. 6. Interview with Prof. Shigeki Goto, Waseda University, Tokyo, 11 June 2003. 7. This section relies heavily on two IETF documents: Bradner 1996: 1–31; IETF 2001: 1–27. I also benefited from the observations of Ye Xia of the University of Florida, Gainesville, and Ken Murakami, Senior Research Scientist, NTT Laboratories.
References Anderson, P. and M. Tushman (1997). ‘Managing Through Cycles of Technological Change’, in M. Tushman and P. Anderson (eds.) Managing Strategic Innovation and Change, New York: Oxford University Press. Besen, S. and J. Farrell (1994). ‘Choosing How to Compete: Strategies and tactics in standardization’, Journal of Economic Perspectives, 8: 117–31. Bradner, S. (1996). The Internet Standards Process—Revision 3, Network Working Group, IETF, http://www.ietf.org/rfc/rfc2026.txt Christensen, C. (1997). The Innovator’s Dilemma: When new technology causes great firms to fail, Boston: Harvard Business School Press. Dore, R. (1987). Taking Japan Seriously, Stanford: Stanford University Press. Fransman, M. (1995). Japan’s Computer and Communication Industry, New York: Oxford University Press. Funk, J. (2002). Global Competition Between and Within Standards, Hounsdsmills, Basingstoke, Hampshire, UK: Palgrave. IETF (2001). The Tao of IETF: A novice’s guide to the Internet engineering task force, http://www.ietf.org/tao.html International Data Corporation (2000). Worldwide Computer Networking Equipment Supplier Shipment Share, Framingham: IDC. Matsuo, T. (2003). Policy Strategies for Knowledge-driven Growth—The OECD approach, Washington DC: OECD, 22 January, L: 1–49. Messerschmitt, D. (2000). Understanding Networked Applications, San Francisco: Morgan Kaufmann. Nihon Keizai Shimbun (2003). ‘NEC, Hitachi Kyoudou Kaihatsu’ (NEC and Hitachi in Joint Development), Nikkei Shimbun, 17 Dec.: 3. Nihon Keizai Shimbun (2004). ‘Fujitsu, Routa Gaibu Choutatsu’ (Fujitsu will Procure Routers from Outside) Nikkei Shimbun, 20 Feb.: 13. OECD (2003). OECD Communications Outlook, Paris: Organisation for Economic Co-operation and Development.
45
Industries, technologies, and value chains Oie, Y., S. Goto, K. Konishi, and S. Nishio (2001). ‘Intanettou Dai 1 sho: Intanettou Nyumon’ (The Internet Volume!: An introduction to the Internet), Tokyo: Iwanami Shoten. Rhoads, C. and C. Hutzler (2004). China Telecom Forays Squeeze Struggling Rivals, WJS.com, http://online.wsj.com/artricle/),SB109459495670311660,00.html ** Semilof, M. (2000). Cisco’s Challengers, 16 October, http://www.crn.com. Shapiro, C. and H. Varian (1999). Information Rules, Boston: Harvard Business School Press. Soumusho (2003). Information and Communication in Japan 2002 White Paper, Tokyo: Ministry of Public Management, Home Affairs, Posts and Telecommunications. Soumusho (2004). Reference on Perception of Superior Technologies in J vs U.S., p. 10. Takemoto, Y. (2004). ‘Fujitsu Expects 16% Rise in U.S. Sales’, International Herald Tribune, 13 May: 13. Temin, P. (1987). The Fall of the Bell System, New York: Cambridge University Press. Tilton, M. (2003) ‘Ideas, Institutions, and Interests in the Shaping of Telecommunication Reform: Japan and the USA’, in L. Weiss (ed.) States in the Global Economy, New York: Cambridge University Press. Vogel, S. (1997). ‘Telecommunication Reform in Japan’, in Japan Information Access Project, Japanese Deregulation: What you should know, Proceedings (April): 143–53. Yamashita, M. (2004). ‘Towards the Emerging Ubiquitous Networks Society’, Japan Economic Currents, 43(April): 1–3, 8. Yamazaki, R. (2003). ‘Kesu Sutadei Shisuko Shisutemuzu: Gyakukyo o bane ni kousei’ (Case Study Cisco Systems: Aggressiveness After The Crisis), Nikkei Business, 26 May: 49–50.
46
3 Modular production’s impact on Japan’s electronics industry1 Timothy J. Sturgeon
Global integration has accelerated the worldwide flow of knowledge and information, causing societies to become embedded in one another in complex ways, even as they retain their distinctive characters. This chapter examines the process of global integration through the lens of national industrial models–the collection of routines and strategies generally shared by corporate managers in a particular society. Some might question the notion of national industrial models, rightly pointing to diversity among firms based in a specific society. All Japanese firms, for example, are not the same (Suzuki 2004). I would agree with Berger’s (2005) assertion that managers face ‘open pathways’ and so can and do choose a range of strategies. Nevertheless, societies continue to have distinct cultures, institutions, and histories, and so differences persist in the face of global integration in ways that profoundly shape corporate strategy. In the course of sustained field research on the locational and organization strategies of more than 600 firms in a variety of industries and countries conducted by a team of researchers at the MIT Industrial Performance Center during the period 1999–2005, such national characteristics were evident.2 At the same time, the managers interviewed were clearly making choices based on what they perceived companies in other societies to be doing. It is this process of outside pressure, reflection, and response that is at the heart of this chapter’s analysis. As a window into the process of global integration, the chapter develops a stylized account of Japanese electronics firms’ response to a new organizational model emanating from the United States: the Modular Production System. The account is stylized both because it is intended to tell a general story about the Japanese electronics industry and because we are required to omit firm-specific data collected during our field interviews to protect the personal and corporate confidentiality of our respondents. There were many differences as well as similarities in the strategies chosen and concerns
47
Industries, technologies, and value chains expressed during our interviews, and an analysis of these differences would doubtless be fruitful, but my aim here is to highlight areas of agreement and similarity.
The evolution and circulation of industrial models Industrial models consist of a range of norms, practices, routines, and tendencies. As such, they are always stereotypical, and firms vary widely in how closely they hew to the stereotype. Nevertheless, industrial models have been characterized at a variety of levels. Some have coalesced at specific historical moments across a wide range of countries, such as the post World War II ‘social contract’ between labour and capital that emerged in the United States, Europe, and Japan in the post World War II period, albeit in different forms. Others have been associated with groups of countries; individual countries; industries in specific countries; regions within countries; and even individual large firms. The core idea that has emerged from this work is that alternative viable forms of industry and corporate organization can co-exist within capitalism (Berger and Dore 1996). But what of the notion that global competition and integration accelerate the degree to which industrial models influence one another? There is a rich body of literature in this area as well, from work that documents the rise and uneven geographic spread of ‘financialization’, the increased responsiveness of publicly traded firms to pressure from financial analysts and large institutional investors (Lazonick and O’Sullivan 2000; Williams 2000) to research on how the collection of ‘best’ industrial practices known as Lean Production (Womack et al. 1990) have been differentially adopted and adapted by managers from different societal home bases and in different industrial sectors (e.g., Abo 1994; Liker et al. 1999). The central message here is that industrial models are not static but evolve with time, and that the pace of transformation tends to accelerate when practices are transferred from one society to another. The Japanese Production System, for example, emerged in the 1950s and 1960s as Japanese firms adapted the principles of ‘Fordist’ mass production to the constraints of the post World War II Japanese economy, namely small markets, scarce capital, and limited consumer spending power (Sayer 1986). Because of the success of Japanese firms in the 1980s, some of the key principals of the Japanese Production System in turn had a profound impact on the organization of industrial production in the United States and Europe in the 1990s. But the elements of Lean Production were introduced into societies with very different institutional structures and industrial histories, and so the process has been one of adaptation and transformation rather than of simple imitation and adoption.
48
Modular production’s impact There is now a rich literature on how the Japanese Production System has been adopted and adapted differentially in various industries, companies, workplaces, and stages of the value chain (e.g., Kenney and Florida 1993; Abo 1994; Liker et al. 1999; Holweg and Pil 2004). American firms did respond to the Japanese Production System, and the MIT book that codified its elements as ‘Lean Production’ (The Machine That Changed the World), was extremely influential among managers in the automotive industry and beyond. North American investments by Japanese firms in the 1980s and 1990s also did much to expose managers and workers at American suppliers to key elements of the Japanese Production System. These lessons resulted in an increased focus on quality at American firms, achieved through systematic and continuous defect reduction programmes and reduced in-process inventories (Cole 1999). In the realm of industry organization, however, the value chain elements of Lean Production that admonished lead firms to ask more from their suppliers dovetailed with other forces in the United States that were both driving and enabling increased outsourcing. I will refer to the industrial model that emerged from this process as the Modular Production System.
Value chain modularity as a response to the Japanese production system Outsourcing became extremely popular in the United States in the1990s, and it was driven by some of the same motivations that exist in Japan: the search for greater flexibility in the face of increased international competition and market volatility through the transfer of fixed assets and inventory to suppliers. A close lead firm–supplier relationship was a key aspect of the Japanese Production System. Japanese lead firms tend to be relatively vertically integrated, and when suppliers are heavily used, they are more likely to be highly dependent on one or a small number of key customer firms. Buyer–supplier relationships have traditionally been canted towards affiliates of the same industrial group, or keiretsu. The qualification process for new suppliers (Japanese and non-Japanese) can be extremely lengthy. Lead firms may make equity investments in their suppliers and can in some cases come to dominate them financially.3 Lead firms often provide the required technical assistance and financial support to help affiliated suppliers adopt asset-specific production technologies, inventory management, capacity planning, and quality control systems. These tight linkages between lead and suppliers have been identified as a source of competitive advantage for Japanese firms (Dyer 1996). While in the United States outsourcing grew beyond anything that had been imagined in Japan, one striking difference was that relationships with
49
Industries, technologies, and value chains suppliers did not change their adversarial tone, but retained much of their arms-length, short-term, and contractual character (Helper 1991). Nevertheless, the challenges of transferring and coordinating complex and sensitive information along the supply chain, reducing in-process inventories, and ensuring quality remained. Here American industry drew on its long history of systems integration, ‘the art of conceiving, designing, and managing the development of large systems involving multiple disciplines and many participating organizations’ (Sapolsky 2003: 31).4 Elsewhere I and others have characterized the new model that emerged in the United States during the 1990s, in part as a response to Lean Production and in part as a response to home grown pressures to ‘re-engineer’ the corporate landscape, as the ‘Modular Production System’. It is based on value chain specialization, formalization of value chain linkages, and an increase in the scale and global reach of each horizontal segment–or ‘module’–of the value chain.5 In modular value chains distinct breaks in the chain of activities tend to form at points where information regarding product and process specifications can be highly formalized. As in modular product design, activities tend to remain tightly integrated and based on tacit linkages within functionally specialized value chain nodes. Between these nodes, however, linkages are eased by the application of widely agreed upon protocols and standards. Discrete nodes of tacit activity can reside within divisions of the same firm, but only when activities are outsourced can scale economies build up beyond the level of the firm (Langlois and Robertson 1995). According to Pavitt (2003), the robustness of systems integration in the face of growing complexity in the realm of commercial products has been enabled by advances in information technology, especially computer simulation technologies that reduce the cost of experimentation and technological search. This has enabled the development of simplified and codified methods for transmitting detailed product and production information along the value chain. Specifically, the key business processes that have been computerized are product design (e.g., computer aided design), production planning, and inventory and logistic control (e.g., enterprise resource planning), as well as various aspects of the production process itself (e.g., assembly, test and inspection, material handling). Furthermore, the Internet has provided an ideal vehicle for sharing the data generated and used by these systems. Such technologies and practices are at the core of the Modular Production System. It is the formalization of information and knowledge at the inter-firm link, and the relative independence of the participating firms that gives value chain modularity its essential character: flexibility, resiliency, speed, and economies of scale that accrue at the level of the industry rather than the firm (Sturgeon and Lee 2005).6 Value chain modularity introduces risks as well as benefits for participating firms. Responsiveness may suffer as contracts are hammered out. There is
50
Modular production’s impact potential for intellectual property and other sensitive information about product features, pricing, production forecasts, and customers to leak to competitors through shared suppliers. The ability of lead firms to innovate and design successive product generations may suffer from the atrophying of manufacturing and component knowledge, a problem that has been referred to by Chesbrough and Kusunoki (2001) as the ‘modularity trap’. Reliance on standard interfaces may lead to the use of standard components, leading in turn to a loss of product distinctiveness. Shared and overlapping inventory resident in supplier organizations can lead to distortions and tracking problems that introduce waste. How the risks and benefits of the Modular Production System balance out depends, like all things in business, on execution. Both lead firms and suppliers must perform their respective tasks well, anticipate problems before they occur, and deal with them effectively when they inevitably do. One unavoidable issue is that independent firms in buyer–supplier relationships often have competing interests. In sum, there appears to be no single best way to organize production. Takeishi and Fujimoto (2003) argue that firms and industries that make products with integral product architectures7 tend to have integral value chains, while firms and industries that make products with modular product architectures tend to have modular value chains. But value chain architecture is not always a function of design architecture. As Baldwin and Clark (2000) have shown, there are many cases in which break points in modular value chains have been willfully engineered. While products with highly modular design architectures such as the personal computer certainly make value chain modularity more likely, even a single modular link in the flow of activities, such as the link between an integral design and manufacturing, can unleash the dynamics of value chain modularity. In addition, firms such as Autodesk, Cadence, and Mentor Graphics have aggressively created new opportunities for value chain modularity by developing and marketing design automation tools that produce files in standard format. The degree of modularity in a given chain of activities thus involves a large measure of strategic choice, and is not a simple function of design architecture. The question, then, is how well a given industrial model fits with the product, industry, and larger institutional and competitive factors that influence firm strategy. What is clear is that national scale institutions heavily influence managerial choices related to industry organization. For example, corporate responses to intensified competition and market volatility depend on the strength of worker protection and how well the institutions of new firm formation function at the level of the nation-state. In the United States, volatility is high, not only because globalization and technological change displace workers in vulnerable industries, but also because worker protection is relatively weak
51
Industries, technologies, and value chains and labour unions have been in serious decline for decades (only 8.5 percent of the United States private sector workforce is unionized). On the other hand, the financial and regulatory mechanisms that enable rapid entrepreneurship and corporate restructuring are very strong, and so volatility has spurred the formation of new kinds of businesses that focus on the pooling and rapid redeployment of workers and machines. Today, lead firms in the United States can lease almost anything, from workers to trucks to entire factories by making a phone call to Adecco, Ryder, or Solectron. These large, specialized suppliers have arisen in direct response to increased volatility. In countries where worker mobility is lower, such as in Japan (Brown et al. 1997), the infrastructure and motivation for new firm formation tends to remain underdeveloped, and so volatility is weakly translated into industry re-organization and modular suppliers of pooled resources have not emerged. In the American electronics industry, value chain modularity took shape during the late 1980s and early 1990s. Because many established firms had in-house manufacturing and components divisions, this change required the breakup of vertically integrated corporate structures and the aggregation of cast-off activities in suppliers. Hewlett Packard and IBM led the way, selling most of their worldwide manufacturing infrastructure to contract manufacturers such as Solectron and Flextronics, or spinning off internal divisions as merchant contract manufacturers, as IBM did with its Toronto manufacturing complex in 1997, creating the contract manufacturer Celestica. Another source of growth in contract manufacturing was increased business from newer firms that never built up internal manufacturing divisions, such as the Internet switch company Cisco and the computer workstation and server firms Sun Microsystems and Silicon Graphics. Circuit board and final product assembly work was mostly transferred to contract manufacturers based in North America, specifically the big five ‘electronics manufacturing services’ (EMS) firms Flextronics, Solectron, Sanmina-SCI, Jabil, and Celestica, while the assembly and even some of the design of notebook computers went to ‘original equipment’ and ‘original design’ (OEM and ODM) contract manufacturers based in Taiwan, such as Quanta, Compal, Inventec, Hon Hai (Foxconn), and the various contract manufacturing arms of Acer. By the end of the 1990s, much of the manufacturing capacity of the Taiwan-based contract manufacturers had shifted to mainland China, and the big five US-based contract manufacturers had established a global-scale network of factories (Sturgeon and Lester 2004). At the level of components, the 1990s was a time of rapid growth among ‘fabless’ semiconductor design firms as well as the semiconductor foundries that serve them, such as the Taiwan-based TSMC and UMC, as well as IBM (Linden and Somaya 2003). Thus, by the end of the 1990s, the Modular Production System in the United States’ electronics industry had become fully developed and global in scope.
52
Modular production’s impact
Competitive challenges to Japan’s electronics industry at the end of the 1990s Value chain modularity came to the attention of Japanese electronics firms in the late 1990s, triggered by the fantastic growth of the Internet and the huge demand for data communications and Internet enabled enterprise computing equipment that came with it. American firms are leaders in nearly all Internetrelated electronics hardware and software product categories; Cisco Systems and Juniper Networks in Internet routers and switches; IBM and Sun Microsystems in powerful computer servers; Dell in personal computers; EMC in storage arrays; Microsoft and Netscape in Internet browsers; Yahoo! and Google in Internet portals and search engines; Amazon and E-Bay in Internet retailing and auctions; and Accenture, Price Waterhouse, McKinsey, and IBM in Internet enabled corporate computing networks. Japanese electronics firms are focused on components, stand alone consumer electronics devices, and proprietary enterprise computing systems that connect client sites through private leased data lines.8 The sudden rise of the Internet, and almost complete lack of any driving role for Japanese electronics firms in this rise, combined with severe financial losses, initiated a period of questioning in the Japanese electronics industry. Cisco Systems, based in California, jumped to an early lead in the market for Internet (TCP/IP) protocol switching equipment. Through a combination of technological excellence and a shrewd and efficient acquisition strategy, Cisco managed to accrue and maintain an 80 percent market share in Internet routers while continuing to drive innovation in the field (Mayer and Kenney 2004). As they rushed to learn about the Internet, Japanese firms looked to Cisco and saw some very striking features. First, Cisco relied almost entirely on third-party systems integrators such as Accenture and McKinsey for the creation of fully functional Internet enabled data networks and enterprise computing systems. Even more striking from the Japanese point of view was that Cisco did not directly produce its own equipment, but relied on contract manufacturers such as Solectron and Flextronics. Cisco’s success was based on its ‘platform leadership’ (Gawer and Cusumano 2002), that is, its ability to drive the standard setting process through technological and market leadership while leveraging the capabilities of its suppliers and customers. The major Japanese electronics firms, on the whole, are much more vertically integrated, with in-house design and manufacturing of many subsystems and components. In 2001 competition from American firms in modular production networks formed only part of the challenge facing Japanese electronics firms. Korean firms such as Samsung, LG Electronics, and Hyundai are highly vertically integrated. Similar to Japanese firms, large Korean electronics firms tend to follow the ‘components plus products’ strategy; they manufacture and sell
53
Industries, technologies, and value chains components on world markets, and use their most advanced components first in their own branded products to the highest degree possible. Until the late 1990s, Japanese companies followed the ‘flying geese’ strategy of licensing older component technologies to less capable firms in Korea and Taiwan and moving to newer technologies without much worry, but by 1999 Korean firms, especially Samsung and LG, began to close the gap in specific consumer electronic and component markets, such as mobile phones, digital cameras, digital televisions, computer monitors, high capacity memory chips, and flat panel displays. In Japan, intensified competitive pressure from both the United States and Korea fostered the widespread impression that Japanese electronics firms were losing pace. This, along with losses at several firms in 1998, focused managerial attention on the practices of rival firms and fostered the consideration of radical shifts in strategy. The build-up of the Internet bubble, and its bursting in 2001, whipsawed Japanese electronics firms along with the global industry, not because Japanese firms were driving innovation in the field, but because they were significant suppliers of components, personal computers, and computer peripheral equipment, the sales of which were being driven by the expanding Internet. As a result, the near moratorium on IT spending that followed the excesses of the Internet bubble deeply affected Japanese firms along with the rest of the industry. But the losses posted in 2001 and 2002, while very large, were this time accompanied by even greater losses at the North American firms that had been most caught up in the mania of the Internet boom, such as Lucent, Nortel, JDS Uniphase, and Solectron. Table 3.1 summarizes the financial performance of the ten largest Japanese electronics firms during the period 1997–2004.
Table 3.1 Net income (loss) of the ten largest Japanese electronics firms, 1996–2004, US$million Firm name
1996
1997
1998
1999
2000
2001
2002
2003
2004
Matsushita Sony Fujitsu NEC Toshiba Hitachi Canon Mitsubishi Sharp Sanyo Top 10
1,228 1,243 411 827 598 800 839 76 433 157 6,612
764 1,812 46 387 60 41 970 (864) 202 101 3,517
107 1,409 (107) (1,190) (109) (2,652) 862 (351) 36 (204) (2,199)
898 1,098 385 94 (252) 152 633 224 253 195 3,681
376 152 77 513 871 946 1,215 1,131 349 366 5,995
(3,427) 123 (3,064) (2,499) (2,035) (3,876) 1,342 (625) 91 11 (13,959)
(160) 948 (1,002) (202) 152 229 1,566 (97) 268 (506) 1,196
374 785 441 364 256 141 2,446 398 539 119 5,862
545 1,527 297 632 429 480 3,200 663 716 (1,599) 6,892
Source : Company reports Notes : Except of Cannon, dates are approximate calendar years ending on 31 March of the year following the year listed. US$ figures were derived from average currency ‘ask’ prices for the period 1 April of the year listed through 31 March of the following year. Currency pricing was obtained from http://www.oanda.com
54
Modular production’s impact
The response to modular production: A new Japanese model?9 This section examines the strategic responses of Japanese electronics firms to Modular Production in the period 2000–04, responses made in the context of the competitive and financial challenges discussed in the previous section. Our interviews (as part of the IPC Globalization Study) reveal that Japanese electronics firms have been strongly influenced by Modular Production but that they have, unsurprisingly, resisted certain aspects of the model while adopting and adapting others. What emerged most powerfully in our interviews was the depth and scope of the questioning taking place within the highest levels of Japanese electronics firms. One respondent summed up the situation in the late spring of 2001 in this way: Mega-competition means we are facing strong companies with narrow core competence, such as Micron and Dell. Such single-function players are very strong. We are an all around diversified player so we cannot fight such players with agility. We have convened a series of one-day meetings to determine how to survive. (Japanese electronics executive, June 2001)
The dilemmas and contradictions facing the largest Japanese electronics companies were great during the interview period, as they continue to be today. Japanese electronics firms are highly diversified and have large numbers of employees both in Japan and abroad. For firms selling enterprise computing systems, key customers in Japan, which prominently include national and local governments, are demanding IT systems comprised of the best hardware and software in the world, and since such systems now must be Internet compatible or even Internet-based, this often means using elements created by non-Japanese companies. For firms selling consumer electronics products and electronic components, competition is intense from low cost producers with modular value chains, such as Dell in personal computers, and with high levels of vertical integration, such as Samsung in mobile phones and flash memory chips. These pressures prompted decision makers at Japanese electronics firms to consider new strategies to rapidly acquire or develop new competencies, increase specialization, and relocate in-house operations to low cost locations such as China. At the same time, the managers we spoke to agreed that it would be politically and strategically impossible to enact the layoffs that would be required if radical restructuring was taken too far. The bursting of the Internet bubble in early 2001 led Japanese managers to step back from the brink of radical transformation. The ‘dot.com’ crash dramatically exposed some apparent weaknesses of the Modular Production System. As a result of over-anticipating demand, Cisco was forced to liquidate US$2.2 billion of finished and in-process inventory, largely held by its contract manufacturers. The company cut 8500 jobs and posted its first loss in its 11 years as a public company (US$2.69 billion) in the third quarter of 2001
55
Industries, technologies, and value chains (Niece 2005). Over the next few years Solectron, Cisco’s most important contract manufacturer, suffered a total of US$6.5 billion in losses and laid off nearly a third of its global workforce of 60,000. However effective these developments were in driving Japanese firms back to their traditional industrial model, managers at Japanese electronics firms have nevertheless made significant breaks with past practices. Only key components, such as system-on-a-chip (SoC)–known in Japan as LSI–semiconductors, leading edge flat panel displays, high capacity batteries, and advanced memory chips are to be produced in Japan, either in-house or in joint ventures with other Japanese firms. In-house final assembly in Japan is largely being limited to high-cost models with advanced features. Low end products are to be produced offshore, especially in China, either by affiliates or by Taiwanese contractors. Divestiture of old, unprofitable, and unrelated businesses and products lines has accelerated, though these moves comprise only an incremental step toward downsizing and specialization. Increased specialization, increased complexity, and the continued importance of foreign component sales has led to increased outside purchasing and heightened dependence on global markets for a wider variety of inputs, including technology inputs.10 The remainder of this section presents evidence of these changes, and their limits, in three areas: alliances, outsourcing, and information technology and communications services.
Alliances The renewal of traditional strategies at Japanese electronics firms has a high price. The fast pace of technological change in the technologies that underlie key components has required a spate of new investments in leading-edge factory production in Japan (see Table 3.2 for some examples). The high cost of many of these new investments has convinced managers to forge an unprecedented set of production sharing alliances. Seven of the 25 factory investments listed in Table 3.2 involve more than one firm. The shift in thinking about alliances is captured by the following statements made by the same top manager in 2001 and 2002: We have a terrace-house style management where we exchange ideas with people in the same house, so we don’t want to sell our factories to other people. (Japanese electronics executive, June 2001) We’re thinking of a smaller terrace house now. And we’re also thinking about having good neighbours. (Same Japanese electronics executive, July 2002)
In contrast to the technology and standards development deals forged with American and European firms in the 1980s, most of these recent agreements have been between Japanese firms. In some cases the deals are simple technology
56
Modular production’s impact Table 3.2 Examples of recent and planned electronics factory investments in Japan Investment ¥B
Planned opening
Firm name
Kind of factory
Location
Renesas Technology (Hitachi-Mitsubishi joint venture spin-off)
Semiconductors (system LSI)
Hitachinaka, Ibaraki
200
Latter half of 2005
Elpida Memory (Hitachi-NEC joint venture spin-off)
Semiconductors (DRAM)
Higashihiroshima, Hiroshima
500
Latter half of 2005
NEC Electronics
Semiconductors (system LSI)
Tsuruoka, Yamagata
100
Latter half of 2005
Toshiba
Semiconductors (flash memory)
Yokaichi, Mie
270
Latter half of 2005
Fujitsu
Semiconductors (system LSI, ASIC)
Kuwana, Mie
160
First half of 2005
Renesas Technology (Hitachi-Mitsubishi-NEC joint venture spin-off)
Semiconductors (system LSI, flash memory)
Kagami, Kochi
200
Undecided
Matsushita
Semiconductors (system LSI)
Uozu, Toyama
130
Latter half of 2005
Sony
Semiconductors (microprocessor)
Isahaya, Nagasaki
200
First half of 2005
Sony
Semiconductors (CCD)
Kyushu (undecided)
100
Undecided
Sharp
Semiconductors (flash memory)
Fukuyama, Hiroshima
50
First half of 2006
Oki
Semiconductors (undecided)
Kiyotake, Miyazaki
100
Undecided
Toshiba Matsushita Display Technology
Liquid crystals for cellular phones
Kawakita, Ishikawa
50
April 2006
Sharp
Liquid crystal displays for TVs
Kameyama, Mie
150
June 2006
IPS
Liquid crystal displays for TVs
Mobara, Chiba
110
2nd Q, 2006
Toshiba-Canon
SED displays
Taiji, Hyogo
180
January 2007
Matsushita (Panasonic)
Plasma displays
Ibaragi, Osaka
60
April 2004
Matsushita (Panasonic)-Toray
Plasma displays
95
Fujitsu Hitachi plasma display
Plasma displays
Pioneer
Plasma displays
Amagasaki, Hyogo Kunitomi, Miyazaki Tatomi, Yamanashi
September 2005 Latter half of 2006 September 2004
Konica Minolta
Polarizing film for Kobe, Hyogo liquid crystal displays
Fuji film
Film for flat panel displays
Kikuyo, Kumamoto
100
Dainihon insatsu
Film for liquid crystal panels
Kitakyushu, Fukuoka
30
End of 2006
Toppan insatsu
Film for liquid crystal panels
Hisai, Mie
50
October 2006
85 26–7 30
Autumn 2006 December 2006
Sumitomo Chemical
Polarized plates
Niihama, Ehime
10
Autumn 2006
Asahi glass
Glass plates for liquid crystal panels
Takasago, Hyogo
25
Autumn 2006
Source : Nikkei Shinbun, various dates: 2004–06
57
Industries, technologies, and value chains development and patent sharing deals between firms with complementary assets and capabilities. In other cases firms have combined component divisions and spun them off as separate companies. In still other cases firms have purchased the divisions of other firms to gain control over needed components or to build larger, more viable divisions, especially in the face of volatile global markets and fierce competition (e.g., DRAMs). Of the greatest interest and significance are eight deals listed in Table 3.3 that involve joint factory investments, where partner firms share output. Such deals require significant investment that heighten risk and make withdrawal difficult. Alliances of this kind create shared factory space, and shared risk. They move the Japanese electronics industry in the direction of Modular Production in that large fixed investments are pooled and shared by a number of industry players. But in this case the number of firms sharing capacity is limited to the members of the alliance, which is typically two and in a few cases three firms. Overall, this restructuring activity is leading the Japanese electronics industry on a path toward greater specialization, concentration, and fixed capital sharing. These are the same goals that American firms have sought as they have moved toward the Modular Production System, albeit pursued in a somewhat different and more partial way. Table 3.3 provides some examples of recent restructuring in the Japanese electronics industry, including mergers, spin-offs, acquisitions, and alliances.
Outsourcing In the realm of outsourcing as well, Japanese electronics firms have taken a partial step in the direction of Modular Production. Dense interactions between design and manufacturing is one of the hallmarks of the Japanese Production System, and much criticism was leveled at the tendency at American firms to ‘throw designs over the wall’ to manufacturing (Kenney and Florida 1993). But this was one lesson of Lean Production that went largely unheeded in the United States. On the contrary, one of the most significant challenges to the traditional Japanese system posed by Modular Production is the notion that manufacturing can be entirely separated from product development. Were American firms simply misguided or had technology enabled new ways of organizing the value chain? One respondent put the question this way: Traditionally we thought that if we don’t keep manufacturing, we can’t keep our core technological competence. US firms threw that out. This is the central question. For ‘analog’ manufacturing, where you have the in-house accumulation of technology, [outsourcing] is dangerous. For ‘digital’ manufacturing, [outsourcing] is OK. But does digital equipment eliminate the accumulation of manufacturing expertise? This is one of my questions. We need at least to keep experimental pilot plants in Japan. For manufacturing technologies, like miniaturization, there is real Japanese strength. What will
58
Modular production’s impact Table 3.3 Examples of recent restructuring in the Japanese electronics industry Partners (% share)
Year announced
Sony-Konica-Minolta
2005
Matsushita-Olympus
2005
Hitachi-Matsushita
2005
Hitachi (50%)-Matsushita (23.4%)-Toshiba (23.4%) NEC-Pioneer
2004 2004
Seiko Epson (55%)-Sanyo (45%) Toshiba-Mitsubishi Sharp-Sony Ericsson
2004 2004 2004
Casio (51%)-Hitachi (49%)
2003
Konica-Minolta
2003
Fujitsu
2002
Hitachi (55%)-Mitsubishi (45%) (Renesas Techology)
2002
Mitsubishi-NEC-Hitachi (Elpida Memory)
2002
Toshiba (60%)-Matsushita (40%) NEC
2001 2001
Sony-Toshiba-IBM
2001
Matsushita-NEC
2001
Matsushita-Toray
2000
Fujitsu-Hitachi
1999
Toshiba (50%)-Canon (50%)
1999
NEC (50%)-Hitachi (50%)
1999
Mitsubishi-Matsushita Electronic Toyota Jido Shokki (50%)Sony (50%)
1998 1997
Products
Type of deal
Digital still cameras (SLR) Digital still cameras (SLR) Flat panel displays (plasma)
Joint product development
Flat panel displays (liquid crystal) Flat panel displays (plasma) Flat panel displays (liquid crystal) Semiconductors Software for cellular phones Cellular phones
Joint product development Joint R&D, production, marketing, and intellectual property sharing Joint production Sale to Pioneer Merger and spin off Sale to Mitsubishi Joint development
Joint product development, design, and purchasing Cameras, printers, and Merger copiers, etc. Flat panel displays Spin off of division (liquid crystal) Semiconductors Merger and spin off of (system LSI) R&D, product development, production, and marketing Semiconductors Merger and spin off of (DRAM) R&D, product development, production, and marketing Flat panel displays Joint production (liquid crystal) Semiconductors (net- Spin off work applications) Semiconductors Joint product development (system LSI) Software for cellular Joint product development phones Flat panel displays Joint venture (plasma) Flat panel displays Joint production (plasma) Flat panel displays Joint R&D and production (SED) Semiconductors Merger and spin off of (DRAM) R&D, product development, production, and marketing Semiconductors Joint product development (system LSI) Flat panel displays Joint venture (liquid crystal)
Source : Trade press publications and Nikkei Shinbun
59
Industries, technologies, and value chains US core competence be if all their manufacturing goes? New technology must combine various technologies and expertise within the company. The question is: can we throw manufacturing out of this mix totally? (Japanese electronics executive, July 2001)
The general strategy in Japan has been to keep the production of leading edge products in-house but embrace outsourcing for high volume, price sensitive products such as low end personal computers, mobile phones, and previous generation video game consoles. But instead of American contract manufacturers, Taiwan-based contract manufacturers have received the lion’s share of these new orders. Taiwanese contract manufacturers were thought to have access to lower cost capital and be willing to tolerate lower returns on investment than American firms. Taiwanese manufacturers of commodity flat panel displays, personal computers, and mobile phones are especially popular. Some use of Taiwan’s semiconductor foundries (TSMC and UMC) was reported as well. Japanese managers have confined the use of contract manufacturers to older and simpler products and components because of the engineering time required to transfer specifications and because they fear the leakage of intellectual property. One respondent put it as follows: Some of our products require special components and finishes. If we used a Taiwanese [contractor], we would also use them as parts vendors, and we’d have to teach them about this [advanced process], and we don’t want to – it would take too much of our engineering time. Eventually they will be able to do everything, but we don’t want to teach them so quickly. It’s a constant decision to figure out how much of our resources to invest in teaching them as opposed to the cost of doing it ourselves. Moreover it leads to the leakage of our intellectual property. Eventually they catch up – but maybe we can delay that (Japanese electronics executive, October 2004)
This statement reveals a deep ambivalence about outsourcing that has not been as evident in American electronics firms, which tend to deal with such problems by codifying complex product specifications and punishing suppliers that try to compete with them by withdrawing business. While not unheard of, American companies have had no appreciable problems with IP leakage to rivals via shared suppliers. Managers of Japanese electronics firms, in contrast, have largely opted to continue traditional strategies that seek to develop and leverage synergies within their organizations. As one respondent put it: In can be an advantage to have both components and [final products] in-house; we can use advanced components in our own products first and introduce new features faster. If manufacturing is outsourced, 100% of the strength of Japanese companies will die. Launching new models quickly is the key. If we don’t have a manufacturing function, we will not be able to launch new products based on new [in-house] technologies, such as batteries, LCDs, and semiconductors, nor could we make modifications to existing products. The ability to make incremental modifications on the factory floor
60
Modular production’s impact is important. Dell doesn’t create. They will have a hard time creating new products because they depend on outside [contract manufacturers] that have no unique technology. Making parts and products is important. When products are commodities, then using [contract manufacturers] is OK, but advanced products are better made in-house. (Japanese electronics executive, June 2001)
Still, it was recognized that the benefits of this strategy were declining with the increased ability to codify product and process information that has come with digitization. One respondent put the problem this way: With digital technology it becomes easier to gain the capability to manufacture. It’s easier to make personal computers than televisions. Everyone can buy the technology. The machines embody the instructions. It’s no longer a ‘black box’; the Japanese advantage when it used to be that way is eroding. For example, the Koreans can simply buy the machines and have the technology. (Japanese electronics executive, July 2001)
On the other hand, in some areas the increasing consolidation of functionality enabled by digitization has created new technical challenges and a greater need to integrate product and component design efforts. The Japanese managers we spoke with believed this to be especially true in the case of system-ona-chip semiconductors.11 As one respondent put it: United States companies specialize in a core competence, a piece of the value chain. We do it all: system LSI, [product] design, manufacturing, production equipment, and marketing. The main business [of our division] is to manufacture digital audio-visual products. To do this we must co-develop with our semiconductor group. We can put all of our knowledge about system design into the LSI design. The system LSI made by a specialist may not work as well or fit as well within the final product. In the past we could buy key components from the outside, but now system LSI determines everything so we buy these from inside. But this is the exception. Other components can be bought from the outside. (Japanese electronics executive, October 2004)
Another respondent listed the benefits of in-house system-on-a-chip semiconductor production as: Speed, cost, and intellectual property protection . . . When outside vendors are used, roadmaps are leaked to competitors. Inside we don’t have that problem. (Japanese electronics executive, October 2004)
Our interviews suggest that by 2003 the questioning on the topic of outsourcing had led only to modest changes. One of the main difficulties was the work-force reductions that would be required for more radical restructuring. If we got rid of manufacturing, we’d have to get rid of 50% of our workforce. We couldn’t survive if we did that because other stakeholders, like the governments who procure our services, couldn’t accept our doing such a thing. (Japanese electronics executive, July 2002)
61
Industries, technologies, and value chains Severe workforce reductions were also seen as problematic because of their effect on morale. As one respondent put it: We can’t just fire people, because if we did, we couldn’t keep the others. This is the Japanese way of business; we can’t just adopt the American way. We have to make full use of Japanese people. If we fire the laggards, the talented promising people will think about their own future and also leave. They would think that [our company] is not a good place to work. We are building up some outside companies like real estate and maintenance to absorb excess people, but all this has to happen slowly. We are thinking of cutting some businesses, but this must be done gradually, according to the Japanese way. (Japanese electronics executive, June 2001)
To sum up, Japanese firms appear to have settled on a mixed model in the realm of outsourcing. Advanced components and products are to be produced in-house or in joint ventures, and older, simpler, and non-strategic components and products lines are to be outsourced. While this strategy might seem clear and decisive on its face, it provides no real guidance on how far to take outsourcing. What comprises a core technology, a key component, or advanced product? How soon should advanced process technologies be transferred to outside suppliers? It was recognized that the definition of core competencies and key components would shift over time. One respondent summarized this point as follows: In regard to outsourcing, we have a mixed model. We make key components in-house. We must choose these key components carefully and engage in constant search and revision. What is considered key will change over time. Then, we must choose our real high tech collaborators; firms that can provide specialties and have special R&D capabilities. (Japanese electronics executive, October 2004)
What this suggests is that Japanese electronics firms face the same strategic challenges that their foreign competitors do and have similarly moved in the direction of Modular Production, in most cases for older product lines but in other cases with the aim of developing high level technological collaborations. But even in the case of older, nonstrategic products, the migration of in-house production to low cost locations, especially China, was mentioned at least as often as outsourcing. What is clear is that the degree and speed of these changes are limited in the Japanese institutional context. The following statement sums up this point well: Suppose we do away with all of our plants and fire all of our workers? If we were driven to this we might do it, but in Japan you can’t do this. It is our policy to protect [manufacturing workers’] jobs. It is part of our mission as a company. So we must continue to develop products that cannot easily be outsourced. Putting parts together is the job of a trading company. We are not a trading company. This is why we cannot do what Apple computer has done [externally sourcing the components and assembly of its iPod digital music player]. (Japanese electronics executive, March 2004)
62
Modular production’s impact
Information technology and communications services As they lost money in the late 1990s, Japanese electronics firms with the breadth to supply large scale corporate computing systems saw a solution to their financial woes through growth in the service side of their business, following IBM’s long success in this area. An expanding information technology and communications (ITC) services business is attractive in many ways, not least in its potential to absorb a large number of employees in high valueadded, knowledge intensive work. But customized ITC systems and IT-enabled business services tend to require a deep and thorough understanding of the end user’s line of business and close collaboration to identify and fulfill the buyer’s needs. Such ‘domain knowledge’ is typically industry-specific, requiring knowledge that is applicable only to relatively narrow ‘vertical’ markets, so there is an obvious knowledge gap.12 Japanese electronics firms have very little experience providing ITC services outside of Japan. There is, moreover, a great deal of competition in the realm of ITC services, both from other integrated electronics hardware and software firms such as IBM and from services-only consulting firms, such as Accenture, Price Waterhouse, and McKinsey. Competing with foreign ITC service firms even in Japan has proved difficult, and Japanese electronics firms have found that their product lines and service offerings are not considered serious contenders. One reason for this is Japanese electronics firms’ continued bias toward using their own products for the ITC systems they sell. The provision of advanced ITC services to Japanese customers, but especially to global clients, introduces a contradiction with the traditional Japanese way of doing business. Japanese electronics firms, when providing complete ITC systems to clients, produce and thus have an incentive to supply a full range of their own hardware products, from components to PCs to servers to large computers to networking equipment, as well as software. With the deployment of global scale data communications systems, and especially since the rise of interoperability based on Internet, or TCP/IP, protocols, it has become much easier and in some cases necessary to integrate hardware and software from a variety of vendors. In fact, many customers, even in Japan, believe that their systems should be built from best products available. It has been very difficult for Japanese electronics firms to adopt this model, not least because of strongly held notions about the learning synergies between various components of large complex systems. In addition, sales forces have little or no experience selling products from outside vendors, and may well have incentive structures that discourage this practice. One executive explained this dilemma as follows: Five to seven years ago, there were no cases of United Sates [firms’] success in selling [IT systems] in Japan, but today, even local governments are choosing whoever has the best integration package. If we try to sell only our own products we’ll lose business. We do have one case where we sold a big system integration solution with no in-house
63
Industries, technologies, and value chains products. It included only American-made hardware and software. Our engineers on that project asked, ‘What company am I working for?’ But pure systems integration like this is profitable. (Japanese electronics executive, July 2001)
Debates about shifting from manufacturing to services at Japanese electronics firms have apparently been quite intense. For now, it seems as if the integrated approach has won the day. This is captured well by the following three statements made by the same high level Japanese electronics executive: Now we want to change from a hardware to a software and services solutions business, so we need more differentiation to fit customers in every country. Will turning away from manufacturing create weakness? We are struggling to find an answer. Even on government programmes, we can’t do it ourselves and so we are using some American firms as sub-contractors. Accenture, Mckinsey, and Price Waterhouse and others have a very good business in Japan and can win bids over us. Our engineers make full use of our products first. So customers prefer to go with American companies because they’ll provide integrated packages using the best components from a variety of vendors. Our engineers are trying to integrate products from Cisco and others but sales and engineering issues force them to use our own products, so we lose some bids. (July 2001) There are two different views at our company. Some say we should simply be good at choosing the best components, but others say anyone could package them in the same way. They would be standardized parts, so anyone could do the same thing. Where’s the competitive advantage for us then? With no differentiating hardware, there is no way of succeeding in a pure software/services business. How would we make profits in such a business? In this view, we need to maintain advanced hardware capabilities. (July 2002) Service companies cannot expect to make profits. We found we cannot make money from just software services. Even IBM is facing losses from its system integration business. Competition is too tough in being a pure provider of services. Therefore my opinion now is that we need to keep making all the necessary hardware in our company. Some people in our company said we should lead in services and software and use the best hardware we can find whether it’s ours or another company’s. Gradually we realized that the company that produces the key hardware in-house can provide customers with the confidence and security they need. (October 2004)
Conclusions The failure of Japanese electronics firms to participate fully in the Internet fueled growth in the global electronics industry during the late 1990s triggered a period of questioning among the top executives in Japan’s leading electronics firms. At the time, the Modular Production System emanating from the
64
Modular production’s impact United States seemed to be providing American firms with significant competitive advantages. Moreover, the key features of the Modular Production System provided a direct challenge to a host of the most cherished strategies of Japanese electronics firms. In Modular Production, manufacturing capacity is pooled in specialized contract manufacturers, freeing lead firms to engage in ‘open innovation’ (Chesbrough 2003) by specializing in specific aspects of technology development and system architecture while depending on outside firms for complementary system elements. The goal of ‘platform leaders’ (Gawer and Cusumano 2002) in the Modular Production System is to attain early market dominance to set standards in emerging technologies, thereby forcing ‘partner’ firms to create products, subsystems, and components that comply with the standards they set. These strategies clash with the strongly held belief among Japanese managers that competitive advantage comes from design collaboration within a diversified organization, tight feedback between internal design and manufacturing, and the first use of internally developed components. The bursting of the Internet bubble in 2001, and the ensuing inventory and financial problems at firms closely associated with Modular Production such as Cisco and Solectron, affirmed the skepticism of Japanese managers regarding the model. Our interviews after 2003 suggested that the pressure for radical moves away from manufacturing had lost momentum and that traditional strategies of vertical integration were being reasserted, especially for advanced products and technologies. In addition, financial performance at many large Japanese electronics companies improved in 2004, driven in significant degree by rising demand from Japanese customers with booming sales to China for products such as steel, ships, and heavy machinery. Only the threat from Korea failed to diminish. So the period of intense questioning came to an end, at least temporarily, as decisions were taken to deepen traditional strategies, especially for advanced products and technologies. Japanese electronics firms continue to have shallow, tactical alliances with foreign firms, and have reasserted their vertically integrated approach by investing in a new round of factory construction–in Japan–for key components such as system-on-chip semiconductors, advanced flat panel displays, high capacity batteries, and high-performance memory chips. There have been partial but significant steps taken in the direction of Modular Production in the form of increased specialization, outsourcing of low end products, and shared factory investments in Japan, but wholesale restructuring has been resisted, at least for now. However slowly it may be moving, restructuring at Japanese electronics firms is indeed underway, and most large firms reported reductions in their global workforces by 10–15 percent since the late 1990s, mostly through attrition. Still, this restructuring is proceeding under the substantial weight of existing organizational routines, investments, and workforces, and is being
65
Industries, technologies, and value chains driven by contradictory pressures. As a result, Japanese electronics firms are simultaneously shedding and protecting jobs, getting out of old business lines and adding new ones, opening their sourcing networks and investing in new in-house component plants, and expanding some facilities and shrinking or closing others, both off- and on-shore, in an effort to rebalance their organizations. While it is too early to determine how successful these changes will be, or if they will add up to a new and distinct industrial model in the end, there are a host of new challenges and opportunities that now face Japanese electronics firms given their recent experiments with joint technology development, production alliances, relocation, and global outsourcing. Whether Japanese electronics firms can simultaneously and quickly shed noncore business, develop new software and service competencies, and carry the weight of the substantial new component manufacturing investments, many of which are being made in the context of new and completely untested alliances, is unknown. Finally, what do recent trends in the Japanese electronics industry tell us about the global circulation of industrial models? First, firms can and do react to external pressures for change, but in a complex and increasingly integrated world, there are multiple models that are viable at any given moment, and even if a particular model is quite popular, firms receive mixed signals from the outside. For Japanese electronics firms, challenges have come not only from the Modular Production System, but also from Korean firms such as Samsung and LG Electronics, which remain highly vertically integrated. Second, the resistance to radical change is typically quite high, so changes tend to be introduced in piecemeal fashion. It is this process of partial application, experimentation, and reversal, however, that could work to create a new model, one well adapted to the industrial history and institutions of Japan as well as to the exigencies of global markets.
Notes 1. This chapter is based on research funded by ITEC (COE) at Doshisha University, Kyoto, and the Alfred P. Sloan Foundation. The field research was conducted by the author, other members of MIT’s Globalization Study Team (see http://ipc-lis.mit.edu/ globalization/main.html) and Yoshiji Suzuki of Doshisha University. Clair Brown and Gregory Linden at UC Berkeley, Martin Kenney at UC Davis, and Mon-Han Tsai and Kazushi Nakamichi at ITEC, provided important insights and valuable support, as did Jun Kurihara of the John F. Kennedy School of government at Harvard University. Hugh Whittaker and Robert Cole provided helpful suggestions for improving the text. All responsibility for the final text, of course, resides with the author. 2. At the time of this writing MIT’s Globalization Study Team had conducted 622 field interviews in 19 countries, including 42 interviews with managers of electronics companies in Japan.
66
Modular production’s impact 3. Although this pattern of cross-holding has been quite strong historically, the keiretsu structure has loosened considerably in the past decade or so, in part driven by the efforts of foreign investors, such as Renault, to drive down the cost of components. See Lincoln, Chapter 12, in this volume. 4. Systems integration developed as a formal practice in the United States during the Cold War in response to a need to coordinate the invention, development, production, deployment, and maintenance of increasingly complex and exotic weapons and aerospace systems. Projects to create complex weapons systems such as ballistic missiles, early warning radar systems, and nuclear submarines were so large and interdisciplinary that detailed knowledge required to design and produce all of the subsystems were far beyond the scope of knowledge and expertise contained within any single military branch, firm, university laboratory, or other single organization. Someone had to make sure the systems worked as intended. At first, the task of systems integration fell to a few aerospace contractors, government agencies, and specially created nonprofit agencies, but over time the approach and methods of systems integration migrated to the private sector as private military contractors gained experience with the approach and systems engineering and management became established, if much maligned, academic disciplines by the 1970s (Johnson 2003). 5. Baldwin and Clark 2000; Sturgeon 2000, 2002; Takeishi and Fujimoto 2001; Langlois 2003; Principe et al. 2003; Gereffi et al. 2005; Sturgeon and Lee 2005. 6. The high volume of nonprice data flowing across the inter-firm link differentiates modular value chains from simple markets. Because of this complexity it is not unusual that additional engineering and coordination be required. The hand off of product and process specifications between firms need not be perfectly clean, but only relatively so for modular value chains to function. 7. Products with integral architectures have tight design interdependencies with components and subsystems of which they are comprised. 8. See Cole, Chapter 2, for a detailed account of the Japanese response to the Internet and the weakness of Japanese firms in the network equipment sector. 9. This section is based on several rounds of interviews with top managers at Japan’s largest electronics firms. The interviews were semi-structured in that the same themes were covered. They were conducted at the respondent’s office, and typically lasted 1–2 hours. The names of the firms and managers are withheld for reasons of confidentiality. The respondents typically, but not always, occupied high level decision making positions at their firm. 10. For example, most large Japanese electronics firms have licensed processor cores, a modular block of design code (or ‘IP block’) for inclusion in SoC semiconductors, from the British firm Advanced RISC Machines (ARM) as a way to stimulate business in Europe, where ARM technology amounts to a de facto standard for embedded communications equipment. 11. This is in contrast to American lead firms, which commonly source their SoC semiconductors externally or do the logic design in-house and outsource the remaining design and fabrication tasks (Greg Linden, personal communication, September 2005). 12. See Rtischev and Cole (2003) for an analysis of the not always wise penchant of large Japanese firms to try to use expanding businesses to absorb redundant labour.
67
Industries, technologies, and value chains
References Abo, T. (ed.) (1994). Hybrid Factory: The Japanese production system in the United States, Oxford: Oxford University Press. Albert, M. (1991). Capitalism vs. Capitalism, New York: Four Wall Eight Windows. Baldwin, C. and K. Clark (2000). Design Rules, Cambridge, MA: MIT Press. Berger, S. and R. Dore (eds.) (1996). National Diversity and Global Capitalism, Ithaca, NY: Cornell University Press. —— (2005). How We Compete; What companies around the world are doing to make it in the global economy, New York: Doubleday. Brown, C., M. Reich, L. Ulman, and Y. Nakata (1997). Work and Pay in the United States and Japan, New York, NY: Oxford University Press. Chesborough, H. (2003). Open Innovation: The new imperative for creating and profiting from technology, Boston, MA: Harvard Business School Press. —— and K. Kusunoki (2001). ‘The Modularity Trap: Innovation, technology phase shifts, and the resulting limits of virtual organizations’, in L. Nonaka and D. Teece (eds.) Managing Industrial Knowledge, London: Sage. Cole, R. (1999). Managing Quality Fads: How American business learned to play the quality game, New York: Oxford University Press. Dore, R. (1986). Flexible Rigidities: Industrial policy and structural adjustment in the Japanese economy 1970–1980, Stanford University Press: Palo Alto. Dyer, J. (1996). ‘Does Governance Matter? Keiretsu Alliances and Asset Specificity as Sources of Competitive Advantage’, Organization Science, 7(6): 649–66. Gawer, A. and M. Cusumano (2002). Platform Leadership; How Intel, Microsoft, and Cisco Drive Innovation, Boston, MA: Harvard Business School Press. Gereffi, G., J. Humphrey, and T. Sturgeon (2005). ‘The Governance of Global Value Chains’, Review of International Political Economy, 12(1): 78–104. Hall, P. and D. Soskice (eds.) (2001). Varieties of Capitalism, Oxford: Oxford University Press. Helper, S. (1991). ‘How Much has Changed between U.S. Automakers and their Suppliers?’ Sloan Management Review, 32 (Summer): 15–28. Holweg, M. and F. Pil (2004). The Second Century: Reconnecting Customer and Value Chain through Build-to-Order. Cambridge, MA: MIT Press. Johnson, S. (2003). ‘Systems Integration and the Social Solution of technical Problems in Complex Systems’, in A. Prencipe, A. Davies, and M. Hobday (eds.) The Business of Systems Integration, Oxford: Oxford University Press. Kenney, M. and R. Florida (1993). Beyond Mass Production: The Japanese system and Its transfer to the United States, Oxford and New York: Oxford University Press. Langlois, R. (2003). ‘The Vanishing Hand: The changing dynamics of industrial capitalism’, Industrial and Corporate Change, (April) 12(2): 351–85. —— and P. Robertson (1995). Firms, Markets, and Economic Change, London: Routledge. Lazonick, W. and M. O’Sullivan (2000). ‘Maximising Shareholder Value: A new ideology for corporate governance’, Economy and Society, 1 February, 29: 13–23. Liker, J., W. M. Fruin, and P. Adler (eds.) (1999). Remade in America: Transplanting and transforming Japanese Management Systems, Oxford: Oxford University Press.
68
Modular production’s impact Linden, G. and D. Somaya (2003). ‘System-on-a-Chip Integration in the Semiconductor Industry: Industry structure and firm strategies’, Industrial and Corporate Change, 12 (3): 545–57. Mayer, D. and M. Kenney (2004). ‘Economic Action Does Not Take Place in a Vacuum: Understanding Cisco’s acquisition and development strategy’, Industry and Innovation, December 11, (4): 299–325. Niece, J. (2005). ‘Cisco’s first glitch’, Journal of Business Research, 58: 1003–5. Pavitt, K. (2003). ‘Specialization and Systems Integration: Where manufacturing and services still meet’, in A. Prencipe, A. Davies, and M. Hobday (eds.) The Business of Systems Integration, Oxford: Oxford: University Press. Principe, A., A. Davies, and M. Hobday (eds.) (2003). The Business of Systems Integration, Oxford: Oxford University Press. Rtischev, D. and R. E. Cole (2003). ‘Social and Structural Barriers to the IT revolution in High-Tech Industries’, in J. Bachnik (ed.) Roadblocks on the Information Highway, Landham, MD: Lexington Books. Saplosky, H. (2003). ‘Inventing Systems Integration’, in A. Prencipe, A. Davies, and M. Hobday (eds.) The Business of Systems Integration, Oxford: Oxford University Press. Sayer, A. (1986). ‘New Developments in Manufacturing: The just-in-time system’, Capital and Class, 30: 43–72. Sturgeon, T. (2000). ‘Turn-key Production Networks: The Organizational Delinking of ¨ ergens (ed.) New Product Development and ProProduction from Innovation’, in U. Ju duction Networks: Global industrial experience, Berlin: Springer Verlag. —— (2002). ‘Modular Production Networks: A new American model of industrial organization’, Industrial and Corporate Change, 11(3): 451–96. —— and J.-R. Lee (2005). ‘Industry Co-Evolution: Electronics contract manufacturing in North American and Taiwan’, in S. Berger and R. Lester (eds.) Global Taiwan: Building competitive strengths in a new international economy, New York: M.E. Sharpe. —— and R. Lester (2004). ‘The New Global Supply-base: New challenges for local suppliers in East Asia’, in S. Yusuf, A. Altaf, and K. Nabeshima (eds.) Global Production Networking and Technological Change in East Asia, Oxford: Oxford University Press. Suzuki, Y. (2004). ‘Structure of the Japanese Production System: Elusiveness and reality’, Asian Business & Management, 3(2): 201–19. Takeishi, A. and T. Fujimoto (2001). ‘Modularization in the Auto Industry: Interlinked multiple hierarchies of product, product, and supplier systems’, International Journal of Automotive Technology and Management, 1(4): 379–96. —— and—— (2003). ‘Modularization in the Car Industry’, in A. Prencipe, A. Davies and M. Hobday (eds.) The Business of Systems Integration. Oxford: Oxford University Press. Williams, K. (2000). ‘From Shareholder Value to Present-day Capitalism’, Economy and Society, February, 29(1): 1–12. Womack, J., D. Jones, and D. Roos (1990). The Machine that Changed the World, New York: Rawson Associates.
69
4 Technology management and competitiveness in the Japanese semiconductor industry1 Takashi Yunogami
After experiencing a dizzying rise in the 1970s and early 1980s, the Japanese semiconductor industry went into a precipitous decline in the 1990s. Various reasons have been advanced for this decline, predominantly related to management failures. Technology, by contrast, is widely considered to have remained competitive. This chapter argues that such a dichotomous characterization, while comforting for engineers and researchers, is misleading, and may be part of the problem itself. It explores the declining competitiveness of the semiconductor industry in Japan as a failure to link technology to changing market conditions, and hence as a failure in technology management, and ultimately in technology itself. It explores the reasons for the failure, and links it to the very factors which propelled the industry to prominence in the first place. It raises the possibility, furthermore, that excessive emphasis on quality, or reliability, which is at the crux of the failure, is not limited to semiconductors, but can be found in other industries as well. (See Robert Cole’s Chapter 2 in this volume for another example.) The chapter first explores reasons given for the decline in the international competitiveness of the Japanese semiconductor industry during the 1990s, including politics, management failure and the excessive quality thesis advanced here. These are not mutually exclusive, of course. There are many contributing factors to the decline, but excessive quality has been overlooked in the past because of the analysts’ use of a dichotomous management vs. technology distinction. This is followed by a brief examination of the main wafer process technologies required for producing semiconductors. I then show that in all three technology phases–elementary process technology, integration process technology, and mass-production technology–the problem was not a lack of technological sophistication, but an excess of it (at least
70
The Japanese semiconductor industry until recently). Japanese manufacturers remained locked into the quality paradigm set by NTT for mainframe semiconductors, even when they were producing for PCs. This substantially increased their costs, both in monetary terms as well as in time. Comparative evidence suggests that the competitors who stole market share from Japanese semiconductor manufacturers pursued technologies which were appropriate for PCs, allowing them to save on development time, to save on costs, and to raise yields rapidly, eventually undermining the Japanese manufacturers with their ‘superior’ technology. This chapter concludes by exploring why this happened and possible implications for other industries. The research which this chapter is based on is ongoing, and the thesis advanced is therefore subject to refinement.
The ‘lost market decade’ The 1990s have been disparagingly called by some, Japan’s ‘lost decade’. For the semiconductor industry, it might be called the ‘lost market decade’. As Figure 4.1 shows, a dramatic rise in Japan’s DRAM share from the mid-1970s to the mid-1980s was followed by an equally precipitous decline, with very little market share remaining by the early 21st century. Needless to say, various reasons have been advanced to explain this decline. In this section I will explore the main explanations. The first type seeks to place the blame on external factors, beyond both corporate-level management and technology. Oyane (2002) attributes the seeds of decline to politics. The very success of the industry, and the decline
100 Japan
Share (%)
80
US
60
Korea
40 20 Europe 0 1975
1980
Taiwan 1985
1990 Year
1995
2000
Figure 4.1 Changing DRAM share by country
71
Industries, technologies, and value chains of the industry in the US, provoked a sharp reaction from the US who tried to impose various constraints on Japanese semiconductor makers under the guise of leveling the playing field (see Murayama in this volume). These constraints were so successful that they eventually took the wind out of the Japanese sails and left it vulnerable to attack from both a resurgent US and emerging Asian competitors. Most analysts, however, place the blame closer to home, even if they have some sympathy for the political constraints view. Most focus on management failures. Itami (1995), for instance, notes that US semiconductor manufacturers such as Intel focused on micro-processing units (MPU) in response to the Japanese onslaught and survived. Korean semiconductor manufacturers focused on DRAMs and made huge and timely investments necessary to become and remain competitive in the industry. Japanese semiconductor manufacturers, by contrast, failed to focus, they spread their investments too thinly and competed without a clear strategy. Kawanishi (1997), a former top semiconductor manager, concurs. These authors note that semiconductor manufacturers were part of large, sprawling business concerns with high overheads and which, in the case of companies like Hitachi, were steeped in heavy electric cultures averse to risk and slow in decision making. Semiconductor businesses require heavy investments in the trough of the silicon cycle; by the time Japanese semiconductor businesses got around to making their investments, they were too late to capture the bulk of profits on the upswing. Indeed, blame is placed with management failure by most industry insiders. In an interview survey by Yunogami and Arikado of 21 engineers from the semiconductor consortium SELETE,2 14 cited a loss of cost competitiveness as the key reason for Japan’s decline and, relatedly, poor management strategies or the absence of strategy (see Appendix, A1). Only two thought that technology was an issue. Indeed, when asked about Japan’s technological strengths compared with competitors in other countries, 20 out of the 21 rated them as equal or superior. Interviews with engineers from the Japan Semiconductor Consortium (A2) carried out in September 2004 elicited similar views. A report by the Semiconductor Industry Research Institute Japan (SIRIJ) in 2003, based on a wide-ranging survey of Japan’s ten leading semiconductor manufacturers, is one of the most authoritative analyses.3 It cites problems with cost competitiveness and lack of focus (semiconductors being part of much broader electric and electronics businesses). With regards technology it noted that Japan’s design technology was superior to that of Korea and Taiwan, but was losing ground to the US. With regards wafer process technology, Japan was at the forefront, along with the US, but Korea was catching up in front-end processes. The report argued that Japanese companies should restructure their businesses, cut costs, strengthen design and systems
72
The Japanese semiconductor industry integration, and so on, but it made no recommendations with regards wafer process technology, again suggesting that this was not an area of concern. In brief, dominant views place the blame at the feet of management and not engineers (in other words, the ‘strong factory–weak headquarters’ syndrome: (Fujimoto 2004). If there are any problems for the latter to address, it is in design. A dissenting view to this orthodoxy is Fujimura (2000), who warned against technology complacency by arguing that a) Japanese semiconductor makers had fallen behind because of inadequate development capabilities in element process technology as well as inadequacies in integration process technology, and b) semiconductor makers of Korea, Taiwan, and the US increased their share by producing low cost DRAMs. Even for Fujimura, however, technological capabilities and cost competitiveness were different issues. A fourth view–that presented here–is that not only is technology a problem, but it is intimately linked with the cost competitiveness problem. This view is implicit in the work of Yoshioka (2004), who stresses the impact on DRAM makers of the shift from mainframes to PCs in the late 1980s (see Figure 4.2). Japanese semiconductor makers, she suggests, made a poor transition because of organizational rigidities. Korean semiconductor makers–particularly Samsung–on the other hand, did not have these legacy problems (leading to rigidity) because they were in an early phase of growth as the PC market developed. They were helped by being able to purchase the newest manufacturing equipment, much of it produced in Japan, in which key elementary process technology was embedded. They developed a production system for 25000
20000 PCs
Mainframes ¥ Bill.
15000
10000
5000
0 1983
1985
1987
1989
1991
1993
1995
1997
Year Figure 4.2 Japanese computer shipments
73
Industries, technologies, and value chains DRAMs made for the PC market, rather than one adapted from DRAMs made for mainframes. As we shall see, this explains another finding of her study, namely that Samsung became able to develop DRAMS more quickly and so was able to surpass its Japanese competitors.4 In sum, various reasons have been advanced for the decline of the Japanese semiconductor industry, ranging from political factors to management failures, technology, and organization. Most accounts, especially by insiders, focus on management failures and argue either explicitly or implicitly that technology was not a significant problem. This distinction is problematic, however. In fact, making this distinction may itself have contributed to the decline. According to this distinction, technology is evaluated according to a relative measure of ‘higher’ (or more advanced, superior) vs. ‘lower’ (inferior). Higher is better. Full stop. But higher is not better if it results in a high cost structure which undermines cost competitiveness where cost is critical, as in DRAMs for PCs. The lack of cost competitiveness of Japanese DRAMs, I contend, did not simply result from excessive overheads in diversified companies–which also existed during the heyday of Japanese DRAMs in the early–mid-1980s–but from the technology deployed to make them. It is dangerous to sever technology from market requirements, and even more dangerous to then take comfort from the fact that one’s technology is ‘superior’ to that of competitors who are gaining market share. In this view, a precondition for restoring competitiveness in the industry is to reconnect technology with market needs.
Wafer process technology for semiconductor production Before trying to show that declining competitiveness was linked to technology and technology management problems, we need to know something about the wafer process technology used in semiconductor production. There are three phases of wafer process technology involved: elementary process technology, integration process technology, and mass production technology in the factory (see Figure 4.3). Let us look at each of these in turn.
Elementary process technologies Semiconductor devices are produced in a number of steps. The basic unit technologies for these steps are called elementary process technologies which include thin film deposition on the silicon wafer by chemical vapor deposition (CVD) or sputtering; lithography by which a resist mask is made on the film; etching which removes nonmasked film using plasma chemical reaction; cleaning technology by which the resist mask and other residue is cleaned off; and inspection technology. Fine processing technology is a combination of lithography and etching.
74
The Japanese semiconductor industry
Mass production technology (yield and quality are important)
Elementary process technology Film depo.
Lithography
Etching
Cleaning
Inspection
Si
Si
Si
Si
Si
Flow 1. 2. 3.
Flow 100. 101. 102.
DRAM DRAM
Yield
Integration process technology (construction of device flow) Flow
Test
Time
498. 499. 500.
Technology transfer Development center
Flow 1. 2. 3.
Flow 100. 101. 102.
Flow
Factory
498 499 500
Figure 4.3 Three phases of wafer process technology
Gate electrode etching, in which the finest processing is demanded, demonstrates the difference between ‘superior’ and ‘inferior’ elementary process technology. Superior etching produces a perpendicular sidewall (B in Figure 4.4a), rather than a tapered or overhanging sidewall (A, C). Precision here has a major impact on transistor performance. And in micro-processing, minimum future size as well as the aspect ratio are critical (see 4.4b: B is superior to A, C is superior to B), as is uniformity (A is technologically superior to B in 4.4c). There is little doubt that Japanese semiconductor manufacturers achieved highly advanced levels in elementary process technologies, which
Figure 4.4 Superior etching technology
75
Industries, technologies, and value chains increased their competitiveness, but this came at a cost, a result of the Japanese companies’ decision to press for frequent customization of equipment.
Integration process technology Integration process technology combines the elementary process technologies to create semiconductor devices on silicon wafers. In the case of the DRAM, over 500 steps are involved. ‘Superior’ integration technology is the ability to integrate the elementary process technologies in such a way as to produce high performance semiconductor devices. Few would doubt that Japanese semiconductor manufacturers succeeded here as well. Quality or performance alone, however, is not sufficient to ensure competitiveness. Time is critical, as well as cost. Indeed, being able to create a given level of performance with the minimum of steps–keeping mask layers to a minimum, for instance–may be considered the ultimate test for superiority of integration process technology. I shall return to this point later.
Mass production technology Mass production technology creates semiconductor devices on silicon wafers according to the process flow built with the integration process technology. Quality and yield are crucial. The three main elements of quality are performance, reliability, and uniformity. The yield indicates the percentage of satisfactory–defect-free–semiconductor devices built on the silicon wafers. Several hundred circular semiconductor devices are created on each silicon wafer, but during the several hundred steps particles from equipment and so on can lead to defects. (It is important to note that if the specifications are increased, the yield will decline, even if the quality distribution remains unaltered. Therefore, a high yield rate per se does not equate with high quality.) ‘Superior’ mass production technology is the ability to mass produce high performance, reliable semiconductor devices uniformly; in Figure 4.5 B is superior to A and C. During the first stages of mass production the yield is normally low. It is crucial to identify problems in the production process and to correct them promptly to raise the yield. The ability to do this reduces the cost and enhances opportunities for profits. Here there is a direct link between cost competitiveness and technology which, as we shall see shortly, has been problematic for Japanese semiconductor manufacturers.
Competitive and excessive technology How competitive is Japanese technology? The question is more easily posed than answered. The simplest way to answer it would be to compare the three phase technologies above across a range of Japanese and non-Japanese
76
The Japanese semiconductor industry
Number of devices
A
B
C
Low
High Quality
Figure 4.5 Technology and quality
semiconductor manufacturers, relative either to each other or to benchmarks. Semiconductor firms, however, are extremely reluctant to allow outsiders in to observe their operations–for obvious reasons–which makes this method impossible. Instead, the following methods have been used in this study, based on interviews with three groups of engineers.5 1 Interviews with equipment manufacturers: Equipment manufacturers deliver equipment to semiconductor manufacturers throughout the world. Their engineers install the equipment, and during the start-up process they have a chance to speak with the engineers at the semiconductor manufacturers. By doing so, they have the opportunity to compare the levels of elementary process technologies at Japanese and non-Japanese semiconductor companies. Three engineers were interviewed, taking care not to compromise their nondisclosure requirements (A3). 2 Interviews with Japanese engineers assigned to foundries abroad: Foundries are businesses which manufacture semiconductors on contract but do not design them. They are frequently located in Taiwan or China. Their customers include Japanese semiconductor manufacturers whose engineers are able to compare wafer process technology between their home base and the foundry during technology transfer of the process flow. Three engineers were interviewed (A4) (A5). 3 Interviews with engineers who have moved from Japanese to non-Japanese semiconductor manufacturers: In spite of ‘lifetime employment’, some engineers– sometimes key engineers–have left Japanese semiconductor manufacturers to join semiconductor firms abroad. Such engineers are also able to compare the element, integration, and production technologies of their current with those of their former employers. One engineer was interviewed (A6).
77
Industries, technologies, and value chains The comparisons of these three groups of engineers provide revealing insights into the three types of technology discussed above.
Elementary process technology The consensus of the engineers from the three groups above was that Japanese elementary process technology was very advanced. Three engineers assigned to an overseas foundry commented: ‘Japanese semiconductor makers’ fine processing technology is far superior’ (A4) (A5). Not only was the fine processing technology held to be superior, but the consensus was that the development strengths were also superior. In the words of an engineer from a dry etching equipment maker, also echoed by engineers of the other groups: ‘The Japanese fine processing technology and development strengths are higher than those found overseas’ (A3). Indeed, one engineer argued that semiconductor manufacturers elsewhere in Asia were ‘incapable of creating new elementary process technology’(A6). The Taiwanese foundries, at least, ‘are dependent on the equipment makers for the elementary process technology’ (A4). By contrast, Japanese semiconductor manufacturers were not content with the performance of standard equipment and placed orders for customized equipment providing higher levels of performance (A3). This is telling. It indicates the existence of high technology and technology development levels for elementary process technology in Japan, but at a cost–at the very least, the cost of customization, which is far from negligible when it comes to semiconductor equipment. A dry etching engineer commented: ‘The elementary process technology and developmental strength in Japan is a bit excessive’. Semiconductor makers elsewhere in Asia, he suggested, were capable of producing their devices on standard equipment. Without superior fine processing technology, they were nonetheless achieving the same level of integration process technology and minimum pattern size as their Japanese counterparts who were using more expensive equipment, which took extra time to produce.
Integration process technology The view on integration process technology was similar: ‘Integration process technology to produce high-performance semiconductor devices in Japan is high’ (A5). Here too, however, it was felt that there was a tendency to create products that go beyond necessary specifications (A5). It is very likely, then, that along with excessive elementary process technology, excessively high performance is set as a goal. Equating the highest specifications with the highest level of technology, and lower specifications with lower level technology is a kind of technological snobbery that ignores market requirements. Producing cost competitive
78
The Japanese semiconductor industry devices requires a sophisticated deployment of technology resources, particularly in the phase of integration process technology. Two engineers were told by engineers at a Taiwanese foundry that the process flow brought over from Japan was too long (A4). It was cut by one-third, prompting the Japanese engineers to question whether the devices produced would perform reliably. In fact, not only did they perform reliably, but the yield went up and profitability improved (A4). Japanese semiconductor manufacturers are said to use more mask layers than elsewhere in Asia (A4) and this has a direct impact on cost competitiveness. Micron Technology of the US allegedly uses only twothirds of the number of masks used by Japanese semiconductor manufacturers, achieving a large reduction in costs (Kanazawa 2000).
Mass production technology High performance and durable semiconductor devices are often equated with superior mass production technology. A Japanese engineer proudly commented on his company’s DRAM mass production facility overseas: ‘This is the only factory producing DRAMs which can be guaranteed for over ten years’ (A7). As the main use for the DRAMs nowadays is in personal computers with a short product life, however, this pride was almost certainly misdirected. Consider Figure 4.6; an impulse to pursue high quality will push the target specification line to the right, reducing the number of devices passed (thus lowering the yield rate). With an emphasis on adequate rather than excessive quality, the target specification line lies further to the left which, all things being equal, increases the yield rate and hence significantly lowers costs. This
Number of devices
Target specification
Pass products
a Low
High Quality
Figure 4.6 Quality and yield
79
Industries, technologies, and value chains is compounded by the fact that less time needs to be spent in reaching the target specification during ramp up. The much vaunted mass production technology of Japan is weak in the area of cost competitiveness. It creates time lags. One engineer noted ‘the slow pace of start up for the yield’ of Japan compared to Taiwan (A5). He also noted that Japanese semiconductor firms need more equipment. Another noted that Japanese semiconductor firms feel a need to increase not only the yield but also the quality level of the devices, while Asian competitors are mainly concerned with just raising the yield. During development, the former want to incorporate new technology, while the latter will not add new technology unless it leads to a rise in the yield. Equipment, processes, or process flows are not changed unless absolutely necessary (A6). Indeed, non-Japanese manufacturers sometimes extend the life of mass production technology deliberately, for example in maintaining KrF lithography for semiconductor devices under 90 nm when competitors had switched to more advanced ArF lithography from the 130 nm generation of semiconductor devices (A6). Even if this is making a virtue out of necessity, if the result of extending the life of equipment is improved cost competitiveness while achieving the required performance and reliability, it can be considered astute technology management. In conclusion, there is little doubt that the elementary process technologies of Japanese semiconductor manufacturers are ‘high’. Their development abilities in this phase also appear to be high. Moreover, the integration process and mass production technologies necessary to produce high quality semiconductor devices are high. However, it is very likely that the elementary process technologies are in fact excessive, and the high levels of integration process and mass production technology lead to the manufacture of DRAMs of excessively high quality for the purposes for which they are required. Imposing special requirements, in turn, results in the need for more equipment. There are more masks and a greater number of steps in the process flow. Costs are raised as a direct consequence, and also as a result of the concomitant difficulties in improving yield and lengthening lead times. The latter are critical when it comes to cost competitiveness and profitability in the DRAM business. Thus, those who argue that the declining world share taken by Japan’s DRAM industry results from management problems but not technology or management of technology are mistaken. Fujimura (2000) correctly argues that technological development is also a problem, but the evidence here suggests that it is not so much one of a decline in the levels of elementary process technology development or integration process technologies, but rather one of excessive technology deployment. To put it bluntly, Japanese semiconductor makers either lacked or did not deploy the technology necessary to be cost competitive. Compared to the
80
rea Ko wan i Ta
Ideal
Japan
Low
Technology for low cost High
The Japanese semiconductor industry
Low High Technology for high quality Figure 4.7 Two evaluation axes of technology
quest for ever greater quality and performance, the pursuit of low cost and mass volume is not glamorous. But it is not easy either. Japanese semiconductor firms which tried to imitate Micron Technology in mask reduction were unsuccessful (A8). Micron Technology had developed a technological competence which was difficult to imitate. Thus the Japanese failure is a failure of technology itself, as well as a failure of technology management.
Why? Why did this happen? It was hardly a matter of technological backwardness. What other plausible reasons are there? The reasons, I suggest, are embedded in the history of the industry, its technology, and its very success.6 The history of the DRAM starts with the invention of the 1K bit DRAM by Intel in 1971. As shown in Figure 4.1, the US held the top share in the DRAM market in the 1970s and the DRAM was developed in this environment. The major Japanese electronics companies invested huge amounts of energy and resources into DRAMs (see Okimoto et al. 1984) and, by the early 1980s, had overtaken the US to claim a dominant share of the global market. The main use for Japanese DRAMs during this period was for large, mainframe computers (see Figure 4.2), and dominance was achieved on the basis of quality (see Nonaka and Nagata 1995). Reliability requirements were especially stringent, with 25 year guarantees sought (A8)! It is said that the reliability demands made were influenced by NTT because of their need for a long-term guarantee for phone systems. This is important because it arguably created a mindset equating rising competitiveness with improved quality, to which all technological efforts were focused. Cost competitiveness was not so
81
Industries, technologies, and value chains much designed into the production system as ensured by relative wage levels and the exchange rate. Let us look briefly at the technology required for a 25 year guarantee. Known as Moore’s law, the degree of integration level of semiconductors has increased by a factor of four every three years. Along with increasing degree of integration, minimum pattern sizes have decreased by a factor of 0.7 every three years. Pattern size reduction not only increases complexity, but increases speed and reduces the minimum necessary power by the scaling rule. In other words, size reduction results in higher performance, and as such, it has continuously been a target for development engineers of DRAMs and other semiconductor devices. Size reduction presents developers with two problems: fine processing and maintaining uniformity. Lithography and etching technology used in the early 1970s were not conducive to miniaturization and uniformity. At the end of the 1970s, however, a new type of stepper was developed by Nikon and Canon which made it possible to create an extremely small resist mask reliably and with uniformity. New dry etching technology known as reactive ion etching (RIE) was also developed at the end of the 1970s and early 1980s by a number of companies, including Nichiden-Anelva, Toshiba, Tokuda Seisakusho, and Hitachi, which made anisotropic fine processing (referred to in Figure 4.2) possible in wet etching (Tarui 1991). With these innovations, Japanese semiconductor manufacturers simultaneously attained fine processing and uniformity. Their engineers continued to push the envelope with respect to the performance of equipment. When they were no longer satisfied with the performance, they would develop new equipment. It was not unusual for semiconductor companies to work hand in hand with the equipment makers to do this. To increase performance and reliability, new processes–in annealing, for instance–were needed. The number of mask layers increased. Inspection was detailed and rigorous. These were all needed to mass produce high quality DRAMs with a 25 year guarantee. These efforts, in turn, set the standards for succeeding generations. Production of high quality DRAMs became the norm for which engineers continued to strive, with strong support and guidance from the quality assurance department, which had a key function. The idea of producing a DRAM inferior in quality to the preceding generation was inconceivable and would have required significant changes to the way semiconductors were designed and produced. By default then, it seems that Japanese semiconductor manufacturers ended up with significantly more inspection steps, transistor forming steps, and heat treatment steps to eliminate damage or defects–perhaps 10 percent more for each of these–than was appropriate for mainframe quality DRAMs. Korean semiconductor manufacturers, by contrast, had not established themselves in the age of the mainframe and they grew with the PC market. They designed sufficient but not excessively high quality into their semiconductor
82
The Japanese semiconductor industry technology. Micron Technology of the US, too, dumbfounded Japanese semiconductor manufacturers with its lower costs: the so-called ‘Micron shock’. The technology requirements in the new era were, simply put, deploying the most cost effective technology in the respective phases. With elementary process technology, this could be done with standard equipment purchased from equipment makers, thereby reducing development costs, the number of processes or steps, and time. Prolonging the useful life of technology where possible is a virtue, not a vice. With integration process technology, it involves designing of a process flow that would produce semiconductors of appropriate quality quickly, reducing where possible the number of masks and steps, aiming for simplicity, and limiting inspection to what is absolutely necessary. Mass production technology involves raising yields at adequate quality as quickly and efficiently as possible. Failure to meet these requirements reduces cost competitiveness, squeezing profits and hence funds for reinvestment. This ultimately undermines competitiveness, even in areas of strength. The reasons for these failures have been suggested, but need amplification. First, whatever their intentions might have been, the Japanese manufacturers continued to make DRAMs largely in the ways they had established in the 1970s and 1980s. They created a technology culture that was resistant to change and, as a result, created path dependence, even as the changing environment made this response inappropriate. They allowed the technology development process to become separated from market requirements. Second, the very fact that they had been so successful no doubt made it harder to change. Success creates big organizations, which in themselves become hard to change, being more introverted and less sensitive to changing market conditions. Success validates certain ways of doing things and norms which then become resistant to change. Top managers have usually had a direct stake in creating the conditions for past success, and so they often are the last to recognize the need for change. The ‘poison of success’ phenomenon is well recognized and shows little regard to nationality.7 Cole and Whittaker treat this same phenomenon in the Introduction to this volume. Third, and somewhat speculatively, there is probably a link with the Japanese productionist culture which, despite its many strengths, also has some blind spots. This productionist culture stresses quality improvement and performance improvement as a way of engaging and motivating engineers. Moreover, it may encourage a view of technology as ‘high’ (good) or ‘low’ (bad), rather than focusing on what customers want. If this is so, the same problems may be present in other industries which have experienced a decline in competitiveness in recent years. Cole’s chapter on telecoms in this volume, for instance, suggests that NTT’s same high quality requirements set standards for its suppliers (major electronic companies) that led to an unwillingness to take TCP/IP protocols seriously, and this proved disastrous for Japan’s competitive position in the emergent networking equipment industry. Further
83
Industries, technologies, and value chains research is needed, both in the semiconductor industry and other industries, before the findings can be generalized, but there is growing evidence that DRAMs are not an isolated case.8
Concluding comments Most analyses of the decline of Japan’s semiconductor competitiveness have stressed general management failures. Technology and technology management have largely been exonerated. The common industry view is that Japan ‘lost in management, strategy and cost competitiveness’ but ‘did not lose in technology’. To some extent this view is correct. But it is also one of the reasons the Japanese semiconductor industry finds itself in such a pinch. The failure to recognize that costs are crucially linked with technology development meant that manufacturers continued to produce semiconductor of excessively high quality when they should have been focusing on developing and deploying technology to achieve cost competitiveness. It also meant, conversely, that efforts to reduce costs through heavy restructuring were only partially effective. In fact, they may have hurt the companies as much as they helped them because many key engineers were lost as a result. (To compound the loss, some were recruited by expanding Asian competitors.) If Japan’s once proud semiconductor industry is to survive at all, it will need to migrate, as suggested by Figure 4.7, by linking technology and technology management more effectively to market requirements.
Appendix: Interview and survey list A1 Results of survey conducted with 21 engineers by the author and Dr T. Arikado, former CEO and first Research Department Head of Semiconductor Leading Edge Technologies, in September 2004. A2 Interviews conducted with five engineers from the Japan Semiconductor Consortium in September 2004. A3 Interview conducted with three engineers at a dry etching equipment manufacturer, 4 May 2004. A4 Interview conducted with two engineers from a Japanese semiconductor manufacturer which consigned work to a foundry, 27 April 2004. A5 Interview conducted with an engineer from a different Japanese semiconductor manufacturer which consigned work to a different foundry, 19 April 2004. A6 Interview conducted by the author and Dr H. Yoshioka with an engineer who had left a Japanese semiconductor manufacturer to work for a semiconductor manufacturer in Asia, 10 July 2004.
84
The Japanese semiconductor industry A7 Interview conducted with an engineer at a Japanese semiconductor manufacturer’s DRAM factory in Southeast Asia, 28 January 2004. A8 Interviews by the author and Dr H. Yoshioka with a former department head of the semiconductor department at Hitachi, 19 August 2004.
Notes 1. Several people have contributed greatly to the research and arguments of this chapter. I am very grateful to Dr T. Arikado, former CEO of Semiconductor Leading Edge Technologies (SELETE) for sharing his valuable insights, as well as to the cooperation of SELETE researchers. Also to Mr H. Mizokami, former head of the Oki Electric DRAM Production Division and former CEO of KLA-Tencor. As will become evident, Dr Hidemi Yoshioka of Kyushu University provided a number of important insights which stimulated the research, and also participated in some of the research itself. I would like to express my deep appreciation to her for her help. Any factual and interpretive errors are the responsibility of the author. Funding and support from the following are also gratefully acknowledged: 1) Ministry of Education, Culture, Sports, Science and Technology (MEXT) 21st Century Centre Of Excellence Program at Doshisha University (ITEC); 2) New Energy and Industrial Policy Development Organization (NEDO) grant ‘Scientific View and Product Development Capabilities in the Semiconductor Industry’; 3) the Murata Science Foundation ‘Comparison of the Technology of Semiconductor Industry in Japan, the US, Korea, and Taiwan: Increasing the International Competitiveness of the Japanese Semiconductor Industry’. 2. SELETE (Semiconductor Leading Edge Technologies) is a consortium of more than 10 of Japan’s semiconductor manufacturers, formed to develop element and module technology for micro-fabrication of next generation transistors, wiring and lithography. SELETE engineers are seconded or transferred from the R&D labs and production facilities of member companies. 3. SIRIJ was established in 1995 to plan and implement programmes to ‘revitalize the Japanese semiconductor industry, increase its international competitiveness and explore the many possibilities of the semiconductor devices’. It is comprised of researchers sent from the member companies, and when necessary research groups are formed with those from the semiconductor and related industries to collect and analyse information. Member companies are Fujitsu, Matsushita, NEC, Oki, Renesas, Rohm, Sanyo, Sharp, Sony, and Toshiba. The Semiconductor Industry Strategy Promotion Committee, established within SIRIJ, was responsible for the report and its recommendations. 4. There are other reasons behind Samsung’s competitiveness vis-a`-vis Japanese semiconductor manufacturers, including organizational factors. Exploration of these are beyond the scope of this chapter. 5. Although the number of interviews introduced here is limited, the interviews are corroborated by the author’s experiences in DRAM research and development (retrospectively) and other informal discussions. 6. I am grateful to Hugh Whittaker for helping me with this section.
85
Industries, technologies, and value chains 7. See for example, Schein 1992; Gerstner 2002; Inagami and Whittaker 2005 (and Whittaker in this volume). 8. Before the arrival of Carlos Ghosn, for instance, Nissan is said to have suffered from similar problems (See Ghosn et al. 2003).
References Fujimoto, T. (2004). Nihon no monozukuri tetsugaku (The Philosophy of Japanese Production), Tokyo: Nihon Keizai Shimbunsha. Fujimura, S. (2000). Handotai rikkoku futatabi (Resurrecting semiconductors), Tokyo: Nikkan Kogyo Shimbunsha. Gerstner, L. (2002). Who Says Elephants Can’t Dance? Inside IBM’s historic turnaround, New York: Harper Business. Ghosn, C, P. Rise, and Y. Takano (2003). Carlos Ghosen keiei o kataru (The management of Carlos Ghosn), Tokyo: Nihon Keizai Shimbunsha. Inagami, T. and D. H. Whittaker (2005). The New Community Firm: Employment, governance and management reform in Japan, Cambridge: Cambridge University Press. Itami, H. (1995). Nihon no handotai sangyo: Naze mitsu no gyakuten wa okottaka? (The Japanese semiconductor industry: Why did three reversals happen?), Tokyo: NTT Shuppan. Kanazawa, T. (2000). Wagakuni handotai sangyo no mezasubeki tokoro (What the Japanese semiconductor industry should be aiming for), Tokyo: Kikaishinko-kyokai. Kawanishi, T. (1997). Waga handotai keiei tetsugaku (Management philosophy of Japanese semiconductors), Tokyo: Kogyo Chosakai. Nonaka, I. and K. Nagata (1995). Nihongata inobeshon shisutemu: Seicho no kiseki to henkaku e no chosen (The Japanese innovation system: Growth miracle and the challenges of change), Tokyo: Hakuto Shobo. Okazaki, S, A. Suzuki, and T. Ueno (2003). Hajimete no handotai risogurafi gijutsu (Introductory semiconductor lithography technology), Tokyo: Kogyo Chosakai. Okimoto, D., T. Sugano, and F. Weinstein (1984). Competitive Edge: The semiconductor industry in the US and Japan, Stanford: Stanford University Press. Oyane, S. (2002). Nichibeikan handotai masatsu (US–Japan semiconductor trade friction), Tokyo: Yushindo. Schein, E. (1992), Organizational Culture and Leadership, 2nd edn., San Fransisco: JosseyBass. Semiconductor Research Institute Japan (2003). Waga kuni handotai sangyo no genjo to kadai (Current situation and issues for the Japanese semiconductor industry), Tokyo: SIRIJ. Tarui, Y. (1991). Handotai rikkoku Nippon: Dokusoteki na sochi ga kizukiageta kiroku (Japanese semiconductors: The record of unique devices), Tokyo: Nikkan Kogyo Shimbunsha. Tokuyama, T. (1992). Handotai durai echingu gijutsu (Semiconductor dry etching technology), Tokyo: Sangyo Tosho. Yoshioka, H. (2004). ‘Consideration of Catching-up of Samsung Electronics in the DRAM Market: From the aspect of change in the demand for DRAM’, Journal of Korean Economics Studies, 4, August: 21–44.
86
5 Global value chains in the pharmaceutical industry Jocelyn Probert
Introduction The configuration of the global pharmaceutical industry’s value chain has undergone substantial change since the molecular biology ‘revolution’ of the 1970s. This has allowed many new specialist firms to emerge as suppliers of technologies or drug candidates to established pharmaceutical companies.1 These new firms have created a market for know-how in various aspects of the discovery research function. The emergence of contract service providers has also encouraged fragmentation of the industry’s value chain in the development, manufacturing, and marketing functions (see Figure 5.1). In contrast to the vertically integrated pharmaceutical companies of the pre-biotechnology age, modern pharmaceutical firms are able to build and draw on dispersed networks of collaborators and alliance partners who share both the costs and the risks of their joint endeavour–a logic that Chesbrough (Chapter 7) describes as the Open Innovation model. This approach, of using externally derived knowledge and technology to complement in-house research and development activities, is common to several of the high tech industries discussed in this volume. But to what extent have Japanese pharmaceutical companies participated in this strategy of fragmentation and reconfiguration? If the short answer to this question is ‘not much until now’, we could well ask why should this be, does it matter, and is their position changing? In this chapter, I compare the strategies of large Japanese pharmaceutical companies with those of their US and European counterparts, looking particularly at interactions with external R&D parties, attitudes towards industry consolidation, and the implications for future growth and performance. The chapter starts with a brief outline of developments in the global industry, then looks at the institutional features
87
Industries, technologies, and value chains Typical alliances/partnerships From biotechs/universities Technology Target identification Lead optimization
From biotechs/other pharma Compounds (all stages): -global rights -territorial rights -co-development
Discovery Research Large pharma company
Target identification Target validation Screening Lead substance -identification -optimization
Development
From biotechs/other pharma Co-promotion Co-marketing
Manufacturing
Marketing
Pre-clinical Phase I Phase II Phase III
To CROs Patient recruitment Phase II trials Phase III trials Data management
To CMOs Biologics mfg Formulation Packaging
To contract marketing organizations Territorial marketing rights
Typical outsourcing/supply contracts
Figure 5.1 Value chain fragmentation options
of the Japanese industry that have been important in shaping firms’ attitudes to the transformation of the R&D process. The third section looks at the activities of individual firms, drawing on interview data from an on-going cross-national research project2 as well as secondary sources, including industry analysis and corporate websites. In conclusion, I discuss the implications of recent strategies for the future competitiveness of the Japanese pharmaceutical industry.
Change in the global pharmaceutical industry and the management of technology The pace of change in biological and chemistry-based sciences presents a challenge to the fully vertically integrated model of pharmaceutical firms. Advances in genetic engineering, cell biology, protein chemistry, and other scientific disciplines as well as the emergence and refinement of technological tools are so diverse that not even Pfizer, the world’s largest pharmaceutical
88
Global value chains in the pharmaceuticals company, aims to possess leading-edge capabilities in all scientific areas. Research and technology collaborations with external parties allow a firm to sample new approaches without immediately having to establish its own research programme. But a firm cannot simply buy in discovery research from outside: it needs internal capabilities to select between competing technologies, and the absorptive capacity to exploit the imported resources effectively (Odagiri 2003). Research partners may be geographically dispersed too, and this adds monitoring and communication skills to the set of capability requirements a firm must possess. In recent decades firms have become more aware of technology not only located outside their own organizational boundaries, but also across national borders. The quest for global reach has encouraged the internationalization of R&D activity, with some pharmaceutical firms operating research sites, sometimes organized as a network of Centres of Excellence (CoEs), outside the home country. More and more, the purpose of such facilities is to tap into the scientific/technological knowledge resources of the host country, rather than to follow the older practice of adapting home country innovations to local market requirements. Roche’s acquisition of a majority stake in the Japanese pharmaceutical firm, Chugai, is an example of the new strategy. Scientific/technology- or market-related factors drive choices around the location and type of site. When firms invest internationally to take advantage of host country skills and knowledge, they are no longer tied exclusively to their native innovation system. But the ease with which they can exploit the advantages of proximity to external resources depends partly on how closely they are tied to their home environment. The national system of innovation in which research-intensive organizations operate influences the type and intensity of research conducted. It reflects past patterns of technological strength and shapes the development of further technological competences (Nelson 1993; Cantwell and Molero 2003). Attitudes to the acquisition of technology affect the sort of linkages made between public and private sector organizations. If there is cutting-edge technological know-how in academic research institutions, firms are more likely to increase the amount of collaborative work they conduct (Hemmert 2004). Corporate strategy and the availability and quality of internal human and capital resources also shape innovation capabilities (Hemmert 2004), as does the organization of knowledge management and internal knowledge transfer (Nonaka and Takeuchi 1995). How successful individual firms are at accumulating scientific and technological competences ultimately determines the competitiveness of an industry at the national level. The cost of accessing and managing the complexities of the wide range of scientific and technological competences necessary for pharmaceutical R&D today partly explains the escalating price tag of bringing a new drug to market– US$800 million in 2000 is a commonly cited figure (DiMasi et al. 2003).
89
Industries, technologies, and value chains A second factor is the heavy regulatory burden imposed by the US Food and Drugs Agency (FDA) and other national or regional authorities, in the form of increasingly complex and extensive clinical trials. The need to recoup R&D expenditure and to maximize returns during the patented life of the product has intensified pressures on firms to globalize their sales and marketing operations. Yet research productivity has not matched the rise in research costs: only 26 new molecular entities (NMEs) reached the market in 2003, compared with 40 NMEs in 1993. The Bain Drug Economics Model 2003, cited by Mertens (2004), shows that the cumulative success rates of drugs moving from the pre-clinical stage through the three phases of clinical trials and onto the market have declined from 14 percent in 1995–2000 to only 8 percent in 2000–02, a record that exposes the increased vulnerability of firms to research failure. Among American and European pharmaceutical firms, strategies to overcome rising costs and pipeline gaps have included mergers and acquisitions to achieve economies of scale, as well as alliance formation and in-licensing. Although M&A activity has apparently not solved their research productivity issues, it has created a set of companies that dwarf the Japanese pharmaceutical industry’s biggest players, both in terms of research budget and marketing muscle. Only three Japanese firms–Takeda, Sankyo and Yamanouchi–appeared in the ranking of the world’s top 20 pharmaceutical companies in 2000. Consolidation increased the world market share of the top 20 firms from 36 percent in 1990 to 69 percent in 2000, yet the three Japanese firms increased their world market share only from 3 percent to 4 percent, while European companies grew their share from 19 percent to 30 percent and US companies from 14 percent to 35 percent (Schofield 2001). In 2003 Pfizer recorded pharmaceutical revenues of US$43 billion, nearly five times greater than Takeda’s US$9 billion, and its research budget of over US$7 billion similarly dwarfed Takeda’s US$1.2 billion. Japanese pharmaceutical industry R&D expenditure as a whole grew much more slowly during the 1990s than in earlier years, reflecting a stagnant domestic market, while R&D expenditures accelerated in the US and Europe.
The institutional background Historically, Japanese firms have been relatively isolated from the rest of the world pharmaceutical industry in terms of their presence in global markets and their access to or usage of external knowledge networks. In market terms, less than 15 percent of Japanese discoveries achieved global status3 in the period 1985–1994, compared with 40–60 percent of products from Swiss, German, British, and US firms (Thomas 2001). More recent data from CMR International (2000) indicate that during the 1990s only one in eight NMEs
90
Global value chains in the pharmaceuticals launched by Japanese companies was sold internationally, compared with one in three European and US drugs. Few Japanese firms have the organizational infrastructure to market their innovative drugs in all major markets. The domestic biotechnology sector is also very immature compared with the sector in the US, the UK and even Germany–the Japan Bioindustry Association (2003) identified 334 firms in 2003, of which only four had gone public. This reduces the possibilities for pharmaceutical firms to access domestic innovations in biotechnology. On the other hand, firms in other industries (such as foods and textiles) are suppliers of know-how and technological competence to an extent not found in other leading pharmaceutical nations. This suggests a rather different pattern of development for the Japanese pharmaceutical industry. Firms have felt rather little pressure to consolidate because of the lack of foreign competition in their home market (Thomas 2001), but also because, as in Germany, shareholder pressure has been limited. As a result, firms have stayed small relative to their American, British, and Swiss peers. Reasons for the relative isolation of the Japanese industry can be traced to Japanese drug pricing policies in the 1980s and 1990s, which helped to keep out foreign firms while also reducing domestic firms’ financial resources, the regulatory approval framework, and the under-developed nature of industry–academic relations. In the following sections, the consequences of each of these factors will be explored in turn.
Pricing policies Drug pricing policies are widely regarded as disadvantaging Japanese pharmaceutical company competitiveness in world markets (Howells and Neary 1995; Thomas 2001). From 1981 onwards near annual reimbursement price cuts penalized firms with older product portfolios; combined with perverse incentives that encouraged doctors to prescribe multiple drugs at each patient consultation (particularly the newer drugs on which the doctors earned high margins), this pricing policy prompted drug companies to launch waves of minor, imitative drugs suited only to the domestic market. Clinically more valuable older drugs disappeared from the market as their prices fell while newly marketed products, irrespective of their innovative value, at least initially earned prices that were high relative to drug prices in the US or Europe (Thomas 2001: 114–15). Only in the 1990s did the health care ministry, under pressure from the finance ministry to rein in overall health care costs, begin to set initial launch prices in the Japanese market closer to global norms, while adding a supplement of 10 percent–increased to 40 percent in 2000 and 100 percent in 2002– for newly listed ‘innovative’ drugs (Motohashi 2004). But because annual or bi-annual reimbursement list price cuts continued, the Japanese market
91
Industries, technologies, and value chains registered virtually no growth between 1990 and 2000 (JPMA 2002), in sharp contrast to the 10 percent annual growth seen in the United States, and 5–6 percent growth in some European markets (Interview Notes 2004). Pricing policies have strongly influenced the ways that firms operated. While the domestic market was rich and growing rapidly, as it did until the early 1990s, they had little incentive to look for new opportunities in foreign markets–particularly the US and Europe, whose more demanding regulatory authorities (see below) were unlikely to approve the sort of drugs that the Japanese pricing system encouraged firms to produce. Although firms began establishing overseas operations in the late 1980s, they paid much more attention to honing their domestic sales activities. After the new ‘innovation’ price supplements were introduced in the 1990s, foreign firms began to compete more aggressively and to take a greater share of the zero growth domestic market. That pushed the more innovative Japanese pharmaceutical companies to start more actively operating in important Western markets.
Regulatory framework The domestic market also stayed relatively isolated because of a prohibition until the mid-1980s on nonJapanese firms applying by themselves for drug approval in Japan (Reich 1990). This encouraged Japanese pharmaceutical companies to seek opportunities to in-license drugs already marketed in the US or Europe, making the identification of such drugs their primary technology strategy (Cockburn et al. 1999). At the same time it protected them from foreign competition. Throughout the 1980s and early 1990s clinical testing standards in Japan emphasized safety over effectiveness, reinforcing the incentives for firms to proliferate and imitate drugs already available (Thomas 2004). And yet full clinical trials on foreign drugs had to be repeated in Japan despite the more rigorous pre-approval testing they had already undergone elsewhere. The high costs of re-testing continued to protect the domestic industry, even after foreign firms were finally allowed to lodge their own applications for marketing approval, since not all firms were prepared to bear the cost of repeating the trials. A consequence of the authorities’ concentration on safety above efficacy was the relatively low percentage, noted above, of Japanese drugs that found global markets. In this protective regulatory environment the Japanese pharmaceutical industry remained highly fragmented compared, for example, with the UK. British firms had faced a challenging competitive environment since the 1950s because the government set high efficacy standards and gave incentives to all firms, foreign and domestic, investing in pharmaceutical research–a policy that meant only firms able to compete at the international level could survive (Thomas 1994; Gambardella et al. 2000). But the international
92
Global value chains in the pharmaceuticals harmonization of standards on good clinical practice gradually eroded Japan’s market isolation. The revisions implemented in 1997 introduced in Japan the principle of informed consent for clinical trials and also established objective standards for drug approval for the first time (Motohashi 2004). Firms could also now conduct small ‘bridging’ studies to complement clinical trial data collected abroad. Since domestic clinical trial costs are high and large scale trials are difficult to organize because the hospital system is so fragmented (Howells and Neary 1995), Japanese firms able to conduct trials overseas were in a better position than those with purely domestic operations. Another regulation also made Japanese firms behave differently than their foreign competitors. Until the Pharmaceutical Affairs Law was amended in 2002, drugs were approved on the basis of applications to manufacture, rather than on marketing applications, as is the case in the US and Europe. This meant that firms had to retain control over their entire value chain, keeping all manufacturing in-house. Since the amendments, several firms have announced the spin-off of their production sites into separate subsidiaries, which will also offer contract manufacturing services to other firms. Contract research organizations (CROs) were also slow to develop in Japan compared with the US and Europe, because they found it hard to break into the tight relationships between pharmaceutical companies and doctors. The strengthening of clinical trial guidelines began to change attitudes–although CROs struggled until recently to recruit good staff (Interview Notes 2004). Also, some of the leading domestic pharmaceutical firms began to gain experience of CROs in foreign markets. These changes in the regulatory environment have gradually supported the emergence of a variety of external service providers to the domestic pharmaceutical industry.
Industry–university linkages The rise of the US biotechnology industry on the basis of start-up firms created by entrepreneurial academics (often with the support of venture capitalists) is a well documented story, repeated on a lesser scale in parts of Europe. Scientists displaced from large firms through mergers are another source of venture creation in Western countries. But in Japan neither of these trends was evident until recently. Instead, pharmaceutical firms and established firms in other sectors have been the prime movers in bio-pharmaceutical activity, and they have dominated the biotech patenting process (Thomas 2001; Kneller 2003). There are no start-up Japanese genome database firms like Celera or Incyte in the US; neither have any important new firms with bioinformatics expertise come out of the universities. Why has Japan’s biotechnology capability not emerged from universitybased research activity? It has been said that suitable conditions to foster standalone biotechnology firms and academic spin-outs did not emerge
93
Industries, technologies, and value chains because the government did not recognize the potential value of biotechnology (Mahadeva 2004). But the slowness of Japanese academic researchers to move into areas such as genomics can also be traced to the poverty of universities,4 compared with the substantial resources available to US life sciences academics (primarily through the NIH’s competitive grant awards). From the 1960s, funding for universities failed to keep up with student growth, putting a squeeze on resources available for research in science and engineering (Whittaker 2003). And although master’s programmes–whose graduates entered corporate R&D laboratories–boomed, there was no equivalent flow of students into PhD programmes (Kobayashi 1997). The tiny number of PhDs given in biology–only 200 each year compared with 6000 in the US, according to Lehrer and Asakawa (2004)–provides a stark contrast. As the academic qualifications of corporate scientists are comparatively low, moreover, it is hard for them to grasp easily the significance of academic advances in biology. Part of the problem lies with Japanese public sector governance mechanisms. Unlike the liberal university–industry institutional environment in the US, in Japan there has traditionally been a clear boundary between academic (non-profit) activity and (profit-seeking) commercial activity. Academic researchers were not allowed to patent discoveries from externally funded research projects, and the commercial use of intellectual property arising from government sponsored research was also difficult (Lehrer and Asakawa 2004). Not surprisingly, the flow of cutting-edge academic knowledge into the commercial sphere was very poor, particularly because public university professors were barred from playing a direct role in private companies. The rules even on collaborative work with companies were so restrictive that researchers sometimes gave their findings to companies in exchange for donations to their laboratories (Whittaker 2003). Instead, relations between industry and university have typically been based around semi-formal and informal research networks and government sponsored research consortia like those around the genome and SNPs in the 1990s and proteomics in 2000. Another problem has been that various government bodies wanted to take the lead in biotechnology throughout the 1980s and 1990s. This picture finally changed in the late 1990s. Several new laws were passed in 1998–2000, including the equivalent of the US’s 1980 Bayh-Dole Act, to stimulate Japan’s participation in global biotechnology research. National science policy reforms focused on producing what Lehrer and Asakawa (2004: 929) call ‘networked scientist-entrepreneurs’, and in 1998 the first university technology licensing office (TLO) to assist in the commercialization of research was established. By January 2002 there were 26 TLOs in operation (Motohashi 2004). The first example of an IPO by a university biotechnology spin-out was AnGes MG, founded by an Osaka University professor and listed on the Mothers section of the Tokyo Stock Exchange in September 2002. Significantly, the performance of national universities and research institutes
94
Global value chains in the pharmaceuticals is now becoming based on quantitative indicators, such as patent filings, publications, and the winning of external grants and research contracts (Motohashi 2004). Although these measures have improved conditions for industry–university linkages, they cannot guarantee either the same mobility of researchers between universities and firms or between firms that has been so important to the US bio-pharmaceutical industry or the availability of venture capital for start-ups. In a society where the social stigma attached to bankruptcy and to frequent job hopping is high, scientists are not prepared to take the risk of joining a start-up firm when the chances of the main project failing are high (as they are in biotechnology). Similar problems hampered the development of the biotech industry in Germany. As for pharmaceutical firms, they apparently appreciate the involvement of companies from other industries (such as Hitachi) in government sponsored research consortia but are concerned about the possible leakage of their own technologies to rivals. Even so, they often feel government pressure to participate. Finally, it is possible that linkages between universities and companies do not work because companies themselves have internal procedures and processes that deter cooperative activity, such as a value system that favours in-house product development and disregards the potential benefits of research alliances. Against this background, we next explore the strategies Japanese firms use to capture knowledge and technology resources.
Firm-level strategies Although the size and wealth of the domestic market provided little incentive for firms to establish overseas activities, larger players including Takeda, Eisai, Yamanouchi, and Fujisawa began in the 1980s to develop international operations. In the 1990s other companies attempted to offset difficult domestic market conditions by tapping into overseas markets. The number of companies with operations (including pharmaceutical bulk materials, medical devices and nutrition tonics, as well as pharmaceutical R&D activities) in the US rose from 22 in 1990 to 34 in 1995 and to 51 in 2000, while in the major European markets (France, Germany, and the UK) they increased from 18 to 36 and then to 46.5 By the end of the 1990s some leading firms were deriving 35– 45 percent of sales from foreign markets and, in some cases, a considerably higher percentage of profits, based on innovative compounds to treat ulcers (Takeda, Yamanouchi), prostate cancer (Takeda), Alzheimer’s (Eisai), and organ transplant patients (Fujisawa). This allowed them to increase their R&D expenditure even though total Japanese pharmaceutical R&D expenditure slowed. Firms with small or no overseas sales struggled to grow. At the same time, foreign competition in the domestic market intensified. Table 5.1 shows the rising importance of foreign companies.
95
Industries, technologies, and value chains Table 5.1 Changes in the domestic landscape, 1992–2002: An increasing Western presence 1992 Overall 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25
Company Takeda Sankyo Sumitomo Yamanouchi Shionogi Fujisawa Daiichi Eisai Tanabe Otsuka Chugai Banyu Kyowa Hakko Dainippon Tsumura Bayer Ono Green Cross Hoechst Taiho Pfizer Schering Sandoz Kaken Yoshitomi
2002 Rx sales
Overall
2608 2450 1950 1825 1617 1500 1383 1367 1158 1017 908 858 808 800 767 755 717 650 633 592 524 519 508 500 483
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25
Company Takeda Sankyo Pfizer/Pharmacia Yamanouchi Daiichi Eisai (New) Mitsubishi Fujisawa Novartis Shionogi Banyu/Merck Chugai/Roche Otsuka Tanabe GSK Sumitomo Ono Kyowa Hakko Aventis Dainippon AstraZeneca Bayer Taiho Meiji Seika Boehringer
Rx sales 5208 2925 2783 2100 2083 1942 1775 1717 1525 1508 1500 1383 1325 1183 1117 1075 1067 1067 1050 958 917 708 683 650 622
Source : Japan PharmaCyclopedic, cited in a presentation by Bill Mattson, The Mattson Jack Group at BioNetwork 2003, Newport Beach, CA, 28 October 2003 Note : The merger between Yamanouchi and Fujisawa placed Astellas second, behind Takeda, in 2005.
The common factors among firms with fast growing overseas businesses were a) innovative drugs that met global standards, and b) their own clinical development facilities in the US and Europe. Having one without the other did not work. Overseas clinical development capabilities allowed firms to capture more of the value of their compounds before (perhaps) assigning marketing licences at a later stage–Sankyo could earn only royalties on US sales of its blockbuster anti-cholesterol drug, because it could not do any overseas development work. These facilities also give firms the option of conducting clinical trials overseas first, and later doing only bridging studies in Japan. More importantly, launching a drug first in the US, where pricing is market-based, allows firms to establish a level from which to negotiate launch prices in controlled markets like Japan (Interview Notes 2004). The competence to manage clinical development in Western markets has become an important dividing line between Japanese pharmaceutical companies. Some firms also began to disperse their discovery research facilities, opening sites in the US and Europe to complement domestic research laboratories, and
96
Global value chains in the pharmaceuticals to tap into local technologies and human capital resources. Eisai opened a facility in Boston in 1989 and one in London in 1990; Yamanouchi ran a basic research institute in Oxford from 1990 until its work was reintegrated into the Tsukuba lab in 2002; and Fujisawa opened a small neuroscience research facility in Edinburgh in 1990 and later a small pharmaco-dynamics unit near Chicago. Eisai’s US research site was responsible for discovering one drug (for sepsis) that by 2004 was in Phase 2 clinical trials. Takeda, on the other hand, chose to establish clinical development and marketing joint ventures with local partners in the US and Europe (which it usually later bought out), while retaining all its basic research in Japan. Sankyo, similarly, has always kept all its discovery research in Japan. Industry consolidation has been slower in Japan than elsewhere. As noted earlier, Japanese companies have not participated in any of the big international mergers, preferring their independence to economies of scale. But since the late 1990s several mergers have taken place, both between domestic companies and between a Japanese company and a foreign partner (Table 5.2). The first merger between leading companies occurred only in 2005 with the creation of Astellas from the former Yamanouchi and Fujisawa. Since then, Sankyo and Daiichi Pharmaceutical have also merged, but both Eisai and Takeda have firmly stated their intention to go it alone. The experience of Mitsubishi Pharma, formed from mergers between four drugs companies, shows that size by itself is no help without competitive new drugs in the pipeline. Internal corporate restructuring measures–led by Takeda–have also focused firms more closely on high margin pharmaceutical business. Nonpharmaceutical activities such as chemicals and animal health have been sold, and noncore functions such as manufacturing transferred to
Table 5.2 Selected consolidation moves in the Japanese pharmaceutical market 1998 1999 2000 2000 2001 2002 2002 2002 2003 2003 2005 2005
Yoshitomi and Green Cross merged to create Yoshitomi (renamed Welfide in 2000) Tokyo Tanabe and Mitsubishi Chemical merged to create Mitsubishi-Tokyo Pharmaceuticals Boehringer Ingelheim (Germany) OTC acquired 51.4% of SS Pharmaceutical (OTC activity) Schering (Germany) acquired Mitsui Pharmaceutical Welfide and Mitsubishi-Tokyo Pharmaceuticals merged to create Mitsubishi Pharma Roche (Switzerland) acquired a 50.1% stake in Chugai Pharmaceutical Abbott bought outstanding 33.3% of Hokuriku Seiyaku Taisho bought 20% of Toyama Chemical, formed R&D alliance, and established joint sales force Suntory merged pharma activities into Daiichi Suntory, a new subsidiary 66% owned by Daiichi Pharma Merck (US) increased its stake in Banyu Pharmaceutical from 51% (in 1984) to 99.4% Yamanouchi and Fujisawa merged to create Astellas Dainippon Pharmaceutical and Sumitomo Pharmaceuticals merged to create Dainippon Pharmaceutical
Source : Company materials and news databases
97
Industries, technologies, and value chains subsidiaries. Some nonpharmaceutical companies have completely withdrawn from drug-related activities. As a result, the landscape of the domestic pharmaceutical industry has changed significantly since the late 1990s. Turning to research strategies, a survey of 16 leading Japanese companies in 1999 found that they were active in as many therapeutic areas (on average, six) as Western counterparts, despite their much smaller R&D budgets (CMR International 2000). An examination of corporate websites and drug development pipelines in 2005 indicates that–in parallel with organizational restructuring– many companies are focusing their research efforts more narrowly on their ‘franchise’ therapeutic areas,6 plus a limited number of additional key areas. The newly formed Astellas, for example, claims a global franchise in urology (from the Yamanouchi side) and the transplant segment of immune disorders (from Fujisawa), while Sankyo declares a franchise in cardiovascular diseases, is prioritizing metabolic, bone and joint, and immunological diseases, and regards oncology and infectious diseases as ‘challenge’ areas. Eisai mentions three areas (neurology, gastrointestinal diseases, and oncology) and Takeda, four (oncology, urology, central nervous system diseases, and ‘lifestyle-related’ complaints–diabetes, obesity and hypertension). All the same, and despite R&D/sales ratios matching those of Western pharmaceutical companies, research budgets remain relatively thinly stretched. Japan’s biggest companies, Takeda and Astellas, spent ¥130 billion (US$1.2 billion) and ¥140.5 billion (US$1.3 billion) respectively on research in 2004, but this is just a fraction of what leading firms in the US each spend.7 Many Western pharmaceutical firms have turned recently to in-licensing compounds at various stages of development, to replenish depleted drug pipelines. Yet in-licensing is not a core strategy for Japanese companies, large or small. They had earned low margins on products in-licensed in the 1980s from foreign firms for marketing in Japan, and this type of arrangement fell out of favour in the early 1990s for two reasons: their desire to reduce dependence on foreign innovations and foreign firms’ lack of interest in marketing their products in Japan because of unfavourable government pricing policies. It has also been suggested (by Kneller 2003) that Japanese companies see partnering and licensing agreements as a way of dividing up geographic markets and dealing with regulatory and marketing challenges, rather than of integrating themselves into a global network, for example by participating in the early stages of drug development. Some research (Takayama et al. 2002) even claims that drug in-licensing is a stop-gap strategy for Japanese firms entering a new therapeutic area, until in-house drug candidates in the same therapeutic area reach the market, at which point the in-licensed products are de-emphasized in the firm’s portfolio. But these comments apply to foreign drugs in Japan. Most Japanese companies would find it hard to compete for a licence to develop and market a foreign drug worldwide, because of the costs involved and the infrastructure they would need. On the other hand, in the domestic
98
Global value chains in the pharmaceuticals market food and chemical firms have for many years been important sources of early stage compounds for Japanese pharmaceutical companies, in effect fulfilling a role similar to the role Western biotechnology firms play vis-a`-vis Western pharmaceutical firms.8 One way for Japanese companies to leverage their relatively small research budgets is to form research alliances and partnerships, allowing them to share both the costs and the risks of research. The proportion of R&D budgets spent outside the firm has been increasing steadily over the last 25 years, although again, not as much as at Western firms. A survey by the Ministry of Public Management, Home Affairs, Posts and Telecommunications showed above average outsourcing expenditure by the pharmaceutical industry in 2001 (some 15.9 percent of R&D spending, compared with 11.1 percent for Japanese industry as a whole). The proportion of outsourced pharmaceutical R&D accelerated since the early 1990s, even though pharmaceutical industry R&D spending in Japan flattened, and foreign companies were the major beneficiaries of this trend. An analysis of inter-corporate alliances by Hirai (2002), which Motohashi (2004) cites, shows that US biotechnology firms are the fastest growing group of collaborators for Japanese pharmaceutical companies, and that target identification alliances account for one-third of all alliances. Research has shown that although the leading firms certainly form more alliances around drug discovery technologies or target identification than for drug candidates in pre-clinical or clinical development, they are much less active in forming alliances than Western firms (Kneller 2003). Another survey, by CMR International (2000), points out that Japanese firms are much less active in making alliances and technology acquisitions than US firms, but not so different from European firms when firm size is taken into account. Odagiri (2003) also finds that 60 percent of the alliances he identified are research partnerships, and the rest are development/marketing agreements. One example of recent in-licensing activity is Takeda’s acquisition of the development and marketing rights in Japan for an obesity drug still in phase two clinical trials at the UK biotech company, Alizyme. Some firms reported to us that forming alliances with European biotechs is easier than with US biotechs, because competition is less intense (Interview Notes 2003). Since the US biotech industry is the most dynamic source of key discovery technologies, firms that do not have a physical presence in California or Boston may find it difficult to identify and build important research networks there. Generally speaking, our own interviews indicate that the number of technology or product partnerships signed by Japanese firms with foreign biotechs is very small compared with the wealth of alliances made by American and British pharmaceutical companies. Japanese pharmaceutical companies were a welcome source of financing for new biotechnology firms in the 1980s, when venture capital markets were tight–Chugai’s full acquisition of Gen-Probe in 1989 is one example.9
99
Industries, technologies, and value chains But senior management at Japanese firms have usually preferred minority stakes in biotech firms rather than outright acquisitions, over-riding the strong recommendations of their scientists (Interview Notes 2003).10 Some companies, including Takeda (and both Yamanouchi and Fujisawa, the precursors of Astellas), have established small VC funds in the US to identify early stage research opportunities among the many biotech firms there. But in Japan, a vibrant biotechnology industry has been slow to develop, partly because there are no substantial domestic VC funds.11 In any case, Japanese pharmaceutical companies seem sceptical that domestic biotechs can play a role in providing important technologies (Kneller 2003). But there are bright areas in the Japanese industry, and the trend among Western firms towards research alliances means that even firms confined until now to the domestic market have some opportunities to develop overseas business if they have promising compounds in development. Western firms recognize the excellent science being conducted in some companies and are forming strategic pharma-pharma research relationships that resemble the more familiar pharma-biotech alliances.12 GlaxoSmithKline established early stage research alliances with both Shionogi and Tanabe in 2001 that it describes as ‘virtual CEDDs’ (Centres of Excellence in Drug Discovery), which it uses to complement its in-house CEDDs (Interview Notes 2004). Tanabe is supplying GSK with potential compounds in several therapeutic areas, while the Shionogi alliance revolves around specific compounds in two disease areas plus exclusive rights over future compounds in the same areas. A joint venture will develop and commercialize the products globally, helping Shionogi to expand its international presence. Tanabe has been almost entirely domestically focused, but is using R&D alliances with GSK, Novartis, and Menarini to expand its overseas operations. Finally, Japanese firms have developed more formal research partnerships with foreign universities than with domestic public research laboratories. Several firms list collaborations formed with US universities in recent years. Eisai’s London research facilities–like Fujisawa’s Edinburgh centre–are embedded in a university department. Collaborative relationships with foreign universities are important vehicles for recruiting scientists, as well as achieving specific research goals. Over time, they may help to overcome the cultural and linguistic mismatches that seem to plague some research relationships. In Japan, secondments of bench-level scientists from companies to universities are not uncommon (Hicks 1993) and could bridge the gaps in understanding between industry scientists and university researchers that Kneller (2003) highlights. In summary, Japanese firms are becoming more integrated into the research networks that global pharmaceutical firms typically build, but they are more cautious than Western firms in forming alliances and partnerships. There has been a general preference for organic growth, but that picture is changing as
100
Global value chains in the pharmaceuticals the domestic industry restructures. Japanese firms now not only seek external partners to plug competence gaps (e.g., in gene expression) but are sought themselves by Western firms as sources of early stage compounds, in recognition of their innovative capabilities.
Conclusions The strategic intent of individual firms is beginning to show through more clearly in response to the many changes in the institutional landscape since the mid-1990s–including growing foreign competition in a previously rather isolated domestic marketplace. Various reconfigurations are taking place, not least the creation of Astellas and the Daiichi/Sankyo merger, yet in global industry terms the leading firms remain medium sized at best. Mega-mergers among US and European firms have not proved that size guarantees R&D success, but they have created giant marketing machines capable of distributing blockbuster drugs worldwide. Japanese drugs have only reached blockbuster status through the marketing muscle of their global licensors. With research costs and complexities rising rapidly, Western (especially US and UK) firms have placed options on various compounds and technologies by forming alliances with a range of external partners. These collaborations are important risk-sharing mechanisms that allow firms to evaluate the potential of a particular technology or research area before spending too much, and to withdraw easily if the arrangement is unfruitful. Japanese firms have been more wedded to organic growth and have been slower to externalize functions such as clinical trials, or to withdraw from unpromising areas of research, probably partly because in-house scientists on ‘lifetime’ contracts need to be kept occupied. The lack of employee mobility between university and firm, and from firm to firm, hindered the creation of the specialized research or platform technology spin-offs that underpin the world’s biotechnology sector and contribute to the rapid commercialization of new scientific ideas. But with corporate restructuring, the development of a small domestic VC industry, and the strong desire of government agencies to put biotechnology at the centre of science and technology policy, a biotechnology sector is beginning to emerge. Yet, as the case of Germany shows, a negative institutional environment takes many years to overcome, and missed opportunities in the past damage current corporate scientific capabilities. An optimistic view of Japan’s pharmaceutical companies is that their cautious attitude to new biotech research tools means they have not wasted large sums of money on approaches that have yet to prove fruitful (as some would say Western firms have); a more negative perspective is that they have lost years of experience working with such tools and integrating them into their research laboratory methodologies. It could take another decade to reveal which is the case.
101
Industries, technologies, and value chains The largest Japanese companies do have innovations in their drug pipelines, like their much larger European and American competitors. But it is unlikely that any Japanese firm will join the group of global mega-companies (unless as part of an international merger, in which case it may lose its identity). Even if it discovers a potential blockbuster drug, it lacks the necessary global marketing strength to create the blockbuster status. But in any case the blockbuster model is under some threat from patent expiries, a lack of suitable candidates, and increasing resistance by health care insurers (in the US) to the high price of these drugs. This suggests that a strategy that continues to devote R&D resources to compounds for low severity diseases where several acceptable treatments already exist (e.g., hypertension, hyperlipidemia or arthritis) is relatively risky. When a blockbuster drug comes off patent, corporate sales will fall sharply unless there are other big drugs following on behind. But a plausible alternative strategy is to focus on high severity conditions with significant unmet needs (e.g., HIV, thrombosis, transplant) and/or smaller disease areas that are unattractive to the mega-players. Because these drugs require smaller clinical trials, the cost of development is lower–regulatory authorities or insurers regard them in better light than another treatment for a ‘lifestyle’ disease. Medium sized European companies seem to pursue niche market strategies with success, and even large US firms such as BMS and Johnson & Johnson are moving towards finding treatments for unmet needs. In Japan, the strategies of Astellas (urology and transplants) and Shionogi (HIV) point to similar thinking. The advantage of such specialist care products is that only small sales forces are needed to achieve complete coverage of important overseas markets. But the key challenge for Japanese companies as well as for American and European firms is to generate important drug candidates. And that means leveraging research resources around the world to their full extent, both by giving suitable autonomy to overseas research laboratories and by being open to potential opportunities for collaboration with other firms.
Notes 1. The focus in this chapter is on producers of ethical pharmaceuticals, i.e. prescription drugs, that are under patent protection. 2. I am indebted to ITEC at Doshisha University and the Cambridge-MIT Institute for financial support for this research, and to Christel Lane (University of Cambridge) for our many discussions of external knowledge sourcing practices. 3. Defined as a product approved in at least two out of the three Triad markets. 4. The education ministry has extensive regulatory and funding powers to control both public and private universities. 5. Asian countries also became important locations.
102
Global value chains in the pharmaceuticals 6. Where they already have significant market presence and more compounds in research. 7. Members of PhRMA, the American industry association, collectively spent US$33.2 billion on research in 2003. 8. In an interesting new development, some smaller Japanese firms are choosing to out-license products to their larger peers for the North American market, rather than to US or European firms. Examples include Mitsubishi Pharma’s diabetic neuropathy drug and Dainippon Pharmaceutical’s gastro-intestinal drug, both of which have been out-licensed to Takeda for the US and Europe; and an acute heart failure treatment licensed by Daiichi Suntory to Fujisawa (now Astellas). 9. When Roche acquired a majority stake in Chugai, US anti-trust laws required the divestment of Gen-Probe in light of Roche’s pre-existing ownership of Genentech. 10. Western pharmaceutical companies have indicated to us that they will acquire product-based biotech firms, but prefer licensing or collaborative arrangements with technology-based firms, in case the technologies are superseded later. 11. German biotech industry growth was also hampered by a lack of VC funding. 12. The comparison is facilitated by the disparity in size between the organizations, although the Japanese firms have greater downstream experience than many biotechs.
References Cantwell, J. and J. Molero (eds.) (2003). Multinational Enterprises, Innovative Strategies and Systems of Innovation, Cheltenham: Edward Elgar. CMR International (2000). Japan in Focus: Strategies for innovation and global drug development–What differentiates Japanese pharma companies from their Western counterparts? CMR International, R&D Briefing No.28, Epsom. Cockburn, I., R. Henderson, L. Orsenigo, and G. Pisano (1999). Pharmaceuticals and Biotechnology, U.S. Industry in 2000: Studies in Competitive Performance, The National Academy Press, www.nap.edu/openbook/0309061792/html/363.html DiMasi, J. A., R. W. Hansen, and H.G. Grabowski (2003). ‘The Price of Innovation: New estimates of drug development costs’, Journal of Health Economics, 22: 151–85. Gambardella, A., L. Orsenigo, and F. Pammolli (2000). Global Competitiveness in Pharmaceuticals: A European perspective, Brussels: European Commission DG Enterprise. Hemmert, M. (2004). ‘The Influence of Institutional Factors on the Technology Acquisition Performance of High-tech Firms: Survey results from Germany and Japan’, Research Policy, 33: 1019–39. Hicks, D. (1993). ‘University–Industry Research Links in Japan’, Policy Sciences, 26: 361–95. Hirai, H. (2002). ‘Alliance of Pharmaceutical Firms in Europe, Japan and United States’ (in Japanese), OPER Report No.4, September. Howells, J. and I. Neary (1995). Intervention and Technological Intervention: Government and the pharmaceutical industry in the UK and Japan, Basingstoke: Macmillan Press. JBA (2003). Statistical Analysis of Japanese ‘Bio-ventures’, Tokyo: Japan Bioindustry Association 20 (1–3).
103
Industries, technologies, and value chains JPMA (2002). The Japanese Pharmaceutical Industry, Tokyo: Japan Pharmaceutical Manufacturers Association. Kneller, R. (2003). ‘Autarkic Drug Discovery in Japanese Pharmaceutical Companies: Insights into national differences in industrial innovation’, Research Policy, 32: 1805–27. Kobayashi, S. (1997). ‘Educational System in Raising Human Capital’, in Japan Commission on Industrial Performance (ed.) Made in Japan: Revitalizing Japanese manufacturing for industry growth, Cambridge MA: MIT Press. Lehrer, M. and K. Asakawa (2004). ‘Rethinking the Public Sector: Idiosyncracies of biotechnology commercialization as motors of national R&D reform in Germany and Japan’, Research Policy, 33: 921–38. Mahadeva, H. (2004). ‘BioJapan 2004 – Promising biotech growth in Japan’, 29 Nov., IMS Health Online Store. Mertens, G. (2004). Targeted Cancer Therapies: Innovative drugs and their impact on the future of oncology, London: Reuters Business Insights. Motohashi, K. (2004). OECD/TIP Project on Biopharmaceutical National Innovation Systems. National Report: Japan, March, OECD/TIP. Nelson, R. (ed.) (1993). National Innovation Systems: A comparative analysis, New York: Oxford University Press. Nonaka, I. and H. Takeuchi (1995). The Knowledge-Creating Company, New York: Oxford University Press. Odagiri, H. (2003). Transaction Costs and Capabilities as Determinants of the R&D Boundaries of the Firm: A case study of the ten largest pharmaceutical firms in Japan, Managerial and Decision Economics, 24: 187–211. Reich, M. (1990). ‘Why the Japanese Don’t Export More Pharmaceuticals: Health policy as industrial policy’, California Management Review (Winter): 124–50. Schofield, M. (2001). ‘The Global Pharmaceutical Industry’, in P. Nightingale (ed.) Globalization: The external pressures, Chichester: John Wiley & Sons. Takayama, M., C. Watanabe, and C. Griffy-Brown (2002). ‘Alliance Strategy as a Competitive Strategy for Successively Creative New Product Development: The proof of the co-evolution of creativity and efficiency in the Japanese pharmaceutical industry’, Technovation, 22: 607–14. Thomas, L. (1994). ‘Implicit Industrial Policy: The triumph of Britain and the failure of France in global pharmaceuticals’, Industrial and Corporate Change, 3/2: 451–89. —— (2001). The Japanese Pharmaceutical Industry: The new drug lag and the failure of industrial policy, Cheltenham: Edward Elgar. —— (2004). ‘Are We all Global Now? Local vs. Foreign Sources of Corporate Competence: The case of the Japanese pharmaceutical industry’, Strategic Management Journal, 25: 865–86. Whittaker, D. H. (2003). ‘Crisis and Innovation in Japan: A new future through technoentrepreneurship?’ in W. Keller and R. Samuels (eds.) Crisis and Innovation in Asian Technology, Cambridge: Cambridge University Press.
104
6 Software’s hidden challenges1 Robert E. Cole
Software is the fastest growing sector of the Information and Communications Technologies (ICT) sectors, and is at the heart of the revolution in information technologies through its increasing criticality in providing functionalities for hardware (OECD 2001: 105). Software-related sales in OECD countries grew at a rate of 16 percent a year from 1992 to 2001 (OECD 2002: 58). Given the growing scale of these activities, and the high probability that firms, industries and even nations approach these activities differently, it would be surprising if such differences didn’t have some strategic consequences. Indeed, research suggests that how firms apply software increasingly serves as a critical element determining competitive success (Rapp 2002). In the past, data seemed to show that Japanese industry invested far less in software than American firms. According to OECD data, for example, Japan was only spending 25 percent of its ICT investment on software versus 46 percent for the US (OECD 2001: 65). As Jorgenson and Motohashi (2003) have recently shown, however, a large part of these differences derive from how software investment was measured. US data include the three basic types of software for investment: prepackaged, custom made, and firms creating software on their own account. However, Japanese software investment is measured only using custom made software. When adjustments are made for the absence of the large number of firms developing their own software (own account) and for the smaller number of purchasers of prepackaged software, the software/ICT ratio is almost comparable for the two countries. To be sure the sectoral location of software investment is an issue. The productivity spurt experienced by the American economy in the last half of the 1990s was stimulated not only by innovation in the ICT sector but also by the widespread adoption of ICT outside the high tech sector (Oliner and Sichel 2002), especially in services. It is in such sectors as financial services and retail that the Japanese have lagged behind the Europeans and the Americans.2 The discussion to follow is based not on the failure of Japanese firms to invest in software or the sectoral
105
Industries, technologies, and value chains location of investment but rather on their mode of investment and the types of organizational change made to support IT investment. If one asks Japanese managers in the ICT sector what are their major competitive problems today few, if any, will mention software as an issue. A variety of Japanese government surveys routinely ask experts in given fields if they are leading or lagging in key technologies. These experts consistently report that Japan is lagging in software and software tools (METI 2003: 162–3). We see then a large gap between managerial perceptions and expert judgments. What accounts for this gap? In part it is because the experts focus on innovation in a broad range of areas, including the ongoing worldwide development of packaged solutions based on standardizing best practices; Japan has not contributed much to these global developments either as developer or user. Managers, however, focus on software usability and in this regard, Japanese managers have been largely content to rely on customized software developed in-house or with the help of outside system integrators. Japanese high tech firms have been slow to act on the potential of the software revolution and its profound implications for future competition. I focus on three dimensions of this transformation with the twin objectives of understanding the reason for the large role of customized software solutions used by Japanese ICT firms and the competitive consequences of these decisions. The first dimension is the emphasis on proprietary systems that grew out of the different trajectories of the U.S and Japanese software industries over the last 30 odd years. The second is the metamorphosis of hardware firms into software firms. The third is what I call the ‘curse of genba shugi,’ which leads managers to insist on customized software that enshrines existing organizational practices, whether or not they confer competitive advantage. In the final section, I discuss the future trajectories of customization and packaged software.
Evolution of the software industry in Japan and the US With regard to the first dimension, the Japanese computer manufacturers of today rose out of the large integrated multi-product firms that originally supplied NTT. These large integrated electronic firms had few incentives to cooperate and competed fiercely with one another as they sought to differentiate product and services and lock customers into their own proprietary standards. The result was fragmented national standards (Cottrell 1996). Originally, this competition got played out in mainframes but the fragmented structure was carried over to PCs (Fransman 1995: 175). The trajectory can be seen most clearly in the experiences of NEC, which once dominated the domestic computer market but saw its shares plummet in the 90s. NEC was most effective, relative to its domestic competitors, in using
106
Software’s hidden challenges its proprietary operating systems to protect its market share, but in doing so isolated itself and the software needed to run its products from world markets. In 1982 it introduced its PC9801 using its proprietary operating system N88 Basic. It slowly migrated its support to MS-DOS, but with a proprietary graphic controller and expansion throttle that was incompatible with the IBM PC. Other leading mainframe producers followed suit. Because the dominant Japanese computer makers saw the PC as a front end to the mainframe, or as an entry level computer, they each developed incompatible standards to lock in their existing mainframe customers (Dedrick and Kraemer 1998: 80–1). Moreover, as Fransman (1995: 175) notes, apart from applications software, NEC retained all the design, manufacturing, and services required for its emergent PC business. Its subsidiaries provided a variety of services including a great deal of software and a variety of system integration services. NEC’s situation as a full service manufacturer and service provider, mimicked by other major Japanese competitors, meant that it was slow to see software as a product distinct from hardware. Software was provided for free as a way of boosting the sales of proprietary hardware systems (see Anchordoguy 2000: 395). Even today, customer firms are reported to be reluctant to pay extra for software. A key development took place in the early 1990s with IBM’s introduction of DOS/V for IBM PC/AT, a bilingual operating system for IBM and its compatibles, and the subsequent rise of application software for the new operating system. A second major development was Microsoft’s introduction in 1993 of a Japanese version of Windows 3.1 that could run both NEC and DOS/V hardware. This further broke down the wall separating the Japanese consumer market and component sourcing from the rest of the world. NEC’s dominant share of the market began to erode. The architecture of the IBM PC/AT was open and, as it evolved, it produced a variety of de facto standards. By the mid-1990s, NEC partially conceded by offering a customized version of Microsoft’s DOS/Windows for its 9800 series PCs. It capitulated in 1997 when it introduced fully compatible Windows machines, but by then its market share had fallen from above 50 percent in 1994 to 27 percent in 1998.3
Unbundling The example of IBM’s unbundling of hardware and software spread in the late 1970s, but Japanese computer manufacturers were slow to follow suit. When they did, they adopted halfway measures that involved assigning software development to their keiretsu contractors. There also arose user software spinoffs, many from financial sector firms. This structure for providing software services tended to stunt the development of independent software houses. The computer maker spinoffs and user spinoffs as keiretsu firms grew expert and quite comfortable at churning out code to strict contractor specifications,
107
Industries, technologies, and value chains instead of learning how to develop products for mass market individual and corporate customers (Hamilton 1993: B4). Corporate spinoffs seldom challenged incumbents with innovative software applications. Their focus was on achieving current usability for an individual firm and not innovation or standardization (except for pursuing standardization within a firm or among a group of cooperating companies). In 1990, independent software houses accounted for only an estimated 28 percent of all software sales, with user spinoff software houses accounting for 26 percent, computer vendor’s software sales accounted for 31 percent, and computer maker spinoffs for 15 percent (Baba et al. 1996: 108). Independent software houses throughout the 1980s found it particularly difficult to compete because the hardware makers were not required to publicize information about their hardware (Anchordoguy 2000: 396). By contrast, the US developed a thriving and diverse set of independent software firms focused on innovation. These firms began to develop in the 1960s with the period between 1965–70 marking the beginning of the independent software industry. It was spurred by the introduction of the IBM 360 with its standard operating system which increased the size of the installed base of mainframe computers that could use packaged software designed to operate specific applications (Mowery 1996: 24–5). Development further accelerated with IBM’s decision to unbundle the pricing and supply of its software and services from its hardware in 1969. Still further impetus for new entries resulted from the development of the microprocessor, IBM’s decision to outsource its components, and the rapid emergence of ‘dominant designs’. In particular, IBM’s effort to encourage a large number of applications and other programs for its PC encouraged new entries. Under increasingly open environments, software solutions were increasingly modularized; this acted as a catalyst for further product development, industry standardization, and encouraged new entries with no ties to hardware manufacturers (Ministry of Internal Affairs and Communications 2002: 4, 12). The result has been a vibrant industry filled with a continuous stream of newcomers. Consider only that between 1996–2001, software venture capital investments totalled 18.2 percent of all US technology venture capital investments (OECD 2002: 123).
The transformation of hardware into software firms With regard to the second dimension of the software revolution, most large US electronic and telecommunication companies started off as hardware companies but are increasingly becoming software companies. They initially treated software as a cost centre, but are increasingly coming to realize software is providing the most value added.4 As hardware commodification proceeds,
108
Software’s hidden challenges many firms seek refuge in software as a source of differentiation and higher margins. EMC, for example, a leader in selling data storage equipment, has been rapidly assembling an array of software capabilities (through acquisitions) for handling data and other tasks. This is, in large part, a response to the drift toward commodification of its hardware products (Forelle 2004: B3). Large telecommunication firms have gone perhaps the furthest in shifting their strategic focus to software. Firms like Alcatel, Nortel Networks, Lucent, and Ericsson, once known as equipment makers, now talk about themselves as software firms. Nortel Networks, the North American telecom equipment supplier, outsources most of the gear it builds and focuses internally on developing the software programs to run that gear. Software now drives most networking functions and allows new features to be added in the field (Delaney 1999: B8). Japanese telecom firms have lagged in these developments. Japanese firms still tend to assign most of their efforts for finding solutions for IT integration problems to hardware rather than software vendors.5 The net result is that software application vendors can expect a very small share of the solution revenue. This stunts the development of an independent software sector and leaves many large Japanese manufacturing companies with hardware-oriented solutions. Not unrelatedly, the fundamental premise at MITI in the 1980s and through the early 1990s was that hardware production was more important than software production and thus government policies to promote industry favoured expansion of hardware production and sales over software (Baba et al. 1996: 122; Anchordoguy 2000: 402). It is not surprising, too, since hardware (and the monozukuri culture that underlies it) provides the basis of current managerial capabilities and skill sets. Of course, many Japanese manufacturing managers are aware of the significance of the software revolution. Yet they find it hard to imagine a successful strategy of differentiation in the marketplace using software that would be competitive with Western hegemony in this area. This being the case, they fall back on their past ability to differentiate based on their hardware capabilities. The relative merits of hardware and software have changed along with the development of innovative new applications, however. Software solutions are often superior in term of cost, time to market, and design and field flexibility; firms can reprogram new features while the product is in the field rather than investing in new hardware (Delaney 1999: B8).6 Better software can often reduce the need for costlier hardware (Bulkeley 2003: A1, A6). Those that recognize these developments early are in a position to accelerate the process and gain competitive advantage from it. To be sure, the shift to software solutions has often been made uncritically, and there are cases where hardware solutions are better (such as when very high reliability is a high priority for users and when speed is critical. Software doesn’t always deliver on shorter time to market. Nevertheless, it is clearly the
109
Industries, technologies, and value chains case that software is increasingly being used to determine hardware functionality. The promise of what is known as adaptive or reconfigurable computing is one in which software is able to effectively redraw a chip’s physical circuitry (Markoff 2003: C1, C4). We can only expect software’s role to grow. An observation by Isao Okawa, the founder of CSK, one of Japan’s few successful software start-ups, summarizes the matter as follows: ‘It is the caliber of software that increasingly determines the preference for hardware’. While a number of perceptive Japanese managers may share his evaluation, few seem to have acted on his vision. One area where huge investment in software has taken place is ‘embedded software’. This technology has become vastly more important for Japanese manufacturers, with annual sales of embedded software products estimated to total US$500 billion a year in Japan (Yarime and Baba 2004: 8). Embedded software is incorporated into products such as computer printers, DVDs, cameras, scanners, and scientific instruments. The huge sales of such goods testifies to the capabilities of Japanese software engineers. Nevertheless, even with embedded software, one sees problems. Second generation phones are a major example where preference for hardware over software solutions, in the context of the specific business models being used, has limited the Japanese presence on global markets. European, American, and Korean handset makers rely on programmable processors using software for expressing features and other modes of differentiation. Their phones are more modular than Japanese handsets, with companies like Nokia developing platform modules (called engines) which can then be used for different models, thereby reaping huge cost savings as basic costs are spread out over many models. Nokia was initially forced into the platform model by its need to serve multiple carriers across the European market. In a similar fashion, Samsung, constrained by a small domestic market, aimed at the global market and thus had strong incentives to develop platforms for various models that would appeal to diverse global customers. By contrast, the Japanese handset manufacturers build their models ‘from the ground up’, both in terms of hardware and software. Carriers such as DoCoMo and Au (the KDDI brand) contract with their vendors to develop distinct phones from the basic hardware and compensate them by guaranteeing high volume sales or in some cases commit to paying a portion of the vendor’s R&D costs. This model reflects the continuing power of NTT vis-a`-vis the handset makers. It is also a rational response to market conditions. The Japanese handset makers, in contrast to the Koreans, are subject to the seduction of the large domestic market. The business model adopted by the carriers and the handset makers, based on custom built handsets, works admirably for the large Japanese domestic market, yielding strong profits to major industry players. In other industries, the Japanese have been able to build on the economies of scale
110
Software’s hidden challenges achieved in domestic markets to launch successful attacks on global markets. But this success assumes that key users in the domestic market are sufficiently in tune with global user needs and that the domestic business models can be applied to global markets. In the case of mobile phones, this has not been the case. Japanese handsets are more fine tuned, usually more compact, and with superior optimization for performance than those of Korean or European handset makers. The use of more expensive hardware than software solutions is a major factor in their ability to optimize designs. That said, the phones are very expensive since costs can not be spread out over many models; the heavier use of hardware further raises costs. Their failure to use more modular designs based on a more software oriented platform makes it difficult for them to spin out a large number of variations from a single model, which would reduce costs.7
Customization and the ‘curse of genba shugi’ We now turn to the third dimension, what I called the curse of genba shugi, which leads managers to insist on customized software that enshrines existing organizational practices, whether or not they confer competitive advantage. As discussed earlier, there are historical factors and subsequent path dependent trajectories that led Japanese firms to emphasize customization. These flowed from the proprietary technologies adopted by each of the hardware makers. Historical practices, however, need to be reinforced by current constraints if they are to continue to survive. The focus on ‘genba’ in Japanese firms–a source of great competitive advantage–has played this role. By genba, I mean a focus on the workplace where the actual work gets done and in particular to the workplace as the focus of production. Genba shugi is said to have its historical antecedents in the early Meiji period, when managers focused on the production shop floor as they tried to bridge the gap between traditional artisanal skill sets and the skill sets required by imported Western technology (Nakagawa 1990: 17–19). I elaborate this argument by first showing the relative weakness of the Japanese packaged software industry and then proceed to illustrate the problem with an analysis of Enterprise Resource Planning (ERP) packaged software.
Packaged software We can document the heavy reliance of Japanese firms on customized development and the limited use of packaged software. In Table 6.1, we see the results of a NTT Data survey commissioned by the Ministry of International Affairs and Communications. Seventeen percent of Japanese firms reported using packaged software with little customization compared to 29 percent of US firms; 26 percent of Japanese firms reported customizing their package
111
Industries, technologies, and value chains Table 6.1 Ways of building information systems
Japan United States
Packaged software with little customization (%)
Customized packaged software (%)
Developed or ordered custom software (%)
17 29
26 49
54 16
Source : Ministry of Internal Affairs and Communications (2004): Appendix, p. 340.
software vs. 49 percent of US firms while 54 percent of Japanese firms reporting developing their own or ordering custom software versus only 16 percent of US firms (Ministry of Internal Affairs and Communications 2004: Appendix, p. 340). The differences are quite striking. Are they significant? Fransman (1995: 188) noted similar differences, but concluded that they don’t reflect any technical competencies but rather the lucrative opportunities offered by the Japanese customized software market. Certainly he is correct as far as he goes, but by emphasizing the development of customized software for their Japanese customers, Japanese software firms have excluded themselves from participating in the global competitive market for packaged software. As such, they have foregone not only the chance to earn profits in this rapidly growing industry but, even more importantly, the opportunity to participate in the standardization of the software in ways that are particularly suited to Japanese practices and customer needs.8 Japan is a huge net importer of software. Its expenditures for software imports are 15 times higher than its receipts for software exports (METI 2003: 224). It is estimated that some 90 percent of packaged software sales in Japan are accounted for by foreign firms, and not just American; 20 of the top 100 software firms in Japan are European (Nezu 2002: 138). In the past, internationally competitive Japanese manufacturing firms had advantages in rapid product development and supply chain management, but a number of these advantages have been eroded in some key sectors in the last decade, in part because Western firms have been quicker at incorporating software applications to improve these processes taking a total systemwide perspective. Japan’s overall slowness in adopting packaged software applications has been documented in a METI survey. As can be seen in Figure 6.1, with the exception of CAD and EDI, Japan shows lower rates of adoption of various application software than competitor nations; these are applications for running the business. Japanese firms meet or exceed foreign competitors in only two of the listed applications: EDI and CAD. With the advent of the Internetbased applications and the development of XML, EDI is rapidly diminishing in importance. In the case of CAD, moreover, Japanese companies are said to have lagged behind US companies more recently in adopting 3d-CAD.
112
Software’s hidden challenges (%) 100
Japan USA Europe Asia (NIEs)
80
60
40
20
0 CAD
ERP
SCM
CRM
KM
EDI
Figure 6.1 International comparison of rate of introduction of IT-related applications Notes : CAD figures are for the manufacturing industry, SCM figures are for the manufacturing, wholesale and retail industries, and the others are for all industries. Source : International Survey of Corporate Management Strategies (METI)
Japanese managers will often claim that their failure to adopt a broad range of packaged applications reflects the uniqueness of their organizational requirements vis-a`-vis Western competitors. It is interesting to observe from Figure 6.1, however, that the rate of adoption of these applications by their direct and highly successful Asian competitors from Korea and Taiwan is significantly higher than for Japan. Do they not also have unique organizational requirements?
Enterprise resource planning (ERP) We turn our attention now to one of the key IT applications of the 1990s– Enterprise Resource Planning (ERP) software. ERP application software originated as a solution to automating information entry and processing across the major functions of a company, integrating this information in a predetermined way, and allowing all of these business functions to access this information in a real-time environment. The software subsequently evolved to include information from supply chain partners and end customers. These kinds of applications are based on modelling the respective business processes and then creating a standardized best practice model.
113
Industries, technologies, and value chains Before going further, we should note an interesting contradiction among US firms. Although adoption of ERP is widespread among US firms, few users say they received the full benefits they expected and there is a widespread view that ERP software projects have a negative return on investment (ROI) (Braue 2003; McKinsey and Company 2003). The potential direct measurable benefits from ERP, as reflected in reduced headcount, and improved performance metrics resulting in greater revenue and/or lower costs, typically do not compensate for the high total cost of ERP acquisition, implementation and maintenance (Davenport et al. 2003).9 Although proponents of ERP argue that its benefits only grow over time, from the perspective of failed expectations, its adoption by large numbers of US firms appears irrational. Assuming reasonably efficient information flows across firms, one would expect that a bad perceived experience by one firm would be quickly communicated to other potential adopters and they would be reluctant to adopt ERP. Why didn’t this occur? Further analysis documents that there are a variety of indirect and more difficult to measure benefits that flow from ERP software projects that are not included in typical ROI calculations. First, there are cumulative enhancement benefits (improvement in existing software functionalities and more functionality), delivered in upgrades, that are not typically experienced in the first year or so after adoption. For example, the standardized ERP package has become more and more configurable (without requiring programming skills) and thus easier to personalize to fit specific corporate needs. Second, the provision of more accurate, more timely, and more accessible business data is a commonly recognized ERP benefit and has a number of potential strategic outcomes. These include, for example, improving customer relationships by allowing better customer management (Davenport et al. 2003), which should translate into increased revenue via higher repeat sales. Improved employee productivity is another, but it is so intertwined with other factors that it is hard to separate out its effects. The largest indirect benefit, however, is probably the facilitation of better corporate strategic decision making, resulting from access to higher quality and more timely data. This allows better and speedier identification and targeting of new strategic objectives and activities. Like any capability (resource) with extremely broad possible applications, some firms will move faster and more effectively than others to extract value from it. Some firms have a higher ‘IT IQ’. This view runs counter to that espoused by Rapp (2004), who argues that capabilities enabled by standardized software confer no competitive advantages because they are available to all,10 but is consistent with Jim Collins (2001: 79) who found that there is no evidence that good to great companies had more or better information than the comparison companies; what set them apart rather was their ability to use information strategically. In summary, despite the absence of strong direct
114
Software’s hidden challenges benefits in terms of reduced costs, we can make a prima facie case that American firms have been more willing to invest in ERP because of the indirect strategic benefits that may ensue. We don’t have comparable data for Japanese ERP adopters, but there is reason to believe that many of the early adopters in Japan also had bad experiences with ERP, that these were also widely publicized, but that in Japan these reports did lead to a reluctance among potential followers to adopt. What might account for the different responses of Japanese managers to the same data as their American counterparts, and ultimately the difference in adoption rates suggested by Figure 6.1 (which probably also applies to the number of ERP modules used in addition to overall adoption)? Adding to our understanding of this matter is data from a Japanese government survey of US and Japanese firms on their perceptions of the effectiveness of their IT investments (a category broader than just ERP, but revealing nevertheless). According to Figure 6.2 the perceived effectiveness of IT investments was much greater for US than for Japanese firms in the following areas: increasing sales, winning new customers, development of higher value added products/ services, increases in customer satisfaction, and in improvement in product and service quality (Ministry of Internal Affairs and Communication 2003: 15). By contrast, when it came to reducing costs and increasing operational efficiency, US and Japanese managers had relative similar perceptions of the
Increase in sales 80
48.3 60
Improvement in the quality of products/services
35.1 70.5 48.5 20
Winning new customers
36.3 24.5
26.5 42.2 53.7 73.4
Increase in customer satisfaction
Development of high added value of products/services Japan
US
Figure 6.2 Proportion of Japanese and US companies that found their IT investment effective
115
Industries, technologies, and value chains effectiveness of IT investments. We can infer from these findings that US firms seem to value and believe in the indirect, hard to measure strategic benefits of IT investment far more than the Japanese firms, which concentrate on direct operational benefits. In short, it makes sense that the Americans would be more willing to invest in ERP software since they seem more focused on and confident that they will be able to extract indirect strategic benefits. The Japanese by contrast are interested primarily in the hard direct benefits relating to operational efficiencies, and when these do not seem to exceed investment costs, they are less willing to invest in ERP. This is consistent with long-term findings that Japanese manufacturing firms orientate themselves more to process than product innovation (Mansfield 1988: 225–6). It is also the case, as we shall see, that Japanese executives had more confidence in customized solutions. One measure of the degree to which IT investment is viewed as strategic is the proportion of firms with a Chief Information Officer (CIO) post. The percentage of firms in all industries in the US with a full-time CIO (56 percent) is reported to be nearly five times higher than in Japan (Yokota 2001: 29). It is also reported that while 51 percent of American CIOs state that they participate in designing corporate strategy, only 13 percent of Japanese CIOs make a similar claim (Ministry of International Affairs and Communication 2003: Appendix, p. 340). Parenthetically, it is plausible that competitive pressures among CIOs in large US corporations, fuelled by success stories in IT and CIO journals and subsequent attempts by individual CIOs to emulate their peers, played a major role in the frenzied adoption of IT in the late 1990s.
Customization We turn now to the flip side of adopting application software–customization. Customized software is far more expensive (according to one estimate, Japanese companies are paying upwards of eight times more for customization and services of ERP-like applications than for a licence for ERP packaged software).11 Of course, there are implementation costs for packaged software, so the gap is not this large (though, as discussed below, long-term maintenance and upgrade costs are much higher for customized software). By customizing so heavily, Japanese firms are getting applications that are finely tuned to their current practices. They are getting these functionalities, however, at a tremendously high cost relative to packaged software, costs which escalate with upgrading and efforts to network with other companies to maintain the customized software’s utility in face of changing circumstances. In the fierce global competitive environment faced by many ICT firms, the additional costs imposed by the use of customized software can be a heavy burden. (See Yunogami, this volume, for a comparable example of the costs imposed by excessive customization of hardware for DRAM production.)
116
Software’s hidden challenges For firms that are industry leaders with finely honed business processes that confer strong competitive advantage, it makes perfect sense to develop customized software internally to reflect and build on these practices. Such is the path taken by many industry leaders like Toyota, Nippon Steel, and Ito-Yokado as well as Western companies like Dell and Nokia (see Rapp 2002. We note, though, that even some of these companies will use packaged software for noncore business processes and even core business processes such as product design.) To emulate these companies, as so many seek to do, firms need to develop superior proprietary business processes; only then typically does it make sense to build your own customized software for those business processes that confer competitive advantage. To jump immediately to customized IT solutions for most business processes may only lock a firm into inferior and expensive solutions. Where the technology and the know-how are widely available, one can make a strong argument that the right approach to closing performance gaps with market leaders is by using reliable low cost off the shelf applications (see Farrell et al. 2003: 4). Regardless of geographical location, the best companies use standard applications packages for routine and certain specialized (e.g., chip design) business tasks and aim to minimize customization. By doing so they put themselves in the position to do inexpensive and rapid upgrades. Infrastructure is kept as uniform as possible to ease integration among departments and geographically diverse sites. In this way, IT spending shifts from maintenance and system integration toward developing new applications–the ultimate driver in the creation of business value (Lohmeyer et al. 2002: 8). While pressures for customization are prevalent among US and European firms, Japanese manufacturing firms, as we have seen, are especially prone to engaging in extensive customization.12 This is typically explained by the desire of firms to differentiate their business from competitors: ‘Because even Japanese organizations that belong to the same industrial sector usually have different management policies and organization, it makes sense for the processes to be customized than to be purchased as prepackaged programs’ (Baba et al. 1996: 124). This reasoning applies just as well to US, European, and Korean firms, however, which are nonetheless much more receptive to using packaged software. If the point of customization is to differentiate, moreover, it is hard to understand the oft-noted observation that Japanese competitors in the same industry seem to excessively mimic each other’s strategies and practices. A more substantive argument for customization is that Western software, with a presumed top–down decision making model, and presuming a different division of labour in work organization, doesn’t fit Japanese decision making practices (Fujimoto 2004: 318–21). Insofar as software firms use leading Western firms (lead users) to model best business processes in designing their software products, there is some truth to this argument.
117
Industries, technologies, and value chains This explanation, we note, is intertwined and reinforced with the ‘curse of genba shugi’. Masahiko Aoki (1988) provides an analytical perspective on these characteristics. He describes the internal information structure of the Japanese firm (the J-firm) as relatively more decentralized and with relatively more autonomous problem solving at individual work units than those of Western firms, while at the same time the J-firm relies more on horizontal communication among functional units rather than hierarchy. These decentralized practices, in turn, are based on worker integrative skills as opposed to the development of highly specialized jobs as more likely occurs in Western firms. So tasks such as responsibility for quality, which evolved in the West as specialized staff jobs devolved in Japanese firms into the workplace as production unit and production worker responsibilities (see also Udagawa et al. 1995; Tsutsui 1998: 190–235). This has indeed been a major strength of Japanese manufacturing, especially in industries where different operational divisions are highly complementary and information is effectively and efficiently shared across these units. The emphasis on delegating operational problem solving tasks to those who have the relevant on-site knowledge makes this decentralized decision making a quite effective approach, as has been stressed in some breathless accounts of Japanese focal factories (see Fruin 1997: 30–1). As the ancient Greeks (e.g., Pythagoras) remind us, however, even beneficial concepts and practices, when taken to the extreme, begin to suffer negative effects. To put it more generally, ‘excess turns virtues into vices’. In this case, it can lead to an emphasis on local optimization at the expense of company-wide optimization, and to turf protecting behaviour. Applied to IT, Japanese manufacturing firms have tended to stress optimizing IT at the plant level (see Shimizu 2001: 136–7) while forgoing many of the benefits that would derive from firm-wide data consolidation. The focus has been on maintaining shop floor usability (Baba et al. 1996: 124), rather than on organization-wide optimization. This is seen most clearly in the slow uptake of applications that aim at supply chain optimization. In the US, improving overall supply chain efficiency is a major driver of ERP and SCM sales. In all firms, regardless of nationality, small business units exhibit a strong tendency to want to use IT to optimize their processes; they tend to resist IT when it doesn’t promise that outcome–even if told that their acceptance will help the whole organization. This resistance is especially fierce in Japan because production units tend to have a good deal more autonomy and power–genba shugi is strong–thereby often sabotaging systemwide optimization objectives. Many firms allocate IT budgets to each division separately and there is usually complete freedom in how they use that budget. Managers of these divisions stress the efficiency of the day to day activity at their genba rather than optimizing at a broader system level or trying to achieve a long-term reduction of overall IT spending. This tends to result in heavy customization.13 It also makes
118
Software’s hidden challenges it difficult to achieve synergies across divisions in this era of system solutions and technology fusion. In one large Japanese electronics manufacturer which I studied there was fierce opposition to the proposed implementation of a packaged ERP system from the shop floor because the proposal was seen to detract from plant level efficiency. The corporate managers driving the implementation process aimed to use ERP to move from monthly production planning to weekly production planning. The plant managers, however, liked the old system of monthly production planning. From their point of view, moving to weekly production planning was not optimal because it detracted from efficiency of their operation. They would have to purchase raw materials and parts in smaller quantities, raising their costs. Moreover, with weekly production, they would have to change schedule frequently, thereby raising changeover costs and increasing probability of defects. Weekly-based production was much more challenging. The plant personnel failed to consider, however, that under the current system, the company had to maintain large amounts of expensive inventory, largely because the products being produced were not closely tuned to actual customer demand as would likely be more the case with weekly production schedules. The plant personnel focused only on optimizing their own unit (genba) and not on systemwide benefits that would result from adjusting their behaviour.14 The core of the problem lies in the assumption that all current practices are the best possible. This is the true curse of genba shugi. When the practices required by new application packages don’t match up, managers manifest an instinctive tendency to customize the application to fit current practices. This also naturally leads to maintaining legacy IT systems. Yet there are huge costs associated with customizing for system integration, and especially for long-term maintenance and future upgrades. And as time goes on, these standardized software packages get better and better at modelling real business processes and are, as noted earlier, becoming more and more flexible in allowing greater configurability to account for specific business needs. To ensure that current practices are assessed for their ability to confer competitive advantage and to use packaged software like ERP to achieve strategic objectives takes strong management leadership. In a large number of Japanese ICT firms, however, ERP implementation is given over to either internal IT department personnel or outside system integrators. The former typically do not have strong clout or understanding to assess the value of existing practices; they lack deep knowledge of functional processes. The external system integrators, on the other hand, typically do not have the incentive or the knowledge to engage in such systematic analysis. Under these constraints, IT managers and outside system integrators are little match for strong functional managers who seek to preserve existing practices. Even when outside system integrators want to do the right thing, they
119
Industries, technologies, and value chains complain that it is very hard to identify suitably strong management leaders who will take the initiative in implementation. Both of these problems, in turn, reflect the lack of pursuit of strategic objectives in adopting IT in many–though certainly not all–Japanese manufacturing firms. In this environment, Japanese firms are probably better off not even attempting ERP implementation. For those exceptional Japanese firms noted for their strategic use of customized software to enhance their competitive status, IT managers are said to have deep functional knowledge (Rapp 2004). Manifestation of resistance to change can be found in the findings of the aforementioned NTT Data survey, which examined the measures taken by Japanese and US companies toward optimizing IT investment effectiveness. As shown in Figure 6.3, only 42 percent of Japanese respondents reported reform of organizational practices to comply with IT system operations, vs. 65 percent of US respondents. Similarly, while almost 83 percent of US respondent firms reported conducting operational process reviews (e.g., establishing workflows) to comply with IT systems, only 69 percent of Japanese respondent firms reported doing so (Ministry of Internal Affairs and Communications 2004: 17). We might further hypothesize that when these changes are undertaken, they are likely to be more far reaching in American firms. What are the costs of using IT to preserve and enshrine current practices? Bruce Harreld, chief strategist at IBM, notes: ‘Just spending money on IT never creates any value. It’s what you do differently in terms of business processes that matter’ (Schlender 2003: 82). If IT installations are followed by heavy customization that preserves existing business processes, not all that much has been accomplished from a competitive point of view. The evidence suggests that for all the success the Japanese have had in using IT to rationalize the shop floor, they are, as noted earlier, more reluctant than their American counterparts to use IT for strategic purposes such as business expansion (see Motohashi 2004). The relative reluctance of Japanese firms to adopt standardized software applications for the many existing business practices that do not confer competitive advantage does not bode well for Japan’s competitive strength in the future. Viewed over the last 30 years, Japanese manufacturing firms are slowly migrating away from closed customized software solutions to open packaged systems. Because of the ‘path dependent’ nature of knowledge and practices, this has been a slow evolutionary process (see Baba et al. 1996: 117). The adoption of the MS–Windows platform was a major step forward toward eliminating fragmented standards in operating systems. Japanese firms will have to move much further and faster, however, if they are going to be able to effectively draw the full benefits of combining standardized solutions, semicustomized and customized ones as strategic tools for running the firm.
120
Re-utilization of generated effects (e.g., reduced cost) in business management (investment in new fields, etc.) Thorough communication of the background of introduction and the post-introduction vision of the IT systems to employees
Verification of cost effectiveness before introduction 73.2 80 73.2
Review of operational processes 82.7 (e.g., establishing workflow) to comply with IT system
82.9
60 57.8 48.4
66.2
40
53.1 33.5
20
68.8
13.5 Reform of organization and systems to comply with IT system operations
Establishment of a quantitative effectiveness index for verification of cost effectiveness
62.4
41.8 Regular quantitative verification of effect after introduction
64.9 64.1
50.5
Selection and concentration of operations (clarification of core competence/labour saving rationalizing and outsourcing measures for noncore operations)
73.3 Top management makes decisions on IT-related investment based on the company’s circumstance
74.4 79.6
Formulation of IT strategy based on business strategy This figure shows the percentages of companies that answered either ‘fully implemented’ or ‘somewhat implemented’ for the respective measures
Japan
121
Figure 6.3 Measures by Japanese and US companies to optimize effectiveness of IT investment
US
Software’s hidden challenges
58.0
Industries, technologies, and value chains
The future Arguments that customized and semi-customized applications win the day (e.g., Rapp 2002) are betting against the history of industrialization. Standardization has been at the heart of productivity improvement since the days of Henry Ford and the development of standardized interchangeable parts. Under most circumstances, standardized components are less expensive than a component designed and produced for one time use. The lower cost is obtained largely through producing in larger volume, and achieving greater learning effects and economies of scale. Standardized components also produce higher performance (for a given cost) than one of a kind designed components (Ulrich 2003: 132–3). These are powerful principles that have demonstrated their power again and again in the course of industrialization. There are no convincing reasons to think that, over time, the software industry will be exempt from these principles. The software application industry is still in its infancy and many of the standardized solutions have initially ill-fit the needs of industry. Five years ago, the IT firms were saying to their customers the equivalents of Henry Ford’s, ‘You can have the Model T in any colour, so long as it’s black’. Now personalization is increasingly the objective of IT firms’ efforts, with the customers more in the driving seat. Japanese firms are only marginal participants in these developments, to their disadvantage. At the same time, developing innovative products and services often does require customized solutions both to create and protect competitive advantage. The future, enabled by web services, may be one in which firms eschew the purchase of large monolithic ERP packages from one vendor. Instead they may subscribe to web services that will give them the specific pieces of needed functionality in modular form. They will choose from multiple vendors competing to show they have the most suitable products. In this fashion, they will build up many capabilities from numerous vendors to create seamless experiences for users. Alternatively, they also have the option of buying prepackaged composites from software vendors. In either case, firms will access these services over the web using http with published Application Programmer Interfaces (APIs). These APIs will enable a smooth integration of the different modules into a larger system and enable firms to cost effectively access these services. Should this vision come to pass, we will witness a process in which the modularization of the PC, and all its attendant consequences for vendor competition described earlier in the chapter, comes to be paralleled by how software is developed, sold, and used. Japanese firms continue to lag in their participation in these evolutionary developments.
122
Software’s hidden challenges
Notes 1. The author is indebted to the Doshisha Business School and Doshisha University’s Center of Excellence (COE) Program for financial support for this research as well as Ford Motor Co. IT Research Grant to the Management of Technology Program, Haas School of Business, UC Berkeley. Funds were also contributed by the Center for Japanese Studies at UC Berkeley. The author benefited from the research assistance of Yasuyuki Motoyama and Toru Ebata. Finally, I would like to especially thank Michael Ger of Oracle USA, Hideyuki Yamagishi of Oracle Japan, and Josh Greenbaum, Principal, Enterprise Applications Consulting (specializing in SAP applications) for explaining the many fine points of Enterprise Resource Planning software. Comments from Dimitry Rtischev, Hitotsubashi Business School (ICS) were also helpful. My colleague at Doshisha University, Toshiro Kita, played a special role in challenging my interpretations. None of these individuals are responsible for the use I have made of their comments. 2. I am indebted to communications with and research documents from Kazuyuki Motohashi of the Research Center for Advanced Science and Technology, University of Tokyo. 3. This section draws heavily from Fransman (1995: 174–5). 4. I am indebted to my colleague David Messerschmitt of the Electrical Engineering and Computer Science Department of UC Berkeley for this formulation. 5. Interviews with selected Oracle managers, Nov. 2002. 6. Conversations with colleagues at Berkeley Wireless Research Center, UC Berkeley, 17 Feb. 2000. 7. Individual Japanese handset makers cannot easily shift to a modular strategy, on the other hand, without sacrificing their profitable domestic market share to existing competitors. This section drew heavily from the observations of Kimio Inagaki, President Jabil Circuit, Japan. I also benefited from conversations with Prof. Jan Rabaey, Scientific Co-Director of the Berkeley Wireless Research Center. Neither of these individuals are responsible for my interpretations. 8. Despite Japan being the second largest economy in the world, only NTT Data (ranked 7th in sales) makes the list from Japan of the 20 largest software firms in the world (defined as having 90 or more percent of sales derived from software and services). Moreover, like many Japanese software firms, and unlike many of the other entries on this list, NTT Data derives much of its sales from customized software (Yamaguchi 2004: 72). 9. We should keep in mind that the studies of ERP’s ROI vary widely in methodologies and what and how they count lending a strong subjective element to these findings (Alter 2003: 20). They also measure ROI at different time periods since implementation. 10. This discussion draws from research on the secondary literature by Ian Larkin, Ph.D. student, Haas School of Business, UC Berkeley. 11. Conversations with selected Oracle managers Nov. 2002. Anchordoguy (2000: 397) also mentions in passing the heavy costs imposed by the Japanese preference for customized software based on closed standards.
123
Industries, technologies, and value chains 12. Interview with Risaburo Nezu, Fujitsu Research Institute, Dec. 2002. When packaged solutions are clearly superior, however, such as for chip design, Japanese firms like NEC eventually use a great deal of packaged software (Rapp 2002: 35). 13. I am grateful to Eugene Y. Kawamoto, IBM Japan, for these observations. 14. See the ‘Naniwa Tech’ case by Cole 2004.
References Alter, A. (2003). ‘ROI Roundtable’, CIO Insight, Oct. 15 (1): 1–20. Anchordoguy, M. (2000). ‘Japan’s Software Industry: A failure of institutions?’ Research Policy, 29: 391–408. Aoki, M. (1988). Information, Incentives, and Bargaining in the Japanese Economy, Cambridge: Cambridge University Press. Baba, Y., S. Takai, and Y. Mizuta (1996). ‘The User-Driven Evolution of the Japanese Software Industry: The case for customized software for mainframes’, in D. Mowery (ed.) The International Computer Software Industry, New York: Oxford University Press. Braue, D. (2003). ‘ERP: Carving a new niche’, ZDNET, http://www.zdnet.com.au/news/ business/0,39023166,202724921,00.htm Bulkeley, W. (2002). ‘IBM to Spend More on Services’, Wall St. Journal, 20 Nov.: B5. —— (2003). ‘For Clues to Why Tech is Still Down, See Mr. Kheradpir’, Wall St. Journal, 3 Mar.: A1, A6. Cole, R. (2004). ‘Naniwa Hi Tech: Implementing ERP’, Business case, Doshisha Business School, Kyoto, Japan. Collins, J. (2001). Good to Great, New York: HarperCollins. Cottrell, T. (1996). ‘Standards and the Arrested Development of Japan’s Microcomputer Software Industry’, in D. Mowery (ed.) The International Computer Software Industry, New York: Oxford University Press. Davenport, T., J. Harris, and S. Cantrell (2003). Better Things Come to Those Who Wait: Enterprise system benefits, Accenture Institute for Strategic Change. Dedrick, J. and K. Kraemer (1998). Asia’s Computer Challenge, New York: Oxford University Press. Delaney, K. (1999). ‘Telecom-Equipment Concerns Focus on Software’, Wall St. Journal, Oct. 18: B8. Farrell, D., T. Terwilliger, and A. Webb (2003). ‘Getting IT Spending Right this Time’, McKinsey Quarterly, http://www.mckinseyquarterly.com/article_print.aspx?L2¼13& L3¼13&ar¼1285 Forelle, C. (2004). ‘EMC to Buy Software Firm Smarts’, New York Times, 21 Dec.: B3. Fransman, M. (1995). Japan’s Computer and Communications Industry. Oxford: Oxford University Press. Fruin, M. (1997). Knowledge Works. New York: Oxford University Press. Fujimoto, T. (2004). Nihon no monozukuri tetsugaku (The Philosophy of Japanese Monozukuri), Tokyo: Nihon keizai shinbunsha. Hamilton, D. (1993). ‘U.S. Companies Rush in to Fill Japanese Software Void’, Wall St. Journal, 7 May: B4. Jorgenson, D. and K. Motohashi (2003). Economic Growth of Japan and the United States in the Information Age, Tokyo: METI, RIETI Discussion Paper Series 03-E-015.
124
Software’s hidden challenges Lee, G. and R. E. Cole (2003). ‘From a Firm-Based to a Community-Based Model of Knowledge Creation: The case of the Linux Kernel development’, Organization Science, Nov.–Dec., 14(6): 633–49. Lohmeyer, D., S. Pogreb, and S. Robinson (2002). ‘Who’s Accountable for IT?’, The McKinsey Quarterly, 4: 1–9. Mansfield, E. (1988). ‘Industrial R&D in Japan and the United States: A comparative study’, American Economic Review Papers and Proceedings, May, 78(2): 223–8. Markoff, J. (2003). ‘Computing’s Big Shift: Flexibility in the chips’, New York Times, 16 June: C1, C4. McKinsey and Company (2003). ‘Enterprise software: Where now’, McKinsey Quarterly, (Summer): 45. METI (2003). Trends in Japan Industrial R&D Activities: Principal Indicators and Survey Data, Tokyo: Ministry of Economy, Trade and Industry; Industrial Science Technology policy and Environment Bureau, Technology Research and Information Office. Ministry of Internal Affairs and Communications (2002). ‘Kokusai kyoso ryoku no tame no kigyo no IT ka senrakyu kenkyukai’ (Research group for promoting IT within firms for the purpose of increasing global competitive power), http://www.soumu. go.jp/s-news/2002/pdf/021220_3_02.pdf —— (2003). Kigyo keiei ni okeru IT katsuyo chosa (IT Practical Survey of Company Management), Tokyo: Ministry of Internal Affairs and Communications. —— (2004). 2003 White Paper on Information and Communication in Japan, http://www. johotsusintokei.soumu.go.jp/whitepaper/eng/WP2003/2003-index.html. Motohashi, K. (2004). Building an Information Infrastructure for Knowledge Based Economy, Part B: ICTusers in Japan, Paper presented to K4D Hitotsubashi Seminar, Tokyo, 13–14 Nov. Mowery, D. (ed.) (1996). The International Computer Software Industry, New York: Oxford University Press. Nakagawa, K. (1990). Nihon kigyo no keiei kozo no hikaku shiteki kosatsu (Reflections on Comparative History in Management Organization of Japanese Firms), in K. Nakagawa (ed.) Kigyo keiei no rekishiteki kenkyu (Historical Study of Firm Management), Tokyo: Iwanami. Nezu, R. (2002). ‘IT sengoku jidai’ (IT’s Warring Period), Tokyo: Chuo koron shinsha. OECD (2001). OECD Science, Technology and Industry Scoreboard, Paris: Organisation for Economic Co-Operation and Development. —— (2002), OECD Information Technology Outlook, Paris: Organisation for Economic CoOperation and Development. Oliner, S. and D. Sichel (2002). ‘Information Technology and Productivity: Where are we now and where are we going?’ Federal Reserve Bank of Atlanta Economic Review, Third Quarter: 15–44. Rapp, W. (2002). Information Technology Strategies: How leading firms use IT to gain advantage, New York: Oxford University Press. —— (2004). ‘Information Technology Strategies: How leading firms use IT to gain an advantage’, Presentation given at the Doshisha Business School, 22 Oct. Schlender, B. (2003). ‘The Next Battles in Tech’, Fortune, 12 May: 80–2. Shimizu, K. (2001). ‘Kodo johoka no jidosha kumitate Shokuba’ (Moving Toward Information-based Activity in an Automotive Assembly Workshop), in K. Odaka and T. Tsuru (eds.) Dejitaru-ka jidai no sosiki kakushin (The Organizational Revolution in the Digital Age), Tokyo: Yuhikaku.
125
Industries, technologies, and value chains Stein, T. (2004). ’Big Strides for ERP’, Information Week, 715: 67. Tsutsui, W. (1998). Manufacturing Ideology: Scientific management in twentieth century Japan, Princeton: Princeton University Press. Udagawa, S. H., K. Nakamura, and I. Nonaka (1995). Nihon kigo no hinshitsu kanri (The Quality Management of Japanese Companies), Tokyo: Yuhikaku. Ulrich, K. (2003). ‘The role of Product Architecture in the Manufacturing Firm’, in R. Garud, A. Kumaraswamy and R. Langlois (eds.) Managing in the Modular Age, Oxford: Blackwell. Yamaguchi, E. (2004). ‘Sofuto sangyo no bokko wo hagukumu Nihonteki jijyo’, (Japanese style circumstances which give rise to barriers for growth of the software industry), Nikkei Biz Tech, 004: 68–74. Yarime, M. and Y. Baba (2004) ‘Dynamics of Embedded Software Development: Coevolution of OS standards and community networks in Japan’, Working Paper, Research Center for Advanced Science and Technology, University of Tokyo. Yokota, T. (2001). ‘Abstract of the White Paper on International Trade 2001’, Journal of Japanese Trade and Industry, July/ August: 25–9.
126
Part 2 MOT in and between enterprises
This page intentionally left blank
7 The open innovation model: Implications for innovation in Japan Henry W. Chesbrough
Introduction Industrial innovation is becoming more open, requiring changes in how firms manage innovation. External sources of knowledge become more prominent, while external channels to market also offer greater promise for utilizing internal knowledge. This elevates the importance of the evaluation of early stage technology projects, which often involve significant technical uncertainty and significant market uncertainty. Companies need to ‘play poker’ as well as chess, in such circumstances. Measurement errors (false positives, false negatives) are likely to arise from judgements about the commercial potential of early stage projects. Most companies’ policies consciously limit ‘false positives’ in assessing a project’s commercial potential, but few companies take steps to manage the risk of ‘false negatives’. New metrics may help a firm focus more upon external sources of innovation to enhance its business model, and enable the firm to salvage value from false negatives that otherwise would be lost. Open innovation should not be understood to mean the lack of any internal mechanism to capture value from innovation. It may be better viewed as an open–closed process, where the openness helps to create value throughout the value chain of the company, its suppliers, its customers, and the ultimate end users. Once value is created, which means that the technology has been embraced by the players in the value chain, the closed-ness helps to claim a portion of that value. Openness also has important implications for the boundary of the firm, particularly in Japan. Technologies should move more fluidly between organizations and, at times, new organizations may be the most effective means of pursuing a new technology opportunity. Spin-offs stand as a mechanism to
129
MOT in and between enterprises manage the risk of false negatives and explore new business models to create and capture value in the Japanese context, though much remains unknown about their effectiveness in practice in doing so.
The shifting process of industrial innovation Not long ago, internal research and development was viewed as a strategic asset, and even a barrier to competitive entry in many industries. Only large companies with significant resources and long-term research programmes could compete. Research-based companies like DuPont, Merck, IBM, GE, and AT&T did the most research in their respective industries. And they earned most of the profits as well. Rivals who sought to unseat these firms had to ante up their own resources, and create their own labs, if they were to have any chance against these leaders. There were significant economies of scale in R&D, and the biggest companies generally developed the best technologies. Vertical integration was the dominant business logic of the last century. Underlying the logic was the belief that valuable knowledge was fundamentally scarce. As a result, companies sought to develop a knowledge advantage that others could not match. This corporate world view brought with it a number of working assumptions: . The company which gets an innovation to market first, will win. . If you create the most, and the best, ideas in the industry, you will win. . The smart people in our field work for us: Companies competed for the best and the brightest graduates and offered these recruits the best salaries and equipment. . If we discover it ourselves, we will get it to market first: Internal R&D was seen as a barrier against smaller competitors. . To profit from R&D, we must discover it, develop it, and ship it ourselves: The rise of companies like DuPont, General Electric, General Motors, IBM, Xerox, Merck, and Procter & Gamble were all fuelled by sustained investment in internal R&D. A by-product of this emphasis was the ‘not invented here’ syndrome, where companies rejected any technology that had come from outside. . We should control our intellectual property so that our competitors don’t profit from our ideas. For most of the twentieth century, this closed innovation model worked, and worked well. We can thank it for a whole range of inventions and developments. It enabled Thomas Edison to invent the phonograph and the electric light bulb among other things. In the chemicals industry, companies like DuPont established central research labs to identify and commercialize an amazing variety of new products such as the synthetic fibres nylon, Kevlar,
130
The open innovation model and Lycra. And Bell Labs researchers discovered groundbreaking physical phenomena and harnessed those discoveries to create a host of revolutionary products, including transistors and lasers, winning a number of Nobel prizes for their research along the way. However, this model of closed innovation ran into severe problems towards the end of the century. By way of illustration, compare Lucent, which inherited the lion’s share of Bell Laboratories after the breakup of AT&T, with Cisco. Bell Labs was perhaps the premier industrial research organization of the last century. Within the old model of innovation, this heritage should have been a decisive strategic weapon for Lucent in the telecommunications equipment market. Yet, Cisco, without the deep internal R&D capabilities of Bell Labs, has consistently managed to stay abreast of Lucent, occasionally beating it to market. Today, Cisco dominates the telecommunications equipment market, while Lucent (which nearly went bankrupt at one point) is a distant follower. How can Cisco’s relative success vs. Lucent be explained? The two organizations were simply not innovating in the same manner. Lucent was a classic example of a closed innovator, devoting enormous resources to exploring the world of new materials and state of the art components and systems, seeking fundamental discoveries that could fuel future generations of products and services. In contrast, Cisco deployed a very different, and far more open, strategy. Whatever technology the company needed, it acquired from the outside, usually by partnering or investing in promising start-ups (some, ironically, founded by ex-Lucent veterans). In this way, Cisco was able to keep up with the R&D output of perhaps the finest industrial R&D organization in the world, and without doing much internal research of its own. The story of Lucent and Cisco is hardly an isolated instance. IBM’s research prowess in computing provided little protection against Intel and Microsoft in the personal computer business. Similarly, Motorola, Siemens, and other industrial titans watched helplessly as Nokia catapulted itself to the forefront of wireless telephony in just 20 years, building on its industrial experience from earlier decades in the low-tech industries of wood pulp and rubber boots. And pharmaceutical giants like Merck and Pfizer have watched as a number of upstarts, including Genentech, Amgen, and Genzyme, have parlayed the research discoveries of others to become major players in the biotechnology industry. These days, the former leading industrial enterprises are finding remarkably strong competition from many newer companies. These newcomers conducted little or no basic research on their own. They have been very innovative, but they have innovated with the research discoveries of others. And there is a legion of other, even newer, companies waiting to supplant these firms, if an opportunity should arise. To make matters worse, some companies that made significant long-term investments in research found that some of the resulting output, however
131
MOT in and between enterprises brilliant, wasn’t useful for them. They found ways to gracefully exit from further funding of these projects, and moved on to more promising work. Then, to their amazement, some of those abandoned projects later turned into very valuable companies. This was the experience of the Xerox Corporation, for example, with its Palo Alto Research Center. Numerous valuable computer hardware and software innovations were developed at PARC, but few of them made any money for Xerox and its shareholders.
The shift to open innovation Over the past two decades, the management of innovation has fundamentally changed. It is still true that no company can grow and prosper without new ideas. It is also clear that the changing needs of customers, increasing competitive pressure, and the evolving abilities of suppliers necessitate continual creative thinking for a company to stay ahead of the pack. The challenge is that the distribution of this critical knowledge has shifted from being locked up in the corporate laboratories of the biggest firms in the industry, to being dispersed among for-profit firms of all sizes, and non-profit organizations like universities and research institutes. This has important implications for how every company thinks about growth and innovation. There are many fewer economies of scale in R&D today. The reasons behind this basic change are many and varied. In the United States one factor was the success of the GI Bill which increased college numbers in the postwar years. Other factors include the rise in the amount and quality of university research, the increased mobility of skilled personnel between companies, and the growth in venture capital and private equity that created a pool of risk capital to fund the development of new ventures. In Japan, the accession to the World Trade Organization, the liberalization of import markets, the recent acceleration in the development of legal systems to promote cooperation among academia, industry (including the conversion of national universities and research institutes into Independent Administrative Institutions), and the government personnel and institutions, the movement toward modularized product architecture and production, and the rise of the Chinese economy have all changed the innovation landscape for Japanese firms. See Probert in this volume for a description of how the landscape has changed for the Japanese pharmaceutical industry. The result in both countries has been an erosion of the carefully created and nurtured knowledge monopolies inside leading industrial corporations. Instead of being retained within corporate walls, knowledge streamed out of centralized R&D to suppliers, customers, start-ups, and spin-offs. A new generation of companies arose, which innovated with ideas brought in from outside. Of course, they added to this knowledge base, and crafted innovative
132
The open innovation model business models around that knowledge. But they did little internal R&D on their own, relying instead on licensing, acquiring, and copying external technology. While many large companies in Japan remain successful within their industries, few have escaped the pressures of stronger foreign competition within Japan, combined with tremendous competition to establish market leadership in emerging economies like China and India. We have moved from closed innovation to a new logic of innovation: open innovation. This new logic builds upon the recognition that useful knowledge is widely distributed across society, in organizations of all sizes and purposes, including nonprofits, universities, and government entities. Rather than reinvent the wheel, the new logic employs the wheel to move forward faster. What accounts for the apparent decline in the innovation capabilities of so many leading companies? We are witnessing a ‘paradigm shift’ in how companies commercialize knowledge from ‘closed innovation’ to ‘open innovation’. Closed innovation is a view that says successful innovation requires control. Companies must generate their own ideas, and then develop them, build them, market them, distribute them, service them, finance them, and support them on their own. This paradigm counsels firms to be strongly self-reliant, because one cannot be sure of the quality, availability, and capability of others’ ideas. Increasingly, however, the closed innovation approach to innovation is no longer sustainable. A paradigm of open innovation is emerging in its place (Chesbrough 2003). The open innovation paradigm assumes that firms can and should use external ideas as well as internal ideas, and internal and external paths to market, as they look to advance their technology. Open innovation assumes that internal ideas can also be taken to market through external channels, outside the current businesses of the firm, to generate additional value. This transition to open innovation will not be easy or painless. Clair Brown’s chapter in this volume explores the human resource challenges involved in implementing a more open innovation approach. People must be recruited from new places, given different assignments, provided different reward systems and new job definitions and roles. There are technical issues as well. Admitting external sources of technology into a company’s innovation process increases the number of possible sources of innovation. This greater complexity places even greater burdens upon the ability to evaluate early stage technologies. It suggests that innovators must address a key concern: measurement error.
The problem of technical and market uncertainty: Measurement error Successful commercialization of a new technology involves managing both technical and market uncertainty. The capability and performance of a fledgling technology often are poorly understood. This technical uncertainty is
133
MOT in and between enterprises compounded by market uncertainty, when early stage technology projects also address an uncertain market. How a technology might be used by customers, and what benefits it might provide to them, are far from clear. Measurement errors (both false positives and false negatives) are inevitable. Yet companies evaluate early stage R&D projects with processes that implicitly assume that the Type II (false negative) error rate is nearly zero. This is because they employ no processes to re-examine earlier negative decisions to discontinue the technology. Evaluating the commercial potential of a new technology is less subject to measurement error when it addresses a current market with a known set of customers. Xerox had little apparent difficulty dealing with even high degrees of technical uncertainty, for example, when those projects directly addressed its copier and printer markets. The company managed to convert its entire technology base from a mechanical base in its early years, to an electromechanical base in its high growth years, to a fully electronic and digital platform in the 1990s (Chesbrough 2003: Ch. 1). Where the innovation challenge frustrated Xerox was where the company had to apply its promising technologies outside of its current markets and customers. Here, the technical uncertainty that they had to contend with was joined to a new market uncertainty: which customers and which uses of its technology would be most valuable. The personal computer industry had to be invented, in order for these PARC technologies to become valuable. Coping with market uncertainty greatly complicates the already difficult challenge of managing technical uncertainty, because resolving the technical uncertainty depends on which market the technology is intended to serve, and vice versa. One cannot anticipate the best path forward from the very beginning. Not only is this path unknown, it is unknowable. No amount of planning and research can reveal the facts, because they simply don’t exist yet. Instead, a firm must experiment, adapt, and adjust, in response to early feedback. This is a fundamentally different process from the usual process of advancing the current business, more akin to a game of poker than to a game of chess.
Playing poker: The management of false negatives A large number of false negatives have emerged over the years, where projects that looked initially very unpromising turned out later to be commercially quite valuable. When Intel first obtained its design win for the 8088 microprocessor for the IBM PC, it did not regard this as even ranking among the top 50 prospects for the chip (Moore 1996). IBM almost abandoned a software project (the XML parser) that today forms the centrepiece of its WebSphere Internet services strategy (Chesbrough 2000a). The compound UK-92480 that was under development as a treatment for hypertension within Pfizer did not
134
The open innovation model achieve sufficiently positive clinical results to warrant further development. Due to a rather unusual side effect, however, UK-92480 gave rise to one of Pfizer’s most profitable compounds today, Viagra. Similarly, Thalidomide, which was driven from the market in the 1960s due to the large number of birth defects encountered by pregnant women taking the drug, has re-entered the market successfully in the late 1990s as the preferred treatment for myeloma, a fatal form of cancer in bone marrow. In this volume, Yamaguchi shows how, one after another of the large Japanese electronic firms and NTT gave up on gallium nitride crystals as the path to creating a blue LED only to find out later that researchers at a small firm outside the mainstream had successfully commercialized the blue LED based on the gallium nitride solution. How can firms manage these false negatives? By their very nature, false negatives are projects that seem unpromising inside a company due to the lack of fit with the business model of that company. As a result, these projects receive no further support. This is as it should be. One cannot continue to support unpromising initiatives or else nothing would get out into the market. How then can one determine whether or not an unpromising project truly lacks value? In these situations, a company must develop a second process for managing innovation, a process for playing poker. The analogy comes from Jim McGroddy, the former head of IBM’s TJ Watson Research Center: When you’re targeting your technology to your current business, it’s like a chess game. You know the pieces, you know what they can and cannot do. You know what your competition is going to do, and you know what your customer needs from you in order to win the game. You can think out many moves in advance, and in fact you have to, if you’re going to win. In a new market, you have to plan your technology entirely differently. You’re not playing chess any more, now you’re playing poker. You don’t know all the information in advance. Instead, you have to decide whether to spend additional money to stay in the game to see the next card.
The metaphor of poker is well suited to situations of high technical and market uncertainty. Not all the information is yet known in these situations, yet companies often manage these situations as though they were just like situations in the main business, where they are playing chess. Xerox was actually very good at chess, at finding technologies to advance its copier and printer business. However, it was a poor poker player, unable to explore the potential options of computing technologies in new markets (Chesbrough 2002, 2003). To play poker, companies need to meter their capital carefully, and to stage their investments in projects upon the receipt of new information. Projects still have to have funding terminated. But now the company
135
MOT in and between enterprises must observe what happens after that decision. How are the researchers responding to the decision to terminate further support? Have they moved onto the next project or are they still committing time to the terminated one? If the latter is the case, have they found any external customers for the project? A second process to play poker is to expose the ‘failures’ to outsiders, to gain their perspective on the potential of these projects. (After all, once you have decided to discontinue their funding there is little at risk for you.) When IBM placed its XML Parser software on its external AlphaWorks website back in 1998, it had discontinued internal funding for the project. However, the number of people who downloaded this particular code from the website was ten times the usual number. To IBM’s credit, they took note of this high interest level and began to probe the technology more closely. They reconsidered their earlier decision, and today the XML Parser is a core element of IBM’s WebSphere Internet services initiative (Chesbrough 2000a). A third approach is to out-license the rejected project, which allows another firm to utilize the ideas and see if they are valuable. This not only provides additional funds to the licensing firm, it can allow that licensor to watch and learn from the experience of the licensee. When Intel originally invented the microprocessor, it did so under a contract from Busicom in Japan. As Intel saw what Busicom was doing, it realized that the microprocessor had great potential, and bought back the licence (Moore 1996).
Spin-offs: A Japanese mechanism to manage false negatives? Organizations can also respond to a potential false negative by creating a new venture to pursue the technology, without being constrained by the current business model. Forming an external spin-off venture allows the technology to develop further outside the originating firm than it would if kept bottled up within. Having an external venture spin-off enables new learning to occur. Moreover, if the venture becomes profitable, the equity owned by the originating firm may become valuable. This organizational strategy attempts to achieve greater decentralization, higher incentives, and greater focus while preserving coordination with the parent firm. Such endeavours have had a checkered past in the US (Burgelman and Sayles 1986; Block and MacMillan 1993; Chesbrough 2000b, 2003), but are commonly done in Japan (Odagiri 1992; Odagiri and Goto 1993). There are indications that forming new subsidiaries is becoming even more prevalent in Japan (Sako 1997). Companies like Fujitsu are themselves the end product of a series of ‘hivings off’, with the Furukawa group partnering with Siemens to form Fuji Electric, and then Fujitsu spinning off from Fuji Electric in the 1930s. (Fanuc would later spin off from Fujitsu in the 1960s).
136
The open innovation model Spin-off formation may be particularly helpful in the Japanese institutional environment, where labour markets are relatively rigid and start-up venture capital is relatively scarce. These conditions make the creation of de novo start-up firms problematic, and comparatively advantage spin-off companies that are ‘endowed’ by the parent firm with human capital and financial resources. These subsidiary firms are managed differently from the parent firm in terms of labour policies, pay packages, and sometimes even the union relationship (Sako 1997). Generally, these subsidiaries do not enjoy the status of the parent firm, and must recruit from second- and third-tier universities for their personnel. As such, observers believe that such firms provide technology that is ancillary to the core technology of the parent firm (Odagiri 1992; Okimoto and Nishi 1994). Because of the pressures on lifetime employment in Japan, observers also feel that these subsidiaries are increasingly important to the Japanese employment system, and that ‘lifetime employment’ is now promised within the enterprise group, rather than within the parent firm (Sako 1997; Kusunoki and Numagami 1998). The work of James Lincoln and his colleagues reported in this volume shows that Japanese firms often utilize an organizational process that will incubate a new technology inside the parent corporation, and then spin it off as a separate entity to commercialize the technology. There are risks to consider with this approach, however. As Ritschev and Cole (2003) report in their research on ‘internal venture capitalism’ at companies like Sony, large companies may constrain the operations of new spinoffs in ways that reduce their chances for success. One type of constraint is to limit the markets for the venture to those deemed ‘strategic’ by the corporate parent. Another different constraint is, as Ritschev and Cole (2003: 145) put it, ‘Many large Japanese manufacturing firms cannot resist the temptation of solving their problem of redundant fifty-year old engineers through new spin-offs’. These transferred personnel may lack the skills needed for success in the new venture, while the young company lacks the time required to retrain them for the new skill requirements. Spin-off formation has important benefits for innovation. The new venture’s activities reveals new information about the potential for a technology that might otherwise remain latent. When Lucent’s New Ventures Group formed Lucent Digital Video as a separate company, it judged that digital video was far from being ready for the market. Once LDV got going, though, it became clear that the market was closer–and bigger–than Lucent originally judged. Lucent found that it was selling hundreds of millions of dollars of telecommunications equipment to the Chinese market, bundled with the digital video encoders from this tiny start-up company. Lucent ended up reacquiring the rest of the venture and hastened its own entry into digital video (Chesbrough and Socolof 2000). Had Lucent not formed the spin-off, it may never have realized the market potential of this technology.
137
MOT in and between enterprises
Metrics for managing open innovation IBM also has embraced the idea of enabling others to utilize their own technology. The company reported royalties of US$1.7 billion in 2001, about 15 percent of its operating income that year. It received these royalties in payment for licensing its technology for other companies to use in their businesses. Procter & Gamble (P&G) similarly has set a policy in place that, if a patented technology had not been picked up by at least one P&G business within three years, that technology would be made available to outsiders–even competitors. P&G rightly assumes that its technology is perishable, and that keeping it on the shelf only dissipates any potential value from the technology. If P&G is not going to use it, it is better to let others do so and profit thereby. P&G is also an active participant in the marketplace for externally generated ideas. It determined that, in 2001, about 10 percent of its pipeline of new products came from external sources. It decided that in order to meet its growth objectives, the percentage of external ideas should rise to 50 percent over the next five years. If the context of industrial innovation is shifting from closed to open, and if there is latent value in managing false negatives, companies will need to alter their usual metrics for managing innovation. These metrics will help companies play poker as well as chess. This was the subject of a workshop held at the Industrial Research Institute’s Spring Meeting in May 2003. Many large Japanese firms were among the attendees, including representatives from Fujitsu, Hitachi, NEC, and Toshiba. In response to the challenges of managing innovation within an open system, and to monitor the opportunities offered by that system, a number of metrics were identified across multiple small groups within the workshop (reported in Chesbrough 2004). These metrics included: 1 What % of your sales of products and services last year came from externally licensed technologies? Is this % increasing or decreasing from 2–3 years ago? 2 What % of your net income last year came from technology licensed out to other companies? Is this % increasing or decreasing from 2–3 years ago? 3 How long does it take for patented ideas inside the company to be put into use via a company’s own products or services (i.e., taken to market via a new product or service)? Has this time interval changed in the past five years? In what direction? 4 What % of your internal ideas are offered for external licence? How much time elapsed between the patenting of ideas and their external licensing? 5 How many projects were terminated in the past year? How many were reviewed at a later date? How many subsequently were offered to external parties for further development?
138
The open innovation model 6 Of the projects tracked in #5, are any of them developing faster technically, and/or growing faster in the market than expected? Are any projects able to raise external capital for further development? Have they signed any major customers? Metrics 1 and 2 focus management attention on the outputs of the open innovation process, whether that be growth in product sales or growth in licensing activity. Participants in the workshop felt that the senior leadership within their own companies needed their R&D organization’s metrics to connect directly to corporate sales and profit measures. Metrics 3 and 4 focused on a second ‘currency’ for R&D, namely time to market for new products and services, either internally as in metric 3 or externally as is metric 4. Shortening the time required for products and services to get to market was viewed as important, as this increased the rate of learning from R&D for the company, and increased the productivity and effectiveness of R&D as well. A more subtle benefit is that the prediction horizon of the marketing organizations in these organizations was shorter than the usual time it took for the R&D cycle to run its course. Reducing the time to market for new technologies increased the chances that the innovation output was still desired by the market (and that the market hadn’t shifted in the meantime). Participants felt that metrics for managing ‘false negatives’ were at an early stage of understanding. No participants reported any internal tracking system that actively monitored the occurrences of false negatives. The typical pattern was that, once a decision was taken to terminate funding support for a given project, no further tracking of that project was done. Initial metrics to manage false negative projects in metric 5, therefore, should focus on recording their incidence and build a tracking system to follow them after the initial decision to terminate further support. Metrics like those in metric 6 should evaluate any further progress of potentially false negative projects against the expectations of the company that terminated further funding support. Most projects will likely cease at this point. When a project continues and makes further progress that significantly exceeds expectations, a re-assessment of the project’s technical and/or market potential is warranted. The ability of a project to raise external capital or to sign a major customer, should act as a strong signal that a false negative may exist. A poker playing company may reverse itself at this stage and find a way to get back into the game. For Japanese companies that are endowed with strong R&D portfolios, it is important that new ways be found to unlock the potential value in these portfolios. Spin-offs are not the only means to do this, but they may be an effective means to explore situations where new business models are needed to commercialize the technologies in new markets. If an established business
139
MOT in and between enterprises model already exists, then the company would do well to license the technology. But many new R&D programmes lack a clear path to market. These situations are where spin-offs can be most useful. At the moment, venture capital markets and entrepreneurship remain under-developed in Japan. Spinoffs may provide the most effective near-term mechanism to rejuvenate the Japanese innovation system, even as longer term initiatives in higher education and the financial and labour markets begin to bear fruit. It should be noted in passing that implementing these concepts of open innovation in Japan will also require more proactive management of intellectual property (IP). Japan is actively exploring how to become a leader in IP management, and many of its largest companies have rich portfolios of patents that may enable more open, proactive utilization of Japanese technologies in a wide variety of industrial contexts.
Open up to a point: Open–closed innovation Mr Hajime Sasaki, chairman of the NEC Corporation, had an interesting and important analysis of open innovation. He argued in an address1 to the Japanese International Intellectual Property Society in Tokyo that the term ‘open innovation’ was a little misleading. He stated that, understood properly, it should be viewed as open–closed innovation. In a forward that he has graciously contributed to the Japanese language version of Open Innovation (Sanno Daigaku Publishers, 2004), Mr Sasaki reminds us that openness is necessary to create value for customers in the innovation process, and to enable a value chain to deliver that value profitably. A certain amount of closed-ness is needed, however, to make a profit from innovation and to be able to continue to innovate in the future. According to Mr Sasaki, at NEC they regard open innovation as an ‘open–closed’ process. Intel also exemplifies the open–closed approach. Much of the internal R&D it undertakes is done to connect the company to external research in its supply chain (through its Components Research Lab) or to its customers and developers (through its Intel Architecture Labs). Intel also spends more than US$100 million annually in funding university research, seeking new ideas that it can bring into its business. Intel does not own these ideas; it does, however, gain early access to them. So Intel is open in these regards. To capture value from these ideas, however, Intel uses its internal labs. Most of Intel’s internal research is concentrated in its Microprocessor Research Lab, which focuses on new generation Pentium technologies and architectures. It is very closed about the activities in this part of its business and it seldom outlicenses any of its work in this lab to other companies. So Mr Sasaki’s point is well taken. To go further, open innovation concepts are not equally applicable to all industries. For example, the nuclear reactor
140
The open innovation model industry depends mainly on internal ideas and has low labour mobility, little venture capital, few (and weak) start-ups, and relatively little research being conducted at universities. Whether this industry will ever migrate towards open innovation is questionable. At the other extreme, some industries have been open innovators for some time. Consider Hollywood, which replaced the studio system (which in its heyday was highly vertically integrated, and rather closed) with a far more open model. Since at least the 1960s the industry has innovated through a network of partnerships and alliances between production studios, directors, talent agencies, actors, scriptwriters, independent producers, and specialized subcontractors such as the suppliers of special effects. And the mobility of this workforce is legendary: every waitress is a budding actress, every parking attendant has a screenplay he is working on. And everyone has an agent. Many industries–including those of copiers, computers, disk drives, semiconductors, telecommunications equipment, pharmaceuticals, biotechnology, and even military weapons and communications systems2–are currently undergoing a transition from closed to open. For such businesses, a number of critically important innovations have emerged from seemingly unlikely sources. Indeed, the locus of innovation in these industries has migrated past the confines of the central R&D laboratories of the largest companies and is now situated among various start-ups, universities, research consortia, and other outsiders. And the trend goes well beyond high technology. Other industries such as automotive, health care, banking, insurance, and consumer package goods have also been moving toward open innovation.
Issues for further research While open innovation suggests a greater external focus to industrial R&D, there may be many paths by which to get there. In the area of spin-offs, for example, it is important to contrast the use of voluntary spin-offs in Japan with the US pattern of spin-offs, most of which are involuntary (from the perspective of the originating firm). Involuntary spin-offs result from engineers and managers moving from one company to a competing company without the permission of the first company. This occurs frequently in the US and is almost legendary in places such as Silicon Valley. For the originating company, this flow can be quite disruptive to the continuity of internal research and development activities (Okimoto and Nishi 1994). While these involuntary spin-offs may be disruptive, they have been quite prolific in many key high tech industries in the US. One can construct a genealogy of disk drive firms from the diaspora of engineers emanating from IBM, Memorex, Control Data, and a few other early entrants into the drive
141
MOT in and between enterprises industry. A similar genealogy of semiconductor firms also could be developed from firms that emerged out of AT&T, Fairchild, and Texas Instruments.3 Involuntary spin-offs face an exciting, but Darwinian, world of high risk and high reward. When individual engineers and managers perceive new opportunities arising from innovation, they can opt to form an involuntary spin-off. If things go well, the new firm will raise capital, begin product shipments, and perhaps achieve an initial public offering (IPO) or be acquired at an attractive profit. If, however, subsequent events prove unfavourable for the venture, its financial backers will shut it down. Voluntary spin-offs, which are much more common in Japan, face different prospects. They can help a large firm focus upon a new market opportunity without creating the disruption that might ensue if that opportunity were pursued inside the firm. If the spin-off fails, there may be some possibility for employees to return to the original company. Initial capital provided by the firm reduces the financial difficulty of raising initial start-up capital. And the spin-off generates new knowledge about the market and the technology. This reduces the risks to established firms. As a result they may enter later, but they can enter with greater confidence that they will be able to protect their investments upon entry. However, creating new spin-off organizations may introduce tensions between the new entity and the parent organization. How these tensions can be managed is an important research question that remains to be answered. While Lincoln’s work is encouraging (Chapter 12, this volume), Ritschev and Cole (2003) offer a more mixed assessment. The effectiveness of voluntary spin-offs is not yet well established, and the impact of forming such spin-offs upon the performance of the parent organization after the spin-off has occurred is similarly unexplored. Clair Brown’s research in Chapter 8 of this volume also examines the costs and benefits of incorporating more open approaches within the company’s core human resource practices. With these different pathways to open innovation, there may also be different metrics required to track the progress of innovation systems. The processes for managing false positives and false negatives are poorly understood at this point. There may be an analogy to the early days of the quality movement, when Juran and Deming advanced their concepts of management responsibility for quality and statistical process control respectively to a recalcitrant US audience, only to find their ideas enthusiastically embraced in Japan. Quality used to be inspected at the very end of the process, until Juran and Deming’s concepts entered into Japanese manufacturing practice. False positives and false negatives also may be identified and managed throughout the innovation process in the future, rather than being identified at the end of the process. In sum, there are clear changes underway in the industrial innovation system, in the US, in Japan, and throughout the leading industrial economies.
142
The open innovation model There is an increasing appreciation that early stage technologies intended for nascent markets suffer from high degrees of measurement error. Both Type I errors (false positives) and Type II errors (false negatives) can arise in the evaluation of these technologies. Companies have designed their R&D evaluation systems to manage the Type I errors, but typically lack any system to manage the risk of Type II errors. Spin-offs stand as one mechanism that offers a means to manage these latter measurement errors, though we have much to learn about their performance in practice.
Notes 1. Keynote Address at the Japan International Patent Licensing Seminar, meeting of the Japan International Intellectual Property Society, Royal Park Hotel, Tokyo, 27 January 2004. 2. In 2000, the Central Intelligence Agency financed a venture capital firm, In-Q-Tel, intended to assist the intelligence agency in identifying promising technologies from start-up companies. The reason for forming this unusual organization was that the defence procurement process is so onerous that most start-up companies avoid selling to the government. Because important new technologies are emerging from start-ups in areas such as software and cryptography, to take two, the CIA decided it needed a new process to access this technology. 3. An important, but seldom recognized, dependency emerges here. Start-up firms that raid the talent of established firms are highly dependent upon the presence of successful, established firms to supply the management and technical talent they require. Start-up firms, and the venture capital that funds these firms, have no interest in paying for training for their people. Without an ample supply of qualified people to hire from these start-ups would be greatly impaired in their ability to grow.
References Block, Z. and I. Macmillan (1993). Corporate Venturing: Creating new businesses within the firm, Boston, MA: Harvard Business School Press. Burgelman, R. and L. Sayles (1986). Inside Corporate Innovation, New York: Free Press. Chesbrough, H. (2000a). alphaWorks: IBM’s technology talent agents, Harvard Business School case #9–601–001. —— (2000b). ‘Designing Corporate Ventures in the Shadow of Private Venture Capital’, California Management Review, Spring, 42 (3): 31–49. —— (2002). ‘Graceful Exits and Foregone Opportunities: Xerox’s management of its technology spinoff organizations’, Business History Review, winter, 76(4): 803–38. —— (2003). Open Innovation: The new imperative for creating and profiting from technology, Boston, MA: Harvard Business School Press. —— (2004). ‘Managing Open Innovation: Chess vs. poker’, Research–Technology Management, Jan.–Feb. 47: 13–16.
143
MOT in and between enterprises Chesbrough, H. with S. Socolof (2000). ‘Commercializing New Ventures from Bell Labs Technology: The design and experience of Lucent’s new ventures group’, Research– Technology Management, March, 43: 1–11. Kusunoki, K. and T. Numagami (1998). ‘Interfunctional Transfers of Engineers in Japan: Empirical findings and implications for cross-functional integration’, IEEE Transactions. Moore, G. (1996). ‘Some Personal Perspectives on Research in the Semiconductor Industry’, in R. Rosenbloom and S. William (eds.) Engines of Innovation: Industrial research at the end of an era, Boston, MA: Harvard Business School Press. Odagiri, H. (1992). Growth Through Competition, Competition Through Growth, Oxford: Clarendon Press. —— and A. Goto (1993). ‘The Japanese System of Innovation: Past, present and future’, in R. Nelson (ed.) National Innovation Systems: A comparative analysis, Oxford: Oxford University Press. Okimoto, D. and Y. Nishi (1994). ‘R&D Organization in Japanese and American Semiconductor Firms’, in M. Aoki and R. Dore (eds.) The Japanese Firm: Sources of Competitive Strength, Oxford: Oxford University Press. Ritschev, D. and R. Cole (2003). ‘Social and Structural Barriers to the IT Revolution in High-Tech Industries’, in J. Bachnik (ed.) Roadblocks on the Information Highway: The IT revolution in Japanese education, New York: Lexington Books. Sako, M. (1997). ‘Forces for Homogeneity and Diversity in the Japanese Industrial Relations Systems’, in M. Sako and H. Sato (eds.) Japanese Labour and Management in Transition, London: Routledge.
144
8 Managing creativity and control of knowledge workers1 Clair Brown
This chapter focuses on how high tech firms structure their human resource and knowledge systems to support and control the creative activities of their knowledge workers. The results are based on a unique survey of R&D engineers in three Japanese and two US semiconductor companies and an intensive study of product development at two leading semiconductor companies, one in the US and one in Japan. The survey of engineers explores how companies vary in the ‘external’ as opposed to ‘internal’ orientation of their HRM and knowledge systems by examining how work organization, incentive systems, and communication systems affect the creation, sharing, and control of knowledge. Following the arguments of Iansiti (1998) and Chesbrough (2003), this study explores whether HRM and knowledge systems with an external orientation facilitate R&D engineers’ access to leading-edge technology compared to systems with an internal orientation. The survey results show that overall the two US companies have a more external orientation of the HRM and knowledge systems than the three Japanese companies, yet their HRM and knowledge systems share many similarities. The product development study shows how two companies successfully manage tension between control and creativity, which are inherent in innovation activities. The management of creativity and control operates differently in Japan compared to the US, and a trade-off appears to exist between supporting team members working together and supporting individual creativity within the team in developing knowledge. Overall these results indicate that Japanese teamwork provides control and may constrain individual creativity, and that Japanese management practices for knowledge workers can benefit from supporting more individual innovation within the team setting.
145
MOT in and between enterprises
Introduction Managing knowledge workers plays an important role in creating competitive advantage in high tech companies, where innovation is critical to long run competitiveness. This chapter explores how human resource management (HRM) and knowledge systems in select semiconductor companies in Japan and the United States affect the creation, sharing, and control of knowledge. The Japanese HRM system has highly developed systems to support specific interfirm knowledge creation and sharing (i.e., the joint development of ideas or the acquisition of knowledge from other firms) and intrafirm knowledge creation and sharing programs (i.e., the joint sharing of knowledge and skills among employees within a team and across groups) (Nonaka and Takeuchi 1995). The US human resource system is better at structuring and rewarding individual as opposed to group initiative and endeavours (Sutton 2001; Malone 2004). Although US companies do not have a long history of interfirm knowledge sharing, US semiconductor companies have been undertaking joint ventures with other companies, including participation in the consortium SEMATECH, largely in response to extremely high capital and research and development costs. In addition, US semiconductor companies have increasingly turned to start-ups or emerging companies as an important source of innovation, and often acquire the company or license the technology instead of developing the technology internally (Rtischev and Cole 2003). HRM and knowledge systems that organize development activities reflect these differences in the US and Japan, and a trade-off appears to exist between supporting information sharing and supporting individual creativity in knowledge development. Precisely those HRM and knowledge systems of the Japanese firm that support team-based learning and problem solving impose constraints on individual initiative and autonomy. Precisely those structures of the US firm that support individual creativity and breakthroughs impose problems of control over the process.
HRM and knowledge systems The HRM system has three major components: 1) work organization; 2) training and skill development; and 3) pay and promotion (MacDuffie 1995; Youndt et al. 1996; Huselid et al. 1997). The knowledge system also has three major components: 1) sources of technical information; 2) communication networks; and 3) intellectual property (IP) controls (Tyre and von Hippel 1997; Teece 2000). The knowledge and HRM systems together produce the organizational structure in which engineers solve problems and create new technology. An engineer’s activities are situated within the project team, and the individual engineer’s work tasks may be undertaken independently and/or
146
Managing knowledge workers with other team members. They may draw from a variety of knowledge sources, including firm-based, publicly available, and private networks. The HRM and knowledge systems structure their employees’ problem solving activities, which are facilitated by their ability to locate and integrate knowledge. Here HRM and knowledge systems are characterized as having an external (outside the firm) or internal (within the firm) orientation. The components of the HRM system can be described as internal or external according to their orientation to firm-based rules or external markets, respectively, in determining how work is organized, skills are learned, and pay and promotions are determined. At one extreme an internal HRM system relies upon bureaucratic rules to organize work in teams, to train, and to structure compensation by seniority and performance. At the other extreme an external HRM system relies upon the external labour market to set pay for individual workers, who are in charge of their own careers, maintain a network of other professionals working in similar areas, and often work independently even within project teams. The components of the knowledge system also can be described as internal or external, and external can be divided into public and private sources. Internally oriented knowledge systems rely primarily on knowledge sources (both personal contacts and documents) inside the firm. When engineers in internal knowledge systems access external knowledge and information, their sources are public in nature (patents, journal articles, reverse engineering, conferences, tradeshows, popular press, and newsletters). In contrast, an externally oriented knowledge system relies on engineers supplementing their firm-based knowledge sources through external private networks with professionals at other companies and through the company’s private collaborative agreements (alliances). A worker’s external knowledge is also expanded through changing jobs as often as every three years, which reflects the semiconductor life cycle. The public–private division of external knowledge is more a function of the age of the knowledge rather than its characterization as tacit or explicit (Nonaka and Takeuchi 1995). Although knowledge may be codified, or embedded in tools, the most up to date and nuanced knowledge is likely to be tacit, such as knowledge that comes with experience or learning by doing. The speed associated with word of mouth interactions versus the time it takes to codify and then disseminate new technical knowledge favours external systems for technology development and problem solving in rapidly evolving industries. Even when knowledge can be codified, however, often it is not documented because it is not expected to be used by people outside the group, or because the knowledge depreciates rapidly, or because engineers do not like to document and would rather explain it to other engineers. Use of documented technology by new users often requires that they already know the previous generation, since complete codification would be cumbersome and often not
147
MOT in and between enterprises worth the time. Although private external knowledge includes both tacit and explicit knowledge, this is less important than the fact that it is state of art knowledge and is transferred among people who specialize in a technology and already have a repertoire of both tacit and explicit knowledge upon which to build. Public external knowledge would be mostly explicit in that it has already been (mostly) codified. For these reasons, private external sources contain state of the art knowledge, while public sources contain more dated material. The importance placed on internal knowledge sources vs. external knowledge sources has changed over time and across countries. In the past, engineers were often reliant on colleagues, or gate-keepers, within their own organizations (Allen 1977), and companies were likely to use a ‘local language’ that impeded the ability of their engineers to communicate effectively with outsiders (Katz and Tushman 1997). Earlier studies (Dore 1973; Whittaker 1990) focused on ‘market orientation’ vs. ‘organization orientation’ as critical features of HRM systems that shaped the organization of work incorporating technology in Britain and Japan, respectively. Today problems and solutions are not likely to be local in nature as common technology is used across products. Also the complexity of technology across the value chain has resulted in specialization at each activity, from design (e.g., QualComm) to manufacturing (e.g., TSMC) to packaging (e.g., Amkor) to assembly (e.g., Flextronics, Dell), which makes it very difficult for an integrated company relying only upon its internal knowledge base to stay competitive (Gereffi et al. 2005). Companies creating new products in an industry characterized by proliferating products with short product generations find themselves relentlessly combining new internal knowledge with external knowledge to keep pace with the industry (Iansiti 1998; Iansiti and West 1999; Eisenhardt and Galunic 2000; Chesbrough 2003). In the semiconductor industry, knowledge specific to a product generation depreciates rapidly, and the need to create the next generation requires access to the latest research and education and is supported by individual creative activities. Although a team is still needed to coordinate and integrate activities, the team is a less useful structure for on the job training and accessing private external knowledge. The individual team members must pursue learning and ideas as he/she creates new knowledge that incorporates external and internal knowledge into the innovation process. Based upon our fieldwork and past literature, we think that engineers who regularly tap into expertise both inside and outside their firms and who are supported by an HRM system that encourages external ties and rewards individual creative performance can be more creative in their innovation activities compared to engineers who work mostly in a team setting and rely upon firmbased expertise and knowledge.
148
Managing knowledge workers
Comparing HRM and knowledge systems in Japan and the United States The two countries under study are known for having different national labour market institutions (Brown et al. 1997; Sattler and Sohoni 1999; Appleyard and Brown 2001) and for occupying different market positions in the semiconductor industry (Macher et al. 1998). Japanese employees are characterized as performing the majority of their work in teams where the building of consensus is important. Japanese firms have mostly struggled to break into highmargin semiconductor markets, which require the continual development of new products, after their dominance was eroded in the low-margin DRAM market, which relied on relentless cost cutting as well as continual innovation. US engineers are characterized as more likely to work independently and to seek career advancement and technical information from their personal contacts outside their firms, which are leaders in the industry, particularly in logic devices and signal processors. To some degree, national institutions constrain the firm’s ability to cultivate external or internal knowledge and HRM systems. For example, company control over intellectual property is influenced by national labour market institutions. In Japan, where professional careers are primarily advanced within a company, and opportunities for advancement outside the company are limited, engineers and their company share the same goal of expanding and protecting knowledge within the company. Japanese companies, which focus on team performance, are concerned with the long-term development of their engineers and with long-term relationships within the company. With low turnover, they are less concerned about internalizing their engineers’ knowledge and protecting knowledge. With increasing job mobility among engineers, especially to other Asian countries, we are seeing an increase in company concerns about losing control over intellectual property. In the United States, where professional careers are often advanced through a succession of jobs at different companies, engineers rely on expanding their own knowledge for advancing their careers. In contrast with the situation in Japan, individual engineers have less incentive to expand and protect knowledge within the company. US engineers often pick jobs on the basis of the technology they will have access to as well as pay, since learning new technologies will advance their careers and affect their attractiveness to future employers. US companies that focus on individual performance are concerned with labour mobility and competing for talented engineers with their competitors. Their focus is on retaining their best workers, protecting their IP, and internalizing their engineers’ knowledge. With short product life cycles, however, the protection of knowledge often is secondary to getting leading-edge products to market. Here we use the results of a unique survey of engineers in three Japanese and two US leading semiconductor companies to describe the basic
149
MOT in and between enterprises characteristics of the HRM and knowledge systems for engineers in major semiconductor companies in the two countries.2 Although our limited sample is not representative, it provides profiles of the systems that are consistent with our extensive fieldwork at semiconductor companies in Japan and the United States. Here the respondents, who are engineers in R&D and fabrication facilities (fabs), include 35 engineers from three leading Japanese semiconductor companies and 27 engineers from two leading US semiconductor companies. All five companies are integrated semiconductor companies. The surveys were collected during the mid-1990s. They represent what might be considered best practices from the industry in the two countries. In the tables presented below, significant differences in the sample distributions between the two countries are reported using Analysis of Variance (Prob>F gives the level of significance). Differences in the sample distributions that are statistically significant will be denoted as ** for p<0.01, * for p<0.05, and þ for p<0.10.
The engineers The Japanese engineers in the sample were six years younger (born on average in 1958) and had more education (two years post BS)3 on average than the US engineers (born on average in 1952 and 0.4 years post BS). ** The US engineers were more mobile than the Japanese engineers*: the US engineers had worked for 1.6 semiconductor companies and spent 87 percent of their career at their current employer compared to the Japanese engineers, who had worked only for their current employer. The US engineers in this sample were much less mobile than semiconductor engineers more generally, since they worked for two of the older, more traditional companies. A large sample of on-going US semiconductor jobs held by workers aged 25–55 years in 2001 is instructive.4 In large growing firms, only one in four had worked at least five years, and one-half had worked less than three years. Mobility was even higher at large shrinking companies and small companies. Since engineers build their private–external knowledge largely through working at different companies, average age being equal, we would expect the orientation of the engineers at the two US companies in the survey to be more internal than of US semiconductor engineers in general.
HRM system W O R K O R G A N IZ AT IO N Almost all Japanese engineers (93 percent) reported working a majority of their time in teams, compared to 42 percent of US engineers (see Table 8.1). When we asked how the teams functioned (i.e., members worked independently,
150
Managing knowledge workers Table 8.1 Time spent working in a team and independently Country
Japan US
A Work majority of time on team**
B % team time spent working independently and/ or sequentially*
C % total time spent working independently**
0.93 0.42
0.42 0.54
0.46 0.81
Notes : Column C ¼ (1-Column A) þ (Column A)(Column B), which assumes the engineers either work 100% time (‘majority’) or 0% time (‘not a majority’) on a team.
sequentially, reciprocally, or as a team), the Japanese members were slightly less likely to work independently or sequentially than the US members. However when we combine this with information on how much of their time they spent working independently (as opposed to working in teams), the US engineers worked independently 81 percent of the time compared to about 46 percent of the time for the Japanese engineers. All US engineers reported being assigned to a team, but as one US engineer quipped, ‘My favorite team is a team of one’. S K I L L DE V E L O P M E N T The engineers spent a substantial portion, approximately one-fourth, of their time in training. However the training for US engineers was evenly split between classroom and on the job training, while the Japanese engineers received three-fourths of their training on the job. Over 85 percent of US engineers had classroom training in problem-solving methods, communication skills, and leadership skills, and over 90 percent reported using this training in their work (see Table 8.2). US engineers were even more likely Table 8.2 Training by current employer Japan
United States
Type of training OJT Technical
Nontechnical
Problem-solving methods Design of experiments Science
Paper writing Communication skills Company orientation Leadership skills % Total time spent in training (previous year)
Class
Used?
OJT
Class
Used?
0.54 0.51þ 0.34
0.49** 0.37** 0.56
0.51** 0.37** 0.37
0.52 0.30þ 0.26
0.85** 0.78** 0.46
0.93** 0.67** 0.56
0.68** 0.49 0.43 0.37 20%*
0.53þ 0.49** 0.63** 0.49** 6%**
0.71** 0.43** 0.43 0.46** –
0.20** 0.33 0.41 0.37 12%*
0.43þ 0.89** 0.93** 0.93** 12%**
0.43** 0.93** 0.63 0.93** –
Notes : **p<0.01; *p<0.05; þp<0.10. Proportion reporting that they received training by topic and where the training occurred–on the job (OJT) or in the classroom (Class), and that they regularly used the training in their work (Used?).
151
MOT in and between enterprises than their Japanese counterparts (93 percent vs. 63 percent) to receive company specific classroom training. At least one-half of Japanese engineers had on the job training in design of experiments, problem-solving methods, communication skills, and writing skills, and they were most likely to use their training in writing. Even with relatively high labour mobility, US firms trained their engineers in basic knowledge as well as job-related nontechnical skills, and provided training that would be used regularly on the job. Some of the classroom training at US firms reflects the need to provide foreign nationals with English and communication skills required on the job. In contrast, the Japanese companies were relying on the formal education of their engineers, who were more likely to have a MS than their US counterparts. PAY AND PR OMO TIO N In our sample, compensation systems were more oriented to company performance in the US than in Japan, and individual performance was rewarded in both countries. Profit sharing was reported by 82 percent of the US engineers and none of the Japanese engineers**. One-third of the US engineers** reported receiving stock options that were worth 20 percent of total pay on average, whereas none of the Japanese engineers received options, which were not legal in Japan before 1998. Although a small part of the bonus paid to workers in Japanese companies reflects company performance, the US engineers’ pay depended much more on company performance than did the Japanese engineers’ pay. Individual performance pay (reported by 46 percent of US and 43 percent of Japanese engineers) was more common than knowledgebased pay (reported by 38 percent of US and 18 percent of Japanese engineers). Because of the importance of seniority, the compensation system for the Japanese engineers was more internal than the system for the US engineers. ‘Creativity and initiative in problem solving’ was ranked the number one criterion for pay and promotion in the both US and Japan. Seniority was the number two criterion for determining pay and promotion in Japan, while seniority was not ranked as an important criterion for promotion or pay by US engineers*. Most Japanese engineers only ranked two criteria–problem solving (no.1) and seniority (no.2)–as important for pay and promotion. The US engineers ranked five criteria as important in promotion and pay decisions: ‘Suggestions and improvements made’ was an important criterion (no.2) for promotion** in the US, followed by ‘meeting production targets’*, ‘skills learned’ **, ‘communication with people outside team but within company’** and ‘team participation’**.5 Other criteria for promotion and pay were significantly different across the two samples, although these criteria were not ranked by a majority of
152
Managing knowledge workers engineers in either country. US engineers were more likely to rank ‘willingness to share knowledge with others’** as a criterion for pay and promotion and to rank ‘developing contacts with technologists in other companies’* as a criterion for pay. Japanese engineers were more likely to rank ‘presented papers at professional conferences’** and ‘published papers in professional journals’* as criteria for promotion. Although these are not major criteria for advancement, these differences highlight the focus on private external sources in the US and on public external sources in Japan.
Knowledge system Next we profile the sources of knowledge and communication networks used by the engineers. S O U R C E S O F T E C H N IC A L I N F O R M AT I O N Engineers in both countries rated their colleagues within the company as their most important source of technical information, with journals and conference presentations as the second and third most important sources, respectively (see Table 8.3). The Japanese engineers rated journals, conference presentations, and patents as significantly more important sources of technical information than the US engineers did. This indicates a greater reliance by the Japanese engineers on external public knowledge, which is consistent with the popular view that US semiconductor companies are further along the technology curve. The US engineers rated two private external sources–material suppliers and benchmarking studies–as significantly more important sources than the Japanese engineers did. However other private external sources– technologists at other companies, equipment vendors, and customers–were not ranked significantly different in the two countries.
Table 8.3 Importance of sources of technical information
Colleagues in own company Journals, books, etc.** Presentation at conferences** Patents** Technologists at other companies Equipment vendors Materials suppliersþ Customers Benchmarking studies*
Japan
US
5.9 5.6 5.1 4.6 4.1 4.0 3.5 3.2 3.0
5.9 4.5 4.3 2.0 3.6 4.4 4.5 4.0 4.1
Notes: **p<0.01; *p<0.05; þp<0.10. Based on 7-point scale from 1¼not important to 7¼very important.
153
MOT in and between enterprises INTE R N A L C O M PANY DOCUM ENTS The large majority of engineers reported that teams kept an archive of documents from previous projects (86 percent of Japanese, 67 percent of US)þ. Slightly fewer reported that the company has a document control system that stores information about previous projects (74 percent of Japanese, 63 percent of US). Only one-third of the Japanese engineers received training on how to control confidential information compared to three-fourths of the US engineers**. This reflects the differences in the need for such training, since US companies must set the rules under which their engineers can share knowledge with colleagues outside the company (including those at the employee’s next job), so that use of informal networks does not turn into a liability. Japanese companies have not been as conscientious about teaching the rules for knowledge sharing because it was not expected to occur. However protection of proprietary knowledge has become a much larger concern for Japanese companies since the late 1990s as some workers left to work for Asian rivals. E X T E R N A L A N D I N T E R N A L C O M M U N I C AT I O N N E T W O R K S Within their own fabs, engineers from both countries rated face to face meetings with individuals as the most important channel, and group meetings or seminars as the second most important channel for finding out useful technical information. US engineers were more likely than Japanese engineers to rely upon email* and electronic memos** and less likely to depend upon internal newsletters*. At another fab within the company, face to face meetings were also the most important way to find out technical information for Japanese engineers, while emailþ was the most important channel for US engineers. Telephone contact was more important than group meetings in both countries. US engineers were again more likely than Japanese engineers to rely upon electronic memos** and less likely to use company newsletters*. Channels of communication with fabs outside one’s company indicate that US engineers operate in a knowledge system more oriented toward external private channels, and Japanese engineers operate in a knowledge system more oriented toward external public channels (see Table 8.4). The top channels of information about other companies for Japanese engineers were conferences, the popular press*, and public newsletters**. In contrast, the top channels of information for US engineers were membership in a consortium**, trade journals, attending conferences, and personal telephone contacts*. Two-thirds of engineers in both countries reported attending at least one conference during the year. For those attending conferences, the Japanese engineers attended 2.4 and US engineers attended 1.7 (not significantly different). Most US and Japanese engineers (73percent and 69percent, respectively) belonged to a professional society.
154
Managing knowledge workers Table 8.4 Channels of technical information acquisition from other semiconductor companies Access
Method
Japan
United States
Attending conferences Popular press* Trade journals Public newsletters**
5.1 4.8 4.6 4.5
4.8 3.8 5.1 3.3
Face to face meetings Visiting other fabs Personal telephone contact* Personal email
3.6 3.4 3.0 3.0
4.3 3.7 4.3 3.7
Reviewing patents** Reverse engineering* Consortium Membership**
4.3 2.9 2.5
2.0 1.9 5.3
External: Public
External: Private
Company-directed
Notes : **p<0.01; *p<0.05; þp<0.10. Based on 7-point scale from 1¼not important to 7¼very important.
In learning information from equipment vendors, engineers in both countries rated face to face meetings as the most important source. This is not surprising, since equipment engineers from the vendor company are usually stationed at the fab. Otherwise US engineers relied more on private channels (telephone calls**, visit vendor’s facilities, and email**), and Japanese engineers relied more on trade shows, the popular press**, conferences, and public newsletters**. Together these information channels indicate US engineers have private access to information about leading technology outside their companies, while Japanese engineers rely on public sources for technical information. PR OBLE M-SO LVING PRO CE SS To understand how the knowledge system influences how engineers do their jobs, we surveyed the chronology of knowledge acquisition activities in solving a specific technical problem that the engineer had worked on recently. We asked the respondents to describe the problem, classify it, and walk through the problem-solving procedures in chronological order (1 ¼ the first source consulted, so a low score reflects earlier consultation). In both countries, the engineers approached a co-worker on the team, the whole team, or a co-worker on another team early in the process of problem solving (see Table 8.5). US engineers were likely to approach someone from outside the company earlier in the knowledge-gathering process**, and Japanese engineers tended to go to their supervisors earlier in the process*. Although use of team and company documents was not ranked highly by most engineers, the US engineers ranked use of these documents even lower than the Japanese engineersþ.
155
MOT in and between enterprises Table 8.5 Sources of knowledge to solve a specific technical problem Rank
Japan
US
1 2 3 4 5
Whole team (3.7)þ Co-worker on team (3.8)þ Supervisor (6.3)* Worker on another team (6.4) Person from manufacturing (6.8)
Co-worker on team (4.2) Whole team (5.7)þ Worker on another team (6.9) Person outside company (7.2)** Person from manufacturing (8.0)
Notes : **p<0.01; *p<0.05; þp<0.10. Score given is the average rank for the sample, where workers ranked each source in the order in which they used that source in solving a recent problem.
Overall the survey results indicate that the engineers at these two leading US semiconductor companies, compared to their Japanese counterparts at three leading Japanese companies, operate under HRM and knowledge systems with many similarities. However, the US engineers compared to Japanese engineers are more likely to work independently (rather than with team members), to be trained in the classroom (rather than on the job) and to use the training in their work, to receive more pay based upon company performance, to have promotion based upon individual contributions (rather than seniority), and to rely more upon private external (rather than public external) knowledge sources. A statistical analysis of the survey data (Appleyard et al. 2005) indicated that external orientation of HRM systems went with quicker problemsolving performance. Now let us take a closer look at how the development process actually works at a leading Japanese producer of semiconductors, especially memory chips, (pseudonym ‘JapanTech’) and a leading American producer of logic chips (pseudonym ‘USTech’) through data collected on several fieldwork trips to USTech and JapanTech in the mid to late 1990s.
A closer look at product development We can see the HRM and knowledge systems at work in the development process at JapanTech and USTech, and see how the management practices are consistent with and constrained by product and labour market environments. As a producer of logic devices, USTech’s goal is to control the market for their devices by maintaining a lead over potential competitors in introducing the next generation. Time to market, more than price competition, is an important part of their strategy. As a producer of memory, JapanTech’s goal is to keep up with their competitors in introducing the next generation and then to reduce costs as prices fall. Since generations are now separated by only two years, the time to market and price competition are both important parts of their strategy. For both companies, research and development
156
Managing knowledge workers activities require development of new process technology and are driven by time pressures. Consistent with the profile presented above, JapanTech’s HRM practices include lifetime employment, annual national wage determination, and an enterprise union. Consistent with the profile above, USTech’s HRM practices include decentralized and individualized wage setting in a nonunionized setting, and employees accustomed to a high degree of autonomy and input into the job assignment process. However compared to the profiles above, USTech’s workforce exhibits more worker autonomy in carrying out work assignment in the development process and less autonomy in carrying out work activities in fabrication. Any semiconductor development team has to simultaneously fulfil several management goals, including the development of a new product that satisfies customers and is designed for low cost, high quality manufacturing (Appleyard et al. 2000). Often management will use specific rules to help guide the development process, such as a ceiling on the number of new fabrication steps and new types of equipment (or conversely, the process must use certain equipment). These types of specific rules governing constraints on process steps or equipment are much easier to identify and implement (as well as negotiate) than management of the creative process itself. In comparing the development process at JapanTech and USTech, we focus on the tension between encouraging individual autonomy and creativity and controlling the direction of the development process, including the use of teamwork for encouraging, evaluating, and controlling individual ideas in the development process (Hatch and Mowery 1998; Thomke and Reinertsen 1998; Bharadwaj and Menon 2000; Hillebrand and Biemans 2004).
Work organization Job assignment reflects past performance and expected future performance at USTech, while job assignment at JapanTech reflects a project’s requirement for skills and knowledge already acquired, and the plan to develop the knowledge of junior engineers. USTech rewards development engineers for outstanding performance by assigning them more responsibility on their next project. JapanTech assigns development engineers to projects on the basis of company needs and requires more rotation among different types of tasks, including fabrication. Although engineers at both companies prefer to do the more challenging development work rather than the mundane tasks such as documenting modifications and calibrating equipment, the problem of using the creative talents of junior engineers and engaging them in critical aspects of development (rather than documenting and calibrating) seems to be more of an issue at JapanTech, since most engineers do not specialize in development or
157
MOT in and between enterprises manufacturing, and they begin work with a BS degree. At JapanTech, engineers (excluding the Central Research Labs) rotate among development and fabrication tasks. USTech’s engineers are more specialized, and their work reflects their education with development and research engineers likely to have advanced degrees. A confrontational style, where engineers criticize each other’s suggestions, offer alternative solutions, and openly challenge each other’s findings, is practised at USTech. However disagreements are put aside after a commitment to an idea is made. Autonomy and creativity are highly prized at USTech in development (but not in manufacturing). Engineers who do not fit into this competitive environment either leave or are fired. Even in development projects, a consensus approach is practised at JapanTech, and teamwork and stability are highly prized.
Incentive or compensation system Both companies use a relative performance ranking system to evaluate their engineers, but the rewards for performance are different. Pay, especially for the first dozen or so years while the engineer is in the enterprise union, is more rigidly set at JapanTech than USTech, which is performance oriented with pay based upon a required ranking of workers by supervisors. Although both companies pay bonuses, the bonus at JapanTech mainly reflects national wage setting, while the bonus at USTech reflects performance at the unit, division, and company levels. Also, USTech engineers can be richly rewarded with stock options, which account for an important part of total pay. JapanTech is struggling with the aging of its work force and the declining demand for managers relative to those eligible. In a two-tier management system, JapanTech is exploring how to provide cost-effective incentives to older professionals who are specialists and do not supervise employees. A younger and faster growing company, USTech only mentioned this as a potential problem. However, USTech’s performance oriented and flexible compensation system, coupled with willingness to fire workers, allows it to deal with changing company needs and employee demographics more easily than JapanTech’s compensation system, which is more dependent on rigid job grades and career ladders.
Skill development and communication system When research engineers with advanced degrees are hired, often straight out of the university, to work for USTech, they are assumed to have the research skills necessary to undertake their own research projects or the manufacturing skills necessary to oversee the operation of specific equipment. Both research and manufacturing engineers go to work for JapanTech after graduating with a
158
Managing knowledge workers BS degree. They are expected to learn on the job through their teamwork, continual firm-based training, and job rotation that usually includes both development and fabrication activities. Some engineers earn advanced degrees while working, either from the company or from an affiliated university. At USTech, junior development engineers are given major responsibility for developing new technologies. At JapanTech, major assignments are given to subteams within a team, and new ideas are evaluated by a test group. Junior engineers are assigned to work with senior engineers and are expected to learn through their work assignments. USTech encourages new ideas by having their development engineers highly specialized and by assigning major responsibilities for solving a specific problem to one or two engineers. USTech controls the development process by requiring design specifications. Engineers who are successful are highly rewarded both monetarily and by their next assignment; those who are not successful are likely to leave the company under pressure. USTech represents an extreme case in the US semiconductor industry in protecting its intellectual property in the 1990s. Engineers rarely made public presentations, published papers, or shared information with outsiders (including vendors), since USTech’s main goal was to protect its intellectual property rather than share knowledge. Patent applications were made only if the knowledge can be learned through reverse engineering. As USTech dramatically enlarged its product offerings since the late 1990s, its emphasis on not sharing knowledge has changed. Now USTech sponsors conferences and has partnerships with universities that enlarge public knowledge, although these programs are set up so that USTech can maintain some control on the knowledge sharing that occurs. USTech development engineers for new products now locate and assess new technology developed by other companies, especially start-ups, and USTech has acquired or licensed the technology of many outside companies. In contrast, JapanTech depends on public presentations to maintain its reputation and to announce the introduction of new devices. Engineers are expected to submit 3–4 patents annually. The publication of papers, patent applications, and conference presentations are important for keeping up with the competition and for the advancement of an engineer’s career. One important external linkage is knowledge sharing with suppliers as part of the equipment development process. Since this study, JapanTech has participated in multiple public and private semiconductor research alliances. JapanTech’s transfer of new technology occurs early in the process after only a few good dice (chips) are produced in the development fab. The group leader decides at an early stage among competing ideas, primarily based on the results from the test team. The junior engineers’ education continues within the company through working with senior engineers on projects and through formal training. JapanTech also believes that job rotations, which include fabrication as well as development assignments are an important part of the
159
MOT in and between enterprises education process. Since an engineer’s career depends on the team’s presentations at conferences and patent applications, individual creativity is less important than team outcomes, and the individual is granted less autonomy and responsibility by the group leader. Overall a system emphasizing individual autonomy, responsibility, and reward for development engineers, along with no knowledge sharing with outsiders, characterized the old USTech, which has since focused on identifying and buying external technologies in new product areas. However USTech still strictly controls its core IP. A system of teamwork, explicit career ladders, and company-based education for engineers, who do not specialize in either development or fabrication jobs, along with required internal sharing of knowledge through required patent applications and presentation of papers, characterizes JapanTech. Consistent with the external environments imposed by their product and labour markets, these approaches resulted in USTech being a top performer in logic and JapanTech being a top performer in memory in the 1990s. Since the research for this case was completed in the late 1990s, the marketplace for USTech and JapanTech has changed considerably, as the product markets for consumer electronics, telecommunications, and computing have been converging. Semiconductors for PCs and simple cell phones have become a smaller source of growth for the industry, and the rise of a variety of networked mobile and audio-visual products has made identifying winners a difficult and risky undertaking (Linden et al. 2003). Japan DRAM makers saw Korean semiconductor companies (especially Samsung) capture market share with low pricing and high capital investment in the early 1990s, during a period that Japanese semiconductor companies dramatically reduced investment. As a result of the changing marketplace, USTech and JapanTech are in the process of changing the employment systems documented above. USTech has become more open to sharing knowledge through presentations and reports, and the company is more actively engaged in the high tech community, especially with universities. USTech has begun to enter new product markets with the proliferation of semiconductor products in the post-PC World (Linden et al. 2003). Creative activities now span a larger knowledge set, and control over engineers’ activities has been relaxed somewhat. The voluntary turnover rate at USTech is lower than at other large US semiconductor companies. If the voluntary turnover rate should rise, then USTech will become concerned with how to control their engineer’s knowledge sharing with the outside world. As JapanTech has struggled to become more innovative in their product line, JapanTech has been working to increase individual incentives for creative performance and to give young engineers more autonomy in their job assignments. However this transformation is proceeding slowly through a consensus process that includes union involvement and a shift in company
160
Managing knowledge workers mindset for everyone from executives to junior engineers. Although JapanTech has made progress in supporting and rewarding individual creativity, teamwork is still its forte in pursuing innovative development. Time will tell if JapanTech is able to recapture market share, and if USTech is able to maintain a leadership position in the emerging networked product markets.
Summary: External or internal orientation profiles Our survey of engineers presents profiles of HRM and knowledge systems that are strikingly similar in Japan and the United States in their internal orientation in some key components: spending substantial time in training, using company colleagues as the primary source of technical information, and relying on co-workers and teams for solving problems. In addition, problemsolving performance was an important criterion in determining pay and promotion in both countries. However significant differences also exist. US engineers usually had worked for more than one employer, whereas Japanese engineers had worked for only one employer. In work organization, US engineers spent the majority of their time (81 percent) working independently, while their Japanese counterparts spent over half (54 percent) of their time working with others. Although US engineers displayed more mobility than the Japanese engineers, the US engineers received more classroom training. They were also more likely to use their training regularly on the job, which indicates their training was targeted to skills needed for their current job assignment. In skill acquisition, the US companies seemed to be developing their workforce rather than relying solely on external hires. Seniority was not an important criterion in pay and promotion for the US engineers, while seniority figured heavily for the Japanese engineers. Most US engineers (82 percent) received profit sharing, and onethird US engineers received stock options that accounted for 20 percent of pay. None of the Japanese engineers received stock options or profit sharing, outside of the loose relationship between company profits and the annual bonus system. For technical information, Japanese engineers were more likely than US engineers to rely upon external public sources (journals, conference presentations, and patents); US engineers were more likely to rely on external private sources (material suppliers and benchmarking studies). For technical information on other semiconductor companies, US engineers were more likely than Japanese engineers to rely on external private contacts (consortium membership and personal telephone calls), while Japanese engineers were more likely to rely on external public sources (popular press, public newsletters, and patents).6 When solving problems, US engineers approached someone outside
161
MOT in and between enterprises the company sooner in the process, and Japanese engineers went to their supervisor sooner. A leading-edge company’s strategy for innovation in development is intertwined with and reflects the relative importance in maintaining market control (i.e., remaining first to market for a particular product) vs. the relative importance of staying slightly ahead of (and never falling behind) the competition in delivering new products and being able to quickly lower the price as needed. USTech is an example of the former situation, and JapanTech is an example of the latter. However in both markets, the companies are mindful that they have a short window for selling the latest generation of a product at a high price to recoup research and development costs, before competition begins to drive down the price. Overall the survey results and case study indicate that US engineers compared to their Japanese counterparts operate in somewhat more externally oriented HRM and knowledge systems, which supports access to knowledge about leading technologies outside their companies. Engineers are more mobile in the US than Japan. Most US engineers will work for several employers over their careers, their job mobility increases their informal network with other engineers working on similar technologies as well as their knowledge of technology and products at other companies. Also US engineers’ need to continually learn new technologies in order to build their careers at their current and next employer facilitates open innovation. During tight labour markets, the mobility of engineers poses serious challenges to firms, which often cannot hire the talent they need or may see competitors poach workers who are in the middle of an important project. Control can be a serious problem for US firms. Japan’s low labour mobility and internal focus provides control but constrains knowledge about the technology at other companies and may hinder individual creativity. Japan management practices for knowledge workers can benefit from increased hiring of experienced workers and from supporting more individual activities within the team setting. With converging and fragmenting product markets, the semiconductor industry is experiencing major restructuring that has changed competitive positions and posed new challenges. Japanese companies have been adopting HRM and knowledge systems that are more external in orientation. The many paths of open innovation, from acquiring start-ups or licensing technology developed by other companies, to outsourcing to innovative companies throughout the value chain, will continue to be pursued by semiconductor companies worldwide. Innovation in product markets, especially through start-ups, and outsourcing noncore activities once provided US chip companies with a competitive advantage in product development. Open innovation practices will continue to spread across countries and companies, and companies will continue to struggle to get it right in a highly competitive industry that punishes mistakes.
162
Managing knowledge workers
Notes 1. This chapter draws upon fieldwork and discussions with my colleagues in the Sloan Semiconductor Program, especially Melissa Appleyard, Greg Linden, Linda Sattler, David Hodges, Rob Leachman, Ben Campbell, Neil Berglund, and David Mowery, and from discussions with my colleagues at Doshisha Business School, especially Yoshifumi Nakata, Eiichi Yamaguchi, and Hugh Whittaker. This work was supported by the Alfred Sloan Foundation, the Institute for Technology, Enterprise, and Competitiveness at Doshisha University and the Center for Work, Technology, and Society, IIR, UC Berkeley. For information on the Sloan Semiconductor Program at UC Berkeley, see http://esrc.berkeley.edu/csm/ and http://iir.berkeley.edu/worktech/ 2. This survey was developed, the sample was collected, and the analysis was done with Melissa Appleyard. The full sample consisted of surveys from 108 integrated circuit (IC) engineers from 1994 through 1998, and is analysed in Appleyard, Brown, and Sattler (forthcoming). 3. Graduate training of the Japanese and US engineers is not precisely comparable, in that PhD graduate work by Japanese engineers is usually undertaken with a local university while still working, whereas both MS and PhD graduate work by US engineers is undertaken at a university before joining a company. 4. Brown et al. (forthcoming). 5. Significance level shown is for pay criteria. For promotion, only team participation and communications were ranked significantly more important in the US than in Japan. 6. The finding that Japanese engineers were not likely to consult people outside of their company differs from Irwin and Klenow’s finding that Japanese memory chip producers do appear to learn from each other and from other countries (Irwin and Klenow 1994).
References Allen, T. (1977). Managing the Flow of Technology, Cambridge, MA: MIT Press. Appleyard, M. and C. Brown (2001). ‘Employment Practices and Semiconductor Manufacturing Performance’, Industrial Relations, July, 40(3): 436–71. —— N. Hatch, and D. Mowery (2000). ‘Managing the Development and Transfer of Process Technologies in the Semiconductor Manufacturing Industry’, in G. Dosi, R. Nelson, and S. Winter (eds.) The Nature and Dynamics of Organizational Capabilities, London: Oxford University Press. —— C. Brown, and L. Sattler (forthcoming). ‘An International Investigation of ProblemSolving Performance in the Semiconductor Industry’, Journal of Product Innovation Management. Bharadwaj, S. and A. Menon (2000). ‘Making Innovation Happen in Organizations: Individual creativity mechanisms, organizational creativity mechanisms or both?’ Journal of Product Innovation Management, November, 17(6): 424–34. Brown, C., Y. Nakata, M. Reich, and U. Lloyd (1997). Work and Pay in the United States and Japan, New York: Oxford University Press.
163
MOT in and between enterprises Brown, C., J. Haltiwanger, and J. Lane (forthcoming) Economic Turbulence: The impact on workers, firms, and economic growth. Chesbrough, H. (2003). Open Innovation: The new imperative for creating and profiting from technology, Boston, MA: Harvard Business School Press. Dore, R. (1973). British Factory–Japanese Factory: The origins of national diversity in industrial relations, London: Allen & Unwin. Eisenhardt, K. and D. Galunic (2000). ‘Coevolving’, Harvard Business Review, Jan–Feb. Gereffi, G., J. Humphrey, and T. Sturgeon (2005). ‘The Governance of Global Value Chains’, Review of International Political Economy. Hatch, N. and D. Mowery (1998), ‘Process Innovation and Learning By Doing in Semiconductor Manufacturing’, Management Science, November, 44(11): 1461–77. Hillebrand, B. and W. G. Biemans (2004). ‘Links between Internal and External Cooperation in Product Development: An exploratory study’, Journal of Product Innovation Management, 21(2): 110–22. Huselid, M., S. Jackson, and R. Schuler (1997). ‘Technical and Strategic Human Resource Management Effectiveness as Determinants of Firm Performance’, Academy of Management Journal, February, 40(1): 171–88. Iansiti, M. (1998). Technology Integration, Boston: Harvard Business School Press. Iansiti, M. and J. West (1999). ‘From Physics to Function: An empirical study of research and development performance in the semiconductor industry’, Journal of Product Innovation and Management, July, 16(4): 385–99. Irwin, D. and P. Klenow (1994). ‘Learning-by-Doing Spillovers in the Semiconductor Industry’, Journal of Political Economy, December, 102(6): 1200–27. Katz, R. and M. Tushman (1997). ‘A Study of the Influence of Technical Gatekeeping on Project Performance and Career Outcomes in an R&D Facility’, in R. Katz (ed.) The Human Side of Managing Technological Innovation, New York: Oxford University Press. Linden, G., C. Brown, and M. Appleyard (2003). ‘The Semiconductor Industry’s Role in the Net World Order’, in Florida and Kenny (eds.) Locating Global Advantage, Palo Alto: Stanford University Press. MacDuffie, J. (1995). ‘Human Resource Bundles and Manufacturing Performance: Organizational logic and flexible production systems in the world auto industry’, Industrial and Labor Relations Review, January, 48(2): 197–221. Macher, J., and D. Mowery (2003). ‘Managing Learning by Doing: An empirical study in semiconductor manufacturing’, Journal of Product Innovation Management, 20(5): 391–410. Macher, J., D. Mowery, and D. Hodges (1998). ‘Reversal of Fortune? The recovery of the US semiconductor industry’, California Management Review, Autumn, 41(1): 107–36. Malone, T. (2004). The Future of Work, Boston: Harvard Business School Press. Nonaka, I. and H. Takeuchi (1995). The Knowledge-Creating Company, Oxford: Oxford University Press. Rtischev, D. and R. Cole (2003). ‘Social and Structural Barriers to the IT Revolution in High-Tech Industries’, in J. Bachnik (ed.) Roadblocks on the Information Highway: The IT Revolution in Japanese Education, New York: Lexington Books. Sattler, L. and V. Sohoni (1999). ‘Participative Management: An empirical study of the semiconductor manufacturing industry’, IEEE Transactions on Engineering Management, November, 46(4): 387–98.
164
Managing knowledge workers Sutton, R. (2001). Weird Ideas that Work, New York: Free Press. Teece, D. (2000). Managing Intellectual Capital, Oxford: Oxford University Press. Thomke, S. and D. Reinertsen (1998). ‘Agile Product Development: Managing development flexibility in uncertain environments’, California Management Review, Fall, 41(1): 8–30. Tyre, M. J. and E. von Hippel (1997). ‘The Situated Nature of Adaptive Learning in Organizations’, Organization Science, January/February, 8(1): 71–83. Whittaker, D. H. (1990). Managing Innovation: A Study of British and Japanese factories, Cambridge: Cambridge University Press. Youndt, M., S. Snell, J. Dean, Jr., and D. Lepak (1996). ‘Human Resource Management, Manufacturing Strategy, and Firm Performance’, Academy of Management Journal, August, 39(4): 836–88.
165
9 Rethinking innovation1 Eiichi Yamaguchi
Japan’s technology companies suffered collective convulsions on 30 January 2004, when the Tokyo District Court ordered Nichia Corporation to pay ¥20 billion to former employee Shuji Nakamura as compensation for his patent (Nakamura 1991a) relating to the blue light emitting diode (LED). The figure represented the full amount claimed by Nakamura, and had he claimed more, it is likely he would have got it. The court estimated his contribution at ¥60.4 billion, or half the benefits the company was expected to earn before its key patents expired. It argued that Nakamura made the invention ‘with his individual power, based on utterly original thinking’ despite the fact that he was ‘working in a poor research environment at a small company’. The court was mistaken on several grounds–the extent to which the invention of the blue LED resulted from Nakamura’s heroic, individual efforts, the poverty of his research environment, the significance of support from the top management, and subsequent investment decisions, and even the significance of Nakamura’s patent in the production process.2 Irrespective of the intellectual property rights dimension, the Nichia/blue LED case also encapsulates some fundamental insights about innovation (in a theoretical and practical sense), and the structure of innovation. This chapter explores these theoretical lessons, how they relate to the decline of innovation in Japan, and offers some preliminary suggestions about what might be done to reverse the decline. It starts by re-examining Christensen’s concept of disruptive technology, categorizing this as ‘performance disruptive technology’, and introducing a further basic type–‘paradigm3 disruptive technology’–to create a two-by-two quadrant. To illustrate paradigm disruptive technology, we return to the innovation process leading to the blue LED, and the breakthroughs which were necessary. Some of these were initiated by researchers in large corporations who had their research terminated, in part because it flew in the face of accepted paradigms. Instead, it was in the tiny company from the island of Shikoku that the final breakthrough was made.
166
Rethinking innovation This, in turn, illustrates the difficulties of pursuing paradigm disruptive innovation in large firms, which were exacerbated in the 1990s when large companies restructured their R&D operations and cut back on their core researchers. Since 80 percent of Japan’s R&D expenditure takes place in private companies, and most of it in large companies, the damaging effects on Japan’s innovation system can easily be imagined. The chapter concludes with some suggestions for rebuilding the innovation system, not by attempting to turn the clock back to the ‘golden age’ of linear innovation, but cognizant at least of conditions conducive to supporting paradigm disruptive innovation.
Innovation dilemmas The old model of ‘linear’ innovation, with great central research labs located away from the hustle and bustle of the ‘real’ world, in splendid isolation, generating new technologies which then make their way through development and production and finally into markets, has become a victim of the times. As a result of IT development and the consequent unprecedented speed of feedback linking markets, technology, and science, extensive efforts have been made to develop new models of innovation which go beyond feedback loops to fundamentally change the process. In the post-linear world, it becomes all the more critical that researchers search for ways to gain exposure to customers needs, real or latent. But listening to the customer is not enough, either, as has been forcefully demonstrated by Clayton Christensen in The Innovator’s Dilemma: When new technologies cause great firms to fail (1997). In fact, Christensen shows, listening carefully to major customers makes the business vulnerable to what he terms ‘disruptive’ technology. Thus market leaders, who would normally be expected to be best placed to capitalize on new technologies because of their close links with customers, often end up becoming victims of them. One of his examples is the hard disk drive, in which successive generations of market leaders were replaced as the size of the disk was reduced. The reason the leaders failed to make the leap to the next generation was not because they were ignoring their customers (or because they were bureaucracy-bound for that matter), but because, by listening to them, they were unwilling to make timely investments in technologies which ‘result in worse product performance, at least in the near-term’. While they were unbeatable in terms of incremental or ‘sustaining’ technological innovation, they were vulnerable to disruptive technology. They were captive to what Christensen calls their ‘value network’.4 Thus, Seagate Technology, which in 1980 manufactured five inch hard drives, incorporated the value network of desktop personal computer users. It conducted customers surveys, and accordingly, placed emphasis on high
167
MOT in and between enterprises storage capacity. The survey was unable to reveal the priorities of potential future customers, who would value portability over storage. Seagate Technology did not enter the 3.5 inch hard drive market until 1988, by which time it had missed the potential of new markets for the smaller drive. Christensen’s thesis is well known, and need not be repeated at length here. Unfortunately, while Christensen was very careful to specify what does and does not constitute disruptive technology, some who have followed have been much less careful. Disruptive technology has become a kind of box into which all manner of technology and innovation-related threats and ailments have been bundled, a shorthand for ‘watch out for the unexpected’, non-incremental technology which may emerge from where you least expect it. I would like to re-open that box here, and to differentiate between two fundamentally different types of disruptive technology. The first is that described by Christensen–a technology whose performance is, initially at least, inferior to existing mainstream technology but develops because there are other features which customers, usually in different markets, value. Let us modify Christensen’s expression and call it performance disruptive technology. The second is a new technology whose performance is not inferior but superior–often strikingly so–than current technology. As such it does not present the dilemmas of performance-disruptive technology. It presents companies with other dilemmas, however, because it is based on different scientific principles than existing technologies. These principles themselves are often little understood, and incremental efforts do not lead to them. Let us call this paradigm disruptive technology. The basic difference between the two is clear. In the former, the performance of the technology itself is initially inferior because the science on which the technologies are based itself is within the current paradigm. An example is micro-processor unit (MPU), which displaced the central processing unit (CPU) of main frame computers. Managers eschew such technologies for business-related reasons. In the latter, the science on which the technology is based offers the prospect of higher–often strikingly higher–performance, but as it is not within the current paradigm. The resulting uncertainties, however, cause managers to reject it. In practice, the distinction between the two is not always easy to make. Paradigm disruptive technology may not be immediately distinguishable from a performance disruptive technology because other factors may inhibit this potential from being realized immediately. Consider the transistor, which is often cited as an example of disruptive technology (by Christensen as well as others). Strictly speaking, it is a paradigm disruptive technology, not a performance disruptive one. Its performance was initially inferior to that of vacuum tubes, not because of the underlying science, but because of technical immaturity. When the transistor was in fact designed as predicted in quantum physics, its frequency response performance was overwhelmingly higher than
168
Rethinking innovation that of the vacuum tube, which operated in accordance with classical physics.5 The initial glitches with the transistor were related to moving to a new paradigm, and not the bold and intentional lowering of performance levels. The reason for making this distinction is not pedantic. The reasons companies fail to meet the challenges of paradigm disruptive technologies are not the same as the reasons they fail in performance disruptive technologies. The former fail not because the technology in question is inferior, and customers in existing value networks do not like it, but because it is difficult for companies to design and sustain R&D activities around paradigm destructive scientific principles which are often poorly understood, which conflict with conventional textbook wisdom, and cannot be pursued with certainty. In other words, they present a different set of challenges for companies, and the intensity of these challenges has increased in recent years, partly as a result of intensified bottom line pressures on managements. We shall explore these challenges shortly. Finally, it is possible to depict both types of technology as axes–performance disruptive technology at one pole of the x-axis, with performance sustaining technology as its polar opposite, and paradigm disruptive technology at one pole of the y-axis, and paradigm sustaining technology as its polar opposite (see Figure 9.1). Clearly, it is easiest for large, successful, established companies
Paradigm disruptive
Transistor (vacuum tube) MOSFET (junction transistor) HEMT (Si-MOSFET) Blue LED (all other LED/LD)
Performance sustaining
Performance disruptive Smaller hard drive (Larger hard drive)
Thin film head/MR head (Ferrite head)
Transistor (Vacuum tube)
Paradigm sustaining
Figure 9.1 Paradigm disruptive innovation and performance disruptive innovation
169
MOT in and between enterprises to pursue innovation in the bottom left quadrant, i.e., innovation which is both performance and paradigm sustaining. According to this representation, the challenges in the bottom right are distinct from those in the top left. By definition, the top right quadrant is empty.
The blue light emitting diode Having introduced the concept of paradigm disruptive innovation, and with these preliminary comments in mind, let us return now to the blue light emitting diode (LED), the technology for which was developed in Japan in the late 1980s and early 1990s, and finally created by a group of young researchers at Nichia Corporation, led by Shuji Nakamura (currently a professor at UC Santa Barbara), Masahiro Senoh, Naruto Iwasa, and Takashi Mukai. It is important to note that this team did not discover the scientific basis which led to the innovation; this was done by others. Moreover, it was eventually made possible because of a close working relationship with the top management of Nichia Corporation, a relationship which other would-be developers lacked. In the 1980s red LEDs had already been developed, but not green and blue LEDs. If green and blue LEDs could be developed, the primary colours could be combined in different ways to create any colour desired. This would open the way for lights to be made that would drastically reduce energy consumption and have a semi-permanent life expectancy. With the potential for enormous paybacks, major corporations invested heavily in green and blue LED research programmes. It was known theoretically that green or blue light emission should be possible through the use of gallium nitride (GaN) crystals or zinc selenide (ZnSe) crystals. In practice, this proved difficult. Crystals are grown using binding blocks that resemble Lego pieces. Non-Lego blocks cannot be placed on other Lego pieces because the junctions do not match. Similarly, when constructing crystals, the junction configuration of the added layer must be the same as that of the substrate crystal (the lattice matching condition). In the case of the GaN crystal, this substrate did not exist. On the other hand, for ZnSe, there was already a known sustance, gallium arsenide (GaAs), which could be used as the substrate. For this reason ZnSe had come to play an important part in crystal growth technology by the end of the 1980s, and ZnSe had become the material of choice for blue LED research in universities and private laboratories worldwide. A small number of researchers, however, defied conventional scientific wisdom, and insisted on using GaN. Isamu Akasaki was one of these. For him, the true meaning of research was to see what has not yet been seen (discovery) and to create what has not yet been created (invention). His company, however, was not sympathetic and ordered him to abandon his research. Akasaki left
170
Rethinking innovation Matsushita Electric and went to Nagoya University to pursue his agenda. In 1985, one of his students, Hiroshi Amano, discovered that placing poorly crystallized aluminum nitride (AlN) on sapphire and then applying GaN improved the crystallization process (Amano et al. 1986). The layer that had not fully crystallized acted as a buffer. The discovery of the buffer layer method by Amano and Akasaki was ultimately to lead to the development of blue LED. Another breakthrough was also achieved by Amano and Akasaki, as a result of an accident. Despite the efforts of many researchers, p-type GaN could not be made. Some theoretical physicists argued that nitrogen defects prevented the creation of p-type GaN. If it was impossible to create a p-type diode, then obviously an LED composed of a pn junction would be impossible. In his efforts to create p-type GaN, Amano conducted many experiments involving doping the acceptors, but all ended in failure. In 1987, he was using an electron microscope to observe acceptor doping when he noticed something unusual. When the electron beam used for observation was projected, the GaN started to glow. The electron beam had activated the acceptor. Capitalizing on their luck, Amano and Akasaki continued to experiment and eventually succeeded in creating p-type GaN in 1988 (Amano et al. 1989). Encouraged by these two developments, Takashi Matsuoka and his team at NTT planned to make alloyed crystals combining GaN and indium nitride (InN). To change the colors to ultra-violet, violet, blue and green, the proportions of InN had to be altered. Current scientific knowledge held that it was impossible to mix GaN and InN. In 1989, however, Matsuoka and his colleagues defied conventional wisdom and succeeded in doing just that (Matsuoka et al. 1990). A third critical obstacle had been removed. Shuji Nakamura at Nichia Corporation, a small company in Tokushima Prefecture then with around 200 employees, also selected GaN, but for a different reason from Akasaki. After entering Nichia, Nakamura was assigned to the semiconductor manufacturing division and spent time visiting customers trying to sell the semiconductor he had made. It was a painful learning experience to see customers opting for products from larger, known companies in preference to little-known companies like his. He realized that using ZnSe, the choice of the larger companies, would not gain him customers and that he had no choice but to try paths large companies spurned. He appealed directly to the company president, Nobuo Ogawa, arguing that unless the company created something new, it would be difficult to survive. Ogawa, who had faced death as a pharmacist on Guadalcanal during World War II, returned to his hometown, and created the company from scratch, was persuaded. He allocated ¥500 million in research funds and gave Nakamura time off to study crystallography at the University of Florida for one year. This was no trivial commitment in a company of that size and created a substantially larger research budget than many researchers in large companies could hope for. During 1990, Nakamura developed a 2-flow method (Nakamura et al.
171
MOT in and between enterprises 1991) that involved the introduction of a source gas (ammonia and tri-methyl gallium) into the reactor from the side while blowing a large amount of nitrogen gas and hydrogen gas onto the sapphire base. The conditional setting parameters for this method are quite wide, so it was difficult to determine the optimal setting. However, during his first trial he was able to grow higher quality GaN crystals than had ever been seen before. When trying to build a p-type GaN, Nakamura initially tried the method discovered by Amano of irradiating a low energy electronic beam onto the surface. However, this did not produce the desired results. He was unsure how to proceed when his assistant Naruto Iwasa simply annealed the magnesiumdoped GaN in a nitrogen atmosphere in an effort to break it down. To their surprise, the GaN was easily turned into a p-type (Nakamura et al. 1992). It was much easier than anyone imagined. Nakamura, meanwhile, continued his meticulous study of indium infusion, encouraged by Matsuoka, who openly shared his knowledge with him. As a result, surprisingly quickly, the blue LED became a reality (Nakamura et al. 1993a, b). There is little doubt that the direct support of the president and the lack of bureaucratic organization at Nichia helped significantly, as well as the research carried out earlier by researchers like Akasaki, Amano, and especially Matsuoka. Matsuoka, on the other hand, had his research terminated by NTT in 1992. The management at his research center had decided that ZnSe was the answer, and that this was where both energy and funds should be concentrated. Senior management would not overturn that decision lightly. It took until 1996 for major corporations like Toshiba, NEC, Matsushita, NTT, and Sony to finally recognize that their decision to select ZnSe was wrong, by which time they had effectively missed the boat. Again, we see parallels to Cole’s Chapter 2 analysis of NTT’s tendency to cling to wrong technology bets. The three steps described leading to the development of blue LED were not the result of incremental knowledge accumulation along established scientific paths. They involved a discontinuous jump. At the same time, they did not lower the performance of existing technology. Thus, they were not performance disruptive innovations, but paradigm disruptive ones. The process can be depicted diagrammatically. In Figure 9.2, the horizontal access represents knowledge creation, or discovery (with discontinuity). The vertical axis represents knowledge realization, or accumulation. Initially, ZnSe was the choice of major companies since it maintained the lattice match premise and could be used for crystal growth. GaN, on the other hand, did not support lattice match theory, so to use it for crystal growth ran counter to an accepted paradigm in crystallography. Why, then, did some researchers persist with GaN? It can perhaps be best described as a hunch, intuition, or tacit knowledge that a hard and strong material like GaN should be used instead of the soft and easily damaged ZnSe. This tacit knowledge was
172
Rethinking innovation Lattice-matched ZnSe MOVPE LED All major corporations withdraw /1997
Blue LED
Blue LED Nakamura: Integration 1992
Nakamura GaN Buffer Layer Method 1991
Nakamura GaN 1989 Nakamura:
Nakamura & Iwasa Anneal Method 1991
2-flow method 1990
Crystallography Amano & Akasaki Amano & Akasaki AlN buffer layer method LEEBI Method 1985 1988
Akasaki: Lattice-mismatched MOVPE 1970s
Akasaki & Amano: Challenge p-GaN
Matsuoka: Realizes InGaN 1989
Matsuoka: Challenges InGaN
Figure 9.2 Innovation process for the blue light emitting diode
rooted in solid state physics. Moving vertically downwards in Figure 9.2 represents a willingness to move against the current, into this domain of tacit knowledge, to search for a new way. The wisdom of moving in this direction is very hard to sell to top managers in large companies. It runs against their demands for theory or empirical-based reasons, for milestones and probabilities which may be used to justify support of research projects, especially in group-based decision making. The critical question is whether the process of paradigm disruptive innovation is purely a matter of chance, or whether it can be managed, and if the latter, what kind of conditions are conducive to achieving it. Let us consider one more example of a paradigm disruptive innovation, early in the history of semiconductors, before we address these issues. The MOSFET (Metal-Oxide-Semiconductor Field Effect Transistor) is a transistor used in LSI (large scale integrated circuits) for computers and all other digital equipment. It is made from silicon and it operates by switching the flow of electrons on the interface between an oxide insulator film and the silicon. It was invented in 1960, but its characteristics were still unstable, and the reason for this remained elusive. By 1964, however, a research team at Fairchild Semiconductor led by Robert Noyce and Gordon Moore was able to identify the cause of the instability, thus potentially opening the way to the integrated circuit of the MOSFET (MOS-IC).
173
MOT in and between enterprises The management of Fairchild Semiconductor’s parent company, however, was opposed to the commercialization of the MOS-IC. The existing bipolar IC was commercially successful in an increasingly competitive market, and they were not ready to allocate resources to develop an unproven product like the MOS-IC. They could not be convinced of its potential. Noyce and Moore eventually left Fairchild Semiconductor, which they had founded, to form a company they named Intel. With the MOS-IC at the core of their business, they had remarkable success, while Fairchild Semiconductor lost an enormous business opportunity. According to accepted quantum mechanics theory, interface states which trap carrier electrons must exist when joining dissimilar materials which are not lattice-matched. In the case of a MOSFET device, interface states must normally exist, according to quantum mechanics. However, where an oxide film is grown on silicon, the concentration of interface states is extremely low. This discovery was serendipity, although, it seems, the researchers were spurred by a hunch, or tacit knowledge. As the management of the parent company did not share this tacit knowledge, they could not support the project, leaving Noyce and Moore no choice but to create a new vehicle for commercializing the paradigm disruptive innovation.6
Paradigm disruptive innovation and large firms Such stories are familiar, and they offer clues as why paradigm disruptive innovation is difficult to carry out in large firms. The tacit knowledge described in these stories is essential for a research team which wishes to undertake paradigm disruptive innovation. It requires what can be called a ‘field of resonance’ (kyomei ba) to create the conditions necessary for breakthroughs. It is extremely difficult to convey the tacit knowledge to senior management, but unless senior managers share and support this field of resonance, it cannot flourish. Here, the field of resonance is defined by the field (ba) (Shimizu 1995) in which the tacit knowledge itself can be transferred. Nonaka and Konno (1998) suggest that there are four types of field (ba), which correspond to the four stages of the so-called ‘SECI model’.7 The originating ba, corresponding to the stage of socialization, is a world where individuals share feelings, emotions, experiences, and mental modes. The field of resonance (kyomei ba) is similar to this type, but is more specific, relating to principles conducive to paradigm disruptive innovation. On the other hand, the Christensen’s performance disruptive innovation has nothing to do with the field of resonance. The individual who generates the field of resonance is always a researcher who confronts a dead end with the current technology. Instead of overwhelming it by an incremental improvement, he goes down to the scientific principles the current technology is based on. It must be noted that
174
Rethinking innovation science, at any time, is not firmly established but fragmented, incomplete and permeated by tacit knowledge. Therefore, there are always scientists who investigate the principles in order to clarify them and make the knowledge explicit. When others join his endeavor, sometimes for different reasons, a field of resonance is born. Each individual recognizes the differences for any other individual’s goals, works, and wills. Finally, this resonance enables the generator to discover the means of paradigm disruption. Participation of top management is ultimately essential. However, attempts at externalization and justification (in the SECI model), which often happens in the project approval and monitoring process, are likely to disturb the generation of paradigm disruptive innovation.8 If the top management had followed the SECI model without discretion, neither Intel nor Nichia would exist as the current world’s top companies in Si electronic devices and GaN photonic devices, respectively. Difficulties in supporting nonincremental innovation in large, successful firms have been noted by a number of researchers. According to life cycle models, they become bureaucratically layered. The more layers, the greater the chances of research proposals incorporating tacit knowledge being killed off, and the less chance there is for top management to share in the ‘field of resonance’. To use the framework presented in the Chesbrough chapter, there will be a significant number of false negatives. At the same time, senior managers tend to lose their entrepreneurial drive and become ‘stewards’ and ‘[t]he compulsion to innovate diminishes and the willingness to violate norms and bear disapproval falls,’ according to Porter (1990: 556). Anderson and Tushman ([1991] 1997) note the difficulties in simultaneously nurturing incremental innovation for ‘today’s’ businesses and nonincremental innovation in preparation for ‘tomorrow’s’ business. These occur at different phases of the technology cycle and require different competences. Organizations must become ‘ambidextrous’ to support both kinds of innovation at the same time.9 It becomes harder, and yet more critical, with the compression of product cycles and the rise of nonlinear models of innovation. In fact, it may be even more difficult since the conditions which support paradigm disruptive innovation may not be the same as those which support performance disruptive innovation. Both require the support of top management, but the former is particularly difficult as hunches and tacit knowledge are unconvincing for hard-pressed top managers. In Japan, the situation is particularly critical. R&D expenditure in Japan is overwhelmingly concentrated in private firms, particularly large firms. Such firms were considered very innovative up until the 1980s. Scholars and policy makers abroad cast an envious eye on the number of Japanese corporations led by engineers when they complained about the ‘short-termism’ of top managers in the UK and US during the 1980s. Crudely put, however, there are two types of engineer in Japan’s major corporations. There are those who
175
MOT in and between enterprises engage in R&D but seldom enter the top management ranks. After they hit the top specialist levels, they tend to move into universities (see Fujimoto 2005: 183). The others are those who work their way up through key operating divisions to top executive posts. In these operating divisions there is greater reliance on explicit, paradigm-sustaining knowledge than in R&D labs.10 Such engineers can be as unsympathetic to tacit knowledge and sites of resonance as accountants. In fact, humanities and social science graduates can be more sympathetic, as they at least know they don’t know the science behind the R&D efforts, and hence may be more willing to give researchers the benefit of the doubt for longer period. This distinction became apparent in the 1990s. It may be that large Japanese companies were still too chaotic until the 1980s to force researchers to try to make explicit prematurely what should have remained tacit. Or conditions were more benign, and hence greater slack was allowed in R&D labs. The slack was taken up in the 1990s, however, with devastating consequences. Ironically, conditions for paradigm disruptive innovation were more benign in smaller companies led by founders, as in Nichia Corporation, or in universities, both of which had fewer resources to devote to it. As large companies restructured their R&D labs, researchers were transferred to other divisions, such as manufacturing or marketing, or were in some cases sent to subsidiaries on ‘loan’ or transfer. Others were lost in a wave of voluntary redundancies at the turn of the century. The number of published papers published by researchers in major organizations was broadly proportional to the number of researchers with PhDs in 1999 (Figure 9.3). As shown in Figure 9.4, the number of academic papers published by researchers in large corporations fell continuously after 1994, suggesting a deterioration of conditions favouring basic research, or knowledge creation. Thus, it can be reasonably estimated that the number of researchers with PhDs also dropped continuously from 1994. Figure 9.5 shows data on the ratio of academic papers published in 2003 relative to 1994 for specific Japanese and non-Japanese companies. Japan’s electronics giants uniformly appear in the shaded area, where the ratio is less than 1. The figure also gives the 2003 to 1994 ratio of total market value of the shares issued by those corporations. There is a strong correlation between the two ratios (correlation coefficient of 0.711). If we accept changes in numbers of academic papers published as a proxy for changes in knowledge creation, a number of explanations might be advanced for the correlation: 1 As a result of strong R&D, successful products were created leading to a rise in the company’s value. Weak R&D led to the opposite effect. 2 There was an increase in motivation in research departments at the companies that hired more researchers and technicians and encouraged the writing of research papers. This motivation spread to other departments
176
Rethinking innovation 1000 NTT
Number of published papers (1999)
1999
Hitachi
AIST NIMS
Fujitsu Mitsubishi Elec. Matsushita
Sony
CRIEPI
Nippon Steel 100
Toyota Mitsubishi Heavy Indus Sumitomo Chemical
Toray
Fuji Film Canon
10 10
100
1000
Number of researchers with PhDs (1999) Figure 9.3 The number of published papers vs. the number of researchers with PhDs in 1999
and increased productivity and product development. Conversely, there was a decline in motivation in the research departments at companies that significantly reduced their researchers and technicians, and declining motivation, which then spread, resulting in a fall in corporate value as productivity and product development vitality suffered. 3 At companies that increased their value due to higher profits, greater resources were allocated to R&D, leading to an increase in the number of researchers and academic papers published. In contrast, in companies that lost value due to an inability to generate profits, cost considerations led to cutbacks in research departments and thus a reduction in the number of academic papers published. Of these, the most likely explanation is a combination of 2 and 3. Widespread reductions in researchers took place across the board, depriving these companies of a critical source of creativity. Motivation in research labs dropped. Either as a result, or in tandem, motivation also fell in manufacturing divisions, resulting in a decline in the overall vitality of the whole industry group.
177
MOT in and between enterprises 120 NTT+J8 Total number of papers (right axis) 3
other J8 NEC
2
90
Hitachi 1
Total number of papers / 104
(Number of papers / Total number of papers) / 10−3
4
NTT
0 1980
1985
1990
1995
60 2000
Year published Figure 9.4 Trends in academic papers by year Notes : The number of papers is normalized by the total number of papers on the respective data bases. J8 ¼ Canon, Fujitsu, Hitachi, Matsushita, Mitsubishi Denki, NEC, Sony and Toshiba. Source : SciSearch, Social SciSearch, respective years
Of course, this did not happen in every large Japanese company, as Figure 9.5 shows, but it was widespread. Recently, these same companies have attempted to restore their R&D capabilities, but they will undoubtedly find that it takes much more time and energy to rebuild R&D dynamism than it took to destroy it.
Towards a new innovation system Let us summarize the characteristics of paradigm disruptive innovation. First, this type of innovation cannot be found on a line extrapolated from known technology A. Second, it is not until researchers ‘burrow down’ to basic scientific principles S that a new paradigm P is discovered. Third, once a new paradigm is discovered, expertise for creating new technology A* becomes explicit knowledge. The process A–S–P–A* is not linear. In addition to the researchers who undertake this process, moreover, it requires a key person on the business
178
Rethinking innovation
Texas Instruments Microsoft
IBM 10
Intel
HP Nissan Samsung
Sony Apple
Honda
Sharp Matsushita 1
Kyocera
Mitsubishi Chemical Mitsubishi Electric
NTT Toray Motorola Mitsubishi Heavy Industries Hitachi
Lucent Fujitsu 0.1 0.1
‘
Toyota
Toshiba
NEC
Japan US/Korea
Decline
Ratio of total market value of shares 2003/ 1994
Canon
1 10 Ratio of Academic papers published 2003/1994
Figure 9.5 Academic papers and company value Note : Lucent denominator is papers for 1996. Source : SciSearch, Social SciSearch, respective years
side who ‘co-owns’ the process to succeed. In the case of Nichia this was Nobuo Ogawa, the president. Noyce and Moore, on the other hand, could not transmit their tacit knowledge to the senior management of Fairchild Semiconductor, and hence were unable to secure the support of such a key person, leaving them no choice but to spin off. As described here, tacit knowledge is transmitted through a ‘field of resonance’. The critical question is how to manage this. It requires firms to develop new kinds of competences. In general, large firms excel at sustaining innovations, but they also work as potential incubators for paradigm disruptive technologies. This requires a willingness to accept diversity and experimentation by researchers with a desire to create what has not yet been created, able to descend to scientific principles. It also requires mechanisms to value and transfer concepts not easily quantified by normal indicators.
179
MOT in and between enterprises Let us reflect briefly on why NTT ordered Matsuoka and his team to abandon their research, and further, why major corporations like Toshiba, NEC, Matsushita and Sony also lagged behind Nichia in the race to create the blue LED. It appears that the managers in these companies did not create a field of resonance with their researchers through which they could co-own the tacit knowledge their researchers had. Nakamura and his team, on the other hand, were able to succeed due to a close working relationship with the management of the company. Although Nakamura and his team did not discover the scientific basis that gave rise to this innovation, Nichia worked as the field of resonance among researchers and management teams, and rapidly commercialized the paradigm disruptive technologies discovered by Isamu Akasaki and Hiroshi Amano of Nagoya University and Takashi Matsuoka of NTT. This offers another lesson. Managers in large companies would do well to recognize that in the emerging system of innovation in Japan universities and small businesses also play an important role. Many large companies have been actively engaged in creating spin-offs or even new ventures, but the former are usually the result of cost-cutting measures, and the latter, by retaining links to the ‘parent’ company generate potential conflicts of interest, since disruptive innovations have the potential for undermining established markets through which large companies obtain their profits. Fields of resonance can be created across corporate boundaries. In fact, in the emerging system of innovation in Japan there are opportunities for small businesses not only to commercialize paradigm disruptive (and performance disruptive) innovations, but to coordinate the various types of expertise and resources which can enable these innovations to happen. Large company managers may miss valuable opportunities if they continue to cling to corporate-centric views of innovation, without looking for potential fields of resonance. In a more flexible industrial order, large firms are not necessarily the originators of paradigm disruptive innovation, nor even the coordinators of it. They may be providers of capital, however, or customers of technology, or undertake contract R&D. The forces which prevent managers from adopting a more flexible mindset to create diverse win-win situations must be addressed. The biggest of these is social–political forces for centralization, which leave little autonomy for regions, or for individuals. Conversely, the quest to recognize and establish new fields of resonance could unleash long frustrated creative forces, ushering in a new era of innovation. The seeds of paradigm disruptive innovation have always been generated at universities as well as corporate research institutes. Unfortunately, under the current industrial model, companies have killed most of them. An ideal model for the 21st century industry is one in which the society and large companies encourage the generators of fields of resonance to start companies as vehicles for it, and provide coordinators to create new markets by networking across current industrial categories with large companies. New architecture for the
180
Rethinking innovation 21st century must be designed so as to maximize the conditions for generating fields of resonance.
Notes 1. I am most grateful to Mr Keigo Shiowaki (Dai-ichi Mutual Life Insurance) who collected the data on total stock issues. I would also like to thank Dr Greg Linden (UC Berkeley) for background information on why Robert Noyce left Fairchild. The research reported here is work in progress, and forms part of ITEC’s activity as a Center of Excellence selected by the Japanese Ministry of Education, Culture, Sports, Science and Technology (MEXT). 2. See Yamaguchi 2004 for details. On 11 January 2005 the suit was settled amicably at the Tokyo High Court, where the judge recognized problems in the earlier Tokyo District Court ruling. The final amount of compensation was ¥600 million not only for the patent of the lawsuit, but all the contributions by Nakamura. The patent portion can be estimated at ¥10 million according to the formulation suggested by the High Court, which is 1/6000 the amount ordered by the Tokyo District Court. See Byosiere, Chapter 10 of this volume for a further perspective on this case, and the treatment of corporate researchers. 3. ‘Paradigm’ is used in the original meaning given by Thomas Kuhn (1962). 4. A value network is ‘the context within which a firm identifies and responds to customers’ needs, solves problems, procures input, reacts to competitors, and strives for profit’ (Christensen 1997: 32). 5. According to William Shockley: ‘One was demonstrated on 20 April 1950 according to my marginal note. This nonphotogenic device did perform according to theory but had a wide base and poor frequency result and provoked little interest’ (Shockley 1984: 1542). The problem was that because of the thick p-layer, it was not reflecting the quantum physics. 6. According to Ross Knox Bassett: ‘When Fairchild bypassed Noyce for the chief executive position, he quit. Gordon Moore, the head of Fairchild R&D, left with Noyce, out of a growing frustration over the difficulties in transferring products from R&D to manufacturing and a belief that any new head of Fairchild would likely undertake a major reorganization’ (Bassett 2002: 172). Resolving the instability of MOSFETs by 1965, Noyce and Moore were scientifically convinced that MOS-ICs would take over bipolar ICs. However, the manufacturing department and even the management did not want to challenge such new products. Furthermore, it was unlikely, in their view, that a new head would be competent in anticipating the long-term future. 7. ‘SECI’ stands for socialization (sharing of tacit knowledge with others based on shared experiences), externalization (conversion of tacit to explicit–codified–knowledge in the form of concepts and models which are then subject to a screening process in which they may be justified), combination (systematizing concepts with knowledge systems through combining different forms of explicit knowledge), and internalization (embodying explicit knowledge back into tacit knowledge, creating shared knowledge which in turn becomes the basis for a new spiral of knowledge creation): Nonaka and Takeuchi 1995: ch. 3.
181
MOT in and between enterprises 8. As a matter of fact, both the reason why the density of interface states for Si-SiO2 systems are so low and the reason why ZnSe produces such deep levels in the bandgap are still unknown in modern physics. 9. This can be called ‘squeezed states’ from the analogy of quantum optics. Here, the original concept of squeezed states is minimum uncertainty states situated in between two opposite states which never coexist due to the principle of uncertainty. 10. Empirical evidence for this argument is being gathered.
References Amano, H., N. Sawaki, I. Akasaki, and Y. Toyoda (1986). ‘Metalorganic vapor phase epitaxial growth of a high quality GaN film using an AlN buffer layer’, Applied Physics Letters, 48: 353–5. —— , M. Kito, K. Hiramatsu, and I. Akasaki (1989). ‘P-type conduction in Mg-doped GaN treated with low-energy beam irradiation (LEEBI)’, Japan Journal of Applied Physics, 28: L2112–14. Anderson, P. and M. Tushman ([1991] 1997). ‘Managing Through Cycles of Technological Change’, Research/Technology Management, May–June: 26–31. Reprinted in M. Tushman and P. Anderson (eds.) Managing Strategic Innovation and Change, Oxford: Oxford University Press. Bassett, R. (2002). To The Digital Age, Baltimore, MD: Johns Hopkins University Press. Christensen, C. (1997). The Innovator’s Dilemma: When new technologies cause great firms to fall, Cambridge, MA: Harvard Business School Press. Fujimoto, M. (2005). Senmonshoku no tenshoku kozo (The Job-changing Structure of Professionals/Specialists), Tokyo: Yushindo. Kuhn, T. (1962). The Structure of Scientific Revolutions, Chicago: University of Chicago Press. Matsuoka, T., H. Tanaka, T. Sasaki, and A. Katsui (1990). ‘Wide-gap semiconductor (In,GaN)’, Inst. Phys. Conf. Ser. 106: 141–3 (Proceedings of Int. Symp. GaAs and Related Compounds, Karuizawa). Nakamura, S. (1991a). ‘Handotai kessho no seicho hoho’ (Methods to Grow Semiconductor Crystals)’, Japan Patent No. 2628404. —— (1991b), ‘GaN growth using GaN buffer layer’, Japan Journal of Applied Physics, 30: L1705–7. —— , Y. Harada, and M. Seno (1991). ‘Novel metalorganic chemical vapor deposition system for GaN growth’, Applied Physics Letters, 58: 2021–3. —— , T. Mukai, M. Senoh, and N. Iwasa (1992). ‘Thermal annealing effects on p-type Mgdoped GaN films’, Japan Journal of Applied Physics, 31: L139–41. —— , M. Senoh, and T. Mukai (1993a). ‘P-GaN/n-InGaN/n-GaN double heterostructure blue-light-emitting diodes’, Japan Journal of Applied Physics, 32: L8-10. —— (1993b) ‘High-power InGaN/GaN double-heterostructure violet light emitting diodes’, Appl. Phys. Lett. 62(19): 2390–2. Nonaka, I. and H. Takeuchi (1995). The Knowledge-Creating Company, New York: Oxford University Press. —— and N. Konno (1998). ‘The Concept of ‘‘Ba’’: Building a foundation for knowledge creation’, California Management Review, 40(3): 40–54.
182
Rethinking innovation Porter, M. (1990). The Competitive Advantage of Nations, Basingstoke: Macmillan. Shimizu, H. (1995). ‘Ba-principle: New logic for the real-time emergence of information,’ Holonics, 5(1): 67–79. Shockley, W. (1984). ‘The Path to the Conception of the Junction Transistor’, IEEE Transactions on Electron Devices, ED-31 11: 1523–46. Yamaguchi, E. (2004). ‘The Dimness of the Blue Diode Lawsuit’, Japan Echo, 31(3): 27–31.
183
10 ‘Microbursts’ of knowledge and creative work in Japan1 Philippe Byosiere
Fundamentally, Japanese culture is based on rice farming. Rice cultivation requires a lot of water, and water must be shared evenly by everyone. Planting rice also requires teams of people walking from row to row, at the same speed. And all of this has meant that uniqueness had to be suppressed.
These are the words of Dr Hideki Shirakawa, Nobel Laureate in Chemistry (New York Times, 7 August 2001), reaching for a time-honoured explanation for the widely perceived aversion to individual initiative in Japan. While appealing in its simplicity, the statement conceals more than it reveals. Japan is one of the extraordinary cases of rapid economic development and the world’s second largest economy. Should we believe that it arrived at this position without the exercise of significant individual initiative? This is not a credible position. First, looking at the matter historically, following crisis periods such as after World War II, old leaders were pushed aside and new entrepreneurial leaders took the stage. There were the individuals that formed companies like Honda (Soichiro Honda) and Sony (Masaru Ibuka and Akio Morita), and those who reinvigorated established companies. More recently, individual creativity is revealed in researchers like those who carried out the steps necessary to create the blue light-emitting diode (LED), most famously Shuji Nakamura, but not confined to him. This is described by Eiichi Yamaguchi in this volume, who emphasizes the importance of (individual) top managers who are capable of communicating with individual researchers or researcher teams and come to ‘co-own’ the tacit knowledge regarding a paradigm disruptive technology, or alternatively displaying such faith in an individual researcher as to support the project without fully understanding its implications. Either way, that individuals play a key role in the creative process in Japan as well cannot be denied. From a different perspective, and from a bastion of Japanese management, Toyota’s vice president and head of technology development, Kazuo
184
Microbursts of knowledge Okamoto, said recently, referring to the company’s famed Chief Engineer system: ‘Automobiles are not made by everyone pitching in together. Unless there is someone with a passion, who says ‘‘This is the car I want to make,’’ it’s not going to succeed’ (Okamoto 2005: 20). The Chief Engineer system had its origins in the development of the Crown, and has been institutionalized as a key feature of product development at Toyota ever since, including its Lexus brand models. Second, Nonaka and Takeuchi (1995), building on examples from Japanese corporations, describe the innovation process as one in which individually acquired knowledge (tacit knowledge) combines with group activities to create a successful innovation process, at least in terms of incremental advances. The remarkable popularity of this book in Western countries suggests that the process they describe is universal and one in which individual insight, intuition, initiative and creativity have a significant role to play when effectively united with group processes. In short, the innovation process typically involves an interplay between individual creativity and team (group) processes. All this is not to say there is no merit in Dr Shirakawa’s observations, far from it. There are particular challenges related to recognizing and unleashing individual creativity and initiative in large corporations in general and in Japanese corporations in particular, as we shall see. Our objective is to explore some of these in this chapter. In the case of the US, these difficulties are compensated, in large part, by a strong start-up culture that is able to release individual drive and to bring to fruition innovative capabilities. Japan, however, is characterized by a weak start-up culture, an economy in which existing interlocking ties leaves little room for newcomers (Ritschev and Cole 2003, although this has not always been the case; see Whittaker 1997). Those who have tried to start their own business, moreover, have often been cast as misfits rather than heroes. The weakness in fostering new ventures in Japan makes it all the more imperative that new research and product initiatives can thrive in large Japanese corporations. My task in this chapter, then, is to delve more deeply into innovation and creativity in large Japanese firms, to identify some of the key challenges, and how they might be addressed. We first consider the issue of recognizing individual creativity through a metaphor from meteorology–the phenomenon of ‘microbursts’.
Ted Fujita and microbursts Tetsuya (Ted) Fujita was born in northern Kyushu in 1920 and was educated in mechanical engineering. He became interested in thunderstorms, but was unable to find sufficient support for his work in Japan and moved to the University of Chicago in 1953 to work with Prof Horace Byers in
185
MOT in and between enterprises the meteorology department. His lifetime of research involving detailed observations of thunderstorms and tornadoes unlocked many mysteries, advanced theoretical understanding in meteorology, and saved many lives through application in airport and aircraft safety, earning him the affectionate accolade of ‘Mr Tornado’.2 One of Fujita’s discoveries was microbursts–small, sudden, severe, downdrafts from thunderstorms that can result in 150 mph winds on or near the ground. The ferocity of the downdraft creates a damaging outward burst of wind on the surface. The discovery was made after observing the pattern of trees uprooted during thunderstorms or tornadoes, and he blamed the Eastern Airlines Flight 66 crash at New York’s Kennedy Airport in 1975 on this phenomenon. The claim was controversial for years, as it was empirically very difficult to verify microbursts, but eventually it was accepted by the majority of meteorologists, leading to the installation of Doppler radar at all airports worldwide. I cite Ted Fujita and microbursts in this chapter for three reasons. First, I believe the microburst is a valuable analogy for considering creativity in organizations. Microbursts are extremely hard to recognize; existing means of observation failed to capture their existence for many years; a number of aircraft accidents since attributed to microbursts were attributed to other causes. Yet the impact of microbursts is significant, and sometimes devastating, not just in terms of the downthrust, but also the lateral surface movement. Individual knowledge creation or creativity is also hard to recognize; it plays a significant yet often ignored role in the thunderstorms of innovation. The impact of this creativity can be extensive and is not necessarily unpredictable. Rather than seeing microbursts of knowledge creation or creativity as hazardous, however, I see them as of great potential benefit if they can be harnessed. An organization which refuses to recognize them, or actively suppresses the conditions which create them, loses as a result. (The analogy is not perfect, of course–microbursts of knowledge are seen as creative, or leading to creative destruction, and their potential benefits may be lost or they are not discovered.) Second, microbursts were discovered by someone who refused to accept common sense and current scientific observation as infallible. He was the ultimate microburst spotter. If organizations could train spotters, or become more skilled in spotting creativity not directly contributing to their current trajectories, in predictable patterns, they would be the better off for it. Third and relatedly, Fujita was clearly an extremely creative Japanese individual himself (colleagues described him as an ‘observational genius’), but one of many who failed to gain sufficient support or recognition in Japan and who ultimately left Japan for an environment where those talents could better flourish (usually the US). Many stay behind, their talents unrecognized and/or unrewarded. A stunning example was brought to light with the
186
Microbursts of knowledge Nobel Prize award to corporate researcher Koichi Tanaka in 2002 for the development of methods for identification and structure analyses of biological macromolecules: ‘(T)he 43-year-old was unknown in domestic academic or government research circles before the award. Even in his own technology company, Shimadzu, the 43-year-old’s talents had failed to lift him beyond the second lowest rank of a promotion ladder, which was–as is still usual in Japan– determined by age.’3 At the time of his discovery in 1987, his company awarded him ¥11,000. After the Nobel Prize award, he was promoted, had a company lab named after him, and was deluged with recognition and offers. This argument shares elements in common with the argument of Eiichi Yamaguchi in the previous chapter, that senior managers in large companies undermine the potential for paradigm disruptive innovation by creative researchers because they make judgments based upon accepted scientific understanding or common sense. A more general form of his argument is that they support innovation which conforms to both accepted scientific principles and the types of innovation the company is used to doing–a path dependence argument. The distinctive contribution I would like to make, however, is that the seeds for significant innovation may already exist, but go unrecognized because they have not been spotted, or are suppressed. This may be because the seeds have been created by someone who is not supposed to create them–a new engineer in sales rather than an established researcher in R&D, for instance–or for a variety of other reasons. Continuing (at the risk of using another metaphor, albeit a familiar one), even if they are recognized as seeds, they may well be discarded because people are too busy to plant and tend for them, or because people do not know the conditions under which they might grow. This leads into the point made by Chesbrough (Chapter 7 in this volume), that companies must become better at dealing with ‘false negatives’–ideas or technologies that are normally killed because the company cannot see a worthwhile product and/or market for them, but which may, under different conditions, flourish beyond all expectation. ‘Getting better’ means identifying their potential earlier in the evolving work processes. I noted earlier that there are particular challenges related to recognizing and unleashing individual initiative in large corporations in general and in Japanese corporations in particular. Recognizing individual creativity is a particular challenge in the latter because of their ‘egalitarian’, group emphasizing community orientation, because of strong expectations of behaviour appropriate for posts, jobs, or seniority, and because of the cultural inclinations referred to by Dr Shirakawa. That does not mean that creativity is absent, however. Relatedly, unleashing individual creativity and initiative is a serious problem as well, and there are reasons to believe the problem has become worse for large Japanese corporations. We now turn our attention to this issue, focusing on R&D workers.
187
MOT in and between enterprises
Knowledge society and knowledge workers Peter Drucker (1968; also Fritz Machlup 1962) talked about the emergent knowledge society and the increasingly important knowledge workers who drive the new knowledge society forward. Drucker’s thesis, which has been largely accepted in subsequent years, is that the productivity of knowledge has become the key to overall productivity, competitive strength and economic development. The application of knowledge in organizations has become the foundation of the modern economy. This entails individual and organizational learning, but it is the individuals creating and utilizing the knowledge that constitutes the basis of the knowledge society (see MEXT 2003: 4). Drucker also suggests, however, that there is an unresolved conflict between individual knowledge workers and their position as employees (Drucker 1973: 276–7). Knowledge workers are paid with salary and pension benefits for applying their knowledge, exercising their judgment and creativity. They have a boss, however, and they also depend on the organization for their income and the opportunity to exercise their skills. The organization, in reciprocal fashion, gives knowledge workers discretion and depends on them to identify promising technologies that have good prospects to become commercial products. Organizational managers, however, make decisions as to which research projects should be pursued, funded and, if funded, at what level, which projects should be renewed or discontinued and finally which ones which be commercially developed. Managers also determine levels of compensation. Herein lies a potential for considerable conflict between knowledge workers and the organization. One aspect of that conflict was captured by Drucker in a well known jingle composed by a Unilever executive: Along this tree From root to crown Ideas flow up And vetoes down
The application of knowledge to create new products and services in large organizations is nowhere more exemplified than among employees in the R&D departments of these companies. It is they who are formally tasked with this objective (though they are in fact not the only avenue, nor is R&D the only venue, for creating new products and services). The challenge of making big breakthroughs, however, most often eludes R&D departments in large organizations. This is nowhere more true than in Japan where in recent years, as noted in the introductory chapter, many have lamented that the high ranking of Japan in R&D investments, patents, and number of researchers has not been matched by a steady flow of new products and services. The kind of data reported in Figure 10.1 have been widely discussed in Japan from about
188
Microbursts of knowledge the year 2000 with the intention of showing that there has been a decline in R&D efficiency. Why is this so? Or to borrow the words of the title of a special section of the MOT magazine, BizTech (2005, vol. 8), referring to similar data, ‘What’s killing engineers?’ To answer these questions we have to look at changes in Japanese companies, and changes on the part of researchers.
Changing companies in Japan Japanese ingenuity has been celebrated in the popular and scholarly management literature since the mid to late 1950s (Abegglen 1958; Drucker 1973; Ouchi 1981; Pascale and Athos 1981). Consider Drucker’s pronouncement in 1973 (p. 799): Change is seen as an opportunity by the Japanese because they are guaranteed their jobs and are not afraid of putting themselves or their colleagues out of work by proposing something new. . . But fear and ignorance are also overcome in Japan by making continuing change the opportunity for personal achievement, for recognition, for satisfaction. The man who, in a Japanese training session, comes up with a new idea receives no monetary reward, even if his idea is a big and profitable one.
One might conclude from Drucker’s analysis that Japanese management had successfully found a solution which had long eluded Western managers– that is, how to manage and motivate workers to optimize their performance. These optimistic observations, though stretched, weren’t entirely wrong for the time period in which they were made, but there are reasons to believe that both Japanese companies and their employees have changed significantly since then. Until the early 1970s, the Japanese economy had been growing at double digit rates, and even into the 1980s was growing at over 5 percent per annum. New products and services were being introduced at a rapid clip. New markets were being entered and new businesses were being established either through existing divisions, new divisions, or spin-offs. Between 1990 and 2003, however, the growth rate slowed dramatically to a 1.1 percent annual increase, and capital investment fell significantly. From a peak in 1992, per company R&D expenditures pretty much flattened out and even fell so that by 2002, the nominal per company expenditures were only about up to their 1992 levels as shown in Figure 10.1. There has also been an increase in the outsourcing of R&D activities. Some 23 percent of firms surveyed by the Ministry of Education in 2001 reported that they had adopted outsourcing of some portion of their research and development operations as part of their R&D strategy (MEXT 2002:16). With growth rates low throughout most of the 1990s, per company R&D expenditures flat or down, and a modest increase in R&D outsourcing, one can easily see that opportunities for new and exciting projects, ones that could
189
MOT in and between enterprises Per company R&D expenditures used in the company (scale on the right) Per company operating profit (scale on the right) R&D efficiency (scale on the left)
600
Amount (¥100 million)
R&D Efficiency (%)
500 400 300 200 Decline 100
00 20
95 19
90 19
85 19
19
80
0
Figure 10.1 Decline in R&D efficiency in manufacturing industry Note : R&D efficiency is defined as the ratio of operating profit (in past 5 years) to R&D expenditure input (in previous 5 years). The assumption is that the lead time of R&D before commercialization is 5 years. Normalized values used. Source: Ministry of Public Management, Home Affairs, Posts and Telecommunications, Statistics Bureau, ‘Report on the Survey of Research and Development’
capture the imagination of researchers, were also down. Correspondingly, the opportunities for promotion for R&D employees would have also declined in this environment. There is another more intangible factor. In the postwar period leading up to the 1990s, the Japanese were the attackers. They were the risk takers challenging established Western firms with bold new investments across a range of industries, from DRAMs to machine tools to consumer electronics. They had a ‘bias toward growth maximization’ as outlined in the Introduction. They were quick to commercialize new technologies like the transistor. As a result of their success, however, they increasingly became the incumbents, with achievements and accumulated wealth to protect. New aggressive competitors willing to take risks challenged them. Despite having similar organizational structures (large vertically integrated firms with broad product lines) Korean firms seemed to act so much more decisively. Moreover, nimble specialist firms in the US were able to move so much more quickly. Large Japanese firms had begun to suffer from middle age, or ‘large firm malaise’ (Inagami and Whittaker 2005). On top of this–or as a result–they restructured in the late 1990s. This had an impact on perceptions of employment. In the past, what individuals achieved at the firm was seen as good for the company and the individual identified with the company community. Individual employees were part of that community and derived benefits from participating in that community, including long
190
Microbursts of knowledge term employment security even in medium size firms. Moreover, for those in large corporations, being a member of the company community gave one status in the outside community. A reciprocity existed between individual contributions to the company and the benefits employees derived from being members of the company community. Restructuring in the late 90s however, included layoffs, incentives and pressures on employees to take early retirement, and relocating old and new jobs to Asia. Unemployment rose to unprecedented levels in the postwar period. Performance wages were introduced in many large companies, but they proved difficult to implement without creating a sense of unfairness and uncertainty among many employees. These developments instilled fear in many employees, who sensed that they could no longer count on the company for their employment security and a fair salary. Identification with the company declined. With the reciprocity in mutual benefits greatly weakened, not surprisingly, employees increasingly want to receive direct and immediate benefits for their innovative contributions, but this was often not forthcoming. In this environment, unleashing the creativity and initiative of individual workers arguably became more problematic. Let us look at this assertion in more detail.
Changing Japanese employees In a survey I carried out of senior and middle R&D managers of a large manufacturing company in 1996–before restructuring–respondents were asked about their satisfaction on a range of choices.4 Most relevant for this discussion were two groups of choices, relating to personal or self-satisfaction (self-development through work, suitability for current work, sense of achievement, and opportunities for success through work), and system satisfaction (career building; confidence your superiors have in you; evaluation system; and opportunities for promotion). Both senior and middle R&D managers rated satisfaction with the former group higher than the latter group, although among the former group, opportunities for success in work was rated lowest. In the latter group, satisfaction with the evaluation system and opportunities for promotion were rated lowest, and the rankings of middle managers were lower than those of senior managers. It is likely that the scores would have been lower had the survey been carried out more recently. The 2002 White Paper on Science and Technology produced by the Ministry of Education, Culture, Sports, Science and Technology (MEXT 2003: 1–13) reports on a survey questionnaire of the degree of satisfaction with company policies among those currently engaged in the research profession. Most
191
MOT in and between enterprises striking in Figure 10.2 is the high level of researcher dissatisfaction with special rewards for research results (59.7 percent) and amount of research expenses (53.6 percent). Moreover, over 30 percent are dissatisfied with their salaries, promotion opportunities, and the mode of vesting patent rights (presumably giving those rights to the firm rather than sharing with the researcher). In a related Ministry survey, 44.5 percent of researchers interviewed wanted improvement in their ‘percentage of ownership of patent rights’ (MEXT 2002: 33). In an independent survey of creative work conducted by Inagami and others in 1996, white collar employees were asked what they received and what they wanted for high performance. There was some overlap, in the form of increased bonus and assignment to more important work, but while the companies tended to give awards or publicity, the employees wanted promotion and an increase in base pay. The same survey found an atmosphere of risk aversion was listed by department managers of creative departments (57 percent) and employees of those departments (50 percent) as the largest impediment of creative work, followed by (for employees) lack of delegation of authority and a range of issues around pressures of the pace and volume of work (Inagami and Whittaker 2005: 60–5). Many of these factors are related to employees’ sense of achievement. R&D employees need opportunities to achieve to fully engage their motivation and capabilities. The 1990s, however, were a period of retrenchment in R&D expenditures that involved reduction of R&D personnel and restructuring of R&D organizations. This was clearly accompanied by an intensification of work.
Salaries (including bonuses) Special reward for research results Amount of research expenses Promotion Vesting of patent rights, etc. Freedom of activities outside their institution Freedom of activities within their institution 0
10
20 Dissatisfied
30
40
Cannot say which
50
60
70
Satisfied
Figure 10.2 Degree of satisfaction with treatment by researchers Notes : Satisfied ¼ ‘Satisfied’ þ ‘Relatively satisfied’; Dissatisfied ¼ ‘Dissatisfied’ þ ‘Relatively dissatisfied’. Source : MEXT, Survey of the State of Japan’s Research Activities (FY2002)
192
Microbursts of knowledge It is hard to see Drucker’s motivated employees in this picture. Two cautions are in order. First, we should not confuse attitudes with behaviour. For example, Japanese employees have long reported strong desires to change jobs without this data showing up in quit rates. Second, we don’t have time series data on these questions. It is quite possible that there was considerable dissatisfaction on these matters in the 70s and 80s as well. That said, the level of dissatisfaction, especially with special rewards for research results and amount of research expenses, is quite high and seemingly inconsistent with the view that Japanese managers have solved the problem of how to manage knowledge workers. These findings are of concern according to the Ministry analysis of the data because, if the results of research activities are not reflected in the treatment of researchers, then the motivation to display creativity and the attraction of the research profession (a set of occupations already starting to experience shortages because of demographic changes) will be greatly reduced. It is easy to dismiss grumblings about pay or dissatisfaction with current circumstances as common to any class of workers. The distinctive condition in Japanese organizations, however, is that the wages of knowledge workers– those occupations requiring specialized professional skills and expertise–by and large, are not particularly high as compared with the average wage. This reflects the egalitarian wage structure (very modest wage spread among different categories of workers) promoted by most Japanese large firms. Even though companies have been moving more toward individually based performance wages, the egalitarian dimension has not been fundamentally changed (Inagami and Whittaker 2005: 65). It is likely that researchers are less sympathetic to the reasons for this egalitarianism than in the past. Shinichi Tanaka, a communications consultant, describes how advertisers in the past believed it sufficient to reach their audience by targeting age and gender demographics and organizational attachments like large firm employees, and self-employed business presidents. Today, however, advertisers place much more emphasis on individual affiliations in recognition of the greater diversity of individual values that is now present. The objective has become to identify the specific values held by people using value segmentation surveys. The source of individual values is seen as residing in individual life styles. Where people feel they belong is changing. With weakened links to work organizations, people develop stronger attachments to individual goals like happiness and sense of achievement. People are trying to find where they belong, now that work organization is less able to perform that role. Increasingly people seek to find happiness in things ‘close to them’.5 Dissatisfaction has increasingly found expression in litigation. The case of Shuji Nakamura, in which he took the novel step (for Japanese researchers) in 2001 of suing his employer Nichia first for control over patent rights (he lost that case) and second for a share of the rewards associated with the successful
193
MOT in and between enterprises use of the patent, sent shock waves through the Japanese corporate establishment. It became a cause ce´le`bre in the Japanese media (see, for example, Yamaguchi 2004) and undoubtedly played a role in heightening researcher consciousness of shortcomings (for them) in the existing reward system. It is a system in which Japanese companies typically list their names above the inventor’s in patent applications, whereas in Western companies typically firms file patents on behalf of their employees and pay them a percentage of the profits emanating from the patent. Nakamura, as had been typical in the reward practices of many large Japanese companies, received roughly US$100 for filing his patent and another US$100 for its being accepted (Hara 2001. It is common for companies to pay researchers again if the patent actually contributes to the company’s business, but the payments have been small.) In contrast, Nichia became a US$ billion company. The Nakamura case highlights another point. Japanese researchers increasingly work in a global environment; foreign assignments to companies and universities are not unusual. As such, these researchers come into contact with Western researchers and become aware of the rewards their contemporaries are able to reap from their research successes. Nakamura has commented how Western researcher associates nicknamed him ‘slave Nakamura’ to capture the failure of Japanese firms to share substantial rewards with successful researchers (Hara 2001). The Nakamura case has highlighted the differences in treatments accorded Western and Japanese researchers and it is only to be expected that this would lead increasingly well-educated, sophisticated Japanese researchers to become more dissatisfied and be reflected in the Ministry survey results discussed above. Since the Nakamura case, some ten other researchers have initiated suits of a similar nature demanding greater rewards for their research achievements. As a result of the unfolding of Nakamura’s continuing litigation, the Ministry of Economy, Trade and Industry (METI) moved quickly to submit a statutory amendment (the National Diet) in early 2004 which was approved by May, perhaps a speed record for IP legislation approval, showing the seriousness with which the issue was regarded. The amended law now reads that a contract or an employment regulation governing inventor compensation should be respected unless it is unreasonable, but there is relatively little guidance on what constitutes unreasonable. The manner in which a firm sets up its compensation arrangements and the nature of its discussions with employees about those arrangements, however, are expected to be important considerations in any future litigation. The Nakamura case has already accelerated a movement among large firms to set up clear compensation arrangements for extraordinary research results achieved by individuals. Parallel developments occurred on the judicial side in which the Japanese courts appeared to cap the compensation granted inventors at 5 percent of the
194
Microbursts of knowledge employer’s profit attributed to the invention (an amount well below what many Western researchers receive). The 5 percent base was the outcome not only in the Nakamura case but two other recent cases as well (Tessensohn and Yamamoto 2005). The courts have not explained how they arrived at the 5 percent figure, but one may presume that they believe it is sufficient to serve as an incentive for corporate researchers but not large enough to hurt company competitiveness. These developments are hardly likely to settle the matter, and indeed may have the reverse effect. There is little doubt among knowledgeable observers that Nakamura’s legal case, by raising fundamental issues of the research employee’s relationship to the firm, has had a strong impact on Japanese industry and society. The seeming hostility of Japanese industry toward researcher expectations for better compensation for their contributions would seem to have unwittingly fuelled further researcher dissatisfaction (Tessensohn and Yamamoto 2005: 4).
Recognizing and unleashing individual creativity Japanese firms are revising their strategies regarding research and development. Many have upgraded and formalized their reward system for extraordinary results achieved by researchers. In a survey of private sector organizations, MEXT (2002: 16) reported the various measures which over 40 percent of the responding companies had undertaken. They included: organizational structure reform within R&D divisions, scrap and build of R&D topics (scrapping existing research projects and selecting what are seen as new more promising ones), research and development efforts in new fields, and joint research with domestic universities and public research organizations. Whether measures like these will foster a significant turnaround in recognizing and unleashing individual creativity and initiative, and ultimately make a bottom line difference in terms of R&D contribution to corporate growth, remains uncertain. As noted, opportunities for the application of knowledge in large organizations typically occur in teams, not only in R&D labs, but throughout the innovation process leading to new products or services delivered to customers. We have already cited Nonaka’s and Takenchi’s (1995) observation that the interaction between the individual’s creativity and his/her tacit knowledge and the team is critical. There are still unanswered questions, however. Clair Brown (Chapter 8 of this volume) documents, in an admittedly small sample, striking differences in the extent to which Japanese and American semiconductor product development engineers participate in individual vs. team activity. Ninety-three percent of the Japanese engineers reported spending a majority of their time in teams versus 42 percent of the US engineers. Brown
195
MOT in and between enterprises believes that Japanese practices, in the context of a more inwardly focused orientation, constrain individual creativity. It may not be a matter of time spent in team activities per se, but an emphasis on getting along in teams, in not upsetting the apple cart, that causes this. To obtain the optimum performance of individuals and control the development process clearly involves employee participation in teams, but what is the optimum interplay between individuals and the team that fosters creativity? In fact, we know very little about these dynamics. Moreover, given embedded social and cultural norms, it is possible that what is optimum in the US or UK might not be optimum in Japan. It is worth noting, however, that in the Inagami et al. survey on creative work, the majority of respondents themselves said they gained their creative insights alone rather than with others, away from work and indeed outside the company, even if they gained stimulation from colleagues (Inagami and Whittaker 2005: 65). This should give pause for thought for those seeking to unleash the creativity of their R&D workers. Relatedly, research on stress has shown that one source is role load, which has a quantitative as well as a qualitative dimension (Kahn and Byosiere 1992). The above survey points to increasing stresses of the quantitative type. Attempts by companies to create breakthrough innovation mechanisms can, however, intensify stress of the qualitative type as well. It is not simply the result of increasing stress levels through work intensification under restructuring (cf. Morishima 2003). Stress also rises through multiple affiliations and tasks in project-based work, which paradoxically is intended to release creativity. In fact, measures to promote creative working can be a double-edged sword. The issue of recognition requires new skill sets, so that the significance of individual creativity can be recognized, and directed in the right direction. To some extent, this is being addressed by an increase in MOT and related business education (see Chapter 15, this volume, by Kaneko, Nakata and Yokoyama). At the very least, senior R&D managers must become more accomplished ‘spotters’ of researchers’ microbursts of knowledge, a requirement of which is greater attunement to external market opportunities. At a more fundamental level, though, it requires a different way of looking at employees–‘people as people’.6 Finally, however, we should also recognize that many of the key decisions on new projects lie outside the R&D group. To be successful, firms will need to find ways to take more calculated risks and to do so in a more decisive fashion. (As they say in sports, ‘the best defence is a good offence’.) If they can do this and find new ways of recognizing and unleashing creativity–which is by no means limited to specific industries like anime (animation)–they will be able to face the future with renewed confidence.
196
Microbursts of knowledge
Notes 1. I would like to thank Denise Luethge and the editors of this book for their comments and suggestions, from which I greatly benefited, and participants in the executive programmes at Doshisha Business School for sharing their insights. 2. See http://www.usatoday.com/weather/tornado/fujita/wfujobit.htm and http:// stormtrack.org/library/people/fujita.htm 3. http://www.guardian.co.uk/japan/story/0,7369,824277,00.html 4. The company had around 25,000 employees. The survey was a self-administered paper and pencil questionnaire, with 986 usable responses from 1030 managers. In R&D, there were responses from 40 senior and 142 middle managers. 5. Interview with Shinichi Tanaka, communication consultant, Tokyo, 28 July 2005. 6. This reference comes from the title of a book by Hideo Fujita (1998), chair of the Organization Reform Research Group, a training organization founded by a former Sony director in 1971: http://www.sokaku.co.jp
References Abegglen, J. (1958). The Japanese Factory, Glencoe, IL: The Free Press. Drucker, P. (1968). The Age of Discontinuity, New York: Harper and Row. —— (1973). Management: Tasks Responsibilities, Practices, New York: Harper and Row. Fujita, H. (1998). Hito o hito toshite: shiji machi ningen o masamesaseru mono (People as People: What makes people who just wait for orders wake up), Tokyo: PHP. Hara, Y. (2001). ‘Researcher Sues Nichia Over Blue LED Patent Rights,’ EE Times, August, 24: 1–3. Inagami, T. and D. H. Whittaker (2004). The New Community Firm, Cambridge: Cambridge University Press. Kahn, R. and P. Byosiere (1992). ‘Stress in Organizations’, in M. D. Dunnette (ed.) Handbook of Industrial and Organizational Psychology, Chicago: Rand McNally. Machlup, F. (1962). Production and Distribution of Knowledge in the United States, Princeton: Princeton University Press. MEXT (2002). Annual Report on the Promotion of Science and Technology, Tokyo: Ministry of Education, Culture, Sports, Science and Technology. —— (2003). Annual Report on the Promotion of Science and Technology, Tokyo: Ministry of Education, Culture, Sports, Science and Technology. Moore, M. (1995). Inside the Tornado, New York: HarperCollins. Morishima, M. (2003). ‘Changes in White-Collar Employment from the Employee’s Perspective’, Japan Labour Bulletin, 42(9): 8–14. Nonaka, I. and H. Takeuchi (1995). The Knowledge-Creating Company, New York: Oxford University Press. Okamoto, K. (2005). ‘Naze Toyota wa kaihatsu taisei o ‘‘kaizen’’ shitsuzukerunoka’ (Why Does Toyota Keep Doing Kaizen to its Development System?), PM Magazine, 1: 20–3. Ouchi, W. (1981). Theory Z, New York: Addison Wesley. Pascale, R. and A. Athos (1981). The Art of Japanese Management, New York: Simon and Schuster.
197
MOT in and between enterprises Ritschev, D. and R. E. Cole (2003). ‘Social and Structural Barriers to the IT revolution in High-Tech Industries,’ in J. Bachnik (ed.) Roadblocks on the Information Highway, Landham, MD: Lexington Books. Schein, E. (1992). Organizational Culture and Leadership, 2nd edn, New York: Jossey Bass. Tessensohn, J. and S. Yamamoto (2005). ‘Inventor Compensation: An early spring for Japanese industry’, World Intellectual Property, March, 19(3): 4. Whittaker, D. H. (1997). Small Firms in the Japanese Economy, Cambridge: Cambridge University Press. Yamaguchi, E. (2004). ‘200 Oku en hanketsu: akamura Shuji wa eiyu ka’ (A 20 billion verdict: Is Shuji Nakamura a Hero?) Bungei Shunji, March, 82(6): 162–9.
198
11 Hitachi’s nascent ‘new production(ist)’ system D. H. Whittaker
This chapter examines the response of one of Japan’s technology champions to a fundamentally changed competitive environment from the 1990s. The environment featured the challenges we noted in the Introduction–loss of competitive advantage through emulation, and new business models associated with modularization and open innovation. Hitachi is Japan’s pre-eminent general electric company, with businesses ranging from power generation equipment to plasma display panels and micro-chips. Its enterprise group spans over a thousand companies, some of them household names in their own right. Profitability at Hitachi declined in the 1990s, culminating in a massive loss in 1998, its first in the postwar period. Japan’s leading economic newspaper commented: Japan’s economic woes have now reached the major electric machine companies which support the nation’s very economic foundations. The picture of this giant battleship Hitachi, losing its way, unable to take effective measures before this massive loss materialized, is the very picture of Japan today. (Nikkei shimbun, 4 September 1998)
In the aftermath of this loss, a series of reforms was embarked upon. Initially, these addressed ‘large firm malaise’ through governance, organization, and human resource management (HRM) measures. Subsequently, however, innovations in technology management, or the production system dimension, began to figure, and to become integrated with the earlier reforms. Three steps are identified–exploratory, strategic, and systemic. One of the goals of this chapter is to assess whether these amount to an emerging new production(ist) system.1 Of course no single firm is representative of Japanese industry: Hitachi’s sheer size and the range of technologies and industries it encompasses make it distinctive for a start. Its very size, however, lends its reforms a
199
MOT in and between enterprises significance far beyond the company’s boundaries. And its organization and HRM changes will be familiar to observers of other large Japanese technology companies. The chapter is organized as follows. The first section describes the crisis and initial series of organization and HRM reforms. The second section looks at explorations in linking technology and new business models through internal ventures. The third section examines the emergence of a technology strategy which becomes central to the business reforms, while the fourth section charts efforts to further link and operationalize this strategy in a systemic way. The final section summarizes features of the emerging system, and speculates on its strengths and weaknesses.
Crisis and reform Background Hitachi traces its founding to 1910, when Namihei Odaira succeeded in making three five horsepower electric motors. Odaira’s dream was that Japan should develop its own technological capabilities, and not have to rely on foreign imports. This dream gave birth to today’s giant, Hitachi. It became a giant through a combination of: . daring entrepreneurship in a growth economy; . internal bottom–up and top–down mechanisms which fostered ‘hard’ and ‘soft’ innovation; . employment relations which fostered skill development and commitment (see Dore’s ([1973] 1990) ‘welfare corporatism’ or the community firm model); and . complementary external supports, including patient capital.2 Many of the features of the production, organization, and skill formation/ HRM systems emerged in the postwar years, and were systematized in 1964– 68. A large number of ‘excellent company’ books were written about Hitachi during the late 1970s and early 1980s. From the second half of the 1980s, however, Hitachi increasingly came to be seen as giant, and bureaucratic. It became increasingly difficult to compete effectively in the wide range of businesses the company had expanded into. This was particularly true in the fast moving electronics and ICT industries, whose needs top management, still typically drawn from heavy electric operations, did not always respond to. Despite its technological prowess and technological firsts, in product markets the company often adopted a follower pattern. Profitability began to slip. Various reforms were attempted, without decisive success. In the mid-1990s divisions were re-organized into a small number of units in an attempt to
200
Hitachi’s ‘new production(ist)’ system decentralize decision making, but it was not clear just what authority the heads of these units had. Spinning out consumer electronic operations in the 1990s cut costs (which were generally lower in subsidiaries), but was often accompanied by negative morale effects. Countermeasures for declining competitiveness in semiconductors were largely ineffective. Inside the company, as well as outside, many came to doubt that real reform was possible. The 1998 loss, however, created a sense of crisis and a consensus that fundamental reforms were necessary. It unleashed reformist energies which had hitherto been kept in check in favour of a more gradualist approach to change. Measures which were planned for the company’s 90th birthday in 2000 were brought forward, and deepened.
Organization reforms In mid-1998 Hitachi announced that it would adopt a holding company structure with business operations. The previous business units would be replaced by 11 internal groups, which would be given substantial independence as virtual companies with their own ‘CEOs’. One of these would be spun out immediately, and others might follow. The new ‘small headquarters’ would restrict its activities to strategy, coordination and monitoring. The number of board members was halved, from 30 to 14, as was the senior executive committee, which was replaced by a management council with 11 members to facilitate substantive debate and faster decision making. An advisory board was set up to bring in outside opinions.3 Critics and sceptics, however, also wanted to see evidence of a focus (‘concentrate and select’) strategy. The medium-term ‘i.e. Hitachi Plan’, announced in 1999, envisaged a transition ‘from a manufacturer to a solutions-providing company’, and a focus on information and electronics (‘i.e.’) through a combination of mergers and acquisitions, internal investment, and spin-offs and divestitures.4 New consolidated accounting requirements in 1999 raised the importance of Group performance, above and beyond that of the core company. A new Hitachi Group Council was established as a ‘virtual board of directors for the whole Group’ to discuss and propose Group strategy.5 This would consider measures to cut costs, which were considered too high, as well as new business creation through synergies among Group companies, which had remained latent because of substantive subsidiary independence. Mergers among subsidiaries and core company units to achieve this followed.6 Inside Hitachi, the organization reforms prompted questioning of business processes and job design. What jobs should corporate staff be doing? What should be done by operations funded ‘business staff,’ and what should be outsourced? This led to a business process reform initiative in indirect departments. Similarly, decentralization was associated with a flattening of reporting
201
MOT in and between enterprises hierarchies, through the removal of ‘deputy’ posts. This was expected to speed up decision making processes.
HRM reforms HRM reforms actually started in 1998, before the organization reforms, but the two were meant to complement each other. In the run-up to the HRM reforms, debates focused on total labour costs, and stimulating white collar productivity. While there were cost-cutting measures, emphasis was placed on the latter, and new value creation by white collar ‘professionals.’ The 1998 measures targeted premanagement ‘planning’ track employees, and featured a new qualification system, clarifying criteria for promotion through 15 job types, and de-emphasizing seniority. This was supported by a new wage system, which placed emphasis on demonstrated ability. The third component was a new work system, which extended flex-time, and sought to promote flexibility in terms of time and place of work, and hence to encourage individual creativity and responsibility. The message of self-responsibility was repeated elsewhere–in a new ‘cafeteria’ welfare and benefit scheme, for instance, as well as in education and training. Along with de-emphasis of long service in both the pension system and long service awards, the company was sending a calculated message that it did not necessarily value loyalty per se, but individual initiative which contributed to company objectives. This was reiterated in the second round of HRM reforms, in 2000, which targeted managers, and introduced HITACHI VALUE, a set of qualities linked to implementation of the i.e. Hitachi Plan which the company expected managers to exhibit. VALUE criteria, as well as MBO (management by objectives) evaluations, would decide their pay, promotion, and job assignments. High performers would gain accelerated promotion, demotion was a possibility. The first two rounds of HRM reforms were in turn systematized and applied to all employees in 2004. Combined, they marked the first major overhaul of the HRM system since the mid-1960s, and attempted to create a certain ‘tension’ in the employment relationship. They did not, however, mark the end of the premise of long-term employment for the majority of employees. In this sense, they were evolutionary rather than revolutionary.7 Additionally, however, more mid-career people were being hired, to promote expansion in financial services, for instance. This and other measures–a new set of family friendly measures, for instance–increased the scope for workforce diversity, modifying the previous emphasis of integration through homogeneity.8 Inagami and Whittaker (2005) characterize the reforms as an attempt to create a ‘new community firm,’ without negating the community itself. They attempted to address Hitachi’s ‘large firm malaise’ as well as new challenges the company was facing from product, financial, and labour markets, and a
202
Hitachi’s ‘new production(ist)’ system rapidly changing legal environment. They were complemented by further reforms featuring technology management concepts.
Technology and new business models9 Institutional reforms by themselves will not make a giant dance, or turn a ‘supertanker’ into a ‘speedboat’.10 There need to be success stories, championed by individuals. Aware of this, the company introduced early executive selection by creating four tiers of high fliers in 1999, half of whom were engineers. The top two tiers were under the direct authority of the new president, and the top tier in particular was given assignments which would previously have been considered radical–outside their specialist areas, and designed to break down internal boundaries. An even more select group of eleven ‘corporate senior staff’ was appointed in 2001, charged with restoring profitability either through rationalizing existing businesses, or creating new ones. The actual direction was left to them.
Hitachi mu solutions One of those selected was Ryo Imura, then head of the Device Research Centre at the Central Research Lab. Imura took up the challenge of creating a new business around the recently created ‘mu-chip’, the world’s smallest RFID (radio frequency identification device).11 Conventionally, this device would have been assigned to an existing operating division, manufactured, and sold as an RFID to interested customers. The problem with this model was that as the price came down to expand the market, and as competitors emerged, margins would become razor thin. This was not the recipe for enhanced profits the company was seeking to build with its technology resources. Table 11.1 Corporate senior staff (2001–03) (Stage1)
1 2 3 4 5 6 7 8 9 10 11
Name
Theme
Current work (April 2003)
I. Yamazaki K. Urayama H. Mizuta H. Mitsumaki H. Kawamoto A. Kaji N. Tomita T. Yamashita M. Sano T. Fukushima R. Imura
Hitachi Group manufacturing Manufacturing Sales Personal healthcare Management Information Asset management, outsourcing Next generation net business Information Parts manufacture in China Mu chip
Hitachi Home & Life Solutions Renesas Technologies Info & Telecom Systems Public systems PHC Venture Coy. Automotive Systems group Retired Hitachi Building Systems Info & Telecom Systems ID Solution Renesas Technologies Hitachi Industrial Systems Div. Info & Telecom Systems, Mu Solutions Div.
Source : Y. Mizuno 2004: 18
203
MOT in and between enterprises Imura decided to commercialize the technology with a new business model, and through a new vehicle–an internal venture, made possible by the launch of a CVC (corporate venture capital) fund by the company in 2000. The business model envisaged a combination of mu-chips, back-end systems, and solutions, sourced from both inside and outside the company. Thin margins on the chip itself could be offset by profits on the systems and solutions. Easy licensing terms would encourage diffusion of the technology, another feature of the business model. Imura secured independence over all aspects of commercializing the product. To avoid common problems of not securing the best–or most appropriate– human resources, he negotiated to assemble his own team. He got his way on this, but the incentive structure for the team resembled that of the parent company more than a Silicon Valley start-up. The significance of the venture is summed up by Isaacs: By creating new autonomous companies . . . perhaps a new method could be established for commercialization of Hitachi technology that would shorten the time-to-market of Hitachi products. Perhaps young ‘intrapreneurs’ at Hitachi would see the effort at business re-engineering taking place at the company, and seek to join these new, riskier quasi-external ventures. It was hoped that a new business model applied to important innovations might help build morale internally, improve Hitachi’s conservative image externally, and bring in new business, all at the same time. (2002: 13)
The mu-chip and associated technology was used in the 2005 Expo at Aichi, for admissions management, reservations, and other services. The package is also being used in museums, factories, warehouses, farms, and even schools (to monitor student arrivals and departures in China), bringing it a long way towards meeting its revenue target (US$150 million in 2005). This, in turn, will enhance its demonstration effect within the company, and underline the importance of combining technology with innovative business models to generate profits.12
Personal healthcare venture company Another corporate senior staff appointee was Hiro Mitsumaki, who began studying merger and acquisition and management buy-out possibilities, but eventually decided these would not change Hitachi. The only way to change the company, or return it to its roots, he believed, was to create an internal venture: Join the company and work your way up to general manager, and eventually factory GM, protect your organization and hand it on to the next generation . . . This is Hitachi’s tradition, a good one, but if all employees are steeped in this culture, we can’t win in fierce competition. If only 10% of employees change their thinking . . . (quoted in Mizuno 2004: 25)
204
Hitachi’s ‘new production(ist)’ system He created the second internal venture, in 2002, around a new technique to measure sugar levels in the blood without puncturing the skin. This is potentially of great use to the 10 percent of diabetics who require daily injections of insulin–some 750,000 in Japan alone. Instead of simply marketing the instruments, Mitsumaki’s business model envisaged selling or leasing them to hospitals and clinics, whose doctors would instruct patients in their use. Data from the blood tests would be stored on IC cards, simplifying doctor–patient interaction, and Hitachi would offer data management and other services.13 Here, too, the model envisaged a service package built around specific products such as this tester (and ‘i-mat’, which analyses heartbeats), this time in the area of medical care. Relatedly, Hitachi joined a consortium of Japanese companies to develop an infrastructure which would expand the market for such services and products.
PET With the rapid aging of Japan’s population and the growing role of IT, healthcare is fast becoming a major business area for Hitachi. From expertise primarily in diagnostics, it is expanding to encompass therapy, information, biotechnology, and medical services.14 A new diagnostic business is being developed, not as a corporate venture, but in the unlikely setting of the Power and Industrial Systems group. It was started by a small team led by Makio Uchida, who worked in cost accounting and marketing and had watched PIS group revenues fatally hit by declining domestic demand and cost cutting by public authorities. Uchida’s team determined to develop a new business model which did not rely on the manufacture of hardware. They decided on a PET (positron emission tomography) service, which is used for detecting Alzheimer’s-type dementia, malignant tumours, and heart disease. For hospitals, a PET system is costly, and considerable new technical expertise is required. The business model envisaged the equipment being manufactured externally, with the team selecting the equipment, financing–through Hitachi Capital and Hitachi Insurance Services–installing, running, and maintaining it, while drawing on expertise in nuclear technologies and the company’s reputation for reliability.15 Prospects for the PET service received a boost in 2002, when national health insurance coverage was extended to cover malignant tumours. In 2003, the first contracts were signed with the Tokyo Women’s Medical University and the Utsunomiya Central Clinic. In 2004 Hitachi combined PET, optotopography, genome testing, and bio-informatics, and teamed up with a bio-venture company and Kurume University’s medical department hospital to offer clinical testing services to pharmaceutical companies.16 Another nascent medical business, also drawing on nuclear technology but in the therapy area, is PBT (proton beam therapy). The system has been installed at the University of
205
MOT in and between enterprises Tsukuba and the University of Texas, M.D. Anderson Cancer Centre. It remains to be seen whether, along with the construction of new small generators, and energy business, such ventures will succeed in breathing new life into Hitachi’s historical core business. The above ventures emphasize services and solutions. At one extreme Hitachi’s own hardware is eschewed, at the other it is crucial. Although the extent differs, then, they represent a different approach to business than the model successfully pursued in the 1970s and 1980s, when Hitachi was seen as a manufacturing and technology company, with strong ‘push’ dynamics. Sales and marketing often played second, if not third, fiddle. There was a widespread view within the company that over time, in fact, it had often lost sight of the markets it was meant to be serving. This tendency was challenged most starkly in the ‘component concept’, formulated by Shoichiro Asai when he became head of the newly created Corporate Technology Office in 2002 (see Mizuno 2004). The concept places sales and marketing on the front line, and manufacturing and development in a support role, as ‘components’ which can be substituted by external purchases, or purchases elsewhere in the group if necessary. In essence, it was a contract manufacturing (EMS) concept, applied to the company and Group.17 Applying this concept to new ventures is one thing, however; applying it to established businesses is much more problematic. This combination of technology with new business models has been exploratory rather than transformational. It does, however, represent one step in the emergence of a new type of technology management at Hitachi.18
Technology strategy The next step might be described as strategic. Technology strategy became explicitly integrated into the corporate reform agenda and i.e. Hitachi Plan in 2002, when a Corporate Technology Office was set up.19 This office mapped out growth areas within Hitachi’s broad business domains on the one hand, and Group technology resources addressing those areas on the other. Areas of potential synergy were identified, as well as gaps. The 30 areas (including the three introduced above) were subsequently grouped into nine fields, half transcending corporate boundaries (Figure 11.1). Here we will look briefly at three of the so-called ‘Inspire A’ businesses.
Automotive products As electronic devices increasingly permeate automobiles, a trend which will be accelerated with hybrid cars and eventually fuel cells, and as Japanese autocompanies take a larger share of the global market, automotive products
206
Hitachi’s ‘new production(ist)’ system Expansionary phase
Incubation phase
Next generation core business aiming for No.1
New businesses that will become future core businesses
Medical, biotechnology
Interdivision businesses Leveraging the collective strengths of the Hitachi Group Production, testing equipment Storage SAN/NAS Hard disk drives Ubiquitous HDDs
Individual businesses Development spearheaded by individual business groups
For semiconductors/ displays
Medical diagnostic systems Medical services
Key technologies, devices Nanotechnology Materials
Social infrastructure systems Urban planning and development Railway systems
Services, solutions Outsourcing e-government
Batteries Lithium-ion batteries Fuel cells
Automotive On-board car information systems Electric powertrains
Consumer Digital home appliances (plasma TVs, etc.)
Figure 11.1 ‘Inspire A’ businesses (Stage 2) Source: Hitachi Ltd mimeo
become increasingly attractive to electronics companies like Hitachi.20 Growth areas identified within this business domain were telematics and electric powertrains, as well as lithium-ion batteries and fuel cells. Until the late 1990s, automotive products constituted a relatively small division within Hitachi. The potential to apply technologies from other operations such as locomotives and industrial machinery was considerable, but latent. Hitachi began strengthening the business in 1999, when it acquired a 17 percent stake in Unisia Jacks from restructuring Nissan. It bought out a car navigation joint venture with Nissan in 2000, and the remaining shares of Unisia in 2002. In October 2004 it reversed the trend of spin-offs by absorbing Unisia, as well as Tokico. In 2004, too, it formed a new company with Hitachi Maxell and Shin Kobe Electric Machinery called Hitachi Vehicle Energy to develop lithium-ion batteries for hybrid cars. After the mergers, Hitachi ranked 24th in the world in terms of automotive product sales, and was only half way to its goal of ¥1 trillion sales in the industry by 2010, but its depth of technology resources in growth areas suggest it may be a powerful competitor or partner in the future, especially if US assemblers strengthen their ties with Japanese electronics companies for hybrid technologies.21
Urban development ‘Urban development’ also transverses internal business group and corporate boundaries. Previously, there was little internal coordination between them. A core group was Building Systems, whose business was focused on elevators and escalators, which are vulnerable to business cycles. Similar to increasing
207
MOT in and between enterprises permeation of electronics in automobiles, however, Hitachi’s expertise in other areas could allow it to contribute value added and capture profit in new growth areas. By reconceptualizing the market, new business opportunities could be found which drew on technology synergies within the Group, and would be service and systems-led, utilizing but escaping from dependence on hardware. Control technology expertise from trains and industrial systems, for instance, could be put to work in building energy optimization. This, in turn, could be combined with IT expertise (including security) and financial expertise for ‘total asset management,’ which could be managed by Hitachi on an outsourcing basis. A new Urban Planning and Development Systems group was established, headed by a former CRL researcher who had moved over to marketing and had scored a great success with the 32 inch plasma TV. His was another unusual appointment, aiming to bring fresh market-attuned eyes in to transform the business. He was joined by a corporate senior staff member with a background in control systems, as well as an author of Hitachi Research Institute’s ‘ubiquitous metropolis’ concept. At the same time, a Group-wide council spanning 17 companies was set up to develop new businesses.
Hard disk drives In 2002, Hitachi announced it would purchase (the bulk of) IBM’s loss-making hard disk drive (HDD) business for US$2.5 billion. The announcement came in the wake of Hitachi’s second loss, bigger than the first and briefly the second biggest in Japanese history. Analysts were incredulous. They saw HDDs as lossmaking, not just in the immediate future, but perhaps forever, as commoditization and cut-throat competition intensified. This was definitely not the way to restore profitability, and flew in the face of the company’s intent to exit lossmaking businesses. Hitachi’s top management, by contrast, saw HDDs as a key element in storage solutions, which was a strategic business domain for integrating a range of technologies. Under the revised ‘i.e. Hitachi Plan II’ of 2003, HDDs were one of a limited number of ‘global products incorporating advanced technology’, in which the company had a chance of establishing a strong position globally.22 A strategic view of technology was thus coming to influence the composition of the business portfolio. It led, not to a narrowed focus, but a broadened one. While the business objective was profitability, and positive ‘FIV’,23 it was not maximization of the return on capital. Indeed, the i.e. Hitachi Plan II announced that the company would withdraw from operations with sales amounting to 20 percent of company turnover, but in February 2004 the president announced that this target would be revised downwards to 10 percent, two-thirds of which was already accounted for by spinning out logic chip operations into Renesas Technologies.
208
Hitachi’s ‘new production(ist)’ system
A new production(ist) system? The second step in Hitachi’s nascent new production(ist) system saw the elaboration of a technology strategy, integrated into the medium-term business strategy. It sought to strengthen super-enterprise coordination of R&D and technology resources for the development of new synergy (Inspire-A) businesses.24 Thus, consolidated management had moved beyond restructuring of the business portfolio to place, by intention at least, technology management at the heart of business development. The third step systematized this trend, and linked it to product development and production, themselves the subject of innovations designed to enhance speed and synchronization in technology commercialization. In April 2004 the Corporate Technology Office was placed in the new Group Headquarters, along with global business strategy and brand management. The objective was a ‘strengthening of R&D through mutual (cross-company) use of Group technology’ and ‘strengthening of product competitiveness through vertical integration of Group technology’ (Hitachi 2004). It would aim to promote frontier (‘paradigm shift’) technology and business models, as well as platform research to raise productivity, decrease development time and strengthen manufacturing technology. To this end, a new materials research lab. and three new centres–testing, simulation, and embedded software–would be created. Patent strategy would include the use of frontier and platform research patents by Group companies. The Group CTO Council, comprised of Chief Technology Officers of internal business groups and 29 Group companies (started in 2002) would oversee strategy and its implementation. Regarding product development and production, a raft of projects, supported by the Monozukuri Engineering Division included: (1) TSCM (total supply chain management) innovations; (2) development of a ‘Worldwide Real-time Management System’ that establishes techniques based on past management know-how, and utilizes intellectual assets (non-financial assets) and simulation technologies; (3) HiSPEED 21 Activities, which are designed to improve design quality and shorten design time through the use of design techniques and integrated digital engineering; and (4) ‘e-Meister Activities’ which pass on the techniques of highly skilled veterans in the workplace using the latest digital media technologies and aim for even greater levels of mastery.25
Technology and the production system increasingly came to interact with organization and skill development/HRM in the third stage (see Figure 11.2). And finally the reforms, which had a primarily domestic focus from 1998–2001, had begun to take on a more global perspective, signalled for instance by the identification of ‘global products incorporating advanced technology’ in the i.e. Hitachi Plan II and the attempt to develop a ‘worldwide real-time management system’, not to mention the announcement in 2004 of planned
209
Degree of overall optimization
• Innovations targeting overall optimization in all processes (Flow of goods, information, money) • Strengthening technical capabilities in development and design
i.e. Integrated companies • Worldwide real-time management systems • Worldwide collaboration, etc.
Total solutions based on Monozukuri technologies Company-wide innovation projects
Promoting company-wide innovations related to Monozukuri Cash flow Improvement projects Reduction in inventory assets
IT promotion projects
Total SCM innovations
Worldwide management systems
Development/design innovations (HiSPEED21)
Total SCM/HiSPEED21
Skill transmission activities (e-Meister)
Manpower training
Strengthening foundations of Monozukuri technologies
SCM Operational innovations
Static SCM: supply chain management
Company-wide cross-cutting functions (technical working groups, etc.)
Rapid prototype solutions
Degree of response (speed) e-PEC: e-professional engineering consultation
Figure 11.2 Monozukuri, MOT, and corporate management (Stage 3) Source: T. Shimizu 2004: 12
On-site technical support (PT)
Technological consultation (e-PEC)
PT: production technology team
Dynamic
MOT in and between enterprises
210 Next generation management systems
Hitachi’s ‘new production(ist)’ system investment of US$1 billion in China over the next three years, aiming to increase sales there from US$4.5 billion in 2003 to US$7.0 billion in 2006, and procurement from US$3.0 billion to US$6.0 billion. R&D in China would also be strengthened, along with that in the US and Europe. Collaboration with universities, both within Japan and abroad, received increased emphasis.
Conclusion It is possible to discern in these developments the emergence of a systemic response not just to ‘large firm malaise’ as in 1998–99, but to the changed competitive environment and challenges outlined in the Introduction of this book. It began with HRM and organization reforms, but then encompassed technology management in a step-wise evolution which increased strategic as well as operational coordination. This nascent system seeks to mobilize and integrate resources rapidly for production, product development, and even more fundamental technology development. It is more open than in the past, but differs significantly from Silicon Valley-type ‘open networks’. External alliances are common (witness the numerous announcements on Hitachi’s website), but the enterprise Group plays an important role. It may be described as a ‘Galaxy system’ with Hitachi (and large listed Group companies) at the centre, key inner core companies, outer orbit companies, as well as external alliances and collaborations. Key features of the new model, extrapolated from the examples and initiatives described in this chapter, are: 1 A strategic view of technology at the heart of business strategy 2 Increased emphasis on services and solutions without abandoning monozukuri (production(ist)) orientation 3 Balancing of technology and production-push with customer-pull dynamics 4 Innovation in business models for commercializing technology 5 Increased emphasis on profitability, though not necessarily shareholder return maximization 6 Pursuit of Group synergies to leverage technology resources, supplemented by external alliances 7 Mobilization of resources and synchronization of their input on a more global scale 8 Information and decision making processes more conducive to speed and risk taking 9 Complementary changes balanced by continuity in organization and skill formation/HRM (cf. the ‘new community firm’). There is continuity with the previous system, in the continued importance of manufacturing, for instance, and in-house R&D. It is less open, more
211
MOT in and between enterprises centrally organized than the US/Silicon Valley model described in this volume. There is less emphasis on focus, more on (mainly organic) growth in new technology frontiers. This is not surprising, given the context in which it has emerged. Japan’s universities, for instance, are only beginning to play the role in Japan that they came to play in the US innovation system from the 1980s (see for example Arimoto’s chapter). Labour market mobility is more muted, the role of start-ups more limited. Is this new system a half-baked one, or does it combine the merits of the former system with similar benefits which accrued in the new US model? One measure, of course, is profitability. We have seen that Hitachi’s loss in 1998 was followed by an even bigger loss in 2001. It remains to be seen whether the gradual upward trajectory in profits since then can be strengthened in the coming years. This will depend on a number of factors, including internally, how far the reforms and initiatives actually take root, as well as external factors, as noted above. The question of whether the quest for ‘paradigm shift’ innovation will bear fruit, and declining competitiveness in key ICT and electronics markets can be reversed, is an even more open one. Other Japanese electronics and manufacturing companies have gone through restructuring and reforms which share at least some of the features of Hitachi’s reforms. They, too, have been seeking to apply technology management concepts. The results to date may be modest, but the perceptual changes relative to the 1980s and even mid-1990s are substantial.
Notes 1. Production(ist)’ denotes both the system and its association with value creation and values implicit in the term monozukuri. See also Note 22. 2. A fuller account of this section can be found in Inagami and Whittaker, 2005, Part II, based on roughly 150 interviews in Hitachi and Group companies between 1996 and 2004. 3. In 2003, these governance bodies were in turn replaced as Hitachi became one of the first Japanese companies to adopt the US-inspired committee system of corporate governance. 4. ¥300 billion would be allocated for acquisitions, equity participation, and alliances, ¥200 billion for existing IT businesses and ¥50 million for a new Internet-based solutions business. Divestitures were not specified. The goal was to achieve 70% of sales in information-related sectors by 2003, including 25% in services. The subsequent bursting of the IT bubble, of course, undermined the premises of this plan. 5. ‘Group’ is written with a capital ‘G’ to avoid confusion with internal groups (‘virtual companies’). There were about 1200 companies in the consolidated Hitachi Group in the late 1990s. 6. These included Kokusai Electric, Hitachi Denshi, and Yagi Antenna to create Hitachi Kokusai Electric in telecommunications, and Hitachi Credit and Hitachi Lease to
212
Hitachi’s ‘new production(ist)’ system
7. 8.
9.
10. 11.
12.
13. 14. 15. 16. 17.
18.
19. 20.
21. 22.
create Hitachi Capital, Japan’s largest leasing company, with growing business in retail, credit cards, securitization, and business outsourcing. Additional measures raised compensation for patents, and in 2005 internal award, negotiation and disputes procedures for inventions were systematized. The new internal groups, too, were encouraged to develop their own HRM systems, according to their business needs, introducing another kind of diversity. The ICT groups were the first to introduce their own system in 2001. Information in the following three sections draws on interviews/discussions with Hitachi managers in 2003–05, and various written sources, especially Mizuno (2004) in this section, and www.hitachi.com/rev/ Kanter 1989; Welch with Byrne 2001; Gerstner 2002. The mu-chip measured a mere 0.4mm x 0.4mm x 0.1mm, with 128 bits in ROM, 100 of which were available for a unique ID activated by an external RF signal. The name ‘mu’ comes from the initials of the project leader (Mitsuo Usami), and can additionally be read as the Greek letter m, meaning microscopic, or a micrometre (one millionth of a metre). See Isaacs (2002) for a business case on the mu-chip. We should note that the venture was, in fact, absorbed into Hitachi as a new division in 2004. Also that Hitachi is leading METI’s ¥1.8 billion ‘Hibiki’ project, a quest to develop a mass producable ¥5 rewritable IC tag by June 2006. Mu-chip prices in 2004 were said to be in the range of ¥10–19. Nikkei shinbun, 22 February 2004. PHVC, too, aims for sales of around ¥15 billion by 2007. Clinical trials were expected to be completed in 2005. See, for instance, Hitachi Review, December 2003, 54(4). www.hitachi.com/rev/archive/2003/2003341_12873.html See Izumida et al., in Hitachi Review, 52(4). Nikkei shimbun, 27 July 2004. The analysis is all conducted in English in anticipation of orders from non–Japanese pharmaceutical companies. Regarding technology-push versus customer-pull dynamics, there has been no serious debate within Hitachi about abolishing the Central Research Lab. To be sure, there was criticism that CRL researchers pursued their research irrespective of its commercial potential, and that they were too cut off from the ‘real world’ to recognize commercial potential, anyway. In response, a ‘management academy’ was set up to study business models for commercializing technology, as well as open days for interaction with customers. Relatedly, a nine month Advanced Course for Executive Engineers was introduced to develop leaders capable of proposing and developing new technology-based businesses. Engineers were also added to the corporate audit function, to audit management from an MOT perspective. Electronics was estimated to account for 22 percent of auto content by value in 2001, and this is expected to rise to 40 percent in 2010: see K. Francis, Automotive News, 29 Oct. 2001, p.3, based on Delphi Estimate. Main customers of the merged operations were Nissan 45 percent, GM 17 percent, and Ford 8 percent: Nikkei sangyo shimbun, 29 March 2004. Estimates of global HDD share ranged from 17 percent to 19 percent in 2003, but for small (2.5 inches or smaller) HDDs, Hitachi was estimated to have around half of global market share: Nikkei shinbun, 22 June 2004; 1 July 2004.
213
MOT in and between enterprises 23. ‘Future Inspiration Value’–an economic value added (EVA) measure in which the cost of capital is deducted from after tax profit. Hitachi aimed to achieve positive FIV in 2005–6. 24. Overall, R&D spending decreased significantly. On a consolidated basis, it was over 6 percent of turnover until 1999–2000, when it plunged to 4.3 percent; it briefly recovered to around 5 percent before falling again to 4.3 percent in 2003–04 and 2004–05. (Capital investment, on the other hand, rebounded to 9–10 percent, and was maintained even during the loss of 2001–2.) 25. T. Shimizu, GM of Monozukuri Engineering Division, in Hitachi Technology 2003– 2004, p.13. The Monozukuri Engineering Division emerged from the corporate production engineering function in 2000 with a broader remit, as the quote suggests. Shimizu defines monozukuri as ‘the crafting of things of quality’ and states: ‘Creating the truly high-quality products that customers want is the speciality of the Japanese production industry, and we want to promote a renewed awareness that Japan has the technology and intellectual power to back this up’ (p.14).
References Best, M. (2001). The New Competitive Advantage: The renewal of American industry, Oxford: Oxford University Press. Dore, R. ([1973] 1990). British Factory–Japanese Factory: The origins of national diversity in industrial relations, London: Allen and Unwin, and Berkeley: Univ. of California Press. Gerstner, L. (2002). Who Says Elephants Can’t Dance? Inside IBM’s historic turnaround, New York: Harper Business. Hitachi (2004). ‘Kenkyu kaihatsu oyobi chiteki zaisan hokokusho 2004’ (Report on Research and Development and Intellectual Property 2004), Tokyo: Hitachi Ltd. —— (various dates), Hitachi Review, Tokyo: Hitachi Ltd. Inagami, T. and D. H. Whittaker (2005). The New Community Firm: Employment, governance and management reform in Japan, Cambridge: Cambridge University Press. Isaacs, A. (2002). ‘The Mu Chip’, UC Berkeley, Haas School of Business case study, 20 October. Kanter, R.M. (1989). When Giants Learn to Dance, New York: Simon & Schuster. Mizuno, Y. (2004). Hitachi: Gijutsu okoku saiken e no ketsudan (Hitachi: Rebuilding the technology kingdom), Tokyo: Nihon keizai shimbunsha. Shimuzu, T. (2004). ‘Contributing to the Recovery of the Production Industry Through the Knowledge of Monozukuri Technologies’, Hitachi Technology 2003–2004, Tokyo: Hitachi Ltd. Welch, J. with J. Byrne (2001). Jack: Straight from the Gut, New York: Warner Books.
214
12 Interfirm networks and the management of technology and innovation in Japan James R. Lincoln
Introduction This chapter reviews a research programme on the role of keiretsu networks in Japanese technology development and management that I with various colleagues and students have pursued in the last ten years. The conceptual theme running through these projects is that of ties and networks (Powell 1990). Technological innovation and its commercialization in Japan business have been embedded in and moulded by networks to a degree extraordinary by American economic norms (Dore 1983; Granovetter 1985; Uzzi 1996). The Silicon Valley model of innovation and entrepreneurship is often described as network-based, but this is less the durable interfirm ties typical of Japan and more the informal encounters of individuals ‘doing lunch’ and otherwise transacting business in a bounded geographic space (Saxenian 1994). The distinctive flavour of the ties that bind Japanese firms to one another and the broader networks into which those ties are woven both shape and in turn are shaped by innovation strategies. At a micro-level, the trust, reciprocity, and stability typical of customer–supplier dyads in Japanese industrial goods markets have facilitated cooperation, synergy, and knowledge-sharing in product and process development. At a more macro-level, webs of crossshareholdings, director transfers, and preferential trade and lending flows have functioned both as information systems and as governance structures to disseminate and while conserving and protecting knowledge assets (Williamson 1996; Lincoln and Gerlach 2004). By the same token, the breaks and ‘holes’ in the network (Burt 1992)–e.g., between direct competitors or rival groups–have at times presented formidable barriers to Japanese firms’ collaboration and learning (Lincoln et al. 1998).
215
MOT in and between enterprises At the core of any treatment of Japanese corporate networks is the keiretsu phenomenon. The term references stable clusters of corporations, and financial institutions, defined by the abovementioned ties and acting to share risks and pursue opportunities. Keiretsu are of two main types: horizontal–the diverse constellations of large financial institutions, trading companies, and manufacturers, some of which descended from the prewar zaibatsu; and vertical–the upstream supplier networks and downstream distribution channels of a large manufacturing firm. Keiretsu are commonly classed as ‘business groups’. While a fitting label for the Korean chaebol, Russian financial conglomerates, and the kinship-based business empires of Brazil, Chile, and India– the ‘group’ label fails to capture the complexities of the Japanese case (Granovetter 2004). The keiretsu are bona fide networks in the sense that their boundaries are vague and permeable, and (unlike the zaibatsu) no one centre exercises full control (Podolny and Page 1998; Lincoln and Gerlach 2004). How have keiretsu networks contributed to the creation, diffusion, and commercialization of Japanese technology? And how has that contribution changed over time? I begin by reviewing how keiretsu ties, particularly the vertical variety, have conditioned product and process development in Japan. Much attention has been given to the symbiosis that distinguishes Japanese manufacturer–supplier relations, enabling the manufacturer to tap the supplier’s expertise and creativity at an early product development stage. By contrast, critics see the arms-length and adversarial flavour of American manufacturer–supplier exchange denying the former access to the latter’s competencies and thus the synergies observed in Japan (Helper 1991). Our case study of Hitachi Omika shows how tight-knit customer–supplier ties facilitated tacit knowledge-sharing around a complex, customized product technology and the generation of new product ideas. Reports of how keiretsu-esque–i.e., close, lasting, reciprocal–procurement relations support organizational learning, innovation, and efficiency permeate the scholarly and journalistic literature on Japanese business. Yet much of this literature neglects the nuanced and evolving character of such relations. First, the causality is symmetric: ties are consequence as well as cause of innovation strategy. Two papers in our programme argue the point. Michael Gerlach and I document how keiretsu genealogies arose from the distinctive Japanese corporate strategy of commercializing innovation by founding and incubating, then spinning off as satellite enterprise, new product divisions (Gerlach and Lincoln 2001). Work I have done with Didier Guillot and Christina Ahmadjian on Matsushita Electric further uses case study materials in examining how Matsushita’s cultivation of a keiretsu network among its top tier suppliers enabled it to better share the fruits and burdens of product and process innovation. Second, the evidence is ample that economic change has altered the historical equation between keiretsu and innovation. Many now see keiretsu networks
216
Interfirm networks having lost their functionality as conducive infrastructure for fast product development and efficient manufacturing process. The final section reviews two of our studies that by and large concur. The first details how Toyota’s drive to acquire expertise in automotive electronics caused the fraying of its theretofore close and fruitful tie to long-term keiretsu supplier (and former division) Denso. The second is a quantitative analysis of strategic alliance foundings in the Japanese electronics industry. Our evidence shows that, unlike the 1980s when keiretsu networks seemed a supportive platform for strategic partnerships of any sort, R&D alliances in the 90s formed in general disregard of keiretsu ties.
‘Taking the role of the other’: High trust ties in interfirm learning Interorganizational ties of the keiretsu sort foster learning and innovation by enabling one company to access the routines and tap the tacit knowledge pools of another. Like the corporate culture of a single firm, such interpenetration of habits, skills, and values allows each partner to adapt to the unspoken, even unrealized, requirements of the other. Imai et al. (1985: 351) influential paper on Japanese product development documents learning of this sort at Fuji Xerox. They quote a manager as follows: We ask our suppliers to come to our factory and start working together with us as early in the development process as possible. The suppliers also don’t mind our visiting their plants. This kind of mutual exchange and openness about information works to enhance flexibility. Early participation on the part of the supplier enables them to understand where they are positioned within the entire process. Furthermore, by working with us on a regular basis, they learn how to bring in precisely what we are looking for, even if we only show them a rough sketch. When we reach this point, our designers can simply concentrate on work requiring creative thinking.
This interfirm process can be characterized as a kind of ‘taking the role of the other’ (Mead 1962). Employees of Firm A immerse themselves in the organization of Firm B to a degree that knowledge transfer occurs without a prior conversion of tacit into explicit knowledge codes. We observed such ‘taking the role of the other’ in Hitachi Omika, a Hitachi division that manufactures computer control systems to the specifications of large customers such as Japan Rail and Tokyo Power. The complexity of Omika’s products present customer and vendor alike with learning challenges. Upon taking possession of a Hitachi system, customers lack the ability to operate it proficiently. Omika’s customers are ‘high reliability’ organizations that run nuclear power plants or rail systems on which the lives of multitudes depend. A mistake in managing the technology could have disastrous consequences (Perrow 1984; Roberts 1990). An Omika manager commented:
217
MOT in and between enterprises Our designers cannot work effectively with ‘white gloves.’ We have to know the corporate culture in the place of production (genba). At the time we introduce the equipment to the customer, it is not fully operational but can still be used to perform many functions. We then monitor what the customer wants to do. It is our responsibility to provide the customer with all available information about the use of the machine. This means comprehending the user’s potential demands. In testing the machine, the user may not fully realize his own needs. Hitachi employees must determine what the customer really wants to do even if the customer is not clear on it. Sometimes it is very difficult to completely understand and solve a customer problem immediately, but we can ‘hold and keep’ the problem while building trust and working together to solve it. So a longterm relationship with the customer is critical. The key point is that success in this review requires more than our people thinking that the project turned out fine and the customer walking away feeling good.
After Hitachi ships a system, its engineers may work for months onsite, fine tuning it, monitoring the customer’s requirements, and training staff. Hitachi, of course, needs to maintain a strong reputation for customer service in order to attract new and repeat business. Additionally, our Hitachi informants said, such immersion in the customer fed new ideas back to Hitachi’s voracious product development machine. As another Omika manager put it: We have to have a keen sense for users’ needs. This provides the seeds for new products. Such awareness of needs and seeds is required (must be cultivated) in the same person (i.e., the Hitachi employee). To make use of information (on new product ideas), our people have to have a keen sense for information whenever they have dealings with the customer.
Research on customer–supplier transactions in intermediate product markets often highlights the facilitating role of trust, but it is often vague on what trust means and how it works to the advantage of the transaction partners. The Omika case, in highlighting how people and processes are meshed in the transfer of tacit skills and knowledge between a customer and a supplier organization, sheds some light on the workings of high trust, tight knit interfirm exchange.
Innovation strategy and the engineering of networks Our Hitachi Omika inquiry, Imai et al.’s (1985) discussion of Fuji Xerox, and other studies of close ties and cohesive networks in Japanese technology management take interfirm ties as given and look to consequences for the partner firms. Two of our studies turn the causal arrow around: both cast innovation strategy as driver of network formation. The first reviews Michael Gerlach’s and my historical treatment of how Japanese companies craft keiretsu networks by delegating to a ring of ancillary enterprises the task of commercializing product innovations (Gerlach and Lincoln 2001). The second examines Matsushita Electric’s efforts to build a collaborative and innovative network among the elite members of its domestic supplier base.
218
Interfirm networks
Corporate spin-offs and the commercialization of innovation in Japan1 Product innovation has followed a two pronged process in Japanese business: at the firm level in terms of the creation and spin-off of new enterprise; and at the interfirm level via the integration of innovators, capital providers, and users of new technologies. Our focus was the evolution of the Japanese electrical and electronics machinery industry. The sector is distinctive both for its long-term strategic importance and its strong global competitiveness. It has been Japan’s most innovative. Through the 1980s, office computing and accounting machines generated the most new patents among Japanese industries, communications equipment ranking third (The Economist, 20 May 1989). In just two other industries (photography and photocopying, and prime movers and power engineering) was there a higher ratio of Japanese to American patents (Frame and Narin 1990). The organization of innovation in Japanese electronics has been distinctive in two main ways. First, within keiretsu networks, the boundaries of firms have been vague and permeable, a structural property that facilitated interorganizational knowledge flows. Such networks were often born of business and government leaders’ efforts to limit corporate size and scope by divesting noncore operations. Satellite firms proliferated where: 1) innovations posed technological and market opportunities that departed from the parent firm’s mainline businesses; 2) localized attention to and incentives for development and commercialization demanded decentralization; and 3) close ties to strategic partners were desired. This process of generating ancillary firms that stay within an orbit of keiretsu control meant that the parent could remain focused on core competencies yet still sew the seeds and reap the rewards of new technologies and markets (Hamel and Prahalad 1994). Second, once in place, the division of labour between core and satellite distributed the burdens of innovation and development in a fashion quite different from the US electronics industry. Silicon Valley is much noted for the roles that independent inventors and their venture-funded start-ups play in innovation and enterprise creation. Entrepreneurial firms are often founded by engineers fleeing the bureaucratic rigidities and incentive limitations of established companies. But while such venture businesses may convert their founders’ research into patentable innovations, far fewer succeed at downstream commercialization. The latter demands the financial, human, and material resources of large players who appropriate control of the start-up after acquiring, licensing, or copying its business model and intellectual assets (Teece 1986). In the Japanese electronics industry, research and development have been the province of large firms. Yet those same firms often delegated the commercialization stage to satellite enterprises focused on manufacturing, user driven
219
MOT in and between enterprises services, and custom applications. Such ‘venture businesses,’ were less involved than their US counterparts in innovation than commercialization. Corporate spin-offs are not, of course, unique to Japan. Divestiture of product divisions to reduce costs and better focus on a core business is common American practice. AT&T’s spin-off of its former manufacturing arm, Western Electric, as Lucent is a notable example, as are Hewlett Packard’s disposal of its instrument business as Agilent; and Ford’s and General Motors’ spin-offs of their parts divisions as Visteon and Delphi, respectively. In the early 1990s, IBM converted to semi-autonomous companies (‘Baby Blues’) its business units responsible for personal computers, printers, operating-system software, multimedia software, and microprocessors (Business Week, 3 May 1993). Even so, the spin-off phenomenon figures much more prominently in the organizational evolution of the Japanese than Western firm. Major companies in the communications and computer industries generated hundreds of enterprises. IBM produced fewer than two dozen. Another contrast is the keiretsu tie maintained between Japanese satellite and parent. The US parent firm either held onto the spin-off as a wholly owned subsidiary (e.g., the IBM PC, until its sale to Lenovo in 2005) or gave it full independence (e.g., the AT&T split up). In Japan, keiretsu coordination was the rule. The parent pursued significant but not exclusive or controlling financial, commercial, and personnel involvement that (assuming the offspring’s success) diminished with time. Japanese corporate design over most of the 20th century shunned full-scale vertical integration, broad diversification, and great scale. The share of the economy controlled by the largest Japanese firms actually declined through the interwar years and again in the postwar period (Nakamura 1983). This trend accelerated with the oil shocks of 1973 and 1979, which ushered in the era of genryoh keiei, or ‘weight-reduction management’ (Uekusa 1987). In thus limiting scale and scope, however, core firms’ reliance on the spin-off strategy grew. Aoki (1987) reports that in 1965 major manufacturers reinvested 11 percent of their own paid-in capital in satellite firms, a figure rising to 44 percent by 1984. In the automobile and electrical machinery industries, this growth rate reached 86 percent and 73 percent (from an original 13 percent and 14 percent), respectively. Thus, the individual Japanese firm remained a moderately sized and rather specialized entity, while distributing a substantial portion of its business across a loosely linked keiretsu network (Fruin 1992). Indeed, many of Japan’s largest companies themselves began life as growth and innovation oriented spin-offs. In the prewar period, the zaibatsu grew by diversifying beyond their traditional strengths in mining, finance, and trade to form new steel-making, chemicals, electrical machinery enterprises at Japan’s technological frontier. Postwar, the pattern persisted in autos and other growth industries, such prominent firms as Denso (Japan’s largest automotive parts maker) originating as divisions of larger corporations (in this case, Toyota).
220
Interfirm networks The keiretsu bonding of parent and satellite gave the former a degree of administrative control over movements of research personnel that in Silicon Valley, for example, are market driven. The seniority pay and promotion system typical of large Japanese firms limits the heaping of rewards on talented people. Managers and engineers transferred to the spin-off enjoyed entrepreneurial opportunities denied product division staff. Compared to the internal product division of a diversified ‘M-form’ American firm, to which resources are allocated in accordance with a corporate budget and strategic plan (Williamson 1985), the spin-off firm exercised considerable discretion in human resource policy, business strategy, and operating procedure. The cap on corporate scale had other implications. Smaller size facilitated team camaraderie and employee identification with the company and its products (Florida and Kenney 1990: 48), important considerations given the vaunted role these play in Japanese corporate culture (Lincoln and Kalleberg 1990). Aoki also (1987: 285) sees the spin-off process (‘quasi-disintegration’) as motivated by management concerns with keeping employee groups homogeneous in career prospects, labour conditions, and work values so as to maintain workforce cohesion. An important difference between the operations of spun-off satellites and those of parent firm divisions concerned the pattern of growth. While division growth was often restricted by the limits on salary and discretion imposed by the corporate administrative structure, the spin-off could in principle expand without bound. As corporate scale is strongly correlated in Japan with status and monetary reward, success as a spin-off translates not only into sales and profit growth but also better human resources, lower capital costs, and higher stature in the community. The satellite–parent tie also became more symmetric over time, as the flow of personnel and equity reversed direction. In some cases, the spin-off grew to surpass the parent as did Toyota Motor, a divestiture of Toyoda Automatic Loom Works, today a minor company and member of the Toyota keiretsu. In a world of Schumpeterian creative destruction, strong incentives to found new business coexist with market conditions that doom the majority of entrepreneurial combinations to failure. The spectre of failure looms large in Japan, where business deaths have run at roughly double the US rate (Ramseyer 1991). It fuelled the diffusion of the spin-off strategy: the core firm could diversify into uncharted territory yet protect its reputation and mainline business should the satellite venture prove a flop. Moreover, the continuing association between parent and satellite mitigated some of the costs of satellite failure. Personnel seconded to an unsuccessful spin-off were generally taken back. Such absorption of employee risks ensured that talented people would not lose their ‘place’ in corporate Japan’s rigidly structured internal labour markets and status hierarchies. This division of labour between parent and satellite along the innovation chain is evident in the evolution of the Furukawa group, of which Fujitsu
221
MOT in and between enterprises (Japan’s largest computer manufacturer) is a prominent member. As with keiretsu genealogies more generally, the Furukawa network arose from an evolutionary process of sequential diversification over several generations of companies. The Furukawa group’s original business was in copper mining, beginning with the extraction and processing of basic metals, chemicals, and other materials needed by the electrical industry. The founding company, Furukawa Corp., is still active in this industry. Furukawa moved out of copper in stages. First, it invested in and acquired sales rights to a wire products operation in Yokohama before merging that business with its own electric copper refining facilities to form Furukawa Electric. Next it collaborated in 1923 with Germany’s Siemens, a company with which Furukawa had longstanding ties, to form a joint venture, Fuji Electric, designed to transfer Siemens’ electrical machinery technology to Japan. (The name, Fuji–which looks to be a Japanese word–derives from the Furukawa ‘fu’ and the Siemens ‘si’.) In 1935, Fuji Electric divested its own fast-growing communications equipment operations, forming Fujitsu. Fujitsu in turn spun off its numerically controlled equipment operations in 1972 to create Fujitsu Fanuc, later Fanuc. Each generation of spin-off represented a distinctive technological and market juncture at which the maintenance of independent organizational identities was deemed strategically essential to the growth of the overall enterprise. In three cases (Furukawa Electric, Fujitsu, and Fanuc), factories or divisions were converted into new firms. In the Fuji Electric case, a new joint venture was founded to which the parent company committed substantial personnel and other resources. In all such cases, the satellite business targeted new product areas, ranging from component wires, cables, and optical fibres (Furukawa Electric) to heavy machinery, measuring and control equipment, vending machines and other electrical products (Fuji Electric) to advanced data processing and communications equipment (Fujitsu) and finally to machine tools and robots (Fanuc). Essentially no product line overlap persisted across spin-off generations. Japanese keiretsu networks are often viewed as institutional ‘givens’ that constrain and enable corporate strategies and performance outcomes such as innovation, growth, or profitability. Our analysis of the spin-off phenomenon makes clear that the emergence, shape, and functioning of those networks at times in the course of Japanese business history were driven by, not drivers of, those same strategies and outcomes. As Chesbrough notes in his contribution to this volume, the Japanese spin-off strategy has provided a solution to the problem of ‘false negatives’ in product innovation. Japanese firms face a real challenge now, however, in adapting it to such dynamic and discontinuously evolving sectors as IT and biotech. It seems unlikely that in these environments spin-offs can substitute for entrepreneurial start-up activity in generating and commercializing innovation.
222
Interfirm networks
Matsushita’s kyoei-kai The creation and spin-off of product divisions represent one prominent way that Japanese firms have constructed keiretsu networks in the enactment of innovation strategy. It is not the only way. Another of our electronics industry studies examined Matsushita’s effort to organize its suppliers as a cohesive network in order to offload to them some of the burden of technological innovation and product/process development. Matsushita’s reputation was one of relatively arms-length transactions with suppliers, maintaining a ‘dry’ (cold and calculating) and kibishii (strict, harsh) posture toward them (Shimotani 1997; Whittaker 1997). Matsushita managers we interviewed were emphatic that they ‘had no keiretsu’. Unlike NEC and Toshiba, Matsushita eschewed personnel and equity ties and the sharing of supplier risks (Fruin 1997). Yet in the mid-90s Matsushita revised its approach to supply chain management. The headquarters Corporate Purchasing Division (CPD) launched a campaign to bring top suppliers together in a cooperative network–the kyoei-kai (mutual prosperity association)–and through its activities upgrade the capabilities of suppliers and contribute to MEI’s research and development strategy. In 1995, Matsushita bought parts and materials from some 10,000 vendors, accounting for roughly one-half the value of its total sales. Forty percent of these were proprietary suppliers from whom Matsushita purchased off-theshelf standard products. The rest were supplier-subcontractors whose production was tailored to Matsushita’s needs. Of these, 270, chosen for the quality and reliability of their products and accounting for a third of Matsushita’s outsourced parts and materials, became the kyoei-kai. Many Matsushita products are complex technologically, requiring large capital investments and intensive engineering to develop and produce. The kyoei companies were generally small and lacked the plant, equipment, and workforce skills to design and manufacture high-tech components. Due to the kyoei program’s heightened standards and Matsushita’s support, however, kyoei designs and processes began matching and in some cases exceeding the sophistication of Matsushita’s own. Matsushita conceived of the kyoei-kai as a complement to its globalization strategy. Like other major Japanese manufacturers, Matsushita was moving lower value parts and materials procurement to China, Malaysia, and other offshore sites. Its message to the domestic suppliers was clear: unless they acquired some specialized high-tech capabilities, Matsushita had no use for them. The CPD was charged with designing and leading the kyoei programme, the goals of which were laid out in a 1993 ‘Revitalization Plan’. To assist the kyoei companies in acquiring sophisticated technological capabilities and participating in Matsushita’s research and development, the CPD measured and monitored suppliers’ performance, while coaching them in: a) raising quality; b) reducing costs by rationalizing (gorika) production; c) improving delivery;
223
MOT in and between enterprises and d) making design and development proposals. The programme used an elaborate system of grading suppliers on multiple performance dimensions. Matsushita is famous among Japanese firms for its diverse product line and divisionalized corporate structure (jigyobusei; see Beer and Spector 1981). Unlike the auto industry, where purchasing is mostly a corporate function, Matsushita product divisions are generally responsible for their own procurement. CPD’s new responsibility to develop the domestic supplier network and increase division–supplier collaboration signalled a corporate shift toward centralized supply chain management. Consistent with its past reputation for arms-length dealings with suppliers, Matsushita’s practice had been to design parts on its own and give suppliers blueprints for making them. As the kyoei programme advanced, however, Matsushita began delegating early stage product and process design decisions to suppliers. Each Matsushita product division, assisted by the CPD, solicited input from suppliers in three areas: concept-in activity, comprehending targets, and aligning with the longer term product trajectory of the division. Matsushita sought to better communicate its expectations to suppliers (through publication of specs, formal training, personnel transfers or shukko, etc.) The product divisions also assisted kyoei companies in preparing design proposals. The air conditioner division, for example, might produce rough drawings or wood models then turn to the kyoei-kai for modifications of product form and function. The suppliers benefited (e.g., gained competitive advantage) from the early inside information they received on Matsushita’s development plans and procurement needs. Didier Guillot and I (2005) contrasted Matsushita’s supplier relations with those of Sanyo Electric, another prominent Osaka-based electronics firm. Unlike Matsushita’s keiretsu strategy of transferring knowledge to suppliers and combining them in a network, Sanyo handled supply relations in one-off fashion—treating procurement as a series of dyadic exchanges versus a single interface with the network as a whole. With data from the two companies on the responsibility given and stage of inclusion of each supplier in the product development and production, we found that Matsushita’s network model yielded deeper and earlier cooperation than did Sanyo’s dyadic approach.
Technological innovation and the dissolution of keiretsu ties While the Matsushita case shows how the pressures of global product competition spawned a keiretsu supply network around a company formerly distinguished by its lack of one, similar pressures were elsewhere eroding longstanding keiretsu ties. The final two studies reviewed here have this theme. The first examines how product and process innovation combined with globalization pressures to weaken the Toyota vertical keiretsu
224
Interfirm networks (Lincoln et al. 1998; Ahmadjian and Lincoln 2001). The second is a quantitative analysis of the decline of keiretsu constraints on strategic alliance formation in the Japanese electronics industry.
Toyota and Denso: How Toyota’s move into electronics weakened a keiretsu tie Toyota’s evolving relationship with Denso Corp.–the car maker’s former division and long-term keiretsu supplier of automotive electronic components– offers an instructive lesson in the impact of technological and global change on procurement networks in Japan (Odaka et al. 1988). Toyota’s drive to acquire strategic capability in automotive electronics has weakened one of the oldest and most successful vertical keiretsu partnerships in Japan. Toyota’s tie to Denso had long been a standout illustration of the efficiency and flexibility advantages of keiretsu sourcing in Japanese manufacturing (Nishiguchi and Beaudet 1998; Dyer and Nobeoka 2000). In the world auto industry, Toyota has been a highly focused company whose core competencies were the design and assembly of sedans. Yet Toyota made deft use of an array of partner firms to manufacture both the components with which it assembled cars and (through consignment or itaku arrangements) to broaden Toyota’s product line to include trucks (Hino), minicars (Daihatsu), and other specialty vehicles (Shioji 1995). Toyota lacked the expertise to produce in-house all the components it sourced from suppliers and incorporated in its cars. Electronics, in particular, were sufficiently afield of Toyota’s skill set that it entrusted their development and design to external subcontractors/vendors, chiefly Denso. Toyota was thus relieved of the need to invest in complex technological capabilities. It depended on Denso but so did Denso depend on Toyota; this reciprocity in their exchange relation functioning as governance structure to mitigate the hazards of opportunism and holdup that dependencies of this sort are known to pose (Williamson 1985; Lincoln et al. 1992). Toyota’s Hirose plant, founded in 1988 near Toyoda City, has four electrical engineering divisions. They specialize in a) the design and planning of electronic parts; b) antilock brake systems (ABS); c) car navigation systems; d) semiconductors. The Hirose plant’s product line thus overlapped and competed directly with Denso’s. By the summer of 1997 Denso was supplying 50 percent of Toyota’s total electronic component needs. Data on Toyota’s purchases from Denso during 1984–93 show the 50 percent figure to be down from four years previous (see Table 12.1). They also show Toyota’s purchases from Denso shifting to lower-tech commodity parts (e.g., alternators, starters, distributors, coils). Steep declines are evident as well in Toyota’s sourcing from Denso of such cutting-edge technologies as traction control and antilock brake systems.
225
MOT in and between enterprises Table 12.1 Denso’s share of Toyota’s total inputs of selected electronic parts by year Part name
1984
1987
1990
1993
Electronic fuel injection Injector Diesel fuel injection Oxygen sensor Nitrous oxide sensor Alternator Starter Distributor Voltage regulator Ignitor Ignition coil Spark plug Glow plug Cruise control system Traction control system Anti lock brake system Speedometer Power relay Keyless entry system Corner sensor Air conditioner Clock Navigation system Flasher Horn
100 . 100 . . 100 100 100 . . . 65 . . . . 80 90 . . 100 0 . 100 45
100 60 100 100 50 100 100 100 . 100 100 60 100 . . 100 68 70 . . 100 0 . 100 45
100 60 100 100 38 100 100 100 100 100 100 60 100 . 67 71 61 70 . . 100 0 . 100 45
100 52 100 98 42 100 100 100 100 95 100 50 90.3 70.6 56 47.9 61.3 70 65 77 100 4.8 26.3 100 50
Note : Empty cells indicate that data are not available. Source : IRC Co. Ltd (Industry Research & Consulting)
These shifts corroborate Hirose managers’ assertions to us that Toyota’s move into electronics was largely a learning strategy and that Toyota continues to use outside suppliers for parts that offer little of value to learn. Why did Toyota take on the daunting challenge of acquiring mastery of electronics? Its contribution to the value of a car has risen markedly in recent years (10 percent at the time of our interview and projected to rise to 40 percent), and Toyota felt it could not remain ignorant of such critically important technology and ‘black-box’ dependent on external suppliers, however trusted they might be. Moreover, electronic systems were becoming so integral to automotive design and manufacture that it was no longer possible to develop and produce them in modular independence of the mechanical side. Toyota told us that its overriding rationale for outsourcing was lower cost. Toyota well understood the parts and subassemblies it purchased from suppliers but could not produce them as cheaply. In the case of high-tech electronics, however, this reasoning did not hold. Toyota felt compelled to learn the technology ‘by doing’–i.e., designing and producing the parts on its own. Also, Toyota’s dependence on Denso had increased to the point of violating a central tenet of Toyota procurement policy: that every part or material be sourced from two rival suppliers (Asanuma 1989). The option
226
Interfirm networks of buying from a second source gave Toyota the leverage to impose its rigorous schedule of cost reductions on the primary supplier. For a number of electronic parts, however, no alternative to Denso existed. Toyota management concluded that Toyota itself must become the second source and thus compete with Denso for Toyota’s own parts business. Finally, the fraying of the Denso–Toyota tie was cause as well as consequence of Toyota’s quest for competence in electronics. Its status as Toyota spin-off and keiretsu satellite notwithstanding, Denso always had other customers, something envisioned at the time of its founding in 1949, hence the name Nippon- as opposed to Toyota-Denso (Odaka et al. 1988). But Denso’s keiretsu obligation to Toyota was sufficiently strong that Denso stopped short of sales to Toyota’s chief competitors: Nissan and the US automakers. That changed in the 90s, partly in response to US pressure on the Japanese car industry to eliminate exclusionary keiretsu ties. Denso’s North American business with the American big three and the Japanese transplants was growing rapidly. In 1996 GM named Denso’s North American subsidiary its supplier of the year. In diversifying its supplier base to include US and other foreign parts makers, Toyota had actively encouraged Denso to seek new customers. Yet Denso’s rapid success in doing so, and thereby reducing its dependence on Toyota, had produced some consternation in Toyota management ranks. The worry was that Denso’s deep knowledge of Toyota operations might spill over to competitors. Toyota was particularly troubled by Denso’s assistance to Chrysler in developing the Neon, a car marketed in Japan and initially thought to be competitive there. Toyota’s ambitious initiative to learn automotive electronics thus arose from the confluence of two forces: 1) Toyota was reluctant to maintain black-box dependence on Denso for technology when that technology was becoming so integral to automobile design; 2) Toyota’s reluctance was amplified by Denso’s global expansion and growing ties to Toyota’s competition. Toyota’s path to expertise and capacity in electronics faced significant hurdles, however. While some 30 percent of its mid-1990s recruitment of engineers was in electronics–a three-fold increase over ten years before–Toyota was at a disadvantage in the competition with Japan’s major electronics firms for EE graduates. Denso was the obvious candidate to assist Toyota in its quest to learn electronics, but that relationship was strained by Toyota’s decision to integrate into electronics in the first place.2 Denso did provide some help early on, shukkoing six technicians to Toyota in 1994 to run a training programme (Lincoln and Ahmadjian 2001). By the next year, however, the Denso people were gone. Conveying satisfaction with the outcome, a Toyota human resource manager told us, ‘We have graduated from the Nippondenso phase’. Toyota has held about a quarter of Denso’s stock and maintained two to three directors on its board. Intensifying frictions between the two firms
227
MOT in and between enterprises around the Hirose division, however, led Denso to assign those people marginal jobs (e.g., public relations) and exclude them from important meetings. The Toyota-Denso story, then, like Matsushita’s, is one of technological innovation strategy reconfiguring a keiretsu network. The difference between the cases is in the outcome: a weakened keiretsu procurement tie in the Toyota instance; a stronger one in the Matsushita case. The final study I review–a statistical analysis of the changing role of keiretsu in shaping new strategic alliances in the Japanese electronics industry of the 1980s and 1990s–is like Toyota’s a tale of innovation strategy breaking down keiretsu networks in a changing Japan.
Keiretsu and strategic alliances in Japanese electronics Keiretsu ties facilitate cooperation and exchange between Japanese companies in a variety of ways. They enable firms to join forces easily in a new undertaking, pooling complementary assets to create synergy–value beyond the sum of the parts. Such ties economize on ‘hard’ governance forms–legal contracts, strong monitoring, or full blown acquisition. Given the numerous costs and risks associated with two firms scoping out one another and committing to an alliance, the option of turning to a known and trusted keiretsu partner has its understandable attractions. Yet the risks to seeking a strategic alliance partner within the ranks of one’s own keiretsu are significant as well, particularly in light of growing evidence that keiretsu ties are withering in Japan’s evolving economy. Despite their merits in trust, familiarity, and alignment ease, intragroup alliances are likely to suffer from insufficient heterogeneity of resources, processes, and cultures. The tradeoff firms face here is an obvious one—tough going at the outset in fashioning a productive partnership yields dividends down the road in long-term capabilities that neither partner can garner on its own. The problem is particularly acute for alliances that concern R&D. Ritshchev and Cole (2003), for example, comment that: ‘sometimes the predilection toward intra-keiretsu R&D in Japan precludes potentially more beneficial fusion across alliance boundaries’. Didier Guillot and I (2005) posed the following question: to what extent do vertical keiretsu relationships constrain Japanese electronics firms’ choice of strategic alliance partners? More specifically: 1) is the propensity for intragroup selection smaller for R&D than nonR&D alliances? 2) Has the keiretsu effect diminished over time? The keiretsu constraint might work in two ways: first, firms are more likely to form strategic alliances if they are affiliated with the same group. Second, firms are less likely to form a strategic alliance if they are affiliates of different groups. The one outcome does not require the other: members of the same keiretsu might more easily mesh routines and collaborate in a new venture.
228
Interfirm networks Alternatively, the gains to intra-keiretsu partnering could be few, but competitive rivalries nonetheless drive firms to shun cross-group alliances. The taboo on cross-sourcing long honoured by the Toyota and Nissan supply keiretsu illustrates in the context of procurement relations. We addressed these hypotheses with a longitudinal data set on strategic alliances in the Japanese electronics industry between 1985 and 1997. From the population of Tokyo, Nagoya, and Osaka stock exchange listed electronics firms, data were obtained on a sample of 128, including every such company that had entered at least one alliance, whether domestic or international, over the 11-year period. We coded information on alliances from press reports published in the five largest economic/industrial Japanese newspapers (Japanese Economic Newspaper, Japanese Industrial Newspaper, Daily Industrial Newspaper, Japanese Economy and Industry Newspaper, Japanese Distribution Newspaper). Each firm’s keiretsu affiliation was coded from Kigyo keiretsu soran, an annual publication that records and describes keiretsu ties among Japanese firms. The financial data used in the analysis derive from the Japan Development Bank Corporate Finance Data Bank, providing unconsolidated and consolidated accounting data on companies (excluding finance and insurance) listed on the Tokyo, Osaka, and Nagoya Stock Exchanges. Thus, every publicly held Japanese electronics firm ‘at risk’ of an alliance falls in the sample. From these data we constructed two panels: 1985–91 and 1992–97. The dependent variable was the likelihood that a pair of firms–a dyad–formed an alliance in a given year (Gulati 1995). Thus, there are N(N-1)/2 ¼ 8192 dyads in each year or 56,896 dyad-years in the 1985–91 interval and 48,768 dyad-years in the 1992–97. We used a random effects probit to model the probability that an alliance was reported for an included pair of firms in each year. Alliance events were divided into two classes. Alliances formed for the purpose of joint development of a new product or technology were coded as R&D alliances. NonR&D alliances were geared to production, distribution, marketing, or supply. The results were clear. In the first period, 1985–91, firms were more inclined to form alliances if they were members of the same group. It made no difference whether the alliance was R&D-based or not. In the second panel, 1992–97, the vertical keiretsu constraint disappears from R&D alliances but remains strong on nonR&D alliances. Thus, from the 80s to the 90s a change took place in how Japanese electronics firms organized their innovative partnerships. In the latter period, such alliances were being formed without regard to keiretsu ties. The lesson here is similar to that taken from the Toyota-Denso case. Compared to the 1980s, the strategies of innovation pursued by Japanese firms in the 1990s and beyond have been largely oblivious to if not outright destructive of keiretsu ties.
229
MOT in and between enterprises
Conclusions: Networks and innovation in Japan Product and process innovation in Japan has been bound in a number of ways to the country’s famously thick and convoluted corporate networks. The research reviewed in this paper reveals that bond to be more complex and temporally bound than a casual read of the extensive literature on these topics might suggest. Close customer–supplier ties, as the Hitachi, Matsushita, and spin-off studies show, facilitate cooperation, mutual learning, and synergistic creativity, consistent with the claims of past research. But networks need not operate this way, and in Japan some of those that do were engineered expressly for that purpose. Firms spin off new product divisions as keiretsu satellites in order to maintain a climate of independence, flexibility, and entrepreneurship (higher powered incentives) within which innovative enterprise can flourish. Such strategic crafting of networks is more and more the norm in Japanese business— while the legacy networks of the past (main bank ties, big six horizontal groups, ministry amakudari) increasingly fade away (Lincoln and Gerlach 2004). Moreover, while the trajectories of Japanese innovation have been shaped and channelled by networks, it is equally true that Japanese purchase–supply and other networks have been sculpted by the contours of technology and innovation strategy. Hitachi Omika’s practice of immersing itself in–‘taking the role of’–the customer in order to adapt its products to the customer’s needs and school it in the proper use of them springs in part from the complexity, cost, and custom nature of Omika’s products. To be sure, long-term hand holding by a vendor of complex systems to facilitate a customer’s use of those systems is not uniquely Japanese. Product technology of this sort everywhere constrains transactions to be somewhat close and collaborative. Yet Japan’s strong traditions of customer service and exchange partner trust (Dore 1983; Sako 1992)–augmented by such Japan-specific corporate network institutions as preferential bidding, cross-shareholding, and employee exchange–impart a relational texture to Omika’s purchase-supply transactions that is uncommon elsewhere and is irreducible to technological constraint. The Denso-Toyota case is a lesson in how the dependence of a customer on a supplier’s technological knowhow both conditions and is in turn conditioned by the keiretsu bond. It suggests that keiretsu networks are unravelling and less able to contain interfirm resource dependencies in an era of accelerated technological change and global expansion and competition. Under the ‘old’ Toyota vertical keiretsu regime, where Toyota sourced less from nongroup suppliers and Toyota suppliers sold less to Toyota’s competitors, the risks posed by ‘black box’ reliance on Denso’s electronics capability were manageable. Technological change played a role as well: were the technical complexity and strategic importance of electronics not rising so steeply, Toyota would have been under less pressure to hedge its dependence on Denso by mastering the technology and becoming its own second source.
230
Interfirm networks Globalization also figured in Matsushita’s strategy of organizing its best domestic suppliers in an active learning network, designed to get them more involved at earlier stages in product and process development. Thus, in contrast with Toyota’s shift toward Denso, Matsushita actively sought dependence on some suppliers while crafting the keiretsu infrastructure within which it could manage that dependence. The difference between the Matsushita and Toyota stories is: first, the stage of network evolution. The Toyota Group has long been Japan’s most developed vertical keiretsu, while Matsushita had no supply keiretsu to speak of at the time it launched the kyoei programme. There is, second, a difference of scale, strategic capability, and dependence: despite its origin as Toyota spin-off, Denso was a large company that had successfully diversified its customer base both in and outside Japan and so had substantially lowered its dependence on Toyota. Finally, we conclude from our quantitative analysis of strategic alliances in the electronics industry that, while Japanese interfirm relations built around technology and innovation retain distinctive features, the highly institutionalized vertical and horizontal keiretsu networks are less constraining or facilitating of strategic innovation than they were. In the 1980s keiretsu ties were leveraged in support of new strategic alliances, whether R&D or not. But in the 1990s electronics firms showed a new (and clearly rational) willingness to go outside their historical groups for partners to strategic alliances that had an innovation thrust. There is a good deal of recent evidence that the keiretsu form is disappearing from the Japanese economic scene. Lincoln and Gerlach’s (2004) extensive longitudinal analysis finds both the structure and functioning of these historically important networks weakening substantially since the early 1980s. We (2004: 103, 354) write: Business ties in Japan over most of the last half-century served to cement and symbolize the stable macro-structural commitments that keiretsu represent. They thus presented themselves to firms as fixed constraints on business dealings. The post-bubble network is different: more micro-strategic–i.e., firm- and dyad-specific–less macro-keiretsu. Japanese capitalism remain(s) network capitalism . . . but its fundamentals are changing. In a way uncommon in the past, firms are leveraging relationships to advance their business agendas–seeking partnerships that endow the individual firm or transacting pair with competitive advantage but are less ‘embedded’ in–that is, integrated with and consequential for–the keiretsu group and corporate network as a whole.
The case studies reviewed here do not dispute this statement but they do require some modification of it in the context of innovation management. First, they make clear that the engineering and leveraging of keiretsu networks in the service of Japanese corporate strategy is by no means a new thing. Vertical keiretsu alliances in the globally critical industries of electronics and autos were often moulded and remoulded to advance the innovation agendas
231
MOT in and between enterprises of firms and industries. In dynamic perspective there is no inconsistency between the observations that: a) keiretsu networks may be crafted and harnessed in the service of innovation strategy; and b) they may impose unproductive constraints on firms’ choice of partner and flexibility of action. Like any organizational form, keiretsu alliances can outlive their usefulness and harden into entanglements that firms need to be free of in order to innovate and compete. But as old alliances decay, new ones can and should be created. And, somewhat at odds with Lincoln and Gerlach’s claim that Japan’s new networks are more micro-dyadic than its old ones, some of our evidence–that on Matsushita’s kyoei-kai in particular–reveals that strategically designed macro-networks retain a place in the modernizing Japanese economy and that innovative firms are sometimes well advised to build and use them.
Notes 1. This section follows the disscussion in Gerlach and Lincoln (2001). 2. While Denso’s direct role in Toyota’s electronics learning has been small, Toyota’s keiretsu ties to Daihatsu, Hino, and Kanto Auto Works enabled these companies to acquire capabilities in electronics. Toyota said it had no formal programme to teach electronics to Toyota Group firms, but as those firms were codesigning, supplying, and assembling Toyota products, Toyota must help them with the electronics. Another Toyota affiliate, Aisin Seiki, has moved independently into electronics, producing car navigation systems that competed with Denso’s.
References Ahmadjian, C. and J. Lincoln (2001). ‘Keiretsu, Governance, and Learning: Case studies in change from the Japanese automotive industry’, Organization Science, 12: 683–701. Aoki, M. (1987). ‘The Japanese Firm in Transition’, in K. Yamamura and Y. Yasuba (eds.) The Political Economy of Japan, Vol.1, Stanford, CA: Stanford University Press. Asanuma, B. (1989). ‘Manufacturer–supplier Relationships in Japan and the Concept of Relation-specific Skill’, Journal of the Japanese and International Economies, 3: 1–30. Beer, M. and B. Spector (1981). Matsushita Electric Industrial, Harvard Business School Case #9-481-146. Burt, R. (1992). Structural Holes: The social structure of competition, Cambridge, MA: Harvard University Press. Clark, K. and T. Fujimoto (1991). Product Development Performance, Boston, MA: Harvard Business School Press. Dore, R. (1983). ‘Goodwill and the Spirit of Market Capitalism’, The British Journal of Sociology, 34: 459–82. Dyer, J. and K. Nobeoka (2000). ‘Creating and Managing a High-performance Knowledgesharing Network: The Toyota case’, Strategic Management Journal, 21: 345–67.
232
Interfirm networks Florida, R. and M. Kenney (1990). The Breakthrough Illusion: Corporate America’s failure to move from innovation to mass production, New York: Basic Books. Frame, D. and F. Narin (1990). ‘The United States, Japan and the Changing Technological Balance’, Research Policy, 19: 447–55. Fruin, M. (1992). The Japanese Enterprise System: Competitive strategies and cooperative structures, New York: Oxford University Press. —— (1997). Knowledge Works: Managing intellectual capital at Toshiba, New York: Oxford University Press. Gerlach, M. and J. Lincoln (2001). ‘Economic Organization and Innovation in Japan: Networks, spinoffs, and the creation of enterprise, in I. Nonaka, G. von Krogh, and T. Nishiguchi (eds.) Knowledge Creation: A new source of value, London: Macmillan. Granovetter, M. (1985). ‘Economic Action and Social Structure: A theory of embeddedness’, American Journal of Sociology, 91: 481–510. —— (2004). ‘Business Groups and Social Organization’, in N. Smelser and R. Swedberg (eds.) Handbook of Economic Sociology, 2nd edn., Princeton, NJ: Princeton University Press. Guillot, D. (2001). ‘The Permeability of Network Boundaries: Strategic alliances in the Japanese electronics industry in the 1990s’, American Sociological Association, Anaheim, CA, August. —— and J. Lincoln (2005). ‘Dyad and Network: Models of manufacturer–supplier collaboration in the Japanese TV manufacturing industry’, in A. Bird and T. Roehl (eds.) Advances in International Management: Special Issue on Changing Japan, Greenwich, CT: JAI Press. Gulati, R. (1995). ‘Social Structure and Alliance Formation Patterns: A longitudinal analysis’, Administrative Science Quarterly, 40: 619–52. Hamel, G. and C. Prahalad (1994). Competing for the Future, Boston, MA: Harvard Business School Press. Helper, S. (1991). ‘How Much has Really Changed Between U. S. Automakers and their Suppliers?’ Sloan Management Review, 32: 15–28. Imai, K., I. Nonaka, and H. Takeuchi (1985). ‘Managing the New Product Development Process: How Japanese companies learn and unlearn’, in K. Clark, R. Hayes, and C. Lorenz (eds.) The Uneasy Alliance: Managing the productivity-technology dilemma, Boston, MA: Harvard Business School Press. Lincoln, J. and C. Ahmadjian (2001). ‘Shukko (Employee Transfers) and Tacit Knowledge Exchange in Japanese Supply Networks: The electronics industry case’, in I. Nonaka and T. Nishiguchi (eds.) Knowledge Emergence: Social, technical, and evolutionary dimensions of knowledge creation, New York: Oxford University Press. —— and M. Gerlach (2004). Japan’s Network Economy: Structure, persistence, and change, New York: Cambridge University Press. —— and A. Kalleberg (1990). Culture, Control, and Commitment: A study of work organization and work attitudes in the U.S. and Japan, New York: Cambridge University Press. —— C. Ahmadjian, and E. Mason (1998). ‘Organizational Learning and Purchase–Supply Relations in Japan: Hitachi, Matsushita, and Toyota compared’, California Management Review, 24: 241–64. —— M. Gerlach, and P. Takahashi (1992). ‘Keiretsu Networks in the Japanese Economy: A dyad analysis of intercorporate ties’, American Sociological Review, 57: 561–85.
233
MOT in and between enterprises Mead, G. (1962). Mind, Self, and Society, Chicago: University of Chicago Press. Nakamura, T. (1983). Economic Growth in Prewar Japan, New Haven, CT: Yale University Press. Nishiguchi, T. and A. Beaudet (1998). ‘The Toyota Group and the Aisin Fire’, Sloan Management Review, 40: 49–60. Nonaka, I. and H. Takeuchi (1995). The Knowledge-Creating Company: How Japanese companies create the dynamics of innovation, New York: Oxford University Press. Odaka, K., K. Ono, and F. Adachi (1988). The Automobile Industry in Japan: A study of ancillary firm development, Tokyo: Kinokuniya. Perrow, C. (1984). Normal Accidents: Living with high-risk technologies, New York: Basic Books. Podolny, J. and K. Page (1998). ‘Network Forms of Organization’, Annual Review of Sociology, 24: 57–76. Powell, W. (1990). ‘Neither Market nor Hierarchy: Network forms of organization’, in B. Staw and L. Cummings (eds.) Research in Organizational Behavior, Greenwich, CT: JAI. Ramseyer, J. (1991). ‘Legal Rules in Repeated Deals–Banking in the shadow of defection in Japan’, Journal of Legal Studies, 20: 91–117. Ritschev, D. and R. Cole (2003). ‘The Role of Organizational Discontinuity in High Technology: Insights from a U.S.–Japan comparison’, in J. Bachnik (ed.) Roadblocks on the Information Highway, Lanham, MD: Rowman Littlefield Publishers. Roberts, K. (1990). ‘Managing High Reliability Organizations’, California Management Review, 32: 101–13. Sako, M. (1992). Price, Quality, and Trust: Inter-firm relations in Britain and Japan, Cambridge: Cambridge University Press. Saxenian, A. (1994). Regional Networks: Industrial adaptation in Silicon Valley and Route 128, Cambridge: Harvard University Press. Shimotani, M. (1997). ‘Matsushita Denki ‘Kyoei-kai’ no rekishi to genzo’ (Matsushita Electric’s ‘Kyoei-kai:’ past and present), Kyoto University Economic Review. Shioji, H. (1995). ‘ ‘Itaku’ Automotive Production: An aspect of the development of fullline and wide-selection production by Toyota in the 1960’s’, Kyoto University Economic Review, 65: 19–42. Smitka, M. (1991). Competitive Ties: Subcontracting in the Japanese automotive industry, New York: Columbia University Press. Teece, D. (1986). ‘Profiting from Technological Innovation: Implications for integration, collaboration, licensing, and public policy’, Research Policy, 15: 285–305. Uekusa, M. (1987). ‘Industrial Organization: The 1970s to the present’, in K. Yamamura and Y. Yasuba (eds.) The Political Economy of Japan, Vol.1, Stanford, CA: Stanford University Press. Uzzi, B. (1996). ‘The Sources and Consequences of Embeddedness for the Economic Performance of Organizations’, American Sociological Review, 61: 674–98. Whittaker, D. H. (1997). Small Firms in the Japanese Economy, Cambridge: Cambridge University Press. Williamson, O. (1985). The Economic Institutions of Capitalism, New York: The Free Press. —— (1996). The Mechanisms of Governance, New York: Oxford University Press. Womack, J., D. Jones, and D. Roos (1990). The Machine that Changed the World, New York: Rawson Associates.
234
Part 3 Transforming Japan’s innovation system
This page intentionally left blank
13 Innovation policy for Japan in a new era Tateo Arimoto
With the new century already six years old, it is time to take stock of where the society and economy of Japan are headed after the ‘lost decade’, a decade in which the world at large underwent great changes. In particular, the innovation system which served effectively for the first postwar half-century needs fundamental reform, now that Japan has evolved out of its catch-up phase to being a front runner. But what specifically needs to change? While the post-Cold War international community has seen great changes, Japan has been preoccupied with dealing with the after effects of the bubble, to the neglect of the badly needed structural reforms of its polity, economy, and society, and its innovation system. Many seemed not to look for deeper causes but to assume that once the financial sector had been put back into shape all would be well and Japan could resume its position as a great economic power. But that is clearly not the case. How far the recent recovery of the private sector will take Japan remains uncertain. Meanwhile the Japanese state has an accumulated debt of ¥700 trillion, way above its ¥500 trillion annual GDP. The crisis of the 90s had much more complex roots than the previous two crises– the oil shock in the 70s, and the Plaza Agreement and the revaluation of the yen in the mid-80s–but the preoccupation with the weakness of the banking sector prevented cool, clear analysis of those causes and what should be done about them. It is only recently that people have come to take reform of the innovation system seriously. The major recent changes are: the introduction of new systems of technology management in Japanese firms; the incorporation of former national universities and national laboratories as independent corporations which have to be managed with a view to making them competitive; and emphasis on university–business collaboration and the development of regional clusters (see the Introduction in this volume). But in my view, the fact that Japan has become a technology front runner requires fundamental changes in
237
Transforming Japan’s innovation system the innovation policies that served so well since the war–indeed, since the Meiji Restoration. What is needed above all is the training of talented researchers of high quality and in large numbers, and also policies to seamlessly nurture innovative capacity at every stage of the innovation process, from upstream in basic science to downstream in the marketplace. The next decade, during which world competition will undoubtedly intensify, will be the test: can Japan develop an new innovation system and an environment for it which is capable of continuous development? This chapter is divided into five sections. The first section deals with Japan’s choices in the context of a changing and increasingly competitive world; the second section is concerned with reform in the universities which are responsible for the crucial task of producing scientific and technological talents and seeds of innovation; the third section emphasizes the role of science and technology in responding to social needs, and the increasing importance of technologies which fuse hitherto unrelated disciplines or are of basic importance across disciplines; the fourth section discusses the central issue of the spirit which should animate and the norms which should guide researchers in a front runner country; and the final section deals with communication between the scientific community and society, and the need to build relations of trust.
Choices for Japan in the 21st century All discussion of science and technology policy in Japan must begin with an appreciation of how much the world environment has changed. During the catch-up phase, Japan’s objectives were clear. But now she has to find her own path in a world in which America is the overwhelming leader in advanced science and technology, in which China’s importance as workshop of the world is rapidly increasing and Europe is being revitalised by the enlarged European Union. As compared with five years ago, there is a much greater awareness of the problems, not only among those directly concerned in industry, the universities and government, but also among the public at large. Put simply, during the catch-up period, firms’ strategy focused on quality and cost competition, but that is no longer acceptable for front runners. What is necessary is core technologies and business models of world-class quality, creating value by originality and product differentiation, and sustaining competitiveness not by single innovations but by streams of innovations. For that to happen, we need a national vision and concrete policies. The circumstances are ripe for cool and objective discussion of what those policies should be and what role innovation policies can play in the long-term in making Japan a country which can attract talent, resources, capital, and information from all over the world, and sustain quality living conditions for the next generations.
238
Innovation policy for Japan in a new era
Social change The notion that this century will be the century of the knowledge-based society has become widely accepted over the last five years, with the penetration of information technology and increased discussion of intellectual property. On the one hand the general public has shown a much enhanced concern for safety and awareness of the dark side of technological progress: global warming, the spread of terrorism and infectious diseases such as HIV and bird influenza, shortages of energy and resources, and unusual climatic aberrations have all had their effect. On the other, people look to science for solutions. Scientific knowledge should make it possible to reconcile environmental protection and economic development. Many things cast a shadow over the future of Japan. Company restructuring and social instability during the 15 years which were dominated by the task of dealing with the after-effects of the bubble have brought in their wake a loss of the sense, sustained throughout the postwar period, that Japan was a country with an egalitarian income distribution. Class divisions have become more apparent and the attenuation of a sense of belonging to, and of having moral obligations to, organizations or local communities has spread from older adults to the young. Suicide among men and the rapid increase in incidents involving children point to deep social problems. At the same time, it should be noted that there has been a growth in volunteer work, in efforts to do something positive to build the public sphere at the local level.
Changing ideas and perceptions The 2004 Science and Technology White Paper stressed the increasingly intricate links between science and society. It reported, notably, that 80 percent of respondents to a survey said that the development of science and technology should bring spiritual as well as material benefits. But most worrying was the finding that a professed interest in science and technology was on the wane; compared with five years ago the proportion who claimed an interest fell by five percentage points among the general population, and by 10 points among people in their 20s. People think more about the long-term prospects for the country, if only because of the expected onset of population decline beginning in 2005, thanks to the fall in the birth rate, and the problem that poses for the pension system. Some forecasts have GDP itself declining persistently from 2009. How to maintain the standard of living of future generations, both materially and culturally, is the greatest challenge to our science and technology capability. If, as forecast, the population does peak at 127 million in 2005 and declines to 90–100 million by 2050 (Figure 13.1), that means also a rapid decline in the
239
Transforming Japan’s innovation system
Population (million)
150
Aged 65+ 15−64 0−14
100
50
0
1960 1970 1980 1990 2000 2010 2020 2030 2040 2050 Year
Figure 13.1 Projected population decrease in Japan Source : Compiled from National Institute of Population and Social Security Research Institute figures
number of scientific researchers and engineers, if the system of recruitment and training remains unchanged (Figure13.2). More of that later.
Change in scientific methodology The four century history of modern science has been one of progressive specialization. Specialization was appropriate for detailed understanding of nature and for its transformation, but it now threatens the continued existence of mankind. Doubts about the specialization of knowledge have been expressed from the last half of the 20th century, and continue to grow. What is required for the continued survival and prosperity of mankind and the solution of its problems is a change of direction towards the unification of knowledge. In the fusion of information technology, biotechnology, and nanotechnology we can look forward to the development of promising new scientific fields and the creation of economic and public values (Figure 13.3). And if the 21st century is indeed to be the century of the life sciences, it is necessary to mobilize a whole
240
Innovation policy for Japan in a new era 2.6 Number of scientific researchers and engineers Ratio of researchers/engineers in general population
2.
2.8
2.
2.4 Million
% 2.
2.0
1.
1.6
1.
2004
~ Year
2050
Figure 13.2 Projected decrease of researchers Source: Cabinet Office (2005). Keizai Zaisei hakusho (White Paper on Economic and Fiscal Policy), p. 289
range of disciplines–mathematics, physics, chemistry, biology, medicine, computer science, engineering, and the human and social sciences.
Organizational changes affecting science and technology The foremost of the big changes that have occurred in the organizational structure of science and technology was the creation, in the administrative reforms of 2001, of the Council for Science and Technology Policy as one of two major councils within the Cabinet Office, chaired by the Prime Minister. (The other is the Council for Economic and Fiscal Policy.) It has oversight over the whole field of S&T policy. At the same time the merger of the former Ministry of Education and Culture with the Science and Technology Agency created a ministry (the Ministry of Education, Culture, Sports, Science and Technology–MEXT) with competence over the whole field of science training, from primary school to the post-doctoral level, and administrative responsibilities covering the major part of the innovation cycle from basic science to university–industry collaboration.
241
Transforming Japan’s innovation system
BT
• Theory • Experiment • Computer science
Biology, medicine
modeling prediction visualization
New promising field Creation of new values
NT Physics, chemistry, and material science
IT Mathematics, information, and communication
Figure 13.3 Convergence of disciplines
Researcher organizations have also undergone change. Reform of the Science Council of Japan is currently under way. The Council was launched in its new form in October 2005, no longer a lobbying body, but a vehicle for bottom–up representations on policy, thereby performing a vital check and balance function in policy formation. It should, too, help to give Japanese science and technology dignity and cultural maturity. A further big change was the transformation of the national universities, as from April 2004, into independent corporations, the most fundamental change in a system basically unaltered for a century, and one expected to have positive effects for science and technology. The same corporate form has been given to a large number of state laboratories and special public corporations with a research mission, posing for all the new corporations the challenge to develop efficient management and deliver performance. The implication of all these changes over the last five years should be profound, and will determine the success of efforts to sustain development over the next half-century. Big changes are under way in the formulation of S&T policy, and in university–industry collaboration, which has got well beyond the stage of wary encounters to the evolution of systematic and ruleordered contracting. There are similar advances in intellectual property strategies and the creation of regional clusters. We expect the greater freedom of the new corporate universities to accelerate all these trends.
242
Innovation policy for Japan in a new era
The goals of national policy The goals of national policy need to be thought out clearly, with a 30-year perspective. After the end of the Cold War, ordinary Japanese citizens see a world of great uncertainty. They live in what was once a growth society and now is a stagnating society. Income inequality grows, and there is a fragmentation of aims and values. What is needed for a real rebirth of Japan is a clear long-term vision of common values and goals to which all can subscribe, not just innovation in the economy and industry but something that will provide the nation with a centripetal focus, something with spiritual and cultural as well as economic promise; bringing together intellectual and cultural, economic and industrial, social and public values, a vision of a secure country at peace and trusted and respected among nations. That international dimension should not be forgotten; to contribute to the establishment of an East Asian community prefigured at the 2004 ASEANþ3 Summit should be an important Japanese objective. In short, I believe we must redefine the objectives of our innovation policy: we must realize that there is soft as well as hard power, and conjoin to our economic priorities considerations of culture, tradition, ethics, and the public welfare.
Human resources, basic research, and the universities With a wide spread of the lower slopes, broad and solid middle slopes, and high peaks: the Mount Fuji metaphor is useful in thinking about the cultivation of scientific talents. Figure 13.1 shows an estimate of the likely decline in researcher numbers after the Japanese population peaks in 2005. To minimize the decline at all levels and to secure the quality and quantity of human resources, we need first to improve access to the scientific talent market–for women, older people, and foreigners as well as for young researchers–and, second, we need to reform the education system to ensure that young students are attracted to maths and science. Researchers typically follow a variety of career paths as shown in Figure 13.4. Hitherto university science education has been too much directed towards training of academic researchers, but already there are initiatives to give greater importance to training for a wider range of professions essential for innovation: general management, management of technology, intellectual property protection, journalism, etc. Special courses are being developed for these careers, and the Nippon Keidanren and the MEXT are promoting extended internship schemes for students at undergraduate and graduate level. Already the amalgamation of the former Ministry of Education and the STA is proving fruitful in the improvement of science education in schools, in
243
Transforming Japan’s innovation system • Science policy head • Industry CEO
• University president • Faculty member
• Science communication: journalist, editor, producer
• Independent investigator POSTDOCTORATE
• Gov., univ., institution administration
DOCTORATE
• MOT, law, bioethics, patent management
MASTERS SCIENCE DEGREE
• Lab manager, staff, technician
FIRST DEGREE SCIENCE MAJOR
Non-academic SCIENCE OUTREACH • Public literacy programme, media, museum exhibits
• Research team member • University, professional school teaching • Professional schools • School teacher
Academic PRE-UNIVERSITY SECONDARY ELEMENTARY
SCIENCE EDUCATION • Science enrichment • Science curriculum development
Figure 13.4 The pipeline and the tree: a new framework for training and career development in the natural sciences Source : MEXT (2004), Kagaku gijutsu hakusho (White Paper on Science and Technology), p. 104
strengthening links between high schools and universities, and in improving university curricula. International competition to attract scientific talent is increasing. Not only the advanced industrial countries, but countries such as China, Singapore, Thailand, and Malaysia are redoubling their efforts to bring back home their talented nationals trained in the advanced countries. From the other end, the American National Science Board in a 2002 report remarked on the way in which the influx of the world’s best brains over the last half-century had helped to build the basis for America’s prosperity, and warned that greater efforts to develop native born talent were needed, now that America could no longer profit from the brain drain from the rest of the world. The new century calls for a commitment of science and technology, not only to the production of new knowledge, but also to the application and regulation of knowledge. To ensure that Japan’s S&T has the underlying strength provided by the trust and support of the general public, we need to develop satisfactory and rewarding career systems for all the other professionals who
244
Innovation policy for Japan in a new era can contribute to that end: science communicators, university administrators, bioethics specialists, science policy managers and staff.
Basic research for front runners Japan’s successful innovation model for the catch-up period was clear: the seed of technology were imported. We added value to that core with improvements and competed in international markets on quality and cost through process innovation. Now, with America in front and China on our coat-tails we need to be able to generate that steady stream of innovations that can give world uniqueness in core fields. That is where basic science comes in. History shows that catch-up countries, in order to get over trade and technology friction and build the basis for sustained competitiveness, require full cooperation among industry, universities, and government in basic research capabilities. It was thus in Germany and the United States challenging the supremacy of Britain at the turn of the last century. The Imperial Physical and Engineering Laboratory and the Kaiser Wilhelm Institutes in the one, and the National Bureau of Standards, the Carnegie Foundation and the Rockefeller Institute, the National Institutes of Health, and the research grant and contract system in the other, were essential instruments. Japan, too, experienced frictions with the US in the 1980s. As for the response, there were moves at the time of the bubble for the private sector to provide the funds for strengthening Japan’s research base, but they came to little as the bubble burst. Now that the after effects of the bubble have been cleared away, the time is ripe for the cooperative effort to begin. Michael Porter (1990) speaks of universities as of daily increasing importance in innovation in the advanced industrial countries. Universities are increasingly involved in international competition in which their success (in education, in research and in contributions to society) is directly linked to their ability to enhance and revitalize their nations and regions and firms. The pace of reform in Japanese universities–their incorporation, the establishment of graduate law and other professional schools, the creation of evaluation systems–is something that could not have been imagined five years ago. The next five years will be the test of the system as Japan’s 700 universities compete and differentiate their functions (Figure 13.5). In the catch-up phase, what industry asked of the universities was, above all, to produce large numbers of averagely able students capable of cooperative endeavour. Core technologies they got from abroad. Moreover, firms themselves could afford the resources to continue graduate level education after employment. The limits of enterprise self-sufficiency in training and research have been reached in the world of intensified international competition. Firms have allowed their training functions to run down. Now, the responsibility for developing talent lies with the universities, and the
245
Transforming Japan’s innovation system
Knowledge-based society; global competition among universities; Japan now front runner The functions of universities • World-class research and education • Training high level professional manpower • Training for a broad range of occupations • Comprehensive cultural education • Special fields: arts and physical education • Links to community and local industry • Adult extension education
Values for society
Management strategy for universities Making universities attractive in new era of university corporations: • Functional specialization • Characteristic identity Organization, finance, talent recruitment, new attitudes, management performance Consolidation of university resources Cross-disciplinary centres and platforms Intellectual property, industry, community, and international collaboration
→ Training scientific manpower: quality, quantity, diversity
Education
→ Coordination of science education: school to university
→ Contribution to culture and civilization
→ Contribution to public welfare: Research
policy advice, security
→ Economic contribution: collaboration with industry, technology diffusion, retraining engineers
Links To society
Inter-university links
→ Links to local community: intellectual hub, public sphere environment, disaster prevention, welfare, culture, scientific communication
[Universities for knowledge] + [Universities in and for society] [Creation and transmission of knowledge]
[Utilization and control of knowledge]
Figure 13.5 Universities, society, and S&T in the 21st century
talents required are above all creativity, a broad knowledge base and flexible openness to ideas. Industry–university collaboration has been stressed in recent years, but this trend is criticized by some industry leaders on the grounds that universities have become overly concerned with short-term research and development expected to bring results in two or three years. Quite wrong, they say. What the universities should concentrate on is basic research with a ten to 15 year horizon. That is why the importance of universities is increasing by the day.
The deepening relation between universities and society–university social responsibility In an age of open access, with half of each age group attending universities, and universities adding service to their local community to their traditional teaching and research missions, functional specialization is inevitable. Universities must create their own special colour and characteristics, create their own special niche, and their contributions to the local economy, environmental protection, disaster prevention, culture, and public services will offer fruitful fields for such specialization. For enterprises, too, in a context in which the expansion of domestic demand becomes all important, positive involvement in the creation of social value can be not only a means of fulfilling corporate
246
Innovation policy for Japan in a new era social responsibilities, but also of enriching their creativeness. Cooperation between firms and universities in developing courses for the re-education and continuing education of their employees can be another positive factor in the evolution of universities themselves.
The opportunity The transformation of universities into independent bodies should have a big impact on the innovation system. At present the new style universities are feeling their way, working out their own individual organizational, personnel, and financial management systems, not forgetting as they seek to establish their own distinctive identities the importance of raising the quality, career prospects, and sense of involvement, not just of teachers and researchers, but also of administrators and research support staff. The new style universities can be expected to intensify their collaborations with industry, but it is important to recognize both that university and enterprise differ both in mission and management style, and that bridging the two cultures is necessary. For that purpose what is needed is business–university round tables, both at national and regional levels, for the promotion of mutual understanding and trust, and for cooperative action. In order to create cultural, economic, and public values by science and technology, with more research becoming cross-disciplinary and more requiring large-scale equipment, there is a need for a restructuring of the national innovation system and the governmental research grant portfolio in order to plan appropriate allocations to the various stages of basic, applied, and development research, and to give proper weight to high risk and exploratory technology development and large-scale projects.
Science for society The World Science Conference held in Budapest in 1999 under the sponsorship of UNESCO and the International Council for Science (ICSU) brought together some 2000 researchers, administrators and journalists from all over the world. Its week-long deliberations on the general theme ‘New commitments for science in the 21st century’ resulted in the so-called Budapest Declaration which urges commitment not only to 20th century notions of science as the pursuit of knowledge–‘science for knowledge’–but also ‘science for peace’, ‘science for sustainable development’, and ‘science in and for society’. Its preamble stresses that scientific and technological knowledge must always be used responsibly and never abusively, a timely reminder of our responsibilities to future generations.
247
Transforming Japan’s innovation system When Japan set out its second Basic Plan for Science and Technology in 2001, there was a strong push for policies that would produce quick short-term solutions to the prolonged economic stagnation which dominated thinking at the time. Now, from industry too, we hear more and stronger pleas for emphasis on the long-term build-up of innovation capacity. The new Basic Plan period, beginning in 2006 will have to give greater emphasis to quality of life issues, security and cultural fulfilment (see Murayama’s chapter in this volume), not least because they are important elements in the expansion of domestic demand and the improvement in productivity in a period when the population is rapidly ageing. It will have to give emphasis not only to discrete technologies in various fields–information, biotechnology, nanotechnology– but also to the combination of technologies to build social systems in the fields of medicine, energy, communications, transport, environment, and security. Sustainable development for all mankind is the challenge of the 21st century. Japan tended to be inward looking during its lost decade. Now it needs to mobilize its resources and experience for that international task, building intellectual networks particularly in Asia, seeking to gain international dignity and trust on the basis of its own sustained progress. Prime Minister Koizumi gave a strong speech at the Science and Technology in Society Forum in Kyoto, November 2004, in which he insisted that environmental protection and economic development could be made compatible by science and technology capability. But scientific knowledge alone is not enough to solve mankind’s problems. We also need insights from the humanities and social sciences. And for that we need a large-scale dynamic vision.
Setting priorities The strategic priorities for resource allocation to the four priority fields of bioscience, IT, nano-science and environment in the second Basic Plan for Science and Technology (2001–05) have played a big part in ensuring the strategic character of funding and concentrating research capacities. At the same time there have been criticisms. It is alleged, for instance, that in bio-science and communications research where the ‘lower slopes’ offer a wide range of promising research topics, the concentration of resources on a few of them may produce short-term results, but impoverish the development of the field as a whole. The proper balance between concentration and diversity must clearly be a consideration for the next plan. The priority funding of the last five years has made it possible to construct projects with a clear sight of the technology developments and the social outcomes which can be expected in their fields. This highly desirable development requires cooperation among industry, universities and government, with a clear division of labour.
248
Innovation policy for Japan in a new era Another emphasis is on the fusion, rather than the translateral linking, of scientific disciplines. As research agendas develop, methodologies and research styles change. With the completion of the human genome project and great advances in analytical equipment and computing, for example, bio-science has moved from being a qualitative science to being a quantitative and precise science over the past decade. Structures can be modelled, their functions predicted and even visualized in what is, in effect a fusion of information technology, biotechnology, and nanotechnology, which promises to be the source of much future innovation. Japan was late in getting started on human genome research, but its strengths in IT and nanotechnology should ensure that in protein and cell research, and their applications, it operates at world class levels. The concept of ‘critical technology’ is recently emerging from the Council for Science and Technology Policy and is likely to form an important strategic perspective in future. It has to do with comprehensive national security, the projection of soft power, and the maintenance of international competitiveness. Examples might be the high-end supercomputer, space technology, earth observation system and new radiation sources which exploit Japanese leading edge expertise in sensors, optics, and semiconductors. It could involve not only research to develop the separate elements, but also the development of large-scale research facilities, setting up a virtuous cycle of development, procurement and operation, providing shared access for both industry and academia and producing a continuous stream of innovation.
The right mind-set Realizing that catch-up has ended The catch-up recipe guided Japan’s innovation policy for 150 years. Though it ceased to function, Japanese people, having lost their self-confidence, continued to import American or British methods. But there is room for variety in capitalisms and innovation systems (Hall and Soskice 2001). As Michael Porter (1990) says, competitive advantage is created and sustained by national characteristics; differences in economic structures, value systems, cultures, institutions, and history all contribute to competitive success. We need to think coolly about what of the systems we have built we need to preserve, what to reform (Dore 2004). The first thing to get straight is, what are the differences between catch-up and being a front runner (Figure 13.6). Because they could import their technologies in already half-complete form, enterprises did not have to face the problem of choosing among high-risk and exploratory technology development projects. The universities, too, besides producing educated manpower, have conducted basic research within an ivory tower. They did not have to experience the ‘valley
249
Transforming Japan’s innovation system
20th century Century of catch-up Imported technologybased innovation
21st century Century of mega-competition Original science-based innovation Original technology at top world levels
Research level
* Continuous innovation
Imported technology Homegrown technology Core technology, basic and peripheral patents, production technology
Originality differentiated speciality
* Sustained competitiveness
Own core technology + open innovation + intellectual property regime,
management, markets Diversity of basic research: Advanced research facilities * Raising the technology level: fusion of fields cost and speed * Innovation eco-system: talents, info, capital, regulation, tax, etc. from basic research to market * State and local innovation system, hot spots, clusters * Industry, university, government collaboration
Figure 13.6 Innovation system for Japan after catch-up
of death’ phase when projects seem to be getting nowhere. As a consequence, as we saw over the distribution of funds at the time of the second five-year Basic Plan, there was a straight zero-sum struggle between basic research and industry. But front runners cannot afford such zero-sum games. They need both–in complementarity. And firms, too, need a new model to replace the quality/cost emphasis, now that product quality has generally risen across the board and costs have been reduced to the minimum. Industry–university collaboration today extends across frontiers. Talent, technology, intellectual property, and university basic research can be bought in world markets. Yet, long-term national security and competitiveness requires the retention of national capacity to do sustained original research. And that requires a comprehensive innovation system seamlessly geared to every stage of the innovation process and encompassing not only the funding system itself but patent policy and tax measures, deregulation measures, the training of R&D managers and a stepping up of industry–university collaboration. In the comprehensive system that we need, state support should be provided not simply in the flow of research money, but also in the provision for common use of the most advanced facilities for computing, measurement and analysis, and databases, to provide a solid stock of research capital. The notion of regional clusters is primarily aimed at regions, some hitherto highly dependent on public works construction projects, seeking to wean
250
Innovation policy for Japan in a new era them off such funds and revitalize them with science and technology. Unlike earlier efforts to attract large firms, this is an ambitious project which relies on collaboration between local firms and universities, and is aimed at developing industries which can be internationally competitive. There are already some success stories of the marrying of local traditional techniques, cultures and skills to university-derived advanced technology. Such self-reliant efforts promise to prevent the hollowing-out of local areas and to create attractive places to live and invest.
A world class research base Dr Tonegawa, the Nobel Prize winner has said to us: ‘however hard you try, if you’re doing research with equipment ordered out of a catalogue you are doing second-class research’. The American presidential science advisor Dr John Marberger III spoke in a recent visit to Japan of research funding priorities: first equipment and computers, then talent, and third ideas. The US Department of Energy attracted attention with its announcement of the 28 research facilities that are going to be important over the next 20 years (‘Facilities for the Future of Science–A Twenty-Year Outlook’ 10 Nov. 2004, U.S. DOE). The National Institutes of Health, likewise has published a road map showing the equipment needed and the necessary collaboration with physicists. World unique research facilities have played a big part in the creation of new innovatory fields of research. Japan has the Subaru Observatory, Super-Kamiokande, the Spring 8, the Earth Simulator, all of world class level and producing a stream of world top research and, indeed, Nobel Prizes. Such facilities not only produce academic research; their creation also induces a good deal of unique innovation, of which we can expect more in the future.
The researcher’s moral stance Reform of systems and institutions can induce change from outside, but it is only the individual scientist or technologist who can actually plan and guide reform of the innovation process in any particular piece of research. Quantitative controls of funds and personnel can be imposed from the outside, but what counts for the quality of ideas and outcomes is the spirit that moves and the norms that guide the individual researcher. To do independent and original work, researchers need to establish their own firm set of beliefs about science and technology and human nature and to be in an environment where they can test them against the beliefs of others. In other words they need to see what they are doing in the light of a sense of responsibility towards society and future generations and a sense of history. Qualitative breakthroughs come from researchers who work from a base of high ideals and values and self-propelled initiative.
251
Transforming Japan’s innovation system The London Economist published a long article in 1986 about the Japanese– American high-tech war, then at its height. It concluded: ‘The differences in style mirror the differences in ideas that the two peoples hold dear. The Japanese have a saying: ‘‘The nail that stands up will be hammered flat.’’ The Americans say : ‘‘Let the daisies grow.’’ . . . And that puts the advantages decidedly on the side of Yankee ingenuity’ (‘The high-tech titans’, Economist, 23 August 1986; cf. also Byosiere’s chapter, this volume). Twenty years later we can see that this prophecy was correct. It is clear that if we want a steady flow of original innovations we must make sure that ‘nails that stick out don’t get knocked; develop your strong points’ culture is fostered in Japanese society as a whole and in the research community in particular. The relation of trust and confidence between the research community and the rest of society is also important. Consider the central themes of the last three Davos conferences: Leadership in Fragile Times; Building Trust; Partnering Security and Prosperity. That succession of titles in a period of ever intensifying competitive marketism sends us back to the thoughts of some of the founding fathers of modern economics–Adam Smith’s assertion that morality is at the root of economic prosperity, Keynes arguments that the state of confidence of entrepreneurs and consumers is the crucial factor in economic stability and growth. Moral sense, confidence, trust–these are all accumulated in society over long periods of time. Their importance for society as a whole is mirrored by their importance for the community of researchers. It is that trust and confidence underwritten by the unique history and cultural traditions of their country that researchers need, now that they have to make their own way in a front runner country, and innovate into the unknown. The last century and a half of our modernization history has been a story of a marriage between imported Western modern science and technology and the scholarship and technology indigenous to Japan and mediated by the developments of the Edo period. Our researchers need to be fully aware, both of that indigenous tradition and of the strenuous efforts and the great achievement of their predecessors such as S. Kitasato, H. Nagaoka, S. Toyota, H. Noguchi, J. Takamine, and H. Yukawa who saw those two traditions melded through a constant process of conflict and conciliation. This should provide them with the inspiration, the energy, and the guiding compass that will enable them, in this age of intensifying competition, to gain the support of the general public by truly substantial innovative achievements.
Public understanding of science The need for communication That support is important. Japan has been backward in fostering close communication between researchers and society at large. As last year’s Science and
252
Innovation policy for Japan in a new era Technology White Paper analysed in detail, the events of the last 15 years have only intensified the impact of science on society, yet at the same time, as mentioned earlier, public interest in science is declining. The agenda to be dealt with has no limits–the environment and bioethics, food and drug safety, abuse of the Internet, science education in schools and scientific literacy, the development of scientific journalism . . . Mutual dialogue is necessary if scientific innovations are to be accepted, engage, and activate the general public. The Science Council of Japan held a symposium in May of 2004 on the topic ‘Fostering public understanding of, and sympathy for, science and technology’. This sparked off a series of such meetings throughout the country–dialogues with citizens, students, and teachers, lectures in which scientists recounted their experiences. The Nippon Keidanren issued a guideline paper: ‘The role that industry should play in furthering the understanding of industrial technology’. Discussions between businessmen, including top managers, and citizen or student groups have been increasing.
From understanding to resonance and support We have our 100 famous mountains and our 100 famous lakes, but the scientists and engineers who have contributed to modernizing Japan over the last 150 years tend to remain faceless, without any similar attempts to make them known. It is a matter for celebration that the National Science Museum is planning a series of exhibitions around Japanese scientists and engineers, and one hopes that it will make a big impression on young people and their parents who go to it. But it is also important to provide opportunities for the general public to personally meet scientists and engineers based in their local area. A number of initiatives are under way in Japan in connection with the UN’s World Year of Physics celebrating the centenary of Einstein’s theory of relativity. There are also moves to send Japanese highschool students to the International Science Olympics for maths, chemistry, physics, and biology. It is hoped that all these developments will serve to enlarge the ‘catchment area’ for science among the younger generation and give people a proper sense of science and technology, a sense that as Japanese citizens they share in Japan’s needs and problems and that they can draw strength from Japan’s traditions.
Conclusion I cannot do better by way of conclusion than quote from the 2001 Basic Plan for Science and Technology: As Japan commenced its modernization efforts earlier than any other countries outside of the West, it has gained long experience in harmonizing modern S&T and traditional
253
Transforming Japan’s innovation system cultures and knowledge. Japan should use this experience to help create an environment in which the various peoples of the world, particularly in developing countries, can thoroughly enjoy the benefits of modern S&T while maintaining their own cultures and value systems. (Kagaku gijutsu kihon-keikaku, Cabinet decision 30 March 2001)
Neither should the spiritual and public, as well as the purely economic, goals be neglected in science and technology policy, nor should we forget the importance of respecting other cultures and contributing to international development–of particular importance in the Asian region where, as the EU example shows, science and technology will have a big role to play in moves towards the establishment of an East Asian community agreed upon at last year’s ASEANþ3 Summit meeting. At last there are signs of recovery from the after effects of the bubble, but this is a recovery which leaves us with a huge public debt. To find the funds for the full reform and strengthening of our innovation system, not just understanding but sympathetic support from the general public is essential.
References Dore, R. (2004). ‘Nihon-teki keiei no nani ga nokoruka’ (What will be Left of JapaneseStyle Management?), Gakushi-kai kaiho, VI, Tokyo. Hall, P. and D. Soskice (2001). Varieties of Capitalism: The institutional foundations of comparative advantage, Oxford: Oxford University Press. Keynes, J. M. (1971–79). The Collected Writings of J. M.Keynes (XX), D. E. Moggridge (ed.), London: Macmillan, for the Royal Economic Society. Porter, M. (1990). The Competitive Advantage of Nations, London: Macmillan. Smith, A. ([1776] 1937). An Inquiry Into the Nature and Causes of the Wealth of Nations (Book 5, Part 4), New York: The Modern Library.
254
14 Security and techno-systems: A comparative analysis Yuzo Murayama
Various attempts have been made by scholars interested in national systems of innovation (NSI) to identify explicitly the dimensions in which national patterns differ. Charles Edquist1 offers a list of seven sources of variation in innovation patterns: 1 2 3 4 5 6
Affected by the learning process Holistic and interdisciplinary Affected by history Achieved thorough nonoptimal (or nonequilibrium) processes Affected by the interaction between firms, universities, and government Show a coincidence between organizational and product technology innovations 7 affected by other institutions National security factors have hitherto played little part in the study of NSI. Yet, taking only the dimensions listed 2, 3, 5, and 7 in the above list, it is obvious that the military sector must play a part. No comparison between NSI in Japan and North Korea, for instance, could miss the fact that the development of nuclear weapons and missiles have been the main driving force in the latter, whereas military technologies have received only a limited share of resources in Japan, with the bulk going to commercial technologies. This analysis of what I shall call the Japanese techno-system, with special emphasis on the security dimensions, attempts to fill this gap by looking at the attitudes and consequent policies of Japanese government and industry towards security. The analysis is carried out in comparison with the innovation system of the United States, both in order to clarify the distinct characteristics of Japanese NSI, and to aid in understanding the impact of technology transfer to Japan and technology conflict between the two countries.
255
Transforming Japan’s innovation system The main focus of this study of techno-systems is on relations between government, industry, and academia, with government–for obvious reasons in the field of security–taking the lead, and with the government’s relations to industry and universities being an important determinant of the system. The emphasis is on the high-tech sector, mainly the semiconductor, computer, and telecommunications industries. In the following section, the techno-systems of the United States and Japan during the Cold War period are explained and compared. In the second section, technology-related friction between the two countries arising in the 1980s is described and the reasons are interpreted from a techno-system perspective. The third section deals with the transformation of the United States techno-system in the 1990s and 2000s, and in the final section the emerging techno-system in Japan is examined and its significance and implications are discussed.
US and Japanese techno-systems during the Cold War The US techno-system The role played by such military innovations as radar, proximity fuses, solid fuel rockets, and the atomic bomb in World War II gave rise to the conviction in the US that military-related R&D should be continued even during peacetime,2 and this conviction was strengthened by the onset of the Cold War. Huge sums, a large proportion of them on the Department of Defence budget, were allocated to R&D funding for military research. The total R&D spend for 1969 (civilian and military) was US$25.6 billion (60 percent of it government expenditure) compared with only US$11.3 billion for West Germany, France, Great Britain, and Japan combined.3 Three critical new technologies that were to change the landscape of postwar military technology were invented during the same period. They were the jet engine (the first jet airplane flew in 1939 in Germany and the first flight in the US was in 1942, copying a British model), the digital computer (the ‘electronic numerical integrator and computer’ ENIAC was completed in 1945) and the semiconductor (invented in 1947 by Bell Laboratory scientists). Although these new technologies had great potential, they were too risky and costly for private companies to commercialize alone. Government funding for military development filled the gap, with the companies which received the funds moving to commercial developments later, and universities such as MIT and Stanford using their large military contracts to attract top quality faculty and graduate students, thus developing into prestigious technological universities.4 The interest of government in national security, of private companies in developing new technologies, and of universities in raising their status,
256
Security and techno-systems came together through military R&D funding to form a unique technosystem. This system nurtured and developed such military technologies as the ICBM, jet fighters, and surveillance satellites, and contributed greatly to the technological superiority that the United States maintained during the Cold War. It also produced the seeds of new commercial technologies in such fields as telecommunications and computers, and laid the basis for high-tech industrial clusters such as Route 128 in Boston and Silicon Valley near San Francisco. It was a techno-system which conformed to what Harvey Brooks called the ‘linear model of innovation’,5 and it covered the whole chain–basic research, applied research, development, design, production, and marketing. There was a smooth progression from military to commercial applications; the majority of companies that obtained military contracts intended to use them for commercial applications. What was atypical was that the driving force of the system was the security threat from the Soviet Union and the huge volume of government funding.
The Japanese techno-system The Japanese technological community was devastated by World War II, after suffering from a decade of isolation from developments in the United States and Europe before and during the war. The urgent postwar priority was to revive Japanese technology, both to undertake reconstruction and to catch up with the US and Europe. The chief means was imports, informed and supplemented by visits to the United States, and study of the technological and scientific literature. In the 1950s and 1960s the United States was willing to transfer at least its second-rate technologies. By the 1970s, computer and semiconductor technologies had developed to the extent that the future course of technological development became easier to predict, thanks to IBM’s dominance in mainframe computers and the DRAM’s appearance as a technology driver. These technology trends helped Japanese industry to accelerate its technological catch-up. The Ministry of International Trade and Industry (MITI) was one of the major players in the Japanese techno-system. It cooperated with the technological community to identify potentially promising technologies and nurture them through R&D subsidies, tax breaks, and low-interest financing.6 Hightech companies, too, made massive investments in what they had identified as promising technologies. The emphasis was on applied research and development, taking off from basic technologies imported from abroad. Companies funded 75–80 percent of the R&D effort, while MITI was responsible for the diffusion of promising technologies. One consequence was that all the major high-tech companies tended to develop similar technologies, and as a result competition in specific technologies became extremely severe.
257
Transforming Japan’s innovation system Compared with MITI, the Japan Defense Agency (JDA) made only a minor contribution. The main priority for Japanese technological development was economic growth, not national security, and defense took only 5–15 percent of the R&D spend rather than approximately half as in the US. Moreover, even the JDA took the possibility of commercial spin-offs into account in commissioning defense research.7 Japanese defense production was carried out in dedicated departments of large companies such as Mitsubishi Heavy Industry and Ishikawajima Harima Industries. Arms exports were banned so defense production lots tended to be small. This, and the commercial emphasis, favored the integration of defense and commercial production, as Samuels argued.8 Licensed production of USdesigned weapons provided essential learning opportunities for the Japanese defense industry, and also had commercial spin-offs. The role of Japanese universities was in marked contrast to that of their American counterparts. There was no JDA funding for universities, no defense research contracts. Even in the training of scientists and engineers, their role was eclipsed by that of the large companies, though the universities provided the foundation training for–and through the system of stiff entrance examinations for the high-prestige universities identified the most talented among– large numbers of engineers working in the applied research and development areas to absorb and improve on imported foreign technologies. To sum up the contrast with the US techno-system, the driving force was economic growth not national security. The dividing line between military and commercial technologies was blurred and technology flowed easily between the military and commercial sectors. As a catch-up country Japan did not need the whole linear model; it could skip basic research and concentrate on applied research and development. A salient characteristic was the consensual identification of promising technologies orchestrated by MITI, which led to similarity in the research programs of major companies and intense competition among them, a feature which helped to make them extremely competitive in the world market.
Interpreting technology friction from the techno-system perspective The differing nature of their techno-systems and the changing linkages between economics and national security affected overall technological relations between Japan and the United States. Specifically, the reliance on imported basic technology and concentration on down-stream commercialization and the orchestrating role of MITI in the Japanese techno-system became sources of friction. This friction intensified during the 1980s, and by the early 1990s had become one of the most sensitive issues in overall Japan–US relations.
258
Security and techno-systems
A brief history of friction in the high-tech sector Friction over dual-use technologies between the United States and Japan developed for the first time in the early 1980s. In 1981 when AT&T invited an open bid for construction of sections of a fiber optic transmission line between Boston and Richmond, Fujitsu’s bid was the lowest among eight companies and was expected to win the contract. However, Fujitsu’s bid was rejected and Western Electric, a subsidiary of AT&T, was awarded the contract, reportedly on national security grounds.9 This was the first in a series of incidents which arose from an American concern to keep sensitive technologies out of Japanese hands, first because they might be exported to the Eastern bloc, and second because Japanese loyalties allegedly could not be counted on in case of emergency mobilization of the American war machine.10 In 1983 when Dexel, Kyocera’s US subsidiary, started to produce electronic products for military uses, the Department of Defense put pressure on Kyocera to replace the Japanese with American managers. Kyocera finally sold the subsidiary to an American Company.11 The claim of a threat to national security was often a cover for, or mixed with, fear of, Japanese industrial competition. In March 1983 the US National Machine Tool Builder’s Association (NMTBA) petitioned the Secretary of Commerce claiming that imports of Japanese-made machine tools had weakened the US industry to the extent that the US machine tool industry could no longer meet the national security needs of the country. The Department of Commerce initiated an investigation, which took a new turn in 1985 when the Pentagon started to express concern about the dependency of machine tool technology on Japan, moving the issue from production capacity to research and development capacity.12 In October 1986, moreover, Fujitsu was poised to acquire Fairchild Semiconductor. The combination of Fujitsu, which was trying to establish a solid basis for its US operations, and Fairchild whose profitability was under pressure and whose parent company, French-owned Schlumberger, was looking for a buyer, made good sense from a business point of view. Congress and other administrative branches such as the DoD, the Department of Commerce, and the US Trade Representative (USTR), however, opposed the acquisition on national security grounds. They argued that if Fairchild were to be acquired by Fujitsu, the US military would have to depend on Japanese sources for such sensitive semiconductors as emitter-coupled logic devices.13 Due to the politicization of the issue, Fujitsu abandoned the acquisition in March 1987. Responding to the Fujitsu-Fairchild incident, which exposed the inability of the US legal framework to block corporate acquisitions with national security implications, the US Congress passed the Exon-Florio Clause in the Omnibus Trade and Competitiveness Act of 1988. This clause increased the responsibilities of the Committee on Foreign Investment in the United States (CFIUS) and
259
Transforming Japan’s innovation system gave the President authority to take action to suspend or prohibit foreign acquisitions or investments when national security was considered to be threatened.14 Friction also arose over the transfer of military technology in which the point at issue was not security, but commercial competition between the US and Japan. The first incident of this kind occurred when the Japanese government decided to purchase the Aegis defense system from the United States. The US Congress opposed the deal and in August 1988 Senators argued in heated discussions that instead Japan should be sold a whole Aegis cruiser, as a package, because the Aegis system was an outstanding technology that gave the US shipbuilding industry a competitive edge, and which it would lose to Japan if only the technology were sold. They were clearly responding to pressure from the shipbuilding industry.15 The FSX affair revolved even more dramatically around the same issues. After long negotiations the US and Japanese governments had finally reached agreement to co-develop the FSX fighter in November 1988. This project was, however, opposed by Congress which argued that co-development would transfer aircraft-related technology to Japan which could be used for the development of its commercial aircraft industry. Under Congressional pressure the administration renegotiated the project and technology transfer to Japan was partially restricted as a result.16
Techno-systems and technology friction These frictions, involving both security and industrial competition, began in the early 1980s. Until then the differing nature of the US and Japanese technosystems did not give rise to major problems in their technology relations. Security and economic issues were politically separated and in the Cold War environment security was the dominant aspect of the overall Japan–US relationship. Military technology transfers from the United States to Japan were welcomed in both countries. Strengthening the military technology base of its most important Asian ally was seen as in America’s interest. Even when Japan used its military technology for strengthening its commercial technology base–a real possibility because of the nature of Japan’s techno-system–that was still acceptable to the United States because it wanted an economically healthy Japan. For example, what the Japanese learned from the licensed production of F104 fighters was applied to developing such systems as the anti-vibration structure and brake systems for bullet trains (Shinkansen) without any complaint from the US.17 But these perceptions changed in the latter half of the 1980s as Japan gained world level competitiveness in the high-tech sector. The high-tech apprentice which had done so well commercially out of the transfer of military technology became a hi-tech economic threat. Complaints about the lack of
260
Security and techno-systems reciprocity were especially sensitive in the case of transfer of military technology, such as the FSX, because they had been developed using US citizens’ taxes. Japan came to be seen as a free-rider making commercial gain from technologies developed for military purposes in the United States without paying the price for the necessary basic research. Friction regarding Japanese acquisition of US companies and the US military dependence on Japanese-made components also expanded from security to competitiveness in the late 1980s, The targeted strategy coordination between government and industry, which were characteristic of the Japanese technosystem, helped to create sectors that were truly competitive worldwide and sometimes dominant, such as DRAM memory chips. The fears aroused by the prospect of Fujitsu acquiring Fairchild were, first, that that market dominance would be reinforced; second, that certain armaments might become dependent on Japanese-produced parts which might not be available in an emergency; third, that the defense industry was becoming dependent on Japanese producers for semiconductors;18 and, fourth, that the Japanese dominance in production might lead to dominance in R&D, and America’s capacity for advanced research for military purposes might wither away. Differences in techno-systems intensified the technology friction. The friction itself–over both security and competition–significantly influenced military technology transfer, corporate acquisitions by Japanese companies, and US military use of Japanese technology.
Transformation of the US techno-system The new techno-system The prediction that the Japanese high-tech sector would dominate world markets in the 21st century was not fulfilled. The irony was that it was the United States which dominated newly created information technology, while Japan experienced a ‘lost decade’ during the 1990s. Behind the dramatic revival of the United States was the emergence of a new techno-system that replaced the DoD-dominated Cold War system. Industry, not government, led the so-called Information Technology (IT) revolution and contributed greatly to the forming of the new techno-system. To be sure, the seminal research leading to the Internet was funded by the government with national security concerns in mind. The Internet was operationalized initially as a government-based infrastructure for universities and national laboratories. In the early and mid-1980s, commercial applications were allowed on this Internet and it was gradually privatized (Messerschmitt 2000: 128). Diffusion of the Internet and its uses, however, occurred largely through commercial firms such as Netscape, Intel, and Microsoft, with the
261
Transforming Japan’s innovation system latter two going on to obtain a dominant presence in the personal computer sector and greatly increasing their power in the Internet world. The expansion of the network environment also produced a set of new players producing both hardware and software–Cisco Systems, 3Com, Bay Networks, and Ascend Communications in telecommunications; Google, Netscape Communications and Yahoo with their Web search engines; Internet-service providers such as American Online (AOL) and various electronic commerce companies, such as e-Bay. Established players in computer and telecommunications such as IBM and AT&T also restructured their businesses and developed new strategies for the new environment created by the Internet.19 The government role in the new techno-system was mainly confined to rule setting. During the senior Bush administration there were heated discussions regarding industrial policy, and concerning the role government should play in promoting industrial competitiveness. Some urged the government to take responsibility for installing a nationwide fiber optic network. The Clinton administration, however, opted, not for such proactive policies, rather for improving the legal framework for such matters as intellectual property rights, protection of privacy, and conflict resolution to cope with problems arising from the new network environment.20 The government also tried to extend these legal frameworks and rules to foreign countries, both through bilateral negotiations and by promoting them through international institutions. This policy gave clear advantages to US companies in the IT field in enlarging their overseas business opportunities.21 US universities also switched their dependence for research funds from the government to the private sector. The percentage of DoD funding to universities in all federal university research funds had been declining for many years. In 1958 (at the peak of the Cold War) it had reached 44 percent, but was down to 8 percent by 1975.22 The ratio of academic R&D supported by the federal government itself declined from 70.5 percent in 1970 to 59.0 percent in 1992.23 In looking for alternative funds from the private sector, universities were helped by changes in the legal framework. The Bayh-Dole Patent Act of 1980 allowed government agencies to grant excusive licenses for inventions made with government funds. It was followed by the Stevenson-Wydler Act of 1980, encouraging university–industry collaboration.24 Thus the incentives for universities to utilize their patents for commercialization greatly increased. In addition, companies increasingly looked elsewhere for basic research in order to concentrate on applied research and product development. Thus, universities and companies discovered mutual opportunities to add value and started to work together. This new techno-system, in which the private sector leads the way, started to function in the 1990s, and soon saw the United States establishing domination over world markets in information and telecommunication technologies.
262
Security and techno-systems
Impact on the security sector Transformation of DoD technology development and acquisition procedures started during the Clinton administration. One focus was the use of advanced commercial technologies in the military sector. There was an awareness that commercial, not military technologies, had started to lead overall technological advancement during the senior Bush administration and discussions started as to how the DoD should react to the new environment. The concept which emerged in the Clinton administration was civil–military integration, promoted at three different levels. The first was component procurement for military systems. Under the Cold War techno-system, the DoD set its own specifications, more advanced and sophisticated than those currently in use for commercial products. Now, however, commercial technologies had become superior to military applications, causing DoD to abolish most military specifications and to acquire commercial components directly for military purposes.25 The second change was at the R&D level. A much higher proportion of the research promoted had both commercial and military uses. And, third, at the manufacturing level, the DoD required the defense industry to adopt the best manufacturing practice available in the commercial sector. This was the burden of what the DoD called the Single Process Initiative. One outcome of this new approach applying commercial technologies in the military sector was the so-called Revolution in Military Affairs (RMA). The Gulf War marked the first use of such sophisticated weapons as precision guided missiles. When US troops fought in Iraq for the second time in 2002, the weapons used were even more sophisticated; the so-called C4ISR (Command, Control, Communication, Computers, Intelligence, Surveillance, and Reconnaissance) utilized a great deal of technology developed for commercial use.
The emerging Japanese techno-system: The concept of anzen-anshin Collapse of the old techno-system By the 1990s, it had become obvious that the old Japanese techno-system no longer functioned. The success of the old system had been based on accurate predictions of the future direction of technological development, such as the development of mainframes led by IBM in computers and the elaboration of ever more sophisticated DRAMs in the case of the semiconductor industry. The Japanese government and industry had assessed these technology trends correctly and targeted them for industrial growth. However, Japan was unable to adjust to the newly emerged technology trends in the 1990s. In the computer industry, personal computers instead of mainframe computers started to dominate. MITI’s strategy was encapsulated in the Fifth-Generation Computer
263
Transforming Japan’s innovation system Project (1980–92), in which the main focus of research on artificial intelligence was to supplant IBM’s domination by developing advanced large-scale computers not based on IBM software systems.26 Microprocessors became the technology driver in the semiconductor market and Intel, which had set the industry standard in this field, dominated the market by the 1990s. Japanese companies, however, persevered with DRAMs and could not adjust to the new trend. They experienced two setbacks as a result. They could not penetrate the growing microprocessor market due to Intel’s effective strategy in protecting its intellectual property rights. At the same time new players, especially Korean companies, entered the DRAM market and squeezed their profits with cheaper prices (see Yunogami, in this volume). Furthermore, Japan failed to predict the rapid diffusion of new network technology and the Internet and could not adjust to this totally new telecommunication environment (see Cole’s chapter, this volume). It is inherently difficult to predict market trends in the telecommunication sector, making it unsuitable for targeting type industrial policies. As Rosenburg observed: ‘[T]he history of the telecommunications industry is in fact the history of these technological discontinuities and their unexpected impacts’, hence, ‘policy should be constructed to ensure that the technological path is as flexible as possible’.27 One of the worst aspects of Japan’s ‘lost decade’ was the state of its defense industry. After the end of the Cold War, the world defense industry went through a restructuring process due to the declining trend in defense budgets. The US defense industry was reorganized into three major companies, Lockheed-Martin, Boeing, and Raytheon. European countries also each established one major defense company, and these national champions in the industry then started to merge, leading to a new structure in the European defense industry. In Japan, however, the defense industry was slow to adjust to the new post Cold War environment. The first instance of restructuring in the industry was the sale of Nissan’s defense and aerospace sector to Ishikawajima Harima more than ten years after the end of the Cold War.28 In addition to suffering from a declining defense budget and failing to adjust to the new environment, the Japanese defense industry was slow to adopt acquisition procedures that reflected the new technology environment in which commercial rather than military technologies lead technological development. Due to its failure to adjust, the question is now not whether the industry will remain competitive, but whether it will survive.
Anzen-anshin: an emerging techno-system? An important factor behind Japan’s ‘lost decade’ of the 1990s was that Japanese technology policy lost direction and was unable to devise effective strategies to replace targeting and catch-up style policies. Recent developments are worth exploring, however.
264
Security and techno-systems A reorganization of governmental departments and agencies took place in January 2001 and the Council for Science and Technology Policy (CSTP) was founded within the Prime Minister’s cabinet office. Its aim was to synthesize and implement various functions of science and technology policies under one umbrella. Under the CSTP, a revised Science and Technology Basic Plan (STBP) appeared in March 2001 which offered three main visions or concepts which should guide Japanese science and technology policy: 1) to be a nation contributing to the world through the creation and utilization of scientific knowledge; 2) to be a nation with international competitiveness and a capacity for sustainable development; and 3) to be a nation which gives its citizens security and a high quality of life.29 Initially, priority was given to the first two; the third, was not emphasized and no political initiative was taken to further it. The terrorist attacks on the World Trade Center and the Pentagon on 11th September 2001, however, changed this. The Japanese government realized the importance of anti-terrorist measures and started to consider how technology could contribute in that area. The rising crime rate in Japan was also a factor. The Japanese government therefore started to upgrade the importance of the third guiding concept. The Ministry of Education, Culture, Sports, Science and Technology (MEXT) was the most active in this, and created a Study Group on Science and Technology Policy for a Safe and Secure Society in April 2003. After intense discussion in this study group, a final report was published on April 2004.30 The study group was unusual (especially for what was predominantly the former Education Ministry) in including international relations experts, political scientists, and criminologists, brought together to examine how science and technology could contribute to the creation of a safe and secure society and to identify significant policy measures to deal with terrorism, crime, and cyber attacks as well as traditional elements of safety such as natural disasters, traffic accidents, and chemical pollution. The Cabinet Office CSTP also initiated a series of study sessions on science and technology policy for a safe and secure society in April 2003, and started a project team in December 2004 to consider how to integrate these elements into its third Basic Plan. The CSTP adopted broader perspectives than MEXT in its discussions, including such topics as border controls to prevent terrorism and international crimes. Several researchers in international relations and politics were invited to discuss issues directly related to national security. One interesting aspect of CSTP discussions is its attempt to redefine ‘dual-use’ not, as traditionally, as technologies with both commercial and military applications, but technologies which are both commercial and able to promote security and improve the safety of society and individual’s life. These are the technologies which it believes should have priority. There is an external aspect to this too. The Council believes that Japan, as a leading economic power in the region, has a responsibility to promote safety and security in Asia where natural
265
Transforming Japan’s innovation system disasters such as earthquakes, tsunami, and floods occur so frequently, and which forms the ‘arc of instability’ where conflict may occur in future.31 It has hitherto been difficult to discuss national security issues in the context of science and technology policy in Japan. This is partially due to Japan’s pacifist sentiment in general and partially to the specific reluctance of the Japanese technology community to enter into discussions related to national security. It is possible, therefore, that these new developments could herald a new security-related direction for Japanese technology policy. The results of the studies of both groups are to be incorporated in the new Basic Plan that will start from 2006.
Implications for techno-systems These discussions need to be observed closely because it is the first time since World War II that Japan has tried to create a security-related technology policy, and this movement may lead to a new techno-system in Japan. One reason is Japan’s promising potential in this anzen–anshin area. ‘Ubiquitous technology’, for instance, is a field in which Japan is on the world’s cutting edge. The radio frequency IC tag, a classic example of ubiquitous technology, can be used to deter terrorism. When an IC tag is attached to the boarding passes and luggage of air travelers it would make it only possible to load a suitcase onto an aircraft after the passenger matching that tag also was on board. Technology for which Koichi Tanaka won the Nobel Prize, mass spectrometric analysis, can be used to detect substances at the molecular level, and could be used for antiterrorism purposes if applied to detecting explosives or harmful chemicals.32 There are other outstanding Japanese technologies such as censors and robotics that have the potential to contribute to security and safety. In military technologies, on the other hand, it possesses no cutting-edge expertise. It is in the former, therefore, that Japan could most realistically establish itself as a technological leader, and so also contribute internationally to security and safety. A counter argument could be made that rising security concerns, coupled with growing nationalism and geo-political tensions in east Asia, will create a far less benign trajectory for the emergence of a new techno-system. Developments like joint R&D with the US for missile defense systems and moves to relax Japan’s three principles against weapons exports might be cited in support. However, the defense budget is unlikely to grow significantly in the foreseeable future, and the three principles, while under some strain, are combined with a continued aversion on the part of many civilian companies to become involved in military-related technology development. The counter argument, therefore, is not persuasive, at least in terms of a new techno-system. Rather, it is the anshin-anzen trajectory which will be spelled out in greater detail in the third STBP, and which will then stimulate new
266
Security and techno-systems developments, for instance, in university–industry–government links. And this trajectory, ironically perhaps given its contrast with the current US emphasis on mobilizing commercial technologies to reinforce its military might, is likely to create less friction in US–Japan technology relations. Indeed, when the US Department of Homeland Security approached the Japanese government to advance R&D cooperation in developing anti-terrorism technologies, the Japanese side insisted on broadening the perspective of the research to include a wider range of issues involved in its anshin-anzen approach. An agreement was reached in July 2004 to advance research cooperation in such areas as infectious diseases, terrorism, cyber security, and border control.33 In these defensive areas of technology, even if Japan were to develop better technologies than the US, it is less likely that they would give rise to major technology friction. It cannot be denied, of course, that the United States would put political pressure on Japan if it were to gain an overwhelming lead in such technologies. Judging from the recent climate surrounding the US–Japan relations, however, it is more likely that objective assessment of the costs and benefits of cooperation would prevail. It is yet to be seen whether the anshin-anzen approach will develop into a new security-related techno-system. There are, however, technological, political, and international relations reasons to believe that it has the potential to do so.
Notes 1. 2. 3. 4. 5. 6. 7. 8. 9.
Edquist 1997: 1–35. For the postwar debate on science policy, see, for instance, Penick et al. 1992. Mowery and Rosenberg 1989: 125. See Leslie 1993. Brooks 1996. For instance see Goto and Wakasugi 1984. See Chinworth 1992: ch. 2. Samuels 1992. New York Times, 12 December 1981; Mainichi shimbun, 3 November 1981; Nihon keizai shimbun, 30 April 1982. 10. Cf. ‘Sale of New Hampshire Ball Bearings, Inc. to the Japan-Based Minebea Company’, hearing before Subcommittee on Preparedness of the Committee on Armed Services, US Senate, 98th Congress, 26 September 1984; ‘Administration Wins a Battle but Worries about the Export Control War,’ National Journal, No.31, 30 July 1983. 11. Nihon keizai shimbun, 26 February 1983 and 3 July 1984. Cf. also Nippon Steel’s aborted attempt in 1983 to acquire Special Metals Corp., a subsidiary of Allegheny International Inc., which made special alloys for fighter aircraft (Nihon keizai shimbun, 3 July 1984). Minebea, however, successfully acquired New Hampshire Ball Bearings in 1984, in spite of similar DoD pressure (Nihon keizai shimbun, 12 Sept. and 5 Oct. 1984).
267
Transforming Japan’s innovation system 12. Prestowitz 1988, p.410. 13. Nihon keizai shimbun, 24 Oct. 1986; New York Times, 12 Jan. 1987. 14. Three cases were investigated by the CFIUS involving Japanese acquisitions and investments and US companies. Two of them–Tokuyama Soda’s acquisition of General Ceramics on March 1989, and Fanuc Machine Tool’s investment in Moore Special Tool on September 1990–were about nuclear-related technologies. In the third, Nippon Sanso’s offer to buy Semi-Gas Systems was opposed by SEMATECH (of which Semi-Gas Systems was a member) on the grounds that technologies developed in SEMATECH would pass to Nippon Sanso. After investigating the claim CFIUS approved the acquisition, despite strong opposition in Congress (McLoughlin 1991; Hearing before the Subcommittee on Science, Technology, and Space of the Committee on Commerce, Science, and Transportation, US Senate, 101st Congress, 10 October 1990; ‘Decline of the U.S. Electronics Industry,’ Hearing before the Subcommittee on Science, Technology, and Space, US Senate, 101st Congress, 1 August 1990). 15. Congressional Record, 15 August 1988, 134(116): S10936–S10948. 16. Cf. Teshima and Ohtsuki 1991, and Honda 1991, for a detailed account of the FSX episode. 17. Murayama 2000a: 157. 18. For instance Heginbotham 1990. 19. Murayama 2000b: 163–71. 20. Murayama 2000b: 176–84. 21. The United States started to be actively engaged in this kind of economic diplomacy in the telecommunication field in the latter half of the 1990s (Marayama 2000b: 15–21). 22. Sapolsky 1990: 137. 23. Tech 1998: 89. 24. Branscomb and Florida 1998: 17. 25. Murayama 2000b: 193–203. 26. For the Fifth Generation Project, see Matsuda 1995: 351–83. 27. Rosenberg 1994: 219 and 228, italics in the original. 28. The sale took place in April 2000 (Kubota 2002: 112). 29. The words translated here as security-anshin–anzen–literally mean ‘peace of mind and safety’ (see also Arimoto’s chapter, this volume). 30. Monbu kagaku sho, Anzen anshin na shakai no kochiku nishisuru kagaku gijutsu seisaku nikansuru kondankai, hokokusho, April 2004. 31. Anzen ni shisuru kagaku gijutsu suishin purojekuto chiimu, Anzen ni shisuru kagakugijutsu no igi, mokuhyo, oyobi hoshin ni tsuite, March 2005. 32. Murayama 2004. 33. Nikkei shinbun, 19 July 2004.
References Branscomb, L. and R. Florida (1998). ‘Challenge to Technology Policy in a Changing World Economy’, in L. Branscomb and J. Keller (eds.) Investing in Innovation: Creating a research and innovation policy that works, Cambridge, MA: MIT Press.
268
Security and techno-systems Brooks, H. (1996). ‘The Evolution of U.S. Science Policy’, in B. Smith and C. Barfield (eds.) Technology, R&D, and the Economy, Washington, DC: The Brookings Institution. Chinworth, M. (1992). Inside Japan’s Defense: Technology, economics and strategy, Washington: Brassey’s (US). Edquist, C. (1997). ‘Systems of Innovation Approaches–Their emergence and characteristics’, in C. Edquist (ed.) Systems of Innovation: Technologies, institutions and organization, London: Pinter. Goto, A. and R. Wakasugi (1984). ‘Gijutsu seisaku’ (Technology Policy), in R. Komiya, M. Okuno, and K. Suzumura (eds.) Nihon no sangyo seisaku (Japanese Industrial Policy), Tokyo: Tokyo daigaku shuppankai. Heginbotham, E. (1990). Dependence of U.S. Defense Systems on Foreign Technologies, Alexandria, VA: Institute of Defense Analysis. Honda, M. (1991). Nichibei FSX Senso (The U.S.–Japan FSX War), Tokyo: Ronsosha. Kakugi kettei (2001). Kagaku gijutsu kihon keikaku (Science and Technology Basic Plan), 30 March. Kubota, Y. (2002). ‘Nihon no boei sangyo no tokushitsu: Sangyo kozo to anzen hosho seisaku ga ataeta eikyo no bunseki’ (Characteristics of Japan’s Defense Industry: Analysis of the influence of industrial structure and security policy), Kokusai seiji, October. Leslie, S. (1993). The Cold War and American Science: The military–industrial–academic complex at MIT and Stanford, New York: Columbia University Press. McLoughlin, G. (1991). ‘The Semi-Gas Systems Sale: Technology and national security issue’, CRS report for Congress, 12 March. Matsuda, M. (1995). ‘Soshikiteki kenkyu kaihatsu no mekanizumu: Dai go sedai konpyuta no kenkyu kaihatsu o megutte’ (Mechanism for Organizational Technology Development: The case of technology development for the fifth-generation computer), in I. Nonaka and A. Nagata (eds.) Nihongata inobeshon shisutemu: Seicho no kiseki to henkaku e no chosen (Japanese Innovation System: History of its development and challenges for Reform), Tokyo: Hakuto shobo. Messerschmitt, D. (2000). Understanding Networked Applications, San Francisco: Morgan Kaufmann. Mowery, D. and N. Rosenberg (1989). Technology and the Pursuit of Economic Growth, Cambridge: Cambridge University Press. Murayama, Y. (2000a). Keizai anzen hosho o kangaeru: Kaiyo kokka Nihon no sentaku (On Economic Security: Alternative security policy of Japan), Tokyo: NHK Shuppan. —— (2000b). Tekuno shisutemu tenkan no senryaku: San kan gaku renkei e no michisuji (Strategy of Techno-System Transformation: Path to Effective Alliance between Industry, Government and University), Tokyo: NHK Shuppan. —— (2004). ‘ ‘‘Defensive’’ Technology Right up Japan’s Alley’, International Herald Tribune, 9 September. Penick, J. with C. Pursell, M. Sherwood, and D. Swain (eds.) (1992). The Politics of American Science, 1939 to the Present, Cambridge, MA: MIT Press. Prestowitz, C. (1988). Trading Places: How we are giving our future to Japan and how to reclaim it, New York: Basic Books. Rosenberg, N. (1994). Exploring the Black Box: Technology, economics, and history, Cambridge: Cambridge University Press.
269
Transforming Japan’s innovation system Samuels, R. (1992). Rich Nation, Strong Army: National security and the technological transformation of Japan, Ithaca: Cornell University Press. Sapolsky, H. (1990). Science and the Navy: The history of the Office of Naval Research, Princeton: Princeton University Press. Tech, A. (1998). ‘The Outlook for Federal Support of University Research’, in R. Noll (ed.) Challenges to Research Universities, Washington, DC: Brookings Institution Press. Teshima, R. and S. Ohtsuki (1991). Nippon FSX o ute (Shooting Down the Nippon FSX), Tokyo: Shinchosha.
270
15 Technology management training in Japan: Government initiatives and their effects Atsushi Kaneko, Yoshi-fumi Nakata and Muneaki Yokoyama
Background to the introduction of MOT 2005 is an important year in the Japanese technology calendar, the last year of the 2001–05 Science and Technology Basic Plan (budget ¥24 trillion) and the year in which the basic directions for the next STBP are to be decided. Main themes of the Plan will be: priority emphasis on the four fields of the life sciences, information technology, environment, and nanotechnology; promoting the application of research results in production and business; and promoting an understanding of the relation of science and technology to society. In short, the theme song will be the building of a national innovation system with dynamic feedback, capable of getting the achievements of R&D swiftly and efficiently translated into social and industrial activity in full awareness of the possible effects on society (see Arimoto’s chapter, this volume). We shall sum up what this national planning background means for the management of technology in Japanese corporations in the subsections that follow.
Management of the uncertainties of advanced technology The commercial application of new technologies becomes more difficult as science and technology becomes more advanced and more narrowly specialized, and the procedures for embodiment in production become more complex and uncertain (see Chesbrough, Chapter 7, this volume). Although the latent possibilities for creating markets are good, what is often needed is a multi-dimensional approach which combines several unrelated fields. This is especially so in
271
Transforming Japan’s innovation system the emerging technology fields–not only the four enumerated above, but also in the fields of energy, production technology, social infrastructure, and what has come to be known as ‘human frontier science’. Many corporations are deficient in the MOT skills which are needed here–the discernment to judge and evaluate markets and technologies, and the capacity to build and operate cross-disciplinary and cross-divisional organizations. It is in this sense that the uncertainties of advanced technology create a big challenge.
R&D systems under international competitive pressure The need to be internationally competitive requires speeding up the commercialization process at a time when the process is becoming ever more difficult. In today’s competitive environment what is essential to secure greater speed and a greater quotient of successes in feedback to the research process from society, and the ability to take strategic research management decisions, including the creation of alliances. But Japanese corporations are too addicted to the ‘linear model’ in which the central lab does everything from basic to applied science, producing ideas for the production divisions to implement.1 This robs the R&D process of speed and the capacity to react flexibly to change. Three observations are commonly made about Japanese R&D: 1 That there is a shortage of exploratory research, the phase preceding basic research which seeks to determine the most promising themes for industrial research. Exploratory research needs a sensitivity to trends in science and in society and should provide compass bearings in a situation of uncertainty with respect both to markets and technology. It has more of the character of a public good, using the store of scientific knowledge for the benefit of industrial society as a whole, rather than something of benefit to individual companies, and hence should be undertaken by universities. 2 That the marketing function plays too small a part in the whole R&D process.2 All the way through from exploratory research to production application there needs to be continuous feedback, with consequent adjustment, not only from knowledge of advanced technology trends, but also from ‘society’–customers and their surrounding environment. The need has long since been recognized, but practice still falls short. 3 That there is a need for a more appropriate division of labour among business, academia and government.3 The effectiveness of the linear model is constrained by the limits of a single corporation’s resources and the financial burden involved. A more efficient model would be for universities to undertake the basic research with firms providing them with the necessary marketing information and then taking over the application and development stages, or for universities to do the exploratory research, and national laboratories to be responsible for basic and perhaps applied research with firms taking over from there. Better coordination is, in any case, needed.
272
Technology management training in Japan
A profit-oriented competitive strategy The third problem comes after entering the market–the difficulty of meeting the competition and making profits. In today’s market environment in which technology advances with such dazzling rapidity, even first movers are by no means assured of their profits, given the challenge of imitators and producers of something even newer. Hence the need not only for a constant stream of new products, but also a patenting strategy, and a search for new uses of already developed technologies. While settling on and concentrating on what is to be their core competence, firms need the strategic imagination to flexibly exploit external resources through close collaboration not only with other large firms, but also with smaller firms, venture firms, universities, and government laboratories. None of this is easy in a situation in which uncertainty prevails, both about markets and technologies.
Using science and technology to serve society The fourth issue is the growing awareness of how important it is that science and technology should work to the benefit of society as a whole. An understanding of, and due consideration for, the effects on man and society are not only important for answering the demand for ‘security and peace of mind’ (anzen-anshin, see Murayama’s Chapter 14, this volume), but also as a prerequisite for the building of a ‘science-based society,’ both major themes of the STBP. Firms cannot stand aside from these concerns but must take aboard their role as prime movers. And that means having a clear idea of how applying science to industry produces an innovative society, how science relates to technology, and what are the conditions for creating innovative organizations and institutions. These are not matters susceptible to snap judgements, and they have to be taken seriously with the emergence of every new technology.
Maintaining and strengthening the Japanese management style The fifth question is how one keeps, and builds on, the management methods traditionally labelled as ‘Japanese management’. It was, indeed, the efficiency of operations management in Japanese firms which directly led to the establishment of MOT programmes in the United States. At the heart of that style of management was the idea that the human being is important, that the firm needs to protect its human resources from shareholders and other outside forces, and foster the development of the creative skills of its employees. That is why making sure of the firm’s ‘contribution to society’, its ‘sincerity’ as an organization, by maintaining the traditional shopfloor orientation and thereby fostering those skills and providing incentives for their full use
273
Transforming Japan’s innovation system through such devices as teamwork should, together with making profits, be considered the objectives of MOT. It is often observed that these traditions are disappearing. Hence, an important MOT issue is to discover the weaknesses of the mechanisms for intergenerational transmission of these enterprise resources, and find some way of restoring them (cf. Hitachi’s ‘e-Meister Activities’ reported by Whittaker, this volume). To tackle all the above five problem areas, first a capacity for taking a comprehensive overview of the most advanced fields of science is needed, as well as deep insight to discern what can be translated into useful technology. And the other requirement for taking managerial decisions about R&D investment which are informed by powers of judgement at once responsive and strategic, is a good sense–right from the early stages of scientific exploration–of where global market advantage is to be found. That is what modern MOT is about–that and strengthening the efficient shopfloor oriented management methods for which Japanese corporations have been renowned the world over. So much for the background to the introduction of MOT. In the sections that follow we explore the question of the personal qualities, the ideal profile of the technology manager, and then look at how the problem of developing such personnel resources is tackled in Japan and other countries.
The profile of the ideal technology manager As described in the Introduction, the Ministry of Economy, Trade and Industry (METI) launched a number of initiatives in 2002 to promote the training of technology managers. A project called ‘Measures to speed the introduction of technology manager training programmes’ was contracted out to the Mitsubishi Research Institute. One element of the project was to study and delineate the skills and personal qualities required for MOT as a means of establishing a community of ideas between training organizations and firms. The study used interviews with firms, literature searches–including the search for best practice in casestudy reports–and expert opinion. Its conclusions form the basis of what follows. The requisite skills are divided in the first instance into those specific to technology management, and those more generally needed in all management, such as leadership, interpersonal skills, value creation, flexible response to change, and a moral sense. The specific skills were subdivided, as shown in Table 15.1, into strategic ability in management, business and technology, executive capacity, entrepreneurship, mobilization of intellectual capital and mobilization of external resources. It will be apparent that the skills required to tackle the challenges listed above are to be found herein.
274
Technology management training in Japan Table 15.1 Management skills for MOT Skill category
Skill Management
Drawing up a management ‘vision’, strategy Integrating technology in a management/business strategy
Business
Choice of business, settling on core business Assessing business prospects Drawing up a business plan Defining the business/product concept Translating market needs/wants into R&D objectives Grasping those market needs/wants Making technology the formal or de facto standard
Technology
Drawing up a technology ‘vision’, strategy Settling on core (and noncore) technology Planning your technology platform (forming your R&D process structure) Getting a grasp of technology resources: profile management Creating a technology roadmap Setting R&D objectives Discerning new uses for technologies, exploiting latent demand Technology forecasting and evaluation Understanding trends and findings in the most advanced fields of research
Strategic
Executive skills
Creating the most appropriate team or organization for the task in hand Managing technology experts Managing complex and multi-functional projects (in a constantly changing environment) Innovation management Managing technology transfer inside and outside the firm
Entrepreneurship
The ability to promote start-ups, spin-offs, spin-outs The ability to engage in start-ups, spin-offs, spin-outs
Exploiting intellectual resources
Securing, reinforcing, and getting best value out of intellectual property Evaluating intellectual property Advising on commercialisation from a legal point of view
Exploiting external resources
Evaluating pros and cons of licensing own technology and devising a strategy for this Acquiring technology from outside (alliances, M&A)
Source : METI (2002). ‘Report on Entrepreneurship Education Programme’
The list in Table 15.1 is comprehensive and the ideal is to combine them all, but in practice one can differentiate certain essential roles: . The CTO, with the ability to take responsible decisions on technology strategy in full awareness of uncertainty and risk . The strategist, with a planning capability which integrates business and technology strategy . The MOT lawyer, capable of strengthening the competitiveness of the firm’s technology through legal protection of intellectual property . The MOT administrator, able to use the sale and purchase of technology or businesses to enhance the firm’s business activities and profitability
275
Transforming Japan’s innovation system . The MOT creator, able to get the best out of the firm’s technology with a stream of winning products . The liaison engineer, who samples and develops promising technologies . The project manager, who knows how to organize the planning and production of products of the highest design and performance. Graduate education designed to foster such practical MOT talent is beginning to emerge based on cooperation between business and educational institutions, in full awareness of the problems listed earlier. However, these are still exploratory beginnings both in terms of their historical depth and their educational technology. We still have to establish the guiding principles. For that reason we take a look at foreign examples before turning to a consideration of Japan’s special circumstances and the problems that lie ahead.
Approaches to MOT training overseas The United States In America of the 1980s the twin deficits–in trade and in public finance– weighed heavily on the political agenda. At the same time the rise of other manufacturing powers, notably Japan, threatened America’s competitive supremacy. At the prompting of a business community suffering from a severe sense of crisis the US government set in motion a programme of studies and consequent institutional reforms designed to improve the nation’s international competitiveness. Two landmark studies were the Young Report of 1985 and the Made in America study by an MIT group in 1989. Among the various suggestions, a frequent theme was the need for training in the commercialization of new technologies. The response was a reform of higher education and particularly of business schools to accommodate this need. This provided the impetus for the launch and expansion of MOT programmes in the US. One of these, started in 1985, was an MOT programme at MIT, and in 1988, the Leaders for Manufacturing programme. On the west coast, UC Berkeley created a joint programme between the Haas School of Business and the Engineering School and Information Technology Faculty in 1987. In 2000 a new programme was started at Stanford as a result of a series of organizational tie-ups (Table 15.2). The Leaders for Manufacturing programme consisted not just of classroom work, but also a six-month internship, many site visits and a thesis.4 The Sloan School (business) and the Engineering School work together to make it as practical and as comprehensive as possible. The internship programme also functions as a means whereby firms get the opportunity to recruit talented individuals. The MOT programme at Berkeley has a good reputation in industry
276
Table 15.2 Types and characteristics of MOT programmes in the US and Europe
Background
Response of educational institutions to industry’s sense of crisis and declining competitiveness
Diverse sponsors, common concerns, collaboration with industry, straddle business, and engineering
Sponsoring body
Joint programme
Engineering school
• MIT MOT programme • Northwestern University Technology Industry Management programme • SIMT Technology and Innovation Management • EPFL MOT programme
• UCB MOT programme • MIT LFM programme • University of Pennsylvania MOT programme
• Stanford University MS&E • University of Sussex Technology and Innovation Management
Curriculum
Basic business studies, technology studies, innovation-related studies integrated in MOT as a new interdisciplinary study
Teaching methods
• Internships, PBL, etc. as means both of intellectual training and imparting leadership/teamwork skills • Casestudies for training in problem identification, hypothesis-framing, decision making
Teaching resources
• Deployment of teachers well versed in the ways of business • Deployment−and training−of teachers capable of giving practical training
277
Note: These summary characterizations are merely indicative and should not be taken as comprehensive. Source: MRI document
Technology management training in Japan
Business school
Transforming Japan’s innovation system for producing graduates with teamwork skills, analytical skills and problemsolving ability, possibly because the programme includes participation in the actual R&D projects of technology firms.5 The Stanford programme includes a high quotient of project-based learning, which provides a catalyst for closer relations between firms and the university, inasmuch as firms frequently provide cases which they want the course to tackle.6 Apart from these examples, there is a great diversity of MOT programmes in American universities. Some, like the above, are joint programmes, others are based in engineering schools and emphasize technology, and others are based in business schools, with a greater emphasis on management and leadership. The dominant themes also shift with the times; universities flexibly change focus depending on where they think the major problems lie. The MOT course at MIT, for instance, has shifted from an early emphasis on R&D management, to such themes as technology transfer, innovation, technology strategy, and corporate venturing.
Europe Europe, too, undertook a number of strategic measures to improve its competitiveness, after it lost ground vis-a`-vis the US and Japan in the years after the oil shock. After the introduction of the single currency in 1999, too, there has been lively discussion of the need for greater competitiveness, centring on questions of finance, taxation, employment, and industrial restructuring, but also on promoting innovativeness and market orientation in R&D policies, not to mention calls for a European MIT. This has meant, in spite of differences in the circumstances of individual countries, a common awareness both of the way that the globalization of corporate activity intensifies competition, and of the importance of MOT, of the role that R&D plays in profitability, and the need to direct firms’ technology strategies with profit in mind. While there are fewer actual MOT programmes than in the United States, well-known examples include SIMT in Germany, SPRU in Sussex, UMIST in Manchester, and EPFL in Switzerland. There are no marked differences from American programmes; the main source of diversity lies in problem focus and whether they are primarily based in business schools or engineering schools.
Features of MOT programmes in Europe and the US The first common characteristic is that the core course content is similar to that of an MBA–the full range of corporate finance, marketing, competitive strategy, operations management, management audit, corporate strategy, etc. The main point may be ‘solving management problems in a technological society’, but students are expected to acquire an understanding of basic business concepts regardless of background. Optional subjects explore some
278
Technology management training in Japan aspects of this core in depth, with courses on entrepreneurship and on specialized technology fields such as IT, computer science, engineering systems, environmental engineering, etc. (There are no MOT departments at the undergraduate level, though there frequently are undergraduate STS–Science Technology and Society–courses.) As for teaching methods, there is considerable use of internship and projectbased learning, designed at one and the same time to impart knowledge of engineering, production management, information technology, innovation management, etc. and to develop through experience the skills in leadership and teamwork. With the synergies produced by a curriculum which combines and integrates management and engineering, the aim is a hybrid product; the graduate with both management and engineering skills. In addition to practical projects, there is extensive use of casestudies designed to hone the ability to identify problems and form hypotheses. Other features are the chance to meet the top management of international corporations and hear their practical wisdom, and the provision of follow-up continuing education programmes for graduates. The faculty usually consists of people with a good deal of practical knowledge, or experience in business, and efforts are made to school the faculty in down to earth teaching methods. Close relations between business and the academy are an essential key to all of this, based on the mutual advantages they obtain–for the universities high quality teaching and research, for the firms a supply of first rate talent and the opportunity to learn from academic approaches to problem solving.
MOT programmes in Asia MOT programmes have been established in a number of East Asian countries, including China, South Korea, and Singapore. In China they began with the economic reforms of the early 1980s when the need for talented managers became apparent. At first it was mostly a matter of direct import of MBA programmes in collaboration with American or European universities, but programmes independently designed to deal with Chinese circumstances were subsequently developed. There were frequent declarations from the Chinese government recognizing the importance of MOT, and the need for people with a good technical background who could deploy their skills for economic development. Practically this meant introducing technology elements into MBA courses and promoting MOT-related research. In South Korea in the1980s MOT training was limited to a small number of universities, but provision expanded in the 1990s. With the growth of the IT industry and an increase in the number of venture businesses there was a growing appreciation of the fact that competitiveness depended on the ability to produce new products with advanced technology. The Korea Advanced
279
Transforming Japan’s innovation system Institute for Science and Technology (KAIST) began its Techno-MBA programme in the mid-1990s and MOT courses were started at Seoul National University, Yonsei University, etc. Samsung was heavily involved in the KAIST Techno-MBA programme, and continues to provide an example of academic– business cooperation by sending its researchers to the Institute. Such fully developed arrangements are rare, however, and in most universities the ability to respond to changes in the industrial environment is limited. The Singapore government, in the context of the growing strength of the surrounding Asian economies, has launched a Technopreneur-21 programme, setting goals for the year 2010 and with a heavy emphasis on encouraging entrepreneurial start-ups. The ‘Global Schoolhouse’ part of this aims to invite the world’s top universities to set up campuses for training over the whole range of science and business. MIT, INSEAD, Stanford, Chicago, and Pennsylvania have already responded and, given the direction of policy and the industrial characteristics of Singapore, one can expect them to develop MOT programmes. Such programmes are also being introduced at the National University of Singapore, Nanyang Technological University, and Singapore Management University.
MOT in Japan MOT training in Japan is still in its infancy. Here we describe how it has developed, and what are the problems. The main thrust of MOT programmes was initially on the production management side that has been Japan’s forte. The problems of MOT in conditions of uncertainty were dealt with only by smallscale on-the-job training programmes in a limited number of individual firms.7 In this context, and changes in the industrial environment, the initiative was taken not by industry or academia, but by the central government. From 2002 it began the development of educational programmes designed to systematize the available knowledge in the MOT field. By 2004 some 113 programmes had been developed by universities and other educational institutions, and considerable effort had been expended to develop capable teaching faculty and to spread an awareness among firms of the importance of MOT training.8 The educational foundations were laid and incentives provided, and as a result there has been a rapid increase in the number of educational institutions offering MOT programmes.9 At the same time the involvement of industry and university has been very much trial and error, and more passive than active. There is some doubt as to whether most of these rapidly-created programmes really serve industry’s needs and whether they really produce people who can play a useful industrial role. A hundred flowers are blooming in the MOT field, and one can expect to see a process of natural selection, sharp competition for survival.
280
Technology management training in Japan We conclude this chapter by taking a look at the present state of and problems facing, on the one hand the educational institutions responsible for training MOT talent, and on the other industry which provides the organization in which that talent can be deployed.
Educational institutions10 MOT programmes have mushroomed in Japan, but many of them have difficulty recruiting students. There are problems with curriculum design, teaching methods, and teaching resources. A good course needs an educational philosophy based on an appreciation of the nature and problems of industry and with a clear idea of the profile of skills it seeks to produce. That is what should guide the framing of the curriculum and teaching methods. But all too often MOT courses are introduced without such guiding principles, sometimes because they have been introduced simply as a survival strategy induced by the intensity of the current inter-university competition, sometimes because the university just has not had enough contact with industry to acquire the necessary understanding, and sometimes because the course has been promoted by one or two enthusiastic individuals who have lacked the organizational clout to get full understanding and cooperation from the university. The educational philosophy might be right and still the course content insufficient to produce the right set of skills and knowledge. This may be a case of the learning objectives for individual subjects, or the means of attaining them, not being properly thought out. Or it may be the case that the interrelation between different subjects is inappropriate and full educational effectiveness thereby lost. This can easily happen because there is a need to revamp individual subjects to adapt them to the essentially interdisciplinary field of MOT, and it can prove difficult for several reasons–if there is not enough of a research base, if there is organizational resistance, or simply because of a lack of pedagogic talent. The majority of programmes end up as a simple yoking together of established technical courses and established business courses. There is another problem which comes from the multiplicity of the objectives of MOT courses. They seek to impart knowledge. They seek to stimulate and motivate creativity. They seek to impart the ability to make value judgements based on a full understanding of the relation of science and technology to society, and they seek to perceive where the problems of developing particular technologies lie, to develop hypotheses and to test them. These objectives cannot be achieved simply by the one-way transmission of knowledge from teacher to student. They require a great deal of learning of concepts and abilities that cannot be expressed in words and that come from the nature of the relation between teacher and student and of the student with other students. Japanese universities have been too tied to the one-way transmission practice to have
281
Transforming Japan’s innovation system acquired adequate skills to do those things well. And yet this is precisely what people who have already had some work experience–who make up the majority of the clientele for these courses–want. The gap between expectation and offering may be one reason why many courses have difficulty filling their places. As for educational resources, there is indeed a shortage of teachers who can offer courses of suitable content in a suitable manner. Most of the teachers presently engaged in such courses have a business or a technical background, but most lack the understanding or awareness of corporate organizational dynamics necessary for the new discipline of MOT. Those with a business background often lack an appreciation of the changing place of technology in the value chain, or of the technology product cycle in the market, while those with a technical background have difficulty digesting and reasoning about the significance of technologies as seen from the management point of view. And where the teacher has little experience at the ‘coal face’ and is faced with students from industry, one often hears complaints from the students that they ‘can take it in with the head, but can’t feel it in the belly’. The problems are thus considerable. What is clearly needed is to develop training for MOT educators before the hopes and expectations that are placed on MOT training begin to diminish.
Industry Japanese firms continue to attract world attention for their strengths in efficient production management. And some firms are, in trial and error fashion, developing MOT training and competencies. But there are problems both on the production side and on the training side. On the production side the problem often is a lack of appreciation of the value of scientific management–a matter of the corporate culture, the absence of common management norms. This means that even when someone who has the right skills arrives, he is not able to be fully effective. They are not able to diffuse their skills, and the seeds of further development are stunted. Again, even firms that do practice MOT usually do so on the basis of tacit knowledge, which limits the possibilities of transmission and diffusion. Not everything can be made formally explicit, but in order for tacit knowledge to be passed on widely and easily, there needs to be collaboration with academic institutions skilled in making the tacit systematically explicit with good teaching materials. In view of the impending retirement of large numbers of senior technology managers, this is a matter of some urgency. On the personnel development side, what is missing is training practice which starts from a careful profiling of the qualities demanded by the problems the firm faces. Management training for technical staff and half day courses on MOT are fairly common, but they are rarely the product of fully worked out planning, and hence are of limited efficacy.
282
Table 15.3 Types and characteristics of MOT programmes in Japan
Background
State initiative, early 2000s, to increase social, industrial returns to R&D, through aid to educational institutions Majority based in engineering faculties with developed patterns of collaboration with industry
Business studies
Engineering
• Doshisha U. Business School • Waseda U. Graduate School of Asia-Pacific Studies • Kyushu U. Dept of Business and Technology Management • Nihon U. Business School
Course content
For the most part adaption of existing courses, with little effort to re-structure as a new interdisciplinary MOT study
Teaching methods
• Some experimentation, in initial stages, with internships and Project-based learning • Casestudy method encouraged but used mostly as examples with little educational effect
Teaching resources
• Few educators with experience of industry, practitioners with little experience of teaching • Few teachers skilled in practical teaching methods, and training courses in their infancy
• Shibaura Institute of Technology • Yamaguchi U. Graduate School of Technology Management School of Technology Management • Tokyo U. of Science, Management • Tokyo U. of Agricultural Technology, Graduate School of of Science and Technology Technology Management • Tokyo U. of Technology. Graduate School of Innovation Management • Nihon Institute of Technology, Graduate School of Technology • Ritsumeikan U. Graduate School of Management Technology Management
283 Note: These summary characterizations are merely indicative and should not be taken as comprehensive. Source: MRI document
Technology management training in Japan
Sponsoring body
Transforming Japan’s innovation system A few firms are starting fully fledged MOT training, but usually it is on a small scale. What bothers the training departments of such firms is not knowing how the training they give will be exploited–what the pay-off from the training investment will be. Good answers to that question are not forthcoming because there are no well charted career paths for technology managers. Establishing such career paths–profiling the personal requirements for the functions their firm needs performing and adapting the firm’s organization so that those who have such qualities can deploy them–is where the effort has to be applied. In short, the importance of developing the skills required for MOT is generally recognized. Practical implementation is beset by multiple unsolved problems, however. The basis for good MOT training has been more or less laid with help from the government. Now the task is to build on that base and raise its quality through collaboration and competition among the three partners of industry, government, and academe.
Notes 1. This addiction survives recent re-organizations of corporate research; work from the central lab is commonly moved to an operating division lab without changing the premise of the linear model (MEXT 2004: figure 2.3.1). 2. Early feedback from consumers was chosen by only 10.2% of responding firms as the reason for re-organizing research activities (MEXT 2004: figure 2.4.1). 3. See the section 2 of MEXT(2001). 4. See http://lfm.mit.edu/ 5. See http://mot.berkeley.edu/ 6. See http://www.stanford.edu/dept/MSandE/ 7. Only 12 percent of the firms surveyed have offered MOT programmes whereas another 20 percent of them have a plan to offer in the future (MOT Consortium 2005: 4). The most frequently cited method is OJT, which was chosen by 74 percent of respondents (ADL Japan 2005: 14). 8. See http://www4.smartcampus.ne.jp/ 9. In contrast to the active role played by METI, MEXT’s involvement in MOT education has been negligible. This may change in the future; in MEXT’s mid-term report toward the 3rd Basic Plan for Science and Technology, MOT and intellectual property were mentioned as areas for promotion in the 3rd BPST period (MEXT 2005: 19). 10. The basic perception of the state of MOT programmes at Japanese educational institutions is based on Goto and MRI (2005).
References ADL (Arthur D. Little) Japan (2005). MOT riida ikusei housaku no chousa kenkyu (Research in the Methods of MOT Leader Training), Tokyo: ADL Japan. Dertouzos, M., R. Lester, R. Solow, and the MIT Commission on Industrial Productivity (1989). Made in America: Regaining the productive edge, Cambridge MA: MIT Press.
284
Technology management training in Japan Goto, K. and Mitsubishi Research Institute (2005). Gijutsu keiei kyoiku jigyo yoin chosa hokokusho (Research Report on the Factors for Developing MOT Educational Initiatives), Tokyo: MOT Consortium. MEXT (Ministry of Education, Culture, Sports, Science and Technology) (2001). Shinjidaino sangakukan renkei no kochiku ni mukete (Toward a New Form of Industry, Government and Academic Collaboration), Mid-term Report of Council of Science and Technology, Tokyo: MEXT. —— (2004). Minkan kigyo no kenkyu-katusdo ni kansuru chosa hokoku (Survey on the Research Activities of Japanese Firms), Tokyo: MEXT. —— (2005). Daisanki kagaku gijutsu kihon keikaku no juyo seisaku (Policy Priorities of the 3rd Science and Technology Basic Plan), Tokyo: MEXT. MOT Consortium (2005). Kyoiku ni kansuru ankeito chosa: Kekka gaiyo (Survey on Company-sponsored MOT Education: Summary), Tokyo: MOT Consortium. Report of the President’s Commission on Competitiveness (the ‘Young Report’) (1985). Washington DC.
285
16 Electronic government in Japan: Towards harmony between technology solutions and administrative systems Toshiro Kita
The e-Japan strategy was announced in January 2001 when the Japanese economy was struggling to emerge from the long tunnel of the so-called ‘lost decade’ of the 1990s. For those used to bureaucracy-led policy making, the announcement was striking. Not only did it present a clear action plan and set quantitative targets, but it was drawn up by the IT Strategic Headquarters headed by the Prime Minister.1 Its successor, the e-Japan Strategy II was announced in July 2003, and aimed to realize an ‘energetic, worry-free, exciting and more convenient society through effective IT utilization’. In February 2004, moreover, the IT Strategic Headquarters proposed an e-Japan Strategy II Acceleration Package in order to realize the goal of ‘Japan becoming the most advanced IT nation in the world by 2005’. The Network System for Basic Resident Registers2–or Juki-net, as it is commonly called–is a core system for e-Japan. It was constructed by the Ministry of Internal Affairs and Communications (MIC) to enable the identification of individual residents anywhere in Japan based on four basic data (name, address, date of birth, and sex) and the resident register code number (Juki-code). It was intended to fulfil a similar function in local administration services to the employee database in private corporations. It is a very large-scale network computing system connecting 48 prefectures and more than 3000 municipalities. As such, it is the most important system to realize e-Government and e-Local Government, which are targets of the e-Japan strategy. In reality, the Juki-net is in danger of being ignored, if not forgotten, after being dogged by controversy before and after its introduction.3 In this chapter I would like to stand back from my former role as the project manager for Jukinet’s systems integration (until March 2004), discuss the controversy, and offer some suggestions as to how Juki-net might be resurrected.
286
Electronic government in Japan
Juki-net and Juki-card: The system The Juki-net was constructed to provide a standard for resident identification in governmental and municipal administration procedures. Its functions and cutover (phase-in) dates are regulated by the Revised Basic Resident Registers Law, passed in August 1999. The initial phase, for administrative use, was implemented on 5 August 2002. On 25 August 2003, construction was completed, and smartcard services for residents using Juki-cards were simultaneously introduced. There are two types of registration in Japan. One is the Basic Resident Register (jumin kihon daicho) for domicile certification, and the other is the Family Register (koseki) for certification of kinship. In the former, 16 items of information are recorded, including name, present address, birthday, sex, relation to household head, date of latest moving, and resident code. The Juki-net records only six items of data: name, address, birthday, sex, resident code, and data history. Among these, name, present address, birthday, and sex constitute the ‘Basic Four Data’ because most individuals can be specified using this dataset. In case two people with the same name, birthday and sex should live in the same address, however, a Juki-code with 11 digits, created by a random process, is assigned to each resident. The Juki-code provoked a fierce reaction in some quarters that the government was imposing an Orwellian system of state control. While each local government had used a specific resident number for the sake of efficiency in the past, it was not contentious because its use was limited to the local jurisdiction, and was not transferred to another local government at the moving of the resident. To allay Big Brother government control fears, the following assurances were given: 1 The Juki-code can be utilized only by governmental agencies specified in the Basic Resident Registers Law. 2 Use of the Juki-code for the collation of information on individuals is prohibited. 3 Use of the Juki-code by private companies is prohibited. In terms of security, moreover, the system is advanced. Figure 16.1 shows a schema of the Juki-net. It has three levels, consisting of city/town/village systems, prefecture systems, and the national centre system. As mentioned, only six data from the Basic Resident Registers managed by cities/towns/villages are stored in the national centre server via the communication servers in local government and prefecture servers. Each server is protected with a firewall and connected by IP-VPN. All data are enciphered, and all firewalls are checked 24 hours to prevent both illegal network access and physical attack. The Juki-card services started simultaneously with the second phase or cutover on 25 August 2003. The Juki-card is one of the most advanced and
287
Transforming Japan’s innovation system
National centre
Administrative agencies
IP-VPN
Prefecture
Personal information IP-VPN
City town village
Figure 16.1 Configuration of the Juki-net
secure smartcard systems based on the ISO 14443 type-B standard. Its specifications include tamper resistance and firewalls between the application (AP) areas in the IC chip.4 There are two types of Juki-card, one with a photo and one without. The former can be used as an authorized ID card similar to a driver’s licence or passport. The functional construction of Jukicard is shown in Figure 16.2. In the Juki AP area, the Juki-code, password, and symmetric cipher keys are stored. Those are used for resident identification or the issue of a Juki-card. By installing special AP software in the multiuse AP areas, furthermore, local governments can provide additional services, such as the reservation of public facilities, local currency for welfare services, etc.
Present status of the Juki-net One year after the second phase cutover on 25 August 2003, the status and use of the Juki-net could be summarized as follows: 1 Among roughly 3000 local governments, three–Yamatsuri Town, Suginami Ward, and Kokubunji City–refused to connect with the Juki-net. Suginami Ward, in fact, had undertaken litigation to secure the right to decide whether to register its Juki-dataset with the Juki-net. 2 A number of citizen’s groups had also filed lawsuits seeking to prevent the connection of local government resident register systems with the Juki-net.
288
Electronic government in Japan
• Separation of APs
AP
AP
Juki-AP
System manager • AP install • AP uninstall
AP
AP firewall
Operating system Hardware
Figure 16.2 Configuration of the Juki-card
3 Otherwise, utilization of the Juki-net in Japanese government and local governments administrative affairs had progressed steadily. In 1998, ten ministries and administrative agencies were given permission to use Jukinet’s register ID data for 93 administrative procedures. In 2002, three egovernment-related bills were passed in the National Diet which extended the use of Juki-net’s ID data to 264 administrative procedures. Local government regulations, furthermore, were amended to enable the data to be used in 31 local procedures, such as address checks of taxpayers. 4 On the other hand, in spite of MIC’s efforts, including financial assistance for procurement by local governments, the spread of Juki-cards had stalled. According to the Mainichi shinbun (4 July 2004), the number of cards issued by all local governments was about 250,000 at the end of March 2004, which was less than one tenth of MIC’s target figure. Regarding the security of Juki-net, whose apparent weaknesses or security holes had so concerned anti-Juki-net groups, no incidents had been reported. System upgrading, including security patch and anti-virus software installation, had been executed continually. Checklist-based system audits for system configurations and operation procedures had been performed by all local governments, and additionally penetration tests had been carried out to check the safety of network appliances in Sapporo, Shinagawa Ward, and so on.5
289
Transforming Japan’s innovation system
Confrontation between stakeholders Tensions over the construction and implementation of the Juki-net system have been alluded to. The major tensions have been between: 1 Juki-net promoters and anti Juki-net groups 2 Administrative agencies and residents The former have been very intense, have attracted widespread media attention,6 and have fanned tensions in the latter. In particular, disputes over Juki-net security between MIC and Nagano Prefecture were reported by television networks and major newspapers.7 While tensions between administrative agencies and residents have not attracted much media attention, they may in fact be more significant. Let us look at these two confrontations in more detail.
Juki-net promoters vs. anti Juki-net groups The problems with Juki-net, as expressed by Ms Yoshiko Sakurai, one of the most progressive anti-Juki-net campaigners and head of the Anti-National ID Forum, can be summarized as follows: . Japanese residents should not be identified by 11 digit numbers . Juki-net and Juki-cards have fatal security problems . Cost performance of Juki-net is poor. Merits for residents are not clear in spite of the huge expense. The Japan Federation of Bar Associations (Nichibenren), states its reasons for opposition as follows: . All personal information possessed by governmental agencies can be searched using the Juki-code . No legal system to manage large network systems like the Juki-net exists, sanctions against misuse are unclear . Local governments have no right to control the Juki-net Putting cost-effectiveness to one side for the moment, the key concerns boil down to protection of personal information (privacy) and security problems. Anti-Juki-net campaigners assert that the introduction of a national ID system like the Juki-code constitutes an infringement of privacy by governments and governmental agencies, and hence it infringes on fundamental human rights. They have pursued this argument in the courts. It is often held as axiomatic rather than open to logical debate. On the other hand, privacy of personal information is already compromised to some degree, even for such opponents. Driver’s licences are demanded (and often photocopied) even when purchasing a cellular phone, and these show
290
Electronic government in Japan not only the name and address, but also records of violation. The critical issue is how the government and local governments, private corporations or organizations, and individuals handle such personal information from the viewpoint of ‘fundamental human rights’ and ‘public welfare’. Unfortunately, with regards the Juki-net, proponents have not devoted sufficient attention to allaying fears on these issues, and there has not been a constructive dialogue. In the absence of such a dialogue, emotions hold sway, as in the catch phrase: ‘Humans are not lumps of beef. The numbering of residents by the government is not acceptable’.8 Next, concerns over security precede the first cutover on 5 August 2002. Many of the concerns are based on misapprehensions and a lack of technical knowledge, which have fed sensationalist articles in major newspapers. In many cases, responsibility for misapprehensions rests with the MIC itself; its efforts at clarifying and publicizing Juki-net technology security have been woefully inadequate. As a result, it became widely believed that the Juki-net had a number of fatal security problems. The governor of Nagano Prefecture accused MIC of laxness over Juki-net security. This resulted in an open forum about Juki-net security between MIC and Nagano Prefecture, but both sides’ opinions were entrenched and diametrically opposed.9 Such squabbles unfortunately substituted for reasoned debates about security for preventing fraud from the perspective of protecting personal information. And while such squabbles continued, to date the Juki-net has operated without any reported technical security troubles or leaks of personal information. This may mean that the Juki-net has technically sophisticated security functions, and that the security education for local government workers and the system audit directed by MIC are effective. There is no room for complacency, however. The most likely and dangerous sources of security holes and leakage of personal information are likely to come from inside. Preventing these is crucial to gaining acceptance by the majority of residents, even if campaigners will never be placated.
Administrative agencies vs. residents Prof. Takashi Kobayashi of Tokai University, who used to work in the Information Strategy Office of Yamato City, Kanagawa Prefecture, analysed the situation of Juki-card utilization in 248 cities with populations of more than 100,000. His report, which was uploaded on the Nikkei-BP portal site, begins: Not only has the Juki-card lost users, but it has already become a forgotten existence. It seems a midsummer night’s dream, that sensational debate about the Juki-net raged between the anti Juki-net people and the Japanese government around the formal phasein date of Aug. 25th, 2003. The article by the Mainichi shinbun on July 4th, 2004 which reported the number of Juki-cards issued at less than 250,000–less than one-tenth of MIC’s expectation–came as a shock to someone who had worked on the Juki-net in a
291
Transforming Japan’s innovation system local government. Since then, there has been silence. The news value of Juki-net seems to have become lower and lower. Even the anti Juki-net groups seem not so active in their objections to the Juki-card. Most local government workers who are directly involved in the Juki-card may feel that this quiet situation is better. Or, some local governments may claim that the MIC is in charge of issuing the Juki-card, and not them . . . 10
As this report suggests, the Juki-net/Juki-card is not widely accepted by residents. It is a fact, however, that the Juki-net has already become a mission-critical system for many administrative agencies. As mentioned, legislation in 2002 enabled the Juki-net to provide residents’ ID data to central administrative agencies for 264 governmental administrative procedures and, through local government regulations, 31 procedures in local government. While most residents are not conscious of it, the utilization of the Jukinet is advancing. This situation is similar to that of the employee database in corporations, which is an infrastructure of the information system for corporate activities which employees are not conscious of in their ordinary work. They implicitly accept that personal information, including that of salary, health conditions, family, etc. is stored in it, and that the corporation can use it almost without any restriction. Residents, on the other hand, do not necessarily accept such use of their personal information, even if it is strictly limited. Comparisons may be drawn between the Juki-code and the social security number in the United States. Legally neither the Juki-code nor the social security number are ‘National ID’ numbers. According to the Electronic Privacy Information Center (EPIC), most American banks and credit card companies have created a customer database based on social security numbers, and those data have been sold illegally.11 However, the social security number is accepted widely by US citizens because it is used in their daily lives as an important social system. By contrast, the Juki-code is the target of criticism, in spite of strict limitations to its use. Differences between the Juki-net and corporate databases and the social security number system should be elucidated. Why is the social security number indispensable for US residents’ daily life now? How about the Juki-net: does the Juki-net only increase the efficiency of governmental agencies or does it benefit residents as well? Answers to these questions, especially the last, may offer a key to the revival of the Juki-net.
Proposals for the Juki-net As we have seen, the major issues raised by anti-Juki-net groups revolve around privacy protection and Juki-net security. In fact, the latter is not really a technical issue. It derives from a vague distrust of administration agencies. Moreover, this vague distrust has a close relationship to the privacy issue. Just
292
Electronic government in Japan after the first cutover, it was pointed out that a fatal mistake had been made in not disclosing sufficient information to residents on: . . . .
Who can use ID data How it can be used When it is used To whom it is sent
As a result, the procedures were improved by making it possible for residents to ask for disclosure of their Juki-net processing log information. Many residents, however, felt that this was not sufficient. What then is necessary for residents to utilize the Juki-net in safety, comfort, and convenience? A key concept is ‘self control rights of personal information’. This phrase appeared in ‘Privacy Basics: The OECD Guidelines’,12 and was discussed in the National Diet deliberations over the law for protecting personal information. Nonetheless, only one aspect of ‘self control rights of personal information’–the protection of personal information from illegal use by governments, organizations, and corporations–has been emphasized. The purpose of self-control rights is surely not only to restrict the use of personal information by others, but also to provide use of personal information for self-benefit. Consider the following. Moving house involves not just administrative procedures but many other time consuming tasks as well, including notifying the new address to life-line services, credit card companies, banks, and so on. Presently, it is prohibited for private corporations to use Juki-net information. If residents could update their data simply by giving permission for the Juki-net to send the relevant information to designated corporations, using their self-control rights, however, the benefits would be remarkably increased. It would be extremely useful if this could be applied to the many procedures which require a domicile check. Figure 16.3 shows a framework to realize the above-mentioned service. It is a ‘C2G2B’ model, where the Juki-net data (G) is provided to a private corporation (B) under the control of the resident (C). Technically, this service may be realized by using ‘Single Sign-On’, ‘Attribute Exchange’, etc. as specified by the Liberty Alliance.13 Another approach to increasing the benefits for residents is to deploy multipurpose smart cards based on Juki-card technology. The Juki-card can be used as a multi-purpose smart card onto which software applications can be downloaded. Under the present framework, however, this is limited to services specified by local government regulations, and only local governments can issue cards which have Juki-card functions. But software for Juki-cards can be developed by anyone. The Juki-card basic specifications are based on ISO 14443 Type-B, and details of its specifications are open for all smart card vendors. As shown in Figure 16.2, the Juki-card has a very secure functional structure, such as the application firewall. The Juki-AP software
293
Transforming Japan’s innovation system 3
Enhancement of utilization of Juki-net
Data utilized by corporation: • Name • Sex • Birthday • Address • Telephone no. • email • Company name • Office address • Office telephone no. • Bank account no. • Credit card no. etc.
Resident (C)
1
Provision of Juki-data controlled by resident
I) Application for service
VI) Provision of service
III) Request of decision
IV) Decision by resident Juki-net data: Name Sex • Birthday • Address • •
II) Request for Juki-data
Corporation (B)
Juki-net (G) V) Provision of Juki-data 2
Utilization of Juki-data with a sense of security
Figure 16.3 ‘C2G2B’ service model
for Juki-services is completely separated from other applications, and no application shares the data. Present regulations, though, do not permit the placement of Juki-AP data on private smart cards, even those which satisfy Juki-card specifications. Is this sensible? A smart card boom has now come to Japan. The East Japan Railway Company has issued about ten million smart cards. Tokyo Mitsubishi Bank has started a smart card bank service and other major banks are following. NTT DoCo-Mo developed ‘osaifu keitai’–a cellular phone with a smart card chip which enables cashless purchase and ticket reservation services. A huge number of corporations are adopting the ID smart card system based on ISO14443 Type-B. Juki-cards have been issued to only 0.2 percent of the Japanese population, while issues of traffic cards, bank cards, and corporate ID cards are mushrooming. Imagine if Juki-AP software was installed in corporate ID smart cards which could also be used for banking, credit, and travel. Not only would this be convenient, but it would have another merit for cash-strapped local governments. It costs about ¥1000 to issue each Juki-card. By using private smart cards installed with the Juki-AP, not only would the diffusion rate increase, but it could be done with a low fiscal burden for local governments. Of course, some people might be even more anxious about the security of private cards than Juki-cards, but if they are done to the same security specifications such fears might be allayed. In short, a solution to the Juki-net and Juki-card impasse may be found not simply by pursuing ‘self-control rights of personal information’ defensively, but also from the viewpoint of resident self-interest or utility.
294
Electronic government in Japan
Concluding comments We have looked at the clashes between Juki-net promoters and anti-Juki-net groups, as well as administrative agencies and residents. These two confrontations expose a clash between technology solutions and administrative systems, a far cry from the trust and efforts to promote societal support for science and technology called for by Arimoto in this volume. In order to resolve this clash, and to bring about a revival of the Juki-net, two proposals were made. Technically, these two proposals are feasible, even at the present time, but they require additional administrative or legal measures. Under the current framework, utilization of the Juki-net by private corporations is completely banned. And only local governments can issue Juki-cards. Juki-AP software cannot be installed on private smart cards. Rather than allaying residents’ fears over the Juki-net, such a situation may in fact intensify the vague sense of mistrust over the government’s initiative by failing to demonstrate any direct utility to residents themselves. Rather than exacerbating residents’ fears over security, conversely, relaxation of the regulatory framework–though not the security–might in fact demonstrate utility of the system and ultimately allay many of the fears. The MIC needs to reconstruct the Juki-net’s institutional framework to maximize the residents’ benefits by sharing the personal information under their self-control rights, while providing a sense of security. There needs to be an active discussion, involving both partners and opponents, and not merely squabbles. As mentioned above, the Juki-net is based on the Revised Basic Resident Registers Law. In the National Diet deliberations for this law, the importance of the relationship between the Juki-net and the Personal Information Protection Act were recognized. However, handling of personal information in the Jukinet was hardly discussed at all as the scope of the Personal Information Protection Act itself was increasingly narrowed and its purpose changed. Even opponents of the Juki-net agree that some indications of personal information are necessary for daily living. If so, it is necessary to re-argue the following issues: . How to handle Juki-net ID data among the government, local governments, public organizations, private corporations, and individuals from the viewpoint of ‘fundamental human rights’ and ‘public welfare’. . How to protect the Juki-net ID data and the Juki-net system technically and operationally. Unfortunately, barren arguments continue.14 MIC is to blame for the greater part of them. Residents are bemused by the arguments between MIC and the anti-Juki-net groups. MIC should argue positively and productively, and demonstrate the importance of the Juki-net for residents.
295
Transforming Japan’s innovation system The confrontations observed with the implementation of the Juki-net are in fact observed in almost every e-Japan programme. The e-Japan Strategy has been highly rated because of the large increase of broadband and mobile Internet subscribers. But effective IT utilization was given a very low evaluation by the Expert Committee on IT Strategy Evaluation in March 2004.15 Concerning the promotion of e-Government and e-Local Government, especially, this Committee cited not only low utilization of the e-Government portal site for e-applications and e-notifications by residents and corporations, but also low user satisfaction. Japanese residents and corporations, it concluded, have not been able to reap the full benefits of the IT revolution. The increase in broadband use can hardly be attributed to government initiative. It is more due to the huge increase of ADSL subscription carried out by private companies like Yahoo BB, which was able to catch user needs. Parallels exist between the e-Government/e-Local Government systems ignored by residents and the notorious public works schemes–highways in particular–in underpopulated areas where ‘the number of cars is less than the number of bears’ (a famous quip from the Highway Authority privatization debate).16 The reason is clear. Pushing supply-side logic and blindly believing in the efficiency of technology without having to modify local government systems produces similar results. In order to maximize residents’ benefits, the policy making process should be changed to a demand-based logic and harmony between technology solutions and administrative systems pursued from this perspective.
Notes 1. http://www.kantei.go.jp/foreign/policy/it/index_e.html 2. http://www.soumu.go.jp/english/c-gyousei/index.html 3. http://www.japantimes.co.jp/weekly/ed/ed20030705a1.htm and http://www.nichibenren.or.jp/jp/katsudo/sytyou/kaityou/00/2002_8.html 4. MIC notification No.392, 5 May 2003. 5. http://www.soumu.go.jp/c-gyousei/daityo/ 6. http://dailynews.yahoo.co.jp/fc/domestic/resident_register_network/ 7. http://www.pref.nagano.jp/soumu/shichoson/jyukisys/touron.htm and http:// www.pref.nagano.jp/soumu/shichoson/jyukisys/singikai/siryo12-1.pdf 8. Traceability of beef in Japan is carried out by a ten-digit number. 9. http://www.soumu.go.jp/c-gyousei/daityo/ http://www.pref.nagano.jp/keiei/seisakut/happyou/kaiken/s-kaiken.htm 10. http://premium.nikkeibp.co.jp/e-gov/special/2004/sp040901main.shtml 11. http://www.epic.org/ 12. http://www.oecd.org/document/18/ 0,2340,en_2649_34255_1815186_1_1_1_1,00.html 13. http://www.projectliberty.org/
296
Electronic government in Japan 14. http://www.computerworld.com/securitytopics/security/story/ 0,10801,97600,00.html 15. http://www.kantei.go.jp/jp/singi/it2/hyouka/dai4/4siryou1.pdf 16. ‘On a highway in Hokkaido which was built by the favour of a politician, a traffic accident occurred between a car and a brown bear. When investigated, it became clear that the number of bears crossing the highway is more than the number of cars on it’, claimed Mr. Nobuteru Ishihara, Minister of Land, Infrastructure and Transport, at a town meeting held on Nov. 2001.
297
17 Conclusions and reflections: Emergent models D. Hugh Whittaker and Robert E. Cole
The last two decades have seen Japan change from catch-up industrial upstart to a leading–mature–knowledge-based economy. As with other industrialized economies, the transitions have not been easy, economically, socially, and politically. Complicating these transitions was the post-bubble ‘lost decade’ of the 1990s, which prompted a widespread questioning of the ‘Japanese model’. This occurred both at the corporate level and in terms of the national innovation system. In its early, pre-1998 stages, much concern was focused on the health of the financial sector, although some concern was also expressed about declining profits in manufacturing. The questioning coincided with the US resurgence, the rise of increasingly strong Asian competitors, and the global tide of corporate governance reform, which intensified in the wake of the Asian Crisis in 1997. When many manufacturers set about restructuring in 1998, there was an initial emphasis on restoring profitability and reforming top corporate executive bodies and boards. There was more resistance to business media and analysts’ calls for ‘focus’ and disposal of noncore activities; there was a good deal of spin-off and consolidation activity but this was mostly to stabilize markets and cut losses. The path chosen initially for restoring profitability, across a broad span of the production chain, often involved doing more of the same better. The shift of commoditized manufacturing operations offshore, especially to China, accelerated. Managers also worked on modifying employment relations, introducing performance-based management for regular employees, early retirement or voluntary redundancy schemes, and enhancing flexibility through hiring more nonregular workers. Another pivotal period followed in 2001–02, following high profile corporate scandals such as Enron, which took some of the shine off the US model of corporate governance, and the collapse of the IT bubble in 2001. The red ink
298
Conclusions experienced by companies like Cisco and Sun Microsystems had a similar effect (wisely or unwisely) among those Japanese firms considering the embrace of modular manufacturing (Sturgeon, Chapter 3 in this volume). As managers contemplated global competition, and often sliding global market share in the ICT sector, many increasingly came to the conclusion that their future lay in monozukuri manufacturing and the astute management of technology resources, for which they would need to raise their game. Just how to do that in a way that adapted monozukuri to the new competitive environment was not so clear. The ground was prepared for the ‘MOT boom’, which began in 2002. We described this boom in the Introduction; here we will assess the results of the boom, based on the perspectives of the intervening chapters and on some broader perspectives. Coinciding with the MOT boom was a flurry of activity aimed at promoting university–industry linkages, fostering start-ups, technology transfer, and clusters. These initiatives were aimed at jolting Japan’s national innovation system into a more decentralized, entrepreneurial direction, with universities and start-ups playing a much more prominent role, as in the US and subsequently the UK. We offer a provisional analysis of this flurry of activities as well.
Dilemmas and limitations of the ‘knowledge-creating company’ Before we do so, however, we would like to offer some reflections arising from the individual chapters concerning knowledge management in Japanese manufacturing companies. This is not so say that all the strengths and weaknesses can be reduced to knowledge management, but the observations are worth highlighting. They are relevant for understanding the respective merits of the two models which follow. Nonaka and Takeuchi published their influential book The Knowledge-Creating Company in 1995. The research on which the book was based was done mostly in the 1980s when Japanese firms were riding high in global markets. The theoretical models the authors produced were based on the experiences of primarily successful Japanese firms which they believed personified best practices. At the time they published their book, the full depth of Japan‘s problems in high technology industries were not yet known, with much of the blame for its problems at the time, as noted, assigned to the financial sector. With the passage of time, specific limitations have become more visible in the knowledge creating activities of Japanese firms as well as the ability of large hightech firms to convert these activities into profitable commercial products. This should not be surprising, since as we noted in the Introduction, the very success of strong competitive models tends to create blind spots or weaknesses, which become manifest when the competitive environment changes. These limitations can be seen in the observations of several of our contributors. We summarize these limitations as follows roughly using the
299
Recovering from success Nonaka-Takeuchi SECI model (socialization, externalization, combination, internalization–see Yamaguchi, Chapter 9) of the spiral process of knowledge creation and conversion. 1 Japanese manufacturing firms have displayed a tendency to over-emphasize tacit knowledge generated in the first stage, and a reluctance or inability to make such knowledge explicit. Cole’s chapter on software pointed to a strong resistance by individual operating units to standardization through packaged software such as ERP. At least part of this resistance derives from the fear that standardization will result in a loss of local practices conferring competitive advantage.1 However, not all tacit local knowledge is equally valuable and some may reflect outmoded processes which are not subject to scrutiny. Moreover, local optimization and control, and hence scope for benefiting from shared tacit knowledge, often comes at the expense of firm-wide optimization. Modularization presents similar dilemmas. Modularization advances through standardization and formalization or codification of key points in the value chain and in the module itself. Modularization with publicly disclosed standards has been especially threatening to incumbents because it exposes long-term relations with customers and trusted suppliers to broader competition and broadens the basis of innovation. It was primarily this stable (closed) world that was seen by Nonaka and Takeuchi as the springboard for most of the innovative activities they describe. The reluctance of many Japanese companies to embrace a more open model may also be traced to the fear that it will expose their core technology assets to others and devalue their strong coordination capabilities generated in the ‘originating ba’ of socialization (Nonaka and Konno 1998). Hence the preference for integral architecture which requires close cooperation across and among product designers and production units (Fujimoto et al. 2001). Notwithstanding, the growing modularity of many high-tech sectors from computers and semiconductors to telecom–including the Internet–and pharmaceuticals, provides a set of continuing challenges. 2 Large Japanese manufacturing firms have tended to favour incremental innovation rather than discontinuous or ‘paradigm disruptive’ innovation, the latter sometimes associated with the problem of forced externalization. Yamaguchi (Chapter 9) points to the latter problem in R&D, namely failure to establish a ‘field of resonance’ in R&D operations in large firms in which tacit knowledge is shared. In this case, the evolution of tacit knowledge is stifled, through pressures for early externalization and warranting of knowledge. This is somewhat ironic in view of point 1 above, but can be partly explained as a response to increasing risk and expense of R&D operations (managers need to know what to fund, and seek ‘objective’ measures to evaluate R&D expenditures to justify decisions as early as possible), or pressures of
300
Conclusions restructuring, or a combination of both. Some of the problems encountered in this area may also be due to the reduced risk taking that increasingly characterized Japanese management behaviour in the 1990s. On top of these general explanations, however, there is a specific problem when it comes to ‘paradigm disruptive innovation’. Since this involves uncovering new scientific principles, which only certain individuals sense is necessary and feasible, it may be difficult to create shared tacit knowledge through socialization even with fellow researchers, let alone top managers. The processes of socialization and externalization, then, become biased towards ‘paradigm sustaining’ or incremental innovation. Byosiere’s chapter further points to difficulties in creating ‘originating ba’ in R&D when the creativity of individuals is not recognized or even suppressed. Indeed, it seems that the majority of discontinuous innovations that have reshaped global competition in the high-tech arena over the last two decades such as the Internet or organizational and product architectural innovation have come from non Japanese firms. We saw in detail in Cole’s telecom chapter the difficulty key Japanese firms had in coming to terms with the Internet, constrained as it was by long-term trust relationships between NTT and its suppliers, inappropriately high reliability standards, and a limited start-up culture. Despite the celebration of knowledge creating innovation of Japanese firms by Nonaka and Takeuchi, their reported innovations tend to be stand alone innovations which are limited in scope and competence enhancing (rather than disruptive or systemic) in nature. 3 Japanese manufacturing firms have had difficulties in accessing external tacit knowledge outside of existing enterprise groups and long-term relational partners. Nonaka and Takeuchi (1995: 200–12) cite the development of the Nissan Primera to show that Japanese firms are quite capable of absorbing overseas tacit knowledge through a process of socialization. There is evidence to suggest that this case may not be representative, however. Brown’s survey yields a very interesting finding in relation to accessing (and as a consequence, integrating) external knowledge. US semiconductor engineers tended to access private (largely tacit) external knowledge, while Japanese semiconductor engineers tended to access public (codified) external knowledge. This emphasis of the Japanese engineers is in direct contrast to their alleged extensive internal use of tacit knowledge (in the SECI model). This discrepancy is not accidental. The socialization of the engineers into the ‘company as a community’, designed to promote sharing of tacit knowledge, creates social and psychological boundaries.2 These boundaries implicitly represent the extent of belonging, where shared experiences can be expected to generate tacit knowledge. This does not apply to interactions outside such boundaries, although the boundaries can be extended to encompass suppliers with whom the firm has a shared history, common
301
Recovering from success language, and problem solving routines, exchange of personnel, and sometimes, ownership ties (Ahmadjian 2004: 230–4). The strength of ‘company as community’ and its quasi-community enterprise group form (Dore 1973; Inagami and Whittaker 2005) create different boundaries for tacit knowledge sharing from ‘communities of practice’ (Brown and Duguid 2000; Wenger 1998). The latter transcend corporate (or enterprise group) boundaries in localities like Silicon Valley. The links are based on weak ties and the locus is not the firm but the region or industry. The weak ties are generated by fluid labour markets, shared education, and interaction among a variety of institutional actors such as start-ups, universities, and venture capital firms. These different types of community, moreover, may foster different types of innovation. The former, Ahmadjian suggests, have been recreated by Japanese firms in the high-tech sector and tend to generate incremental innovations, whereas the latter–communities of practice–because of the diversity of outputs, enables novel combinations of knowledge, or the creation of completely new kinds of knowledge. Other contributors (such as Probert, Chapter 5) point out that Japanese researchers and engineers are missing out on knowledge generated in global networks, or are accessing it late because of the time it takes for the knowledge to become codified. Even if such knowledge is codified, as Brown points out, the codification is likely to be partial, and of limited value. A related problem, as Cole’s account of the skunkworks team at NTT suggests, is that even if global knowledge networks (and communities of practice) are tapped by individual researchers, they often have trouble getting this external knowledge accepted into their organization, especially if it challenges the conventional wisdom. A further related problem is that the failure to participate in global networks puts Japanese companies at a disadvantage when it comes to standard setting which emerges out of communities of practice (Cole, Chapter 2).3 Large Japanese high-tech firms have built up extensive networks of global R&D sites. Common socialization and training, and reliance on tacit knowledge and unspoken shared mutual understandings on the part of their domestic R&D workforces, however, has made it difficult to fully incorporate and optimize contributions of foreign R&D employees, despite their growing importance.4 While Nonaka and Takeuchi (1995: 82, 86, 103) stress the importance of ensuring ‘requisite variety’ in product development teams, they don’t deal with the very real problems Japanese firms increasingly face in effectively introducing foreign researchers into the mix. 4 Japanese firms display an increasing problem of tapping into global knowledge networks. The previous discussion (in point 3) demonstrates a critical problem in the conception of the SECI model itself, which is induced from Japanese practice. Significantly, the model begins with ‘socialization’ and
302
Conclusions the ‘originating ba’ inside the company, and unfolds in a spiral from this. Increasingly, however, companies are confronted with the need to tap and integrate cutting edge knowledge which has been externally generated, but which is at least partially tacit, and whose usefulness for competitive advantage depreciates rapidly. Insofar as the model accurately captures the knowledge creation process at Japanese companies, they appear relatively poorly equipped to handle this challenge. 5 Japanese top management’s leadership of the innovation process has been problematic. As a by-product of focusing on success stories, the critical role of top management is not fully explored. Emphasis is rightly placed on the importance of middle managers and on top managers setting new directions, creating challenges for the organization and aligning the knowledge ascending from middle managers with corporate strategy. Top management can stifle knowledge creation in the middle of their organizations, however, through creating a climate of risk aversion and caution. This appears to have been a problem, highlighted by Byosiere. The implication of these reflections is that prospects for a revival in corporate innovation are at least partially linked to overcoming these limitations. Thus it is not surprising to find, with regards points 3 and 4 at least, a movement towards greater openness, and with regards point 1 a relative decline in the importance of shopfloor-based activities. As far as we can tell, there is no systematic response to point 2 as yet, despite rhetoric invoking ‘new paradigm’ R&D. There are, however, signs of greater appreciation of, and attempts to address, point 5 at individual companies. There has also been some growth of R&D outsourcing.
Between open and closed innovation: The ‘reformed Japanese/large firm model’ If we conceive an axis with ‘organization orientation’ at one pole and ‘market orientation’ at the other, large Japanese companies could traditionally be located towards the former in terms of employment relations (cf. Brown, Chapter 8). Similarly, interfirm relations could be depicted as ‘relation oriented’ rather than spot contract or market oriented. Indeed many features of the Japanese model exhibited similar orientations, with a ‘logic of commitment’ opposed to a ‘logic of exit’ (Dore 1983, 2000; Kagono and Kobayashi 1994). There does, however, seem to have been some movement in the past few years into the intermediate ground between the two poles, in the market direction (though not nearly as much as in the US and UK). In other words, large companies have been trying to maintain the benefits of community while gaining new flexibility and stimulation from market forces.5
303
Recovering from success If engineers from large Japanese companies are asked which column of Table 17.1 best represents their firm, they have little hesitation about replying ‘the left column’. A lot of researchers, too, have characterized corporate innovation in Japan–especially technology development–as closed, or autarchic (Kneller 2003; Sakakibara 2003). Several chapters in this book, however, suggest that large companies may be attempting to move towards the intermediate ground here as well, in the direction of open innovation pioneered in the US, most notably in Silicon Valley, without going all the way. We have represented this diagrammatically, in Figure 17.1. This figure suggests that there is an emerging ‘reformed Japanese model’ as well as a ‘nascent network model’ in this intermediate ground, the latter being closer to the open innovation pole, but much less clearly formed. We shall look at these in turn, and then speculate on how the two models might interact.
Table 17.1 Open and closed innovation orientations Closed innovation
Open innovation
. The smart people we need work for us or our partners . To profit from R&D, we must discover, develop, and ship it . If we discover it, we will get to market first
. We need to work also with smart people outside the company network . External R&D and alliances can create value for us . We don’t have to originate research to profit from it . Building a better business model wins . Winners make the best use of internal and external ideas . We can profit from others’ use of our IP; we can buy others’ IP if it advances our business model . Can we benefit from hiring those with skills acquired from others
. The firm that gets to market first wins . If we create the most and best ideas, we will win . We must control our own IP so competitors don’t profit from our ideas . We must train our own employees
Source : Adapted from Chesbrough 2003
Reformed Japanese model
Closed innovation (Traditional ‘Japanese model’)
Semi-open, new monozukuri system, group coordination, etc.)
Figure 17.1 Dual innovation system
304
Nascent network model constructive
(Technology oriented relationship? startups, coordinators, university spin-outs
Open innovation (Silicon Valley model)
Conclusions
Reformed Japanese/large firm model Indeed, the chapters in this volume point to a response to the competitive problems on the part of many Japanese companies, in a broadly similar direction, though obviously with varying degrees of change and consistency. By emphasizing similarities rather than differences, we can extrapolate a ‘reformed Japanese/large firm model’ which is not a new model per se, but a series of reforms to the traditional model with the expectation that they will gain some degree of internal consistency. The main features of the model, including some dilemmas and potential impediments to its success, may be summarized as follows: ‘THE SHOPFLOOR’ (1) Under the system of manufacture refined in the postwar years, production was emphasized as the source of value creation, and the ‘soul’ of the company lay in the genba, especially the production shopfloor and directly related activities such as product development and design. Importance was placed on bottom–up improvement activities, such as represented by quality control circles. The strengths of genba shugi are well known, but there were also potential blind spots, including resistance to company-wide optimization (see the previous section, and Cole’s treatment of software in Chapter 6) and a focus on ever-increasing quality and technological sophistication which can get separated from market demands (see Yunogami’s treatment of DRAMs in Chapter 4). When the postwar system was being formed in the 1950s and 1960s, blue collar employees typically comprised half or more of the labour force, and harnessing their potential–creating cooperative industrial relations, training, and motivating them–was crucial to corporate success. Now there are much fewer of them, especially in electronics and ICT companies. They are still prominent in labour unions, but even here their influence is on a downward trend. If the genba is no longer the real ‘soul’ of the manufacturing company, however, sometimes it is no longer clear where the soul resides, or if there is one. The emergent reformed model sees a relative decline in the importance of genba activities. They are still critical, but decisions as to what products to make become as important as how to make them, and even the latter become a matter of strategic choice: outsourcing production is an option, as is spinning out production operations, or merging them with those of other companies, using non-regular workers, or having production lines operated by workers employed by another company (sometimes comprised of former employees). Alternatively, some Japanese high-tech companies seeking to emulate IBM now speak of themselves as service solution providers, though as noted in the Introduction, actual movement in this direction has been limited.
305
Recovering from success M O N O Z U K U R I (2) Recent emphasis on monozukuri (production(ism)) activities within manufacturing companies may be seen to refute these assertions, or alternatively, they may be seen as an attempt to address the spiritual void to which we alluded. Even in companies which have taken on IBM’s rhetoric of being service solutions providers, the norms and the incentives still emphasize the monozukuri ethic of producing and selling hardware, and for the dyed in the wool monozukuri engineer, those doing software and systems integration (involving other companies’ hardware) are frequently viewed as lesser contributors. When Japanese managers in high-tech industries are queried as to the relative importance of product versus business model for future success, they are likely to stress product. One focus of monozukuri activities has been skills. For monozukuri to survive and flourish, there has to be a steady stream of new workers with the necessary skills. However, skill transmission between generations is in question, in view of the ‘2007 problem’ (the onset of retirement of the baby boomers who were critical in establishing the postwar system) and the labour force implications of population decline. These very problems, though, point to the possibility of a further erosion of the importance of the production shopfloor. The shift toward modularized product design and manufacturing in many high-tech sectors has challenged the traditional monozukuri model by devaluing the kind of coordination at which Japanese firms have been so good. Managers are continuing to struggle with this challenge. Can they find markets where their high quality, high functionality products can still command a premium price? Can they maximize margins by creating products that contain internally integral product architectures but which are externally modular (as seen and used by the customer)? Modularized design and modularized production are not technically big challenges for Japanese manufacturers; for many the challenges or dilemmas lie elsewhere–in geographic dispersion of production chains, with implications for the domestic labour force, and its negation of hitherto advantage conferring manufacturing practices. In the ICT sector, an even bigger challenge is constructing business models for modular products based on open systems and open standards. There are relatively few examples of Japanese success globally in doing this. Modularization puts a premium on systems integration skills (Sturgeon’s Chapter 3). It may be that large Japanese high-tech firms’ deep experience and success with building their competitive strength on proprietary corporate technology ill-prepares them for this kind of challenge (see also Best 2001). The danger for some managers is that a monozukuri ‘revival’ will allow them to take refuge in a golden age of productionism. Others recognize that for monozukuri to survive, it must evolve. Hitachi’s Monozukuri Engineering Division has reinterpreted its mission to create a rapid response global resource mobilization and management system (Whittaker, Chapter 11). It espouses productionist values (‘honest work should be rewarded’; ‘making things is
306
Conclusions about making people’), but its focus is no longer the shopfloor: ‘We now have to look at the environment, and globally, on a timely basis’ (interviews quotes, June 2005). In other words, it sees its role as creating a platform for management which draws upon production, rather than simply creating an efficient production platform (which is the point made in 1 above), and implies a shift in the decentralization–centralization balance). TECHNOLO GY DEVELOPMENT (3) The transition from ‘catch up’ innovation to being a front runner (Arimoto, Chapter 13), and competitive pressures, have forced Japanese manufacturers to shift from their traditional emphasis on process technology to expending more resources upstream, on R&D, and new technology development. This increases uncertainty and risk, which is compounded by corresponding market uncertainties downstream, especially when new markets must be created. These uncertainties in general have been one of the driving forces towards open innovation (Chesbrough, Chapter 7), and necessitate different management strategies and skills–‘poker’ in addition to ‘chess’ in Chesbrough’s analogy–and the ability to address false negatives. Some large Japanese companies, simultaneously facing these challenges and deteriorating competitiveness, saw a downward spiral in performance in the 1990s. In response, some closed their central research labs and/or restructured their R&D activities (Yamaguchi, Chapter 9). Others retained their CRLs, but it does appear recently that their role is changing, with attempts to strengthen external monitoring and responsiveness to market signals. This is consistent with a move towards greater openness in innovation, as is the increased emphasis during the recent upswing in R&D spending on strengthening global R&D organization and networks (Nikkei Weekly, 8 and 22 August 2005). Strategies to deal with false negative risk include alliances, mechanisms to identify, transfer and monitor intellectual property and technology within enterprise groups as part of, or ancillary to, core technology strategies. In addition, as Chesbrough and Lincoln (Chapter 12) note, a strategy favoured by Japanese companies has been spin-offs. They were frequently used to take advantage of market growth opportunities during the high growth period, but subsequently have been used defensively to facilitate core firm restructuring. More recently they have been used in conjunction with technology strategies with medium-term time horizons (as seen in the growth of ‘carve-outs’ and carve out funds, discussed below). H R M AN D K N O W L E D G E MA N AG E M E N T ( 4 ) Employment relations have been undergoing change. The psychological or social ‘contract’ of lifetime employment was shifted in the 1990s–not officially and often subtly, but towards a stance nearer ‘employability’. With increasing
307
Recovering from success spin-offs, transfers of undertakings, mergers and acquisitions, secondments and transfers, the chances of employees spending their whole working life in the same company have been reduced. It can be argued that the company’s responsibility in this case is to facilitate career development and raise ‘employability’. But this change also puts greater emphasis on employees taking responsibility for their own career development. Loyalty per se has become less valued, and communitarian aspects of employment have become less encompassing. Probert (2002) has described this as a shift in employment relations from ‘embrace’ to ‘handshake’. Wage and promotion systems have been altered to suppress seniority considerations, and to reflect role, responsibility and current contributions. As Byosiere argues, there is a reinforcement loop here in that high performing researchers, who have seen fewer resources made available to them and fewer opportunities to deploy those resources, and who have sensed a weakening of community as a result of restructuring, now sometimes want greater–and quicker–recognition of their contributions. Finally, homogeneity as a basis for sharing tacit knowledge, promoted through recruitment and socialization, has also begun to change with increases in mid-term hiring, and increased flexibility in work and employment practices, designed in part to increase contributions from female employees. NonJapanese personnel have also been appointed to strategically important roles, although it would be an exaggeration to say that diversity has been embraced. As there has been no explicit break in the psychological contract of employment, moreover, it cannot be said that employment relations have shifted as described in toto, or that ‘communities of loyalty’ have suddenly been replaced by ‘communities of purpose’ (Heckscher 1995). We noted in the previous section a number of dilemmas and limitations of the ‘knowledge creating company’, and mentioned attempts to address at least some of these. Large manufacturers have been seeking to lower corporate boundaries and increase access to external tacit knowledge through a number of means, including mid-term hiring and boundary spanning activities, outlined below. Increased investment has been directed to universities outside Japan. Leading pharmaceutical companies have focused on strengthening and leveraging global R&D networks.6 There are signs that, with a growing shortage of domestic R&D personnel, firms are recognizing the need to draw on the global talent pool, not simply to adapt their home country technology to local markets, but to use it for knowledge creation.7 G R O U P /K E I R E T S U RE LATIO NS (5) Keiretsu relations have weakened at the macro-level, especially for R&D-related activities, and at other levels they are undergoing transformation. They have become less based on status or legacy (a development parallel to the decreased
308
Conclusions emphasis on loyalty in employment relations), and more conditional and strategically oriented, serving a competitive purpose (Lincoln, Chapter 12; see also Lincoln and Gerlach 2004). While in some cases technology and competitive strategies of core companies have resulted in strains and a weakening of keiretsu relations (cf. Toyota and Denso), in other cases they have resulted in strengthening relations (e.g., Matsushita). Interestingly, even car manufacturers taken over by foreign companies in the 1990s have reversed the process of dissolving keiretsu ties: in 2005 Mitsubishi Motors re-launched its keiretsu group, Nissan increased investment in, and Mazda increased technology ties with, key suppliers (Asahi shinbun, 26 June 2005). Restructuring has resulted in a redistribution of business activities among group companies in the electronics sector–generally from inner core to outer core companies–but subsequently, beginning with the quest for synergies in business activities, there has been a greater interest in strategic collaboration, which has been extended to technology and the mobilization of IP resources. Management and IT systems are being developed to achieve this (Whittaker, Chapter 11). Enterprise group or keiretsu relations (and associated personnel assignments) span corporate boundaries but are not fully ‘open’. They are by their very nature an intermediate form of organization between markets and hierarchies, and they will be an important feature of semi-open innovation, perhaps in the ‘galaxy’ form suggested by Whittaker in which coordination is carried out by core firms, but not exclusively.
AL LIA NCE S (6) Strategic interfirm relations are not limited to group or keiretsu companies, moreover. Amidst restructuring around the turn of the century, electronics companies began to create alliances, many with erstwhile domestic rivals, in a way inconceivable just a few years earlier (Sturgeon, Chapter 13). Some were for consolidation, as in semiconductors, some were for multiple purposes such as R&D cost savings and complementarities and procurement cost savings, and others attempted to create industry standards. They, too, attest to a decline in legacy keiretsu ties, and a growing strategic emphasis on interfirm relations. In some cases they also involve outsourcing of manufacturing and OEM contracts. (A Japanese twist to this takes the form of mutual OEM alliances.) Few companies have used alliances as strategically as US firms like IBM, however. When asked, Japanese engineers in large firms will say ‘we [ Japanese firms] are bad at alliances’, a common view which flows from their firms’ relatively inward looking orientations and practices. For many, at a deep level, alliances are a necessary evil which entail a risk that someone else will end up profiting from their technology. Many still find it hard to accept and
309
Recovering from success act on the fact that their products have limited value when used alone, but gain in value when used along with complementarities provided by partners (Gawer and Cusumano 2002). Sony is a company with deep experience in alliances but it has had great difficulty using even its own complementary businesses (media entertainment business) to enhance the development of its electronic products (most notably an i-Pod like product). U N I V E R S IT Y – I N D U S T RY RE L AT I O N S ( 7 ) U–I relations have become more diverse and organized. Faced with a pinch between the increasing importance of R&D activities, increasing uncertainty, risk, need for speed, and a recognition that they can no longer do it all themselves, large Japanese companies have raised their funding for university research. While investing more funds abroad, they have also been actively engaging with domestic universities. They have been exploring these relations for newly discovered education needs as well. This modifies their former selfreliant stance, but problems must be overcome in order for these relations to expand further. A large number of universities, supported by government funds, have opened up MOT programmes. Many are suffering from a shortage of students, along with a shortage of trained teachers truly knowledgeable in principles guiding the management of technology (Kaneko et al., Chapter 15). Some have recruited as professors retired company managers who share their ‘memories’ and ‘war stories’ but have little to offer in the way of an analytic framework. Moreover, some managers are reluctant to see employees take time off to attend classes because of the press of daily work. These are real human resource bottlenecks that slow the change process.8 M A NA G E M E N T A N D CORPORATE GOVERNAN C E (8 ) Management and corporate governance are undergoing change as well, although the latter has not been addressed in this book. Facing global competition, senior managers have been under pressure to react quickly and decisively. (For the most part, these managers are still recruited internally, but in some cases the necessary skills sets have not been developed internally–for change management, or dealing with M&A, for instance–and they are being supplemented by external hires.) They also face new demands over the environmental impact of their activities, corporate social responsibility, and shareholder interests. The decline in stable, mutual shareholding on the one hand, and the rise of foreign shareholders on the other, as well as high profile cases such as Livedoor’s attempt to take over Nippon Broadcasting Corporation and Rakuten’s attempt to take over TBS in 2005, have brought an unprecedented tension into shareholder relations.9 It is still too early to say how far this will pull such managers towards shareholder favouring capitalism, but their
310
Conclusions preferred position is to balance stakeholder interests, maintaining sufficient leeway to keep shareholders at bay, while retaining sufficient funds for investment in future technology and business development; in other words, a shift towards the middle ground, but not too far (see Jacoby 2004).10 In sum, we see a number of shifts among large manufacturers–changing decentralization–centralization balance, movement along the community/organization–market continuum, and along the closed–open continuum–with enough consistence internally, and sufficient prevalence across companies, to merit the label ‘reformed model’, although in many cases, of course, change is partial. What evidence is there that MOT concepts have played a part in these developments? This is hard to quantify since there is no MOT canon, and the mechanisms by which MOT concepts have entered companies and been diffused are not always clear. Media attention means that some of the concepts have been ‘in the air’. We have to concede that our evidence is limited, except for Hitachi, where MOT assumed increasing importance in taking forward the reforms after 2002. Companies’ responses to the MOT Consortium and executive seminars on MOT, too, suggest that MOT concepts have played a part in ‘unfreezing’ earlier practices and fostering recognition that researchers and engineers need new skill sets, including new business model concepts to deploy with technology. They have no doubt played a part in the emergence of IP strategies, and it is likely that they have played a part in encouraging greater openness in technology development. We can say with reasonable confidence, therefore, that the impact of MOT, while difficult to quantify, is real, and continues to develop along with the unfolding of the reformed model. Finally in this section, we should consider the role policy and legislation have played in these shifts. First, a series of laws was enacted or revised to facilitate–encourage–corporate restructuring, including a removal of the prohibition on holding companies (1997), simplification of procedures to divide companies (2000) and its accompanying labour legislation (2001); and a wide range of measures under the Industrial Regeneration Law (1999). Second, numerous labour laws were enacted or revised, bringing a shift from promoting long-term employment to a more neutral stance (Araki 2002). Third, there have also been numerous revisions to the Commercial Code, one of the effects of which has been to erode the former German basis of company law in favour of an American principle (Iwahara 2000). The Commercial Code changes in 2002, in particular, made it possible to reform boards of directors and auditors, and offered the choice of a US-inspired corporate governance system, which Sony and the general electric companies Hitachi, Toshiba, and Mitsubishi Electric opted for (though not many other companies). Finally, we have noted the passage of the Monozukuri Law (1999) and a series of measures to train and retain engineers in industry, to educate 10,000 MOT specialists, and specialists for other areas of focus such as biotechnology.
311
Recovering from success Even a cursory examination of these laws and policies reveals that they are not always mutually or philosophically consistent. When considered alongside others such as financial liberalization and related policies pursued under the Koizumi government, they suggest an urge to dismantle the ‘Japanese model’ and promote ‘global standards’ on the one hand, and a weaker urge to shore it up, with modifications, on the other. They have prodded large Japanese companies to change, but the main impetus for constructing a ‘reformed Japanese model,’ in a proactive sense, has come from the companies themselves. In fact, it is possible to argue that the main focus of much policy effort is not the ‘reformed Japanese model’ at all, but the ‘nascent network model’ of Figure 17.1, which we shall now examine.
The nascent network model The ‘nascent network model’, as its name suggests, exists only in nascent form, but it has been a focus of much policy activity in recent years. Inspiration for the policy makers has been largely drawn from outside Japan–from Silicon Valley, Austin and other regional high-tech concentrations with dynamic start-ups and university and corporate spin-outs. If the inspiration is drawn from abroad, however, the model is attractive because it appears to offer the prospect of a decentralized innovation system with substantial bottom–up initiative through start-ups and diverse forms of entrepreneurship, and through this, increased economic and social participation. Achieving this would alleviate a number of ailments, including de-industrialization and depopulation in some regions. Japan already has a huge small firm sector, parts of which exhibit considerable dynamism, but there are three major problems for policy makers as far as this base is concerned. First, Japan’s pattern of industrialization saw much of its industry concentrated in the major metropolitan centres. Industry spread to the hinterlands during the postwar high growth period through to the 1980s, but regional factories subsequently bore the brunt of the shift of manufacturing abroad in the second half of the 1980s and the 1990s. Thus, there is often a weak base for the development of new regional industry. Regional stagnation, unchecked, feeds upon itself, as young people leave in search of jobs (or stay and do not work, or engage in crime), accelerating ageing and the decline of services and community. Second, in contrast to some of the other major economies which saw an upsurge in start-up activity during the 1980s (Sengenberger et al. 1990) and 1990s, Japan’s start-ups rate declined progressively from the early 1970s through to 2001. Conversely, closures increased, and in the 1990s overtook start-ups in many industries as postwar founders retired, or fell victim to the
312
Conclusions shift of manufacturing abroad or the declining competitive position of larger firms they subcontracted for (cf. Whittaker 1997). Third, the proportion of private sector R&D expenditure accounted for by SMEs, officially at least, is relatively low, especially compared to the US. They also file a lower share of patents (12 percent compared to 40 percent in the US: Odagiri forthcoming). As a result of gradually accumulated know-how some have advanced process and product technology, but the government is keen to promote ‘science-based’ start-ups as a source for new industry. Thus there have been concerted efforts not just to shore up regional industry and SMEs, but to create regional-based innovation centres through cluster projects, regional U–I summits, etc. from 2001, attempts to increase start-ups following the ‘SME Diet’ and associated legislation in 1999, and the goal of doubling the start-up rate in five years and raising the number of university spin-offs to 1000 within three years under the Hiranuma Plan.11 When put alongside the substantial increase in public funds being made available to science and technology and related attempts to strengthen the science base since the passing of the Science and Technology Basic Law in 1995 (including the Basic Law on IP, 2002), it all adds up to an ambitious attempt to reshape Japan’s innovation system. The university spin-off target was reached (with a little help from some universities discovering spin-offs they had previously overlooked, thus bolstering their U–I credentials). Science-based start-ups have increased, raising for instance the number of biotech start-up companies from 60 in 1998 to 387 in 2003.12 At the time of writing, MEXT and METI were evaluating the first phases of their cluster projects. They will probably find some initial success–as in the Technopolis projects before them, where there are already concentrations of industry and networks to build on–but the success is likely to be patchy, and while recent figures are lacking, there is little chance that the start-up doubling goal will be achieved, despite tax incentives and removal of the capital constraint for starting a new business. This points to problems in realizing the alternative model through policy leadership. These problems start with policy makers themselves; changing the innovation system requires that they change, too, as Kita (Chapter 16) argues in the context of e-Government/e-Japan.13 Jurisdictional boundaries solidified during the postwar high growth period and subsequently when the former innovation system flourished. Current circumstances require a more open and collaborative system not just for companies, but for the ministries as well. Nowhere is this more apparent than in the role of universities and the promotion of science-based industry. Bringing universities into the innovation system means METI’s interests overlap with those of MEXT. We noted METI’s attempt to create an MOT MBA in all but name, as part of its drive to educate MOT specialists. This was fertile ground for collaboration between the ministries, but little appears to have occurred.
313
Recovering from success In this context, we should note that attempts to reshape the innovation system have shown that skills or human resources are a crucial bottleneck, and they have emerged as a key area of policy effort. One manifestation is the promotion and rapid growth of professional schools–law schools, business schools, etc.–and other graduate programmes at the graduate and mature student level (cf. Arimoto, Chapter 13, and Kaneko et al., Chapter 15). Attempts to strengthen biotechnology, nanotechnology, IP, MOT, and monozukuri all have a substantial training and education component, which are being addressed at this level. Until the 1990s, such education needs could be taken care of by undergraduate programmes and in-house company training and education, but this is no longer the case. Even for METI, the significance of universities lies as much in their role in providing skills and competent, creative professionals as in technology transfer.14 If the supply of human resources is a major bottleneck, it is just the tip of the iceberg for matters which require a coordinated approach. As Arimoto notes, there has been declining youth interest in science and engineering, and growing public scepticism that science and technology by themselves will improve living standards. For this reason the new S&T Basic Plan of 2006 places greater emphasis on quality of life, security, and cultural fulfilment. Here security is used not just in the sense of defence, but as Murayama (Chapter 14) notes, anzen-anshin, a more comprehensive concept involving a sense of well-being. Ultimately, however, even if a more joined-up approach can be constructed in pursuit of these goals, serious obstacles lie in the path of their implementation. They cannot be achieved by top–down policies alone, but can only come about by initiatives on the ground which tap into the spontaneous energy of wide segments of the population. We have noted obstacles in the way to achieving the goal of education 10,000 MOT specialists. There are similar problems in monozukuri–attracting young people to acquire craft or engineering skills, and to stay at jobs which require them–and entrepreneurship as well. That Japan is far from stimulating the energy and participation of large segments of the population is even more clear from the rapid growth over the last ten years in the number of ‘freeters’ (generally young males and females aged 15–34 in temporary or part-time jobs), those of the same age completely unemployed but looking for work, and NEETs (young people not in employment, education or training).15 The harsh job search environment over the decade took a toll on a large number of young people who appeared to have lost or were unable to realize their career ambitions (though not all freeters fall into this category, and youth unemployment has started to fall). Moreover, with Japan having one of the lowest birth rates among OECD countries, ‘social sustainability’ is as much of an issue as environmental sustainability (cf. Whittaker 2004). On the other hand, changes in social relations in Japan could support the nascent network model in a number of ways. Organizations tend to adopt
314
Conclusions patterns of social relations which reflect those of the wider society at the time they are formed (Stinchcombe 1965; Dore 1973). This was true of the community model of the Japanese firm, but Japanese society has been changing. There is less attachment to the notion of lifetime employment in a single firm, and greater receptiveness to alternative career paths and identities, which could promote networks–communities of practice–which cross company and other organizational boundaries. This applies to attitudes towards software, modularization, and standardization as well. Thus some of the problems identified in our discussion of the ‘knowledge creating company’ and only partially addressed under the large company model may well be addressed more effectively under the nascent networking model. Cluster and U–I promotion activities appear to be accelerating the construction of such alternative networks on a local and regional basis. Then what of the relationship between the two models? By definition, large firms do not play a core, coordinating role in the nascent network model. It is unclear to what extent systems integration practices described by Sturgeon (Chapter 3) will diffuse to Japan, but coordination in the network model is likely to be achieved by a combination of market signals, flexible networking, and/or specialist small firm coordinators. Different activities can be coordinated by different actors. Competition arises between these variations, potentially leading to further innovations in systems integration, which could confer upon the network model further advantages to those outlined in the Introduction and Chapter 3. A re-invigorated large firm sector, however, could exploit the resources and dominate coordination of the nascent network model, much as it was accused of doing under the so-called dual economy up until the end of the 1970s, with the effect of stunting its growth. Alternatively, specialization and greater openness of the reformed large firm model could lead to a more constructive and flexible pattern of relationships. The latter possibility is suggested by the recent popularity of ‘carve-outs’–a category of spin-offs–and carve-out funds. The Development Bank of Japan and Mitsubishi Corporation started an Innovation Carve-Out Fund in April 2005 to fund and support technology spin-outs. What is novel is corporations’ understanding the need for outside organizational specialist partners who can help with the whole process of selecting and spinning off promising ventures. The Hitachi Group launched a fund with the Organization for SME and Regional Development shortly after: The Hitachi Group is in the top class, with 73,000 domestic patents. But there are many technologies it has no prospect of using internally. SMEs will be encouraged to develop new business with them. It is hoped that if this succeeds, it will create a new kind of industrial structure linking large and small firms through (technology and product) development. (Asahi shinbun, 19 August 2005)
315
Recovering from success
‘Kyoto model’ Some have held up Kyoto as a model of enterprise, and an alternative to the traditional Japanese (or ‘Tokyo’) model. The ‘Kyoto model’ is a reasonably close approximation of the network model, and Kyoto’s leading electronics companies experienced contrasting performances with the Tokyo-based giants during the 1990s. The essence of the ‘Kyoto model’ is specialization, technology focus, international orientation, rapid decision making, profitability, and arguably network relations and flexible relations with universities (there are over 30 universities in the city; cf. Suematsu 2002; also Ishikawa and Tanaka 1999). In other words, it is closer to the open innovation pole than the giant companies with their enterprise groups. The analytic principle here is that specialization allows a deepening of knowledge but to be effective in the marketplace specialized products require integration with complementary assets to provide customers a ‘whole product’ and ‘one stop shopping’. This is the role of networks (Piore 1992: 430–44; cf. Suematsu 2002) which, as noted in the Introduction, has been increasingly prominent in the ICT sector. Beneath the top tier of companies said to exemplify the Kyoto model–Omron, Kyocera, Murata, Rohm, Nidec, and Samco are frequently mentioned, and sometimes others such as Shimadzu, Horiba and Nichicon–moreover, there are many niche firms which are not start-ups or university spin-offs, but which have quietly accumulated technological expertise over many years, making specialized process equipment, for instance for semiconductor production, which they sell to global companies. These smaller Kyoto companies are specialized, too, and often outsource a lot of their own manufacturing through local or more dispersed networks, and sell a relatively high proportion of their products in global markets from early stages of their development.16 Kyoto is perhaps the most well-known city for these phenomena, but to varying extents they can be found elsewhere as well, where small firms have reduced reliance on subcontracting through developing their own products, or were never involved in subcontracting. ‘Global niche’ (cf. Probert’s Chapter 5 on pharmaceuticals) has become a buzzword in recent years. We should note, however, that even Kyoto does not have a high start-up rate; in fact it was well under the prefectural average in 1999–2001, and closures were above the average (Kokumin kin’yu koko 2004: 304). In his concluding section–‘Towards a new innovation system’–Yamaguchi identifies tensions between innovation in large firms, and what we have called the nascent networking model. He expects paradigm disruptive innovation to come from the latter, in which ‘fields of resonance’ can more easily be created within firms and across them. There is a tension in his analysis between the creativity of individual researchers, and the bureaucratic structures of large firms which often stifle this creativity. Byosiere adds that on top of growing
316
Conclusions rigidities, there are the constraints imposed on researchers by management’s risk avoidance which saps individual initiative. In these views, Japan’s success up to the late 1980s was not simply attributable to a clever catch-up game, but that as aggressive challengers, manufacturing companies were better at harnessing individual creativity than when they matured. There is probably a range of views among the contributors as to whether the ‘reformed Japanese model’ is capable not just of recreating community, but also a better balance between individuals and that community. Yamaguchi and Byosiere are implicitly somewhat pessimistic on this score, as well as whether a constructive balance between the two models can be achieved. Whittaker is somewhat more optimistic on both scores. To achieve a constructive balance is one thing, to achieve competitive success is another. Are the ongoing corporate reforms sufficient to enable Japanese high-tech manufacturing firms to return to the challenger status that they held up through the 1980s, or will they continue to defend existing market positions and extend existing technologies? We hesitate to predict, though it does seem quite unlikely that they will be able to recreate their triumphant status of the late 1980s. Too much has changed for that to happen, including new powerful actors coming on stage. They can, however, realistically aspire to playing a major role in critical component and stand alone device markets and in selected materials markets. Indeed, some are already doing this. They are well positioned to play a leading role in emergent nanotech technologies from both a process and product perspective. A number of Japanese firms will undoubtedly play a dominant industry role in the global automotive industry for some time to come, building on traditional monozukuri strengths and adapting to changing competitive conditions. Even here, though, they could be challenged should, for example, electric motors become commercially feasible. The possibility of a Chinese challenge in this sector cannot be ignored. The nascent network model will bear some fruit, but it is still too early to say just what type.
Notes 1. This does not mean that there is a universal aversion to standardization–the Toyota production system produced some very powerful tools to advance this process, which were diffused to varying degrees in Japanese manufacturing companies, but in ways partially compatible with local optimization (and control). Nor was there much resistance to factory automation per se. 2. There is an extensive literature on sharp distinctions drawn between ‘inside’ and ‘outside’ relations in Japanese society and organizations: Nakane 1970; Sugimoto 1997. Boundaries have been flexible, however, in that they are extended to encompass other parties through ‘relational contracting’ (Dore 1983), especially but not exclusively within keiretsu.
317
Recovering from success 3. When they are at the cutting edge and competing fiercely, the chances of competing standards increase, as has recently been the case with next-generation DVDs. 4. Interview with Nomura Research Institute consultants specialized in global management, 8 Sept. 2005. 5. This argument is more fully set out in Inagami and Whittaker, 2005, ch. 14. 6. Cf. Nikkei Weekly, 8 August 2005: ‘Global R&D next big step for drug firms’. 7. Cf. Nikkei Weekly, 17 January 2005: ‘Firms tap overseas engineers to hone edge’. 8. Interview with Shuzo Fujimura, Tokyo Institute of Technology and Norimasa Fujii, MOT Project Office, Mitsubishi Research Institute, Berkeley, CA, 12 August 2005. 9. In March 2005 foreign investors owned more than 30 percent of shares in 104 listed companies, up from 75 a year earlier. Some of these were technology companies, such as Canon (51.7 percent), Nitto Denko (49.5 percent) and Rohm (48.7 percent): Nikkei Weekly, 4 July 2005. 10. Cf. Keizai doyukai’s 2003 White Paper, entitled, ‘Market Evolution and Corporate Social Responsibility Management: Toward Building Integrity and Creating Shareholder Value’. 11. Hiranuma was the METI Minister in 2001. Legal measures to strengthen new business and start-up support include overhaul of the SME Basic Law (1999), the Law on Supporting Business Innovation of SMEs (1999), and its successor (which also superseded the SME Creative Activity Promotion Temporary Law of 1995) the Law for Facilitating the Creation of New Business (2005). 12. Japan Biotechnology Association, cited by Odagiri forthcoming. 13. Kita’s chapter also points to an even more generic problem–the failure in many service industries, including public services, to adopt a customer-oriented perspective. Cf. Clark and Kay (2005), who argue that this is the product of a ‘hardware’, ‘manufacturing’ orientation, even in services. 14. Other ministries such as MHLW also have strong interests in skills, too, of course. 15. Freeters are estimated to total some 1.92 million, the unemployed some 1.68 million and the NEETs 640,000, with some estimates being even higher (Japan Institute of Labour Policy and Training 2005: 2). 16. This is in part because they have difficulties, without a track record, in selling to large Japanese companies. These observations are based on interviews in 2004–05.
References Ahmadjian, C. (2004). ‘Inter-organizational Knowledge Creation: Knowledge and Networks’, in H. Takeuchi and I. Nonaka (eds.) Hitotsubashi on Knowledge Management, Singapore: Wiley. Araki, T. (2002). Labour and Employment Law in Japan, Tokyo: Japan Institute of Labour. Best, M. (2001). The New Competitive Advantage: The renewal of American industry, New York: Oxford University Press. Brown, J. S. and P. Duguid (2000). ‘Mysteries of the Region: Knowledge dynamics in Silicon Valley’, in C. M. Lee, W. Miller, M. Hancock, and H. Rowen (eds.) The Silicon Valley Edge: A habitat for innovation and entrepreneurship, Standford: Stanford University Press.
318
Conclusions Burt, R. ([1993] 2000). ‘The Network Entrepreneur’, in R. Swedburg (ed.) Entrepreneurship: The social science view, Oxford: Oxford University Press. Chesbrough, H. (2003). Open Innovation: The new imperative for creating and profiting from technology, Boston: Harvard Business School Press. Clark, T. and C. Kay (2005). Saying Yes to Japan: How outsiders are reviving a trillion dollar services market, New York: Vertical. Dore, R. (1973). British Factory–Japanese Factory: The origins of national diversity in industrial relations, London: Allen & Unwin. —— (1983). ‘Goodwill and the Spirit of Market Capitalism’, The British Journal of Sociology, 34 (4): 459–82. —— (2000). Market Capitalism, Welfare Capitalism: Japan and Germany versus the AngloSaxons, Oxford: Oxford University Press. Fujimoto, T., A. Takeishi, and Y. Aoshima (2001). Bijinesu akitekuchia, seihin, soshiki, purosesu no senryakuteki sekkei nijumon (A Primer on Business Architecture, Strategic Design of Products, Organizations, and Processes) Tokyo: Nihon keizai shinbunsha. Gawer, A and M. Cusumano (2002). Platform Leadership: How Intel, Microsoft, and Cisco drive industry innovation, Boston, MA: Harvard Business School Press. Heckscher, C. (1995). White Collar Blues: Management loyalties in an age of corporate restructuring, New York: Basic Books. Inagami, T. and D. H. Whittaker (2005). The New Community Firm: Employment, governance and management reform in Japan, Cambridge: Cambridge University Press. Ishikawa, A. and K. Tanaka (1999). Kyoto moderu: Gurobaru sutandado ni idomu nihonteki keiei senryaku (The Kyoto Model: Japanese-style management strategies challenging ‘global standards’), Tokyo: Pearson Education. Iwahara, S. (2000). ‘Reform of Company Law’, Junkan shoji homu (Commercial Law Review), August, 1569: 4–16. Jacoby, S. (2004). The Embedded Corporation: Corporate governance and employment relations in Japan and the United States, Princeton: Princeton University Press. Japan Institute of Labour Policy and Training (2005). Labour Situation in Japan and Analysis: Detailed exposition 2005/2006, Tokyo: The Japan Institute of Labour, Policy and Training. Kagono, T. and T. Kobayashi (1994). ‘The Provision of Resources and Barriers to Exit’, in K. Imai and R. Komiya (eds.) Business Enterprise in Japan, Cambridge, MA: MIT Press. Keizai doyukai (2003). Dai 15 kai kigyo hakusho: Shijo shinka to shakaiteki sekinin keiei (The 15th Corporate White Paper: Market Evolution and CSR Management), Tokyo: Keizai Doyukai. Kneller, R. (2003). ‘Autarkic drug discovery in Japanese pharmaceutical companies: insights into national differences in industrial innovation’, Research Policy, 32: 1805–27. Kokumin kin’yu koko (ed.) (2004). Shinki kaigyo hakusho (White Paper on Start-ups), Tokyo: Chusho kigyo risachi senta. Lincoln, J. and M. Gerlach (2004). Japan’s Network Economy: Structure, persistence and change, Cambridge: Cambridge University Press. Nakane, C. (1970). Japanese Society, London: Weidenfeld and Nicolson. Nonaka, I. and H. Takeuchi (1995). The Knowledge-Creating Company, Oxford: Oxford University Press.
319
Recovering from success Nonaka, I. and N. Konno (1998). ‘The Concept of ‘‘Ba’’: Building a foundation for knowledge creation’, California Management Review, 40 (3): 40–54. Odagiri, H. (2006). ‘National Innovation System: Reforms to Promote Science-based Industries’,. in T. Shibata (ed.) Japan, Moving Toward a More Advanced Economy: Assessment and Lessons, Washington DC: World Bank Institute. Piore, M. (1992). ‘Fragments of a Cognitive Theory of Technological Change and Organizational structure’, in N. Nohria and R. Eccles (eds.) Networks and Organizations: Structure, Form and Action, Boston: Harvard Business School Press. Probert, J. (2002). ‘Organizational Change and the Strategic Renewal Process: Innovation, stability and inertia in Japanese companies’, PhD dissertation, Cambridge: University of Cambridge. Sakakibara, K. (2003). ‘Kenkyu kaihatsu wa keiei seika to musubitsuiteiruka’ (Is R&D Linked With Management Results?), Discussion Paper 03-01, Gijutsu kakushingata kigyo sosei project, Keio University. Sengenberger, W., G. Loveman, and M. Piore (1990). The Re-emergence of Small Enterprises, Geneva: International Institute for Labour Studies. Stinchcombe, A. (1965). ‘Social Structure and Organizations’, in J. March (ed.) Handbook of Organizations, Chicago: Rand McNally. Suematsu, C. (2002). Kyo yoshiki keiei: Mojuruka senryaku (Kyoto-style Management: Modularization strategies), Tokyo: Nihon keizai shinbunsha. Sugimoto, Y. (1997). An Introduction to Japanese Society, Cambridge: Cambridge University Press. Wenger, E. (1998). Communities of Practice: Learning, Meaning, and Identity, New York: Cambridge University Press. Whittaker, D. H. (1997). Small Firms in the Japanese Economy, Cambridge: Cambridge University Press. —— (2004). ‘Unemployment, Underemployment and Overemployment: Re-establishing social sustainability’, in Japan Labour Review, 1 (1): 29–38.
320
Index
Accenture 53, 63, 64 Acer 52 adjustment difficulties, Japan 13–14 Agilent 220 Ahmadjian, Christina 216 Akasaki, Isamu 170–1, 180 Alcatel 109 Alizyme 99 alliances/partnerships electronics industry 56–8, 66 pharmaceutical industry 88, 90 reformed Japanese/large firm model 307, 309–10 Amano, Hiroshi 171, 172, 180 Amazon 53 American Online (AOL) 262 Amgen 131 Amkor 148 Anderson, P. 175 anzen-anshin techno-systems 264–6, 267, 273 Aoki, Masahiko 118 APIs (Application Programmer Interfaces) 122 Arimoto, Tateo 24, 271, 295, 307, 314 ARPANET 37 artificial intelligence 264 Asahi Glass 51–2 Asai, Shoichiro 206 Ascend Communications 262 Asian Crisis (1997) 298 Astellas 97, 98, 101, 102 AT & T 33, 35, 44, 130, 131, 142, 220, 259, 262 ATM (Asynchronous Transfer Mode) technology 21, 35–6, 37, 38–9, 40, 41, 42, 43, 44
Au 110 Autodesk 51 automobile industry 2, 49, 220, 317 and Hitachi 206–8 and the emulation challenge 6 Toyota and Denso 225–8, 309 Bain Drug Economics Model (2003) 90 Baldwin, C. 51 Basic Law for the Promotion of Monozukuri Base Technology 20 Basic Plan for Science and Technology 248, 253–4, 265, 266, 271, 314 Bayh-Dole Patent Act (1980) 18, 262 Bell Labs 131 Berger, S. 47 biotechnology industry 22, 101, 222, 311, 314 and industry–university linkages 93–5 and Japan’s innovation policy 240, 248, 249 and open innovation 131, 141 pharma-biotech alliances 99–100 BizTech magazine 189 blue LED innovation 23, 135, 166, 170–4, 180, 184 BMS 102 boundaries, for tacit knowledge 301–2 Brooks, Harvey 257 Brown, Clair 22, 133, 142, 195–6, 301, 302 Budapest Declaration 247 Building Systems 207 Busicom 6, 72, 131, 134, 136 buyer–supplier relationships, and the Japanese Production System 49–50, 51 see also Keiretsu networks
321
Index Byers, Horace 185–6 Byosiere, Philippe 23, 24, 252, 316–17 CAD (Computer Aided Design), Japanese rate of adoption 112 Cadence 51 capital expenditure in Japan 5 career development 147, 149–50, 159–60, 162, 243–7, 284 Casio 59 catch-up, and Japan’s innovation policy 238, 249–51 CEDDs (Centres of Excellence in Drug Discovery) 100 Celestica 52 cell phones 7, 43, 44, 57, 59 CEO appointments, age of 14 Chesbrough, Henry W. 22, 51, 87, 133, 134, 135, 145, 187, 222, 307 China contract manufacturers 52 and global telecom markets 34 Hitachi investment in 211 and innovation 133 and Japan in the 21st century 238, 245 and the Japanese electronics industry 55, 56, 62, 65 and the Japanese semiconductor industry 77 and Matsushita’s 223 MOT programmes 279 and mu-chip technology 204 scientific researchers 244 shift of manufacturing operations to 6, 25, 298 Christensen, C. 166, 167–8, 174 Chugai 89, 99 CIOs (Chief Information Officers), and IT investment 116 Cisco 32, 39, 40, 44, 52, 53, 55–6, 64, 65, 131, 262, 299 Clark, K. 51 class divisions in Japan 239 closed innovation 130–2, 303–4 open–closed innovation 140–1 shift to open innovation 132–3
322
clusters 299, 313 clusters, regional 19, 237, 242, 250–1, 299 Cold War, and US and Japanese technosystems 256–8, 263, 264 Cole, Robert E. 15, 21, 22, 83, 137, 142, 228, 264, 300 Commercial Code 311 communication networks, comparing Japanese and US semiconductor engineers 154–6, 158–61 Compal 52 compensation systems, comparing Japanese and US semiconductor engineers 152, 158 Hitachi’s reforms to 202 and for Japanese engineers and researchers 192, 193, 194 competency-enhancing innovation 23 competitive strategy, need for profitoriented 273 competitiveness 3, 4, 10, 31, 70, 72–4, 76, 79–81, 83–5, 87–9, 91, 146, 163, 195, 201, 209, 212, 219, 238, 245, 249–50, 259–62, 265, 275–9, 307 computer industry 8, 9–10 see also PCs (personal computers); software industry computer simulation technologies, and value chain modularity 40 conference attendances, comparing Japanese and US semiconductor engineers 154 contract manufacturers and the Japanese electronics industry 55–6, 60 and the semiconductor industry 12 and value chain modularity 52 contract research organizations (CROs), and the pharmaceutical industry 93 cost competitiveness, and the Japanese semiconductor industry 73–4, 80, 81–2, 84 Council for Science and Technology Policy (CSTP) 241, 265–6 creativity see individual creativity critical technology 249
Index CROs see contract research organizations (CROs) CSNet (Computer Science Net) 37 customized software 106, 111–21, 122 arguments for customization 116–21 enterprise resource planning (ERP) software 113–16, 118, 119–20, 122 Daiichi Pharmaceutical 97, 101 Davos conferences 252 defence, and Japan’s innovation system 24 Dell 10, 53, 55, 117, 148 Delphi 220 Denso 217, 220, 225–8, 229, 230, 231 Dexel 259 digital cameras 7 disruptive technology 166, 167–70 DoCoMo 53, 54, 55, 66, 73, 110 Dore, Ronald 13, 26, 34, 148, 200, 215, 230, 249 Dosi, G. 13 DRAMs 10, 11, 21, 58, 71, 79, 80–3, 84, 116, 149, 160, 190 and the Japanese techno-system 257, 261, 263, 264 lack of cost competitiveness 72–3 Drucker, P. 188, 189, 193 DuPont 130–1 E-Bay 53, 262 e-government see Juki-net Earth Simulator 251 East Asia establishment of community in 243, 254 MOT programmes in 279–80 Economist, on the Japanese–American trade friction 252 EDI, Japanese rate of adoption 112, 113 Edison, Thomas 130 Edquist, Charles 255 education 3–4, 243–44, 253, 291, 302, 310 educational R&D expenditure 3–4 higher 2, 241, 245, 246–7, 279, 281
MOT 24–5, 196, 271–84, 314 of Japanese Engineer 148, 150, 152, 158, 159, 160 Einstein’s theory of relativity 253 Eisai 95, 97, 98, 100 electronic government in Japan 24–5, 286–96 and the ‘lost decade’ 286 see also Juki-net electronics industry 2, 21, 47–66 alliances 56–8, 66 competitive challenges 53–4 financial performance (1997–2004) 54 and the genba 305 information technology and communications services 63–4 and keiretsu networks 219–24, 228–9 low-end products 56 outsourcing 58–62, 66 recent restructuring 59, 65–6 response to modular production 55–64 shifting from manufacturing to services 63 see also Hitachi; semiconductor industry elementary process technology, and the semiconductor industry 74–6, 78, 80, 83 Elpida 11 email use, comparing Japanese and US semiconductor engineers 154 embedded software 110–11 EMC 53, 109 employees changing Japanese 191–5 see also HRM (human resource management) employment protection 21 in the United States 51–2 EMS (electronics manufacturing services) 52 emulation challenge 1, 2, 5–8, 13 and Hitachi 199 Enron 298 enterprise resource planning (ERP) software 113–16, 118, 119–20, 122, 300
323
Index equipment manufacturers, and the Japanese semiconductor industry 77 ERATO programme 18 Ericsson 40, 54, 109 ERP (enterprise resource planning) software 113–16, 118, 119–20, 122, 300 etching technology, in the semiconductor industry 74–6, 82 Ethernet 35, 37, 38, 40 Europe industrial models in 48 MOT training in 278–9 pharmaceutical industry 90, 102 European Union 238, 254 external communication networks, comparing Japanese and US semiconductor engineers 154–6 fabless semiconductor design firms 11, 52 fads, and the birth of MOT 14–15 Fairchild Semiconductor 142, 173–4, 179, 259 false negatives and open innovation 129, 134–7, 139, 143, 187, 222 and the reformed Japanese model 307 fields of resonance, and paradigm disruptive innovation 174–5, 179–80, 181, 316 Fifth-Generation Computer Project 263–4 financialization, and industrial models 48 Flextronics 52, 53, 148 Ford 220 Fordist mass production, and the Japanese production system 48 foundries, semiconductor 11, 52, 77, 79 FSX affair, and technology friction 260, 261 Fuji Electric 131, 136, 222 Fuji Xerox 217, 218 Fujimoto, Takahiro 8, 12, 51 Fujimura, S. 73, 80 Fujita, Ted 185–7
324
Fujitsu 32, 33, 36, 38, 39–40, 57, 59, 136, 138, 221–2 and technology friction 259 Fujusawa 95, 97, 100 Furukawa 136, 221–2 gate electrode etching, and wafer process technology 75 GE 9–10, 16, 40, 52, 53, 63, 130 Gen-Probe 99 genba shugi 305 curse of 111, 118–20 Genentech 131 General Electric 130 General Motors 130, 220 Genzyme 131 Gerlach, Michael 216, 218, 231, 232 Germany biotech industry 95, 101 history of innovation in 245 GlaxoSmithKline 100 global integration, and the electronics industry 47–8 global knowledge networks 22, 302–3 global value chains, in the pharmaceutical industry 21–2, 87–102 global warming 239 globalization 5, 16, 51, 223–4, 278 and keiretsu networks 230–1 and MOT 16 Google 53, 262 Goto, A. 3, 16 Goto, Dr Shigeki 37, 41 government agencies 2 see also METI, MEXT, MIC, MITI Guillot, Didier 216, 224, 228 hard disk drives (HDDs), and Hitachi 208 hardware firms, transformation into software 108–11 Harreld, Bruce 120 Harvard Business Review 16 healthcare, Hitachi and new business models 204–6 Hewlett Packard (HP) 40, 52, 220 Hibino, Masao 38
Index Hiranuma Plan 313 Hitachi 14, 23, 57, 199–212, 311 Corporate Technology Office 209 crisis and reform 200–3 decentralization 201–2 declining profitability 199 ‘excellent company’ books about 200 founding of 200 Group Council 201 HRM reforms 23, 199–200, 202–3, 209 and Internet-related technologies 32, 34, 39 and large firm malaise 23, 199, 202, 211 mergers and acquisitions 201, 204, 207 and MOT 23 and the nascent network model 315 new production(ist) system 199, 209–11, 306–7 and open innovation 138, 199 organization reforms 201–2 and the pharmaceutical industry 95 restructuring 69 and the semiconductor industry 72, 82 technology management innovations 199 technology and new business models 199, 200, 203–6 technology strategy 206–8 Urban Planning and Development Systems 208 and the US/Silicon Valley model 212 Hitachi Maxell 207 Hitachi Omika 216, 217–18, 230 Hitachi Vehicle Energy 207 Hollywood film industry, and open innovation 141 Hon Hai (Foxconn) 52 Honda, Soichiro 184 Horiba 187, 316 horizontal keiretsu networks 216 HRM (human resource management) 22, 133, 145–62, 298 career development and training in science and technology 243–7
comparing Japanese and US semiconductor producers 150–3, 157–61 components of 146 Hitachi and HRM reforms 23, 199–200, 202–3, 209 internally/externally oriented systems 16, 22, 145, 147, 149, 161–2 and the nascent network model 314 and the reformed Japanese/large firm model 307–8 human frontier science 272 human genome project 249 human rights, and Juki-net 290–1 Hyundai 53 Iansiti, M. 145, 148 IBM 9–10, 16, 40, 52, 53, 63, 220 and Hitachi 208 and industrial innovation 130, 131 and open innovation 134, 136, 138, 141 and the reformed Japanese model 305, 306, 309 and the software industry 107, 108, 120 and techno-systems 262, 263, 264 Ibuka, Masaru 184 ICT (information and communication technology) 1, 2, 21, 299 and the emulation challenge 6 equipment exports 4 and Hitachi 200 and Japanese electronics industry 63–4 and Japan’s innovation policy 240, 249 and the reformed Japanese model 305, 306 see also software; telecommunications industry IDMs (Integrated Device Makers) 11 IETF (Internet Engineering Task Force) 21, 41–3 Imai, K. 217 Imura, Ryo 203–4 Inagami, T. 14, 192, 202 incremental innovation 300 India, and innovation 133
325
Index individual creativity 22, 23, 184–96, 202 and innovation policies 24 and knowledge workers 188–9 knowledge workers and the management of 145, 146 recognizing and unleashing 187, 195–6 industrial innovation 129, 130–2 see also university–industry (U-I) relations industrial models, evolution and circulation of 48–9 Industrial Regeneration Law (1999) 311 Industrial Research Institute 138–9 infectious diseases 239, 267 innovation blue LED innovation 23, 135, 166, 170–4 different patterns of 255 dilemmas 167–70 and individual creativity 185 industrial 129, 130–2 and Japan as a front runner 237–54 Japanese system of 1–2, 237–54, 298–317 and keiretsu networks 23, 215, 216–28, 230–2 linear model of 167, 257, 272 national systems of (NSI) 255 performance-disruptive 23, 166, 168–70, 174, 180 top management and leadership of the innovation process 303 see also closed innovation; open innovation; paradigm disruptive innovation innovation policy 24 integral architecture 8 integration process technology, and the semiconductor industry 76, 78–9, 83 Intel 6, 72, 131, 134, 136, 140, 261–2 intellectual property 11, 239 and knowledge systems 149, 160 leakages 51, 60 legislation 194 and MOT 18 and the pharmaceutical industry 94
326
and the telecommunications industry 39 interfirm networks see keiretsu networks internal communication networks, comparing Japanese and US semiconductor engineers 154–6 internal HRM systems 22, 145, 147 International Council for Science (ICSU) 247 Internet 301 broadband 296 ‘dot.com’ crash 55–6, 298 and the Japanese electronics industry 53, 63–5 and techno-systems 261–2 Internet networking equipment 21, 31, 33–43 and ATM technologies 21, 35–6, 37, 38–9, 40, 41, 42, 43, 44 and NTT 21, 33–8, 39, 40–1, 43, 44 setting standards 41–3 Inventec 52 Investment capital/factory 7, 8, 56, 57, 58, 65, 66, 72, 160, 189, 214, 233 foreign direct 25, 49, 211, 259, 268 in ICT software, 36, 37, 40, 105–6, 108, 110, 114–6, 120 R&D 4, 5, 130, 131, 166, 167, 188, 257 involuntary spin-offs, and open innovation 141–2 Isaacs, A. 204 ISDN 21, 36, 38, 44 Ito-Yokado 6, 117 ITU-T (International Telecommunication Union-Telecommunication Standardization Sector) 39, 41, 42 Iwasa, Naruto 170, 172 Jabil 52 Japan choices in the 21st century 238–43 national policy goals 243 post-Cold War changes 237–8 Japan Bioindustry Association 91 Japan Defense Agency (JDA) 258
Index Japan Development Bank Corporate Finance Data Bank 229 ‘Japanese model’ 2–3, 298 see also innovation, Japanese system of Japanese Production System 48–9 and value chain modularity 49–52 JDA (Japan Defense Agency) 258 JDS Uniphase 40, 54 job satisfaction, and Japanese employees 191–5 Johnson & Johnson 102 JPC-SED (Japan Productivity Centre for Socio-Economic Development) 17 Juki-net 24, 286–96 Basic Resident Register 287 configuration 289 confrontation between stakeholders 290–2, 295 and the e-Japan Strategy 296 Family Register 287 and Juki-card 287–8 present status 288–9 privacy issue 291–2, 292–3 Juniper Networks 53 KAIST (Korea Advanced Institute for Science and Technology) 279–80 Kaneko, Atsushi 24, 196 Kato, T. 14 Kawanishi, T. 72 keiretsu networks 23, 215–32 and Hitachi Omika 216, 217–18 horizontal and vertical 216 and innovation strategy 218–24 and interfirm learning 217–18 Matsushita’s kyoei-kai 223–4, 228, 231, 232 and product and process development 216 and R&D 23, 217 and the reformed Japanese model 308–9 and the software industry 107–8 and strategic alliances in electronics 228–9 technological innovation and the dissolution of 224–8
Keynes, John Maynard 252 Kita, Toshiro 24–5, 313 Kitasato, S. 252 Kneller, R. 100 knowledge management, and the reformed Japanese model 307–8 knowledge systems 145–62, 239 comparing Japanese and US semiconductor engineers 153–6 components of 146–7 internal and external 147–8, 162 knowledge workers and individual creativity 188–9 pay of 193 ‘knowledge-creating company’, dilemmas and limitations of the 299–303, 315 Konica-Minolta 59 Konno, N. 174 Korea electronics industry 53–4, 61, 65, 66 and the emulation model 5, 6, 7 and Japanese firms 190 MOT programmes 279–80 semiconductor industry 72, 73, 82–3 software industry 113 Korea Advanced Institute for Science and Technology (KAIST) 279–80 Kusunoki, K. 51 Kyocera 316 Kyoto model 316–17 labour unions, in the United States 52 ‘large firm malaise’ 14, 23, 190 and Hitachi 23, 199, 202, 211 large firms and paradigm disruptive innovation 174–8, 180 see also reformed Japanese/large firm model; Hitachi LCD industry, and the emulation challenge 6, 7–8 LDV (Lucent Digital Video) 137 lean production 48, 49, 50, 58
327
Index LED (light emitting diode), blue LED innovation 23, 135, 166, 170–4, 180, 184 LG Electronics 53, 54, 66 Lincoln, James R. 23, 137, 142, 231, 232, 307 linear model of innovation 167, 257, 272 lithography, in the semiconductor industry 74, 82 litigation, and Japanese employees 193–5 ‘lost decade’ 237, 298 and e-government 286 and MOT (management of technology) 1, 2 and techno-systems 261, 264 Lucent 40, 54, 109, 131 Digital Video (LDV) 137 McGroddy 135 McKinsey 53, 63, 64 Malaysia and Matsushita’s kyoei-kai 223 scientific researchers 244 management and corporate governance, reformed Japanese model 310–12 management style, maintaining and strengthening Japanese 273–4 managers, and the software industry 106 Marberger, John III 251 market uncertainty, and open innovation 133–4 mass-production technology, and the semiconductor industry 76, 79–81, 83 Matsuoka, Takashi 171, 172, 180 Matsushita 23, 57, 171, 180, 216, 218, 309 kyoei-kai 223–4, 228, 231, 232 MBO (management by objectives), and Hitachi 202 media, Japanese media interest in MOT 18, 19–20 Memorex 9–10, 16, 40, 52, 53, 63 memory chips 7 see also DRAMS Menarini 100 Mentor Graphics 51 Merck 130, 131
328
mergers, and the telecommunication industry 32 Mertens, G. 90 Messerschmitt, D. 33 METI (Ministry of Economy, Trade and Industry) Cluster Plan 19, 313 and IP legislation 194 and the Monozukuri White Paper 20 and MOT 17 and software imports 112 and technology managers 274 and the telecommunications industry 32 and universities 314 MEXT (Ministry of Education, Culture, Sports, Science and Technology) 241, 243, 244, 265, 313 MIC (Ministry of Internal Affairs and Communications), and Juki-net 286, 290, 295 microbursts, and individual creativity 186–7 Micron Technology 79, 83 Microsoft 107, 131, 261–2 MIT Industrial Performance Center 47 Made in America study 276 MITI 109 and the Japanese techno-system 257–8, 263–4 Mitsubishi Electric 59, 311 Mitsubishi Heavy Industries 258 Mitsubishi Motors 309 Mitsubishi Pharma 97 Mitsumaki, Hiro 204–5 mobile communications technology 43–4 mobile phones, and embedded software 110–11 see also cell phones modular architecture 8–9 and the semiconductor industry 12–13 modular production/manufacturing 21, 49, 299 and the Japanese electronics industry 55–65 value chain modularity 49–52
Index modularity trap 51 modularization challenge 8–9, 13, 14, 21 and Hitachi 199 and packaged software 22 modularization dilemmas 300 monozukuri manufacturing 20–1, 299, 306–7 Moore, Gordon 173, 174, 179 Moore’s law 82 Morita, Akio 184 Mosaic 38 MOSFET (Metal-Oxide Semiconductor Field Effect Transistor) 173–4 MOT (management of technology) 1–2, 5, 14–21, 24, 237, 271–84, 299 and creativity 196 emergence in Japan 16–18, 271–4 fads and the birth of 14–15 ideal technology manager profile 274–6 and Japanese management style 273–4 life cycle and appeal of 19–20 and the monozukuri boom 20–1 problem areas and new skill sets 24 and the reformed Japanese model 311 training in Japan 280–4 in the United States 14–16, 276–8, 278–9 and university–industry (U-I) relations 18–19, 280–4 Motohashi, K. 91, 93, 99 Motorola 131 Mowery, D. 4, 18 MPUs (micro-processor units) as disruptive technology 168 and the semiconductor industry 72 MRI (Mitsubishi Research Institute), and MOT 17 mu-chip 203–4 Mukai, Takashi 170 Murakami, Ken 37 Murata 316 Murayama, Yuzo 24, 248, 273, 314 Nagaoka, H. 252 Nakamura, Shuji 166, 170, 171–2, 184, 193–4, 195
Nakata, Yoshi-fumi 24, 196 nanotechnology 240, 248, 249, 271, 314 nascent network model 312–15, 316, 317 national policy goals 243 National Science Museum 253 national systems of innovation 255, 271, 298 see also innovation, Japanese system of natural disasters, and technosystems 265–6 NEC 57, 172, 180, 223 and open innovation 138, 140 restructuring 59 and the software industry 106–7 and the telecommunications industry 32, 34, 38–9 Netscape 261, 262 networks see keiretsu networks new molecular entities (NMEs) 90–1 Nichia Corporation 166, 170, 171, 179, 180, 193–4 Nichicon 187, 316 Nichiden-Anelva 82 Nidec 316 Nippon Keidanren 243 Nippon Steel 6, 117 Nissan 207, 264, 309 Noguchi, H. 252 Nokia 110, 117, 131 Nonaka, I. 174, 185, 195, 299, 300, 301, 302 Nortel 54 Nortel Networks 109 North Korea 255 Novartis 100 Noyce, Robert 173, 174, 179 NTT and the blue LED innovation 171, 172, 180 and Internet networking equipment 21, 33–8, 39, 40–1, 43, 44, 83 and the Japanese semiconductor industry 71, 81 and open innovation 135 and the software industry 105, 111–12, 120 nuclear reactor industry 140–1
329
Index Odagiri, H. 3, 16 Odaira, Namihei 200 ODMs (original design manufacturers) 52 OEMs (original equipment manufacturers) 15, 52, 309 Ogawa, Nobuo 171, 179 oil shock (1970s) 237 Okamoto, Kazuo 184–5 Okawa, Isao 110 Oki Electric 34, 57 Omron 316 open innovation 22, 23, 129–43, 303–4 and the boundaries of a firm 129–30 and false positives 129, 143 and Hitachi 199 and HRM and knowledge systems 162 issues for further research 141–3 and the Japanese electronics industry 65 management of false negatives 129, 134–7, 138, 139, 143, 187, 222 and measurement errors 129, 133–4, 143 metrics for managing 138–40 open–closed innovation 140–1 and the pharmaceutical industry 87 shift to 132–6 and spin-off formation 136–7, 140, 141–2, 143 and the value chain 129 ‘open standards’ architecture 9, 13 outsourcing/supply contracts electronics industry 58–62, 66 pharmaceutical industry 88 Oyane, S. 71 packaged software 111–13, 116, 117, 300 paradigm disruptive innovation 23, 166, 167, 168–70, 178–81, 300–1, 316 and the blue LED innovation 173, 180 and individual creativity 184, 187 and large firms 174–8, 180 pay comparing Japanese and US semiconductor engineers 152–3, 158
330
see also compensation systems and job satisfaction 193 PCs (personal computers) global market for 5–6 and the Japanese semiconductor industry 71 and the software industry 106–7 see also software industry performance disruptive innovation 23, 166, 168–70, 174, 180 PET (positron emission tomography) 205–6 Pfizer 88–9, 90, 131, 134–5 pharmaceutical industry 2, 21–2, 87–102, 132 alliances/partnerships 88, 99–100 Centres of Excellence (CoEs) 89 clinical trials 93, 97 and contract service providers 87 domestic biotechnology sector 91 industry–university linkages 93–5 institutional background 90–5 and the molecular biology ‘revolution’ 87, 88–9 and new molecular entities (NMEs) 90–1 and open innovation 87 outsourcing/supply contracts 88 Pharmaceutical Affairs Law 83 pricing policies 91–2 and R&D 88, 89–90, 95, 98, 99, 100, 101, 102 regulation 90, 92–3 strategies to capture knowledge and technology resources 95–101 and technology management 88–90 Pioneer 23, 57 ‘platform leaders’, and the Japanese electronics industry 65 Plaza Agreement 237 ‘poison of success’ 83 poker metaphor, and open innovation 135–6 population decline in Japan 239–40 Porter, Michael 175, 245, 249 Price Waterhouse 53, 63, 64
Index Probert, Jocelyn 21–2, 132, 302, 308 problem-solving processes, comparing Japanese and US semiconductor engineers 155–6, 161–2 Procter & Gamble (P&G) 130, 138 product development, comparing Japanese and American semiconductor producers 156–61, 162 productionist culture/system and the semiconductor industry 83 Hitachi 209–11 promotion age of 14 comparing Japanese and US semiconductor engineers 152–3 Hitachi 202 QoS (Quality of Service) benchmarks, and the telecommunications industry 35, 36 QualComm 148 quality 1, 2, 3, 5–6, 13, 14–6, 49, 116, 118, 142, 209, 223, 238, 245, 250 in semiconductor industry 70–1, 76, 77, 79–84 in telecommunications industry 21, 35, 40–1 quality challenge 14–15 Quanta 52 R&D and disruptive technology 169 efficiency decline in manufacturing 190 employees and job satisfaction 191–5 and the emulation challenge 6 expenditure 3–4, 5, 189–90 global networks 302 and Hitachi 209, 212 and individual creativity 187 individual researchers 23 and industrial innovation 130, 131 and the innovation system 167 and keiretsu networks 23, 217, 228, 229, 231
and knowledge systems 145, 150 and knowledge workers 188–9 and MOT programmes 16, 278 and open innovation 132–3, 134, 139–40, 141, 143 and paradigm disruptive innovation 175–8, 180 and the pharmaceutical industry 88, 89–90, 95, 98, 99, 100, 101, 102 and the reformed Japanese/large firm model 307, 308, 310 rewarding individual creativity 195–6 systems under international competitive pressure 272 and techno-systems 262, 263, 266 and telecommunications 32, 34, 35 training of managers 250 and the US techno-system 256–7 R&D engineers, survey of 145, 150–62 Rapp, W. 114 reformed Japanese/large firm model 304, 305–12, 317 alliances 309–10 and group/keiretsu relations 308–9 HRM and knowledge management 307–8 management and corporate governance 310–12 monozukuri 306–7 the shopfloor 305 technology development 307 and university–industry relations 310 regional industry in Japan 312 relationship contracting 34 reliability see quality Renesas Technology 57, 208 MOT education projects 17 researchers, scientific decrease in numbers of 239–40, 241, 243 moral stance of 251–2 restructuring 617–8, 29, 23, 48, 191–2, 196, 239, 247, 264, 278, 298, 300–1, 307–7, 311
331
Index in electronics industry 57–2, 55–6, 58, 65–6, 84, 162 at Hitachi 207, 209, 211 Japanes companies and 109–1 in pharmaceutical industry 96, 98, 101 reward practices, see compensation systems RFIDs (radio frequency identification devices), mu-chip technology 203 RIE (reactive ion etching) 82 Ritschev, D. 137, 142, 228 Roche 89 Rohm 316 Rosenberg, N. 4, 18, 264 S&T (science and technology) Basic Plan for Science and Technology 248, 253–4, 265, 266, 271, 314 fusion of scientific disciplines 249 and innovation policies 24, 241–3 public understanding of science 252–3 and society 247–9, 273 training and career development 243–7 see also security and techno-systems Sakurai, Yoshiko 290 Samco 316 Samsung 53, 54, 55, 66, 73, 110, 160, 280 Samuels, R. 258 Sankyo 90, 97, 98, 101 Sanmina-SCI 52 Sanyo Electric 224 Sasaki, Hajime 140 Schein, E. 15 Schumpeter, Joseph 44 Science Council of Japan 242 science and technology see S & T (science and technology) Science and Technology in Society Forum (2004) 248 Science and Technology White Paper (2004) 239, 252–3 scientific management 20–1, 282 scientific methodology, changes in 240–1 Seagate Technology 167–8
332
SECI model (socialization, externalization, combination, internalization) 174, 175, 299–300, 301, 302–3 security and techno-systems 24, 255–67 anzen-anshin techno-system 264–6, 267, 273 and the Cold War 256–8 collapse of the old techno-systems 263–4 implications for techno-systems 266–7 and technology friction 258–61 Seiko Epson 59 SELETE (semiconductor consortium) 72 self-control rights, in Japan 293 SEMATECH 146 semiconductor industry 6, 8, 10–13, 21, 70–85 comparing three groups of engineers 77–81 competitive and excessive technology 76–81 and contract manufacturers 12 and cost competitiveness 73–4, 80, 81–2, 84 declining competitiveness 70–1 and elementary process technology 74–6, 78, 80–1, 83 excessive quality thesis 70, 71 fabless 11, 52 foundries 11, 52, 77, 79 HRM and knowledge systems 146, 148, 149–62 and individual creativity 195–6 and integration process technology 76, 78–9, 83 and mass-production technology 76, 79–81, 83 and modular architecture 12–13 and open innovation 142 and tacit knowledge 301 and techno-systems 264 and technology management failure 70, 73, 84 and wafer process technology 72, 74–6 see also DRAMs; electronics industry
Index Senoh, Masahiro 170 Shimadzu 187, 316 Shin Kobe Electric Machinery 207 Shionogi 102 Shirakawa, Hideki 184, 185, 187 the shopfloor, in the reformed Japanese/ large firm model 305 Siemens 131, 136 Silicon Graphics 52 Silicon Valley model 302, 312 and Hitachi 212 and keiretsu networks 215, 219, 221 Singapore MOT programmes 279, 280 scientific researchers 244 SIRIJ (Semiconductor Industry Research Institute) 72–3 skill development comparing Japanese and US semiconductor engineers 151–2, 158–61 see also training technology managers 274–6 skills, and MOT 24 smart cards, and Juki-net 293–4, 295 SMEs (small and medium-sized enterprises) 312–13 Smith, Adam 252 social change, and Japan in the 21st century 239 social contract 48 social relations, and the nascent network model 314–15 social responsibility, and universities 246–7 society, and science and technology 247–9, 273 Softbank 43 software industry 22, 105–22, 300 and APIs (Application Programmer Interfaces) 122 customized software 106, 111–21, 122 evolution in Japan and the US 106–8 and Hitachi 201 investment 105–6
Japanese expenditures on software imports 112 and techno-systems 256 transformation of hardware into software firms 108–11 unbundling of hardware and software 107–8 Solectron 52, 53, 54, 56, 65 SONET (Synchronous Optical Networking) 37, 40 Sony 57, 172, 180, 184, 311 specialist firms, rise of 8–9 spin-offs 22, 298, 307, 308 and the commercialization of innovation 219–22 and open innovation 136–7, 140, 141–2, 143 standardization 300, 306, 309 in electronics industry 50, 56, 58, 65, 82, 83, 223 in pharmaceutical industry 92–3, 96 insoftwareindustry22, 106–7, 112, 120, 122 in telecommunications industry 39, 41–3, 300 standards, see standardization startups 23, 299, 302, 312–13 stress, and creativity 196 Sturgeon, Timothy J. 21, 299 Subaro Observatory 251 suicide in Japan 239 Sun Microsystems 52, 53, 299 Super-Kamiokande 251 supplier relationships, and the Japanese Production System 49–50, 51 supply chain management 15, 50 see also value chains sustainable development 248 systems integration 50 tacit knowledge 300, 301–2 and individual creativity 195 and paradigm disruptive innovation 174–5, 179–80, 184 Taiwan contract manufacturers 52, 60 and the emulation model 5, 6
333
Index Taiwan (cont.) and the Japanese electronics industry 56 semiconductor industry 72, 73, 77, 80 software industry 113 Takamine, J. 252 Takeda 90, 95, 97, 98, 100 Takeishi, A. 51 Takeuchi, H. 185, 195, 299, 300, 301, 302 Tanabe 100 Tanaka, Koichi 187, 266 Tanaka, Shinichi 193 TCP (Transmission Control Protocol/IP (Internet Protocol) 21, 35–6, 37–8, 39, 40–1, 63, 83 teamwork 22, 274, 275, 278, 279 technical information sources, comparing Japanese and US semiconductor engineers 153, 161 technical uncertainty, and open innovation 133–4 techno-systems see security and technosystems technology see S & T (science and technology) technology development, and the reformed Japanese/large firm model 307 see also R&D technology friction, US and Japan 258–61 technology management see MOT (management of technology) technology strategy, and Hitachi 206–8 telecommunications industry 31–44, 83 Internet networking equipment 21, 31, 32, 33–43 Japan’s trade balance in telecommunications equipment 31–2 mergers 32 and mobile communications technology 43–4 and software 108–9 and techno-systems 256, 264 see also ICT (information and communication technology) terrorism 239 security and techno-systems 265, 267
334
Texas Instruments 142 Thailand, scientific researchers 244 Thalidomide 135 TIM-Japan (Japan Research Centre for Technology and Innovation Management) 17 TLOs (university technology licensing offices) 94 Tokico 207 Tokuda Seisakusho 82 Tonegawa, Dr 251 top management, leadership of the innovation process 166–81 inter alia 200, 201, 303 Toshiba 57, 59, 82, 138, 172, 180, 223, 311 Toyota 6, 23, 117 Chief Engineer System 184–5 and keiretsu networks 217, 220, 221, 225–8, 229, 230, 231 Toyota, S. 252 TQM (total quality management) 16 trade liberalization, and innovation in Japan 132 training comparing Japanese and US semiconductor engineers 151–2 and innovation 238, 314 MOT training overseas 276–80 scientific researchers 243–7 technology managers 274–6 transistors, as disruptive technology 168–9 TSMC 52, 60, 148 Tushman, M. 175 UMC 52, 60 uncertainty and open innovation 133–4 uncertainties of advanced technology 271–2 unemployment 191, 314, 318 UNESCO 247 Uniphase 54 Unisia 207 United Kingdom (UK), pharmaceutical companies 92 United Nations, World Year of Physics 253
Index United States electronics industry 52 Food and Drugs Agency (FDA) 90 HRM and knowledge systems 149–62 innovation system 4, 24 and Japan in the 21st century 238, 245 and Japanese electronics firms 54, 58, 60–1, 64–5 and the Japanese Production Model 49, 50 and the Japanese semiconductor industry 72 and MOT 14–16, 276–8, 278–9 and open innovation 132 pharmaceutical industry 90, 102 semiconductor industry 11, 79, 83, 146, 149–62 social security numbers and the Jukinet 292 software industry 106–7, 108, 114–16 techno-systems 255, 256–7, 258–63 telecommunications industry 32, 33, 43, 44 US-Japan trade/technology friction 245, 252, 256, 258, 259–61, 267 and value chain modularity 51–2 see also Silicon Valley model universities 2, 313–14 and innovation policies 24, 245–7 and the Japanese techno-system 258 and the Kyoto model 316 MOT education projects 17 R & D expenditure 3–4 and techno-systems 262 university–industry (U-I) relations 18–19, 250, 299 and MOT 18–19, 280–4 and the nascent network model 315 in the pharmaceutical industry 93–5 and the reformed Japanese/large firm model 310 and science and technology policy 241, 242, 246 and techno-systems 262, 267
urban development, and Hitachi 207–8 value chain modularity 21 and the Japanese production system 49–52 and Japan’s electronics industry 53–4 vBNS 37 venture business 93, 132, 136, 137, 180, 219, 220 venture capital firms 2 vertical integration 130 vertical keiretsu networks 216 Viagra 135 Visteon 220 voluntary spin-offs, and open innovation 141, 142 wafer process technology, in semiconductor production 72, 74–6 Western Electric 259 Whittaker, D. Hugh 14, 23, 83, 202 wireless technologies, and mobile communications technology 43–4 work organization, comparing Japanese and US semiconductor engineers 150–1 worker employment laws 311 see also employment protection World Science Conference (Budapest 1999) 247 World Trade Organization 132 Xerox Corporation 130, 132, 134 Yahoo 53, 262 Yahoo BB 296 Yamaguchi, Eiichi 23, 135, 184, 187, 300, 316, 317 Yamanouchi 90, 95, 97, 98, 100 Yokoyama, Muneaki 24, 196 Yukawa, H. 252 Yunogami, Takashi 21, 116
335