Foreign Investment, Development, and Globalization: Can Costa Rica Become Ireland?

Foreign Investment, Development, and Globalization This page intentionally left blank Foreign Investment, Developme...

Author: Eva Paus

10 downloads 1277 Views 1MB Size Report

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

Report copyright / DMCA form

DOWNLOAD PDF

Foreign Investment, Development, and Globalization

This page intentionally left blank

Foreign Investment, Development, and Globalization Can Costa Rica Become Ireland? Eva Paus

FOREIGN INVESTMENT, DEVELOPMENT, AND GLOBALIZATION

© Eva Paus, 2005. All rights reserved. No part of this book may be used or reproduced in any manner whatsoever without written permission except in the case of brief quotations embodied in critical articles or reviews. First published in 2005 by PALGRAVE MACMILLAN™ 175 Fifth Avenue, New York, N.Y. 10010 and Houndmills, Basingstoke, Hampshire, England RG21 6XS Companies and representatives throughout the world. PALGRAVE MACMILLAN is the global academic imprint of the Palgrave Macmillan division of St. Martin’s Press, LLC and of Palgrave Macmillan Ltd. Macmillan® is a registered trademark in the United States, United Kingdom and other countries. Palgrave is a registered trademark in the European Union and other countries. ISBN 1–4039–6983–3 Library of Congress Cataloging-in-Publication Data Paus, Eva. Foreign investment, development, and globalization : can Costa Rica become Ireland? / Eva Paus. p. cm. Includes bibliographical references and index. ISBN 1–4039–6983–3 1. Investments, Foreign—Ireland. 2. International business enterprises—Ireland. 3. Investments, Foreign—Costa Rica. 4. International business enterprises—Costa Rica. 5. High technology industries. 6. Globalization. I. Title. HG5450.3.A3P38 2005 332.67⬘3⬘097286—dc22

2005045940

A catalogue record for this book is available from the British Library. Design by Newgen Imaging Systems (P) Ltd., Chennai, India. First edition: October 2005 10 9 8 7 6 5 4 3 2 1 Printed in the United States of America.

Contents

List of Tables

vii

List of Figures and Graphs

ix

Acronyms and Abbreviations

xi

Acknowledgments

xv

1. Foreign Direct Investment in the Global Age: New Opportunities for Development?

1

2. High-Tech FDI-Led Growth in Small Latecomers: An Analytical Framework

11

3. The Rise of the Celtic Tiger

45

4. The Development of Knowledge-Based Assets in Ireland

83

5. From Coffee to Computers: High-Tech FDI in Costa Rica

135

6. The Tico Tiger That Hasn’t . . .

173

7. Lessons from Ireland and Costa Rica

189

Notes

207

References

225

Index

239

This page intentionally left blank

List of Tables

1.1 2.1 2.2 2.3 2.4 3.1 3.2 3.3 3.4 3.5 3.6 3.7 3.8 3.9

3.10 3.11 4.1

4.2

4.3 4.4

Distribution of Global Foreign Direct Investment Inflows, 1970–2002 World Trade by Factor Intensity and Demand, 1985 and 2000 Effective Tax Rates for Subsidiaries of U.S. TNCs Abroad, 1999 and 2002 Number of ISO 9000 Certificates, 1993–2000 The Top Four Contract Manufacturers Basic Economic Data on Ireland, 1960–2002 Structure of the Irish Manufacturing Sector Based on Output and Employment, 1991 and 2000 Employment Distribution in the Irish Economy, 1994 and 2002 Taxes and Government Revenue as a Share of GDP, OECD Countries, 1995 and 2002 U.S. Direct Investment Position Abroad, 1982–2002 Marshall Aid as a Share of Recipients’ GDP, 1948–1952 Allocation of EU Funds for Ireland, 1989–2006 School Enrollment Rates in Ireland, 1970–1999 Average Compensation and Annual Hours per Worker in Majority-Owned Affiliates of U.S. TNCs, 1982, 1989, 1994 IDA Ireland—Grant Payments by Type, 1992–2001 IDA Grant per Job Sustained, 1981–2001 Materials Produced in Ireland as a Percentage of All Materials Purchased by Foreign-Owned Enterprises (FOEs), 1985–2002 Materials Produced in Ireland as a Percentage of All Materials Purchased by Indigenously-Owned Enterprises (IOEs), 1985–2002 Percentage of Manufacturing Plants Using Selected Management and Control Systems in Ireland, 1999 Channels and Extent of Knowledge Transfers between TNCs and Suppliers in Ireland

4 14 20 31 36 48 54 58 60 62 65 66 68

69 73 73

86

89 91 92

viii

4.5 4.6 4.7 4.8 4.9 4.10 4.11 4.12 5.1 5.2 5.3 5.4 5.5 5.6 5.7 5.8 5.9 5.10 5.11 5.12 6.1 6.2 6.3 6.4 6.5

LIST OF TABLES

Cost per Job Sustained in Companies Supported by IDA and EI, 1987–2002 Constraints on Local Sourcing by TNCs Composition of Employed Workers in High-Tech Industries, Ireland, 1991 and 1999 Business Expenditures on Research and Development, FOEs and IOEs in Ireland Exports and Export-Intensity in the Irish Manufacturing Sector, 1991–2001 Productivity and Wages in FOEs Relative to IOEs, 1991 and 2000 Ireland’s R&D Efforts in Comparative Perspective Factors Hampering Innovation, 1998–2000 Basic Economic Data: Costa Rica, 1970–2002 U.S. Aid to Costa Rica, 1980–2001 Net Foreign Direct Investment in Costa Rica by Sector, 1997–2003 Costa Rica’s Export Composition, 1991–2003 Net Foreign Exchange Generation by Special Export Regimes: An Approximation Government Revenue/GDP: Costa Rica, 1991–2003 Costa Rica’s Manufactured Exports, 1997 and 2003 Export Decomposition by Technology Intensity, Selected Countries, 1985 and 2000 Manufacturing Output and Employment: Costa Rica, 1997–2003 Export Intensity of Costa Rica’s Manufacturing Sector, 1997–2003 Structure of the Costa Rican Manufacturing Sector, 1997 and 2003 Labor Costs in the Production of Medical Devices in 2002 Use of Domestic Inputs in Costa Rica’s Free Zones, 1998–2002 Survey Results on Imported Input Share in Costa Rican Manufacturing, 2001 Average Age of Machinery Used in the Production Process, Costa Rican Manufacturing Firms, 2001 (Survey Results) Value of U.S. Content in U.S. Imports under HTS 9802, 2001–2003 Availability of Training Systems in Costa Rican Manufacturing Firms (Survey Results)

97 109 114 116 123 124 127 128 139 141 144 146 147 148 151 152 153 154 155 160 175 175 178 181 182

List of Figures and Graphs

Figures 2.1 2.2 3.1 3.2

High-Tech FDI in SLCs: Critical Contingencies From High-Tech FDI in SLCs to Industrial Upgrading High-Tech FDI in Ireland in the 1990s: Critical Contingencies The Impact of the Celtic Tiger on Ireland’s Location-Specific Advantages 5.1 High-Tech FDI in Costa Rica in the Late 1990s: Critical Contingencies

21 23 76 78 169

Graphs 3.1 Employment in Foreign-owned and Irish-owned Manufacturing Enterprises, 1972–2001 3.2 GNI/GDP Ratio, Ireland, 1960–2002

47 52

This page intentionally left blank

Acronyms and Abbreviations

ABSEI ATC AVL BEA CACM CAFTA CAT CBI CENPRO CEPAL CFR CINDE CIP CM CNC CODESA CRP CSF CSO EC EEC EI EMS ESIM ESRI EU FAS FDA FDI FIAS FOE FOU

Annual Business Survey of Economic Impact Agreement on Textiles and Clothing Approved Vendor List Bureau of Economic Analysis Central American Common Market Central American Free Trade Agreement Certificados de Abono Tributario Caribbean Basin Initiative Centro de Promociones de Exportaciones e Inversiones Comisión Económica para América Latina y el Caribe Code of Federal Regulations Coalición Costarricense de Iniciatívas para el Desarrollo Census of Industrial Production Contract Manufacturer Computer Numerical Control Corporación Costarricense de Desarrollo Sociedad Anónima Costa Rica Provee Community Support Framework Central Statistics Office European Community European Economic Community Enterprise Ireland European Monetary System Employment Survey of Irish Manufacturing and International Services Economic and Social Research Institute European Union Ireland’s National Training and Employment Authority Food and Drug Administration Foreign Direct Investment Foreign Investment Advisory Service Foreign-Owned Enterprise Foreign-Owned Unit

xii

ACRONYMS AND ABBREVIATIONS

FZ GATT GCF GDP GFCF GNI IADB ICE ICSTI ICT IDA IFSC IICE IEES IMF IOE IRA ISA ISI ISO IT ITCR MINEX NAFTA NIERC NLP OECD OBM ODM OEM PC PLN PMA PROCOMER PROFOVE PUSC SFI SLC

Free Zone General Agreement on Tariffs and Trade Gross Capital Formation Gross Domestic Product Gross Fixed Capital Formation Gross National Income Inter-American Development Bank Instituto Costarricence de Electricidad Irish Council for Science, Technology, and Innovation Information and Communication Technology Industrial Development Authority, later Industrial Development Agency International Financial Services Center Instituto de Investigaciones en Ciencias Económicas Irish Economy Expenditure Survey International Monetary Fund Indigenously Owned Enterprise Irish Republican Army The International Federation of the National Standardizing Associations Import-substituting industrialization International Organization for Standardization Information Technology Instituto Tecnológico de Costa Rica Ministério de Exportaciones North American Free Trade Agreement Northern Ireland Economic Research Centre National Linkage Program Organization for Economic Cooperation and Development Original Brand Name Manufacturer Original Design Manufacturer Original Equipment Manufacturer Personal Computer Partido Liberación Nacional Premarket Approval Promotora de Comercio Exterior Proyecto de Fomento al Vínculo de Empresas de Régimen de Zona Franca y Empresas Nacionales Partido Unidad Social Cristiana Science Foundation Ireland Small Latecomer in the Development Process

ACRONYMS AND ABBREVIATIONS

SMEs TNC UATI UNCTAD UNDP U.S.-AID USITC VMI WTO

Small and Medium-sized Enterprises Transnational Corporation Unidad de Asistencia Industrial United Nations Conference on Trade and Development United Nations Development Program United States Agency for International Development United States International Trade Commission Vendor Management Inventory World Trade Organization

xiii

This page intentionally left blank

Acknowledgments

n conducting this research project I have incurred debts to many people who generously gave of their time and shared their thoughts with me. They have contributed greatly to what is good in this book; I take responsibility for the rest. My sincere thanks go to the Instituto de Investigaciones en Ciencias Económicas (IICE) at the University of Costa Rica, for hosting me as a visiting researcher for parts of the summers of 2002 and 2003. I am grateful to Justo Aguilar, Anabelle Ulate, and José Cordero at the IICE for their support and insightful conversation and to Freddy Rojas for sharing information from IICE’s quarterly surveys. I am indebted to PROCOMER and CINDE for their support in arranging interviews. Hugo Chaves at Costa Rica Provee provided an interesting insider’s perspective on promoting indigenous linkage capability. My thanks also go to the Keough Center for Irish Studies, University of Notre Dame, in Dublin, which hosted me during my research in Ireland during parts of the summers of 2002 and 2003. I am indebted to the Center’s director, Kevin Whelan, for his support and willingness to share his deep knowledge of Ireland’s culture and society. Declan Lyons at Enterprise Ireland provided valuable support by arranging interviews and sharing his rich understanding of industrial policy and the electronics sector in Ireland. I owe a great deal to Marcus Breathnach at Forfas for his consistent willingness to provide data and to comment on parts of the writing. The academic colleagues I met with and the political and corporate decision-makers I interviewed are too numerous to list here. I am grateful to all of them for taking time out of their busy schedules to discuss with me aspects of high-tech foreign investment and development from their different vantage points. The many perspectives—from the shop floor to company headquarters to political offices—enriched, questioned, and modified my theoretical understanding of the foreign investment–development nexus. I thank Barbara Schmidt for her wonderful hospitality and insights into life on the Green Isle, and Lowell Gudmundson for sharing his deep

I

xvi

ACKNOWLEDGMENTS

understanding of Costa Rica’s history and politics and for his feedback on parts of the writing. I am indebted to the Whiting Foundation and Mount Holyoke College for financial support at various stages of the research project. I gratefully acknowledge the excellent research assistance of Raluca Cata and Antoniya Ganeva, the unfailing administrative support of Dawn Larder, and the indispensable help in preparing the final manuscript by Jean Costello. This book would not exist without the loving support of my family; Ben’s humor, Nina’s music, and Mike’s intellect made all the difference.

1

Foreign Direct Investment in the Global Age: New Opportunities for Development?

lobalization is at the heart of economic change in the twenty-first century. Its impact on the future well-being of people in developed and developing countries has been the subject of much debate and controversy among scholars, policy-makers, and activists. Some identify globalization as the major source of economic problems, others as the panacea for their solution. Yet others, including this author, see it as a powerful process whose social and economic outcomes are not preordained, but contingent upon policies and preexisting conditions. The increasing fragmentation of the production processes across national borders and the ease with which transnational corporations reorganize their value chains on a global level are the distinctive characteristics of the current globalization process. As TNCs are expanding their global networks, an ever-growing number of developing countries are eager to attract their investments. The zeal with which developing countries are pursuing transnational corporations attests to their high expectations for the contribution Foreign Direct Investment (FDI) can make to development. At the macro level, it is hoped that FDI will increase investment, employment, foreign exchange, and tax revenue. At the micro level, positive spillovers in technology, marketing, and business practices are sought. But the real story is often very different. Some developing countries have been unable to attract any FDI, and for many of those that have, the anticipated benefits have not materialized. Development is fundamentally about structural change, with the core of economic activity shifting from agriculture to industry to services.

G

2

FOREIGN INVESTMENT, DEVELOPMENT, AND GLOBALIZATION

Development is about an expansion of a country’s knowledge-based assets, which constitutes the foundation of these transformations. In this book I investigate—theoretically and empirically—the contingencies that mediate the impact of foreign direct investment on the long-term growth prospects of developing countries in the age of globalization. Concentrating on high-tech FDI and on small latecomers in the development process (SLCs), I focus on the conditions under which high-tech FDI promotes, or fails to promote development of the indigenous knowledge-based assets upon which new opportunities to upgrade industry or even to leapfrog stages in the development process ultimately depend. To emphasize contingent explanations of the actual performance of developing economies is to critique the unequivocally positive assessments of globalization widespread in the economic development literature and in the corridors of political power and policy-making. The unwavering supporters of economic globalization portray it as a kind of magic bullet for underdevelopment, which holds out the promise of higher growth and lower poverty rates. They argue that integration into the world economy through trade and investment will engender greater efficiencies in resource allocation, expand indigenous knowledge through technology transfer, and increase exports to a virtually unlimited international market. But a categorically positive assessment of the development impact of globalization in general and of FDI in particular does not hold up well in light of developing countries’ experiences over the last two decades. The adoption of free market policies aimed at full integration into the world economy has led to disappointing growth outcomes in many Latin American and African countries during the 1990s (e.g. Stallings and Wilson 2000; UNCTAD 2001a). A better understanding of the diversity of possible development outcomes in a global world requires that we explain how the interaction of different conditions and policies can have different consequences for the nature and results of the globalization–development nexus in individual countries. The core argument of this book is that SLCs will reap long-term benefits from high-tech FDI only if TNCs’ strategic global interests and SLCs’ capabilities coincide. Small latecomers will attract high-tech FDI only if the strategic interests of TNCs coincide with SLCs’ location-specific advantages; the positive knowledge spillovers to domestic producers can emerge only if TNCs’ interests coincide with indigenous linkage capability in SLCs; and TNCs will move up the value chain in the host country (i.e. produce more sophisticated products) only if SLCs are able to expand the domestic knowledge base in a way that meets the demands of increasingly sophisticated production. The analytical framework I propose to understand these contingencies focuses on the determining forces behind

FDI IN THE GLOBAL AGE

3

three main categories: TNCs’ strategic global needs, SLCs’ location-specific assets, and SLCs’ linkage capability. TNCs’ strategic needs are shaped by global market dynamics and industry-specific characteristics. TNCs’ interest in host country sourcing is influenced by how their value chains are organized on a global level, which determines who produces where the different parts of the value chain. Host countries’ location-specific assets are the result of their current policies and past development strategies. Path dependency, the cumulative effect of past policy outcomes, matters for the development of locationspecific assets. Finally, SLC producers need to have a minimum threshold of capabilities (technical, risk-taking, financing, information availability, etc.) to have a chance of becoming input suppliers for high-tech TNCs and benefit from positive spillovers which would enhance their technological capabilities. Most indigenous producers started during the import substitution era and have not had to operate in an environment of global competition with its intense price pressures and exacting standards for quality and delivery. The pervasiveness of market failures in developing countries means that most indigenous producers are not in a position to compete with TNCs or supply them with inputs. Widespread market failures mandate proactive government policies aimed at enhancing SLCs’ locationspecific assets, fostering a continuing match between TNCs’ needs and SLCs’ assets, and promoting the expansion of indigenous capabilities for linkages and technology absorption.

Globalization and Small Countries Three major economic forces have been the drivers behind the current process of globalization. The first one is the rapid increase in international trade. World exports and imports as a share of Gross Domestic Product (GDP) increased from 40 to 50 percent during the 1990s. Growing financial flows across borders have been the second driver with daily international transactions now well above $1 trillion. And the third force has been the quantitative increase and qualitative change in foreign direct investment flows. In quantitative terms, global FDI flows tripled and FDI flows to developing countries increased fivefold between 1990 and 2002 (see table 1.1). One of the key factors behind the phenomenal increase in FDI flows to industrialized countries has been the rise in cross-border mergers and acquisitions of large TNCs. Between 1990 and 1995, the value of crossborder mergers and acquisitions in developed countries was equivalent to 75 percent of FDI inflows into those countries; in 2000, it was nearly

Table 1.1

Distribution of Global Foreign Direct Investment Inflows, 1970–2002 4

Millions of current U.S. dollars 1970

1980

1990

2000

2001

2002

World Developed Countries EU Ireland US Developing Countries incl. Central & Eastern Europe Developing Countries Developing Countries w/o China Africa South America Central America and Caribbean Costa Rica Asia Central and Eastern Europe

12,926 9,477 5,127 32 1,260 3,449

54,923 46,530 21,317 286 16,918 8,393

208,501 171,076 96,773 622 48,422 37,426

1,392,957 1,120,528 683,893 26,447 314,007 272,430

823,825 589,379 389,432 15,681 143,978 234,446

651,189 460,334 374,380 19,033 30,030 190,854

3,449 – 928 522 1,063 26 799 –

8,358 8,301 392 3,631 3,854 53 362 35

36,955 33,468 2,429 4,874 4,826 162 24,260 471

246,057 205,285 8,489 57,248 38,110 409 142,091 26,373

209,431 162,585 18,769 39,693 44,033 454 106,778 25,015

162,145 109,445 10,998 25,836 30,183 642 94,989 28,709

World Developed Countries EU Ireland US Developing Countries incl. Central & Eastern Europe Developing Countries Developing Countries w/o China Africa South America Central America and Caribbean Costa Rica Asia Central and Eastern Europe

100.0 73.3 39.7 0.3 9.8 26.7

100.0 84.7 38.8 0.5 30.8 15.3

100.0 82.1 46.4 0.3 23.2 18.0

100.0 80.4 49.1 1.9 22.5 19.6

100.0 71.5 47.3 1.9 17.5 28.5

100.0 70.7 57.5 2.9 4.6 29.3

26.7 – 7.2 4.0 8.2 0.2 6.2 –

15.2 15.1 0.7 6.6 7.0 0.1 0.7 0.1

17.7 16.1 1.2 2.3 2.3 0.1 11.6 0.2

17.7 14.7 0.6 4.1 2.7 0.03 10.2 1.9

25.4 19.7 2.3 4.8 5.3 0.1 13.0 3.0

24.9 16.8 1.7 4.0 4.6 0.1 14.6 4.4

Percentage distribution

Source: Based on UNCTAD Foreign Direct Investment Database. ⬍http://www.unctad.org/Templates/Page.asp?intItemID⫽1923&lang⫽1⬎ accessed May 2004.

FDI IN THE GLOBAL AGE

5

90 percent (UNCTAD 2002). In developing countries, on the other hand, widespread privatization of large publicly-owned enterprises and an open arms policy toward foreign investment were important factors behind the influx of FDI during the 1990s. FDI flows have been highly concentrated, in terms of destinations and TNCs involved. In 2000, the top 10 host countries received 73 percent of all FDI inflows. The largest recipient was the United States with 25 percent. The only latecomers among the top 10 were China, which received 4.1 percent of global FDI flows, and Hong Kong, which got 3.4 percent. And the top 50 TNCs—out of a total of roughly 60,000—accounted for 52 percent of outward FDI stock in the United States, 64 percent in Canada, and 96 percent in Australia (UNCTAD 2002, 52). The qualitative change in foreign direct investment flows is the growing importance of de-verticalized production processes and global production networks under TNC auspices. They constitute the critical aspect of the globalization process that distinguishes the current era of globalization from earlier periods of increased cross-border economic transactions. Increased competitive pressures forced TNCs to intensify their search for cost reductions on a global scale, be it through relocation of parts of the value chain they controlled themselves or through integration of other companies’ production and services into their own global networks. Drastic reductions in the cost of communications and transportation allowed them to view all countries as potential sites of production, while the need for closeness to the customers of consumer goods required them to be present in all major income regions of the world. The confluence of these three tendencies has been driving the growing de-verticalization of the production process of TNCs and the rise of global production networks. Out of the three drivers behind economic globalization, FDI has the greatest potential to bring about significant advances in the host country’s technological capabilities. Since development is essentially about expanding the production of knowledge-based assets within a country, FDI offers new possibilities for developing countries to advance in the industrialization process. In recent years many case studies of developing countries have focused on the Asian Tigers (South Korea, Taiwan, Singapore, and Hong Kong), other large newly industrializing countries like Brazil, Mexico, and Argentina, and more recently on the emerging giants China and India. The literature has paid less attention to small latecomers in the development process, in part undoubtedly because they are not important players on the global stage of power. The development potential of SLCs was severely constrained under import-substituting industrialization (ISI) due to the lack of a sizeable home market. But now that TNCs are expanding global

6

FOREIGN INVESTMENT, DEVELOPMENT, AND GLOBALIZATION

networks, allocating key business functions (design, technology, development, manufacturing, and marketing) to agents in different parts of the world, it is possible for SLCs to attract high-tech foreign investment and become part of TNCs’ global value chains. Through spillovers and linkages with domestic input suppliers, foreign investment has the potential to raise domestic technological capabilities and increase the possibility of sustained industrial upgrading in SLCs. The emphasis on high-tech FDI stems from the fact that high-tech production—which generally does not exist in SLCs—has the greatest potential for generating knowledge spillovers in the host country. The production of high-tech goods spans a gamut of different processes, some of which are highly routinized and standardized. To the extent that SLCs can tap into the global network of TNCs’ high-tech production, they will start with parts and processes that are relatively standardized. The critical questions are whether indigenous SLC producers can indeed become competitive input suppliers and whether, over time, both foreign affiliates and domestic input producers will move up the value chain. The analytical framework developed in chapter two incorporates elements from different bodies of economic thought into a heterodox conceptual framework to analyze the contingencies for FDI-linked industrial upgrading and development. It draws on (1) the literature on the development of indigenous technological capabilities (e.g. Amsden 2003; Amsden and Chu 2003; Lall 2001), especially Amsden’s conceptualization of knowledge-based assets and Lall’s differentiation of technology intensity across products; (2) the mainstream economics literature on foreign direct investment and economic development (e.g. Blomstroem et al. 2001; Caves 1996; Dunning 1998) with its emphasis on TNC core competencies, linkages, and spillovers; (3) the literature on the dynamics of international production networks (e.g. Dicken 1998; Sturgeon and Lester 2003; Yusuf 2003) that analyzes the impact of global competition and technological innovations—in logistics, transport, information technology (IT), and so on—on the ways in which TNCs organize their global value chains; and (4) the literature on cluster formation (e.g. Porter 1990 and 1998), which stresses the need for a simultaneous improvement in key aspects of a country’s location-specific assets to engender cross-fertilization among companies in the development of cluster formations.

Why Ireland and Costa Rica? Given the emphasis on the interaction of contingent factors in the FDI-development nexus, large cross-country panel analyses can offer only

FDI IN THE GLOBAL AGE

7

limited insights. The econometric studies testing for FDI spillovers in host countries are evidence of these difficulties. Supply chain linkages and positive spillovers through education and training are the most important channels through which FDI can broaden the technological capabilities of host country producers. The empirical research, based principally on econometric studies of cross section and panel data sets, provides inconclusive evidence for FDI-generated positive spillovers. The inconclusiveness of the results suggests that the development potential of FDI is contingent upon a set of factors that regression analysis is not able to capture in a systematic way. While a country analysis is of interest in its own right, the ultimate objective of any case study is to explore lessons for other countries, separating the unique from the replicable factors regarding country, time, and policy variables. A comparative study of two countries at different levels of development can provide particularly interesting insights, since it allows us to identify commonalities and differences in how the theoretical contingencies play out under different conditions and in how policies affect development outcomes. The fact that Ireland looked a lot like a developing country not too long ago and then took off to become the Celtic Tiger makes it a good case to study in search of broader lessons.1 Only a few decades ago Ireland shared many characteristics of today’s SLCs: agriculture dominated the livelihood and exports of the country; production in the manufacturing sector was heavily concentrated in basic consumer goods; and poverty was still fairly widespread. A large influx of high-tech FDI in the 1990s was the major driver behind the country’s high growth, transforming it from the poor stepchild of the European Community to its star performer. Yet, not all was well with the Irish model, since linkages with indigenous producers were scant. Only when the Irish government adopted proactive policies in support of indigenous linkage capability did more significant linkages develop. The very success of the Irish economy engendered a dynamic that changed some of the key location-specific assets that had attracted TNCs in the first place, most importantly, the cost of labor. In the context of increasing global competition and growing pressures on production costs, Ireland has ceased to be a competitive site for the production of standardized components and products. A number of TNCs have relocated production from Ireland, primarily to Central European countries and China. As a result, Ireland is now embarking on an R&D-based development strategy, a new phase in which TNCs are expected to play a critical role as well. Costa Rica makes for an interesting case study in juxtaposition to Ireland, because it was such an unlikely candidate to be chosen by Intel

8

FOREIGN INVESTMENT, DEVELOPMENT, AND GLOBALIZATION

as its one investment site in Latin America. In 1996, Intel announced that it was going to set up a microchip factory in San José, Costa Rica, with an initial investment of $300 million. To most people in the development community, this country of three and a half million people had been known for its coffee and banana exports, for its peacefulness amidst the civil wars in the other Central American countries during the 1970s and 1980s, and for its biodiversity and growing ecotourism, but not as a potential site for high-tech direct foreign investments. As in the case of Ireland in 1990, Intel’s investment had a huge signaling effect for other TNCs in high-tech industries, especially in the sector of medical devices. But Costa Rica has not become Ireland. To date, Costa Rica has reaped only limited benefits from high-tech FDI. Part of this failure is due to the downturn in the technology market and part to the fact that it takes time to attract large amounts of FDI. But the limited availability of external and internal funds and the absence of a cohesive government strategy and proactive policies play an important role as well.

Methodology and Findings The empirical analysis of the two country experiences focuses on several overlapping issues: (1) the reasons why high-tech TNCs invested in the two countries; (2) the impact of high-tech FDI on the macroeconomy and structural change in the host economies; and (3) the extent to which indigenous technological capabilities expanded through high-tech FDI. The comparative analysis engages three other issues that cut across the first three: (1) the implications of industry specificity (with the main focus on the electronics industry and medical instruments); (2) the implications of dynamic changes over time (at the country, industry, and the global level); and (3) the role of government policies (in attracting FDI, fostering location-specific assets, and enhancing indigenous linkage capabilities). The comparative case analysis is based on quantitative information and qualitative evidence, secondary data and literature as well as primary data and author interviews. The qualitative evidence, which complements the quantitative data, is used throughout the country-specific chapters to support and illustrate many of the arguments. It stems mainly from in-country interviews with key individuals involved in past and current policymaking on foreign investment and domestic linkage capabilities as well as with private sector representatives from transnational and indigenous companies. In both countries, I conducted extensive interviews with highlevel representatives of TNCs in the computer and medical devices sectors. They provided insights into the dynamics of the industries and the factors

FDI IN THE GLOBAL AGE

9

influencing the decisions of these companies to invest in the host country, to stay there, and, in some cases, to finally leave. Interviews with indigenous producers offered unique insights into the challenges of becoming an input supplier for a high-tech TNC and the possibilities of expanding the market and the product base beyond that initial linkage. I conducted the interviews to get a better understanding of the importance of different factors and forces behind the FDI–development nexus. For insights from statistically representative samples I am relying on the results of existing studies. The comparative country study suggests important lessons for the likely development impact of FDI on today’s developing countries. Two sets of lessons emerge—one regarding SLC-internal factors and policies and the other regarding changes in global dynamics. The first set underscores the pervasiveness of market failures that prevent the automatic development of indigenous linkage capability and cluster formation and call for new industrial policies. It underlines the key role of path dependency, that is the cumulative effect of past policies especially in education and infrastructure, behind the nature of location-specific assets. And it emphasizes the critical importance of access to external and internal funding sources to finance the necessary upgrading of location-specific assets. The second set of lessons stresses the global factors which are making it challenging for developing countries, especially for SLCs, to benefit from high-tech FDI, even if they have the right location-specific assets. Of particular importance are the implications of the intensification of global competition as more countries vie for low- and high-tech FDI, especially with China and India as key players. TNCs will keep reducing the number of global suppliers on which they rely, thus increasing their reliance on large-scale suppliers and suppliers with multidimensional capabilities along the value chain. And as TNCs continue to relocate a large share of commodity production in the high-tech areas to low-wage Asian countries, which have large supplies of unskilled and skilled workers, SLCs need to tap into those parts of high-tech TNC production that are not sensitive to economies of scale, and where quality and intellectual property rights matter. Challenging as these global changes are for SLCs, they have no choice but to embrace them. Increased global competition with intensified pressure on prices and heightened expectations for supplier capabilities underscore the urgency for developing the country’s knowledge-based assets. Such a development provides the only chance for taking the high road to competitiveness, that is through higher productivity and wage growth. The low road of low productivity growth and wage reductions is always the default option.

10

FOREIGN INVESTMENT, DEVELOPMENT, AND GLOBALIZATION

Whether or not an SLC can make high-tech FDI a successful component of its development path under these conditions depends not only on the right conjuncture of factors. Ultimately, a high-tech FDI-linked development strategy also takes a bit of luck and a lot of political will to succeed.

2

High-Tech FDI-Led Growth in Small Latecomers: An Analytical Framework

Industrial Upgrading, Technological Capabilities, and Small Country Size Development is a process of increasing the production of knowledge-based assets within a country, thus shifting production from primary goods to goods of increasingly higher value. Amsden (2001, 3) defines a knowledgebased asset as “a set of skills that allows its owner to produce and distribute a product at above prevailing prices or below market costs.” For a country, the advancement of such skills, managerial and technological, develops over time and cumulatively, but by no means automatically. How to overcome the deficit in skills is the very challenge of development. In the development process, the advancement of knowledge-based assets unfolds first and foremost in the expansion of the manufacturing sector. Increasingly, the set of skills necessary to move up the value chain finds its application in the service area as well, and it becomes harder and sometimes meaningless to separate between manufacturing activities “proper” and service provision. Yet, there is no precedent for a country that has skipped the manufacturing-based production phase and jumped straight from primary goods production to high value-added service provision; the site and starting point for industrial upgrading for latecomers in the development process continues to be the manufacturing sector. For latecomers, the development of technological and managerial skills is first and foremost about the adoption of existing technologies to the specific circumstances of the country. It is only when the country has developed a substantial basis of technological capabilities that innovation— development and diffusion of new technology—becomes the driving force

12

FOREIGN INVESTMENT, DEVELOPMENT, AND GLOBALIZATION

behind ongoing technological change, and thus productivity growth and economic growth. The mastery and adaptation of existing technologies involves a number of different areas, from efficient ways of organizing the production and distribution process, to the optimal choice of inputs and machinery used in the production process, to the introduction of appropriate quality control and product-delivery standards. There is a continuum of complexity for all these processes and requisite skills, which is generally reflected in different kinds of products. Latecomers typically start out at the bottom of this continuum, trying to develop the skills necessary to become competitive in the production of goods with standardized technologies, before being able to face the challenge of producing more technology-intensive products. Moving up the scale of technology-intensive products is fundamental for development. It reflects the expansion of a country’s knowledge-based assets and provides the basis for sustained growth in the future.1

Technology Intensity and Growth There are two widely used typologies for the technology intensity of manufactured products, which produce very similar results. Lall (2000, 342) distinguishes between resource-based products, low-technology products, medium-technology products and high-technology products. Lowtechnology products are simple and labor intensive and based on locally available resources, notwithstanding the fact that some parts can be very capital intensive. Low-technology products embody stable and wellestablished technologies. Medium-technology products use more complex technologies and moderate levels of R&D and need higher skills and longer learning periods. And in high-technology products, technologies change fast, high R&D investments are required, and product design is of utmost importance. A shift in the structure of a country’s manufacturing sector and its manufactured exports from low- to medium- to hightechnology products is thus viewed as an indication that the country is developing and that its technological capabilities have expanded. The Organization for Economic Cooperation and Development (OECD) uses the importance of R&D expenditures as the measure of an industry’s technology intensity. In 1999, for example, R&D expenditures in 12 OECD countries amounted to 2.6 percent of the value of manufacturing output and 7.2 percent of manufacturing value added. But around this aggregate average, the R&D intensity varied considerably, from a high of 10.3 for aircraft and spacecraft to a low of 0.3 for food products, textiles, and footwear (OECD 2003, 156).

HIGH-TECH FDI-LED GROWTH IN SMALL LATECOMERS

13

Porter (1998) argues that there are no low-tech industries, but only low-tech companies. It is certainly true that at any particular point in time a variety of technologies are used for the production of a specific product, ranging from outdated technology to international best practice technology. But we can think of this heterogeneity within industries as the distribution around an average. And this average technology-intensity, as measured by R&D intensity for example, differs across industries. It is likely that the variance of the distribution will decline over time, as competition forces less competitive firms to adopt more up-to-date technology or leave the industry. The importance of knowledge production and learning-by-doing has been identified as the driving force behind economic growth, from Marx in the nineteenth century to Schumpeter in the early twentieth century and Rosenberg in the late twentieth century. But it was only with the rise of the New Growth Theory that these ideas became part of formal economic modeling. As the generation of new ideas and positive externalities of human capital became incorporated into economic growth models, technological change became endogenous, opening up the black box to which it had been relegated in neoclassical growth theory. Since products differ, on average, in the extent of knowledge generation, the impact of their production on the long-term growth potential of a country can be expected to vary with their technology-intensity. Due to the R&D intensive nature of high-tech products, their production is likely to contain a greater element of knowledge production and thus generate higher productivity growth. And since the manufacturing of high-tech goods involves a broader and more sophisticated set of skills, it is also more likely to generate greater positive knowledge spillovers to the rest of the economy. In other words, the very process of developing and manufacturing high-tech goods will generate new ideas that will find application in a variety of other areas and lead to a sustained increase in productivity. As a result, we can infer that, ceteris paribus, the potential for sustained long-run growth is higher for countries with a greater concentration of manufacturing production in high-technology products than for countries, which produce primarily low-tech products. A number of studies provide empirical support for this hypothesis. According to the Global Competitiveness Report 2002, regression analysis shows that ceteris paribus “primary commodity-based economies indeed grew less rapidly in the past decade (and since 1970) than did more technology-based export economies” (World Economic Forum 2002, 46). And the Inter-American Development Bank (IADB 2001, 49) found that only exports of medium and high technology goods, and not of exports per se, generated a statistically significant increase in growth in developing countries.2

14

FOREIGN INVESTMENT, DEVELOPMENT, AND GLOBALIZATION

Productivity growth and positive spillovers are, however, only one reason for the positive correlation between technology intensity of manufactured products and economic growth. Another reason, at least during the last two decades, has been the faster growth in world demand for medium and hightech goods than for primary products and low-tech commodities. A decomposition of world exports by demand and factor intensity between 1985 and 2000 suggests a positive relationship between the growth rate of exports and their technology intensity, notwithstanding the fact that not all subcategories in each group registered the same trend (see table 2.1).3 Primary and resource-based products reduced their share in world exports from 42.7 percent in 1985 to 28 percent in 2000, while the market share of low and medium technology products increased slightly. The share of high technology products, in contrast, increased dramatically from 11.6 percent to 23 percent. Trade developments of OECD countries confirm the superior performance of high-tech industries during the 1990s. High-tech industries grew at an average annual rate of 10.6 percent during the second half of the Nineties, compared to a growth rate of 2.6 percent for medium high-tech industries, 2.0 for medium low-tech industries, and 1.1 percent for low-tech industries Table 2.1

World Trade by Factor Intensity and Demand, 1985 and 2000

Export category by factor intensity

Share of world exports in 1985 (%)

Share of world exports in 2002 (%)

Number of products

24.1 18.5

13.6 14.4

48 62

4 17

44 45

7.73 10.7 14.2 6.6 7.6 28.5

6.6 7.7 15.6 6.9 8.7 29.6

35 27 44* 20 24* 58

9 8 23 12 11 27

26 19 20 8 12 31

9.0 6.8 12.8 11.6 7.9 3.7

9.1 6.7 13.9 23.0 17.7 5.3

5 22 31 18 11 7

3 10 14 14 9 5

2 12 17 4 2 2

Primary products Resource-based products agriculture/forest based other resource-based Low technology manufactures textiles/fashion cluster other low technology Medium technology automotive products process industries engineering industries High technology manufactures electrical products technology products

Rising Declining market market share share

* In “other low technology manufactures,” market share status was incalculable for one commodity group (ISIC 675), which is why the numbers do not add up to 24. Source: Calculated from CEPAL (2002).

HIGH-TECH FDI-LED GROWTH IN SMALL LATECOMERS

15

(OECD 2003, 192). In addition, trade ratios increased much more dramatically for high-tech than for low-tech industries. Between 1991 and 1999, the export ratio of high-technology industries in OECD countries rose from 30 percent to 43 percent, while the import penetration ratio increased from 32 percent to 43 percent. In low-tech industries, in contrast, the export ratio only increased from 12 percent to 15 percent, while the import penetration ratio grew from 15 percent to 19 percent (OECD 2003, 179–180). The production of high-tech goods can normally be expected to have a greater positive impact on the productivity and growth performance of a country than the production of low-tech products. But it is important to keep in mind that the production of high-tech goods spans a gamut of different processes, some of which are very routine and standardized. Understanding this heterogeneity is especially important for latecomers in the development process. To the extent that they produce high-tech products, they will start with those parts and aspects that are more standardized. The key question and development challenge is whether they will get stuck at that level or whether—over time—they move up the value chain. Take the example of microchips, which are generally considered a high-tech product. The high-tech part of the product is the design of the chip, while the assembly of the chips is substantially more standardized and lower-tech. The former happens primarily in the industrialized countries, while the latter takes place mainly in latecomers.4 So, is a country better off, ceteris paribus, if a lot of production is concentrated in the high-tech manufacturing processes of a product that is considered low-tech on average (e.g. sophisticated clothing), or in the lowtech manufacturing processes of a product that is considered high-tech on average (e.g. semi-conductors)? In which of the above two cases is the longterm growth potential higher? How this static scenario unfolds over time is a critical determinant of the answer. If a country gets stuck at the low-end part of a high-tech product, then the long-term growth potential may not be any higher or may even be lower than if it was producing high-end clothing with sophisticated technology. But if the production of the standardized parts of high-tech goods teaches workers skills that will make it possible to move up the value chain of that product, then producing high-tech goods holds out the greater growth potential. It can contribute to the expansion of a country’s knowledge-based assets, even if high-tech production starts at the low end.

The Creation of Knowledge-Based Assets All latecomers face the challenge of how best to create knowledge-based assets and eventually become competitive in the production of high-tech

16

FOREIGN INVESTMENT, DEVELOPMENT, AND GLOBALIZATION

goods. In The Rise of the Rest, Amsden (2003) analyzes the driving forces behind successful industrial upgrading in latecomers in the second half of the twentieth century. She divides countries in the “Rest” into two categories: those that pursued a nationalistic strategy of innovation (China, India, Korea, and Taiwan) and those that followed an FDI-based strategy. She argues that the followers of a nationalistic strategy moved the furthest in building up knowledge-based assets, whereas the countries, which relied on FDI spillovers, became good at producing mid-level technology goods.5 At the beginning of the twenty-first century it has become much more difficult for any latecomer to pursue a nationalistic strategy. Many of the very policies that were an integral part of the nationalistic strategy (e.g. export performance requirements) are no longer sanctioned under World Trade Organization (WTO) rules. But even if WTO regulations were more flexible, countries would find the pursuit of a nationalistic strategy more challenging now due to the high mobility of capital on a global scale with the resulting increase in competitive pressures. If the strategy used by the most successful latecomers of the last 50 years cannot be replicated today, then the question of whether and how latecomers can successfully build knowledge-based assets based on an FDI-based strategy acquires heightened significance. That is especially true for small latecomers. In many ways, SLCs never had the option of using a nationalistic strategy for industrial upgrading. Under certain conditions, globalization and FDI now provide an opening for industrial upgrading that simply was not there before. There is no magic population number that defines a country as small. A rough approximation of the number of small countries with some possibility for attracting hightech FDI might be middle-income countries with a population between 1 and 10.5 million. In 2000, there were 17 lower middle-income countries and 15 upper middle-income countries.6 Under ISI it was difficult for an SLC to develop the indigenous technological capabilities necessary for industrial upgrading because of restricted financial and manpower resources and a market size which limited the level and growth of effective demand for manufacturing output. And since the demand constraint often fell short of the minimum scale necessary for profitable production, manufacturers in small countries had little incentive to invest in technological learning and manufacture commodities with increasingly higher value added. The anti-export bias of ISI policies was not conducive to producers viewing exporting as a desirable option, while the limited home market size did not allow them to go through the necessary learning process to become competitive internationally. As a result, under ISI small countries developed specialized and weak manufacturing structures, producing standardized consumer goods like food products, clothing, and simple metal and electrical goods.

HIGH-TECH FDI-LED GROWTH IN SMALL LATECOMERS

17

In the era of globalization, where anti-export biases and trade barriers have been reduced significantly, and where multinational corporations continuously search the globe for new production sites, many expect small country size to be much less of a barrier to industrial upgrading and development. Bellak and Cantwell (1998, 47), for example, argue that: “if we allow for more openness than just trade, for example integration and foreign direct investment—then the constraints of smallness and the role of size lose their significance further.” And Klein, Aaron, and Hadjimichael (2001, 5) suggest that: “countries can use foreign firms as catalysts that allow them to leapfrog stages in the development of local firms. FDI can thus speed up the structural shift in the economy that allows a country to catch-up with advanced economies.” Although they do not specifically refer to SLCs, the universal claim is clearly meant to apply to small latecomers as well. Based on empirical evidence UNCTAD (2002, 58) argues, “many developing countries . . . have entered areas of technology-intensive activity previously out of their reach by attracting the labor-intensive end of high-technology TNC manufacturing.” There is little question that SLCs would want to use FDI as a catalyst for industrial leapfrogging. But it is a completely different matter whether they will actually be able to do so. That will depend on the right match between MNCs’ strategic needs and SLCs’ location-specific assets as well as on the dynamic development and interaction between the two. In the remainder of this chapter I analyze the main factors that shape these contingencies.

Can SLCs Attract High-Tech FDI? Critical Contingencies A host country needs to have the right location-specific assets to be a serious contestant for FDI (e.g. Caves 1976; Dunning 1977; UNCTAD 2001b). Generally speaking, stable property rights, political stability, peaceful labor relations, cost advantages (wages, taxes, utilities, transportation, grants), and appropriate infrastructure (especially transportation and telecommunications) are critical location-specific assets. Empirical evidence supports the importance of these factors for attracting TNCs. Taylor (2000), for example, found a positive and statistically significant relationship between changes in the assets of U.S. TNC affiliates between 1983 and 1997 and host countries’ openness to FDI.7 In an attempt to assess and compare countries’ ability to attract FDI, UNCTAD (2002) constructed an index of “Inward FDI Potential.” The index is an unweighted average of normalized values of eight variables: the rate of growth of GDP, p.c. GDP, the share of exports in GDP, telephone lines per 1,000 inhabitants, commercial energy use per capita, the share of R&D expenditures in gross national income, the share of tertiary students in the

18

FOREIGN INVESTMENT, DEVELOPMENT, AND GLOBALIZATION

population, and country risk. The index captures key insights that have been found to be empirically robust: FDI tends to follow growth, rather than jumpstart it; most FDI goes to developed countries; and the host country needs to have conditions propitious to investment, particularly regarding infrastructure and stability. In 1998–2000, the United States had the highest index with .666, and Sierra Leone had the lowest index with .078. For SLCs hoping to attract high-tech FDI, three location-specific assets are of particular importance. The first one is proximity to major markets. The limited internal market of an SLC means that TNCs do not invest in the country to service the host market; rather they use the country as a platform from which to supply other markets. In other words, TNCs are not market-seeking, but efficiency-seeking, pursuing lower unit cost. The second locational asset is a sufficient supply of the right human capital at competitive prices. The production of high-tech goods requires a labor force with the right technical skills and a fairly good command of English, even if the particular production process in the SLC is to be at the lower end of technology intensity for the high-tech good. And the third important factor is financial incentives. During the period of import substituting industrialization, many latecomers were highly suspicious of foreign corporations and imposed all kinds of restrictions on foreign investors. But their approach toward TNCs changed dramatically, as they began to pursue market-liberalizing policies in the 1980s. Consequently, many more countries have been vying for FDI, competing fiercely with one another, particularly with financial incentives. In a survey of special incentives for TNCs in 1995, McKinsey (2003, 26) found that 65 percent of the 103 countries surveyed used tax holidays to attract FDI, 61 percent exempted TNCs from import duty, 48 percent offered duty-drawback schemes, 46 percent provided accelerated depreciation, and 25 percent granted investment allowances. From an economic point of view, it only makes sense to offer special financial incentives to foreign investors if the latter generate sufficiently large positive externalities, for example in the area of technology transfer. Based on empirical estimates, a number of authors have argued against the use of financial incentives. Haskel, Pereira, and Slaughter (2002), for example, find that the positive spillovers from TNCs on the productivity of British manufacturing firms was smaller than the value of the grants the government had provided to these foreign companies. Analyzing the impact of TNC investment in four large latecomers (Brazil, China, India, and Mexico) McKinsey (2003) comes to a similar conclusion. Based on detailed case studies of five major economic sectors, the McKinsey Report argues that financial incentives did not create economic value, that financial incentives were not the primary driver behind foreign investment, that, in some

HIGH-TECH FDI-LED GROWTH IN SMALL LATECOMERS

19

cases, financial incentives led to over-investment and a decline in productivity, and that they generated a “race-to-the-bottom” dynamics in a couple of cases. Whether or not high-tech foreign investment has a considerable impact on the industrial advancement of a latecomer depends critically on the extent to which it generates spillovers and linkages. One problem, of course, is that host country governments contemplating financial incentives will find it difficult to know in advance the extent to which such benefits will materialize. The larger problem, however, is that competition for FDI among all countries, including SLCs, makes it impossible for SLCs not to offer financial incentives, if they want to attract high-tech FDI. Financial incentives can consist of cash grants that may be tied to specific performance standards, cost reductions through the provision of requisite infrastructure or reduced utility rates for example, or reduced tax rates on profits. In contrast to the Chinas and Brazils of this world, whose large internal market gives them considerable bargaining power vis-à-vis transnational corporations, SLCs have practically no bargaining power. The absence of bargaining clout puts more pressure on SCLs to participate in a game of financial incentives to sway a TNC’s investment decision in their favor. Financial incentives do not have to be the most critical factor in a TNC’s investment decision in order to play an important role among the various factors a TNC is weighing in its location decisions. But if all other factors are the same or similar among a set of competitor countries, it may well be that the extent of financial incentives will tip the investment decision to one country rather than the other. As capital has become ever more mobile across national borders, the relative power of capital has been increasing, while national governments’ ability to tax capital is being steadily eroded. Analyzing tax returns of U.S. TNC affiliates in the manufacturing sector, Altschuler, Grubert, and Newlon (1998) found that the elasticity of real capital to changes in after-tax returns rose from 1.5 in 1984 to 2.8 in 1992. It is likely that this elasticity has increased further since then, as transnational corporations have become more sensitive to tax rates in different countries. Average corporate tax rates in the 30 richest countries of the world have fallen from 37.5 percent in 1996 to 30.8 percent in 2003 (Swann 2003). Nonetheless, in the early twenty-first century, the range of effective tax rates across countries is still quite large (see table 2.2). Financial incentives pose a clear dilemma for SLCs. The development needs of the countries are normally such that their revenue-strapped governments can ill-afford to use grants to attract foreign investors to their shores. Thus, many have resorted to establishing special zones, where TNCs are exempt from different kinds of taxes and tariffs. Offering low or

20

FOREIGN INVESTMENT, DEVELOPMENT, AND GLOBALIZATION

Table 2.2 Effective Tax Rates for Subsidiaries of U.S. TNCs Abroad, 1999 and 2002 1999 (%) Ireland Bermuda Netherlands United Kingdom Canada Luxembourg Switzerland Japan Mexico Singapore Belgium Hong Kong Germany France Australia Spain Italy China Denmark Malaysia Cayman Islands Portugal South Korea New Zealand All Other Countries All Countries

8 3 10 22 30 2 6 45 31 12 27 9 27 36 24 26 41 21 24 13 3 22 31 37 32 22

2002 (%) 8 2 9 31 26 1 4 39 37 11 12 8 27 34 29 13 41 17 8 18 5 9 28 10 49 20

Source: Sullivan (2004b).

no taxes on profits deprives countries, of course, of the very revenue they need to improve infrastructure, education, and health, that is the very areas that will promote people’s well-being and keep the country competitive internationally. A real solution to this dilemma can ultimately only be provided by international agreements that regulate the use of financial incentives for the attraction of FDI and thus limit the destructive competition among countries. Until such time, however, a host country government needs to weigh the pros and cons of different financial incentives for TNCs given the particular circumstances of the country and the likelihood that the cost of the financial incentives will be outweighed by the long-term development benefits of FDI. And it is upon the host country government to do everything within its power to increase the likelihood that these potential benefits will actually materialize.

HIGH-TECH FDI-LED GROWTH IN SMALL LATECOMERS

21

Financial incentives are but one element in the constellation of factors that determine the location-specific advantages of an SLC. Their relative weight compared to other factors like infrastructure, proximity to markets, highly skilled labor, and so on will vary depending on the TNC’s strategic needs and—like many location-specific factors—on industry specificity. The relative importance of tax incentives varies with product/industry characteristics. Low tax rates play a particularly important role in industries with high profit margins like the pharmaceutical industry. The key contingencies under which an SLC can attract high-tech TNCs are summarized in figure 2.1. Our discussion has brought to the fore a number of important insights. First of all, it has become clear that path dependency is critical. The requisite infrastructure and technical skills of the labor force needed to attract high-tech FDI are the outcome of past development policies. An SLC needs to have already a certain level of development to be considered as a potential site for high-tech investment, which

Structure of global value chain: TNCs’ strategic needs

SLC’s location-specific assets

– Proximity to major markets – Stable property and production relations – property rights – political stability – labor relations – Cost factors – skilled labor with proficient knowledge of English – financial incentives – transportation, telecommunications – Clustering

High-tech FDI in SLC

Figure 2.1 High-Tech FDI in SLCs: Critical Contingencies.

22

FOREIGN INVESTMENT, DEVELOPMENT, AND GLOBALIZATION

means that governments in the past must have invested in raising educational outcomes and building roads and telecommunications. In contrast to financial incentives, which can be implemented fairly quickly, the availability of a skilled labor force is the result of the cumulative effects of past policies in the education and training field. An analogous, and equally important, argument applies to the existence of stable property rights, political stability, and peaceful capital–labor relations. Past investments in education and infrastructure are, of course, dependent both on the government’s ability to raise the requisite funding and the government’s willingness to muster the political will to use available funding for development purposes. Second, SLCs need to be proactive in attracting high-tech TNCs. Proactive policies are required particularly in the field of education and in the area of infrastructure improvement. They are also needed to pursue high-tech foreign investors in a deliberate and selective fashion. The small size of the country will never allow it to produce competitively as wide a range of products as a larger country. A small country needs to specialize much more than a large country. And the host country government can play a key role in helping to shape the niche of specialization within the high-tech field—be it in the area of consumer electronics, semiconductors, or medical instruments. For example, education policies have to focus on matching the anticipated demand of technical and technological skills in a specific cluster of industries. Governments must maximize the benefits of their limited resources and use financial incentives to target those industries and companies that promise to generate the greatest spillovers. If left to its own devices, the market will not generate the best social outcome in the presence of positive externalities. Proactive government policy becomes all the more important, if those externalities are substantial. Many economists have stressed the importance of cluster developments, as the geographical proximity of companies producing similar or complimentary products generates economies of scope and agglomeration (e.g. Krugman and Venables 1995 and 1996; Porter 1998). Incentives give governments some leeway in promoting the development of clusters in areas that hold out high promise of successful niche specialization. If the relevant host country agency follows a targeted strategy in its pursuit of potential foreign investors, it will raise the probability of attracting the first “big fish” among high-tech foreign investors, which will then put the SCL on the radar screen of other potential foreign investors and help set it apart from other countries vying for high-tech foreign investment. TNCs’ limited information about an SLC as a possible investment site will be expanded considerably by the demonstration effect of the investment of one large corporation. In sum, it will be challenging for an SLC to attract high-tech FDI. An SLC needs to be able to offer a set of location-specific assets, whose existence

HIGH-TECH FDI-LED GROWTH IN SMALL LATECOMERS

23

depends partially on having adopted the right policies in the past and on having (had) access to resources. Furthermore, the host government of the SLC has to have strong administrative capabilities to adopt and execute the needed proactive policies to attract FDI and to coordinate policies in the different areas that will allow location-specific assets to develop in the coherent context of an integrated, well-articulated overall development strategy. From High-Tech FDI to Industrial Upgrading: The Challenges When high-tech TNCs invest in an SLC, the investment can have a highly positive impact on the country’s development prospects. But that will not happen automatically. In this section, I analyze under which conditions high-tech FDI can lead to industrial upgrading in the host country, or take place in an enclave-like setting, making little contribution to an expansion of the host country’s knowledge-based assets and thus its long run development. High-tech FDI, like all FDI, can have an important impact at Structure of TNCs’ global value chain

SLC’s linkage capability

– internalized vs externalized production – technological sophistication and preferred supplier arrangements

– technological capabilities – scale considerations – manageable risk – finance

Channels for technology transfer – linkages – training – demonstration effect

Expansion of indigenous knowledge-based assets

Figure 2.2 From High-Tech FDI in SLCs to Industrial Upgrading.

24

FOREIGN INVESTMENT, DEVELOPMENT, AND GLOBALIZATION

the macroeconomic level as well as at the microeconomic level of an SLC. Here we are interested primarily in FDI’s ability to expand the knowledgebased assets of an SLC, and thus in the microeconomic implications of high-tech FDI. Nonetheless, it is important to understand also the potential impact on macro variables because of the potential effect on growth and because of the dynamic interactions between the micro and macro levels in the achievement of development (see figure 2.2).

Impact of High-Tech FDI on Macroeconomic Variables At the macroeconomic level, high-tech FDI can enhance an SLC’s development potential in a number of ways. It can lead to an increase in investment, tax revenue, foreign exchange and employment. The mode of entry of high-tech FDI is a key factor behind a TNC’s capital contribution to the host economy. If a TNC enters through greenfield investment, it adds more to the domestic capital stock than if it enters through the takeover of existing indigenous firms. Because of the absence of any significant indigenous high-tech production in SLCs, most high-tech FDI will initially be greenfield investment and thus lead to an increase in investment in the SLC. Over time, the greenfield nature of high-tech FDI may decline, as high-tech companies—indigenous or foreign—are taken over by other foreign companies. By the same token, the net positive impact of greenfield investment on the SLCs’ overall investment will be reduced over time, to the extent that income generated in the country will leave via profit repatriation.8 The fact that SLCs often use low tax rates as a way to attract foreign investment has two significant implications. First, low tax rates may entice transnational corporations to use the country partly as a tax haven, shifting part of the corporation’s global income from high-tax countries to the low-tax host country via transfer pricing. To the extent that profits are subsequently repatriated, the SLC’s gross national product can be significantly smaller than its gross domestic product. Second, low tax rates or outright financial grants may deprive the SLC of revenue that it needs to upgrade the country’s infrastructure, education, health system, and so on. No revenue is lost, of course, if foreign corporations would not have invested at all in the absence of low taxes. The cost of grants is real in the short run, but whether it is a cost in the long run depends on the benefits generated by the foreign investors, primarily at the micro level. With respect to foreign exchange generation and the balance of payments, high-tech FDI in SLCs will lead to increased capital inflows and exports and thus an improvement in the balance of payments. However, along with

HIGH-TECH FDI-LED GROWTH IN SMALL LATECOMERS

25

the rise in exports comes a rise in imports, as the requisite inputs for high-tech products may be too technologically sophisticated to be available domestically, or if they are available, the domestic supply base may be too limited. And profit repatriation is another possible drain on foreign exchange. The net impact of these four flows on the balance of payments is not obvious a priori and needs to be analyzed in a country-specific context. Finally, high-tech FDI can be expected to have a positive impact on employment, especially if it is greenfield investment. However, the impact will probably be smaller than what would result from increases in indigenous investment, as the empirical evidence suggests that foreign investment is more capital-intensive than indigenous investment.

High-Tech FDI and the Expansion of SLCs’ Knowledge-Based Assets High-tech FDI in a small latecomer will inevitably lead to a change in the structure of the SLC’s manufacturing output and exports. As an indicator of a country’s knowledge-based assets, such structural change is only of limited value though. To gauge whether high-tech FDI is driving the expansion of an SLC’s indigenous knowledge base or contributing to it, we need to investigate whether it generates positive spillovers to indigenous producers. In the absence of spillovers, foreign-owned production will have all the trappings of enclave production, with no benefits at the micro level and limited benefits at the macro level. The appropriate theoretical construct for understanding how FDI can lead to industrial upgrading in an SLC is the input-supplier-oriented model. Most of the mainstream literature focuses on the competitive model, where the presence of foreign corporations forces domestic producers of the same final goods to become more competitive. As indigenous companies reduce X-inefficiencies and incorporate new technology, they raise their productivity and that of the industry as a whole. In the case of high-tech FDI in SLCs, this direct competition effect is not important, since the limited market and the level of the country’s development make it very unlikely that there will be national producers in the same final goods market as the high-tech TNCs. Instead, the question is whether indigenous producers can become competitive suppliers of inputs for high-tech TNCs. Supplier-oriented industrial upgrading will come about, if potential indigenous input suppliers master the challenge of providing internationally competitive inputs and then over time reduce their dependence on their initial TNC clients to supply a broader range of international customers and move into higher

26

FOREIGN INVESTMENT, DEVELOPMENT, AND GLOBALIZATION

value-added activities. A number of analysts attribute the success of the smaller East Asian latecomers to their ability to become competitive input suppliers and move up the value chain over time, e.g. Ernst (2003a and b) and Yusuf (2003). In supplier-oriented upgrading, indigenous firms start by supplying the needs of local affiliates of TNCs; they become original equipment manufacturers (OEMs). As they upgrade their technological capabilities, they eventually become adept and competitive in the design of products as well and become original design manufacturers (ODMs). And finally they may be designing, manufacturing, and selling products under their own brand name and become original brand name manufacturers (OBMs). There are different channels through which high-tech FDI can generate spillovers and broaden the technological capabilities of indigenous producers, thus initiating a process of supplier-oriented upgrading. The most important channel for the transmission of technological knowledge is the supply chain linkage. TNC affiliates may help indigenous input producers to upgrade their technological capabilities, directly through assistance with technology acquisition and sharing of relevant production knowledge or indirectly through the expectation of high quality standards and feedback on technical specifications of suppliers’ output. In addition to the backward linkage effect, there is the human capital effect. TNCs train workers and provide them with new knowledge and technical skills, which they will carry with them when they work for a domestic company or establish their own businesses. And finally there is the demonstration effect, as domestic producers are exposed to TNCs’ products, marketing strategies, and different production processes. In addition to different channels of knowledge transfer, there are also different types of knowledge that TNCs can transfer to indigenous producers. These are process-related know-how, generalized management know-how, or product design know-how. In practice, product design know-how is the least likely knowledge piece to be transferred, since it constitutes part of TNCs’ core competencies. Enhancement of indigenous technological capabilities through backward linkages is key for successful input-supplier-oriented industrial upgrading. Hewitt-Dundas et al. (2002, 6) argue that backward linkages are critical for small countries, because “the more strongly an inward investor is linked through material or service purchases to the local economy, the greater the potential and motivation for knowledge transfers to local firms.” Economists have modeled the benefits of backward linkages in a number of ways. Rodriguez-Clare (1996) proposes an interesting model where TNCs’ increased demand for inputs allows for the production of more specialized inputs, which in turn leads to an overall increase in

HIGH-TECH FDI-LED GROWTH IN SMALL LATECOMERS

27

efficiency. Other models are less applicable in the current context, as they focus primarily on the competition effect in the final goods market.9 FDI can also generate important positive spillover effects via training and education. Slaughter (2002), for example, argues that the long-term effects of high-tech FDI may be particularly important, as a rising wage premium for skilled workers will entice students to seek more education. And higher wages also generate more taxes, which allow the government to spend more on education. Fosfuri, Motta, and Ronde (1999) offer a model for spillovers through worker training. They distinguish between pecuniary spillovers, where TNCs have to pay their workers higher wages to retain them, and technological spillovers, where the worker who was trained by the TNC makes use of the acquired knowledge when working for an indigenous company. Training and work experience with TNCs may also provide the training ground for indigenous entrepreneurs, and, in that way, raise the productivity and technological capabilities of the indigenous sector. For the case of Ghana, for example, Goerg and Strobl (2002b) find that firms, whose owners had worked for a TNC in the same field before, exhibited higher productivity growth than other local firms.10 Empirical studies on spillover effects of FDI generally do not focus on high-tech FDI or on SLCs specifically. They normally do not distinguish between an input supplier model and a competitive model; they do not test for the relative importance of the three different channels for the enhancement of technological capabilities (linkages, training, and demonstration); and they do not investigate whether the impact varies with the mode of FDI entry, that is greenfield versus take-over. Rather, to gauge the spillover effects of FDI, most studies estimate econometrically whether—at the firm or the industry level—productivity is higher in sectors where TNCs are more prominent.11 Direction of causality is a problematic issue in a number of studies, as it is not clear whether TNCs increase the productivity level or growth in the industry or whether they self-select into higher productivity industries. In any case, at the level of generality of the vast majority of the currently available studies, the empirical evidence for productivity spillovers of FDI is mixed. There are at least as many studies finding a negative impact as there are studies reporting a positive impact.12 A statistically positive impact was found for example by Kokko (1996) for Mexico in 1970, but only when enclave sectors are excluded; by Sjoeholm (1999) for Indonesia in 1991, by Haskel et al. (2002) for the United Kingdom, and by Dimelis (2003) for Greece. Others, however, have found a negative impact of FDI on productivity in an industry; for example Aitken and Harrison (1999) for Venezuela, Haddad and Harrison (1993) for Morocco, and Konings (2000) for Bulgaria and Rumania.

28

FOREIGN INVESTMENT, DEVELOPMENT, AND GLOBALIZATION

Some of the reasons for the mixed results will be differences in the definition of key variables, in the specification of the model, and in estimation techniques. It matters whether foreign ownership is defined as foreigners owning 10 percent, 20 percent, or more than 50 percent of a company’s assets. For example, Dimelis and Louri (2002) find a positive spillover effect only from minority-held foreign firms in Greece.13 On the other hand, Blomstroem and Sjoeholm (1999) encounter no difference in the positive spillovers between minority- and majority-owned foreign firms in Indonesian manufacturing in 1991. It also matters which control variables are included in the regression analysis. Dimelis (2003) finds that, in the case of Greece, spillover effects from TNCs on indigenous firms’ productivity growth are positive, but statistically not significant. However, once she controls for firms’ initial productivity level, the coefficient increases in size and becomes statistically highly significant. And, finally, it makes a difference whether regression results are based on cross-section data, which tend to find positive spillovers, or on panel data sets, which tend to generate negative spillover effects.14 While differences in model and variable specifications undoubtedly account for some of the inconclusive empirical results, the different findings also suggest that the studies are not comprehensive enough to tell us what we want to know. Critically for our purposes, they do not tell us about the relative importance of the different channels for technology transfer. Those are normally explored in case studies. For example, based on interviews and questionnaires of 8 TNCs and 16 local subcontracting firms in Singapore’s electronics sector Wong (1992) found that technological know-how was transferred primarily through indirect mechanisms like information disclosure/exposure and learning facilitation rather than through direct intended technology transfer on the part of the TNC. We need to understand, more generally, the factors that determine the extent and nature of technology transfer from high-tech FDI, especially through backward linkages. To that effect I propose an analytical framework that highlights the importance of the confluence of two factors as a necessary condition for positive spillovers. First, the host country needs to have the capability for linkage development. That capability is determined by the existing technological capabilities of indigenous firms, scale requirements, and other factors. And second, the TNCs must have an interest in sourcing in the host country. The extent of such interest is shaped decisively by the way in which TNCs organize the global value chain of their products. Domestic Linkage Capability Analytically, the differences in FDI-related technology spillovers across countries are linked to underlying institutional and economic differences, principally to differences in domestic linkage capability. Domestic linkage capability

HIGH-TECH FDI-LED GROWTH IN SMALL LATECOMERS

29

is particularly important for firms’ ability to absorb technological spillovers from foreign corporations. The model-theoretic literature does not generally consider the domestic capability necessary for backward linkages to develop. For example, the Markusen and Venables model (1999) assumes that domestic input producers increase their output, when TNC demand increases. But what if domestic input producers are not in a position to respond to latent demand from TNCs? The technology gap may be too big for them to meet the TNC affiliates’ demand in terms of cost, product quality, or ontime delivery. If the technological gap between the average host country firm and the average TNC is too large, then indigenous firms may not be able to benefit from new technology and will not increase the supply of inputs. Or indigenous producers may not know about the necessary technology, or they may not have access to credit to upgrade their production facilities, or they may consider the risk too high. In these cases, domestic producers will not become input suppliers for high-tech TNCs, and the latter’s production will not become embedded in the domestic production structure. Given the small size of an SLC, scale factors may also be an important factor hindering linkage development. The quantity of inputs needed may be too small to meet a minimum size of efficiency. Or it may be too large and outstrip local companies’ capacity. In a survey article on the determinants of host country spillovers from FDI, Blomstroem, Kokko, and Globerman (2001) mention the possibility that indigenous firms in developing countries may not be in a position to benefit from latent spillovers, but they do not dwell on the issue. The few econometric studies that address domestic absorptive capacity support the claim that spillovers will only be forthcoming (or that they will be much larger) if the difference between the technology level of foreign- and indigenously-owned firms is small. Haddad and Harrison (1993) found a positive spillover effect on the productivity of domestic firms in Morocco in cases where the technology gap was small. Similar results were found for Uruguay by Kokko et al. (1996), for the United Kingdom by Girma et al. (2001 as cited in Dimelis 2003), and for Greece by Dimelis (2003). On the other hand, Sjoeholm (1999) finds that spillovers in Indonesia were larger in firms where the technology gap was greater suggesting the possibility for leapfrogging. An SLC needs to have a minimum threshold of capabilities (technical, risk-taking, financing, information availability, etc.) to be able to become an input supplier for a high-tech TNC and benefit from positive technological capabilities-enhancing spillovers. If an SLC does not have the linkage capability due to different market failures, then there is a need for government policies to address those failures. The ability to absorb new technology is developed through a process that involves learning by doing. It is not simply about copying a blueprint and

30

FOREIGN INVESTMENT, DEVELOPMENT, AND GLOBALIZATION

following a specific set of instructions, but about developing the understanding and absorption of the tacit, non-embodied knowledge involved. These learning processes have to happen in a context with the right skills, information, financing, and networks, and they will not happen automatically when there are coordination failures, imperfect information, and missing markets. One of the key requirements that a local producer has to meet to become a serious candidate for supplying inputs to a TNC is the ability to provide high-quality components or products. The importance of standardization and quality control has grown hand in hand with the expansion of world trade in components and the globalization of production. A transnational corporation needs to know that a company in China or Costa Rica or Singapore or Ireland is able to produce the same part to the exact specifications provided, over and over again, before even considering whether to have that part sourced from any of these countries. The International Organization for Standardization (ISO), which was founded in 1947, has developed thousands of different standards over the years.15 They provide a reference framework with specifications and criteria for a particular material, process, or product. Though ISO standards are voluntary, they have become increasingly important in recent years, as the globalization of production has intensified. One of the most widely known and used standards is the ISO 9000 family, which lays out the rules for quality management, applicable across products and industries. ISO 9000 is not a quality certification of the product, but it is a certification of consistency in the production process, where the company documents on a daily basis that it followed all the necessary specifications. It takes about nine months for a firm to learn all the steps and to become certified. Increasingly, a TNC will not buy inputs from a company unless the latter can show ISO 9000 certification. Between the beginning of 1993 and the end of 2000, the number of ISO certifications worldwide skyrocketed from 27,816 to 408,631 (see table 2.3). The number of countries with companies that had ISO certification rose from 48 to 158 during that time period. ISO 9000 certification is quickly becoming an absolute necessity for competing as a local and global producer. In sum, the capability to absorb and maximize potential spillovers from high-tech TNCs does not develop automatically. In the presence of imperfect information and creative market failure, governments need to intervene with deliberate and targeted policies to increase the capabilities of domestic firms. That is especially true, if the demand on technological and managerial capabilities is large. And it is especially true in small countries, where the probability for creative market failure is much higher. Ocampo (2002, 18) argues that: “in small economies, a competitive group of small firms cannot be expected to emerge spontaneously because of the lack of

HIGH-TECH FDI-LED GROWTH IN SMALL LATECOMERS

Table 2.3

31

Number of ISO 9000 Certificates, 1993–2000

Africa/W.Asia Share in % No. of countries Latin America Share in % No. of countries North America Share in % No. of countries Far East Share in % No. of countries Australia/New Z. Share in % No. of countries Europe Share in % No. of countries World Share in % No. of countries Ireland Israel Singapore Costa Rica

1/1993

6/1994

12/1995

12/1996

12/1997

12/1998

12/1999

12/2000

951 3.4 7 27 .1 3 1,201 4.3 3 683 2.5 9 1,862 6.7 2 23,092 83.0 24 27,816 100.0 48 100 110 243 0

1,855 2.6 16 475 .7 9 4,915 7.0 3 3,091 4.4 11 4,628 6.6 2 55,400 78.7 34 70,364 100.0 75 1,132 279 662 2

3,378 2.7 27 1,220 1.0 15 10,374 8.2 3 9,240 7.3 13 10,526 8.3 2 92,611 72.7 36 127,349 100.0 96 1,617 526 1,180 2

6,162 3.8 37 1,713 1.0 19 16,980 10.4 3 18,407 11.3 14 9,478 5.8 2 109,961 67.6 38 162,701 100.0 113 2,056 1,709 1,808 n.a.

8,668 3.9 40 2,989 1.3 23 25,144 11.3 3 28,878 13.4 16 12,946 5.6 2 143,674 64.3 42 223,299 100.0 123 2,534 2,303 2,909 7

12,150 4.5 48 5,221 1.9 28 33,550 12.3 3 37,920 14.0 18 16,751 6.2 2 166,255 61.1 42 271,847 100.0 141 2,854 3,700 3,000 12

17,307 5.0 49 8,972 2.6 29 45,166 13.1 3 56,648 16.5 20 25,302 7.4 2 190,248 55.4 47 343,643 100.0 150 3,100 4,600 3,140 33

20,185 4.9 52 10,805 2.6 30 48,296 11.8 3 81,919 20.1 21 27,299 6.7 2 220,127 53.9 50 408,631 100.0 158 3,330 6,140 3,900 79

Source: ISO (2003, 12–15).

adequate externalities and the inherent learning processes involved . . . Policy intervention is required to stimulate strategic change and promote alliances between existing firms that will lead to the formation of dynamic and competitive clusters.” The Organization of Global Value Chains While domestic linkage capability is a necessary condition for the development and sustainability of backward linkages, it is not a sufficient condition. The other critical element is the global sourcing strategy of the TNC, which is shaped by the structure and changes of global value chains of particular products and TNC production strategies. At the beginning of the twenty-first century, we are a far cry from the time when a corporation like Ford Motor Company produced all the parts and inputs for its final product itself. Technological developments during the last 30 years have made it possible for companies to divide the value chain of their products into distinct parts and to produce each part where it is most cost-efficient given the corporation’s overall strategic considerations. More than ever

32

FOREIGN INVESTMENT, DEVELOPMENT, AND GLOBALIZATION

before, TNCs are developing and managing global networks where key business functions (design, technology development, manufacturing, and marketing) are allocated around the globe, and reallocated, if required by the competitive dynamics in the industry. Different parts of the value chain may be produced by TNC affiliates themselves (normally referred to as internalized production) or they may be produced by unaffiliated companies through arms-length contracts (referred to alternatively as externalized production, de-verticalization, or vertical specialization). The extent of internalized versus externalized production varies from TNC to TNC, depending on which activities a TNC considers to be part of its core capabilities. Generally speaking, research and development (R&D), product design, and marketing under the company’s brand name are core competencies of TNCs.Yet, a number of authors have pointed out that there is an increasing trend toward subcontracting not just standardized parts, but also aspects of design, marketing, and overall coordination (Dicken 1998; Sturgeon and Lester 2003; UNCTAD 2002; Yusuf 2003). Increased competitive pressures on TNCs have fostered the growing trend toward de-verticalization beyond the manufacturing process proper, as product life cycles have become shorter, and the complexity and costs of new products have risen. The mirror image of the growing fragmentation of production across national borders has been the increase in international trade of parts and components. Vertical FDI is, by definition, trade creating. As the value chain is spliced on a global scale, intermediate inputs may be crossing borders several times, incorporated in parts with higher value added before reaching the point where they are integrated into the final product. It is difficult to quantify the exact extent of international value chain splicing, since available international trade statistics tell us nothing about who the final importer of a part or component is, and whether the product is part of a global production network. Nonetheless, several scholars have attempted to approximate the phenomenon empirically. Classifying goods by enduse category Feenstra (1998) estimated that trade in parts and components accounts for around 30 percent of trade in manufacturing. Hummels, Ishii, and Yi (2001) found for 10 OECD countries that the share of imported value added in exports had grown by a third between 1970 and 1990, reaching 21 percent in 1990.16 In China, that share rose from 25 percent in 1988 to 41 percent in 1994 (Naughton 1996 cited in Hummels, Ishii, and Yi 2001, 87). As TNCs internalize parts of the value chain across national borders, trade among TNC affiliates constitutes a significant part of global trade. Data for the United States illustrate the importance of intra-TNC trade. In 1998, U.S. multinational parent companies accounted for two thirds of

HIGH-TECH FDI-LED GROWTH IN SMALL LATECOMERS

33

total U.S. exports. Half of the TNCs’ exports went to the TNCs’ foreign affiliates or related companies (Navaretti, Halland, Venables 2002, 12). Developing countries have become increasingly important participants in global production chains. Their share in world exports of parts and components rose from 4 percent in 1981 to 21 percent in 2000. Most of this increase has been concentrated in a small number of countries: China, Mexico, South Korea, Malaysia, and Thailand accounted for 78 percent of developing countries’ exports in parts and components in 2000, while the next five largest latecomers made up around 14 percent (World Bank 2003, 61). TNC affiliates have different strategic options for acquiring needed inputs. Their interest in sourcing in the host country depends on the degree of internalized production, use of global sourcing, and technological requirements. If production is highly internalized, the TNC affiliate will have little interest in sourcing in the SLC host country beyond non-tradable services and very standardized inputs like packaging materials. The degree of internalization is partially influenced by domestic linkage capability. If such capability exists, the likelihood of sourcing may increase. Furthermore, if the production of inputs requires a high degree of technological sophistication, a TNC may opt for inputs from suppliers with whom it has already developed a long-standing relationship and who have a track record of high quality production.17 Kaminski and Smarzynska (2001, 13) argue that as long as locally produced inputs are available at internationally competitive prices, a firm— irrespective of ownership—will buy domestically rather than import. But that is not necessarily the case if strong linkages to preferred global suppliers exist already. To the extent that TNCs have come to rely more on preferred input suppliers, who are able to supply them wherever their affiliates are producing, it will be more unlikely that linkages in the host country will develop. The existence of preferred suppliers does not prohibit the development of local linkages, but it does raise the threshold for potential indigenous input producers to become actual input suppliers. It requires indigenous producers to become more proactive and to have already the level of technological know-how that in the past they might have hoped to acquire from the TNC. Bellak and Cantwell (1998, 54) fear a potential vicious cycle where in light of the growing prevalence of preferred suppliers and strategic alliances “it is difficult for the Latecomer to enter such networks, since they would have to provide beforehand what they actually seek to gain from such a network.” Even if linkages do not develop with indigenous producers, they may develop with foreign-owned input suppliers located in the SLC. If a TNC sources from a preferred supplier who provides it with inputs globally, it does not necessarily mean that the TNC affiliate in the SLC will import the

34

FOREIGN INVESTMENT, DEVELOPMENT, AND GLOBALIZATION

needed inputs. Rather, increasing demands for just-in-time delivery and the desire to have suppliers in close geographical proximity might mitigate against imports and entice foreign suppliers to establish production facilities close to the TNC affiliate in the host country. If a high-tech TNC in an SLC buys inputs from a foreign producer who has colocated with the TNC, then the possibility for linkage development with local producers becomes further removed. Now the foreign supplier becomes the possible source for linkages, with the domestic producer possibly becoming a second or third tier sub-supplier. In that way, the possibilities for knowledge development of indigenous companies become further removed. The use of sub-suppliers on a global level (offshore outsourcing) has become more prominent in a number of industries, across the spectrum of technology intensities. It has been on the rise in labor-intensive industries, like shoes and clothing (e.g. Gereffi 2002), medium-tech industries, like automobile production (Sturgeon 2000), and more high-tech industries like computer electronics (Sturgeon and Lester 2003; Yusuf 2003). Since we are focusing on high-tech industries here, and since foreign investment in the electronics industry has played a critical role in both Ireland and Costa Rica, it is instructive to understand how the structure of global production has evolved in that industry. The Global Value Chain in the Electronics Industry Dicken (1998, 353) calls the electronics industry the “industry of industries,” not because of its absolute size, but because of the application of electronics in a wide and increasing range of products. With its use of offshore assembly, the industry was the first to which the label “global factory” was applied. The industry comprises a broad array of products, from semiconductors (memory chips and microprocessors) to different types of electronic equipment and a vast assortment of consumer electronics. Though the production is generally very capital- and technology-intensive, parts of the production process are much more standardized and labor-intensive, albeit skilled-labor-intensive.18 In the semiconductor industry, for example, the design of the semiconductor, the production of silicon crystal, and the wafer fabrication require large numbers of scientific and engineering personnel, whereas the testing and assembly demands less technical expertise, though by no means unskilled labor. This has led large semiconductor producers (e.g. Intel) to keep design and wafer production in the developed countries and have assembly and testing done in lower-wage locations, which are able to provide the requisite production conditions (e.g. less exacting clean rooms) and offer technically skilled labor at a fraction of the cost in industrialized countries. As a result of the low weight of the wafers,

HIGH-TECH FDI-LED GROWTH IN SMALL LATECOMERS

35

air transport of wafers to assembly and test sites around the globe is not a big cost item in this high-margin business; labor costs are relatively more important. Technological changes, which made it possible to pack an ever larger number of circuits on a single chip, have been driving down production costs in the electronics industries over the last decades contributing to an ever growing demand for electronic products.19 The computer industry provides an excellent example of these developments. During the 1970s, computers were mainly used in the industrial and research context, with large mainframe computers—primarily from IBM—dominating the industry. In the 1980s, the personal computer (PC) took off; prices fell and demand exploded. Between 1985 and 1996, the price of microprocessors fell at an average annual rate of 35 percent per transistor, and for memory chips the respective price decline was 20 percent per year (Kenney and Curry 1999, 11). The development of a mass market in PCs provided the impetus for growth of many complementary products like printers. Subsequently in the 1990s, the expansion of the Internet brought a huge increase in the demand for networking equipment and peripherals. Outsourcing and offshore production became significant during the 1960s and expanded further in the 1970s and 1980s. They were supported and fostered by U.S. tariff provisions (807.00), which required tariff payments only on the value added abroad, not on the value of the original inputs and parts provided by the U.S. TNC, which reimported them in processed form.20 The explosive growth in the demand for electronic products led the large producers to become truly global companies establishing a production presence in all three high-income markets: the United States, Europe, and Japan. Increased competition among TNCs in the 1990s induced a number of brand-name companies like IBM, and the OEMs to move to a global integrator supply chain management system, where the brand name company concentrates on R&D, design, and marketing, and leaves the production to contract manufactures (CMs).21 CMs started out as safety valves for OEMs accommodating their overflow. But as they established their reliability and garnered larger production contracts from OEMs, the latter saw it increasingly to their advantage to outsource the manufacturing of electronic equipment to CMs on a larger scale, as a way of reducing costs as well as risk. Since CMs produce equipment for a number of different clients in the same industry, they are able to leverage larger volumes in negotiations with input suppliers on a global scale and thus achieve lower unit costs. In addition, they can offer OEMs shorter product-time to market. The four largest CMs (Flextronics, Solectron, Sanmina-SCI, and Celestica) are all based in North America, but since the mid-1990s, they have become

36

FOREIGN INVESTMENT, DEVELOPMENT, AND GLOBALIZATION

increasingly more global. On the one hand, they have been colocating with affiliates of OEMs in the major markets, either because the OEMs required them to become their global supplier or because the CM planned to expand its customer base. On the other hand, CMs became global in search of the lowest cost input suppliers. In 2002, for example, Sanmina-SCI had 37 facilities in North America and 63 offshore, and Celestica had 14 plants in North America and 30 offshore (Roberts, 2003a). Contract manufacturers have also moved aggressively into other parts of the value chain, besides manufacturing. They have entered design and engineering on the one end and logistics on the other. And they have entered increasingly other areas besides computers. For example, in 2002, the product mix of Sanmina-SCI, the third largest contract manufacturer, included computers (47 percent), communications (37 percent), industrial products (1 percent), medical devices (4 percent), military (2 percent), and consumer products (9 percent) (Roberts 2003a). In 2002, the top 4 contract manufacturers combined had 254,000 employees and revenues of $42.4 billion (see table 2.4). Their growth has been nothing short of spectacular. Solectron, for example, had revenues of only $300 million in 1989, and Flextronics, the largest contract manufacturer in the world in 2002, had revenues of only $100 million as recently as in 1993 (Roberts 2003b). CMs have grown very rapidly through aggressive acquisitions of suppliers, OEM facilities, and competitors. Solectron and Flextronics made 75 acquisitions between 1997 and 2002 (Roberts, 2003b). Roberts (2003b) observes that: “27 of Flextronics’ acquisitions were other CMs or OEM plants. The rest were a mix of logistics providers, materials suppliers, software developers, and design firms. 24 of Solectron’s acquisitions were OEMs or CMs.” In spite of their growth, we have to remember that CMs still only account for a minority share of electronic output, though their share is Table 2.4

The Top Four Contract Manufacturers

Flextronics (managed from San Jose, but incorporated in Singapore) Solectron (Milpitas, CA) Celestica (Toronto, Canada) Sanmina-SCI (San Jose, CA)

2002 sales (billions)

2002 (employees)

$13.6

95,000

Ericsson, Dell, Microsoft, Alcatel, Siemens, HP

$12.3

73,000

$8.3

40,000

$8.2

46,030

HP, Nortel, Cisco, Apple, IBM, Ericsson IBM, Lucent, Sun Microsystems IBM, HP, Cisco, Dell, Nokia

Source: Roberts (2003a) and ⬍www.hoovers.com⬎

Customers include

HIGH-TECH FDI-LED GROWTH IN SMALL LATECOMERS

37

clearly on the rise. Hyot and Lee (2001, 2) estimated that CMs accounted for 13 percent of the total cost of goods sold by OEM companies in electronics in 2000, and Lyons (2002a) projected that share to increase to about one third by 2005. Not all TNCs rely on contract manufacturers to the same extent. In the telecom industry, for example, Nokia produces all its mobile phones in-house, whereas Ericsson has outsourced all production (UNCTAD 2002, 141). Either choice, however, does not bode well for industrial upgrading in SLCs. If production is internalized, the TNC affiliate will have no interest in sourcing inputs locally. And if the production of standardized components is outsourced, the limited capacity in an SLC will make local sourcing very difficult, if not impossible. As product life cycles in the computer industry have become shorter and shorter, the efficient management of space and time along the value chain has become more and more important. In the PC industry, product life cycles were about one year in the mid-1980s, but only three months by the late 1990s (Kenney and Curry 1999, 8). As a result, suppliers have been increasingly under pressure to maximize logistics efficiency, to guarantee shorter and shorter delivery times. The intensity of competition has led OEMs to become ever more demanding of input suppliers, whether of CMs or others. And the CMs, in turn, ask increasingly more of their own input suppliers, including the ability to provide a multiplicity of tasks rather than just one specific element (e.g. Sturgeon and Lester 2003; Yusuf 2003). Consequently, it has become harder for indigenous producers to participate in global value chains, as the threshold requirements for participation have increased considerably.

Beyond the Existence of Linkages If the right constellation of factors does indeed lead to the development of backward linkages, then the next question is what kind of technology transfer is likely to materialize as a result. As discussed earlier, TNCs may provide direct technical assistance or indirect help by demanding stringent quality control or by providing information to local producers on an ongoing basis. Wong (1992, 34) hypothesizes that the potential for technological development of indigenous producers through subcontracting relationships is highest when the TNC is interested in long-term relationships, when the subcontracted products are specialized and not standardized, and when TNC in-house know-how of supplier technology is high. While supply chain linkages are the most effective channel for technology transfer from TNCs to indigenous input suppliers, there are two other ways in which the indigenous knowledge base can be broadened: through the

38

FOREIGN INVESTMENT, DEVELOPMENT, AND GLOBALIZATION

human capital effect and through the demonstration effect. When hightech TNCs establish production facilities in SLCs, it will normally involve the low-end part of the value chain. Nonetheless, technology spillovers can occur in a number of ways. More highly trained workers can acquire important information and training regarding the production process, the marketing of the product, and the type of inputs needed. And they may decide to strike out on their own and produce inputs or supply services knowing that they will have a first ready-made customer for their output. For other workers, the change may not be as dramatic. Through learningby-doing and on-the-job training they can upgrade their technical skills, which will make them more marketable for jobs requiring greater knowhow and may also induce the TNC affiliate to move the production in the SLC further up the value chain. Finally, actual or potential domestic input suppliers can acquire new information about marketing, technology, or production processes through the demonstration effect of TNCs. It takes time for high-tech FDI presence to translate into an expansion of indigenous knowledge-based assets. Whatever the channel of technology transfer is, in each and every case it takes time for the process of skill and knowledge expansion to get going and to develop. Along similar lines, Porter (1998, 85) argues that the available evidence suggests that it takes a decade or longer for clusters to develop and generate the desired benefits. An obvious implication of the time dimension is that the possibility for spillovers is smaller, the more footloose the TNCs are, that is the more readily they are able and willing to shift production to another country. It is well documented in the literature that TNCs tend to keep the more R&D intensive parts of the value chain close to headquarters in the home country, and that the more routinized parts of the production process are located in latecomers. A number of authors have argued that such a global allocation of the different parts of the value chain offers only limited potential for spillovers through training in latecomers. To analyze this issue in greater detail we need to move away from a static context and explore what will happen to the technology intensity of TNC production in SLCs over time. What is the likelihood that TNCs will change the allocation of parts of the value chain across borders and move from producing the low end part of their high-tech products in SLCs to producing more technology and research intensive parts there? The answer to this question depends critically on the development of the competitive environment in the host SLC versus the other countries in the world. If relative production costs change and make production of standardized parts in other countries more attractive than in the host SLC, then TNCs will relocate production away from the SLC. During the last

HIGH-TECH FDI-LED GROWTH IN SMALL LATECOMERS

39

ten years, the competitive environment for standardized products has changed considerably making it increasingly more challenging for SLCs to compete in the production of standardized products. The appearance of India and China as powerful players in the global market has opened up vast new possibilities for TNCs for cheap production of standardized goods in countries with a large skilled labor force.22 Furthermore, the Central and Eastern European economies, which joined the European Union (EU) in May of 2004, offer possibilities for lower cost production in the context of the largest economic union in the world. But when TNCs relocate standardized parts to new and more competitive areas of the globe, that does not necessarily mean that they will leave the SLC altogether. The factors that may keep them there and induce them to move up the value chain are many of the same factors we discussed earlier in the context of attracting high-tech FDI in the first place: for example the “right” human capital, this time at an even more technically skilled level, financial incentives, proximity to major markets, and so on. What is different from the initial investment of the high-tech TNC is that the transnational has accumulated experience in the SLC. It knows the advantages of producing there, it knows the institutional context, and it has experienced the stability of property and production relations. How likely is it that the TNCs will move up the value chain rather than relocate altogether? And are TNCs more or less likely than indigenous companies to produce more technologically advanced parts or products in the SLC? That is the question I turn to next.

Does Nationality of Ownership Matter? Whether the nationality of ownership matters or not is fundamentally a question about possible differences between indigenous producers and TNCs in the likelihood that they will move up the technology value chain within the SLCs. Let us explore different dimensions of the answer by focusing on two separable issues: the likelihood that a high-tech TNC might produce the really high-tech elements of a product in an SLC, and the probability that an indigenous company that has become a successful competitor in high-tech products will stay in the home country rather than relocate altogether. In The Work of Nations, Robert Reich (1991) argues that the nationality of companies no longer matters, that there is no reason to privilege domestic over foreign capitalists, since both are equally footloose, with no particular loyalty to any particular nation. Thus, he advocates, it is important for government policies to focus on promoting skills and providing an environment for the expansion of technological capabilities to keep and attract

40

FOREIGN INVESTMENT, DEVELOPMENT, AND GLOBALIZATION

companies in the production of high-tech products, irrespective of the nationality of their ownership. Reich’s analysis suggests that if a government pursues educational, training, and technological infrastructure policies that make it possible for companies to produce the more high-tech elements of high-tech products, it will raise the possibility that TNCs will move up the value chain within that country. Historically, TNCs have tended to keep the ultimate high-tech activities—R&D and product design—in their home country. Over the last two decades or so, the trend has started to change though. According to the U.S. Bureau of Economic Analysis (BEA), foreign affiliates of U.S. TNCs increased their share in these corporations’ worldwide R&D from 6.4 percent in 1982 to 11.5 percent in 1994 (cited in World Bank 2003, 63). This new trend holds out some hope for the possibility that TNC affiliates will upgrade their production toward the high-end of the spectrum of technology intensity, if the right human capital and technological infrastructural environment is present. But it is not clear to what extent such upgrading will happen in SLCs rather than in industrialized countries and larger latecomers. There is a noticeable absence of theoretical and empirical inquiry into this issue. One of the interesting questions to be investigated is whether the TNC decision on where to conduct more R&D-intensive activities is influenced by the nationality of the top decision-makers of a TNC affiliate in an SLC. There are two reasons why an indigenous, rather than foreign, CEO of a TNC affiliate may be more likely to have an interest in pushing production up the technology value chain. First, she/he may be less footloose than the capital she oversees and wants to keep her job in her own country. And secondly, commitment to fostering the development of one’s own country may motivate an indigenous CEO of a foreign TNC affiliate to expand indigenous knowledge-based assets. Underlying Reich’s argument is the assumption that companies, irrespective of nationality of ownership, have similar characteristics regarding technological capabilities and scale. While these assumptions may hold true in the context of industrialized countries, they are not applicable for small latecomers and for most of the larger ones as well. There are critical differences in technological capabilities and in scale between foreign and indigenous firms.23 The dearth of indigenous knowledge-based assets puts TNCs and indigenous companies on a very different footing. This gap is in many ways the very essence of underdevelopment. And TNCs, especially in the high-tech sector, are supposed to play a critical role in closing that gap. Let us assume that technological upgrading via TNCinduced linkages and spillovers actually happens, that indigenous producers in the SLC become successful input suppliers to high-tech TNCs, and,

HIGH-TECH FDI-LED GROWTH IN SMALL LATECOMERS

41

if they follow the supplier-oriented upgrading model, move toward broadening their clientele and serve the international market, eventually becoming TNCs themselves. Will they be any more likely to care about the expansion of domestic knowledge-based assets in the SLC than current foreign-owned TNCs in the country? And if foreign high-tech TNC affiliates were more likely to upgrade their production in the SLC than any indigenous producer could do in the medium run future, would we then still care as much about indigenous companies moving from OEM to ODM to OBM? There are two main reasons why indigenous companies in SLCs are likely to behave differently from TNC affiliates and will be less likely to leave completely. First, if companies tend to keep their R&D in the home country, then it is more likely that indigenous companies will keep their research base in the SLC compared to foreign companies bringing part of their research base to the SLC. Amsden and Chu (2003, 3) claim that to compete in hightech sectors, companies in latecomers need to develop assets, which relate to project execution, product engineering, and R&D. They argue that national entities have a stronger interest than TNCs “to invest in the specific assets that are required to compete at this stage of development” because their opportunity costs are lower. The argument that indigenous companies have a keener interest in expanding the domestic knowledge base and technological capabilities loses some of its power, to the extent that TNC affiliates are interested in moving their production in the SLC up the value chain. The second reason why SLC companies will be less footloose overall is that—on average—the scale of their operations is smaller than that of TNCs. As long as they are still fairly small—even as they are growing— they simply do not have the same global network and reserves of skills to be able to relocate production with the same ease. There is a dearth of empirical studies investigating the link between nationality and plant closure, and the evidence of the few existing studies is inconclusive. In a comparison of indigenous Indonesian manufacturing companies and companies with foreign participation, Bernhard and Sjoeholm (2003) find that between 1975 and 1989, companies with foreign participation were 20 percent more likely to close in any given year than indigenous companies, once you control for plant size and productivity. On the other hand, enterprise survival rates in Ireland show that foreign companies, which had started operations between 1980 and 1994, were more likely than Irish-owned firms to be still there in 2000.24 In sum, there are good reasons to believe that the nationality of ownership matters in the dynamic industrial development of SLCs. Because of scale factors and the tendency for companies to keep R&D operations in their home country, indigenous companies in the high-tech sector will

42

FOREIGN INVESTMENT, DEVELOPMENT, AND GLOBALIZATION

likely keep key parts of their operation in the SLC. While TNCs may be more willing to relocate some parts of R&D to affiliates in other countries, it is not clear to what extent SLCs are a viable option for that, not least again because of the limitations of scale. The right educational and technological infrastructure is clearly important. Whether the nationality of the top decision-makers of a TNC affiliate has any bearing on the issue is a question that requires empirical investigation.

Focal Points in the Comparative Analysis of Ireland and Costa Rica The analytical framework developed in this chapter has laid out key contingencies under which high-tech FDI may lead to industrial upgrading in small latecomers. The next chapters analyze how these contingencies are articulated in the concrete cases of Ireland and Costa Rica. The analysis in the country chapters will focuses on three key overlapping sets of questions. The first set of questions aims to draw out the reasons why high-tech TNCs invested in the SLC. What were their strategic needs? What were the main country-specific assets that met those needs? What was the relative weight of the financial incentives offered by the host country? How important was path dependency in generating the country-specific assets? How did the country generate the resources necessary to build up and expand its location-specific assets? The second set of questions focuses on the implications of FDI at the macro level and for structural change. What was the impact of FDI on investment ratios, on economic growth and on employment growth? How did FDI affect the generation of fiscal revenue? The third set of issues addresses the expansion of technological know-how through positive spillovers from high-tech FDI to local input producers. This is the most important part of the analysis for our purposes, since the generation of knowledge-based assets lies at the very heart of the success of an FDI-led development strategy. Did any spillovers occur? What were the main channels, backward linkages, or training effects? Did spillovers occur primarily directly or indirectly? In addition, the analysis focuses on several issues, which cut across these three sets of questions. The first and most important one is the role and nature of government policies in establishing the country’s locationspecific assets, in attracting and keeping high-tech FDI, and in developing the domestic linkage capability. Key factors are the availability of resources, the existence of effective government institutions, and the coherence and coordination of government policies. The second issue is the question of

HIGH-TECH FDI-LED GROWTH IN SMALL LATECOMERS

43

industry specificity. The study concentrates on the electronics industry and the medical-instruments industry to explore the existence and implications of industry specificity in the FDI-industrialization nexus. Both industries are key high-tech industries, comprising a range of products that vary in technology intensity. One of the main reasons for choosing these two cases is that both Ireland and Costa Rica have attracted TNCs in these industries; in a number of instances the same TNCs. The pharmaceutical sector is not discussed in as much detail, since it does not play any role in Costa Rica, though it is an important sector in Ireland. Another important question investigated in the country chapters is how the amount and nature of high-tech FDI in the host country changed in response to the dynamics generated by TNCs in the SLC as well as by changes in the global competitive environment. In sum, commonalities and differences in the Irish and Costa Rican cases are analyzed along three axes: the country level, the industry level, and the time dimension. Costa Rica and Ireland tapped into high-tech FDI and the possibility of industrial leapfrogging at different stages in their respective development process and at different stages in the globalization process. An analysis of the complex web of interactions among these levels will be particularly illuminating for understanding the extent and possibilities of future high-tech FDI-driven industrial upgrading, in these two countries and other SLCs.

This page intentionally left blank

3

The Rise of the Celtic Tiger

uring the 1960s, real GDP per capita in Ireland increased by 38 percent; during the Seventies it rose by 35 percent; in the course of the Eighties it grew by 28 percent, but in the Nineties it skyrocketed by 90 percent, from $15,084 in 1990 to $25,622 in 1999. The high economic growth rates in the 1990s—which earned Ireland the title Celtic Tiger—were driven by large inflows of foreign investment into the electronics, software, medical appliances, and pharmaceutical sectors.1 By the early twenty-first century, however, the Celtic Tiger started to linger. Growth rates tapered off to 3 to 4 percent, some TNCs in the electronics industry started to leave the isle, and unemployment was on the rise again. In this chapter I argue that the impressive growth rates of the 1990s were the result of the dynamic interaction of Ireland’s location-specific assets, TNCs’ strategic needs, and the nature of the global competitive environment. Changes in these three key factors have led to a new critical juncture for Ireland in the late 1990s and have given rise to a whole new set of development challenges facing Ireland today. Whether this new juncture spells the end of the Celtic Tiger depends critically on the success of the country’s newly adopted strategy of R&D-led growth. The extent to which FDI has led to the advancement of indigenous knowledge-based assets and the ability and willingness of high-tech TNCs to move their production in Ireland up the value chain are two critical determinants of the success of the new R&D-based strategy. They will be analyzed in chapter 4.

D

Background on Ireland’s Development Policies Any analysis of the Irish FDI-led growth experience in the 1990s has to start with a brief synopsis of the country’s development history. Grounding in the proper historical context is particularly important, since FDI promotion started in the late 1950s. Path dependency was critical for Ireland, as

46

FOREIGN INVESTMENT, DEVELOPMENT, AND GLOBALIZATION

a number of the country’s location-specific assets in the early 1990s resulted from the cumulative effects of deliberate and fortuitous policies adopted during earlier decades. After the establishment of the Free Irish State in 1922, the Irish government pursued a free-trade strategy with emphasis on agriculture, followed by an import-substitution strategy in the early 1930s, with high protection and limitations on foreign ownership of domestic productive assets.2 Average tariff rates rose from zero percent in 1924 to 9 percent in 1931 to 45 percent in 1936 (Whitaker 1983, 62). In addition, the Control of Manufactures Act of 1932 restricted majority ownership of production facilities to Irish citizens. By the 1950s ISI had run into severe difficulties. Balance of payments problems were becoming chronic, and the economy was stagnating.3 GDP grew at an annual rate of less than 1 percent between 1951 and 1958 (Mac Sharry and White 2000, 21), and over 400,000 people left the country during that period, a reflection of their difficulty to find gainful employment in the industrial sector. Based on a report on the state of economic development in Ireland by the then Secretary of Finance T.K.Whitaker, the Irish government opted for a new development strategy in 1958, which emphasized a reduction in tariff protection, the promotion of manufactured exports, and the attraction of FDI. The strategy contained many of the same elements that would be advocated for developing countries 30 years later by the World Bank and the International Monetary Fund (IMF). But it differed from the Washington Consensus in its emphasis on financial incentives to stimulate trade and FDI and on proactive government policies to attract FDI. Trade liberalization also unfolded more gradually than it would for many developing countries during the 1980s and 1990s. From the beginning, the Irish strategy of integration into the world economy combined an open-economy approach with selective interventionist government policies. The Irish government offered a 100 percent remission of taxes on export profits in order to stimulate exports of manufactured goods.4 Even though this tax incentive was available to national and foreign producers alike, it was predominantly TNCs that benefited from it. Numerous government institutions were put in charge of implementing the new strategy, most importantly among them the Irish Export Board for the promotion of exports and the Industrial Development Authority (IDA). Originally, the IDA was charged with the promotion of industry regardless of ownership. Nonetheless, it spent most of its energy and resources on the attraction of FDI, and in 1993, its charge was officially limited to FDI.5 Tariff liberalization started with unilateral tariff reductions in 1963 and 1964, followed by the Anglo-Irish Free Trade Agreement in 1965 and Ireland’s subscription to the General Agreement on Tariffs and Trade in 1967.

THE RISE OF THE CELTIC TIGER

47

300,000 250,000 200,000 150,000 100,000 50,000

19 72 19 74 19 76 19 78 19 80 19 82 19 84 19 86 19 88 19 90 19 92 19 94 19 96 19 98 20 00

0

Foreign

Irish

All Ownership

Graph 3.1 Employment in Foreign-owned and Irish-owned Manufacturing Enterprises, 1972–2001. Source: Based on Annual Employment Survey of Irish Manufacturing Firms and International Services, data facilitated by Breathnach, Forfas.

Trade liberalization reached a climax in 1973, with Ireland’s accession to the European Economic Community (EEC). Dramatic tariff reductions led to a profound restructuring of the manufacturing sector, as many national producers found themselves unable to compete with the onslaught of imports.6 As a result, national manufacturing production declined considerably. The drop in output was exacerbated by the decline in overall economic growth in the early 1980s, when Ireland finally had to face up to its foreign debt problem. In response to the oil shock of 1973, the Irish government, not unlike governments in many developing countries, had opted for accommodating the increased oil prices rather than for restructuring, thus accumulating foreign debt and raising inflation.7 Between 1973 and 1986, 44 percent of Irish manufacturing firms closed, including three-quarters of the clothing firms, two-thirds of the textile firms, and half of the metal/engineering firms (O’Hearn 1998, 42). Employment in Irish-owned manufacturing enterprises was more or less stagnant during the 1970s, declined during the 1980s, and did not start to recover until the early 1990s (see graph 3.1). While EEC membership led to a substantial contraction of the indigenously-owned manufacturing sector, it also brought considerable benefits. First, the Common Agricultural Policy of the EEC with its price floor for agricultural products led to a substantial increase in rural incomes. Second, EEC membership offered Ireland duty-free access to a huge market. That was an important factor in TNCs’ decisions to use Ireland as an export

48

FOREIGN INVESTMENT, DEVELOPMENT, AND GLOBALIZATION

platform from which to supply the continental EEC market. In the 1970s and early 1980s, a number of important electronics companies invested in Ireland; e.g. Mostek (microchips), Wang Laboratories and Verbatim (wordprocessing programs) and Apple (personal computers). And major medical companies invested in Ireland as well such as Abbot, Baxter, Bausch and Table 3.1

Basic Economic Data on Ireland, 1960–2002

Foreign direct investment (net inflows) (millions of current U.S.$)

Net FDI inflows as a % of GDP

Net FDI inflows as a % of GCF

Gross capital formation as % of GDP

Annual growth rate of real GDP (based on 1995 US$)

Real GDP per capita (based on Unemployment 1995 US$) rate

1960s average

–

–

–

–

4.4

6,479

–

1970s average

205

1.7

6.4

25.9

4.7

9,204

–

286 203 242 170 121 164 ⫺40 89 92 85

1.4 1.0 1.2 0.9 0.6 0.8 ⫺0.2 0.3 0.3 0.2

5.2 3.8 4.4 3.7 2.9 4.2 ⫺0.8 1.7 1.6 1.3

26.7 27.2 26.6 22.8 21.8 19.1 18.0 16.4 15.9 18.0

3.1 3.3 2.3 -0.2 4.4 3.1 ⫺0.4 4.7 5.2 5.8

10,895 11,120 11,253 11,148 11,551 11,870 11,816 12,346 13,051 13,891

– 10.5 12.1 14.0 15.6 16.7 17.2 17.0 16.4 15.2

141

0.7

2.8

21.3

3.1

11,894

15.0

627 1,357 1,442 1,121 838 1,447 2,618 2,743 11,035 18,615

1.3 2.8 2.7 2.2 1.5 2.2 3.6 3.4 12.7 19.5

6.3 14.8 16.5 14.8 9.5 11.9 18.0 15.6 53.1 80.4

21.0 19.2 16.3 15.1 16.1 18.4 19.9 22.0 23.9 24.2

8.5 1.9 3.3 2.7 5.8 9.9 8.1 10.9 8.8 11.1

15,084 15,288 15,697 16,056 16,942 18,453 19,778 21,724 23,364 25,622

13.0 14.8 15.2 15.8 14.8 12.2 12.0 10.3 7.8 5.7

4,184

5.2

24.1

19.6

6.9

18,801

12.2

22,778 9,578 24,697 25,238

23.9 9.6 20.3 17.0

97.5 39.7 – –

23.9 23.6 – –

10.0 5.7 6.9 1.8

28,981 30,551 30,888 –

4.3 3.7 4.2 –

1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 1980s average 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 1990s average 2000 2001 2002 2003

Source: Based on World Bank World Development Indicators ⬍http://devdata.worldbank.org/dataonline/⬎ accessed October 2004.

THE RISE OF THE CELTIC TIGER

49

Lomb, and IMED (see Mac Sharry and White 2000, 201–203). As a result of the inflow of FDI into manufacturing, TNCs increased their share in manufacturing employment from 26.9 percent in 1973 to 34.6 percent in 1980.8 During the 1970s, real GDP grew at an average annual rate of 4.7 percent (see table 3.1), but as a result of the 1979 oil-price shock and rising fiscal imbalances in the first half of the 1980s, the Irish economy experienced prolonged slow growth. Employment in foreign-owned manufacturing enterprises stagnated; employment in Irish-owned firms declined precipitously; unemployment reached 17 percent in 1986 and net inflow of FDI was negative that year.9 Only after macroeconomic stability was reestablished and a new social contract of national pay agreements was signed in 1987 did growth resume in the latter part of the 1980s. FDI and Economic Performance in the 1990s: Impact at the Macro Level Starting in 1990, net inflows of FDI took off reaching unprecedented levels for Ireland by the end of the decade. While the average annual FDI inflow was $141 million during the 1980s, it rose to $4.2 billion during the 1990s (see table 3.1). Intel’s arrival in Ireland in 1990 marked the beginning of the surge of FDI inflows. In the manufacturing sector, FDI went primarily to the IT, pharmaceutical and medical instruments sectors. And in the services sectors, it went to the establishment of call centers and in the late 1990s increasingly to the financial offshore center, International Financial Services Center (IFSC) in Dublin, which had been founded in 1987. In 2000, the stock of FDI in Ireland was estimated at $59.4 billion (Forfas 2002, 3). U.S. TNCs have been the dominant foreign investors, with U.S. affiliates accounting for about three-quarters of Ireland’s industrial exports in 1999.10 In the same year, U.S. firms employed 29 percent of the workforce in manufacturing (NACE 15–27), but 66 percent in the electronics industries (NACE 30–33) and 51 percent in chemicals (NACE 24).11 Since I am interested in the impact of FDI on the productive structure of the economy and the advancement of knowledge-based assets in Ireland, I focus primarily on foreign investment in the manufacturing sector. The vast FDI inflows of the late 1990s clearly overstate their potential impact on the productive structure of the country, as large amounts of investment went to the IFSC. Barry, Goerg, and McDowell (2001, 3) point out that FDI inflows to the IFSC basically constitute a transfer of funds by TNCs to their subsidiaries in the IFSC, which then reinvest the funds in portfolio investment overseas. Expansion of the IFSC and of call centers did have a significant impact on the labor market though, through an increase in the demand for labor. In 2001, around 500 Irish and foreign

50

FOREIGN INVESTMENT, DEVELOPMENT, AND GLOBALIZATION

companies were operating under the IFSC. They provided a range of financial services, employed 8,500 people and managed funds of over $150 billion (Forfas, 2002, 25). In addition, a significant amount of FDI went to the establishment of call centers and the centralization of TNCs’ back office services in Ireland for the regional or global level. Both service areas have created significant employment, but have not contributed much to the expansion of knowledge-based assets. The share of foreign direct investment going to the IFSC fluctuated between 44 percent and 79 percent over the period 1998–2002.12 Nonetheless, non-IFSC flows were still huge by historical standards: $3.6 billion in 1998, $6 billion in 1999, $8.8 billion in 2000, $1.3 billion in 2001, and $12.5 billion in 2002.13 UNCTAD’s Inward FDI performance index provides another good indication of the dramatic increase in TNC investment in Ireland in the 1990s. The index measures the ratio of a country’s share in global FDI to its share in global GDP. For Ireland that ratio increased from 0.7 in 1988–1990 to 5.1 in 1998–2000. As a result, Ireland’s ranking among all countries went from 71 to 4 (UNCTAD 2002, 25, 27)! During the 1990s, net FDI inflows rose considerably as a percentage of GDP and of gross capital formation (GCF), even if we exclude the latter years of huge investments in the IFSC (see table 3.1). In absolute terms, real gross capital formation rose as well, though not until 1995. Honohan and Walsh (2002, 22) attribute the increase of the investment share in the second half of the 1990s to the large investments in housing.14 Initially, high-tech investment was predominantly in the form of greenfield investment. But by the late 1990s, a number of foreign companies, especially input suppliers, established a beachhead in Ireland through the purchase of existing companies, indigenous as well as foreign-owned. Global contract manufacturers became important investors in Ireland, as they expanded their global operations through the acquisition of already existing companies. Celestica, for example, one of the big global players in electronics contract manufacturing, entered Ireland in 1998 through the acquisition of Madge Networks, a U.K.-based company.

FDI and Transfer Pricing: The Growing Gap between GDP and GNI The TNCs have taken advantage of the low tax rates in Ireland to overstate the share of profits from their global operations that was actually generated in the country. By under-invoicing its imports and over-invoicing its exports, a TNC can inflate its profits in Ireland thus shifting declared profits

THE RISE OF THE CELTIC TIGER

51

from a high-tax country to a low-tax country, a phenomenon commonly referred to as transfer pricing. Rules governing transfer pricing are a murky area of tax law, and it has been notoriously difficult to prove transfer pricing outside those stipulations, that is illegal activities, as it involves transactions between different affiliates of the same TNCs. The pervasiveness of transfer pricing in Ireland distorts key macro variables like GDP and trade, making it impossible to interpret the official data as an accurate reflection of the real values produced and traded by TNCs. Among all the countries where U.S. TNCs operated in 2002, Ireland ranked first with respect to the amount of pretax profits declared by U.S. foreign affiliates. The U.S. TNCs declared $28.6 billion profits in Ireland, roughly 10 percent of all profits of U.S. TNC affiliates worldwide (Sullivan 2004b). Official data show that Ireland’s real GDP grew at an average annual rate of 6.9 percent during the 1990s, more than twice as high as during the 1980s and more than 50 percent higher than in the 1960s and 1970s. Honohan and Walsh (2002) calculated that during 1995–1998 GDP growth adjusted for transfer pricing would have been two percentage points lower, 6.2 percent instead of 8.2 percent.15 Transfer pricing is also one of the reasons why the share of manufacturing value added in GDP is so much larger in Ireland (33.3 percent in 2000) compared to the EU average (19.5 percent) or OECD average (18.3 percent) (OECD 2003, 191). The flipside of transfer pricing is profit repatriation. When TNCs overstate their profits in Ireland, they are interested in reaping the tax benefits, but they are often interested in investing the after-tax profits outside the host country. The outflow of profits, or profit repatriation, has driven an increasing wedge between GDP and gross national income (GNI) (see graph 3.2). Historically, income remittances of the extensive Irish expatriate community have made Ireland’s GNI larger than its GDP. But in 1980, GNI became smaller than GDP, as a result of interest payments on foreign debt and profit remittances, and by 2000, GNI was only 85 percent of GDP. For most countries in the world GNI and GDP are roughly equal. It is highly unusual to have a difference of 15 percentage points between the two. In 2000, there were only three countries in the world with a GNI to GDP ratio smaller than Ireland’s.16 All three were low-income countries in Africa: Angola with a ratio of .8, the Republic of Congo with .7, and Equatorial Guinea with .33. Other SLCs at Ireland’s level of development present a different picture. The ratios for Israel and Singapore, for example, were .96 and one, respectively. Even though transfer pricing inflates Ireland’s GDP figures, it does not negate the reality of the Celtic Tiger. Real GNI still grew at an annual rate of 5.8 percent during the 1990s, rather than the 6.9 percent for real GDP.

52

FOREIGN INVESTMENT, DEVELOPMENT, AND GLOBALIZATION 1.20 1.10 1.00 0.90 0.80 0.70 0.60 1960 1965 1970 1975 1980 1985 1990 1995 2000 GNI/GDP

Graph 3.2 GNI/GDP Ratio, Ireland, 1960–2002. Source: Calculated based on World Bank World Development Indicators ⬍http://devdata.worldbank.org/ dataonline/⬎ accessed December 2003.

One of the clearest indications of real economic growth in Ireland is the decline in the unemployment rate. Although the unemployment rate remained stubbornly high in the first half of the 1990s, it had fallen to 4.7 percent in 2000, the lowest rate in 20 years. At the end of the twentieth century, Ireland’s historically chronic unemployment problem had disappeared, and after more than a 100 years of emigration the country has become a net recipient of immigrants, of returning Irish expatriates as well as non-nationals. Employment grew by 5 percent annually between 1993 and 2000, making the Irish record of job creation in the 1990s more outstanding than the frequently highlighted employment boom in the United States (Clinch, Convery, and Walsh 2002, 46). FDI, Trade, and Structural Change TNCs have had a significant impact on Ireland’s international trade and the balance of payments. Since they have used Ireland as an export platform, the country’s trade openness has grown considerably with the influx of FDI. The share of exports of goods and services in GDP increased from an average of 32.8 percent during the 1960s to 39.8 percent in the 1970s to 52.8 percent in the 1980s.17 Between 1990 and 2000, the export share rose from 57 percent to 94.9 percent, while the import share increased from

THE RISE OF THE CELTIC TIGER

53

52.4 percent to 80.7 percent.18 While the balance of merchandise trade has continued to be highly positive, the balance of service trade has become increasingly more negative in the course of the 1990s, primarily due to TNC-related service imports. The deficit on the service trade balance went from €651 million in 1990 to €13.1 billion in 2000, the equivalent of 12.6 percent of GDP (Forfas 2002, 3). And in 2000, the three largest categories of service imports accounted for nearly 70 percent of total service imports. They were royalties and licenses, miscellaneous business services (e.g. accounting and legal services), and trade-related services. In 2000, 62 percent of Ireland’s exports went to other EU countries. Great Britain continues to be the single most important destination for Ireland’s merchandise exports, accounting for 22 percent. Nonetheless that is a dramatic reduction of Ireland’s complete export dependence on the British market, with nearly 90 percent of Irish exports destined for Great Britain in the 1950s. Of the 38 percent of exports, which went to non-EU countries in 2000, 17 percentage points went to the United States, mainly in Information and Communication Technology (ICT) goods and pharmaceuticals. Since over a third of Irish exports are destined for non-EU countries and over 20 percent for the United Kingdom, which is not a member of the European Monetary System (EMS), exports are susceptible to exchange rate movements of the Euro vis-à-vis the British pound and the US dollar.19 The vulnerability to exchange rate changes is mitigated to the extent that TNC affiliates in Ireland export to sister affiliates in other EMS member countries. The FDI boom in Ireland in the 1990s has led not only to a dramatic increase in exports, but also to a remarkable change in their structure. The share of high-tech exports grew by 25 percentage points in the course of the 1990s, accounting for nearly 60 percent of all manufactured exports in 2001. That is substantially larger than the EU and the OECD averages, which were 23.5 percent and 26.4 percent respectively. The difference is mainly accounted for by the export share of pharmaceuticals, media, and computing equipment. In 2001, exports of those three high-tech sectors accounted for 52.8 percent of Ireland’s manufactured exports, compared to 16 percent for the EU average. The high export share reflects the significant impact of FDI on high-tech production in Ireland, but also, once again, the implications of extensive transfer pricing. The structural change in manufacturing production has been equally pronounced. Because of the influence of transfer pricing on output data, changes in employment allocation in the manufacturing sector give a better indication of the extent of structural change (see table 3.2).20 The chemical and pharmaceutical sector (NACE 24) and the electrical, electronic, and optical sectors (NACE 30–33) each doubled their share in real gross output between 1991 and 2000. Together they accounted for

Structure of the Irish Manufacturing Sector Based on Output and Employment, 1991 and 2000 Real gross output

15–16 Food, drink, and tobacco All local units Foreign-owned local units Irish-owned local units 17–18 Textiles and textile products All local units Foreign-owned local units Irish-owned local units 20 Wood and wood products All local units Foreign-owned local units Irish-owned local units 21–22 Paper, print, and publishing All local units Foreign-owned local units Irish-owned local units 24 Chemicals and chemical products All local units Foreign-owned local units Irish-owned local units 25 Rubber and plastic products All local units

54

Table 3.2

Persons engaged

Absolute numbers (in 1,000s of €)

% distribution

Absolute numbers (in 1,000s of €)

% distribution

1991

2000

1991

2000

1991

2000

1991

2000

8,494,168 2,579,280 5,914,888

11,087,202 4,988,703 6,098,500

35.1 19.9 52.4

15.8 9.1 39.8

44,747 12,683 32,064

48,102 13,170 34,932

22.7 14.6 29.1

18.8 10.7 26.3

812,707 454,786 357,921

573,487 272,485 301,001

3.4 3.5 3.2

0.8 0.5 2.0

21,028 9,573 11,455

10,265 3,546 6,719

10.7 11.0 10.4

4.0 2.9 5.1

238,967 58,200 180,767

512,057 141,864 370,193

1.0 0.4 1.6

0.7 0.3 2.4

4,285 480 3,805

6,249 1,111 5,138

2.2 0.6 3.5

2.4 0.9 3.9

1,593,955 840,950 753,004

6,624,108 5,413,768 1,210,340

6.6 6.5 6.7

9.4 9.8 7.9

16,761 3,336 13,425

23,816 7,457 16,359

8.5 3.8 12.2

9.3 6.1 12.3

3,646,519 3,028,118 618,402

24,185,187 23,203,287 981,900

15.0 23.4 5.5

34.4 42.2 6.4

14,668 11,346 3,322

23,198 17,874 5,324

7.5 13.1 3.0

9.1 14.5 4.0

532,471

890,915

2.2

1.3

8,106

10,846

4.1

4.2

366,502 524,412

2.4 2.0

0.7 3.4

4,360 3,746

3,951 6,895

5.0 3.4

3.2 5.2

1,246,866 223,339 1,023,527

2.9 1.1 5.0

1.8 0.4 6.7

10,103 1,848 8,255

11,166 1,584 9,582

5.1 2.1 7.5

4.4 1.3 7.2

1,465,559 482,191 983,368

4.2 3.4 5.1

2.1 0.9 6.4

12,876 3,637 9,239

16,884 3,554 13,330

6.5 4.2 8.4

6.6 2.9 10.0

1,272,682 683,657 589,025

3.7 4.7 2.6

1.8 1.2 3.8

12,380 7,001 5,379

14,396 6,436 7,960

6.3 8.1 4.9

5.6 5.2 6.0

28,182,406 26,602,924 1,579,482

20.2 35.3 2.9

40.1 48.4 10.3

32,483 27,260 5,223

68,967 55,861 13,106

16.5 31.4 4.7

27.0 45.4 9.9

16,036,552 15,597,062 439,490

10.8 19.5 0.8

22.8 28.4 2.9

8,019 3,322 1,252

20,723 18,303 2,420

4.1 3.8 1.1

8.1 28.4 1.8

1,992,351 1,546,475 445,876

2.7 4.1 0.8

2.8 2.8 2.9

10,278 7,811 2,467

15,141 9,438 5,703

5.2 1.0 2.2

5.9 7.7 4.3

5,028,519 4,745,009 283,510

2.3 4.0 0.4

7.1 8.6 1.8

4,887 4,128 759

14,993 12,785 2,208

2.5 4.8 0.7

5.9 10.4 1.7

Continued

55

Foreign-owned local units 305,854 Irish-owned local units 226,617 26 Non-metallic mineral products 710,523 All local units 147,586 Foreign-owned local units Irish-owned local units 562,937 27–28 Basic and fabricated metals All local units 1,010,167 Foreign-owned local units 439,678 570,489 Irish-owned local units 29 Machinery and equipment n.e.c. All local units 894,835 Foreign-owned local units 605,323 Irish-owned local units 289,513 30–33 Electrical and optical equipment All local units 4,895,709 Foreign-owned local units 4,571,577 Irish-owned local units 324,132 30 Office machinery and computers All local units 2,622,257 Foreign-owned local units 2,529,968 Irish-owned local units 92,290 31 Electrical machinery 654,576 All local units Foreign-owned local units 526,446 Irish-owned local units 92,290 32 Radio, tv, and communication equipment All local units 566,655 Foreign-owned local units 522,229 Irish-owned local units 44,426

Continued

56

Table 3.2

Real gross output

33 Medical, precision, and optical instruments All local units Foreign-owned local units Irish-owned local units 34–35 transport equipment All local units Foreign-owned local units Irish-owned local units 15–37 All manufacturing All local units Foreign-owned local units Irish-owned local units

Persons engaged

Absolute numbers

% distribution

Absolute numbers

1991

2000

1991

2000

1991

2000

1991

2000

886,285 843,262 43,023

3,372,659 3,036,149 336,510

3.7 6.5 0.4

4.8 5.5 2.2

9,299 8,554 745

18,110 14,114 2,504

4.7 9.8 0.7

7.1 11.5 1.9

426,572 96,674 329,899

705,321 477,395 227,926

1.8 0.7 2.9

1.0 0.9 1.5

8,940 1,590 7,350

9,610 5,365 4,245

4.5 1.8 6.7

3.8 4.4 3.2

24,231,023 12,947,217 11,283,806

70,343,646 55,008,128 15,335,518

100.0 100.0 100.0

100.0 100.0 100.0

196,878 86,869 110,009

255,644 122,978 132,666

100.0 100.0 100.0

100.0 100.0 100.0

Note that the percentages do not add up to 100, since NACE 36–37 (Manufacturing n.e.c.) is not included. Source: Calculated based on the CSO Census of Industrial Production and the wholesale price index, facilitated by Breathnach, Forfas.

% distribution

THE RISE OF THE CELTIC TIGER

57

74.4 percent of the output value in 2000. However, the fact that the employment share of the electronics sector rose considerably in the course of the decade (from 16.5 percent to 27 percent), while the employment share of pharmaceuticals increased only slightly (from 7.5 percent to 9.1 percent), suggests extensive profit shifting by TNCs in the pharmaceutical and chemical industry. Pfizer’s production of Viagra in Cork constitutes the single most important product in the chemical sector and likely contributes significantly to the output in that sector. Worldwide, the pharmaceutical industry generates high profits, and the decision on where to declare those profits has become increasingly more sensitive to tax rates. Sullivan (2004a) found that the non-U.S. share of worldwide profits of six top U.S. pharmaceutical companies increased from 37.6 percent to 65.1 percent between 1994 and 2003, even though the companies’ physical production did not see a similar shift. While Sullivan does not focus specifically on Ireland, he points out that all the top nine U.S. multinationals analyzed have operations in low-tax Ireland. Transfer pricing aside, the change in employment shows that Ireland’s manufacturing sector underwent remarkable structural change during the 1990s. In 2000, 36 percent of manufacturing employment was generated in the high-tech sectors (NACE 24 and 30–33), compared to 25 percent at the beginning of the 1990s. The high-tech sectors accounted for 60 percent of all employment by foreign-owned local units in 2000 (up from 43 percent in 1991), and foreign firms employed 80 percent of the workers in those sectors. Irish-owned companies, in contrast, have continued to dominate the more traditional sectors. They generated over 70 percent of employment in food, drink, and tobacco (NACE 15–16) and paper and printing (NACE 21–22). And employment in indigenously-owned firms in these two sectors accounted for around 55 percent of total employment of indigenouslyowned manufacturing companies. In spite of the increase in employment in both foreign- and Irish-owned manufacturing firms, Ireland has followed the path of all developed countries, where an increasing share of total employment has been generated in the service sector. Employment in manufacturing, mining, and quarrying made up 20.7 percent of total employment in 1994, but only 17.3 percent in 2002 (see table 3.3).

FDI and Taxes The low tax rate on corporate profits in Ireland has been a major attraction for TNCs. It goes back to the change in development strategy in the 1950s, when the government placed major emphasis on bringing foreign

58

FOREIGN INVESTMENT, DEVELOPMENT, AND GLOBALIZATION

Table 3.3

Employment Distribution in the Irish Economy, 1994 and 2002 1994

Agriculture, forestry, fishing Manufacture, mining, quarrying Construction Wholesale & retail distribution Hotels and restaurants Transport, storage, communication Financial & other business services Public administration & defense Education Health Other services Total

2002

1,000s

%

1,000s

%

146.8 251.9 91.5 168.9 68.2 55.8 114.1 66.3 80.4 100.9 74.1

12.0 20.7 7.5 13.9 5.6 4.6 9.4 5.4 6.6 8.3 6.1

120.7 302.9 181.1 245.9 104.8 110.2 229.1 89.2 110.0 157.0 99.0

6.9 17.3 10.3 14.1 6.0 6.3 13.1 5.1 6.3 9.0 5.7

1,218.8

100.0

1,749.9

100.0

Source: Department of Enterprise, Trade, and Employment (2003, 110).

companies to Ireland. The Export Sales Tax Relief Act of 1956 provided for a 50 percent tax exemption on profits from export sales. That was extended to 100 percent in 1958. While the exemption was applicable to both foreign and Irish-owned enterprises, it was more beneficial for foreign enterprises, since their export share was much larger. When Ireland joined the EEC in 1973 it faced growing pressure to change the tax rate, as it was seen as giving an unfair advantage to exporters and thus violated the EEC requirements of nondiscrimination. The Irish government finally abandoned the zero tax rate on export profits in 1981. It adopted a uniform 10 percent tax for all manufacturing activities, effective for 20 years. In 1990, the regime was extended till 2010. However, inequity in corporate tax rates across sectors persisted, as the tax rate for services was 40 percent, with the exception of the IFSC, which also enjoyed a 10 percent rate. To provide a level playing field for all, the government introduced a universal rate of 12.5 percent, starting in January of 2003. But since the effective tax rate in Ireland is still a lot lower than in the other EU member countries (see table 2.2 again), there is continuing and increasing pressure on Ireland to raise its corporate tax rate. Total tax revenues are, of course, the product of the tax rate and the total value of profits. While the tax rate has been low since 1981, the influx of FDI combined with profit shifting has led to a huge increase in the value of declared profits during the 1990s. As a result, corporate profit taxes as a share of government revenue tripled during the 1990s, reaching levels that were clearly exceptional by historical standards for the country. Corporate profit taxes as a share of tax revenue were, on average, 3.5 percent during the

THE RISE OF THE CELTIC TIGER

59

1950s, 4.1 percent in the 1960s, 4.4 percent in the 1970s and 4.8 percent in the 1980s. In the Nineties, the average share rose to 10.5 percent, reaching 17.7 percent by 2002.21 And TNCs’ share in Ireland’s corporate taxes has risen commensurately with their growing importance in the Irish economy. Whether Ireland’s corporate tax share is high or low in comparison with other countries depends on the point of reference. The estimated corporate tax share of 17.7 percent for 2002 compares very favorably with Sweden and the United States, where it was 8 percent.22 In the United Kingdom, on the other hand, it was 20.7 percent.23 TNCs’ net tax contributions are somewhat diminished by the grants that Ireland’s Industrial Development Authority (IDA) offers to TNCs. Yet, since the total budget of the IDA is equivalent to about one tenth of the taxes paid by TNCs, TNCs’ net contribution to tax revenue is indeed considerable. The IDA (2003, 5) estimates that its client companies paid around €2.2 billion in corporate taxes in 2002. Yet, the rising TNC tax contribution has to be seen in the context of an overall tax ratio, which is fairly low in comparison with other OECD countries. As a share of GDP, taxes and overall government revenue in Ireland are at the lower end among OECD countries. They are closer to the U.S. ratios than those of most Western European countries (see table 3.4). In sum, the discussion above has shown that the large inflow of direct foreign investment into the chemical, electronics, and medical device industries in the 1990s resulted in substantial growth of manufacturing output, exports, and employment. It also promoted ongoing structural change toward high-tech sectors. Transfer pricing and profit repatriation have driven a growing wedge between GDP and GNI in Ireland in the 1990s, which makes it difficult to pinpoint the exact impact of TNCs on key macroeconomic variables.24 Compared to most other countries, corporate profit tax rates have been very low in Ireland. Nonetheless, the large influx of FDI and the large amounts of declared TNC profits in the 1990s have generated substantial tax revenue for the Irish government and are the key factor behind the tripling of the corporate tax share in the course of the 1990s.

Why Ireland? Global Value Chains, the EU, and Location-Specific Assets Ireland’s success in attracting large amounts of high-tech FDI in the 1990s has its origin in a critical conjuncture of external and internal factors. The dramatic increase in demand for electronics products in Europe in the 1980s, especially for computers and related products, coincided with the implementation of the directives of the Single European Act of 1986,

60

FOREIGN INVESTMENT, DEVELOPMENT, AND GLOBALIZATION

Table 3.4 Taxes and Government Revenue as a Share of GDP, OECD Countries, 1995 and 2002 Tax receipts/GDP

Australia Austria Belgium Canada Czech Rep. Denmark Finland France Germany Greece Hungary Iceland Ireland Italy Japan Korea Luxembourg Mexico Netherlands New Zealand Norway Poland Portugal Slovak Republic Spain Sweden Switzerland UK US

Government revenue/GDP

1995

2002

1995

2002

– 26.3 29.3 30.6 23.2 47.5 31.1 24.4 22.6 21.3 – 31.5 27.2 27.5 18.2 – 30.2 14.5 23.4 34.1** 32.8 28.6 22.6 27.2 20.6 35.8 – 28.2 20.9

31.5 29.0 31.1 29.4 21.7 47.1 33.4 27.2 22.8 24.1 25.7 34.8 23.8 29.0 16.2 – 30.3 15.8* 25.0 32.8 34.4 21.4 24.7 19.5 22.9 35.2 – 29.3 19.6

35.5 51.9 48.5 43.9 49.9 58.0 55.7 49.7 46.1 40.9 – 40.7 39.4 45.8 31.1 24.6 47.6 – 47.3 44.6 54.9 39.6 53.3 38.4 60.8 60.8 33.3 39.1 33.9

36.8 50.9 50.5 43.4* 48.1 57.4 54.4 50.2 45.0 45.3 43.4 44.9 33.1 45.6 30.3 30.3 46.7 – 45.9 – 56.5 43.2 45.3 39.9 58.1 58.1 35.9 40.1 32.5

* 2001 ** 1997 Source: OECD General Government Accounts 2004, ⬍http://www.oecd.org/document/5/0,2340, en_ 2825_495684_33785349_1_1_1_1,00.html⬎ accessed October 2004.

which aimed to bring about a truly borderless economic area, with 340 million inhabitants by the end of 1991. Those factors shaped TNCs’ desire to open or expand production facilities within the European Union, and Ireland was the country with the right location-specific advantages at that particular juncture. It could offer a well-educated, English-speaking, and relatively cheap labor force, a sufficiently developed national infrastructure, macroeconomic and political stability and stable capital–labor

THE RISE OF THE CELTIC TIGER

61

relations, a proactive government with highly effective institutional capacity, and grants and low taxes. Once major brand name corporations had invested in Ireland, others followed. Some were reassured by the stamp of approval that those first companies had put on Ireland as a lucrative investment location. Others were attracted by potential agglomeration economies. And still others followed as global input suppliers to their TNC clients who had established production facilities in Ireland. But none of them would have invested in Ireland, if the country had not had location-specific assets that matched the specific needs of the companies.

Critical External Factors Outside Ireland’s Control During the 1980s the PC became a mass consumption item. The European Community (EC) was an important consumer market targeted by U.S. corporations. Both U.S. electronics exports to the EC and sales of U.S. TNC affiliates within the EC increased considerably in the course of the 1980s. U.S. exports of computers, peripherals, and semiconductors to the European Community increased from $4.9 billion in 1982 to $12 billion in 1988, accounting for 38 percent of U.S. global exports in that category (Fieleke 1989, 22). And the sales of electric and electronic equipment of U.S. TNC affiliates in the EC grew from $11.9 billion in 1982 to $24.1 billion in 1986 (Fieleke 1989, 23). A widely shared assessment at the end of the 1980s was that the European PC market would grow much faster than the U.S. market in the years to come.25 By the 1980s, U.S. TNCs were well established in the electronics industry and other industries in EC member countries. Yet, in the course of that decade, the EC became a more important destination for direct foreign investment from the United States. Between 1982 and 1989, the EC increased its share in the U.S. foreign investment position from 36.2 percent to 41.6 percent (see table 3.5). Part of the rise was due to the fact that Spain and Portugal joined the EC in January of 1985. But excluding Spain, the share of the original 10 member countries was still 39.9 percent for all industries, and 44 percent for manufacturing, a substantial increase over 1982.26 The increase was partly a response to the Single European Act of 1986, which provided a road map for the elimination of remaining intra-EC technical, physical, and fiscal barriers and the establishment of a truly free market by 1992.27 “Opportunity Europe” motivated higher U.S. foreign investment in the EC, because foreign companies saw the opportunities an enlarged market had to offer. A number of U.S. PC producers like Compaq and Apple, for example, started shifting more production to their European subsidiaries (Kirkland 1989, 260).

62

FOREIGN INVESTMENT, DEVELOPMENT, AND GLOBALIZATION

Table 3.5 U.S. Direct Investment Position Abroad, 1982–2002 (on a Historical Cost Basis) All industries millions of $ 1982 All countries EC (10) Ireland 1989 All countries EC (12) Ireland 1994 All countries EU (12) Ireland 2000 All countries EU (15) Ireland 2002 All countries EU (15) Ireland

Manufacturing

% share

millions of $

% share

207,320 74,951 1,796

100.0 36.2 0.9

83,500 – –

– – –

380,068 158,169 4,449

100.0 41.6 1.2

146,907 68,585 2,912

100.0 46.7 2.0

612,893 253,242 7,239

100.0 41.3 1.2

200,996 93,268 3,830

100.0 41.3 1.9

1,316,247 609,674 35,903

100.0 46.3 2.7

343,899 157,099 9,658

100.0 45.7 2.8

1,520,965 699,970 41,636

100.0 46.0 2.7

392,553 188,353 13,197

100.0 48.0 3.4

Source: U.S. Department of Commerce (1998, 1992, 1985), for 1994, 1989, and 1982 respectively: For 2000 and 2002: ⬍http://www.bea.doc.gov/bea/di/di1usdbal.htm⬎ accessed December 2003.

But TNCs also saw the specter of “Fortress Europe,” as the EC adopted more stringent “rules of origin” and local content requirements, which meant that some non-EC producers would be disadvantaged and potentially excluded from the European market. The semiconductor industry is a case in point. In early 1989, EC policy changed in that henceforth microchip production would only satisfy the “rule of origin” requirement for tariff-free intra-EU trade, if the core activity was undertaken in an EC member country. The core activity in that case was defined as the chip diffusion process, where the circuits are laid on the chip. The change in legislation was one reason why U.S. microchip manufacturers increased their investment in the EC.28 A case in point is Intel whose decision to invest in Ireland in 1990 marks the beginning of the investment boom in the electronics industry in the 1990s. In 1988, 24 percent of Intel’s revenue came from sales in Europe, where its sales were increasing twice as quickly as in the United States (Wall Street Journal 1989). Intel had been considering a major investment in an EC country in anticipation of “Opportunity and Fortress Europe,” but the

THE RISE OF THE CELTIC TIGER

63

actual search for the appropriate investment site was precipitated by the change in the rule of origin requirements.29 Semiconductors were the fastest growing products in international trade between 1985 and 2000, with world exports increasing from $26 billion to $235 billion, and Intel was becoming the world’s leading semiconductor manufacturer, increasing its sales 43 fold between 1983 and 2000 (UNCTAD 2002, 26).30 Hufbauer (1990, 18–19) refers to the results of the change in rules of origin for semiconductors as “the outstanding example of a seemingly small change in technical rules triggering production decisions that, in time, will very likely shift the balance of the world’s trade in semiconductors.” While Hufbauer may not have known that Intel was looking to produce within the EU anyway, the quotation highlights how changes completely outside the control of any national government can nonetheless have important implications for the destination of foreign investment flows. As a result of Intel’s rising leadership in the industry, its investment in Ireland had an important signaling effect for other potential foreign investors. On the one hand, it increased foreign investors’ awareness of Ireland as an attractive investment site, as they were looking to establish production sites within the European Union. And on the other hand, it is likely that the agglomeration benefits of a critical number of foreign investors in the same location had a snowballing effect and brought in additional companies. A representative of Intel argued, “the most effective thing that Intel has done for Ireland was to move the country from a second-tier investment place to a first-tier investment place. It had a demonstration effect for HP, IBM, Xerox. In the first couple of years we used to have endless trails of visitors asking what we were doing here.”31 Agglomeration benefits refer to the efficiency gains that can come from knowledge spillovers, a concentration of industry-specific skills, and potential intercompany linkages. Barry, Bradley, and O’Malley (1999, 65) cite “surveys of executives of newly arriving foreign companies in the computer, instrument engineering, pharmaceutical and chemical sectors [which] indicate that their location decision has been strongly influenced by the fact that other key market players are already located in Ireland.” Asked why Hewlett Packard invested in Ireland in 1995, a representative of the company named the fact that Intel was there already as one important factor.32 In an empirical study of the importance of agglomeration-efficiency effects for the clustering of foreign companies in Ireland, Barry, Goerg, Strobl (2001) find that agglomeration benefits played a statistically significant role for U.S. companies, but not for U.K. companies. U.S. companies are mainly concentrated in the high-tech sectors, while British TNCs are more prominent in food production, textiles, shoes, and paper.

64

FOREIGN INVESTMENT, DEVELOPMENT, AND GLOBALIZATION

In a number of cases, foreign input suppliers and service providers to the big TNC computer and medical device companies followed them to Ireland. Based on their established business relationships they had an advantage in providing the same inputs or services there that they had been supplying already in the TNC home country. Other companies, like some of the contract manufacturers in electronics who came in the second half of the 1990s, invested in Ireland in the hope of acquiring more large TNC customers that had established production facilities there. “Europe 1992” had another, quite different, but equally important, implication for the rise of the Celtic Tiger. It led to substantially higher levels of net transfer from Brussels to Dublin. The transfers financed a significant share of the improvements in education and infrastructure as well as the training grants of the IDA, all of which played an important part in the attraction of FDI. EC policies have always reflected the belief that a true common market can only be sustainable if ever deeper market liberalization among member countries is combined with a net transfer of development funds to the poorer member countries and regions within countries to speed up income convergence within the European Community. With the accession of Spain and Portugal to the EEC in 1985 and in preparation for Europe 1992, the EEC completely restructured its funding programs, which resulted in much larger transfers to Ireland. The structural funds, which provide funding through four different channels, were doubled progressively between 1989 and 1993.33 With the Treaty of Maastricht in December 1991 the structural funds were increased further. In addition, the Cohesion Fund was established in the early 1990s, especially to help the four poorest countries of the EC face the challenges of further economic and monetary union.34 And the poorest countries at the time were Ireland, Greece, Portugal, and Spain. To obtain structural funds assistance, an eligible EC member country has to submit a plan to the European Commission, specifying its development priorities and how EC grants and domestic public and private resources will be used to meet them. The plan forms the basis for the Community Support Framework (CSF), an agreement between the European Community and the member country on development priorities and funding. Throughout the 1980s, EC structural funds for Ireland amounted to around 1.5 percent of GNP. But during 1991–1993 they increased to nearly 3.5 percent, and for the rest of the 1990s, they hovered around 2.5 percent of GNP (Fitzgerald 1998, 681). That is a remarkable infusion of funds. Indeed, it is of a magnitude equivalent to the Marshall Aid that Ireland and other European countries received after World War II. Between April 1, 1948 and June 30, 1952, Marshall aid amounted to 2.7 percent of Ireland’s

THE RISE OF THE CELTIC TIGER

65

Table 3.6 Marshall Aid as a Share of Recipients’ GDP, 1948–1952*

Austria Belgium Denmark France Germany, Fed. Rep. Greece Ireland Netherlands Norway Portugal Sweden United Kingdom

Total

Grants

Loans

4.92 1.79 1.86 2.25 1.21 8.86 2.68 4.56 2.20 3.70 0.44 2.02

4.92 1.57 1.63 2.06 1.02 8.86 0.35 3.86 1.86 1.09 0.36 1.78

0.00 0.22 0.23 0.19 0.19 0.00 2.33 0.70 0.34 2.61 0.08 0.24

* Marshall aid covered the period from April 1, 1948 to June 30, 1952. The GDP figures for the two years have been adjusted accordingly. Where GDP figures were not available for specific years (for Germany for 1948 and 1949, and for Portugal for all years except 1950), the Marshall aid figures were adjusted accordingly. Sources For Marshall Aid: ⬍http://www.marshallfoundation.org/about_gcm/marshall_ plan.htm⬎ For all GDP data in local currency (except France): Mitchell (1998), Table J1. For GDP data in local currency for France: IMF (1954). For exchange rate data: IMF (1953).

GDP, and only .35 percent, when we exclude the loan element (see table 3.6). Structural funds and resources from the Cohesion Fund were of critical importance behind the rise of the Celtic Tiger. The most immediate positive effects were demand effects that provided a temporary boost to economic growth. But the long term and lasting positive effects were the supply effects, as structural funds were used to improve the basis for higher growth over time. Between 1989 and 1999, Ireland received 8.1 billion Irish pounds (in 1994 prices) through the CSF as well as additional non-CSF funding (Barry, Bradley, and Hannan 1999, 108). During the 1990s, most of the funding went to the private sector in various types of investment support, followed by funding for the improvement of human resources and physical infrastructure (see table 3.7). Barry, Bradley, and Hannan (2001) estimate that through the demand effects structural funds may have contributed as much as half a percentage point per year to GNP growth in the 1990s.

66

FOREIGN INVESTMENT, DEVELOPMENT, AND GLOBALIZATION

Table 3.7

Allocation of EU Funds for Ireland, 1989–2006 Community support framework

Aid to productive sectors Human Resources Physical infrastructure

CSF 1989–1993

CSF 1994–1999

CSF 2000–2006

56.0 25.0 19.0

47.0 32.0 21.0

16.0 36.0 48.0

Source: Bradley (2004, 35).

But there were other benefits from the structural funds, which were not and cannot be captured by these numbers: the right timing and the role they played in attracting TNCs and improving planning and budgeting capabilities. Fitzgerald (1998, 683) points out that—with the benefit of hindsight—it was particularly fortuitous that the increase in structural funds happened at the beginning of the 1990s: “The increase in funding under the first CSF encouraged the government to raise public investment from its extremely low level in the late 1980s. Without such a stimulus Ireland could have found itself suffering from under-investment in the face of rapid growth.” In addition, the improvement in human resources and infrastructure as well as aid to private investment bolstered Ireland’s location-specific assets during the very period when TNCs were increasingly interested in the country as an investment site. The CSF process also forced the government to formulate a development plan and introduce evaluation procedures for the different projects (Bradley 2004; Fitzgerald 1998). And the preparation of national development plans induced multiannual budgeting for capital purposes, which produced “more efficient delivery of investment and a higher rate of return” (Economic and Social Research Institute [ESRI] 2003). In sum, structural funds played a very important role in the rise of the Celtic Tiger. While the demand effects of the transfers from Brussels can be estimated and the direct supply effects can be simulated, it is the immeasurable benefits that are equally, if not more, significant. EU transfers allowed Ireland to enhance its location-specific assets through improvements in human capital and infrastructure, and those improvements were important in attracting and keeping FDI. Ireland’s Location-Specific Assets The boom in the PC industry in conjunction with Europe 1992 explains why US companies were looking to invest in a European Union member

THE RISE OF THE CELTIC TIGER

67

country in the 1990s. But these factors do not explain why they chose to invest in Ireland. In 1994, the member states of the EU accounted for 41.3 percent of the U.S. stock of foreign investment abroad. By 2000, the share had risen to 46.3 percent, with Ireland accounting for a substantial part of that increase (see table 3.5).35 The signaling effects of the early large investors with international name recognition were undoubtedly important, since in the late 80s—according to several CEO surveys at the time— Ireland did not register as a top investment site.36 But, in the end, it was the location-specific assets that Ireland had to offer, particularly to the hightech industries, which made it a desirable investment location. In many ways, the country was in the right spot at the right moment. Education and Labor Cost High-tech production requires the availability of a labor force with the right skills. That means a large number of workers who are technically proficient for the more standardized parts of the production process as well as a sizeable number of specialized engineers for the technologically demanding processes. In the early 1990s, Ireland could offer both, as a result of major educational investments and reforms dating back to the late 1960s. High unemployment and national pay agreements with moderate wage increase meant that labor was relatively cheap. That English was the national language was an added benefit, especially for high-tech industries. And a highly elastic labor supply due to favorable demographics, rising female labor force participation rates, and positive net immigration could meet the growing demand for labor in the course of the Nineties. The Irish government started major educational reforms in the late 1960s in response to an OECD report on education in Ireland, which highlighted the poor state of education in the country, especially in comparison with other OECD member countries. It introduced free second level education, raised the minimum school leaving age from 14 to 15 in 1972, and expanded third level education, especially in the non-university sector. In the late 1980s, the government introduced a means-tested third-level grant system and expanded the development of Regional Technical Colleges, which focused primarily on engineering, the sciences, and business (Durkan, Fitz Gerald and Harmon 1999; Smyth and Hannan 2000). As a result, enrolment rates in secondary and tertiary education expanded dramatically (see table 3.8). In 1997, 29 percent of the working population had completed third level education, compared to 21 percent in 1991 (FAS 2002). In addition to the broad-based commitment to expanding education opportunities at the secondary and tertiary levels, the government was very responsive to skills demands from TNCs. In the late 1970s, when

68

FOREIGN INVESTMENT, DEVELOPMENT, AND GLOBALIZATION

Table 3.8

School Enrollment Rates in Ireland, 1970–1999

Year

Gross primary school enrolment rate

Gross secondary school enrolment rate

Gross tertiary school enrolment rate

Scientists & engineers in R&D per million people

1970 1975 1980 1985 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999

107.4 104.2 99.9 99.9 102.8 102.8 103.7 104.0 104.2 103.8 104.5 140.7 – 118.9

73.8 87.1 89.7 97.8 100.5 102.6 111.3 113.5 115.0 115.9 117.6 109.1 – –

11.9 16.8 18.1 22.3 29.3 32.7 34.7 37.3 38.0 39.6 41.0 43.0 44.5 46.2

– – – – 1,317 1,463 1,563 1,362 1,476 1,600 1,873 2,132 2,080 2,190

Source: World Bank World Development Indicators ⬍http://devdata.worldbank.org/dataonline⬎ accessed October 2004.

Wang, Texas Instruments, Verbatim, and Mostek invested in Ireland, it became obvious that the lack of qualified engineers was a major obstacle to the expansion of the electronics sector. White (2000, 284), then the head of the IDA, argues that “as a short-term solution they [the government and the colleges] proposed to convert science graduates to electronics qualifications via one-year conversion courses, and for the medium term they expanded or added new courses. By autumn 1979, the Higher Education Authority was given more funds and fourteen special conversion courses and 58 new or expanded courses in electrical engineering were under way.” The broadening and deepening of the skill level of the labor force does not happen overnight; it obviously takes time, since the effect of education policies is slow and cumulative over a period of time. In the poor economic conditions of the 1980s, a considerable number of Irish engineers left to work in other EU member countries and the United States. Yet, in the 1990s, Ireland could offer a supply of educated workers that was the result of policies started 20 years earlier. It also included the return of Irish engineers working abroad. The desire of some of the Irish expatriate engineers to come back home was a critical factor in the IDA’s ability to assure Intel that there would be enough engineers to meet the company’s needs.37 Furthermore, a number of Irish workers had acquired industry-specific skills in the late 1970s, with the

THE RISE OF THE CELTIC TIGER

69

first wave of FDI in the electronics sector. Even though many of the foreign companies of the first wave had since relocated their operations to other countries, the training of Irish workers paid off in the third wave of the 1990s. The fact that Ireland was a labor surplus economy, with the unemployment rate remaining in the double digits well into the 1990s, meant that skilled labor and engineers were relatively cheap compared to the other EU members. Data for majority-owned affiliates of U.S. TNCs show that the compensation of Irish production workers was considerably below the European average, while their working hours were considerably above the average. Between 1982 and 1994, the average compensation per hour worked by production workers in Ireland was 30 to 35 percent below the European average, but the average annual hours worked were 4 to 8 percent higher (see table 3.9).38 Important factors that made the overall increase in the Irish labor supply in the 1990s possible were demographics and rising female labor force participation rates. The baby boom of the 1960s and 1970s translated into a considerable increase in the labor force, as did the growing number of women entering the work force.39 During the period 1993–2000, the population above 15 years of age increased at an annual rate of 1.6 percent, while the employment ratio grew from 44.5 percent to 56.4 percent (Clinch, Convery, and Walsh 2002, 45). Particularly dramatic was the rise in the female labor force participation rate. It rose from 29 percent in 1988 to 62 percent in 2002 for women aged 33–44 (Walsh 2003, 5).40 Ireland’s English-speaking population has been an additional benefit for the attraction of FDI. As one TNC representative explained in an interview: “we are dealing with complex technology, and transferring it is a major challenge. If you cross the language barrier, you add to the challenge.” National Infrastructure Physical infrastructure like roads and ports were improving considerably in the 1990s—not least due to large financial support from the EU Table 3.9 Average Compensation and Annual Hours per Worker in Majority-Owned Affiliates of U.S. TNCs, 1982, 1989, 1994 Average annual hours per worker

1982 1989 1994

Compensation of workers per hour (in current US$)

Europe

Ireland

Europe

Ireland

1,740 1,763 1,783

1,820 1,904 1,859

9.44 14.45 18.80

6.15 10.17 13.22

Source: U.S. Department of Commerce (1998, 1992, 1985).

70

FOREIGN INVESTMENT, DEVELOPMENT, AND GLOBALIZATION

structural funds. Privatization of the airline industry led to lower transportation costs in the 1990s (Barry 2000). Most importantly, it was the up-to-date telecommunications infrastructure that was particularly attractive for foreign investors in the high-tech areas and IT-related services like call centers. When the Irish government decided in the late 1980s that the telecommunications system in Ireland was in dire need of improvement, the infrastructure was so outdated that it allowed the country to leapfrog straight into a digitally-based telecommunications network, positioning the country well at a time when telecommunications was becoming increasingly important in the operations of TNCs’ global production chains. Political Stability and Stable Capital–Labor Relations The political instability of the 1970s, with the violence of the Northern Ireland conflict spilling over to the South, had calmed down by the late 1980s.41 It was equally important that the conflictive capital–labor relations from the recession years of the first half of the 1980s had given way to a new social contract in 1987. Tri-partite corporatist agreements between the government, the employers, and the unions determined the increases of the base pay on a national level. Employers were allowed to give an additional 3 percent wage increase based on local negotiations. The overall thrust of the agreements has been moderation in wage increases in exchange for tax reductions for wage earners. There were six social partnership agreements between 1987 and 2003: the Programme for National Recovery (1988–1990) which recommended a nominal pre-tax pay increase of 2.5 percent; the Programme for Economic and Social Progress (1991–1993) with a recommendation of 3.58 percent; the Programme for Competitiveness and Work (1994–1996) recommending 1.82 percent; the Partnership 2000 for Inclusion, Employment and Competitiveness (1997–2000) suggesting 2.04 percent; and the Programme for Prosperity and Fairness (2000–2003) (Lane 1998, 230). The national pay agreements in conjunction with a high unemployment rate well into the later 1990s meant that—from the TNCs’ point of view— wage increases were small and predictable. Between 1987 and 1996, average real compensation (wages plus social security contribution) in the business sector grew at an annual rate of 2.3 percent (Lane 1998, 30). The increase was larger for after-tax real wages, since individual tax rates declined. For a worker earning the average industrial wage, the tax rate fell from 35.4 percent in 1987/1988 to 30 percent in 1995/1996; at 50 percent of the average wage, the decline went from 22 to 17.4 percent; and at 500 percent of the average wage, the tax rate dropped from 54.9 to 46.8 percent (Lane 1998, 231).

THE RISE OF THE CELTIC TIGER

71

The flip side of the moderate wage increases was a considerable rise in the profit rate and share. The rate of return on capital increased from 8.6 percent in 1987 to 15.4 percent in 1996; and the profit share went from 25 percent to 34.8 percent during this period (Lane 1998, 227). Part of the shift in factor shares toward capital is a reflection of underlying transfer pricing; yet it is highly unlikely that transfer pricing explains the whole change. During the 1990s the IDA and others touted a substantially higher rate of return on capital in Ireland compared to other EU countries as one of country’s key location-specific assets. Effective Institutional Capacity and Financial Incentives The effective institutional structure, which Ireland had mounted to attract FDI, was an important factor behind the foreign investment surge of the 1990s. Given the high unemployment rates in Ireland, the IDA initially went after any foreign investment that would create jobs. But by the late 1970s/early 1980s, the agency started to focus on FDI with high-tech content, primarily in electronics, computer software, and healthcare (Mac Sharry and White 2000, 207). The IDA has always worked hard to identify potential foreign investors in the prioritized areas and then courted them aggressively.42 Providing a one-stop shop, the IDA worked with potential investors to overcome anticipated hurdles, from ensuring the availability of the requisite skilled labor (recall the example of bringing back expatriate Irish engineers in the case of Intel) to providing the best site, with ready access to airports and seaports. Being a small country was advantageous in this context, since it was easier to bring together all decision-makers, who play an important role in addressing the concerns of TNCs. When the level of coordination among concerned institutions is high, the outcome can be a very speedy and effective process from beginning to end, the opposite of lengthy bureaucratic procedures that TNCs often encounter in potential host countries. Once a TNC has decided to invest in Ireland, the IDA continues to work with it in finding solutions when problems arise, especially in the area of skills shortages. As Kieran McGowan, head of the IDA during the 1990s, put it: “As head of the IDA, I spent at least three quarters of my time looking after the companies that were already here, because they can be very powerful in either direction.”43 In 1997, for example, Intel, IBM, and HP came together to talk to the government about the foreseeable skills shortage and how best to address it. A number of factors have contributed to the effective administrative capacity. First, the relevant government agencies had many highly educated people in their ranks since employment opportunities in the private

72

FOREIGN INVESTMENT, DEVELOPMENT, AND GLOBALIZATION

sector were so limited. Second, successive governments were committed to a strategy of aggressive pursuit of foreign investment, thus providing ongoing continuity to the strategy. And finally, given the high priority accorded to FDI in the Irish development strategy, the IDA always had a soft budget constraint, which gave it leeway in its aggressive pursuit of foreign investors. Several of the IDA’s grant programs were supported quite generously by EC funds in the course of the 1990s.44 White (2000, 181) argues that “the availability of European Social Funds for training of staff in new foreign companies, coupled with the positive training programmes they funded, were among a small number of crucial determinants that influenced companies to choose to locate in Ireland.” From the beginning, the Irish government provided very generous financial incentives to attract foreign corporations. In addition to the low corporate tax rates the IDA could offer generous cash grants. In nearly all of the interviews I conducted with TNC representatives in Ireland, the company representative highlighted the low tax rate as one key factor of attraction for choosing Ireland as the investment site. For companies with high margin products, like those in the pharmaceutical industry, the tax rate played a particularly important role. Nonetheless, low taxes are not always a big draw. One TNC contract manufacturer in the electronics field told me that in their particular case the effective tax rate at the company level would have been lower than the 10 percent in Ireland; so taxes were not important. Rather, access to a customer base (i.e. the potential TNC clients already located in Ireland) and global reach were the main drivers behind the investment decision of that particular company. Whereas the tax rate of 10 percent (now 12.5 percent) was applied uniformly to all corporations in the manufacturing sector, cash grants were negotiated on a case-by-case basis, with grant support given primarily in exchange for employment creation. The grants have been used as a major tool of industrial policy, particularly with respect to the targeting of high-tech TNCs. Between 1992 and 2001 IDA grants totaled €1.39 billion (see table 3.10). Total annual grant payments, in current Euros, increased until 1996, but declined steadily thereafter. Counting only the jobs that were created and sustained over a seven-year period, we see that the grant per sustained job in constant Euros has become substantially smaller over the last 20 years. And in 2001, it was roughly only a third of what it had been in the early 1980s (see table 3.11). The size of the IDA grants has been criticized repeatedly in the course of recent Irish economic history. The Telesis Report of 1982, the first major study of industrial policy commissioned by the government, argued that the grants to TNCs were too large and should be reduced, together with a switch in industrial policy to the promotion of indigenous companies.45

THE RISE OF THE CELTIC TIGER

73

Table 3.10 IDA Ireland—Grant Payments by Type, 1992–2001 (1,000s of Current Euros) Capital grant 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001

Employment grant

68,198 72,488 66,448 79,003 101,727 95,149 63,862 42,174 42,970 38,844

7,467 8,496 17,175 25,437 52,268 64,988 70,197 105,347 91,886 60,105

All other grant types

All grant types

15,522 21,146 18,803 18,858 31,466 22,757 42,117 19,839 17,711 9,815

91,187 102,130 102,426 123,298 185,461 182,894 176,176 167,360 152,567 108,764

Source: IDA Planning Division, Dublin.

Table 3.11 IDA Grant per Job Sustained, 1981–2001 (in Euros, Constant 2001 prices) Years 1981–1987 1982–1988 1983–1989 1984–1990 1985–1991 1986–1992 1987–1993 1988–1994 1989–1995 1990–1996 1991–1997 1992–1998 1993–1999 1994–2000 1995–2001

IDA grant 35,341 30,972 26,194 23,191 22,167 20,743 19,309 16,879 15,766 15,592 15,281 15,472 14,417 13,521 13,375

Source: IDA Planning Division, Dublin.

The Culliton Report, the next major assessment of industrial policy 10 years later, advocated a further reduction in the grants for foreign companies— “internationally mobile industry”—and more forceful support for cluster formation of indigenous companies (Culliton 1992). Girma, Goerg, and Strobl (2003) investigate the effectiveness of IDA grants through an analysis of their impact on plants’ longevity and

74

FOREIGN INVESTMENT, DEVELOPMENT, AND GLOBALIZATION

employment growth. They use firm level data for the period 1983 to 1998 and analyze the separate impact of employment grants, capital and technology grants, and other grants. The econometric results show that capital and technology grants had a significant impact on plants’ longevity, with exit hazards 50 percent smaller than in plants that did not receive grants. On average, these grants also increased the unconditional demand for labor by 16 percent. The results provide statistical support for the effectiveness of key IDA grants. We see in chapter 4, that grants by Enterprise Ireland for indigenously-owned plants were equally, if not more, significant, with respect to longevity as well as employment creation. Ireland is certainly not the only country, which uses grants in an attempt to sway TNCs toward investment in its economy. How important the grants are in influencing the location decision of TNCs varies from company to company. Ceteris paribus, that is if all the other locationspecific assets of countries are the same, grants can clearly play an important role. But then, other factors rarely are the same. At times, government officials think that their grants have to be higher to compensate for disadvantages vis-à-vis competing countries; and at other times they can be lower, because of relative strengths in key assets compared to other countries. An executive of a prominent TNC computer company in Ireland characterized the importance of tax incentives as follows: “Grants are important, once you start to do the numbers. You don’t get people to the table with grants. You need to have the other factors. But then when you don’t have grants at the table as well, they might go away again; but you don’t get investors just with grants.”46 Survey results reported in the Telesis Report indicate that grants affected the location decisions of foreign companies only marginally (Tomaney 1995, 102). But “marginally” should not be interpreted as “unimportant,” as the quotation above indicates. It is useful to think about TNC location decisions as a two-stage process. In the first stage, a TNC makes a shortlist of the countries with the right assets to meet its strategic needs. In the second stage, when it makes the final decision about location, the grant amount matters. The key question, of course, always is whether the same investment could have been attracted with a smaller grant. Even though the TNC’s bargaining power is likely to outstrip that of the small host country in nearly all circumstances, it does not necessarily mean that the SLC will lose. The critical question from the host country’s perspective is whether the benefits generated by the foreign investment exceed the costs. The IDA bases its grants on a cost-benefit analysis of each project and claims that it never funds a project where the net discounted benefits are negative.47 Tax revenues from foreign companies figure most prominently on the benefits side.

THE RISE OF THE CELTIC TIGER

75

As Ireland moved toward full employment in the late 1990s, employment creation was no longer a top priority in grant support. Rather, it became substantially more important to retain foreign corporations in Ireland and support/induce them to move up the value chain toward more technology and research-intensive parts of the production process. As a result, the IDA has been wrestling with how best to define criteria, measurability, and accountability for financial incentives in support of a move up the value chain.

Summary The main argument in this chapter was that the rise of the Celtic Tiger was due to the fortuitous conjuncture of TNCs’ strategic needs and Ireland’s location-specific assets. Ireland had the right location-specific assets to attract high-tech FDI at a moment when a change in high-tech TNCs’ strategic needs induced them to increase their direct investment in the EC. I have summarized these factors and their interactions in figure 3.1. The demand boom in the PC industry and the desire to be closer to the final consumer market drove TNCs to increase their investment in Western Europe. The interest to invest in the EEC was strengthened considerably by ‘Opportunity and Fortress Europe,’ which emerged from the Single European Act of 1986. It was in this context that Ireland had the right assets to offer: a cheap, skilled, English-speaking work force, appropriate infrastructure, macroeconomic and political stability, stable capital–labor relations, and very attractive financial incentives in the form of low tax rates and generous grants. The dramatic increase in EU structural funds, another implication of ‘Europe 1992,’ contributed significantly to the formation/expansion of Ireland’s location-specific assets, particularly through large-scale funding of education and infrastructure projects. It is impossible to separate out the relative contribution of each of the external factors and the location-specific assets to the FDI boom in Ireland. We cannot write the historical counterfactual and undertake a controlled experiment simulating what would have happened to FDI in Ireland without one or the other factor in place, not least because these factors interact in different ways generating cumulative results. Surveys conducted by Forfas (2002, 34) shed some light on the relative importance of Ireland’s assets. In 2001, 57 percent of TNCs surveyed listed “appropriate skills” among the most important factors that attracted them to Ireland, 47.5 percent pointed to the favorable tax regime, 3 percent to the English language, and 30 percent to “motivation and loyalty of the staff.” In

76

FOREIGN INVESTMENT, DEVELOPMENT, AND GLOBALIZATION

– PC boom – Proximity to markets – Following TNC clients – Cluster seeking

Global value chains: TNCs’ strategic needs

Ireland’s locationspecific assets

FDI in EC

Opportunity and fortress Europe

EC Development Funds

IDA & coordinated strategy, effective civil service

– Cheap, skilled, english-speaking work force – Adequate infrastructure – Political stability – Stable capital– labor relations – Grants and low tax rates

– High unemployment – Demographics – Rising female LFPR – Education policies – Training by TNCs in the past – Telecommunications – Air transport National Pay Agreements

Figure 3.1 High-Tech FDI in Ireland in the 1990s: Critical Contingencies.

the end, it is the cumulative interaction among some of the factors that matters for a particular foreign investment decision. The relative importance of each of the factors also varies by sector. Low taxes are especially important for the pharmaceutical industry, while the availability of a highly technically skilled labor force is particularly significant in the IT sector. A representative of Intel emphasized in an interview: “after you have invested your $1 billion, you are only as good as the engineers you can hire, and that really means high-tech engineers.” But then Intel may not be representative of the sector, since a core part of its value chain is located in Ireland, in contrast to other IT companies. Once Ireland was elevated to the status of a first-tier investment location in the eyes of potential investors, a dynamic was set in motion that

THE RISE OF THE CELTIC TIGER

77

attracted further foreign investment. Some companies invested because they were following their client TNCs who demanded colocation, while others were attracted by the agglomeration economies of growing clusters. The IDA was very proactive in targeting and pursuing high-tech TNCs and providing continued support once they were in Ireland. Their determination, resources, and savvy were important factors in the matchmaking between TNCs’ needs and Ireland’s assets. Indeed, the coherence and effectiveness of the institutional structure behind the pursuit of an FDI strategy and the retention of high-tech FDI were decisive for the rise of the Celtic Tiger.

Is the Celtic Tiger Dead? The late 1980s constituted a critical juncture for Ireland, with specific external and internal forces coming together to give rise to the Celtic Tiger. In the course of the 1990s, the very tiger unleashed a dynamic, which has transformed the location-specific advantages Ireland held in the early 1990s. And that transformation in conjunction with a change in the global environment for foreign investment has led to a new critical juncture for Ireland at the beginning of the twenty-first century. Some of the important location-specific advantages have ceased to be advantages or are threatening to disappear (see figure 3.2). Skilled labor, which used to be relatively cheap, has become substantially more expensive in comparison with other countries competing for the production of standardized high-tech TNC production. Increases in real wages are a reflection of the country’s development achievements, and productivity has risen along with wage increases. As a result, unit labor costs have actually fallen in the 1990s. In 2000, average unit labor costs in the manufacturing sector were 30 percent lower than in 1995 (Department of Enterprise, Trade, and Employment 2003, 57). The fact that Ireland is no longer able to compete with the rest of the world in the production of standardized electronics products in spite of the considerable drop in unit labor costs is the result of the rise of competitor countries vying for FDI. New low-cost latecomers have emerged full force on the global scene presenting attractive alternatives for high-tech FDI. The most important ones are China and India, where very low wages for unskilled as well as technically skilled workers together with the possibility for large-scale production offer an attractive set of locationspecific assets for high-tech foreign investment. Another set of new

78

FOREIGN INVESTMENT, DEVELOPMENT, AND GLOBALIZATION Early 1990s

Early 2000s

Cheap, skilled, English-speaking work force

unemployment disappears

More expensive, skilled, English-speaking workforce

Adequate infrastructure

growth outpaces infrastructure

Infra structural deficiencies

Political stability

uneven distribution of benefits

Challenge to existing social contract

Stable capital–labor relations

inflation

Threats to continuation of National Pay Agreements

Growth and low tax rates

EU implications

Lower grants and low tax rates under continued pressure

Figure 3.2 The Impact of the Celtic Tiger on Ireland’s Location-Specific Advantages.

competitors, which are equally important in the European context, are the Central European countries, especially the ones that joined the EU in May 2004: the Czech Republic, Hungary, Latvia, Lithuania, Poland, the Slovak Republic, and Slovenia. Several of these countries have become very attractive foreign investment sites exhibiting characteristics not unlike the ones Ireland had in the late 1980s: a cheap, well-educated labor force, many of whom speak English; proximity to major markets, political stability, docile labor unions, low tax rates, and soon an inflow of European structural funds to enhance infrastructure and human capital further. Where Ireland’s infrastructure was adequate to attract high-tech FDI in the early 1990s, it is now showing considerable deficiencies. Infrastructure upgrading has not kept pace with the rapid growth of the 1990s. A congested road system, the absence of a low-cost national broadband system, and high electricity costs are among the major obstacles, which need to be

THE RISE OF THE CELTIC TIGER

79

overcome for Ireland to continue to be a competitive location for doing business.In 2003, the Minister for Enterprise, Trade, and Employment commissioned a study on the challenges Ireland needs to address to be successful in an increasingly global economy. The commission chaired by Eoin O’Driscoll released its report in 2004. The Enterprise Strategy Group (2004, 5) recommends “a systematic programme to lower the cost and improve the quality of infrastructure and public utilities—telecommunications, energy, roads and ports.”48 The need for larger investments in Ireland’s infrastructure comes at a time when funding support from the EU is declining sharply. As Ireland’s GNP per capita converged on and then surpassed the average GNP per capita of EU member countries, structural funds from Brussels have become smaller. During the first three years of CSF III (2000–2006), Ireland received €1.7 billion in cofinancing of the national development plan (ESRI 2003).49 That is still a large injection of outside funds for development projects. But it amounts to “only” 0.6 percent of GNP, a far cry from the 2.5 to 4 percent range of the 1990s. With the enlargement of the EU eastwards and Ireland’s increase in GDP per capita relative to the EU average, EU funding after 2006 will become much smaller and disappear altogether before too long. Political stability, another asset of Ireland in the 1990s, continues to be strong at the beginning of the twenty-first century. Yet, the uneven distribution of the fruits of the Celtic Tiger and the inadequate provision of key public services, especially health and low income housing, raise the possibility of greater political discontent in the future. While real incomes have risen across income groups, relative income poverty has increased. If the 60 percent threshold of median household income in 1994 is indexed to inflation, we find that the percentage of people below that threshold declined from 16 percent in 1994 to 2 percent in 2001. But if we look at the actual 60 percent income threshold in each year, the percentage of the population below that threshold rose from 16 percent to 22 percent (Layte, Nolan, Whelan 2004, 5–6). Ireland’s distribution of income remains one of the most unequal among EU countries. In 2001, the ratio of the income of the top 20 percent to the bottom 20 percent of income earners was 4.5 in Ireland, compared to 3.1 in Denmark, one of the lowest in the EU (European Commission 2004, 33). Since the establishment of the National Pay Agreements in 1987, capital–labor relations have been peaceful overall, and wage increases have been fairly moderate in exchange for tax reductions. The tri-partite agreement came under strain during the last round of negotiations in 2003–2004, as inflation had been rising faster than the agreed upon nominal wage increases, and the possibility for income tax reductions had become

80

FOREIGN INVESTMENT, DEVELOPMENT, AND GLOBALIZATION

smaller. While the consumer price index increased between 1.5 to 2.5 percent per year during 1993 to 1999, it rose at twice that rate between 2000 and 2002. In the end, the partnership survived in 2003, with moderate wage increases under the agreement Sustaining Progress. Whether the partnership agreements will survive the fallout of the structural changes in the Irish economy during the coming years remains to be seen. Financial incentives were among the most important location-specific assets Ireland had to offer to foreign investors. The grant element of these incentives will change in the future, directed more toward higher valueadded activities and less toward employment generation and investment amounts. But the total amount will probably have to decline in the future, not least due to reduced funding from Brussels. Low profit tax rates of 12.5 percent, on the other hand, will persist and remain a very strong attraction for foreign investors, especially if they generate high value added products. While European Union members will increase the pressure on Ireland to raise tax rates, the Irish government is likely to defend them at just about any cost. The changes in some of the location-specific advantages, which made the Celtic Tiger possible, have serious implications for Ireland’s future development. They are all the more serious because of the changed external context in which they are taking place. There have been crucial changes in external factors over which Ireland has no control, but which have profound repercussions for Ireland’s future growth and development (see figure 3.1). I already highlighted the increase in important competitor countries vying for foreign investment, especially high-tech foreign investment. Another important change in the external environment results from the developments in the electronics industry over the last years. The collapse of the technology bubble in 2000 and the slowdown in global economic growth have led to a radical restructuring of the industry, with TNCs in aggressive pursuit of the lowest-cost production locations around the globe. The confluence of these internal and external changes has spelled the departure from Ireland of a number of TNCs in the electronics industry. Employment in the manufacturing sector fell from a high of 253,200 in March 2001 to a low of 227,400 in June 2003 (CSO 2003c). Two thirds of that decline was due to a reduction in employment in the optical and electrical equipment industries (NACE 30–33). The major transnational contract manufacturers who had entered Ireland in the second half of the 1990s had left the country again by the end of 2003. What does all this mean for Ireland’s future? Ireland will not be able to grow at the same rate as in the 1990s, simply because it has full-employment now. Where the supply of labor grew at an annual rate of

THE RISE OF THE CELTIC TIGER

81

3.4 percent during 1995–2000 and 2.2 percent over the period 2000–2005, it is projected to increase at a rate of 1.7 percent during 2005–2010 and 0.7 percent over the period 2010–2015 (Department of Enterprise, Trade, and Employment 2003, 52). In the future, the elasticity of the labor supply will depend greatly on the country’s emigration policies. But while emigration can compensate to some extent for the changing demographics, which have made the supply of indigenous labor considerably less elastic, it will likely be limited by cultural and political factors rather than by economic factors.50 Labor supply growth can and will not be a major factor behind future economic growth in Ireland. Ireland also does not need to attract the same amount of FDI as in the 1990s. It needs to attract enough investment to replace the relocation of production out of Ireland, whether because of TNCs moving to Eastern Europe, or Irish companies using offshore outsourcing to lower wage countries to remain competitive. To achieve a new path of sustained growth in the twenty-first century Ireland needs a development strategy based on productivity growth. And that means a strategy where companies, foreign and indigenous, move up the value chain, do not produce standardized products, and generate more R&D-related activities. It is a testimony to the Irish government’s proactive approach to development policies that it ordered a science and technology foresight study in 1998, at a time when the economy was booming and huge amounts of FDI were entering the country. The study recommended a focus on R&D in ICT and biotechnology as the basis for the next phase of growth and development in Ireland. The implementation of that recommendation commenced in 2000 with the establishment of Science Foundation Ireland (SFI) with a budget of €650 million. In light of the new challenges Ireland faces at this critical juncture, it is uncertain whether the country can make the transition to R&D-driven growth and development. A reemergence of the Celtic Tiger depends on the country’s ability to overcome the problems with its location-specific assets. I will not address these problems here further. Rather I will focus only on one aspect of the possible success of the new strategy, where economic growth is based primarily on productivity growth rather than factor growth. And that is the expansion of knowledge-based assets in Ireland. Ultimately, sustained productivity growth depends on the state of the country’s indigenous knowledge-based assets and on the willingness of TNCs to move production in Ireland up the value chain into the high-end of high-tech production. Those are, of course, the very issues which are at the heart of this book. What has been the impact of high-tech FDI on the

82

FOREIGN INVESTMENT, DEVELOPMENT, AND GLOBALIZATION

development of the host country’s knowledge-based assets? And, in a dynamic context, does a high-tech TNC move up the value chain in its production of high-tech products in an SLC or does it relocate when conditions change? These are the questions I investigate in the next chapter for the case of Ireland.

4

The Development of Knowledge-Based Assets in Ireland

or many countries, a change in the structure of manufacturing production and manufactured exports toward high-tech industries and products reflects a substantial advancement of the country’s knowledgebased assets. Such an inference is, however, not warranted for an SLC like Ireland. The structural change in Irish manufacturing, which we saw in the previous chapter, has been driven primarily by foreign investors and not by Irish companies. If foreign companies in the high-tech sectors were to leave tomorrow, as some have been doing in the electronics industry, how would Irish manufacturing look like? The answer depends critically on how knowledge-based assets in Ireland developed during the 1990s, on whether the linkages that were formed with indigenous companies generated learning processes that will allow indigenous producers to compete globally and to move increasingly up the value chain, into the more R&D intensive parts of the value chain. An SLC like Ireland, with its small internal market and need to specialize, will not be able to depend only on its own knowledge-based assets for future growth and development. Nor is it desirable in this age of globalization, where technology spillovers and knowledge sharing through international trade and investment provide possibilities for leapfrogging. In the end, sustainable growth and development have to be based on a positive and dynamic symbiosis between indigenous and foreign knowledge-based assets. The expansion of indigenous knowledge-based assets is important, since an advancement of indigenous technological capabilities reduces the dependence on TNCs, increases the possibility for some reliance on indigenously-sustainable growth, and supports firms’ ability to adjust in the continuously changing global market. But in a dynamic context, it is

F

84

FOREIGN INVESTMENT, DEVELOPMENT, AND GLOBALIZATION

equally important, that TNC affiliates move up the value chain and carry out more research-intensive parts of the value chain in Ireland rather than in their home countries, a dynamic that is intimately linked to the on-going upgrading of indigenous capabilities. What makes high-tech FDI such a potentially potent impetus for industrial upgrading, and maybe even leapfrogging, is the possible development of indigenous technological capability through linkages and human capital spillovers. The analysis in this chapter shows that, in Ireland, technology transfer happened through both channels. Nonetheless, transfers did not happen automatically, but needed the impetus and support of proactive government policies. Whether or not the development of the indigenous learning processes was sufficient or not to contribute to the success of Ireland’s next development phase remains to be seen.

Linkages with Indigenous Firms The available studies and data indicate that high-tech FDI in Ireland generated linkages with indigenous firms and that a number of Irish companies indeed followed a supplier-based upgrading model. Yet, the overall effect is smaller than suggested by some of the official data, and there do not seem to have been additional extensive linkages in the 1990s, the very decade when high-tech FDI took off in Ireland. There are also indications that the willingness of TNCs to help indigenous producers become competitive input suppliers has declined, while indigenous capabilities have grown; the room for indigenous sourcing of standardized products has declined, while the potential for niche production has increased.

Empirical Evidence Linkages at the Aggregate Level A common way to gauge the extent of backward linkages from foreign to indigenous companies is to analyze the sourcing behavior of TNCs over time. If TNCs increase their purchases of material inputs and services from indigenous companies and thus reduce their imports accordingly, then linkages with domestic companies have grown. I will analyze the extent of foreign-owned enterprises’ (FOEs) local sourcing in absolute terms, that is the absolute amount of Euros spent on domestic inputs, and in relative terms, that is domestically sourced materials and services as a share of total inputs. If domestic sourcing increases in absolute terms, I will call it an

DEVELOPMENT OF KNOWLEDGE-BASED ASSETS

85

“extension of linkages.” And if it grows in relative terms, I call it an “intensification of linkages.” The two types of linkage developments do not have to move in the same direction. If extensive linkages grow at a slower rate than the increase in total input purchases, then intensive linkages will decline. Is one or the other linkage growth preferable from the vantage point of an SLC? Is one or the other a better indicator of FOEs’ embeddedness in the domestic economy? By definition, an intensification of linkages means that FOEs source more in the host country for a given amount of total inputs and thus for a given amount of output. To the extent that this implies that not only more, but also more sophisticated products are sourced domestically, an intensification of linkages would indicate a greater contribution to the development of knowledge-based assets. The implicit assumption is that if, over time, foreign corporations purchase more domestically produced inputs per unit of output, domestic companies must be developing competitiveness in parts and products that demand greater technological capabilities, or FOEs would have purchased the inputs before. Thus, I argue that intensification rather than an extension of linkages is more suggestive of an advancement of domestic knowledge-based assets. In Ireland, extensive linkages between TNCs and indigenous companies grew throughout the decade of the 1990s. However, intensive linkages increased mostly from the second half of the 1980s to the early 1990s. They didn’t grow any further under the Celtic Tiger, and actually declined toward the end of the 1990s. The two data sources from which we can extract information about linkage development in Ireland paint a rather different picture. They are the CIP and the Annual Business Survey of Economic Impact (ABSEI). The CIP data indicate considerably higher intensive and extensive FOE-generated linkages in manufacturing than the ABSEI data. They also suggest that FOEs intensified their local sourcing during the 1990s, in contrast to the ABSEI, which shows a decline in sourcing intensity during the second half of the decade. The CSO conducts and publishes the CIP, which includes information on inputs and imports of firms located in Ireland. It reports the results from the CSO questionnaire, which asks producers to indicate the value of materials purchased from other parties and to estimate the percentage that was imported (CSO 2003b, 164). Since the estimates of the import share seem to be flawed for the 1990s, I am relying here on the ABSEI data for an analysis of linkages.1 ABSEI is conducted and published by Forfas, the government’s policy and advisory board for Enterprise, Trade, Science, Technology, and

86

FOREIGN INVESTMENT, DEVELOPMENT, AND GLOBALIZATION

Innovation. ABSEI has been undertaken since 2000, when it became the successor survey to the Irish Economy Expenditure Survey (IEES).2 The time series in table 4.1 presents the ABSEI data together with recalculated grossed-up estimates of the former IEES. Some of the fluctuations in the series are due to the fact that the response rate fluctuated over the years, as companies were not required to fill out the IEES questionnaire. While that makes the figures for any particular year somewhat less representative, it should not matter as much with respect to the trend over time, which smoothes out the ups and downs. In the ABSEI questionnaire companies are asked to specify their total expenditures on raw materials and to estimate the proportion produced in Ireland. The results indicate that TNCs in the non-food manufacturing sector increased their share of domestically produced inputs considerably during the second half of the 1980s. The share remained fairly stable at around 20 percent in the course of the 1990s, and then started to decline again in the late 1990s. The data only tells us the percentage of inputs FOEs sourced from companies in Ireland. But it does not differentiate between purchases from foreign-owned enterprises and purchases from Irish-owned companies. Table 4.1 Materials Produced in Ireland as a Percentage of All Materials Purchased by Foreign-Owned Enterprises (FOEs), 1985–2002

1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002

Manufacturing NACE 15–37

Non-food manufacturing NACE 17–37

Electrical & electronic equipment NACE 30–32

Medical instruments NACE 33

23.2 23.4 28.6 28.1 26.6 28.8 29.0 29.2 28.7 27.1 25.6 24.6 24.7 26.4 25.1 21.0 22.4 20.4

12.6 14.1 16.4 16.4 17.2 18.9 19.1 19.4 19.9 19.7 20.0 18.9 20.0 22.5 21.8 18.5 19.9 17.8

7.7 8.6 10.3 11.0 11.2 14.1 17.3 18.6 18.6 17.9 18.6 16.3 18.9 23.0 23.1 17.8 20.0 16.6

17.7 18.5 21.7 20.2 25.3 23.5 22.8 20.0 23.9 26.0 28.1 26.4 25.5 28.4 21.4 19.1 17.4 18.8

Source: Forfas 2003a.

DEVELOPMENT OF KNOWLEDGE-BASED ASSETS

87

Based on a Forfas study, Breathnach and Kelly (1999, 22) report that, in 1995, two thirds of locally sourced inputs in the electronics sector came from other foreign-owned companies. It is quite likely that the arrival of the major U.S. contract manufacturers in Ireland in the second half of the 1990s increased this figure even further. TNC affiliates often ask their foreign input providers to colocate with them because they want more control or just-in-time delivery. Geographical proximity allows them to make more frequent inspections, demand greater flexibility from suppliers with regard to last minute changes, and require shorter lead times. Services integral to the production process have to be provided on site anyway, of course. Other times, colocation of foreign input suppliers is completely voluntary, as they set up production close to a cluster of potential TNC clients, in the hope of winning new customers. Another complication in getting a precise handle on the extent of the indigenous sub-supply sector is the fact that ownership changes over time. For example, a transnational corporation acquired the Irish-owned company, which pioneered the concept of carotid stenting—a showcase of indigenous research capability. A very successful Irish producer of medical devices told me in an interview that he had been approached three times already by a foreign company offering to buy him out. Breathnach and Kelly (1999, 22) claim that during 1994–1996 “many of the better sub-supply companies were taken over by multinational component manufacturers seeking a presence in Ireland in order to supply the substantial MNC presence in the country. Approximately 10 Irish-owned manufacturing and international services companies per year are sold to multinational corporations.” On the other hand, there is anecdotal evidence of Irish companies that got started through a managed buy-out of a TNC affiliate, primarily in cases where the latter had decided to close operations in Ireland. To summarize: FOEs in the non-food manufacturing sector intensified their linkages with domestic companies substantially in the second half of the 1980s, kept them at a steady level throughout the 1990s and reduced them again in the late 1990s. However, in the high-tech industries the intensification of linkages kept growing during most of the 1990s— though at a slower rate than in the 1980s—before declining in the late 1990s. But even though intensive linkage creation declined in the late 1990s, extensive linkages kept growing until 2001. FOUs increased their purchases of material inputs from €376 million in 1985 to €2.7 billion in 2001. Purchases of domestic services were even higher. They grew from €710 billion in 1985 to €3.8 billion in 2002. A substantial share of the domestically sourced goods and services were purchased from foreign-owned input suppliers located in Ireland, rather

88

FOREIGN INVESTMENT, DEVELOPMENT, AND GLOBALIZATION

than from Irish-owned input suppliers. As a result, the development of intensive and extensive linkages with Irish companies has been considerably smaller than suggested by the ABSEI data. In 1996, 174 of the Irish manufacturing companies associated with the national linkage program were sub-suppliers to TNCs. That corresponds to 6.5 percent of all Irish companies with more than ten employees, or 15.5 percent, when we only include firms in those sectors where there actually was sub-supply to TNCs.3 While the importance of sub-suppliers was undoubtedly larger in value terms, the percentages indicate that in the mid-1990s, linkage development with indigenous companies was still limited. Whether TNC affiliates source from Irish-owned companies or from other foreign enterprises located in Ireland does not make a difference with respect to employment, since the overwhelming majority of employees in all manufacturing companies, irrespective of ownership, are Irish. It makes a difference though with respect to the expansion of indigenous knowledge-based assets via backward linkages, if the TNC sub-suppliers— like their TNC clients—are less likely to source domestically. The ABSEI data show that indigenously-owned enterprises (IOEs) buy a much greater share of their inputs domestically (see table 4.2). Where FOEs in the chemical sector bought 15 percent of their material inputs domestically in 2002, IOEs sourced 58 percent. In the electronics industries, the respective shares were 17 and 34 percent and in medical devices 19 versus 45 percent. In absolute terms, the story is somewhat different though. Compared to their foreign counterparts, IOEs spent 70 percent more on local material inputs in 2002, but 56 percent less on local services. Differences in Domestic Sourcing Across Subsectors In Ireland, TNC affiliates across industrial sectors are similar in their domestic sourcing behavior in that they make considerable use of domestic packaging and printed materials like labels. The same holds true for services, which range from cleaning to maintenance to cafeteria services to legal and accounting services. Beyond that, however, the substance of domestic sourcing differs by sub-sector. In the chemical and pharmaceutical sector (NACE 24), FOEs decreased their domestic sourcing share during the 1990s. The ABSEI/IEES survey data show that after remaining at around 20 percent between 1985 and 1997, the domestic sourcing share in NACE 24 had dropped to 15 percent by 2000 (and 2002). Nonetheless, extensive linkages kept increasing, with expenditures on material inputs more than doubling between 1997 and 2002 and expenditures on services more than tripling. The possibilities for sourcing material inputs from indigenous companies are limited though, as TNCs are engaged in bulk production and are sourcing in bulk globally (e.g. Pfizer’s production of Viagra and Allergan’s production of Botox).

DEVELOPMENT OF KNOWLEDGE-BASED ASSETS

89

Table 4.2 Materials Produced in Ireland as a Percentage of All Materials Purchased by Indigenously-Owned Enterprises (IOEs), 1985–2002

1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002

Food/drink/ tobacco NACE 15–16

Non-food manufacturing NACE 17–37

Electrical & electronic equipment NACE 30–32

Medical instruments NACE 33

86.4 85.0 83.3 83.6 83.4 84.5 83.9 82.3 84.0 84.5 85.0 85.9 87.2 87.1 87.1 79.5 86.6 86.1

33.0 36.7 39.1 37.5 35.6 36.7 35.5 38.5 40.0 39.5 37.1 37.9 38.7 37.3 36.1 44.9 51.0 51.0

28.4 34.0 36.4 35.0 29.7 34.2 39.9 28.7 26.3 28.0 30.3 28.4 29.5 26.4 28.7 24.9 34.0 33.5

35.5 25.9 13.1 18.5 18.4 27.0 47.0 45.4 51.2 43.1 48.0 42.0 53.6 54.5 62.0 32.1 43.7 44.9

Source: Forfas 2003a.

Electronics is the sector where indigenous sub-supply to TNC affiliates has developed the most. According to Enterprise Ireland (2003b), there are around 100 Irish firms in the electronics sub-supply sector, making boxes and components of various kinds, plastic fabrication and different tools. A small number of Irish companies also provide capacity for overflow production and repair of electronic devices. Indigenous electronics companies sold €1 billion worth of output in 2000, 42 percent of which was sub-supply to TNCs. The electronics sub-supply sector accounted for 57 percent of the total IOE employment of 8,575 in the sector (Lyons 2002a). Most of the companies use medium level technology, and a small number of companies seem to have made the jump from production to design. Since the beginning of the twenty-first century, the electronics industry in Ireland (and globally) has been undergoing a process of restructuring, consolidation, and relocation of standardized components to low-wage cost countries. In that context, the indigenous electronics companies, particularly in the sub-supply sector, are faced with the challenge of finding a new niche for themselves, as some of their transnational clients have

90

FOREIGN INVESTMENT, DEVELOPMENT, AND GLOBALIZATION

moved abroad. Some have made the switch from producing inputs for electronics corporations to inputs for medical instrument corporations. Others do not have the wherewithal to survive such a restructuring process, and still others who have successfully hooked into the global value chain network of their TNC clients are moving with them to new locations. There are currently several Irish-owned companies who, supported by Enterprise Ireland, have established manufacturing sites in Central European countries close to the TNC affiliate they used to supply in Ireland. At this juncture, it is too early to assess the relative importance of these different response channels. The software manual printing industry, though technologically less sophisticated, stands as an example of the difficulties faced by sub-suppliers, when the industry of the TNC customer base undergoes changes. The industry arose as a sub-supply industry to TNCs locating in Ireland. Between 1987 and 1992, it grew at an annual rate of around 30 percent (Jacobson and Mottiar 1999, 438). In the course of the 1990s, however, software manuals were increasingly being replaced by manuals on the CDRom itself or by on-line delivery, forcing firms in the industry to adjust. Jacobson and Mottiar (1999, 440) recount the success story of Mount Salus, which switched to the printing of small CD booklets that fit into the plastic covers. However, they also suggest that Mount Salus has been the exception rather than the rule. The emergence of sub-suppliers in the medical-devices industry is a more recent and incipient phenomenon. Although several of the top TNCs in the field have been in Ireland since the 1970s (e.g. Abbot and Baxter), the large increase in the output of medical devices of different kinds is primarily a phenomenon of the 1990s. The medical-device and diagnostics industry has developed in clusters, most prominently around the Galway area in the west of Ireland, and around Dublin in the east. A small, but growing number of Irish-owned companies have developed in these clusters, producing medical equipment, components for medical devices, diagnostics and different types of interventional products. The requisite technological sophistication is quite high, and a number of Irish-owned companies have been active in design and product innovation. In 2002, IOEs in the medical-device industry exported 66 percent of their total sales of €180 million (Forfas 2003a). The software industry is another sector whose considerable growth is at least partially due to TNC presence and demand in Ireland. According to O’Riain (1999, 21), 35 percent of Irish software companies had contracting alliances with international firms in the first half of the 1990s. FOEs in software development reduced the share of their domestic sourcing, of both materials and services, quite substantially during the second half of

DEVELOPMENT OF KNOWLEDGE-BASED ASSETS

91

the 1990s. Nonetheless, the sales of software IOEs grew nearly ten-fold during that period, an indication that they were successful in their own right.4 Iona is the most well-known example of homegrown success in the software industry. Founded in 1991 by members of the computer science department at Trinity College, Dublin, Iona provides sophisticated web services software linking different computer systems together. Iona is listed on the NASDAQ, and in 2002, had sales of $123 million.5 Channels of Knowledge Spillovers through Linkages The available evidence suggests that FOEs in Ireland’s high-tech sectors generated positive spillovers through linkages. The use of best practice management and production techniques, for example, was widespread among TNCs, but much less common for Irish firms. All indications are that the gap between FOEs and IOEs that is shown in table 4.3 for 1999 was larger at the beginning of the decade, and that positive spillovers from TNCs contributed to the closing of that gap. One cannot become a subsupplier of a TNC without learning about best-practice logistics, supply chain management techniques, quality standards, and processing efficiencies. Some of these spillovers will have happened through the learning by doing of everyday business, and others through deliberate support from the TNC customers. In the second half of the 1980s and the early 1990s, TNCs were quite willing and eager to work closely with domestic input suppliers, especially in the electronics sector. The National Linkage Program initiated in 1985 had a lot to do with the promotion of such cooperative behavior, as did competitive global pressures. Systematic evidence on the channels of technology transfer through backward linkages is scarce. A study by Breathnach and Kelly (1999) indicates that the technological capabilities of Irish companies were enhanced through supply chain linkages. Analyzing the innovative capability of Irish indigenous firms during 1994–1996, they did not find a statistically significant difference for a number of indicators of innovative performance Table 4.3 Percentage of Manufacturing Plants Using Selected Management and Control Systems in Ireland, 1999

Total quality control management Just in time delivery Computer integrated manufacturing Computer aided manufacturing Computer aided design Source: Hewitt-Dundas et al. (2002, 47).

TNCs

All plants

89.0 64.0 67.0 57.0 75.0

30.5 30.1 20.1 37.9 55.3

92

FOREIGN INVESTMENT, DEVELOPMENT, AND GLOBALIZATION

between Irish sub-suppliers and Irish non-sub-suppliers. For other indicators there was a difference though. Sub-suppliers of TNCs were more innovative as measured by patent applications, learned a lot through joint product development, were engaged much more intensively in design, and operated in a more competitive environment. Although the sample of the study is limited—41 sub-suppliers and 112 non-sub-suppliers—the results suggest that TNCs contributed to the advancement of the technological capabilities of their Irish suppliers, that linkages served as a conduit for technology transfer.6 Based on extensive interviews with 61 large TNCs in Ireland in 2000, Hewitt-Dundas et al. (2002) provide detailed information on the channels and extent to which foreign companies have been transferring knowledge to domestic suppliers (see table 4.4).7 Of the respondents, 66 percent had monthly contact with local input suppliers on technical issues and 58 percent helped with the introduction of quality control systems. In many ways, TNCs’ behavior toward local suppliers was not very different from their behavior toward nonlocal suppliers. An exception to this similar treatment of suppliers is “collaboration on product development,” which is considerably higher for nonlocal suppliers. Nonetheless, nearly 80 percent of the TNCs interviewed collaborated with their local suppliers Table 4.4 Channels and Extent of Knowledge Transfer between TNCs and Suppliers in Ireland Proportion of TNC plants engaged in knowledge-transfer activity with local and non-local suppliers Local suppliers

Non-local suppliers

20.5

17.8

65.9

51.1

12.5

11.2

45.0

47.7

78.8

91.7

62.9

59.3

58.2

49.0

64.4

71.1

Weekly contact on technical issues Monthly contact on technical issues Monthly feedback on performance Quarterly feedback on performance Collaboration on product development Audit suppliers’ manufacturing Assist with introduction of quality control systems Information on other business opportunities Source: Hewitt-Dundas et al. (2002, 66).

DEVELOPMENT OF KNOWLEDGE-BASED ASSETS

93

on product development, an important indication that technology transfer was not just limited to process development. Product development requires a different set of technological skills and reflects a company’s ability to move up the value chain. Since the Hewitt-Dundas study does not distinguish between Irish suppliers and foreign suppliers located in Ireland, its results do not provide direct evidence for technology transfer to indigenous companies through domestic sourcing. Nonetheless, since the results are compatible with the insights from the Breathnach and Kelly study, we have no reason to believe that they would not apply to indigenous companies as well.

Domestic Linkage Capability and Global TNC Strategies: A Pas de Deux The analysis above has shown that in the course of the 1990s intensive and extensive linkages between high-tech TNCs and local input producers grew considerably in the electronics industry, while intensive linkages did not grow for the manufacturing sector as a whole. FDI in the electrical and optical industries was a key driver of the economic boom of the 1990s. Thus, we have the greatest extension and intensification of linkages in the high-tech sector where FOE employment expanded the most, doubling in the course of the 1990s to constitute 27 percent of total manufacturing employment in 2000.8 Market forces by themselves did not bring about the desired outcome. The mere presence of TNCs in high-tech areas is a necessary but not a sufficient condition for linkage formation. Backward linkages only expand when there is a coincidence of domestic linkage capability and TNC interest in sourcing domestically, when both come together at the right moment, in the right way. In the case of Ireland, domestic linkage capability was low in the 1970s and 1980s. It started to improve considerably with the National Linkage Program of 1985, when the government decided on a deliberate program to help small potential Irish input suppliers overcome the myriad of market failures they were facing. One of the reasons the program succeeded to the extent it did was that TNCs saw it in their interest at the time to cooperate. However, as the strategic needs of TNCs changed over time, the meaning of requisite linkage capability changed as well. And if domestic input suppliers do not expand their ability to meet those new needs, prior linkages might deteriorate and disappear. In the course of the 1990s, many TNCs changed the organization of their value chains. In the electronics industry, increased reliance on

94

FOREIGN INVESTMENT, DEVELOPMENT, AND GLOBALIZATION

contract manufacturers and on regional or global suppliers together with a greater demand for shorter lead times and just-in-time delivery began to militate against sourcing from suppliers who did not have global or regional reach. As a result, linkage capability now often requires the availability of a regional or global presence, something that only few Irish input suppliers have achieved. The question is whether technology-intensive niche production can provide enough of an alternative that will allow Irish sub-suppliers and producers in high-tech areas to survive and thrive in the future. The Need for Proactive Linkage Formation Policies We saw in the last chapter that the Irish institutional structure played a critical role in the attraction of potential foreign investors. However, the IDA’s aggressive pursuit of FDI stands in stark contrast to the scant attention it paid to helping indigenous companies become competitive input suppliers. In 1982, the Telesis Report commissioned by the Irish government to assess its industrial policy highlighted the weak links between foreignowned and indigenously-owned manufacturing companies. It criticized the government for paying insufficient attention to the development of a competitive indigenous sector. In the aftermath of the Telesis Report, the National Linkage Progam (NLP) was established in the mid-1980s. The next major evaluation of industrial policy, the Culliton Report of 1992, still criticized the government’s insufficient attention to the indigenous manufacturing sector after claiming that “it is only a mild exaggeration to say that most of the new foreign firms operate here as essentially an industrial enclave” (Culliton 1992 quoted in Tomaney 1995, 107). The report highlighted infrastructure, education, and training as important areas of noncompetitiveness (e.g. O’Sullivan 1995, 372). The Report’s recommendations led to a restructuring of the institutional support structure for linkage development with the establishment of Enterprise Ireland (EI) in 1994. EI was charged with the promotion of competitiveness of the indigenous manufacturing sector, while the IDA’s responsibility was restricted to the pursuit and cultivation of foreign corporations.9 For those close to the action it must have been clear early on that TNC affiliates were not developing extensive backward linkages with indigenous companies. Why then was the Irish government so disinterested in the active promotion of linkages and industrial upgrading until the mid1980s, when proactive policies vis-à-vis TNCs had already been the rule for 25 years?10 The answers suggested in my interviews with different past and current members of IDA and Enterprise Ireland center primarily on political factors. The FDI strategy was initially motivated and driven by the keen desire to address the problem of high unemployment in Ireland.

DEVELOPMENT OF KNOWLEDGE-BASED ASSETS

95

Since there was significant skepticism of large IDA grants for TNCs, highlevel authorities felt under pressure to perform. And performance was defined as increasing the number of jobs for Irish workers. It was considered both politically more advantageous and strategically easier to get 1,000 new jobs through new foreign investment than through increased competitiveness of indigenous companies. It was politically more advantageous for government officials, since the ribbon cutting ceremony at the opening of a new foreign-owned plant with 1,000 new jobs carries a substantially higher public relations value than the opening of many small electronics firms, which may have created 10, 20, or 50 jobs at a time. And it was strategically easier, since the understanding of how to promote linkages and a competitive indigenous sector was (and is) much more cumbersome and difficult. Linkage promotion is painstaking, nittygritty, and nonglamorous work, and returns may take a fairly long time to materialize. Two other factors played a role in the initial hands-off approach to the promotion of competitiveness of indigenous producers. One was the perception that domestic capitalists had squandered the chance provided by tariff protection under ISI. After all, between 1932 and 1958, foreigners were not allowed to have majority ownership in productive enterprises, and Irish producers were protected by fairly high tariffs. While economic growth was quite considerable in the 1940s, it was negligible in the 1950s. That was all the more noticeable, since the rest of Western Europe was experiencing high growth rates after World War II. A second and related reason may have been that the generally free market approach, which the country pursued, was seen as incompatible with government involvement at the micro level in support of specific enterprises. After all, increased competition from imports and foreign producers was supposed to force indigenous companies to become competitive. At times, history does repeat itself. Thirty years after Ireland embarked on free-market policies, a lot of developing countries followed suit. Under the double pressures of the exhaustion of the old ISI model and the foreign debt burden, the International Monetary Fund and the World Bank were in a very powerful position to push market liberalization as a development strategy in the 1980s and 1990s. For the developing countries in point, support for indigenous companies was not part of the model either. And the results have not been encouraging, as we see in the case of Costa Rica. It was only with time that policy-makers realized that foreign investment did not automatically generate all the benefits that were expected for Ireland. O’Riain (1999, 15) argues, “it took the massive social and economic crisis of the 1980s to de-legitimize the IDA’s role as the sole bearer of the task of Irish industrial transformation. It was into this restricted

96

FOREIGN INVESTMENT, DEVELOPMENT, AND GLOBALIZATION

institutional space that the alliance of Irish technical professionals and the previously marginalized ‘science and technology’ state agencies stepped to support indigenous industry.” And O’Malley (1998, 267) suggests that the switch in official industrial policy rhetoric in the White Paper on Industrial Policy in 1984 stemmed from “a recognition that there were limits to the benefits that could be expected from foreign investment and that the relatively poor performance of indigenous industry called for a greater focus on addressing that problem.” The change in policy manifested itself in the establishment of the National Linkage Program (NLP) in 1985. Set up outside the structure of the IDA, the NLP was charged with identifying opportunities with TNCs, locating potential indigenous suppliers and helping them become competitive input suppliers, that is, able to meet TNCs’ criteria in terms of quality, on-time delivery, and cost.11 The NLP started out with some twenty odd people from different institutions and recent work experience with a TNC. They were brought together under the leadership of Bill Chambers. As an Irish industrialist and the spokesperson of the Irish electronics industry, Chambers carried considerable cache with the CEOs of foreign affiliates in Ireland. The newly formed team audited all Irish companies with some engineering capabilities, “anybody who knew how to put some wires together,” and thus benchmarked the indigenous sector. The enormous benchmarking effort revealed that potential input suppliers existed, but that their competitiveness had to be created. The members of the NLP started with 60 of the roughly 300 audited companies and provided a wide range of support services to help them become competitive. They were engaged in capability building consulting on everything from quality standards (e.g. ISO 9000), customer expectations in a vendor audit, financial support, technological requirements, and logistics. TNCs assisted as well, providing technical advice and solving problems more generally. Between 1985 and 1987, 250 TNCs became actively involved with the NLP (Lyons 2002b, 2). And TNCs in the electronics sector doubled their share of material sourcing from domestically based companies between 1985 and 1990 (see table 4.1). During the 1990s, Enterprise Ireland broadened the focus beyond linkage building to more generalized capability building for indigenous producers. The institution offered an array of programs (grants, loans, consulting, equity involvement) to help start-up companies (especially in high-tech areas) and support existing companies to become and remain competitive (see table 4.4). In 2002, EI’s client companies employed 144,000 people and exported goods worth €10.7 billion, 42 percent of their total sales. Software, international services, and industrial products (e.g. electronics) accounted for nearly 30 percent of these exports

DEVELOPMENT OF KNOWLEDGE-BASED ASSETS

97

(Enterprise Ireland 2003a), indicating that indigenous producers had come a long way since the inception of the NLP in 1985. They achieved competitiveness in areas in which Irish companies had not been active 10 years earlier, a clear reflection of advancement in indigenous technological capabilities. EI’s goal to serve as matchmaker and information disseminator is demonstrated on its website. Anybody interested in buying particular goods or services from an Irish company can access the database of EI client companies online and search by industrial category for companies that may meet their needs. We do not have the information for a detailed cost-benefit analysis of EI’s expenditures, but we have some indications of costs and benefits. EI’s cost per job created and sustained by its client companies over the period 1996–2002 was €9,846. That was about half of what it had been in the first half of the 1990s (see table 4.5), and less than the analogous figure for the IDA. The main benefit is the advancement of the competitiveness of Irish companies and, with that, of their technological capabilities and their abilities to upgrade production in the future. Girma, Goerg, and Strobl (2003) approach the issue of EI grant effectiveness by investigating the impact of grants on firms’ longevity and employment creation. Using plant level data for 1983–1998, they find that grants for capital, technology, and research had a significant impact on both plant survival and employment. Plants that received capital, technology, and research grants had exit hazards of only about 37 percent of those plants that did not receive grants, and increased their demand for labor by 11.7 percent.12 Table 4.5 Cost per Job Sustained in Companies Supported by IDA and EI*, 1987–2002 (Constant 2002 Prices)

1987–1993 1988–1994 1989–1995 1990–1996 1991–1997 1992–1998 1993–1999 1994–2000 1995–2001 1996–2002

EI-supported companies

IDA-supported companies

n.a. n.a. 14,858 15,200 14,385 15,116 14,522 10,777 9,408 9,846

25,417 20,625 18,960 18,428 17,965 17,166 15,124 14,076 14,017 15,897

* Only jobs created during the period and sustained at the end of the period are included. Source: IDA (2003) for IDA data, EI (2003) for EI data.

98

FOREIGN INVESTMENT, DEVELOPMENT, AND GLOBALIZATION

Imperfect information and creative market failures are the two main reasons why targeted government action is needed to foster linkage formation. On the one hand, TNCs did not have enough information about the existence of potential suppliers and their potential capabilities. Thus, imperfect information led to selection bias and reliance on previous known suppliers. On the other hand, many potential indigenous suppliers did not have the necessary technological and managerial capabilities to become suppliers, or sufficient information to know how to upgrade their capabilities, or the resources needed to undertake a new, risky venture. The risk must have seemed all the more challenging, as many of the indigenous firms were small. Structural change during the 1970s had led to a decline in the number of larger indigenous firms in internationally traded goods and an increase in the number of small firms (O’Malley 1998, 261). Smaller firms have fewer resources for gathering requisite information of various kinds. And smaller firms are also less likely to have the capacity to produce a requisite quantity of a product. The Telesis Report of 1982 identified insufficient scale as one of the main reasons for the lack of linkage development in Ireland (Ruane 2001).13 The existence of these different barriers, which prevented many small potential indigenous producers from becoming actual input suppliers to TNCs, meant that the market could not fulfill its creative function among indigenous producers, that is to generate growth and technological change.14 The National Linkage Program, and subsequently Enterprise Ireland, addressed these market failures through the provision of information, different types of information for TNCs and indigenous companies, assistance on technological process, and marketing issues, and financial support. Matchmaking and support for start-ups and expansions of indigenous firms have been the main functions of Enterprise Ireland up to this day. While the general focus has remained on capability and capacity building, policies have changed, as the nature of the challenges facing indigenous companies has been changing. In recent years, EI—like the IDA—has placed major emphasis on the promotion of R&D and support for high potential, high-tech start-up companies. Within Enterprise Ireland, the linkage program has become much smaller, and in recent years its focus has shifted in response to changes in TNC strategies. In 1998, the International Business Linkage Division was established within EI to help indigenous companies meet the challenges posed by the changed organization of global value chains of the TNCs in the electronics industry. The division assists competitive Irish-owned companies to increase exports, become sub-assemblers rather than sub-suppliers and colocate with TNCs to establish affiliates in eastern European countries.

DEVELOPMENT OF KNOWLEDGE-BASED ASSETS

99

Assistance for colocation means assistance for Irish companies to become multinational and relocate part of their production out of Ireland to countries where labor costs are lower. It is an innovative approach to the dynamic promotion of indigenous companies, but it raises the question why Irish taxes should be spent on helping Irish companies outsource production to other countries. The answer is that Ireland in the twenty-first century is no longer a competitive site for the production of standardized products and that those commodities will have to be produced elsewhere, by foreign firms as well as indigenous firms. The hoped-for payoff of this international linkage policy is that the IOEs, which establish production facilities in Eastern Europe, will leave their headquarters in Ireland. And the headquarters is where R&D and design activities remain concentrated, generating considerable positive spillover to the rest of the economy. TNCs’ Interest in Domestic Sourcing The ability of indigenous producers to offer requisite inputs at competitive prices is a necessary condition for TNC sourcing in the host country. But it is not a sufficient condition. Foreign affiliates’ operations in Ireland are an integral part of the parent companies’ global strategies and the organization of their value chain at the global level. Both circumscribe TNCs’ interests in sourcing in the host country. Different parts of the value chain may be produced by TNC affiliates themselves (internalized production) or they may be produced by unaffiliated companies through arms-length contracts (externalized production or de-verticalization or vertical specialization). The extent of internalized versus externalized production varies from TNC to TNC, depending on which activities a TNC considers part of its core activities. There is a growing trend across industries toward consolidation of suppliers at the global level. And there are commonalities across firms in the same industry, like the increasingly common use of contract manufacturers in the electronics industry. Nonetheless, not all firms make use of contract manufacturers to the same extent, as they define their core competencies somewhat differently. Furthermore, TNCs’ interest in outsourcing, the use of contract manufacturers, and relocation to lower-wage countries of standardized production processes are greatly influenced by the regulatory environment in which they operate. In contrast to producers of electrical and electronic products, manufacturers of medical devices—and of pharmaceutical products—are faced with stringent regulatory requirements regarding product safety and production processes from the Food and Drug Administration (FDA) in the United States and its European counterparts. The costs of complying with those standards, both in terms of time and money, impose limitations

100

FOREIGN INVESTMENT, DEVELOPMENT, AND GLOBALIZATION

on the willingness and ability of transnational companies in that sector to relocate production processes quickly. The Electrical and Electronics Industry When the NLP focused on the promotion of backward linkages in the mid-1980s, a number of TNCs in the electronics industry were interested and willing to buy from domestic input suppliers and help them become competitive. According to one expert who has been with the NLP since its inception, TNCs in the computer industry were interested in sourcing domestically at that time, because competition in the industry had been increasing and margins, though still high, were coming down. Thus, TNCs had an interest in sourcing more inputs domestically and lowering their costs through a reduction in transportation costs, as long as they could be assured that Irish companies could deliver the requisite quality at a reasonable price. The desire to maintain good relationships with the Irish government, which was pushing the NLP, may have been another reason for some TNCs’ willingness to explore local sourcing. A high-ranking representative of one of the large foreign computer companies stressed in an interview that “good corporate citizenship” had been the main reason his company had responded to the NLP and cooperated in bringing local suppliers up to speed. The same representative also pointed out that the company had stopped the extensive support for local suppliers: “We are not sitting there, handholding any more. If they want to play the game, they need to step up to the plate.” While this statement is not necessarily representative of the sector as a whole, it does indicate a rather drastic change in attitude, the result of a sea change in the industry globally. Based on a detailed analysis of the new order of global production networks in the electronics industry, Sturgeon and Lester (2003, 46) argue,“international customers are looking for suppliers who are already able to make the products—not for firms that can be brought up to the needed level of performance.” By the second half of the 1990s, many of the OEMs were outsourcing the manufacturing part to the CMs. All the important CMs in the electronics sector came to Ireland in the course of the 1990s: Solectron, Flextronics, Sanmina, and Celestica. And many of the OEMs in Ireland outsourced to these CMs. Contract manufacturers are often required to choose an input supplier from the client’s approved vendor list (AVL). And it is difficult for an Irish input supplier to get on an AVL, even more so as the importance of scale increases. CMs can produce more cheaply than their clients because they bundle the business from a number of clients and leverage the larger volume in their negotiations with suppliers from around the globe to achieve lower

DEVELOPMENT OF KNOWLEDGE-BASED ASSETS

101

input costs. It is estimated that CMs can reduce unit costs for OEMs by 15–20 percent, by aggregating business from many different OEMs and managing global supply chains. CMs execute locally, but they manage on a global level. As discussed in chapter 2, the electronics industry has been undergoing considerable restructuring over the last several years in response to the drop in demand in the IT and telecom sector. The restructuring process has involved the consolidation of activities, the relocation of standardized production processes to low wage countries, and a further focus on the high-tech elements of the industry in the more developed countries. Before the downturn in the IT sector the rule of thumb for CMs was to have roughly 30 percent of their business in low-cost countries and 70 percent in industrialized countries. The shake-up at the beginning of the twentyfirst century led to a reversal of that maxim: 30 percent in developed countries and 70 percent in low cost countries, primarily Asia. Not surprisingly, by 2003, all the large CMs, Solectron, Flextronics, Sanmina, and Celestica had closed plants in Ireland and relocated to Eastern Europe and Asia. A few examples of the rise and fall of the large CMs in Ireland illustrate the vagaries of foreign investment, when it has no ties to the host country and when standardized commodities are its main ware. Celestica came to Ireland in February 1998, when it took over the production facility of Madge Networks, a network supplier to users of information technology (Mac Carthaigh 1998). In December of 2000, Celestica acquired Motorola’s manufacturing plants in Dublin and Iowa, which produced mobile phones and pagers (Smyth 2000). After reaching a high of 1,100 employees in 2001 and having just expanded its production facilities, Celestica was hit by the collapse of the technology boom and the downturn in demand. The company started laying off workers in late 2001, and closed the plant in Ireland in the early summer of 2003 (Smyth 2003). Solectron’s experience has been very similar. Solectron entered Ireland in 1998 through the acquisition of the manufacturing facilities of NCR, whose production included computer servers and retail scanners (Mac Carthaigh 1997). Like Celestica, it expanded its production facilities and just when it was ready to step up production, the circumstances in the computer industry changed. In May 2002, Solectron announced that it was closing its Irish operations just four years after the company had started producing in the country (Brennan 2002). Flextronics and Sanmina-SCI, the other two global CMs which had arrived in Ireland in the second half of the 1990s, were still operating in Ireland by the end of 2003, though at a much reduced scale. When Flextronics closed its plant in Tullamore, Ireland, in the spring of 2002, the company’s director of operations provided the following explanation: “A significant

102

FOREIGN INVESTMENT, DEVELOPMENT, AND GLOBALIZATION

proportion of our business which serves a particular customer is required to transfer to Flextronics in Asia, and China particularly, and in order for us to still keep that business and stay competitive and meet that particular customer’s requirements, we’re going to have to move that business to Asia” (Brennan 2002). The trend by OEMs toward reducing the number of key suppliers has brought with it the emergence of local hubbing, so that global supply can be made compatible with just-in-time delivery. The supplier from China, for example, has to warehouse the product locally and then supply it as needed, using vendor management inventory (VMI). He/she will be forced to absorb all the costs, since the client will not pay until, for example, two hours before the inputs are needed. This model of VMI global supply chain management is one of the emerging models in the electronics industry (and others). To the extent that scale and the ability to manage regional or global supply chains have become increasingly important, it has become much more difficult for Irish input producers to compete, since the bar has been raised considerably. The existing econometric studies on the determinants of linkages in the Irish manufacturing sector do not consider the organization of global value chains. But their findings are compatible with the analysis here. All studies find a statistically significant negative correlation between firm size and linkages (Goerg and Ruane 2000; McAlesse and McDonald 1978; O’Farrell and O’Loughlin 1981).15 In separate estimations for foreign and indigenous firms, Goerg and Ruane find, however, that while the correlation coefficient is negative and statistically significant for foreign firms, it is positive and statistically significant for indigenous firms. In other words, as IOEs grow in size, they will tend to source more domestically, while FOEs will tend to source less, the larger they are. Goerg and Ruane also find that domestic sourcing for both FOEs and IOEs is non-linear over time following a concave trend, increasing first and then declining. If scale requirements already mitigated against TNCs’ interest in extensive backward linkages in Ireland in the past, then that will certainly be the case in the new environment of global supply chain management. Not all foreign companies in the electronics industry have intensified their use of outsourcing and contract manufacturing. Intel is an example of a company at the opposite end of the spectrum, where many of the processes along the value chain are internalized. Wafer fabrication is a high-tech production process, and—given the high intellectual property content—clearly one of the core competencies of the company. The material inputs, which are very high-tech and require high volume production, come from companies outside the country. They are not produced in Ireland. Foreign companies also provide the sophisticated on-site support

DEVELOPMENT OF KNOWLEDGE-BASED ASSETS

103

services. While those companies employ Irish workers, it is the more basic support services (ground maintenance, restaurants, etc.) that are furnished by Irish companies. It has become increasingly less profitable to produce standardized products and parts in Ireland, as the competitive pressures on the world market have changed, wages in Ireland have risen, and new opportunities for investment in lower-wage countries have opened up. The intensified price pressure from lower-wage countries makes it increasingly difficult for Ireland—and eventually impossible—to remain a competitive location for the production of these commodities. One Irish producer of input components in the electronics industry provided a vivid description of the dynamics at work: OEMs started to reduce the number of suppliers, to reduce overhead in purchasing and reap benefits from greater economies of scale. The now global procurement manager is looking for the best price, and the best price for completely standardized items is in China. So he can demand the same best price in Ireland. Automatically that squeezes the price down in the higher labor cost areas. And then after two to three years, the price in Ireland has been squeezed down to its ultimate limit; the operation has become unprofitable, and then it has to close because it is cheaper to bring the product from China.

As a result of the intense price pressures, many of the companies that are in the business of producing standardized products have closed their production facilities in Ireland. And that is not just true for contract manufacturers. For example, in 2000, Panasonic closed its disk drive factory in Ireland, the last remaining disk drive factory in Europe. Competitive pressures had intensified dramatically, after the EU decided to remove tariffs on the imports of disk drives in 1997. To illustrate the price pressures that led to the eventual closure of Panasonic’s facility, its manager stressed that in 1992 the company had sold 160 MB disk drives for $153, while in 1999 it sold 15 GB disk drives for $77 (Keogh 2000). Overall employment in the electrical and electronics industry had fallen from a peak of more than 50,000 jobs at the beginning of 2001 to 31,900 in September 2003 (Taylor 2003). The Industry of Medical Instruments and Devices Some of the giants in the sector of medical instruments and devices, like Abbot and Baxter, have had production facilities in Ireland since the 1970s. In the course of the 1990s, however, the sector expanded considerably in Ireland, with large increases in investment from both TNCs operating in Ireland already as well as companies investing in Ireland for the first time. Today, Ireland is one of the centers in Europe for medical device production.

104

FOREIGN INVESTMENT, DEVELOPMENT, AND GLOBALIZATION

Out of the world’s top 20 companies in the field, 13 are producing in Ireland and several have chosen Ireland as the location from which to service their worldwide markets.16 The medical-device industry used to be more intensive in the use of domestic inputs than the manufacturing sector as a whole, and the electronics sector in particular. In 1985, TNCs in the medical-device industry sourced more than twice as much in relative terms as those in the electronics industry, 17.7 versus 7.7 percent of material inputs (see table 4.1 again). As pointed out earlier, the data do not allow us to make a distinction between supplies from Irish-owned companies versus supplies from foreign-owned companies, which colocated in Ireland to supply the TNC clientele. The percentage of locally sourced materials increased throughout the 1990s, but by the late 1990s, it started to decline, just like in the other sectors. And today, the share of domestically produced material inputs is about the same as for electronics and the manufacturing sector overall. In addition to plastic packaging and paper cartons of various types, some of the TNCs also source different plastic and metal-sheet-based components in Ireland. But generally the purchases of components and products from indigenous companies are limited. The prevalence of regulatory requirements and the importance of proprietary knowledge are important explanatory factors for the limited domestic sourcing of material inputs. The possibility for customized production of relatively small runs of components has opened possibilities for Irish producers who have already established the requisite production expertise. At the same time, the move toward a smaller number of suppliers at a regional and global level and a growing, though still cautious, trend toward contract manufacturing militate against increases in local sourcing and are probably the main driving factors behind the relative decline in local sourcing of material inputs in recent years. If quality is important in the supply to electronics corporations, it is critical for medical devices and instruments. If a chip in an electronics product is bad, it is a nuisance for the customer and a cost for the producer. But if the design or manufacturing of a medical device is flawed, it may possibly lead to the death of the patient. Because of questions of life and death, regulation and validation are pervasive in the medical-device industry. The medical-device sector encompasses a range of heterogeneous products and components, ranging from Band-Aids to test equipment in laboratories to catheters to artificial limbs. To ensure the effectiveness and safety of medical devices, the U.S. Food and Drug Administration (FDA) imposes strict regulations. Any company that wants to sell a medical device in the United States needs to comply with the regulatory requirements of

DEVELOPMENT OF KNOWLEDGE-BASED ASSETS

105

the FDA. The FDA classifies medical devices into three classes, with regulatory requirements becoming more stringent, as the potential risk to the patient increases. Class 1 devices do not present a risk to the patient if they fail; for example, Band-Aids and stethoscopes. Class 2 devices are mostly products used for a procedure, for example blood pressure reading devices and IV sets. And Class 3 devices comprise all devices that go into the body and stay there, sustaining or supporting life, and whose failure may be fatal, for example heart valves and stents. The different FDA regulations deal with design, manufacturing, packaging, labeling, commercial distribution, and post-market surveillance of the devices.17 For most Class 1 and Class 2 devices that means submitting a Premarket Notification, 510 (k), where the applicant has to demonstrate functional equivalency of the new product to an already existing product. Some Class 2 devices and all Class 3 devices require Premarket Approval (PMA), which is a much more elaborate process involving clinical trials. On average, the FDA takes about one to three months to review and decide 510 (k) submissions, whereas PMAs can take two years or more. Generally speaking, medical device producers have to adhere to both ISO 9001 and the FDA’s Quality System Regulation. The quality regulations in ISO 9001 are more product driven aiming for a company’s consistent adherence to specified technical requirements, while the FDA’s Quality System Regulation is more process driven, where the company has to demonstrate its capability to meet high safety standards in the design and supply of a product. The whole gamut of medical device manufacturing is controlled by validation, where a company has to establish documented evidence that a system actually does what it is meant to do (Shotwell and Carr 2003). If inputs and machinery are changed it often necessitates revalidation, a process that is costly in terms of both time and money. The pervasiveness of validation has worked against TNCs assisting potential indigenous input producers to become competitive suppliers. Upholding high quality standards is absolutely critical in the context of life and death, and a company can ill afford to take a risk by bringing a potential input supplier along. For many medical device TNCs, proprietary knowledge lies not only in the design of a new medical device, but also in the production of that device. When production constitutes part of the core competencies, internalized production becomes prevalent. One affiliate of a large U.S. TNC in Ireland defined the plant’s core competencies as the ability to integrate backward into molding and extrusion, but also to integrate forward into the sterilization, packaging and finishing up the product. Traditionally, corporations that source a large part of their components, have been dealing with many individual suppliers. The supplier base of the medical-device industry in the industrialized countries has typically

106

FOREIGN INVESTMENT, DEVELOPMENT, AND GLOBALIZATION

been very fragmented, with suppliers concentrating on one specialized part, be it metal wire forming, plastic molding or sheet metal stamping. Few companies have offered capabilities across materials or processes (Huckman 2003, 2). That is due to the fact that components for different medical devices are different and custom-made. A heart catheter with a stent at the end is completely different from an ambulatory pump delivering IVs or from an artificial hip. What the three devices have in common is that they contain some plastic parts and that they are all manufactured under clean room conditions and in compliance with US or European regulations. But the basic items are different and their components are different. Each one is custom-made. The fact that component production is often customized and on a relatively limited scale makes it feasible for an Irish producer to link up with a TNC, assuming he/she can provide competitive pricing and quality. Nonetheless, OEMs tend to stay with particular input suppliers and production sites for the life span of a product, since any changes may require a costly revalidation. Thus, new suppliers have traditionally had a better chance with new products, rather than existing products. Like in the electronics sector, the existence of approved vendor lists makes it difficult to break into that market. The vice president of market development at MedSource Technologies, an aspiring contract-manufacturing firm in the medical device sector, offers an insightful summary of the risks for the TNC: “If you are an OEM who is buying $1 million worth of components from supplier A to support a product that generates $50 million in revenue, are you really putting the whole $50 million at risk for the sake of saving $100,000 a year?” (Huckman 2003, 3). Regulatory requirements and validation together with the heterogeneous nature of the medical-device sector have so far limited the use of contract manufacturing in the sector. Contract manufacturers in the electronics industry are able to produce at a lower unit cost, because they bundle the orders from several clients and leverage the higher volume in their negotiations with input suppliers globally. They can and will switch input producers and production sites at a fairly rapid pace, if they find a company or location with lower production costs. In the medical industry that is much less of an option. When inputs are validated and registered in a site, a producer does not simply switch to another input supplier. That would require revalidation and entail additional risks and costs. As a result, producers tend not to change input suppliers and location for the life span of a product. Developments in the health-care sector have started to change the organization of global value chains in the medical-device sector though. The continued rise in health-care costs has increased pressures for cost containment on hospitals, the main clientele for medical devices. And

DEVELOPMENT OF KNOWLEDGE-BASED ASSETS

107

large hospital groups have used their buying power to exert pricing pressure on producers of medical devices. To counter the downward pressure on margins in the industry, medical-device companies have been seeking to reduce the number of suppliers and to pursue contract manufacturing more seriously. Given the large number of single input suppliers and the utmost importance of quality, contract manufacturing in the medical-device industry is more challenging. Since contract manufacturers cannot leverage the larger volume of multiple clients in the same way as a CM in the electronics industry, the cost savings are supposed to come primarily from combining multiple capabilities offering a one-stop source for device design and manufacturing. In all industries, acquisition of existing facilities and capabilities is an important driver behind the growth of contract manufacturers. The difficulties of standardizing and coordinating across different platforms are formidable in the medical-device industry.18 One of the reasons existing CMs have become increasingly more interested in addressing those challenges is the fact that the restructuring in the computer sector—and in electronics more generally—has forced them to broaden their pursuits. Among the top 100 contract manufacturers, there are still very few corporations for whom medical devices constitute a significant share of the finished product mix. And the ones for whom medical devices are important are all very small by CM standards. Most of them have less than 1,000 employees (Roberts 2003a). As intense competitive pressures persist in the market for medical devices, producers will continue to pursue aggressively possibilities for cost reductions and to reduce the time to market for new products. As a result, they will continue to look toward expanding contract manufacturing. And as contract manufacturers in the medical-device sector go through a learning phase and the new competitor countries on the global stage improve adherence to quality standards and their enforcement, it is likely that contract manufacturing will grow in the industry. The characteristics and development of the organization of global value chains in the medical-device sector have several implications for the possibilities of extensive and intensive backward linkages in Ireland. First, the pervasiveness of custom-made component production on a relatively small scale means that scale is not a prohibitive factor for potential Irish input suppliers. That contrasts with the electronics industry, where many more components are standardized and can be produced profitably at a large scale and in the context of a global network, both of which exceed the capacity and capability of most Irish producers. Second, TNCs have a high interest in geographical proximity of input producers, because they want greater control due to quality concerns, quick responses to changes, and

108

FOREIGN INVESTMENT, DEVELOPMENT, AND GLOBALIZATION

shorter lead times. Thus, the ability to produce locally, for example in the medical cluster around Galway, is advantageous for Irish suppliers. Third, potential Irish producers have to pursue opportunities aggressively, demonstrating that they already have the capability to produce high quality, and that they can do it more inexpensively. After all, they are operating in a structural context where TNCs will tend to prefer tried and trusted suppliers, that is, the ones on their approved vendor lists, and where the trend is toward reducing the number of suppliers. Irish-owned contract manufacturers—of which there is a small number—are, of course, also bound by AVLs. In response to my inquiry about the types of inputs that can be sourced in Ireland, an Irish contract manufacturer of medical devices responded that in his particular case everything could be sourced in Ireland. But, he said: “when we take a job from a U.S. company for example, then the suppliers for that company will be in the U.S. You can’t change because they have been validated.” Of course, if the cost difference with Irish producers is large enough, the client will weigh the benefits of revalidating the local source. And if there are two or three local suppliers together, it may be worth the company’s while to revalidate. If an upgrade of the product were in the offing, then revalidation would be needed anyway. In order to become a competitive supplier, it will be increasingly important to be able to sell multiple capabilities along different parts of the value chain. Irish input suppliers who were supported by EI and who grew up on the back of the TNCs in the electronics sector have acquired significant manufacturing capabilities. As the nature of electronics is changing dramatically in Ireland, leaving Irish input producers in the lurch, the challenge for them is to switch to supplying the medical sector and to expand the range of their capabilities. Summary The growth of intensive and extensive linkages in the Irish high-tech industries was the result of both EI policies and TNC interest. It was not until the NLP got under way that linkage formation in the electronics industry took off. But it only could take off because TNCs were sufficiently interested in domestic sourcing. Without one or the other element, EI policies and TNC interest, linkage formation would not have developed as far as it did. That became obvious as circumstances changed. With changes in market competition, technology, and political–economic dynamics, the needs of TNCs changed as well, reducing the degree of linkages and posing new challenges for Irish producers. In some cases, the ability to provide inputs at a regional or global level has become a prerequisite for being a competitive input supplier, which presents a big challenge for some

DEVELOPMENT OF KNOWLEDGE-BASED ASSETS

109

Irish-owned producers and former input suppliers to TNCs. In other cases, the capability to address multiple aspects of the supply chain has become increasingly important. Was it scale limitations, lack of indigenous capabilities by Irish producers, or TNCs’ desire to source from within their own corporations or their established purchasing network that limited material input sourcing of FOEs in Ireland? The answer varies somewhat by company and industry, as we saw earlier. A study by Hewitt-Dundas et al. (2002) sheds some additional light on this question. It is based on extensive interviews with 61 TNCs in Ireland’s manufacturing sector, the vast majority of them in high-tech industries. The study shows that TNCs’ internalized production and IOEs’ lack of capacity played an important role in limited sourcing from local producers. One in five TNCs indicated a preference for TNCinternal sourcing, and one in three pointed to limited supplier capacity. But the most important constraint was the non-existence of particular inputs at the local level. An astounding 86 percent of TNCs indicated that the lack of local input availability was a constraining factor on local sourcing (see table 4.6). Our analysis together with the results of the Hewitt-Dundas study suggests that the small size of the Irish economy ultimately poses a limit for the possibilities for extensive and intensive linkages. That is, of course, not surprising, since the need for specialization is much larger in a small economy, and Irish producers could not possibly produce profitably all inputs for TNCs, even if the latter had the desire to source as much as possible domestically, and Irish producers had the capability. Where supply constraints were already a key factor behind limited domestic sourcing in the year 2000 (the year when the Hewitt-Dundas study was conducted), the limitations of size will become even more pressing in the future and will hold out even greater challenges. The average size Table 4.6

Constraints on Local Sourcing by TNCs

Source of constraint on local sourcing

Percent of TNCs highlighting constraint on local sourcing

Prefer group supplier Technical competence Delivery times Quality issues Too expensive Supplier capacity Local availability

21 25 10 15 38 31 86

Source: Hewitt-Dundas et al. (2002, 38).

110

FOREIGN INVESTMENT, DEVELOPMENT, AND GLOBALIZATION

of the Irish-owned manufacturing unit did not change in the course of the 1990s. Measured in terms of employment, an IOU employed on average 29 persons in 1991 and 30 persons in 2000. The average enterprise size across NACE sectors ranged from 29 to 58 in 2000. With the average size of indigenous production units unchanged, the increased scale requirements in some industries, such as the electronics, make it so much harder for Irish firms to compete in the future, unless they move up the value chain, specialize in niche production, and export more. A few Irish-owned companies in the high-tech sectors stand out among the rest for their growth and mastery of technological capabilities. I already mentioned Iona in the software industry. ELAN in the pharmaceutical industry is another example. It is focused on drug delivery and specialty pharmaceuticals. It was started in 1969, had sales of over $1 billion in 1999, and its shares are traded on the New York, London, and Dublin stock exchanges. In a report on future industrial policy Forfas (2003d, 17) argues that “the trend is towards the creation of niche production centers linked into a global value chain, with the global market being the end customers.” The software industry provides an example of growth based on successful niche production. The success of the indigenous Irish software industry had a lot to do with the characteristics of the sector: low barriers to entry and ample opportunity for niche production serving segmented markets (e.g. O’Malley and O’Gorman 2001). The software development industry is the indigenous high-tech sector, which has grown the most. According to ABSEI, the sales of the indigenous software development sector grew from €57.4 million in 1990 to €1.3 billion in 2002, compared to an increase from €235.5 million to €883.9 million for the electrical and optical industries (Forfas 2003a, 27). The software sector was able to grow on its own more than any of the manufacturing subsectors. While TNCs have at times been useful in the seeding of indigenous software ventures and as a source of software demand, very low barriers to entry and the generation of computer engineers by the Irish education system have allowed for significant non-TNC related growth. At this juncture though, Irish software companies—like other Irish companies—face the challenge of tapping into international networks. The challenge for software companies is to establish credibility with global systems integrators in order to expand on the international market. Human Capital Formation and Indigenous Knowledge-Based Assets The previous section showed that intensive and extensive linkages between TNC affiliates in high-tech industries and Irish input producers did grow

DEVELOPMENT OF KNOWLEDGE-BASED ASSETS

111

in the 1980s and 1990s, but that they were ultimately limited by changing TNC interests and needs, Irish company capacity, and small country size. The available evidence strongly suggests that knowledge transfers happened through the linkages that did develop. In the mid-1980s when EI was promoting domestic linkage capability, especially in the electrical and electronic industry, many TNCs in those industries were very cooperative and helpful in transforming potential Irish input suppliers into actual input suppliers. They provided assistance in a number of different areas, from technology to quality standards to information on sourcing. As competitive circumstances changed, many TNCs became less interested in direct and close cooperation with Irish suppliers regarding their ongoing competitiveness. But suppliers were probably less in need of the kind of support that they had required initially. The development of regional capacity at the supra-national level has now become the more critical challenge for indigenous manufacturers. The development of supply chain linkages between TNCs and Irish input producers is not the only way through which the expansion of indigenous knowledge-based assets can happen. The extent of knowledge transfer from high-tech FDI to domestic producers also depends on the amount of on-the-job training, the type of skills conferred, and the spillovers to other companies. Regarding the skill requirements of TNC production in Ireland’s hightech sectors, there clearly is a bifurcation. At one end of the spectrum, FOE activities comprise a lot of assembly type operations, even though workers may be operating sophisticated machinery. At the other end of the skills spectrum, FOEs employ a large number of workers who are technically highly educated, from technicians to engineers to scientists with different specializations. After all, the availability of well-educated labor at relatively low cost was one of the factors that attracted high-tech TNCs in the first place. Several authors have suggested that the training and knowledge spillovers from high-tech FOEs have been rather small. O’Sullivan (2000, 1995) and O’Hearn (1998) stress that the vast majority of jobs in the hightech enterprises are standardized low-tech jobs, which do not equip workers to master more sophisticated production processes. O’Sullivan (1995, 388) argues that most of the TNC training of Irish workers “is designed to improve dexterity in routine operations rather than to develop multiple, flexible skills that could serve as a basis for innovative activities.” While this argument is certainly supported empirically, it takes insufficient account of the dynamics of skill composition and skill development over time. As an SLC, Ireland has by definition not had a history of extensive indigenous R&D, with the requisite institutional infrastructure. So, we

112

FOREIGN INVESTMENT, DEVELOPMENT, AND GLOBALIZATION

cannot expect that TNCs in the high-tech sector in Ireland invested in production activities that focused primarily on the high end of the value chain. The critical question is, whether once in Ireland, TNCs moved up the value chain over time, thus increasing the demand for more highly skilled workers as well as providing some of the upgrading of skills themselves. And here we see that the skill intensity of FOEs increased in the course of the 1990s. Nonetheless, the high-tech sectors do not present a uniform picture. At the industry-wide level, higher skill intensity has grown in the chemical, electrical, and medical instruments sector, but it has fallen in the computer industry. It is harder to tell what happened to the R&D intensity of FOEs. The data indicate that it fell throughout the decade, but they may well look differently if we could factor out the impact of transfer pricing. With respect to IOEs, we find an upgrading of skills, as indicated both by employment composition and research intensity. Training and R&D in the Manufacturing Sector The very fact that many FOEs, especially in the electronics industry, have increasingly relocated the production of standardized components and products from Ireland to lower wage countries supports the argument that the skills transferred in many foreign-owned production sites in Ireland were at the lower end of the skill spectrum. At that level, the cost of labor is the critical determinant of competitiveness, rather than the availability of very high-skilled labor. The composition of the labor force and changes in the composition over time provide some insights into the skill distribution in an industry. The CSO’s CIP provides disaggregated information about employment in the manufacturing sector. The survey distinguishes three categories: industrial workers and apprentices, clerical and other office staff, and administrative, technical and other staff. It also provides separate data for male and female workers. The literature on the gender composition of employment in assembly type production suggests that female workers constitute the majority of the labor force in assembly-intensive production, which requires the manual dexterity that O’Sullivan is referring to (Elson and Pearson 1981, Standing 1989). Yet, the Mexican maquiladoras, for example, have witnessed an increasing masculinization of the labor force over the last 20 years, which may indicate that the industry has become more capital intensive, with greater technical skill requirements (Wilson 2002).19 Thus, if female workers are relatively concentrated in the production of hightech goods, it may suggest that the skills are predominantly of the type described in O’Sullivan’s quotation above.

DEVELOPMENT OF KNOWLEDGE-BASED ASSETS

113

When we analyze the gender breakdown of employment categories in the high-tech sectors in Ireland, we find that the percentage of women among production workers in the electrical and electronics industries is indeed considerably higher than for the manufacturing sector as a whole. This holds especially true for the production of medical instruments: 62 percent of production workers are female, compared to 32 percent for the average in manufacturing. One of the Irish producers of medical instruments explained in an interview that most of his production workers were female, because they had the requisite dexterity and nimble hands.20 Nearly all FOEs provide formal training for their workers. According to an ABSEI survey (Forfas 2003a), 93 percent of FOEs in the manufacturing sector provided formal structured training in 2002. Although only 81 percent of IOEs spent resources on formal training programs, they did spend relatively more, 1.7 percent of payroll cost compared to 1.3 percent for FOEs. Among IOEs, the medical instruments sector stands out for its high training intensity. Nearly all IOEs in the sector (96 percent) had a structured training program in 2002, and the training expenditures in the industry amounted to 2.4 percent of total payroll cost. Since the figures tell us nothing about the types of training undertaken and the kinds of skills conferred, they do not allow for inferences regarding upgrading. For that, we would need more detailed information at the firm level. At the anecdotal level we know, for example, that the cumulative effect of the training has played an important role in attracting new waves of foreign companies in the electronics industry. Additional insights into TNCs’ development of skills can be gleaned from a look at the other end of the skills spectrum. The CIP’s data on “technical and administrative staff ” provide the best proxy available for the use of more highly skilled labor. Comparing the composition of employment in 1991 and 1999 we find that, in both years, the high-tech sectors are substantially more high-skill intensive than the manufacturing sector as a whole (see table 4.7). Technical and administrative staff comprised 16 percent of employment in the manufacturing sector in 1999, but 25 percent in the chemical industry and 20 percent in the electrical and optical industries. When we distinguish between indigenously-owned enterprises and foreign-owned enterprises though, we find that the two sets of companies followed different trends over time. The share of more highly skilled workers in the IOEs in high-tech sectors held steady or declined over the course of the 1990s, but it was more than twice as large as for IOEs in the manufacturing sector as a whole, with the exception of the electrical equipment sector. On the other hand, the share of more skilled labor in FOEs in hightech sectors was considerably above the respective average for the manufacturing sector in 1991, but was about the same by the end of the decade

114

FOREIGN INVESTMENT, DEVELOPMENT, AND GLOBALIZATION

Table 4.7 Composition of Employed Workers in High-Tech Industries, Ireland, 1991 and 1999 Industrial workers and apprentices

Clerical and other office staff

Administrative, technical and other staff

IOEs and FOEs

1991

1999

1991

1999

1991

1999

24 Chemicals

IOEs FOEs Total

61.3 63.6 63.1

66.2 60.4 61.5

15.7 12.5 13.3

13.8 13.6 13.7

23.0 23.8 23.6

20.0 26.0 24.8

30–33 Electronics

IOEs FOEs Total

69.5 70.1 70.0

69.7 64.8 65.6

10.1 12.8 12.4

9.9 15.8 14.8

20.4 17.1 17.6

20.4 19.4 19.6

30 Computers

IOEs FOEs Total

60.3 53.7 54.7

57.8 59.4 59.2

13.0 20.4 19.3

15.0 21.0 20.3

26.7 25.8 26.0

27.3 19.7 20.6

31 Electrical

IOEs FOEs Total

78.3 80.8 80.2

79.4 71.4 74.0

9.0 8.0 8.3

8.8 8.0 8.3

12.7 11.2 11.5

11.8 20.5 17.7

32 Radio, TV

IOEs FOEs Total

69.3 65.7 66.3

68.0 55.2 56.7

7.7 13.5 12.6

7.2 25.1 23.1

23.0 20.7 21.1

24.8 19.6 20.2

33 Medical

IOEs FOEs Total

56.6 76.2 74.7

63.0 74.7 73.0

11.0 10.4 10.4

8.9 7.1 7.4

32.4 13.5 14.9

28.1 18.2 19.6

15–37 Total Manufacturers

IOEs FOEs Total

76.9 73.6 75.4

74.9 66.1 70.3

21.4 36.9 28.3

11.9 15.9 13.9

10.6 11.7 11.1

13.6 18.0 15.8

NACE 2-digit

Source: Calculated from unpublished data from the Census of Industrial Production 1991, 1999 provided by Breathnach, Forfas.

(with the exception of the chemical industry). That was due to the fact that the average high-skill intensity of FOEs in manufacturing had increased significantly in the 1990s. Of particular interest for our purposes here are the trends in the employment composition in FOEs in the computer industry (NACE 30) and the medical instruments industry (NACE 33). On average, FOEs in the computer industry have become less intensive in the use of highly skilled labor, with the share of more skilled workers declining from 26 percent in 1991 to 20 percent in 1999. Nevertheless, the absolute numbers of more highly skilled workers in the computer industry grew from 4,750 in 1991 to 10,481 in 1999. The relative decline might reflect the fact that FOEs in the computer sector have expanded their activities more into services in the 1990s.21 Some FOEs have established call centers where they provide

DEVELOPMENT OF KNOWLEDGE-BASED ASSETS

115

technical customer support or marketing services. Dell, for example, has a multilingual telesales and technical support center in Ireland; IBM has an e-commerce center; and Oracle has its European business center for customer and technical support there.22 For other TNCs, Ireland has become the site for a shared services center, where different aspects of national administrative centers in a number of European countries are merged into one central site, as for example in financial administration, procurement, and logistics. Examples include Apple, Dell, Honeywell, and Xerox. In contrast to the development of the skills composition in the computer industry, the share of more highly skilled labor in the medical instruments industry has increased throughout the 1990s, suggesting that companies in the industry have been moving up the value chain. In both industries though, IOEs have been considerably more intensive in more skilled labor than their foreign counterparts. This suggests an expansion of knowledge-based assets among indigenous firms and bodes well for their future ability to compete. Companies’ engagement in R&D activities provides another indication of the extent to which more advanced skills are fostered and required by FOEs and IOEs. Forty-seven percent of FOEs claimed to have spent some money on in-house R&D activity in 2002, compared to 54 percent of IOEs. Since IOEs are, on average, considerably smaller than FOEs, it is not surprising that the percentage of IOEs claiming to have spent on in-house R&D declines markedly when a minimum expenditure of €127,000 is imposed.23 In that case, the share of IOEs falls to 20 percent, while the share of FOEs falls to 33 percent (Forfas 2003a). Nonetheless, even though we do not know what kind of activities companies classify as R&D, the fact that more than 50 percent of IOEs are engaged in some kind of R&D activities is ground for cautious optimism regarding the abilities of indigenous companies to go beyond the production of standardized products. On average, FOEs’ expenditures on R&D per person employed were nearly twice as high, €3,600 per person, compared to €1,900 for IOEs. But, here again, we see that the performance in the high-tech sectors is not uniform. Although FOEs in the electronics and the medical instruments sectors spend more on R&D per person employed than the respective average for the manufacturing sector as a whole, their expenditures pale in comparison with the IOEs in these sectors. In high-tech industries, IOEs are much more research intensive than FOEs. In the electrical and electronics industry, FOEs spent €5,700 per employee on R&D in 2002, compared to €10,000 spent by IOEs. For the medical instruments sector, the comparable figures are €4,700 for FOEs and €8,500 for IOEs (Forfas 2003a). Expenditures on R&D as a percentage of output are commonly used as an indicator of the R&D intensity of economic production in a country. In high-tech sectors, R&D expenditures as a percentage of output are above

116

FOREIGN INVESTMENT, DEVELOPMENT, AND GLOBALIZATION

the average for the manufacturing sector as a whole; and they are always higher for IOEs than for FOEs. In absolute amounts, FOEs’ R&D spending dwarfs the domestic spending, of course (see table 4.8). But R&D spending has not become relatively more important for high-tech FOEs in Ireland in the 1990s. On the contrary, the share of R&D expenditures in total output has fallen, for the whole manufacturing sector and in each of the high-tech industries. While these data suggest that TNCs have become less research intensive, the existence of transfer pricing and inflated profits, and thus overstated output values, has to temper such an inference. In sum, the available evidence suggests that TNCs in the high-tech sectors have increased the use of more highly skilled workers in the course of the 1990s: in absolute numbers in all high-tech sectors, but in relative terms only in the chemical, electrical, and medical-instrument sectors. It is difficult to interpret the decline in FOEs’ R&D intensity, since transfer pricing distorts the data. Nonetheless, as I argue later, the skill formation and learning process provided by TNCs has made a significant contribution to the skill and research intensity of IOEs. Knowledge Spillovers from High-Tech TNCs The clearest evidence that TNCs generated positive knowledge spillovers and that they contributed to an expansion of IOEs’ technological Table 4.8 Business Expenditures on Research and Development, FOEs and IOEs in Ireland Electrical & electronic equipment

Pharmaceuticals

Medical instruments

Total manufacturing

Mill.€

% of output

Mill.€

% of output

Mill.€

% of output

Mill.€

% of output

FOEs 1993 1995 1997 1999 2001

102.1 132.2 206.1 257.9 290.3

1.8 1.2 1.5 1.2 1.2

57.8 72.2 70.1 77.5 62.1

7.6 4.6 3.8 2.6 1.2

16.9 19.9 23.4 25.2 51.1

1.7 1.7 1.5 1.0 1.2

216.2 294.5 362.2 422.3 464.8

1.2 1.1 1.0 0.8 0.8

IOEs 1993 1995 1997 1999 2001

15.4 27.8 44.1 58.2 53.6

3.7 5.0 5.1 6.7 4.2

4.4 3.9 4.8 5.0 8.6

6.5 4.8 4.8 4.6 3.3

1.8 4.2 4.9 13.7 8.2

2.4 4.4 3.3 9.0 1.8

87.7 127.5 142.2 164.9 161.8

0.7 0.9 0.9 0.9 0.8

Source: Forfas (2003c).

DEVELOPMENT OF KNOWLEDGE-BASED ASSETS

117

capabilities is that many IOEs were founded by Irishmen (and some women), who were former TNC employees. Through their work experiences with foreign high-tech companies they had acquired skills and insights into production processes and the workings of the global market that allowed them to venture out on their own. Many of the owners of Irishowned companies in the medical-instruments field worked with one or several of the big TNCs in the country, for example Abbot and Baxter, before opening their own businesses. And the same holds true for a sizeable number of highly skilled ex-employees of FOEs in the electronics industry. A couple of examples illustrate the dynamics. Having worked at Dell, IBM, Intel, and other TNCs, allowed Irish engineers to develop ideas for a product or service that they could produce in a small company of their own. Often, EI provided the support for a feasibility study and for overcoming early growing pains. For example, one Irish producer with 15 years of work experience with different multinationals in the electronics sector, decided to set up a company that would offer rework and repair services to TNCs in the electronics sector. Besides providing expertise in the area and the context, which generated the new business idea, the TNC connection brought other benefits. For one, the Irish producer was keenly aware of the paramount importance of quality standards in the industry. Thus, one of his first actions was to send his training supervisor to the United States to become fully trained in the requisite standards of production for his business. He knew that production based on ISO 9000 was a necessary prerequisite for becoming a TNC supplier. Furthermore, his prior work history with TNCs provided him with an “in” with his former employers. It gave him greater legitimacy with his clients and a ready-made network. Since its inception around the mid1990s, the company has expanded its production/service repertoire as well as its client base. With the structural transformation in the electronics industry, the company has shifted more of its business toward the medical instruments sector, and—at the time of our interview in August 2002—it was planning to set up a production facility in Hungary. It is a good example of a company, which has followed a supplier-based upgrading model. Another Irish producer with more than a dozen years of work experience with TNCs in the medical sector saw a niche for contract manufacturing and struck out on his own. He began with two people in the mid-1990s and now has over 50 employees, supplying a number of new clients beyond the original network he started out with. His characterization of the dynamics of the industry and his business is insightful: “The business is opportunity driven. You have to find the opportunities and seize them; they won’t come to you. The medical device business is changeable. A lot of companies are trying to grow with new technology. So the old technology,

118

FOREIGN INVESTMENT, DEVELOPMENT, AND GLOBALIZATION

which is old to them [TNCs] and high-tech to everybody else, is being discarded to companies like us.” This particular company has been very successful. It sources globally, and it exports three quarters of its output. There are many anecdotal stories about engineers setting up their own enterprise after having worked in a multinational company for a few years. But there is no systematic evidence on how many Irish-owned companies have started and developed along the lines of the two examples described above. Based on interviews and the available general data, it is most likely that their number in the manufacturing sector is fairly small, probably a few dozen and not a few hundred. One of the industries that clearly has benefited from knowledge spillovers through labor mobility is the software industry, one of the few sectors in which IOEs have thrived in the 1990s. Barry, Bradley, and O’Malley (1999, 65) refer to a study on clustering in indigenous manufacturing which found that one third of entrepreneurs of IOEs in the software sector had worked in foreign firms immediately before the startup of the new firm, while two thirds of entrepreneurs had worked at some stage in their careers in foreign firms. Referring to a more recent study “Ireland’s Software Cluster 2002,” EI (2002) states that “only one quarter of all Irish software start-ups are set up by former employees of multi-national technology companies, with one in six originating in an academic environment—either a University or Research Institution.” Summarizing the findings of their interviews with indigenous companies in the software industry O’Malley and O’Gorman (2001, 312) report that the vast majority of them highlighted the quality of the labor force as critical for their competitive success, a quality which resulted from an educational system that produced computer engineers and from on-the-job training that was provided in part by TNC affiliates. There are numerous ways in which TNCs can have a positive impact on improving the competitiveness of Irish firms and the expansion of technological capabilities in the long run. Most of them defy quantification. One way in which TNCs can contribute to IOEs’ competitiveness is through the provision of contacts and references. For example, if a small Irish software company wants to get into the global market and get a share of the business of a European-wide systems integrator, its chances will depend a lot on its ability to establish credibility. If the company can sell successfully to a big-name TNC in Ireland, it will provide an acceptable international reference. We have seen that small country size imposes limitations on the possibilities of Irish-owned producers. But there are contexts in which small size is advantageous. In a country like Ireland, everybody knows everybody else in the business. And as long as he has provided good quality services,

DEVELOPMENT OF KNOWLEDGE-BASED ASSETS

119

an Irish software producer will find it easier to get one of the higher-ups in a client TNC to pick up the phone and make new contacts or provide the critical reference than it would be in a less tight-knit business community. One example of TNCs’ unmeasurable impact on the country’s longterm technological capability is the influence they have had on the development of the education system at the technical, post–high school level. Over the years, TNCs in high-tech sectors have conferred with different government institutions, with concerns about scarcities of different kinds of high-level technical education, providing important input into educational planning.

Linkages and Knowledge Spillovers: Econometric Evidence While it is useful to distinguish, theoretically and empirically, between linkages and human capital spillovers as two separate conduits for technology transfer, the Irish experience demonstrates that the two channels can be inextricably interconnected. In a small latecomer (SLC), where—by definition—high-tech industry does not exist before the arrival of hightech FDI, work experience with a TNC at the higher end of the skill level has given rise to Irish entrepreneurs who mostly started out as input providers for their former employers. An analysis of the contribution of high-tech FDI to the formation of a new indigenous entrepreneurial class would make a fascinating sociological study. In chapter 2, we saw that the econometrically based empirical literature provides mixed evidence for the existence of positive productivity spillovers from TNCs to indigenous producers. The inconclusive evidence is due to differences in the level of aggregation (firm level data versus industry level at different degrees of aggregation), differences in the definition of foreign ownership (majority versus minority foreign ownership), differences in time periods, and differences in the extent to which threshold levels for foreign investment or indigenous capabilities are taken into account. Econometric studies of spillovers from FDI in Irish manufacturing are no exception, even though they all focus on the same country. The studies that do not consider threshold levels in the estimation tend to find no statistically significant evidence for productivity spillovers, while the studies that include thresholds, do find such evidence. Without including any threshold levels in the estimation, Barrios et al. (2002), Ruane and Udur (2000), and Ruane and Udur (2002) found no statistically significant evidence that the relative importance of TNCs in an industry (measured by TNCs’ employment share) generated positive

120

FOREIGN INVESTMENT, DEVELOPMENT, AND GLOBALIZATION

productivity spillovers.24 Two other studies without threshold inclusion did generate statistically significant positive spillovers (Goerg and Strobl 2002a, Kearns 2000 cited in Ruane and Udur 2002). Kearns (2000) found a statistically significant positive relationship between TNC presence and labor productivity of indigenous firms across manufacturing subsectors.25 Based on Markusen and Venables (1999), Goerg and Strobl (2002) investigate whether TNC presence in a sector induced indigenous firms to enter that sector. Using panel data for the period 1974–1995, they find a statistically positive effect of TNC presence on the net entry rate of national firms in the same industry and upstream industries. Nonetheless, the estimated effect is very small.26 Furthermore, changes in firm numbers do not tell us anything about the relative importance of indigenous firms in terms of output. Nor do they speak to the issue of productivity growth. On the other hand, there are two studies, which suggest that statistically significant productivity spillovers exist only under certain conditions. Ruane and Udur (2002) find empirical support for their hypothesis that there has to be a critical mass of FDI to generate spillovers to indigenous companies. When they used the absolute level of TNC employment in an industry, rather than the commonly used share of industry employment accounted for by TNCs, the econometric results show a statistically significant impact of TNC presence on the productivity of indigenous firms. The results are, however, sensitive to the level of aggregation at the industry level.27 While Ruane and Udur (2002) focus on the importance of a critical mass of FDI in an industry, Barrios et al. (2002) highlight the absorptive capacity of indigenous firms. They find a positive and statistically significant spillover effect from TNCs to those indigenous firms with the capacity to absorb spillovers, which they define as all firms with exports and some R&D expenditures.28 The findings of the econometric studies are compatible with the analysis in this chapter, namely that spillovers do not happen automatically, but that their development is conditional. The study by Barrios et al. is particularly important in its statistical support for indigenous linkage capability as a condition for linkage development.

Putting the Pieces Together: High-Tech FDI and Technology Transfer It is not possible to quantify exactly the extent to which high-tech FDI in the Irish manufacturing sector has contributed to the development of indigenous technological capabilities, both through linkages and through direct human capital formation and its spillovers. A main reason is that it

DEVELOPMENT OF KNOWLEDGE-BASED ASSETS

121

is impossible to distinguish in a systematic way between TNC sourcing from Irish companies and TNC sourcing from foreign companies located in Ireland. Another reason is the multidimensionality of spillovers from high-tech FDI, many of which defy quantification. Nonetheless, the different indicators analyzed in this chapter—when taken together—provide strong evidence that high-tech FDI advanced indigenous knowledge-based assets in Ireland. The sub-supply sector in the electronics industry grew considerably with the rise of the Celtic Tiger. In the year 2000, Irish-owned companies in the electronics industry sold €1 billion worth of output, with subsuppliers to TNCs accounting for 42 percent of sales and 57 percent of employment of IOEs in the industry (Lyons 2002a). Knowledge transfers via linkages occurred through assistance with the implementation of quality control systems, information on business opportunities, advice on technical issues, and in some cases joint product development. Furthermore, the production of FOEs has become more skill intensive during the 1990s as evidenced by the rising share of more skilled workers in their work force, albeit with different manifestations in the computerand medical instruments industries. And TNCs turned out to be a training ground for local entrepreneurs, providing knowledge about markets, process, and product technology. A small number of Irish-owned companies in the electrical and electronics sector have followed the supplierbased upgrading model, starting out as sub-suppliers to TNCs and then becoming multinational companies themselves. In order to assess the different pieces as a more coherent whole, I look at key aggregate performance indicators of technological upgrading among Irish-owned companies in the manufacturing sector. Not everything that has happened in the IOE sector can, of course, be attributed to TNCs. To assess the technological basis of indigenous production I look at key indicators of skill and technology upgrading at the aggregate level: changes in the output and export structure of IOEs toward more high-tech industries; a decline in the productivity gap between TNCs and IOEs; an increase in IOE engagement in research and development; and an increase in IOE investment abroad. With respect to the output and export structure of the IOE sector, we find that the weight of the high-tech sectors has increased considerably, but that it is still fairly small. In the context of a growing IOE sector in the 1990s, indigenous firms in the electrical and optical industries (NACE 30–33) more than doubled their share in IOE employment, from 4.7 percent in 1991 to 9.9 percent in 2000 (see table 3.2). Over the same period, the export share of the sector grew from 3.9 percent to 13.8 percent.29 The indigenous chemical industry (NACE 24) became more important too,

122

FOREIGN INVESTMENT, DEVELOPMENT, AND GLOBALIZATION

though the changes were not as impressive. Its share in IOE employment rose from 3 to 4 percent, and its participation in exports grew from 3.9 to 6.2 percent. Most telling with respect to the better performance of indigenous high-tech sectors is the fact that their export intensity increased considerably in the late 1990s. In 2001, the share of exports in output was 62.3 percent for indigenous companies in the electronics/medical-device industries, whereas it was only 37.3 percent for the manufacturing sector as a whole (see table 4.9). The available data suggest that productivity in IOEs has increased and that the gap with FOEs may have declined. The widespread practice of transfer pricing among TNCs makes it difficult to analyze the development of the productivity gap between FOEs and IOEs in the standard way. Since inflated profits invariably distort net output figures for FOEs, we cannot tell to what extent increases in labor productivity (net output per employee) are due to improvements in real productivity or to overstated profits.30 It is hard to believe that the ratio of net output per employee in FOEs relative to IOEs more than doubled between 1991 and 2000 as indicated by the data (see table 4.10). While transfer pricing affects the size of profits, it should not distort wages. Thus, on the assumption that there was some relationship between wage changes and changes in real productivity over the span of a decade, we can approximate the development of the productivity gap through the wage ratios between FOEs and IOEs. For the manufacturing sector as a whole, the wage ratio between FOEs and IOEs remained unchanged between 1991 and 2000. The average wage and salary per employee remained 24 percent higher in foreign-owned enterprises. In the electrical and optical industries, however, the gap shrank considerably, from 1.3 to 1.09. And in the computer industry (NACE 30) and the medical-instruments sector (NACE 33), the average wage in IOEs was actually larger than in FOEs. Since transfer pricing does not affect the output figures for IOEs, we can take their labor productivity figures at face value. For the indigenous manufacturing sector, productivity increased at an average annual rate of 1.5 percent during the 1990s. And in the electrical and optical industries, the growth rate was more than twice as high, with 3.5 percent in the computer industry and 4.2 percent in the medical instruments sector. Regarding R&D, we see that the IOE manufacturing sector increased its research intensity in the course of the 1990s, from .7 percent of output in the early 1990s to .9 percent in the late 1990s (see table 4.8). And IOEs in the high-tech sectors spent considerably more per employee on research and development than their foreign-owned counterparts. Companies in the electronics and medical instrument sectors doubled their share in total

Table 4.9

Exports and Export-Intensity in the Irish Manufacturing Sector, 1991–2001 Total manufacturing FOEs

1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001

Chemicals/pharmaceuticals IOEs

FOEs

Electronics and optical

IOEs

FOEs

IOEs

1,000s of current Euros

Export intensity (%)

1,000s of current Euros

Export intensity (%)

1,000s of current Euros

Export intensity (%)

1,000s of current Euros

Export intensity (%)

1,000s of current Euros

Export intensity (%)

1,000s of current Euros

Export intensity (%)

12,191,011 13,720,192 15,797,029 19,006,603 24,731,712 27,338,196 32,740,926 40,844,882 52,214,521 66,262,204 69,808,894

86.0 87.1 85.7 87.4 89.0 89.3 90.0 91.8 92.1 91.7 93.1

4,299,399 4,175,163 4,649,185 4,754,957 5,328,806 5,255,304 5,561,752 5,795,883 5,616,669 6,684,784 7,232,261

34.8 33.0 35.3 35.1 35.9 34.0 34.3 34.1 31.3 33.2 37.3

2,687,011 3,375,823 4,107,397 5,070,283 6,070,473 7,040,925 9,113,493 14,008,823 17,497,957 22,959,902 24,228,912

94.9 95.4 96.3 96.4 97.1 97.4 95.5 98.7 98.4 98.6 98.5

205,929 140,555 148,889 161,116 196,027 259,284 205,743 238,042 344,132 413,530 413,912

35.6 25.1 26.2 26.8 29.6 32.4 29.0 34.3 43.3 41.9 43.3

4,377,447 4,711,381 5,651,760 6,654,212 10,145,789 11,026,885 13,355,471 15,362,663 20,657,606 26,580,683 28,045,370

95.9 94.2 89.2 91.3 91.7 92.6 93.6 94.4 92.2 90.3 93.7

167,826 219,054 223,461 260,025 298,726 349,161 509,815 717,992 515,425 924,636 1,056,722

51.9 56.0 53.3 54.1 53.2 48.0 49.4 55.9 44.3 52.9 62.3

Source: CSO Census of Industrial Production various years, provided by Breathnach, Forfas.

123

124

FOREIGN INVESTMENT, DEVELOPMENT, AND GLOBALIZATION

Table 4.10 Productivity and Wages in FOEs Relative to IOEs, 1991 and 2000 Net output per employee Wages and salaries per employee FOEs/IOEs FOEs/IOEs NACE

1991

2000

1991

2000

15–16 17–18 20 21–22 24 25 26 27–28 29 30–33 30 31 32 33 34–35 15–37

3.53 1.37 2.72 5.94 3.10 1.36 1.06 1.68 1.69 2.90 4.62 1.58 2.27 1.97 1.34 2.74

6.05 2.01 2.15 14.99 15.23 1.23 1.17 1.59 1.70 4.47 5.16 2.60 6.95 1.85 1.90 6.59

1.52 1.24 1.70 0.96 1.16 1.36 1.16 1.48 1.32 1.30 1.31 1.26 1.25 1.36 0.85 1.24

1.56 1.24 1.57 1.08 1.20 1.19 1.13 1.39 1.22 1.09 0.92 1.25 1.15 0.93 1.31 1.24

Source: Calculated from unpublished data from the Census of Industrial Production, provided by Breathnach, Forfas.

IOE expenditures on R&D, from 20 percent at the beginning of the 1990s to 40 percent by the end of the decade. One indicator of the productiveness of investment in R&D is the number of patents that are being generated. In 2001, IOEs applied for 948 patents and 397 patents were granted.31 The respective figures for FOEs were 750 and 397. While absolute patent numbers do not speak to the long-term impact and marketability of the innovations, it is noteworthy that a sizable number of the IOEs’ patents were in high-tech sectors: 53 in electrical and electronic equipment, 61 in software, and 26 in pharmaceuticals (Forfas 2003c, 43). Finally, we have seen a dramatic increase in recent years in Irish investment abroad. After averaging $400 million annually from 1988 to 1993 FDI outflows increased steadily in the course of the 1990s reaching $6.6 billion in 2001 (Forfas 2003b and 2001). The growth of FDI outflows has been even larger than of FDI inflows in recent years (Barry, Goerg, McDowell 2001). But most of the outward investment has been in services (banking, insurance, and property development) and from traditional manufacturing sectors (construction materials, paper and packaging, and agribusiness). Even though the CSO does not keep data on outward investment by

DEVELOPMENT OF KNOWLEDGE-BASED ASSETS

125

industrial sector of origin, the available information indicates that outward FDI from Irish companies in high-tech sectors has grown. The most important companies investing abroad are Elan, Iona Technologies, Baltimore Technologies, Trinity Biotech, Lionbridge, and CBT Systems (Forfas 2001, 18). In sum, in the course of the 1990s, high-tech sectors increased their share in output and exports of the IOE sector; the wage gap between FOEs and IOEs remained unchanged at the aggregate level, but nearly disappeared in the computer and medical instruments sectors; IOEs’ labor productivity grew at an annual rate of 1.5 percent, and at more than twice that rate in the computer and medical-device industries; the R&D intensity of the IOE sector increased; IOEs in high-tech sectors generated 151 approved patents in 2001; and IOE investment abroad increased considerably, including investment by IOEs in the “new technology” fields. The discussion in this chapter has shown that FOEs made an important contribution to the improvement of these indicators, serving as the handmaiden of advancements in the technological capabilities of indigenous companies.

Ireland at a New Critical Juncture: Will the Celtic Tiger Come Roaring Back? Over the last 50 years, the Irish development strategy has been modified a few times, in response to problems that either were not addressed by the prevailing strategy or resulted from it. The most drastic change occurred in the late 1950s, with the switch from import-substituting industrialization to FDI-driven industrialization, when the lack of economic growth under ISI in the 1950s was to be remedied through competition and technology transfer via international trade and foreign capital. Initially the IDA was indiscriminate in its pursuit of FDI, since the overarching goal was to solve the country’s endemic unemployment problem. In the early 1980s, the search for FDI became more focused, concentrating on attracting technologically advanced TNCs, which could generate growth and domestic purchases. Electronics, computer software, and healthcare were the main targeted areas. But just as the presence of FDI in high-tech sectors does not automatically lead to positive spillovers through linkages and human capital formation, so does the expansion of linkages, once they exist, not guarantee their longevity. Neither TNCs nor indigenous companies keep producing the same products, forced to change by the very dynamics of competition and internal and external changes. I investigated those dynamics here for two industries: the electronics industry and the medical instruments industry. While both have been important destinations for foreign investment in

126

FOREIGN INVESTMENT, DEVELOPMENT, AND GLOBALIZATION

Ireland, they have not been the only ones, of course. Our goal was not to analyze every aspect and sector of Irish development under the Celtic Tiger, but rather to use the electronics and medical instruments sectors to explore sector-specific differences in linkage developments in the dynamic interaction between global and host country factors. And as linkage developments in Ireland have been changing, so has the very foundation of Ireland’s development strategy. At the beginning of the twenty-first century, Ireland finds itself at a new critical juncture. The government’s response has been to change the development strategy. The new strategy still emphasizes the importance of FDI to Ireland’s future growth and development, but it focuses on R&D and higher value added, rather than factor accumulation, as the basis for growth. The obstacles to success are formidable. The country needs to meet the challenges posed by its changed location-specific assets, as discussed at the end of the last chapter. Out of the many factors that need to come together for a reemergence of the Celtic Tiger, I want to highlight the ones at the heart of this book, the expansion of the country’s knowledge-based assets and its link to high-tech FDI. The success of the R&D based strategy depends critically on the country’s ability to make the move up the value chain across the board. And that, in turn, depends fundamentally upon advancing simultaneously in a number of different areas: TNCs’ interests in moving up the value chain—especially in bringing more R&D intensive production to Ireland; further broadening and deepening of indigenous technological capabilities; the ability of the educational system to generate the requisite number of computer science and engineering graduates, ongoing improvements in infrastructure, and Science Foundation Ireland’s success in pushing research and development in key biotechnology and ICT areas. In 1998, when the Irish economy was booming, the Irish Minister of Science and Technology requested a science and technology foresight study to identify the bases for future growth of the Irish economy. The results of the study were released in 1999.32 The study identified ICT and biotechnology as the key engines of future global growth. It argued that Ireland was well positioned to benefit from a strategy that focuses on biotechnology and ICT, given its recent history in pharmaceutical, computer, software, and medical instruments production.33 But the report identified the critical element missing for Ireland’s successful participation in the next growth wave: topnotch research capability in selected niches of these two technologies. As a latecomer, it is not surprising that Ireland’s research efforts should be lagging in comparison with other industrialized countries, especially since there is no indigenous defense industry. R&D expenditures as a share of GDP increased considerably over the last 20 years, from .7 percent to

DEVELOPMENT OF KNOWLEDGE-BASED ASSETS

127

1.2 percent, but they are still considerably below the EU and OECD average (see table 4.11(a)). The gap is not as large, but still existent, when we use GNP rather than GDP as the point of reference.34 The number of researchers is relatively smaller than in the average EU and OECD country as well, and a much smaller percentage of R&D expenditures is devoted to basic research (see table 4.11(b) and (c)). We saw earlier that the research efforts of TNC affiliates and of indigenous companies in Ireland have been growing, but are still limited, due to the relatively low research intensity of the former and the small absolute size of the latter. In addition, surveys indicate considerable barriers to innovation by businesses. High perceived economic risk, excessive costs and lack of financing sources are cited most frequently as barriers (see table 4.12). As a result, it is not likely that the requisite jump in research

Table 4.11 Ireland’s R&D Efforts in Comparative Perspective (a) Gross domestic expenditure on R&D as a % of GDP 1981 1985 1991

1995

2001

Ireland EU

0.68 1.69

0.77 1.86

0.93 1.90

1.28 1.80

1.17 1.93

Total OECD

1.95

2.26

2.23

2.1

2.33

(b) Researchers per 10,000 labor force by sector of employment Business Higher enterprise Government education

Overall

Ireland 1991 1999

15.7 31.3

2.6 1.8

18.3 13.5

44.0 49.0

EU 1991 1999

22.1 26.8

7.1 7.4

14.4 18.3

47.0 56.0

Total OECD 1991 1999

35.0 39.9

5.3 5.2

12.9 16.5

56.0 64.0

(c) Basic research as a percentage of total R&D activities and as a percentage of GDP, 1995 As a % of all R&D activities Ireland United States Hungary

12.0 16.1 27.9

As a % of GDP 0.12 0.40 0.18

Sources: (a) OECD (2003, Table A.2.1, 164); (b) OECD (2003, Table A.5.2, 172, and Table A.9.2.1, 175); (c) OECD (2003, Table A.6.3.1, 173).

128

FOREIGN INVESTMENT, DEVELOPMENT, AND GLOBALIZATION

Table 4.12 Factors Hampering Innovation, 1998–2000 (Percentage of Companies Surveyed Highlighting the Factor)

Excessive perceived economic risks Innovation costs too high Lack of appropriate sources of finance Organizational rigidities within the enterprise Lack of qualified personnel Lack of information on markets Insufficient flexibility of regulations or standards Lack of customer responsiveness to new goods

Manufacturing

Services

26 31 33 24 26 19 21 22

34 41 42 26 28 23 36 29

Source: Forfas (2003c, Figure 10.1, 46).

capability could be generated by the private sector alone. Based on the technology foresight reports, the government created a technology foresight fund of €646 million, and established the SFI to administer the fund and spearhead the strategy of R&D advancement. SFI started as a sub-board of Forfas in 2000, but became a separate legal entity in August 2003. The SFI has concentrated on awarding highly competitive grants in the two prioritized areas to top researchers and research teams with international recognition and on promoting research cooperation between universities and the private sector through the establishment of Centers of Excellence located at Irish universities. To realize its goals in the shortest time possible, the SFI does not limit itself to Irish research talent. Rather, since it was founded, it has been intent on bringing the best international talent in selected research fields to Ireland. Attraction of foreign research excellence, promotion of indigenous talent, and focus on the industrial applicability of research results are at the core of the SFI strategy (SFI 2002). The import of brainpower started with the organizational structure of the SFI. The head of the SFI, Bill Harris, is a former director of the National Science Foundation in the United States, where he was responsible for building the science and technology programs which connected the NSF with industry, and also a former vice president for research at the University of South Carolina. The first director of the ICT section, Alastair Glass, was a former vice president of photonics research at Bell Labs, where he played an instrumental role in changing the institution’s identity from working in the ivory tower to working with industry.35 Why would topnotch foreign researchers want to come to Ireland? The hope of the SFI is that the size of the available research grants, together

DEVELOPMENT OF KNOWLEDGE-BASED ASSETS

129

with the excitement of being in it from the start, and the desirability of Ireland as a place to live provide a sufficient draw for foreigners. While the SFI can show some impressive initial achievements only time will tell whether or not it will be able to realize its ambitious goals in biotechnology and ICT. The goals for 2007 include the recruitment of 50 highpowered researchers or research teams to Ireland, the expansion of the country’s research infrastructure, and the establishment of centers with important research links between Irish universities and companies in Ireland (indigenous and foreign-owned). The SFI has already generated some first high-profile links and collaborations. It succeeded in bringing in a whole high-powered research team in photonics from Ipswitch/England to the University of Cork, where it promised to build a new facility (Murray Brown 2003). Siemens decided to establish a new research facility at the Royal College of Surgeons in Ireland, where researchers from the two entities will collaborate in the fields of biotechnology, bioinformatics, and pharmaceuticals. And Lucent Technologies’ Bell Labs has announced the establishment of two research and development centers in Ireland. At Lucent’s facility in Dublin, one center will be the global headquarters for supply-chain technologies. And at Trinity College, Dublin, another establishment will focus on telecommunications value-chain-driven research, in collaboration with nine leading Irish institutes and universities.36 In recent years a growing amount and share of direct foreign investment in Ireland has gone into internationally traded services (e.g. Google and eBay opened operations in 2004), into the Financial Services Center, and toward the establishment of call centers and centralization of TNCs’ back office services in Ireland for TNCs’ regional or global operations. Some TNCs in the high-tech manufacturing sector have left Ireland or relocated the standardized parts of the production process out of Ireland. Some have expanded their services operations in Ireland, from technical support centers to financial administration to logistics. A few TNCs have kept manufacturing in Ireland, and have been able to remain competitive through ongoing reorganization of the supply chain and automation. An example is Dell, which in contrast to Apple, Gateway, and IBM is still making personal computers in Ireland. In fact, Dell’s plant in Limerick is the company’s only manufacturing plant in Europe. The company carries only enough inputs for two hours of production, with the result that most components go directly from the container to the assembly line. The company has a full-time staff of over 50 whose job it is to find ways of improving the process (Murray Brown 2004, 6). And some companies have indeed moved up the value chain in their Irish operations, expanding into areas of greater value-added production

130

FOREIGN INVESTMENT, DEVELOPMENT, AND GLOBALIZATION

and research intensity. In the biotechnology pharmaceutical field, Wyeth is building one of the largest integrated biotechnology campuses in the world in Ireland, with an investment of more than $1.5 billion. And in the ICT area, the most important TNC decision in support of Ireland’s R&D based strategy has come, yet again, from Intel. Since its first investment in Ireland in 1990, Intel had invested a total of $4.5 billion in the country, when it announced in June 2004 plans for an additional $1.6 billion investment in a new plant at its Leixlip facilities, FAB 24–2. The first Intel plant outside the United States to use 65-nanometer technology, FAB 24–2 is scheduled to open in 2006. It remains to be seen how many TNCs will follow Intel’s example this time around. The main reasons for the move toward much more high-value-added investments by these and other TNCs are familiar from our analysis in the last chapter. First there is the availability of highly skilled labor, though with concerns about future supply in specific areas. Second, low tax rates have continued, and those are particularly attractive, the higher the valueadded activities. Now, companies have the added stimulus of a tax credit for R&D as well as for headquarter offices for white-collar operations. Thirdly, IDA grants continue, with a particular focus on supporting R&D. The IDA is said to have paid $120 million to tip Intel’s decision on where to locate FAB24–2. And finally, but equally importantly, is the full-fledged government support for R&D activities across the board and the proactive approach of the requisite government institutions. And how will indigenous companies fare in this next phase of Ireland’s economic development? Indigenous producers in high-tech industries can survive and thrive, if technological, management, and marketing capabilities have developed sufficiently to make the jump to greater R&D-intensive production possible. We see in this chapter that indigenous knowledgebased assets expanded under the Celtic Tiger. In many cases it happened through linkage developments with high-tech TNC and government support programs, for example in the electronics and medical instruments sectors. In other cases, it happened through positive spillovers more generally. And in still other cases, expansion of technological capabilities and output was primarily homemade. That seems to have been especially true in the software industry where niche production and very low barriers to entry allowed indigenous computer-skilled entrepreneurs to thrive. Linkage opportunities in the electronics industry will be shrinking, while they are likely to persist for a longer period in the medical instruments sector. In the electronics industry, ever-shorter product cycles and cut-throat global competition, together with the standardization of the key components have led to increased outsourcing of component production (or for some OEMs of the whole manufacturing process) to countries with lower-wage costs

DEVELOPMENT OF KNOWLEDGE-BASED ASSETS

131

for semiskilled workers. Limitations of size and lack of multidimensional capabilities have forced some Irish input producers to shut down. Others have been making the jump to becoming multinational companies, and still others are refocusing from computer companies as their clientele to companies in the medical instruments sector. The medical instrument sector offers greater opportunities to Irish producers at this juncture. It is in the very nature of the sector that there are no standardized components across the industry, that component production can take place on a relatively small scale, but with very high quality expectations. High regulatory standards and the traditional fragmentation of the supplier base in the medical-device sector will keep a rapid trend toward CMs and relocation to lower-wage countries at bay over the next few years. And growing price pressures on TNCs in the sector might make TNCs more willing to look at Irish input suppliers, especially given their technological capabilities. That potential willingness together with indigenous linkage capability offers a window now for further expansion of linkages, especially higher up the value chain and especially as the indigenous companies are developing a more regional presence. How long the window of opportunity will remain open depends, among other things, on the speed with which CMs can overcome the challenges to large-scale outsourcing in the industry, and on the time that lower-wage countries will need to establish a credible guarantee that indigenous producers meet high quality standards. I noted earlier that the IOEs in the medical instruments sector stand out for their high productivity growth and research intensity (R&D expenditures per employee), both of which indicate some potential for moving along the supplier-oriented path of industrial upgrading. But it is still a small sector, and one that is not necessarily representative of the rest of the indigenous manufacturing and services industry. In a recent assessment of Ireland’s place in the global economy, the O’Driscoll Report (Enterprise Strategy Group 2004) argued that Irish companies have acquired significant expertise in manufacturing operations, but that they are still lacking at the front and back end of the supply chain process, namely R&D and technology on the one hand and sales and marketing on the other. They recommend an even greater focus within EI on supporting technology and R&D development in indigenous companies. One of the greatest challenges for Irish companies is the limitation imposed by small size, not country size, but company size. As possibilities for indigenous linkages become smaller, it is paramount for Irish companies in high-tech sectors to start selling a greater share of their output in the international market, in addition to colocating with their TNC clients, as we saw earlier in the context of the electronics industry. In developing

132

FOREIGN INVESTMENT, DEVELOPMENT, AND GLOBALIZATION

the marketing expertise to conquer the international market, small size is a real barrier. To overcome this obstacle, EI needs to intensify its efforts in helping companies to overcome information deficits and reduce risks, whether it is related to linkages, as in the International Business Linkage Division or to exports more broadly. An additional way to overcome the size constraint is through the promotion of cluster development and networks among smaller indigenous companies, whether in manufacturing or in the service sector. Government policy has to focus more fully on advancing the knowledgebased assets of the indigenous manufacturing sector in a more explicit and concentrated form. The Telesis Report of 1982 chastised the Irish government for putting excessive emphasis on TNCs at the expense of the development of the indigenous industry. In response the NLP was established, and it was rather successful, especially at that conjunctural moment. The Culliton Report of 1992 criticized the government for insufficient emphasis on improving the overall business conditions, and the establishment of EI with broader company support was the outcome. Informed by production relocation of standardized electronics products and the need to move up the value chain in Ireland, the O’Driscoll report of 2004 points out that indigenous industry has come a long way in becoming competitive in production, but that it lacks in R&D and marketing expertise. If the government had given the support the O’Driscoll Report demands now to indigenous producers 10, 20, or 30 years ago, indigenous knowledgebased assets could well be further along than they are today. It is unrealistic to expect that companies could have gone from significant deficiencies in competitiveness in production process, quality and delivery standards to competitiveness in design and development without going through the necessary learning process. Nonetheless, a much greater and concentrated focus on cluster formation and networks with their synergies and agglomeration benefits for indigenous companies (in addition to international corporations) should have come earlier. The next five years will be critical in Ireland’s pursuit of a new R&Dbased strategy. Key elements are in place for its success, with the initiatives pursued by SFI, some large and important high-tech TNCs moving up the value chain in Ireland with billion dollar investments, and recognition that more direct and targeted support for technology and marketing development for indigenous enterprises is crucial. There is a recognized need to increase the number of students studying science subjects and the number of specialized engineers (EGFSN 2003). And the improvement of the physical infrastructure, particularly the roads program, forms an important part of the National Development Plan. To move forward on all these fronts is a daunting challenge, especially at a time when external financing will become more restricted.

DEVELOPMENT OF KNOWLEDGE-BASED ASSETS

133

At the same time, development policies over the next decade need to address the distributional inequities that the Celtic Tiger has perpetuated and maybe increased. Ireland’s experience in the 1990s shows that high economic growth does not necessarily lead to a reduction in income inequality. Thus, deliberate government policies are needed to achieve improvements in equality and social welfare more generally.

This page intentionally left blank

5

From Coffee to Computers: High-Tech FDI in Costa Rica

s the so-called Switzerland of Latin America Costa Rica has long held a special position among Latin American countries, as it successfully combined economic growth and social advancement. Unlike the other Central American countries with their history of high inequality and dictatorships, Costa Rica built a more social–democratic society in the twentieth century, providing broad access to health care and education.1 Yet, like its neighbors, Costa Rica has traditionally depended on the exports of bananas and coffee, and later clothing, as the main force behind economic growth. Intel’s decision in 1996 to establish a microchip factory outside of San José marked a watershed in Costa Rica’s development path. It highlighted and opened up the potential for profound structural change toward hightech FDI-led economic growth and development. By 2004, this potential has only been realized incipiently. High-tech FDI brought about significant structural change in the manufacturing sector. Yet, economic growth has continued to fluctuate and a dual manufacturing sector has developed with a stagnant national sector and a growing high-tech foreign-owned sector, though with mixed performance within each of them. While the computer-related sector has proved to be rather susceptible to the ups and downs of the global business cycle, the medical instruments sector has shown greater immunity to it. Linkages with national producers have been very limited though in both cases. In this chapter and in chapter 6, I analyze why high-tech investment has not unleashed a Tico Tiger. I argue that the legacy of Costa Rica’s past policies, especially in education, in conjunction with current efforts to attract FDI led to high inflows of high-tech FDI in the late 1990s, in a global context

A

136

FOREIGN INVESTMENT, DEVELOPMENT, AND GLOBALIZATION

where U.S. TNCs were looking for more cost-effective production sites outside U.S. borders. In addition to Intel, several world-class medical supply producers established production facilities in Costa Rica. Yet, sustained high-tech FDI-led growth has not happened to date. There are a number of reasons for that, some outside the country’s control, and others the result of domestic structural and institutional problems. First, it takes time for an FDI-led strategy to translate into sustained higher growth, since it takes a cumulative process for such a strategy to bear fruit. Second, some external factors and forces have not been conducive to increasing growth. The global downturn in the IT industry at the turn of the century hit just at the time when Intel had started production in Costa Rica putting a brake on Intel’s output expansion. Another important external obstacle has been the absence of the equivalent of European Structural Funds for Costa Rica. Rather than receiving external development grants, there was actually a net outtransfer of development funds from Costa Rica during most of the 1990s. The low level of funds is a more significant impediment to sustained growth than the decline in Intel’s output, because it is a structural, and not cyclical, problem. Significant net inflows of development funds would make it a lot easier to improve the country’s human capital and physical infrastructure. The final set of barriers to sustained growth is homemade. The lack of a cohesive network of proactive government institutions to implement a comprehensive development strategy, the absence of a significant institutionalized effort to build national linkage capabilities and promote cluster formation, and a very low tax ratio, which does not generate sufficient resources for the required advancement in education and infrastructure, are deep-rooted structural problems. They may be the hardest to overcome, since they are embedded in the political–economic constellation of power that has evolved after the dismantling of the import substitution model during the 1980s.

Background on Costa Rica’s Development Policies Latecomers in the development process have historically depended on one or two major commodities to generate foreign exchange for economic growth. Costa Rica is no exception to this mono-export dependency with all the well-known implications for foreign exchange volatility and economic growth. In the mid-nineteenth century, coffee started to dominate Costa Rica’s production and exports. The fact that the country’s major cash crop was largely cultivated by small farmers and not on large estates,

FROM COFFEE TO COMPUTERS

137

like in many other developing countries, has been an important factor behind the relatively equal distribution of income in Costa Rica.2 In the late nineteenth century, bananas became the second major export commodity, with the construction of the Atlantic Railroad and the arrival of the United Fruit Company. The 1940s constituted a critical juncture in Costa Rica’s political and economic development. Following a brief civil war, the temporary junta under José Figueres (1948–1949) legislated major reforms with farreaching implications for Costa Rica’s subsequent development path. The creation of autonomous institutions enhanced the capability for effective policy-making and implementation, particularly after the introduction of a merit-based civil service in 1953. The nationalization of the banking system gave the government extraordinary influence over resource allocation, which became especially significant in the 1970s, when the government became a major producer in its own right. The suppression of strong unions led to a structurally weak position of labor in the private sector, with solidarity organizations (comparable to company unions) becoming the principal form of worker organization. And the abolition of the national army freed up resources for investment in education, health, and housing by successive governments.3 Costa Rica embarked upon ISI at a time when Ireland had just abandoned it. A balance of payments crisis in the late 1950s led to the adoption of protective measures for Costa Rican manufacturing. But ISI only became possible in the regional context, when Costa Rica joined Honduras, Guatemala, El Salvador, and Nicaragua in the early 1960s in the Central American Common Market (CACM).4 With manufacturing value added growing at high rates during the 1960s and 1970s, the share of manufacturing in GDP increased from 13.2 percent in 1960 to 22 percent in 1979, while manufactured exports as a share of total exports grew from 2.4 percent to around 30 percent (Buitelaar, Padilla Pérez, Urrutia-Álvarez 2000, 12). Under the regionally based ISI strategy, the Costa Rican economy grew at impressive rates. Real GDP increased at an average annual rate of 5.9 percent during the 1960s and 6.4 percent during the 1970s.5 Yet by 1980, the development strategy had run into trouble, as the CACM had collapsed with civil wars in Nicaragua, El Salvador, and Guatemala, while foreign debt had accumulated to potentially unsustainable levels. In the wake of the 1973 oil crisis, the government had easy access to international loans, which it used to increase public investment as well as social spending.6 With the establishment in 1972 of the Costa Rican Development Corporation (Corporación Costarricense de Desarrollo Sociedad Anónima, CODESA), the Figueres Administration expanded the role of the government

138

FOREIGN INVESTMENT, DEVELOPMENT, AND GLOBALIZATION

from regulator to producer. But the proliferation of new enterprises with insufficient regard for efficiency foreshadowed economic problems and generated increasing political opposition.7 The 1979 development plan of the Carazo government had proposed a change in strategy, with more emphasis on agro-exports and a reduced role for the state (Wilson 1998, 103).8 The foreign debt crisis in the early 1980s highlighted the need and provided the occasion for a change in development policies. With banana and coffee prices plummeting and interest rates on foreign debt soaring, Costa Rica declared a moratorium on its foreign debt in July 1981. Debates over the appropriate development direction during the Monge Administration (1982–1986) were resolved in favor of a gradual adoption of the neoliberal Washington Consensus, with its emphasis on a sharply circumscribed role for the state, broad market liberalization and a full integration into the international economy.9 Initially, government policies focused primarily on economic stabilization (1981–1985). The precipitous decline in economic growth was halted due to the combination of a sharp devaluation of the peso, a restructuring of the foreign debt, a cut in government spending and public-sector employment, as well as new funding from international organizations, especially large inflows from the United States Agency for International Development (U.S.-AID). After GDP had declined by 10 percent between 1980 and 1982, the economy started growing again in 1983 (see table 5.1). Economic reforms under the new economic model were introduced gradually, and not through shock therapy. The financial sector was liberalized, leading to rapid growth of private banks, and—with the liberalization of the capital account in 1992—to a renewed increase in foreign debt in the late nineties. CODESA was abolished in 1990 and state-owned companies were privatized. One important exception is the electric company ICE (Instituto Costarricence de Electricidad). Political opposition—primarily in the streets—has so far prevented attempts to privatize it, most recently under the Rodríguez Administration (1998–2002).10 Reforms aimed at changing the relative prices in Costa Rica vis-à-vis the international economy unfolded slowly and, at times, in a way that did not conform to the Washington Consensus, but was more reminiscent of export promotion under ISI. On the import side, nontariff barriers were converted into tariff equivalents and then reduced progressively. The average tariff rate fell from slightly over 60 percent in 1985 to 11.7 percent in 1995 to 5.8 percent in 2004.11 It was on the export side that Costa Rican policies were more heterodox, as the government increased rather than reduced its involvement. In order to promote nontraditional exports, the government offered subsidies in the form of tax credits (CATs, Certificados de Abono Tributario).

139 Table 5.1

Basic Economic Data: Costa Rica, 1970–2002 Net FDI inflows

Year

Millions of current US$

% of GDP

GFCF as a % of GDP

Real GDP growth based on constant pesos

Investment rate % of GFCF

Average 1960s

–

–

–

17.5

5.9

1970 1971 1972 1973 1974 1975 1976 1977 1978 1979 Average 1970s

– 22 26 38 46 69 62 63 49 44 44

2.7 2.1 2.1 2.5 2.8 3.5 2.6 2.0 1.4 1.1 2.3

13.8 9.3 9.5 11.1 11.6 16.0 11.0 9.1 6.0 4.1 10.1

19.5 22.1 21.9 22.2 24.0 22.0 23.4 22.4 23.0 26.2 22.7

7.4 6.9 8.2 7.8 5.6 1.9 5.5 8.9 6.3 5.0 6.4

1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 Average 1980s

53 70 29 61 56 70 61 80 122 101 70

1.1 2.7 1.1 1.9 1.5 1.8 1.4 1.8 2.7 1.9 1.8

4.6 11.0 5.5 10.7 7.6 9.2 7.4 9.0 14.0 9.4 8.8

23.9 24.1 20.3 18.0 20.0 19.3 18.7 19.8 18.9 20.5 20.3

0.7 ⫺2.4 ⫺7.3 2.9 7.9 0.7 5.5 4.8 3.4 5.6 2.2

1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 Average 1990s

163 178 226 247 298 337 427 408 613 620 352

2.8 2.5 2.6 2.6 2.8 2.9 3.6 3.2 4.4 3.9 3.0

12.7 14.0 13.4 12.5 14.4 15.1 21.0 17.6 21.3 21.8 16.0

22.4 17.8 19.7 20.5 19.5 19.0 17.1 18.0 20.4 18.0 19.0

3.5 2.3 9.1 7.6 4.9 3.7 0.9 5.3 8.4 8.5 5.0

2000 2001 2002

409 454 662

2.6 2.8 3.9

14.4 15.3 20.6

17.8 18.1 19.1

1.4 1.4 2.8

Source: World Bank World Development Indicators ⬍http://devdata.worldbank.org/dataonline⬎ accessed July 2004.

140

FOREIGN INVESTMENT, DEVELOPMENT, AND GLOBALIZATION

In 1984, the government established the export contract and increased the CAT from 15 to 30 percent of export value. It also modified the Temporary Admission Regime (Régimen de Admisión Temporal), which had existed since 1972 and allowed producers to import duty-free inputs used in the production of exports, as long as they exported all their production to non-regional markets. Although the CAT was successful in stimulating non-traditional exports, it led to over-invoicing and fictitious exports as well as an increasing fiscal burden. As a result, CATs were reduced in the early 1990s and abolished in 1999.12 In 1996, the Temporary Admission Regime and the Export Contracts were replaced by the Active Processing Regime (Régimen de Perfeccionamiento Activo, PA), whose characteristics were similar to the Temporary Admission Regime. It allows for duty-free import of inputs used in the production of exports. In the early 1980s, Costa Rica also established free zones (FZs) with the goal of attracting foreign investment and fostering nontraditional exports. After the initial failure of a free-zone area outside Limón, legislation was expanded in 1984 to allow for private operators of FZs and in 1990 to include service provision as well. FZs offer duty-free imports and—in contrast to the Temporary Admission Regime—also a variety of tax exemptions, most importantly for profits. In 1996, there were 196 firms in the FZs, up from seven in 1986. The largest number was in textiles and footwear (48), followed by services (22), and electronics (20) (Ketlhoehn and Porter 2002, 19). Costa Rica cemented its commitment to full integration into the world economy, when it joined the General Agreement on Tariffs and Trade (GATT) in 1990. By the mid-1990s, Costa Rica had implemented Washington Consensus policies to about the same extent as the average across Latin American countries. Morley, Machado, Pettinato (1999) calculated indices of the degree of reform implementation in key areas. Each index ranges from 0 to 1, with a higher number indicating a greater extent of reform implementation. In 1995, the value of Costa Rica’s general reform index was .848 compared to .821 for the Latin American average. The IMF and the World Bank played a crucial role in the adoption of Washington Consensus policies by developing countries, which dealt with a debt and foreign exchange crisis in the early 1980s. Costa Rica is no exception. It entered a standby agreement with the IMF in 1982 and two structural adjustment programs with the World Bank, one in 1985, the other in 1987. But the international institution, which had arguably the most decisive influence on shaping Costa Rica’s development path in the 1980s, is the U.S. Agency for International Development—U.S.-AID. When the Sandinistas succeeded in overthrowing the Somoza dictatorship in Nicaragua in 1979, and civil wars continued in El Salvador and Guatemala, Central America assumed heightened strategic importance for

FROM COFFEE TO COMPUTERS

141

the conservative U.S. Reagan Administration. Costa Rica was assigned a key role in U.S. foreign policy, given its strategic location as a democratic country, which in the South neighbored Panama with the Canal, and in the North provided a potential base for the Contras, who were trying to overthrow the Sandinistas. Aiming to turn Costa Rica into an economic success story, the U.S. government provided massive resources and promoted policies in Costa Rica, which were deemed likely to generate positive economic outcomes. Costa Rica’s need for foreign capital to service its foreign debt in conjunction with the debate over the direction of development policies under the Monge Administration provided the perfect opening for U.S.AID’s powerful entry on the Costa Rican stage.13 Between 1982 and 1990, U.S. aid to Costa Rica totaled $1.34 billion, equivalent to a remarkable 3.15 percent of Costa Rica’s GDP over that period (see table 5.2).14 Financial aid from all sources surged to 9 percent of GNI in 1983, after being consistently below 2 percent in the previous two decades.15 The relative magnitude of U.S. aid to Costa Rica during the Table 5.2 U.S. Aid to Costa Rica, 1980–2001 Year

In millions of current US$

As a % of Costa Rica’s GDP

1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001

15.9 15.2 51.8 214.1 169.8 220.1 162.7 181.2 120.3 121.9 95.3 44.9 26.7 27.6 12.1 6.2 2.1 0.1 0.7 1.1 0.5 0.5

0.6 0.6 1.6 5.8 4.3 5.0 3.6 3.9 2.3 2.1 1.3 0.5 0.3 0.3 0.1 0.1 0.0 0.0 0.0 0.0 0.0 0.0

Source: U.S.-AID (2004), Costa Rica’s nominal GDP data from World Development Indicators, online, accessed July 2004.

142

FOREIGN INVESTMENT, DEVELOPMENT, AND GLOBALIZATION

1980s was similar to the inflow of European Structural Funds to Ireland during the 1990s. Yet, the similarity in the relative size of external funding stands in stark contrast to the different uses to which these funds were put in the two countries. Where the Irish government used them for the development of infrastructure and education, AID funding in Costa Rica was used to support economic stabilization and to build up new institutions. Initially, a good part of U.S. aid was used for balance of payments support, which enabled Costa Rica to come out of the recession in the early 1980s sooner than most other Latin American countries and to do so while mitigating some of the social costs.16 Other funds were used to channel credit to the private sector, a key element in the rise of the private banking sector in the second half of the 1980s. And still other aid was used to fund new institutions conceptualized by AID, most importantly in 1982 the Costa Rican Coalition for Development Initiatives, CINDE (Coalición Costarricense de Iniciatívas para el Desarrollo), in charge of attracting FDI. AID was also instrumental in establishing a Ministry for Exports, MINEX (Ministério de Exportaciones) in 1986, and paid for all its expenses during the first three years (Clark 1997). Building CINDE outside the existing structure of government institutions allowed continuation in policy direction irrespective of the government in power, but it also contributed to institutional disarticulation, which constitutes a major challenge for FDI-led development in Costa Rica in the early twenty-first century. In addition, AID funding also supported the “voluntary labor mobility program” in the early 1990s, which led to a substantial reduction in public sector employment. Equally important, it led to the dismantling of a high-quality civil service, as the most qualified workers were most able to avail themselves of opportunities in the private sector.17 The strategic role of Central America for the Reagan Administration was also the driving force behind a new U.S. trade agreement for the Caribbean, the Caribbean Basin Initiative (CBI), which went into effect in 1984.18 The trade initiative allowed beneficiary countries to export a range of non-traditional commodities to the U.S. market without facing tariffs, so long as 35 percent of the product value was produced in one or more of the beneficiary countries. Most textiles and footwear were, however, explicitly excluded from the CBI. That changed only in 2000 with the United States–Caribbean Basin Trade Partnership Act, the extension of CBI. It corporated products that had previously been excluded, most importantly clothing, and loosened the rules on U.S. inputs to be used. Another U.S. trade preference system, which was equally important for the development of Costa Rica’s nontraditional exports, was the shared production regime under regulation 807 of the U.S. tariff system, later HTS 9802 provisions. It allowed for the reimport of U.S. materials assembled abroad, with duty paid only on the value added outside the

FROM COFFEE TO COMPUTERS

143

United States. A response to increased foreign competition in labor-intensive industries, these tariff provisions allowed U.S. producers to outsource the labor-intensive parts of production processes to countries with lower labor costs, for example the assembly of precut fabric into clothing.19 In 1993, 84 percent of Costa Rica’s apparel exports were exported to the United States under HTS 9802.00.80 (Mortimore and Zamora 1999, 67).20 In the context of U.S. trade preferences via HTS 9802 and CBI, Costa Rica’s export-promotion policies led to a rapid increase in the country’s nontraditional exports. Nontraditional exports are all exports except coffee, bananas, sugar, and meat. They grew from an annual average of $390 million in the first half of the 1980s to $584 million in the second half, and $1.14 billion in the first half of the 1990s. As a result, nontraditional exports increased their share in total exports from 41 percent to 59 percent over this period (Mortimore and Zamora 1999, 31). Surprisingly, structural change in the manufacturing sector was fairly moderate during this time span, at least at the 2-digit ISIC level.21

FDI and Economic Performance Foreign direct investment inflows grew considerably under the new economic model. After slow growth during the 1970s and 1980s, net FDI inflows to Costa Rica grew substantially faster in the first half of the 1990s, and then reached unprecedented levels with and after Intel’s investment in 1997 (see table 5.1). Net FDI inflows grew from an annual average of $44 million in the 1970s to $70 million during the 1980s, and then $352 million in the 1990s. In most Latin American countries, increased FDI inflows in the 1990s were the result of large-scale privatization efforts, particularly of utilities. That was not the case in Costa Rica. While the largest share of FDI went to agriculture and services in the 1980s and early 1990s, the manufacturing sector became the main destination for FDI in the later 1990s, just as it had been under ISI in the second half of the 1970s.22 Between 1997 and 2003, 65 percent of FDI inflows went to industry (see table 5.3). About a third of the investment in industry was in the electronics sector.23 While there had been foreign investments in the electronics industry before, most notably Motorola in the 1970s and DSC Communications Corporation in 1995, the big jump in FDI in electronic equipment came with Intel’s $300 million investment. In the medical-device industry, Abbot was the biggest investor in the late 1990s. Even though Baxter had been in Costa Rica since 1987, it was only with Abbot’s investment and the publicity following Intel’s investment that FDI in the medical-supply sector surged. The tourist sector was the second most important target for FDI. Foreign companies invested in hotels and related services, as Costa Rica

144

FOREIGN INVESTMENT, DEVELOPMENT, AND GLOBALIZATION

Table 5.3

Net Foreign Direct Investment in Costa Rica by Sector, 1997–2003 1997

1998

1999

2000

2001

2002

2003*

Millions of US$ Agriculture Agroindustry Commerce Industry Services Financial system Tourism Other

38.1 6.5 17.6 270.6 ⫺7.3 ⫺0.2 79.3 2.3

41.9 14.7 39.3 423.5 6.6 22.1 61.4 2.1

49.9 10.4 9.2 355.9 12.7 93.4 84.7 3.3

⫺11.2 11.5 17.4 296.2 14.6 27.1 52.1 0.9

1 5.2 8.3 231.4 58.9 36.3 111.5 1

⫺8.6 2.8 15.2 482.7 52.1 19.8 76 22

⫺29.8 10.6 0.5 365.5 111.1 10.1 88.3 39.6

81.3 61.7 107.5 2425.8 248.7 208.6 553.3 71.2

Total

406.9

611.6

619.5

408.6

453.6

661.9

586.9

3749.0

Percentage breakdown Agriculture 9.4 Agroindustry 1.6 Commerce 4.3 Industry 66.5 Services ⫺1.8 Financial system 0.0 Tourism 19.5 Other 0.6

6.9 2.4 6.4 69.2 1.1 3.6 10.0 0.3

8.1 1.7 1.5 57.4 2.1 15.1 13.7 0.5

⫺2.7 2.8 4.3 72.5 3.6 6.6 12.8 0.2

0.2 1.1 1.8 49.4 12.6 7.7 23.8 0.2

⫺1.3 0.4 2.3 72.9 7.9 3.0 11.5 3.3

⫺5.1 1.8 0.1 62.3 18.9 1.7 15.0 6.7

2.2 1.6 2.9 64.7 6.6 5.6 14.8 1.9

100.0

100.0

100.0

96.8

100.0

100.0

100.0

Total

100.0

1997–2003

* Preliminary Source: Ministerio de Comercio Exterior, Costa Rica ⬍http://www.comex.go.cr/estadisticas/inversion/ default.htm⬎ accessed July 2004.

became an increasingly well-known tourist attraction, with beautiful beaches on the Pacific and Atlantic and ample opportunities for ecotourism. In the service sector, the big FDI projects went to the establishment of call centers (e.g. Skyes Latin America, established in 1999 and Western Union), and administrative services for global operations (e.g. Procter & Gamble, 2000). As in the past, U.S. corporations have been the largest foreign investors. Between 1997 and 2003 U.S. FDI accounted for 63 percent of all inflows, followed by Mexico with 7.1 percent and Panama with 5.3 percent (Ministerio de Comercio Exterior, Costa Rica). The vast majority of FDI has been greenfield investment. A notable exception is Heineken’s purchase of a stake in the Costa Rican beverage company Florida Bebida in 2002, which made up nearly a third of the country’s FDI inflows in that year (Banco Central de Costa Rica 2003, 3). In comparative perspective, FDI flows to Costa Rica during the 1990s were relatively less important than in the rest of the world. UNCTAD’s FDI Performance Index, which measures a country’s share in global FDI relative to its share in global GDP, fell from 2.6 in 1988–1990 to 1.0 in 1998–2000 in Costa Rica, while it remained unchanged at .99 for developing countries overall (UNCTAD 2002, 25, 27). But given its status as a middle-income

FROM COFFEE TO COMPUTERS

145

country, Costa Rica does stand out for its ability to attract high-tech FDI. At the end of the twentieth century, it ranked first among Latin American countries in the share of high-tech products in total exports.

Impact at the Macro Level With the increase in FDI, TNCs have become much more significant for investment and economic growth in Costa Rica. The share of FDI in gross fixed capital formation (GFCF) increased from an average of 10 percent in the 1970s to 16 percent in the 1990s (see table 5.1).24 Nonetheless, FDI’s increased participation in GFCF has not been matched by an increase in the overall investment share. During the 1990s the average investment share was 19 percent, slightly below the 20 percent rate of the 1980s, and well below the rate of 22.7 percent in the 1970s. At first glance, the movement of the foreign and national investment shares in opposite directions suggests that foreign investment crowded out national investment. But on closer scrutiny, the crowding out hypothesis does not hold up; correlation does not imply causation.25 The decline in the national investment share in the 1990s seems to have been primarily a reflection of national companies’ struggle against the larger competitive pressures of a more open economy. Direct competition from FDI played a much less important role, since most TNCs manufacture goods that are not produced by national companies. The decline in the public investment share under the new economic model is another factor behind the overall lower investment rate in the 1990s (and 1980s) compared to the 1970s.26 The increase in FDI has had positive implications for economic growth. About 45 percent of all TNC output comes from the Free Zones, and FZs have increased their participation in the Costa Rican output from .5 percent in the early 1990s to 8 percent in 2003 (Monge, Rosales Tijerino, and Arce Alpizar 2004, 15). Of singular importance is the production of Intel, the Goliath in the Costa Rican economy.27 In 1999, GDP grew by 8.5 percent, but when Intel’s production is excluded, it only grew by 3 percent. Intel’s powerful impact on economic growth works in both directions. The downturn in the world economy in 2000, exacerbated by the collapse of the IT boom, brought home the fact that demand for information technology-related output behaves pro-cyclically.28 Between 1996 and 1999, Costa Rica’s total exports increased by $2.9 billion, with FZ exports accounting for 102 percent of that increase. In the following two years, total exports dropped by $1.7 billion, 77 percent of which was due to a decline in FZ exports (see table 5.4). And by 2003, when exports had recovered, Intel’s exports made up 23 percent of Costa Rica’s exports (Monge, Rosales Tijerino, and Arce Alpizar 2004).

146

Table 5.4 Costa Rica’s Export Composition, 1991–2003 (in Millions of Current US$)

In millions Coffee Bananas Industrial exports* Active processing regime Free zones Total exports In percentages Coffee Bananas Industrial exports* Active processing regime Free zones Total exports

1996

1992

1993

1994

1995

263.6 396.6 518.4 266.6

201.6 562.6 664.0 391.5

201.6 764.8 726.5 485.2

307.6 561.0 845.4 420.8

417.3 385.4 402.3 409.4 288.7 272.0 161.8 165.1 193.6 680.2 631.1 577.3 667.5 623.5 546.5 516.0 477.5 553.1 951.8 1,107.2 1,121.0 1,244.4 1,134.7 1,090.2 1,029.3 1,023.2 1,057.3 485.2 378.8 427.2 444.5 396.1 398.9 366.4 354.1 331.7

145.0

234.1

273.6

343.4

434.6

643.0

1997

1998

1999

2000

2001

2002

2003p

1991

891.3 1,936.2 3,588.8 2,956.3 2,247.4 2,630.6 3,303.3

1,899.3 2,385.2 2,625.5 2,878.3 3,475.9 3,758.4 4,205.5 5,525.6 6,662.4 5,849.7 4,921.4 5,264.0 6,102.2 13.9 20.9 27.3 14.0

8.5 23.6 27.8 16.4

7.7 29.1 27.7 18.5

10.7 19.5 29.4 14.6

12.0 19.6 27.4 14.0

10.3 16.8 29.5 10.1

9.6 13.7 26.7 10.2

7.4 12.1 22.5 8.0

4.3 9.4 17.0 5.9

4.6 9.3 18.6 6.8

3.3 10.5 20.9 7.4

3.1 9.1 19.4 6.7

3.2 9.1 17.3 5.4

7.6

9.8

10.4

11.9

12.5

17.1

21.2

35.0

53.9

50.5

45.7

50.0

54.1

100.0

100.0

100.0

100.0

100.0

100.0

100.0

100.0

100.0

100.0

100.0

100.0

100.0

* Industrial exports do not include the Active Processing Regime and the Free Zones. p Preliminary Source: Banco Central de Costa Rica, ⬍http://websiec.bccr.fi.cr/indicadores/sector.web?sector⫽1⬎ accessed July 2004.

FROM COFFEE TO COMPUTERS

147

Exports of FZ companies other than Intel have increased at a steady rate since 1997. Fortunately for Costa Rica, international demand for other high tech goods has turned out to be much less susceptible to swings in the business cycle. Exports of medical devices have continued to grow steadily since the late 1990s, from $83 million in 1998 to $401 million in 2002.29 In spite of the overall fluctuations in export earnings, the surge in FDI in Costa Rica in the second half of the 1990s has led to a considerable increase in the country’s exports and export intensity. The share of exports of goods and services in GDP had grown from an average of 24.5 percent in the 1960s to 29.6 percent in the 1970s to 34.9 percent in the 1980s. By 2002, it had reached 42.4 percent.30 At the macro level, all developing countries face two critical constraints on economic growth: a foreign exchange constraint and a government revenue constraint. FDI has played an important role in alleviating the foreign exchange constraint in Costa Rica. While net foreign investment inflows and exports generate foreign exchange for Costa Rica, the repatriation of profits as well as imports of intermediate inputs and capital goods reduce foreign exchange. We do not have specific data for all these flows for TNCs in manufacturing, but based on reasonable assumptions, we estimate that the net foreign exchange contribution of TNCs in the manufacturing sector was $4.9 billion during the period 1991–2003 (see table 5.5).31 That is the equivalent of 65 percent of the cumulative current account deficit during the period. Table 5.5 Net Foreign Exchange Generation by Special Export Regimes*: An Approximation (in Millions of US$) Net FDI 65% of net Exports– Imports of inflows Profit FDI inflows— Exports of imports of equipment (all repatriation 65% of profit Special Special of Special sectors) (all sectors) repatriation Zones Zones Zones 1991 178.4 1992 226.0 1993 246.7 1994 297.6 1995 336.9 1996 427.0 1997 406.9 1998 611.7 1999 619.5 2000 408.6 2001 453.6 2002 661.9 2003 576.7 1991–2003 5,451.5

187.6 216.8 241.5 138.4 225.2 186.5 254.5 461.4 1,806.9 1,233.8 760.3 498.3 827.3 7,038.5

⫺6.0 6.0 3.4 103.5 72.6 156.3 99.1 97.7 ⫺771.8 ⫺536.4 ⫺199.4 106.3 ⫺162.9 ⫺1,031.6

411.6 625.6 758.8 764.2 609.8 1,021.8 1,318.5 2,380.7 3,984.9 3,355.2 2,713.8 2,984.6 3,635.0 24,564.5

⫺9.1 177.9 186.7 71.9 ⫺149.0 259.5 282.8 751.6 2,022.4 1,468.3 676.7 764.6 1,414.7 7,919.0

34.9 52.4 57.0 71.5 71.8 85.0 118.5 256.6 224.6 246.5 249.9 268.8 262.7 2,002.2

Net foreign exchange generation of Special Zones ⫺50.0 131.5 133.1 103.9 ⫺148.2 330.8 263.4 592.7 1,026.0 685.4 227.4 602.1 989.1 4,887.3

* Special Export Regimes include Free Zones and Temporary Admissions Regimes. Source: Calculated from Central Bank of Costa Rica, ⬍http://websiec.bccr.fi.cr/indicadores/sector.web? sector⫽1⬎.

148

FOREIGN INVESTMENT, DEVELOPMENT, AND GLOBALIZATION

FDI has done very little, however, to alleviate the government’s revenue constraint. In Costa Rica, the FZs offer a number of attractions to producers. In addition to providing 100 percent exemption from import duties and sales and consumption taxes on local purchases of goods and services, the free zone regime offers 100 percent exemption from profit taxes for eight years, and 50 percent exemption for another four years.32 Under these conditions, FZ corporations contribute to an increase in tax revenue not through profit taxes, but rather through income taxes paid by the employees. And that only constitutes a net increase in tax revenue, if those workers were not employed before or if they were paid higher wages than they would have been paid outside the FZs. In a cost–benefit analysis of FZs Monge, Rosales Tijerino, and Arce Alpizar (2004) identify salaries and social security contributions as the largest source of the positive contributions of FZs.33 The upshot, however, is that—in spite of the surge in FDI—the ratio of government revenue to GDP has remained fairly unchanged during the 1990s, at around 13 percent (see table 5.6).34 The low revenue ratio constitutes one of the major obstacles to Costa Rica’s future development, as it does not allow for the needed improvements in infrastructure and education. In contrast to Ireland, transfer pricing has not been a problem for Costa Rica, at least not judging from the national income accounts. Over the last 40 years, the ratio of GNI to GDP has generally been slightly below one. It Table 5.6

1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003

Government Revenue/GDP: Costa Rica, 1991–2003

Total government revenue/GDP

Customs tax/total government revenue

Income tax/total government revenue

Sales & consumption taxes/total government revenue

11.5 12.4 12.1 11.5 12.0 12.3 12.2 12.3 12.1 12.2 13.1 12.9 13.3

46.6 49.0 50.7 48.8 46.7 45.6 46.3 47.5 41.9 44.6 41.7 46.2 36.3

12.1 11.2 14.1 15.6 16.1 15.1 15.4 17.3 21.7 20.4 21.7 21.7 23.5

25.8 21.2 22.0 21.9 21.2 28.1 28.1 24.1 21.0 19.9 23.0 20.1 18.1

Other revenue/total government revenue 15.5 18.6 13.2 13.6 16.0 11.2 10.3 11.1 15.4 15.2 13.6 12.0 22.1

Source: Calculated from Central Bank of Costa Rica, ⬍http://websiec.bccr.fi.cr/indicadores/ sector.web?sector⫽1⬎ accessed August 2004.

FROM COFFEE TO COMPUTERS

149

dropped considerably in the 1980s with the increased interest payments on foreign debt, but then recovered again. And while it dropped to .88 in 1998, the first year after Intel started operations in Costa Rica, the ratio increased again to .97 in 2002. FZ expansion in Costa Rica has not only meant increased production, but also increased employment and higher wages. While employment in the FZs was only 7,000 in 1990, it had grown to 35,000 by 2002, 16 percent of the country’s total employment in manufacturing. In the late 1990s, the average wage in the FZs was 25 percent higher than in the non-FZ manufacturing sector. And for workers in the FZs’ high-tech enterprises, it was 35 percent higher, primarily due to the greater skill intensity (Monge, Rosales Tijerino, and Arce Alpizar 2004, 50).35 In sum, at the macro level, FDI, and especially high-tech FDI, has had several positive implications for economic growth. It has generated employment opportunities for more highly skilled workers at higher wages. It has had a positive impact on investment; though in the context of a falling national investment share, it meant that FDI prevented the domestic investment share from falling further, rather than increasing it further. And FDI has also made a significant contribution to alleviating the foreign exchange constraint. On the other hand, the ups and downs of Intel’s production have been mirrored to some extent in the ups and downs of Costa Rica’s GDP. FDI has not done much to alleviate the government revenue constraint though. It is quite possible that in the absence of tax exemptions high-tech foreign investors might not have invested in Costa Rica in the first place, and thus there is no issue of foregone tax revenue. But a small latecomer like Costa Rica is facing a real dilemma here. The country may need low tax rates to attract FDI, but it also needs revenue to expand its location-specific assets, whether that revenue comes from foreign investors, from international development agencies or from national producers and employees. FDI’s impact on the development potential of the country is not just the result of the macro outcome, but also, and in the long run more importantly, of the outcomes at the micro level. What FDI’s impact has been on the expansion of Costa Rica’s knowledge-based assets is the subject of analysis in chapter 6.

FDI, Trade, and Structural Change High-tech FDI in the 1990s has led to significant structural change in Costa Rica’s manufacturing sector. However, the change has manifested itself primarily in the changing composition of foreign-owned production,

150

FOREIGN INVESTMENT, DEVELOPMENT, AND GLOBALIZATION

which now dominates the national picture, and much less so in the nationally-owned manufacturing sector. In fact, over the last few years, there has been a growing dualism between the nationally-owned and the foreign-owned manufacturing sector. At the aggregate level, the nationallyowned manufacturing sector has been stagnant, while the foreign-owned sector has been growing, though different subsectors have grown at different rates due to differences in demand elasticities over the international business cycle. Since the Costa Rican industrial statistics do not provide the same detailed sectoral data for foreign- and nationally owned producers as the Irish statistics, I assess structural change by focusing on production statistics for nationally owned companies and on export data for foreign-owned companies. Given the nature of the available data I am making the assumptions that developments in the manufacturing sector outside the Special Investment Regimes capture the development of the national manufacturing sector (notwithstanding the fact that a few foreign-owned companies may be part of that sector), and that the performance of manufacturing firms in the Free Zones reflects the economic behavior of Transnational Corporations in the Costa Rican manufacturing sector (notwithstanding the fact that some national companies are in the Free Zones as well).36 The influx of high-tech FDI has led to a significant change of Costa Rica’s export profile. Non-traditional exports increased from 65 percent of total exports in 1995 to 87 percent in 2003. Part of the shift stems from the decline in coffee and banana exports, as the prices of these two commodities dropped precipitously.37 But the main reason for this shift is a genuine increase in nontraditional exports, even taking into account their temporary drop in 2000–2001. PROCOMER’s export data show that—at a fairly aggregate level—all nontraditional export categories grew, with the sole exception of shoes. Electrical machinery registered the most dramatic growth, rising from $45 million in 1997 to $1.4 billion in 2003. As a result, this sector, which includes most of the high-tech electronic and medicalinstrument products, increased its share in total exports from 1.4 percent to 24.6 percent (see table 5.7). Due to the surge in high-tech products from Costa Rica, the country ranks number one in Latin America in the share of high-tech goods in total exports. In 2000, Costa Rica’s share was 34 percent, more than double the Latin American average of 14 percent (see table 5.8). But this structural change is not a reflection of improved indigenous capabilities. It is driven predominantly by developments in the Free Zones. In 2003, FZ exports accounted for 56 percent of Costa Rica’s total exports (up from 10 percent in 1997), and for nearly 100 percent of all exports of electrical machinery.

FROM COFFEE TO COMPUTERS

Table 5.7

151

Costa Rica’s Manufactured Exports, 1997 and 2003 FZ exports/total exports

Total exports 1997

2003

2003

%

%

%

in US$

%

Agriculture and fishing Food industry Chemical industry Plastics Leather products Apparel Shoes Basic metal products Electrical machinery Metal mechanics Others

1,364,122,424

42.1

1,265,833,590

21.5

0.3

2.9

313,470,639 164,650,715 61,017,265 34,439,916 195,134,353 13,794,750 88,789,475 45,587,522 214,282,259 745,768,103

9.7 5.1 1.9 1.1 6.0 0.4 2.7 1.4 6.6 23.0

478,256,922 354,895,911 128,494,011 50,298,109 585,072,986 745,823 132,607,392 1,449,202,369 497,399,606 954,373,333

8.1 6.0 2.2 0.9 9.9 0.0 2.2 24.6 8.4 16.2

0.9 3.4 4.3 24.4 54.4 10.8 1.6 15.0 59.5 7.9

47.2 32.1 24.6 69.6 53.0 0.0 28.5 97.1 83.9 71.6

Total exports

3,241,057,421

100.0

5,897,180,052 100.0

10.1

55.9

Source: Calculated from PROCOMER, jsp?val⫽Panel⬎ accessed September 2004.

in US$

1997

⬍http://servicios.procomer.go.cr/estadisticas/Estadisticas.

Non-traditional exports outside the Free Zones, on the other hand, showed no overall growth. There the most significant structural change was the rise in apparel exports. However, most non-FZ apparel exports come from companies operating under the Active Processing Regime exporting to the United States under HTS 9802. And many of them are foreign-owned. The difference in growth between the Free Zone and the non-Free Zone manufacturing sector becomes even clearer, when we look at industrial statistics. Total manufacturing output was 29 percent higher in 2003 than in 1997. But whereas FZ production increased by 200 percent during this period, non-FZ output declined by 4.6 percent (see table 5.9). Output in small enterprises, which are not included in these data, stagnated as well.38 A further indication of the decreasing competitiveness of the non-FZ manufacturing sector is the drastic decline in its export intensity, from 28 percent in 1997 to 8 percent in 2003 (see Table 5.10). And with the decline in exports and output, employment decreased as well. Employment in the manufacturing sector remained unchanged between 1997 and 2003, with the increase in employment in the FZ countered by a decline in the nonFZ sector.39 Given the scant overall employment growth in the manufacturing sector, manufacturing has become a rapidly declining source of employment in Costa Rica. The sector accounted for 18 percent of the

152 Table 5.8 Export Decomposition by Technology Intensity, Selected Countries, 1985 and 2000 Primary products and resource-based products

Latin America Argentina Bolivia Brazil Chile Colombia Costa Rica Dominican R. Ecuador El Salvador Guatemala Honduras Jamaica Mexico Nicaragua Panama Paraguay Peru Trinidad & Tobago Uruguay Venezuela Asia Bangladesh China India Indonesia Korea Malaysia Philippines Singapore Sri Lanka Thailand Vietnam

Low technology intensive products

Medium technology intensive products

High technology intensive products

1985

2000

1985

2000

1985

2000

1985

2000

73.6 81.1 89.1 63.2 91.9 87.5 75.1 48.0 97.3 76.3 84.3 92.0 80.7 63.7 94.2 47.8 94.6 86.5 89.3 41.6 95.0 52.8 36.6 48.6 60.0 95.2 14.1 70.8 51.5 42.6 60.9 67.9 95.6

43.6 71.7 80.7 52.7 88.9 72.7 37.6 13.5 91.4 23.2 48.5 27.9 68.9 17.5 55.8 39.2 71.9 77.4 76.9 57.2 90.0 19.2 9.0 11.6 45.0 56.6 13.7 20.2 10.5 16.0 24.3 27.3 50.0

7.7 8.9 1.9 13.5 1.3 4.5 14.5 28.2 0.4 9.0 6.7 5.1 11.7 5.8 1.1 7.7 4.4 7.1 2.9 22.2 1.8 25.0 61.0 39.7 30.5 3.0 48.7 5.0 18.5 8.7 34.8 17.8 3.1

14.1 8.5 13.4 11.6 3.0 10.4 17.1 62.7 3.4 62.7 39.9 65.2 25.6 14.7 36.7 18.1 8.7 14.5 7.9 22.0 2.6 29.1 88.8 47.6 39.0 22.4 16.9 9.4 11.9 4.9 67.7 19.5 41.4

12.4 6.6 0.4 19.4 3.6 5.4 6.5 9.9 1.5 4.2 5.1 1.7 5.3 18.0 1.3 34.7 0.5 3.9 5.9 5.1 2.3 10.2 1.4 7.7 6.4 1.1 21.7 6.1 6.5 17.5 2.9 5.9 0.7

24.9 16.2 1.1 24.3 5.7 10.4 8.3 17.5 2.9 5.9 7.4 4.1 2.9 38.5 2.5 26.7 0.8 2.4 13.4 13.7 6.2 17.2 1.6 17.3 10.9 10.9 29.2 13.5 7.1 15.2 3.2 17.7 5.7

4.1 2.6 0.3 2.9 0.4 0.6 3.2 1.1 0.4 9.8 3.2 0.6 0.7 9.9 0.4 5.6 0.2 0.4 0.6 0.7 0.1 10.8 0.5 2.6 2.3 0.5 14.4 16.9 21.4 27.9 0.5 7.7 0.1

14.3 2.4 1.0 8.9 0.7 2.0 34.3 3.5 0.9 6.4 2.5 1.3 0.4 25.3 0.3 12.0 0.7 0.5 0.3 2.8 0.4 32.9 0.4 22.4 4.3 9.3 38.4 55.3 68.1 60.0 4.2 33.8 2.5

Percentages do not add up to 100, as a few SITC positions are not included in Lall’s (2000) classification. Source: Calculated based on CEPAL (2002).

Table 5.9

Manufacturing Output and Employment: Costa Rica, 1997–2003 Employment index (4th semester 1996 ⫽ 100)

Index of manufacturing output (4th semester 1996 ⫽ 100) Total manufacturing output

Manufacturing without FZR

Free zone regime

110.4 122.8 145.3 138.0 130.6 132.8 142.4

110.2 109.8 112.2 109.9 107.4 104.9 105.4

111.6 199.0 338.6 302.7 265.9 296.2 n.a.

Total manufacturing output

Active processing regime

Manufacturing without FZR

109.2 113.6 102.0 103.4 93.6 91.6 n.a.

1997.0 1998.0 1999.0 2000.0 2001.0 2002.0 2003.0

103.0 106.8 105.4 104.3 101.8 103.0 102.5

103.8 103.7 100.0 94.4 90.6 89.1 87.2

Distribution of manufacturing production (based on constant colones) 1997 1998 1999 FZ 14.8 23.8 34.2 Sector w/o FZ 85.2 76.2 65.8

2000 32.2 67.8

2001 29.9 70.1

2002 32.7 67.3

2003 37.0 63.1

Total

100.0

100.0

1997 1998 1999 2000 2001 2002 2003

100.0

100.0

100.0

100.0

100.0

Distribution of employment 1997 FZ 26.0 Sector w/o FZ 74.0

1998 29.1 70.9

1999 31.2 68.8

2000 29.8 70.2

2001 22.6 77.4

Total

100.0

100.0

100.0

100.0

100.0

Active processing regime

100.8 116.0 121.0 132.8 136.6 142.8 n.a.

105.3 103.1 94.2 92.2 94.1 90.8 n.a.

153

Source: Proyecto Coyuntura Industrial, Instituto de Investigaciones en Ciencias Económicas, Universidad de Costa Rica.

Free zone regime

154

FOREIGN INVESTMENT, DEVELOPMENT, AND GLOBALIZATION

Table 5.10 Export Intensity of Costa Rica’s Manufacturing Sector, 1997–2003 1997

1998

1999

2000

2001

2002

2003

100.0 27.8

100.0 21.4

100.0 14.2

100.0 13.7

100.0 14.7

100.0 11.6

100.0 7.8

Distribution of exports (based on constant colones) FZ 39.4 60.1 Sector w/o FZ 60.6 39.9

79.1 20.9

77.9 22.1

74.6 25.4

81.0 19.0

88.6 11.4

100.0

100.0

100.0

100.0

100.0

Exports as a share of output FZ Sector w/o FZ

Total

100.0

100.0

Source: Proyecto Coyuntura Industrial, Instituto de Investigaciones en Ciencias Económicas, Universidad de Costa Rica.

country’s total employment in 1990, but for only 13.9 percent in 2003 (Proyecto Estado de la Nación 2004, 400). Within this picture of overall stagnation in the non-FZ manufacturing sector, structural change did happen, though by far not as pronounced as in the FZs. Food and drink continued to account for over 40 percent of non-FZ production (see table 5.11). Traditional consumer goods like textiles, clothing, and leather declined in importance, while metal, mechanic, and electrical products gained in importance. The share of plastic, non metallic mineral, metal, and electrical products increased from 22.5 percent of manufacturing output in 1997 to 27.6 percent in 2003 (see table 5.11). But even though some branches have been growing, an indicator of greater competitiveness, their growth has, so far, been insufficient to compensate for the decline in the other branches. There are a number of examples of good-sized national companies though, which have grown through increasing their exports, especially at the regional level. The stagnation of the manufacturing sector, the decline in employment, and the modest structural change, are the result of rising competitive pressures that came with tariff liberalization and the abolition of the CAT, stagnation in internal demand, difficult access to finance capital, and, most importantly, lack of government support.40 The Costa Rican government has not had an industrial policy to support national manufacturers in the transition from a protected market to a market with international prices and competition. In that respect, the Costa Rican government is similar to the Irish government, which neglected the promotion of the national industrial sector for several decades before realizing the importance and necessity of proactive government policies to overcome the market failures confronting national producers. I discuss this issue at greater length in the next chapter, in the context of the scarce linkages between foreign and national production. Here I now

FROM COFFEE TO COMPUTERS

155

Table 5.11 Structure of the Costa Rican Manufacturing Sector, 1997 and 2003* Percentage breakdown based on real colones

Food and drink Textile products Clothing Leather products Wood products Paper products Printing, publishing Chemical products Rubber & plastic products Non-metallic mineral products Basic metal products Metal products Electrical machinery Transport equipment Furniture Active processing regime Total

1997

2003

40.8 3.3 1.7 1.0 1.2 3.5 1.5 11.9 7.2 6.3 4.9 2.2 1.9 0.2 1.0 10.7

41.4 1.6 1.5 0.5 0.7 4.2 2.0 9.5 9.1 8.0 5.9 2.2 2.4 0.2 1.4 8.6

100.0

100.0

* Only includes companies with more than 20 employees. Does not include FZ and PA. Source: IICE Proyecto Coyuntura Industrial.

turn to a discussion of the reasons behind high-tech foreign investment in Costa Rica. What motivated Intel, Inamed, Abbot, or any of the other hightech companies to set up production in this small country in Central America? Why Costa Rica? The factors which motivated Intel to invest in Costa Rica are fundamentally the same as those for other high-tech foreign investors.41 What is different about Intel’s investment decision is the huge impact it has had on the small country’s ability to attract high-tech FDI. As in the case of Ireland in the early 1990s, Intel’s investment put Costa Rica on the radar screen of potential foreign investors. It bestowed a seal of approval signaling that the country met the requisite conditions for profitable high-tech production. To be sure, Costa Rica and Ireland are at different levels of development, and thus the technological complexity of foreign production, which

156

FOREIGN INVESTMENT, DEVELOPMENT, AND GLOBALIZATION

the two countries attract, is different. After all, Intel built a wafer production plant in Ireland, but a testing and assembly plant in Costa Rica. Starting out at a lower technological level is a “natural” by-product of being at a lower level of development and should not be seen a priori as antithetical to a high-tech FDI-led development strategy. The critical issue is whether, over time, the production of high-tech foreign companies moves up the value chain in Costa Rica or not.

The External Context We saw in chapter 3 that “Europe 1992” and the surge in demand for computers were critical factors which motivated TNCs in the IT industry to establish production facilities inside the European Community in the 1990s. Ireland had the right location-specific assets at that particular juncture to become a top recipient for TNCs searching for the best location in the EC. In the case of Costa Rica, we cannot pinpoint a similar conjuncture of specific forces. Rather, Costa Rica’s location-specific assets provided a match for foreign investors’ needs in the context of the ongoing process of internationalizing production. Several characteristics of this process are particularly important for understanding why Costa Rica became a possible investment location for high-tech FDI. First, transnational corporations with global production and distribution networks, like Intel and Abbot, are always on the lookout for new investment sites.42 The search for new profitable sites is an integral part of their global expansion and restructuring strategies, as risk diversification and global commodity-chain management change over time. Thus, it becomes important for a country to be on the radar screen of these corporations. But it is not just the large TNCs, which are now looking for new production sites abroad. The second important trait of the ongoing internationalization process is that small and medium-sized companies from the industrialized countries are also increasingly looking to relocate production to countries with lower labor costs. In the case of technologically more sophisticated production that means lower wages for higherskilled workers. A good number of the high-tech foreign companies, which have invested in Costa Rica in the last few years, are indeed small or medium-sized. And for many of them, the investment in Costa Rica was their first or second investment abroad. For small and medium-sized potential foreign investors, the demonstration effect of large TNC investments in a specific country is particularly important, since their more limited resources translate into more limited knowledge of the pros and cons of potential production locations

FROM COFFEE TO COMPUTERS

157

abroad. My interviews with TNC representatives and government officials in Costa Rica left no doubt that Intel’s investment played a critical role for many of the other high-tech foreign companies, which came afterwards. Abbot’s subsequent investment had a signaling function as well, and it also made some investors in the medical field remember that Baxter had already been there for a long time. In some cases, the investment of the top TNC brought the country to the attention of a potential foreign investor; in other cases it confirmed a hunch they had had already about the attractiveness of the country’s location-specific assets; and in still other cases it was an important reason for the investment itself, with companies hoping for the benefits of agglomeration effects, as clusters of complementary activities develop in a particular industry. A by-product of the ongoing internationalization of production process has been the expansion of an internationally mobile management class, whose members have firsthand knowledge of different global production sites and of the historic development of global commodity chains for particular companies. This tacit knowledge provides a much less obvious, but nonetheless important, channel for a small latecomer to emerge on the radar screen of potential foreign investors. Anecdotal evidence involving Baxter illustrates the point. In 1987, Baxter established a plant in Costa Rica with 32 employees. It manufactured components for a large Baxter plant in Puerto Rico, taking advantage of shared production provisions offered under the Caribbean Basin Initiative. In the late 1990s, a manager of a U.S. medical-device company, in search of a production site abroad, remembered Baxter’s component production in Costa Rica, because he had worked for Baxter during the 1980s. He tracked down the current general manager of the Baxter company in Costa Rica, visited to scope out the possibilities, and now his company is producing medical devices in Costa Rica. A CEO of the same Baxter plant in Puerto Rico later established his own company of precision mechanical instruments on the U.S. mainland. When he was looking to establish a production facility outside the United States in the late 1990s, he, too, remembered Costa Rica, because of Baxter’s component production there. And his company is now producing in Costa Rica as well. Such stories illustrate how the informal network of contacts and knowledge that invariably develops in the course of internationalization has served Costa Rica well in placing it onto the radar screen of potential foreign investors. The tourist boom in the 1990s increased Costa Rica’s visibility as a potential investment site as well.43 It is, of course, a huge step from being on the radar screen of a TNC and actually becoming the foreign investor’s final choice of site. The former is

158

FOREIGN INVESTMENT, DEVELOPMENT, AND GLOBALIZATION

a necessary condition to attract FDI, but the latter will only be achieved, if the country has the right location-specific assets. Costa Rica’s Location-Specific Assets Costa Rica’s location-specific assets are in some ways similar to the ones Ireland had to offer, in terms of location, human capital, infrastructure, financial incentives, and stability. Costa Rica has a strategic location, as a bridge between the United States and Latin America, as well as the Atlantic and the Pacific. It has a highly productive workforce, where most workers have a good basic education, and some have technical training as well. It has an acceptable infrastructure in telecommunications and transportation. It offers exemptions from profit taxes and other taxes for a considerable number of years. It provides stability, in the broader economic and political sense as well as with respect to capital–labor relations. And it can claim a rather effective institution in charge of marketing the country to potential foreign investors, namely CINDE. While these advantages constituted a sufficient incentive for Intel and other foreign investors, major challenges loom in each of these areas, as I discuss later. The two most important obstacles for moving forward are the absence of a coherent development strategy for FDI-linked industrial upgrading, and the low tax ratio which makes it very difficult to meet the development needs for upgrading infrastructure and education and also maintaining the social services Costa Rica has developed over the last few decades. The Quality and Cost of Human Capital Human capital is the single most important factor in attracting high-tech FDI to a small latecomer. The availability of qualified labor is the result of an educational system that provides good general education and specialized higher education and on-the-job training. Costa Rica has had a long tradition of broadening access to education, starting with obligatory and free primary education in the nineteenth century and mandated secondary education in the 1940s. In the 1970s, the government moved to expand third level education, which meant not only an increase in the number of universities and technical schools, but, equally importantly, the establishment of a loan fund to support students of lesser means in higher-level education.44 Costa Rica’s adult literacy rate is close to 100 percent, and primary school enrolment is nearly universal. Net enrolment in secondary schools has been increasing from 37 to 51 percent in the course of the 1990s, and gross tertiary enrolment was 21 percent in 2001 (World Development

FROM COFFEE TO COMPUTERS

159

Indicators, online). In 1987, the Omar Denge Foundation was established to promote the introduction of computer instruction in elementary schools, an initiative that has been considered rather successful. A considerable number of people have some knowledge of English, the result of the introduction of foreign-language teaching in primary schools in 1994 as well as of the prevalence of English-speaking tourists. Nonetheless, school completion and repetition rates are a matter of concern at both the primary and secondary level.45 And Costa Rica still has considerable deficiencies in technical education at the tertiary level. Availability of a sufficiently large pool of technically trained labor was of major concern for Intel when considering Costa Rica as a potential production site. Costa Rica’s education authorities proved extremely flexible and forthcoming in developing programs to meet Intel’s needs. Close collaboration among Intel, the Costa Rican Institute of Technology (Instituto Tecnológico de Costa Rica— ITCR), and the Ministry of Education helped identify the gaps and programs were established to fill them, most importantly a one-year technical certificate program at the ITCR and a one-year associate degree program, which Intel and the ITCR designed jointly (see, for example, Spar 1998). The flexibility in higher technical education clearly paid off, as it assuaged Intel’s concerns about the right human capital availability in Costa Rica. Nonetheless, education of more engineers in a sufficiently wide number of areas remains an ongoing challenge. One cannot continue to attract large TNCs in the high-tech area without a commensurate increase in the supply of engineers. Part of the problem is that there are not enough students in the sciences. In the mid-1990s, only 18 percent of students in tertiary education were in the sciences, compared to 43 percent in Chile and 30 percent in Argentina and Ireland (UNDP 2004, 176). It is a common complaint among Costa Rican producers that the country has too many lawyers and not enough engineers. Low wages often are developing countries’ main attraction for foreign investors. In an econometric analysis of the determinants of exportoriented foreign investment in the Caribbean Basin in the 1980s, Woodward and Rolfe (1993) found that wage rates had indeed a negative impact on export-oriented FDI. This result is a reflection of the dominance of foreign investment in maquila clothing production under the CBI, where unskilled labor predominated and productivity differences across countries were low.46 But when skilled labor is involved, focusing just on low labor cost is insufficient. What ultimately matters to corporations is unit labor cost, in other words, wage rates and productivity levels. While labor in Costa Rica is relatively expensive compared to competitor countries for high-tech FDI at similar income levels, it is also considered to be more productive. The social insurance contributions, which support the relatively

160

FOREIGN INVESTMENT, DEVELOPMENT, AND GLOBALIZATION

high social services in Costa Rica, are a main reason for the comparatively high wages, but also for the comparatively higher productivity. In interview after interview, managers of high-tech TNC affiliates stressed that Costa Rican workers learn new production processes quickly and that the training periods had been substantially shorter than they had planned for, making the cost of labor training considerably lower in Costa Rica. And some stated explicitly that higher productivity at their Costa Rican plants compensated for wages that were higher than in other affiliate plants in Asian countries. When CINDE advertises the low cost of labor in Costa Rica as an important location-specific advantage, it uses countries at higher income levels as reference points, for example Ireland, Mexico, and Puerto Rico (see table 5.12).47 But whether Costa Rica is indeed the cheaper production site compared to those countries depends on the relative productivity levels. Low turnover rates on the job are another location-specific asset for Costa Rica. Several high-tech TNCs highlighted an average turnover rate of less than 7 percent as an important attraction, especially compared to Mexico, which is currently the only other major contender for high-tech FDI on the Central American isthmus. Location Since all high-tech TNCs in Costa Rica produce exclusively for the export market, Costa Rica’s proximity to the United States is a distinct advantage. About half of Costa Rica’s exports in 2003 went to the United States.48 It takes less than three hours by plane from San José, Costa Rica, to Miami, which considerably reduces the time and cost of transportation to the U.S. market. The latter is probably more important than the former. While the rising volume of ships between Asia and the United States may lower the cost differential for a container from Singapore to Los Angeles compared to a container from Costa Rica to Miami, Costa Rica will always retain the advantage of more timely delivery. Costa Rica’s proximity to the United States Table 5.12 Labor Costs in the Production of Medical Devices in 2002 Annual gross remuneration index

Costa Rica Ireland Mexico, Baja California Puerto Rico

Production assembler

Manufacturing engineer

100 330 148 160

100 190 200 235

Source: CINDE, ⬍http://cinde.or.cr/cinde/home.nsf/pages/medical⬎ retrieved July 5, 2004.

FROM COFFEE TO COMPUTERS

161

will become even more important in the future, if the proposed U.S.Central American Free Trade Agreement (CAFTA) is passed. In that case, ceteris paribus, more TNCs are likely to invest in Costa Rica to take advantage of duty-free access to the U.S. market. Infrastructure Costa Rica has been able to offer the needed infrastructure regarding telecommunications and transportation, and in the case of infrastructure deficiencies, government agencies have shown considerable adaptability and flexibility to overcome them. The country has a relatively well-developed telecommunications system. While there is a lot of grumbling about the government monopoly in electricity and telecommunications, problems manifest themselves mainly outside the Free Zones. ICE reserves lines for companies in the FZs, providing sufficient access to broadband, digital, voiceover, and so on. FZ managers establish parallel agreements with ICE, where they buy the equipment and ICE operates it, with adjustments made in the operating bills. Flexibility and speed in meeting Intel’s concerns about electricity and road transportation allayed the corporation’s concerns in those areas. ICE built a special electricity substation for Intel (though with funds advanced by the company), and the Ministry of Transportation took charge of improving road transportation between the Intel facility and the airport. A new airport outside of San José has markedly improved air transportation for all companies. Nonetheless the preferential treatment of TNCs in the Free Zones with respect to telecommunications and electricity makes it more difficult for indigenous companies to become competitive. Major challenges in infrastructure lie ahead for the future development of the country. In addition to the need for greater capacity in the telecommunications area, improvements in the road system generally and the traffic system in San José specifically are critical to lower transportation costs in the country, as well as car-induced air pollution in the capital.49 The low tax ratio constitutes a major obstacle to the government’s ability to address these challenges. Outside development funding from multilateral agencies will be important, but it is not likely to breach the whole gap. U.S.-AID funding plummeted in the early 1990s and was terminated in the mid-1990s, when Costa Rica reached the threshold GDP per capita level for AID funding. After that, foreign aid flows became much smaller, and net flows actually turned negative for a few years. To be sure, Costa Rica has always received support for development projects from international agencies, most importantly from the Inter American Development Bank (IADB). Over the last 40 years the IADB has invested $1.2 billion in ICE to extend electricity to all corners of the country.50 Between 1991 and

162

FOREIGN INVESTMENT, DEVELOPMENT, AND GLOBALIZATION

2001, IADB approved 19 loans for a total of $951 million. But since most projects were approved before 1993 and since the aid was in the form of loans (not grants like in the 1980s), net financial flows to the country were actually slightly negative in the second half of the 1990s, as amortization and interest payments dominated over new inflows.51 The negative net aid flows from abroad, during a time when high-tech FDI inflows were most prominent, stands in stark contrast to the Irish experience in the first half of the 1990s when EU structural funds amounted to 3.5 percent of Ireland’s GDP. Those funds came in the form of grants and not loans, and they were used strategically for development purposes. IADB funding for Costa Rica was used for development goals as well. But their extent was much more limited, and since they came in the form of loans rather than grants, net inflows could actually become negative. Stability, Transparency, and Governability Costa Rica is one of the few countries in Latin America, which has not had a military government in the last 50 years. Political stability and weak company unions in the private sector provided an attractive climate for potential foreign investors. In addition, the transparency of rules for foreign investors and the relative absence of special deals have sent strong signals about respect for laws and willingness to enforce them. The lack of transparency and the pervasiveness of “under the table payments” in Mexico versus Costa Rica was one factor, which tilted the balance of Intel’s site decision against Mexico and in favor of Costa Rica (Spar 1998; Larrain, Lopez-Calva, and Rodríguez-Clare 2000). Several other high-tech foreign companies made a similar point during my interviews. However, there are numerous challenges to continuing economic and political stability in Costa Rica. Macroeconomic stability is becoming more tenuous with rising fiscal deficits. While the inflation rate has moved into the single digits over the last few years, the fiscal deficit has actually increased. During the 1990s, the public deficit was, on average, 3.4 percent of GDP, but in 2002 it had reached 5.9 percent of GDP (IADB 2003b, 2). Such a deficit is unsustainable and poses a clear threat to macroeconomic stability. Another, and more fundamental, challenge in the long run stems from the absence of a clearly-defined and accepted new social contract supported by a strong coalition of key sectors of society. A new contract is needed, both because the demise of the leading coalition under the former ISI model has left a void of contention and because the social services, which form such an integral part of the social–democratic model, Ticostyle, are coming increasingly under siege. The new economic model dismantled the former leading coalition of the state, segments of domestic capital, and public sector trade unions by

FROM COFFEE TO COMPUTERS

163

reducing the economic role of the state and weakening public sector unions, while empowering financial capital and transnational corporations.52 Unless a new leading coalition can be established, which brings labor back in and includes a chief role for government-initiated and coordinated development policies, economic growth can be neither generated nor sustained in the long term.53 A further challenge to political stability, that needs to be addressed is the strain on social services. It is the result of limited resources, and additional pressures from a large immigrant population from Nicaragua that is estimated at around 350,000 people, nearly 10 percent of the Costa Rican population. In addition, the bribery scandals of 2004 cast a potentially long shadow on Costa Rica’s image of a country with no backroom deals. At the end of 2004, former president Rafael Calderon (1990–1994) was serving a ninemonths sentence for a conviction on corruption charges involving the social security fund and a loan from Finland; and former President Rodríguez (1998–2002) and former President Figueres (1994–1998) stood accused of accepting large sums from the French telecommunications company Alcatel. It is not clear what the fallout of these corruption scandals will be for future FDI, as the actual and alleged wrongdoings seem to involve mainly the upper echelons of power and not the overall institutional structure of government. Financial Incentives Like Ireland in the 1960s and 1970s, Costa Rica offers de facto zero percent profit taxes, as investors under the Free Zone Regime are given a 100 percent exemption from taxes for 8 years, and a 50 percent exemption for another 4 years.54 Yet unlike Ireland, Costa Rica does not offer any additional grants to attract foreign investors. While zero tax rates play an important part in foreign investors’ calculations, the significance of this factor within the whole set of locationspecific assets varies considerably among companies. Tax holidays play a more significant role for larger companies than for smaller ones. In contrast to the larger corporations I interviewed, several smaller companies did not even mention the tax exemption as a factor behind their decision to invest in Costa Rica. That does not mean that small companies do not care about the level of the tax rate, only that they tend to weigh this factor less heavily. How much of a deal breaker the tax exemption has been becomes a rather important question, as WTO rules mandate their elimination by the end of 2007.55 What should the new tax rate on profits be? A nominal rate of 30 percent, for example, would make Costa Rica an outlier among its competitors, deterring large new foreign investors and probably motivating

164

FOREIGN INVESTMENT, DEVELOPMENT, AND GLOBALIZATION

some TNCs to leave Costa Rica. In response to my question what an acceptable tax rate would be from the business perspective, the manager of a large TNC pulled a folder off his desk with details about tax breaks and nominal and effective tax rates for all countries in which the TNC was operating and insisted that the company would move if the effective tax rate went above 9 percent. In 2002, the average effective tax rate for subsidiaries of U.S. transnational corporations was 8 percent in Ireland and 11 percent in Singapore (see table 2.2 again). Another company, on the other hand, left room for a higher rate, pointing out that, after all, creative use of transfer pricing could compensate to a certain extant for tax rates that are considered too high. As productive capital has become increasingly global, governments in all countries have found it ever more difficult to tax capital, with the result that the tax share from profits has gone down in many countries. It is impossible for one country on its own to challenge the trend of declining corporate taxes. An SLC will be better off aiming for a tax rate that is cognizant of this reality and of the possibility of transfer pricing than for a tax rate that is meant to compensate for the many years when TNCs did not pay any taxes. What matters in the end, both to the TNC and the host country, is the effective tax rate. In the case of Ireland, we saw that shifting from a zero tax rate to a 10 percent tax rate in 1981 did not lead to a decline in foreign investment or the corporate tax share. On the contrary, the corporate tax share actually tripled between 1980 and 2000, as FDI inflows surged. The lesson for a country like Costa Rica is that FDI inflows are not likely to decline, ceteris paribus, when the effective tax rate moves from zero to positive, as long as it is in line with tax rates in the countries that are the country’s most immediate competitors in the pursuit of high-tech FDI.

Costa Rica’s Institutional Structure The two most important institutions dealing with transnational corporations in Costa Rica are CINDE, the Costa Rican Coalition for Development Initiatives, and PROCOMER (Promotora de Comercio Exterior), the export promotion agency. CINDE is charged with marketing the country to potential foreign investors, while PROCOMER is providing technical and strategic information and services for exporters and importers. Established at the end of 1996, PROCOMER defines itself as a public service organization, which operates like a private company.56 The organization has streamlined the administrative procedures by providing onestop services for companies engaged in international trade. It has also

FROM COFFEE TO COMPUTERS

165

improved logistics through greater speed and flexibility for exports and imports. But it has clearly exhibited preferential treatment for foreign corporations compared to national producers.57 While it can take up to eight hours for a local company to retrieve imported inputs from customs at the airport, it takes one hour for a company in the FZ. At port facilities, the retrieval process takes 4–5 hours for a FZ corporation, while it takes at least two days for a company that operates outside the special investment regimes.58 Established and funded by U.S.-AID in 1982, CINDE has been a nongovernmental organization since its inception, with no funding from the government, and no accountability to it either.59 CINDE’s nongovernmental status has freed it from the potential vicissitudes of a change in agendas from one government to the next. One of CINDE’s former directors for Investment Promotion praised the organization’s private sector status for allowing greater flexibility in its actions and closer relations with the foreign investors. But private sector status has also become a problem, as it has meant that the responsibility for FDI has been located outside of existing government institutions. CINDE was very well-endowed when it started in 1982. Established with $21 million and annual budgets between $4 million and $8 million in the 1980s, CINDE opened seven international offices in Europe, Asia, and the United States, and employed 300–400 people at the height of its operations. Initially, CINDE went after any investment it could attract, which turned out to be primarily in clothing, as we already saw in the discussion of CBI and HTS 9802. But when the other Central American countries became more peaceful and thus attractive sites for maquila clothing operations in the first half of the 1990s, CINDE became aware that Costa Rica needed to attract FDI in other areas. The organization began to focus more on potential foreign investors in the electronics sector, particularly after a commissioned study identified electronics as an area where Costa Rica had the requisite location-specific assets (FIAS 1996).60 Assessments differ of the extent to which CINDE has pursued a targeted strategy in electronics and medical devices in the past. CINDE did not have a blueprint, which was drawn up proactively and then implemented. But it did have some idea that these areas were important, and when a big player in those industries invested in Costa Rica, that is Intel and Abbot, it led CINDE to become more strategic in those particular areas. By mid-2003, CINDE had an action plan in three areas: electronics, business services, and medical devices. Global Park, a free zone outside of San José, has been the site for the development of a cluster in the medical device industry. The Free Zone was

166

FOREIGN INVESTMENT, DEVELOPMENT, AND GLOBALIZATION

established in 2000 to accommodate Abbot’s investment, and by the summer of 2003, 10 smaller foreign investors in the medical device industry had established production in the same FZ. The medical cluster continued to grow slowly in the following year, with three more medical device companies deciding to establish production facilities in Costa Rica, Boston Scientific among them. Like INAMED, which assembles breast implants in Costa Rica, Boston Scientific has production facilities in only one other country outside the United States, namely in Ireland. As U.S.-AID turned its attention to post-civil war reconstruction in Central American countries in the early 1990s, funding for CINDE declined, and then dried up in the mid-1990s, when Costa Rica was no longer eligible for AID funding, having reached AID’s funding thresholds. Since then CINDE has been financed from the proceeds of capital that CINDE had invested before and from grants from a foundation promoting development in Costa Rica—itself partially financed from an endowment AID had set up before leaving Costa Rica. In 2002, CINDE had a budget of only $1.5 million. As a nongovernmental institution, CINDE has not been in a position to implement coordinated movement forward among the requisite institutions to execute a full-fledged FDI-led strategy of industrial upgrading, because it has neither the political clout nor the resources to do so. And since the government has not spent resources to support CINDE, it also has not had to make a real commitment to a strategy that could turn FDI into a more important force behind development. An amount of $1.5 million is not enough to be effective in cluster formation. But ultimately the more important problem is the absence of an articulated strategy and vision, in which CINDE would be one actor with a specific role. Since the adoption in the early 1980s of the Costa Rican-style neoliberal model, successive governments have not had a full-blown development agenda, which would have articulated a development vision for the country defining the role of industry, of CINDE and of high-tech FDI in the country’s economic future. There has not been a strategy for industrial upgrading, agreed upon, adhered to, and institutionalized from one government to the next. There has been a continued belief in the benefits of economic opening and market liberalization and no governmental intervention in the economy. And there has been a lot of rhetoric about the importance of attracting high-tech FDI. But there have not been institutionalized policies which recognize that FDI-generated and sustained growth does not happen automatically, that the different forces behind the development of a country’s knowledge-based assets have to move ahead in synch, and that interaction between the requisite institutional bodies needs to be coordinated from the top. A nongovernmental institution like CINDE cannot be in charge of industrial policy, even if it wanted to be.

FROM COFFEE TO COMPUTERS

167

Over the last 20 years, Costa Rica has not had a government that has specified the steps needed for making FDI a driving force behind the country’s economic future and translated them into an institutionalized strategy. The government of José Maria Figueres (1994–1998) seems to belie this claim, as this president played a critical role in bringing together all the necessary actors to attract Intel to Costa Rica. But his insights and actions did not subsequently translate into an institutionalized strategy. After CINDE had made a presentation at Intel headquarters in November 1995, an Intel team visited Costa Rica for the first time in April 1996.61 President Figueres took a strong personal interest in bringing Intel to Costa Rica and took an active role in the negotiations. In an interview with Ketelhoehn and Porter (2002, 14) Figueres summarized: “Ever since the first meeting, Intel became priority number one for me. There was no week, in which Intel’s issues were not discussed and worked in my office. I met with all Intel missions that arrived to the country.” After the first encounter, Figueres appointed Jose Rossi, Minister of Foreign Trade, as the point person to coordinate all efforts needed to satisfy Intel’s demands and make the TNC decide in favor of Costa Rica. The small size of the country turned out to be an advantage, as it allowed for quick integrated, collective action, once the political will and the necessary power to enforce it were there. The Minister of Foreign Trade brought together the heads of the Ministry for Environment and Energy, the Costa Rican Electric Utility Company, the Ministry for Transport and Public Works, the Ministry of Finance, the Ministry of Science and Technology, the Ministry for Education, and the Technical Institute of Costa Rica (Spar 1998, 17). And it was this cooperation and concerted effort that allowed the Costa Rican government to address successfully Intel’s concerns about infrastructure (improved road access to the airport and uninterrupted access to electricity at reasonable prices), about education (availability of a sufficiently large labor force with the requisite technical skills), and about financial incentives (nonapplicability of a 1 percent tax on assets on corporations in the Free Zones). The Intel case makes very clear the possibilities and limitations of CINDE. The agency was instrumental in getting Costa Rica on Intel’s radar screen and in coordinating the first steps with other institutions, once the TNC showed a clear interest in investing in the country. But it did not have the political muscle to do what in the end needed to be done, and was done, by an institution closer to the higher echelons of political power. From Costa Rica’s perspective, Intel’s arrival was the outcome of a fortuitous set of circumstances, where the President’s involvement and his directives to make things happen were critical. But unfortunately, the grand collective effort faded away, once Figueres left office in 1998. The experience with Intel did get different institutions to talk to each other, but

168

FOREIGN INVESTMENT, DEVELOPMENT, AND GLOBALIZATION

informal talks, while important, need to be anchored in formally agreed upon modes of cooperation, which do not depend on who happens to be in office or heading a particular ministry. The fate of Programa Impulso, the outline of a strategic vision under the Rodríguez Administration (1998–2002), is another case in point for the importance of institutionalizing coordination efforts in support of a comprehensive strategy. Andres Rodríguez-Clare, a Stanford-Ph.D. economist, chief of the Council of Presidential Advisors, and son of the president, was the driving force behind Programa Impulso. Given that Programa Impulso was run out of the president’s office, it had the potential to move the country forward. But given that it was anchored in an individual and not institutionalized, it did not survive with the change in government. When Rodríguez left office, Programa Impulso was moved to the Ministry of the Economy, where it died quietly. A development agenda for the twenty-first century came out of a study commissioned by the Central American presidents in 1996 from the newly established Latin American Center for Competitiveness and Sustainable Development at INCAE and the Harvard Institute for International Development (Robles 1999). Using Porter’s “Diamond Theory” of competitiveness as the conceptual framework, the study analyzed the diamond’s four areas of competitiveness (factor conditions, demand conditions, competitive environment, and related industries) and their interrelations. The report seems to have informed Programa Impulso as well as a joint action plan in 2002 for the development of a cluster around medical devices (CINDE, INCAE, Program Impulso 2002). But it was not adopted by the government as the official development strategy and action plan. And the government did not offer an alternative vision either. While the report discusses a number of the challenges highlighted above, it fails to stress, let alone mention—in its 130 pages—the need for strategic and proactive intervention on the part of the government to move ahead in the requisite areas in a coordinated and synergistic way. That is all the more surprising given the report’s explicit choice of South Korea as the ideal country to emulate (Robles 1999, 54). Most analysts of the South Korean development experience concur on the critical importance of strategic government action and proactive industrial policies for that country’s miracle performance (e.g. Amsden 1989; Wade 1994).

A Long Way from a Tico Tiger This chapter analyzed the different elements that led to the matchmaking between TNC’s strategic needs and Costa Rica’s specific assets to result in

FROM COFFEE TO COMPUTERS

169

high-tech FDI in this small Central American country. Figure 5.1 summarizes them in schematic form. The legacy of past policies in education, social services, and infrastructure combined with the special treatment afforded to corporations in the Free Zones under the new economic model (no taxes, speedy processing in customs, and special accords with ICE) provided sufficiently attractive conditions for foreign investors in the hightech field to establish production sites in Costa Rica. Intel’s decision to establish an assembly and testing facility for microchips was critical, not only because of the investment itself, but also because of the seal of approval it bestowed on a country, which most people knew only for its coffee and ecological richness. The demonstration effect was an important element in attracting other companies. The investments by Intel and Abbot as well as by the smaller companies that followed them suggest that Costa Rica may indeed have the possibility to use high-tech FDI as a handmaiden or engine of future economic growth. But that possibility will remain remote, if policies continue on their current trajectory. The main force behind the Celtic Tiger in the 1990s was the sustained increase in high-tech FDI. Costa Rica has not yet seen a sustained increase – Ongoing process of internationalizing production – Large TNCs – Small, medium TNCs – Following TNC clients – Cluster seeking

Global value chains: TNCs’ strategic needs

Costa Rica’s locationspecific assets

Tacit knowledge of int’l management

– Cheap, productive skilled labor CINDE

– Location – Acceptable infrastructure

Tourist boom

– Political & economic stability, transparency U.S.-AID

– Zero tax rates

Figure 5.1 High-Tech FDI in Costa Rica in the late 1990s: Critical Contingencies.

170

FOREIGN INVESTMENT, DEVELOPMENT, AND GLOBALIZATION

in high-tech FDI that could lead to a Tico Tiger. The absence of a critical mass of high-tech FDI also means that not many international input suppliers have followed their TNC clients to Costa Rica. One of the reasons is that the IT industry went into decline soon after Intel had geared up in Costa Rica. The demand for microchips turned out to be very susceptible to the ups and downs of the international business cycle, as computers and electronics more generally are less of a necessity to daily life. The medicaldevice industry, in contrast, has proven to be much more recessionimmune. And foreign investments in that sector have increased steadily, but slowly, in Costa Rica. Supply constraints of key location-specific assets, rather than TNCs’ willingness to invest in the country, constitute the more important constraints on large FDI inflows in high-tech areas in Costa Rica. The country is not at a level of development where it can absorb huge amounts of additional high-tech FDI, since that requires a lot of technically trained workers and engineers. There is a dialectical relationship between growth in hightech FDI and growth in a country’s location-specific assets. And it is currently primarily the latter, which limit the growth of the former. The focus on building a cluster in medical devices, primarily with smaller and medium-sized companies around one big company (Abbot) fits the current availability of human capital well. In that way, high-tech FDI will continue to make a contribution to economic growth, together with tourism and agriculturally based exports, but it cannot generate the quantitative and qualitative leap that would provide the impetus for a Tico Tiger. The two main problems Costa Rica has to overcome are the shortage of resources, fiscal revenue and external funding, and the absence of a proactive government strategy. Costa Rica needs to sustain its location-specific assets in order to provide the foundation for sustained growth and development. At the same time, the very nature of the current model of FDI attraction, with its preferential tax treatment for FZ corporations, limits the government’s ability to overcome the challenges in education and infrastructure and effective institutional support. The bifurcated model with one set of rules for the FZs and another set of rules for everybody else is not sustainable, because it is predicated on no profit taxes in the FZs. The low tax ratio of 13 percent puts a major stranglehold on the government’s ability to finance the needed expenditures in education and infrastructure, not just to attract further large-scale high-tech TNCs, but also to address the development needs of the country more broadly. A solution to this impasse needs to include a positive effective tax rate on companies in the free zones. Since WTO rules stipulate the elimination of preferential treatment for FZs, those reforms will happen in Costa Rica, though not until after 2007.

FROM COFFEE TO COMPUTERS

171

Improvement in tax collection generally, however, does not have to wait until then. The Costa Rican legislature debated a tax reform in 2004 aimed at increasing fiscal revenue. It remains to be seen, however, whether the reform will be implemented. The dearth of tax revenue, which is also one source of the macroeconomic disequilibrium stemming from the fiscal deficit, is not the only problem the country needs to overcome. An even more important problem is the absence of an institutional structure, where the key institutional actors move forward in a coordinated way, and the government articulates the goal and the path of a coherent development strategy. The lack of resources for CINDE, the partially overlapping mandates of some institutions, turf fights, and more generally the lack of an institutionalized, well-articulated and coordinated development strategy from the top together with a multiagency coordinated network to implement it constitute important obstacles for the emergence of a Tico Tiger. The articulation and institutionalization of such a strategy has to come from the office of the president, because the presidency commands the power to bring the relevant institutions together for collaborative action. Individual initiatives, which do not become institutionalized, but wither away with the individual’s move out of office—like Figueres’ all-out efforts to attract Intel—do not serve as the institutional base which needs to anchor a hightech FDI-led strategy. A lot of individual actions by different institutions, public or nonpublic, do not add up to a coherent strategy, unless they are articulated and executed and coordinated in the context of the overall development strategy. The most serious obstacle to the realization of a Tico Tiger is the absence of a sustained leading coalition behind Costa Rica’s economic model. Costa Rica’s developmentalist state of the 1960s and 1970s was eliminated in the 1980s and 1990s, rather than modified by correcting its excesses and mistakes. The void left by the disintegration of the old leading coalition has been filled by parties that can no longer be identified with particular interest groups. And scandals and blatant corruption have led to disenchantment with the government’s ability to be an effective agent of change. In such a context, it is not clear who will heed the call for a developmentalist state for the twenty-first century. Of course, Costa Rica today is not where it was 20 years ago; Intel and Abbot have put it on the map. But that does not mean that Costa Rica no longer needs an agency in charge of attracting foreign investment. Imperfect information by potential foreign investors combined with fierce competition from the many countries, which are vying for FDI generally and high-tech FDI specifically, necessitate proactive government policies. A proactive government strategy is needed to combine the right

172

FOREIGN INVESTMENT, DEVELOPMENT, AND GLOBALIZATION

improvements in the country’s location-specific assets, infrastructure, education, and so on to continue attracting high-tech FDI. Such a strategy is all the more critical in an environment where interSLC competition is increasingly fiercer and external developments have posed new challenges. The abolition in 2005 of the industrialized countries’ quota system of textile and clothing imports will lead to a dramatic restructuring of the global value chains in those industries. It will pose a major challenge for clothing exports from Costa Rica, and have a negative impact on the country’s economic growth.62 In addition, the emergence of Central and Eastern European countries and China as key locations for high-tech foreign investment has increased the competitive pressures for the right location-specific assets considerably.

6

The Tico Tiger That Hasn’t . . .

NCs’ most valuable potential contribution to the development of a small latecomer lies in the advancement of the country’s knowledgebased assets. Through linkages and spillovers, high-tech FDI can expand the technological, production, and marketing capabilities of national producers thus allowing them to become more competitive over time, expand and export their own production, and move up the value chain in their respective areas of production. Whether such positive developments actually occur or not depends critically on whether there is a match between TNCs’ desire for linkage formation on the one hand and indigenous capabilities to meet such desire on the other. The analysis in chapter 5 showed that, at the beginning of the twentyfirst century, the manufacturing sector in Costa Rica exhibits clear similarities to the Irish manufacturing sector twenty years earlier. The competitive import pressures, which had come with Ireland’s entry into the European Economic Community in 1973, brought about a decline and restructuring of the national manufacturing sector. In Costa Rica, the competitive pressures, which resulted from more gradual tariff liberalization and the abolition of export subsidies, led to a restructuring and stagnation of the national manufacturing sector as well. There are also striking similarities between Ireland and Costa Rica in the initial scant development of linkages. Just as supply does not create its own demand, the mere presence of high-tech foreign investment does not create backward linkages. In the early 1980s, high-tech TNCs in Ireland basically operated in an enclave. The Telesis Report of 1982 highlighted the limited linkages between TNCs and national input suppliers in Ireland. In Costa Rica, by 2004, linkage development between TNCs in the Free Zones and national producers has been equally unsatisfactory. Lack of domestic linkage capability, rather than TNCs’ interest in domestic sourcing, has been the most critical obstacle to linkage formation.

T

174

FOREIGN INVESTMENT, DEVELOPMENT, AND GLOBALIZATION

The pervasiveness of market failures, from coordination failures to asymmetric information to imperfect competition, works against the “automatic” development of indigenous linkage capability in small latecomers. Targeted government policies are needed to overcome these market failures and allow indigenous producers to become competitive input suppliers. In Ireland, it was not until the government adopted aggressive proactive policies in the second half of the 1980s that a number of indigenous producers became competitive and supplied TNCs with some of the needed inputs. Costa Rica is only starting to promote domestic linkage capabilities more actively. The establishment of the linkage promotion program Costa Rica Provee (CRP, Costa Rica Provides), in 2001, has generated some initial success. But its long-run success remains uncertain in light of the limited resources at its disposal. Other spillovers have been more significant in Costa Rica, as TNCs have trained workers, demanded high standards in terms of quality and on-time delivery, and influenced the direction in higher education in the technical areas. Production in the Free Zones has also become somewhat more skillintensive during the 1990s, an indication of some potential for technological spillovers.

Limited Linkages between TNCs and Indigenous Firms Backward linkages between TNCs and input suppliers in Costa Rica have been very limited to date. Between 1998 and 2002, there was no increase in extensive or intensive linkages. TNCs in the Free Zones bought domestic inputs worth $54.6 million in 2002, the equivalent of two percent of the value of their output (see table 6.1), and around 2.5 percent of the total value of their inputs.1 The high-tech sectors, electronics and medical instruments, turn out to be even less connected with the domestic economy than the clothing sector, which is notorious for the limited use of local inputs, as foreign producers operate under HTS 9802. Outside the services, packaging, and printing industries only a few national firms have succeeded in becoming competitive suppliers to high-tech TNCs, and only a small number of international input suppliers have followed their TNC clients to Costa Rica. Foreign exporters in Costa Rica import a much higher percentage of their inputs than national producers. A survey of 300 manufacturers in 2001 showed that 54 percent of exporters imported more than 75 percent of their inputs, versus 22 percent of non-exporters. And 10 percent of exporters imported none of their inputs, compared to 40 percent of nonexporters (see table 6.2). Among exporters, foreign producers imported

175 Table 6.1 Use of Domestic Inputs in Costa Rica’s Free Zones, 1998–2002 (in Millions of US$) 1998

1999

2000

2001

2002

All Companies Exports Imports Local purchases Purchases from FZ companies

1,960.6 1,539.11 67.5 6.5

3,624.9 1,858.7 54.9 6.6

2,998.0 1,794.2 72.2 3.9

2,376.7 2,014.7 56.6 3.4

2,657.3 2,210.3 51.6 3.0

Electronics (HS 84–85) Exports Imports Local purchases Purchases from FZ companies

1,300.9 919.5 n.a. 0.1

2,795.0 1,213.9 14.1 0.9

2,021.8 1081.15 16.0 1.0

1,204.9 1,305.2 15.7 2.2

1,286.8 1,459.2 4.1 0.4

Medical instruments (HS 90) Exports Imports Local purchases Purchases from FZ companies

83.9 67.5 n.a. 0

117.7 42.9 2.7 0

200.7 35.0 2.6 0

329.6 40.7 3.8 0.3

401.3 74.9 1.6 0

Clothing (HS 50–63) Exports Imports Local purchases Purchases from FZ companies

320.9 310.2 n.a. 1.1

372.2 333.8 3.9 0.7

387.0 330.9 6.2 2.4

372.4 310.2 4.3 0.9

399.2 276.6 2.7 2.6

n.a.: not available Source: PROCOMER.

Table 6.2 Survey Results on Imported Input Share in Costa Rican Manufacturing, 2001 0

1–25

26–50

51–75

76–99

100

Total

“N”

Non-exporters ⬍30 employees

39.8 40.9

13.9 12.9

18.5 19.4

5.6 4.3

12.0 12.9

10.2 9.7

100 100

108 93

Exporters National Foreign Mixed

9.7 12.8 3.7 8.3

15.4 15.4 14.8 16.7

11.8 12.8 11.1 8.3

8.7 10.3 3.7 12.5

35.9 35.0 37.0 37.5

18.5 13.7 29.6 16.7

100 100 100 100

195 117 54 24

Exporters, ⬍30 employees National Foreign Mixed

19.3 20.0 7.7 40.0

15.8 17.5 15.4 0.0

12.3 10.0 23.1 0.0

7.0 10.0 0.0 0.0

22.8 17.5 30.8 40.0

22.8 22.5 23.1 20.0

100 100 100 100

57 40 13 5

Source: IICE (2001).

176

FOREIGN INVESTMENT, DEVELOPMENT, AND GLOBALIZATION

a higher percentage of their inputs than national producers. To the extent that TNCs source domestically, they purchase lower-tech services and packaging and printing materials. Although the limited degree of domestic sourcing is partly due to TNC sourcing strategies, it is primarily the result of lack of indigenous linkage capability. The large TNCs (Intel, Abbot, and Baxter) source the major inputs from the company-internal global network, that is, a lot of the production is internalized among the affiliates spanning the globe. In addition, many of the key inputs simply cannot be produced in Costa Rica at the current juncture, either because the requisite scale is too large or the technology is too sophisticated. Most important for some TNCs, however, is the fact that Costa Rica’s manufacturing sector does not have the linkage capability necessary to supply high-tech TNCs, even with lower-tech material inputs, where scale is not an issue. Domestic linkage capability is a necessary, but not a sufficient, condition for linkages to develop. But if it did exist, it might induce some TNCs to source additional inputs in Costa Rica. Baxter in Costa Rica, for example, professes an interest in expanding domestic sourcing. And there would probably be some room for doing so, if some of the inputs were available domestically at a competitive quality and price. While protection of know-how and quality control will keep production of most key parts in-house at Baxter’s Costa Rica plant, the company’s affiliates in Singapore and Puerto Rico source some of those parts locally, because there is no quality difference between in-house production and local producers.2 Many of the small and medium-sized TNCs in Costa Rica are very eager to buy inputs domestically. Smaller TNCs, by definition, do not have the global reach to internalize production across many borders. Thus, they either import from independent suppliers abroad or buy the inputs domestically. The absence of a global network together with a relatively smaller demand for inputs means that smaller companies have much less clout to bargain for reduced input prices internationally. And it means that they have a greater interest in achieving cost reduction through local sourcing, technological sophistication, quality, and scale permitting. But even though some smaller TNCs have a greater motivation to source in Costa Rica, extensive anecdotal evidence, from my interviews and otherwise, indicates that they have encountered considerable difficulties in finding local suppliers. One small foreign company, which took the initiative to seek local suppliers for six machine shop parts, started out with the Yellow Pages. It eventually identified three suppliers willing to take on the job. But only one of them mastered the technical side of the production process, and it did not deliver on time. At that juncture, and after a process that had lasted 18 months, U.S. headquarters told the Costa Rican

THE TICO TIGER THAT HASN’T . . .

177

subsidiary to abandon all attempts at local sourcing. The operations manager who had been so keen on developing his company’s supply chain in Costa Rica summed up the experience with the comment “in the end, we are not responsible for making technical suppliers in this country.”3 But not all attempts at backward linkages have ended in failure. Another small foreign company, whose manager I interviewed, is producing specialty equipment for film cameras and very determined to source as much as possible domestically, since labor costs are so much cheaper than in Europe, where the company is headquartered. The manager estimated that the costs of producing parts locally is 40 percent lower, and the company has had an excellent experience so far in finding local producers who adopted ISO 9000 quality standards and learned the requisite technology. One national input supplier for TNCs in the medical field, Electroplast, pursued ISO 9000 certification early on, realizing that it was a sine qua non for being even considered a potential player. The company was the first among domestic producers in the field to build a clean room, positioning itself at the forefront of domestic input suppliers. Yet another local company I visited, Fema, had highly modernized CNC equipment and was in the process of building its second machine for the production of precision parts. The Femas and Electroplasts of the Costa Rican manufacturing sector demonstrate that some companies have indeed become successful material input suppliers to high-tech TNCs, producing specialty moldings and parts of metal and plastics. But these companies are far and few between. They are the exception rather than representative examples of the state of competitiveness of the indigenous manufacturing sector. Where there are success stories, TNCs have often played an important role in raising the technological capabilities of their Costa Rican suppliers. In response to my question about the most important benefits from dealing with TNCs, successful domestic input suppliers highlighted TNC support for the installation of quality control systems (e.g. ISO 9000), assistance with the acquisition of specialized machinery, and trouble shooting over production problems as they arose. In a thorough analysis of the indigenous Costa Rican metal-mechanics sector, Buitelaar, Padilla Pérez, and Urrutia-Álvarez (2000) found that quality across producers varied tremendously. Most companies are small; in fact they identified only one national company that could be characterized as large: Atlas. It produces refrigerators and stoves and has worked with a license from Westinghouse and Frigidaire since the beginning of the 1980s. Bitar (2002) estimates that, in 1998, small and mediumsized enterprises accounted for 45 percent of the 4,856 enterprises in the industrial sector.

178

FOREIGN INVESTMENT, DEVELOPMENT, AND GLOBALIZATION

But scale is not the only factor that limits the potential for backward linkages. Most potential domestic input suppliers are unable to provide the requisite quality; they do not have ISO certification or the technological know-how. Some may lack the drive of the Schumpeterian entrepreneur, a frequently voiced concern in response to my interview question why national manufacturers were having difficulties in becoming competitive input suppliers.4 Others have not attempted to become TNC input suppliers, as it is a high-risk undertaking in a Catch-22 context. In order to become actual input suppliers, national producers need to invest in ISO certification and new machinery. But finance capital is expensive in Costa Rica, and the national manufacturer will not know for certain if a contract from a TNC will indeed be forthcoming if he goes through with the investment. On the other hand, the TNC is rarely willing to commit to a firm contract, without knowing that the Costa Rican producer can deliver the requisite quality, consistently and in the time frame required. Pervasive market failures have hampered the development of indigenous linkage capabilities. Imperfect information about technology and markets, lack of access to finance capital and its high cost, and the high-risk nature of investment for many of the small Costa Rican producers are evidence that the market has been failing in its creative and coordinating functions. Such failures need to be addressed through deliberate government policies, which focus on providing information about potential clients, technical assistance, access to investment capital, and risk reduction. The government needs to deal with the competitiveness of the national industry in a systemic way, in the area of institutions, finance, infrastructure, and education. Access to finance capital is especially important, as companies need to update older and technologically more obsolete machinery. Manufacturing producers in Costa Rica who do not export have a considerably older capital stock than those who export, be they national or foreign. In 2001, nonexporters indicated that 43 percent of the machinery they used was more than 10 years old, compared to 31 percent for exporters (see table 6.3). Table 6.3 Average Age of Machinery Used in the Production Process, Costa Rican Manufacturing Firms, 2001 (Survey Results) Response to the question: What is the average age of the machinery used in the production process?

Non-exporters Exporters National Foreign Source: IICE (2001).

⬍5 years (%)

5 to 10 years (%)

⬎10 years (%)

32.0 23.1 20.7 32.1

25.0 46.2 50.9 37.5

43.0 30.8 28.4 30.4

THE TICO TIGER THAT HASN’T . . .

179

In chapter 5, I identified the absence of a coordinated institutionalized industrial development strategy in Costa Rica as a major obstacle to the realization of FDI-linked industrial upgrading. This absence has to be understood against the backdrop of the new economic model that advanced a generally free market approach and a steady retrenchment of the economic functions of the government. Under the new model successive governments in Costa Rica—as in other Latin American countries—aimed to dismantle the structure of government intervention of the ISI era rather than reform it. As a result, it is not surprising that Costa Rican governments have not done much to strengthen national linkage capability.5 Over the years, there has not been a shortage of calls for proactive linkage-creation policies in Costa Rica as well as incipient moves in that direction. A study by the Foreign Investment Advisory Service (FIAS 1995), for example, which CINDE had commissioned in the mid-1990s to explore the country’s potential competitiveness in the electronics sector, noted the need for a strategy that is more proactive in linkage development and provides a systematic approach to FDI-led development. It argued for the establishment of a “broader economic development structure.” A number of nongovernmental institutions have tried, independently of one another, to get their own program for strengthening national manufacturing off the ground. At one point in the second half of the 1990s, there were at least three somewhat competing programs: CINDE had established Programa MIL, the Chamber of Commerce and PROCOMER had started PROFOVE, and the Chamber of Industries had Unidad de Asistencia Industrial (UATI).6 As these programs were partly competing with one another, not coordinated, and certainly not integrated into a larger proactive strategy, which did not exist, Programa Mil and PROFOVE did not last very long. It was not until the launching of Costa Rica Provee (CRP) in 2001 that institutionalized strengthening of indigenous linkage capability became a real possibility. Costa Rica Provee got under way in October 2001 as a pilot project co-funded by the IADB, CINDE, and PROCOMER. Its charge was to turn potential national input suppliers into actual input suppliers of high-tech TNCs in the Free Zones.7 Four people started out to do the grunt work of matchmaking between a high-tech TNC’s interest in local sourcing and a local company, which might be able to meet that interest, after technical and financial assistance. In 2002, CRP organized a major conference for TNCs and potential input suppliers in the medical supply sector that provided a valuable opportunity for making contacts and exchanging information. By mid-2004, CRP had been instrumental in facilitating more than fifty linkages. Of them, 27 were in the medical sector and had led to total sales of $22.5 million by local companies, $8.5 million in products,

180

FOREIGN INVESTMENT, DEVELOPMENT, AND GLOBALIZATION

and $14 million in services (Costa Rica Provee 2004). CRP has also established a web presence offering its services and a listing of national suppliers organized by product (www.crprovee.com). And most importantly, in 2004, the agency became formally integrated into the export promotion-agency PROCOMER, a critical step toward grounding the program institutionally. It remains to be seen whether CRP will bring about a qualitative jump in domestic-linkage capability. The process of developing linkages is slow and gradual, as evidenced by the experience in Ireland and other countries. But for it to be successful, to the extent that it is possible in a small country, requires that the promotion of linkage capability becomes a priority, financially and institutionally. On the one hand, that means increased funding, as a handful of people working in the program are not sufficient to generate a critical mass of competitive input suppliers. On the other hand, creation of linkage capability requires synergetic, institutionalized cooperation among all the relevant players: CINDE which deals directly with current and prospective TNCs, PROCOMER which has detailed information about TNCs in the Free Zones, the Chamber of Industries which could provide technical assistance through UATI, its technical assistance unit, and the universities which have the requisite know-how and equipment in some of the technical areas.

Limited Linkages and Linkage Capability in Latin America The absence of significant linkage capability is not limited to Costa Rica, but exists in other Central and Latin American countries as well. TNC sourcing has been much less extensive in Latin America than in other parts of the world. For example, U.S. preferential trade agreements involving shared production have led to the decline of the textile industry in Costa Rica and the rise of the clothing industry, as TNCs used foreign inputs, primarily from the United States. One can observe the same trend in other Central American countries, but not necessarily in other developing areas of the world. Data on U.S. trade under HTS 9802 show that U.S. imports from Central America contain a much higher percentage of U.S. value added than U.S. imports from many other regions of the world (see table 6.4). Especially the newly industrializing countries in East Asia, with China as the latest and most important newcomer, have been able to compete across a wider range of the value chain, offering integrated services beyond the manufacturing of parts. For the smaller Latin American countries, size is an obvious limitation on competitiveness across a broader spectrum of the value chain. But the absence of proactive government policies is

THE TICO TIGER THAT HASN’T . . .

181

Table 6.4 Value of U.S. Content in U.S. Imports under HTS 9802, 2001–2003 Selected country groupings and countries

2001

2002

2003

Central America Costa Rica ASEAN China Eastern Europe OECD World

60.1 64.5 40.1 16.1 60.2 16.4 22.2

59.2 59.8 41.7 16.3 59.2 13.5 18.6

58.6 59.8 42.7 16.3 58.6 9.7 15.1

Source: United States International Trade Commission, Interactive Tariff and Trade Dataweb ⬍http://dataweb.usitc.gov/⬎ accessed November 2004.

another reason for the difference in linkage capability in some Latin American and Asian countries. In some Asian countries, governments have adopted very deliberate and aggressive policies to promote linkages between export zone producers and local suppliers. As a result of these policies, local supplies to TNCs in Taiwan, for example, rose from 8 percent of imports in 1969 to 46 percent in 1979 (Esquivel et al. 2000, 5).8 In Latin America, in contrast, governments have, as a rule, not been active in fostering linkages between foreign producers in free zones and indigenous industry. They have tended to consider Free Zones as sites for employment and foreign exchange creation, rather than as a way of upgrading, energizing, and diversifying local manufacturing. Under the new economic model the absence of proactive government policies is also the result of an excessive faith in free-market policies. The wholesale rejection of import-substitution policies in Latin America in the course of the 1980s led to a drastic retrenchment of the government as an economic actor and to a rollback of any interventionist policies, which smacked of industrial policies. The development of TNC sourcing in maquila production in Mexico, for example, has been very similar to that in Costa Rica. In 2000, local inputs for maquila-based exports in Mexico constituted a mere 2.2 percent of output value, only slightly higher than the 1.4 percent in 1993, the last year before the North American Free Trade Agreement came into force (Buitelaar and Padilla Perez 2000, 1631). U.S. TNC affiliates in the electronic and electric equipment industry sourced much less in Mexico than in other areas of the world. In 1994, U.S. TNC affiliates in Mexico imported goods for further processing equivalent to 66.5 percent of affiliates’ total sales in Mexico, compared to 22 percent for U.S. affiliates worldwide (Hanson, Mataloni, and Slaughter 2001, 265).

182

FOREIGN INVESTMENT, DEVELOPMENT, AND GLOBALIZATION

Lack of linkage capability is evidence of the limitations of indigenous knowledge-based assets. In Latin America, these limitations are also reflected in the continued dominance of primary products in the region’s exports. In 2000, after 15 years of neoliberal reforms, primary and resourcebased products still accounted for more than 70 percent of South American exports, with the exception of Brazil and Uruguay, where the shares were in the 50 percent range.9 Limited domestic linkage capability restricts TNCs’ options to source inputs in the host country. And TNCs will only go so far in trying to develop that capability, even if they have some initial interest. Knowledge Spillovers from High-Tech FDI Even though high-tech TNCs in Costa Rica have not generated many knowledge spillovers through backward linkages, they generated some spillovers through training and education. In addition to having a more up-to-date stock of machinery (see table 6.3 again), a higher percentage of foreign-owned companies have a training system in place. In 2001, 55 percent of exporters reported to have a training system, compared to 16 percent of companies producing for the domestic market only (IICE 2001). And among exporters, the percentage of foreign companies with a training system was nearly 20 percentage points larger than for national companies (see table 6.5). High-tech producers employ a higher percentage of skilled workers, which offers the potential for greater knowledge spillovers. In 2000, skilled labor constituted 13.4 percent of employment in the FZs without Intel, and 20.1 percent, when Intel is included. In metal manufacturing, where most high-tech FDI is concentrated, the percentage of skilled labor was 18.7 percent without Intel and 29.6 percent with Intel (Larudee, CorderoPeña, Arroyo-Abad 2001). At times, foreign companies provide training in Table 6.5 Availability of Training Systems in Costa Rican Manufacturing Firms (Survey Results) Response to the Question: Do you have a training system in place?

Non-exporters Exporters National Foreign Source: IICE (2001).

Yes (%)

No (%)

“N”

15.7 55.3 50.8 67.8

84.3 44.7 49.2 32.2

108 206 124 59

THE TICO TIGER THAT HASN’T . . .

183

skill areas that are not widely available in the host country. A foreign producer of precision instruments in Costa Rica, for example, emphasized that his company is one of the few in the country that uses CNC (computer numerical control). And a representative of Intel commented that even though some of the technical workers were overqualified for the fairly routine operations of assembling and testing microchips, they gained valuable experience in technical problem solving in this area. A 2004 survey of local input suppliers to TNCs showed that 6.2 percent of their managers, 27.6 percent of their engineers, and 31 percent of their technicians had previously worked for a transnational corporation (Monge, Rosales Tijerino, and Arcee Alpizar, 2004). In addition to the accumulation of skills and knowledge in the production process and the positive spillovers that come with labor mobility, foreign high-tech companies have had some positive influence on business practices, from safety standards in the work environment to planning and administrative tools, to expectations regarding quality standards and on-time delivery. Furthermore, they have had an important impact on upgrading the technical curriculum at Costa Rican universities. That is particularly true for Intel. Both the engineering school at the University of Costa Rica and the Technological Institute of Costa Rica have contracts with Intel to collaborate on the development of curricula for technical careers. The goal is to provide the most up-to-date skills and knowledge needed at different levels, from technicians to engineers. After Intel had decided to invest in Costa Rica, the TNC brought Costa Rican professionals to other Intel plants in the United States and elsewhere. And it brought faculty from the Technological University to universities in the United States to receive training and information on production processes and curricula for the type of technicians and engineers needed by Intel. The upgrading of the technical curriculum is particularly significant for the growth potential of high-tech production in Costa Rica, as it supports an expansion of the sector and, potentially, a move up the value chain within the sector as well. The fact that Costa Rica is a middle-income developing country means that high-tech production in the country is at the low-end of the high-tech spectrum. The key question is whether the value added of production in the high-tech area increases over time. If it does, and if the increase in value added is not due to transfer pricing, it would indicate a larger potential for spillovers and possibly deeper roots of the TNC affiliates in the host country. Even though the time period of more substantial high-tech TNCs in Costa Rica has been relatively short, there are first indications of an incipient upward movement along the value chain. Intel-Costa Rica, for example, invested in a component design group and in software development

184

FOREIGN INVESTMENT, DEVELOPMENT, AND GLOBALIZATION

for industrial applications for other Intel facilities. The company found that there was a good degree of expertise in software development and VLSI design both in Costa Rica as in other Latin American countries, and suggested to headquarters that they could staff a group locally to provide remote design services.10 Baxter, on the other hand, is expanding more into the centralization of back-office services, announcing the opening of a Shared Services Center in early 2005 (CINDE 2004). Technical training in Costa Rica, especially in computer technology, has given rise to a growing national software industry. Its origin predates the arrival of Intel and other high-tech TNCs, but its subsequent development was aided by the presence of high-tech TNCs, as these corporations expanded the availability of trained personnel. It is not coincidental that the software industry in Costa Rica—like in Ireland—has developed primarily on its own. With no barriers to entry and plenty of opportunities for niche production, it is easier to establish oneself on a small scale, given requisite training and access to funding. The lack of venture capital has been a problem though in Costa Rica, as in many other developing countries. A joint project of the IADB and domestic institutions, “Prosoftware” provided some support for the software sector through the provision of a venture capital fund as well as other assistance. The program which ran between 2000 and 2003 provided direct benefits for 80 companies, and at the end of 2003, Costa Rica had a pool of 4,000 IT professionals (Costa Rica Digest 2004). Most of the software services are tailor-made solutions for clients in the domestic market. Exports of the software industry are still small; they amounted to only $90 million in 2003. The most notable success story in software production has been ArtinSoft. Founded in 1993, it is now one of the world leaders in automated software migration based on artificial technology intelligence. It has partners around the world, and has provided services for several of the widely known TNCs in the IT field. While the national software sector has been growing fast over the last few years, it is still small and somewhat fragile, and a far cry from the software exports of $1.2 billion envisioned for 2010 (Costa Rica Digest 2004).

Dilemmas of Development and Political Economy The absence of targeted and coordinated government policies in support of a high-tech FDI-linked development strategy is key to understanding why high-tech FDI has not been made more successful in contributing to development in Costa Rica. The most important contribution high-tech FDI can make to the development of a host country is through technological

THE TICO TIGER THAT HASN’T . . .

185

and marketing spillovers that lead to the eventual advancement of indigenous knowledge-based assets. In SLCs, spillovers can occur through backward linkages, training and education, and demonstration effects. There have been some spillovers in Costa Rica through the education and demonstration channels, but spillovers via backward linkages have been very limited. Linkages do not develop automatically. They did not develop automatically in Ireland during the 1960s and 1970s, and they have not developed automatically in Costa Rica either. When the gap between domestic linkage capability and TNC expectations regarding input quality and price is too large, the possibilities for spillovers via linkages are very small. Domestic linkage capability is not a sufficient condition for linkage development, but it is a necessary condition. One expression of the size of the gap between linkage capability and TNC expectations is the extent of market failures under which indigenous producers operate. Imperfect information, high risk, barriers to entry, and maybe insufficient Schumpeterian animal spirits are huge obstacles to the development of a competitive indigenous supply sector, notwithstanding the occasional successful local supplier. Some might argue that the absence of significant linkage development is a reflection of the relatively short period of time that high-tech TNCs have been producing in Costa Rica. While it is certainly true that it takes time for linkages to develop, market failures do not tend to disappear with the passage of time. Lack of access to financing does not disappear automatically over time, but only as the result of deliberate government policies. Entry barriers do not disappear by themselves either. The one market failure that may indeed become less significant over time is the lack of information about requisite technology and production standards, assuming that demonstration and training effects work to make such knowledge more widely available. Costa Rica cannot afford to wait for that eventuality to materialize. The government needs to adopt a proactive strategy in order to achieve higher growth sooner rather than later, not least because higher growth is the surest path to reducing the high poverty rates. In 2004, 21.7 percent of households lived below the poverty line, compared to 20.3 percent in 1995 and 27.1 percent in 1990 (La Nación 2004 and Proyecto Estado de la Nación 2004). Another reason for the urgency behind adopting a more proactive strategy is the need to minimize the fallout and maximize the benefits from upcoming changes in international trade policies. Of primary relevance for Costa Rica are the expiration of the international quota system of trade in clothing and textiles and the possible ratification of CAFTA. World trade in textiles and apparel has been regulated since 1968, first

186

FOREIGN INVESTMENT, DEVELOPMENT, AND GLOBALIZATION

under the Multi Fiber Agreement and since the Uruguay Round under the Agreement on Textiles and Clothing (ATC). The ATC stipulates that all quotas be phased out by the beginning of 2005. The result will be a massive restructuring of production and trade of clothing on a global scale, with China bound to become the tailor of the world in the long run.11 Given Costa Rica’s proximity to the United States, the country may well continue for a while in its current line of specialization, the production of higherend undergarments for the U.S. market. Yet, in contrast to outer apparel, these products are less beholden to the fashion mandate of quick turnaround and may well be relocated to the Far East in the medium term. A likely reduction in employment in Costa Rica heightens the need for other sources of employment and economic growth. In May 2004, the Central American countries and the United States signed the CAFTA. At the beginning of 2005, the agreement still needs to be ratified by the legislatures in most of the participating countries. Ratification of CAFTA would make proactive policies all the more urgent. On the one hand, they are needed to mitigate the onslaught of import competition for small-scale producers in agriculture and manufacturing and thus to address the ramifications for employment. On the other hand, proaction is necessary to realize the potential benefits of larger inflows of FDI. In the wake of a ratification of CAFTA, some foreign companies are likely to relocate from Mexico, Canada and the United States or from outside the North American Free Trade Agreement area (NAFTA) to take advantage of Costa Rica’s location-specific assets and free access to the U.S. market. Proactive government policies in education and infrastructure aimed at improving Costa Rica’s location-specific assets would raise the likelihood of increased high-tech FDI. They would also be important to help retain foreign producers operating in the country already. Policies to expand indigenous linkage capabilities and to integrate indigenous firms into emerging clusters would enhance the development-inducing potential of larger inflows of high-tech FDI. Under the new economic model there have been two major barriers between the need for proactive coherent government policies and their actual implementation: the absence of a coherent development strategy supported by key sectors of society and the limitations on financing. The buy-in from the major groups of society is all the more important, as the new economic model has led to growing income inequality. Between 1995 and 2003, the Gini coefficient in Costa Rica increased from .377 to .425 (UNDP 2004). As if the conundrum posed by the political economy of the new economic model were not enough of a challenge, there continues to be another large obstacle to the implementation of coherent proactive polices. And that is

THE TICO TIGER THAT HASN’T . . .

187

the shortage of financial resources, internally and externally, as we already saw in chapter 5. An effective tax reform is needed that will lead to an increase in the tax ratio. The same is true for the imposition of a profit tax on companies in the Free Zones after 2007. As long as the effective rate is in line with the one charged by Costa Rica’s competitors, it is likely that TNCs will not relocate and prospective foreign investors will not turn away on account of a positive tax rate. Costa Rica needs to meet the challenges posed by both a nondevelopmentalist state and a very low tax ratio in order to have a chance at high-tech FDI-facilitated industrial upgrading and economic growth. Meeting the challenges is a necessary condition for a successful high-tech FDI-linked strategy, but it is not a sufficient condition. Success also requires TNCs with an interest in domestic sourcing and upgrading. Large TNCs like Intel, Abbot, and Baxter have their own global networks from which they source the inputs they need. However, as we have seen, there is a growing number of small and medium-sized companies in the developed countries that are looking to offshore outsourcing as a way of reducing costs. Costa Rica has a real advantage there, given its location, the similarity in culture to the United States, and the pervasiveness of a working knowledge of English among its skilled workers. If Costa Rica can continue to provide skilled, English-speaking labor at relatively high rates of productivity and if it can continue to improve its infrastructure, there can be a reasonable expectation that small and medium-sized TNCs will continue to invest in Costa Rica. That is especially true in the area of medical devices. Given the industry’s high standards for quality, safety, and clean room production, foreign producers are less likely to rely on global contract manufacturers in the near future. Costa Rica’s enormous richness in biodiversity should provide the basis of a strong comparative advantage in the future, particularly in the context of an expanding sector of medical devices. One of the emerging global trends in that industry has been the combination of medical devices with pharmaceutical products. Continued attraction of smaller, rather than larger, TNCs would be in line with the rate at which Costa Rica’s technical labor force and infrastructure can expand. A successful high-tech FDI-linked strategy could make a huge contribution to the development of the country’s knowledge-based assets. The obstacles to achieving it are substantial. But Costa Rica—as any other country like it—needs to aim for the high road to competitiveness, based on productivity and wage growth. And thus it needs to confront these obstacles, because high-tech FDI in combination with domestic linkage capability provides a shot at the high road. The low road of insufficient productivity growth and declining wages is not a road to development, but it is de facto the default option. It is important to remember though

188

FOREIGN INVESTMENT, DEVELOPMENT, AND GLOBALIZATION

that even in the best-case scenario, high-tech FDI can be but one engine among several to generate economic growth and especially employment. Nontraditional agriculture, tourism, and the indigenous software sector are the most likely other sources of growth. Costa Rica has been able to attract high-tech FDI because of the legacy of its past policies in education, infrastructure, social peace, and its more current policies of aggressive pursuit of FDI and skillful, but dispersed administrative capacity. That set of location-specific assets attracted Intel. And Intel’s decision put the country on the map for potential investors at a time when a growing number of medium-sized TNCs from developed countries started to look for opportunities for offshore outsourcing. However, business as usual will not keep TNCs coming, and it certainly will not lead to the desired spillover effects for the expansion of indigenous knowledge-based assets. As global competition keeps increasing, with China, India, and Eastern Europe aggressively vying for high-tech FDI, proactive policies will become ever more urgent: from targeting specific TNCs to supporting cluster formation, to expanding technical education, to improving access to telecommunications, to aggressive support policies to develop indigenous linkage capability and integrate national firms into emerging clusters. It is a tall order, but in the absence of such policies, there is no chance that the potential promise of high-tech FDI-linked economic upgrading and development can be realized.

7

Lessons from Ireland and Costa Rica

s the twenty-first century begins, a large and growing number of developing (and developed) countries are competing fiercely for foreign direct investment. They hope that FDI will provide a major impetus for economic development by increasing total investment and by generating positive spillovers through the transfer of technology, marketing, and business practices. The realization of such expectations is not automatic. Developing countries will only reap long-term benefits from FDI if there is a coincidence between TNCs’ strategic global interests and host country capabilities, at a particular point in time as well as over the long term. In this book, I have focused on high-tech FDI and small latecomers, because high-tech FDI offers the greatest potential for industrial advancement for small countries whose small market size imposes constraints on indigenously grown industrialization. I offered a conceptual framework for analyzing these contingencies and examined how they played out in the cases of Ireland and Costa Rica. Both are small countries that have been successful in attracting high-tech FDI. But they were at different levels of development when they began to pursue high-tech FDI, and have experienced very different growth outcomes to date. Large inflows of high-tech FDI were a critical driver behind the transformation of the Irish economy into the Celtic Tiger during the 1990s. A fortuitous conjuncture of TNCs’ strategic needs and Ireland’s locationspecific assets opened the possibility for a surge in economic growth. The demand boom in the PC industry and the desire to be closer to the European consumer market drove TNCs to increase their investment in Western Europe in the late 1980s. Their interest was heightened by “Opportunity and Fortress Europe,” which resulted from the Single European Act of 1986. It was in this context that Ireland had the right location-specific assets to offer: a cheap, skilled, English-speaking workforce, appropriate

A

190

FOREIGN INVESTMENT, DEVELOPMENT, AND GLOBALIZATION

infrastructure, macroeconomic and political stability, stable capital labor relations, and very attractive financial incentives in the form of low tax rates and generous grants. The dramatic increase in EU structural funds contributed significantly to the formation and expansion of Ireland’s location-specific assets, particularly through the funding of education and infrastructure projects. Ireland provides a fascinating case study of changing dynamics in a high-tech FDI-linked development strategy, as the global economy and the country’s location-specific assets changed over time. In the course of the 1990s, the Celtic Tiger unleashed a dynamic that altered the locationspecific advantages Ireland had at the beginning of the decade. As the country moved from a high unemployment to a full employment economy, previously cheap skilled labor became more expensive, in absolute terms and in comparison with other countries vying for the production of standardized high-tech TNC production. At the global level, increased competitive pressures together with a slowdown in economic growth and the collapse of the technology bubble in 2000 led to a radical restructuring of the electronics industries. TNCs aggressively relocated commodity production to other low-cost sites in the world, especially to more recent competitors for FDI: China, India, and countries in Central and Eastern Europe. The confluence of these internal and external changes led to the departure from Ireland of a number of TNCs in the electronics industry and to a considerable slowdown in economic growth. In anticipation of these changes, the Irish government began to modify its development strategy, focusing on how to make productivity growth rather than factor accumulation the driving force behind the revival of the Celtic Tiger. The new strategy aims to promote R&D-based growth and to support foreign and indigenous companies in their move up the value chain. But it faces huge challenges, as the country has to move on many fronts at the same time. One of the critical conditions necessary for success is in place: a proactive government with a vision of where to go and the steps that need to be taken to get there. The next five to ten years will show whether the challenges can indeed be met. The main reason high-tech FDI can provide a powerful impetus for industrial upgrading is its potential impact on the development of indigenous knowledge-based assets. In the case of Ireland, technology transfer happened through linkages as well as human capital spillovers. But linkages did not happen automatically. The existence of pervasive market failures negates the possibility of an automatic emergence and expansion of indigenous linkage capabilities. Linkages eventually started to develop in the mid-1980s, after the government adopted proactive policies and

LESSONS FROM IRELAND AND COSTA RICA

191

when TNCs in the computer industry were willing to source some material inputs domestically. But linkage intensity reached a plateau in the second half of the 1990s and then started to decline somewhat, mainly due to changed TNC strategies. TNC-generated spillovers definitely contributed to an expansion of indigenous knowledge-based assets. Yet, indigenous producers’ ability to adjust to the current restructuring of the economy and to be successful participants in the new development strategy would probably be further along if the government had focused on the establishment of linkage capability earlier than it did, and if it had focused also more broadly on cluster formation, as a number of analysts had urged (e.g. O’Sullivan 1995 and 2000). The fact that the indigenous software industry in Ireland has emerged and grown largely on its own underscores the point that indigenous capabilities do not arise in the presence of pervasive market failures. Software producers operate in a high-tech industry where market failures are not very significant. There are no barriers to entry, niche-production is possible, and the education system is at a level high enough to generate computer and software engineers. As a result, it is not so hard for an enterprising individual with the requisite knowledge and a limited amount of seed money to get started. Down the road, size does become a challenge though, as companies need to establish a global presence to remain competitive. Can a small country at a lower level of development than Ireland represented in the 1960s and 1970s emulate that country’s economic success with high-tech FDI, or are the factors and contingencies behind the Celtic Tiger unique? The answer is not straightforward. Some of the country, time, and context-specific factors are indeed unique. But uniqueness does not necessarily mean that they are the only possible factors that can engender a positive high-tech FDI-development nexus. Some of the policies are replicable, and some of the policy mistakes are avoidable. In this concluding chapter I discuss these factors in greater detail, informed by the insights of the comparative study of Ireland and Costa Rica. Costa Rica met the first challenge in the high-tech FDI development sequence when it succeeded in attracting the investment of one of the world’s leading high-tech TNCs. As in the case of Ireland, Intel’s investment had a huge signaling value, demonstrating to international producers that Costa Rica was a promising location for high-tech production. But at this juncture, Costa Rica is a long way from joining the tiger club. There are multiple reasons for Costa Rica’s lackluster economic growth, not all of them related to FDI. One reason is simply bad luck: no sooner was Intel gearing up production in San José in 1997, when the technology bubble burst in 2000 and a global recession developed. But that is a cyclical effect, which will disappear with an upturn in the cycle. Another reason is

192

FOREIGN INVESTMENT, DEVELOPMENT, AND GLOBALIZATION

the relatively short period of time during which the country has been on the radar screen of high-tech TNCs. Even though Costa Rica has attracted foreign investment in the electronics industry and the medical devices sector, most of these investments have been relatively small, with the exception of Abbot’s investment in 2000. The quantity of high-tech FDI Costa Rica has attracted so far has not been enough to generate a surge in economic growth. At the same time, the supply of highly skilled labor is also not so large that the country could sustain huge inflows of FDI, even if TNCs wanted to invest. But the main challenges Costa Rica faces in advancing a high-tech FDI-linked development agenda are financial constraints and the absence of a cohesive government strategy with wellintegrated proactive policies to foster linkage capability and advance key aspects of the country’s location-specific advantages. The comparative analysis of Ireland and Costa Rica leads to important insights for the conditions under which high-tech FDI can contribute to enhanced growth and development. Lessons for other SLCs need to be extracted with care, as a comparative study of two countries provides a limited basis from which to generalize. Nonetheless the proposed lessons can be assessed against the experiences of other country cases and can inform future cross-country studies. And they provide food for thought for policy-makers interested in FDI-linked development strategies. Even though this book concentrates on the nexus between high-tech FDI and small countries, the discussion that follows suggests that some of its conclusions are relevant for any type of FDI and any size of host country.

Location-Specific Assets for High-Tech FDI in SLCs: Geography, Education, and Infrastructure There is a basic set of location-specific assets, which any country needs to attract FDI, regardless of its size or the type of FDI it hopes to attract: stable property rights (though not necessarily all private, as the case of China demonstrates), political stability (though not necessarily in the context of democracy, as evidenced by the example of Vietnam), stable capital–labor relations, macroeconomic stability, and minimal levels of infrastructure in terms of public utilities, transportation, and communication. Geography which has regained importance in analyses of underdevelopment in recent years is particularly important for SLCs in search of high-tech foreign investors. By definition, TNCs do not invest in small countries to service the host country market, but rather to use the country as a base from which to service larger markets. Both Costa Rica and Ireland are blessed with proximity to major markets, with Costa Rica close to the

LESSONS FROM IRELAND AND COSTA RICA

193

United States and Ireland a member of the European Union and close to the European mainland. Ceteris paribus, small countries that are far away from large markets will have a hard time attracting high-tech FDI. Beyond the basic set of location-specific assets and physical location, SLCs in pursuit of high-tech FDI need to have a more highly developed infrastructure, especially of air transport and telecommunications, and a comparatively cheap educated and skilled labor force. The prerequisites of better infrastructure and widespread education are another way of saying that developing countries have to be at a certain level of development already to have a chance of attracting high-tech FDI. At the same time, there is not a one-to-one mapping from GNP per capita to education and skill levels. A country at a given income level with a comparatively more educated labor force has an obvious advantage, assuming that the advantage is not negated by an equivalently higher unit labor cost. Path dependency is critical for the creation of such favorable circumstances. Costa Rica’s historical emphasis on education and greater income equality is responsible for the existence of a labor force in the 1990s that was more educated, skilled, and productive than that of many other SLCs at the same level of GNP per capita. The absolute number of skilled and educated workers in an SLC is one factor that ceteris paribus influences the amount of high-tech FDI an SLC can attract. One of the reasons Ireland was able to attract large inflows of high-tech FDI was the relatively large number of engineers and other highly trained people it could provide. Even though only a relatively small percentage of the workers in the high-tech sector actually were highly skilled technical workers, the total demand for them was still considerable. The supply of highly skilled technical labor was expanded by the existence of an educated expatriate community, many of whom were eager to return to Ireland, as job opportunities became available. Because Costa Rica is at a much lower level of development, it would be in no position to accommodate a similarly large inflow of high-tech FDI. The experiences of Ireland and Costa Rica suggest that unless an SLC has levels of educated and skilled labor that are exceptionally high for its GNP per capita, the amount of high-tech FDI cannot be as large as in the case of Ireland. Given the heavy emphasis that past and present socialist countries have put on education generally and science education in particular, some of them are in an excellent position to compete for high-tech FDI, at least insofar as the availability of an educated labor force is concerned. In the mid-1990s, the latest date for which comparable data are available, the share of students at the tertiary level who studied science was 34 percent in the Czech Republic, 32 percent in Hungary, and 43 percent in Slovakia, compared to 30 percent in Ireland and 18 percent in Costa Rica

194

FOREIGN INVESTMENT, DEVELOPMENT, AND GLOBALIZATION

(UNDP 2004, 176). It was 53 percent for China, the highest percentage among all the countries listed in the Human Development Report 2004. That is a powerful indicator of China’s potential as a competitor in high-tech products, and not just labor-intensive commodities.

Can SLC Abilities Meet TNC Needs? Whether high-tech TNCs invest in SLCs and whether—once there—they buy inputs from indigenous producers and build backward linkages depends on the coincidence between SLC capabilities and TNC needs. The study of Costa Rica and Ireland showed that TNCs’ strategic global needs are not invariant across time and industry, with respect to initial investments, continuing investments, and sourcing in the host country. International competition and the internationalization of production have intensified considerably over the course of the last decade. The increase in the number of FDI-seeking low wage countries with large amounts of unskilled and skilled labor has offered huge new opportunities to companies in the more developed countries for cost reductions through offshore outsourcing. Due to its size, China has played a particularly important role since its entry into the WTO in December of 2001. One of the consequences has been a restructuring of global value chains, with the relocation of a significant part of commodity production to Central and Eastern Europe and China, of products and components, that is, for which quality differentiation does not matter, but economies of scale and market power bring huge gains. In the case of Ireland we saw how contract manufacturers in the electronics industry came and went in the second half of the 1990s, first attracted by the Celtic Tiger, but then—as conditions changed—relocating to Central Europe and China. Considerations of risk and market diversification will limit the share of commodity production that will move to China and other developing countries. But, apart from that, SLCs will not be in a position to attract production of standardized high-tech components and products because they cannot offer the requisite scale and probably not the low labor costs either. For Ireland, the restructuring of the global commodity chain in the electronics industry heightened the pressure on companies to switch to higher value added parts, R&D and design on one end and marketing on the other. As long as they can offer the right location-specific assets, SLCs will continue to have opportunities to attract high-tech foreign producers. Those opportunities will be in areas where scale economies are limited or not important, and where quality and intellectual property rights are particularly

LESSONS FROM IRELAND AND COSTA RICA

195

important. A country like Costa Rica can continue to be an attractive location to small and medium-sized companies in subsectors of the electronics industry for example. Cultural affinity to the United States and Western Europe and a supply of skilled workers with some knowledge of English are particularly attractive to first or second time foreign investors. And SLCs can also attract affiliates of large TNCs in production areas where scale economies are less important, as in medical devices. In my comparative country study I explored whether industry differences mattered for the high-tech FDI-development nexus. Using the electronics and the medical-device industries as two examples, I found that differences in the nature of the industries were reflected in differences in the organization of the value chains on a global level. In contrast to the computer and electronics industry, the medical-device industry encompasses a range of heterogeneous products and components, none of which is exactly like the other, thus making for smaller runs and limited possibilities for economies of scale and global contract manufacturers. In addition, quality and intellectual property rights are of paramount importance in the production of medical devices. To ensure the effectiveness and safety of medical devices, the U.S. FDA and its counterparts in other countries impose strict regulations. The importance of FDA validation and of the protection of companies’ proprietary knowledge, which is often embedded in the production process itself, means that TNCs in the medicaldevice industry are not as apt to change production sites quickly. And they are not likely to have production or sourcing sites in countries where the enforcement of intellectual property rights is questionable. Another important factor influencing TNCs’ investment decisions are changes in trade rules. How they affect SLCs depends on the specifics of the rules. “Europe 1992” provided an important impetus for U.S. TNCs to invest in the EU and, with that, in Ireland. China’s entry into the WTO led to a surge in FDI inflows, from a level that was already very high. After the introduction of the Caribbean Basin Initiative, FDI flows into Central American and Caribbean countries rose considerably. And after the North American Free Trade Agreement went into effect in 1994, net FDI flows to Mexico more than doubled. If CAFTA is passed, then ceteris paribus Costa Rica can expect to be a more attractive location for FDI. By the same token, the new members of the European Union will likely see increased inflows of FDI, particularly from the older member countries. The attraction of high-tech FDI is only the first step in the realization of an FDI-development link. The second, and ultimately more important, step is to obtain positive spillovers in the host economy. The analysis showed that material input sourcing by high-tech TNCs in Ireland and Costa Rica has been limited. One reason, of course, is that a large number

196

FOREIGN INVESTMENT, DEVELOPMENT, AND GLOBALIZATION

of requisite inputs are beyond the scale or technological sophistication of indigenous producers. But there are other inputs that could potentially be sourced in the SLC. If they are not, it is either because indigenous producers do not have the capabilities, because TNCs have internalized the production of the inputs, or because TNCs are buying the inputs from a list of approved vendors. The interests of large TNCs in sourcing in the host SLC have declined even further over the last decade as TNCs have aimed to lower costs by reducing the number of global input suppliers, thus requiring a scale that indigenous producers in SLCs cannot meet. The existence of approved vendor lists makes it harder for indigenous producers to become competitive suppliers, as it raises the barriers to entry considerably. Since the small and medium-sized TNCs do not have a well-developed global value chain, they are most likely to have an interest in indigenous sourcing. Metal and plastic parts are the primary areas for possible sourcing. But if indigenous producers cannot meet TNCs’ expectations in terms of quality, on-time delivery, and price, then most TNCs will not make the effort to invest in the development of indigenous linkage capability. The manager of the small foreign electronics company in Costa Rica who said, “in the end, we are not responsible for making technical suppliers in this country,” conveys the same sentiment as the high-level official of a U.S. computer company in Dublin who said, “if they want to play the game, they need to step up to the plate.”

Advancement of Indigenous Knowledge-Based Assets The core issue of this book is the impact of high-tech FDI on the productive structure of the host economy and its knowledge-based assets. High-tech FDI dramatically changed the profile of manufacturing output and exports in Ireland and Costa Rica. Whether that absolute and relative shift toward high-tech production will eventually reflect greater indigenous technological capabilities is the key question. With its small internal market and the need to specialize, an SLC will never be able to depend only on its own knowledge-based assets for future growth and development. Self-reliance is neither feasible nor desirable in this age of globalization. At the end of the day, sustainable growth has to be based on a positive and symbiotic dynamic between foreign and indigenous knowledge-based assets. The expansion of indigenous knowledge-based assets is important, since an advancement of indigenous technological, production, and marketing capabilities reduces the dependence on TNCs, increases the possibility for some reliance on indigenously based growth, and supports firms’ ability to adjust to the requirements of

LESSONS FROM IRELAND AND COSTA RICA

197

a continuously changing global market. In a dynamic context, upgrading of indigenous capabilities has to happen along with a move up the value chain in TNCs’ production in the host country, not least because in the long run, foreign companies will always tend to be more mobile and footloose. Due to scale factors and the tendency for companies to keep R&D operations in their home country, indigenous companies in the high-tech sector will likely keep key parts of their operations in the SLC and they will be less likely to leave the SLC completely. The extent and type of skilled labor an SLC can offer influences not only the quantity but also the type of high-tech FDI it may be able to attract. Intel built a wafer production facility in Ireland that requires a large amount of specialty engineers, but it built a testing and assembly chip factory in Costa Rica that demands far fewer highly trained people. From the vantage point of an SLC like Costa Rica, the main point is to enter the high-tech value chain at whichever point it can and then—over time— move up. You cannot go straight to the top, but need to start on the ground floor. Ireland and Costa Rica are at different levels of development, but for development in both countries it is important that TNC production moves up the value chain: on the one hand because it increases the potential for greater spillovers, and on the other because the rest of the world does not stand still and competitive advantages change. But whether there will indeed be a movement up the value chain depends among other things on the host country’s ability to expand the supply of the requisite technically trained and specialized workforce, on upgrading of the infrastructure, and also on indigenous technological capabilities. FDI can enhance indigenous technological capabilities through three channels that can be distinguished analytically, but in reality often overlap: the demonstration effect, the training effect, and the linkage effect. In the case of high-tech FDI these effects can be particularly powerful. My analysis of high-tech FDI in Ireland and Costa Rica reveals interesting similarities in the importance of the different spillover channels. In both countries, TNC sourcing of indigenous inputs was basically limited to packaging materials, printing materials, and services. Most of the services were linked to cleaning, meals, and transportation, but in Ireland services included as well some that were business-related such as accounting and legal services. Sourcing of other material inputs has basically not happened in Costa Rica, and it did not happen in Ireland until the government adopted policies targeted at promoting indigenous linkage capabilities. When the technological gap, broadly defined, between TNC producers and potential indigenous input suppliers is too large, backward linkages will not materialize unless and until active policies are adopted to narrow the gap. Hightech TNCs have exacting standards for input suppliers regarding price,

198

FOREIGN INVESTMENT, DEVELOPMENT, AND GLOBALIZATION

quality standards, and delivery. In Ireland, indigenous producers who met those demands often received support from Enterprise Ireland and sometimes from TNCs as well, especially in the introduction of quality control systems and in advice on problems that arose. In both countries, high-tech TNCs have had an important impact on shaping the nature of technical education, primarily at the post-secondary level. There clearly is a bifurcation in the skill requirements of high-tech TNC production. At one end of the spectrum, production activities comprise a lot of assembly-type work, even though workers may be operating sophisticated machinery. At the other end of the skills spectrum, high-tech TNCs employ a number of highly educated technical workers, from technicians to engineers to scientists. The skill-intensity in the high-tech sectors was above the average for the manufacturing sector in both countries. The fact that it increased over time in Ireland suggests that there was some upward movement in the value chain of technology-intensive TNC production in Ireland. TNC training of employees can also have positive inter-temporal spillovers. By upgrading the skills of the work force, TNCs facilitate the attraction of the next round of foreign investors. That was the case in Ireland, where high-tech FDI in the 1990s was the third wave of FDI inflows since the early 1970s, and it was true in Costa Rica as well, though to a much lesser extent. TNCs also provide a training ground for local entrepreneurs. Through their work experience with foreign high-tech companies, Irish men (and a few women) acquired skills and insights into production processes and the workings of the global market that allowed them to venture out on their own. A small number of Irish-owned companies in the electrical and electronics sector started out as sub-suppliers to TNCs and then became multinational companies themselves, thus following the supplier-based upgrading model. In Ireland, indigenous producers in the high-tech sectors have come a long way from where they were in 1985, when the NLP started, but they still have a long way to go. The O’Driscoll Report found that Irish companies have acquired significant expertise in manufacturing operations, but that they are lacking at the front and back end of the supply chain, namely R&D on the one hand, and sales and marketing on the other. In the first decade of the twenty-first century, the manufacturing sector in Costa Rica looks rather similar to Ireland’s manufacturing sector in the 1970s. The indigenous sector is stagnating, while the foreign-owned sector is growing. The Costa Rican companies that have become input suppliers to high-tech TNCs are few and far between. In part, that is the result of the limited amount of high-tech investment in the country so far. But a more

LESSONS FROM IRELAND AND COSTA RICA

199

important reason has been the neglect of the indigenous manufacturing sector under the new economic model. And in that respect, Costa Rica and Ireland have indeed been remarkably similar. Ireland started to pursue an open-economy model in the 1960s, but the national linkage program did not start until the mid-1980s. Costa Rica embarked on the new economic model in the 1980s, and it was not until 2001 that the first serious, albeit very small, attempts got under way to institutionalize policies aimed at the promotion of national linkage capability.

Industrial Policies for High-Tech FDI-Linked Development One of the main lessons that the Irish and Costa Rican experiences suggest for other SLCs is the importance of proactive government policies. They are needed to attract FDI, to promote indigenous linkage capabilities, and to enhance key location-specific assets on an ongoing basis in the context of a coordinated policy framework. With the establishment of the IDA in the late 1950s, Ireland was at the forefront of establishing an agency charged with aggressively pursuing potential foreign investors. The rationale behind such an agency is threefold: to overcome the problem of imperfect information that transnational corporations have about the potential host country; to reduce transaction costs for TNCs by providing one-stop support throughout the investment exploration process; and to target specific TNCs for the attainment of specific development goals. Overcoming imperfect information is particularly important when a country first tries to attract FDI. But as the potential comparative advantages of the country become clearer and the country’s location-specific advantages become better known, policies can become more targeted. Attracting the first high-profile high-tech TNC is of key importance; its demonstration effect is invaluable. That was true for Intel’s investment in Ireland in 1990 and for its investment in Costa Rica in 1997. Such an investment bestows a seal of confidence on the host country, which reduces other TNCs’ uncertainty about the country’s potential as an investment site. Many countries have followed the Irish example and established the equivalent to the IDA. Costa Rica was one of the first to do so, with an IDA official serving as a consultant in the formation of CINDE. Many of the countries in Central and Eastern Europe established similar agencies in the 1990s, for example, Czech Invest, the Polish Information and Foreign Investment Agency, the Hungarian Investment and Trade Development Agency, and Invest Bulgaria Agency. Other agencies have been set up with

200

FOREIGN INVESTMENT, DEVELOPMENT, AND GLOBALIZATION

an explicit focus on high-tech FDI, like Chile Invest for example, which was established in 2000 as a high-technology investment program. From the perspective of institutional effectiveness, the small size of a country can be an advantage in the negotiation process with TNCs, since it can be easier to bring together all decision-makers, who play an important role in addressing the TNC concerns related to different aspects of a possible investment. When the level of coordination among the relevant institutions is high, the outcome can be a speedy and effective process from beginning to end. But that advantage will only work, if there is effective multiagency cooperation. And such cooperation is more likely to exist in countries with a developmentalist state, a state with a concrete vision of development goals and of the needed steps to move toward their achievement. In Ireland, in contrast to Costa Rica, the government was committed to an FDI-linked development strategy and had an agenda for its implementation. Nonetheless, while the Irish government for decades has been unwavering in its pursuit of FDI-linked development, it did not assign a significant role to the advancement of indigenous industry in the medium and hightech technology areas for a long time. I have argued throughout this book that the pervasiveness of market failures that small indigenous producers face prevents the development of indigenous linkage capability. Imperfect information, coordination failures, barriers to entry, and lack of access to finance capital require policy responses in each of these areas. The small size of the home market makes it imperative for indigenous producers to set their sights beyond the domestic market almost from the beginning. The presence of high-tech TNCs in the country is an opportunity to learn to become competitive “on home territory,” at least in areas where there is TNC interest in local sourcing. It presents an opportunity that indigenous companies are more likely able to take advantage of, if the government adopts policies to overcome market failures. At the same time, the small size of the average indigenous company makes it important to promote networks and collaborations among indigenous firms and as a part of larger clusters with foreign companies. INCAE has been at the forefront of cluster promotion in Costa Rica. In the context of Programa Impulso, INCAE, CINDE, and the President’s Office under Rodriguez developed a plan for a cluster in medical devices. But the plan did not include a strategy on how best to integrate indigenous firms. Costa Rica is not alone in the neglect of domestic producers. Based on an analysis of 12 clusters in Latin America, Pietrobelli and Rabellotti (2004) argue, “the major shortcoming of the current policy approach in most countries is the lack of an integrated and consistent vision of local SME [small and medium-sized enterprise] development and upgrading.”

LESSONS FROM IRELAND AND COSTA RICA

201

Promoting indigenous linkage capability has to be an integral part of industrial policies more generally. The very same market failures that prevent the automatic development of indigenous linkage capability are also the ones that prevent advancement and diversification of indigenous economic activity more generally, in the manufacturing sector and otherwise. Adopting proactive policies to overcome market failures does not mean that governments have a better understanding than the private sector of the economic areas of profitable engagement. Rather, it is about public–private cooperation in identifying where the key obstacles lie, in our case with respect to linkage formation and integration into clusters. Then policies to address these obstacles must be devised and implemented. Such policies need to have built-in control mechanisms, with clear rules of accountability, reciprocity, and enforcement. There is a growing literature on how best to structure such policies to increase the likelihood of their success (e.g. Amsden 2003; Mortimore and Peres 1998; Rodrik 2004). The specifics depend on the particular country context, that is, available resources and institutional capabilities. The challenges to the adoption of industrial policies are primarily political, not technical. One of the reasons why governments have not focused on policies to enhance indigenous manufacturing capabilities in specific areas has been the prevalence of Washington Consensus policies during the last twenty years. Industrial policies under ISI were often not successful, when they only provided incentives—like subsidies or protection without a clear date of termination—but no enforceable accountability. But they were very successful when they included both, as evidenced by the experience of the Asian Tigers. The Washington Consensus failed to distinguish among conditions that determine the success of industrial policies. The adoption of free-market policies and the focus on “getting the government out of the economy” and “getting the prices right” were meant to solve problems seen to be the result of past government intervention. But the embrace of Washington Consensus policies cannot be the only explanation for the absence of industrial policies, since they were rarely adopted uniformally across the board. The adoption of export subsidies (CAT) in Costa Rica during the 1980s and the preferential treatment of foreign corporations, through IDA and CINDE, through tax exemptions—and in the case of Ireland generous grants—are evidence that governments did not completely shy away from intervention. Today, SLCs are in a situation where the disappointing results of the Washington Consensus in many countries have generated a renewed emphasis on the nature and role of market failures and on what policies are required to overcome them. That discussion constitutes an important context of policy legitimization for governments to move toward more

202

FOREIGN INVESTMENT, DEVELOPMENT, AND GLOBALIZATION

proactive policies aimed at national producers. In the end, political economy and politics will determine whether SLC governments will actually embrace the need for new industrial policies in the age of globalization. Sánchez-Ancochea (2004, 13) articulated this challenge succinctly with the comment that the question is “how to design policy interventions that are both consistent with each country’s class structure and simultaneously shift that class structure toward a more ‘development-friendly’ one.” In a dynamic context, host countries’ location-specific assets have to keep improving over time to provide the conditions for high-tech TNCs to stay in the country and move up the value chain, and for indigenous producers to advance as well. The imperative to improve location-specific assets over time raises two important issues. The first is the need for coordinated and targeted policies regarding education, infrastructure, and so on. That need speaks again to political and institutional issues, as it takes a government with a vision to plan them and institutional structures to implement them. The second to which I now turn to, is funding.

High-Tech FDI and the Resource Constraint of SLCs If SLCs want to keep high-tech TNCs, they need to expand their locationspecific assets, especially in education and infrastructure. Such a need is, of course, integral to the pursuit of economic growth and development more generally. The great challenge for any latecomer in the development process is to muster the necessary resources to finance expansions and improvements in education and infrastructure. The larger the quantity of resources that can be mobilized, either internally through increased tax revenue or externally through grants or loans, the more the fiscal constraint on growth can be loosened. And the more quickly it can be done, the more quickly location-specific assets can be advanced. Ireland was in the very fortunate position of receiving large inflows of funding from Brussels through the EU Structural Funds. An injection of external funding of the magnitude of the Marshall Funds was a key reason for Ireland’s ability to grow so quickly in the 1990s and to attract such a large amount of high-tech FDI. Marshall Aid, which for Ireland came primarily in the form of loans, amounted to 2.7 percent of the country’s GDP. EU Structural Funds, which are in the form of grants, were equivalent to 1.9 percent of GNP during the 1980s, 3.5 percent of GNP in the early 1990s, and 2.5 percent during the rest of the 1990s. Funds were geared mainly toward infrastructure projects, education, and partially also toward grants for TNCs.

LESSONS FROM IRELAND AND COSTA RICA

203

The members of the EC have always believed that the achievement of a true common market can only be politically sustainable with income convergence among the member countries. Redistribution of funds among the member countries was to speed up the convergence process, with a net transfer of funds from the richer to the poorer countries and regions. The newest members of the EU who joined in May 2004 will be the beneficiaries of net transfers in the future, though probably not to the same extent as Ireland in the 1990s. By the end of 2004, EU members had agreed to freeze structural funds for 2006–2012 at the current absolute level, but they had not yet agreed on the exact formula for their distribution. Other SLCs are not as fortunate as the members of the European Union. The rules governing free trade associations, like NAFTA or CAFTA for example, are focused on trade issues (tariffs, rules of origin, etc.), but they do not include any income transfers. Income convergence is left to the market. Thus, non-EU SLCs are limited to multilateral organizations, such as the IADB as possible sources of outside funding. Given the level of development they have reached already, funding will be in the form of loans, where interest payments and amortization reduce future net inflows. In the extreme, net inflows can become negative, as did indeed happen in Costa Rica in the second half of the 1990s. The more limited access to external development funding for non-EU SLCs has two major implications. First, ceteris paribus, improvement of education and infrastructure will be less rapid, thus generating less rapid growth. And second, it will increase the pressure to raise funds internally through higher taxes. The Costa Rican experience shows that SLCs find themselves somewhere between a rock and a hard place in their attempts to attract high-tech FDI and to raise tax revenue. SLCs at the income level of Costa Rica are in no position to use grants to attract FDI. Thus to be able to compete for FDI, many have offered tax incentives including zero profit taxes and tariff-free imports. It is difficult to argue that high-tech TNCs have reduced tax revenues, since that would imply that they crowded out investments by domestic producers who would have paid taxes, which is not likely. Rather, the main result is that TNC production in the country does not contribute to increasing tax revenue, except for social security contributions and taxes paid by its employees. Since WTO rules stipulate an end to preferential treatment of exports, for example, zero profits in export processing zones, countries like Costa Rica have to establish a national policy of positive taxes on TNCs. As long as new tax rates do not surpass the ones in competing host countries, countries like Costa Rica should see increased tax revenue from TNCs. But there is every reason to believe that the race to the bottom will continue with respect to tax rates on TNCs. On the one hand, the mobility

204

FOREIGN INVESTMENT, DEVELOPMENT, AND GLOBALIZATION

of capital across national borders gives TNCs enormous bargaining power, especially as developing countries are competing for their attention. And on the other hand, countries can adopt other preferential tax-related policies that do not violate WTO regulations. Countries can offer tax credits for R&D, for example, for whom all producers are eligible, but which de facto will benefit TNCs primarily. Or special incentives can be provided at the regional or local level. While this has been the case in developed countries for a long time, it is increasingly also the case in the more developed of the developing countries. Song (2004) reports, for example, that Ochang, a city of 100,000 in South Korea, wants to build a technology cluster, as existing industries can no longer compete with China. Ochang offers high-tech foreign investors free land for up to 50 years, zero taxes for seven years, and cash grants. Unless there are international rules that put an end to tax-based competition for investment, it is hard to see why a tax race to the bottom would stop. For SLCs (and other countries) this implies heightened pressure on the national tax base. For Costa Rica, an SLC where tax revenue accounts for only 13 percent of GDP, effective tax reform becomes imperative, for development generally and for the ability to provide a context in which high-tech TNCs can move further up the value chain. High-Tech FDI-Linked Development in SLCs: Concluding Thoughts Can Costa Rica become Ireland? Can a small country at a lower level of development than Ireland in the 1960s and 1970s emulate Ireland’s economic success with high-tech FDI or are the factors and contingencies behind the Celtic Tiger unique? My analysis suggests that the type of production that high-tech TNCs will undertake in SLCs will be different from Ireland’s initial experience, since TNCs will continue to relocate a large share of commodity production in the high-tech areas to low-wage Asian countries, which have large supplies of unskilled and skilled workers. Nonetheless, there are still plenty of opportunities for SLCs to tap into those parts of high-tech TNC production that are not sensitive to economies of scale, and where quality and intellectual property rights matter. To the extent that a lower level of GNP per capita goes along with a more limited supply of skilled and highly trained labor and that net transfer of funds from abroad will be more limited (and in loan rather than grant form), SLCs will be likely to attract smaller amounts of hightech FDI. Those can make smaller, but nonetheless, positive contributions to economic growth, though not as fast as Ireland’s growth in the 1990s.

LESSONS FROM IRELAND AND COSTA RICA

205

The new EU member countries in Central and Eastern Europe will possibly have a different experience, since they have exceptionally high levels of skilled labor and will soon benefit from injections of EU Structural Funds, which will help maintain or expand their location-specific advantages. For high-tech FDI to translate into long-term contributions to development, it is imperative that SLCs adopt industrial policies to enhance indigenous linkage capability and support the formation of clusters, which include transnational corporations as well as national input suppliers. SLCs need to avoid Ireland’s mistake of paying only belated attention to indigenous linkage capability and to cluster promotion. Whether SLCs can move in that direction is foremost a political challenge, and only secondarily a technical question of how to devise such policies. The other major challenge SLCs face is limited capacity to increase tax revenues to upgrade and expand education and infrastructure. Here the major challenge is political as well. Moving up the value chain inevitably means an increase in capital intensity. That holds for foreign as well as domestic producers. For that reason, high-tech FDI is no longer in a position to solve unemployment problems, as it did originally in Ireland. Even with the right proactive policies in place, adopted by a developmentalist state, it is important to see high-tech FDI as only one pillar of development and to use other pillars to address the need for creating jobs, whether that is in tourism, or call centers, or other areas.

This page intentionally left blank

Notes

1 Foreign Direct Investment in the Global Age: New Opportunities for Development? 1. Throughout this book Ireland refers to the Republic of Ireland only.

2

High-Tech FDI-Led Growth in Small Latecomers: An Analytical Framework

1. Analyzing the relationship between exports per capita and revealed comparative advantage in 1985 and 1998 in a number of countries Lall (2000, 350) found that the correlation coefficient is positive and significant for medium-technology products and high-technology products, but negative and significant for lowtechnology and resource-based products. 2. Regression analysis found that “a country where medium and high technology content exports represent 10% of GDP tend to grow between .1 and .2 percentage points more than another where, other factors being equal, there are no exports of this nature” (IADB 2001, 49). 3. To analyze the performance of export commodities by technology intensity, I used Lall’s classification (2000) to group trade positions into five categories: primary products, resource-based products, low technology products, medium technology products, and high technology manufactures. The calculations in Table 2.1 are based on the CAN data series (Comisión Económica para América Latina y el Caribe (CEPAL) 2002), which start in 1985. 4. In 2000, high-tech goods accounted for one third of Japanese exports, but for two thirds of Philippine exports. There can be little doubt that the indigenous knowledge in the high-tech exports of these two countries was of a very different nature. While production in the Philippines was at the low end of the high-tech spectrum, often standardized production outsourced by foreign corporations, production of high-tech goods in Japan was at the high-tech end of the continuum. 5. The countries included in Amsden’s analysis are Argentina, Brazil, Chile, China, India, Indonesia, Malaysia, Mexico, South Korea, Taiwan, Thailand, and Turkey. They all had pre–World War II manufacturing experience, which Amsden identifies as a key ingredient for their subsequent success. She excludes countries

208

6.

7.

8.

9.

10. 11.

12. 13.

NOTES

whose manufacturing experience was mainly based on processing a single primary product, e.g. sugar in the case of Cuba, as well as countries which are small (e.g., Singapore and Hong Kong) or have a unique historical experience (the former Soviet bloc countries) (2003, 15). The lower middle-income countries included, from smallest to largest, Swaziland, Namibia, Macedonia, Latvia, Jamaica, West Bank and Gaza, Albania, Bosnia and Herzegovina, Jordan, Paraguay, El Salvador, Honduras, Bulgaria, Bolivia, Dominican Republic, Tunisia, and Belarus. The upper middle-income countries included Mauritius, Gabon, Trinidad and Tobago, Estonia, Botswana, Oman, Paraguay, Uruguay, Lithuania, Costa Rica, Lebanon, Croatia, Slovak Republic, Hungary, and the Czech Republic. We exclude countries with less than one million people based on the assumption that they do not have the critical mass for a diversified manufacturing sector of any kind. In 2000, there were 27 developing countries with a population of less than one million people. And we exclude low-income countries, since they will—by definition— not have the minimal requisite conditions to attract high-tech FDI. None of the countries in Amsden’s analysis of latecomers had less than 10.5 million people. Taylor’s measure of openness (2000, 641) is based on the average score on a number of survey questions from the 1993 World Competitiveness Report, namely ease of hiring and firing, security, price controls, state of the justice system, nature of antitrust regulations, transparency, restrictions on FDI, state of intellectual property regime, and difficulty of cross-border ventures. Caves (1996) provides an excellent survey of the empirical findings on the impact of FDI on savings, investment, and growth in developing countries and concludes that “the statistical studies [. . .] bear strong imprints of their authors’ prior beliefs about whether a negative or positive relationship would emerge” (238). Markusen and Venables (1999), for example, develop a model where TNCs’ entry generates two effects in the host country: a competition effect for local producers of the same final good and a linkage effect where the increased demand for the output of intermediate goods’ producers makes entry into that sector more attractive for domestic producers. As the supply of inputs increases their price falls thus making final goods producers more competitive. The analysis is based on firm level data for the manufacturing sector for 1991–1997. Haskel, Pereira, and Slaughter (2002), for example, use plant level data for UK manufacturing between 1973 and 1992 and find a statistically significant positive correlation between an indigenous plant’s total factor productivity and the share of employment in that plant’s industry accounted for by foreignowned enterprises. Goerg and Greenaway (2003) provide an excellent summary of 40 econometrically-based empirical studies on FDI and productivity spillovers. This is confirmed by Barrios et al. (2002) who find that there is no effect from majority-owned foreign firms in Greece.

NOTES

209

14. Dimelis (2003) argues that nearly all the studies that find positive and significant spillovers in developing countries are based on cross-sectional studies, while those which find negative and significant results are mainly from panel firm data. 15. Since the acronym for the organization would be different in different languages, the organization decided on ISO as its acronym, for the Greek word “iso”, which means “the same.” The ISO is the successor organization of the ISA, the International Federation of the National Standardizing Associations, which had been operating since 1926. ⬍http://www.iso.ch/iso/en/aboutiso/introduction/ index.html#two⬎. 16. Navaretti, Halland, Venables (2002, 7) claim that “the average share of trade in parts and components of total EU manufacturing is approximately 15%, rising to 35% in “networking” industries, those for which trade in parts and components can be distinguished from trade in finished products.” 17. Though using a different analytical framework, Rodriguez-Clare seems to point to similar factors. He argues that the linkage effect is higher, the larger the cost of communication with the home country, the more developed the host country (because it is then assumed to produce a greater variety of sophisticated inputs), and the more intermediate inputs are used by TNCs. 18. The discussion in this section draws on Dicken (1998, chapter 11). 19. In 1971, Intel’s first commercial microprocessor contained 3,500 transistors. In the mid-1990s, a microprocessor could hold as many as 5.5 million transistors (Dicken 1998, 362). 20. Fairchild established the first offshore facility in the semiconductor industry in 1962 in Hong Kong (Dicken 1998, 373). 21. For an excellent discussion of the dramatic changes in electronics manufacturing over the last decade see for example Hyot and Lee (2001); Sturgeon (2000); Sturgeon and Lester (2003); and UNCTAD (2002). 22. During the 1990s, both India and China became much more open economies, reducing tariff levels, increasing exports and welcoming FDI with open arms. Between 1991 and 1997, the average import tariff decreased from 87% to 20% in India and from 43% to 20% in China (McKinsey 2003, 3). In addition, China became a member of the WTO in 2001. 23. When Mason and Encarnation (1994) ask, “does ownership matter?” they too are only focusing on developed countries. Analyzing the different behavior of Japanese and U.S. multinationals in Europe, they found that in the early 1990s Japanese TNCs were still much more likely to set up distribution networks in the EU and then import, compared to U.S. companies which were much more prone to set up production networks. But while Japanese and U.S. TNCs were organizing their global value chains differently, there was no difference in technological capabilities and scale. 24. Department of Enterprise, Trade, and Employment (2003, 9) refers to this study. It is not clear though whether the analysis controlled for size and productivity of the firms. And foreign firms are substantially larger than indigenous companies in Ireland, as we see in chapter 3.

210

NOTES

3

The Rise of the Celtic Tiger

1. Kevin Gardiner of Morgan Stanley is said to have coined the term “Celtic Tiger” in 1994 (Mac Sharry and White 2000, 360). Honohan and Walsh (2002) suggest that the title “Irish Hare” would be more appropriate, since tigers are not really native to the Green Isle. 2. For a detailed discussion see for example Haughton (1995) and Whitaker (1983). 3. In 1951, the external deficit was equivalent to 15% of GNP (Whitaker 1983, 8). 4. The tax remission on profits from exports started at a rate of 50% in 1956 and was increased to 100% in 1958. In 1969, the de facto zero rate of profit on manufactured exports was extended until 1990. 5. The IDA had originally been established in 1949. 6. According to O’Sullivan (1995, 365), the average nominal tariff for industrial goods fell from 25% in 1973 to 5% in 1983. Trade liberalization was accompanied by changes in the exchange rate regime. In 1979, Ireland joined the EMS leaving behind the link to the British pound at parity, which had prevailed for decades. 7. Foreign debt increased from 127 million pounds in 1972 to 5.2 billion pounds in 1981. Over this period, national debt (foreign and domestic) rose from 1.4 billion pounds to 10.2 billion pounds (Whitaker 1983, 109). 8. The calculations are based on Forfas’ Employment Survey of Irish Manufacturing Firms and International Services. 9. For a succinct discussion of the growing fiscal imbalances starting in the mid-1970s, see e.g. Honohan and Walsh (2002). 10. According to Forfas (2002, 12) U.S. firms exported 79% of Ireland’s industrial exports in 1999, followed by Irish firms with 10%, U.K. firms with 4% and German and Canadian companies with 2% each. 11. The percentages are based on Forfas’ annual employment survey of manufacturing firms. 12. These percentages are based on data from the Central Statistics Office of Ireland (CSO). CSO data are based on balance of payments information and thus not directly comparable to the data in table 3.1, which are based on UNCTAD information. For example, in 1999, foreign direct investment flows were $17.8 billion according to CSO data, and $18.6 billion according to UNCTAD data. 13. The data for 1998–2000 are based on Forfas (2002, 31) and for 2001–2002 on CSO (2003a, 9). 14. Honohan and Walsh (2002) attribute much of the decline of the investment share in the second half of the 1980s to the reduction in public investment. 15. They also estimate that labor productivity in manufacturing would be reduced by 5 percentage points (from 8.6 to 3.8%), and overall labor productivity by 2 percentage points (from 3.4 to 1.4%). They arrive at their calculation assuming that value added in the FDI-intensive sectors in Ireland was the same as value added for the EU average in those sectors. 16. Based on all countries, for which the World Bank’s World Development Indicators provided information on-line in November 2003.

NOTES

211

17. The ratios are calculated based on the World Bank’s World Development Indicators online. 18. Honohan and Walsh (2002) estimate that manufactured exports in 1999 were 25% lower, once we adjust for transfer pricing. But even if we take transfer pricing into account, the export share still rose considerably during the 1990s, given that the official data show a 66% increase in the share. 19. All the figures in this paragraph refer to the year 2000 and come from Forfas (2002, 14). 20. The real output data in table 3.2 were calculated based on the nominal output data of the Census of Industrial Production (CIP) and the wholesale price index. I deflated each NACE category by the wholesale price index, which was the closest fit for the respective category. I am indebted to Marcus Breathnach at Forfas for providing me with the nominal CIP data and details on the wholesale price index. The employment data in table 3.2 are not strictly comparable with those in graph 3.1, since they come from different sources. The employment data in graph 3.1 are based on Forfas’ Employment Survey of Irish Manufacturing and International Services (ESIM), which covers all enterprises assisted by the main development agencies (IDA, Enterprise Ireland, Shannon Development, and Údará na Gaeltachta). It covers a high proportion of total employment and provides a useful series because it goes back to 1972. The employment data in table 3.2, in contrast, come from the annual Census of Industrial Production (CIP), which covers all manufacturing enterprises with three and more employees. In 1991, data reporting in the CIP was changed to NACE Rev. 1. As a result, a consistent time series based on the same NACE classification only goes back to 1991. That is the reason why the tables in this book based on CIP data have 1991 as the starting year. The difference between the two sources, ESIM and CIP, can be significant at times. For example, in 1999, total employment in NACE categories 30–33 was 64,462 according to the CIP, but only 57,675 according to the ESIM. 21. I calculated the averages based on Hanlon (1998) and Ireland’s Department of Finance ⬍www.revenue.ie⬎. I am grateful to Frances Ruane for providing me with the Hanlon paper. 22. Based on data from the Swedish Ministry of Finance ⬍http://www.finans. regeringen.se/inenglish/swedishmoney/taxes.htm#5⬎ and the U.S. Department of the Treasury and the Office of Management and Budget ⬍http://w3. access.gpo.gov/usbudget/fy2004/sheets/b80.xls⬎. 23. Based on data of the United Kingdom Internal Revenue, ⬍http://www. inlandrevenue.gov.uk/about/annual_reps.htm⬎. 24. According to Keating (2000, 12) the 34 largest TNCs contributed 16.4% to Ireland’s GDP in 1998, and 5.7% to GNP. But those figures are obviously limited and only take into account the direct effect, and not the multiplier effects. 25. Kirkland (1989) cites several studies predicting substantially faster growth in the European market compared to the U.S. market. 26. The U.S. investment position in Portugal is not disclosed for 1989.

212

NOTES

27. The data in table 3.8 contradict Lipsey’s assertion back in 1990 that in spite of all the hype around “Europe 1992”, there was little evidence up to that point that the pace of U.S. FDI in the EC had picked up. Lipsey (1990) based his claim on annual data about U.S. capital flows, whereas the data in table 3.10 for 1982 and 1989 are based on the benchmark survey the U.S. Department of Commerce undertakes periodically. 28. Hufbauer (1990, 18) cites the NEC corporation as an example of a company that after the ruling stopped buying U.S.-made computer chips for its assembly of computer printers in Great Britain. 29. Based on an interview with Frank Turpin, Academic Relations Manager at Intel, on August 2nd, 2002 in Dublin. 30. In 2000, Intel’s sales were 29.7 billion in a semiconductor industry with total sales of $204.4 billion. After the collapse of the technology market in 2001, Intel’s sales had fallen to $22.7 billion, but sales in the industry as a whole had dropped to $139 billion (UNCTAD 2002, Table V.2, 126). 31. Interview with the Academic Relations Manager, Intel, Dublin, August 2002. 32. Interview with the Public Affairs Manager, HP, Dublin, July 2002. 33. The four sources of financing are the European Regional Development Fund, the European Social Fund, the European Agriculture and Guarantee Fund, and the Financial Instrument for Fisheries Guidance. 34. The Treaty of Maastricht, which established the European Union, aimed at bringing about further economic and social cohesion within the EU. It laid out the steps for the creation of a single European currency and of the European Monetary System. 35. The EU expanded its membership from 12 to 15 during this period, with Austria, Finland, and Sweden joining in January 1995. Excluding their data from the 2000 EU figures, the share of the original 12 members had risen to 44% in 2000. 36. Jacobson and Andreosso (1990, 314–315) cite three major surveys of chief executives in 1989; one by the Bank of Boston ranking Ireland very positively, the other two by Fortune Magazine and Business International ranking Ireland very low as a potential investment site. 37. According to Mac Sharry and White (2000, 214), when the IDA negotiated with Intel, it hired a specialist-consulting firm to track down expatriate Irish engineers in the semiconductor industry. “Within five weeks, over 300 Irish engineers, mainly in the US, had been identified and individually contacted; each of them had between three and seven years’ experience in the production of semiconductors. The formal report handed to Intel had the positive finding that over 80% of the expatriate engineers would return to Ireland if given a good career opportunity with a quality company. This report was crucial in satisfying Intel that Ireland could satisfactorily host an advanced microchip plant.” 38. Ireland had the longest workweek among all European OECD members. In 1988, the average weekly hours worked in Ireland were 43.6, compared to 43 in Portugal, for example, and 39.5 in Spain and Belgium (OECD Observer 1990, 13).

NOTES

213

39. In 1988, Ireland had the second youngest population among OECD members. The age group of under-15-year-olds constituted 28.2% of the population, second only to Turkey with 36.6% (OECD Observer 1990, 7). 40. A substantial part of the increase stems from a rise in part-time employment, which the labor market surveys suggest was mostly voluntary. 41. In 1975, the Irish Republican Army (IRA) kidnapped the managing director of the largest foreign employer in Ireland; in 1976 the British ambassador to Ireland was assassinated, and so was Lord Mountbattan in 1979 (Mac Sharry and White 2000, 199). 42. For an excellent summary of IDA policy see Ruane and Goerg (1997). As an example of the IDA’s strategic, persistent, and patient pursuit of high-tech companies they point to the fact that the IDA had apparently targeted Hewlett Packard, Intel, and IBM for over a decade with regular presentations and visits making sure that Ireland was on their radar screen, when the TNCs finally decided on major investments in Europe. 43. Interview with the author, July 29, 2002. 44. According to White (2000, 181), the reimbursement rate for training grants for new or expanded foreign industry was 50% after Ireland entered the EEC. It was raised to 55% in 1978, to 65% in 1989 and to 75% in 1994. “In the five years ending in 1998, the average annual cash refund to the IDA for training and employment in foreign companies was 5 million pounds. . . . The government, too, was able to secure substantial recoupment for various elements of the foreign-investment programme, such as new industry grants and advance factories.” 45. See, for example, Tomaney (1995). 46. Interview in August 2002. 47. The model currently used by the IDA assumes that 65–80% of the benefits would have occurred anyway, without IDA support. 48. Based on different surveys of the cost of doing business in EU countries, the O’Driscoll Report claims that in 2004, “Ireland is ranked the third most expensive of nine countries for industrial electricity costs, third most expensive out of 10 countries for landfill costs, and fourth most expensive of 16 countries for insurance premia” (Enterprise Strategy Group 2004, 21). 49. That does not include funding under the Common Agricultural Policy. 50. In the early 2000s, work permits were issued to non-EU immigrants at an annual rate of about 2% of the labor force (Walsh 2003, 9).

4 The Development of Knowledge-Based Assets in Ireland 1. The CIP data show that the share of material inputs (which includes industrial services) imported by foreign-owned units (FOUs) fell throughout the 1990s. For FOUs in the manufacturing sector as a whole, the share declined from 68.5% in 1991 to 52.2% in 2000. In the electronics sectors (NACE 30–33) the share of imported materials dropped from 78.3% in 1991 to 51.7% in 2000.

214

2.

3. 4. 5. 6.

7. 8. 9.

10.

11. 12.

NOTES

This trend is so different from the one suggested by the ABSEI data, that the CSO reevaluated the interpretation of the firms’ answers to its questionnaire. The main reason for the CSO’s larger domestic sourcing data seems to have been that any blank company response to the requested estimate of the imported share of materials was interpreted to mean that all materials were sourced domestically. For 2001, the CSO reports the share of imported materials in the electronics sector at 83.4%, which is in line with the ABSEI findings. Forfas conducted the IEES from 1983 until 1999. The IEES tracked the sourcing behavior for a sample of firms with more than 30 employees starting in 1983 and with more than 20 employees as of 1990. In 1999, the IEES was amalgamated with Enterprise Ireland’s Annual Business Survey. ABSEI covers the client base of Enterprise Ireland, IDA Ireland, Shannon Development, and Údará na Gaeltachta, and includes businesses with more than 10 employees. The ABSEI time series data in this section are based on ABSEI 2000–2002 and recalculated grossed-up estimates of the former IEES. They were provided by Marcus Breathnach at Forfas. These data are based on Breathnach and Kelly (1999). Total sales of Irish-owned companies in software manufacturing increased from €141.4 million in 1995 to €1.2 billion in 2000 (Forfas 2003a, p.27). ⬍http://media.corporate-ir.net/media_files/nsd/iona/IONA_Fact.pdf⬎. 35% of the sub-suppliers had applied for patents during the period considered, compared to 21% for non-sub-suppliers. And 33% of sub-suppliers indicated that they were developing products with others, as opposed to 15% for nonsub-suppliers. Furthermore, a much larger percentage of sub-suppliers were engaged in industrial design, 44.2% compared to 19.6% for non-sub-suppliers. And a much larger percentage listed a reduction in labor costs as an important motivator for innovation, 64% in contrast to 43% for non-sub-suppliers. The latter finding suggests that sub-suppliers are experiencing stronger pressures to reduce costs. Of the 61 companies, 52.5% were in engineering and chemicals, 37.7% in electrical and electronic engineering, and 9.8% in other manufacturing. Because of the distortion of output data due to transfer pricing, we use employment data as the main indicator of the industry’s growth. The Industrial Development Act of 1993 created two agencies: the IDA (now renamed as the Industrial Development Agency) in charge of attracting FDI and Forbait in charge of promoting indigenous industry. A year later Forbait was merged with the Trade Association and became Enterprise Ireland. Since the mid-1970s a linkage unit had existed with the IDA. But it was never given the necessary attention and resources that would have allowed it to be effective on a significant scale. The details in this paragraph are based on an interview with Thomas Kennedy at Enterprise Ireland in August 2002. The results refer to the long-term effect, with the estimation taking into account grants over the last four years. Long-term employment grants and other grants are significant as well with respect to employment generation, but not with respect to longevity.

NOTES

215

13. Other obstacles to linkage formation highlighted in the Telesis Report were insufficient technological capability on the part of potential indigenous producers, distrust and lack of knowledge of potential suppliers on the part of TNCs. 14. See Arndt (1988) and Kaldor (1972) for an excellent discussion of the allocative and creative functions of the market and the possibilities for market failures in the two different senses. 15. Both McAleese and McDonald (1978) and O’Farrell and O’Loughlin (1981) base their estimates on a cross section of firm-level data, for 1974 and 1976, respectively. Goerg and Ruane (2000), in contrast, use firm level panel data for 1982–1995. 16. See ⬍http://www.ida.ie/industry/medical_industry.asp⬎. 17. For an excellent description of FDA regulations of the medical device sector, see for example Lincoln (2003). Most medical device regulations are under Title 21 of the Code of Federal Regulations (CFR), Parts 800–1299. 18. Huckman (2003) provides a vivid description of the difficulties MedSource Technologies faced in becoming a successful contract manufacturer for the medical device sector, which was the stated goal of the company when it was founded in 1999. 19. In 1975, 21.7% of the workers in Mexican maquiladoras were men. By 1998, the share of male employment had increased to 43.4% (Wilson 2002, 4). Wilson also suggests other reasons for the masculinization of the maquila work force, namely demographics and the appearance of docile, nimble-fingered males. 20. Women applying for a job at this company are asked during the job interview whether they sew or knit. 21. The employment data are based on the CSO’s enterprise survey for the CIP, where all activities of an enterprise are attributed to that NACE category which accounts for the majority of the enterprise’s generation of value. 22. For more details see the IDA webpage at: ⬍http://www.ida.ie/industry/ teleservices_companies.asp?industry⫽teleservices.⬎. 23. That corresponds to previous threshold values of 100,000 Irish pounds. 24. Based on panel level data for 1992 and 1997 from the Irish Economy Expenditure Survey, Barrios et al. (2002) find no statistically significant impact of TNC presence on labor productivity in the Irish manufacturing sector. Using CIP data, Ruane and Udur generate the same conclusions using 2- and 4-digit industry level data for 1991 and 1997 in Ruane and Udur (2000), and plant level data for 1991 and 1998 in Ruane and Udur (2002). 25. Like Barrios et al. (2002), Kearns uses panel data from the IEES, but for 1984–1998. 26. For example, a one percentage point increase in the share of TNCs in the total number of firms in a sector was associated with an increase in the net entry rate of indigenous firms of .2 percentage points. 27. Ruane and Udur (2002) use firm-level data that is aggregated to the 2-, 3-, and 4-digit industrial level. The statistically significant result holds at the 2- and 4digit level, but not at the 3-digit level. 28. The significant coefficient may not tell us anything about spillovers, if there is a high correlation between minority foreign ownership and firms with

216

29. 30.

31. 32.

33. 34. 35.

36.

NOTES

absorptive capacity. In that case, minority- and majority-owned TNCs may simply coexist in the sectors with higher productivity levels and growth. Calculated based on CIP data. Net output is defined as gross output minus material inputs. Net output is equal to value added in the industrial sector. But it still includes nonindustrial services, which are not included in material inputs. Given the lag between application and approval, it is not clear how many of the patents granted relate to the patents applied for in 2001. The Irish Council for Science, Technology, and Innovation, which was established in 1997, released reports on eight different areas in April 1999: on chemicals and pharmaceuticals, information and communication technologies, manufacturing and processes, life sciences, natural resources, energy, transport and logistics, and construction and infrastructure. The reports are available at: ⬍http://www.forfas.ie/icsti/statements/tforesight/intro.htm⬎. For example, biotechnology companies are joining with medical device companies in developing new hybrid technologies like drug-coated stents. R&D expenditures as a percentage of GNP are 1% in Ireland, 1.2% for the EU average, and 1.6% for the OECD average. Glass left his position in September 2004 to become chair of the board of the new National Institute for Networking and Electronic Research in Cork, Ireland. For details see ⬍www.sfi.ie⬎.

5 From Coffee to Computers: High-Tech FDI in Costa Rica 1. While the social democracies in the European countries have historically been grounded in working class and union support, social democracy “Tico style” developed in the 1940s on an antilabor platform in the private sector (though pro-white-collar public-employees unions, which became a key source of support for the government). One indication of Costa Rica’s achievement on the social front is the fact that it ranks more highly on the UNDP’s human development index than on a simple GDP per capita index. In 2002, Costa Rica ranked 45 on the Human Development Index. The only Latin American countries ranked above Costa Rica were Argentina (34) and Chile (43), both of which have a higher GDP per capita (UNDP 2004, 139). 2. Nonetheless, the coffee-producing sector was hardly homogeneous. Gudmundson (1995) provides a lucid analysis of the dynamics of class differentiation in the coffee sector between 1850 and 1950. 3. For an excellent discussion of these issues see for example Wilson (1998) and the compilation of essays in Edelman and Kenen (1989). 4. Wilson (1998, 93) argues that the dominant agro interests in Costa Rica were dragging their feet on the CACM and that the country finally joined in November 1963, after the U.S. government threatened to withhold Alliance for Progress benefits.

NOTES

217

5. Manufacturing value added in Costa Rica grew at an average annual rate of 10.6% between 1960 and 1973, and at a rate of 7.2% between 1973 and 1979, as civil war was raging in Nicaragua, El Salvador, and Guatemala (Buitelaar, Padilla Pérez, Urrutia-Alvarez 2000, 12). 6. According to Wilson (1998, 12) per capita public expenditure on health increased from $2.20 in 1970 to $8.10 in 1979, and on education from $20.60 to $35.10, respectively (all in prices of 1970). 7. Figueres headed the Costa Rican government three times. In 1948–1949, he was head of an 18-months junta, and in 1953–1958 and 1970–1974, he ran the government of the National Liberation Party (Partido Liberación Nacional, PLN), which had its base primarily among non-elite coffee growers and public sector employees. 8. The Carazo government (1978–1982) of the PUSC (Partido Unidad Social Cristiana—Social Christian Unity Party), the other major party in Costa Rica during the second half of the twentieth century, had run on a platform to cut funding for CODESA and reduce the size of the state. 9. The essays in Ulate Quiros (2000) provide an excellent description and analysis of the reforms. They constitute the Costa Rican part of the CEPALsponsored multi-country assessment of the Washington Consensus in Latin America. For a summary of the CEPAL study see Stallings and Peres (2000). IADB (1997) provides another excellent assessment of the results of the Washington Consensus in Latin America. 10. Other areas, like most of health care provision, have not been privatized either, but continued public ownership in those areas has been much less controversial. 11. The figure for 1985 is from the IADB (1997, Figure 17, 43) and the data for 1995 and 2004 from the Costa Rican Ministry of Foreign Trade, ⬍http:// www.comex.go.cr/estadísticas/aranceles/arancel_promedio.pdf⬎. 12. In 1988, for example, the cost of the CAT amounted to 3.5% of government expenditures (Ulate Quiros 2000, 32). 13. Clark (1997) provides an interesting account of the coincidence of factors, which allowed for the grand-scale influence of U.S.-AID in Costa Rica. He argues that widely discredited state institutions (e.g. the Center for the Promotion of Exports and Investment, CENPRO—Centro de Promociones de Exportaciones e Inversiones), disenchantment with CODESA’s overreach, the collapse of the CACM and the inability of some sectors to obtain credit from the state banking system had led to the search for a new, anti-ISI, proliberalization strategy by a segment of the economic and political elite. The split within the PLN over structural adjustment provided a vacuum in policymaking, and the foreign debt crisis made the promise of US-funds very attractive and allowed for an emerging alliance between disenchanted Costa Rican producers and US-AID. For a detailed discussion of the political economy of the development of the new alliances under the neoliberal model see SánchezAncochea (2004). 14. In comparison, total U.S. assistance to Costa Rica over the period 1946–1981 amounted to $288 million. Edelman and Kenen (1989, 189) point out that in the first half of the Eighties . . .“Costa Rica became the second largest recipient

218

15.

16.

17.

18. 19. 20.

21.

22.

23.

24.

25.

26. 27.

NOTES

of U.S. aid in Latin America, after El Salvador, and at one point was the second highest per capita recipient of U.S. assistance worldwide, after Israel.” According to the World Development Indicators, foreign aid as a share of GNI constituted on average 1.9% of annual GNI during the 1960s, and 1.4% during the 1970s. Edelman and Kenen (1989, 189) also point to the willingness of the IMF and other international organizations to put up with Costa Rica’s noncompliance with IMF targets, due to the country’s strategic political importance. The Costa Rican government had agreed with the IMF to reduce public sector employment by 20%. The voluntary labor mobility program, which replaced the forced public employment reduction program, offered public sector workers a number of benefits, if they left the public sector, including a one-time payment of three-months pay (Sánchez-Ancochea 2004, 198). As a result, the bargaining power of public sector employees declined, as did the rate of unionization, which fell from 15 to 10% in the course of the 1990s (SánchezAncochea 2004, 199). CBI is officially known as the Caribbean Basin Economic Recovery Act, signed into law on August 5, 1983. See Mortimore and Zamora (1999) for an excellent development of this argument. In 1993, the textile and apparel exporters mostly operated under one of three special export regimes. Of the 138 registered exporters (out of a total of 700 firms in the sector), 44.2% exported under the temporary admission regime, 21.7% were located in FZs, and 18.1% operated under the Export Contract Regime (Mortimore and Zamora 1999, 67). Between 1980 and 1994, the food sector and wood products declined in importance, while paper products, chemicals, and nonmetallic mineral products slowly increased their shares in manufacturing output (Mortimore and Zamora 1999, 25). According to Mortimore and Zamora (1999, 27), the manufacturing sector attracted 35% of FDI inflows during 1970–1974, 59% during 1975–1979, 12% during 1980–1984, 36% during 1985–1989, and 28% during 1990–1993. According to CINDE foreign companies invested $650 million in the electronics sector in Costa Rica between 1997 and 2002 ⬍http://www.cinde.or.cr/ cinde/home.nsf/pages/electronics⬎ accessed July 2004. To the extent that FDI went into acquisitions rather than greenfield investment, its economic impact on investment is, of course, somewhat overstated by the ratio of FDI inflow to GCF. Cordero (2000) raises the same concern, even though his econometric work shows a statistically significant negative relationship between FDI and national investment. Cordero (2000) provides a detailed discussion of the decline in the public investment share under the neoliberal model. Spar (1998, 1) points out that Intel’s annual global sales are three times as large as Costa Rica’s GNP.

NOTES

219

28. When Intel’s exports from Costa Rica dropped by 35% in 2000, the manufacturing sector declined by 2.9%; and when it dropped by another 44.9% in 2001, manufacturing output fell by 8.4% (OECD and IADB 2003, 27). 29. The data were provided by PROCOMER. 30. The calculations are based on data from the World Bank’s World Development Indicators, online. 31. Data for net foreign investment inflows and profit repatriation are only available for all TNCs in the aggregate and not for TNCs in manufacturing separately. Since foreign investment in industry accounted for 65% of all FDI during 1997–2003, we use that percentage to approximate manufacturing TNCs’ net foreign investment flows and profit repatriation during 1991–2003. We also assume that the trade data for Free Zones and the Temporary Admissions Regime provide a good proxy for the trade of TNCs in manufacturing. 32. For companies in less developed areas, the length of time is 12 and 6 years respectively. 33. The cost-benefit analysis only considers the short-term implications of the Free Zones and excludes considerations of benefits derived from positive externalities via linkages, technology transfer, and training. 34. The ratio of government revenue to GDP was 13.3% on average in the 1970s and 14.7% in the 1980s. 35. In September 1999, the average wage at Intel was $615, while the average wage for the manufacturing sector was $406 (Rodríguez-Clare 2001, 12). 36. Disaggregated data for the manufacturing sector are only provided for companies operating outside the special investment regimes. Industrial statistics provide disaggregated information for national companies with more than 20 employees, but only aggregate industrial information for small enterprises and corporations operating under Special Export Regimes. The data were provided by the Institute for Economic Research at the University of Costa Rica (Instituto de Investigaciones en Ciencias Económicas, IICE). The IICE data is based on quarterly surveys, whereas the information from the Central Bank, the other source for industrial data, is based on annual values. In contrast to the Central Bank, the IICE survey also provides information on employment, which is the main reason I am using IICE data here. The trends in production data for the two sources are very similar. The IICE changed its survey and sample at the end of 1996, which means that a consistent annual data series starts only in 1997. Data on FZ exports provide more disaggregated information about TNCs in the manufacturing sector. The data come from the Costa Rican Institute for Export Promotion (Promotora de Comércio Exterior), which collects information about corporations in the Free Zones. Since PROCOMER was founded only in 1996, its data series also goes back to 1997 only. Since corporations in the FZs export 100% of their production, exports are the same as total sales, and thus a good proxy for production.

220

NOTES

37. Between 1995 and 2003, the price per sack of coffee fell from $149.4 to $72.6, while the price per ton of bananas dropped from $680 to $553 (BCCR, ⬍http://websiec.bccr.fi.cr⬎). 38. In 2000, small companies (pequenˇa industria) accounted for 8.3% of manufacturing output outside the special investment regimes. Its production was slightly higher than in 1997, but the same as it had been in 1991 (calculated on the basis of data provided by the Central Bank of Costa Rica). 39. In the non-FZ sector, employment declined in all branches except for nonmetallic minerals and basic metals. 40. Between 1997 and 2003, imports outside the special investment regimes increased from $3.8 billion to $5.2 billion. (⬍http://www.comex.go.cr/ estadistícas/importaciones/régimen.pdf⬎). Unfortunately, there are no separate data on imports of manufactured goods, at the aggregate or disaggregated level, which would make it possible to analyze import penetration at the ISIC level. 41. There are a number of accounts of Intel’s decision to establish a production facility in Costa Rica. Spar (1998) provides the most detailed and extensive study, which provides the basis for most subsequent studies, e.g. Larrain, Lopez-Calva, Rodríguez-Clare 2000; Egloff 2001; Gonzalez 2001; Ketelhoehn and Porter 2002. 42. Spar (1998) points out that Intel is establishing a new production facility somewhere in the world about every 18 months. 43. Between 1987 and 1997, the number of foreign tourists in Costa Rica increased at an annual rate of 8% (Ulate Quiros 2000, 39). In one of the small high-tech companies I visited in Costa Rica, the general manager explained that they had explored Costa Rica as a potential site when their lawyer brought it to their attention after having spent his vacation in the country. 44. In 1977, the Figueres Administration established the National Loan Commission for Education (Comisión Nacional de Prestamos para Educación). 45. According to IADB (2003b, 14) “16 percent of students repeat a grade between first and seventh grade, and only 33 percent of 20-year-olds have made it through high-school.” 46. The analysis is based on 187 TNCs in the Caribbean Basin using data for the period 1984–1987. 47. In a presentation to a U.S. TNC in the computer industry in late 2001, CINDE pointed to relatively low wages in Costa Rica, with the average hourly compensation for a technician at $3.62 in Costa Rica, $6.54 in Mexico, and $25 in Nashville. For an assembler, the average compensation was given as $1.73, $2.48, and $12.62, respectively (CINDE internal memo). 48. In 2003, the share of Costa Rican exports going to the United States was 46.6% for all exports and 52.3% for Zona Franca exports. 49. While only 27% of the road system was in bad shape in 1993, the percentage had increased to 48% by 1998, a result of the decline in public investment during the period (Cordero 2000, 213). 50. Interview with Bertus Meins, head of the IADB office in Costa Rica, July 2003.

NOTES

221

51. The Office of Evaluation and Oversight writes in 2003, “the net negative financial flow has ranged from 0.4 percent to 0.6 percent of GDP in recent years and interest and fees paid to the [Inter American Development] Bank have absorbed between 24 percent and 70 percent of the primary fiscal surplus since 1996” (IADB 2003a, vi). 52. Sánchez-Ancochea (2004) provides a wonderfully insightful and persuasive presentation of this argument. 53. Sánchez-Ancochea (2004, 224) analyzes how the Rodriguez Administration (1998–2002) embarked on a process of “concertación nacional,” in an attempt to bring capital and labor together. He concludes that most agreements were not ratified, because the process had become too broad, the opposition party was not included sufficiently, and there was not enough technical support and involvement of public institutions. He argues, “one of the most severe failures of ‘concertación nacional’ was its inability to reach an agreement on what to do with the ICE and other public institutions.” 54. The minimum investment requirements for operating under the FZ regime are $150,000 if the company is located within an FZ area, and $2 million, if it locates outside of such an area. 55. Under WTO rules, any export subsidies in countries whose GDP per capita exceeds $1,000 have to be eliminated by 2007. That includes tax-free treatment in free zones. The deadline for phasing out preferential treatment for exports had originally been set for 2003; but it was later extended to the end of 2007. 56. PROCOMER’s Board of Directors is headed by the Minister of Foreign Trade. It also includes three members of the Executive, the heads of the Chamber of Commerce, the Chamber of Industries, and the Chamber of Exporters, as well as representatives from small and medium-sized exporting firms. ⬍http://www. procomer.com/index.cfm?queHacer⫽showQuienesSomos⬎ 57. PROCOMER is the outcome of the fusion of the previous export promotion organization (CENPRO) and the Free Zone Corporation (Corporación de Zonas Francas). 58. Based on an interview with the director of export development and logistics at PROCOMER, May 2002. 59. Clark (1997) provides a fascinating account of the dynamics behind the establishment of CINDE. Clark recounts that Daniel Chaij, who had become the director of the AID office in Costa Rica in 1982, proposed the idea of an AID-funded export promotion center in his twice-weekly meetings with local business and political leaders. The handpicked members of the first CINDE Board would go on to become government officials in subsequent years. Foreign consultants were brought in from the get-go: first the architects of Operation Bootstrap in Puerto Rico, Teodoro Moscoso and Leo Suslow, who recommended that CINDE focus on manufacturing exports and FDI and that a new ministry of foreign trade should be created; and in 1985 an official from the IDA of Ireland and from the U.S. Consulting Firm Carana, who oversaw the establishment of CINDE’s offices abroad, were brought in. Clark (1997, 89) claims that AID had to give permission for any CINDE expenditures above $20,000.

222

NOTES

60. The study concluded that Costa Rica had a comparative advantage in electronics products that required small runs and skilled labor for set-ups and testing. 61. The details in this paragraph are based on the description by Spar (1998). 62. Based on a detailed assessment of the Costa Rican clothing industry, Mortimore and Zamora (1999, 70) come to the conclusion that while the competitive situation of the clothing industry improved a lot between 1980 and 1995, it is not built on a solid basis, because of the financial incentives and the lack of the development of an integrated strategy in that industry. So with rising wage costs, the survival of the industry is at stake. While wage pressures moved production in the industry to higher value-added products (see Larudee, Cordero-Peña and Arroyo-Abad 2001), the abolition of the quota system will pose a major challenge for clothing production in Costa Rica.

6 The Tico Tiger That Hasn’t . . . 1. There are no publicly available time series data on the value of inputs of Free Zone companies. We only have data on imports, which are not a good proxy for inputs since they also include the value of imported capital goods. Both PROCOMER and CINDE always refer to domestic input sourcing of about 7%. 2. Interview with Baxter’s general manager, June 2003. 3. Interview with company’s CEO in San José, May 2002. 4. The common Spanish phrase used in this context was “tienen otra cultura.” 5. Mortimore and Zamora (1999, 34) describe how the Arias Administration (1986–2000) tried to put an industrial reconversion program in place. It was not geared toward linkage development specifically, but more broadly toward increasing the competitiveness of the national manufacturing sector, as tariffs were reduced and import competition increased. Workgroups in three sectors were to identify problems and propose solutions. But there never were any proposals for concrete actions, and the funds were eventually used for other purposes. 6. Programa MIL stood for Mejoramiento de la Industria Local (program for the improvement of local industry), PROFOVE for Proyecto de Fomento al Vínculo de Empresas de Régimen de Zona Franca y Empresas Nacionales (linkage formation project between FZ companies and national companies), and UATI for Unidad de Asistencia Industrial (unit for industrial assistance). 7. The agreement between the IADB and the Costa Rican government had originally been signed in December of 1999, but because of organizational and administrative difficulties the program did not effectively get underway until late 2001. The pilot program was to run until December 2002, with a budget of $1.5 million, 60% of which came from the Multilateral Investment Fund of the IADB and the remaining 40% from Costa Rican counter funds.

NOTES

223

8. Esquivel et al. (2000, 5) argue, “in Taiwan, under government guidance, personnel from firms in the zones were placed at potential suppliers’ factories to offer advice in production methods and quality control.” 9. For a more detailed discussion see for example Paus (2004) and Reinhardt and Peres (2000). 10. Based on interview with Danilo Arias, spokesperson for Intel-Costa Rica, June 2003. 11. See e.g., USITC (2004).

This page intentionally left blank

References

Aitken, Brian J. and Ann Harrison. 1999. Do domestic firms benefit from direct foreign investment? Evidence from Venezuela. American Economic Review 89 (3): 605–618. Altshuler, Rosanne, Harry Grubert, and T. Scott Newlon. 1998. Has U.S. investment abroad become more sensitive to tax rates? National Bureau of Economic Research, NBER Working Paper No. 6383. Amsden, Alice 1989. Asia’s next giant: South Korea and late industrialization. New York: Oxford University Press. ———. 2003. The rise of the “rest.” Oxford and New York: Oxford University Press. Amsden, Alice and Chu Wan-wen. 2003. Beyond late development. Taiwan’s upgrading policies. Cambridge and London: MIT Press. Arndt, Heinz. 1988. Market failure and underdevelopment. World Development 16 (2): 219–229. Banco Central de Costa Rica. Indicadores Económicos. Sector Externo. ⬍http:// websiec.bccr.fi.cr/indicadores/sector.web?sector⫽1⬎. ———. 2003. Informe trimestral sobre los flujos de inversión extranjera directa en Costa Rica, 1997–2003. San José, Grupo Interinstitucional de Inversión Extranjera Directa, Banco Central de Costa Rica, September. Barrios, Salvador, Sophia Dimelis, Helen Louri, and Eric Strobl. 2002. Efficiency spillovers from foreign direct investment in the EU periphery: A comparative study of Greece, Ireland and Spain. Madrid: Fundación de Estudios de Economía Aplicada, Madrid, Working Paper No. 2002–02. Barry, Frank. 2000. Convergence is not automatic: Lessons from Ireland for Central and Eastern Europe. The World Economy 23 (10): 1379–1394. Barry, Frank, John Bradley, and Aoife Hannan. 1999. “The European Dimension: The Single Market and the Structural Funds,” in Understanding Ireland’s economic growth, ed. Frank Barry, 99–118. Houndmills et al.: Palgrave. ———. 2001. The single market: The structural funds and Ireland’s recent economic growth. May, mimeo. Barry, Frank, John Bradley, and Eoin O’Malley. 1999. “Indigenous and Foreign Industry: Characteristics and Performance,” in Understanding Ireland’s economic growth, ed. Frank Barry, 45–74. Houndmills et al.: Palgrave. Barry, Frank, Holger Goerg, and Andrew McDowell. 2001. Outward FDI and the investment development path of a late-industrializing economy: Evidence from Ireland. Centre for Economic Research, University College Dublin, Working Paper 01/08, April.

226

REFERENCES

Barry, Frank, Holger Goerg, and Eric Strobl. 2001. Foreign direct investment, agglomerations and demonstration effects: An empirical investigation. Centre for Economic Research, University College Dublin, Working Paper 01/04, March. Bellak, Christian, and John Cantwell. 1998. “Globalization Tendencies Relevant for Latecomers. Some Conceptual Issues,” in Latecomers in the global economy, ed. Michael Storper, Stavros B. Thomadakis, and Lena J. Tsipouri, 40–75. London and New York: Routledge. Bernhard, Andrew and Fredric Sjoholm. 2003. Foreign owners and plant survival. National Bureau of Economic Research, NBER Working Paper No. 10039, October. Bitar, Javier. 2002. Programa de desarrollo de proveedores para empresas multinacionales de alta tecnología. Evaluación de medio termino y metodologías de monitoreo, San José. October 23, mimeo. Blomstroem, Magnus and Fredrik Sjoeholm. 1999. Technology transfer and spillovers: Does local participation with multinationals matter. European Economic Review 43: 915–923. Blomstroem, Magnus, Ari Kokko, and Steven Globerman. 2001. “The Determinants of Host Country Spillovers from Foreign Direct Investment: A Review and Synthesis of the Literature,” in Inward investment, technological change and growth, ed. Nigel Pain, 34–65. Houndmills et al.: Palgrave. Bradley, John. 2004. Experiences in small European countries and regions. Committing to growth. Dublin: Economic and Social Research Institute, June, mimeo. Breathnach, Marcus, and Daragh Kelly. 1999. Multinationals, subcontracting linkages and the innovative performance of indigenous firms: Some Irish evidence. Paper presented at the European Network on Industrial Policy International Conference, Dublin, December 9–10. Brennan, Ciaran. “Worst Fears of Workers Realized as Rumours Prove Well Founded,” Irish Times, May 2, 2002. ———. “370 Jobs Cut in Offaly as Flextronics Concentrates on Asia,” Irish Times, March 7, 2003. Buitelaar, Rudolf, and Ramón Padilla Pérez. 2000. Maquila, economic reform and corporate strategy. World Development 28 (9): 1627–1642. Buitelaar, Rudolf, Ramón Padilla Pérez, and Ruth Urrutia-Álvarez. 2000. Costa Rica: Sistema nacional de innovación. No. 82, July. Santiago de Chile: CEPAL, Serie Desarrollo Productivo. CEPAL. 2002. Trade CAN, CD-ROM. Santiago de Chile: CEPAL. Caves, Richard E. 1976. Multinational enterprise and economic analysis. London: Cambridge University Press. ———. 1996. “Multinationals in Developing Countries,” in Multinational enterprise and economic analysis, 214–242. Cambridge: Cambridge University Press. CINDE (Coalición Costarricense de Iniciatívas para el Desarrollo). Medical Devices. ⬍http://www.cinde.or.cr/cinde/home.nsf/pages/medical⬎. ———, INCAE, and Programa Impulso. 2002. Caracterización inicial del cluster de dispositivos médicos en Costa Rica. San José, mimeo.

REFERENCES

227

CINDE. Baxter to expand its operations in Costa Rica, June 7. ⬍http:// www.cinde.or.cr/cinde/home.nsf/pages/20040607⬎ retrieved September 2004. Clark, Mary. 1997. Transnational alliances and development policy in Latin America: Nontraditional export promotion in Costa Rica. Latin American Research Review 32 (2): 71–97. Clinch, Peter, Frank Convery, and Brendan Walsh. 2002. After the Celtic Tiger: Challenges ahead. Dublin: The O’Brien Press. Cordero, José Antonio. 2000. “El Crecimiento Económico y la Inversión: El Caso de Costa Rica,” in Empleo, crecimiento y equidad. los retos de las reformas económicas de finales del siglo XX en Costa Rica, ed. Anabelle Ulate Quiros, 199–282. San José: Editorial de la Universidad de Costa Rica. Costa Rica Digest. Costa Rica’s software sector exports growth. December 2004. ⬍http://www.talkabouttravelling.com/group/alt.travel/messages/52376.html⬎ retrieved December 2004. Costa Rica Provee. 2004. Informe de progreso, San José: CRP, mimeo. CSO (Central Statistics Office). 1994. Census of industrial production 1991. Dublin. ———. 2003a. Balance of international payments, Qtr 2, September 30. Dublin. ———. 2003b. Census of industrial production 2000. Dublin. ———. 2003c. Industrial employment, September 30. Dublin. Culliton. 1992. A time for change: Industrial policy for the 1990s. Report of the Industrial Policy Review Group. Dublin: Stationery Office. Department of Enterprise, Trade, and Employment. 2003. Review of industrial performance and policy 2003. Dublin: Stationery Office. Dicken, Peter. 1998. Global shift: Transforming the world economy. Third edition. New York and London: Guilford Press. Dimelis, Sophia. 2003. Firm productivity growth: FDI, financial and market structure effects. Paper presented at the annual meeting of the Eastern Economic Association, New York City, February 21–23. Dimelis, Sophia and Helen Louri. 2002. Foreign ownership and production efficiency: A quantile regression analysis. Oxford Economic Papers 54: 449–469. Dunning, John H. 1977. “Trade, Location of Economic Activity, and the MNE: A Search for an Ecclectic Approach,” in The international allocation of economic activity, ed. Bertil Ohlin, Per-Ove Hessleborn and Per Magnus Wijkman, 395–418. London: Macmillan. ———. 1998. Globalization, trade and foreign direct investment. New York: Pergamon. Durkan, Joseph, Doireann Fitz Gerald, and Colm Harmon. 1999. “Education and Growth in the Irish Economy,” in Understanding Ireland’s economic growth, ed. Frank Barry, 119–135. Houndmills et al.: Palgrave. Edelman, Marc, and Joanne Kenen. 1989. The Costa Rica reader. New York: Grove Weidenfeld. EGFSN (Expert Group on Future Skills Needs). 2003. Responding to Ireland’s Skills Needs. The Fourth Report of the Expert Group on Future Skills Needs. ⬍http:// www.forfas.ie/publications/egfsn031002/webopt/egfsn_4skills_report.pdf⬎. Egloff, Enrique. 2001. La inversión de Intel y “políticas micro” para fortalecer la competitividad en Costa Rica. Santiago de Chile: CEPAL, March 15, mimeo.

228

REFERENCES

Elson, Diane and Ruth Pearson. 1981. Nimble fingers make cheap workers: An analysis of women’s employment in third world export manufacturing. Feminist Review Spring: 87–107. Enterprise Ireland. 2002. Economic review of Ireland. Dublin: Enterprise Ireland, mimeo. ———. 2003a. Enterprise Ireland annual reports and accounts 2002. Dublin: Enterprise Ireland. ⬍http://www.enterprise-Ireland.com/connect-profiles.asp? sectorid⫽14⬎ Retrieved November 2003. ———. 2003b. Sector profiles. electronics. Dublin: Enterprise Ireland. ⬍http:// www.enterprise-Ireland.com/connect-profiles.asp?sectorid⫽14⬎ retrieved November 2003. Enterprise Strategy Group. 2004. Ahead of the curve. Ireland’s place in the global economy. Dublin: Forfas Secretariat. Ernst, Dieter. 2003a. How sustainable are benefits from global production networks? Malaysia’s upgrading prospects in the electronics industry. East–West Center Working Paper No. 57, June. ⬍http://www.eastwestcenter.org/stored/pdfs/ econwpo57.pdf⬎ retrieved November 2003. ———. 2003b. Upgrading perspectives in east Asia’s electronics industry. Harvard University, Science, Technology, and Innovation Program. ⬍http://www.cid. harvard.edu/cidbiotech/comments/comments162.htm⬎ retrieved November 2003. Esquivel, Gerardo, Mauricio Jenkins, and Felipe B. Larrain. 2000. Export processing zones in Central America. Harvard University, INCAE, and the Central American Bank for Economic Integration, Development Discussion Paper, Central America Project Series. ESRI (Economic and Social Research Institute). 2003. The mid-term evaluation of the national development plan and community support framework for Ireland, 2000–2006: Final report to the Department of Finance. Summary. ⬍www.esri.ie/ pdf/MTEsummary.pdf⬎ retrieved November 2003. European Commission. 2004. The social situation in the European union. 2004 overview. ⬍http://www.finfacts.ie/biz10/eusocialsituationEuropereport.pdf⬎ retrieved August 2004. FAS. 2002. The Irish labour market review 2002. ⬍http://www.fas.ie./ FAS_Review/LMO.html⬎ retrieved December 2003. Feenstra, Robert. 1998. Integration of trade and disintegration of production in the global economy. Journal of Economic Perspectives 13 (4): 31–50. FIAS (Foreign Investment Advisory Service). 1996. A strategy for foreign investment in Costa Rica’s electronics industry. Washington, DC: International Finance Corporation and the World Bank. Fieleke, Norman. 1989. Europe in 1992. New England Economic Review May–June: 13–25. Fitzgerald, John D. 1998. An Irish perspective on the structural funds. European Planning Studies 6 (6): 677–694. Forfas. 2001. Statement on outward direct investment. October. ⬍http://www. forfas.ie/publications/outward_direct_invest_01/outward_direct.pdf⬎ retrieved October 2003.

REFERENCES

229

———. 2002. International trade and investment, report 2001. A report by Forfas to the trade advisory forum and the foreign earnings committee. January. ———. 2003a. Annual business survey of economic impact for 2002, mimeo. ———. 2003b. International trade and investment report 2002. Dublin: Forfas. ———. 2003c. Research and development in the business sector 2001. Dublin: Forfas. ———. 2003d. World trade organisation negotiating objectives for Irish enterprise policy. ⬍http://www.forfas.ie/publications/0302_forfas_WTO_72dpi_s.pdf⬎ retrieved January 2004. Fosfuri, Andrea, Massimo Motta, and Thomas Ronde. 1999. Foreign direct investment and spillovers through workers’ mobility. London: Centre for Economic Policy Research, Discussion Paper No. 2194, August. Gereffi, Gary. 2002. Outsourcing and changing patterns of international competitiveness in the global apparel commodity chain. Paper presented at the Conference on “Labor and Globalization of Production,” New York City, New School University, March 8. Girma, Sourafel, Holger Goerg, and Eric Strobl. 2003. Government grants, plant survival and employment growth: A micro-econometric analysis. Institute for the Study of Labor, Bonn, Discussion Paper Series No. 838. Girma, Sourafel, David Greenaway, and Katharine Wakelin. 2001. Who benefits from foreign direct investment in the UK? Scottish Journal of Political Economy 48: 119–133. Goerg, Holger and David Greenaway. 2003. Much ado about nothing? Do domestic firms really benefit from foreign direct investment? Bonn: Institute for the Study of Labor, Discussion Paper Series No. 944, November. Goerg, Holger and Frances Ruane. 2000. An analysis of backward linkages in the Irish electronics sector. The Economic and Social Review 31 (3): 215–235. Goerg, Holger and Eric Strobl. 2002a. Multinational companies and indigenous development: An empirical analysis. European Economic Review 46 (7): 1305–1322. ———. 2002b. Spillovers from foreign firms through worker mobility: An empirical investigation. Bonn: Institute for the Study of Labor, Discussion Paper Series No. 591. Gonzalez, Anabel. 2001. Key drivers for investing in Costa Rica: The Intel case. Presentation at the OECD Global Forum on International Investment, Mexico City, November 2001, mimeo. Gudmundson, Lowell. 1995. “Peasant, Farmer, Proletarian: Class Formation in a Smallholder Coffee Economy, 1850–1950,” in Coffee, society, and power in Latin America, ed. William Roseberry, Lowell Gudmundson and Mario Samper Kutschbach, 112–150. Baltimore and London: John Hopkins University Press. Haddad, Mona and Ann Harrison. 1993. Are there positive spillovers from direct foreign investment? Evidence from panel data for Morocco. Journal of Development Economics 42: 51–74. Hanlon, Deirdre. 1998. Developments in corporation tax policy in Ireland. Dublin: Department of Economics, Trinity College, mimeo.

230

REFERENCES

Hanson, Gordon. 2001. Should countries promote foreign direct investment? UNCTAD, G-24 Discussion Paper Series No. 9, February. Hanson, Gordon, Raymond Mataloni Jr., and Matthew Slaughter. 2002. “Expansion Strategies of U.S. Multinational Firms,” in Brookings trade forum: 2001, ed. Susan M. Collins and Dani Rodrik, 245–294. Washington, DC: Brookings Institution Press. Haskel, Jonathan E., Sonia C. Pereira, and Matthew Slaughter. 2002. Does inward foreign direct investment boost the productivity of domestic firms? National Bureau of Economic Research, NBER Working Paper No. 8724, January. Haughton, Jonathan. 1995. “The Historical Background,” in The economy of Ireland: Policy and performance of a small European country, ed. John O’Hagan, 1–48. New York: St. Martin’s Press. Hewitt-Dundas, Nola, Bernadette Andreosso-O’Callaghan, Mike Crone, John Currary, and Stephen Roper. 2002. Learning from the best: Knowledge transfers from multinational plants in Ireland: A north–south comparison. Belfast: Innovation Lab: NIERC (Northern Ireland Economic Research Centre). Honohan, Patrick and Brendan Walsh. 2002. Catching up with the leaders: The Irish hare. Washington, DC: Brookings Panel on Economic Activity, April 4–5. Huckman, Robert S. 2003. MedSource technologies. Harvard Business School, Case 9-603–081. Hufbauer, Gary Clyde. 1990. Europe 1992: Opportunities and challenges. The Brookings Review 8 (3): 13–22. Hummels, David, Jun Ishii, and Kei-Mu Yi. 2001. The nature and growth of vertical specialization in world trade. Journal of International Economics 54: 75–96. Hyot, David and Hau Lee. 2001. Solectron: From global manufacturer to global supply chain integrator. Graduate School of Business, Stanford University, Case number: GS-24, November. IADB (Inter-American Development Bank). 1997. Latin America after a decade of reforms. Economic and social progress, 1997 report. Washington, DC: John Hopkins University Press for the Inter-American Development Bank. ———. 2001. Economic and social progress in Latin America. Competitiveness: The business of growth. Washington, DC: John Hopkins University Press for the Inter-American Development Bank. ———. 2003a. Country program evaluation: Costa Rica 1990–2001. Office of Evaluation and Oversight. Washington, DC: IADB. February 19. ———. 2003b. IDB strategy with Costa Rica. Washington, DC: IADB. May. IDA (Industrial Development Authority). 2003. Annual report 2002. Dublin: IDA. IICE (Instituto de Investigaciones en Ciencias Económicas). 2001. Survey of characteristics of exporting and non-exporting firms. University of Costa Rica, mimeo. IMF (International Monetary Fund). 1953. International financial statistics. 6 (5). ———. 1954. International financial statistics. 7 (8). ISO (International Organization for Standardization). 2003. The ISO survey of ISO 9000 and ISO 14000 certificates, tenth cycle up to and including 31 December 2000. ⬍http://www.iso.ch/iso/en/iso9000-14000/pdf/survey10thcycle.pdf⬎ retrieved November 2003.

REFERENCES

231

Jacobson, David and Bernadette Andreosso. 1990. “Ireland as a Location for Multinational Investment,” in The Single European Market and the Irish Economy, ed. Anthony Foley and Michael Mulreandy, 307–334. Dublin: Institute of Public Administration. Jacobson, David and Ziene Mottiar. 1999. Globalization and modes of interaction in two sub-sectors in Ireland. European Planning Studies 7 (4): 429–444. Kaldor, Nicholas. 1972. The irrelevance of equilibrium economics. Economic Journal 82 (328): 1237–1255. Kaminski, Bartlomiej and Beata K. Smarzynska. 2001. Foreign direct investment and integration into global production and distribution networks: The case of Poland. Washington, DC: World Bank, Working Paper No. 2646. Kearns, Allan. 2000. Essays on the consequences of research and development for manufacturing firms in Ireland. Department of Economics, University of Dublin, Trinity College. Keating, William. 2000. Measuring the economy: Problems and prospects. Journal of the Statistical and Social Inquiry Society 30: 1–32. Kenney, Martin and James Curry. 1999. Beating the clock: Corporate responses to rapid change in the PC industry. California Management Review Reprint Series 42 (1): 8–36. Keogh, Elaine. “Panasonic Blames Low PC Prices for Closure of Factory,” Irish Times, February 19, 2000. Ketelhoehn, Niels W. and Michael E. Porter. 2002. Building a cluster: Electronics and information technoloy in Costa Rica. Harvard Business School, Case 9-703–422. Kirkland Jr., Richard I. 1989. Europe goes wild for yankee PCs. Fortune, June. Klein, Michael, Carl Aaron, and Bita Hadjimichael. 2001. Foreign direct investment and poverty reduction. Washington, DC: World Bank, Working Paper No. 2613. Kokko, Ari. 1996. Productivity spillovers from competition between local firms and foreign affiliates. Journal of International Development 8 (4): 517–530. Kokko, Ari, Ruben Tansini, and Mario C. Zejan. 1996. Local technological capability and productivity spillovers from FDI in the Uruguayan manufacturing sector. Journal of Development Studies 32: 602–611. Konings, Jozef. 2000. The effects of foreign direct investment on domestic firms: Evidence from firm level panel data in emerging economies. London: Centre for Economic Policy Research, Discusion Paper No. 2586, October. Krugman, Paul R. and Anthony Venables. 1995. Globalisation and the inequality of nations. Quarterly Journal of Economics 110: 857–880. ———. 1996. Integration, specialization, and adjustment. European Economic Review 40: 959–967. Kynge, James. “Investment in China Reaches Record High of $54 Billion,” Financial Times, November 16, 2004. La Nación Week in Review, “42,000 Households Fall into Poverty,” December 2, 2004. Lall, Sanjaya. 2000. The technological structure and performance of developing country exports, 1985–1998. Oxford Development Studies, 28: 337–368. ———. 2001. Competitiveness, technology, and skills. Cheltenham and Northampton: Edward Elgar Publishing.

232

REFERENCES

Lane, Phillip. 1998. Profits and wages in Ireland, 1987–1996. Journal of the Statistical and Social Inquiry Society of Ireland 28, Part V: 223–252. Larrain, Felipe B., Luis F. Lopez-Calva, and Andres Rodríguez-Clare. 2000. Intel: A case study of foreign direct investment in Central America. Center for International Development, Harvard University, Working Paper No. 58, December. Larudee, Mehrene, José Antonio Cordero-Peña, and Leticia Arroyo-Abad. 2001. Climbing the skill ladder: Costa Rica’s export processing zone, 1995–2000, mimeo. Layte, Richard, Brian Nolan, and Christopher Whelan. 2004. Explaining poverty rates in Ireland during the boom. Irish Banking Review Summer: 1–13. Lincoln, John E. 2003. The 510 (k). Its purpose, compilation, and submission. Institute of Validation Technology. ⬍http://www.ivthome.com/free/se/ 510k.htm⬎ retrieved October 2003. Lipsey, Robert. 1990. American firms face Europe: 1992. National Bureau of Economic Research, NBER Working Paper 3293, March. Lyons, Declan. 2002a. The changes facing electronics sub-contracting. Dublin: Enterprise Ireland, mimeo. ———. 2002b. Linkage policies in Ireland—maximizing the opportunities arising from multinational investment. Dublin: Enterprise Ireland, mimeo. Mac Carthaigh, Sean. “Solectron Will Create 350 Jobs at NCR Plant,” Irish Times, December 20, 1997. ———. “Canadian Firm Announces 500 Jobs for Swords,” Irish Times, September 18, 1998. Mac Sharry, Ray and Padraic White. 2000. The making of the Celtic Tiger. Dublin: Mercier Press. Markusen, James and Anthony Venables. 1999. Foreign direct investment as a catalyst for industrial development. European Economic Review 43: 335–356. Mason, Mark and Dennis Encarnation, eds. 1994. Does ownership matter? Japanese multinationals in Europe. Oxford: Clarendon Press. McAleese, Dermot and Donogh McDonald. 1978. Employment growth and the development of linkages in foreign-owned and domestic manufacturing enterprises. Oxford Bulletin of Economics and Statistics 40 (4): 321–329. McKinsey Global Institute. 2003. Multinational company investment: Impact on developing countries. McKinsey and Company. ⬍http://mckinsey.com/ knowledge/mgi/newhorizons/pdf⬎. Ministerio de Comercio Exterior de Costa Rica. Estadísticas. Inversión extranjera directa. ⬍http://www.comex.go.cr/estadisticas/inversion/default.htm⬎. Mitchell, Brian R., ed. 1998. International historical statistics, Europe 1750–1993. 4th edition. Houndmills et al.: Stockton Press. Monge Gonzalez, Ricardo, Julio Rosales Tijerino, and Gilberto Arce Alpizar. 2004. Análisis costo-beneficio del régimen de zonas francas. Impactos de la inversión extranjera directa en Costa Rica. Report prepared for PROCOMER (Promotora de Comercio Exterior de Costa Rica), mimeo. Morley, Samuel A., Roberto Machado, and Stefano Pettinato. 1999. Indexes of structural reform in Latin America. Vol. 12. Santiago de Chile: CEPAL, Serie Reformas Economicas.

REFERENCES

233

Mortimore, Michael and Wilson Peres. 1998. Policy competition for foreign direct investment in the Caribbean Basin: Costa Rica, the Dominican Republic, and Jamaica. Working Paper No. 49, January. Santiago de Chile: CEPAL, Serie Desarrollo Productivo. Mortimore, Michael and Ronney Zamora. 1999. The international competitiveness of the Costa Rican clothing industry. Working Paper No. 46, April. Santiago de Chile: CEPAL, Serie Desarrollo Productivo. Murray Brown, John. “Ireland Extends its Hospitality to Top Scientists,” Financial Times, November 28, 2003. ———. “How Dell Keeps going in Euorope,” Financial Times, June 1, 2004. Naughton, Barry. 1996. China’s emergence and prospects as a trading nation. Brookings Papers on Economic Activity 2: 273–344. Navaretti, Giorgio Barba, Jan I. Halland, and Anthony Venables. 2002. Multinational corporations and global production networks: The implications for trade policy. London: Centre for Economic Policy Research. Ocampo, José Antonio. 2002. Small economies in the phase of globalization. William G. Demas Memorial Lecture at the Caribbean Development Bank, Cayman Islands, May 14. OECD (Organization for Economic Cooperation and Development). National Accounts of OECD Countries: General Government Accounts. ⬍http://www. oecd.org/document/5/0,2340,en_2825_495684_33785349_1_1_1_1,00.html⬎. ———. 1990. “OECD in figures—1990 edition,” in OECD Observer. Vol. 164, June/July. ———. 2003. OECD science, technology and industry scoreboard. 2003 edition. Paris: OECD. OECD and IADB. 2003. Caribbean rim investment initiative. Business environment report. Costa Rica, April 29. ⬍http://www.oecd.org/dataoecd/62/20/2635563. pdf⬎ retrieved July 2004. O’Farrell, P. N. and Brian O’Loughlin. 1981. New industry input linkages in Ireland: An econometric analysis. Environment and Planning 13 (3): 285–308. O’Hearn, Dennis. 1998. Inside the Celtic Tiger: The Irish economy and the Asian model. London and Sterling, VA: Pluto Press. O’Malley, Eoin. 1998. “Industrial Policy in Ireland and the Problem of Late Development,” in Latecomers in the global economy, ed. Michael Storper et al., 254–274. London and New York: Routledge. O’Malley, Eoin and Colm O’Gorman. 2001. Competitive advantage in the Irish indigenous software industry and the role of inward foreign direct investment. European Planning Studies 9 (3): 303–321. O’Riain, Sean. 1999. The flexible developmental state: Globalization, information technology and the Celtic Tiger. Paper presented at conference “Global Networks, Innovation, and Development Strategy: The Informational Region as a Development Strategy,” Center for Global, Regional, and International Studies, University of California Santa Cruz, November, 11–13. ⬍http://www2.ucsc.edu/cgirs/publications/cpapers/oriain.pdf⬎ retrieved October 2003.

234

REFERENCES

O’Sullivan, Mary. 1995. “Manufacturing and global competition,” in The economy of Ireland: Policy and performance of a small European country, ed. John O’Hagan, 361–396. New York: St. Martin’s Press. ———. 2000. The sustainability of industrial development in Ireland. Regional Studies 34 (3): 277–290. Paus, Eva. 2004. Productivity growth in Latin America: The limits of neoliberal reforms. World Development 32 (3): 427–445. Pietrobelli, Carlo and Roberta Rabellotti. 2004. Upgrading in clusters and value chains in Latin America. The role of policies. Washington, DC: IADB. Porter, Michael. 1990. The competitive advantage of nations. New York: Free Press. ———. 1998. Clusters and the new economies of competition. Harvard Business Review November–December: 77–90. PROCOMER (Promotora del Comercio Exterior de Costa Rica) Datos de Comercio Exterior en Línea. ⬍http://servicios.procomer.go.cr/estadisticas/Estadisticas. jsp?val⫽Panel⬎. Proyecto Estado de la Nación. 2004. Décimo informe del estado de la nación. San José: Costa Rica. Reich, Robert. 1991. The work of nations: Preparing ourselves for 21st century capitalism. New York: Alfred Knopf. Reinhardt, Nola and Wilson Peres. 2000. Latin America’s new economic model: Micro responses and economic restructuring. World Development 28: 1543–1566. Roberts, Bill. 2003a. Top 100 contract manufacturers. Electronic Business, September. ———. 2003b. The ups and downs of contract manufacturing. Electronic Business, September. Robles, Edgar. 1999. Agenda para la competitividad de Costa Rica hacía el siglo XXI. ⬍www.incae.ac.cr/ES/clacds/investigacion/pdf/agendacostarica.pdf⬎ retrieved December 2003. Rodríguez-Clare, Andrés. 1996. Multinationals, linkages, and economic development. American Economic Review 86 (4): 852–873. ———. 2001. Costa Rica’s development strategy based on human capital and technology: how it got there, the impact of Intel, and lessons from other countries, New York: UNDP, mimeo. Rodrik, Dani. 2004. Industrial policy for the twenty-first century. John F. Kennedy School of Government, Harvard University, KSG Working Paper No. RWP04–047, November. Rosenberg, Nathan. 1994. Exploring the black box. Technology, economics, and history. Cambridge and New York: Cambridge University Press. Ruane, Frances. 2001. Reflections on linkage policy in Irish manufacturing—policy chasing a moving target? Paper presented at the UNECE/EBRD Expert Meeting on Financing for Development: Enhancing the Benefits of FDI and Improving the Flow of Corporate Finance in the Transition Economies, Geneva, December 3. Ruane, Frances and Holger Goerg. 1997. Reflections on Irish industrial policy towards foreign direct investment. Dublin: Trinity College, Trinity Economic Paper Series, Policy Paper No. 97/3. Ruane, Frances and Ali Udur. 2000. Foreign direct investment and productivity spillovers in Ireland—some preliminary results. Paper delivered at the annual conference of the Irish Economic Association, Waterford, March 21–April 2.

REFERENCES

235

———. 2002. Foreign direct investment and productivity spillovers in the Irish manufacturing industry: Evidence from firm level data. Dublin: Trinity College, Trinity Economic Paper Series, Policy Paper No. 02/06. Sánchez-Ancochea, Diego. 2004. “Leading coalitions” and patterns of accumulation and distribution in small countries. Ph.D. Dissertation, New School University, April. SFI (Science Foundation Ireland). 2002. Vision 2003–2007: People, ideas and partnerships for a globally competitive Irish research system. Dublin: SFI. Shotwell & Carr, Inc. 2003. Validation of manufacturing processes. ⬍http://www. shotcarr.com/sm2.pdf⬎ retrieved October 2003. Sjoeholm, Fredrik. 1999. Technology gap, competition and spillovers from direct foreign investment: Evidence from establishment data. The Journal of Development Studies 36 (1): 53–73. Slaughter, Matthew. 2002. Skill upgrading in developing countries: Has inward foreign direct investment played a role? Paper presented at the Conference on “Labor and the Globalization of Production” at the Center for Economic Policy Analysis of the New School University, New York, March 8. Smyth, Emer and Damian F. Hannan. 2000. “Education and inequality,” in Bust to boom? The Irish experience of growth and inequality, ed. Brian Nolan, Philip O’Connell and Christopher T. Whelan, 109–126. Dublin: Institute for Public Administration. Smyth, Jamie. “Deal With Celestica To Cost 750 Jobs At Motorola Plant In Dublin,” Irish Times, December 7, 2000. ———. “More Than 300 Tech Jobs Lost In Dublin,” Irish Times, May 2, 2003. Song, Jung-A. “Small Town Sets Out to Become a Silicon Centre,” Financial Times, December 1, 2004. Spar, Debora. 1998. Attracting high technology investment. Intel’s Costa Rica plant. Foreign Investment Advisory Service, Occasional Paper 11, Washington, DC: World Bank. Stallings, Barbara, and Wilson Peres. 2000. Growth, employment, and equity. The impact of economic reforms in Latin America and the Caribbean. Washington, DC: UNECLAC and Brookings Institution Press. Standing, Guy. 1989. Global feminization through flexible labor. World Development 17 (7): 1077–1095. Sturgeon, Timothy. 2000. How do we define value chains and production networks? MIT IPC Globalization Working Paper 00–010. Cambridge, MA: MIT Industrial Performance Center. Sturgeon, Timothy and Richard K. Lester. 2003. The new global supply-base: New challenges for local suppliers in East Asia. Paper prepared for the World Bank’s Project on East Asia’s Economic Future, Industrial Performance Center, MIT, January 14. Sullivan, Martin. 2004a. Drug firms park increasing share of profits in low-tax countries. Tax Notes, September 20: 1136–1343. ———. 2004b. Economic analysis: Data show U.S. companies shifting profits to tax havens. Tax Notes International, September 20: 1035–1046. Swann, Christopher. “Individuals Will Feel the Pinch as Corporate Tax Take Driven Down,” Financial Times, May 5, 2003.

236

REFERENCES

Taylor, Christopher T. 2000. The impact of host country government policy on U.S. multinational investment decisions. The World Economy 23 (5): 635–648. Taylor, Cliff. “Industrial Employment Falls by 14,800,” Irish Times, December 23, 2003. Tomaney, Rolf. 1995. Recent developments in Irish industrial policy. European Planning Studies 3 (1): 99–113. Ulate Quiros, Anabelle. 2000. “Reformas Económicas en Costa Rica: Un Desafío para el Empleo y el Crecimiento con Equidad,” in Empleo, crecimiento y equidad. Los retos de las reformas económicas de finales del siglo XX en Costa Rica, ed. Anabelle Ulate Quiros, 15–65. San José: Editorial de la Universidad de Costa Rica. UNCTAD (United Nations Conference on Trade and Development). Foreign Direct Investment Database. ⬍http://www.unctad.org/Templates/Page.asp? intItemID⫽ 1923&lang⫽1⬎. ———. 2001a. Economic development in Africa. Performance, prospects, and policy issues. ⬍http://www.unctad.org/en/docs//pogdsafricad1.en.pdf⬎. ———. 2001b. World investment report 2001. New York: United Nations. ———. 2002. World investment report 2002. New York: United Nations. UNDP (United Nations Development Program). 2004. Human development report New York: United Nations. U.S. Department of Commerce, Bureau of Economic Analysis. U.S. Direct Investment Abroad: Balance of Payments and Direct Investment Position Data. ⬍http://www.bea.doc.gov/bea/di/di1usdbal.htm⬎. ———. 1985. U.S. direct investment abroad: 1982 benchmark survey. U.S. Government Printing Office, Final Results, December. ———. 1992. U.S. direct investment abroad: 1989 benchmark survey. U.S. Government Printing Office, Final Results, October. ———. 1998. U.S. direct investment abroad: 1994 benchmark survey. Final Results, May. USAID (United States Agency for International Development). 2004. Overseas loans and grants, obligations and loan authorizations. “The Greenbook.” ⬍http://qesdb.cdie.org/gbk/⬎ retrieved August 2004. USITC (United States International Trade Commission). Interactive Tariff and Trade Dataweb. ⬍http://dataweb.usitc.gov/⬎. ———. 2004. Textiles and apparel: Assessment of the competitiveness of certain foreign suppliers to the U.S. market. Investigation No. 332–448, Publication 3671. Wade, Robert. 1994. “Selective Industrial Policies in East Asia: Is ‘The East Asian Miracle’ Right?” in Miracle or design? Lessons from the east Asian experience, ed. Albert Fishlow, et al. Overseas Development Council ed., 55–79, Washington DC. Wall Street Journal, “Intel Plans to Build Irish Computer Plant at $ 400 Million Cost,” October 4, 1989. Walsh, Brendan. 2003. When unemployment disappears: Ireland in the 1990s. CESIFO Working Paper No. 856. Whitaker, T. K. 1983. Interests. Dublin: Institute of Public Administration. White, Padraic. 2000. “EEC membership: A Boon to Foreign Investment,” in The making of the Celtic Tiger, ed. Ray Mac Sharry, Padraic White, 178–182. Dublin: Mercier Press.

REFERENCES

237

Wilson, Bruce. 1998. Costa Rica. politics, economics, and democracy. Boulder and London: Lynne Rienner Publishers. Wilson, Tamar Diana. 2002. The masculinization of the mexican maquiladoras. The Review of Radical Political Economics 34: 3–17. Wong, Poh-Kam. 1992. Technological development through subcontracting linkages: Evidence from Singapore. Scandinavian International Business Review 1 (3): 28–40. Woodward, Douglas, and Robert J. Rolfe. 1993. The location of export-oriented foreign direct investment in the Caribbean basin. Journal of International Business Studies 24 (1): 121–144. World Bank Group. 2003. World development indicators online. ⬍http://www. worldbank.org/data/onlinedatabases/onlinedatabases.html⬎. World Economic Forum. 2002. Global competitiveness report 2001–2002. New York and Oxford: Oxford University Press. Yusuf, Shahid. 2003. Innovative East Asia. The future of growth. A World Bank Publication. New York and Oxford: Oxford University Press.

This page intentionally left blank

Index

Tables and Figures are in Bold Face. Aaron, Carl, 17 Abbot, 165, 169, 170, 171, 187 Active Processing Regime, 151 administrative capacity, factors in effectiveness of, 71–2 agglomeration benefits, 63–4 Agreement on Textiles and Clothing (ATC), Uruguay Round, 186 Aitken, Brian, 27 Altschuler, Roseanne, 19 Amsden, Alice, 6, 11, 16, 41, 168, 201, 207–8nn5, 6 Annual Business Survey of Economic Impact (ABSEI), 85–6, 88, 110, 113, 214n2 Apple (personal computers), 48, 115 approved vendor list (AVL), 100, 108, 196 Arce Alpizar, Gilberto, 145, 148, 149, 183 Arias administration, 222n5 Arndt, Heinz, 215n14 Arroyo-Abad, Leticia, 182 ArtinSoft software, 184 Barrios, Salvador, 119, 120, 215nn24, 25 Barry, Frank, 49, 63, 64, 70, 118, 124 Baxter, 143, 157 Bell Labs, 128, 129 Bellak, Christian, 17, 33 biotechnology goals for in Ireland, 126, 129 partnership with medical devices industry, 187, 216n33

Bitar, Javier, 177 Blomstrom, Magnus, 28, 29 Bradley, John, 63, 64, 65, 118 Breathnach, Marcus, 87, 91, 93 Brennan, Ciaran, 101, 102 Buitelaar, Rudolph, 177, 181 Calderon, Rafael, 163 call centers, 49, 129, 144, 205 Cantwell, John, 17, 33 capital access to finance, 178 difficulty of taxing global, 164 high mobility of, 16 increased rate of return on, 71 intensity, increase in, 205 relative power of, 19 taxation of, 19 capital-labor relations, 70–1, 79–80 Carazo government, Costa Rica, 138, 217n8 Caribbean Basin Initiative (CBI), 142, 143, 157, 195, 218n18 Caves, Richard E., 17, 208n8 Celestica, 35, 36, 50, 100, 101 Census of Industrial Production (CIP), 85–6, 112, 113, 211n20 Central America, civil wars in, 137, 140–1 Central American Common Market (CACM), 137 Central American Free Trade Agreement (CAFTA), 161, 185, 186, 195, 203

240

INDEX

Central European countries, as competitors, 77, 172, 190, 194 Certificado de Abono Tributario (CAT), 154, 201, 217n12 Chamber of Industries and Unidad de Asistencia Industrial (UATI), 179–80 Chambers, Bill, 96 China, as competitor country, 7, 9, 39, 77, 172, 180, 190, 194, 195 CINDE, see Coalition for Development Initiatives Clark, Mary, 142, 217n13, 221n59 Clinch, Peter, 52, 69 cluster development, 22, 77, 132, 165–6, 170, 187, 191, 200, 205 Coalition for Development Initiatives (CINDE), Costa Rica, 142, 158, 164, 165–8, 179–80, 199, 201, 221n59 Community Support Framework (CSF), Ireland, 64, 66, 67 components customized production of, 106, 107 increased international trade in, 32, 33 computer industry cyclical demand in, 145 demand boom in, 75 development of, 35 employment composition of FOEs in, 114–15 product life cycles in, 37, see also electronics industry contract manufacturers (CMs), 35–7, 36 common use of, 99 in Ireland, 50, 100–1, 190 in medical device industry, 106, 107 Convery, Frank, 52, 69 Cordero-Peña, José Antonio, 182 Costa Rica, 7–8 absence of proactive government strategy in, 166–7, 170, 180–1, 186–7

agriculture in, 135, 136–7, 150, 188, 216n4, 220n37 civil service in, 137, 218n17 development strategy, political support for, 162–3, 171 disarticulation, institutional, in, 142 domestic sourcing in, 175, 176–7, 187 economic data on, 139 educational system in, 158–9, 167, 170, 182, 183, 193 employment in, 142 engineers in, 159 export composition, 146, 150, 151 foreign debt in, 138, 210n7 free zones in, 140, 145–7, 148, 161, 170–1, 219n36, 221n54 government stability in, 162–3 greater income equality in, 137, 193 human capital in, 158–60, 183–4 infrastructure of, 138, 161–2, 167, 170, 202 labor relations in, 137 lack of tax reform, implications of, 187 location-specific assets of, 158–64 manufacturing value added in, 217n5 metal-mechanics sector in, 177 monoexport dependency of, 136–7 national investment share, 145 negative net aid flows in, 161, 162 private banking sector, rise of, 138, 142 public investment, decline of, in, 210n14, 220n49 social democracy in, 216nn1, 6 software industry in, 184, 188 spillovers in, 158–60, 174, 182–4 strategic location of, 141, 142, 218n16 tariff liberalization in, 138, 154 tourism sector in, 143–4, 188, 205 Utilities, privatization of, 138, 148 U.S. aid to, 141–2, 141, 217–18nn13, 14 wage, average, 219n35, 220n47

INDEX

Costa Rican Development Corporation (CODESA), 137–38 Costa Rica Provee (CRP), 174, 179–80 Crone, Mike, 26, 92–3, 109 Culliton Report, 73, 94, 132 Currary, John, 26, 92–3, 109 Curry, James, 35, 37 Dell, Irish PC production by, 115, 117, 129 demonstration effect of FDI, 26–7, 37–8, 63–4, 156–7, 191, 197 development, economic defined, 2, 5, 11 limited external funding for, in Costa Rica, 8, 136 and political economy, dilemmas of, 184–8 policies, Costa Rica, 136–43, 171 policies, Ireland, 45–9, 64, 66, 132 strategy based on productivity growth, 13, 14, 81, 190 transition to R&D driven, Ireland, 7, 81–2, 126–7, 127 developmentalist state, 200, 205 Dicken, Peter, 32, 34, 209nn18, 19, 20 Dimelis, Sophia, 27, 28, 29, 119, 120, 215nn24, 25 Dunning, John H., 17 Durkan, Joseph, 67 Eastern European countries as competitors, 172, 190, 194 Irish IOE production facilities in, 99 ELAN pharmaceuticals, 110, 125 electronics industry, Ireland, 43 characteristics of, 34 downturn in, 136 explosive growth of demand in, 14, 35, 59 global value chains in, 34–7 growth of sub-supply sector in, 121 production costs in, 35 quality standards in, 117 restructuring of, 89–90, 100–3, see also computer industry

241

Elson, Diane, 112 Enterprise Ireland, 89, 90, 94, 96–7, 97, 98, 117, 132, 198 Enterprise Strategy Group, Ireland, 79, 131 Ernst, Dieter, 26 Esquivel, Gerardo, 181, 223n8 “Europe 1992,” 61, 64, 75, 195, 212n2 European Community (EC) and European Economic Community (EEC) benefits of membership in, 47 Common Agricultural Policy of, 47–8 as consumer market, 61 convergence in, 203 as destination for FDI, 61 Ireland’s entry into, 173 policies and true common market, 64, 203 structural funds from, for Ireland, 64–5 see also European Union (EU) European Monetary System (EMS), Ireland’s entry into, 210n6 European Union (EU), 39 Cohesion Fund, 64, 65 current decline in funding support from, 79 and global value chains and location-specific assets, 59–77 Irish location-specific assets within, 60–1 new member countries of, 205, 212n35 Structural Funds, 62, 64, 66, 72, 75, 202, 205 see also European Community and European Economic Community exchange rate fluctuations, vulnerability to, 53 export decomposition, by technology intensity, 152 Export Sales Tax Relief Act of 1956, 58

242

INDEX

FDI, see Foreign Direct Investment FDI inflows, 4 change in, 3–5 in Costa Rica, 139, 143, 144–5, 144, 203 in developing countries, 5 high concentration of, 3 in Ireland, 49–50 Feenstra, Robert, 32 Fieleke, Norman, 61 Figueres, José Maria, 137, 163, 167, 217n7 Fitz Gerald, Doireann, 67 Fitzgerald, John D., 64, 65 Flextronics, 35, 36, 100, 101–2 FOEs and IOEs in Ireland management gap between, 91, 91 productivity gap between, 122, 124 wage ratio between, 122, 124 See also foreign-owned enterprise; indigenously-owned enterprise Foreign Direct Investment (FDI) -in Costa Rica: economic performance and, 143–55, 218n24; and generation of foreign exchange, 147, 147, 149; Index of Inward FDI Potential in 17–18, 50, 144; macro-level impact of, 145, 149; motivation for, 155–68; share of, in GFCF, 145; and trade and structural change, 149–55 -in Ireland: conditions propitious to, 18; and generation of foreign exchange, 24–5; implications of, at macro level, 42; in internationally traded services, 129; mode of entry of, 24, 27; need for critical mass of, 120; pursuit of, 125; reasons for, 42; reorganization of, 1; and taxes, 57–9; and trade and structural change, 52–7; and transfer pricing, 50–2; of U.S., 62 see also FDI inflows; high-tech FDI Foreign Investment Advisory Service, 179

Foreign-owned enterprise (FOE), Ireland increased skill intensity in, 112, 121 local sourcing by, 84–5, 86, 88 lower R&D intensity in, 112 see also FOEs and IOEs in Ireland Forfas, 53, 75–6, 85–6, 90, 110, 115, 124, 128 Fosfuri, Andrea, 27 free zones (FZs), Costa Rica, 140 expansion of, 149 exports of, 145–7, 219n36 investment requirements for, 221n54 limited uses of, 181 positive contributions of, 148 preferential treatment of TNCs in, 161, 169–71 General Agreement on Tariffs and Trade (GATT), 140 Geretti, Gary, 34 Girma, Sourafel, 29, 73, 97 Glass, Alastair, 128, 216n35 globalization, economic, meaning of, 1, 2, 3–6 global value chains changed threshold for participation in, 33, 37 in electronics industry, 34–7 the EU and location-specific assets, 59–77 in medical device industry, 107 organization of, 28, 31–4, 102, 107, 195 R&D, location of, in, 40, 41, 42, 99 restructuring of, 172, 195 see also value chains Globerman, Steven, 29 Goerg, Holger, 27, 49, 63, 73, 97, 102, 120, 124, 208n12 government policies -Costa Rica: barriers to proactivity in, 138–40, 180–1, 186–7; development, 136–43, 171;

INDEX

in education, 158–9; linkageformation policies, 179–80; location-specific assets and, 158–62; revenue/GDP, 148–9, 148, see also Costa Rica -industrial: for FDI development, 199–202; political challenges to adoption of, 201; for promotion of indigenous industry, 29, 98 -Ireland: development, 45–9, 64, 66, 81; in education, 22, 67–8; and free market approach, 95; to improve equality, 133; infrastructure, 22; linkage-formation policies, 94–9, 215n13; location-specific assets and, 42–3; role of, 173–4; see also Ireland -need for proactive strategy in, 185, 199 Greenaway, David, 29, 208n12 greenfield investment, 24, 144 growth, economic, 12–15, 208n8 constraints on, in Costa Rica, 147–9 New Growth Theory, 13 technology intensity and, 12–13 see also productivity growth Grubert, Harry, 19 Gudmundson, Lowell, 216n2 Haddad, Mona, 27, 29 Hadjimichael, Bita, 17 Halland, Jan I., 33, 209n16 Hannan, Aoife, 65 Hannan, Damian F., 67 Hanson, Gordon, 181 Harmon, Colm, 67 Harris, Bill, 128 Harrison, Ann, 27, 29 Harvard Institute for International Development, 168 Haskel, Jonathan, 18, 27, 208n11 health-care costs, pressures for containment of, 106–7 Hewitt-Dundas, Nola, 26, 92–3, 109

243

Hewlett Packard, 63, 71 high-tech FDI in Costa Rica, contingencies for, 135, 136, 169, 169–70 development, industrial policies for, 199–202 English-speaking labor and, 67, 69, 187 and expansion of knowledge-based assets, 25–37 human capital effect, 26–7, 37–8, 110–20, 158–60 impact of on macroeconomic variables, 24–5 importance of, in Ireland, 126, 189 in industrial upgrading, 23–42 influence of, on indigenous producers, 26 in Ireland, contingencies for, 76 in SLCs, 21, 21–3, 192–4, 202–5 spillover effects of, 26–8, 42, 182–4 and technology transfer in Ireland, 38, 120–5 as training ground for entrepreneurs, 27, 121, 130, 191, 198 high-tech industries export share of, 53, 207nn2, 4 growth of, 14–15 knowledge spillovers from, 116–19 trade ratios for, 15 high-tech production heterogeneity of, 15 labor force required for, 18, 182–3 potential of, 6 R&D intensive nature of, 13 standardization of, 6 surge in, in Costa Rica, 150 world demand for, 14, 35, 59 Honohan, Patrick, 210nn9, 14 HTS 9802, 142–3, 151, 174, 180, 181 Huckman, Robert S., 106 human capital formation, 26–7, 37–8, 110–20, 158–60 Hummels, David, 32 Hyot, David, 37

244

INDEX

IBM, 71, 115, 117 import-substituting industrialization (ISI), 5, 16, 18, 95, 125, 137, 138 INCAE, 168, 200 incentives, 18–19 financial, in FDI, 19, 71–5, 163–4, 167 for higher value-added activities, 80 tax, 19–20, 21, 24, 72, 73, 138–40, 203, 204 India, as competitor country, 39, 77, 188, 190, 194 indigenously-owned enterprises (IOEs) barriers to responding to TNC needs, 29 impact of creative market failures on, 98 imperfect information about, 98 insufficient scale of, 98, 102, 131–2, 200 linkages with, 83, 84–110, 205 output and export structure of, 121–2 R&D in, 122–4 sourcing by, 88, 89, 176 see also FOEs and IOEs in Ireland Industrial Development Authority (IDA), 201, 210n5 cash grants made by, 59, 72–4, 73 indiscriminate pursuit of FDI by, 125, 213n44 rationale behind, 199 restricted responsibility of, 94 role of, 59, 68, 71 industrial upgrading, technological capabilities and small country size, 11–17 infrastructure, see Costa Rica; Ireland; SLCs innovation, 16, 116–18, 127–8, 128, 214n6 input-supplier-oriented model, 25–6 input suppliers, 29, 33–4, 36, 178, 196 and colocation, 77, 87, 98

Intel -in Costa Rica, 7–8, 135, 136, 145, 149, 157, 161, 167, 183–4, 197, 218n27, 219n28, 220n41 -in Ireland, 8, 49, 62–3, 71, 76, 102, 119, 130 Inter American Development Bank (IADB), 13, 161–2 and agreement with Costa Rica, 222n7 as source of funding, 203 International Business Linkage Division, Enterprise Ireland, 98, 132 International Financial Services Center (IFSC), 49–50 internationalization process by smaller companies, 156 mobile management class in, 157 International Monetary Fund (IMF), market liberalization and, 95, 140 International Organization for Standardization (ISO), 30 derivation of acronym, 209n15 ISO 9000 certification, 30, 31, 105, 117 quality standards, 177 Iona Technologies, 91, 110, 125 Ireland, 7 allocation of EU funds to, 66 contract manufacturers in, 50, 100–1, 190 cost of doing business in, 213n48 decreased competitiveness of, 103 development policies in, 45–9, 64, 66, 81–2, 190 distributional inequalities in, 79, 133 economic data on, 48 economic growth of, 45, 51 education in, 22, 67–70, 68, 126, 132 emigration policies, 81 employment in, 25, 47, 58, 94–5, 212n38 engineer supply in, 68–9, 110, 191, 212n37

INDEX

exports, manufactured, 53, 211n18 external factors influencing, 61–6 FDI inflows in, 49–50 GNI/GDP ratio, 45, 51–2, 52, 218n15 grants, cash, in 72–4, 214n12 human capital spillovers in, 110–20, 190, 191 inadequate public services in, 79 ICT, goals for in, 126, 129 infrastructure of, 22, 69–70, 78–9, 126, 132, 202 investment of abroad, 124–5 as labor surplus economy, 69 linkage capability, indigenous, in, 7, 93–9, 190–1 location-specific assets of, 42–3, 66–75, 156, 189–90 low tax rates in, 57 medical device production in, 43, 48–9, 103–8 openness to trade in, 52–7 pay agreements, national, in, 60, 67, 70, 79–80 plant closures in, 41 R&D-based development strategy in, 7, 81–2, 126–7, 127 small size of economy of, 109 software industry in, 90–1, 110, 118, 191, 214n4 youthful population of, 213n39 see also electronics industry, Ireland; government policies, Ireland Irish Council for Science, Technology and Innovation, 216n32 Irish Economy Expenditure Survey (IEES), 86, 88, 214n2 Irish Republican Army (IRA), 213n41 Ishii, Jun, 32 ISI, see import-substituting industrialization Jacobson, David, 90 Jenkins, Mauricio, 181, 223n8

245

Kaldor, Nicholas, 215n14 Kaminski, Bartlomiej, 33 Kearns, Allan, 120 Kelly, Daragh, 87, 91, 93 Kenney, Martin, 34, 37 Keogh, Elaine, 103 Ketlhoehn, Niels W., 140, 167 Kirkland, Richard I., Jr., 61, 211n25 Klein, Michael, 17 knowledge production of, 13 spillovers, linkages and, 91–3, 119–20 transfer of, 26, 92, 92, 111 knowledge-based assets -advancement of indigenous, 2, 6, 11, 15–17, 37–9, 173, 190, 196–99 -in Costa Rica, 182–84 -in Ireland, 83–133 Kokko, Ari, 27, 29 Konings, Jozef, 27 labor force participation, female, 67, 69, 112–13 labor supply, 7, 67, 69, 77, 81, 112, 113–14, 114, 193, 197 Lall, Sanjaya, 12, 207nn1, 3 Lane, Phillip, 70, 71 Larrain, Felipe B., 162, 181, 223n8 Larudee, Mehrene, 182 Latin America cluster development in, 200–1 linkages and linkage capability in, 180–82 Latin American Center for Competitiveness and Sustainable Development, 168 Layte, Richard, 79 leapfrogging, 17, 83, 84 Lee, Hau, 37 Lester, Richard K., 32, 34, 37, 100 linkage capability, 28–31, 33, 93–110, 174, 177, 180–2, 190, 191 development, 29, 84–5, 125–6

246

INDEX

linkage—continued effect, backward, 26–7, 88, 91–2, 173, 174–80, 185, 197–8, 209n17 and firm size, 102 formation, industrial, 28, 98 formation, proactive policies for, 94–9, 215n13 with indigenous companies, 83, 84–110, 205 intensification of, 85, 87–8 and knowledge spillovers, 91–3, 119–20 limited, 110–11, 135 local, and preferred input suppliers, 33–4 opportunities for, in medical device industry, 130–1 promotion, political resistance to, 95 location-specific assets -basic, 3, 9, 17, 21, 42–3, 192–3, 202 -in Costa Rica, 158–64 -in Ireland, 66–75, 78, 189–90 Lopez-Calva, Luis F., 162 Louri, Helen, 28, 119, 120, 215nn24, 25 Lyons, Declan, 37, 89, 96, 121 Maastricht, Treaty of, 64, 212n34 McAlesse, Dermot, 102 MacCarthaigh, Sean, 101 McDonald, Donogh, 102 McDowell, Andrew, 49, 124 McGowan, Kieran, 71 Machado, Roberto, 140 McKinsey Report, 18–19 MacSharry, Ray, 71 Madge Networks, 50 manufacturing sector -Costa Rica: age of machines in, 178, 178; data for, 219n36; as declining source of employment, 151–4, 153; dualism in, 135, 150–5; export intensity of, 151–4, 154; free zone, differences from non-free zone in, 151–4; imported input

share in, 174–6, 175; medical devices, 165–6, 170, 187, 200; output, types of, 218n21; structural change in, 53, 54–6, 57, 149–50, 155; training systems in, 160, 182, 182 -Ireland: exports and export intensity in, 123; gender composition of assembly production in, 112–13; training and R&D in, 112–16, 122–4; see also electronics industry, Ireland; medical device industry Marcusen, James, 29, 120, 208n9 market domestic, limited size of, 200 income convergence dependent on, 203 international, expertise needed for, 131–2 liberalization of, as development strategy, 95 market failures, 9, 29, 98, 174, 178, 185, 200, 201 Marshall Aid, 64–5, 65, 202 Mataloni, Raymond, Jr., 181 medical device industry characteristics of, 131, 195 cluster development of, in Costa Rica, 165–66, 170, 187, 200 difficulties of standardizing in, 107 domestic sourcing in, 104 employment composition of FOEs in, 114–15 FDA regulations in, 105, 106, 215n17 fragmented supplier base of, in Ireland, 105–6 high productivity growth of, 131 investing in Ireland, 43, 48–9, 103–8 labor costs in, 160 linkage opportunities in, 130–1 need for proximity of input producers in, 107–8 organization of global value chains in, 107

INDEX

partnering with biotech firms, 187, 216n33 research intensity of, 131 use of contract manufacturers in, 106, 107 medical devices, validation of, 105, 106 medical instruments, see medical devices Mexico, 165, 181, 195, 215n19 Ministry for Exports (MINEX), Costa Rica, 142 Monge Administration, Costa Rica, 138, 141 Monge Gonzalez, Ricardo, 145, 148, 149, 183 Morley, Samuel, 140 Mortimore, Michael, 143, 201, 218nn21, 22, 222nn62, 5 Mostek, 48, 68 Motta, Massimo, 27 Mottiar, Ziene, 90 Multi Fiber Agreement, 186 Murray Brown, John, 129 National Development Plan, Ireland, 64, 66, 132 nationality and plant closure, link between, 41 National Liberation Party (PLN), 217n7, 13 National Linkage Program (NLP), Ireland, 93, 94, 96, 98, 100, 132, 198 National Loan Commission for Education, 220n44 National Science Foundation, 128 Naughton, Barry, 32 Navaretti, Giorgio Barba, 33, 209n16 Newlon, T. Scott, 19 niche production, 110, 184, 191 Nolan, Brian, 79 North American Free Trade Agreement (NAFTA), 186, 195, 203 Ocampo, José Antonio, 30 O’Driscoll, Eoin, 79

247

O’Driscoll Report, 131, 132, 198 O’Farrell, P.N., 102 O’Gorman, Colm, 101, 118 O’Hearn, Dennis, 47, 111 O’Malley, Eoin, 63, 96, 98, 101, 118 Omar Denge Foundation, 159 “Opportunity and Fortress Europe,” 75, 189 Organization for Economic Cooperation and Development (OECD), 12, 50, 51 O’Riain, Sean, 90–1, 95 original brand name manufacturers (OBMs), 26 original design manufacturers (ODMs), 26 original equipment manufacturers (OEMs), 26, 100, 102 O’Sullivan, Mary, 94, 111, 112 outsourcing, 100, 102–3, 181, 187 ownership, nationality of, 39–42 Padilla Pérez, Ramón, 177, 181 path dependency, 3, 9, 21, 45–6, 193 Pearson, Ruth, 112 Pereira, Sonia C., 18, 27, 208n11 Peres, Wilson, 201 Pettinato, Stefano, 140 Pietrobelli, Carlo, 200 plant closure and nationality, link between, 41 Porter, Michael E., 13, 140, 167, 168 production assembly-type, gender distribution in, 112–13 costs, 35 customized, 104 externalized, 32, 99 high-tech, bifurcation of skills in, 198 indigenous, role of, 25–7 internalized, 32, 33, 99, 105 internationalization of, 194 of knowledge, 13 offshore, 35 standardization of, 30, 81, 89, 90, 107, 112

248

INDEX

production networks, 5, 6, 132 production processes collaboration in, 92–3 de-verticalized, 5, 99 fragmentation of, 1 productivity growth, 13, 14, 81, 131, 190 profit repatriation of TNCs in Ireland, 51, 59, 219n31 Programa Impulso, 168, 200 Programa MIL, 222n6 Promotora de Comercio Exterior (PROCOMER), 164–5, 179–80, 219n36, 221nn56, 57 quality control, importance of, 30, 150 quality standards, electronics industry, 117 Quality System Regulation, 105 quota system, in textiles and clothing imports, 172 R&D, see research and development Rabellotti, Roberta, 200 Regional Technical Colleges, development of, 67 Reich, Robert, 39–40 research and development (R&D) as basis for growth, 81–2, 132, 190 importance of expenditures for, 12 in IOE manufacturing sector, 122–4 Irish expenditures for, 112–16, 126–7, 127 level of engagement in, 115–16, 116 location of, in global value chains, 40, 41, 42, 99 tax credits for, 204 Roberts, Bill, 36, 107 Robles, Edgar, 168 Rodríguez administration, Costa Rica, 221n53 Rodríguez-Clare, Andrés, 26, 162, 168, 219n35 Rodrik, Dani, 201 Rolfe, Robert J., 159 Ronde, Thomas, 27

Roper, Stephen, 26, 92–3, 109 Rosales Tijerino, Julio, 145, 148, 149, 183 Rossi, José, 167 Ruane, Frances, 98, 102, 119, 120, 215nn24, 27 Sánchez-Ancochea, Diego, 202, 221nn52, 53 Sanmina-SCI, 35, 36, 100, 101–2 Science Foundation Ireland (SFI), 81, 126, 128–9, 132 Single European Act of 1986, 59, 61, 75, 189 Sjoeholm, Frederik, 27, 28, 29, 41 skill requirements, bifurcation of, in Ireland, 111 Slaughter, Matthew, 18, 27, 181, 208n11 SLCs, see small latecomers small country size advantages of, 118–19 industrial upgrading and technological capabilities, 11–17 as limitation on competitiveness, 180 small latecomers (SLCs) capabilities of, 2–3 creation of competitive environment in, 38–9 globalization and, 3–6 immobility of indigenous companies in, 41 imperative use of financial incentives by, 19–20 internal factors and policies in, 9 market failures in, 9, 29, 98, 174, 178, 185, 200, 201 multiagency cooperation in, 200 need for education and infrastructure improvement by, 202 need for proactive policies in, 185, 199 scale of operations in, 41 Smarzynska, Beata, 33

INDEX

Smyth, Emer, 67 Smyth, Jamie, 101 software industry, see Costa Rica; Ireland Solectron, 35, 36, 100, 101 Song, Jung-A., 204 sourcing, domestic, TNCs’ interest in, 33, 99–100 in Costa Rica, 175, 176–7, 187 in Ireland, 84–5, 86, 88–91, 99–108, 213–14n1 reasons for limitations on, 108–9, 109 South Korea development experience of, 168 tax incentives in, 204 Spar, Debora, 159, 162, 167, 218n27, 220nn41, 42 spillovers in Costa Rica, 174 domestic absorptive capacity for, 29 FDI, limitations of econometric studies of, 7 human capital, 84 government intervention in, 30–1 knowledge, 13, 14, 91–3, 116–19 productivity, 27–8, 119–20, 208n12 in SLC economy, 195–6 training, 111, 112, 182 Special Export Regimes, 147, 147, 149, 218n20, 219n36 Standing, Guy, 113 Strobl, Eric, 27, 63, 73, 97, 119, 120, 215nn24, 25 Sturgeon, Timothy, 32, 34, 37, 100 sub-supply sector, 34, 87, 89–90, 121 Swann, Christopher, 19 Taiwan, local supplies to TNCs in, 181 tax collection, need for improvement of, Costa Rica, 171 credits, 138–40, 163, 204 and government revenue as share of GDP, 60

249

incentives, implications of, 24 policies and FDI, 57–9 rates, corporate, 20, 24, 57, 72, 80, 164, 203–4 reform, effective, Costa Rica, 187 remission, on profits, Ireland, 210n4 revenues, Ireland, 58–9 see also incentives Taylor, Christopher T., 17, 208n7 Taylor, Cliff, 103 technological capabilities development of indigenous, 6, 26, 27 gap in, in Costa Rica, 29, 40 industrial upgrading and small country size, 11–17 Telesis Report, 72, 74, 94, 98, 132, 173 Temporary Admission Regime, 140 timing differences, between Ireland and Costa Rica, 199 TNCs, see transnational corporations Tomaney, Rolf, 74, 94 trade international, 3, 32 intra-TNC, 32–33 rules, changes in, 195 world, by factor intensity and demand, 14 training in Costa Rica, 160, 180, 182 high-tech FDI and, 27, 197 in Irish manufacturing sector, 112–16, 122–4 spillovers in, 182 transfer pricing, 50–2, 59, 122, 183 transnational corporations (TNCs) affiliate decision-makers, nationality of, 40, 42 coexistence of minority and majority owned, 215–16n28 core competencies of, 32, 105 cross-border acquisitions of, 3–5 effect of entry of into host country, 209n9 as efficiency seeking, 18

250

INDEX

transnational corporations—continued global value chains of, 31–34, 196, 209n23 global value chains of, in electronics industry, 31–4 government response to skills demands of, 67–8 influence of, on business practices, 83 influence of, on educational systems, 119, 159, 198 needs of and location-specific assets, 75, 158 net contribution of, to Irish tax revenue, 59 relocation of, 204 size of, 187–88 skill requirements of production in, 111–12 sourcing behavior of, 84–5, 99–100, 176–77, 181 special incentives for, 18–19 strategic needs of, 2–3, 93–110, 194–6 tax rates for, 20, 24, 57, 72, 80, 164, 203–4 as training grounds for local entrepreneurs, 27, 121, 130, 191, 198 Udur, Ali, 119, 120, 215nn24, 27 unemployment rate, 52, 67, 70, 94–5, 125 United Fruit Company, 137 United Kingdom, spillover effect in, 27, 29 United Nations Conference on Trade and Development (UNCTAD), 5, 17, 32, 37, 50, 144 U.S. Agency for International Development (U.S.-AID) in Costa Rica, 138, 140–2, 161, 162, 166 U.S. Bureau of Economic Analysis (BEA), 40

United States–Caribbean Basin Trade Partnership Act, 142 U.S. Food and Drug Administration (FDA), 99–100, 104–5, 195, 215n17 U.S. tariff system, 35 regulation 807 of, 142–3 Urrutia-Alvarez, Ruth, 177 validation, FDA, 105, 106, 195 value chains incentives for moving up, 75, 81, 205 internalized processes in, 102 reorganization of, 1, 93–4 segmentation of, 31–2; see also global value chains Venables, Anthony, 29, 33, 120, 208n9, 209n16 vendor management inventory (VMI), 102 Verbatim (word processing programs), 48, 68 vertical specialization, 32, 99 Wade, Robert, 168 Walsh, Brendan, 52, 69, 210nn9, 14 Wang Laboratories, 48, 68 Washington Consensus, 46, 95, 138, 140, 181, 195, 201, 217n9 Whelan, Christopher, 79 White, Padraic, 68, 71, 72 Wilson, Bruce, 138 Wilson, Tamar Diana, 112 Wong, Poh-Kam, 28, 37 Woodward, Douglas, 159 World Bank, 33, 40, 95, 140 World Trade Organization (WTO), 16 rule on elimination of preferential tax treatment, 170, 203, 221n55 Yi, Kei Mu, 32 Yusuf, Shahid, 26, 32, 34 Zamora, Ronney, 143, 218nn21, 22, 222nn62, 5