Innovation and industrialization in asia

Notes on Contributors vii1 Explaining technological catch-up in Asia 2 Innovation and learning in the integrated circuits industry in Taiwan and China 3 The computer software industry

of Malaysia and Thailand, and the transitional least developed country of Myanmar.

Chapters 1-7 of this book were originally published as a special issue of Journal of the Asia Pacifi c Economy.

Rajah Rasiah holds the Khaza nah Nasional Chair of Regulatory Studies and is Professor

of Technology and Innovation at University of Malaya, Malaysia He is also a Professorial Fellow at UNU-MERIT His research specialization includes science and technology policy, fi rm-level learning and innovation, healthcare services, foreign investment, cluster mapping and designing technology roadmaps

Yeo Lin is Executive Director of the Industrial Development Research Center and Faculty

member at the College of Public Administration, Zhejiang University, China Her research focuses on industrial clusters, industrial development and policy, and on low carbon economy

Yuri Sadoi is Professor at Meijo University, Nagoya, Japan She specializes on automotive

and electronics supplier networks

Innovation and Industrialization

in Asia

Edited by

Rajah Rasiah, Yeo Lin and Yuri Sadoi

Routledge

Taylor & Francis Group

LONDON AND NEW YORK

by Routledge

2 Park Square, Milton Park, Abingdon, Oxon, OX14 4RN

Simultaneously published in the USA and Canada

by Routledge

711 Third Avenue, New York, NY 10017

Routledge is an imprint of the Taylor & Francis Group, an informa business

© 2012 Taylor & Francis

This book is a reproduction of Journal of the Asia Pacifi c Economy, volume 15,

issue 3 The Publisher requests to those authors who may be citing this book to state, also, the bibliographical details of the special issue on which the book was based.

All rights reserved No part of this book may be reprinted or reproduced or utilised in any form or by any electronic, mechanical, or other means, now known

or hereafter invented, including photocopying and recording, or in any

information storage or retrieval system, without permission in writing from the publishers.

Trademark notice: Product or corporate names may be trademarks or registered

trademarks, and are used only for identifi cation and explanation without intent to infringe.

British Library Cataloguing in Publication Data

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

ISBN13: 978-0-415-50545-1

Typeset in Times New Roman

by Saxon Graphics Ltd, Derby

Publisher’s Note

The publisher would like to make readers aware that the chapters in this book may

be referred to as articles as they are identical to the articles published in the special issue The publisher accepts responsibility for any inconsistencies that may have arisen in the course of preparing this volume for print.

Notes on Contributors vii

1 Explaining technological catch-up in Asia

2 Innovation and learning in the integrated circuits industry in

Taiwan and China

3 The computer software industry as a vehicle of late industrialization:

lessons from the Indian case

Balaji Parthasarathy 28

4 Samsung’s catch-up with Sony: an analysis using US patent data

5 Ownership and technological capabilities in Indonesia’s automotive

parts firms

6 Are electronics firms in Malaysia catching up in the technology ladder?

9 Malaysia’s manufacturing innovation experience

v

Abdusy Syakur Amin is a lecturer in Industrial Engineering and Management Department,

University of Pasundan, Bandung, Indonesia He holds a PhD in Economics from University of Malaya His teaching and research interests focus on technology and innovations, industrial organization and strategic management His recent research includes technological capability upgrading in Indonesian automotive and electronic industries

Si Hyung Joo is a doctoral student in the Technology Management, Economics and Policy

Program (TEMEP) at Seoul National University His research investigates changes in technological and knowledge regimes and how these relate to innovation strategies and the performance of fi rms

Xinxin Kong holds a PhD in industrial economics from the Institute of Industrial

Economics at the Chinese Academy of Social Sciences (CASS) She is an associate researcher at the Institute of Science, Technology and Society at the Chinese Academy of Science and Technology for Development (CASTED), at the Ministry of Science and Technology, China Her research specialization includes industrial innovation and national and corporate innovation strategies

Keun Lee is a professor of economics at Seoul National University and the director of the

Center for Economic Catch-Up He is the editor of Seoul Journal of Economics and one of the editors of Research Policy His area of research is in the economics of catch-up, with

a focus on innovation and corporate strategies, which can be downloaded from www.keunlee.com

Yeo Lin is a professor at the College of Public Administration and the director of the

Industrial Development Research Center, Zhejiang University, China She had previously served as director of the Industrial Economics Research Center at Industrial Technology Research Institute (ITRI, Taiwan, and director of Industry and Market Research Department at Applied Science and Technology Research Institute (ASTRI), Hong Kong Her research specialization includes technology, industry and capacity building

Myo Myo Myint works as a research offi cer at the Friedrich-Ebert Stiftung offi ce in Yangon She holds a PhD in Economics from University of Malaya and her research interests include technological capability building in Myanmar, garment manufacturing, and the business and institutional environment in which fi rms are embedded

vii

Balaji Parthasarathy holds the ICICI Chair and is associate professor at the International

Institute of Information Technology, Bangalore His teaching and research interests focus

on the relationship between technological innovation, economic globalization and social change with particular focus on the emerging economies

Rajah Rasiah holds the Khazanah Nasional chair of regulatory studies and is professor of

technology and innovation policy at the Faculty of Economics and Administration, University of Malaya He is also a Professorial Fellow at UNU-MERIT His research focuses on the impact of regulatory frameworks, transnational corporations and intellectual property rights on domestic industrial upgrading and innovations His latest publications

include the edited book of Malaysian Economy: Unfolding Growth and Social Change (Kuala Lumpur: Oxford University Press, 2011) and the co-edited book of The New Political Economy of Southeast Asia (Cheltenham: Edward Elgar, 2010).

V G R Chandran is a Principal Analyst of Economics and Innovation Policy at

Malaysian Industry-Government Group for High Technology, Malaysia He is also a Senior Research Fellow at University of Nottingham, Malaysia and Adjunct Professor at Management and Science University, Malaysia He holds a PhD in Economics and his research interest is in industrial development, technology and innovation policy

Yuri Sadoi is a professor in the Faculty of Economics at Meijo University, Japan She

received her Ph.D in human and environmental studies from Kyoto University She is the

author of Skill Formation in Malaysian Auto Parts Industry (UKM Press, 2003) She is also a co-editor of Production Networks in Asia and Europe: Skill Formation and Technology Transfer in the Automobile Industry (Routledge, 2004) and of Multinationals, Technology and Localization: Automotive and Electronics Firms in Asia (Routledge,

2008)

Peter Wad is associate professor at the Department of Intercultural Communication and

Management, Copenhagen Business School, Denmark He is a sociologist with a research interest in global and local Asian automobile industries, and corporate off -shoring and

outsourcing of production, and innovation activities to Asia He is a co-author of Strategies

in Emerging Markets A Case Book on Danish Multinational Corporations in China and India (Copenhagen Business School Press , 2010).

viii

aRajah Rasiah, bYeo Lin and cYuri Sadoi

a Faculty of Economics and Administration, University of Malaya, Kuala Lumpur, Malaysia;

b College of Public Administration and Industrial Development Research Centre, Zhejiang University, Hangzhou, Zhejiang Province, China; c Faculty of Economics, Meijo University,

Nagoya, Aichi, Japan

This chapter provides the theoretical guide and introduction to a selected list of papers evaluating the drivers of technological catch-up experiences from Asia It departs from neoclassical preoccupation with markets as the sole or dominant institution of economic allocation by arguing that the evidence supports the evolutionary logic of macro, meso and micro interactions between several institutions, depending on the actors involved, structural location and taxonomic and trajectory elements of technical change

1 Introduction

Asia has become the hotbed for the analysis of rapid economic growth and structural change experiences among the emerging economies since the 1980s China, India, Korea, Taiwan and the Southeast Asian economies have increasingly driven developing Asia’s contribution to global gross domestic product While the primary sectors of agriculture and mining have been important in a number of them, manufacturing has been the engine

of growth and structural change There has, however, been little consensus over the drivers

of these processes of growth and structural change Also, the pace of technical progress, industrial specialization and its consequent eff ects on gross domestic product has varied among the fast growing Asian developing economies

This volume seeks to showcase the technological and economic catch-up experiences of selected fi rms and industries from Asia with a view to explaining the key drivers The next section discusses the main arguments of the book before the outline is presented in the

fi nal section

2 The main arguments

Standard neoclassical explanations of Asian industrialization success have been attributed

to export growth, trade liberalization and increased focus on the private sector Despite heavy use of industrial policies Korea, Singapore and Taiwan have generally been classifi ed by mainstream economists as liberal economies (Balassa 1982, Krueger 1997) The other high-performance economies of Asia have also been argued to have enjoyed rapid growth following liberalization initiatives China’s open-door policy since 1978, Malaysia’s return to liberal policies in 1986 following heavy industrialization initiatives

in 1981–1985, Thailand’s and Indonesia’s liberalization eff orts since 1986, Vietnam’s Doi Moi in 1989 and India’s liberalization since 1991 have been considered watersheds in the



advent of market forces and its positive role in transforming them into rapid developers Whereas the World Bank (1993) for the fi rst time conceded that the Northeast Asian economies of Korea and Taiwan had developed rapidly through interventions the conclusions simply dismissed their relevance, claiming that they were neither possible anymore because of the changed pressures in global markets nor worthy of the risks

associated with government failure (Fishlow et al 1994).

Contrary to mainstream accounts, heterodox economists advocating industrial policy and catch-up argue that all latecomers have achieved rapid growth and improvements in living standards only through focusing on the increasing returns activities characterized by manufacturing These accounts go back to Smith (1776), Hamilton (1791), List (1885), Young (1928), Abramovitz (1956), Kaldor (1960, 1967), Gerschenkron (1962), Johnson (1982), Amsden (1989), Wade (1990), Chang (2003) and Reinert (2007) Selective government intervention to support fi rm-level technical change and competitiveness has been the critical explanation in these accounts Whereas Marx and later Schumpeter focused on identifying technology as well as its diff erentiation into Department I and Department II (in the case of Marx) and Mark I and Mark II (in the case of Schumpeter)

as the driver of growth, Nelson and Winter (1982) focused on the institutions that drive technical change (including incremental engineering and upgrading) and catch-up Central

to Nelson’s (2008) argument is that the evolutionary processes of technical change and the critical institutions that eff ect it not only are non-linear in nature but also vary with each diff erent industry

Taken together, the chapters here seek to discuss important policy-relevant learning and innovation experiences from the rapidly growing Asian economies Captured through Nelson’s (2008) evolutionary lenses, these experiences have been driven, on the one hand,

by a combination of, policy instruments by governments, fl ows of knowledge from multinationals directly (through foreign direct investment) and indirectly (through technology licensing and experiential knowledge gained by human capital) and, on the other hand, technological catch-up by industrial fi rms Some of the fi rms from these economies have even become shapers of global technology – e.g Samsung in dynamic random access memories and Taiwan Semiconductor Manufacturing Corporation in logic

circuits (see Mathews and Cho 2000, Lee and Mathews 2008, Rasiah et al 2008) To

provide a suffi ciently broad coverage of the processes of catch-up and technological change, the book addresses fi rm level catch-up and leapfrogging issues as well as at both

a more aggregate level and the fi rm level the interface between policy instruments and technological capability development

3 Outline of the book

The case experiences were carefully selected from Asia on the basis of signifi cance of particular industries to national growth, as well as at least some technological catch-up Si Hyung Joo and Keun Lee discuss Korea’s Samsung’s catch-up strategies against Japan’s Sony Yuri Sadoi examines the nature and extent of technology transfer from foreign to local fi rms in the automotive industry in Thailand The remaining papers discuss the extent

of learning, innovation and technological catch-up in the semiconductor industry in China and Taiwan, the software industry in Bangalore, the electronics industry in Malaysia, the automotive industry in Indonesia and the garment industry in Laos

In this introductory chapter, we, as guest editors, have provided the theoretical anchor for the evolutionary approach to examining technological catch-up The six chapters that follow underscore the view that technological catch-up is shaped by institutions and



institutional change through conscious eff orts at the macro government-policy level, the meso organizational level and the micro fi rm level The process of technological catch-up

is thus uneven, varying with time and geographically with industries and institutions.The second chapter, by Rajah Rasiah, Xinxin Kong and Yeo Lin, discusses how the initial incorporation in multinational value chains through direct operations in Taiwan in the 1960s and in China in the 1980s only transformed into the higher value-added wafer fabrication and designing operations in semiconductor manufacturing when the government launched eff ective high-tech support through labs and design houses As fi rms mature into higher value-added wafer fabrication and designing activities, most of Taiwanese labour-intensive semiconductor assembly operations have been relocated to China and Southeast Asia Because of its large population, semiconductor manufacturing and exports in China are still dominated by labour-intensive foreign-dominated assembly operations However, local fi rms have increasingly increased their participation in wafer fabrication and design activities, suggesting that these set of local fi rms may be on the Taiwanese path of technological catch-up in semiconductor manufacturing

Balaji Parthasarathy argues, in the third chapter, that industry must learn to deploy borrowed technology effi ciently in production and compete internationally beyond the subsidies that states provide This paper is particularly interesting from an evolutionary standpoint, as it provides evidence to show that the marginal cost of producing software is negligible from the outset and hence lumpy investments and huge high-tech labs are not a necessary investment to promote software development

In the fourth chapter, Si Hyung Joo and Keun Lee discuss Samsung Electronics’ technological catch-up with Sony Using the US patents of the two fi rms, the authors show that Samsung Electronics’ catch-up with Sony happened around the early 1990s in qualitative terms and mid-1990s in quantitative terms In addition, Samsung Electronics has also become independent of Sony by producing novel knowledge, quickening the techno-cycles and knowledge appropriation

Rajah Rasiah and Abdusy Syakur Amin examine in the fi fth chapter the development of technological capabilities in local automotive parts fi rms’ vis-à-vis foreign fi rms, following increased liberalization from the late 1990s The authors argue that there were no obvious statistical diff erences in human resource and process technology capabilities between foreign and local fi rms in 2006 The empirical evidence they provide shows that local

fi rms have invested more in research and development than foreign fi rms to compensate for the superior product technologies accessed by the foreign fi rms from their parent fi rms Although foreign fi rms showed higher export intensities, the evidence they have amassed also shows that the liberalization experience has driven rather than discouraged stronger initiatives in local fi rms to raise technological capabilities

In the sixth chapter, Rajah Rasiah examines the development of technological capabilities and economic performance in electronics fi rms in Malaysia Although the evidence shows substantial improvement of technological capabilities in both electronics and the specialized semiconductor fi rms, the incidence of participation of fi rms in the highest levels of knowledge-intensity activities was very low The evidence also shows that technological deepening through increments in skills and research and development personnel will raise labour productivity in the industry

The seventh chapter by Yuri Sadoi analyzes the development of technological capabilities of engineers in automobile parts suppliers in Thailand The evidence shows that Thailand’s latecomer fi rms accumulated their technology primarily from foreign direct investment, particularly from the Japanese carmakers Rising demand and competition have driven some amount of technological catch-up in the industry To



upgrade further the Thai automobile fi rms, government policy should focus on strengthening the automotive cluster by stepping up the supply of high-tech human capital

to provide the technological deepening necessary for the suppliers to upgrade

V.G.R Chandran, Rajah Rasiah and Peter Wad use the evolutionary perspective in the eighth chapter to investigate the innovation experience of Malaysian manufacturing fi rms, including spillover from foreign multinational corporations (MNCs) to local fi rms The evidence shows that innovation intensities diff er among states and sectors owing to diff erences location-specifi c institutional coordination Although Malaysia is not an important location for the off shoring of R&D activities by MNCs, they have driven innovation spillovers in industries such as electronics While process innovation is conducted by most foreign subsidiaries, a handful of them undertake product design activities in Malaysia However, weaknesses in the meso organizations have restricted technological upgrading towards higher value-added activities in the country

In the fi nal chapter, Myo Myo Myint and Rajah Rasiah analyze foreign capital infl ows and the development of garment manufacturing in Myanmar using an adapted version of the global value chain approach The exercise discusses how foreign and national fi rms managed to break in into international garment value chains through its large reserves of cheap labour Despite being gripped by political instability, economic sanctions and poor infrastructure, exports have remained important because of expansion into new markets such as Japan However, the assessment also shows that the poor embedding environment has reduced fi rms operations to low value added and low wage activities and undermined their capacity to raise competitiveness

References

Abramovitz, M., 1956 Resources and output trends in the United States since 1870 NBER

Occasional Papers, vol 52 New York: National Bureau of Economic Research

Amsden, A., 1989 Asia’s next giant: South Korea and late industrialization New York: Oxford

University Press

Balassa, B., 1982 Development strategies in semi-industrial economies Baltimore, MD: Johns

Hopkins University Press/World Bank

Chang, H.J., 2003 Kicking away the ladder: development strategy in historical perspective London:

Anthem Press

Fishlow, A., et al., 1994 Miracle or design: lessons from the East Asian experience Washington,

DC: Overseas Development Council

Gerschenkron, A., 1962 Economic backwardness in historical perspective Cambridge: Belknap

Press

Hamilton, A., 1791 Report on manufactures [online] Available from: http://www.oberlin.

edu/~gkornbl/Hist258/ReportMfres.html, accessed 13 December 2005

Johnson, C., 1982 MITI and the Japanese miracle Stanford, CA: Stanford University Press Kaldor, N., 1960 Essays on economic stability and growth London: Duckworth.

Kaldor, N., 1967 Strategic factors in economic growth Ithaca, NY: Cornell University Press Krueger, A., 1997 Trade policy and economic development: how we learn American economic

review, 87 (1), 1–22.

Lee, K and Mathews, J.A., 2008 Upgrading in the same industry and successive entries in new industries for sustained catchup: cases of Korean and Taiwanese fi rms Paper presented at Catch-

Up Workshop, Mexico City, 20–21 September

List, F., 1885 The national system of political economy London: Longmans, Green & Company Mathews, J.A and Cho, D.S., 2000 Tiger technology: the creation of a semiconductor industry in

East Asia Cambridge: Cambridge University Press.



Nelson, R.R., 2008 Economic development from the perspective of evolutionary theory Oxford

development s tudies, 36 (1), 9–21.

Nelson, R.R and Winter, S.G., 1982 An evolutionary theory of economic change Cambridge:

Harvard University Press

Rasiah, R., Kong, X.X., Lin, Y and Song, J.Y 2012 Variations in the catch up experience in the semiconductor industry in China, Korea, Malaysia and Taiwan Malerba, F and Nelson R (eds), Economic development as a learning process: diff erences across sectoral systems, Cheltenham: Edward Elgar

Reinert, E., 2007 How rich countries got rich and why poor countries stay poor New York:

Carroll & Graf

Smith, A., 1776 An inquiry into the nature and causes of the wealth of the nations London: Strahan

& Cadell

Wade, R., 1990 Governing the market: economic theory and the role of government in East Asian

industrialization Princeton, NJ: Princeton University Press.

World Bank, 1993 The East Asian miracle New York: Oxford University Press.

Young, A., 1928 Increasing returns and economic progress Economic journal, 38 (152), 527–542.



in Taiwan and China

Rajah Rasiah,a∗Xinxin Kongband Yeo Linc

a Faculty of Economics and Administration, University of Malaya, Kuala Lumpur, Malaysia;

b National Research Center for Science and Technology for Development, China; c Industrial

Development Research Center, Zhejiang University, Hangzhou, China

Using the evolutionary framework of inductive screening, this paper seeks to examinethe drivers of technological catch-up in the integrated circuits (ICs) industry in Taiwanand China The paper shows that IC manufacturing began with multinationals relocat-ing export-oriented assembly operations in the 1960s in Taiwan and in the 1980s inChina, but serious technological catch-up took place when, with the assistance of thegovernment, local firms began to participate in wafer fabrication and designing activi-ties While foreign ownership and export markets were critical in initiating connections

in the global IC value chain, the paper argues that the role of the government throughfunding, research and development laboratories and development of human capital werecritical in local firms’ technological catch-up process in both the countries

1 Introduction

The evolution of integrated circuits’ (ICs’) manufacturing was very much driven by thesearch for control in the US military However, because of its complementary and enablercharacteristics, it has diffused extensively into the manufacture of many different productsand processes It is therefore very common to find ICs driving central panel control systems

in steel and cement manufacturing and cad–cam machines in garment manufacturing, itoring captive salmon, storing of graphic memory in digital cameras and computers andproviding control to computer numeric control machines Although the processes of pro-ducing varies in sophistication, the design and fabrication of ICs remains a high-technologyactivity Hence, catch-up attempts in the industry have required lumpy investments in largephysical plants, machinery and equipment, design, human capital and effective demand tosupport scale-intensive activities

mon-Unlike in the pioneering economy of the US where the government-led military andlater the domestic market were critical in the origin and spread of ICs using silicon (theprime material used in IC devices) and gallium arsenide (see Marsh 1981),1 the maindrivers of demand in China and Taiwan have largely been export markets, though domesticmarkets have increasingly become important in the former Between the two countries,China is the only one that had a military plan targeted at the computer and microchipindustry during Chairman Mao’s administration The government created the Ministry of



Electronics Industry (MEI), but the subsequent growth in IC production from the 1980shas had little link with both instruments.

Large-scale foreign-driven IC assembly emerged in Taiwan and China in the 1960sand 1980s respectively, following the opening of export-processing zones (EPZs) Multi-nationals seeking low-wage, literate and disciplined workers in locations with good basicinfrastructure and security relocated assembly and later test operations in these countries.Special EPZs were created and coordinated in these countries to attract IC firms Employ-ment creation started as the prime policy aim of the host governments in the initial phase(see Lim 1978)

Taiwan moved ahead from export processing and assembly to participate in designingand research and development (R&D) activities in IC production, following the opening

of the Industrial Technical Research Institute (ITRI) in 1974 and the acquisition of RadioCompany of America (RCA) that later became the United Microelectronics Company(UMC) in 1980 (see Rasiah and Lin 2005) As wafer fabrication began to mushroom outsidethe developed economies from the 1980s, Taiwan and later China began to launch morefabrication and designing activities in primarily logic circuitry and application-specific ICs(ASICs) (Rasiah 1995, Brown and Linden 2009) The fabrication of most ICs, diodes andtransistors are increasingly being outsourced to Taiwanese and Chinese firms Intel has onememory wafer fabrication plant and announced plans in 2007 to build another in China.Taiwan’s UMC has contract fabrication operations in Singapore

The governments in China and Taiwan took important steps to promote IC ing and therefore offer a unique set of catch-up experiences for other developing countries.These countries also provide enough diversity to examine variations China with a popula-tion of 1.4 billion in 2006 has a large domestic market and has since 1978 been integrating

manufactur-an essentially centrally plmanufactur-anned economy into the capitalist world system With a population

of 19 million people in 2006, Taiwan is a small economy, where domestic demand neveracted as the major stimulant of rapid manufacturing growth Yet, in Taiwan SemiconductorManufacturing Corporation (TSMC), not only did Taiwan start its first contract manufac-turer of IC chips that separated chip design from chip fabrication, but it has evolved to be

at the technology frontier in the logic industry

This paper seeks to explain innovation and learning in the IC industry in Taiwan andChina The rest of the paper is organized as follows Section 2 presents the theoretical guide.Section 3 evaluates the catch-up trajectories of IC firms in Taiwan and China Section 4analyses the drivers of catch-up Section 5 presents the conclusions

2 Theoretical guide

The historical documentation of technological catch-up in manufacturing starts with theascendance of garment manufacturing in Britain from 1485 until firms reached the technol-ogy frontier by the eighteenth century (see Reinert 2007) Subsequent catch-up accountsfor industrial policy promotion involving the US, Germany, the Scandinavian countries,Italy, Australia and Japan to cement the heterodox argument that interventions are neces-sary to stimulate economic catch-up (see Hamilton 1791, List 1885, Gerschenkron 1962,McFarlane 1981, 1984, Johnson 1982, Freeman 1989) Amsden (1985, 1989), Chang (2003)and Amsden and Chu (2003) extended this argument with evidence from Korea and Taiwan.The rich literature mentioned above is complemented by the evolutionary school thatuses inductive methodologies to map technological catch-up taking account of the diversity,uniqueness of catch-up to vary with time and locations and complexities in technologiesand inter-relationships between economic agents, meso-organizations and macroeconomic



settings The evolutionary school provided departures from the standard broad-brushmacroexperiences to distinguish particular firms or industries through effective macro–micro interactions – which often eventually become horizontal – between firms andgovernment policy with the coordinating influence of meso-organizations to support tech-nological and economic catch-up The sources of catch-up are considered to be differentwith each experience, non-linear and uneven and often driven by a blend of influence fromthe institutions of markets, government and cooperation (Nelson 2008).

What is obvious in the catch-up experiences documented by evolutionary economistsare diverse examples where the critical institutions and organizations effecting learningand innovation are somewhat different Contrary to neoliberal policies of leaving economicagents to market forces or simply augmenting market signals, successful catch-up experi-ences posit specific interventions – at the various different levels – to drive innovation andlearning For example, the button manufacturing experience of Qiaotou started as a marketinitiative but only managed to achieve technological deepening in new design and materialsdevelopment with strong support from the Yongjia County Government and collaboration

from the universities of Lanzhou and Huanen (Rasiah et al., forthcoming) A combination

of markets, government and cooperation eventually played critical roles in the tion of the Qiaotou Town into a mature and composite button cluster This developmentalso supports the industrial district arguments on clustering (see Brusco 1982, Piore andSabel 1984, Becattini 1990, Rasiah 1994, Rasiah and Lin 2005)

transforma-Unlike button manufacturing, where scale is not critically important at the level ofindividual firms, IC fabrication shares specificities associated with lumpy investments.Scale economies have not fallen despite continued miniaturization and the decomposition

of IC production vertically into chip design, chip fabrication, assembly and test Even

in Taiwan, Amsden and Chu (2003) and Rasiah and Lin (2005) have argued that scalerequirements have driven up firm size Despite similarities, the sources of learning andinnovation in the industry as articulated by Malerba (2005) can be expected to be differentfrom the routes taken by firms in other industries Malerba (1992), Malerba and Orsenigo

(1997) and Malerba et al (2001) aptly address the specificity of firms and industries in

understanding technological catch-up from the evolutionary perspective

Taxonomically, catch-up in process and product technologies varies considerably ICfirms in Korea, for example, shifted from imitation to innovation through creative duplica-tion (Kim 1997) A lot of such innovations have evolved through user–producer interactions(see Lundvall 1988, 1992) Similarly, a wide range of industrial layouts, process layouts,organizational structure, machinery and equipment and product makes are adapted or mod-ified to raise process efficiency or produce new products without a need for seeking totallynew stocks of knowledge These changes in process and product technologies often help

to raise the boundaries of technical efficiency, but because they do not require new paths

of knowledge, they remain within the Schumpeterian Mark I system of creative tion These types of knowledge are part of innovation that latecomers often appropriatemore than the first mover A wide range of innovations achieved in Taiwan are part ofsuch innovations (see Rasiah and Lin 2005) Rasiah (1994, 1996) and Hobday (1995) alsodocumented a wide range of innovations in Malaysia and East Asia respectively, including

destruc-in multdestruc-inationals

Hence, an open inductive approach is used in the paper to trace the drivers of logical catch-up in the IC industry in Taiwan and China Given the evolutionary fondness toexamine each of the components and nature of influences as firms move up the technologicalladder, a deliberate attempt is made to map and trace the catch-up experiences of IC firms

techno-in Taiwan and Chtechno-ina



3 Technological catch-up paths

IC manufacturing had become extremely important in Taiwan from the 1980s and inChina from 2000, which can be seen from the growth in exports from these countries(see Table 1) Although initiatives to start IC manufacturing started first in China inthe 1950s, it was in Taiwan that the first substantive manufacturing operation began inthe 1960s and subsequently a catch-up into wafer fabrication and IC designing activities

in the 1980s Despite the closing of the gap with Taiwan by 2009, China’s huge populationhas produced an ecosystem of firms that, on the one hand, are dominated by foreign firms inassembly activities and, on the other hand, have an increasing share of local firms engaged

in wafer fabrication and designing activities

Foreign firms relocated back-end assembly operations in EPZs in the 1960s to start ICmanufacturing in Taiwan Government policy was instrumental in making the shift fromsimple assembly and test activities into front-end activities when the Electronic Researchand Service Organization (ERSO) was established among the ITRI in 1974

However, the creation of ITRI initially did not produce significant results for the ICindustry, as no grants were given to stimulate participation in R&D activities The small-firm-structure operations based on Marshallian knowledge flows did spur upgradation

in a highly capital- and knowledge-intensive industry (Rasiah 2008) In addition to thescience and technology project funds that were disbursed from 1979, ITRI started todevelop incubators to stimulate the birth of high-tech firms ERSO, the division within ITRIinvolved in supporting R&D in electronics activities, became a key driver of incubationfrom 1979, but particularly after 1983 when the projects introduced a matching frameworkwhen providing grants (Rasiah and Lin 2005) The government’s extensive investment to

Table 1 World IC and electronics component exports, selected economies, 1990–2006

Value (million US$) Share in nationalexports

provide the high-tech infrastructure (e.g R&D laboratories and standards organizations) inthe Hsinchu Science Park was also instrumental in driving incubators to world-class firmssuch as UMC, TSMC, ASUS, Winbond, Ase and Vanguard.

Large firms such as TSMC, which is the world’s leading contract manufacturer ofsemiconductor wafers, have subsequently forged strong collaboration with foreign technol-ogy R&D laboratories, universities and purchasers to move up the technology trajectory

In 2006, TSMC was fabricating cutting edge 12-in wafers using 0.13-μm chips usingnanotechnology with R&D support from a range of foreign collaborators that were alsoproviding the design support The firm also announced plans to fabricate microprocessors

in 2008 (Shilov 2007)

In both countries, the catch-up started with entry into the assembly of ICs by foreignmultinationals in EPZs: in 1960s in Taiwan and in 1980s in China Also, the technologicalregime of ICs can be characterized by high-velocity high-frequency devices using light-emitting devices that are expensive (using gallium arsenide material base) and are used inmobile phones and related products and low-frequency devices that do not require muchlight-emitting functions and are also cheap and abundant (silicon as the base) and are used tofabricate memories and microprocessors As a specialized contract manufacturer, Taiwan’sTSMC is engaged in all categories

Semiconductor devices are fitted into final goods electronics products such as consumerappliances, industrial electronics and information communication products (see Figure 1).Figure 2 shows selected semiconductor devices with non-scale-based presentation of theadded value enjoyed by each of them The IC devices the value chain typically involved sixstages, viz capacity implant development and specifications, chip design, wafer fabrication,chip assembly, packaging and test and sales and marketing (see Figure 3) The activesemiconductor components provide the control for these products Whereas multinationalsbegan assembly and test operations of memory chips, the entry into chip fabrication andchip design started with government-supported programmes China had the first government

Active semiconductors Transistors, microprocessors, random access memories, read-only memories, application

Passive semiconductors Diodes, capacitors, resistors, light-emitting devices

Plating

Plastic and ceramic packaging materials

Burn-in and test

Lead frame

Moulds, tools,

precision machinery

Bonding wires

Testing materials (e.g acids) Backward supplies

Industrial electronics Consumer electronics

Contract manufacturers

Incentives and grants

Human capital

R&D laboratories

Specialized technology parks Forward users

Government olicy

Unfabricated and

fabricated wafers

Universities

Standards organizations Military electronics

Figure 1 The IC cluster, 2009 Source: Rasiah (2008)



Microprocessors

Figure 2 Selected semiconductor devices Not drawn to scale Passive component devices includediodes, resistors, capacitors and light-emitting devices; electron flow is the central dimension of activecomponents; DRAM, dynamic random access memory; ICs refer to the integration of active semi-conductor devices to expand their capacity and functions; ASICs are ICs designed with specializedfunctions

Capacity implant development

Sales Marketing Value added

Stages in value chain

Figure 3 Value chain of semiconductors, 2008



programme in the 1950s, but the development of the industry from the 1980s has little linkwith that programme Taiwan followed next, launching formal programmes to develop ICmanufacturing through the establishment of ERSO among the ITRI in 1974 (Mathews andCho 2000, Amsden and Chu 2003, Rasiah and Lin 2005).

Unlike foreign affiliates already in possession of the requisite intellectual property rights(IPRs), local firms bought licenses or firms to access both product and process technologies.The regulatory framework on IPRs first became pronounced with the 1989 WashingtonTreaty that legalized industrial layouts in addition to industrial designs and patents Thegovernance regime of IPRs were included in the 1995 World Trade Organization’s TradeRelated Intellectual Property Rights agreement Local firms’ first access to most producttechnologies in Taiwan and China came through licensing and acquisition of foreign firms.The Taiwanese began with the acquisition of RCA in 1979 For a long time, microprocessormanufacturing was dominated by Intel until AMD won a legal suit to enter production

in the 1990s It appeared that American firms would not contract out microprocessorfabrication until in 2008 when TSMC announced plans to manufacture its first output ofmicroprocessors by the end of the year (Shilov 2007)

Unlike foreign assemblies that started operations in the 1960s, local Taiwanese ICfirms went directly into integrated operations from the late 1970s UMC was started in

1980 (see Lee, Liu, and Wang 1994, Mathews and Cho 2000) following the acquisition

of the semiconductor division of RCA in the second half of the 1970s (see Rasiah andLin 2005) RCA offered UMC ASICs, diodes and transistor technology The TSMC (49%)and Phillips (51%) merger in 1987 gave the Taiwanese firms access to more sophisticatedmemory chips

Semiconductors got strategic status when computers and semiconductor devices wereclassified under the category of national industries for research during Mao Ze Dong’sleadership when the MEI was created The initial stage development of IC industry inChina could be traced from the mid-1960s The first IC device called digital logic circuitwas developed successfully in 1965, which led to the successful development of TTL, ECL,PMOS, n-type metal oxide semiconductor and complementary metal oxide semiconductor(CMOS) technologies The basic R&D elements related to materials, equipment, manufac-turing and techniques were largely developed before the 1980s and occurred in the MEI,the Chinese Academy of Social Sciences and the Ministry of Spaceflights Apart frommicroprocessors, it can be argued that prior to 1980, Chinese IC technology caught up withfrontier firms However, in the period 1980–1995, IC industrial development began to fallbehind that of the firms at the technology frontier (Kong 2008)

There are many reasons to explain the stagnation in China in 1980–1995 One nation is that the initial focus was on the relocation of low-end foreign IC firms in Chinafrom the 1980s China enjoyed its first large-scale manufacturing of ICs following the re-location of American plants in EPZs Flagship firms such as Intel, National Semiconductor(Fairchild now), Motorola (Freescale now) and Chippac relocated operations in China, butregulations requiring that non-joint ventures must export all output meant that these firmshad to target export rather than domestic markets China’s share of global IC and electroniccomponents exports rose from 0.2% in 1990 to 1.7% in 2000 and 7.4% in 2006 (computedfrom Table 1, using global export figures as the base)

expla-While a strong platform led by foreign direct investment was evolving from the 1980s,the Chinese Government also launched instruments to encourage R&D in IC production and

to assist the opening of local firms in strategic industries that included ICs, computers andtelecommunication equipments The acquisition of the computer manufacturing division ofIBM worldwide by the Chinese firm Lenova and the expansion of Taiwanese-owned Acer



and American-owned Dell and HP into China heralded a major breakthrough for Chinese

IC firms, which now have the market potential to sell chips to major users IBM was alreadymanufacturing computers in China before the Lenovo takeover

Horizontal user–producer links grew between IC firms and suppliers and buyers inthe US, Japan and Germany Several machinery technology IC firms co-evolved, as stronginteractions helped machinery firms automate, refine, remodify and manufacture moreefficient and effective machinery and equipment Taiwan’s world-class machinery industryfacilitated similar transition, while China’s ecology of electronics and supplier firms hasalso evolved to support clustering

Process technology

Process technology refers to the processes that are undertaken to process or assembleproducts It includes adaptations or reorganization in machinery and equipment, layouts,inventory and quality control systems, production organization and firm structures Leadfirms are able to drive creative destruction as they are able to lower defects, delivery timesand costs while raising quality levels Falling profit margins often drive latecomers lackingproduct innovation rents typically to either drive out high-cost incumbents or fill up thevacuum left behind by closing or migrating firms Apart from industrial layouts of chips andchemical processes, cutting-edge machinery and equipment and materials in most processtechnologies are not subjected to IPRs regulations, and hence their diffusion is far quickerand easier than that of product technologies

Rapid growth of user–producer-driven products and processes, defect-free output anddelivery times drove closer interface and technology coordination between IC firms andbuyer firms (e.g computers, avionics, consumer electronics and mobile phones) in Taiwanand China IC firms in Taiwan and China are engaged in state-of-the-art development ofprocess technologies Between the two, Taiwan led in the take-up of process patents in ICdevices issued by the US patent office over the period 2002–2006 with 2503 patents Chinahad only 27 patents over the same period

Taiwan

The initial source of process efficiency improvements was recorded through technologytransfer by employees gaining experiential knowledge working in American and Japaneseconsumer electronics firms and training provided by machinery and equipment suppliers inthe 1960s and 1970s (Lin 2003) Taiwanese firms then internalized training and inventoryand quality control systems in-house once the suppliers became reluctant to supply thelatest technologies Like in the other countries, kaizen practices manifesting in differentforms – e.g small group activities, just-in-time, quality control circles and six sigma – weredeveloped in these firms to appropriate throughput efficiency and make production agileand flexible to meet volatile fluctuations in demand and prices

The domestic machinery industry adapted strongly in Taiwan to the needs of IC ufacturing, including wafer fabrication to manufacture cutting-edge machinery and equip-ment and to support modifications in firms Kaizen practices to lower throughput time,

man-reduce defects and meet customer requirements (the original equipment manufacturers,

or OEMs, that achieved global service provider status faced added pressure to remaininnovative) among Taiwanese firms – part of Schumpeterian Mark I system (creative de-struction; see Malerba 1992, Rasiah and Lin 2005)

Taiwanese domestic firms – e.g UMC, ASE, TSMC, Windbond, Asus and Vanguard– upgraded and relocated supply base at all major buyer locations and also introduced



and refined their capacity to anticipate changes in demand from buyer firms Networkcohesion facilitated strong differentiation and division of labour in Taiwan to supportlarge-scale manufacturing of OEM computers (only Acer is a major local original brandmanufacturer (OBM) computer manufacturer in Taiwan), scanners, monitors, motherboardsand components.

China

Foreign-owned multinationals in China relocated significant aspects of process technology,including just-in-time systems through their subsidiaries to facilitate better coordinationbetween them and buyers Hence, foreign firms were the initial transmission channel forthe movement of process technologies to Chinese IC firms Because the first wave of ICfirms to China relocated in the 1980s when flexible production systems, automation andcontinuous improvement benchmarks were absorbed by European and American firms,these techniques were already carried out from the outset among export-oriented firms.Given its huge size and population, China has a complex structure with both operationsdominated by multinational company operations and specialized high-tech parks dominated

by local firms (Kong 2008) Intel, Freescale, National Semiconductor, Texas Instrumentsand Chippac have large-scale assembly and test of ICs in China, targeting export markets.These firms also have back-end designing and other operations that are important forimproving production performance

Although the main machinery and equipment in assembly, test and fabrication are stillimported, considerable adaptations have already started in China By the end of the 1990s,Chinese firms had already started supplying robotics and automated machinery to foreignand local IC firms

Product technology

Unlike in process technology, the catch-up process in product technology is much moredifficult owing to the introduction of intellectual property rights, huge investment and theleaps in path-dependent knowledge that are required to sustain participation in the develop-ment of products facing rapidly shortening product cycles Firms in Taiwan and China havemanaged to move up the product technology trajectory in several IC product technologies

Taiwan

Local Taiwanese IC firms went directly into integrated operations when UMC was started

in 1980 (see Mathews and Cho 2000) from the acquisition of the semiconductor division

of RCA in the late 1970s (see Lin 2003) RCA offered UMC ASICs, diodes and transistortechnology

The merger between TSMC (49%) and Phillips (51%) in 1987 gave the Taiwanese firmsaccess to dynamic random access memory (DRAM) technology ERSO helped the incuba-tion and creation of several high-tech firms – including with R&D and wafer fabricationcapabilities – e.g Windbond, ASUS, Vanguard and ASE from 1983

The collapse of RCA coincided with the acquisition of the company’s semiconductordivision by ERSO in the late 1970s ERSO gave birth to UMC in 1980, which startedproducing ASICs for consumer electronics firms (see Mathews and Cho 2000, Ernst 2002,Mathews 2006) TSMC started subsequently in 1987 TSMC was the first contract ICmanufacturer to separate chip design from chip fabrication to specialize in the latter Waferfabrication with strong R&D support from ITRIs of ERSO led to the starting of eventually



front-end operations such as R&D and wafer fabrication Front-end did not integrate withthe old back-end firms in Taiwan, as the companies were different In the more integratedTaiwanese IC companies such as UMC and ASE, the back-end activities of assemblyand test were eventually relocated in China and Malaysia UMC has also relocated waferfabrication abroad in Singapore and the US.

Although UMC was the first Taiwanese fabrication plant to open in Taiwan, TSMC hasbecome the leading Taiwanese fabrication house since the 1990s, and hence the focus here

is on the latter Once the government successfully negotiated a joint venture with Philipsholding 51% of the share, Maurice Chang who had gained tacit and experiential knowledgeworking in American firms and left Texas Instruments as its senior vice president wasappointed as its founding chief executive officer Using his knowledge and linkages withR&D centres, universities buyers, suppliers and rivals played a key role in charting thedirection of the firm

Not only that, the leading local IC manufacturers in Taiwan (UMC, TSMC, ASEand Winbond) and China have linked with R&D and wafer fabrication that is targetedtowards sale of important manufacturers – especially computers (e.g Acer and Lenova andthe contract manufacturing firms such as Tatung, Vanguard and Asus) as well as exports(see Mathews and Cho 2000, Amsden and Chu 2003, Rasiah and Lin 2005) Some largeAmerican multinationals have also set up R&D and wafer fabrication plants in China andTaiwan Especially in Taiwan, considerable R&D offshoring has occurred from the late1990s – something that began with TSMC, a joint venture between Phillips and Taiwanesecapital that was started in 1986 The top 3 leading IC manufacturers in Taiwan are also inthe top 10 DRAM producers in the world Taiwan’s world-class machinery industry hasalso helped the complementary development of IC machinery and equipment

By 2007 TSMC had become completely Taiwanese owned From being the world’s firstindependent contract manufacturer of ICs, entering subsequently into the manufacturing of12-in wafers using nanotechnology, the firm announced plans to fabricate microprocessorsthrough contract arrangements in 2008 (TSMC, 20 February 2008) TSMC was the first

to specialize in IC fabrication, separating it from IC design, and is considered the worldleader in logic chips

China

Both foreign and local firms in China are engaged in assembly, test, fabrication and R&Dactivities on non-optical, optical, discrete, analog, logic, memories, application-specficstandard products and ASICs While much of the R&D is confined to ASICs, some firmsundertake DRAM fabrication and R&D (e.g Intel, Freescale and Qimonda) Government-funded R&D laboratories in China’s high-tech parks are working on DRAM R&D Govern-ment initiatives propelled China’s move to stimulate firms’ movement up the technologytrajectory in the 1990s

The key drive took place under Project 909 (launched in 1996), where with the support

of both the federal and Shanghai Government, Hong Hua was established through a strategicalliance with NEC of Japan to start an 8-in IC production line, design centre and R&Dcentre in 1997 (see Table 2) By 2003, Hong Hua had more than 10 subsidiaries supplyingICs and design services for smart cards, power metres, home networking, car electronicsand development of related application software Hong Hua developed China’s first radiofrequency identification chip in 2001, followed by chips for subscriber identification modulecards in 2002 and high-security microcontroller chip in 2003 In 2004, the company wasauthorized as the chip supplier of the China National ID Card Generation II Project



Table 2 Technology trajectory of lead local firms, China and Taiwan, 1975–2008.

2008 Announced plans to fabricate microprocessors

Source: Compiled from authors’ interviews (2007).

In 2007, Hong Hua was recognized internationally as having world-class capability inlayout engineering of digital, digital/analog, electrical erasable, programmable and read-only memory (EEPROM), embedded MCU and deep submicron digital ICs

Although there were more than 500 IC design houses in China in 2007, most of thesefirms focused on ASICs and little on the more lucrative market provided by DRAMs andmicroprocessors The acquisition of IBM computers by Lenovo is considered widely to

be driven by initiatives to quicken entry and acquisition of IC technology and markets.Hence, although the Chinese experience so far remains behind the frontier, the aggres-sive acquisitions alongside strong government support for learning has given Chinesefirms the confidence to scale the routes of upgrading in IC design, fabrication and R&Dquickly

Overall, ICs are at the technology frontier in process technology in both Taiwan andChina However, only firms in Taiwan are at the product technology frontier Although

IC firms in Taiwan are not involved in integrated operations and are still heavily buthorizontally reliant on strategic alliances for markets and technology, they are engaged incutting-edge product technologies in the segments they have entered China has increasedits share of designing activities in IC manufacturing through largely the operations of localfirms

4 Drivers of catch-up

Having established the technological paths, this section analyses the drivers behind themusing the broad inductive methodology advanced by Malerba 2005 The unfolding of thesepaths will help explain the differences, if any, of the catch-up patterns of IC firms in Chinaand Taiwan



Access to foreign knowledge

IC assembly and test began with the transfer of technology by multinationals to theirsubsidiaries in Taiwan from the 1960s and China from the 1980s Flagship IC firms such

as Intel, AMD, National Semiconductor, Hitachi and HP relocated the assembly and test

of the cutting-edge production technology to China The early phases were associated withemployment generation in EPZs with little focus on catch-up in both countries

Three routes to accessing foreign sources of knowledge in the IC catch-up track can

be identified from Taiwan and China In the first route, Taiwan and China ‘used’ foreignfirms as training grounds to access tacit knowledge In the second route, Taiwanese andChinese firms accessed foreign technology through licensing In the third route, Taiwaneseand Chinese firms merged or acquired foreign firms to access technology and markets.Taiwan led the way among the two countries in driving a catch-up in the IC industry.Taiwanese IC companies established strategic alliances with leading foreign companies.UMC grew from ITRI’s acquisition of RCA in 1979 TSMC started as a joint venture withPhilips in 1987 (Rasiah and Lin 2005) All wafer fabrication plants in Taiwan enjoy strongstrategic alliances with other IC and design firms

Technological partnerships between foreign and Taiwanese companies have taken placesince 1970s The early fabrication houses of UMC and TSMC in Taiwan merged first withRCA and Philips, respectively, in 1979 and 1987 before acquiring them However, the highlyde-verticalized Taiwanese firms remain strongly but horizontally linked to their strategicpartners Taiwanese firms took advantage of the mid-1980s, downswing in the industry:prices of 64K DRAM chips had fallen from$50 to 50 cents in 1980–1985 and the EEPROMfrom$18 to $4 in 3 months in 1985 (Rasiah 1988) This Schumpeter (1934) Mark I entry –the displacement of ailing incumbents (see Malerba 1992, Mathews 2006) such as Mostek(sold subsequently at a low price to Thomson CSF before International Device Technologyacquired it) using more efficient technology drawn from existing stocks of knowledgeand the phasing out of old product lines in AMD, Intel, Texas Instruments and NationalSemiconductor – coincided with the entry of Taiwanese contract IC manufacturers Much

of the early Taiwanese forays into foreign technology was done through ERSO The earlyacquisitions in Taiwan allowed a key point of entry, which was then transformed to drivefrontier research especially in memories (see Malerba 1992)

Chinese firms have also accessed foreign technology through licensing agreements butwith less success However, whereas the focus in Taiwan has been on accessing foreigntechnology by local firms moving up the product technology trajectory, in China foreignfirms are still dominated exports largely in the low-value-added production stages of assem-bly, packaging and test operations Initiatives to follow the Taiwanese framework started inChina in the 1990s as technological deepening became important Hence, both local and for-eign firms have managed to attract incentives to start R&D operations and wafer fabrication.The dramatic rise in technological catch-up took place in China, as there were morethan 500 local firms engaged in IC design and more than 10 firms in wafer fabricationplants in 2007 (Kong 2008) Chinese firms are better placed with strong domestic demandand committed government support The acquisition of IBM personal computer division byLenovo has also given Chinese firms a major boost for the acquisition of technology andmarkets

Demand conditions

Export markets were the critical initiator of large-scale IC manufacturing in Taiwan andChina and have remained important (see Table 3 and Figure 4) Domestic demand has also



Table 3 Trade balance, ICs and electronics, Taiwan and China, 2006.

Value (billion US$) Trade balanceExport (X) Import (M) (X− M)/(X + M)ICs

Source: World Trade Organization (2007, Tables 11.36 and 11.48).

risen strongly in China, as its IC market had become the third largest in the world by theend of the 1990s

Taiwanese consumer and information hardware electronics firms purchase the bulk oftheir ASICs and DRAMs from local IC firms User–producer relations have been strong

in driving wafer fabrication in Taiwan and China, as lock-ins have played a key role in thewafer fabrication start-ups Booking–billing ratios are important in low-margin fabrication,where yield is critical because of the lumpy nature of investment involved The bulk of ICchips are sold to lock-in contractors and the rest in open markets Strong interface between

IC firms and electronics firms that use chips has been an important element driving andshaping the flow of knowledge between them

The acquisition of IBM’s computer manufacturing division has also expanded theChinese-owned Lenovo’s market share in computers and with that the demand for Chinese

IC firms Strong export demand has led to the trade balance, involving the industry to reachpositive figures in Taiwan (see Table 3) The trade balance of China was negative The

2 3

2 0

Table 4 World market share of top 15 IC firms based on sales, 2005–2006.

Source: Global Sources, http://www.globalsources.com

high import figure for China is accounted by imports of wafers (especially by foreign firmslocated in EPZs) and machinery and equipment Nevertheless, the hugely positive tradebalance in the overall electronics industry in China shows that consumer and industrialelectronics firms acquire most of the components produced in their countries

Whereas IC firms in China have largely remained in assembly and test activities, theircounterparts in Taiwan have upgraded strongly to wafer fabrication, designing and R&Dactivities In addition to export demand, the OEM contract manufacturers – especially

in export-oriented computers and peripherals – provide considerable demand for the sale

of high-value-added chips in Taiwan Hence, domestic firms (including Taiwanese firmsthat have relocated front-end activities in China) have since the second half of the 1980sbecome major buyers of Taiwanese DRAMs and ASICs Both the Malaysian type of low-end foreign-dominated assembly and the Taiwanese types of increasing specialization inwafer fabrication and designing activities evolved in China (see Rasiah 2009)

Continuous upgrading has driven export shares in world exports of ICs in the period2000–2005 Exports as a share of world ICs exports from China rose from 1.7% to 3.7%

to 5.9% during 1990, 2000 and 2005 respectively (see World Trade Organization 2006,Table 4.59) The commensurate figures for Taiwan were 7.1%, 7.3% and 8.7% respectively.While export markets essentially provided the demand for IC firms in Taiwan andChina, the rapidly expanding domestic market has become an important consumer of ICoutput in China Wafer fabrication plants begun in Taiwan and China after the lock-in dealswere struck Whereas low-end IC assembly operations have been relocated out of Taiwan

to Southeast Asia and China from the 1990s, both low-end assembly and high-end waferfabrication and designing have evolved in China

The role of the government

Governments have played critical roles in the technological catch-up process in Taiwanand China From simply offering incentives to attract MNCs direct operations in the 1960sand early 1970s, the Taiwanese Government promoted directly the growth of local firms



from the 1970s and 1980s The government-led ITRIs that were started in 1974 (and theelectronics laboratories at ERSO) played a central role in driving R&D in local firms InTaiwan, the science and technology project grants of 1979 (particularly after they wereturned into matching grants with the private sector in 1983) and the Hsinchu Science Park

also offered tremendous R&D synergies (see Ernst et al 1998, Mathews and Cho 2000,

Amsden and Chu 2003, Rasiah and Lin 2005) In China, the MNC route to IC assembly andtest started in the 1980s and expanded sharply to drive Chinese IC exports Unlike Taiwan,some foreign firms began to participate in chip design and wafer fabrication activities inChina However, the large-scale transformation of the local IC firms into wafer fabricationand designing relied extensively on government support

Taiwan

Foreign multinationals were attracted initially with incentives in the late 1960s to undertakeexport-oriented assembly and test activities These firms provided jobs that were importantduring the 1960s RCA was one of the firms to relocate operations in Taiwan The gov-ernment then targeted the semiconductor division of RCA for the acquisition of strategictechnology The government was also involved in the development of incubators at ERSO

to undertake wafer fabrication activities which started IC manufacturing in Taiwan UMCwas the first to be launched by ERSO ERSO subsequently helped the incubation and cre-ation of several high-tech firms – including R&D and wafer fabrication capabilities – e.g.Windbond, ASUS, Vanguard and ASE from 1983

Grants initiated through incubators in the Hsinchu Science Park – which started nally in 1979 but became successful after a 1:1 matching condition was attached from 1983and the hiring of professionals brought back under the brain gain programme from the US –were instrumental in the launching of TSMC, Winbond, Vanguard, Asus and UMC Hence,

origi-in Taiwan, front-end wafer fabrication and R&D eventually took over from the origorigi-inalassembly and test operations undertaken by foreign multinationals from the late 1960s.TSMC had become the fourth largest seller of IC devices in the first half of 2006, risingfrom eighth place in 2005 (see Table 4)

Mathews (2006) argues that the government also timed its intervention to enter the ICindustry during an industry-wide global crisis when incumbent firms were facing severedownswings in the mid-1980s The government subsequently in 1991 formally approvedrelocation of labour-intensive low-value-added assembly and test activities to mainlandChina, which helped Taiwanese firms lower manufacturing costs while coordinating marketexpansion better (Rasiah and Lin 2005)

Taiwan’s technology transfer agreements also actively screened ex ante, monitored the use and diffusion and undertook ex post appraisal to ensure that licensing fees were brought

down, that diffusion occurred and that mistakes were not repeated The government inTaiwan launched an active education policy – at one level driving expansion in humancapital supply from technical schools and universities and at another level imposing levies

on unskilled labour imports to pressure firms to upgrade (Chen and Lee 2004) Governments

in both Taiwan and China also launched a brain gain programme to attract back citizensenjoying experiential and tacit knowledge

The role of the government through the operations of the ITRI laboratories has beencentral in the rapid movement of ICs firms towards the technology frontier The differentstages in the technological metamorphosis of the IC industry through specific government-funded ITRI projects are shown in Figure 5 The dynamic movement of Taiwanese firmsfrom simple transistors and ASICs in the early 1980s to advanced sub-micrometre 12-in



electronic project Sub- Micron project

Micro-Computer project

Computer project

Computer&

Telecom’n

Computer prod’t packaging Telecomn electronics

Advanced sub-micrometre project

Figure 5 ITRI’s IC projects Source: ITRI’s annual report, various issues

wafer DRAM chips in the 1990s owes very much to the R&D undertaken in the funded ITRI laboratories Access to foreign knowledge through multinationals and thereturning diaspora complemented government initiatives

China’s large labour force supports assembly and test in MNC operations in locationssuch as Pearl River Valley and Shenzen and local integrated firms located in high-techparks High-tech infrastructure, public R&D laboratories and universities play an importantrole in supporting the knowledge base of local firms in the high-tech parks in China.Several domestic IC manufacturing plants were launched through the national seventh,eight and ninth 5-year plans By the end of 1999, there were five large domestic and joint-venture companies – e.g Huajing, Huayue, Beiling, Xianjin and Shougang NEC Electronics– and by 2007, there were over 500 IC design houses in China (Kong 2008) In addition,there were also 10 specialized and 871 electronics factories in the GanSu Province, enjoyingsupport from the Microelectronics Industrial Centre of Chinese Academy of Sciences ICmanufacturing industry developed rapidly from late 1990s Shougang NEC Electronics

Co Ltd started production of 0.56 64M DRAM in 1996 Huajing started production in its



Table 5 Composition of IC value chain, China, 2001–2006 (billion yuan).

Year Design Manufacturing Packaging and testing Total

on mainland semiconductor industry’, CCID-MRD, P43; Editing committee of China Industrial Maps, China Economy Booming Inspection and Foresight, China Industrial Maps 2004–2008, Social Sciences Academic Press.

0.96 CMOS line in 1998 Shanghai Huahong started production of 0.58 MOS line in 1999.Compared with the IC manufacturing industry and the IC design industry, the IC packagingand testing industry was much larger in China

Huada IC Design Centre – which was set up in 1986 as the first IC design company inChina – became important after the government resumed interest in promoting local firmsfrom 1996 Document 18 issued by the State Council in 2000 has been instrumental indriving IC design development in China IC design output rose from 1.48 billion yuan in

2001 to 18.62 billion yuan in 2006 – expanding by around 13 times (see Table 5) By theend of 2006, the number of IC design firms in China reached 488 The share of IC designindustrial sales in overall IC industrial sales reached 18.5% in 2006 IC design humancapital also increased in numbers from less than 5000 to more than 20,000 in 2006 Theshare of the low-value-added packaging and testing activities in total sales of ICs in Chinafell from 79.1% in 2001 to 49.3% in 2006

Because it costs around$3 billion to ramp up a wafer fabrication plant and around $10billion for the plant to enjoy the scale necessary to make profits, in 2000 the governmentshave either directly or indirectly funded the founding of the early fabrication plants inTaiwan and China The Taiwanese Government took 49% equity in 1987 when TSMCwas founded, and since 2008 it has taken control of the firm The Chinese provincialgovernments have invested extensively to start IC design and fabrication houses

Human capital

IC manufacturing became knowledge-intensive from the late 1970s Hence, minimumstatistical numeracy and communicative, cognitive and judgemental skills are importantbefore firms hire even operators Intel’s super-operators in China enjoyed a wage premium

in 2008 The use of just-in-time and kaizen-type practices has raised the demand for skilledworkers Schooling in Taiwan and China provide such labour In addition, participation inwafer fabrication and new product development requires strong supplies of R&D engineersand technicians

The share of R&D scientists and engineers in the population of Taiwan rose from

3326 in 1996 to 4159 in 2006, falling only in 2002 (see Table 6) However, labour kets for skilled labour tightened in Taiwan from the 1980s Taiwanese firms relocatedlabour-intensive assembly and test in Southeast Asia since 1985 and China since 1991

mar-As IC firms upgraded to participate in wafer fabrication and designing, the intensity ofR&D scientists and engineers – drawn from both a massive expansion in domestic supply



Table 6 R&D scientists and engineers per million people, selected economies, 1996–2006.

Source: Computed from UNESCO (various issues); World Bank (2010).

and the returning diaspora – rose strongly in Taiwan Particularly, human capital endowedwith tacit knowledge from working education and experience in the US helped Taiwanesefirms make the first separation between the three UMC, TSMC, ASE and Windbond suc-cessfully specialized in wafer fabrication without evolving R&D capabilities internally.Taiwan supplemented demand for such human capital by stepping up their supply as well

as through talent-attracting programmes Taiwan is by far the most successful between thetwo countries, as it has managed to attract back specialists who had gained tacit and experi-ential knowledge by working especially in American multinationals, R&D laboratories anduniversity R&D centres TSMC, ASUS, Vanguard and Winbond are examples of high-techfirms run by these experts Maurice Chang has been the chief executive officer of TSMCfrom its founding until now

China’s conversion from centrally planned economy to markets has also provided enoughhuman capital to drive catch-up in both product and process technologies The engineeringintensity of GanSu High-Tech Park in China exceeded that of Hsinchu Science Park inTaiwan in 2005 As a share of the population, R&D scientists and engineers in China rosefrom 657 in 1996 to 1143 in 2006, recording a fall only in 1998 (see Table 6) The successfulexperience of Taiwan and Korea in attracting their diaspora led to China adopting the samepolicy to attract back its talent from abroad

Networks, alliances and consortiums

Consortiums, strategic alliances and the co-evolution through industry and laboratory flows of knowledge have been critical in Taiwan in the catch-up experience of

cross-IC firms They have not been as significant in China The independent local designingand fabrication plants of Taiwan have relied extensively on formally registered strategicalliances with foreign firms and R&D centres for key technologies and markets

Strategic alliances with foreign firms and laboratories were critical in the catch-up stage

of Taiwanese firms Strategic alliances are also critical in Chinese firms, but given theirinfancy it is too early to predict their direction Nevertheless, initiatives are being taken bythe provincial and county governments to establish collaboration and technology tie-upsbetween local Chinese and foreign IC firms (Kong 2008)

Taiwan provides the alternative de-verticalized route to reaching the technology frontier.The establishment of TSMC’s Design Center Alliance has been highly strategic in attractingaccess to new technology, which is vital for it to specialise in foundry-based fabrication.Accent joined this alliance in 2004 to supply complex customer designs for TSMC’s processtechnologies Membership in TSMC’s Design Center Alliance allows Accent to service



companies wishing to utilize TSMC’s foundry operations Accent supplies design supportfor OEM, fabless and chipless small- to medium-sized enterprises, as well as start-ups.Recent projects by Accent for TSMC include the analog-mixed signal IC, capacitiveinterface sensors, a multi-million gate IC for wireless applications, the hardening of anARM central processing units, systems with large memories and analog blocks and avery complex IC for networking applications, which include 0.13-, 0.18- and 0.25-μmtechnologies Accent also supplies its very deep-sub-micrometre design for the delivery ofincoming 90- and 45-nm designs.

TSMC announced in February 2008 to establish production lines to start fabricatingcentral processing units on a 45-nm high-end process for low-cost personal computers with

a target of start sales by the second half of the year AMD is widely believed to have giventhe contract to TSMC to develop a low-cost method to produce microprocessors, which haveopened the way for catch-up from Asian firms without which the stringent copyright andpatent laws have for long threatened to restrict new entrants in the product line Entry intomicroprocessors also offers Taiwanese firms the opportunity to leapfrog Korean, Germanand American IC firms

Taiwanese firms have by far benefited most from the co-evolution of industries.Machinery and chemicals had emerged on a large scale by the early 1980s in Taiwan(see Fransman 1985, Rasiah and Lin 2005) The co-evolution of these industries as well

as the plastics industry through support from ITRI helped provide the complementaritiesessential for process and product technology improvements in Taiwan’s IC firms Indeed anumber of latest technologies that were developed at ERSO quickly found their place in thechips fabricated by Taiwanese firms A number of such complementary technologies weredeveloped in the chemical, plastics and machinery and equipment laboratories One suchexample is the thin membrane developed to separate transistors in wafers

In Taiwan, the IC industry simultaneously has been the driver and is being driven byother industries On the one hand, major developments in chemicals and plastics and inmachinery and equipment have enhanced IC product technologies and process technologiesrespectively On the other hand, the proliferation of IC technology as the driver has benefitedenormously the development of automotive parts, computer numeric control and electronicdevice machinery and fish tracking equipment Microchips fabricated by Taiwanese firmsalso power the industries that manufacture chemicals and plastics

Local machine tool and plastic injection moulding firms co-evolved with upgradingand flexibilization in IC firms in Shenzen, China, from the 1990s However, despite risingproduction domestically, fabricated wafers are also primarily imported into China Hence,the components of the IC cluster in China are still not as compactly developed as inTaiwan

5 Conclusions and implications

Despite acquiring the OEM capability to fabricate cutting-edge DRAMs, Nand Flash andother CMOS products on 12-in wafers, Taiwanese IC firms still rely considerably on R&Dand designing support from abroad Chinese firms are in the same situation but have notacquired the OEM capabilities to fabricate the latest chips Taiwanese firms have relocatedmuch of their assembly and test activities in China and Southeast Asia, while Chinese firmsstill largely specialize in these activities

The path taken by IC firms in Taiwan is very different from that of IC firms in China.Taiwanese firms have remained highly specialised in wafer fabrication or vertically inte-grated operations within ICs while continuing to depend strongly on R&D support from



ERSO and foreign firms Taiwanese firms also continue to enjoy strong collaboration, inparticular participation in R&D activities, among themselves as well as foreign researchlaboratories and designing companies In addition, some Taiwanese firms such as TSMCand UMC have continued to specialize in contract wafer fabrication foundries withoutparticipation in assembly and testing These firms have also continued to rely extensively

on strategic alliances for both technological support and markets Different firms in wan participate in chip design and software systems, which is similar with local Chinesefirms

Tai-Export demand initiated IC manufacturing in Taiwan and China Tai-Exports continue todominate IC demand of Taiwanese firms, though the low-end labour-intensive stages ofassembly and test have been relocated to China and Southeast Asia Domestic demandonly became important in China as expansion into consumer and industrial electronicsattracted significant production for use in firms located ICs among the two countries in

2006 China experienced a negative trade balance in 1990–2006 Whereas the development

of technological capabilities in wafer fabrication and other inputs helped Taiwanese firmsrecord a positive trade balance, a short period of participation primarily explains the negativebalance in China Nevertheless, China recorded a positive trade balance in the broaderelectronics industry, demonstrating that domestic demand has grown for ICs

While multinationals started the first large-scale assembly and test of IC devices, withstrong government support local firms spearheaded the first critical mass of fabrication anddesigning activities in both countries Despite the differences, the evidence amassed in thepaper using the lenses provided by the sectoral innovation theory (Malerba 2005) showssignificant similarities in the channels that drove catch-up The three major similaritiesacross the two countries are the role of government, integration in multinational valuechains and local firms spearheading technological catch-up

Government policy has been central in the emergence and growth of the added activities of wafer fabrication and designing in Taiwan and China However,the nature of state intervention has differed Taiwan shifted from its 1960s policies ofattracting foreign direct investment to create jobs to support the development of localfirms from the 1970s Hence, the leading semiconductor firms in Taiwan are local firms,while those in China are foreign firms Also, Taiwanese firms have successfully accessedR&D laboratories directly from ERSO and the other ITRI laboratories for complementarytechnologies Government has also been a critical pillar in the provision of human capital(including attracting experts from abroad), R&D grants and in Taiwan the provision ofhigh-tech facilities such as R&D laboratories Chinese wafer firms demonstrate bettercapacity to upgrade because of large reserves of human capital and better coordinationwith performance standards However, having emerged only after 1997, Chinese designfirms are still engaged in the manufacture of low-end EEPROMS, ASICs and ASSCs andhave yet to participate in the higher end of semiconductors

high-value-Only in Taiwan the evidence is clear that the co-evolution of other industries and gic alliances have been important Strategic alliances through R&D consortiums have beenimportant in driving the utilization of state-of-the-art technology to support de-verticalizedstructure of Taiwanese IC firms The participation of ITRI’s world-class laboratories inseveral different industries and its smooth integration with firms has also facilitated the ap-propriation of complementary knowledge synergies from chemical, machinery and plasticslaboratories The de-verticalized local firms of China appear to be following the Taiwaneseframework, but it only appears promising in the former The rising share of sales con-tributed by IC design and fabrication in overall IC sales over the period 2001–2006 is agood indicator that IC firms in China are catching up in the value chain

strate-

Amsden, A., 1985 The division of labor is limited by the rate of growth of the market: the Taiwanese

machine tool industry Cambridge journal of economics, 9 (4), 271–284.

Amsden, A., 1989 Asia’s next giant: South Korea and late industrialization New York: Oxford

University Press

Amsden, A and Chu, W.W., 2003 Beyond late development: Taiwan’s upgrading policies Cambridge,

MA: MIT Press

Becattini, G., 1990 The Marshallian industrial district as a socio-economic notion In: F Pyke, G Becattini, and W Sengenberger, eds Industrial districts and inter-firm cooperation in Italy.

Geneva: International Labour Organization, 37–51

Brown, C and Linden, G., 2009 Chips and change: how the crisis reshapes the semiconductor

industry Cambridge: MIT Press.

Brusco, S., 1982 The Emilian model: productive decentralization and social integration, Cambridge

journal of economics, 6 (2), 167–184.

Chang, H.J., 2003 Kicking away the ladder London: Anthem Press.

Chen, T.J and Lee, J.S., eds., 2004 The new knowledge economy of Taiwan Northampton: Edward

Elgar

Ernst, D., 2002 Global production networks and the changing geography of innovation systems

Implications for developing countries Economics of innovation and new technology, 11 (6),

497–523

Ernst, D., Ganiatsos, T., and Mytelka, L., eds., 1998 Technological capabilities and export success:

lessons from East Asia London: Routledge.

Fransman, M., 1985 International competitiveness, technical change and the state: the machine tool

industries in Taiwan and Japan World development, 14 (12), 1375–1396.

Freeman, C., 1989 New technology and catching-up European journal of development research, 1

(1), 85–99

Gerschenkron, A., 1962 Economic backwardness in historical perspective: a book of essays

Cam-bridge: Belknap Press of Harvard University Press

Hamilton, A., 1791 Report on manufactures [online] Available from: http://www.oberlin.

edu/∼gkornbl/Hist258/ReportMfres.html [Accessed 13 December 2005]

Hobday, M., 1995 Innovation in East Asia Cheltenham, UK: Edward Elgar.

Johnson, C., 1982 MITI and the Japanese miracle Stanford: Stanford University Press.

Kim, L., 1997 From imitation to innovation Cambridge, MA: Harvard Business School Press.

Kong, X.X., 2008 The development of the semiconductor industry in China Unpublished graph Beijing

Mimeo-Lee, M.L., Liu, B.C., and Wang, P., 1994 Education, human capital enhancement and economic

development: comparison between Korea and Taiwan Economics of education review, 13 (4),

275–288

Lim, L.Y.C., 1978 Multinational firms and manufacturing for export in less-developed countries: thecase of the electronics industry in Malaysia and Singapore Unpublished thesis University ofMichigan, Ann Arbor

Lin, Y., 2003 Industrial structure, technical change and the role of government in the development

of the electronics and information technology industry in Taiwan Working Paper Manila: Asian

Development Bank



List, F., 1885 The national system of political economy London: Longmans, Green & Company.

Lundvall, B.A., 1988 Innovation as an interactive process: from user–producer interaction to the

national system of innovation In: G Dosi, C Freeman, G Silverberg, and L Soete, eds Technical

change and economic geography London: Frances Pinter, 349–369.

Lundvall, B.A., 1992 National systems of innovation: towards a theory of innovation and interactive

learning London: Frances Pinter.

Malerba, F., 1992 Learning by firms and incremental technical change Economic journal, 102 (413),

845–859

Malerba, F., 2005 Sectoral systems of innovation: a framework for linking innovation to the knowledge

base, structure and dynamics of sectors Economics of innovation and new technology, 14 (1–2),

63–82

Malerba, F., Nelson, R.R., Orsenigo, L., and Winter, S., 2001 Competition and industrial policies

in a ‘history friendly’ model of the evolution of the computer industry International journal of

industrial organization, 19 (5), 635–664.

Malerba, F and Orsenigo, L., 1997 Technological regimes and sectoral patterns of innovative

activi-ties Industrial and corporate change, 6 (1), 83–117.

Marsh, H., 1981 Silicon chip book London: Pelican Press.

Mathews, J.A., 2006 Electronics in Taiwan: a case of technological learning In: V Chandra, ed.

Technology, adaptation and exports: how developing countries got it right Washington DC:

World Bank, 83–126

Mathews, J.A and Cho, D.S., 2000 Tiger technology: the creation of a semiconductor industry in

East Asia Cambridge: Cambridge University Press.

McFarlane, B., 1981 Australian capitalism in boom and depression Kuala Lumpur: APCD

Rasiah, R., 1988 The semiconductor industry in penang: implications for the new international

division of labour theories Journal of contemporary Asia, 18 (3), 44–69.

Rasiah, R., 1994 Flexible production systems and local machine tool subcontracting: electronics

transnationals in Malaysia Cambridge journal of economics, 18 (3), 279–298.

Rasiah, R., 1995 Foreign capital and industrialization in Malaysia Basingstoke: Macmillan.

Rasiah, R., 1996 Institutions and innovations: moving towards the technology frontier in the

elec-tronics industry in Malaysia Industry and innovation, 3 (2), 79–102.

Rasiah, R., 2008 Ownership, institutions and technological intensities: automotive and electronics

firms in east and Southeast Asia In: J Eatwell and P Arestis, eds Issues in finance and industry:

essays in honour of Ajit Singh London: Pelgrave Macmillan, 137–166.

Rasiah, R., 2009 Growth and slowdown in the electronics industry in Southeast Asia Journal of Asia

pacific economy, 14 (2), 123–137.

Rasiah, R., Kong, X., and Vinanchiarachi, J., forthcoming Moving up the value chain in button

manufacturing in China Asia Pacific business review.

Rasiah, R and Lin, Y., 2005 Learning and innovation: the role of market, government and trust in the

information hardware industry in Taiwan International journal of technology and globalization,

1 (3–4), 400–432

Reinert, E., 2007 How rich countries got rich and why poor countries stay poor New York: Carroll

& Graf Publishers

Schumpeter, J., 1934 The theory of economic development Cambridge: MIT Press.

Shilov, A., 2007 TSMC confirms entering microprocessor manufacturing business [online] Available

from: http://www.xbitlabs.com/news/cpu/display/20070727063402.html [Accessed 22 ber 2009]

Decem-World Bank, 2010 Decem-World development indicators Washington, DC: Decem-World Bank Institute.

World Trade Organization, 2006 International trade statistics Geneva: World Trade Organization World Trade Organization, 2007 International trade statistics Geneva: World Trade Organization.



lessons from the Indian case

in-to deploy borrowed technology efficiently in production and compete internationally.The state must also curtail consumption to build physical and educational infrastructure.However, current theories of late-late industrialization, which mostly draw from the EastAsian manufacturing experience, demand reexamination thanks to the characteristics ofinformation and communication technologies, especially software These characteristicsinclude rapid technological change, which can preclude learning from borrowed tech-nology, and the need to encourage consumption, since information and communicationtechnologies are general purpose technologies that can potentially enhance aggregateproductivity Further, as the marginal cost of production of software is negligible, effi-cient manufacturing is not an issue This paper shows how our understanding of late-lateindustrialization may be broadened by drawing on the experience of the Indian softwareindustry

1 Introduction

How less-affluent countries experience economic growth and social change by izing, to ‘catch-up’ with their prosperous counterparts, is a question of much intellectualand policy relevance The work of Gerschenkron (1962) on nineteenth-century Germanyand Russia, and that of Johnson (1982) on mid-twentieth-century Japan, showed that lateindustrialization cannot be left to ‘free’ markets; instead, societies must develop the institu-tional means to construct comparative advantage rather than relying exclusively on naturalendowments On the basis of the studies of South Korea and Taiwan, Amsden (1989) andAmsden and Chu (2003) propose a general theory of industrial transformation in the latetwentieth century (or late-late industrialization) that emphasizes technological upgrading.They argue that reciprocity between the state and industry is a key premise oflate-late industrialization amid economic globalization There is a pronounced need forstate intervention to subsidize those sectors in which it wants to build comparative ad-vantage by distorting prices and combining financial incentives with tariff protection In

industrial-

exchange for subsidies, the Korean state, for instance, imposed performance standards onindustry in the form of export targets Firms receiving subsidies imported technology andlearned to use it efficiently by world standards Efficiency gains were made by buildinglarge production facilities to exploit economies of scale and by reverse-engineering im-ported products and production processes Besides providing subsidies, the state must alsocurtail current consumption and invest the savings to build physical infrastructure and todevelop an education system This is because late-late industrialization is led by a laborforce that is much better educated than the ones which led the industrialization process inearlier eras Production engineers, rather than researchers, are critical to learning, for whichthe shop floor, rather than the corporate research and development (R&D) laboratory, is theprimary site This educated labor force is paid relatively low wages, at least by the standards

of industrialized countries The ability to deploy low-wage but increasingly well-educatedlabor makes it possible to efficiently produce and export a range of low-cost products thatrely on borrowed technology

Rasiah and Hing (2009) contrast the export-led Northeast Asian model of late-lateindustrialization with the import-substitution-led industrialization (ISI) model pursued inSoutheast Asia.1The domestic focus of the ISI model denied the benefits of scale economies

in relatively small domestic markets, while the close relationship between the state andlocal capital in protected environments only led to cronyism, which dampened competitivepressures for technological upgrading After the limits of this model were exposed bythe financial crisis of the late 1990s, Southeast Asia also began to shift toward foreign-capital-led export-oriented industrialization, a strategy long pursued by Singapore Buteven in Singapore, ‘government intervention was crucial to strengthen the quality of thehigh-technology environment [and] stimulate upgrading in multinationals’ (Rasiahand Hing 2009, p 109)

Thus, the underlying premise of late-late industrialization, about the importance ofstate–society partnership, has gained wide acceptance, from scholars such as Wade (1990),Evans (1995) and Kohli (2004), to international organizations (United Nations Indus-trial Development Organization 2009) While concurring with that premise, this paperargues, however, that an understanding of late-late industrialization based on the EastAsian manufacturing experience must be broadened to accommodate the characteristics ofmicroelectronics-based information and communications technologies (ICTs) for at leastthree reasons First, while it is possible to rely on learning from borrowed technology forcertain segments of the sector (Mathews and Cho 2000), technological change in the sector

is rapid, with new product and process cycles set in motion even before a product and itstechnology are standardized

Second, not only have ICTs emerged as a new sector, but they are also at the heart of

a third industrial revolution because they are general purpose technologies that have thepotential ‘to transform an economy by finding new applications and fusing with existingtechnologies to rejuvenate other, preexisting sectors of the economy’ (David 2000, p 75)

In other words, encouraging consumption, not only of ICTs but in every domain of theeconomy where ICTs find application, is essential to improving productivity and aggregatestandards of living To invoke Bell (1999), the utility of information-processing capabilities

is limited by the social context to understand the structure and content of information.This point is also made for the software industry by Schware (1992), who argues for theadoption of a ‘walking on two legs’ strategy Such a strategy entails developing a domesticmarket for various software application domains not only to automate various sectors butalso to help firms hone their expertise and experience locally before serving global marketsand to minimize the risk of being confined to low-value-added work The reference to



software is particularly important because while advances in design and manufacturingtechnologies have physically shrunk and enabled the inexpensive mass production ofinformation-processing devices, as predicted by Moore’s law, the functional versatility

of ICTs comes from being able to program them to undertake a range of tasks In otherwords, although it has no materiality, it is software that increasingly gives contemporaryICTs their revolutionary character Lacking materiality, software is electronically dupli-

cated, and production costs are a nonissue (Varian et al 2005) Instead, it is design and

development efforts that determine the cost of software

2 The case of the Indian software industry

To understand how the characteristics of the software industry can broaden our standing of late-late industrialization, this paper will draw on the Indian experience TheIndian software and services industry grew from less than$100 million in 1985–1986 tonearly$50 billion in 2006–2007 (Table 1) With nearly 80% of this revenue coming fromexports, India became the world’s largest exporter of ICT services (World Bank 2009).Besides the quantitative expansion of the industry, it has also undergone qualitative shifts:from being a provider of low-wage, even if high-skill, services, to a source of innovation

under-As Table 2 shows, in 2006–2007, R&D and engineering services accounted for 15.7% ofsoftware and service exports How did a country, among the poorest and most illiterate in

Table 1 India’s software and service revenues and exports, 1985–1986 to 2006–2007

Table 2 Software exports – sectoral contribution, 2001–2002 to 2006–2007.

R&D service Embedded systemsITES-BPO exports exports as a exports as a

as a percentage percentage total percentage oftotal software and software and R&DYear services exports services exports service exports

Note: ITES, information technology enabled services; BPO, business process outsourcing.

Sources: NASSCOM (www.nasscom.in).

the world, achieve such export growth?2This paper will explain the quantitative expansionand the qualitative shifts in Indian software service exports by examining how changingtechnological conditions and policies have created a historically specific organization ofproduction for the industry in the country This paper will show how these changes can beunderstood in five phases

As Section 2 will describe, until 1984 (Phase I), there was virtually no software industry,thanks to an autarkic, public sector enterprise (PSE)-dominated, ISI policy regime thatdiscouraged entrepreneurship and foreign investment and proved inimical to innovation(Ahluwalia 1985) Indeed, when India established the Department of Electronics (DoE);(the Department of Information Technology since November 2000) in 1971 to formulateelectronics policy, India’s electronics production was larger than that of Korea (Sridharan1996) Yet, a decade later, Korean production and exports dwarfed India’s With efforts

to liberalize these policies from the mid-1980s, there emerged an export-driven softwareindustry However, prior to 1990 (Phase II), exports involved little more than bodyshopping,

or the practice of providing inexpensive on-site (i.e at customer locations overseas) labor

on an hourly basis, for low-value-added programming services such as coding and testing.Section 3 will explain why it was only in the 1990s (Phase III), coinciding with theprovision of data communication facilities and greater openness to the world economy,that a territorially grounded industry emerged in India It was also in this period thatBangalore became the leading region for the industry, thanks to its endowment of skilledlabor, and came to be referred to by terms such as ‘Silicon Plateau’ (Fineman 1991) or

‘India’s Silicon Valley’ (International Data Group 2001) Yet, this section will also explainwhy, despite becoming the center of the industry in the liberal economic climate of the1990s, Bangalore lacked the innovative milieu of the Silicon Valley Indeed, with little localinteraction between export-oriented firms and other institutions, the industry responded

to surging labor market pressures and the spread of communications infrastructure withdecentralization away from Bangalore

However, as Section 4 discusses, after 2001 (Phase IV), alongside decentralization,forces of agglomeration began to assert themselves in Bangalore With the global decline indemand for software professionals, following the overinvestment in information technology,there was a reverse flow of expatriate Indians with entrepreneurial and technical skills,mostly to Bangalore Their skills built on the existing labor pool in Bangalore to help move