Multinationals and Foreign Investment in Economic Development

Multinationals and Foreign Investment in Economic Development Edited by

Edward M. Graham

Multinationals and Foreign Investment in Economic Development This is IEA conference volume no. 141

This page intentionally left blank

Multinationals and Foreign Investment in Economic Development Edited by

Edward M. Graham Institute of International Economics, Washington, DC, USA

in association with INTERNATIONAL ECONOMIC ASSOCIATION

© International Economic Association 2005 All rights reserved. No reproduction, copy or transmission of this publication may be made without written permission. No paragraph of this publication may be reproduced, copied or transmitted save with written permission or in accordance with the provisions of the Copyright, Designs and Patents Act 1988, or under the terms of any licence permitting limited copying issued by the Copyright Licensing Agency, 90 Tottenham Court Road, London W1T 4LP. Any person who does any unauthorized act in relation to this publication may be liable to criminal prosecution and civil claims for damages. The authors have asserted their rights to be identified as the authors of this work in accordance with the Copyright, Designs and Patents Act 1988. First published in 2005 by PALGRAVE MACMILLAN Houndmills, Basingstoke, Hampshire RG21 6XS and 175 Fifth Avenue, New York, N.Y. 10010 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 13: 978–1–4039–4940–0 ISBN 10: 1–4039–4940–9 This book is printed on paper suitable for recycling and made from fully managed and sustained forest sources. A catalogue record for this book is available from the British Library. Library of Congress Cataloging-in-Publication Data International Economic Association. World Congress (13th : 2002 : Lisbon, Portugal) Multinationals and foreign investment in economic development / edited by Edward M. Graham. p. cm. “Selected from papers presented at the Thirteenth Congress of the International Economic Association, held in Lisbon, Portugal, in September 2002”–P.. Includes bibliographical references. ISBN 1–4039–4940–9 1. Investments, Foreign – Developing countries – Congresses. 2. International business enterprises – Developing countries – Congresses. 3. Economic development – Finance – Congresses. I. Graham, Edward M. (Edward Montgomery), 1944– II. International Economic Association. III. Title. HG5993.I572 2002 332.67
3
091724—dc22 2004059167 10 9 8 7 6 5 4 3 2 1 14 13 12 11 10 09 08 07 06 05 Printed and bound in Great Britain by Antony Rowe Ltd, Chippenham and Eastbourne

Contents The International Economic Association

vii

Acknowledgements

ix

List of Contributors

xii

List of Abbreviations and Acronyms

xiii

1 Introduction: Foreign Direct Investment in Developing Countries – Where Do We Now Stand? Edward M. Graham

1

2 On Intra-firm Trade and Multinationals: Offshoring and Foreign Outsourcing in Manufacturing Ashok Deo Bardhan and Dwight Jaffee

26

3 Foreign Direct Investment, Externalities and Economic Growth in Developing Countries: Some Empirical Explorations Nagesh Kumar and Jaya Prakash Pradhan 4 Technological Spillovers and Export-platform FDI Kjetil Bjorvatn and Carsten Eckel 5 Foreign Direct Investment in South Asia: Impact on Economic Growth and Local Investment Pradeep Agrawal

42

85

94

6 The Internationalization of Korean Firms: Strategic Interaction and Tariff-jumping when Quality Matters Bénédicte Coestier and Serge Perrin

119

7 Foreign Direct Investment, Trade and Regional Integration in Mercosur Marta Castilho and Soledad Zignago

145

8 The Effect of Exchange-rate Uncertainty on Foreign Direct Investment in the United Kingdom Matteo Iannizzotto and Nigel J. Miller

163

v

vi Contents

9 The Role of Foreign Direct Investment and Natural Resources in Economic Development José De Gregorio

179

10 Multinationals and Foreign Direct Investment in India and China Subrata Gupta

198

11 Portuguese Investment in Brazil – the Challenges of an Iberian Logic Cézar Miranda Guedes and Mario Gomez Olivares

212

The International Economic Association A non-profit organization with purely scientific aims, the International Economic Association (IEA) was founded in 1950. It is a federation of some sixty national economic associations in all parts of the world. Its basic purpose is the development of economics as an intellectual discipline, recognizing a diversity of problems, systems and values in the world and taking note of methodological diversities. The IEA has, since its creation, sought to fulfil that purpose by promoting mutual understanding among economists through the organization of scientific meetings and common research programmes, and by means of publications on problems of fundamental as well as of current importance. Deriving from its long concern to assure professional contacts between East and West and North and South, the IEA pays special attention to issues of economies in systemic transition and in the course of development. During its nearly fifty years of existence, it has organized more than a hundred round-table conferences for specialists on topics ranging from fundamental theories to methods and tools of analysis and major problems of the present-day world. Participation in round tables is at the invitation of a specialist programme committee, but thirteen triennial World Congresses have regularly attracted the participation of individual economists from all over the world. The Association is governed by a Council, composed of representatives of all member associations, and by a fifteen-member Executive Committee which is elected by the Council. The Executive Committee (2002–5) at the time of the Lisbon Congress was: President: Vice-President: Treasurer: Past President: President-elect: Other members:

Professor János Kornai, Hungary Professor Bina Agarwal, India Professor Jacob Frenkel, Israel Professor Robert Solow, USA Professor Guillermo Calvo, Argentina Professor Maria Augusztinovics, Hungary Professor Eliana Cardoso, World Bank Professor Duardo Engel, Chile Professor Heba Handoussa, Egypt Professor Michael Hoel, Norway vii

viii The International Economic Association

Professor Andreu Mas Colell, Spain Professor Kotaro Suzumura, Japan Professor Alessandro Vercelli, Italy Advisers: Professor Fiorella Kostoris Padoa Schioppa, Italy Professor Vitor Constancio, Portugal Secretary-General: Professor Jean-Paul Fitoussi, France General Editor: Professor Michael Kaser, UK Sir Austin Robinson was an active Adviser on the publication of IEA Conference proceedings from 1954 until his final short illness in 1993. The Association has also been fortunate in having secured many outstanding economists to serve as President: Gottfried Haberler (1950–3), Howard S. Ellis (1953–6), Erik Lindahl (1956–9), E. A. G. Robinson (1959–62), Ugo Papi (1962–5), Paul A. Samuelson (1965–8). Erik Lundberg (1968–71), Fritz Machlup (1971–4), Edmund Malinvaud (1974–7), Shigeto Tsuru (1977–80), Victor L. Urquidi (1980–3), Kenneth J. Arrow (1983–6), Amartya Sen (1986–9), Anthony B. Atkinson (1989–92), Michael Bruno (1992–5), Jacques Drèze (1995–9) and Robert M. Solow (1999–2002). The activities of the Association are mainly funded from the subscriptions of members and grants from a number of organizations. Support from UNESCO since the Association was founded, and from its International Social Science Council, is gratefully acknowledged, particularly for specific help for the Lisbon Congress.

Acknowledgements As Editor I wish to make a personal acknowledgement of gratitude to John L. Marshall, MD, who has kept me alive long enough to complete the task (and, one hopes longer) and to Professor Robert Solow, who stood by me as choice of editor even when it was not clear I could finish the job. ∗ ∗ ∗ The Congress was held from 9–13 September 2002 in the Centro Cultural de Belém, Lisbon, at the invitation of the Ordem dos Economistas de Portugal, and was attended by 1,100 registered participants. The Opening Session was addressed by the President of the Republic of Portugal, HE Senhor Jorge Sampaio, and by the newly-appointed Minister of Finance, HE Senhor Manuela Ferreira Leite; the IEA President, Professor Robert M. Solow, delivered a paper, ‘Is Fiscal Policy Possible? Is it Desirable?’ The programme comprised twenty invited lectures and three invited panels – on ‘Growth in Developing and Transition Economies’ (arranged by the Global Development Network); on ‘Poverty Dynamics and Insurance’ (organized by the European Development Research Network); and on ‘The Turkish Financial Crisis’ (prepared by the Turkish Economic Association). There were 198 contributed papers, a selection of which have been included with Invited Lectures in the four volumes of the Congress proceedings: Bina Agarwal and Alessandro Vercelli (eds) Psychology, Rationality and Economic Behaviour: Challenging Standard Assumptions; Alan V. Deardorff (ed.) The Past, Present and Future of the European Union; Edward Graham (ed.) Multinationals and Foreign Investment in Economic Development; Robert M. Solow (ed.) Structural Reform and Macroeconomic Policy. Studies generated by the Global Development Network are published in Gary McMahon and Lyn Squire (eds) Explaining Growth: A Global Research Project (IEA Conference Volume No. 137). ix

x Acknowledgements

The scientific responsibility for the selection of papers was in the hands of an International Programme Committee chaired by Robert Solow, with the following members: Bina Agarwal, India Maria Augusztinovics, Hungary Victor Becker, Argentina Miguel Beleza, Portugal Enrique Bour, Argentina

Gene Grossman, USA Seppo Honkapohja, Finland Peter Howitt, Canada Andrea Ichino, Italy Fiorella Kostoris Padoa Schioppa, Italy Juan Camilo Cardeñas, Colombia Valery Makarov, Russian Federation Elinana Cardoso, Brazil Andreu Mas-Colell, Spain Vitor Constâncio, Portugal Mustapha Nabli, Tunisia Vittorio Corbo, Chile Ademola Oyejide, Nigeria Jacques Drèze, Belgium Adrian Pagan, Australia Jean-Paul Fitoussi, France Luis Servén, USA Marc Flandreau, France José Silva Lopes, Portugal Augustin Fosu, Kenya António Simões Lopes, Portugal Jacob Frenkel, UK Hans-Werner Sinn, Germany Hans Gerbach, Germany Kotaro Suzumura, Japan A National Scientific and Organizing Committee was convened by the Ordem dos Economistas de Portugal, under the chairmanship of its President, António Simões Lopes, who, with Amilcar Theias, Carlos Queiroz and Luisa Ahrens Teixeira (Executive Director of Mundiconvenius) formed an Executive Committee: Mário Abreu José Freire de Sousa Luis Miguel Beleza José Silveira Godinho Daniel Bessa Manuela Ferreira Leite Miguel Cadilhe Emâni Rodrigues Lopes Teodora Cardoso Isabel Almeida Lopes Eduardo Catroga Manuel de Oliveira Marques Maria José Constâncio Manuela Morgado Vitor Constâncio Isabel Almeida Mota Vitor Pereira Dias José de Almeida Serra Erlânder Estrela Francisco Soares João Ferreira do Amaral The IEA is most grateful to the Ordem dos Economistas de Portugal; the Banco de Portugal; the Caixa Geral de Depósitos; the European Commission; the Fundação Calouste Gulbenkian; Portugal Telecom; and other sources in Portugal which generously ensured most of the funding of the Congress. Among forty-six other funders, mention must be made

Acknowledgements xi

particularly (in alphabetical order) of the Banca d’Italia; the Bank for International Settlements; the European Central Bank; UNESCO; and the World Bank. Cultural events were supported by the Fundação Calouste Gulbenkian for a concert at its Headquarters; by the Casino do Estoril for a Gala Dinner at the Casino; and by the SECIL Corporation for a Dinner for speakers at the Convento da Trindade. A Welcome Cocktail was offered on the opening evening at the Maritime Museum, Belém, and the publishers of the IEA conference volumes, Palgrave Macmillan, gave a reception on the second evening to commemorate the Fiftieth Anniversary of the series, hosted by Amanda Watkins and Pooja Talwar. The Instituto Vinho do Porto provided a lecture and a tasting of port wines. Day tours in the region around Lisbon were arranged in the three days after the Congress for participants and accompanying persons. The logistics of the Congress were handled efficiently by the staff of Multiconvenius, co-ordinated by Luisa Ahrens Teixeira, its Executive Director. The staff of the Ordem dos Economistas do Portugal furnished additional assistance under the management of Carlos Quiroz. The President of the IEA, Robert Solow, was Congress Editor. The IEA editorial team comprised Maureen Hadfield and Michael Kaser; the latter was responsible for the present volume.

List of Contributors Pradeep Agrawal, Institute of Economic Growth, Delhi, India Ashok Deo Bardhan, Haas School of Business, University of California, Berkeley, USA Kjetil Bjorvatn, Norwegian School of Economics and Business Administration, Bergen, Norway Marta Castilho, Universidade Federal Fluminense (UFF) and Institute for Applied Economic Research, Rio de Janeiro, Brazil Bénédicte Coestier, Université de Paris X, Nanterre, France Carsten Eckel, Universität Göttingen, Germany Cézar Miranda Guedes, Universidade Federal Rural de Rio de Janeiro, Brazil Edward M. Graham, Institute of International Economics, Washington, DC, and Columbia University, New York, USA José De Gregorio, Banco Central de Chile, Santiago, Chile Subrata Gupta, Jogamaya Devi College, Calcutta University, India Matteo Iannizzotto, University of Durham, UK Dwight Jaffee, Haas School of Business, University of California, Berkeley, USA Nagesh Kumar, Research and Information System for the Non-Aligned and Other Developing Countries, New Delhi, India Nigel J. Miller, University of York, UK Mario Gomez Olivares, Universidade Técnica de Lisboa, Portugal Serge Perrin, Agence Française de Devéloppement, Paris, France Jaya Prakash Pradhan, School of Social Sciences, Jawaharal Nehru University, New Delhi, India Soledad Zignago, Université de Paris I, France

xii

List of Abbreviations and Acronyms

ADR ASEAN BCB BEA CAN CEP CET CPI D–P ECLAC EIM EMU ETDZ EPZ EU FCCB FDI FERA FIAS FIE FIPB FTA FTAA GDR GDP GFCF GMM IADB INTAL

American Depository Receipt Association of South-East Asian Nations Banco Central do Brasil Bureau of Economic Analysis (USA) Andean Community Centro de Estudios para la Producción (Argentina) Common External Tariff (Mercosur) consumer price index Dixit–Pyndick Economic Commission for Latin America and the Caribbean European Investment Monitor database European Monetary Union Economic and Technical Development Zone (India) Export Processing Zone (India) European Union foreign-currency convertible bond foreign direct investment Foreign Exchange Regulation Act (India) Foreign Investment Advisory Service foreign-investment enterprise (China) Foreign Investment Promotion Board (India) free trade area Free Trade Area of the Americas Global Depository Receipt gross domestic product gross fixed capital formation generalized method of moments Inter-American Development Bank Instituto para la Integracion de America Latina y el Caribe (IADB) IV instrumental variable JV joint venture M&A mergers and acquisitions MIGA Multilateral Investment Guarantee Agency Mercosur Southern Cone Common Market xiii

xiv List of Abbreviations and Acronyms

MNE NAFTA NAICS NIC NRI OECD OLS PPP R&D RBI RIS

multinational enterprise North American Free Trade Area North American Industry Classification System newly-industrialized country non-resident Indian Organization for Economic Cooperation and Development ordinary least squares purchasing power parity research and development Reserve Bank of India Research and Information System for the Non-aligned and Other Developing Countries RL real lending rate SC Schwartz information criterion SD standard deviation SEZ Special Economic Zone SIA Secretariat of Industrial Approvals (India) TFP total factor productivity TRIMs Agreement on Trade-Related Investment Measures TRIPs Agreement on Trade-Related Intellectual Property UNCTAD United Nations Conference on Trade and Development USAID United States Agency for International Development VAR vector auto-regression VCR video-cassette recorder WTO World Trade Organization ZIP zero-inflated Poisson

1 Introduction: Foreign Direct Investment in Developing Countries – Where Do We Now Stand? Edward M. Graham Institute for International Economics, Washington, DC, USA

1

Introduction

This volume is about foreign direct investment in developing countries. The chapters were selected from papers presented at the Thirteenth Congress of the International Economic Association, held in Lisbon, Portugal in September 2002. To be included in this volume, a paper had to address some aspect of the topic, but some addressed it rather obliquely, and justification for their inclusion is noted below. The IEA papers on this topic could not have come at a more opportune time. The past decade has seen the study of foreign direct investment (FDI) and multinational enterprise enter increasingly into mainstream economics, especially mainstream econometrics. This has been, generally, quite good news for a veteran such as myself who has tried to work on the interface between economics and FDI for something like thirty years and who has found that, for much of this time, FDI simply was not taken very seriously by most economists. It is not entirely good news, however, in part because some of the new research fails to account fully for insights and discoveries by veterans who investigated FDI and multinational enterprises in the days before such study became ‘mainstream’. This has created a number of problems and shortcomings in the present body of work, and one thing that this introduction attempts to do is to identify these. The reader should realize, however, that I look upon the new work for the most part with approval; the contributions made by this work are substantial and likely to endure. My only intent in raising shortcomings is to suggest ways in which future research can continue to be useful. 1

2 Introduction: FDI in Developing Countries

One can guess that the new attitude of the economics profession is occasioned by the fact that, beginning in about 1985, world flows of FDI began to rise sharply, such that FDI simply became too large and important a phenomenon in the world economy to be ignored. Thus let us begin by reviewing some basic facts and figures regarding these flows, with special emphasis on that component of these flows going to developing nations. 1.1 On the rising importance of FDI and its importance to developing countries One might expect, or at least so on the basis of the classical theory of capital movements, that capital would flow from places where it is relatively abundant, and hence relatively cheap, to places where it is relatively scarce and hence where returns to capital are relatively high. However, this expectation is generally not borne out by foreign direct investment, which is one specific form of capital flow. Most FDI flows are not from developed nations having relatively high capital to labour ratios to developing nations having relatively low such ratios. Rather, the vast majority of FDI flows since 1985 have gone from one developed country to another developed country, such that, as a percentage of total such flows, little FDI has flowed from capital-rich countries to the developing world (see Table 1.1). One reason doubtless is that much of the enormous expansion of FDI since 1985 has resulted from cross-border mergers and acquisitions (M&A), whereby a firm in one advanced country acquires or merges with another firm in some other advanced country (again, see Table 1.1). Alas, however, this ‘reason’ in fact explains very little. This is because the phenomenon of cross-border merger and acquisition remains under-studied, so what motivates these transactions is not really well understood. This is unfortunate because this phenomenon has been the newest and arguably the most prominent feature of the much-discussed ‘globalization’ of the world economy that has taken place since the later 1980s. Even so, one finds that relatively little space, even in recent popular books on ‘globalization’, is devoted to this phenomenon (for example, Stiglitz, 2003; Wolf, 2004). But it is accurate to say that the huge ‘wave’ of cross-border mergers and acquisitions of recent times is, from an historical perspective, a first occurrence – that nothing like it has been witnessed in earlier history.1 Although large cross-border mergers and acquisitions occurred sporadically prior to the beginning of this wave, neither the numbers of transactions nor the size of the largest individual transactions of the past have been even remotely close to those of recent years.

Edward M. Graham 3 Table 1.1 Foreign direct investment inflows by category of FDI and by category of host nation (US$ billions or percentages)

Total FDI inflows all countries Cross-border M&A all countries Residual FDI all countries Total inflows developed countries Cross-border M&A developed countries Residual FDI developed countries Total inflows developing countries Cross-border M&A developing countries Residual FDI developing countries Developing country percentage share of world FDI inflow Developing country percentage share of world M&A Developing country percentage share of residual FDI

Average 1980–5

Average 1991–6

1997 1998

49.8

254.3

481.9 686.0 1079.1 1393.0 823.8 651.2

—

130.6

304.8 531.6

766.0 1143.8 594.0 369.8

—

123.7

177.1 154.4

313.1

37.2

154.6

269.7 472.3

824.6 1120.5 589.4 460.3

—

112.0

232.1 443.2

679.5 1056.1 496.2 307.8

—

42.6

12.6

91.5

—

15.9

—

75.6

25.3

36.0

40.1

27.9

21.2

17.7

25.4

24.9

—

12.2

22.0

12.1

9.7

6.2

14.4

12.0

—

61.0

71.3

70.3

49.6

70.4

53.7

41.8

37.6

1999

29.1

145.1

193.2 191.3

229.3

67.0

82.7

74.0

126.2 108.6

155.3

2000

2001 2002

249.2 229.8 281.4

64.4

93.2 152.5

246.1 209.4 162.1 70.6

85.8

44.5

175.5 123.6 117.6

Notes: (1) Cross border M&A is by country of seller (sale of assets in an M&A transaction to a foreign investor is, by definition, an inward FDI); (2) Residual FDI inflow is calculated as the difference between total FDI inflow and cross-border M&A; (3) ‘Developing countries’ and ‘Developed countries’ exclude Russia and Eastern Europe, which do not appear in this table. Sources: Data for 1991 and later: UNCTAD (2003), annex table B.1 (data on total FDI inflows), annex table B.7 (data on cross-border mergers and acquisitions, and author’s calculations (data on residual FDI and percentages). Data for 1980–5, UNCTAD (1992), annex table 1.

Past ‘waves’ of M&A transactions have taken place largely within the boundaries of one nation – the United States of America. In fact, we really have no good theory to explain even these past waves of M&A,2 and we certainly have no theory that would have predicted the huge volume of cross-border M&A transactions during the 1990s.

4 Introduction: FDI in Developing Countries

However, cross-border M&A among advanced nations is not the topic of this volume; perhaps a future IEA volume should be devoted to this subject. Let us simply note that even if cross-border M&A did dominate the FDI flows since the 1980s, these transactions have not been the whole show, nor have they been totally confined to the advanced countries. Our focus here is thus mostly on that portion of FDI that is not accounted for by such M&A. Indeed, the chapters of this volume do not really focus on M&A even where at least one of the merging parties is located in developing nations. Rather, most of the focus, even if in some chapters this focus is implicit, is on FDI generated by activity other than M&A. It is tempting to term this latter portion – the difference between reported FDI inflow and reported M&A transactions – ‘greenfields’ FDI. But, in fact, this difference includes reinvested earnings by incumbent affiliates of MNEs and not just investment in newly-created affiliates or facilities, where the latter is the standard definition of greenfields FDI. Alas, international data do not enable one to disaggregate FDI, even after subtracting that created by M&A, by greenfields investment and by retained earnings. Thus here we shall term the difference between reported FDI inflow and reported M&A transactions simply as ‘residual’ FDI. What the international data do reveal (see Table 1.1) is that the proportion of this residual flowing to developing countries is much higher than the proportion of total FDI going to these countries. It automatically follows, then, that M&A-generated FDI is a relatively smaller portion of total FDI to these countries than to advanced countries. Indeed, in most recent years (the exceptions are 1999 and 2002), the majority of the world’s residual FDI flows have been to developing nations. Given this, it is only a small leap to claim that the majority of greenfields FDI flows worldwide seems to have been going to these countries in recent years; this claim, to be substantiated, would only require that the ratio of greenfields investment to reinvested earnings is the same in developing countries as in developed ones, not a wholly implausible assumption, albeit not a verifiable one either. Thus, if one leaves cross-border mergers and acquisitions out of the story, the expectation that FDI flows mainly from developed to developing nations is borne out to a greater extent than if one includes these M&A. Alas, this latter none the less does not quite square with a classical explanation of capital flow: a disconcertingly high percentage of residual FDI still flows from developed nations to other developed nations. This is particularly evident if one looks at residual FDI flow per capita (see Tables 1.2 and 1.3): while in most recent years the percentage

5 Table 1.2 FDI flows to the developing countries, by region and by category of FDI, 1997–2002 (US$ billions) 1997

1998

1999

2000

Africa Total FDI inflow Cross-border M&A Residual

10.7 4.3 6.4

8.9 2.6 6.3

12.2 3.1 9.1

8.5 3.2 5.3

18.8 15.5 3.3

11.0 4.6 6.4

Middle East and Central Asia Total FDI inflow Cross-border M&A Residual

9.0 2.7 6.3

9.8 0.2 9.6

3.2 0.4 2.8

3.4 1.1 3.3

9.2 1.3 7.9

6.4 0.6 5.8

South, East and SE Asia Total FDI inflow Cross-border M&A Residual

100.1 18.6 81.5

90.1 15.8 74.3

105.3 28.4 76.9

138.7 21.1 117.6

97.6 22.1 75.5

88.6 16.8 71.8

Of which to China Total FDI inflow Cross-border M&A Residual

44.2 0.8 43.4

43.8 1.3 42.5

40.3 0.1 40.2

40.8 0.5 40.3

46.8 0.5 46.3

52.7 1.0 51.7

Remainder of South, East and SE Asia Total FDI inflow Cross-border M&A Residual

65.9 17.8 48.1

46.3 14.5 31.8

65.0 28.3 36.7

97.9 20.6 77.3

50.8 21.6 29.2

35.9 15.8 20.1

Latin America and Caribbean Total FDI inflow Cross-border M&A Residual

73.3 41.1 32.2

82.0 63.9 18.1

108.3 42.0 62.3

95.4 45.2 50.2

83.7 35.8 47.9

56.0 22.4 33.6

Of which to Brazil Total FDI inflow Cross-border M&A Residual

19.0 12.1 6.9

28.9 29.4 −0.5

28.6 9.4 19.2

32.8 23.0 9.8

22.5 7.0 15.5

16.6 5.9 10.7

Remainder of Latin America and Caribbean Total FDI inflow Cross-border M&A Residual

54.3 29.1 25.2

53.1 34.5 18.6

79.7 32.6 47.1

62.6 22.2 40.4

61.2 28.8 32.4

39.4 16.5 22.9

Source:

As Table 1.1.

2001

2002

Average residual FDI inflow, 1998–2002 (US$ billions)

Average total FDI inflow per capita, 1998–2002 (US$)

Average residual FDI inflow per capita, 1998–2002 (US$)

Population, 2000 (millions)

Average FDI inflow, 1998–2002 (US$ billions)

Developed countries

854

693.43

96.89

812.0

113.4

Developing countries

5779

207.64

136.38

35.9

23.6

Africa Sub-Saharan Africa Kenya South Africa

807 628 30 43

14.7 13.0 2.1 48.8

7.5 5.6 2.1 −43.5

Middle East and Central Asia

11.88 8.16 0.062 2.10

6.05 3.51 0.062 −1.87

327

6.41

5.67

19.6

18.5

3190 131 1017 1262

104.06 0.155 2.79 44.88

23.06 0.148 1.72 35.6

32.6 1.2 2.7 42.8

25.4 1.1 1.8 33.9

47 4.07 23 48

5.91 10.40 2.83 0.30

Nil 8.30 2.11 0.30

126.6 2,555.2 123.1 6.3

Nil 2,039.3 91.9 6.3

Latin America and Caribbean Argentina Brazil Bermuda Mexico

607 37 170 0.061 98

85.08 9.43 25.85 9.43 15.89

43.20 1.09 10.92 6.02 9.50

140.2 254.8 152.0 154,590.2 162.2

71.2 29.3 64.2 98,688.5 96.9

Eastern Europe and Russia

333

25.54

23.92

76.7

71.8

South, East and SE Asia Bangladesh India China Hong Kong SAR South Korea Singapore Malaysia Burma (Myanmar)

Sources: Population data from World Bank (2002), Table A21, or from national data sources. Other data: author’s calculations from FDI-related data from same sources as Table 1.1 and population data from this table.

6

Table 1.3 Average annual total FDI flow to developing countries, average annual residual FDI to them, and average annual per capita total and residual flows, 1998–2002

Edward M. Graham 7

of worldwide residual FDI flows to developing nations is indeed higher than to developed countries, the per capita residual flows to developing nations are much lower than to the developed ones. Moreover, of FDI going to developing countries, the flows have been very uneven in recent years, with some countries and regions receiving much more FDI than others, even after adjusting for populations. In particular, Africa, the Middle East and Central Asia, all of which are capital-poor regions, have received very low shares of total or even residual FDI flows in recent times. Also, although not shown in the tables, and with the exception of a few relatively developed nations within these regions such as South Africa, an unusually high proportion of the FDI that has gone to Africa and the Middle East has been in the oil and gas sectors. The main reason for this has been simply the availability of these particular natural resources. These sectors are ones that many development experts believe do not generally foster any spillover effect that might enhance local economic development (but on FDI in natural resources more generally, see Chapter 9 by De Gregorio in this volume, discussed later in this chapter). Latin America, by contrast, a region with a total population much less than that of all of Africa, the Middle East and Central Asia, has received relatively large amounts of FDI. In Latin America, a much higher percentage of FDI has been in the form of cross-border mergers and acquisitions than in other parts of the developing world. This cross-border M&A activity in Latin America largely follows from private foreign participation in the privatization of formerly state-held assets. By 2004, this privatization process had wound down, because most public assets that might have been candidates for privatization had already been sold. One result has been a significant reduction in FDI flows into Latin America in recent years, because of decreased M&A. However, it is not Latin America that has received the lion’s share of the FDI to the developing world in recent years, but rather Asia (excluding Central Asia), with China leading the list of individual recipient countries. Arguably, however, some of the FDI flows to this region as listed in Table 1.2 might be misclassified, because certain of the recipient countries or entities arguably should no longer be properly classified as developing countries – for example, Singapore, Hong Kong, Taiwan and South Korea. Also, some of the FDI into this region may be doublecounted (for example, FDI listed as flowing into Hong Kong might be to a headquarters operation that reinvests the funds in mainland China;

8 Introduction: FDI in Developing Countries

both the flows to Hong Kong and China are listed as ‘inward FDI’ while, in reality, Hong Kong is only an intermediary).3 However, even after making adjustments, East Asia has received the lion’s share of FDI to developing nations. Even so, from the tables, it is evident that, within this broad region, the flows of FDI have been very uneven. For example, FDI flow to India has been much less than to China, even after taking into account China’s larger population (see Chapter 5 by Pradeep Agrawal and Chapter 10 by Subrata Gupta in this volume, on India). Moreover, the distribution of FDI within some major nations is also uneven; in China, for example, most of inward FDI has gone to four coastal provinces, leaving much of the interior of China, even areas that are densely populated, without significant amounts of FDI (Lemoine and Ünal-Kesenci, 2004). 1.2

Why does not more FDI flow to developing countries?

The expectation drawn from classical theory of capital movement that capital generally will flow from capital-rich to capital-poor countries is thus borne out not particularly well in the case of foreign direct investment flow. Why is this? To begin with, this expectation is, of course, based on certain ceteris paribus assumptions, and one reason why the expectation is not fully borne out certainly must be that, indeed, not everything is equal. In particular, the assumption is made in the case of classical theory that technological capabilities of producers around the world are more or less the same, such that absolute efficiencies (and hence total factor productivities) are the same for most countries, or at least not wildly different. Also, the assumption is made that markets, both for final goods and services and for factors are competitive, or at least ‘close enough’ to competitive that competitive market conditions can be assumed. But the total factor productivity (TFP) of the typical developing country is clearly much lower than that of the typical developed country, reflecting differences in the technological sophistication of individual producers in these differing groups of countries, and manifesting itself in large wage differences between the two groups of countries, caused surely by differences in marginal labour productivity that also must be rooted in technological differences (see, for example, discussion by Romer, 1990, 1994). Moreover, beginning with Stephen Hymer’s doctoral dissertation at MIT (completed in 1960 but published as Hymer, 1977), it has been generally recognized that FDI occurs in sectors or markets characterized by imperfect competition. This is one of those insights noted above

Edward M. Graham 9

that is not always taken into account by recent research and can wreak havoc with econometric specification when this specification assumes that something like perfect competition prevails. This is a matter to which we return below. But does, in these instances, ‘not ceteris paribus’ explain why developing countries receive less of the FDI flow than might be expected? Imperfect competition might reign in those sectors in which FDI is prominent, but how does this account for lower than expected FDI in developing nations? I do not see a clear answer to these questions. Indeed, in literature dating back to the 1970s, numerous writers from developing countries worried that the market power of multinational firms would lead to these firms dominating their economies, such that prosperity, if it came, would be clouded by foreign control over these economies. This led, in many countries, to the adoption of policies to regulate or even, for some sectors at least, to ban foreign direct investment. Such policies could, of course, in and of themselves greatly reduce the amount of foreign direct investment that entered the relevant nations, relative to the potential of the nation to draw such investment. In this volume, in Chapter 10, Subrata Gupta argues that this has in fact been the case for India in particular. In support of this hypothesis that ‘policy does matter’, Gupta points out that, following the relaxation of many restrictions on FDI in 1991, China experienced a huge surge of inward foreign direct investment. Moreover, Chapter 5 by Pradeep Agrawal reinforces Gupta’s claims that ‘policy matters’ with hard econometric evidence; after relaxation of some of the direct and indirect official restrictions on FDI following 1991, flows of FDI to India increased markedly. There were, however, across-the-board policy changes in India that, among other things, caused India’s growth rate to rise. Thus it is a little difficult to sort out whether increased FDI to India in the years following 1991 was a response in a change in official policy towards FDI per se, or a response to an overall more favourable economic situation. Indeed, it is also true that, since 1991, quite a number of developing nations have relaxed or removed regulations and restrictions on this investment, and some of these liberalizations went far deeper than those in China or India (see, for example, Brewer and Young, 2001, ch. 5). But few of these nations have experienced a rise in FDI inflow of anything like the magnitude of those to China, or even the less pronounced rise to India. One is tempted to conclude that the complex relationship between government policy and the ability of a country to attract FDI is simply not well understood. In particular, although it certainly is clear

10 Introduction: FDI in Developing Countries

that country policy can discourage FDI, it is not so clear that an open policy is sufficient to attract this investment. Moreover, the fact that total factor productivity and technological capabilities are not at as high levels in developing countries as in developed countries might, if anything, reinforce an expectation that FDI would flow mainly from the latter to the former countries. I would argue that this would be true in particular if, as is now fashionable, technology is treated as an ‘endogenous’ factor – that is, it is specific to firms rather than being just a part of the ‘economic ether’ in which all the world’s firms operate. If technology is specific to firms, this is presumably because there is a return available to a firm from creating a technology that is not held by rival firms (and hence also an incentive for the innovator firm to withhold the technology from rivals or potential rivals). And, if technology were generally to be relatively scarce in developing nations, there would be some incentive for these firms to exploit their firm-specific technologies in markets in these countries where, plausibly, the return to the technology would entail some premium not available in developed nations. Of course, such ‘exploitation’ could come about via the export of products embodying the technology into these countries; but the exploitation could also possibly come about through the creation of local operations within those same countries – that is, via FDI. Indeed, there is an evolving theoretical literature on the choice between FDI and exporting where a firm does hold some sort of proprietary, firm-specific technology; the pioneering article on this is Markusen (1984). One reason why this might not happen is a failure of these nations to provide adequate protection to intellectual property, so that multinational investors are loath to transfer their proprietary technologies to these countries. This unwillingness could also account for some of the negative evidence regarding ‘technology spillovers’ discussed below. In fact, one basis on which the World Trade Organization’s (WTO) Agreement on Trade-Related Intellectual Property Rights (TRIPs) was ‘sold’ to developing nations was that implementation of the obligations of this agreement by these nations would lead to greater FDI inflow. To date, this does not really seem to have occurred but, also, most developing nations have not yet fully implemented the TRIPs obligations (and, indeed, under the Agreement, these nations were granted a lengthy ‘grace’ period before they were required to do so) – see, on this set of issues, Maskus (2000a, 2000b). This volume does not, however, take us into this issue. Rather, the point to be established here is simply that assumptions that technology is ‘endogenous’, and that producers of goods and services in developing

Edward M. Graham 11

countries do not, in general, operate at levels of total factor productivity as high as potentially rival firms in developed countries, do not serve adequately to explain directly why less of the world’s recent FDI flows is to developing countries than might be expected under a classic model of international capital flow. However, this particular ‘not everything is equal’ factor might have some significant indirect explanatory power. In particular, technological differentials between developed and developing nations, and the relative paucity of technological capabilities among the latter, might explain why in much of the developing world’s economic growth is lacklustre (on this, see Solow, 1994; and Romer, 1994). There is long-standing empirical evidence that a nation’s economic growth and that nation’s attractiveness as a locus of FDI-generated activity are associated with one another – that is, that fast-growing nations attract more FDI, even after controlling for other factors, than do slow-growing nations (for an early study, see Kobrin, 1976). Of course, no causality is necessarily implied by such an association – that is, it is not immediately clear whether relatively fast growth causes FDI to flow to the nation in question, or the FDI causes the growth. Indeed, whether the latter is true or not is one issue that is addressed in this volume, and there is more on this shortly. For the moment, let us simply note that, if relatively low technological capabilities in at least certain developing countries are a causal factor of relatively low growth, then it is possible that the surfeit of these capabilities is a reason for these countries failing to draw FDI. If so, this might create something of a vicious circle, or at least so if we posit that the association between FDI and growth embodies a ‘two-way’ causal relationship: greater FDI could induce greater growth (indeed, as argued in this and other chapters in this volume, in part by raising the technological capabilities of the relevant nations), but low growth deters FDI. There is also another indirect explanation based at least in part on technical capability. This derives from observations associated with Raymond Vernon that FDI is concentrated in industries that themselves are technology-intensive, in the sense that research and development (R&D) as a percentage of value-added of the relevant products is higher than average (Vernon, 1966, 1971). On this, around the time that Vernon was writing, FDI was concentrated in the manufacturing sector, and the main home country of this investment was the United States. Thus, in particular, Vernon’s observations might not hold for the service sectors that have since witnessed considerable FDI-related activity. However, FDI in developing nations is angled more towards manufacturing than is

12 Introduction: FDI in Developing Countries

FDI in high-income nations, and so this insight might still hold for the former. Why might this deter FDI from flowing to developing nations? There are two reasons. First, demand for the products associated with technology-intensive industries seems to be highly income-elastic, so demand in lower-income countries is relatively low (a lower percentage of national income is expended on them than in higher-income countries). Second, if the manufacture of these products requires high human capital, as seems plausible, such capital might simply be lacking in many developing nations. Thus, FDI in developing nations is deterred because (i) demand for the products produced by firms that typically engage in FDI is lacking; and/or (ii) because a factor of production needed by these firms is scarce in these nations and, thus, by the Heckscher–Ohlin theorem, these nations would not possess comparative advantage in these sectors. It would therefore not make sense for multinational firms to invest in these nations; or, in other words, there might be intrinsic disadvantages possessed by these nations in the sectors in which FDI is concentrated. Of course, this assumes causality that FDI is concentrated in certain sectors that are R&D-intensive, because R&D intensity leads to firms’ becoming multinational. While there is a certain logic to this assumed causality (Cantwell, 1989), it is also possible that there are barriers to FDI in non-technology-intensive sectors that inhibit FDI from flowing to developing nations in these sectors. A prime candidate would be trade protection by developed countries that restrict imports into these countries from developing countries in labour-intensive goods and services. Thus, it has long been assumed that the reason for historically little FDI in developing nations in textiles and (especially) garment-making is that, given in particular that these sectors are not R&D-intensive, multinational firms have scant technical advantage over local ones. The multinational firms might have other advantages, e.g., privileged access to marketing and distribution channels, but these can be exploited via subcontracting relationships where the local firm supply finished (and often differentiated) products to the multinational. On this, recent evidence from China indicates considerable FDI in these sectors that has been strongly export-orientated (Lemoine and Ünal-Kesenci, 2004), where the exports have been achieved in spite of trade-restrictive measures, and where the ‘home’ to the FDI has been other Asian nations. Perhaps the experience in China warrants a re-examination of at least some of the ‘received wisdom’ regarding FDI and technology. This latter matter notwithstanding, I consider that Vernon’s early insights and later developments of these can still probably explain quite

Edward M. Graham 13

a lot regarding the low rates of FDI to developing nations, but, alas, these insights have not been always taken into account sufficiently by the recent empirical literature. It does seem, for example, that these insights might help to explain why China and Brazil have received much more FDI than the Middle East or Africa, because the two former countries do have large numbers of technically-skilled workers embedded in their populations, whereas the latter regions do not. Moreover, China and Brazil have larger numbers of middle-income consumers than do the Middle Eastern or African nations, and these might create demand for technology-intensive products. I do not know of recent research where this type of proposition has been tested in the form just stated. However, Chapter 7 of this volume, by Marta Castilho and Soledad Zignago, does move in this direction. What they show is that recent FDI into the Southern Cone Common Market (Mercosur) has generated significant amounts of additional imports into the region, but not an equivalent amount of exports. They do not test directly whether the latter might be the result of trade restrictions by developed countries, but speculate that this might be so. They do note that most FDI into the region seems to be meant to serve local rather than export markets (indeed, unlike in most developing nations, a rather high percentage of recent FDI into this region has taken the form of crossborder acquisition; see also Chapter 11 in this volume, by Cézar Miranda Guedes and Mario Gomez Olivares, on this). Interestingly, Castilho and Zignago also find that multinational operations in Brazil are often orientated towards the entire Mercosur, while multinational operations in the other nations within this common market aim towards national markets only. This would seem to run against a theoretical argument presented in Chapter 4 of this volume, by Kjetil Bjorvatn and Carsten Eckel and discussed later in this chapter. Interestingly, if the proposition that FDI is associated with highertechnology sectors does hold up, and that this places developing nations in a disadvantaged position with regard to their ability to attract this investment, it would reinforce the conclusion of Chapter 10 that policy does matter. After all, India is known to have one of the largest and deepest pools of highly-skilled workers, including professionals with very advanced skills and, again, Gupta argues persuasively that India has received far less than its potential FDI. The explanation would not thus appear to be low endowment of human capital, and policy does seem to be at least a candidate explanation. Fortunately, in spite of some words above that might be interpreted as ‘unkind’, the literature is not bereft of the ‘technology factor’ in FDI, and

14 Introduction: FDI in Developing Countries

Chapter 2 in this volume, by Ashok Bardhan and Dwight Jaffee, both explores some recent contributions to the literature and adds to these contributions. Specifically, Bardhan and Jaffee document a trend among multinational corporations to fragment their operations in order to make specific intermediate products in nations where comparative advantage for that product might reside. Thus, even if the Vernon insights do hold up to contemporary scrutiny, there is hope for more FDI to nations where one might expect comparative advantage to lie in those products that involve intensive production using relatively low-skilled labour. Bardhan and Jaffee note that the resulting ‘outsourcing’ by multinationals is often in fact done internally within the firm. Some sort of economy of internalization has long been noted as a necessary condition for FDI to take place (on this, see Buckley and Casson, 1977, 1985). What follows from Chapter 2, then, is that even if multinational firms largely produce what overall are highly technology-intensive products, it is likely that some stages of production, or the production of some intermediate goods, do not require the input of highly-skilled workers, and that this production can be done economically in countries where these workers are in short supply. The authors note that this phenomenon is extending into the service sectors as well as the manufacturing sector. Also, as is becoming well-known, in the services sector – for example, in the development of software – outsourcing is reaching into the developing world (for example, into India), where the required input is highly-skilled workers (Mann, 2003). One contention that is true is that it is not a necessary requirement for inward FDI that a developing nation builds upon existing comparative advantage to create export sectors. The case of South Korea is exemplary in this regard. Korea has, in the space of less than forty years, evolved from quite a low-income country to one that can really no longer be considered as ‘developing’ but rather is now in the ranks of the high-income developed countries. Much of this was achieved by Korea’s developing strong export sectors and, moreover, as Korea’s income has risen and the internal characteristics of its economy changed, so has its ‘revealed comparative advantage’: whereas Korea once exported mainly goods embodying intensively low-skilled labour, such as textiles and apparel, it now exports a wide range of increasingly technology-intensive goods – for example, it is now the world’s largest exporter of advanced memory chips. Moreover, Korea accomplished this with very low FDI; until the late 1990s, when Korea was admitted into the OECD, signalling its transition from developing to developed status, Korea had one of the lowest ratios of FDI stock to GDP in the world.4

Edward M. Graham 15

Chapter 6 of this volume, by Bénédicte Coestier and Serge Perrin, thus concerns a topic that stands somewhat apart from other topics dealt with in this volume: that a rising developing country (in this case Korea), as its economy evolves to create comparative advantage in products where no such comparative advantage had existed historically, might for a number of reasons seek to become an outward foreign direct investor in precisely those sectors in which this new comparative advantage has arisen. They present both a theoretical and an empirical treatment to show that there are circumstances where the producing firm in the developing country is a new entrant, and where overcoming barriers to entry (for example, the need to establish a reputation for meeting quality standards and a need to offer price discounts until their reputation barrier is overcome) can lead to such ‘reverse’ direct investment, that is, FDI from a developing to a developed nation. Korea is, in fact, becoming a direct investment home nation, where investment is indeed flowing from Korea to advanced nations such as the USA and certain EU nations in highly technologyintensive sectors such as semiconductors. Again, though, it is of some importance to note that this started to occur after Korea appeared to have negotiated the cusp dividing the developed nations from the developing ones, and thus recent Korean investment to developed nations is probably best seen as ‘intra-developed nations’, rather than as investment from a developing to a developed nation. This notwithstanding, Korea now is where most developing nations hope someday to be, and the reasons why a nation that only recently was in the cadre of ‘developing’ might become a outward direct investor in technology-intensive sectors are of interest. This contribution by Coestier and Perrin thus sheds new light on the reasons why developing nations can be home as well as host to FDI, an issue that has received periodic attention since a pioneering collection of articles edited by Wells (1983). Before closing this section, it should be noted that there are other, non-economic factors that affect FDI flows to developing nations. One of these invokes a rather old hypothesis that foreign direct investors go to countries where language and cultural characteristics are the same, or similar, to those of the investor’s home country (on this, see Buckley and Casson, 1991). This at least seems to be true for those investor firms that are rather new to investing in and operating subsidiaries outside their home countries. Originally, this hypothesis was applied to US-based firms, which indeed did seem to have a predilection to invest disproportionately in English-speaking countries during the large ‘wave’ of US outward foreign direct investment that occurred during the twenty or so years following the Second World War

16 Introduction: FDI in Developing Countries

(Vernon, 1971) (by ‘disproportionately’, it is meant that investment flows to English-speaking countries were greater than to non-English-speaking ones, controlling for such factors as size of economy, per capita income and so on). In this volume, in Chapter 11, Miranda Guedes and Gomez Olivares explore whether the cultural and linguistic characteristics of Brazil and the Spanish-speaking ‘Southern Cone’ countries of South America was an important factor in newly-multinational firms based in Portugal and Spain investing in these countries. As has already been noted, these countries received rather large shares of FDI, when these shares were measured as a proportion of all FDI to developing countries, during the 1990s. Miranda Guedes and Gomez Olivares do note that these were the most developed of Latin-American countries and that, with the exceptions of Bermuda and Mexico, which they argue to be special cases, these countries received more FDI per capita than other Latin-American countries. Thus, cultural and linguistic characteristics, along with economic characteristics, seem to have created some sort of ‘Iberian logic’ that led to large amounts of FDI to the Southern Cone (Mercosur) and Brazil from the two Iberian nations. Ultimately, in this volume, there is no final answer to the questions ‘Why doesn’t more FDI flow to developing nations?’ and ‘Why do some of these nations receive considerably more FDI than do others?’ Here, we have offered what clearly are partial answers plus, we hope, food for thought. Moreover, the discussion above does raise several important issues that are not covered by the two questions raised just here, ones that are addressed in certain chapters of this volume. These are: first, does FDI create an ‘exogenous’ positive effect on economic growth (that is, an effect that cannot be explained by other factors)? And, second (and related), does FDI create technological ‘spillovers’? In the next section we look at these two issues.

2 Foreign direct investment, spillovers and economic growth When foreign direct investment takes place and the investor does hold proprietary technology, does this investment lead to ‘spillovers’ – that is, effects on the local economy not directly attributable to the operations of the foreign investor? This issue has been explored over several decades, beginning with a seminal work by John Dunning (1958) who found evidence for productivity spillovers resulting from US direct investment in the manufacturing sector in the United Kingdom. Spillovers could

Edward M. Graham 17

conceivably generate either positive or negative externalities. In turn, positive spillovers could be of two varieties: to rival firms and/or workers not employed directly by the multinational operation (in which case the spillover would create a positive externality, where the externality to workers would take the form of wages higher than would otherwise be received) or to local firms that act as suppliers to the foreign-invested operation (in which case the spillover might or might not create a positive externality, depending on whether these same firms acted as suppliers to firms other than the foreign-invested one, but it would certainly enhance efficiency within the local economy). Such spillovers would most probably be created by technology transfer to unrelated firms and, in either case, the expected effect would be the enhancement of total factor productivity. But, as already noted, it is also possible that spillovers might create negative externalities or other negative effects. These again could be reflected in negative externalities – for example, wages might be suppressed, or local R&D activity that creates its own externalities might be curtailed. Whether FDI creates spillovers has proved to be an elusive issue or, to be more precise, whether these are predominantly positive or negative, is not clear. Starting with work by Blomström (1986, 1989) and Kokko (1992), this issue has been subjected to rather intense econometric scrutiny. A recent survey by Görg and Strobl (2001) of twenty-one empirical studies shows mixed evidence, where studies based on cross-sectional data tend to reveal negative externalities while those based on timeseries data tend to reveal positive ones. Lipsey and Sjöholm (2004) note, however, that this survey omitted at least two empirical studies based on time series that revealed negative externalities. Moreover, the studies do not define consistently and precisely what is meant by a ‘spillover’. One should then note that evidence has been inconclusive, not only with respect to whether spillovers to developing nations exist, but also whether they are positive or negative, a perplexing situation indeed. Indeed, one chapter in this volume – Chapter 3 by Nagesh Kumar and Jaya Pradhan – contributes to this perplexity. The authors investigate one possible externality that can be either positive or negative, and notably whether FDI seems to induce or suppress capital formation by indigenous firms in developing countries. Their findings are that it depends on which country one investigates; in some cases, FDI is associated with greater capital formation than would otherwise be expected (a positive externality) but in some cases less capital formation (a negative externality). This chapter also looks at FDI and growth, and this aspect is discussed below.

18 Introduction: FDI in Developing Countries

The Kumar and Pradhan result thus adds to a conclusion that seems to emerge from the literature to the effect that ‘it all depends upon which country you look at, and what type of spillover you are attempting to measure’. A recent contribution by Lipsey and Sjöholm (2005), amplified by Keane (2005), points out that that there might be problems in modelling frameworks in which technological spillovers are assumed to do nothing but shift TFP (although this shift would be an expected result). This is because spillovers modify the production functions of existing firms in ways more subtle and complex than a simple shift of the TFP parameters of these functions. In particular (Keane, 2005), the spillovers might enable firms to take advantage of previously unrealized economies of scale, and this suggests that the spillover affects the underlying technology of the firm. In particular, in order to take advantage of a scale economy, the firm might alter significantly its capital-to-labour ratio. More generally, this can lead to technological heterogeneity within narrowly-defined sectors, leading to econometric estimation problems; such problems are discussed in some detail in the appendix to Keane (2005). Keane notes that his own research with Feinberg and Bognanno (1998) tends to confirm the existence of such heterogeneity; we note here that the likelihood of such heterogeneity was observed by authors as early as Dunning (1958), Vernon (1971) and Hymer (1977). Having noted all of this, in this volume the issue of whether spillovers in fact exist is not addressed beyond the findings of Kumar and Pradhan noted above. But a further issue is addressed in another chapter: the effects of spillovers on multinational firm strategy in developing nations. On this latter issue, specifically, Chapter 4 by Bjorvatn and Eckel examines, via a theoretical framework, how the existence of spillovers affects the choice by investors between locations within a region for an operation. In this matter, the authors do assume that there is technological heterogeneity between the multinational firm and its local rivals, but that this is reflected only in a difference in marginal costs of producing an identical product. The main conclusion is that, if the technological gap between the multinational and local rivals is high, this will drive the firm to invest in a smaller nation and to use this nation primarily as an ‘export platform’ to other nations with larger markets in the region. In part this is because of expected spillovers to rival firms, such that over time the cost advantage of the multinational will become eroded. But if the gap is small, the multinational firm will invest in the biggest market in the region and export to smaller markets. As has already been noted, this conclusion is somewhat at variance with evidence regarding the Mercosur nations of Latin America,

Edward M. Graham 19

as presented in Chapter 7. On this discrepancy, it must be noted that Chapter 5 presents a theoretical result that has not been tested empirically, while Chapter 7 presents an empirical result that might not hold for other regions. Evidently, more work in this area is called for. The issue of whether FDI exerts an autonomous effect on growth has in very recent times become more controversial than before because of the findings of Carkovic and Levine (2005) that the answer might be ‘no’, and this in spite of earlier findings that, by and large, suggest ‘yes’. These earlier findings (for example, Balasubramanyam, Salisu and Sapsford, 1996, and Borensztein, De Gregorio and Lee, 1997), and others are mostly based on panel data wherein growth over time is compared across a group of countries, and where, of course, the effort is made to control for factors other than FDI that might affect rates of growth. The former study concludes that FDI has positive effects on growth, but only if trade and investment policy in the host country is open. The latter study concludes that this relationship is positive, but only in countries where there exists some threshold level of human capital (as argued earlier, such a threshold might be a necessary condition even for a nation to receive significant amounts of FDI). Also, Alfaro et al. (2001) find that growth in developing countries is enhanced by FDI only if financial markets are open. These findings of a positive relationship between FDI and economic growth based on cross-sectional data are buttressed by several studies of single countries – for example, Dayal-Gulati and Hussain (2000), Graham and Wada (2001), and the Kumar and Pradhan chapter of this book. The studies from 2000 and 2001 both look at FDI and economic growth in China, and both conclude that there is a positive relationship between them, while Kumar and Pradhan look at this in the context of India – this is discussed below. Carkovic and Levine (2005) take issue mainly with the panel-databased studies, arguing that such studies have both consistently ignored a country-specific factor that might affect growth rates and, more importantly, that the ‘estimators’ used for regression analysis are of the ordinary least squares (OLS) variety, whereas a more appropriate estimator is the Arellano–Bond estimator (after Arellano and Bond, 1991), a modified version of a generalized method of moments (GMM) estimator. Carkovic and Levine show that, using a panel data-set incorporating data for seventy-two countries over the period 1965–95, and controlling for possible country-specific effects using first differencing as an instrument, the Arellano–Bond estimator fails to show a robust exogenous effect of FDI on growth, whereas using the same data and specification, an OLS estimator does yield a robust effect. Given that the former estimator

20 Introduction: FDI in Developing Countries

is better suited for this type of data, the authors conclude that there is no robust exogenous effect of FDI on growth that can be detected, and that the OLS-based studies thus yield what amount to a ‘false positive’. They do readily admit that their results could be wrong for individual countries and that they might not have accounted fully for countryspecific effects via their instrumental variable approach (on this, see the discussion below of Chapter 5 of this volume).5 But, in another recent paper, Blonigen and Wang (2005) show, using panel data and ordinary least squares (OLS) estimation techniques, that a robust relationship between FDI and growth is observed when the data are restricted to the poorer countries, but not when the data are restricted to rich countries. This could in fact be because, as noted in the previous section, FDI flows in recent years to the rich countries are dominated by cross-border M&A, while FDI flows to poorer countries are dominated by ‘residual’ flows, although this possibility is not tested. Thus they argue that the Carkovic and Levine result might obtain from the pooling of data for rich countries with that for poor countries. However, because the main argument of Carkovic and Levine is that it is the estimator, and not the data, that create what they consider to be a false positive, the issue would seem not to be settled. Unfortunately, the IEA meeting on which the chapters of this volume are based was held before either the Carkovic and Levine, or the Blonigen and Wang, findings were known, and only the chapter by Kumar and Pradhan addresses the issue of what is the correct estimator to be used for panel-data testing of the relationship between FDI and growth, as raised by Carkovic and Levine. Even so, this volume stirs the pot because, in Chapter 3, using both OLS and Arellano–Bond methodology, where the former is done with a specification similar to that of Balasubramanyan, Salisu and Sapsford (1996), and the latter is modified to include lagged dependent variables, Kumar and Pradhan do find in both specifications a positive relationship between FDI and growth in a panel data set for 107 developing countries covering the years 1980–99. But they also find, for the first specification, that the normally assumed causality – that FDI induces growth, might be wrong! Instead, their findings show that it might be economic growth that induces FDI, a finding consistent with the discussion above in this chapter. Further muddying the waters, in Chapter 5, Pradeep Agrawal finds evidence that FDI in South Asia (India, Pakistan, Bangladesh, Sri Lanka and Nepal) has both contributed to growth and created positive externalities since 1990, when overall economic policy throughout the region was liberalized, but not before that time, when this policy was much

Edward M. Graham 21

more restrictive with respect to both trade and direct investment. This finding generally bolsters a finding of Balasubramanyan, Salisu and Sapsford (1996) that trade openness bolsters the effect of growth on FDI. And, at least in the mind of this editor, the finding also presents a challenge to Carkovic and Levine, notably that Agrawal in effect finds that a change in what amounts to country-specific factors (trade and investment policy) changes the outcome of the analysis. Carkovic and Levine, in controlling for a country-specific factor, as noted above, essentially assume that this factor is time-invariant for all countries. However, Agrawal uses an OLS estimator and thus Carkovic and Levine, if given the chance (they are not given such a chance here and so I am doing it for them!) could argue back that the use of this type of estimator is inappropriate even for Agrawal’s limited panel data-set. This, again, leads to the inevitable conclusion that yet more work needs to be done. One virtue of this volume is that it makes no claim that anything is as yet fully resolved, at least with respect to the effects of FDI in developing nations on economic growth or on the creation of positive or negative spillovers.

3

Some other issues

Two chapters in this volume cover issues that are (i) important; and (ii) do not quite fit into the discussion above. Thus, in this concluding section, we examine each of these. The first chapter, Chapter 9 by José De Gregorio, examines the issue of whether FDI in the natural resource sectors is less beneficial to developing nations than FDI in other sectors – for example, manufacturing. A standard hypothesis that appears more in the political science literature than that of economics is that natural resource FDI is often quite negative in its attributes (see, for example, Moran, 1974, which, like De Gregorio, examines the case of Chile but at a different point in time). Reasons given include effects that are macroeconomic (for example, claimed overvaluation of exchange rates that might suppress development of export sectors that otherwise would exist) and sociological (for example, that FDI in natural resources leads to more corruption than otherwise would exist). De Gregorio examines the case of Chile, his home country, which has received considerable FDI in natural resource sectors and concludes that this investment has, in terms of measurable negative effect, ‘gotten something of a bum rap’. The supposed negatives do not seem to infect present-day Chile, and tangible benefits can be identified. The second chapter, Chapter 8 by Matteo Iannizzotto and Nigel Miller, looks at the issue of whether appreciation or depreciation of a currency or

22 Introduction: FDI in Developing Countries

currency exchange rate volatility has measurable effects on FDI inflows of the relevant country. In some ways, this chapter might seem out of place in this volume, because Iannizzotto and Miller do not examine developing countries at all, but rather the United Kingdom. In spite of this arguable shortcoming, their contribution is appropriate because the currencies of almost all developing nations are either partly pegged to a so-called ‘international currency’ such as the US dollar or the euro, or are subject to restrictions in terms of convertibility. For the purposes of testing whether or not currency value changes or exchange rate volatility have any measurable impacts on FDI flows, it is desirable that the currency transactions not be restricted and that the currency not be tied to another currency, and this is the case for the British pound sterling. The finding of the authors is that sterling appreciation caused a marked slowdown in FDI inflow to the United Kingdom, but that changes in volatility have had no such measurable effect. Having noted this, the editor feels that the reader has now been subjected to enough introduction to this volume or, indeed, far too much. The reader is thus encouraged to proceed to the individual chapters.

Notes 1 In fact, cross-border M&A activity began to accelerate sharply following 1985 and grew each year until 1990, but then waned during 1991–3. It accelerated again in 1994, and grew each year until 2000. Both 2001 and 2002 showed significant slowdown in this activity, but to levels that were still very high relative to any time prior to the mid-1980s. Data for 2003 and 2004 are not available at the time of writing, but early indicators are that another period of acceleration might be taking place. Thus, arguably, we have witnessed two or even three ‘waves’ of this activity in rapid succession. But however we view this phenomenon – as three waves in succession or one big wave with fluctuations – it is without historical precedent. 2 Prior to the wave (or waves) of the past twenty years or so, three earlier waves of mergers and acquisitions can be identified in the USA, one that began in the late 1880s, extended through the 1890s and into the first years of the twentieth century; a second that took place during the 1920s; and a third during the 1960s. Interestingly, only the first of these waves can be explained by what is arguably the most ‘straightforward’ motivation – the enlargement of horizontal market power (this was the era of the formation of the ‘trusts’ in key industrial sectors, where these trusts clearly were intended to create monopolies or near-monopolies in the relevant markets). Most of the M&A of the 1920s were in the utilities sector, and few of these appear to have increased the market power of the firms. The mergers of the 1960s were often to create ‘conglomerates’ – that is, firms that sell into several unrelated markets. There is considerable consensus now that this organizational form (conglomerate firm) was a failure, and many of what were, at the time, highly ballyhooed firms

Edward M. Graham 23 of this sort have since gone out of business or been reduced to shadows of their former selves (for example, Litton Industries, Textron, ITT, Sperry-Rand). On these earlier waves of M&A, see Scherer and Ross (1990), ch. 5. 3 This type of problem can exist for FDI to areas other than Asia, but there are reasons to suspect that the problem is more acute in the Asian data than in data for other areas. No reliable estimates exist as to the magnitude of doublecounting that is created. 4 On this, see Graham, 2003, and mainly the references therein that collectively tell the story of Korea’s rise from a poor to a rich country. 5 Specifically, first-time differencing data to eliminate a country-specific variable, as done by Carkovic and Levine, is a valid means of dealing with such an unobservable (and unmeasurable) variable if, and only if, that variable does not change over time. However, there are reasons to believe that at least some of the elements of what such a variable is meant to capture have, in the Carkovic and Levine panel data, indeed changed over time – for example, in many countries, there have been major changes in official policy towards FDI. Such changes would thus not be properly controlled by their approach.

References Alfaro, Laura, Areendarn Chanda, Sebnem Kalernli-Ozcan and Selin Sayek (2001) ‘FDI and Economic Growth: The Role of Local Financial Markets’, Harvard University Graduate School of Business Administration, Working Paper no. 01-083. Arellano, M. and S. Bond (1991) ‘Some Tests of Specification for Panel Data: Monte Carlo Evidence and an Application to Employment Equations’, Review of Economic Studies, vol. 58(2), pp. 277–97. Balasubramanyam, V. N., M. Salisu, M. Sapsford and D. Sapsford (1996) ‘Foreign Direct Investment and Growth in EP and IS Countries’, Economic Journal, vol. 106, pp. 92–105. Blomström, Magnus (1986) ‘Foreign Investment and Productive Efficiency: The Case of Mexico’, Journal of Industrial Economics, vol. 15, pp. 97–110. Blomström, Magnus (1989) Foreign Investment and Spillovers (London: Routledge). Blonigen, Bruce A. and Miao Wang (2005) ‘Inappropriate Pooling of Wealthy and Poor Countries in Empirical FDI Studies’, in Theodore H. Moran, Magnus Blomström and Edward M. Graham (eds) (2005), pp. 241–3. Borensztein, E., J. De Gregorio and J. W. Lee (1998) ‘How Does Foreign Investment Affect Growth?’, Journal of International Economics, vol. 45(1), pp. 115–35. Brewer, Thomas L. and Stephen Young (2001) The Multinational Investment System and Multinational Enterprises (Oxford: Oxford University Press). Buckley, Peter J. and Mark Casson (1977) The Future of the Multinational Enterprise (London: Macmillan). Buckley, Peter J. and Mark Casson (1985) The Economic Theory of the Multinational Enterprise (London: Macmillan). Buckley, Peter J. and Mark Casson (1991) ‘Multinational Enterprises in Less Developed Countries: Cultural and Economic Interactions, in Peter J. Buckley and Jeremy Clegg (eds) Multinational Enterprises in Developing Countries (London: Macmillan), pp. 27–55.

24 Introduction: FDI in Developing Countries Cantwell, John A. (1989) Technological Innovation and Multinational Enterprise (Oxford: Basil Blackwell). Carkovic, Maria and Ross Levine (2005) ‘Does FDI Accelerate Economic Growth?’, in Theodore H. Moran, Magnus Blomström and Edward M. Graham (eds) (2005), pp. 125–220. Dayal-Gulati, Anuradha and Aasim M. Hussein (2000) ‘Centripetal Forces in China’s Economic Takeoff’, International Monetary Fund Working Paper no. WP/00/86. Dunning, John H. (1958) American Investment in British Manufacturing Industry (London: George Allen & Unwin and New York: Arno Press, 1970). Feinberg, S., M. Keane and M. Bognanno (1998) ‘Trade Liberalization and “Delocalization”: New Evidence from Firm-Level Panel Data’, Canadian Journal of Economics, vol. 31, pp. 749–74. Görg, Holger and Eric Strobl (2001) ‘Multinational Companies and Productivity Spillovers: A Meta-analysis’, Economic Journal, vol. 111, pp. F723–39. Graham, Edward M. (2003) Reforming Korea’s Industrial Conglomerates (Washington, DC: Institute for International Economics). Graham, Edward M. and Erika Wada (2001) ‘Foreign Direct Investment in China: Effects on Growth and Performance’, Institute for International Economics Working Paper no. 01-03. Hymer, Stephen H. (1977) The International Operations of National Firms (Cambridge, Mass.: MIT Press). Keane, Michael P. (2005) ‘Econometric Challenges in Estimating FDI Spillover Effects on Productivity’, in Theodre H. Moran, Magnus Blomström and Edward M. Graham (eds) (2005), pp. 179–91. Kobrin, Stephen (1976) ‘The Environmental Determinants of Foreign Direct Investment: An Ex Post Empirical Analysis’, Journal of International Business Studies, vol., pp. 29–42. Kokko, Ari (1992) Foreign Direct Investment, Host Country Characteristics, and Spillovers (Stockholm: Economic Research Institute of the Stockholm School of Economics). Lemoine, Françoise and Deniz Ünal-Kesenci (2004) ‘Assembly Trade and Technology Transfer: The Case of China’, World Development, vol. 32(5), pp. 829–50. Lipsey, Robert and Fredrik Sjöholm (2004) ‘Foreign Direct Investment, Education and Wages in Indonesian Manufacturing’, Journal of Development Economics, vol. 73, pp. 415–22. Lipsey, Robert and Fredrik Sjöholm (2005) ‘Host Country Impacts of Inward FDI: Why Such Different Answers?’, in Theodore H. Moran, Magnus Blomström and Edward M. Graham (eds) (2005), pp. 23–43. Markusen, James R. (1984) ‘Multinationals, Multi-plant Economies and the Gains from Trade’, Journal of International Economics, vol. 16, pp. 205–16. Maskus, Keith E. (2000a) ‘Intellectual Property in the New Round’, in J. J. Schott (ed.), The WTO after Seattle (Washington, DC: The Institute for International Economics), pp. 137–58. Maskus, Keith E. (2000b) Intellectual Property Rights in the Global Economy (Washington, DC: Institute for International Economics). Moran, Theodore H. (1974) Multinational Corporations and the Politics of Dependence (Princeton, NJ: Princeton University Press).

Edward M. Graham 25 Moran, Theodore H., Magnus Blomström and Edward M. Graham (eds) (2005) Does FDI Promote Development? (Washington, DC: Institute for International Economics). Romer, Paul M. (1990) ‘Endogenous Technological Change’, Journal of Political Economy, vol. 98, pp. 70–102. Romer, Paul M. (1994) ‘The Origins of Endogenous Growth’, Journal of Economic Perspectives, vol. 8(1), pp. 3–22. Solow, Robert M. (1994) ‘Perspectives on Growth Theory’, Journal of Economic Perspectives, vol. 8(1), pp. 45–54. Stiglitz, Joseph E. (2003) Globalization and its Discontents (New York: W. W. Norton). UNCTAD (1992) World Investment Report 1992 (Geneva: United Nations Conference on Trade and Development). UNCTAD (2003) World Investment Report 2003 (Geneva: United Nations Conference on Trade and Development). Vernon, Raymond (1966) ‘International Investment and International Trade in the Product Cycle’, Quarterly Journal of Economics, vol. 80, pp. 190–207. Vernon, Raymond (1971) Sovereignty at Bay: The Multinational Spread of US Enterprises (New York: Basic Books). Wells, Louis T. (ed.) (1983) Third World Multinationals (Cambridge, Mass.: MIT Press). Wolf, Martin (2004) Why Globalization Works (New Haven, Conn.: Yale University Press). World Bank (2002) World Development Indicators 2002 (Washington, DC: The World Bank).

2 On Intra-firm Trade and Multinationals: Offshoring and Foreign Outsourcing in Manufacturing Ashok Deo Bardhan and Dwight Jaffee Haas School of Business, University of California, Berkeley, USA

1

Introduction

For advanced industrialized economies, the era of globalization has created key roles for both the foreign outsourcing of intermediate inputs and intra-firm trade.1 Recent papers, including Feenstra and Hanson (1996) and Brainard (1997), have treated these subjects separately, but their interaction and possible intersection (namely, transnational intrafirm trade in intermediate inputs) have received little attention.2 Lowcost foreign outsourcing has long attracted many firms, whether part of a multinational enterprise or acting as independent companies. Increasingly, however, organizational and other considerations have motivated firms to use imported inputs from affiliates abroad, instead of inputs from arm’s-length domestic manufacturers; this activity amounts to vertical integration across borders. This process of intra-firm offshoring seems to be particularly intense in the case of high-tech sectors. Indeed, one of the signal attributes of a manufactured high-tech product is the extensive nature of its value-chain, the number of intermediate products and services, and the global, fragmented, nature of the final output.3 Progress in transportation, communications and standardization has significantly increased the fragmented nature of production. The high-tech valuechain is now a multilateral, multinational production mosaic, involving many countries but often just one firm or a group of affiliated firms. The phenomenon of foreign outsourcing is no longer restricted to the manufacturing sectors. Business-process and business-services outsourcing is gathering momentum, and jobs and occupations ranging 26

Ashok Deo Bardhan and Dwight Jaffee 27

from medical transcription to stock-market research are being outsourced to other countries. As with manufacturing, this kind of outsourcing also encompasses both outsourcing to arm’s-length firms, as well as to subsidiaries and affiliates in foreign countries. Although similar issues arise in both goods and services, data limitations constrain us to consider only goods trade here. In this chapter we look at intra-firm trade and imported intermediate inputs, with a special focus on the high-tech computer industry. The main questions we pose are: (i) What are the cross-country determinants of US imports of intermediate inputs? (ii) What is the relationship between imported intermediate inputs and intra-firm imports? (iii) Does the importance of intra-firm and intermediate input trade vary across industry lines? The foreign affiliates of US multinational enterprises (MNEs) can provide either distribution or production facilities for their parent companies, and here we focus on affiliates functioning as production centres. The output of these affiliates can be directed in several ways: (i) to the MNE parent; (ii) to customers in the home country of the MNE; and (iii) to worldwide customers of the MNE. Category (i) is one form of intra-firm imports for the USA. The other, symmetric, form of intra-firm US imports occurs when a foreign-based MNE ships goods to its US-based affiliate. Both types of intra-firm trade will be influenced by industrial organization factors such as transactions costs, as well as specific international trade factors such as tariffs, longdistance transportation costs, worldwide marketing, and issues related to taxation and exchange rate hedging. Intra-firm trade can cover both final and intermediate goods. This chapter focuses on intermediate goods. The use of imported intermediate inputs in manufacturing depends on the industrial organization and international trade factors just mentioned, as well as on supply chain management tools that control demand, supply and quality variability. Global economic integration has allowed MNEs to create fragmented, sequential production processes by locating their intermediate production activities in various parts of the world. Together with such

28 On Intra-firm Trade and Multinationals Table 2.1 Imports into the USA by trade categories (percentages of total imports) 1992

1997

Intermediate inputs/final goods Percentage intermediate inputs Percentage final goods

37 63

38 62

Intra-firm/arm’s-length Percentage intra-firm (a) US MNEs (b) Foreign MNEs Percentage arm’s-length

43 17 26 57

52 30 22 48

Total imports (US$ billion)

505

748

Source:

See Appendix on page 37.

fragmented production comes an intensive trade in intermediate inputs for the purpose of production of the final manufactured good.4 Table 2.1 shows aggregate US data on intermediate-input imports and intra-firm imports, for 1992 and 1997. Approximately three-eighths of all US goods imports have been intermediate inputs (the remainder are final goods). About 43 per cent of all US goods imports arrived through intra-firm channels in 1992, rising to 52 per cent in 1997 (the remainder came through arm’s-length channels). The computer industry is a particularly telling example of both intermediate input and intra-firm trade. The complexity and sophistication of the end products of this sector dictate a wide range of specialized production activities and stages, and hence a large number of intermediate inputs. The geographical spread of the production base of most of the large multinational firms in this industry results in a brisk international trade in intermediate products. The empirical tests in this chapter focus on US imports, because of the importance of US MNEs in international trade, and because the US Bureau of Economic Analysis (BEA) has provided high-quality tabulations of several especially relevant data sets. First, the BEA publishes detailed data on intra-firm trade in goods, involving both US and foreign MNEs and their respective affiliates. These data are based on extensive benchmark surveys taken every five years, as well as smaller annual surveys. The BEA also publishes related data on US foreign direct investment both abroad in the USA. Finally, US imported intermediate inputs can be computed by combining three BEA data sets: (i) an input–output data set, based on the 1992 and 1997 US Census of Manufactures, is applied to determine the total quantity of intermediate inputs by industry;

Ashok Deo Bardhan and Dwight Jaffee 29

(ii) industry import data are used to estimate imported intermediate inputs; and (iii) import data by industry and country of origin are then used to estimate imported intermediate inputs by country of origin. See the Appendix (page 37) for how we determine imported inputs by industry and country of origin. Previous studies involving imported intermediate inputs have applied only steps (i) and (ii) of this methodology. Also, these studies primarily use industry data and have focused on the labour market impacts or exchange rate exposures.5 Our analysis, in contrast, focuses on a country cross-section, and applies the data to study the intersection and interaction of imported intermediate inputs and intra-firm trade.

2

Literature review

The issue of intra-firm trade is linked inextricably to the study of multinationals and of foreign direct investment (FDI). A large part of the FDI literature deals with its country-wise determinants, such as size, relative endowments, and trade and investment costs (Carr et al., 2001). The literature also studies relative rates of return (Chernotsky, 1987) and, in the case of foreign investment in R&D activity, the size of the scientific base (Kuemmerle, 1999) and company strategies to tap locally-embedded expertise or to develop an organizationally complex international network for technological learning (Cantwell and Santangelo, 1999). There is also a literature that assesses the impact of FDI on the local host economy in terms of its impact on innovation (Glass and Saggi, 2002), on benefits accruing from increased competition and efficiency gains (Graham and Wada, 2001), and on economic growth (Zhang, 2001; NairWeichert and Weinhold, 2001) or, on the other hand, the lack of robust influence of FDI on growth (Carkovic and Levine, 2005). For our purposes, the literature that deals with transnational vertical integration and intra-firm trade is of even greater relevance. For example, Wilamoski and Tinkler (1999) show that there was a rise of intra-firm exports and imports between the USA and Mexico as a result of US FDI in Mexico. Other studies of multinational firms have looked at the motivation behind investment abroad, and whether FDI complements or substitutes trade (Konan, 2000; Roy and Viaene, 1998). Konan’s theoretical model, in particular, shows that intra-firm trade in intermediate goods implies that vertical investment complements rather than substitutes for trade. Lipsey and Weiss (1981) show that foreign production by a US firm does not, on balance, substitute for exports by that firm to the

30 On Intra-firm Trade and Multinationals

area in which the production occurs, and that a firm’s output in a foreign area and the firm’s exports from the USA to that area are positively correlated, particularly for exports of intermediate goods. Another branch of the intra-firm trade literature deals with its determinants. For example, Helpman (1984) develops a model that generates shares of intra-firm trade as a function of relative nation size and variations in relative factor endowments. A large literature also exists on transfer-pricing and taxation issues, and their relationship with intrafirm trade (Taylor, 2002), while Madan (2000) shows how different levels of taxation in the host-country give rise to a different mix of intra-firm trade in final and intermediate goods. Turning to outsourcing, Grossman and Helpman (2002) study the determinants of outsourcing locations in a global economy, using a general equilibrium trade model. Costly searches and incomplete contracts are critical in this model. The relative thickness of markets for input suppliers, relative search costs, and the contracting environment have an impact on the extent of global outsourcing. Countries with an active inputs market and reliable contracting environment would be relatively dependable sites for outsourcing. In an empirical study, Andersson and Fredrikson (2000) show that internal imports of intermediate goods by Swedish firms were dependent on the international organization and concentration of production, market size and R&D expenditures. The large volume of literature reviewed in this section confirms the importance attached to the separate topics of intra-firm trade and imports of intermediate inputs. On the other hand, the combination and integration of these two key aspects of globalization appears not to have been studied. This intersection of intra-firm trade and imports of intermediate inputs is thus the focus of our empirical tests, to which we now turn.

3

Analysis

To start, it is useful to clarify the relationship between imported intermediate inputs and intra-firm imports. This relationship is illustrated in Figure 2.1. The full 360-degree circle represents the total amount of goods imported by the home country from any given foreign country in any given year. The right hemisphere (quadrants 1 and 2) show intra-firm imports, representing transactions between a MNE and its affiliate, either from a foreign affiliate to a home country MNE, or from a foreign MNE to its home-country-based affiliate. Intra-firm trade can occur in either intermediate inputs or final goods, represented by quadrants 1 and 2

Ashok Deo Bardhan and Dwight Jaffee 31

4

1

Intra-firm

Arm’s length

Intermediate inputs

Final goods 3

2

Figure 2.1 Total imports into the home country classified as intermediate inputs or final goods and as intra-firm or arm’s-length trade

of the circle, respectively. The left hemisphere (quadrants 3 and 4) represents imports from arm’s-length trading partners, meaning that these imports are not carried out within the same firm. Arm’s-length trade can also occur in either intermediate inputs or final goods (in quadrants 4 and 3, respectively). The quadrants in Figure 2.1 are of equal size only for graphical convenience. In fact, a primary goal of this study is to determine the average size of these quadrants, and to determine the factors that cause the quadrant sizes to vary across countries. As already noted in Table 2.1 for 1997, however, we do have information concerning the size of the two sets of hemispheres in Figure 2.1: ●

●

38 per cent of all US imports were intermediate goods and 62 per cent final goods; and 52 per cent of all US imports were intra-firm imports and 48 per cent arm’s-length.

Also, about three-fifths of all intra-firm imports were carried out by United States MNEs, the remainder being imports by the US affiliates of foreign multinationals. The size of the four quadrants individually cannot be derived from aggregate data that only separate intra-firm from arm’s-length trade, and intermediate from final goods trade. We need three pieces of independent information, in addition to the total amount of imported goods (the size of the circle), to determine the size of each quadrant. Data that separate (i) intra-firm and arm’s-length trade; and (ii) intermediate and final goods trade, provide only two pieces of independent information.

32 On Intra-firm Trade and Multinationals

In fact, we know of no standard data set that provides separate quadrant sizes. However, the information we have illustrated in Figure 2.1 and summarized in Table 2.1 for aggregate imports into the USA are also available on a disaggregated basis by country of origin (that is, the exporting country). These data have the potential to provide substantially more information about the distribution of import flows across the four quadrants. As a simple example, assume the split between intra-firm and arm’slength trade is 50–50, and the split between intermediate input and final goods trade is also 50–50. With no further information, we cannot know the size of each of the four quadrants shown in Figure 2.1. Now also assume that intermediate input and intra-firm imports always occur together, and that final goods and arm’s-length imports also always occur together, although the two factors may vary across countries. This pattern is still consistent with an 50–50 aggregate split between intrafirm and arm’s-length trade, and between intermediate inputs and final goods trade. But the disaggregated patterns provide additional information. In particular, we now know that quadrants 2 and 4 of Figure 2.1 must be empty, and that exactly 50 per cent of the trade would appear in each of quadrants 1 and 3. This example illustrates why data disaggregated by country of origin may provide insights into the aggregate data that are not available from the aggregate data directly. It is, of course, a stylized case. In the real world, the best hope is to find that the cross-country correlations for the various import categories are sufficiently informative to allow us to decipher the true structural features with a reasonable level of confidence. Table 2.2 shows total imports and input imports into the USA from the major countries of origin in 1997. The table also provides input imports/total imports ratios and intra-firm imports/total imports ratios by country; for all data, see the Appendix on page 37. Much greater variation is apparent in the intra-firm import ratios. More than 70 per cent of the exports to the USA from countries such as Japan are carried out through intra-firm trade, while at the other end of the spectrum, imports from Taiwan are primarily of an arm’s-length nature. The table also reflects the diverse nature of the countries shipping intermediate inputs to the USA, covering developing and developed countries, and European and Asian countries alike: a true testimonial to globalization. Table 2.3 shows total imports and intermediate input imports for the four industries with the largest amount of total imports among all US 3-digit NAICS industries in 1997. The table also shows the top three countries of origin for intermediate inputs for each industry. Computers

Ashok Deo Bardhan and Dwight Jaffee 33 Table 2.2 1997 imports into the USA by trade categories and selected major countries Total imports (US$ billions)

Input imports/ total imports (ratio)

Intra-firm imports/ total imports (ratio)

168 121 86 43 63 33 33 21 23 19 748

0.40 0.42 0.36 0.43 0.29 0.43 0.43 0.46 0.36 0.40 0.38

0.47 0.71 0.35 0.60 0.10 0.47 0.08 0.38 0.22 0.67 0.52

Canada Japan Mexico Germany China UK Taiwan France Korea Italy US total Source:

Table 2.3

Imported inputs into the USA, by largest importing industries, 1997

Computers and electronics NAICS 334 Transportation equipment NAICS 336 Machinery except electrical NAICS 333 Chemicals NAICS 325 Source:

See Appendix on page 37.

Total imports (US$ billions)

Input imports (US$ billions)

Input imports/ total imports (ratio)

173

68

0.39

Japan, Taiwan, Mexico

149

72

0.48

Canada, Japan, Mexico

65

35

0.54

Japan, Canada, Germany

51

26

0.51

Canada, Japan, Germany

Top three countries of origin for inputs

See Appendix on page 37.

and Electronics (NAICS 334) and Transportation Equipment (NAICS 336) are first and second, respectively, with regard to both total imports and input imports. The input import ratio of 39 per cent for NAICS 334 is somewhat lower than the other industries shown, since many computer and electronic products are fully assembled abroad and imported as final goods.

34 On Intra-firm Trade and Multinationals

4

Regression estimates

We now turn to regression tests on intra-firm and intermediate input imports into the USA. Our dependent variable is the log of US intermediate input imports from a cross-section of countries of origin. The descriptions for all data are given in the Appendix on page 37. We estimate multivariate cross-section regressions for the years 1992 and 1997 separately to determine which factors are correlated most highly with the observed cross-country pattern. Our specification starts with a one-direction version of the gravity model, since we are looking only at the imported intermediate inputs from each trading partner to the USA (see Feenstra et al., 2001, for a recent survey).6 We then modify the standard model by including intra-firm and arm’s-length goods imports by country as additional explanatory variables. We also separate intra-firm trade into imports from the foreign affiliates of US MNEs, and imports from foreign MNEs to their US affiliates. The following are the primary independent variables (all variables apart from the Asian dummy are measured in logs):7 IFUSA IFFOR

US imports from the foreign affiliates of US-based MNEs; US imports from foreign-based MNEs to their US-based affiliates; ARML US imports sent to arm’s-length recipients (= total US imports − IFUSA − IFFOR). GDPPC Gross domestic product (GDP) per capita of the country of origin of imports; POP Population of the country of origin; DIST Great Circle distance between largest city of foreign country and Kansas City, Missouri;8 and ASIAN Dummy variable: 1 for Asian countries of Asia-Pacific Economic Cooperation. The results in Table 2.4 are divided into three parts. Part A has the log of imported intermediate inputs as the dependent variable, and the regression is estimated on a cross-section of forty-eight countries for which data are available for 1992. Equation 1 is a standard gravity model, based on per capita GDP, population (as a measure of size), distance, and a dummy variable for the Asian countries. The adjusted R2 is over 75 per cent, indicating that an important share of the cross-country distribution of US imports of intermediate inputs can be explained on the basis of gravity variables alone. We also tested a variety of other gravity variables, but

Ashok Deo Bardhan and Dwight Jaffee 35 Table 2.4 Regression results Eqt #

Constant

Part A 1

Imported intermediate inputs is the dependent variable −5.31 0.92 0.77 ∗∗ ∗∗ ∗∗ (2.15) (11.94) (8.99) −5.31 0.008 0.06 0.80 0.49 0.18 ∗∗ ∗∗ ∗∗ (2.56) (0.12) (0.74) (5.34) (2.72) (1.65)

– 1992 data −1.00 1.90 ∗∗ ∗∗ (3.97) (7.27) −0.19 0.47 (0.77) (1.18)

Imported intermediate inputs is the dependent variable −7.96 .93 0.87 ∗∗ ∗∗ ∗∗ (9.15) (8.38) (2.51) −2.93 0.17 0.24 0.36 0.24 0.22 ∗ ∗∗ ∗∗ ∗∗ ∗∗ ∗∗ (1.98) (2.98) (4.53) (6.17) (2.97) (2.89)

– 1997 data −0.89 1.65 ∗∗ ∗∗ (3.8) (4.88) −0.13 0.46 ∗∗ (0.90) (2.63)

2 Part B 3 4

IFUSA

IFFOR

ARML

GDPPC

POP

DIST

ASIAN

Adj. R2

0.779 0.895

0.763 0.950

Part C High-tech (NAICS 334) imported intermediate inputs is the dependent variable – 1997 data −27.93 1.54 1.34 −0.55 3.39 0.58 5 ∗∗ ∗∗ ∗∗ ∗∗ (3.36) (5.96) (5.46) (0.98) (4.15) 6 −23.25 0.77 0.38 0.06 0.50 0.39 0.92 1.54 0.731 ∗∗ ∗∗ ∗ (2.33) (2.50) (1.11) (0.26) (1.12) (0.93) (1.74) (1.87) Notes: Ordinary least squares with White heteroskedasticity adjustment. Absolute values of t -statistics shown in parentheses; ∗ significance at 10%, ∗∗ significance at 5%. All regressions are estimated on a cross-section of countries; 48 countries in 1992; 38 countries in 1997. All data are in logs except for the Asian dummy. See Appendix (p. 37) for detailed description of data series. Source:

Authors’ calculations.

none were consistently significant. Our basic results would be unaffected by including any of these variables. Equation 2 in Table 2.4 adds three disaggregated import flows as potential determinants of imported intermediate inputs. Their coefficients measure the elasticity of imported inputs with respect to each of the US import categories. The results indicate that intra-firm trade was related primarily to final good imports as of 1992. We also tested for a direct effect of US foreign direct investment in each country, but it did not provide an independent effect over and above that of the related intra-firm trade flows. We also estimated all the equations in Table 2.4 using instrumental variables, but none of the results were changed in any substantive way.9 Part B of Table 2.4 repeats the estimation of Part A for 1997 data. The sample size is now only thirty-eight countries, since the 1992 data rely on a special tabulation carried out by the Bureau of Economic Analysis (see Zeile, 1997), and a comparable tabulation is not yet available for the 1997 data. Equation 3 provides estimates based on the gravity specification alone, with results similar to those obtained for the 1992 data. Equation 4

36 On Intra-firm Trade and Multinationals

adds the same three import variables used in Equation 2. The coefficient estimates for 1997 for these variables indicate a significant increase in the importance of intra-firm trade, related to both US and foreign MNEs, and a corresponding reduction in the importance of arm’s-length trade, as a determinant of imported intermediate inputs. This is an important result, since it confirms the view that MNEs are increasingly using intrafirm offshoring as they decentralize their production processes.10 Part C of Table 2.4 repeats Parts A and B (with 1997 data), but the dependent variable is now the log of imports of only high-tech intermediate inputs, defined here as NAICS code 334.11 Equation 5 begins with the gravity model variables, but with two notable differences from the results in Equations 1 and 3. First, the distance variable is now much less important. This is understandable, since high-tech imports are commonly referred to as ‘weightless’ in terms of their value-to-weight ratio, implying that they are much less sensitive to transportation costs. Second, the Asian country dummy is much more important than it was in the earlier equations. This too makes sense, since there is other evidence that the Asian countries are of increasing importance as sources of intermediate inputs for US high-tech industries. Equation 6 of Table 2.4 adds the import variables to the basic gravity model for high-tech imported inputs. Compared with Equations 2 and 4, Equation 6 indicates that US imports of high-tech intermediate inputs depend primarily on intra-firm trade (and not arm’s-length transactions), and especially on imports by US MNEs. Indeed, imports by US MNEs are now the predominant source of imported high-tech intermediate inputs into the USA. This result provides empirical verification of the view that offshoring has become especially important for US MNEs in high-tech industries.

5

Conclusions

High and growing levels of intra-firm trade, and trade in intermediate inputs, are among the stylized facts of international trade, and are thought to play key roles in the new era of globalization. Although, individually, they have been studied intensively, little attention has been paid to their interaction – that is, intra-firm trade in intermediate inputs. One major problem has been the absence of data that could measure the amount of intra-firm trade involving intermediate inputs, or vice versa. Our study offers two primary innovations. First, we have developed a data set of imported intermediate inputs by both industry and country of origin. Second, we have used estimates from a regression model with

Ashok Deo Bardhan and Dwight Jaffee 37

intermediate inputs imported into the USA as the dependent variable to determine the absolute and relative importance of intra-firm imports as a determinant of trade in intermediate inputs. Our key results, from Table 2.4, are: (i) Intra-firm imports were a relatively unimportant source of intermediate imports as of 1992. Most US intermediate goods imports at that time were the result of arm’s-length trades. (ii) By 1997, intra-firm trade, by both US and foreign MNEs, had become very important as a source of imported intermediate inputs. However, arm’s-length trade also remained a significant determinant of US intermediate input imports. (iii) Standard gravity model variables were found to be important determinants of US imports of intermediate inputs, in addition to the key role of intra-firm trade variables. (iv) Estimates were also derived for high-tech intermediate input imports, defined as NAICS code 334, which represents computers and electronic products. These additional results were: (a) Transportation costs, measured by distance, were not a major hindrance to high-tech intermediate imports, consistent with the high-value, low-weight, character of these goods. (b) Intra-firm trade (not arm’s-length transactions), and especially imports by US MNEs, were the key determinants of high-tech intermediate input imports, consistent with the view that offshoring has become especially important for US MNEs in hightech industries. In particular, US MNEs were responsible for more than two-thirds of all imports of high-tech intermediate inputs into the US. For further research, we plan to investigate intra-firm trade flows and outsourcing in services, as well as to study the possibility of spillover effects of intra-firm trade, and trade in intermediate inputs on trade overall.

6

Appendix

6.1 Computation of imported intermediate inputs by country of origin To calculate imported intermediate inputs by sector and by country of origin, we applied the following formulae to each 6-digit input sector in US manufacturing

38 On Intra-firm Trade and Multinationals (all amounts in US$ billions):
 Mi IIi = Ii (Pi − Xi + Mi )

(1)

Where Ii = amount of sector i goods used as inputs in all US manufacturing (from US Census of Manufacturing Input/Output data for 1992 and 1997, respectively); IIi = imported inputs of sector i goods; Mi = total imports of sector i goods; Pi = US production of sector i goods; and Xi = US exports of sector i goods. The basic assumption here is that, for any input sector, the percentage that imports of intermediate input represent of total intermediate inputs is the same as the percentage that imports represent of all net sources of that commodity (= Pi − Xi + Mi ).12 IIic = Mic IIi

(2)

Where IIic = imported intermediate inputs of sector i from country c; and Mic = sector i imports from country c as a proportion of US total sector i imports. The basic assumption here is that country c’s share of imported intermediate imports of sector i goods equals that country’s share of all imports of sector i goods.

6.2

Data sources

6.2.1

Trade data by countries and industries

All import data by countries and industries are from US International Trade Commission’s Trade DataWeb website: http://dataweb.usitc.gov.

6.2.2

Intra-firm trade imports by country of origin

Data for 1992 are from Zeile (1997); data for 1997 are from Mataloni (1999) for US MNEs, and from Zeile (1999).

6.2.3

Gravity model variables

The distance data have been calculated using Encarta. Gross domestic product (GDP), Gross domestic product per capita (GDPPC) and Population (POP) are from the World Bank database: http://devdata.worldbank.org/data-query.

Notes 1 See Markusen (1995), Douglas (1996) and Feenstra (1998) for surveys of the basic patterns for international trade, with particular attention to intra-firm trade and international outsourcing. 2 Other recent studies of trade and sales by multinational firms and their affiliates include Zeile (1997), Markusen and Maskus (2001) and OECD

Ashok Deo Bardhan and Dwight Jaffee 39

3

4 5 6

7

8

9

10 11

12

(2002). Other recent studies of international outsourcing include Hummels, Raporport and Yi (1997), Campa and Goldberg (1997) and Swenson (2000). Outsourcing is not a sequential concept in the sense that the final assembly is necessarily done in the developed country after importing the intermediate inputs; sometimes the final assembly is done abroad after a series of crossborder transactions and trade. See Arndt and Kierzkowski (2001) for a collection of papers on fragmented production. Work on this method was pioneered by Campa and Goldberg (1997), Feenstra and Hanson (1996) and Hummels, Raporport and Yi (1997). The gravity equation has been derived by economists from basic principles of international economics. The equation postulates that bilateral trade between two countries would be proportional to the product of their respective outputs and declining in distance between them. Deardorff (1995) shows how the basic Heckscher–Ohlin model of international trade can lead to a gravity specification for bilateral trade. Harrigan (2001) reviews the theoretical and empirical literature on gravity models, and similarly stresses the role played by relative as well as absolute transportation, and other trade and transactions costs. We also note that the following identity holds among the trade variables (also see Figure 2.1): Total imports = intermediate input imports + final good imports = IFUSA + IFFOR + ARML. We have also calculated distances to some of the larger countries, such as Russia, India and Australia, by taking into account the country’s size, rather than the distance to the largest city, with insignificant impact on the results. The two instruments were: (i) an index of competitiveness that took into account the investment climate and availability of skilled labour, among other measures, and (ii) the share of high-tech exports in total exports. It is useful at this point to note again that our import data correspond only to goods imports. We hope to consider trade flows in services at a later time. Specifically, NAICS 334 is defined as Computers and Electronic Product Manufacturing, and includes semiconductors, scientific instruments and telecommunications equipment. Components imported for sale in the aftermarket as spare parts do not count as intermediate goods but are considered final goods, since they are not used as inputs in production.

References Andersson, T. and T. Fredrikson (2000) ‘Distinction between Intermediate and Finished Products in Intra-Firm Trade’, International Journal of Industrial Organization, vol. 18(5), pp. 773–92. Arndt, S. W. and H. Kierzkowski (eds) (2001) Fragmentation: New Production Patterns in the World Economy (Oxford: Oxford University Press). Brainard, S. L. (1997) ‘An Empirical Assessment of the Proximity–Concentration Tradeoff between Multinational Sales and Trade’, American Economic Review, vol. 87(4), pp. 520–44.

40 On Intra-firm Trade and Multinationals Campa, J. and L. Goldberg (1997) ‘The Evolving External Orientation of Manufacturing Industries: Evidence from Four Countries’, Federal Reserve Bank of New York, Economic Policy Review, vol. 3(2), pp. 53–81. Cantwell, J. and G. D. Santangelo (1999) ‘The Frontier of International Technology Networks: Sourcing Abroad the Most Highly Tacit Capabilities’, Information Economics and Policy, vol. 11(1), pp. 101–24. Carkovic, M. and R. Levine (2005) ‘Does Foreign Direct Investment Accelerate Economic Growth?’, in Theodore Moran, Magnus Blömstrom and Edward M. Graham (eds), FDI Policy in Developing Countries: New Methods of Research and a Future Research Agenda (Washington, DC: Institute for International Economics). Carr, D. L., J. R. Markusen and K. Maskus (2001) ‘Estimating the KnowledgeCapital Model of the Multinational Enterprise’, American Economic Review, vol. 91(3), pp. 693–708. Chernotsky, H. I. (1987) ‘The American Connection: Motives for Japanese Foreign Direct Investment’, Columbia Journal of World Business, vol. 22(4), pp. 47–54. Deardorff, A. (1995) ‘Determinants of Bilateral Trade: Does Gravity Work in a Neoclassical World?’, NBER Working Paper no. 5377. Douglas, I. (1996) ‘The United States in a New World Economy? A Century’s Perspective’, American Economic Review, vol. 86(2), pp. 41–51. Feenstra, R. C. (1998) ‘Integration of Trade and Disintegration of Production in the Global Economy’, Journal of Economic Perspectives, vol. 12(4), pp. 31–50. Feenstra, R. C. and G. H. Hanson (1996) ‘Globalization, Outsourcing and Wage Inequality’, American Economic Review, vol. 86(2), pp. 240–5. Feenstra, R. C., J. Markusen and A. Rose (2001) ‘Using the Gravity Equation to Differentiate Among Alternative Theories of Trade’, Canadian Journal of Economics, vol. 34(2), pp. 430–47. Glass, A. J. and K. Saggi (2002) ‘Licensing versus Direct Investment: Implications for Economic Growth’, Journal of International Economics, vol. 56(1), pp. 131–53. Graham, E. M. and E. Wada (2001) ‘Foreign Direct Investment in China: Effects on Growth and Economic Performance’, Institute of International Economics Working Paper no. 01–03. Grossman, G. and E. Helpman (2002) ‘Outsourcing in a Global Economy’, NBER Working Paper no. 8728. Harrigan, J. (2001) ‘Specialization and the Volume of Trade: Do the Data Obey the Laws?’, New York Federal Reserve Staff Report no. 140. Helpman, Elhanan (1984) ‘A Simple Theory of International Trade with Multinational Corporations’, Journal of Political Economy, vol. 92(3), pp. 451–71. Hummels, D., D. Rapoport and K. M. Yi (1997) ‘Globalization and the Changing Nature of World Trade’, Federal Reserve Bank of New York, Economic Policy Review, vol. 3(2), pp. 53–81. Konan, D. E. (2000) ‘The Vertical Multinational Enterprise and International Trade’, Review of International Economics, vol. 8(1), pp. 113–25. Kuemmerle, W. (1999) ‘The Drivers of Foreign Direct Investment into Research and Development: An Empirical Investigation’, Journal of International Business Studies, vol. 30(1), pp. 1–24. Lieberman, M. B. (1991) ‘Determinants of Vertical Integration: An Empirical Test’, Journal of Industrial Economics, vol. 39(5), pp. 451–66. Lipsey, R. E. and M. Y. Weiss (1984) ‘Foreign Production and Exports of Individual Firms’, Review of Economics and Statistics, vol. 66(2), pp. 304–9.

Ashok Deo Bardhan and Dwight Jaffee 41 Madan, V. (2000) ‘Transfer Prices and the Structure of Intra-firm Trade’, Canadian Journal of Economics, vol. 33(1), pp. 53–68. Markusen, J. R. (1995) ‘The Boundaries of Multinational Enterprises and the Theory of International Trade’, Journal of Economic Perspectives, vol. 9(2), pp. 169–89. Markusen, J. R. and K. Maskus (2001) ‘Multinational Firms: Reconciling Theory and Evidence’, in M. Blomström and L. S. Goldberg (eds), Topics in Empirical International Economics: A Festschrift in Honor of Robert E. Lipsey (Chicago: University of Chicago Press), pp. 71–95. Markusen, J. R. and A. J. Venables (1995) ‘Multinational Firms and the New Trade Theory’, NBER Working Paper no. 5036. Mataloni, R. J. Jr., (1999) ‘U.S. Multinational Companies: Operations in 1997’, Survey of Current Business, July. Nair-Reichert, U. and D. Weinhold (2001) ‘Causality Test for Cross-country Panels: A New Look at FDI and Economic Growth in Developing Countries’, Oxford Bulletin of Economics and Statistics, vol. 63(2), pp. 153–72. OECD (2002) ‘Intra-Industry and Intra-Firm Trade and the Internationalization of Production’, Economic Outlook, vol. 71, ch. vi. Riker, D. and S. L. Brainard (1997) ‘US Multinationals and Competition from Low Wage Countries’, NBER Working Paper no. 5959. Roy, S. and J. M. Viaene (1998) ‘On Strategic Vertical Foreign Investment’, Journal of International Economics, vol. 46(2), pp. 253–79. Swenson, D. L. (2000) ‘Firm Outsourcing Decisions: Evidence from U.S. Foreign Trade Zones’, Economic Inquiry, vol. 38(2), pp. 175–89. Taylor, V. A. (2002) ‘Analytic Framework for Global Transfer-Pricing’, Journal of American Academy of Business, vol. 1(2), pp. 308–13 Wilamoski, P. and S. Tinkler (1999) ‘The Trade Balance Effects of U.S. Foreign Direct Investment in Mexico’, Atlantic Economic Journal, vol. 27(1), pp. 24–37. Zeile, W. J. (1997) ‘U.S. Intrafirm Trade in Goods’, Survey of Current Business, February. Zeile, W. J. (1999) ‘Foreign Direct Investment in the United States: Preliminary Results from the 1997 Benchmark Survey’, Survey of Current Business, August. Zhang, K. H. (2001) ‘Does Foreign Direct Investment Promote Economic Growth? Evidence from East Asia and Latin America’, Contemporary Economic Policy, vol. 19(2), pp. 175–85.

3 Foreign Direct Investment, Externalities and Economic Growth in Developing Countries: Some Empirical Explorations∗ Nagesh Kumar Research and Information System (RIS) for the Non-aligned and Other Developing Countries, New Delhi, India

and Jaya Prakash Pradhan School of Social Sciences, Jawaharlal Nehru University, New Delhi, India

1

Introduction

Foreign direct investment (FDI) emerged as the most important source of external resource flows to developing countries over the 1990s and has become a significant part of capital formation in those countries despite their share in the global distribution of FDI remaining small or even declining. FDI usually flows as a bundle of resources including, as well as capital, production technology, organizational and managerial skills, marketing know-how, and even market access through the marketing networks of multinational enterprises (MNEs) that undertake FDI. These skills tend to spill over to domestic enterprises in the host country. Therefore, FDI can be expected to contribute to growth (more than proportionately) compared to domestic investments in the ∗

This chapter represents a substantial development of a paper prepared initially for the World Bank, as a background paper for ‘Global Development Finance 2002’. Earlier versions have been presented at seminars at the United Nations University, Institute for New Technologies (UNU/INTECH); Maastricht; RIS; and at the International Economic Association Congress in Lisbon. We thank Lynn Mytelka, Shoko Negishi, K. L. Krishna and Monty Graham, among others, for their feedback and comments. The views expressed are personal and should not be attributed to RIS, the World Bank, or UNU/INTECH. 42

Nagesh Kumar and Jaya Prakash Pradhan 43

host country. There is now a body of literature that has analysed the effect of FDI on growth in inter-country frameworks and another analysing knowledge spillovers to domestic enterprises from MNEs (see, for example, De Mello, 1997; Kumar and Siddharthan, 1997; Saggi, 2000, for recent reviews of the literature). However, the mixed findings reached by these studies on the role of FDI inflows in host country growth and on knowledge spillovers from MNEs suggest that these relationships are not unequivocal. A major reason for expecting a more favourable effect of FDI on growth is the externality of MNE entry for domestic firms. But externalities such as spillovers may not take place in some cases because of poor linkages with the domestic enterprises or poor absorptive capacity, for example. FDI projects vary in terms of generation of linkages for domestic enterprises. There is also a possibility of MNE entry affecting domestic enterprises adversely, given the market power of their proprietary assets such as superior technology, appeal of brand names and aggressive marketing techniques. Therefore, FDI may crowd-out domestic investment and may thus be immiserizing (Fry, 1992; Agosin and Mayer, 2000). The crowding-out effect may be sharper when the technology gap between foreign and domestic firms is too wide to be bridged. Further, because FDI may be attracted to a country by high growth rates, among other factors, the observed relationships between FDI and growth rate may suffer from causality problems. Another problem that may have affected the existing studies is that they are all made within a comparative static framework, while the effect of FDI on domestic investment and growth is likely to be of a dynamic nature. There may be two rounds of effects of MNE entry on domestic investment. The initial round may be felt by domestic firms in the industry where the foreign entry has taken place. Because of a superior asset bundle brought by the foreign entrant, domestic enterprises may be affected adversely as their market share is eroded. The second round of effects may be more favourable, with domestic rivals absorbing spillovers of knowledge (demonstration-based learning) as well as diffusion of knowledge through vertical linkages with domestic enterprises. The net effect of FDI on domestic investments would depend on the relative weights of these two rounds of effects. Given the dynamic nature of the effect of FDI on domestic investment and growth, analysis in a comparative static framework may yield biased results. Against this backdrop, this chapter proposes to make some quantitative explorations into the nature of the relationships between FDI, domestic investment and growth, taking special note of the possible dynamic nature of the effects using a panel data set for 107 developing

44 FDI: Some Empirical Explorations

countries for the period 1980–99. The structure of the chapter is as follows: Section 2 presents some stylized facts regarding the mechanism of impact of FDI on growth in developing countries, and reviews the existing literature on the subject. Section 3 presents new empirical explorations into the relationships between FDI, growth and domestic investments in the framework of an augmented Solow model. It also examines the direction of causality between FDI and growth for the sample countries. Section 4 presents results of quantitative explorations into the relationship between FDI and domestic investment, which hold the key to its effect on growth. Finally, in Section 5, the chapter concludes with some remarks on policy lessons.

2 FDI inflows, externalities and growth: mechanisms of impact and evidence In the neo-classical model, growth results from technological progress, growth of the labour force – both of which are treated as exogenous – and capital accumulation that is subject to diminishing returns. However, new growth theories incorporate the role of knowledge or technology endogenously as a factor of production in its own right, and provide for the possibility of non-diminishing returns to capital (see Romer, 1994; Grossman and Helpman, 1991). The recognition of the role of knowledge in economic growth has also led to a renewed interest in the analysis of the role of FDI in growth. Romer (1993, p. 548) has argued that, by bringing new knowledge to their host countries, MNEs may help to reduce the ‘idea gaps’ between developed and developing countries that are sources of growth. Thus FDI’s effect on growth in host countries could be more valuable than its direct generation of output, by complementing domestic investments. The indirect effect of FDI on growth in the host country may comprise the sum total of its externalities on domestic investments through knowledge spillovers and vertical linkages. The externalities of FDI on a host economy include positive as well as negative effects. Among the positive externalities are vertical linkages and knowledge spillovers for domestic enterprises. A foreign entrant may generate demand for intermediate goods and may crowd-in domestic investment to deliver it. It may also help to diffuse new skills and knowledge brought into the host economy. As observed earlier, FDI inflows are generally accompanied by inward transfer of valuable resources such as technology, organizational capability, managerial skills and marketing know-how. The knowledge spillovers associated with FDI could be

Nagesh Kumar and Jaya Prakash Pradhan 45

classified into two broad categories, namely intra-industry spillovers and inter-industry spillovers. Intra-industry spillovers are absorbed by competitors of foreign entrants who are prompted to respond to new, improved processes or products introduced by technology-importing firms by upgrading their own technology. In certain cases the demonstration effect from foreign firms may speed up the diffusion of new technologies. Yet another source of spillovers could be through the increased competition from foreign entry which forces local firms to become more efficient users of existing technologies or to explore new ones. Among the mechanisms of technology spillovers of this sort are reverse engineering by competitors, increased rivalry through R&D and product development, and the mobility of employees trained in new technologies by foreign firms. Another mechanism of diffusion of technology imported within the host economy is through the generation of vertical inter-firm linkages. The suppliers and customers of foreign firms may benefit from the knowledge brought in the course of their dealings with it. MNEs may demand higher specifications, retooling and technology updating from their component suppliers, forcing technology effort on their part. In quite a few cases they may actually be passing on new designs, drawings and specifications that may be significant sources of technology diffusion. Similarly, certain elements of knowledge may be passed downstream to customers of foreign firms by embodiment in products. The diffusion of knowledge through this channel could be particularly significant in the case of equipment manufacturers. For example, a foreign investment spent on making more efficient looms may play an important role in diffusing the new technology within the textile industry of the host country. However, the most immediate externality of an MNE entry on domestic enterprises in the industry of the entrant is negative, as foreign entry erodes their market share (Markusen and Venables, 1997; Agosin and Mayer, 2000). In recent years, acquisition of domestic enterprises has become an increasingly popular form of MNE entry in some regions such as Latin America. In the case of acquisition, foreign entry can entirely crowd out domestic investment. Besides eroding the market share of domestic enterprise, foreign entry could have an adverse effect on domestic investment in the industry by its entry-raising conduct. It has been argued that MNE affiliates, with their dowry of intangible assets such as internationally-known brand names, captive access to technology and reservoirs of technical, managerial and organizational skills, are likely to pursue non-price modes of rivalry to maximize the revenue productivity of these assets. With the greater emphasis on

46 FDI: Some Empirical Explorations

Domestic firms in final goods industry

I2

I1

I3 B3

B

F

B1

F1

B2 I2

I3

A

I

F F1 I1 Domestic firms in intermediate goods industry

Figure 3.1 Source:

I

Effect of an MNE entry on domestic firms

Adapted from Markusen and Venables (1997).

product differentiation and other modes of non-price rivalry, the entry of new domestic firms to the industry is impeded by the ‘contrived entry barriers’ (see Kumar, 1990, 1991, for evidence). Therefore, MNE entry may crowd-out domestic investment in the industry of the entrant more than by merely eroding the market share of existing firms. Markusen and Venables (1997) provide a simple conceptual framework for analysing the effects of an MNE entry on the domestic investment in the host economy, as discussed above, in the framework of a twoindustry model, where one industry produces final consumer goods and another produces intermediate goods. MNE entry takes place in the finalgoods-producing industry. This is depicted in Figure 3.1, where the FF curve represents a locus of numbers of firms at which there are normal profits for the final goods industry. Above the curve FF there are more firms than can operate profitably, and below there is room for entry. Similarly II is a locus of normal profit for the intermediate goods industry and B represents a point of equilibrium (for more details, see Markusen and Venables, 1997). The entry of an MNE into a final goods industry first produces a competition effect – namely, it crowds out domestic enterprises. Therefore, the FF curve shifts downwards to F1 F1 . The second effect counters it by generating vertical linkages for the intermediate goods industry. The generation of backward linkages by the MNE entrant shifts the II curve to I1 I1 or I2 I2 depending on whether the MNE entrant

Nagesh Kumar and Jaya Prakash Pradhan 47

generates the demand for intermediate goods in the same proportion as domestic enterprises, or at a lower level. The new equilibrium will be at B1 or at B2 , respectively. Note that in the former case, there is no effect for the domestic intermediate goods industry, as B and B1 are at the same level on the horizontal axis. The domestic investment in intermediate goods shrinks in the second case – that is, at B2 . However, if the MNE entrant produces for export or substitutes the imports of final goods so that the FF curve does not shift (that is, no crowdingout of domestic firms), then the new equilibrium could be at B3 , which represents a net crowding-in through backward linkages. Therefore, the net effect of foreign entry for domestic investments depends on whether the foreign entrant produces for the domestic market, substitutes imports or produces primarily for exports, and whether it generates a similar amount of backward linkages as domestic firms, or lower. This simple framework highlights the fact that the nature of the FDI project has much to do with its effect on the host economy. That the extent of externalities generated by FDI depends on the nature of the project has been recognized elsewhere (Fry, 1992; De Mello, 1997). Kumar (2002) argues that MNE entry in modern-knowledgeintensive or intermediate goods industries may generate more favourable externalities for the host economy than those in matured consumer goods industries. Similarly, he expects export-orientated investments by MNEs, especially those having product mandates to serve third-country markets, to have more favourable externalities than do domesticmarket-orientated activities. He finds such so-called ‘quality’ FDIs more concentrated in fewer countries than FDI in general.

2.1

FDI, growth and domestic investment: empirical evidence

Although a number of studies have analysed the relationship between FDI inflows and economic growth, the issue is far from settled in view of the mixed findings reached. These studies have typically adopted a standard growth accounting framework to analyse the effect of FDI inflows on the growth of national income, along with other factors of production. A number of early studies generally reported an insignificant effect of FDI on growth in developing host countries. For example, Singh (1988), who found FDI penetration to have a little or no consequence for economic or industrial growth in a sample of seventy-three developing countries, or Hein (1992) reporting an insignificant effect of FDI inflows on mediumterm economic growth of per capita income for a sample of forty-one developing countries.

48 FDI: Some Empirical Explorations

Fry (1992) examined the role of FDI in promoting growth in the framework of a macro model for a pooled time series cross-section of sixteen developing countries for the period 1966–88. The countries included in the sample are Argentina, Brazil, Chile, Egypt, India, Mexico, Nigeria, Pakistan, Sri Lanka, Turkey, Venezuela, Indonesia, Korea, Malaysia, the Philippines and Thailand. For his sample as a whole he did not find that FDI exerted a significantly different effect from domestically-financed investment on the rate of economic growth, as the coefficient of FDI, after controlling for the gross investment rate, was not significantly different from zero in statistical terms. FDI had a significant negative effect on domestic investment, suggesting that it crowds-out domestic investment. Hence FDI appears to have been immiserizing. However, this effect varies across countries, and in the Pacific Basin countries FDI seems to have crowded-in domestic investment. Blomström, Lipsey and Zejan (1994) found that FDI inflows had a significant positive effect on the average growth rate of per capita income for a sample of seventy-eight developing and twenty-three developed countries. However, when the sample of developing countries was split between two groups based on level of per capita income, the effect of FDI on the growth of lower-income developing countries was not statistically significant, although still with a positive sign. They argue that least-developed countries learn very little from MNEs because domestic enterprises are too far behind in their technological levels to be either imitators or suppliers to MNEs. Borensztein, De Gregorio and Lee (1995) for a sample of sixty-nine developing countries for the period 1970–89 find that the effect of FDI on host-country growth is dependent on the stock of human capital. They infer from it that the flow of advanced technology brought by FDI can increase the growth rate only by interacting with a country’s absorptive capability. They also find that FDI stimulates total fixed investment more than proportionately. In other words, FDI ‘crowds-in’ domestic investment. However, the results are not robust across specifications. Balasubramanyam, Salisu and Sapsford (1996) find the effect of FDI on average growth rate for the period 1970–85 for a cross-section of forty-six countries as well as the sub-sample of countries that are deemed to pursue export-orientated strategies to be positive and significant, but not significant and sometimes negative for the sub-set of countries pursuing inward-orientated strategies. Pradhan (2001) finds a significant positive effect of lagged FDI inflows on growth rates only for Latin-American countries in a panel data estimation covering the 1975–95 period for seventy-one developing countries. The effect of FDI was not significantly different from zero for the overall sample and for other regions.

Nagesh Kumar and Jaya Prakash Pradhan 49

De Mello (1999) has conducted time series as well as panel data estimation for a sample covering fifteen developed and seventeen developing countries for the period 1970–90 of the relationships between FDI, capital accumulation, output and productivity growth. The time-series estimations suggest that the effect of FDI on growth or on capital accumulation and total factor productivity (TFP) varies greatly across countries. The panel data estimation suggests a positive impact of FDI on output growth for developed and developing country sub-samples. However, the effect of FDI on capital accumulation and TFP growth varies across developed (technological leaders) and developing countries (technological followers). FDI has a positive effect on TFP growth in developed countries and a negative effect in developing countries, but the pattern is reversed in the case of the effect on capital accumulation. De Mello infers from these findings that the extent to which FDI is growth-enhancing depends on the degree of complementarity between FDI and domestic investment. The degree of substitutability between foreign and domestic capital stocks appears to be greater in technologically-advanced countries than in developing countries. Developing countries may have difficulty in using and diffusing new technologies of MNEs. Findings of Xu (2000) for US FDI in forty countries for the period 1966–94 also corroborate the finding of De Mello, that technology transfer from FDI contributes to productivity growth in developed countries but not in developing countries, which he attributes to a lack of adequate human capital. Agosin and Mayer (2000) analyse the effect of lagged values of FDI inflows on investment rates in host countries, to examine whether FDI crowds-in or crowds-out domestic investment over the 1970–95 period. They find that FDI crowds-in domestic investment in Asian countries and crowds it out in Latin-American countries, while in Africa the relationship is neutral (or one-to-one between FDI and total investment). Therefore, they conclude that the effects of FDI are by no means always favourable, and simplistic policies are unlikely to be optimal. These regional patterns tend to corroborate the findings of Fry (1992), who also reported East Asian countries to have a complementarity between FDI and total investment. 2.2

Knowledge spillovers and productivity improvements

Other stream of studies have related the productivity levels across industries or firms within a country with the extent of foreign presence, in an attempt to evaluate the presence of knowledge spillovers, following Caves (1974). These studies have also resulted in mixed findings. Blomström (1989, ch. 4) has found a strong positive association between the labour productivity of local enterprises and the foreign share in

50 FDI: Some Empirical Explorations

employment in 1970 in Mexican manufacturing industries. However, foreign entry was not found to be related either to changes in the technological frontier or to changes in labour productivity of the least efficient plants. A simple relationship between productivity levels and foreign ownership as examined by these studies has the limitation of the potential overestimation of the positive impact of a foreign presence if the FDI was concentrated in more productive industries. Blomström and Wolff (1989), in further work on twenty 2-digit Mexican industries for the period 1965–84 found increasing convergence of the productivity levels of locally-owned firms to those of foreign-owned firms, thus suggesting the presence of knowledge spillovers. Haddad and Harrison (1993), in the case of Morocco’s manufacturing sector using a firm-level panel data set for 1985–9 found no significant relationship between higher productivity growth in domestic firms and greater foreign presence in a sector. Aitken and Harrison (1999), in a similar exercise for Venezuela, found that foreign ownership affected the productivity of domesticallyowned plants adversely, and the negative effects of FDI were large and robust. Similar results were reported for Indonesia, but the negative effect on the productivity of domestic plants was slightly weaker. Therefore, the authors inferred that the benefits of FDI are limited to direct effects on productivity improvements, with improved technology by enterprises receiving foreign participation. The spillovers to other local enterprises are negligible and do not justify the incentives granted by host governments to foreign investors. Kokko (1994), examining Mexican data, found no evidence of spillovers in industries where the foreign affiliates had a much higher productivity and larger market shares than local firms. In other industries, there appeared to be a positive relationship between a foreign presence and local productivity. This result suggests that spillovers from foreign enterprises are dependent on the local capability in the industry. If the local firms are too weak they will not be able to absorb spillovers and might disappear in the face of competition from foreign firms. Similar findings were obtained by Kokko, Tansini and Zejan (1996) in Uruguay, and Kathuria (1998) in India. The existing literature therefore suggests that the host country may not benefit from knowledge spillovers when the technology gap between foreign and domestic firms is too wide, as is generally the case in poorer countries. The literature also found the effect of FDI on growth to be dependent on the presence of skills that facilitate the absorption of new knowledge (Borensztein, De Gregorio and Lee, 1995; UNCTAD, 1999).

Nagesh Kumar and Jaya Prakash Pradhan 51

In view of the relatively low levels of skill accumulation, low-income countries are not able to experience the more favourable effects of FDI. Some studies have observed an insignificant or adverse effect of FDI on low-income countries and a more favourable effect on middle-income countries (Blomström, Lipsey and Zejan, 1994; De Mello, 1999; Xu, 2000). Therefore, not only is FDI concentrated in relatively richer countries, but these countries are also able to experience its more favourable effects than are poor countries.

3 FDI and growth in developing countries: new empirical evidence In what follows, we attempt some fresh explorations on the effect of FDI on growth with an up-to-date panel data set for a sample of ninety-eight developing countries covering the 1980–99 period, followed by tests on the direction of causality for sample countries. 3.1

Analytical framework

The effect of FDI on economic growth is analysed in the standard growth accounting framework. Initially, the capital stock is assumed to comprise two components, namely, domestic- and foreign-owned capital stock. So: Kt ≡ Kdt + Kft We adopt an augmented Solow production function that makes output a function of stocks of capital, labour, human capital and productivity (see Mankiw, Romer and Weil, 1992; Benhabib and Spiegel, 1994, among others). However, we specify domestic- and foreign-owned capital stock separately in a Cobb–Douglas type production function: β

γ

α λ Yit = Ait Kdit Kfit Lit Hit

(1)

where Y is the flow of output, Kd , Kf represent domestic- and foreignowned capital stock, respectively, L is labour and H is human skills capital stock. A is total factor productivity that explains the output growth not accounted for by the growth in the specified factors of production. Taking logs and differentiating Equation (1) with respect to time, we obtain the familiar growth equation: yit = αit + αkdit + λkfit + βlit + γ hit

(2)

52 FDI: Some Empirical Explorations

Lower-case letters represent the growth rates of output, domestic capital stock, foreign capital stock, labour and human capital, while α, λ, β and γ represent the output elasticity of domestic capital stock, foreign capital stock, labour and human skill capital, respectively. In a world of perfect competition and constant returns to scale these elasticity coefficients can be interpreted as respective factor shares in total output. Equation (2) is the fundamental growth accounting equation, which decomposes the growth rate of output into growth rate of total factor productivity plus a weighted sum of the growth rates of capital stocks, human capital stock and the growth rate of labour. Theoretically, α, β and γ are expected to be positive, while the sign of λ would depend on the relative strength of competition and linkage effects and other externalities that FDI generates in the development process, as discussed in Section 2. Following the established practice in the literature (see, for example, Bosworth and Collins, 1996), Kd and Kf are proxied by domestic investment to GDP ratio (Id ) and FDI to GDP ratio (If ), respectively, in view of problems associated with the measurement of capital stock. Graham (1995) has for example, pointed out the limitations in FDI representing capital formation. The justification for using the rate of investment comes from the assumption of a steady-state situation or a linearization around a steady state. Therefore, the final form of Equation (2) could be written as follows. yit = ai + αIdit + λIfit + βlit + γhit + εit

(3)

Where εit is an error term. 3.2

Data set and estimations

The data set for estimations covers 107 developing countries representing Africa, Asia, Latin America and the Caribbean for the period 1980–99 for most countries (see Appendix on page 68 for details). Because of missing values of certain variables, nine countries had to excluded from estimations, reducing the sample to ninety-eight. The data on the growth rate of GDP, gross investment rate, FDI to GDP ratio, and labour force are from World Development Indicators 2001 (World Bank, 2001). The domestic investment rate (Id ) is obtained by subtracting the FDI to GDP ratio from the gross investment rate. This avoids the double-counting of FDI in domestic investment from which a number of previous studies have suffered. The measurement of human skill stock has been a challenge,

Nagesh Kumar and Jaya Prakash Pradhan 53

with different studies employing literacy rates, gross enrolment rates or other measures of educational attainment as indicators of skills stock (see Benhabib and Spiegel, 1994; Barro and Sala-i-Martin, 1995, for problems in the measurement of skills). Here, human skill stock has been constructed by multiplying the mean years of schooling for the total population aged 25 years and over (Barro–Lee data set) by the total population in the year. The Barro–Lee database that provides the educational attainment variable has, however, two limitations. First, the average years of schooling are given for every five years and not annually. This has been resolved by interpolating the values for the intermediate years on the basis of the growth rate over the five-year period. The second problem with the data set relates to the country coverage. The Barro–Lee data set covers only sixty-five of the ninety-eight countries in our sample. Therefore, we report two sets of estimations: one for a sample of sixty-five countries with the full model, and a second with the complete sample of ninety-eight countries minus the human capital variable. Two methods of estimation have been employed. One is OLS with White’s heteroskedasticity consistent covariance estimation. The other is panel data estimation with fixed effects. In contrast to the random effect model, which treats country effects as random variables, the fixed effect assumes that the differences across countries reflect parametric shifts in the regression function. The Hausman test (1978) strongly favours the fixed effect estimation throughout. The estimations of Equation (3) for our data set are presented in Table 3.1. As observed above, two sets of results are presented, one for the sample of sixty-five countries for which we have the human skill variable (Equation 1.1 in columns 2 and 3), and the other for the full sample of ninety-eight countries without the skill variable (Equation 1.2 in columns 4 and 5). The estimations reported in Table 3.1 explain between 14 per cent and 33 per cent of total variation in the data set, which is reasonable considering the diverse cross-section covered, and are always significant at the 1 per cent level of confidence in terms of the F-test. The estimations suggest that the bulk of the growth of sample countries has been contributed by growth in the labour force, but also that domestic and foreign investments have contributed to it in significant measure. In particular, the finding pertaining to FDI is significant in view of previous studies, which resulted in mixed findings regarding the role of FDI in promoting growth, especially in poorer countries. The present estimations provide fresh support for the role that FDI potentially plays

54

Table 3.1 FDI and economic growth in developing countries, 1980–99 Independent variables

Coefficients (t-values) Equation 1.1 OLS

If Id Labour growth Skill growth If∗ DAfrica

0.2630∗∗∗ (5.39) 0.1805∗∗∗ (8.5) 0.8795∗ (1.73) 0.1051∗∗ (2.19)

Fixed effects

Equation 1.2 OLS

0.3721∗∗∗ 0.2495∗∗∗ (7.23) (11.88) 0.1906∗∗∗ 0.1903∗∗∗ (6.94) (6.53) 1.4131∗∗∗ 0.9507∗∗ (8.68) (2.31) 0.1264∗ (1.85)

Fixed effects 0.2561∗∗∗ (7.82) 0.2251∗∗∗ (9.13) 1.2527∗∗∗ (7.76)

If∗ DAsia If∗ DMiddle income Constant F-value R2 No. of countries No. of observations

−3.663∗∗∗ (−2.51) 26.04 0.1509 65 1204

−5.562∗∗∗ (−7.7) 47.08 0.2402 65 1204

−3.539∗∗∗ (−2.76) 47.55 0.1395 98 1640

−5.025∗∗∗ (−7.54) 56.40 0.3308 98 1640

Equation 1.3 OLS

Fixed effects

0.2382∗∗∗ (5.65) 0.1755∗∗∗ (7.93) 0.8916∗ 1.76 0.0960∗∗ (1.95) −0.0063 (−0.07) 0.1332∗ (1.63)

0.4647∗∗∗ (6.43) 0.1882∗∗∗ (6.85) 1.4238∗∗∗ (8.74) 0.1252∗ (1.83) −0.1695∗ (−1.67) −0.2317 (−1.23)

−3.564∗∗ (−2.42) 22.22 0.1524 65 1204

−5.497∗∗∗ (−7.58) 32.00 0.1974 65 1204

Equation 1.4 OLS

Fixed effects

0.2316∗∗∗ 0.4812∗∗∗ (5.68) (5.94) 0.1871∗∗∗ 0.2187∗∗∗ (6.3) (8.93) 0.9698∗∗ 1.2976∗∗∗ (2.36) (8.09) 0.0036 −0.304∗∗∗ (0.08) (−3.5)

0.1179∗ (1.72) −3.559∗∗∗ (−2.79) 29.33 0.1409 98 1640

0.2014 (1.43) −5.313∗∗∗ (−8.01) 39.75 0.1922 98 1640

Notes: Figures in parentheses are t -ratios; ∗∗∗ , ∗∗ and ∗ respectively indicate 1 per cent, 5 per cent and 10 per cent level of significance. Estimated using statistical package STATA 7.0. Source:

Authors’ computations as described in the text.

Nagesh Kumar and Jaya Prakash Pradhan 55

in fostering growth in its host countries. Furthermore, the impact of FDI is observed to be higher than the impact of domestic investment. For example, a 1 per cent increase in the FDI to GDP ratio is observed to lead to an increase in the growth rate of about 0.37 per cent whereas the increase is 0.19 per cent in the case of domestic investment (Equation 1.1 in Table 3.1). The growth of human skill stock also comes up with a positive and significant effect, suggesting that the accumulation of skills does contribute significantly to growth.

3.3

Shifts across regions and income levels

There are reasons to believe that the contribution of FDI to economic growth varies across different developing regions. First, the composition of FDI inflows into developing countries is highly regionally concentrated in Latin America and Asia (Dunning, 1998; Lipsey, 1998). In the 1990s, the concentration of FDI in Latin America was further strengthened because of its strong progress towards regional economic integration, privatization, debt-equity swaps and so on. Africa, given its poor levels of development and its political instability, has remained less attractive vis-à-vis other developing regions. Second, there is considerable variation in the patterns of FDI across regions. For example, a greater proportion of FDI in Latin America than in other regions has come in recent years through privatization and the acquisition of existing enterprises. Regions also differ with respect to policies towards facilitating the vertical linkages of MNEs with domestic enterprises. UNCTAD (2001), for example, provides case studies of policies adopted by some East Asian countries to foster linkages between foreign and domestic enterprises. Finally, previous studies such as that by Fry (1992) showed that FDI was more productive in Pacific Basin countries than in others. To examine any systematic differences in the role played by FDI in the explanation of growth across regions, two differential slope dummies have been included in the estimation (Equation 1.3 in Table 3.1), with Latin-American countries being treated as the base category. The estimations suggest that FDI has been less productive in African developing countries than in Latin-American countries, as the dummy for Africa is negative and statistically significant at the 10 per cent level. The differential slope dummy for Asia has a positive and significant coefficient in the OLS estimation, suggesting that the growth impact of FDI is statistically more favourable in the Asian region than in Latin America. However, the coefficient is not significantly different from zero in the panel data estimation.

56 FDI: Some Empirical Explorations

Further, as is argued above, FDI may be less productive in lowincome countries in view of the technology gap or a lack of capacity among domestic enterprises to absorb favourable knowledge spillovers. To examine this, a differential slope dummy for middle-income countries was included, with low-income countries treated as the base category in Equation 1.4 in Table 3.1. In the OLS estimations the middle-income dummy achieves a positive coefficient that is significant at 10 per cent level, but is not significantly different from zero in fixed-effect estimations. It would appear to lend weak support to the hypotheses that FDI is more productive in middle-income countries than in low-income countries. To sum up, therefore, the panel data estimations reported above highlight a positive effect of FDI on their host country’s growth, along with other factors of production such as labour deployment, domestic capital and skill accumulation. Although the results of the estimations are quite robust across different estimations and specifications, these do suffer from some limitations. For example, as argued in Section 2, the FDI-togrowth relationship may suffer from causality bias. That is, rather than causing growth, the observed relationship might be a result of growth attracting FDI. Further, FDI’s effect on growth was posited to be of a dynamic nature comprising two rounds of effects. A contemporaneous analysis in the growth-accounting framework might have limitations in capturing fully the dynamic effects. Finally, the literature suggests that the effect of FDI on growth varies across countries depending on, among other factors, the nature of the effect on domestic investment, and backward linkages and knowledge spillovers generated, which in turn are determined by the nature of the FDI received, the local absorptive capacity, and the technology gap between domestic and foreign enterprises (see Fry, 1992; De Mello, 1999). In what follows, further tests are conducted to determine the direction of causality between FDI and growth. 3.4

FDI and growth: which comes first?

To further understand the relationship between FDI and economic growth, and to resolve the possible causality bias between FDI and growth, we have used the Granger causality test in a bivariate VAR framework. FDI would be considered ‘Granger-causing growth’ only if the lagged values of FDI contribute significantly to the explanation of current growth. Therefore, this test essentially looks at the predictive performance between variables, to determine the existence or direction of causality between them. Given the fact that it takes into account the

Nagesh Kumar and Jaya Prakash Pradhan 57

effect of lagged values of the causing variable on the current value of the dependent variable, it takes care of the dynamic nature of FDI’s effect on growth that we have postulated. A Granger causality test was performed for all the eighty-one countries in the sample for which adequate observations are available. The Schwartz information criterion (SC) has been used to determine the optimal lag length in the test. The detailed findings are reported in Appendix Table A3.1 on page 69, and summarized in Table 3.2 here. Out of 81 countries included, the causality test between FDI and economic growth suggests existence of causality only in the case of 28 countries. Unidirectional causality from FDI to economic growth was observed in 12 countries. The growth rate is found to attract FDI in 11 countries. Feedback causality, i.e. two-way interaction, has been detected in five countries. In the other 53 countries the direction of causality is not pronounced and hence the Granger test is not able to determine its direction. This analysis therefore suggests that the nature of the FDI–growth relationship varies across countries. The predominant pattern emerging from this analysis is that the direction of causality in the majority of cases could not be determined, and even in the remaining cases it is not always from FDI to growth. Therefore, caution needs to be applied before drawing definitive inferences from the findings of panel data estimations on the role played by FDI in determining growth. FDI is attracted by growth in an equal number of countries as the reverse.

4 FDI and domestic investment: complements or substitutes? The above analysis suggests that the effect of FDI on growth varies across countries, and that in some countries growth may in fact be pulling the FDI rather than the FDI contributing to growth. Hence the findings of panel data estimations need to be read with caution. Conceptually, we expect the effect of FDI to differ from that of domestic investment because of the potential of FDI to crowd-in domestic investment through vertical linkages, or to improve their productivity through knowledge spillovers, or to substitute for domestic investment. The effect of FDI on growth would be more favourable than domestic investment if it crowds-in more domestic investment than it crowds-out. We have also posited that these effects might have a dynamic dimension. In this section, we explore the nature of the relationship between FDI and domestic investment.

Unidirectional causality FDI → growth

Growth → FDI

Cameroon Colombia Guinea-Bissau Jamaica Mexico Paraguay Senegal St. Lucia Swaziland Trinidad and Tobago Uruguay Zambia

Argentina Belize Congo Republic Congo Democratic Republic Ecuador El Salvador Guatemala Guyana Mauritania Tunisia Kenya

Source:

Based on Appendix Table A3.1.

58

Table 3.2 Causality between FDI and economic growth, 1980–99 Feedback causality FDI ↔ growth Côte d’Ivoire Indonesia Malawi Pakistan Thailand

Granger neutral

Bangladesh, Barbados, Benin, Bolivia, Botswana, Brazil, Burkina Faso, Burundi, Central African Republic, Chad, Chile, China, Comoros, Costa Rica, Cyprus, Dominica, Dominican Republic, Egypt Arab Republic, Ethiopia, Fiji, Gabon, Gambia, Grenada, Ghana, Haiti, Honduras, India, Republic of Korea, Lesotho, Madagascar, Malaysia, Mali, Mauritius, Morocco, Mozambique, Nepal, Niger, Nigeria, Panama, Peru, Philippines, Rwanda, Seychelles, Sierra Leone, Singapore, Solomon Islands, Sri Lanka, St. Kitts and Nevis, St. Vincent and the Grenadines, Turkey, Vanuatu, Venezuela, Zimbabwe

Nagesh Kumar and Jaya Prakash Pradhan 59

The nature of FDI and domestic investment is first examined in the framework of a simple model in which the current values of domestic investment are made a function of current and past values of FDI, as well as lagged values of it itself (dependent variable) and lagged growth variable. That is: Id,it = λi0 + λ1 Id,it−1 + λ2 Id,it−2 + λ3 If ,it + λ4 If ,it−1 + λ5 If ,it−2 + λ6 gy,it−1 + εit

(4)

where Id and If are domestic investment and FDI, both expressed as a percentage of GDP of the host economy, and gy,it−1 is the lagged growth rate; λi0 is the country effect and is assumed to be time invariant; εit is the classical disturbance term. The inclusion of present and lagged values of FDI in Equation (4) enables us to capture the possibly dynamic nature of the effect of FDI on domestic investment, as argued above. In particular, we posit that the initial effect of FDI on domestic investment may be negative, because it erodes the market share of domestic investors. However, in the subsequent period, it could have a positive effect on domestic investment through the generation of backward linkages. 4.1

Dynamic panel data estimations

The inclusion of lagged dependent variables in the specification makes Equation (4) a dynamic panel data model. For such models, conventional estimation techniques, namely OLS and panel data (both fixedand random-effects), are not appropriate. The OLS estimates are biased and inconsistent, because the lagged dependent variables are correlated with the error term, violating a fundamental assumption. In the typical panel data setting with large cross-sections and short time series such as the present one, the fixed-effects estimator is both biased and inconsistent because the Within transformation wipes out the individual effects but does not resolve the problem of a correlation between the differenced lagged variable and the error term. Anderson and Hsiao (1982) suggested an instrumental variable (IV) method for the estimation of dynamic panel data models. The IV estimates, even though consistent, however, are not efficient, as they do not utilize all the available moment conditions. Arellano and Bond (1991) have proposed one-step and two-step generalized method of moments (GMM) framework that utilizes the orthogonality conditions that exist between the lagged values of a dependent variable and the disturbances. The method takes the first-difference of the model to eliminate the individual effects, and then

60 FDI: Some Empirical Explorations Table 3.3 Estimations capturing the effect of FDI on domestic investment Independesnt variables

Coefficients (t-values) Equation 1.1 OLS estimation

Id,t−1 Id,t−2 If ,t If ,t−1 If ,t−2 gy,t−1 Constant F-value Sargan test: Chi-square Serial correlation of 1st order Serial correlation of 2nd order R2 No. of countries No. of observations

0.6345∗∗∗ (10.89) 0.1811∗∗∗ (4.05) −0.6579∗∗∗ (−15.6) 0.4774∗∗∗ (7.29) 0.2156∗∗∗ (3.13) 0.1187∗∗∗ (4.01) 3.0387∗∗∗ (5.8) 227.97

0.7207 107 1667

Arellano–Bond GMM dynamic panel data estimationa 0.3503∗∗∗ (44.46) 0.0131∗∗∗ (4.2) −0.6464∗∗∗ (−114.85) 0.2830∗∗∗ (43.52) 0.1731∗∗∗ (50.87) 0.0457∗∗∗ (12.81)

99258.95 98.94 −2.96 0.97 0.6290b 107 1559

Notes: Figures in parentheses are t -ratios; ∗∗∗ , ∗∗ and ∗ indicate, respectively, 1 per cent, 5 per cent and 10 per cent level of significance. a Using the Arellano–Bond two-step GMM estimation estimated using STATA 7.0. In the estimation, the current period FDI has been treated as a predetermined variable rather than a strictly exogenous variable. b Obtained from the one-step GMM differenced residuals. Source: Authors’ computations as described in the text.

estimates it by using two or higher period lagged dependent variables as instruments, following Hansen’s optimal GMM framework (see Baltagi, 1995, for more details). We follow the Arellano–Bond GMM method for the estimation of Equation (4). The estimation results using OLS (with White’s heteroskedasticity consistent covariance estimation) as well as the Arellano–Bond GMM estimations are summarized in Table 3.3. For these estimations, data are available for all the 107 developing countries in our panel for the period 1980–99. Irrespective of the techniques used, the estimated

Nagesh Kumar and Jaya Prakash Pradhan 61

models are found to be highly significant in terms of the F-test, and explain over 70 per cent of variation in the dependent variable in the case of OLS and over 63 per cent in the GMM estimation. The Sargan test from the two-step estimator cannot reject the null hypothesis that the over-identifying restrictions are valid. Further, it is not possible to reject the null hypothesis of no second-order autocorrelation, which suggest that the obtained estimates are consistent. It can be seen that FDI inflows in current period and in the previous two years have a significant effect on domestic investment in the current year besides lagged domestic investment and lagged growth rate. However, the signs of the effect of FDI inflows in current period and the two past years are different. FDI in the current period has a strong negative effect on domestic investment, while the lagged inflows have a positive effect. Keeping in mind the first-difference form of the model in dynamic panel data estimation, the following interpretation can be provided. A 1 per cent increase in the FDI ratio in the current period decreases, on average, current domestic investment ratio by 0.65 per cent. A 1 per cent increase in the FDI ratio in the previous two years, on average, is followed by an increase of about 0.28 per cent and 0.17 per cent in the current period domestic investment ratio, respectively. The pattern observed tends to corroborate our proposition that the effect of FDI on domestic investment is of a dynamic nature, and that the nature of effects may differ over time. The important point to note here is that, irrespective of the method of estimation, the negative effect of current FDI is larger than the positive effect of FDI in previous years. Therefore, it appears that for the sample covered here, the crowding-in effect of lagged FDI is weaker than the crowding-out effect of current period FDI. 4.2

Country-wise estimations

While the estimations for the panel of ninety-eight countries reported above do indicate that, on the whole, crowding-out dominates the effect of FDI on domestic investment, there may be important intercountry differences depending on their ability to attract FDI of better ‘quality’ – namely those that generate more favourable externalities for domestic investment. Hence, Equation (4) was re-estimated for each of the eighty-three countries for which we had at least seventeen observations. The optimal lag structure was determined with the help of the Schwartz criterion. The results of estimations are presented in the Appendix Table A3.2 on page 69. Table 3.4 summarizes the findings of these estimations for fifty-two countries for which the FDI variable had

62 Table 3.4 Countries with a significant coefficient of FDI in the investment equation Country

Sign of the coefficient of FDI (t)

Argentina Bangladesh Barbados Belize Bolivia Botswana Brazil Burkina Faso Cameroon Chad∗ Chile Colombia Costa Rica Côte d’Ivoire Cyprus Dominica Ecuador El Salvador Fiji Gambia Ghana Grenada Guyana Haiti Honduras India∗ Jamaica Korea, Republic of Lesotho Mali Mauritania Mauritius Mexico Morocco Nepal Nigeria Panama Papua New Guinea Paraguay Peru Philippines

FDI (t − 1)

−1.581 4.996

−1.059 −0.86 −1.0538

0.64 6.4846 0.962 1.575 0.8613 1.455

−4.373 0.961 3.682 −0.7966 −0.765 −0.826 −0.6966 1.42 −1.121 −0.929

−1.3717 −12.4973 −0.8213 −1.6768

FDI (t − 2) 2.059

5.374 −1.4231

−2.3571 0.5392 0.319 0.7869 −1.109 1.073 0.916 8.0696 3.0889 −5.2697 1.5725 17.3632

1.1499 −1.644 −0.499 −4.351 0.912

−1.773 3.27 −1.018

−1.088

1.6746 3.567 −1.635 14.8534 −1.0687 −0.6461 −0.9432 −1.255 −1.2983 −1.5022

−2.513

0.8545

Net effect of FDI

Crowding in Crowding in Crowding in Crowding out Crowding out Crowding out Crowding out Crowding in Crowding in Crowding in Crowding out Crowding out Crowding out Crowding out Crowding in Crowding out Crowding out Crowding out Crowding out Crowding in Crowding in Crowding out Crowding out Crowding in Crowding in Crowding out Crowding in Crowding in Crowding out Crowding out Crowding in Crowding in Crowding out Crowding out Crowding in Crowding out Crowding in Crowding out Crowding out Crowding out Crowding out

Nagesh Kumar and Jaya Prakash Pradhan 63 Table 3.4 Continued Country

Sign of the coefficient of FDI (t)

Rwanda∗ Senegal Sierra Leone Singapore Sri Lanka St. Kitts and Nevis St. Lucia∗ Swaziland Thailand Uganda Uruguay

FDI (t − 1)

FDI (t − 2)

4.539 −0.959 −1.067 −0.7589 −0.725 −2.367 −2.6068 −1.9627 −1.9017

0.75 1.072 0.9511 1.2265 0.5047 0.656 3.2418 1.7744 2.1955

Net effect of FDI

−0.404

Crowding in Crowding in Crowding in Crowding out Crowding in Crowding out Crowding out Crowding out Crowding in Crowding out Crowding in

Notes: ∗ Denotes cases where the estimated model is not significant even at the 10 per cent level. Blank cells indicate that the estimated coefficient is not significantly different from zero in statistical terms. Source:

Authors’ calculations from Table A.1.

a significant coefficient. In other cases, FDI did not have a significant effect on domestic investments. The general pattern of effects of FDI on domestic investment is that current values of FDI have a negative effect on domestic investment in the current period, whereas a positive effect dominates the relation with one period lag. This is exactly the pattern observed in panel data estimations. By taking into account the sign and magnitude of the FDI coefficient for the current and lagged periods, the nature of the net effect of FDI can be determined for each country. Of the fifty-two countries that are shown in Table 3.4 to have a significant coefficient of FDI, twenty-nine experience a net crowding-out effect from FDI and twenty-three experience the reverse. In order to examine regional patterns of the nature of the effect of FDI on domestic investment, we cross-tabulate in Table 3.5 the countries according to their regions and the nature of the effect. Crowdingout appears to dominate the relationship between FDI and domestic investment in the Latin America and Caribbean region, with seventeen countries in this group, and only seven report a crowding-in. In Asia and Africa, the patterns of crowding-in and crowding-out are distributed more evenly. These regional patterns appear to be consistent with the observations of Fry (1992) and Agosin and Meyer (2000).

Asia

Africa

Latin America and the Caribbean

Crowding in

Bangladesh Korea, Republic of Nepal Sri Lanka Thailand

Argentina Barbados Haiti Honduras Jamaica Panama Uruguay

Crowding out

Fiji India∗ Papua New Guinea Philippines Singapore

Burkina Faso Cameroon Chad∗ Gambia Ghana Mauritania Mauritius Rwanda∗ Senegal Sierra Leone Botswana Côte d’Ivoire Lesotho Mali Morocco Nigeria Swaziland Uganda

Belize Bolivia Brazil Chile Colombia Costa Rica Dominica Ecuador El Salvador Grenada Guyana Mexico Paraguay Peru St. Kitts and Nevis St. Lucia∗

64

Table 3.5 Summary patterns of relationships between FDI and domestic investment

FDI coefficients not significantly different from zero

Notes: Source:

China Indonesia Malaysia Pakistan Turkey

Algeria Benin Burundi∗ Central African∗ Republic Comoros Congo Democratic Republic Congo Republic∗ Egypt, Arab Republic∗ Ethiopia∗ Gabon∗ Guinea-Bissau Kenya Madagascar∗ Malawi∗ Mozambique Niger Seychelles∗ Togo∗ Tunisia Zambia∗ Zimbabwe∗

Dominican Republic∗ Guatemala St. Vincent and the Grenadines Trinidad and Tobago Venezuela∗

∗ Denotes cases where the estimated model is not significant even at the 10 per cent level.

As for Table 3.4.

65

66 FDI: Some Empirical Explorations

5

Concluding remarks and policy implications

In this chapter we have analysed the relationships between FDI, growth and domestic investment for a sample of 107 developing countries for the 1980–99 period. It has been argued that the effect of FDI on growth could be of a dynamic nature in that there may be two rounds of effects, namely a competition effect for domestic enterprises in the industry of the foreign entrant, which is generally negative, and a subsequent round that could include a generally favourable externality on domestic investment because of backward linkages. The net weight of these effects would depend on the nature of FDI projects or the quality of FDI, which is known to vary greatly in different types of investments. Our panel data estimations in the standard growth accounting framework using Solow’s augmented production function suggest a significant positive effect of FDI on growth in developing countries. However, this finding should be treated with caution, given the causality bias and limitations of a contemporaneous estimation in capturing a possibly dynamic relationship. The tests of causality suggest that in a majority of cases the direction of causation is not pronounced. Furthermore, in poor countries, the direction of causation seemed to be running from growth to FDI in as many cases as from FDI to growth. Thus, in a substantial number of cases, the growth rate of an economy acts as a signalling mechanism for FDI. Given the fact that the nature of the relationship between FDI and domestic investment is at the heart of the former variable’s effect on growth, we analysed the effect of current and lagged values of FDI on domestic investment in each current year. The findings of these estimations corroborate our proposition that FDI affects domestic investments in a dynamic manner, with the initial effect being negative and the subsequent effect positive for the panel data, as well as for most of the countries individually. Furthermore, in net terms, the effect of FDI on domestic investments appears to be negative for the pooled estimations as well as for the majority of countries. So, in general, FDI appears to crowd-out domestic investment in net terms. However, for a number of countries, FDI is seen as crowding-in domestic investment. Therefore, some countries have been able to benefit from FDI more than others. A more detailed examination of the factors that explain the greater success of some countries in experiencing more favourable effects of FDI is clearly warranted. What are the policy lessons from the above analysis for poorer developing countries? The finding that causality runs from growth to FDI

Nagesh Kumar and Jaya Prakash Pradhan 67

in a substantial proportion of cases suggests that the poorest countries need to pursue alternative strategies to get their process of development going, rather than to wait for MNE investment to stimulate the process of their industrialization and development with incentives and policy liberalization. They would do better to focus on human resources, improving infrastructure, developing local entrepreneurship, and creating a stable macroeconomic framework and conditions conducive to productive investments to kick-start the process of development. The restoration of concessional development finance should help in complementing the meagre domestic resources that can be invested. Once the pace of development picks up, FDI will probably be attracted and help in carrying the process forward. Building absorptive capacity will facilitate the absorption of the knowledge brought in by FDI. It is clear that the effects of FDI on domestic investments and growth depend very much on the nature or quality of FDI. Certain types of FDI tend to have more favourable developmental externalities than others. In this context, attention needs to be paid by host countries to the quality of FDI inflows besides attracting greater magnitudes of FDI. Recent work has shown that host-country policies have an important bearing on the quality of FDI inflows received (see Kumar, 2002, among others). Governments have employed various measures to improve the overall quality of FDI inflows, including selective policies to target more desirable FDI inflows. Many governments – in developed as well as developing countries – have imposed performance regulations such as local-content requirements on MNEs to intensify the generation of local linkages or export obligations to ‘[trigger] a burst of export-focused investments’ (see Moran, 1998; Kumar, 2003, for examples). Some of the performance requirements, such as local content requirements, have since been phased out under TRIMs Agreements of the WTO, but some others such as export obligations can still be imposed. Some countries have employed incentives such as pioneer industry programmes to attract FDI in industries that have the potential to generate more favourable externalities for domestic investment (see UNCTAD, 1999, 2001, for examples). Similarly, because MNE entry through the acquisition of domestic enterprises is likely to generate less favourable externalities for domestic investment than are greenfields investments, some governments discourage acquisitions by foreign enterprises (see Agosin and Mayer, 2000, for examples). Another sphere where governmental intervention may be required to maximize gains from globalization is in the diffusion of knowledge

68 FDI: Some Empirical Explorations

imported by foreign enterprises. An important channel of knowledge diffusion brought in by MNEs in the host economy is vertical inter-firm linkages with domestic enterprises. Host governments could consider employing proactive measures that encourage foreign and local firms to deepen their local content, as a number of countries (for example, Singapore, Taiwan, Korea and Ireland) have done so successfully (see Battat et al., 1996; UNCTAD, 2001). Finally, an implication of the above discussion is the importance of preserving the policy flexibility for host governments to improve the quality of FDI in multilateral trade negotiations.

Appendix: data sources and variable measurements The data on GDP growth (annual percentage), gross fixed capital formation (percentage of GDP), net FDI inflows (percentage of GDP) and total labour force have been obtained from the World Bank’s World Development Indicators 2001. The mean years of schooling has been collected from the Barro–Lee Human Capital Appendix, Table A3.2. Variables used in the study: yit is the growth rate of GDP (annual percentage); Id,it is the domestic investment rate. Obtained as the difference between the total investment rate and the FDI ratio of the host economy; If ,it is the FDI ratio. It is defined as the net FDI inflows as a percentage of the GDP of the host country; lit is the growth rate of labour force (annual percentage); and hit is the growth rate of human capital stock. This variable, as mentioned above, is constructed as the mean years of schooling multiplied by the population of the country. In the Barro–Lee data set, mean years of schooling are available only at five-year intervals. For intermediate years, this study calculated an arithmetical growth rate of the form: Sh = [((Ht+5 − Ht )/Ht )∗ (1/5)] for mean years of schooling and filled in the missing cells. This methodology is consistent with the methodology of mid-year estimation of population adopted worldwide. Sample coverage: The sample covers 107 developing countries from Africa, Asia, and Latin America and the Caribbean. The period covered is 1980–99. However, for some countries data are available for fewer years.

Table A3.1

Granger causality between FDI and economic growth

Country

Null hypothesis

Argentina

FDI dngc growth Growth dngc FDI FFDI dngc growth Growth dngc FDI FDI dngc growth Growth dngc FDI FDI dngc growth Growth dngc FDI FDI dngc growth Growth dngc FDI FDI dngc growth Growth dngc FDI FDI dngc growth Growth dngc FDI FDI dngc growth Growth dngc FDI FDI dngc growth Growth dngc FDI FDI dngc growth Growth dngc FDI FDI dngc growth Growth dngc FDI FDI dngc growth Growth dngc FDI

Bangladesh Barbados Belize Benin Bolivia Botswana Brazil Burkina Faso Burundi Cameroon Central African Republic

Obs.

Number of lags

F-statistic

Probability

16 16 15 15 19 19 19 19 18 18 15 15 17 17 19 19 15 15 18 18 15 15 15 15

4 4 5 5 1 1 1 1 2 2 5 5 3 3 1 1 5 5 2 2 5 5 5 5

0.86109 2.78975 1.18812 1.08368 0.01724 0.66855 0.51942 7.25640 2.16644 0.63803 1.77561 0.59889 1.27003 0.43915 0.39344 0.62399 2.47322 1.30007 1.22198 1.47023 7.17057 1.32717 0.51271 1.39778

0.53094 0.11206 0.44618 0.48243 0.89717 0.42557 0.48149 0.01597 0.15416 0.54410 0.29897 0.70862 0.33682 0.72999 0.53935 0.44112 0.20050 0.41131 0.32636 0.26569 0.03973 0.40343 0.75941 0.38387

Conclusion Growth→FDI∗ Granger neutral Granger neutral Growth→FDI Granger neutral Granger neutral Granger neutral Granger neutral Granger neutral Granger neutral FDI→Growth Granger neutral

69

Continued

Continued

70

Table A3.1 Country

Null hypothesis

Chad

FDI dngc growth Growth dngc FDI FDI dngc growth Growth dngc FDI FDI dngc growth Growth dngc FDI FDI dngc growth Growth dngc FDI FDI dngc growth Growth dngc FDI FDI dngc growth Growth dngc FDI FDI dngc growth Growth dngc FDI FDI dngc growth Growth dngc FDI FDI dngc growth Growth dngc FDI FDI dngc growth Growth dngc FDI FDI dngc growth Growth dngc FDI FDI dngc growth Growth dngc FDI FDI dngc growth Growth dngc FDI

Chile China Colombia Comoros Congo Republic Congo Democratic Republic Costa Rica ˆ d’Ivoire Cote Cyprus Dominica Dominican Republic Ecuador

Obs.

Number of lags

F-statistic

Probability

13 13 17 17 19 19 15 15 18 18 15 15 14 14 17 17 19 19 19 19 19 19 19 19 15 15

5 5 3 3 1 1 5 5 1 1 5 5 5 5 3 3 1 1 1 1 1 1 1 1 5 5

7.60652 0.66739 1.28365 0.35388 0.49255 1.72484 4.10456 2.65268 0.29381 0.23655 0.26409 4.20146 0.38865 11.3386 1.31129 0.23472 3.18900 3.33134 0.03257 0.17527 0.37563 2.04492 1.15932 1.03687 0.76995 10.3374

0.12026 0.69087 0.33266 0.78746 0.49287 0.20759 0.09801 0.18290 0.59575 0.63374 0.91164 0.09458 0.83288 0.03654 0.32440 0.87015 0.09310 0.08670 0.85904 0.68103 0.54856 0.17195 0.29757 0.32370 0.61698 0.02097

Conclusion

Granger neutral Granger neutral Granger neutral FDI→Growth Granger neutral Growth→FDI Growth→FDI Granger neutral FDI ↔ Growth Granger neutral Granger neutral Granger neutral Growth→FDI

Egypt Arab Republic El Salvador Ethiopia Fiji Gabon Gambia Grenada Ghana Guatemala Guinea-Bissau Guyana Haiti Honduras India

FDI dngc growth Growth dngc FDI FDI dngc growth Growth dngc FDI FDI dngc growth Growth dngc FDI FDI dngc growth Growth dngc FDI FDI dngc growth Growth dngc FDI FDI dngc growth Growth dngc FDI FDI dngc growth Growth dngc FDI FDI dngc growth Growth dngc FDI FDI dngc growth Growth dngc FDI FDI dngc growth Growth dngc FDI FDI dngc growth Growth dngc FDI FDI dngc growth Growth dngc FDI FDI dngc growth Growth dngc FDI FDI dngc growth Growth dngc FDI

19 19 15 15 13 13 19 19 18 18 15 15 15 15 16 16 16 16 15 15 19 19 19 19 19 19 16 16

1 1 5 5 5 5 1 1 1 1 4 4 5 5 4 4 4 4 5 5 1 1 1 1 1 1 4 4

0.00938 1.20982 0.75399 11.1952 4.78835 2.09977 0.87121 0.43705 0.00587 0.29490 0.33771 2.81828 1.22387 0.29303 1.27125 0.05621 2.11801 3.09723 44.9040 0.84221 1.72872 4.55976 1.07701 0.14570 1.46714 0.01629 2.17488 0.36002

0.92407 0.28764 0.62498 0.01818 0.18174 0.35334 0.36449 0.51796 0.93994 0.59508 0.84361 0.12397 0.43462 0.89469 0.36572 0.99276 0.18167 0.09151 0.00132 0.58225 0.20711 0.04854 0.31480 0.70770 0.24338 0.90003 0.17398 0.82982

Granger neutral Growth→FDI Granger neutral Granger neutral Granger neutral Granger neutral Granger neutral Granger neutral Growth→FDI FDI→Growth Growth→FDI Granger neutral Granger neutral Granger neutral 71

Continued

Continued

72

Table A3.1 Country

Null hypothesis

Indonesia

FDI dngc growth Growth dngc FDI FDI dngc growth Growth dngc FDI FDI dngc growth Growth dngc FDI FDI dngc growth Growth dngc FDI FDI dngc growth Growth dngc FDI FDI dngc growth Growth dngc FDI FDI dngc growth Growth dngc FDI FDI dngc growth Growth dngc FDI FDI dngc growth Growth dngc FDI FDI dngc growth Growth dngc FDI FDI dngc growth Growth dngc FDI FDI dngc growth Growth dngc FDI FDI dngc growth Growth dngc FDI

Jamaica Kenya Korea, Republic of Lesotho Madagascar Malawi Malaysia Mali Mauritania Mauritius Mexico Morocco

Obs.

Number of lags

F-statistic

Probability

15 15 19 19 19 19 19 19 18 18 18 18 16 16 16 15 19 19 17 17 16 16 15 15 15 15

5 5 1 1 1 1 1 1 1 1 2 2 4 4 5 5 1 1 3 3 4 4 5 5 5 5

3.97237 4.65320 2.78987 0.02609 0.55010 7.07330 0.04370 1.02613 0.01248 0.09040 1.12079 0.88676 4.57002 2.84920 0.53972 0.44451 0.79501 0.29637 1.15596 2.88033 2.09364 2.57050 17.4180 0.10424 3.55836 3.04920

0.10299 0.08073 0.11431 0.87370 0.46903 0.01713 0.83704 0.32614 0.91252 0.76779 0.35558 0.43550 0.03945 0.10766 0.74321 0.80123 0.38580 0.59367 0.37401 0.08924 0.18510 0.13036 0.00807 0.98551 0.12130 0.15128

Conclusion FDI ↔ Growth FDI→Growth∗ Growth→FDI Granger neutral Granger neutral Granger neutral FDI ↔ Growth Granger neutral Granger neutral Growth→FDI Granger neutral FDI→Growth Granger neutral

Mozambique Nepal Niger Nigeria Pakistan Panama Papua New Guinea Paraguay Peru Philippines Rwanda Senegal Seychelles Sierra Leone

FDI dngc growth Growth dngc FDI FDI dngc growth Growth dngc FDI FDI dngc growth Growth dngc FDI FDI dngc growth Growth dngc FDI FDI dngc growth Growth dngc FDI FDI dngc growth Growth dngc FDI FDI dngc growth Growth dngc FDI FDI dngc growth Growth dngc FDI FDI dngc growth Growth dngc FDI FDI dngc growth Growth dngc FDI FDI dngc growth Growth dngc FDI FDI dngc growth Growth dngc FDI FDI dngc growth Growth dngc FDI FDI dngc growth Growth dngc FDI

14 14 15 15 15 15 15 15 18 18 19 19 15 15 15 15 18 18 15 15 15 15 19 19 19 19 19 19

5 5 5 5 5 5 5 5 2 2 1 1 5 5 5 5 2 2 5 5 5 5 1 1 1 1 1 1

2.03587 0.55698 2.90376 3.05504 2.44477 1.20615 1.23298 1.89667 6.17275 8.42949 0.00765 0.00382 1.40814 5.86900 5.37542 2.06865 1.32841 0.28650 1.04485 2.02853 2.98154 0.52770 3.35404 1.24827 0.09464 0.00189 1.40588 0.91350

0.29637 0.73543 0.16187 0.15088 0.20351 0.44029 0.43174 0.27744 0.01304 0.00449 0.93140 0.95145 0.38111 0.05558 0.06415 0.25053 0.29858 0.75551 0.49690 0.25646 0.15607 0.75039 0.08572 0.28038 0.76233 0.96589 0.25305 0.35341

Granger neutral Granger neutral Granger neutral Granger neutral FDI ↔ Growth Granger neutral Growth→FDI FDI→Growth Granger neutral Granger neutral Granger neutral FDI→Growth Granger neutral Granger neutral 73

Continued

74

Table A3.1

Continued

Country

Null hypothesis

Singapore

FDI dngc growth Growth dngc FDI FDI dngc growth Growth dngc FDI FDI dngc growth Growth dngc FDI FDI dngc growth Growth dngc FDI FDI dngc growth Growth dngc FDI FDI dngc growth Growth dngc FDI FDI dngc growth Growth dngc FDI FDI dngc growth Growth dngc FDI

Solomon Islands Sri Lanka St. Kitts and Nevis St. Lucia St. Vincent and the Grenadines Swaziland Thailand

Obs.

Number of lags

F-statistic

Probability

19 19 19 19 14 14 19 19 14 14 19 19 16 16 18 18

1 1 1 1 5 5 1 1 5 5 1 1 4 4 2 2

0.16051 0.00548 2.67682 1.74463 1.02192 3.40110 0.56361 0.08468 8.12832 1.84872 0.93278 0.09826 5.04083 0.42202 15.5694 6.41221

0.69399 0.94191 0.12134 0.20513 0.52731 0.17128 0.46371 0.77479 0.05746 0.32511 0.34851 0.75797 0.03126 0.78873 0.00035 0.01155

Conclusion

Granger neutral Granger neutral Granger neutral Granger neutral FDI→Growth Granger neutral FDI→Growth FDI ↔ Growth

Togo Trinidad and Tobago Tunisia Turkey Uruguay Vanuatu Venezuela Zambia Zimbabwe

Notes: Source:

FDI dngc growth Growth dngc FDI FDI dngc growth Growth dngc FDI FDI dngc growth Growth dngc FDI FDI dngc growth Growth dngc FDI FDI dngc growth Growth dngc FDI FDI dngc growth Growth dngc FDI FDI dngc growth Growth dngc FDI FDI dngc growth Growth dngc FDI FDI dngc growth Growth dngc FDI

15 15 15 15 18 18 15 15 18 18 13 13 18 18 15 15 15 15

5 5 5 5 2 2 5 5 2 2 5 5 2 2 5 5 5 5

0.76754 3.83461 4.31291 1.27217 0.21620 3.81828 0.90743 2.04355 4.15796 1.37770 1.85462 5.98104 1.41419 0.75008 13.4937 0.53429 2.41404 0.52322

0.61818 0.10860 0.09085 0.41964 0.80841 0.04960 0.55290 0.25421 0.04018 0.28664 0.38630 0.14942 0.27816 0.49171 0.01295 0.74645 0.20685 0.75308

Growth→FDI FDI→Growth Growth→FDI Granger neutral FDI→Growth Granger neutral Granger neutral FDI→Growth Granger neutral

∗ Indicates a borderline case, that is, conclusion is valid at 11 per cent level of significance; ‘dngc’ stands for ‘does not Granger cause’.

Authors’ computations as described in the text.

75

76

Table A3.2

FDI and domestic investment relationship for developing countries Coefficients (t-value)

Country Dinv(t − 1) Dinv(t − 2) Algeria Argentina Bangladesh Barbados Belize Benin Bolivia Botswana Brazil Burkina Faso Burundi Cameroon Central African Republic

0.7528∗∗∗ (4.381) 0.515∗∗ (2.566) 0.775∗∗ (2.839) 0.507∗∗∗ (5.254) 0.4308∗∗ (2.5014) 0.3116 (1.2856) 0.469∗∗∗ (3.159) 0.978∗∗∗ (5.174) 0.509∗∗ (2.272) 0.6987∗∗∗ (3.5504) 0.3715 (1.4506) 0.263 (1.295) 0.2042 (0.7907)

0.238 (1.309) −0.341 (−1.554) 0.071 (0.305)

−0.896∗∗∗

(−3.961) −0.476∗∗ (−2.773) −0.160 (−0.389)

0.255∗ (1.842)

FDI(t)

FDI(t − 1)

FDI(t − 2)

−16.6589 0.9421 (−0.8231) (0.2691) −1.581∗∗∗ 0.249 2.059∗∗∗ (−13.201) (0.417) (3.271) 4.996∗∗ −1.062 3.334 (2.906) (−0.467) (1.293) ∗∗ −3.288 5.374 0.096 (−0.539) (2.362) (0.048) −1.4231∗∗∗ 1.1499∗∗ (−2.9996) (2.3703) −1.9516 2.3944 (−0.8264) (0.9599) 0.209 0.480 −1.644∗ (0.634) (0.804) (−2.042) −1.059∗∗∗ 0.640∗∗ −0.499∗ (−4.741) (2.304) (−2.020) −1.158 2.801 −4.351∗∗ (−1.109) (1.547) (−2.698) −4.5179 6.4846∗∗ (−1.5247) (2.2764) 3.8178 0.4254 (0.9692) (0.0946) −0.860∗∗∗ 0.962∗∗ 0.912∗ (−4.028) (2.844) (2.069) −1.6539 0.5804 (−1.1988) (0.3848)

Growth(t − 1)

Constant

−0.1561 (−0.7159) 0.131∗∗ (2.580) 0.142 (1.206) 0.269∗ (1.961) 0.3279∗∗ (2.2953) 0.3376 (1.3228) −0.726∗∗ (−2.187) 0.356∗ (1.944) 0.132 (0.617) −0.0449 (−0.2679) 0.1071 (0.5160) 0.161 (1.634) 0.1398 (0.9973)

7.6123 (1.436) 3.181 (1.595) 9.283∗∗∗ (3.447) 4.836 (0.915) 11.591∗∗∗ (2.972) 9.2023∗∗ (2.3486) 21.850∗∗∗ (3.956) 11.071∗∗ (2.568) 14.176∗∗ (2.618) 6.6318∗ (1.7239) 7.9891∗∗ (2.1161) 7.910∗∗∗ (4.303) 9.0832∗∗∗ (2.9709)

Adjusted F-value D–W statistic No. of observations R2 19

0.6225

18

0.8915 24.285

2.130

18

0.8046 12.6678

2.109

18

0.5499

4.4625

2.554

19

0.5234

5.9427

1.7348

19

0.0154

1.0705

2.2678

18

0.6535

6.3429

2.651

18

0.7873 11.4854

2.357

18

0.4975

3.8049

2.681

19

0.5052

5.5939

2.6088

19

0.0155

1.0711

2.3498

18

0.9020 27.0812

19

−0.0144

8.4207

0.9362

1.8672

2.130 1.8542

Chad Chile China Colombia Comoros Congo Democratic Republic Congo Republic Costa Rica ˆ d’Ivoire Cote Cyprus Dominica Dominican Republic Ecuador Egypt, Arab Republic El Salvador

0.002 0.415 (0.006) (1.119) 0.2725 (1.001) 0.3025∗ (2.0718) −0.756∗∗∗ 0.984∗∗∗ (4.986) (−3.220) 0.6271∗∗∗ (2.9086) 0.3259 (1.1323) 0.2954 (1.1105) 0.295 (1.372) 0.805∗∗ (2.409) −0.025 (−0.089) 0.3257∗ (1.7146) 0.4641∗∗ (2.1040) 0.0974 (0.3930) 0.6836∗∗∗ (3.7108) 0.6766∗∗∗ (3.6304)

0.076 (0.337) −0.117 (−0.429) −0.845∗∗ (−2.628)

−1.381 1.575∗∗ (−1.461) (2.474) 0.8613∗ −1.0538∗∗∗ (−2.882) (1.688) 0.1845 −0.3333 (0.8175) (−1.5586) −0.305 1.455∗∗∗ (−0.803) (3.299) −1.7277 0.1994 (−1.2011) (0.1321) −1.1143 0.0878 (−0.9514) (0.0738) −1.6366 (−0.2355) −4.373∗∗ (−2.819) 0.048 (0.100) 0.001 (0.001) −0.7966∗∗∗ (−3.0179) −0.6686 (−1.3043) 0.7317 (0.5752) 0.7413 (0.8026) −0.7650∗∗∗ (−3.5116)

−1.6625 (−0.3048) −0.846 (−0.898) 0.961∗∗ (2.545) 3.682∗ (2.039) 0.2526 (0.8311) 0.4425 (0.5483) −2.3571∗ (−1.7753) 1.0220 (0.9979) 0.5392∗ (2.0384)

−1.041 (−1.309)

−1.773∗∗∗

(−5.270)

3.270∗ (1.775) −1.018∗∗ (−2.271) 0.245 (0.383)

−0.281∗∗∗ (−3.572) 0.5488∗∗ (2.650) 0.3395∗∗∗ (4.8630) 0.584∗∗ (2.678) 0.3848 (0.8916) 0.2923 (1.5874)

8.178∗ (2.139) 11.270∗∗ (2.6199) 21.046∗∗∗ (4.2838) 12.602∗∗∗ (4.016) 8.2163 (1.5257) 6.8084∗∗ (2.2183)

0.8799∗ (1.7409) 0.271∗∗ (2.103) 0.080 (0.289) −0.212 (−1.307) 0.0626 (0.1686) 0.2070 (1.1455) 0.2566 (1.2834) −0.1187 (−0.3132) 0.0997 (1.0325)

20.5879∗∗ (2.5212) 16.992 (1.587) 2.962∗∗ (2.264) 43.156∗∗∗ (5.905) 19.1856∗∗∗ (2.9133) 10.5240∗∗ (2.14269) 18.1994∗∗∗ (3.3895) 3.9747 (1.2571) 4.7557∗ (1.7279)

16

0.3173

2.1618

1.832

19

0.6027

7.8270

2.0823

19

0.6788 10.5109

2.2425

18

0.7128

8.0329

2.292

18

0.4939

5.1467

2.3805

18

0.4283

4.1835

2.0779

19

0.0975

1.4862

2.0718

18

0.8195 13.8646

1.364

18

0.7241

8.4370

2.295

17

0.3879

2.6906

2.240

19

0.5247

5.9677

1.3534

19

0.1625

1.8734

1.8046

19

0.3294

3.2106

2.0754

19

0.6547

9.5326

1.3823

19

0.6455

9.1944

1.9844

77

Continued

78

Table A3.2 Continued Coefficients (t-value)

Country Dinv(t − 1) Ethiopia Fiji Gabon Gambia Ghana Grenada Guatemala Guinea-Bissau Guyana Haiti Honduras India Indonesia Jamaica

0.0395 (0.1292) 0.277∗ (1.849) 0.3674 (1.4353) 0.6181∗∗∗ (4.0102) 0.382 (1.309) 0.736∗∗∗ (5.116) 0.4059∗ (1.8532) 0.3953 (1.5404) 0.630∗∗ (2.304) 0.4651∗∗∗ (3.0706) 0.7713∗∗∗ (5.8180) −0.0579 (−0.2471) 0.028 (0.081) 0.6316∗∗∗ (3.6274)

Dinv(t − 2)

0.321∗∗ (2.131)

0.510∗ (1.743) −0.472∗∗ (−2.330)

−0.838∗∗∗ (−3.647)

0.062 (0.285)

FDI(t)

FDI(t − 1)

−0.2543 (−0.1863) −0.826∗∗∗ (−5.998) −0.3926 (−0.5459) −0.6966∗∗ (−2.2244) 1.420∗∗ (2.232) −1.121∗∗∗ (−3.837) −0.5736 (−1.4416) −1.9174 (−0.8110) −0.929∗∗∗ (−10.890) 3.1069 (1.1978) −0.7084 (−0.6038) 3.9794 (1.3537) 0.574 (0.785) −1.3717∗∗∗ (−2.9205)

1.3102 (0.9157) 0.319∗∗ (2.510) 0.4227 (0.5889) 0.7869∗∗ (2.2890) −1.109∗∗∗ (−3.133) 1.073∗∗ (2.378) 0.2588 (0.6138) −3.7014 (−1.4811) 0.916∗∗∗ (3.296) 8.0696∗∗ (2.4510) 3.0889∗ (1.7916) −5.2697∗ (−1.7918) 0.364 (0.221) 1.5725∗∗∗ (2.9182)

FDI(t − 2)

0.124 (0.751)

0.311 (0.535) −0.239 (−0.431)

−1.088∗∗∗ (−5.264)

−0.871 (−0.523)

Growth(t − 1)

Constant

0.1348 (1.2943) 0.158∗ (1.959) −0.3014 (−1.1417) 0.1013 (0.5781) 0.175 (1.181) 0.044 (0.204) 0.4319∗ (2.2146) 0.1553 (0.6502) 0.329 (1.454) −0.1329 (−1.1761) 0.3794 (1.0738) 0.0281 (0.1536) −0.043 (−0.218) 0.2274 (1.1934)

12.9390∗∗∗ (3.1332) 4.936∗∗∗ (4.431) 18.9411∗∗ (2.6135) 6.0619∗ (1.8875) 1.936 (1.610) 22.345∗∗∗ (2.890) 7.1827∗∗ (2.4134) 19.5804∗∗ (2.3169) 36.358∗∗∗ (7.057) 3.1134∗∗ (2.1065) 1.7924 (0.6293) 24.2511∗∗∗ (4.6409) 26.332∗∗∗ (3.093) 9.2119∗∗ (2.3209)

Adjusted F-value D–W statistic No. of observations R-squared 17 18

−0.0379

0.8538

0.8948 25.095

2.0714 2.271

19

−0.0119

0.9468

1.7512

19

0.5340

4.6896

1.9563

18

0.8868 23.1922

2.279

18

0.5227

4.1032

2.453

19

0.4893

5.3115

2.0468

19

0.2223

2.2913

1.7682

18

0.8311 14.9517

1.611

19

0.8247 22.1672

1.7374

19

0.7567 14.9991

1.8553

19

−0.0214

0.9057

1.8905

18

−0.1418

0.6481

2.052

19

0.5736

7.0535

2.2733

Kenya Korea, Republic of Lesotho Madagascar Malawi Malaysia Mali Mauritania Mauritius Mexico Morocco Mozambique Nepal Niger Nigeria Pakistan

0.2543 (0.9644) 0.6986∗∗∗ (5.5755) 0.7483∗∗∗ (4.0524) 0.1476 (0.6441) 0.0169 (0.0647) 0.4800∗ (1.7938) 0.8539∗ (5.1749) 0.6970∗∗∗ (4.8132) 0.581∗ (1.935) −0.091 (−0.277) 0.591∗ (1.923) 0.5707∗∗∗ (3.5049) −0.1957 (−0.6114) 0.4694∗∗ (2.7072) 0.6902∗∗∗ (3.1681) 0.9411∗∗∗ (3.1915)

−0.047 (−0.208) −0.022 (−0.102) −0.140 (−0.541)

3.1547 (1.0828) −12.4973∗∗∗ (−6.0480) −0.8213∗∗∗ (−3.0940) 1.1605 (0.7246) −1.1282 (−0.8783) 0.2877 (0.2787) −1.6768∗∗ (−2.6088) 1.8548 (1.4722) 3.567∗∗∗ (2.884) −1.635∗∗∗ (−3.909) −1.629 (−1.333) 0.1518 (0.2208) 14.8534∗ (2.6472) −0.5077 (−0.5112) −1.0687∗ (−2.0659) −1.1981 (−0.9826)

2.8229 (1.1962) 17.3632∗∗∗ (5.5224) 0.4406 (1.2435) −0.4506 (−0.1819) −1.8046 (−1.1715) 0.3744 (0.3927) 0.9286 (1.4004) 1.6746∗ (2.0279) −1.010 2.131 (−0.571) (1.454) 0.348 1.477 (0.480) (1.631) 2.060 −2.513∗ (1.393) (−1.946) 0.7387 (0.5644) 5.0722 (1.0551) 0.9502 (1.1151) 0.4315 (0.9177) 1.4644 (1.1131)

0.0586 11.1774∗∗∗ (0.1689) (2.6775) 0.0113 8.9673∗∗ (0.0844) (2.0537) 0.3383 11.6481 (0.7140) (1.4116) 0.1794 8.6206∗∗∗ (1.1045) (3.4231) 0.0285 18.2894∗∗∗ (0.1477) (3.4622) 0.3350 9.8914 (0.9927) (1.3958) 0.0889 3.4451 (0.6633) (1.0858) 0.1456 3.6178 (0.5452) (1.0092) 0.074 8.078 (0.212) (1.337) 0.188 22.259∗∗∗ (1.625) (7.919) −0.099 13.937∗∗ (−0.805) (2.462) 0.2209∗∗ 4.6182∗∗ (2.3348) (2.6797) 0.1696 22.2131∗∗∗ (1.3740) (3.6019) −0.1206 5.2513∗∗ (−0.6176) (2.3297) 0.3572∗ 5.9896 (1.8016) (1.5384) 0.2556∗∗ −0.6699 (2.2251) (−0.1179)

19

0.2766

2.7209

2.0327

19

0.8163 20.9978

1.9398

18

0.7063 11.2194

2.2046

19

−0.0131

0.9415

2.1917

19

0.1796

1.9852

2.0202

19

0.4755

5.0792

2.0105

19

0.5857

7.3612

2.0385

19

0.6830 10.6968

2.0764

18

0.5043

3.8827

1.479

18

0.5578

4.5746

2.218

18

0.3332

2.4163

2.221

18

0.8139 19.5987

1.8526

19

0.6352

8.8361

2.1124

19

0.2559

6.1316

2.0739

19

0.3158

3.0773

2.0205

19

0.4966

5.4398

2.1692

79 Continued

80

Table A3.2 Continued Coefficients (t-value)

Country Dinv(t − 1) Panama Papua New Guinea Paraguay Peru Philippines Rwanda Senegal Seychelles Sierra Leone Singapore Sri Lanka St. Kitts and Nevis

Dinv(t − 2)

0.5780∗∗∗ (3.5351) 0.5899∗∗ (2.5843) 0.226 (1.012) 0.5296∗∗∗ (3.009) 0.4783∗∗ (2.7869) −0.107 (−0.582) 0.383∗∗ (2.104) 0.5139∗ (1.9888) −0.427 (−0.960) 0.8479∗∗∗ (6.0252) 0.4793∗∗∗ (3.6898) 0.5834∗∗ (2.3743)

FDI(t) −0.6461∗∗

−0.139 (−0.727)

−0.251 (−1.257) 0.603∗∗ (2.764) 0.887∗∗ (2.227)

(−2.3267) −0.9432∗∗ (−2.2412) −1.255∗∗ (−2.647) −1.2983∗ (−2.020) −1.5022∗ (−1.9864) 4.539∗∗ (2.821) −0.607 (−1.412) −1.3041 (−1.0799) −0.959∗∗∗ (−9.163) −1.0670∗∗∗ (−4.0889) −0.1378 (−0.2161) −0.7589∗∗∗ (−3.1773)

FDI(t − 1)

FDI(t − 2)

0.8545∗∗∗ (3.1042) 0.7202 (1.4311) 0.488 −0.536 (0.968) (−1.020) 0.4707 (0.720) 0.1804 (0.2460) −2.377 −1.060 (−1.441) (−0.956) 0.168 0.750∗∗ (0.553) (2.894) 1.6206 (1.2877) −0.363 1.072∗∗ (−0.791) (2.484) 0.9511∗∗ (2.6566) 1.2265∗ (1.7368) 0.5047∗ (1.9395)

Growth(t − 1) 0.4003∗ (2.0003) −0.0784 (−0.44) 0.115 (0.996) 0.4317∗∗ (2.854) 0.4863∗∗ (2.5312) 0.009 (0.705) −0.006 (−0.063) −0.1946 (−0.5692) 0.062 (0.674) 0.0204 (0.0878) 0.2381 (0.7616) 0.4816 (1.0840)

Constant 6.5635∗∗ (2.0979) 8.8526 (1.4981) 21.573∗∗∗ (5.755) 10.4686∗∗ (2.3965) 11.0815∗∗ (2.6255) 19.250∗∗∗ (6.370) 0.597 (0.324) 9.1289 (1.1294) 4.640∗∗ (2.948) 4.7499 (0.6788) 10.1621∗∗∗ (3.1917) 12.3284 (1.4289)

No. of Adjusted F-value D–W statistic observations R-squared 19

0.64083 9.0289

1.8286

19

0.3503

3.4258

1.7288

18

0.4295

3.1329

2.933

19

0.5030

5.5549

2.5544

19

0.5931

7.5581

1.8567

18

0.2783

2.0926

1.905

18

0.6505

6.2727

1.615

19

0.1109

1.5611

1.7936

15

0.9068 23.7077

2.284

19

0.7812 17.0705

2.1351

19

0.5119

5.8699

2.0141

19

0.3197

3.1143

1.8355

St. Lucia St. Vincent and the Grenadines Swaziland Thailand Togo Trinidad and Tobago Tunisia Turkey Uganda Uruguay Venezuela Zambia Zimbabwe

0.579 (1.710) 0.4088 (2.3462) 0.373 (1.254) 0.5381∗∗∗ (4.8662) 0.3853∗ (1.6784) 0.7487∗∗∗ (4.3181) 1.162∗∗∗ (5.131) 0.839∗ (1.800) 0.9887∗∗∗ (6.1403) 0.5901∗∗∗ (4.5438) 0.1105 (0.4159) 0.2191 (0.8484) 0.3041 (1.1431)

−0.515∗ (−1.909) −0.763∗ (−1.848)

−0.642∗∗ (−2.347) −0.437 (−0.910)

−0.725∗ (−2.083) −1.0059 (−8.2525) −2.367∗∗∗ (−4.239) −2.6068∗∗∗ (−4.7596) −0.4470 (−0.4399) −0.0148 (−0.0520) −0.522 (−0.569) −2.711 (−0.356) −1.9627∗∗ (−2.5418) −1.9017∗ (−1.9229) −1.2352 (−1.1430) −0.6888 (−1.0980) −0.3625 (−0.6206)

0.656∗∗ (2.403) 0.5930 (2.9760) 0.272 (0.364) 3.2418∗∗∗ (5.1769) 0.7390 (0.7872) −0.0131 (−0.0434) 0.572 (0.534) 8.608 (0.918) 1.7744∗∗ (2.1952) 2.1955∗∗∗ (4.2342) 0.0925 (0.0882) 0.0455 (0.0744) −0.8717 (−1.4658)

−0.404∗∗∗ (−3.330)

−0.640 (−0.862)

−0.012 (−0.013) 1.309 (0.208)

0.184 (1.452) 0.0838 (0.5615) −0.174 (−0.459) 0.8502∗∗∗ (4.8324) 0.0406 (0.2544) 0.0605 (0.2520) −0.470 (−1.352) −0.300 (−0.895) 0.0308 (0.3221) 0.1805∗∗ (2.3448) 0.3543 (1.2349) −0.2190 (−0.9035) 0.0625 (0.3888)

18.513∗∗∗ (3.119) 15.4568∗∗∗ (2.7945) 43.515∗ (2.101) 8.1951∗∗∗ (2.6280) 9.2353∗∗ (2.1597) 3.5756 (0.9734) 14.444∗∗ (2.490) 12.188∗ (2.016) 0.9093 (0.5977) 5.1702∗∗∗ (2.6576) 16.8329∗∗∗ (2.9389) 11.1620∗∗ (2.4796) 12.8313∗∗ (2.5709)

18

0.1691

1.5766

19

0.8766 32.9761

1.6483

18

0.7385

1.760

19

0.9066 44.6431

2.4211

19

0.0966

1.4817

2.1867

19

0.5853

7.3518

2.0496

18

0.4802

3.6175

2.829

18

0.3881

2.7967

2.089

19

0.8232 18.4619

2.4850

19

0.7299 13.1643

1.7962

19

0.0587

1.2805

1.9449

19

0.0867

1.4269

2.2165

19

0.1113

1.5633

1.8721

9.0002

2.096

Notes: The Schwarz information criterion has been used to determine the lag length in the estimation. ∗∗∗ , ∗∗ , ∗ respectively indicate 1, 5 and 10 per cent levels of significance Source: Authors’ computations as described in the text.

81

82 FDI: Some Empirical Explorations

References Agosin, M. R. and Ricardo Mayer (2000) ‘Foreign Investment in Developing Countries: Does It Crowd in Domestic Investment?’, UNCTAD Discussion Paper no. 146 (Geneva: UNCTAD). Aitken, Brian and Ann E. Harrison (1999) ‘Do Domestic Firms Benefit from Direct Foreign Investment?’, American Economic Review, 89(3), pp. 605–18. Anderson, T. W. and C. Hsiao (1982) ‘Formulation and Estimation of Dynamic Models Using Panel Data’, Journal of Econometrics, vol. 18, pp. 47–82. Arellano, M. and S. Bond (1991) ‘Some Tests of Specification for Panel Data: Monte Carlo Evidence and an Application to Employment Equations’, Review of Economic Studies, vol. 58, pp. 277–97. Balasubramanyam, V. N., M. Salisu and David Sapsford (1996) ‘Foreign Direct Investment and Growth in EP and IS Countries’, The Economic Journal, vol. 106, pp. 92–105. Baltagi, B. H. (1995) Econometric Analysis of Panel Data (Chichester: John Wiley). Barro, Robert J. and Xavier Sala-i-Martin (1995) Economic Growth (New York: McGraw-Hill). Battat, Joseph, Isaiah Frank and Xiaofang Shen (1996) ‘Suppliers to Multinationals: Linkage Programs to Strengthen Local Companies in Developing Countries’, FIAS Occasional Paper no. 6. Benhabib, Jess and Mark M. Spiegel (1994) ‘The Role of Human Capital in Economic Development: Evidence from Aggregate Cross-country Data’, Journal of Monetary Economics, vol. 34, pp. 143–73. Blomström, Magnus (1989) Foreign Investment and Spillovers: A Study of Technology Transfer to Mexico (London: Routledge). Blomström, Magnus and Edward N. Wolff (1989) ‘Multinational Corporations and Productivity Convergence in Mexico’, NBER Working Paper no. 3141. Blomström, Magnus, Robert E. Lipsey and Mario Zejan (1994) ‘What Explains the Growth of Developing Countries?’, in William J. Baumol, Richard R. Nelson and Edward N. Wolff (eds), Convergence of Productivity (Oxford/New York: Oxford University Press), pp. 243–56. Borensztein, Eduardo, José De Gregorio and Jong-Wha Lee (1995) ‘How Does Foreign Direct Investment Affect Economic Growth?’, NBER Working Paper no. 5057. Bosworth, Barry P. and Susan M. Collins (1996) ‘Economic Growth in East Asia: Accumulation versus Assimilation’, Brookings Papers on Economic Activity, no. 1, pp. 135–203. Caves, R. E. (1974) ‘Multinational Corporations, Competition and Productivity in Host-country Markets’, Economica, vol. 41, pp. 176–93. Chen, Edward K. Y. (1983) Multinational Corporations, Technology and Employment (London: Macmillan). de Mello, Luiz R., Jr. (1997) ‘Foreign Direct Investment in Developing Countries and Growth: A Selective Survey’, Journal of Development Studies, vol. 34(1), pp. 1–34. de Mello, Luiz R., Jr. (1999) ‘Foreign Direct Investment-led Growth: Evidence from Time Series and Panel Data’, Oxford Economic Papers, vol. 51, pp. 133–54. Dunning, J. H. (1998) ‘Changing Geography of Foreign Direct Investment: Explorations and Implications’, in Nagesh Kumar, Jamuna P. Agarwal,

Nagesh Kumar and Jaya Prakash Pradhan 83 John H. Dunning, Robert E. Lipsey and Shujiro Urata (eds), Globalization, Foreign Direct Investment and Technology Transfers: Impact on and Prospect for Developing Countries (London/New York: Routledge), pp. 40–66. Fry, Maxwell J. (1992) ‘Foreign Direct Investment in a Macroeconomic Framework: Finance, Efficiency, Incentives and Distortions’, PRE Working Paper (Washington, DC: The World Bank). Graham, Edward M. (1995) ‘Foreign Direct Investment in the World Economy’, IMF Staff Studies for the World Economic Outlook, September, pp. 120–35. Grossman, Gene and Elhanan Helpman (1991) Innovation and Growth in the Global Economy (Cambridge, Mass.: MIT Press). Haddad, M. and A. Harrison (1993) ‘Are There Positive Spillovers from Direct Foreign Investment? Evidence from Panel Data for Morocco’, Journal of Development Economics, vol. 42, pp. 51–74. Hausman, J. A. (1978) ‘Specification Tests in Econometrics’, Econometrica, vol. 46(6), pp. 1251–71. Hein, Simeon (1992) ‘Trade Strategy and the Dependency Hypothesis: A Comparison of Policy, Foreign Investment, and Economic Growth in Latin America’, Economic Development and Cultural Change, vol. 40(3), pp. 495–521. Kathuria, V. (1998) ‘Technology Transfer and Spillovers for Indian Manufacturing Firms’, Development Policy Review, vol. 16, pp. 73–91. Kokko, A. (1994) ‘Technology, Market Characteristics and Spillovers’, Journal of Development Economics, vol. 43, pp. 279–93. Kokko, A., Ruben Tansini and Mario C. Zejan (1996) ‘Local Technological Capability and Productivity Spillovers from FDI in the Uruguayan Manufacturing Sector’, Journal of Development Studies, vol. 32(4), pp. 602–11. Kumar, Nagesh (1990) ‘Mobility Barriers and Profitability of Multinational and Local Enterprises in Indian Manufacturing’, Journal of Industrial Economics, vol. 38, pp. 449–61. Kumar, Nagesh (1991) ‘Mode of Rivalry and Comparative Behaviour of Multinational and Local Enterprises: The Case of Indian Manufacturing’, Journal of Development Economics, vol. 35, pp. 381–92. Kumar, Nagesh (2002) Globalization and the Quality of Foreign Direct Investment (New Delhi: Oxford University Press). Kumar, Nagesh (2003) Performance Requirements as Tools of Development Policy: Lessons from Experiences of Developed and Developing Countries for the WTO Agenda on Trade and Investment, RIS Discussion Paper no. 52, available at www.ris.org.in. Kumar, Nagesh and N. S. Siddharthan (1997) Technology, Market Structure and Internationalization: Issues and Policies for Developing Countries (London/New York: Routledge and UNU Press). Lipsey, Robert E. (1998) ‘The Internationalization of US MNEs and Its Impact in Developing Countries’, in Nagesh Kumar, Jamuna P. Agarwal, John H. Dunning, Robert E. Lipsey and Shujiro Urata (eds), Globalization, Foreign Direct Investment and Technology Transfers: Impacts on and Prospects for Developing Countries (London/New York: Routledge), pp. 197–212. Mankiw, N. Gregory, David Romer and David N. Weil (1992) ‘A Contribution to the Empirics of Economic Growth’, Quarterly Journal of Economics, vol. 107(2), pp. 407–37. Markusen, J. R. and Anthony J. Venables (1997) ‘Foreign Direct Investment as a Catalyst for Industrial Development’, NBER Working Paper no. 624.

84 FDI: Some Empirical Explorations Moran, Theodore H. (1998) Foreign Direct Investment and Development (Washington, DC: Institute for International Economics). Pradhan, Jaya Prakash (2001) ‘Foreign Direct Investment and Economic Growth: The Case of Developing Countries’, Unpublished M.Phil. dissertation submitted to Jawaharlal Nehru University, New Delhi. Romer, Paul M. (1993) ‘Ideas Gaps and Object Gaps in Economic Development’, Journal of Monetary Economics, vol. 32(3), pp. 543–73. Romer, Paul M. (1994) ‘The Origin of Endogenous Growth’, Journal of Economic Perspectives, vol. 8(1), pp. 3–22. Saggi, Kamal (2000) ‘Trade, Foreign Direct Investment, and International Technology Transfer: A Survey’, issued as WT/WGTI/W/88, 19 September 2000 (Geneva: World Trade Organization). Singh, R. D. (1988) ‘The Multinationals’ Economic Penetration, Growth, Industrial Output, and Domestic Savings in Developing Countries: Another Look’, Journal of Development Studies, vol. 25(1), pp. 55–82. World Bank (2001) World Development Indicators 2001 (Washington, DC: The World Bank). UNCTAD (1999) World Investment Report 1999 (New York: United Nations). UNCTAD (2001) World Investment Report 2001 (New York: United Nations). Xu, B. (2000) ‘Multinational Enterprises, Technology Diffusion, and Host Country Productivity Growth’, Journal of Development Economics, vol. 62, pp. 477–93.

4 Technological Spillovers and Export-platform FDI Kjetil Bjorvatn Norwegian School of Economics and Business Administration, Bergen, Norway

and Carsten Eckel University of Göttingen, Germany

1

Introduction

The literature on foreign direct investment (FDI) suggests that technological spillovers from multinational firms to local firms are, at least potentially, significant (Blomström and Kokko, 1998). Typical channels for spillovers associated with the activities of multinational enterprises (MNEs) include backward and forward linkages between foreign affiliates and local firms, demonstration effects, and labour turnover in the host country. Audretsch and Feldman (1996), Bransetter (2001) and Keller (2001) report that such spillovers are primarily local in nature – that is, intranational, rather than international. Hence, technological spillovers can affect a firm’s mode of entry into a foreign market. Analyses of entry choice in the presence of spillovers include Ethier (1984), Ethier and Markusen (1996), Siotis (1999), Fosfuri (2000), Fosfuri, Motta and Ronde (2001) and Markusen (2001). One insight from the existing literature is that spillovers reduce the incentive for technological leaders to invest in foreign markets. Typically, the literature suggests that, in order to limit spillovers, the MNE may be more inclined to choose exports or perhaps to expose its foreign competitors to a less advanced version of the MNE’s technology. Spillovers may also induce so-called technology-sourcing FDI, where the less advanced firm invests in the home country of its more advanced competitors in order to learn from them. We discuss the issues of technology sourcing by the less advanced firm and investment strategies by the more advanced firm in Bjorvatn and Eckel (2005). 85

86 Technological Spillovers and Export-platform FDI

In this chapter we focus on a different aspect of spillovers and FDI and analyse how spillovers affect not the mode of entry, but rather the MNE’s choice of location within a region. We model a multinational firm that has decided to invest in a region in order to service demand in that region. Trade costs in exporting from the MNE’s home plant to the destination region are assumed implicitly to be such that exporting to the region is not a profitable option for the MNE. The variables we focus on in our analysis are market size, intra-regional trade costs, spillover intensity, and the technological gap between the MNE and local firms. We find that, with low spillovers, the multinational invests in the larger market and, depending on fixed costs, services the smaller market through exports or through a second plant. With high spillovers, the MNE invests in the smaller market only and uses its plant as a platform for exports to the larger market.

2

The model

Consider a region consisting of two countries, A and B. Demand in country J = A, B is given by Q J = αJ (1 − pJ )

(1)

where Q J is the quantity demanded, pJ is the market price in country J (both markets are completely separated), and αJ is a parameter denoting the size of the market. In each market, there is one local firm, firm a in country A, and firm b in country B. The local firms are assumed to service local demand only. We assume constant marginal costs, denoted by ci for firm i = a, b, m, where subscript m refers to a multinational firm that is not yet established in the region. The multinational firm produces a good that is identical to that of the local firms. In order to service demand in the region, the multinational has to set up at least one subsidiary in the region. Therefore, the multinational firm considers investment in either country A or B, or in both countries. Trade costs associated with exporting from the multinational’s home country to the region are assumed to be prohibitively high, so that exporting is never a profitable strategy. The reason for this assumption is simply that the trade-versus-investment choice is well understood from the literature on FDI, and that it is not the focus of our analysis. The question we wish to analyse is the following: given that a multinational firm has chosen to invest in a region, in which country will it choose to locate its affiliate plant? If the MNE chooses to invest in

Kjetil Bjorvatn and Carsten Eckel 87

a single location and service the other country by exports, we call this export-platform FDI. However, we also allow for the MNE to establish plants in both locations. Setting up a plant requires a fixed cost, F. In the case of export-platform FDI, the multinational also faces an intraregional trade cost t per unit transported from one country to the other. The MNE is assumed to be technologically more advanced than the local firms, which in turn are assumed to be equally efficient. In the model, this means that cm < c, where c ≡ ca = cb . When the MNE sets up a plant in a country, some of its technological advantage spills over to the local competitor in that country (technological spillovers). Spillovers are assumed to be local in nature. This means that they occur only in a market where the MNE is present with a production facility. In our model, spillovers are simply a fraction s ∈ [0, 1] of the difference in marginal production costs. Thus, post-spillover marginal production costs of a local firm are (1 − s)(c − cm ). If s = 0, there are no spillovers, whereas if s = 1, spillovers are complete in the sense that post-spillover marginal costs are identical between the MNE and the local competitor. The market outcome is derived in two stages. First, the MNE decides how to penetrate the two markets. It has three options: FDI in both countries; export-platform FDI in country A; or export-platform FDI in country B. In the second stage, the firms compete on quantities – that is, Cournot competition. Since there are simultaneous moves at the production stage of the game, there is no strategic interaction between the players. In order to reduce the number of variables, let cm = 0. This means that c is both the marginal cost of the local firms a and b as well as a measure of the technology gap between the MNE and the local firms. Profits for the MNE in both markets are described in Table 4.1. We assume that 0 < c < 21 to ensure that the local firms make positive profits, and t < 21 so that exporting from the export platform is always a profitable strategy for the MNE. The advantage with investment in both A and B is that the MNE saves on trade costs as it is present in both markets. However, the presence requires fixed costs in both markets (2F), and spillovers (s) also increase Table 4.1 Profits of a multinational enterprise Options

Profits in Market A

FDI in A and B X platform in A

2 1 9 αA (1 + (1 − s)c) 2 1 9 αA (1 + (1 − s)c)

X platform in B

1 9 αA (1 + c

− 2t)2

Profits in Market B −F −F

2 1 9 αB (1 + (1 − s)c) 2 1 9 αB (1 + c − 2t)

−F

2 1 9 αB (1 + (1 − s)c)

−F

88 Technological Spillovers and Export-platform FDI

the competitiveness of its local competitors in both markets. Export platform FDI has the advantage that the MNE invests in only one location and therefore saves on fixed costs (F) relative to multi-plant investment. In addition, since it locates in only one country, spillovers are restricted to only one local firm. The disadvantage with the export-platform strategy is that the MNE has to incur trade costs (t) from the export platform to the export destination. The choice between export-platform production based in A or B – that is, the choice of location, essentially depends on differences between the two locations. In this chapter we focus on one source of difference between the two countries, namely market size. Specifically, in the remainder of the analysis we shall assume that country A is larger than country B, so that αA > αB .

3

Multi-plant investment

Let us first consider the MNE’s choice between FDI in both locations (multi-plant investment, MPI) and export platform FDI. If the MNE invests in both countries, profits are given by: MPI = 19 αA (1 + (1 − s)c)2 − F + 19 αB (1 + (1 − s)c)2 − F

(2)

where the superscript of  indicates that the MNE invests in both markets (MPI). If the MNE sets up an export platform (XP) in market A, profits are given by: 2 2 1 1 XP A = 9 αA (1 + (1 − s)c) − F + 9 αB (1 + c − 2t)

(3)

If the MNE sets up an export platform in market B, profits are given by: 2 2 1 1 XP B = 9 αA (1 + c − 2t) + 9 αB (1 + (1 − s)c) − F

(4)

The MNE is indifferent between export platform in A and FDI in both countries if MPI = XP A . This condition implies that: 1 α (1 9 B

+ (1 − s)c)2 − F = 19 αB (1 + c − 2t)2

(5)

which can be rewritten as: F = 19 αB [2(1 − t) + c(2 − s)](2t − sc) ≡ FB∗

(6)

Kjetil Bjorvatn and Carsten Eckel 89

Similarly, the MNE is indifferent between setting up an export platform in B and FDI in both countries if MPI = XP B . This condition is fulfilled if:

F = 19 αA [2(1 − t) + c(2 − s)](2t − sc) ≡ FA∗

(7)

Naturally, the decision between setting up an export platform in country A and investing in both locations – that is, adding an extra plant in country B – depends on characteristics of market B (and vice versa). At the critical level of fixed cost FB∗ (FA∗ ), the MNE is indifferent between setting up an export platform in country A(B) and investing in both countries. If F < min(FA∗ , FB∗ ), the MNE favours plants in both locations, and if F > min(FA∗ , FB∗ ), it chooses to invest in only one country. Two things should be noted with respect to Equations (6) and (7). First, FA∗ , FB∗ > 0 requires t > 21 sc. This means that if t < 21 sc, the MNE will always prefer setting up an export platform over investing in both markets as long as F > 0. Intuitively, if spillovers are large relative to trade costs, the multinational will choose to restrict the effects of these spillovers by investing in only one country and exporting to the other. Second, note that FB∗ = ααAB FA∗ . Hence, given that 2t − sc > 0, αA > αB implies that FA∗ > FB∗ . This means that the critical level of fixed costs below which the multinational chooses to invest in both countries is given by FB∗ , which is determined by the size of the smaller market, B. Both FA∗ and FB∗ are plotted in F, t space in Figure 4.1.

F F *A F *B

1 2

Figure 4.1

t sc

Critical levels of fixed costs

90 Technological Spillovers and Export-platform FDI

Our findings are summarized in Proposition 1: Proposition 1 If t < 21 sc, the MNE will always set up an export platform. If t > 21 sc, the MNE will set up an export platform if F > FB∗ and choose multi-plant investment if F < FB∗ .

4

Export platform FDI

Let us now consider the location decision in the export platform case – that is, when F > FB∗ . If the MNE chooses to set up only one plant in the region and uses it as an export platform for the other market, in which country would the MNE locate this plant? The profit differential between setting up an export platform in country A and setting up an export platform in country B is given by: XP 2 2 1 π ≡ XP A − B = 9 (αA − αB )[(1 + (1 − s)c) − (1 + c − 2t) ]

(8)

which can be reduced to: π = 19 (αA − αB )(2t − cs)[2(1 − t) + c(2 − s)]

(9)

As 2(1 − t) + 2c(1 − s) > 0, the MNE will locate in country A and export to country B if π > 0, that is, if: (αA − αB )(2t − cs) > 0

(10)

Of course, if the two countries are identical (αA = αB ), the MNE is always indifferent between the two locations. According to Equation (10), the MNE’s decision depends on the ranking of market sizes and on the size of spillovers relative to trade costs. In our example, where αA > αB , the MNE will set up the export platform in country A if 2t > cs, whereas it will choose to invest in country B if cs > 2t. Note that under the latter condition we already showed that FB∗ < 0, so we know that the MNE will limit its engagement in the region to one plant. Also note that the absolute difference in market size is rather irrelevant. The sign of (αA − αB ) is important, but not the value. In the present model, therefore, it does not matter how much larger is country A relative to country B. Figure 4.2 illustrates these results.

Kjetil Bjorvatn and Carsten Eckel 91

t XP (A) if F > F *B t=

MPI if F < F *B

1 2

sc

XP (B ) sc Figure 4.2

Spillovers and the location of the export platform

F XP (A) XP (B)

F *B

MPI

1 2

Figure 4.3

t sc

The MNE investment decision

The intuition behind this result is that if spillovers are high, the MNE is concerned about its competitiveness in the larger market. By setting up the export platform in the smaller country, it keeps its superior technology completely out of the larger market and limits spillovers to the smaller market. Naturally, the importance of this argument rises with the difference in technology between the MNE and its local competitors (c) and with the spillover rate (s). However, by exporting to the larger market, the MNE also has to incur the trade costs in the larger market. Thus, when deciding where to locate the export platform, the MNE compares spillovers (sc) and trade costs (t). High spillovers are an argument for investment in the smaller country, whereas high trade costs are an argument for investment in the larger country. Clearly, high spillovers are also a reason to limit the production presence to just one location. Hence, if spillovers are large (sc > 2t), the MNE never chooses to invest in both countries. But if spillovers are small (sc < 2t), multi-plant production is an option, and the MNE compares fixed costs F with the critical level FB∗ . The MNE’s investment decision is summarized in Figure 4.3.

92 Technological Spillovers and Export-platform FDI

These results can be summarized as in Proposition 2: Proposition 2 (i) For high spillovers relative to trade costs such that sc > 2t, the MNE will always prefer to set up an export platform rather than invest in both markets. The export platform will be located in the smaller country. (ii) If spillovers are low relative to trade costs, such that sc < 2t, the MNE either invests in both markets or sets up an export platform in the larger market, depending on the level of fixed costs. If F < FB∗ , it chooses multiplant production, whereas if F > FB∗ , it chooses export platform investment in the larger market.

5

Conclusion

Our study analyses an MNE’s location decision within a region when technological spillovers are an important factor. We show that the MNE’s location decision is driven by the trade-off between the size of the market and the size of the spillovers. If spillovers are small, the MNE tries to minimize trade costs and either produces in both countries or sets up an export platform in the larger market. But if spillovers are large, the MNE tries to keep spillovers to a minimum. In this case, the MNE entertains only a single production presence in the region and locates in the smaller market. These findings indicate that the traditional proximity-concentration trade-off theory (Horstmann and Markusen, 1992) has to be amended if spillovers play a part. Fixed costs are an argument for concentrating production, trade costs are an argument for proximity to consumers, and spillovers are an argument for remote production. In our framework, trade costs (t) and spillovers (sc) determine whether a potential export platform would be located close to consumers in the large country or at a remote location in the small country. The relationship between trade costs and fixed costs (F) (or between spillovers and fixed costs) determines whether the MNE will decide to invest in both markets or to concentrate production and limit spillovers to a single location. This model can be extended in various directions. One option is adding policy competition between local governments for FDI. What kinds of investment policy are governments likely to introduce (that is, tax or subsidy), and in what way will these policies affect the location choice of the MNE? We leave these interesting questions for future research.

Kjetil Bjorvatn and Carsten Eckel 93

References Audretsch, D. B. and M. P. Feldman (1996) ‘R&D Spillovers and the Geography of Innovation and Production’, American Economic Review, vol. 86(3), pp. 630–40. Bjorvatn, K. and C. Eckel (2005) ‘Technology Sourcing and Strategic Foreign Direct Investment’, forthcoming in Review of International Economics. Blomström, M. and A. Kokko (1998) ‘Multinational Corporations and Spillovers’, Journal of Economic Surveys, vol. 12(3), pp. 247–78. Bransetter, L. G. (2001) ‘Are Knowledge Spillovers International or Intranational in Scope? Microeconometric Evidence from the U.S. and Japan’, Journal of International Economics, vol. 53, pp. 53–79. Ethier, W. J. (1984) ‘The Multinational Firm’, Quarterly Journal of Economics, vol. 101, pp. 805–34. Ethier, W. J. and J. R. Markusen (1996) ‘Multinational Firms, Technology Diffusion and Trade’, Journal of International Economics, vol. 41, pp. 1–28. Fosfuri, A. (2000) ‘Patent Protection, Imitation and the Mode of Technology Transfer’, International Journal of Industrial Organization, vol. 18, pp. 1129–49. Fosfuri, A., M. Motta and T. Rønde (2001) ‘Foreign Direct Investment and Spillovers through Workers’ Mobility’, Journal of International Economics, vol. 53, pp. 205–22. Horstmann, I. J. and J. R. Markusen (1992) ‘Endogenous Market Structures in International Trade (Natura Facit Saltum)’, Journal of International Economics, vol. 32, pp. 109–29. Keller, W. (2001) ‘International Technology Diffusion’, NBER Working Paper no. 8573. Markusen, J. R. (2001) ‘Contracts, Intellectual Property Rights and Multinational Investment in Developing Countries’, Journal of International Economics, vol. 53, pp. 189–204. Siotis, G. (1999) ‘Foreign Direct Investment Strategies and Firms’ Capabilities’, Journal of Economics and Management Strategy, vol. 8(2), pp. 251–70.

5 Foreign Direct Investment in South Asia: Impact on Economic Growth and Local Investment Pradeep Agrawal Institute of Economic Growth, Delhi, India

1

Introduction

Foreign direct investment is increasingly becoming an important source of investment funds in developing countries. Many economists have hailed it as an important source of new technology and management know-how and a useful link to world markets (see, for example, Balasubramanyam, Salisu and Sapsford, 1996; Fry, 1993). On the other hand, some concerns have also been voiced in developing countries about whether the multinational firms might have an adverse effect on the development of domestic firms, or otherwise be a source of economic exploitation of developing countries. Thus it is important that we understand better the economic role of foreign direct investment (FDI) in developing countries in general, and in South Asia in particular. Two of the important issues of contention are: (i) Does FDI crowd out domestic private investment or does it increase it by fostering various backward and forward linkages with domestic firms? (ii) Does FDI increase GDP growth by creating jobs, increasing exports, bringing in new management and production techniques, or does it lower GDP growth in the long run by taking excessive profits out of the developing country. These issues can be analysed econometrically. A few cross-country studies exist which suggest that, by and large, FDI is economically 94

Pradeep Agrawal 95

beneficial to developing countries (see, for example, Fry, 1993; Balasubramanyam, Salisu and Sapsford, 1996; Borensztein, De Gregorio and Lee (1998) and other chapters in this volume). However, additional work is needed to carry out more careful and country/region-specific studies to eliminate doubts that the cross-country results might be driven by a few extreme cases and might not apply to the specific country or region in question. In this chapter, we analyse the above-mentioned issues econometrically. We undertake time-series cross-section (panel) analysis of data for five South Asian countries: India, Pakistan, Bangladesh, Sri Lanka and Nepal. This is reasonable, since the South Asian countries generally have similar economic structures and have followed similar policies towards FDI. The analysis uses about 25–32 years of data over the period 1965–96 from each of the five countries. Our goal is to provide a clearer understanding of the economic impact of FDI in South Asia. The rest of this chapter proceeds as follows. Section 2 provides a brief overview of policies towards foreign direct investment and presents data on inflows of FDI into India (the largest country in South Asia). Section 3 analyses the effect of FDI on local investment. Section 4 analyses the effect of FDI on GDP growth in South Asia. Main conclusions are summarized in Section 5.

2 Foreign direct investment policies and inflows: The case of India To give an idea of the kind of policy debates and changes regarding FDI that have taken place in South Asia, this section provides a brief overview of the policies towards FDI in India, the largest and most important South Asian country. A discussion of the FDI inflows into India and its breakdown by country of origin and by industry is also included. 2.1

A brief history of India’s policies towards FDI

India’s policy towards foreign direct investment can be classified into four phases: Phase I (1957–67) had a cautiously welcoming approach towards FDI; Phase II (1967–79) can be characterized as a restrictive regime; Phase III (1980–90) was marked by the progressive attenuation of regulations brought about in the 1970s; and Phase IV (1991–2001)

96 FDI in South Asia

has a liberal and welcoming approach to foreign investment (Gopinath, 1997). These phases are discussed briefly below. Following independence in 1947, India initially had an ambivalent approach towards FDI, alternating between the nationalist distrust of colonial firms and the hope that new foreign investment could provide the technology and capital essential for rapid industrialization. The positive view had the upper hand during the second and the third fiveyear plans, spanning the period 1957–67. Thus a cautiously welcoming approach was adopted during this period that included features such as non-discriminatory treatment towards FDI, easy repatriability of profits by foreign firms, and an emphasis on exports by foreign-controlled firms. However, an acute foreign exchange crisis began to develop in the late 1960s, and it was felt that foreign firms were contributing to the problem through the import of inputs and repatriation of profits. The government nationalized some major oil-producing and retailing multinational companies in the early 1970s. The Foreign Exchange Regulation Act (FERA) was passed in 1973, marking the tightening of a regulatory regime regarding the management of foreign capital. A process of indigenization and dilution of foreign equity was carried out whereby foreign companies were required to dilute the non-resident shareholding within two years to the level prescribed by the Reserve Bank of India, which was generally set at 40 per cent. A restrictive approach was adopted for non-cash inflows and the use of foreign brand names for sales within India. Domestic firms requiring foreign technology were encouraged to acquire it through technology licensing rather than through joint ventures. However, companies in high-tech and skills areas were allowed foreign shareholdings of up to 74 per cent (see Kumar, 1994). In the 1980s, growing concern about stagnation and technological obsolescence in Indian industry drew attention to the restrictive licensing procedures. The poor quality and high cost of Indian manufactured products contributed to the weak export performance, creating balance of payments problems in the wake of the second oil shock. The need was felt for foreign collaboration to improve exports and the quality of Indian manufacturing. As a consequence, there was an easing of restrictions on FDI that was further strengthened after Rajiv Gandhi took over as prime minister in 1984. FERA restrictions were relaxed for FDI in areas of high technology, and for 100 per cent export-orientated units. Foreign investment equity up to 74 per cent and 100 per cent was allowed in the priority sectors and export-orientated units, respectively, with full repatriation of profits. Four export-processing zones were set up.

Pradeep Agrawal 97

During 1984 and 1985, many capital goods were moved to the open general licence list, allowing easier importation. In 1986, the tax rate on royalty payments was brought down from 40 per cent to 30 per cent. However, bureaucratic bottlenecks remained, and foreign firms probably continued to distrust the Indian government. As a result, foreign equity inflows remained meagre and Indian industry came to rely largely on foreign debt capital to meet its foreign exchange needs over this period. A major change from the previous policy regime came in 1991, when India began a major liberalization of trade and investment policies. Tariffs were reduced sharply on most products, bringing the average weighted tariff rate on imports down from 87 per cent in 1990–1 to 25 per cent by 1994–5, and to 20 per cent by 1997–8. The foreign technology requirement for FDI was discontinued and a large number of additional sectors, including many consumer goods, were opened to foreign investment. The repatriation of profits by foreign-controlled firms was made easier, and the earlier requirement that the dividend payment over the first seven years from the start of commercial production must be balanced by the export earnings by the foreign venture was dropped, apart from in twenty-four consumer-goods industries. In 1994, India became a member of the Multilateral Investment Guarantee Agency (MIGA). As a result, all investments approved by the government of India are insured against expropriation or nationalization. Up to 100 per cent equity was allowed in a number of industries (especially those in export-orientated and high-tech industries). The Reserve Bank of India was authorized to give automatic approvals (usually within two weeks) to proposals in high priority areas where foreign equity does not exceed 51 per cent, and in the mining sector where foreign equity does not exceed 50 per cent. A Foreign Investment Promotion Board (FIPB) was set up, which acts as a single window clearance. It deals with large investment proposals and other proposals such as situations where foreign equity exceeds 51 per cent, the industry is not on the list of high-priority sectors, or foreign equity does not cover the import of capital goods. Proposals are usually cleared in four to six weeks. Unlike the Reserve Bank of India, FIPB can also initiate and carry out detailed negotiations with foreign firms. Overall, it is easy to see that, over the 1990s, the government has adopted a much more welcoming attitude towards foreign investors. For example, the Common Minimum Programme of the Vajpayee government included a goal of nearly tripling the FDI inflows to US$ 10 billion a year.

98 FDI in South Asia

This overview of the history of India’s policies towards FDI makes it clear that there has existed considerable confusion about the true economic impact of foreign direct investment on the national economy. We hope that the analysis of sections 3 and 4 will help to resolve some of the issues of concern in this regard. 2.2

FDI flows into India

Figure 5.1 shows FDI inflows into India and other South Asian countries as percentages of GDP over the period 1965–96. It may be seen that FDI inflows in India have increased sharply since 1992, following the policy change. The figure also shows a comparison with the FDI inflows (as percentages of GDP) into East Asia over the same period. It is quite evident that, despite the sharp increase since 1992, India has had a much lower level of FDI inflows than did the East Asian countries, and even some of the South Asian countries such as Sri Lanka and Pakistan. Next we present in Table 5.1 the breakdown of foreign direct investment in India by country of origin and by industry of affiliate. The FDI is that which was approved during 1991–2000, the period over which the bulk of the foreign direct investment came into India. It shows the USA as the largest investor nation in India, accounting for over 20 per cent of total FDI inflows over 1991–2000. In second place is Mauritius, which is used as a base for investors from various countries, including the USA, because of its special tax treaty with India that grants Mauritius-based companies exemption from Indian taxes. The UK is the third largest investor in India, followed by South Korea, Japan and non-resident (or overseas) Indians. South Korea has emerged as a major investor in India since 1995–6, particularly in areas of passenger cars and consumer durables such as refrigerators, televisions and washing machines. Other significant investor nations in India include Germany, Australia, Malaysia, France and the Netherlands. Malaysia has emerged as a significant investor in India since 1995. The breakdown of FDI and technical collaborations approved during the period 1991 to 2000, by country of origin, is shown in Table 5.2. The main sectors experiencing FDI inflows are seen to be power generation, oil refining, cellular mobile and basic telephone services, telecommunications and the transportation industry (especially the production of passenger cars). There were also significant FDI inflows in financial (banking, insurance, and so on) services, metallurgy, chemicals, and food processing industries.

5

Foreign direct investment as a percentage of GDP

4.5 4 3.5 3 2.5 2 1.5 1 0.5

1996

1994

1992

1990

1988

1986

1984

1982

1980

1978

1976

1974

1972

1970

1968

1966

1964

1962

–0.5

1960

0

Year EA Avg

Figure 5.1

India

Bangladesh

Nepal

Sri Lanka

Pakistan

Ratio of foreign direct investment inflows to GDP

Note: EA Avg refers to a simple average over six East Asian countries (South Korea, Taiwan, Singapore, Malaysia, Thailand and Indonesia).

99

Source: Author’s calculations and chart, based on data from World Development Indicators, World Bank (1998).

100

Table 5.1 Foreign direct investment approvals during 1991–2000 (percentage share of major investing countries) Country a

1991

1992

1993

1994

1995

1996

1997

1998

1999

2000 b

Total Avg. (1991–2000b )

USA Mauritius United Kingdom Korea, Republic of Japan Non-resident Indians Germany Australia Malaysia France Netherlands Israel Italy Singapore

34.80 0.00 6.01 1.15 9.87 3.69 7.83 0.49 0.03 3.62 10.47 0.00 3.33 0.26

31.68 0.00 3.03 1.01 15.70 11.30 2.22 2.00 1.91 0.76 2.49 0.03 2.30 1.55

39.08 1.40 7.03 0.33 2.91 11.78 1.99 0.33 0.10 1.46 3.63 0.02 1.32 0.75

24.59 3.77 9.16 0.75 2.83 3.46 4.01 2.74 0.18 0.63 1.46 0.06 2.76 1.87

22.00 5.64 5.38 0.98 4.72 2.21 4.18 4.69 4.32 1.31 3.01 12.90 1.44 3.09

27.82 6.46 4.22 8.91 4.12 6.06 4.25 2.31 0.12 4.62 2.90 0.04 0.38 0.88

24.72 19.00 8.18 3.56 3.47 3.31 3.93 0.79 3.83 1.30 1.59 0.09 2.18 1.57

11.56 10.27 10.39 1.20 4.16 2.44 2.77 8.56 5.85 1.67 1.61 0.04 0.90 2.49

12.60 13.41 10.45 12.86 5.62 1.60 4.03 2.29 0.41 5.11 2.23 0.03 6.20 2.91

14.63 20.19 1.35 0.17 3.19 3.03 1.48 0.18 0.03 0.31 0.91 0.00 0.41 0.61

21.26 11.57 6.98 4.16 4.22 3.70 3.53 2.82 2.38 2.19 2.10 1.81 1.95 1.84

Belgium Cayman Islands Switzerland Thailand Canada Hong Kong SAR Sweden South Africa Russia Taiwan GDR/FCCB c TOTAL All countries (US$ millions)

0.30 0.00 6.65 0.00 0.91 3.96 1.31 0.00 1.61 0.08 0.00 206.78

0.61 0.00 17.74 0.06 0.02 1.47 1.25 0.00 0.30 0.46 0.00 1 483.79

0.07 0.04 4.82 4.16 0.31 0.99 0.01 0.00 0.02 0.11 0.00 2 823.24

0.05 0.02 0.34 0.07 0.30 1.16 0.08 0.00 0.74 0.07 36.87 4 521.09

0.52 0.54 0.39 0.00 0.02 6.57 0.96 0.44 0.90 6.14 0.21 0.05 4.28 0.54 0.70 1.27 1.41 0.47 1.57 1.47 0.20 0.05 0.16 0.17 0.36 0.00 0.00 0.01 0.22 0.00 3.71 14.56 8.20 9 116.46 10 060.34 13 974.38

10.67 0.00 0.93 0.00 1.02 0.77 0.70 5.54 0.05 0.01 10.46 7 253.65

0.05 0.04 1.03 0.02 0.13 0.16 0.97 0.09 0.03 0.03 11.37 6 522.54

0.10 0.34 0.27 0.00 0.37 1.56 0.40 0.02 0.00 0.02 47.96 5 431.50

1.69 1.59 1.20 1.05 1.06 0.92 0.77 0.81 0.11 0.06 14.59 61 103.79

Notes: a Ranking of all countries according to cumulative total of approvals for the period 1991–2000; b Represents projected figures for 2000 based on data up to 31/8/2000; c Represents proposals for global depository receipts (GDR) and foreign currency convertible bonds (FCCBs). Source: Secretariat for Industrial Assistance (SIA) Ministry of Commerce and Industry, Government of India.

101

102 FDI in South Asia Table 5.2 Foreign direct investment approvals in India, August 1991–March 2000 (percentage share by industry) Sr. No.

Name

Total, All industries (millions of rupees) 1 Fuels Power generation Oil refining 2 Telecommunications Cellular mobile and basic telephone services Telecommunications 3 Transportation Industry Passenger cars Automobile industry Air/sea transport 4 Service sector Financial Non-financial services Electronics 5 Metallurgical Ferrous Mining services 6 Chemicals 7 Food processing industry Food products 8 Hotels and tourism Hotels and restaurants 9 Textiles including dyed, printed 10 Paper and pulp, including paper products 11 Industrial machinery Source:

3

Industry share 2,142.456 29.72 14.83 8.91 17.32 11.61 4.24 8.27 3.36 1.88 1.12 6.91 4.70 1.44 1.27 5.87 3.44 1.13 5.63 4.04 3.99 2.01 1.65 1.51 1.45 1.05

National Informatics Centre.

The impact of FDI on investment by local investors

In order to study the impact of FDI on local investment, we need to estimate an investment function and then analyse the impact of FDI. Thus, in this section, we discuss briefly the variables that affect investment and specify a likely functional form for it. Then we briefly discuss the econometric procedures used, and finally consider the results. 3.1

The investment function

Blejer and Khan (1984, p. 383) describe some of the difficulties of estimating neo-classical investment functions for developing countries, such as the lack of readily-available measures of the capital stock or its rate

Pradeep Agrawal 103

of return. Thus the investment function estimated here is based on the flexible accelerator model as developed in Fry (1998). The accelerator model postulates that the desired real capital stock k∗ is proportional to real output, y: k∗ = αy

(1)

This can be expressed in terms of a desired investment rate (INV /Y)∗ :


INV Y

∗ = αG

(2)

where INV denotes gross domestic fixed investment in current prices, Y denotes GDP in current prices, and G is the rate of growth of real GDP, y. A partial adjustment mechanism allows the actual investment rate to adjust to the difference between the desired investment rate and the investment rate in the previous period:


INV Y




=λ t

INV Y

∗


−

t

INV Y



 t−1

or,


INV Y






t

INV =λ Y

∗ t




INV − (1 − λ) Y

 (3) t−1

where λ is the coefficient of adjustment. The flexible accelerator model allows economic conditions to influence the adjustment coefficient, λ. That is: λ=

β0 + (β1 X1 + β2 X2 + β3 X3 + · · · )      INV ∗ − INV Y t Y t−1

(4)

where Xi are a set of explanatory variables (including an intercept term for a constant depreciation rate) that affect λ, and βi are their respective coefficients. There exists a vast literature on the determinants of the investment rate (see reviews in Fry, 1995, and Schmidt-Hebbel et al., 1996; see also Nickell, 1978; Abel, 1980; Wai and Wong, 1982; Blejer and Khan, 1984; Gonzales and Gerrardo, 1988; Greene and Villenueva, 1991; Fry, 1998). The explanatory variables commonly used in the context of a

104 FDI in South Asia

flexible accelerator model for developing countries (Fry, 1995; SchmidtHebbel et al., 1996) where firms’ investments often suffer from credit and foreign exchange constraints include: GDP growth rate (G) over the previous year; domestic credit availability as share of GDP (CRDT /Y); terms of trade (TOT ) – an index with base year (1990 = 100); and the real exchange rate (RER). Further, since we are interested in evaluating the impact of FDI and of foreign borrowing, we also include net foreign direct investment inflows as share of GDP (FDI/Y), and net total foreign borrowing as share of GDP (TEDtot/Y) as explanatory variables. In addition, the neo-classical theory of investment (see Jorgenson, 1967) emphasizes the role of the cost of capital (proxied here by the real interest rates on bank loans). Rama (1993) argues that, even in developing countries, while some firms may face credit and foreign exchange constraints, others might not – and the neo-classical model would be applicable to this latter group: he therefore recommends using the explanatory variables suggested by both the neo-classical and the accelerator approaches. Thus we have used all of the above-mentioned explanatory variables, which are explained below. The real GDP growth rate (G) over the previous year is an important determinant of investment, since higher growth would imply higher capital requirement and hence a higher investment rate. Further, the growth rate is also a good proxy for the quality of institutions and policies being pursued in an economy at a given time. These, in turn, affect the profit opportunities available in the economy and thereby the investment rate. The availability of bank credit as a fraction of GDP (CRDT /Y) is another important determinant of the investment rate in developing countries (see Blinder and Stiglitz, 1983; Fry, 1995). The quantity of credit is likely to be important in a credit market where the interest rates are controlled at below market clearing levels and/or directed credit programmes exist for selected industrial sectors. In fact, Stiglitz and Weiss (1981) have shown that asymmetric information between borrower and bank regarding the true default probability will lead to credit rationing even in perfectly-competitive liberalized markets. Further, banks specialize in acquiring information on default risk, highly specific to each client. Hence the market for bank loans is a customer market, in which borrowers and lenders are very imperfect substitutes. A credit squeeze rations out some bank borrowers who may be unable to find loans elsewhere and so be unable to finance their investment projects (Blinder and Stiglitz, 1983, p. 300). Also, asymmetric information will lead to credit rationing even in perfectly competitive liberalized markets. Here, therefore, the investment rate INV /Y is influenced by the ratio of domestic bank

Pradeep Agrawal 105

credit to GDP, CRDT /Y. Its first difference, the change in the volume of domestic credit scaled by GDP, (CRDT /Y), which has been emphasized by some authors (see, for example, Blejer and Khan, 1984, p. 389), enters as dynamic terms in both the ECM and DOLS co-integration procedures used here. Now consider the impact of an increase in the real exchange rate, RER, defined as follows: RER = E · P f /P

(5)

where E is the exchange rate (number of domestic currency units per US$), P is the domestic price level (the GDP deflator) and P f is the foreign price level (proxied here by the US GDP deflator, given that the USA is the most important trading partner of the countries being studied). An increase in RER would increase the price of imported capital and intermediate goods and result in a contraction of investment (Serven and Solimano, 1992; Fry, 1995). Van Wijnbergen (1982) develops a two-sector model which shows that the net effect of a real depreciation is ambiguous – investment in tradable goods increases while that in domestic goods declines. An improvement in the terms of trade, TOT (unit price of exports divided by the unit price of imports), leads to an increase in real income for a nation; this in turn can increase investment, especially if capital goods (which are mainly importables in developing countries) become cheaper relative to non-tradable domestic goods. In some situations, however, an improvement in the terms of trade could also decrease investment by decreasing the demand for domestic goods compared to imported consumer goods (see, for example, Cardoso, 1993). Note that, while both TOT and RER are measuring somewhat related effects, they are not identical. In particular, while RER measures the impact of changes in inflation and exchange rates, TOT also includes the effect of changes in international prices and the competitiveness of various tradable goods. True domestic costs of capital are very difficult to measure in the South Asian developing countries examined here because of the lack of data on tax rates, selective credit policies and disequilibrium institutional interest rates. We therefore use the real lending rate of banks (RL) as a proxy for the cost of capital. The real lending rate is obtained by subtracting the average of the current and following year’s inflation rates from the nominal prime lending rate of the banks.

106 FDI in South Asia

Foreign borrowing can be used as a source of funds for investment, although public foreign borrowing does have a tendency to go (partially) into meeting urgent government budgetary requirements as well. Thus, we also include, as an explanatory variable, the net total foreign borrowings as a share of GDP (TEDtot/Y). (Unfortunately, private borrowing in India prior to the liberalization in 1991 was too minuscule to permit a separate analysis.) This variable will allow us to compare the relative effectiveness of foreign borrowing and FDI inflows in promoting investment. Finally, the primary interest here is in studying the impact of FDI inflows on the investments made by local investors.1 The use of this source of funds depends on the various policies towards FDI, such as the maximum level of foreign equity share allowed, regulations relating to repatriation of profits, various other policies and regulations relating to industry, labour and the size of the domestic market. FDI can also promote local investment through backward and forward linkages with the local industries. While we recognize that FDI is a source of funds for investment, and not investment itself, unfortunately data are not available in India for gross domestic capital formation by affiliates of foreign investors. Given this limitation, we use FDI inflows as a proxy for real investment undertaken by these affiliates, and use FDI inflows as a share of the gross domestic product, FDI/Y, as an explanatory variable. We proxy the locally-owned gross fixed investment by gross fixed domestic investment minus the net FDI inflows, denoting it by INVnf . Then the dependent variable is defined as the ratio of locally-owned gross fixed investment to gross domestic product (INVnf /Y). Given the above explanatory variables, Equations (3) and (4) suggest estimating a long-run relation of the type:

INVnf CRDT FDI = b0 + b1 Yˆ + b2 + b3 + b4 TOTt + b5 RER Y Y Y b6 TEDtot + RL + Y

(6)

For our purposes, the crucial variable here is FDI/Y. The coefficient, b3 , of FDI/Y should be zero if FDI has no impact on local investment. If FDI is associated with a decline in investment from local investors, b3 should be negative, while if FDI inflows are associated with an increase in investment by local investors, the coefficient b3 should be positive.

Pradeep Agrawal 107

3.2

Econometric procedure used

The above relation is estimated using pooled time-series, cross-section data for 25–32 annual observations between 1965 and 1996, for India and four other South Asian countries. A dynamic version of the linear model was estimated by including the lagged dependent variable, since the explanatory variables can be expected to determine the change in the investment rate rather than its absolute level. We used the fixed-effects model, which is typically appropriate when, as is the case here, one has a relatively small number of countries and a large number of observations for each country. In this model, dummy variables were included for all but one country, although, for brevity, their coefficients are not reported here.

3.3

The results

The results of the estimation are shown in Table 5.3. It can be seen that most of the variables have the expected sign: higher GDP growth and additional credit availability are associated with increased investment. The impact of net total foreign borrowing as a share of GDP (TEDtot/Y) is positive, and that of real lending rate (RL) on investment is negative (column C). The coefficients of these variables have the correct signs but their magnitudes are not significant. They were therefore dropped from the estimation. Similarly, the changes in the terms of trade (TOT ) and the real exchange rate (RER) have an insignificant effect on investment (see columns C and B), and they too were dropped from the estimated relation. The resulting relation is reported in column A. The crucial variable, the ratio of net FDI inflows to GDP (FDI/Y), has a strongly positive effect with a coefficient in the range of 1.7 to 2.8 in columns A to C. Note that when the lagged dependent variable is included in the regression, as is the case here, the long-run coefficient is obtained by dividing the coefficient of FDI/Y by 1 − the coefficient of the lagged dependent variable.2 Using this result, the long-run coefficient of FDI/Y is found to be in the range of 4 to 5. This implies that a 1 per cent increase in FDI is associated with a 4–5 per cent increase in locally-owned investment in the long run. This suggests that there exists strong complementarity between FDI and local investment, possibly through various backward and forward linkages. However, there is a possibility that this result could be driven by the requirement of less than 100 per cent equity ownership by foreign direct investors (because this requirement leads to local investment in the

108 FDI in South Asia Table 5.3 Impact of FDI inflows on investment rate (net of FDI) for South Asia, 1965–96, OLS panel estimation (fixed-effects model) Explanatory variable Constant GDP growth rate (G) Total domestic credit as share of GDP (CRTOT /Y) FDI (net) inflow as share of GDP (FDI/Y) Terms of trade (TOT ) Real exchange rate (RER)

A

B

C

4.263 (4.366)∗∗ 0.0399 (0.759) 0.0929 (3.770)∗∗ 1.917 (4.005)∗∗ –

5.191 (3.396)∗∗ 0.1034 (1.650) 0.0913 (3.078)∗∗ 1.705 (2.809)∗∗ −0.01003 (−1.260) −0.0115 (−0.216)

5.348 (3.031)∗∗ 0.0977 (1.232) 0.129 (3.770)∗∗ 2.853 (3.847)∗∗ −0.617 (−0.680) −0.263 (−0.041) 0.0312 (0.561) −0.0119 (−0.246) 0.447 (4.656)∗∗ 5.159 0.906 1.569 2.13 69

–

Net foreign borrowing as share of GDP (TEDtotY) Real lending rate (RL) Lagged dependent variable (INVnf /Y(−1)) Long-term coefficient for FDI/Y R2 Standard error of regression Durbin–Watson statistic (DW) No. of observations (N)

0.566 (8.204)∗∗ 4.417 0.873 1.468 1.91 127

0.570 (7.252)∗∗ 3.965 0.890 1.506 1.87 98

Notes: The t -ratios of regression coefficients are given in brackets; ∗∗ /∗ imply significance at 1%/5% confidence level. Source:

Author’s calculations.

affiliates of the foreign investor, which shows up as domestic investment in the national statistics). It is worth noting here that in India (unlike in some other countries), only the part of investment in joint ventures contributed by foreign affiliates is counted as FDI, while the part contributed by the local partner is counted as domestic investment. Requirements of less than 100 per cent foreign equity have been common in most South Asian countries, except in some selected sectors such as those that are highly export-orientated and those in certain hightech areas. For example, before 1992 India did not allow more than 40 per cent foreign ownership of a firm. Given that such foreign equity restrictions in South Asia were gradually relaxed over the 1980s and even further over the 1990s, one would expect this coefficient to decrease over time. To test this hypothesis, we repeated the regression in column A of Table 5.3 over: (i) all available observations over 1965–96; (ii) all

Pradeep Agrawal 109 Table 5.4 Impact of FDI inflows on investment rate (net of FDI) for South Asia for four different time periods, OLS panel estimation (fixed-effects model) Explanatory variable

A All obs.

Constant

4.263 (4.366)∗∗ 0.0399 (0.759) 0.0929 (3.770)∗∗ 1.917 (4.005)∗∗ 0.566 (8.204)∗∗ 4.417

GDP growth rate (G) Total domestic credit as share of GDP (CRTOT /Y) FDI (net) inflow as share of GDP (FDI/Y) Lagged dependent variable (INVnf /Y(−1)) Long-term coefficient for (FDI/Y) R2 Standard error of regression Durbin–Watson statistic (DW) No. of observations (N)

0.873 1.468 1.91 127

B 1980–96 4.535 (1.940)∗ 0.054 (0.722) 0.107 (2.521)∗∗ 1.461 (2.402)∗∗ 0.519 (5.498)∗∗ 3.037 0.893 1.453 2.263 84

C 1985–96 4.956 (1.440) 0.0897 (1.050) 0.0413 (0.767) 0.934 (1.605) 0.660 (5.977)∗∗ 2.747 0.929 1.090 2.099 60

D 1990–6 2.623 (0.565) 0.154 (1.181) 0.0415 (0.518) 0.476 (0.630) 0.759 (5.095)∗∗ 1.975 0.930 1.175 2.007 35

Notes: The t -ratios of regression coefficients are given in brackets; ∗∗ /∗ imply significance at 1%/5% confidence level. Source:

Author’s calculations.

observations over 1980–96; (iii) all observations over 1985–96; and (iv) all observations over 1990–6. The results are shown in Table 5.4, columns A to D, respectively (using the same explanatory variables as in column A of Table 5.3). It can be seen that the coefficient of FDI/Y does indeed decrease from 1.92 for the 1965–96 period to only 0.48 over the 1990–96 period. This implies that the long-term coefficient of FDI/Y declined from 4.42 over the 1965–96 period to 1.98 for the 1990–6 period. This corroborates our hypothesis that the complementarity between FDI and local investment was, at least partly, policy driven. However, it is noteworthy that the coefficient of FDI/Y remains positive even over the 1990–96 period, so that the FDI inflows have a positive effect on the locally-owned investment over all periods of our analysis. Further, the long-run coefficient for gross fixed investment net of FDI when foreign equity is 40 per cent (as in pre-1992 India) would be 1.5, which would decline to 0.33 when foreign equity increases to about 75 per cent (this is a reasonable estimate for India for the 1992–6 period, since foreign equity limits varied from 51 per cent to 100 per cent, depending on the sector of operation). Since the estimated long-run coefficients for these periods are 4.42 and 1.98, respectively, they

110 FDI in South Asia

suggest that the FDI inflows had a positive impact on the locally-owned investment over and above that necessitated by the policy restrictions on foreign equity share. Thus the complementarity between FDI and the locally-owned investment can be expected to continue even if the restrictions on foreign equity share are further liberalized or removed altogether. However, doing so would dilute the complementarity – that is, increases in local investment accompanying a given amount of FDI would be lower. Thus further liberalization of equity restrictions on foreign firms would be especially desirable when doing so attracts considerably larger inflows of FDI, as was the case in India following the 1992 liberalization of restrictions on foreign equity share, and the impact of FDI inflows on economic growth is positive. The latter aspect is considered in the next section.

4

Impact of FDI on GDP growth

In the previous section, FDI was shown to have had a strongly complementary effect on investment, leading to additional investment by host-country investors, which is several times the FDI inflow. However, the more important question is how FDI inflows might affect GDP growth. FDI inflows could promote GDP growth by providing additional employment in a labour-surplus economy, and by improving technological know-how and human capital. On the other hand, as the analysis of Brecher and Diaz-Alejandro (1977) suggests, foreign capital inflows could lead to immiserizing growth when such inflows can earn excessive profits in the host country, which may be particularly likely in economies subject to various trade and financial distortions. India and most other South Asian economies suffered from severe trade distortions while they were following protectionist policies in the 1960s and 1970s; these eased slowly as the countries liberalized gradually over the 1980s and 1990s. Thus the issue assumes added significance in the context of South Asia. To test the impact of FDI on growth, we use the conventional neoclassical production function, but add foreign capital as an additional variable.3 Further, following a large number of empirical studies (Ram, 1985; Salvatore and Hatcher, 1991; Greenaway and Sapsford, 1994; and Edwards, 1996) that have supported the export-led growth hypothesis, we also introduce exports as a variable in the production function. This is done because exports, like FDI, can result in a higher rate of technological innovation and dynamic learning from abroad. They also impose a certain market discipline, thus reducing the rent-seeking ability of special

Pradeep Agrawal 111

interest groups, and thereby minimizing distortions in the economy (Agrawal et al., 2000). Thus the production function can be written as: y = F(L, kd , kf , x, t)

(7)

where y = gross domestic product (GDP) in real terms; L = labour input; kd = stock of domestic capital in real terms; kf = stock of foreign capital in real terms; x = exports in real terms; and t = a time trend that captures the improvement in productivity resulting from technical progress. Assuming the production function to be log-linear (but not necessarily linear homogeneous), taking logs and differentiating with respect to time, we obtain: G = a + bkˆ d + c kˆ f + d xˆ + eLˆ + u

(8)

where a hat on a variable denotes its growth rate – thus, for example, . G denotes the growth rate of real GDP (y), and u denotes a Lˆ = 1L dL dt random-error term consistent with the assumption of a log-linear production function. In the context of the surplus-labour economies of South Asia, growth of the labour force is not likely to be a significant determinant of GDP growth (this was also confirmed by empirical estimations – not shown here, for brevity). Thus this variable was dropped as an explanatory variable. Furthermore, in view of serious difficulties associated with measuring capital stocks (even more so in the context of developing countries), we follow the precedent set in numerous previous studies and approximate the rate of growth of domestic and foreign capital by the ratio of domestic fixed investment (net of FDI) to GDP (INVnf /Y) and the ratio of net FDI inflows to GDP (FDI/Y). Thus the equation to be estimated is: G =a+b

INVnf FDI +c + d xˆ + u Y Y

(9)

For our purposes, the crucial variable here is again FDI/Y. Note that the coefficient c of this variable should be equal to the coefficient

112 FDI in South Asia

b of INVnf /Y if FDI is just as efficient in promoting GDP growth as locally-owned investment. If the greater technological know-how, human capital or exporting capabilities of FDI make it more efficient in promoting growth, the coefficient c can be expected to be greater than coefficient b. On the other hand, if FDI takes excessive profits out of the country without contributing much to the domestic economy in terms of technology, and so on, the coefficient c should be smaller than the coefficient b. Finally, if the coefficient c of FDI/Y were to be negative, it would imply a net negative impact on GDP growth – that is, that immiserizing growth resulted from FDI inflows. The above relation is estimated using pooled time-series, cross-section data for 25–32 annual observations between 1965 and 1996 for the five countries under consideration. We again used a fixed-effects model. The results of the estimation are shown in Table 5.5, column A. It is seen that both a higher rate of growth of exports (ˆx) and a higher investment rate (INV /Y) are associated with a higher rate of GDP growth. The crucial coefficient of FDI/Y is negative, though not statistically significant. This result suggests that, in the case of South Asia, FDI is neither significantly Table 5.5 Impact of FDI inflows on GDP growth in South Asia, 1965–96, OLS panel estimation (fixed-effects model) Explanatory variable

A All obs.

Constant

1.043 (0.617) 0.0452 (2.837)∗∗∗ 0.170 (1.961)∗∗

Growth rate of real exports (ˆx) Fixed investment (net of FDI) as share of GDP (INVnf /Y) FDI (net) inflow as share of GDP (FDI/Y) R2 Standard error of regression Durbin–Watson statistic (DW) No. of observations (N)

B 1980–96

C 1985–96

D 1990–6

2.627 (0.993) 0.0873 (4.042)∗∗∗ 0.107 0.872

2.998 (0.804) 0.0337 (1.403) 0.106 (0.625)

2.301 (0.526) 0.0376 (1.289) 0.102 (0.541)

0.179 (0.200)

0.595 (0.703)

1.378 (1.401)∗

0.121 2.485

0.261 2.063

0.179 1.745

0.161 1.682

2.31

2.385

2.17

2.29

132

85

60

35

−0.302 (−0.361)

Notes: The t -ratios of regression coefficients are given in brackets; ∗∗∗ /∗∗ /∗ imply significance at 1%/5%/10% confidence levels. Source:

Author’s calculations.

Pradeep Agrawal 113

harmful nor beneficial. But then Brecher and Diaz-Alejandro (1997) did suggest that the main case where FDI could result in immiserizing growth was if there were trade and financial market distortions, such that FDI earned excessive profits. In the case of most South Asian countries, there did exist considerable trade and financial market distortions during the 1960s and 1970s. But, as noted, there has been gradual movement over the 1980s and 1990s in the direction of economic liberalization and a reduction in the trade and financial market distortions. Thus it would be of interest to examine whether the economic impact of FDI has changed over time. Accordingly, we re-estimated the above relation with observations only for periods 1980–96, 1985–96 and 1990–6. The results of these three additional regressions are also shown in Table 5.5, columns B, C and D. It is seen that the sign of the coefficient of the FDI/Y variable becomes positive over the period 1980–96 and gradually increases in magnitude as we move to later periods. Over the period 1990–6, it is positive and statistically significant. In fact it is as much as 1.37 – almost thirteen times the coefficient for local investment, INVnf /Y. We thus find that the FDI inflows helped to achieve faster economic growth in South Asia in the liberalized environment beginning in the 1980s, and especially over the 1990s. This may be a result of a reduction in excessive profits because of reduced distortions and a greater contribution to technological knowhow in the more competitive environment over the 1980s and 1990s. If this interpretation is correct, then further FDI inflows would appear to be desirable and should be encouraged. In this context, it is also worth comparing the relative merits of FDI inflows and foreign borrowing (an alternative form of foreign capital). For this purpose, we re-estimated Equation (9) after adding another variable – namely, net total (private and public) additional foreign borrowing as a share of GDP, TEDtot/Y (this was obtained by calculating the change in total external debt in US dollars, converting it to current local units and dividing by nominal GDP). That is, we estimated the following relation:

G =a+b

INVnf FDI TEDtot +c +d + exˆ + u Y Y Y

(10)

This relation was estimated over the same four periods as in Table 5.5. The estimation results are shown in Table 5.6. It is seen that, over the periods since 1980, the coefficient of TEDtot/Y is positive and statistically

114 FDI in South Asia Table 5.6 Impact of FDI inflows on GDP growth in South Asia, 1965–96, OLS panel estimation (fixed-effects model) Explanatory variable

Constant Growth rate of real exports (ˆx) Fixed investment (net of FDI) as share of GDP (INVnf /Y) FDI (net) inflow as share of GDP (FDI/Y) Net foreign debt inflow as share of GDP (TEDtot/Y) R2 Standard error of regression Durbin–Watson statistic (DW) No. of observations (N)

A All obs.

B 1980–96

C 1985–96

1.0033 (0.596) 0.0471 (2.956)∗∗∗ 0.1647 (1.897)∗

2.498 (0.939) 0.08520 (4.080)∗∗∗ 0.1075 (0.872)

1.676 −0.473 (0.430) (−0.109) 0.0352 0.0348 (1.468)∗ (1.271) 0.153 0.212 (0.892) (1.146)

0.225 (0.252) 0.0547 (0.663)

0.7668 (0.895) 0.0906 (1.146)

0.261 2.063 2.385 85

0.199 1.740 2.12 60

−0.288 (−0.345) 0.077 (1.418) 0.135 2.476 2.28 132

D 1990–6

1.840 (1.940)∗ 0.2066 (2.148)∗∗∗ 0.287 1.580 2.16 35

Notes: The t -ratios of regression coefficients are given in brackets; ∗∗∗ /∗∗ /∗ imply significance at 1%/5%/10% confidence levels. Source:

Author’s calculations.

significant. However, it is small (d = 0.21 over 1990–6) compared to the coefficient of FDI/Y (c = 1.84 over 1990–6). This suggests that FDI inflows are more beneficial than foreign borrowing. It may be worth noting that, since 1980, FDI inflows have averaged about 5 per cent of GDP in mainland China and about 3 per cent of GDP in the rapidly growing East Asian economies, while they have averaged only 0.5 per cent in South Asia (see Figure 5.1). Thus, there would seem to be considerable scope for increasing FDI inflows in South Asia, and doing so could increase the South Asian GDP growth rates by a few percentage points.

5

Conclusions

In this study, we have analysed India’s policies and experience with foreign direct investment. Prior to the liberalization of foreign investment regulations in 1992, India’s policies towards FDI had oscillated several

Pradeep Agrawal 115

times between the nationalistic distrust of the foreign investors and the hope that they might help to modernize India’s industrial sector by bringing in modern technology and management practices. This suggested the need to undertake empirical analysis to determine the true economic impact of foreign direct investment on the national economy. Thus, in this chapter, we tried to infer the economic impact of FDI in India and South Asia by using panel data from the five main South Asian countries to estimate the impact of FDI inflows on locally-owned investment and on GDP growth. We find that an increase in the FDI inflows in South Asia was associated with a considerable increase in investment by local investors. This was probably driven partly by the requirement in most South Asian countries of less than 100 per cent equity ownership by foreign direct investors, which leads to complementary investment by a local partner (counted as domestic investment in India). This was corroborated by the gradual decline over the 1980s and 1990s (decades that witnessed a gradual liberalization of the equity restrictions on foreign investors) in the extent of positive impact of FDI inflows on local investment. Nevertheless, the long-run positive impact of FDI on local investors is found to be considerably more than the equity restrictions on foreign investors would imply by themselves. Thus our analysis suggests that there exist positive linkage effects between foreign and local investment over and above those imposed by the restrictions on foreign equity ownership. We also considered the impact of FDI inflow on the GDP growth rate. It was found to be insignificant over the 1970s and 1980s, but became increasing positive over the 1980s, and in particular over the 1990s (which is when a great deal of FDI came into India and South Asia – see Figure 5.1). It is noteworthy that most South Asian countries followed import-substitution policies and had high import tariffs in the 1960s and 1970s. These policies gradually changed over the 1980s, and by the early 1990s most countries had largely abandoned import-substitution strategy in favour of more open international trade and generally marketorientated policies. Thus our econometric results can be understood in terms of the analysis of Brecher and Diaz-Alejandro (1977), who argued that foreign capital would be less beneficial (or even harmful) in the presence of trade and financial market distortions than it would otherwise be. Thus, our analysis suggests that FDI inflows have a positive impact on economic growth under the economic conditions prevailing at the time of writing in South Asia. We also found that, since 1980, FDI inflows contributed more to investment and to GDP growth in South Asia than an equal amount of foreign

116 FDI in South Asia

borrowing. This suggests that, to the extent that some foreign capital is needed in the economy, FDI is preferable to foreign borrowing. The foregoing results provide support for more liberal policies towards FDI. However, it should be noted that we did not find FDI to be necessarily beneficial under all conditions – in fact, we found it to be significantly beneficial in South Asia only over 1990–6. Just as trade and financial market distortions might have led to excessive profits for foreign investors and little improvement in GDP growth in the 1970s and 1980s, it is possible that excessive concessions on taxes, or opening some of the more vulnerable or non-competitive sectors of the economy to FDI, could reduce the positive impact of FDI on the national economy in the future. Therefore, care should be taken not to give excessive concessions to FDI; and vigorous competition should be encouraged in all industrial sectors that are opened to FDI, including among foreign investors (by encouraging many foreign investors to come into each sector), to minimize the possibility of excessive profits being made by the foreign investors. Notes 1 Although the local subsidiaries of multinational firms (or the parent organization itself) may also borrow funds in the national or international financial markets, this is not of much consequence for our analysis and is treated as FDI only since the multinational firm takes responsibility for repayment (just as the money invested by the local firm is treated as investment even if part of it is borrowed). 2 To see the logic behind this, note that, if Yt = aXt +bYt−1 , then in the long-run equilibrium, where Yt = Yt−1 = Y and Xt = X, we have Y = aX + bY, so that Y = [a/(1 − b)]X. 3 This approach has been previously used by Balasubramanyam et al. (1996).

References Abel, A. (1980) ‘Empirical Investment Equations: An Integrated Framework’, Carnegie-Rochester Conference Series on Public Policy, vol. 12, pp. 39–91. Agrawal, Pradeep, Subir Gokavn, Veena Mishra, Kirit Parikh and Kunal Sen (2000) Policy Regimes and Industrial Competitiveness: A Comparative Study of East Asia and India (London: Macmillan). Balasubramanyam, V. N., M. Salisu and D. Sapsford (1996) ‘Foreign Direct Investment and Growth in EP and IS Countries’, Economic Journal, vol. 106, pp. 92–105. Blejer, M. and M. Khan (1984) ‘Government Policy and Private Investment in Developing Countries’, IMF Staff Papers, vol. 31, pp. 379–403. Blinder, Alan S. and J. E. Stiglitz (1983) ‘Money, Credit Constraints and Economic Activity’, American Economic Review, vol. 73, pp. 297–302.

Pradeep Agrawal 117 Borensztein, E., J. De Gregorio and J. W. Lee (1998) ‘How Does Foreign Direct Investment Affect Economic Growth?’, Journal of International Economics, vol. 45(1), pp. 115–35. Brecher, R. A. and C. F. Diaz-Alejandro (1977) ‘Tariffs, Foreign Capital and Immiserizing Growth’, Journal of International Economics, vol. 7, pp. 317–22. Cardoso, Eliana (1993) ‘Private Investment in Latin America’, Economic Development and Cultural Change, vol. 41, pp. 833–48. Edwards, S. (1996) ‘Why Are Latin America’s Saving Rates So Low? An International Comparative Analysis’, Journal of Development Economics, vol. 51, pp. 5–44. Fry, Maxwell J. (1993) Foreign Direct Investment in Southeast Asia: Differential Impacts, ISEAS Current Economic Affairs Series (Singapore: Institute of Southeast Asian Studies, ASEAN Economic Research Unit). Fry, Maxwell J. (1995) Money, Interest and Banking in Economic Development, 2nd edn (Baltimore, MD.: Johns Hopkins University Press). Fry, Maxwell J. (1998) ‘Saving, Investment, Growth and Financial Distortions in Pacific Asia and Other Developing Areas’, International Economic Journal, vol. 12, pp. 1–25. Gonzalez, A. and M. Gerrardo (1988) ‘Interest Rates, Savings and Growth in LDCs: An Assessment of Recent Empirical Research’, World Development, vol. 16, pp. 589–605. Gopinath, T. (1997) ‘Foreign Investment in India: Policy Issues, Trends and Prospects’, Reserve Bank of India Occasional Papers, vol. 18(2&3), Special Issue, pp. 453–70. Green, Joshua and Delano Villenueva (1991) ‘Private Investment in Developing Countries’, IMF Staff Papers, vol. 38(1), pp. 33–58. Greenaway, D. and D. Sapsford (1994.) ‘What Does Liberalisation Do for Exports and Growth?’, Weltwirtschaftliches Archiv, vol. 130, pp. 152–73. Jorgensen, Dale W. (1967) ‘The Theory of Investment Behavior’ Paper presented at Conference of the Universities – National Bureau of Economic Research, ‘Determinants of Investment Behavior’ (New York: Columbia University Press), pp. 129–56. Kumar, Nagesh (1994) Multinational Enterprises and Industrial Organisations – The Case of India (New Delhi: Sage). Nickell, S. (1978) The Investment Decisions of Firms (Cambridge: Cambridge University Press). Ram, R. (1985) ‘Exports and Economic Growth. Some Additional Evidence’, Economic Development and Cultural Change, vol. 33, pp. 415–25. Rama, Martin (1993) ‘Empirical Investment Equations for Developing Countries’, in L. Serven and A. Solimano (eds), Striving for Growth after Adjustment: The Role of Capital Formation (Washington, DC: The World Bank), pp. 107–43. Salvatore, D. and T. Hatcher (1991) ‘Inward Oriented and Outward Oriented Trade Strategies’, Journal of Development Studies, vol. 27, pp. 7–25. Schmidt-Hebbel, K., K. L. Serven and A. Solimano (1996) ‘Private Investment and Macroeconomic Adjustment: A Survey’, World Bank Research Observer, vol. 7, no. 1, pp. 95–114. Serven, Luis and Andres Solimano (1992) ‘Private Investment and Macroeconomic Adjustment: A Survey’, World Bank Research Observer, vol. 7(1), pp. 95–114.

118 FDI in South Asia Stiglitz, J. and A. Weiss (1981) ‘Credit Rationing in Markets with Imperfect Information’, American Economic Review, vol. 71, pp. 393–410. Van Wijnbergen, S. (1982) ‘Stagflationary Effect of Monetary Stabilization Policies: A Quantitative Analysis of South Korea’, Journal of Development Economics, vol. 37, pp. 133–69. Wai, Tun U. and Chorn-huey Wong (1982) ‘Determinants of Private Investment in Developing Countries’, Journal of Development Studies, vol. 18, pp. 19–36. World Bank (1998) World Development Indicators (Washington, DC: The World Bank).

6 The Internationalization of Korean Firms: Strategic Interaction and Tariff-jumping when Quality Matters∗ Bénédicte Coestier University of Paris X, Nanterre, France

and Serge Perrin Agence Française de Développement, Paris, France

1

Introduction

Korean multinational enterprises (MNEs) have recently emerged as overseas investors, not only increasing their penetration of developing markets but also making significant inroads into industrialized markets. This shift from a traditional export orientation – the source of Korea’s rapid economic growth – occurred in a context where Korean manufacturing industries were hard hit by steep wage increases and an appreciating currency, multiple import restrictions imposed by developed countries, and a catching-up process engaged by a second wave of dynamic Asian economies such as Malaysia and Thailand. This chapter analyses Korean firms’ foreign direct investment (FDI) decision in developed countries. In contrast to the literature on strategic investment (Smith, 1987; Motta, 1992), which emphasized the entry-deterring nature of FDI, Korean multinationals’ direct investment in developed countries can be seen as a response to trade restraints which threatened to increase costs vis-à-vis local rivals (tariff-jumping FDI). In addition, Korean latecomers aimed at catching ∗

An earlier version of this chapter was presented as a paper at the EARIE Annual Conference, Madrid, 5–8 September 2002, and at the World Congress of the International Economic Association, Lisbon, 9–13 September 2002. The authors wish to thank Edward M. Graham for helpful comments. The usual disclaimer applies. 119

120 The Internationalization of Korean Firms

up the Japanese leaders, and may have faced competitive pressure to invest. We characterize this situation in a simple model where an emerging multinational/low-quality firm engages in price competition with an incumbent/high-quality firm in a context of vertical product differentiation, and examine the strategic considerations that influence the multinational’s choice between exports and FDI. This follows Belderbos’ (1997a) analysis on tariff-jumping FDI by Japanese multinationals in a Cournot setting. We show that FDI is more likely the lower the product differentiation, the lower the fixed cost associated with investment, the lower the difference in marginal production costs between investing and exporting without tariff, and the higher the marginal production cost of the incumbent firm. Addressing the question of the host-country trade policy, we show that the socially-optimal tariff is determined by the degree of product differentiation. In the case where the socially-optimal tariff induces tariff-jumping, the host country may prefer FDI to free trade. The influences of tariff-jumping and strategic interaction in explaining the internationalization of Korean firms are then investigated in an empirical analysis. Trade restraints and protectionist threats in industrialized countries are expected to have played a significant role in inducing Korean firms to invest instead of exporting. The importance of trade restraints as a major locational factor favouring overseas production has become a standard assumption of the international direct investment literature. Among a wide variety of locational advantages (market size, labour costs and quality, local incentives and so on), the role of trade barrier circumvention has been emphasized as a ‘pull’ factor when it comes to explaining the remarkable growth of Japanese investment in industrial countries since the mid-1980s (Heitger and Stehn, 1990; Hennart and Park, 1994; Azrak and Wynne, 1995; Belderbos, 1997a and 1997b). Similarly to the Japanese experience, Korean firms are likely to have been sensitive to industrial countries’ trade restrictions as their exports were increasingly constrained by restrictive measures. More recently, vigorous efforts by Korean firms to expand their exports to overcome the financial crisis of 1997 faced more obstacles, as industrial countries feared being swamped by price-competitive products because of currency depreciation. Also, there has been some concern about a general resurgence of protectionism, with anti-dumping cases mushrooming around the world. Between 1995 and 2000, the European Union (EU) and the USA launched 218 and 181 anti-dumping cases, respectively. Korea was the target of 120 cases, ranking second only to China, with 207 cases.

Bénédicte Coestier and Serge Perrin 121

The influence of strategic interaction in explaining the FDI behaviour of multinationals goes back to the seminal work of Knickerbocker (1973), who showed that national oligopolies may exhibit imitative behaviour in their foreign investment strategies. The intuition is that, if a leading firm in an oligopoly decides to locate in a foreign market, certain rivals may choose to match the move for fear that the leader might gain a decisive advantage – such as favourable access to a new market – which in turn would affect the domestic oligopoly. In a more formal setting, Motta (1994) also suggested in a vertical product-differentiation model that a ‘follow-the-leader’ pattern, or bunching investment, could occur among national firms. A study by Yu and Ito (1988) confirmed the oligopolistic reaction hypothesis among US multinationals in the tyre industry. In the Japanese case, Kogut and Chang (1991) tested a positive and significant relationship between the extent of Japanese investment in a US industry and producer concentration in Japan, suggesting that high home concentration and rivalry between Japanese firms encourage direct investment in the United States. In the Korean case, we also suggest that oligopolistic rivalry with higher-quality Japanese firms might have accelerated the internationalization process of Korean multinationals, and that leader – follower behaviour occurs at both national and international levels. The empirical analysis, based on unpublished firm-level data, is conducted using a logit model to explain the FDI decision by Korean firms in the EU and the NAFTA. The focus is on the electronics industry, considered as a strategic sector in Korean industrial policy, and where Korean firms have gained a significant competitive edge in world markets. This industry’s share of Korean manufacturing output and exports was 19.6 and 31.8 per cent, respectively, in 1996, a substantial rise from 1985 when it stood at 9.2 and 15.8 per cent, respectively. Thus it should come as no surprise that the electronics industry accounts for a large share of Korean FDI. According to our estimates based on firmlevel data provided by the Korea Federation of Banks, the electronics industry represented 64.2 per cent of total Korean manufacturing FDI (outstanding amount basis) in the European Union as of the end of 1997. The corresponding figure for North America was somewhat less lopsided but the electronics industry was nonetheless the largest sector, with a 45.3 per cent share of total Korean manufacturing investment in this region. The chapter is organized as follows. The theoretical part is presented in Section 2; Section 3 is devoted to the empirical analysis, and we conclude in Section 4.

122 The Internationalization of Korean Firms

2

The model

We consider a simple situation where there are two firms, an incumbent domestic firm offering a high-quality product and an emerging multinational firm offering a product of lower quality. The level of quality is supposed exogenous. The multinational firm can choose between exporting and investing on the local market. We study the entry mode of the emerging multinational firm under price competition with perfect information on product quality. 2.1

Assumptions and notations

There are two firms: an emerging multinational firm (labelled Firm 1) offering a low-quality good s1 , and a local firm (labelled Firm 2) producing a high-quality good, s2 > s1 (si being a real positive number). s ≡ s2 − s1 denotes the quality differential. For the sake of simplicity we assume that the domestic firm (Firm 2) does not export or seek to invest in the home market of the emerging multinational, and Firm 1 sells its entire production on the foreign market (and thus nothing on its own home market). Both firms set their prices simultaneously. Firm 1 exports at a constant marginal cost c1E and Firm 2 produces on the national market at a constant marginal cost c2 . Firm 1 can either export or invest in Firm 2’s market. As both firms have already entered their respective national markets, the sunk cost specific to the Firm is equal to zero. This sunk cost is interpreted in the literature, notably by Hirsch (1976), as R&D costs necessary to elaborate the product. If Firm 1 chooses to export, it incurs, in addition to the variable domestic production cost, a unit cost of transport, τ . Also, a supplementary cost corresponding to a tariff, t, may have to be taken into account. If the potential multinational chooses to invest, it has to incur a fixed cost related to the establishment of a new plant in a foreign environment. After incurring this cost, the multinational firm can produce at a constant marginal cost c1I . We assume that the marginal cost associated with investment is greater than the marginal cost associated with export (transport cost included); c1I > c1E + τ . Hence the foreign firm will have no incentives to invest in the local firm’s market unless the government of the national country sets up a tariff. This hypothesis is consistent with the observed facts. Indeed, Korean firms have penetrated industrialized countries’ markets through exports before starting to invest. As Korean firms’ investments increased, in the middle of the 1980s, production costs were lower in Korea than in the USA or continental

Bénédicte Coestier and Serge Perrin 123

Europe, so Korean products were still competitive as exports. Besides, the remarkable increase of Korean FDI in industrialized countries concurs with frequent trade frictions (Perrin, 2001). Moreover, we assume that t > c1I − (c1E + τ ); we consider tariffs to be greater than the difference in the marginal costs of the multinational firm in the host country and in the home country. This is a necessary condition for the multinational firm to have incentives to invest rather than export when the government of the potential host country imposes a tariff. Hence, from a particular tariff level, namely t = c1I − (c1E + τ ), it will be less costly, in terms of variable production costs, to set up local production to substitute for exports. Consumers’ preferences are described by the following indirect utility function, U = θ s − p, if the consumer identified by the taste parameter θ buys one unit of quality s at price p, and 0 otherwise. U is the surplus derived from the consumption, hence utility is separable in quality and price. For all consumers, at equal price, high quality is preferred to low quality. However, all consumers are not ready to purchase high quality for the same price: θ is uniformly distributed in the interval [θ , θ]. We set θ = 1 + θ. Hence, the density is 1. We consider the usual assumptions of a duopoly model with vertical product differentiation in which all potential consumers make a purchase in equilibrium (Tirole, 1988): Hypothesis 1: (θ − 2θ)s ≥ c1E + τ + t − c2 ; that is, consumer heterogeneity is sufficiently important given the costs structure. This assumption guarantees that, in equilibrium, whatever the entry strategy adopted by the multinational firm (exports with or without tariffs or investment), it has a strictly positive demand. This assumption also implies that, in any type of equilibrium, the high-quality price is greater than the lowquality price. Finally, this assumption can be considered as defining an upper bound for the tariff; t = (θ − 2θ )s + c2 − (c1E + τ ). Hypothesis 2: At equilibrium prices, the market is covered; that is, all potential consumers buy one unit, whatever the entry strategy adopted by the multinational firm. Formally, the surplus derived from consumption by the consumer indexed with the lowest taste parameter value, evaluated at the low-quality equilibrium price (in the case of exports with tariff) is positive: θ s1 ≥ c1E + τ + t + 13 (c2 − (c1E + τ + t) + (θ − 2θ )s). This inequality can be considered as defining a condition on s1 and s2 . The preference index of the consumer who is indifferent between the quality s2 at price p2 and the quality s1 at price p1 is θ˜ = (p2 − p1 )/s. Thus consumers indexed with a taste parameter θ ≥ θ˜ buy a high-quality good,

124 The Internationalization of Korean Firms

and consumers with a taste parameter θ < θ˜ buy from the low-quality firm. Given the uniform distribution of consumers, market demands for the two qualities are: p2 − p1 −θ s p2 − p1 D2 (p1 , p2 ) = θ − s D1 (p1 , p2 ) =

(1) (2)

We now consider the entry decision of the multinational firm. Under what circumstances will the multinational firm stay out of the market, and let the domestic firm behave as a monopoly ? If (s2 /p2 ) ≥ (s1 /p1 ), the quality-over-price ratio is greater for the high-quality good, and every consumer considering buying one unit of the product will prefer the high-quality one. There will be no demand for the low-quality good and the high-quality firm will behave as a monopoly. Hence the domestic firm will charge a monopoly price equal to: pm = 21 (θ s2 + c2 ) = c2 + 21 (θ s2 − c2 )

(3)

and realize a profit equal to: m 2 =

s2 4


θ−

c2 s2

2 (4)

if p1 ≥ (s1 /2s2 )(θ s2 + c2 ). In autarky – that is, when the multinational firm does not enter the market and the national firm behaves as a monopoly – the functioning of the domestic market is inefficient for two reasons: first, the national firm charges a monopoly price that is greater than the marginal cost of production; and second, consumers with a reserve price smaller than the monopoly price do not buy. The welfare of the domestic country defined as the sum of consumer’s surplus and domestic enterprise profit is equal to: Wm =

θ θ˜

(θ s2 − c2 )f (θ)dθ

(5)

where θ˜ = p2m /s2 . This case corresponds to a situation where the lowquality good is ‘dominated’ (Tirole, 1988). We now consider situations where the foreign firm enters the domestic market via either exports or direct investment.

Bénédicte Coestier and Serge Perrin 125

2.2

Duopoly equilibrium: the exports case

We consider a situation where Firm 1 exports in Firm 2’s market. We assume that both firms compete in price and that the market is covered. We derive the Nash equilibrium under free trade and when the domestic market government imposes a tariff. 2.2.1

The free-trade equilibrium

With free trade, there is no tariff imposed by the domestic market government. Profit functions of Firms 1 and 2 are: 
 p2 − p 1 −θ 1 = p1 − − τ s 
p2 − p1 2 = (p2 − c2 ) θ¯ − s 

c1E

(6) (7)

Both firms compete to attract consumers. In the Nash equilibrium, each firm maximizes its profit with respect to price. Using the reaction functions, we find that, when both firms are active, prices are: p1E = c1E + τ + p2E = c2 +

  c2 − c1E + τ (θ − 2θ )s + 3 3

(2θ − θ)s c1E + τ − c2 + > p1E 3 3

(8) (9)

These yield the following demands:   c2 − c1E + τ θ − 2θ + = 3 3s   c2 − c1E + τ 2θ − θ D2E = − 3 3s

D1E

(10) (11)

and the corresponding profits:   s C 2 θ − 2θ + 9 s   s C 2 2θ − θ − E2 = > E1 9 s E1 =

(12) (13)

where C = c2 − (c1E + τ ). Equilibrium prices depend on marginal costs, on consumers’ taste for quality, and on the quality differential. The higher the latter – that is,

126 The Internationalization of Korean Firms

the higher the degree of product differentiation – the higher the market prices will be. In equilibrium, the domestic/high-quality firm charges a higher price than the emerging multinational and makes a higher profit. 2.2.2

Setting a tariff

Consider now a situation where the emerging multinational chooses to export on the foreign market and that the government imposes a tariff, t. The profit function of Firm 1 is:
   p 2 − p1 −θ 1 (p1 , p2 ) = p1 − c1E − τ − t s

(14)

Maximizing profits with respect to price for both firms, and using the reaction functions, we obtain the following prices: p1E,t

=

c1E

  c2 − c1E + τ + t (θ − 2θ )s + +τ +t + 3 3

p2E,t = c2 +

(2θ − θ)s c1E + τ + t − c2 + > p1E,t 3 3

(15) (16)

Because imposing a tariff on exports increases the marginal cost of the multinational firm, both firms charge higher prices, but the price increases are greater for the low-quality/emerging multinational firm (an increase of 2t3 ) than for the high-quality/domestic firm (an increase of 3t ). We obtain the following demand functions:   c2 − c1E + τ + t θ − 2θ + 3 3s  E  c − c 2θ − θ 2 1 +τ +t D2t = − 3 3s

D1E,t =

(17) (18)

and profits:   s Ct 2 θ − 2θ + 9 s   s Ct 2 2θ − θ − = > E,t 1 9 s

E,t 1 =

(19)

E,t 2

(20)

where Ct = c2 − (c1E + τ + t). Imposing a tariff leads to a decrease in demand for the low-quality good, and simultaneously an equivalent increase in demand for the

Bénédicte Coestier and Serge Perrin 127

high-quality good produced by the domestic firm, so that total demand remains unchanged. Consequently, Firm 1’s profit is reduced and Firm 2’s profit increases relative to the situation of exports without a trade barrier. We now move to the decision regarding investing in the foreign market, and in particular we analyse the conditions under which the multinational firm will proceed to tariff-jumping. 2.3

Duopoly equilibrium under investment

We remind ourselves that, when the emerging multinational chooses to invest, it incurs a unit cost of production c1I and a fixed cost G. Both firms maximize the following profit functions with respect to prices: 
  p2 − p1 1 (p1 , p2 ) = p1 − c1I −θ −G s 
p2 − p1 2 (p1 , p2 ) = (p2 − c2 ) θ − s

(21) (22)

yielding the Nash equilibrium: p1I = c1I +

(θ − 2θ )s c2 − c1I + 3 3

(23)

p2I = c2 +

(2θ − θ)s c1I − c2 + > p1I 3 3

(24)

and the following demands: D1I =

c2 − c1I θ − 2θ + 3 3s

(25)

D2I =

c2 − c1I 2θ − θ − 3 3s

(26)

and the corresponding profits
 s CI 2 θ − 2θ + −G 9 s
 s CI 2 2θ − θ − I2 = > I1 9 s I1 =

(27) (28)

where CI = c2 − c1I . As previously, equilibrium prices depend on marginal costs and on consumers’ heterogeneity. The domestic/high-quality firm charges a

128 The Internationalization of Korean Firms

higher price than the multinational/low-quality firm and makes a higher profit. 2.4

Exports versus FDI

Choosing between exporting and investing for the emerging multinational firm results in a trade-off between a reduced marginal cost and a additional fixed cost G. A necessary and sufficient condition for the emerging multinational to invest is that the profit realized under FDI is at least greater than the profit obtained under exports, I1 > E,t 1 . This holds if:

 

       c1E + τ + t − c1I c2 − c1I + c2 − c1E + τ + t 9G (2θ − 4θ) + > s s s (29) This condition, satisfied as an equality, defines an interval for tariff values, [t1 , t2 ], such that for all t ∈ [t1 , t2 ], the emerging multinational invests. The length of this interval depends, among other things, on the quality differential, the unit production costs, and the fixed cost of direct investment. Hence the probability that the emerging multinational will invest is higher, the lower the quality differential, the lower the difference in unit production costs between investing and exporting without tariff, the higher the marginal cost of the domestic firm, and the lower the fixed cost associated with investment. Observe that, in a duopoly situation, not all values of this interval are acceptable. Taking ¯ denoting into account the limit values for t derived previously, t and t, ˜t ≡ Min{t1, t2 }, and considering that t < t˜ < t, one can establish the following proposition: Proposition 1 In an oligopolistic situation, there exists a tariff threshold ˜ such that, for all t ∈ [t, t] ˜ the emerging multinational exports; and value, t, ˜ t] the emerging multinational invests. for all t ∈ [t, To penetrate the foreign market, the emerging multinational considers ˜ as well as the tariff imposed by the governthe tariff threshold value, t, ment. And the firm proceeds to tariff-jump if the government sets a tariff ˜ This threshold value is lower, the at least equal to this threshold value, t. lower the product differentiation, the lower the fixed cost G, the lower the difference in marginal costs between investing and exporting without tariff, and the higher the marginal cost of the incumbent firm.

Bénédicte Coestier and Serge Perrin 129

2.5

Welfare analysis

We now focus on the host country. We suppose that it uses the tariff to maximize domestic welfare defined as the sum of consumers’ surplus, domestic firms’ profits, and tariff revenue. Indeed, when the emerging multinational exports, setting a tariff allows an increase in the high-quality/domestic firm’s profit and gives a tariff revenue to the government of the importing country. At the same time, prices of both qualities will increase and consumer surplus diminish, but the price differential will be lower, the higher the tariff. 2.5.1

The socially-optimal tariff

When the multinational firm exports, consumers’ surplus at the optimum is given by: CSt =

θ˜ θ

(θ s1 − p1 )f (θ)dθ +

θ˜

θ

(θ s2 − p2 )f (θ)dθ

(30)

that is, with a uniform distribution for which the density is 1:     s1 s2 (θ˜ + θ ) − p1 + (θ − θ˜ ) (θ + θ˜ ) − p2 CSt = (θ˜ − θ) 2 2

(31)

with θ˜ = (p2E,t − p1E,t )/s, p1 = p1E,t , p2 = p2E,t . Tariff revenue is equal to the product of the tariff by the demand for the low-quality good: TR = t

θ − 2θ Ct + 3 3s

(32)

and the profit of the high-quality/domestic firm is:

E,t 2

  s Ct 2 2θ − θ − = 9 s

(33)

Defining the host country’s welfare as: Wt = CSt + TR + E,t 2

(34)

130 The Internationalization of Korean Firms

and maximizing with respect to t gives the following first-order condition (the second-order condition being satisfied; d 2 Wt /dt = −7/9s < 0):

 1 1 Ct − τ Ct − τ − t dWt =− 4θ − 5θ + + θ − 2θ + dt 9 s 3 s   Ct − τ 2 2θ − θ − =0 + 9 s

(35)

Such that the socially-optimal tariff, t ∗ , is equal to: t ∗ = s

(36)

We thus have the following proposition: Proposition 2 In an oligopolistic setting with vertical product differentiation and price competition, the socially-optimal tariff from the host country’s point of view – that is, the level of trade barrier that maximizes domestic welfare – is determined by the degree of product differentiation. The more the products are differentiated, the higher the sociallyoptimal tariff. It is interesting to note that the socially-optimal tariff is independent of cost considerations. In a Cournot setting with homogeneous product, Belderbos (1997a, p. 175) establishes that the socially-optimal tariff depends on the difference between the demand curve intercept and the marginal cost of the multinational. Put another way, the optimal tariff is higher the greater the potential profits of the multinational. Observe that, in our setting, the profits of both firms depend positively on the quality differential such that the optimal tariff will be greater the higher the potential profits of both firms. 2.5.2

The host country’s trade policy

We now assume that the socially-optimal tariff, t ∗ , is greater than the ˜ defined previously. With this tariff value, the tariff threshold value, t, emerging multinational prefers to invest and we obtain the duopoly equilibrium under investment. From the host country’s point of view, which situation maximizes domestic welfare? Free trade or FDI? In the duopoly equilibrium under investment, consumers’ surplus is reduced because of the prices increase, but the high-quality/domestic firm’s profit increases. With respect to the free-trade situation, one has

Bénédicte Coestier and Serge Perrin 131

a net positive effect on domestic welfare if, and only if: W = (CSI − CSt=0 ) + I2 − E2 > 0

(37)

where     s1 s2 ˜ ˜ ˜ ˜ (θ + θ) − p1 + (θ − θ ) (θ + θ ) − p2 CSI = (θ − θ) 2 2

(38)

with θ˜ =

p2I − p1I , p1 = p1I , p2 = p2I , s

and: CSt=0 = (θ˜ − θ)



   s1 s (θ˜ + θ) − p1 + (θ − θ˜ ) 2 (θ + θ˜ ) − p2 2 2

(39)

with θ˜ =

p2E − p1E , p1 = p1E , p2 = p2E s

After simplifying, one obtains the following expression: W =

   I 1  I  E c1 − c1 + τ c1 + c1E + τ + 2θ s − 2c2 6s

(40)

The two first terms of this expression being positive, the sign of W depends on the sign of the third term, so that it is positive if: θs > c2 −

c1E + τ + c1I 2

(41)

One has the following proposition: Proposition 3 From the host country’s point of view, FDI is preferred to free trade if, and only if, the difference between the marginal cost of the domestic/high-quality firm and the average of the marginal costs of the emerging multinational firm (marginal costs under free trade and under

132 The Internationalization of Korean Firms

investment) is lower than the monetary value of the quality differential for the lowest demand-for-quality consumer. The welfare increase under investment will be higher the higher the quality differential, and the lower the marginal costs difference. Observe that FDI always increases domestic welfare if the domestic firm has a cost advantage over the multinational firm (that is, c2 < (c1E + τ + c1I )/2). This point may be related to Belderbos (1997a, p. 176), in a homogeneous products environment. He establishes that domestic welfare increases as long as the domestic firm has a competitive advantage in terms of marginal production cost vis-à-vis the multinational. But if the multinational does have a competitive advantage that can be transferred via direct investment, then FDI intended to avoid trade barriers reduces domestic welfare. Finally, is the setting of the socially-optimal tariff by the host country the optimal trade policy? No; if the socially-optimal tariff is circumvented, the second-best trade policy is to set a tariff slightly lower ˜ decreases in profits and consumers’ surthan the tariff threshold value, t, plus generated by this trade policy being more than offset by the tariff revenue.

3

Empirical analysis

The influences of tariff-jumping and strategic interaction in explaining the internationalization of Korean electronics firms are now investigated empirically using firm-level data. 3.1

Methodology

The purpose of the statistical analysis is to explain the decision of Korean firm i producing product j to set up manufacturing plants in the NAFTA or the EU. The dependent variable (FDI) is a dummy variable set equal to 1 if the firm had established a manufacturing subsidiary in the NAFTA or the EU by the end of 1997, and 0 otherwise. In thirty-three cases (17.7 per cent of the total), there was at least one manufacturing subsidiary established in one of the two regions. A logit model is used in which the probability of a Korean firm investing in Europe or North America is explained by the independent variables. Our sample consists of twenty-five electronics products. For each product, Korean manufacturers were identified in the Electronic Industries Association of Korea directory (EIAK, 1997–8), and only firms listed at the Korea Stock Exchange were retained in order to have

Bénédicte Coestier and Serge Perrin 133

some information on firm characteristics. We obtained ninety-three firm–product combinations and a total of 186 observations for the two regions.

3.1.1

Explanatory variables

Trade policy measures. Previous studies on the internationalization of the Korean electronics industry have stressed the importance of trade restrictions in industrial countries as a major cause of the ‘defensive’ nature of FDI in these markets (Jun, 1987; Han, 1992; McDermott, 1992; Bloom, 1994). This was confirmed in a 1993 survey conducted by the Korea Association of Electronics Industry Promotion on the motives of Korean FDI, which showed that ‘evading import restrictions’ was the second most important factor (23 per cent), following ‘accession to new markets’ (37 per cent). Other motives were ‘securing human resources’ (16 per cent), ‘utilizing lower wages’ (14 per cent), ‘sourcing parts and components’ (5 per cent) and ‘technology sourcing’ (5 per cent). A specific survey on the European Community (EC) market by Han (1992) revealed that the main motivations of Korean electronics FDI were to secure market access from protectionist measures (45.5 per cent) and to explore market opportunities from EC integration and/or East European markets (33.7 per cent). More recently, Shin (1999) conducted a detailed survey on Korean consumer electronics firms’ activities in Europe. Among the top motives for advancing into the EU, ‘to carry out part of a globalization strategy’ came first, followed closely by tariff-jumping. Other factors mentioned, such as ‘to protect the existing market in the EU’, were tied directly to concerns relating to trade barriers’ circumvention. Therefore, it is safe to conclude that protectionism, real or anticipated, did influence these early Korean investments in the USA and Europe. Few empirical analyses have attempted to test the role of industrial countries’ trade policies on Korean FDI. Jeon (1992) included non-tariff barriers in his cross-industry analysis of the determinants of Korean FDI, but the results were marginally significant. An analysis focusing on the electronics industry may yield better results as this industry was the primary target for import restrictions, notably in Europe. Van Hoesel (1999) has conducted an interesting econometric analysis of Korean electronics FDI at the firm level, which shows that anti-dumping policy indeed encouraged Korean firms to invest in industrialized countries. However, he did not attempt to measure the specific influences of various import restrictions.

134 The Internationalization of Korean Firms

The prospective influences of several trade policy measures are considered here. Anti-dumping actions are measured by the dummy AD, which equals 1 if Korean exports of product j manufactured by firm i have been facing an anti-dumping investigation in the NAFTA or the EU. The variable TARIFF measures pre-Uruguay-Round US and EU tariffs for the twenty-five products. Non-tariff barriers are accounted for by the variables VER and QUOTA. VER takes on the value 1 in the case of voluntary export restraints on colour televisions in the NAFTA and on video-cassette recorders, microwave ovens, radio-cassette players and colour TV tubes in the EU. The effect of national quotas in France and Britain on Korean colour televisions is controlled for by the dummy QUOTA. The information on industrialized countries’ import restraints against Korean products is in KITA (1997) and tariff levels are taken from Belderbos (1997b). Oligopolistic rivalry. The role of oligopolistic rivalry is examined tentatively at both national and international levels. Regarding domestic rivalry, Han (1992) hinted that the oligopolistic structure of the Korean electronics industry might have been one of the most important factors fuelling the investment of the top three chaebol companies in Europe. This was also suggested by Jun (1988), who included the North American market when examining the early stages of internationalization for Korean consumer electronics firms. He analysed the time lags between the two leading Korean rivals’ moves in their internationalization process using different penetration modes (export, local sales and local production), and found generally short time lags – from one to three years – between two moves. However, since Korean firms were still in the early stages of international investment in the mid-1980s, the study is not comprehensive as far as the local production aspect is concerned. The test performed by Jeon (1992) attempted to demonstrate that domestic industry concentration had a significant influence on Korean firms’ propensity to invest in developed countries, but his findings were not conclusive. This is probably because of the proxy used for estimating the degree of concentration, at the industry level, whereas it would have been more relevant to use a firm- or product-level variable in this case. As to international rivalry, the only empirical evidence available, to the best of our knowledge, of strategic interaction between Korean and Japanese firms in overseas markets can be found in Yoo (2000). His study suggests that Korean MNEs sought to follow their main Japanese competitors in ASEAN markets in order to catch up with them. Also, Japanese FDI

Bénédicte Coestier and Serge Perrin 135

in one ASEAN location may have signalled to Korean firms that it was worthwhile to invest in this market. The influence of domestic market power on the propensity to invest abroad is captured by the variable DOMINANT, which equals 1 if the firm was designated as a market-dominating firm by the Fair Trade Commission in 1995 (FTC, 1996), meaning that the top three firms’ concentration ratio is over 70 per cent at the product level. Rivalry with higher-quality Japanese firms is measured by the variable NRIVALJAP, which gives the number of Japanese firms manufacturing identical products j in the region as of 1992. The more Japanese rivals are present in one region, the more a Korean latecomer firm should have a strong incentive to invest in this region if it aims to catch up with the leaders. We hypothesize here that oligopolistic rivalry is not limited to a host country, but rather to the whole region, as Asian multinationals entering these markets usually have regional ambitions. Firm and product characteristics. Firm-specific variables are included in the analysis to control for firm characteristics. Marketing and technology assets are two factors that have also traditionally been considered as necessary ownership advantages for competing in foreign markets. Despite a recent drive aimed at increasing R&D in order to stimulate the manufacturing of higher-quality products, R&D spending by Korean firms has remained typically low by international standards and can be considered as a structural weakness. Since there were a number of missing observations for R&D expenditure, this variable is left out of the empirical model. The marketing advantage is not posited to be a relevant asset. However, a MARKETING variable – measured by the advertising expenses to sales ratio – is included in the analysis for reference. Firms featuring greater capital intensity (CAPITAL, property, plant and equipment per capita) should be better able to compete in foreign developed markets. These firm-level variables are taken from Korea Investors Service Inc. (1995), and a three-year average is considered (1992–4) to eliminate yearly disparities. In order to control for product characteristics, and notably for Korean products, which are more competitive in world markets, we introduce the dummy COMPJAP that takes the value 1 when product j has a significant world market share and is in competition with Japanese products. This concerns mainly colour TVs (15 per cent of the world market share), VCRs (27 per cent) and microwave ovens (33 per cent), according to KIET (1997). Also, the DOMINANT variable can also be interpreted as a proxy for product competitiveness at the domestic level: if a firm is

136

Table 6.1 Correlation matrix MARKETING MARKETING CAPINT DOMINANT COMPJAP NRIVALJAP TARIFF AD VER QUOTA

1.00 0.32 0.28 0.10 −0.04 −0.09 0.05 0.04 0.00

Source: Authors’ calculations.

CAPINT

1.00 0.35 −0.05 −0.16 −0.22 −0.03 0.00 −0.10

DOMINANT

1.00 0.33 −0.13 −0.11 0.16 0.23 0.06

COMPJAP

NRIVALJAP

1.00 0.40 0.26 0.60 0.57 0.47

1.00 0.44 0.53 0.29 0.19

TARIFF

AD

VER

QUOTA

1.00 0.50 −0.02 0.44

1.00 0.41 0.35

1.00 −0.06

1.00

Bénédicte Coestier and Serge Perrin 137

market-dominating in a product market, it suggests that the firm has been able to develop firm-specific advantages to maximize its market share. Thus DOMINANT gives a joint indication of the influences of concentration and domestic rivalry, and of competitive advantage in the home country. The correlation matrix of the independent variables shows that some trade policy variables are highly correlated with COMPJAP (see Table 6.1). This is not surprising, as the more competitive Korean products are more likely to be affected by industrialized countries’ import restrictions. Also, AD and TARIFF, and AD and NRIVALJAP, present strong coefficients of correlation and are used separately. 3.2

Estimation results

The results of the logistic regression are presented in Table 6.2. A positive sign for the coefficient of the independent variable indicates an increased probability that Korean electronics firm i manufacturing product j invested in the NAFTA or the EU as of the end of 1997. The model has a high overall explanatory power, and correctly classifies between 87.6 and 90.3 per cent of the observations. The tariff-jumping FDI motivation is confirmed by the estimation results. The tariff, voluntary export restraints and anti-dumping duties variables are of the expected sign and highly significant. This indicates that industrialized countries’ import restrictions did have a strong ‘pull’ effect on Korean FDI. Accordingly, trade policies in the EU and the USA have in all likelihood accelerated the internationalization process of Korean electronics firms in these markets. However, national quotas do not appear to play a significant role. The positive and strong influence of DOMINANT also suggests that Korean market-dominating firms are most likely to invest in the EU and North America. This variable also captures the traditional firm-size advantage – that is, the larger a Korean firm is, the more likely it is to manufacture in developed markets. One alternative interpretation for DOMINANT is the prospective fear of action by the Korean antitrust body, the Fair Trade Commission, which in this case would constitute a strong incentive to expand overseas in order to evade domestic regulations. The Korean case suggests that products where market concentration is very high (CR3 > 70) are more likely to feature a leader–follower pattern. This contrasts with the predictions of the oligopolistic reaction hypothesis. Knickerbocker (1973) found that the leader–follower pattern should occur in moderately-concentrated industries, but not in highlyconcentrated or competitive ones. In highly-concentrated industries,

138 The Internationalization of Korean Firms Table 6.2 Results of logistic regression Variable

Coefficients (t-statistic) 1

Constant MARKETING CAPINT DOMINANT NRIVALJAP COMPJAP AD VER TARIFF QUOTA Log likelihood Chi-square Correctness

2

−2.85∗∗∗ 3.58∗∗∗ (−2.79) (−3.81) −0.21∗∗∗ −0.10 (−2.65) (−1.62) 0.39E−04∗∗ 0.57E−04∗∗∗ (2.02) (3.26) 2.24∗∗∗ − (3.89) − 0.15∗ 0.13∗∗ (1.85) (2.11) – – – – 1.25∗∗ – (2.09) – – 1.73∗∗∗ – (2.94) – – – – 0.15 1.69∗ (0.15) (1.92) −62.1∗∗∗ −72.3∗∗∗ 49.7 29.3 88.7 87.6

3 −3.36∗∗∗ (−2.98) −0.17∗∗ (−2.29) 0.39E−04∗∗ (2.05) 2.05∗∗∗ (3.59) – – – – – – 1.67∗∗∗ (2.63) 0.17∗∗∗ (2.61) 0.39 (0.36) −64.8∗∗∗ 44.3 90.3

4 −2.94∗∗∗ (−2.93) −0.21∗∗∗ (−2.82) 0.46E−04∗∗ (2.33) 1.98∗∗∗ (3.18) 0.18∗∗ (2.32) 1.41∗∗ (2.06) – – – – – – −0.13 (−0.12) −62.1∗∗∗ 49.6 89.8

Note: ∗ = significant at the 10% level; ∗∗ = significant at the 5% level; ∗∗∗ = significant at the 1% level. Source:

Authors’ calculations.

firms are extremely interdependent and the result may be mutually destructive if one firm promptly matches the other’s investment. Thus, in this type of industry, firms should have a stronger incentive to co-operate rather than to compete for market share. Even in tight oligopolistic industries, Korean firms seem to compete rather than co-operate. This significant difference from the theory may lie in the small size of the Korean market, which pushes even dominant firms to extend their market share overseas. Foreign markets can therefore be perceived as an extension of the domestic market, with domestic rivalry simply being exported to overseas markets. Apart from the size and domestic concentration aspects, the advertising intensity variable (MARKETING) is significant but negative. This leads us to conclude that conventional marketing advantages appear

Bénédicte Coestier and Serge Perrin 139

not to be relevant in the Korean case. Previous findings for Korea by Jeon (1992), Park (1998) and Van Hoesel (1999) showed that this variable could have either a positive or a negative influence, but was not significant. In the case of Japanese entries to the USA, Hennart and Park (1994) also found a negative but insignificant influence of the intensity of advertising. One possible interpretation is that foreignmarket-orientated Korean firms rely heavily on OEM and are thus less likely to spend on advertising. The coefficient of the other firm-level variable, CAPITAL, is positive and significant, as hypothesized. This may be an indication of Korean investors’ capacity to engage in mass production in industrialized countries in search of economies of scale. At the product level, the variable COMPJAP confirms that Korean firms are more likely to manufacture abroad the products with which they have gained a significant competitive edge in world markets – namely, certain consumer electronics products. However, this variable was not significant when used simultaneously with trade policy variables because of multicollinearity. The proxy for measuring international rivalry with Japanese firms (NRIVALJAP) is positive and significant, albeit only at the 10 per cent level. In Equation 2, the variable AD was dropped, which increased the level of significance for NRIVALJAP. This confirms the hypothesis that oligopolistic rivalry with Japanese multinationals in the NAFTA and the EU influences the investment decision of Korean electronics manufacturers – that is, Korean latecomers tend to follow the Japanese leaders. It suggests that Korean firms aim at competing directly with Japanese rivals in order to catch up in the global competition in the electronics industry. The importance of oligopolistic rivalry in explaining one firm’s FDI decision can be better understood in a Korean context, where the internationalization process may be perceived as a threat to other members of the oligopoly in terms of its impact on domestic competition. First, expansion in foreign markets, especially in the EU and in North America, should allow a firm to achieve economies of scale and to obtain cost advantages in the domestic market vis-à-vis domestic rivals. Second, on the advertising front, a Korean firm manufacturing in developed markets scores major publicity gains over its domestic competitors in terms of reputation. It sends to Korean consumers a signal that the quality of its products is likely to be superior, since they are manufactured and sold in advanced countries’ markets, and contribute to the building of a global brand name. Third, the direct presence of the leading chaebol companies on these markets may, in turn, have an impact on the group’s overall

140 The Internationalization of Korean Firms

image, as the reputation of a group in Korea is associated mainly with the performance of the flagship companies.

4

Conclusions

This chapter has examined some of the main forces influencing the entry of Korean emerging multinationals into industrialized countries. From a theoretical point of view, in a simple duopoly model with vertical product differentiation, we establish that tariff-jumping FDI is more likely the lower the product differentiation, the lower the fixed costs associated with investment, the lower the difference in marginal production costs between investing and exporting without tariff, and the higher the marginal production costs of the incumbent firm. With respect to host-country trade policy, we establish that the socially-optimal tariff is determined by the degree of product differentiation. If the sociallyoptimal tariff induces tariff-jumping, FDI may be welfare-improving with respect to free trade. As expected, trade barriers, notably anti-dumping duties, voluntary export restraints and tariffs, have played a significant role in inducing Korean firms to invest instead of export. The presence of Japanese leaders has prompted herding behaviour by Korean followers, indicating the importance of international rivalry in the investment decision. Also, rivalry between Korean firms had a ‘push’ effect on FDI. Regarding firm-specific variables, it appears that market-dominating and capitalintensive firms in Korea are more likely to invest, but that conventional marketing advantage does not play a part. We suggest that protectionism and oligopolistic competition have accelerated – or even forced – the internationalization of Korean firms. The influence of rivalry is an interesting point as it illustrates one factor of over-investment by the chaebol companies – one of the underlying causes of the financial crisis of 1997. It is reasonable to assume that the leading Korean groups may have transposed their domestic investment to world markets, characterized notably by the pursuit of aggressive market share expansion strategies. This suggests that Korean firms, characterized by centralized decision-making by the owner-managers and a lack of transparency in management, entered overseas markets to sustain their strategy of quantitative growth at the risk of insufficient evaluation of business profitability. Additional external pressure came from Japanese MNEs perceived by many Korean firms as their main competitors. But protectionist pressures have probably also driven Korean firms to

Bénédicte Coestier and Serge Perrin 141

invest more rapidly than they would have done in a more competitive environment, where these firms were motivated solely by ownership advantages. A central issue highlighted by the crisis of 1997 is whether decisions to invest under such pressure increase the risk of failure. As the crisis has prompted the need to reassess Korea’s past overall economic development, it appears more important than ever to evaluate the outcome of Korean firms’ internationalization strategy. Some preliminary survey results indicating an overall low profitability of Korean overseas operations do raise the question of the sustainability of these investments. Several company surveys on the performance of Korean affiliates abroad have been conducted but they have not provided a clear-cut answer. A study by Wang (1998) does suggest that the profitability of Korean overseas manufacturing affiliates in North America and in Europe is quite low compared with South East Asia. Even though detailed data by region and year of entry are not available, this suggests that it is more difficult for Korean firms to support and sustain the ‘cost of foreignness’ in advanced countries. As for the relative performance of Korean firms in Europe and North America, one cannot reach a conclusion as to whether they perform better in a particular region, because of yearly variations. Other surveys by Ha and Hong (1997, 1998) tend to yield lower profitability figures overall for overseas operations, even surprisingly in Asian markets, and the results for 1996 show a deterioration, particularly in the EU and in Europe in general (see Table 6.3). Since the Korean system of corporate governance is not known for fostering profitability, these results should be compared with the performance of parent companies. For example, in 1996, the ordinary income-to-sales ratio of Korean parent companies investing in the USA and EU were 4.4 per cent and 2.5 per cent, respectively – well above the performance of overseas affiliates (Ha and Hong, 1998). Another study, by Lee (2000), shows that Korean affiliates in overseas markets exhibit extreme instability because of capital depletion caused by continuous losses and high debt ratios in 1997 and 1998, and the heavy reliance of foreign affiliates on the parent company for fund-raising. This suggests a high risk that foreign operations will become a direct burden on the domestic economy. The structural profitability problem is now at the core of corporatesector reform in Korea. An optimistic view is that the financial crisis and the credit crunch constraint forced the chaebol companies to restructure and rid themselves of their unprofitable investments, which could

142 The Internationalization of Korean Firms Table 6.3 Profitability of Korean overseas affiliates by region, 1995–6 (relative to sales, as a percentage, and number of valid replies (in parentheses)) Region

1995 Ordinary income

Asia China ASEAN North America USA Latin America Europe EU Middle East Africa Oceania Total Source:

0.6(300) 1.7(174) 3.6(74) 0.0(62) −0.1(57) −0.3(25) 0.9(20) 1.5(13) 0.0(4) −3.4(1) 5.7(6) 0.5 (418)

1996 Net income 0.3(278) 0.9(156) 1.8(70) 0.4(55) 0.3(50) −1.8(25) −0.3(19) 0.2(13) 0.0(5) −0.4(1) 4.5(6) 2.7 (389)

Ordinary income −1.7(378) −0.2(193) −2.1(107) 0.9(77) 0.9(75) 0.2(26) −5.4(32) −9.5(19) 0.2(3) 0.3(3) 0.6(6) −1.1 (525)

Net income −3.2(366) 0.6(184) −5.0(104) 0.3(74) 0.3(72) 0.0(27) −6.4(33) −10.0(20) 0.1(3) 0.3(3) −7.1(3) −1.7 (509)

Ha and Hong (1997, 1998).

otherwise have been sustained for many years through the system of cross-guarantees within business groups (for a detailed analysis of corporate and financial sector reforms in Korea since 1997, see Graham, 2003). The few studies available so far that stress the low performance of Korean MNEs should be interpreted with caution. They must not be an incentive for the Korean government to limit Korean firms’ entry into overseas markets, as was the case until the early 1990s. FDI is now a essential part of Korean firms’ internationalization strategy, following the export promotion stage, and is crucial to Korea’s industrial upgrading. However, the financial crisis clearly showed that domestic creditor financial institutions did not consider and anticipate the effects of foreign operations in the credit management of Korean firms. A system institutionalizing the monitoring of potential risk in foreign investment guarantee should be developed promptly by Korean financial supervisory regulations. References Azrak, P. and K. Wynne (1995) ‘Protectionism and Japanese Direct Investment in the U.S.’, Journal of Policy Modeling, vol. 17(3), pp. 293–305. Belderbos, R. (1997a) Japanese Electronics Multinationals and Strategic Trade Policy (Oxford: Clarendon Press). Belderbos, R. (1997b) ‘Anti-dumping and Tariff-jumping: Japanese Firms’ FDI in the EU and the US’, Weltwirtschafltiches Archiv, vol. 133(3), pp. 419–54.

Bénédicte Coestier and Serge Perrin 143 Bloom, M. (1994) ‘Globalisation and the Korean Electronics Industry: A Chandlerian Perspective’, in H. Schütte (ed.), The Global Competitiveness of the Asian Firm (New York: St. Martin’s Press), pp. 139–52. EIAK (1997–8) Directory of Overseas Electronics Firms (in Korean) (Seoul: Electronic Industries Association of Korea). Fair Trade Commission (1996) Fair Trade Laws and Regulations (Seoul: The Korea Fair Competition Association). Graham, E. M. (2003) Reforming Korea’s Industrial Conglomerates (Washington, DC: Institute for International Economics). Ha, Byung-Ki and Seok-Il Hong (1997, 1998) Survey of Business Operations of Overseas Affiliates (in Korean) (Seoul: Korea Institute for Industrial Economics and Trade). Han, Sung-Taik (1992) European Integration: The Impact on Asian Newly Industrialising Economies (Paris: OECD Development Centre). Heitger, B. and J. Stehn (1990) ‘Japanese Direct Investment in the EC – Response to the Internal Market 1993?’, Journal of Common Market Studies, vol. 29(1), pp. 1–15. Hennart, Jean-François and Young-Ryeol Park (1994) ‘Location, Governance, and Strategic Determinants of Japanese Manufacturing Investment in the United States’, Strategic Management Journal, vol. 15(6), pp. 419–36. Hirsch, S. (1976) ‘An International Trade and Investment Theory of the Firm’, Oxford Economic Papers, vol. 28(2), pp. 258–69. Jeon, Yoong-Deok (1992) ‘The Determinants of Korean Foreign Direct Investment in Manufacturing Industries’, Weltwirtschafltiches Archiv, vol. 128(3), pp. 527–41. Jun, Yong-Wook (1987) ‘The Reverse Direct Investment: The Case of the Korean Consumer Electronics Industry’, International Economic Review, vol. 1(3), pp. 91–104. Jun, Yong-Wook (1988) ‘The Structural Analysis of the Global Consumer Electronics Industry and the Oligopolistic Behaviour of Korean Firms in their Internationalization’, Occasional Paper, no. 88-07, Korea Institute for Industrial Economics and Trade, Seoul. KIET (1997) A Study of Korean Industry (in Korean), Vol. 1 (Seoul: Korea Institute for Industrial Economics and Trade, Ministry of Commerce, Industry and Energy). KITA (1997) Overview of Import Restrictions of Major Industrialized Countries (in Korean), (Seoul: Korea International Trade Association). Knickerbocker, F. T. (1973) Oligopolistic Reaction and Multinational Enterprise (Boston, Mass.: Harvard University Press). Kogut, B. and Sea-Jin Chang (1991) ‘Technological Capabilities and Japanese Direct Investment in the US’, Review of Economics and Statistics, vol. 73(3), pp. 401–13. Korea Investors Service Inc. (1995) Annual Report of Korean Companies (in Korean), Seoul. Lee, Hong-Gue (1993) ‘Globalization of the Korean Electronics Industry’, Working Paper no. 9314, Korea Development Institute, Seoul. Lee, Seong-Bong (2000) ‘Korea’s Overseas Direct Investment: Evaluation of Performances and Future Challenges’, Working Paper no. 00–12, Korea Institute for International Economic Policy, Seoul.

144 The Internationalization of Korean Firms McDermott, M. C. (1992) ‘The Internationalization of the South Korean and Taiwanese Electronics Industries: The European Dimension’, in S. Young and J. Hamill (eds), Europe and the Multinationals (Aldershot: Edward Elgar), pp. 296–31. Motta, M. (1992) ‘Multinational Firms and the Tariff-Jumping Argument. A Game Theoretic Analysis with Some Unconventional Conclusions’, European Economic Review, vol. 36(8), pp. 1557–71. Motta, M. (1994) ‘International Trade and Investments in a Vertically Differentiated Industry’, International Journal of Industrial Organization, vol. 12(2), pp. 179–96. Park, Young-Ryeol (1998) Determinants of Korean Overseas Manufacturing Investment (Seoul: International Trade and Business Institute). Perrin, S. (2001) ‘Korean Direct Investment in North America and Europe: Patterns and Determinants’, in F. Sachwald (ed.), Going Multinational: The Korean Experience of Direct Investment (London/New York: Routledge). Shin, Sang-Hyup (1999) European Integration and Foreign Direct Investment in the EU: The Case of the Korean Consumer Electronics Industry (London/New York: Routledge). Smith, A. (1987) ‘Strategic Investment, Multinational Corporations and Trade Policy’, European Economic Review, vol. 31(1–2), pp. 89–96. Tirole, J. (1988) The Theory of Industrial Organization (Cambridge, Mass.: MIT Press). Van Hoesel, R. (1999) New Multinational Enterprises from Korea and Taiwan (London/New York: Routledge). Wang, Yun-Jong (1998) The Situation and Performance of Korean FDI (in Korean) (Seoul: Korea Institute for International Economic Policy). Yoo, Sung Min (2000) ‘Korean and Japanese Foreign Direct Investment in ASEAN’, Korea Focus, vol. 8(1). Yu, J. and K. Ito (1988) ‘Oligopolistic Reaction and FDI: The Case of the US Tire and Textile Industries’, Journal of International Business Studies, vol. 19(3), pp. 449–60.

7 Foreign Direct Investment, Trade and Regional Integration in Mercosur∗ Marta Castilho Universidade Federal Fluminense (UFF) and Institute for Applied Economic Research (IPEA), Rio de Janeiro, Brazil

and Soledad Zignago University of Paris I, France

1

Introduction

Latin-American countries experienced strong growth in both trade and foreign direct investment (FDI) during the 1990s. The major factors behind this include the recovery of economic growth in the region (mainly because of macroeconomic stabilization) and the multilateral and regional trade liberalization process. Among a number of different initiatives to achieve economic integration in Latin America, the effort to create a customs union among the eastern countries of the ‘southern cone’ of South America (that is, Argentina, Brazil, Paraguay and Uruguay, termed the Mercardo Común del Sur, or Mercosur) is the most dynamic. The major economic result has been the intensification of regional trade, which increased fivefold during the 1990s. In addition, trade flows between Mercosur and non-Mercosur countries have also risen substantially. The main external partners of Mercosur are the EU, the USA and other Latin-American countries. The expansion of trade was accompanied by strong growth of FDI inflows to the region. Annual average inflows of FDI rose from

∗

We would like to thank the Chaire Mercosur of the Ecole des Sciences Politiques de Paris for financial support, and comments from participants of its Working Group on EU–Mercosur Negotiations. We are also grateful for the research assistance of Ana Cláudia Loureiro and Camila Alves. 145

146 FDI, Trade and Integration in Mercosur

US$2 billion in 1990 to US$54 billion in 1999. The main investors have been the EU countries, most notably Spain, and the USA. One can ask if there is a link between the increase in trade and the growth of FDI in the context of trade liberalization and regionalization that occurred in here over the 1990s. In fact, it would be interesting to know, first, if multinational enterprises (MNEs) modified their strategies in the region as a result of Mercosur’s creation and, second, if FDI has had a positive impact on imports and exports. These issues are important not only from a microeconomic point of view but also from a macroeconomic one. The growth of FDI inflows was essential in financing the current balance in Mercosur in the 1990s. However, several issues concerning its long-term contribution to current and trade balance are raised, because the propensity of MNE operating in this region to import has been stronger than their propensity to export. For example, MNEs tend to buy inputs from the parent firm’s international network of suppliers, and thus their entry has been accompanied by a rise in imports. On the other hand, the FDI inflows, strongly focused on service sectors, did not generate many exports. From a regional point of view, if the MNE strategies were to exploit the regional market but not to integrate their Mercosur subsidiaries into the international production network, their implementation would not be likely to have a positive direct impact on exports. However, indirect effects of FDI on trade, such as efficiency improvement, transfers of technology or even the long-term impact of FDI on the financial balance of the relevant host nations (for example, because of transfers between subsidiaries, profit repatriation and so on) can be important. While we do not discuss all these effects in this chapter, we do concentrate on the following issues: the links between FDI and imports; the links between FDI and exports; the influence of Mercosur on MNE strategies, and FDI flows. We do not exhaust these subjects, and, indeed, we are obliged to make some simplified assumptions to derive the conclusions we reach. No consensus on the link between trade and FDI arises in either the theoretical or the empirical literature. From a theoretical point of view, both flows have traditionally been considered as substitutes: FDI has been seen as an alternative to exports in order to penetrate markets protected by strong trade barriers. Nevertheless, if most of the theoretical approaches emphasise substitution links between these two flows, empirical results often show the existence of complementarities. In an earlier work, we estimated the links between FDI, trade and regional integration

Marta Castilho and Soledad Zignago 147

(Castilho and Zignago, 2000). Our results did not show any evidence of a statistically significant link between the regional integration process and FDI inflows, even though they confirm a complementary relationship between FDI and imports. These results were surprising. Mercosur itself is rooted primarily in regional sectoral trade agreements, it is likely that the establishment of a common market influenced the strategies of MNEs in Brazil and Argentina. Porta and Kosacoff (1997), for example, underline the impact of Mercosur on the strategies of firms in the automobile industry. Thus we attempt here to find significant results pertaining to FDI and imports in certain sectors such as automobiles. We then turn in this chapter to analysis based on national disaggregated data.1 We use Argentinian data for the period 1991–9 from the CEP (Centro de Estudios para la Producción), and data from the Central Bank of Brazil, for 1990–4 and 1996–9. After providing an overview of trade and FDI patterns in the region, Section 2 of this chapter outlines our major empirical findings regarding the linkages between trade, investment and regional MNE strategies in Mercosur in the 1990s, and Section 3 presents a gravity model which demonstrates the influence of the creation of Mercosur on FDI flows.

2

Trade and FDI trends in Mercosur

Mercosur was created in 1991 but not implemented until 1995, with member countries liberalizing their mutual trade during the transition period (tariffs were scheduled to decline by 25 per cent annually). On 1 January 1995, the common external tariff (CET) covered around 85 per cent of the tariff lines and a schedule for convergence towards a complete CET was agreed. The tariff varies between 0 and 23 per cent (with an average of 11.3 per cent). Nevertheless, some sectors (such as capital goods, computers, telecommunications equipment, sugar, wheat and automobiles) figured in an exceptions list and were not covered by this CET.2 Some of these are still subject to special regimes, notably the automotive sector. The recent history of Mercosur – and in particular, the CET – has been troubled by macroeconomic difficulties in Argentina and Brazil. Thus, since 1998 the CET has been changed twice: first, when Brazil raised it by 3 points in response to the 1998 financial crisis (followed by the other partners); and second, in 2001, when Argentina launched a package of measures in order to improve its export competitiveness – and was again followed by Uruguay and Paraguay.

148 FDI, Trade and Integration in Mercosur

Moreover, these countries are involved in other regional negotiations; in particular, the FTAA (the Free Trade Area of the Americas), the FTA with the European Union, and an FTA with the Andean Community (CAN). On the one hand, the FTAA negotiations, which were initially conducted by individual national governments, have forced Mercosur countries to reconsider the pertinence of maintaining a small and imperfect Common Market in the context of a larger regional FTA. On the other hand, the European and Andean FTAs tend to reinforce Mercosur because the negotiations are between groups of countries rather than between individual governments. Also, intra-regional tariff reduction in the first part of the 1990s coincided with unilateral liberalization programmes negotiated separately by each of the member states. More fundamentally, the implementation of Mercosur occurred in a context of a large adjustment process and deep economic reforms that pointed out important changes in the development strategies of the Latin-American countries. In addition, those countries also undertook measures of financial liberalization, including major deregulation of controls on capital flows. Another important feature of this process has been domestic deregulation and the privatization of formerly state-owned enterprises. Privatization has affected many sectors and allowed foreign investors to participate in the purchase of national enterprises. This process was launched before in Argentina and has gone further there than in Brazil. These changes, together with the recovery of economic growth and the implementation of a common market, have contributed to a significant rise of FDI flows into the region. The changes in the economic environment since 2001 restored the growth path of FDI flows into Mercosur economies. Mercosur’s trade flows increased substantially during the 1990s, as can be seen in Figure 7.1. This growth was mainly in imports. Between 1990 and 1999, import growth was 190 per cent, while export growth was 60 per cent. As a consequence, from the mid-1990s, Mercosur countries faced a growing balance-of-trade deficit. The rise of imports seems to have reflected domestic economic expansion, trade opening and exchangerate evolution. These trends seem to have benefited mainly the OECD countries (as shown in Table 7.1; the percentage of OECD nations involved in Mercosur imports rose from 40 per cent in 1985 to 67 per cent in 1998), despite the growth of intra-regional trade being even higher. And while the role of OECD as the most important supplier to the Mercosur region was enhanced, its importance as a destination for Mercosur’s exports was weakened (showing the loss of dynamism of Mercosur’s exports).

149

120.0

US$ billions

100.0 80.0 60.0 40.0 20.0 0.0 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999

Imports Figure 7.1 Source:

Exports

Mercosur total trade, 1990–9

IADB.

Table 7.1 Mercosur’s main trade partners (percentages and annual percentage rates) Total flows

European Union NAFTA USA Mercosur Argentina Brazil Paraguay Uruguay Chile Andean Community Asian NICs Japan China Total Source:

IADB-INTAL.

Imports

1986

1990

1995

1999

1999

23.1 26.1 22.8 9.9 3.4 4.0 1.1 1.4 1.6 1.6 2.3 6.3 2.2

25.6 23.4 20.3 11 3.5 4.6 1.2 1.7 2.2 2.2 4.8 5.9 1.1

27.1 20.6 18.0 19.5 7.6 8.3 1.8 1.8 3.2 3.5 6.5 5.2 2.0

27.4 23.0 20.0 19.8 8.3 8.6 1.3 1.7 2.8 3.0 5.0 4.3 2.1

28.7 24.8 21.7 19.3 8.7 8.5 0.7 1.4 1.8 2.3 5.6 4.8 2.4

100

100

100

100

100

Change 1995–9

Exports 1999

Change 1995–9

8 .1 8 .9 11.0 9 .4 4. 2 21.6 −13.7 −5.6 −23.5 12 .8 −11.5 −5.6 55.6

25.9 21.1 18.2 20.4 7.8 8.7 1.9 2.0 3.8 3.8 4.3 3.7 1.7

5.1 32 .8 27.3 5.4 32.4 −1.4 −29.0 1. 4 3 .1 −18.3 −27 .7 −23.3 −24.5

6.0

100

5.6

150 FDI, Trade and Integration in Mercosur Table 7.2

FDI inflows to Mercosur (US$ millions)

Country

1990

1991

1992

1993

1994

1995

1996

1997

1998

1999

Argentina Brazil1 Paraguay Uruguay

1 836 324 77 –

2 439 89 86 –

3 218 1 924 118 –

2 059 801 75 102

2 477 2 035 137 155

3 818 3 475 155 157

4 922 9 644 246 137

5 099 17 879 270 113

4 504 26 346 423 155

22 358 31 235 306 229

Mercosur

2 237

2 614

5 260

3 037

4 804

7 605

14 949

23 361

31 428

54 127

Note: Source:

1

Source BCB (1996–9). IADB.

As indicated above, FDI (together with portfolio investment and short-term capital) played an important role in financing the currentbalance deficit of Mercosur economies in the 1990s. In fact, the 1990s were marked by the return of Latin-American countries to the financial markets after a long absence due to the debt crisis of the 1980s. The growth of FDI inflows was more vigorous than trade growth during the decade of the 1990s, as shown in Table 7.2. As was stressed by Chudnovsky (2001), Mercosur host countries accounted for 6 per cent of FDI flows to developing countries between 1997 and 1999. This is explained by a conjunction of factors, including macroeconomic stabilization and the economic recovery in the region, but also including the change of the regulatory environment, led by the privatization process, and trade and financial liberalization. The formation of Mercosur itself contributed to the growth of FDI, as we shall show below. The main investor nations have been the European countries, followed by the USA, as the individual tables for Argentina and Brazil show in the following sections. As seen in Table 7.2, these two countries absorbed 99 per cent of the total FDI inflows into Mercosur countries in 1999. In terms of sectoral distribution, the service sector has become the main recipient in recent years. At the start of the 1990s the service sector replaced the industrial sector in the leading position in Argentina and, later, this also happened in Brazil. This change is explained mainly by privatization, which brought foreign buyers into public utility sectors such as telecommunications. The geographical and sectoral characteristics of FDI are discussed further in Sections 2.1 and 2.2, with FDI in Argentina and Brazil being analysed separately. 2.1

FDI inflows into Argentina

The Argentinian economy was characterized in the first half of the 1990s by both macroeconomic stabilization and trade liberalization. As a consequence, the recovery of the economy led to a strong growth of imports

1 400

180 160 140 120 100 80 60 40 20 0

1 200 1 000 800 600

FDI

GDP and trade

Marta Castilho and Soledad Zignago 151

400 200 0 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 GDP

Figure 7.2 Source:

Imports

Exports

FDI

Argentina: GDP, FDI and trade (index 1999 = 100)

ECLAC, IADB.

(see Figure 7.2). In 1995, the peso crisis reduced the growth rate of the Argentinian economy, but the impact of this growth slowdown was tempered by the recovery of the Brazilian economy, the most important destination for Argentine exports. Unfortunately, this did not last long, as the growing instability in international financial markets (that is, first the Asian crisis in 1997 followed by the Russian crisis in 1998) affected these countries strongly by causing a slowdown in capital inflows to all developing countries, including those in Latin America. The change in the Brazilian exchange regime in 1999 and economic recession then reduced the price competitiveness of Argentinian exports, because the peso was fixed to the US dollar in 1991 while the Brazilian currency was allowed to depreciate. Consequently, a balance-of-payments problem emerged in Argentina that was to develop into a crisis. This was largely the result of a trade imbalance: exports were growing at a much lower rate than were imports. But, in contrast, FDI flows3 rose by almost three times between 1990 and 1997 (see Figure 7.2), followed by a boom in 1999 with US$22 billions of FDI inflows (explained by the privatization of the national oil company). As Table 7.3 illustrates, the FDI flows were concentrated in four sectors that accounted for 74 per cent of total industrial FDI. The sector that received the most FDI was oil and gas, reflecting the privatization of the national oil firm YPF. The sector receiving the second largest amount of FDI was electricity, where, again, privatization accounts for most of the FDI growth. Food and beverages absorbed 12 per cent of total FDI, followed by the automotive sector with 10 per cent. In contrast to the

152 FDI, Trade and Integration in Mercosur Table 7.3

Argentina: FDI by activity, 1990–9 (US$ millions and percentages) 1990–4 $mn %

Oil and gas Electrical energy Food and beverages Automotive industry Oil and gas products Chemicals Mineral extraction Pulp and paper Petrochemicals Other industries Total Source:

1995–9 $mn %

1990–9 $mn %

3 271 2 452 1 855 755 705 643 89 158 0 847

30.4 22.8 17.2 7.0 6.5 6.0 0.8 1.5 0.0 7.9

22 457 8 054 6279 6 016 3 245 2 988 2 664 1 758 1 448 3 737

38.3 13.7 10.7 10.3 5.5 5.1 4.5 3.0 2.5 6.4

25 728 10 506 8 134 6 771 3 950 3 631 2 753 1 916 1 448 4 584

37.1 15.1 11.7 9.8 5.7 5.2 4.0 2.8 2.1 6.6

10 777

100.0

58 645

100.0

69 421

100.0

CEP.

first two sectors mentioned above, FDI in the latter two industries was mainly greenfields investment. Table 7.4 details the national origin of foreign direct investment in Argentina during the period 1990–9. The USA was the most important investor nation, if countries are taken individually. But, when treated as a single entity, the European Union is the largest foreign investor, responsible for 47 per cent of total FDI. Among European countries, Spain is the most important one (25 per cent), followed by France (8 per cent). Some historically important investors in Argentina such as the UK, Italy and Germany declined in importance. Within Mercosur, Brazil is the only country with direct investments in other countries, but its participation is very weak – it accounts for only 1.7 per cent of total FDI in Mercosur. Indeed, Chile’s investments in Argentina are much more important, corresponding to 6.6 per cent of total investments. 2.2

FDI inflows into Brazil

For the Brazilian economy, the 1990s can be divided in two periods. The first years of the decade were marked by slow economic growth, because of political and economic instability (see Figure 7.3). But, beginning in 1994, a new stabilization programme was implemented, inaugurating a period marked by low inflation and economic growth. The recovery of the economy is marked by strong growth of imports and very strong inflows of foreign direct investment. Exports, however, presented a mediocre performance – a 75 per cent increase over ten years compared to a 170 per cent growth of imports.

153 Table 7.4 Argentina: main investor countries (US$ millions) 1990–9 Cumulated flows USA Canada European Union Spain Netherlands France Italy United Kingdom Germany Mercosur Brazil Uruguay Chile Mexico Switzerland Japan Total

48 731 3 138 59 472 31 873 2 412 10 293 6 249 5 336 2 565 2 363 2 147 216 8 411 1 494 1 296 407

38.1 2.5 46.5 24.9 1.9 8.0 4.9 4.2 2.0 1.8 1.7 0.2 6.6 1.2 1.0 0.3

128 005

100.0

CEP.

350

12 000

300

10 000

250

8 000

200 6 000 150 4 000

100

2 000

50 0

0 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 Real GDP growth

Figure 7.3 Source:

Import

Export

Brazil: trade, FDI and GDP, 1990s, index 1990 = 100

BCB, IADB.

FDI

GDP

FDI and trade

Source:

%

154 FDI, Trade and Integration in Mercosur Table 7.5

Brazil: FDI by activity (US$ millions and percentages) 12/1995∗

Agriculture and mineral extraction Petroleum extraction and related services Extraction of ores

1996–2000 (cum.)

Stock

%

Flows

%

688.6

1.6

1 781.0

1.7

72.0

0.2

861.5

0.8

330.3

0.8

597.1

0.6

23 402.4 2 851.3 4 747.7 2 332.4 589.7

55.0 6.7 11.2 5.5 1.4

18 454.6 4 360.5 3 334.5 2 856.4 1 686.1

17.8 4.5 3.2 2.8 1.6

2 072.3 816.0

4.9 1.9

1 226.6 843.2

1.2 0.8

Manufacturing Motor vehicles and parts Chemical products Food and beverages Electronic materials and communication apparatus Machinery and equipment Non-metallic mineral products Office machinery and data-processing equipment Electrical machinery Rubber and plastic articles Basic metallurgy Services

441.4

1.0

732.7

0.7

1 100.3 1 317.9 2 566.2 18 439.0

2.6 3.1 6.0 43.4

685.2 592.0 506.4 83 452.5

0.7 0.6 0.5 80.5

Total (US$ millions)

42 530.0

100.0

103 688.0

100.0

Notes: year. Source:

∗ Data from FDI Census, BCB (1996). Considered inflows to firms > US$10 million/

BCB (various issues).

From 1995 to 2000, investment inflows increased tenfold. This can be explained partly by the privatization process – the Brazilian Central Bank estimated for 1997 that about 28 per cent of FDI inflows were directed towards privatization – and by the acquisition of Brazilian companies.4 FDI data by activity (see Table 7.5) shows that there were some important changes in the sectoral composition of inflows in Brazil during the 1990s. First, we notice that the inflows directed to the service sector increased throughout the decade. While in 1990 only 24 per cent of total FDI went to services, in 2000 about 80 per cent of FDI was directed to this sector. The industrial sector, which previously had absorbed the largest portion of foreign direct investment, lost this position, even though there was a major increase in investments going into this sector in 2000.5

Marta Castilho and Soledad Zignago 155

The main service sectors receiving FDI were electricity and gas, telecommunications, financial intermediation, and services rendered to companies.6 The increase of FDI inflows to some sectors was linked to privatization, which led to ‘an increased inflow of foreign direct investments, reaching US$5.2 billion in 1997, contrasting with the US$2.6 billion in the previous year’ (BCB, 1998). With respect to the industrial sectors, 85 per cent of foreign investment is concentrated in just five of these sectors. The automotive industry is the major FDI recipient, and in the period 1996–2000, and it is the sector that has maintained the most constant level of foreign direct investment inflows. The other four sectors that are important recipients of FDI are chemical products, food and beverages, electronic and communications materials, and machinery and equipment. As for the investor (home) countries (see Table 7.6), the EU countries, taken together, remain the largest investor countries in Brazil, as has long been the case. However, there were some important changes in

Table 7.6

Brazil: investor countries (US$ millions and percentages) 1990 Stock

1995 %

1996–2000 (cum.)

Stock

%

Flows

%

USA Canada European Union Spain Portugal Netherlands France Italy United Kingdom Germany Mercosur Uruguay Argentina Switzerland Japan Fiscal havens1 Other countries

10 488 1 979 14 715 122 68 1 179 1 928 1 303 2 708 5 615 85 51 34 3 222 3 440 1 478 1 055

28.8 5.4 40.4 0.3 0.2 3.2 5.3 3.6 7.4 15.4 0.2 0.1 0.1 8.8 9.4 4.1 2.9

10 852 1 819 14 336 251 107 1 535 2 032 1 259 1 793 5 828 1 268 874 394 2 815 2 659 4 668 4 114

25.5 4.3 33.7 0.6 0.3 3.6 4.8 3.0 4.2 13.7 3.0 2.1 0.9 6.6 6.3 11.0 9.7

24 536 1 102 57 018 21 548 7 563 9 650 7 902 1 613 2 064 1 676 989 458 531 1 118 1 471 14 675 2 778

23.7 1.1 55.0 20.8 7.3 9.3 7.6 1.6 2.0 1.6 1.0 0.4 0.5 1.1 1.4 14.2 2.7

Total

36 461

100.0

42 530

100.0

103 688

100.0

Note: Source:

1

British Virgin Islands, Cayman Islands, Bermuda, Panama and Bahamas. BCB (various issues).

156 FDI, Trade and Integration in Mercosur

the ranking among these countries that took place during the 1990s. Germany, historically the largest European investor in Brazil, strongly reduced its investment volume. The share of the United Kingdom has also been reduced since the beginning of the 1990s. But these reductions were more than compensated by the growth of FDI from Spain, Portugal, the Netherlands and France. In 2000, Spain became the largest foreign investor in Brazil, accounting for more than 30 per cent of inflows, with Portugal taking second place. One must note that these two countries’ investments are directed mainly towards services – notably, communications and electricity. US direct investment in Brazil fluctuated considerably during the decade, dropping from 29 per cent of the FDI stock in 1990 to 18 per cent in 2000. Japan and other Mercosur countries, have shown declining FDI positions in Brazil. In contrast, however, investments from tax haven countries have become quite important – almost 10 per cent of total FDI. As this latter group comprises investments routed through the tax haven countries from other countries, as well as some special forms of operation (including illegal money laundering), we do not take it into account in our analysis below.

3

FDI and trade integration: a sectoral gravity equation

In a previous work (Castilho and Zignago, 2000), we used a gravity approach to test the correlation between trade, investment and regionalism in Mercosur for the period 1985–97. The complementarity between FDI and trade was confirmed by this study: the coefficient of FDI inflows was positive and highly significant, indicating a positive linkage between FDI and imports. On the other hand, we did not find a positive and significant link between regional integration and FDI. This last result was quite surprising, as there is some evidence that recent FDI in Mercosur is linked to an export strategy to the region, at least in some sectors. As shown by Porta and Kosacoff (1997) and Chudnovsky (2001), some MNEs installed in one of the Mercosur countries showed an interest in exporting towards the regional market. The automotive industry is a good example. The new firms installed in Mercosur manifested their intention to explore the regional market and the benefits of the Mercosur auto regime.7 Perhaps, however, this cannot be generalized to all sectors, and if we want to capture the effects of individual firm strategies, we need to make a disaggregated analysis. In fact, as stressed by Fontagné (1995), a better analysis of the correlation

Marta Castilho and Soledad Zignago 157

between trade and FDI requires a bilateral and sectoral approach. However, the lack of disaggregated data concerning FDI flows toward these countries constrains our ability to do this kind of analysis. In this chapter, we use a disaggregated gravity approach8 in order to examine the linkages between trade and FDI, and the influence of regionalism on MNE strategies in the case of Mercosur. Equation (1) tests the hypothesis that FDI and imports are substitutes. For the gravity variables – size and distance – the expected signs are, respectively, positive and negative. The set of sector-specific dummies controls for these specificities. Concerning the investment variable, a negative coefficient would indicate that an increase in FDI would lead to a reduction in imports – confirming the substitution hypothesis – while a positive coefficient would indicate a complementary relationship between investment and imports. A complementary relationship can occur when the subsidiary is an assembling unit (or responsible for just one part of the production process) and when firms continue to buy their components from their original suppliers. Thus, we can write: ln (Mijkt ) = α + β1 ln Yit + β2 ln Yjt + β3 ln distij + β4 ln(1 + FDIijkt )  βk Dk + uijkt +

(1)

k

where: i = FDI’s host country (Brazil or Argentina); j = trade partners; k = sectors; t = year for the period 1990 – 9; M = bilateral import flows; Y = the current dollar GDP; dist = distance in kilometers; and Dk = sector dummies. Equation (2) tests the linkage between exports and FDI. A positive coefficient for FDI suggests that foreign investments generate exports. One could argue that it might be necessary to consider a temporal lag between investment and exports to take into account the maturation time of the investment – that is, that it might take a certain length of time before any particular investment generates exports. But, as stressed before, a large part of the foreign investments in Mercosur countries in the 1990s were acquisitions of public firms via privatization or acquisitions of private

158 FDI, Trade and Integration in Mercosur

firms and, in either case, there is no reason why a temporal lag should be present: ln(Xijkt ) = α + β1 ln Yit + β2 ln Yjt + β3 ln distij + β4 ln(1 + FDIijkt )  βk Dk + uijkt +

(2)

k

Finally, Equation (3) examines the influence of Mercosur formation on foreign investment strategies. A positive coefficient for the integration variable suggests that the Mercosur integration was an additional factor in attracting FDI in the 1990s. Concerning the gravity variables, the size, from the host country’s point of view, might constitute an important factor of attraction and, from the investing country’s point of view, it might be an indicator of investment capacity. The geographical distance might have a negative linkage with investment. In fact, the more distant are two countries, the greater the transport costs, and so there are more incentives concentrate production in local firms. We use two variables to represent the Mercosur integration. The first is a dummy that equals 1 after 1995, when the Common Market was launched. This measure neglects, however, any advances in integration made before 1995 (for example, because of the 1988 sectoral agreements and the liberalization schedule launched in 1991). We decided thus to represent integration by a kind of ‘revealed trade integration’ indicator, where this is an ex-post variable corresponding to the part of Mercosur intra-regional trade in its total trade. We did this because we felt that indicators of trade protection (tariffs, for example) are inadequate; beyond the problem of the choice of the indicator (weighted or not; which weight; tariffs or no tariff barriers and so on),9 they would probably not separate multilateral liberalization effects from regional ones. Thus: ln(1 + FDIijkt ) = α + β1 ln Yit + β2 ln Yjt + β3 ln distij + β4 reg  βk Dk + uijkt +

(3)

k

where reg is: dms = dummy Mercosur: 0 before 1995, 1 after; or lmskt = ln (intra-Mercosur trade/total trade), by sector and by year. The three equations were estimated separately for Argentina and Brazil, pooling all sectors and years. The tests were done using an ordinary least squares (OLS) estimator with sectoral dummies. Argentina data

Marta Castilho and Soledad Zignago 159

came from the CEP (Centro de Estudios para la Producción) and covers twenty-nine countries, twenty-four sectors – primary and secondary – and the period 1990–9. For Brazil, the data from the Central Bank covered the same period, thirteen primary and secondary sectors, and forty-nine countries. The differences between Argentinian and Brazilian sectoral classifications are of no importance for the results, as we analyse them separately. Table 7.7 shows the results concerning Argentina and Brazil. The gravity variables show the expected signs in almost all estimations. Distance coefficients are negative and significant for all estimates. GDP from the investor country presents positive and significant estimated coefficients in all equations. GDP from the host countries, however, shows negative coefficients in two cases: for Brazil in the FDI equation and for Argentina in the exports equation. The first result must be analysed together with the coefficient of the integration variable, as we show below. The second suggests that export performance in Argentina is not linked to GDP, which simply indicates that exports are not residual to domestic demand. The main results of our analysis indicate: (i) the existence of a positive and significant relationship between FDI and imports in both countries; (ii) the existence of a negative but weak relationship between FDI and exports in both countries; and (iii) an ambiguous effect of integration on investment flows – that is, a positive and statistically significant relationship in the case of Brazil but a weaker or insignificant one in the Argentinian case. The first result confirms our precedent findings and is not very surprising. FDI in the 1990s in Mercosur led to a strong increase in imports because local affiliates of foreign firms bought inputs from their parent firms’ original suppliers. In this instance, the case of the automotive sector is again very illustrative. A detailed analysis of the trade balance shows that imports of components grew during the decade at rates consistent with demand growth by new affiliates of foreign firms. Concerning this second point, Chudnovsky (2001) shows that foreign firms in Mercosur enter this market with intent to explore regional markets, and that exports are not the main objective of these firms. Even if we think that these results must be examined with caution, they can be explained by several factors: first, a significant part of investment flows was directed towards the services sector, which engages typically in non-tradable activities; second, a large part of the FDI resulted from foreign takeovers of public firms undergoing privatization (mainly in the service sector) or in acquisitions of national firms, without producing

160 FDI, Trade and Integration in Mercosur Table 7.7

Econometric results

Dependent variable Estimation results for Brazil Explanatory variables R2 No. of observations Intercept Distance PIB investor country PIB Brazil FDI Mercosur integration

Imports (1) Exports (1)

FDI (1)

FDI (2)

OLS OLS 0.58 0.66 1 981 2 098 −15.89∗∗∗ −2.83 −4.24 −1.04 −1.37∗∗∗ −1.46∗∗∗ −15.23 −23.04 0.88∗∗∗ 1.07∗∗∗ 26.30 30.31 0.36∗∗ 1.16∗∗∗ 4.09 1.72 0.05∗∗∗ −0.01∗ 6.88 −1.74 0.46∗∗ 0.02 2.28 0.15

OLS 0.53 2 214 6.85 0.61 −2.12∗∗∗ −8.28 3.41∗∗∗ 36.75 −2.21∗∗∗ −2.62

OLS 0.52 2 214 3.95 0.34 −2.52∗∗∗ −9.63 3.75∗∗∗ 38.51 −2.04∗∗ −2.30

4.85∗∗∗ 7.92

12.23∗∗∗ 3.93

Estimation results for Argentina R2 0.59 0.64 0.20 0.20 No. of observations 5 869 5 497 6 960 6 960 Intercept −27.68∗∗∗ 14.25∗∗∗ −15.59∗∗∗ −19.43∗∗∗ −17.39 7.27 −6.02 −8.92 Distance −1.65∗∗∗ −2.32∗∗∗ −0.96∗∗∗ −0.96∗∗∗ −42.66 −58.18 −17.51 −17.22 0.98∗∗∗ 0.72∗∗∗ 0.73∗∗∗ PIB investor country 1.18∗∗∗ 60.97 48.70 28.13 27.89 PIB Argentina 2.05∗∗∗ −0.84∗∗∗ 1.40∗∗∗ 1.75∗∗∗ 13.52 −5.24 6.77 10.35 FDI 0.05∗∗∗ −0.02∗ 6.52 −1.92 Mercosur integration −0.75 0.40∗∗∗ 0.24∗∗∗ −0.04 −1.16 5.92 2.62 −0.69 Notes: ∗∗∗ significant at 1% level; ∗∗ at 5% and ∗ at 10%; the t -statistics figure below the coefficient values. Coefficients of sector dummies are not presented. FDI (1): integration represented by Mercosur dummy (equals 1 for t > 1994); FDI (2): integration represented by the part of intra-zone trade in total trade. Source:

Authors’ calculations.

significant changes in existing production capacities; and finally, for the bulk of foreign firms, their strategy in investing in Mercosur countries was intended to exploit the regional market and not to create an export platform like those created by foreign firms in East Asia or in Mexico; thus the local affiliates of these firms did not create products for export outside the Mercosur zone.

Marta Castilho and Soledad Zignago 161

The influence of the creation of Mercosur on FDI flows is confirmed by our estimations, where integration is represented by a dummy or by the ratio of intra-regional trade in total trade. The results of both estimations are very close in terms of sign and value of the estimated coefficients. Moreover, gravity variables yield the expected sign, apart from the GDP of Brazil, the coefficient for which is negative. The only significant difference between the two equations is the coefficient of the Mercosur integration, which is not significant for Argentina when represented by the part of intra-regional trade in total trade. In other words, this result seems to suggest that Mercosur market is more important to MNEs in Brazil than those in Argentina. This raises an important issue about the location of MNEs inside Mercosur. One might expect that international firms established in Argentina would consider the integrated market important, because Argentina is a relatively small economy and hence participation in Mercosur would enable these firms to benefit from the scale of a bigger market. But, for those international firms established in Brazil, the domestic market is already much bigger and therefore exports to other Mercosur countries would be residual. Our results, however, suggest that Mercosur seems to favour the location of foreign firms in Brazil. The scale factor, differences in productivity – and even the effects of the different exchange regimes on competitiveness – might explain this preference for Brazil.

Notes 1 As the data used here come from national sources, a number of problems arise, such as variant sectoral classifications, time coverage and minor methodological differences. 2 Argentina, Brazil and Uruguay were allowed to keep 300 products outside the CET until 1 January 2001, whereas Paraguay can maintain 399 products until 2006; in any event, the list of exceptions was modified by later measures taken by Argentina. 3 The aggregated statistics come from IADB/IRELA (1998) and OECD, so the source is the balance of payments, which explains the differences from the national data. 4 See Bonelli (2001) for a detailed analysis of mergers and acquisitions in Mercosur countries. 5 We cannot neglect the fact that privatization in Brazil subsequently slowed. 6 As stressed in the Central Bank FDI report, ‘it is worth mentioning the sharp increase of participation of “services rendered to companies”, basically because this group comprises companies whose activity is participation in the share capital (holdings) and, as such, distribute the received funds to companies of other sectors’.

162 FDI, Trade and Integration in Mercosur 7 This does not mean that multinational firms plan to export outside Mercosur so that the increase of exports may correspond only – or mainly – to extraMercosur exports growth. 8 See Fontagné and Pajot (1998) and Chédor and Mucchielli (1998) for other uses of gravity equations to analyse the relationship between FDI and trade. 9 For a discussion about measures of protection, see Bouët (2000).

References IADB/IRELA (Inter-American Development Bank and Instituto de Relaciones Europeo-Latinoamericanas) (1998) Inversión extranjera directa en América Latina: la perspectiva de los principales inversores (Madrid: IRELA). BCB (Banco Central do Brasil) (various dates) Boletim do Banco Central (Brasilia: BCB). Bonelli, R. (2001) ‘Fusões e Aquisições no Mercosul’, in R. Baumann (ed.) Mercosul – Avanços e desafios da integração (Brasilia: IPEA/CEPA). Bouët, A. (2000) ‘La mesure des protections commerciales nationales’, Working Paper no. 2000-15 (Paris: CEPII). Castilho, M. and S. Zignago (2000) ‘IDE et commerce: les effets de l’intégration régionale dans le Mercosur’, Revue Economique, vol. 51(3), pp. 761–74. Chédor, S. and J.-L. Mucchielli (1998) ‘Implantations à l’étranger et performance à l’exportation: une analyse empirique sur les implantations des firmes françaises dans les pays émergents’, Revue Economique, vol. 51(3), pp. 617–28. Chudnovsky, D. (ed.) (2001) El boom de inversión extranjera directa en el Mercosur (Madrid: Siglo XXI de Argentina Editores). Fontagné, L. (1995) ‘Les liens entre investissements étrangers directs et échanges’, DSTI/EAS/IND/WP9(95)8 (Paris: OECD). Fontagné, L. and M. Pajot (1999) ‘Investissement direct à l’étranger et échanges extérieurs: un impact plus fort aux Etats-Unis qu’en France’, Economie et Statistiques, no. 326–7, pp. 31–52. Porta, F. and B. Kosacoff (1997) La inversión extranjera directa en la industria argentina. Tendencias y estratégias recientes (Buenos Aires: CEPAL-CEP).

8 The Effect of Exchange-rate Uncertainty on Foreign Direct Investment in the United Kingdom Matteo Iannizzotto University of Durham, UK

and Nigel J. Miller University of York, UK

1

Introduction

According to the theory of purchasing power parity (PPP), exchange-rate movements offset the effect of relative price changes on the terms of trade, thus serving as a stabilizing influence on an economy’s external position, protecting real variables from local shocks. In practice, however, exchange rates have deviated substantially and for lengthy periods from the rates implied by PPP, and in this case, exchange-rate movements, rather than serving to absorb real shocks (as in PPP), are a potential cause of shocks.1,2 An additional but separate issue of concern relates to the volatility of nominal and real exchange rates, a marked feature of recent experience of flexible exchange-rate regimes. Various studies have emerged exploring the impact of exchange-rate volatility (and associated uncertainty) on investment and growth (discussed further below). In this chapter we extend the literature, using an original database to explore the relationship between foreign direct investment and exchange-rate variability, with respect to both the level and volatility of exchange rates, for the case of the United Kingdom between 1997 and 2001. The case of the UK provides a particularly rich test-bed for the theoretical predictions. As is well known, between 1996 and 2000, the price of sterling against the euro was subject to significant and sustained real appreciation (rising 25 per cent at a time when inflation was higher 163

164 Effect of Exchange-rate Uncertainty on FDI in the UK

in the UK than in countries of the euro zone, and the UK had no external surplus). In addition, during this period, there was considerable exchange-risk relative to Euro competitor nations because of the UK’s non-participation in the European Monetary Union (EMU), and the uncertainty over its willingness to join in the future. We examine the effect of these on FDI into the UK. This is an issue of pertinence to policy-makers, given the historic importance of FDI to the UK. In 1997, the UK was by far the most significant recipient of FDI in Europe, attracting approximately 43 per cent of all non-European sourced investment (largely from the USA and Japan). As Figure 8.1 illustrates, this proportion has fallen steadily, and in 2000, it was approximately 23 per cent. This is only marginally higher than the proportion of all projects going to France, which has seen little change in real FDI over the same period. Other European countries that traditionally have attracted little FDI have experienced real increases – Spain, Hungary, the Czech Republic and Poland, for example.3 Summarising our conclusions, we find that the appreciation of sterling is strongly related to a decline in FDI into the UK. There is some evidence that exchange-rate volatility has had an adverse effect on FDI, but much more work is required on this issue, particularly on defining and calculating appropriate measures of investors’ perceptions of exchange-rate uncertainty.

Number of announcements

30

5 000

25 4 000 20 3 000 15 2 000 10 1 000

5 0

Announced capital expenditure ($ millions)

6 000

35

Ja n M ’97 a M r ’97 ay Ju ’97 l Se ’97 p N ’97 ov Ja ’97 n M ’98 ar M ’98 ay Ju ’98 l Se ’98 p N ’98 ov Ja ’98 n M ’99 ar M ’99 ay Ju ’99 l Se ’99 p N ’99 ov Ja ’99 n M ’00 a M r ’00 ay Ju ’00 l Se ’00 p N ’00 ov ’0 0

0

Number of announcements

Capital expenditure

Figure 8.1 Number of announcements of FDI and announced capital expenditure into the UK, 1997–2000 Source:

Ernst & Young European Investment Monitor.

Matteo Iannizzotto and Nigel J. Miller 165

The structure of the chapter is as follows. Section 2 reviews the literature, Section 3 discusses the data, while Section 4 presents the statistical methodology and the estimation results. Section 5 concludes.

2

Literature

The literature exploring the real effects of exchange-rate variability is small and relatively inconclusive. There are a number of studies exploring the relationship between measures of exchange-rate variability and aggregate investment, but the theoretical predictions are ambiguous and the empirical evidence sparse. Broadly, however, the literature indicates that investment is negatively related to an appreciation of the domestic currency and to measures of the volatility of the exchange rate, although firm- and industry-related industry-specific characteristics are important determinants of the relationship between investment and the exchange rate. Key factors include the degree of competition, the reliance on imported inputs relative to exports, and the extent of sunk costs (see Goldberg, 1993; Campa and Goldberg, 1995, 1999; Goldberg and Kolstad, 1995; Bell and Campa, 1997; Darby et al., 1999). In contrast, studies that focus exclusively on FDI (the focus of this chapter) tend to find a positive relationship between exchange-rate volatility and the flow of FDI. The small theoretical work can be divided up into ‘risk-aversion arguments’ and ‘production flexibility’ arguments. In the first branch of the literature, the decision to invest overseas depends on firm preferences with respect to risk, with riskaverse investors choosing to invest overseas as a hedge against uncertain exchange-rate movements. In models that stress ‘production flexibility’ arguments, the motivation to invest overseas lies in the opportunities to re-allocate production to the cheaper plant ex post the realization of an exchange shock. Within the class of models stressing risk-aversion, Cushman (1985, 1988) analyses the effect of real exchange-rate risk on investment for four cases, depending on where inputs and outputs are produced, and where financial capital is acquired and output sold. While the theoretical predictions are ambiguous, in empirical tests of annual bilateral direct investment flows from the USA, there is strong evidence of a negative relationship between FDI and exchange-rate appreciation (and expected exchange-rate appreciation). In Goldberg and Kolstad (1995), investors are risk averse and produce for a foreign market with a combination of domestic and foreign capacity, deciding on the foreign share of investment and production prior to

166 Effect of Exchange-rate Uncertainty on FDI in the UK

the realization of exchange-rate shocks (that is, production factors are fixed). Assuming both real demand and exchange shocks, it is shown that exchange-rate volatility tends to increase the FDI share even with identical costs of production across countries.4 Risk-neutral investors, by contrast, are indifferent to the location of production. Using quarterly US bilateral FDI flows to four countries for the period 1978–91, it is shown that exchange-rate volatility and the share of FDI in total investment are positively related, supporting the theoretical predictions (assuming that investors are risk-averse). There is no evidence that volatility has affected aggregate US domestic investment (that is, an increase in FDI outflows are offset by FDI inflows from abroad). Also testing for the effects of levels of the exchange rate on levels of FDI, they find that depreciations of the source-country currency lead to a reduction in FDI outflows, although this effect is not large. (This result is clearly at odds with Cushman, 1985, 1988.) Another paper investigating the role of exchange-rate variability on the FDI of a risk-averse firm is Bailey and Tavlas (1991), who conclude that exchange-rate risk has an ambiguous effect on FDI. The alternative approach focuses on the role of production flexibility in FDI. As discussed, the key departure with the risk-aversion literature is the assumption that there is at least one variable factor that firms may optimize ex post the realization of exchange shocks.5 The literature can be regarded as an open economy extension of the irreversibility literature of Dixit and Pindyck (1994) (D–P) and Abel (1983). By investing in more than one country and suspending the decision on where to produce, investors purchase a real option whose value increases with exchange-rate variability. An example of this class of model is Sung and Lapan (2000), who explore the impact of exchange-rate uncertainty on the FDI of a multinational enterprise (MNE) which may open a plant at home and/or abroad. It is shown that, with sufficient exchange-rate volatility, it is profitable for the MNE to open plants in two countries and to postpone the decision as to where to produce until after the realization of the exchange shock. This contrasts with the scenario with certainty, in which the MNE invests only in the domestic economy even with risk aversion. In other words, FDI is increasing in exchange volatility (other models of this type include Aizenman, 1992; and Kogut and Kulatilaka, 1994).6 By contrast, Campa (1993) tests a Dixit-type model with irreversibility and finds that exchange-rate volatility is related negatively to the number of foreign investments into the USA in the 1980s. This effect is most pronounced for industries where sunk costs are relatively high. In Roy and Viaene (1998), FDI is motivated by strategic considerations. In their model, intermediate inputs are produced abroad in an

Matteo Iannizzotto and Nigel J. Miller 167

oligopolistic market, and FDI enables firms to bid up the price of inputs to non-foreign investing competitors, thus increasing the opportunity cost of non-investment, and leading to ‘bunching’ of FDI. In this environment, exchange-rate variability has a positive effect on FDI. Others have examined the impact of the level of exchange rates on the flow of foreign direct investment and the results are not clear cut (Kohlhagen, 1977 and Campa, 1993). Kohlhagen (1977) shows that a domestic firm expecting a foreign-exchange depreciation defers foreign investment until after the devaluation, when it would be more profitable relative to exporting. In Itagaki (1981), by contrast, an expected depreciation always increases the incentive for foreign production.

3 3.1

The data The regressand

One of the problems this literature has encountered is the often very poor quality of the data. Investment expenditure is only available very infrequently. As far as FDI data in the UK are concerned, these are only available annually (from the Office of National Statistics). Isolating the effect of any single deviation of the exchange rate from equilibrium is therefore difficult. One alternative is to use count data on, for example, the number of FDI projects or investment announcements. Typically, these are available at a higher frequency, although there are several drawbacks to this kind of data. The main one is that it represents only the inflow of FDI. Real exchange-rate movements also influence the decision to disinvest (that is, the option to abandon an investment) and, ideally, we would want a proxy for the net flow of funds (as in the balance of payments data). For the UK, the data set that is available for more recent years and more frequently is the European Investment Monitor (EIM) managed by Ernst & Young. This database consists of announcements of investments in the financial press. Each record includes information on the parent company, the sector and industry, the location of the investment and a given date at which the announcement has been made. Some records also report the programmed capital expenditure. Unfortunately, the database does not report over how many years the stated sum is intended to be spent, and this information is not recoverable in any way. Figure 8.1 graphs both the number of announcements and the sum of the programmed capital expenditure over the relevant years.7

168 Effect of Exchange-rate Uncertainty on FDI in the UK

It is of interest to ask how these data compare with the balance of payments entries. For the very limited period of overlap (1997 and 1998), the two sets seem to be at odds, as the EIM shows a downward trend while the Office of National Statistics displays an upward trend.8 This contradiction can be reconciled, however, if one considers that there is a delay between the announcement and the actual implementation of the investment. If the general behaviour is first rising and then falling, it may be the case that the interval of observation, 1997–8, just happens to fall on the inversion of the trend on the announcement side, but the expenditure side is still dominated by the implementation of investments announced a long time before. One of the disadvantages of using the EIM database is that a constructed series of the capital expenditure variable that would try to mimic the behaviour of the gross inflow of investment expenditure into the UK is liable to be exceedingly noisy and display several outliers, which are caused in many cases by a single announcement of a remarkable size destined to be implemented over several years, but the full amount is recorded in the data set only on the date of its announcement.9 Therefore it is better to concentrate, as others have done (Campa, 1993; Tomlin, 1997), on the number of announcements. These represent the decisions of the firms at or near the time at which they have been taken, and therefore embody the firms’ evaluation of uncertainty over the level and volatility of the exchange rate. We consider this feature of the data to be one of the strengths of our data set. In their papers on the effect of exchange rate uncertainty on FDI into the USA, Campa (1993) and Tomlin (1997) have in mind a non-US firm that decides to invest in the USA in order to have access to the US market. While access to the UK market is one motivation for FDI in the UK, investment in the UK can also be seen as a step towards gaining access to the wider European Union (EU) market. This is particularly true of a non-EU firm that decides to locate its capital in the UK, rather than, for example, in France or Germany. The effect of exchange-rate variability is therefore applicable in two ways: take, for example, a US firm investing in the UK. On the one hand, there is uncertainty over the variability of the dollar–sterling exchange rate, which particularly affects the costs of the investment. On the other hand, if the investment is undertaken with a view to penetrating the Common Market, there is also uncertainty over the variability of the sterling–DM/euro exchange rate, to take only the most important rate, because this would affect the ability of the new plant to be competitive over the wider arena of the continental market. Arguably, the second concern over exchange-rate variability would

Matteo Iannizzotto and Nigel J. Miller 169

disappear, were the UK to join the EMU. In fact, uncertainty over the outcome of this political process may well constitute a third element of uncertainty that the firm faces in making its investment decision. To summarize, then, in the case at hand, a marked appreciation of sterling, coupled with the volatility of a floating rate, has a negative effect on FDI into the UK because (i) of higher (sterling) costs of implementation; (ii) of loss of competitiveness on the EU-wide market; and (iii) because reason (ii) may or may not disappear, depending on the outcome of the referendum on EMU. In the D–P framework, all three reasons, the third one being related to the first two but somewhat different in nature, raise the opportunity cost of waiting, and hence firms prefer to postpone their investments (or to locate elsewhere). Admittedly there is therefore a problem of identification, but that is necessarily the case in this literature because even in the D–P model the element of uncertainty constitutes a bit of a catch-all variable standing for anything that has a bearing on the profitability of the investment decision, and whose outcome is unknown and fairly unpredictable at the time the decision is made. One may speculate that among all factors that may be captured by this variable, the future level of demand for the firm’s product is likely to be the most important element of uncertainty facing all firms. When one wants, therefore, to discriminate and isolate the effect of other factors, be they the regime of taxation, the introduction of new technologies rendering obsolete the invested capital, or indeed, the variability of the exchange rate, one is necessarily confronted with a problem of identification. The best one can do, then, is to select those firms whose decisions are more likely to be affected by the factor of uncertainty being analysed, while the future level of demand for the product would affect every firm in a non-discriminatory sort of way. For the exchange rate, then, it is better to focus on FDI, as foreign firms are more likely to take into account in a more prominent way the uncertainty over its variability, over and above the uncertainty of any other factor. Also the degree of irreversibility, or the size of the sunk cost10 that a firm needs to undertake for a given investment, is going to play a major role in the selection of the firms – which leads to the conclusion that one ought to select non-European firms investing into the domestic economy in sectors where high sunk costs and a high degree of irreversibility are likely to be present. The literature has identified these industries as being found mainly in the manufacturing and energy sectors. The Ernst & Young EIM database records selected in this chapter are therefore only those pertaining to the manufacturing sector and the

170 Effect of Exchange-rate Uncertainty on FDI in the UK

energy sector, and whose parent companies are located outside the EU. These amount to 625 records over the four years of the database, the vast majority of them being in the manufacturing sector, with only a handful (six over four years) in the energy sector. In order to remain consistent with our interpretation of these announcements as representing the decision to invest, it also makes a difference whether the investment undertaken is an expansion of an existing concern or an altogether new enterprise. It is conceivable that a firm that has already invested in a country is likely to re-invest in that same country to upgrade machinery, say, regardless of what the exchange rate might be doing, whereas an entirely new venture is more likely to reflect the outcome of a decision process such as the one we are trying to capture. We therefore select the records in such a way as to include only new investments, and exclude expansions of existing ventures. The selection of records in these sectors also has the advantage over the data used by Campa (1993) and Tomlin (1997), that one does not need to control for a threshold level of sunk investment – say, in advertising – as Campa (1993) does, because his data were on the wholesale sector and were therefore likely to represent investments with a much higher degree of reversibility than, for example, the building of a new car plant is likely to have. 3.2

The regressors

There are three regressors – the level of the real exchange rate, the volatility of the real exchange rate, and labour costs of the UK relative to competitor FDI locations. Clearly, the measure of the real exchange rate needs to take into account the disparate provenance of the firms that invest. Campa (1993) constructs a weighted average of the relevant nominal exchange rates where the weights are given by the relative proportion of each country’s firms in the sample. Unfortunately, this methodology is open to the criticism that the sample only includes actual investors and provides no information about potential investors. The weights may therefore be biased. We choose instead to use a real effective exchangerate series that has been weighted by the level of trade between the UK and the rest of the world. In this way, the bias of the sample towards actual investors is absent, and we account – as much as it is feasible – for potential investors in the sense that investors, be they actual or potential, are far more likely to come from countries that are already trading with the UK. As is well known, over the period

Matteo Iannizzotto and Nigel J. Miller 171

under consideration, the UK experienced a significant and sustained real appreciation. In this study, because of the lack of consensus in the literature on the ‘best’ approach, we use a variety of measures to represent investors’ perceptions of exchange-rate volatility. Typically, volatility is measured by the average of the standard deviation of the logarithm of the real exchange rate. Where investors are assumed to have perfect foresight, volatility is the standard deviation of the log of future exchange rates (Campa, 1993). However, most studies assume backward expectations – that is, investors’ expectations are determined by previous experience, and the measure of volatility depends on the values of past exchange rates (see Kenen and Rodrik, 1986; Bell and Campa, 1997; Darby et al., 1999; Dell’Ariccia, 1999). There is no consensus on the time horizon of the representative investor. For example, in Darby et al. (1999), volatility is the standard deviation (SD) of the exchange rate over a two-year interval; in Campa and Goldberg (1995, 1999) the interval is three years; while in Campa (1993) it is only one year. Not surprisingly, differences of this sort, in the nature of expectations and the extent of investors’ memory/foresight, make a significant difference to the estimated measure of volatility, although this would appear to have little impact on study results (for a discussion, see Carruth, Dickerson and Henley, 2000). In this chapter, we implement five different measures, considering both the SD and SD/mean measures, and varying the temporal window and the nature of expectations.11 As discussed already, in this study we are concerned primarily with the factors that determine the choice of location for a non-European investor seeking to gain access to the wider European market. In particular, we seek to offer some insight into the recent marked decline in FDI projects in the UK compared to the rest of Europe. One plausible explanation is relative marginal costs. Unit–labour costs are probably the best proxy for this but, unfortunately, they are only available for a limited number of countries. Instead, we construct an index describing monthly changes in real wages of the UK relative to a sample of the UK’s largest European competitors. This is based on nominal wage data available at the quarterly level in the Statistical Compendium of the OECD. A number of countries report wages by sector, and specifically the manufacturing sector. Where this is not available we take wage data for all industries.12 We take France and Germany as the two major European partners of the UK and construct a wage index, where each country is weighted according to the proportion of FDI projects going to that country in 1997.

172 Effect of Exchange-rate Uncertainty on FDI in the UK

One feature of relative real wage movements during 1995–2000 is notable: there is no significant change in real wages in the three biggest recipients of FDI – the UK, Germany and France.

4

The statistical model and estimation results

While Campa (1993) uses a standard tobit model, the tobit only constitutes an approximation. As Greene (2000) points out, to take into account the number of zeros and the discrete nature of the regressand, the appropriate model for count data is a Poisson regression. The approach taken in this chapter is to assume that the number of announcements per month FDI count = Yi is related to the regressors in a log-linear way so that:13 y

PROB(Yi = yi ) =

e−λi λi i yi !

ln yi = b
i xi

(1) (2)

where xi is the vector of regressors comprising an index of the tradeweighted real exchange rate, a measure of its volatility and an index of relative real wages. The Poisson distribution imposes the restriction that the mean and the variance of the distribution are the same. Tomlin (1997) notes that the outcome of 0, that is, no announcement to invest in the UK, can be interpreted in two ways: on the one hand it can mean that the firm has genuinely decided not to invest because of the explanatory variables included in the model; or the firm was never going to consider investing in the UK in any case. Therefore the zero-inflated Poisson (ZIP) model is used. As the standard Poisson model and the ZIP are non-nested, the Vuong test for non-nested models is applied, to ascertain which distribution applies. This test calculates the logarithm of the ratio of the conditional probability of the dependent variable, conditional on the independent variables, for two alternative distribution hypotheses. The specification of the alternative hypothesis bears a great weight in this instance, as the power of the test may be reduced (Greene, 2000). Formally defined as in Greene (2000):
mi = log

f1 (yi |xi ) f2 (yi |xi )

 (3)

Matteo Iannizzotto and Nigel J. Miller 173

Then the Vuong test can be defined as:  √  1 n n n i=1 mi v=   n 1 2 i=1 (mi − m) n

(4)

where n is obviously the sample size. Values of the statistic that are smaller than −1.96 count as evidence against the ZIP model, whereas a value greater than 1.96 is in its favour. We perform the Vuong test on all our models to check for the possibility that a ZIP model might being more appropriate. The results are reported in Table 8.1. In all cases, the test statistic comes out to be negative and with an absolute value greater than −1.96. As a result, the standard Poisson regression model for count data is used. The estimation results are presented in Table 8.2. As stated above, we choose to maintain a degree of agnosticism as to the appropriate measure of the volatility of the exchange rate. As a result, we perform the estimation five times, for as many models as there can be defined measures of the volatility of the exchange rate. As argued in Section 2, with some qualifications, the theoretical prediction is that all parameter estimates are negative – that is, FDI is negatively related to real appreciation, exchange-rate volatility and real wages. The parameter estimates for the level of the exchange rate are all significant at conventional significance, and all have the expected negative sign for all models. The parameters for the index of relative wage costs is only significant in Model 5. The measure of exchange-rate volatility that differentiates the five models one from another, gives somewhat inconclusive results. The forward-looking measure that assumes perfect foresight is positively signed and significant for the one-year horizon. It is not clear how to treat this result. It seems that this method of measuring exchange-rate Table 8.1 Vuong test statistic for non-nested models (Poisson and ZIP) Model 1 Model 2 Model 3 Model 4 Model 5 Source:

−2.5489 −7.3206 −3.3947 −2.5161 −2.4654

Authors’ calculations.

174

Table 8.2 Estimation results Constant

Exchange rate

Model 1

30.4872 15.49

−3.25102 0.7675

Model 2

27.7980 14.03

−5.08847 0.6732

Model 3

28.1148 13.70

−5.07561 0.6687

Model 4

27.1502 13.64

−3.72242 0.9179

Model 5

29.3025 13.76

−5.23758 0.6879

VolPF1

VolB1

VolB2

VolB3

VolB3sd

Wage −24.4548 15.47

47.7790 16.99

−18.8741 13.91

2.02553 17.67 −8.22044 27.99

−19.0592 13.65 −97.3150 46.43

−18.5113 13.60 2.33213 1.426

−20.2806 13.71

Note: The Poisson model of FDI count is modelled on a constant, the level of the exchange rate, various measures of the volatility of the exchange rate according to whether Perfect Foresight (PF) or Backward-looking Expectations (B) over 1, 2 or 3 years of time horizon is assumed, and an index of relative real wage competitiveness vis-á-vis the UK’s European partners. Estimates that are significant at conventional significant levels are in bold type. Source: Authors’ calculations.

Matteo Iannizzotto and Nigel J. Miller 175

volatility yields results that are at variance with the theoretical predictions, unless one is willing to maintain that there should be a positive effect of volatility over FDI. For the backward-looking specification, as stated above, four different specifications are used, but only on the three-year time horizon is the parameter estimate significantly different from zero. These results therefore lend support to the theoretical literature that indicates an appreciation of the exchange rate to depress the influx of FDI.

5

Conclusions

In this chapter we have contributed to the literature on the effects of uncertainty over investment decisions by firms. We have concentrated on the effects of the level and volatility of exchange rates on foreign direct investment into the UK from outside the European Union over the period 1997–2000. Our database is an original one, never before used in the literature, compiled and maintained by Ernst & Young, where the announcements in the financial press of investments are recorded. We extract from this database the records of those firms that are more likely to be affected by the exchange rate. The literature has identified these as being found in the manufacturing and energy sectors. We also select only firms from outside the European Union. This selection is done both for the sake of ensuring that a sizeable sunk cost of investment is assumed, and for the sake of going some way towards solving the problem of identification that an empirical analysis of a Dixit–Pyndick framework necessarily implies. The end result is a dependent variable that represents the number of announcements per month. We also select records so that only genuinely new investments are included, as we consider that these are more likely to be affected by the exchange rate than are reinvestments in going concerns. Given the monthly frequency of the data, we are also able to improve on much of the empirical literature, as we are able to concentrate on a single appreciation episode – namely, that of sterling from 1996. The availability of data on FDI on, at most, a quarterly basis or, more often, on an annual basis, has in fact meant that it was difficult to isolate any particular episode because the number of observations to be found were insufficient to implement a proper estimation. We do not face such a problem.

176 Effect of Exchange-rate Uncertainty on FDI in the UK

We have maintained a degree of agnosticism over which particular measure of the volatility of the exchange rate is for estimation, as a degree of arbitrariness is unavoidable in this choice. Our estimation results, by means of a standard Poisson regression model, lend support to the theoretical literature. Specifically, we find that FDI flows are negatively related to real wages in the UK relative to European competitors. Importantly, this study finds strong empirical evidence that the real appreciation of sterling in recent years has reduced FDI into the UK. The evidence on the effect of exchange-rate volatility is inconclusive, although this result is probably affected greatly by the way one chooses to measure exchange-rate volatility, and there is – at the time of writing – no consensus in the literature on this matter. Notes 1 For evidence, see Froot and Rogoff (1994). 2 Recent models of investment path dependence show that temporary movement in the exchange rate may have permanent effects. 3 As discussed later, this data refers to investment announcements (EIM, Ernst & Young). 4 The relationship between exchange variability and FDI is sensitive to the direction of correlation between demand and exchange shocks. 5 Typically, this is labour. 6 There is a related and extensive literature that explores the relationship between exchange-rate variability and trade, drawing on arguments, used in the investment literature, of risk-aversion and real option value. Although the weight of evidence suggests that volatility is bad for trade, the evidence is inconclusive (see Kenen and Rodrik, 1986). In a recent study, Dell’Ariccia (1999) shows that exchange-rate uncertainty has a negative effect on trade in Western Europe. 7 Only selected records are included; for the criterion that determines this selection, see further. 8 One may have to accept the estimation of a trend in this case, with only two data points, which is certainly methodologically dubious. 9 The outstanding case is the investment undertaken by Motorola in Scotland announced in January 2000 which, by itself, is something like 400 times the median size of announced investment. 10 See Campa (1993) on this issue. 11 Some alternatives to SD measure are proposed. Campa and Goldberg (1995) and Goldberg and Kolstad (1995) use the ratio of the standard deviation to mean the real exchange rate. In Campa and Goldberg (1995) this is apparently highly correlated to the SD measure and has no significant impact on the results. Others use the average absolute difference between the previous period forward rate and the current spot. As Dell’Ariccia (1999) argues, this measure has the advantage of picking up the presence of a ‘peso problem’, or the lack of credibility of the official parity. Another possibility is to use the percentage difference between the maximum and the minimum of the

Matteo Iannizzotto and Nigel J. Miller 177 nominal spot rate over the t years preceding the observation, plus a measure of exchange-rate misalignment (Peree and Steinherr, 1989). 12 Monthly real wages are derived by deflating by the CPI for each country and by smoothing of quarterly data. 13 Greene, 2000.

References Abel, A. (1983) ‘Optimal Investment under Uncertainty’, American Economic Review, vol. 73, pp. 228–33. Aizenmann, J. (1992) ‘Exchange-Rate Flexibility, Volatility and the Patterns of Domestic and Foreign Direct Investment’, IMF Staff Papers, vol. 39(4), pp. 890–922. Bailey, M. and G. Tavlas (1991) ‘Exchange Rate Variability and Direct Investment’, Annals of the American Academy of Political and Social Science, no. 516, pp. 106–16. Bell, G. and J. Campa (1997) ‘Irreversible Investments and Volatile Markets: A Study of the Chemical Processing Industry’, Review of Economics and Statistics, vol. 79(1), pp. 79–87. Campa, J. (1993) ‘Entry by Foreign Firms in the US under Exchange-Rate Uncertainty’, Review of Economics and Statistics, vol. 75(4), pp. 614–22. Campa, J. M. and L. S. Goldberg (1995) ‘Investment in Manufacturing, ExchangeRates and External Exposure’, Journal of International Economics, vol. 38, pp. 297–320. Campa, J. M. and L. S. Goldberg (1999) ‘Investment, Pass-through and Exchange Rates: A Cross-country Comparison’, International Economic Review, vol. 40(2), pp. 287–313. Carruth, A., A. Dickerson and A.nley (2000) ‘What Do We Know about Investment under Uncertainty’, Journal of Economic Surveys, vol. 14(2), pp. 119–53. Cushman, D. (1985) ‘Real Exchange-Rate Risk, Expectations and the Level of Direct Investment’, Review of Economics and Statistics, vol. 67, pp. 302–7. Cushman, D. (1988) ‘U.S. Bilateral Trade Flows and Exchange Risk during the Floating Period’, Journal of International Economics, vol. 24(3–4), pp. 317–30. Darby, J., A. Hallett, J. Ireland and L. Piscitelli (1999) ‘The Impact of Exchangerate Uncertainty on the Level of Investment’, Economic Journal, vol. 106, pp. C55–C67. Dell’Ariccia, G. (1999) ‘Exchange Rate Fluctuations and Trade Flows: Evidence from the European Union’, IMF Staff Papers, vol. 46(3), pp. 315–34. Dixit, A. and R. Pindyck (1994) Investment under Uncertainty (Princeton, NJ: Princeton University Press). European Investment Monitor (London: Ernst & Young). Froot, K. and K. Rogoff (1994) ‘Perspectives on PPP and Long-run Real Exchange Rates’, National Bureau of Economic Research, Working Paper no. 4952 (Cambridge, Mass: NBER). Goldberg, L. (1993) ‘Exchange-Rates and Investment in US Industry’, Review of Economics and Statistics, vol. 75, pp. 575–88. Goldberg, L. and C. Kolstad (1995) ‘FDI, Exchange-Rate Variability and Demand Uncertainty’, International Economic Review, vol. 36, pp. 855–73. Greene, W. (2000) Econometric Analysis (Upper Saddle River, NJ: Prentice-Hall).

178 Effect of Exchange-rate Uncertainty on FDI in the UK Itagaki, T. (1981) ‘The Theory of the MNF under Exchange-Rate Uncertainty’, Canadian Journal of Economics, vol. 14, pp. 276–97. Kenen, P. and D. Rodrik (1986) ‘Measuring and Analysing the Effects of Short-Term Volatility in Real Exchange-Rates’, Review of Economics and Statistics, vol. 68, pp. 311–15. Kogut, B. and N. Kulatilaka (1994) ‘Operating Flexibility, Global Manufacturing and the Option Value of a Multinational Network’, Management Science, vol. 40, pp. 123–39. Kohlhagen, S. (1977) ‘The Effects of Exchange-Rate Adjustment on International Investment’, in P. Clark, D. Logue and R. J. Sweeney (eds), The Effects of ExchangeRate Adjustments (Washington, DC: US Government Printing Office), pp. 194–7. Peree, E. and A. Steinherr (1989) ‘Exchange-Rate Uncertainty and Foreign Trade’, European Economic Review, vol. 33, pp. 1241–64. Roy, S. and J. M. Viaene (1998) ‘On Strategic Vertical Foreign Investment’, Journal of International Economics, vol. 4(2), pp. 253–79. Sung, H. and H. Lapan (2000) ‘FDI and Exchange-Rate Uncertainty’, International Economic Review, vol. 41(2), pp. 411–23. Tomlin, K. (1997) ‘The Effects of Model Specification on FDI Models: An Application of Count Data Models’, Southern Economic Journal, vol. 67(2), pp. 460–8.

9 The Role of Foreign Direct Investment and Natural Resources in Economic Development∗ José De Gregorio Central Bank of Chile

Foreign direct investment (FDI) has long been a topic high on the policy agenda in emerging markets. The contribution of FDI to a country’s external financing and economic growth, the behaviour of multinational corporations, and the extent of regulation of FDI and other forms of capital flows are some of the issues on which policy-makers usually have to take a stand. It is no coincidence that economic research has devoted much effort to exploring these issues. In this chapter I discuss some of the issues regarding FDI in developing countries, and in particular two issues highlighted in recent discussions and research. The first is whether, and through which channels, FDI affects economic growth. The second is the impact on economic growth of the exploitation of natural resources, normally developed by foreign investors. The two questions are related, and from the answers we can derive policy implications for the treatment of FDI in developing countries. In Section 1, I review some stylized facts about the behaviour of FDI, especially in the years since the surge of capital inflows of the 1990s came to a halt. In Section 2, I discuss the evidence showing that FDI does affect economic growth and the mechanisms through which that occurs. In Section 3, I discuss the relationship between the exploitation ∗

This chapter was prepared for an invited lecture at the 13th World Congress of the International Economic Association, Lisbon, Portugal. I am grateful to Edward M. Graham for comments, and Alberto Naudon for valuable assistance and comments. 179

180 Role of FDI and Natural Resources in Economic Development

of natural resources, in which foreign investors often play an important part, and economic growth. The chapter ends with Section 5 and some concluding remarks.

1

Some stylized facts

Figure 9.1 summarizes the recent history of capital inflows to emerging markets. The late 1970s and early 1980s witnessed a surge of capital flows, but this primarily took the form of debt. During the second half of the 1980s, after the debt crisis that began in Latin America in the early 1980s, capital flows to emerging markets almost disappeared. The early 1990s, however, saw a reversal of this trend: a renewed and even stronger upsurge occurred during that decade, reaching a peak on the eve of the Asian crisis of 1997. Interestingly, this process has its mirror image in the current account of the USA: the US current account shifted gradually from near balance in the early 1980s to a deficit of about 3 per cent of GDP by 1988. It had fallen to almost zero again by 1991, but since then it has been increasing steadily, to almost 5 per cent of GDP in 2002. This shows that, although capital flows respond to opportunities created in emerging markets, they also move according to the availability of capital in the developed world. The latter part of the 1990s saw another sharp decline of capital flows to emerging markets. From more than US$300 billion in 1996, inflows have declined to little more than half that magnitude in the early

250

US$ billions

200 150 100 50 0 –50 1978 1980 1982 1984 1986 1988 1990 1992 1994 1996 1998 2000 2002 2004f FDI

Figure 9.1

Debt, bonds and other flows

Official creditors

Net capital flows to emerging markets (billions of 2000 dollars)

Note: f = forecast. Source:

Portfolio equity

Institute of International Finance.

José De Gregorio 181 (a) 1978–89

(b) 1990–6

87%

73%

16%

4% 9% 8% 0%

2%

1%

0%

(c) 1997–2000 83% 7%

8% 2%

Emerging Asia Emerging Europe Developed countries Latin America Other emerging markets

0%

Figure 9.2 Source:

FDI inflows distribution, 1979–2000 (percentages)

IMF Balance of Payments Statistics.

2000s. But another important dimension of the capital flows slowdown has been a change in composition. Whereas, before 1990, emerging economies received only 13 per cent of worldwide FDI in a typical year (see Figure 9.2), by 1996 that share had more than doubled. Emerging economies in Asia claimed the bulk of these increased flows, but their share declined again after the Asian crisis of the late 1990s. As Figure 9.3 shows, however, even as all forms of capital flows to Latin America have declined, the importance of FDI within those flows has increased. Indeed, in all the world’s emerging regions, but especially in Latin America, the relative importance of FDI in capital inflows has been increasing, from about 10 per cent of total inflows in the early 1990s to more than 35 per cent in 2002–3 (see Figure 9.4). At the time of writing it is widely taken for granted that FDI will be the main, and indeed almost

182 Role of FDI and Natural Resources in Economic Development 100 80 60 40 20 0 –20 1990 FDI

Figure 9.3

1992

1994

1996

Debt, bonds and other flows

1998

2000

Portfolio equity

2002

2004f

Official creditors

Net capital flows to Latin America (billions of 2000 dollars)

Note: f = forecast. Source:

Institute of International Finance.

50 45 40 35 30 25 20 15 10 5 0 1990

1992

Emerging markets

Figure 9.4

1994

1996

Latin America

1998

2000

2002

2004f

Emerging Asia ex. China and India

Share of FDI in net capital inflows by region (percentages)

Note: f = forecast. Source:

Institute of International Finance.

the only, vehicle of foreign financing in developing countries for years to come. As can be seen from Figure 9.3, the decline of FDI in Latin America has been greater than in the rest of the world (see Figure 9.1), to a large extent because of the strength of FDI flows to Asia, and in particular to China.

José De Gregorio 183

120 100 80 60 40 20 0 1993

1994

1995

1996

1997

1998

1999

2000

Developed countries Developing countries Latin America and Caribbean Figure 9.5 Source:

Share of mergers in FDI (percentages)

UNCTAD (1999, 2001).

The mid-1990s surge in FDI to emerging economies largely took the form of greenfields investment – that is, the construction of new foreignowned facilities. Much FDI in recent years, however, has come in response to privatization and, more generally, to opportunities for mergers and acquisitions generated by lower asset prices in these economies. In this respect, capital flows in the developing world have come to resemble those among developed countries, where more than 80 per cent of flows of FDI take the form of mergers and acquisitions.1 As Figure 9.5 shows, the share of mergers and acquisitions in FDI inflows into Latin America rose from about 20 per cent in the early 1990s to around 50 per cent in 2000. An important and frequently cited feature of FDI is that it tends to be relatively stable: when a crisis erupts, FDI cannot flee the country as easily as more liquid forms of capital such as portfolio flows and debt. In times of turbulence, these more liquid flows may increase the volatility of the capital account, with adverse consequences for exchange rates and economic activity. In a sense, foreign direct investors share more in the risks of the domestic economy as the prices of their assets adjust to local economic conditions. A simple way to illustrate this point is to examine the persistence of different flows by estimating the autocorrelation coefficient for a series of annual flows. Table 9.1 shows this coefficient, with one and two lags, for annual data from 1978 to 2000. The table shows that FDI tends to be more persistent than other types of flows,

184 Role of FDI and Natural Resources in Economic Development Table 9.1 Persistence of capital flows Autocorrelation of order

Emerging Markets FDI Portfolio Debt, bonds and other flows Official creditors Latin America FDI Portfolio Debt, bonds and other flows Official creditors Emerging Asia (excl. China and India) FDI Portfolio Debt, bonds and other flows Official creditors Source:

1

2

0.93 0.74 0.66 0.28

0.85 0.52 0.23 −0.03

0.9 0.66 0.56 −0.31

0.8 0.29 0.24 −0.3

0.87 0.52 0.61 0.32

0.66 0.27 0.03 −0.13

Author’s calculations.

and especially persistent in Latin America. This evidence suggests that those countries that are able to attract FDI can expect to see these inflows continue; on the other hand, should FDI stop flowing in, the drought may last a long time. The fact that FDI flows tend to be more stable does not mean that there is no instability associated with foreign investors. Indeed, in periods of turmoil, as foreign investors attempt to hedge their exposure in the exchange rate market, they could induce volatility in financial markets. They could even engage in cross-border hedging, which could introduce instability in the foreign exchange market. Although there is no detailed description of this phenomenon, anecdotal evidence shows that this may have been an important source of instability in East Asia, Mexico and more recently in Latin America. However, this is not an issue exclusive to FDI, but rather to the mechanism to protect the economy from turbulence, and hence it is related more to the prudential regulation of financial markets and to the exchange-rate regime. It is well established, and is discussed below, that countries with a history of strong growth performance are able to attract FDI. This might lead one to think, given the evidence reviewed above, that FDI into the Latin-American economies has increased because these economies are performing well. However, this is clearly not the case. Here it is

José De Gregorio 185

80 70 60 50 40 30 20 10 0 Aaa Figure 9.6 Source:

Aa

A

Baa

Ba

B

Share of FDI in total capital inflows by credit rating (percentages)

IMF, Balance of Payments Statistics and Moody’s.

important to recognize that FDI is almost the only form of capital still flowing to the region. Recently, Haussman and Fernández-Arias (2000) have presented convincing evidence against the view that FDI dominates capital inflows in countries that are more promising, safer, and have better institutions and policies. Figure 9.6 shows that those countries whose inflows consist mainly of FDI tend to be countries with lower creditworthiness. Countries classified by Moody’s as B and Ba receive almost 70 per cent of their capital inflows as FDI, whereas for investment-grade countries that value is between 30 and 40 per cent. These data confirm that, although countries with better performance receive more capital inflows, these inflows are less tilted towards FDI than in economies with poorer performances. The recent trends in Latin-American countries’ capital accounts confirm that, as risk and instability have increased in some important parts of the region, FDI has become their most important source of external financing. At the firm level, Desai, Foley and Hines (2002) show that the existence of capital controls distorts the allocation of assets by American multinational corporations. These companies over-invest in physical capital in countries with capital controls, and under-invest in financial assets. Because multinational enterprises (MNEs) enjoy access to world capital markets, they have an advantage over purely domestic firms in investing in activities that are capital-intensive. This may explain why, in less developed economies and in economies less fully integrated into the world economy, there is a bias towards FDI with respect to other forms of capital. It is important to note that this result refers to composition, not volume of capital flows, but it may help to explain why the evidence

186 Role of FDI and Natural Resources in Economic Development

on FDI and growth tend, as discussed below, to have conflicting results. Countries with large FDI may be countries that can grow faster, but if there is a bias towards FDI because of distortions this may weaken the connection between FDI and growth.

2

The effects of FDI on economic growth

The accumulation of capital is an important determinant of economic growth. Since FDI is a component of total investment, it too contributes to growth. But the interesting question is whether FDI exerts a positive growth effect beyond the direct effect through increased investment – for example, because it facilitates the transfer of knowledge from abroad. Addressing this issue from both an analytical and an empirical point of view may lead to important policy implications: if FDI does foster growth beyond its simple contribution to capital accumulation, policymakers may wish to give special consideration to policies that promote the inflow of FDI. There are many reasons why FDI might give rise to beneficial externalities that promote economic growth: FDI might allow a country to bring in technologies and knowledge that are not readily available to domestic investors, and in this way increase productivity growth throughout the economy. It might also bring in expertise that the country does not possess, and foreign investors might have better access to global markets. Indeed, De Gregorio (1992) found support for this view by examining the evidence on economic growth in Latin America during the period 1950–85: increasing aggregate investment by one percentage point of GDP was found to increase economic growth by 0.1 per cent to 0.2 per cent a year, but increasing FDI by the same amount increased growth by approximately 0.6 per cent a year. This indicates that FDI is about three times more efficient than domestic investment. Along similar lines, and for a broader sample of countries, Blomstrom, Lipsey and Zejan (1992) found that FDI has a positive effect on growth under certain circumstances. Comparing samples of low- and highincome countries, they found a positive effect of FDI only in the second group, suggesting that there is a threshold level of income above which FDI has extra effects on economic growth, and below which it does not. The study by Blomstrom, Lipsey and Zejan is consistent with the idea that only those countries that have reached a certain level of income can absorb new technologies and benefit from technological diffusion, thus reaping the extra advantages that FDI can offer. What explains this

José De Gregorio 187

differential response to FDI at different levels of income? The prime suspect is human capital, a variable that is correlated positively with the level of income per capita. It may take a well-educated population to spread the benefits of newly-introduced technologies to the whole economy. This was the idea explored by Borensztein, De Gregorio and Lee (1998), who analysed the growth effect of FDI in a panel data set of sixty-nine developing countries during the period 1970–89. In an initial analysis, FDI was found to have a positive effect on growth beyond the direct investment effect, but not in general a statistically significant effect. However, when the authors constructed a new variable – the FDI variable multiplied by a measure of human capital – the effect turned out to be both positive and significant. This finding showed that the interaction of FDI and human capital had an important impact on growth. But this effect was observed only when the level of human capital, as measured by years of secondary-school enrolment among the male population, was sufficiently high. The intuition behind this result is that, for a country to take advantage of technological diffusion resulting from FDI, it must have a high level of human capital. For example, a permanent increase in FDI equivalent to 1 per cent of GDP in a country with an average educational attainment of 0.91 years of secondary schooling (the average in the sample), would increase the rate of growth by 0.6 per cent a year. More recently, Balasubramanyan, Salisu and Sapsford (1999) confirmed the positive interaction between human capital and FDI. They also found that the more open economies experience greater benefits from FDI. In addition to the cross-country evidence, many studies have attempted to measure directly, using micro data, the existence of spillovers from foreign investors to the whole economy. In an important paper, Aitken and Harrison (1999) studied the spillover effects of foreign firms on domestic firms in Venezuela and found very limited effects at the plant level. Moreover, the small spillover they found came mainly from joint ventures, suggesting that close ties between domestic entrepreneurs and foreign investors generate some spillovers, but that purely foreign plants do not. This evidence is interesting and persuasive, but by its nature it cannot capture spillovers from foreign investors to the whole economy. For example, if foreign investors in manufacturing induced productivity improvements in some way in, say, the financial sector, this would not be captured at the plant level. Indeed, in a recent study of the Indonesian manufacturing sector, Blalock and Gertler (2004) look at spillovers from foreign investors to

188 Role of FDI and Natural Resources in Economic Development

other sectors in the economy. In particular, given that most foreign investors are in the exportable sector, they examine whether there is a spillover upstream to local suppliers. They show that multinationals transfer technology to reduce the costs of inputs and to increase competition among suppliers. Their results show large productivity gains, greater competition and lower prices among local firms upstream from foreign firms. In addition, there is an externality by which producers in other sectors that demand supplies from the same upstream firms also benefit from technological transfers. This chapter is the first study that has looked beyond the sector where foreign firms are producing and shows that spillovers are significant.2 In a recent paper, Carkovic and Levine (2005) took another look at the effects of FDI on economic growth in a large sample of countries. Previous studies had tackled the endogeneity issue – that is, whether the observed positive effect is caused by the fact that fast-growing countries are the ones best able to attract foreign investors – but these authors argued that their procedures were more robust to the endogeneity problem. After using a battery of tests and techniques to control for these problems, Carkovic and Levine concluded that growth and a good macroeconomic environment are what drive FDI, rather than the reverse. Although most of the evidence would point to a positive effect of FDI on economic growth, the evidence is not fully conclusive. As argued above, the fact that FDI sometimes becomes more important than other forms of capital flow may be the result of capital controls or distortions in the tax system, or may simply be the safest way to get into a country with weak institutions (see Figure 9.6), and may weaken the relationship between growth and FDI. However this refers to the composition of inflows rather to than its level. One would like to see the relationship of the aggregate degree of international financial integration to growth. But, the evidence in this regard is also inconclusive. Prasad et al. (2003) argue that there is no relationship between financial integration and growth, and this association only occurs at higher levels of income and of institutional development, because at lower levels the incidence of financial crisis is more important.3

3 Should countries offer incentives to foreign direct investors? The existence of a positive externality from FDI to economic growth remains an unsettled issue, and therefore the policy implications are

José De Gregorio 189

not straightforward. However, even if there is a positive externality, as most cross-country studies and recent micro studies suggest, the implication that there should be special policies to foster foreign investment does not necessarily follow. In this section some problems are discussed that a policy of targeting FDI may have. The discussion is based on the experience with the treatment of FDI in the Chilean economy, but some general conclusions may be extracted from this case. My presumption is that a number of issues call into question the case for special incentives. Indeed, in order to justify special incentives, it is necessary not only to prove that FDI has a positive effect on growth, but also that it is possible to identify policies promoting FDI without inducing distortions that might offset the gains in growth. What is really needed to support a policy of special incentives for FDI is empirical evidence of a positive relationship between discriminatory policies in favour of FDI and economic growth. There are several important reasons to be cautious when considering policies that discriminate in favour of foreigners: ●

●

●

Discriminatory policies open the door to rent-seeking and economic distortions. They create a perverse incentive for domestic investors to seek out foreigners with whom to share the potential subsidies. Such policies also reduce incentives for local entrepreneurship. In summary, discrimination induces other distortions in the economy. Special treatment for some projects or sectors may reduce the net benefits from FDI. In attempting to foster particular sectors or specific investment projects, authorities may negotiate, on a case-by-case basis, special conditions for foreign investors. This is a risky business. In a competitive world, if many countries bid against each other to attract foreign investment from the same source they may end up dissipating all the potential gains from such an investment. Direct incentives to foreign investors generate unfair discrimination between domestic and foreign investors, and this raises serious political economy implications, especially in middle-income countries. Some affected groups will ask why foreigners should enjoy better treatment than domestic investors. One possible answer is that, until the country has built strong institutions and a reputation as a safe location for foreigners to do business, some form of discrimination may be advisable. For example, the country may need to offer some specific guarantees in terms of protection of property rights and rules of the game.

190 Role of FDI and Natural Resources in Economic Development

This has been the route followed by Chile in its efforts to sustain inflows of FDI. Under the 1974 Foreign Investment Statute (DL 600), later made part of the Constitution of 1980, foreign investors can sign a contract with the state of Chile guaranteeing them property rights and some stability in the rules of the game.4 Foreign investors accounting for 85 per cent of total FDI inflows into Chile since 1974 have taken advantage of this mechanism. The principal rules promoting stability are those ensuring full access to foreign exchange, something that traditionally has been heavily restricted in Chile. Foreign investors may also apply for a guarantee of tax invariability, although they will pay a higher tax rate than that currently imposed on profit remittances. These rules were applied in a period of great uncertainty about the protection of property rights and stability of the rules of the game. This is also reasonable in the context of a country that until recently imposed many restrictions on the foreign exchange market. However, there is no reason for Chile to go further in this direction, especially now that it has established its reputation and commitment to fair treatment of foreign investors on the basis of equal treatment of nationals and foreigners. Hence, if anything, the special rules applying to foreigners should be dismantled as time goes on. Only those rules that relate to the essence of being a foreign investor, such as those governing access to foreign exchange, should be kept in place.

4

The effects of natural resources on economic growth

Proposals aimed at attracting foreign investment often seek to promote certain specific sectors over others. In particular, at the time of writing there is a debate about whether the exploitation of natural resources is good for growth. This is a particularly important issue for a country such as Chile, where more than a third of foreign investment has been directed towards the mining sector. But it is an important topic for all countries interested in FDI. The role of natural resources in economic development touches on many issues, from FDI to the environment, and to the level and management of exchange rates. The issue I will address here is the fundamental one of whether natural resources are a blessing or a curse for developing countries. Recent empirical research has found a negative relationship between an abundance of natural resources and economic growth. Most notably, Sachs and Warner (1995) found such a negative relationship using

José De Gregorio 191

cross-country regressions. Their finding was robust to different measures of resource abundance, such as the share of mining production in GDP, land per capita, and the share of natural resource exports in GDP. They found that a one-standard-deviation increase in the share of natural resources exports in GDP was associated with a lower rate of growth in the order of 1 per cent a year. The claimed robustness of this finding has been subjected to severe scrutiny. For example, it has been noted that certain high-income countries, such as Finland, Sweden, Canada, the USA and Australia, continue to rely on their abundance of natural resources, and thus provide some strong historical counter-examples. Lederman and Maloney (2002) re-examined the econometric analysis of Sachs and Warner (both their original 1995 paper and subsequent papers) and argued that, after controlling for omitted variables and endogeneity problems, their finding does not hold. What Sachs and Warner’s natural resource variables were really capturing, according to Lederman and Maloney (2002), was the negative correlation between export concentration and inter-industry trade, and the positive correlation between such concentration and real exchange-rate volatility. Although the empirical evidence may thus be inconclusive, as it was in the case of FDI, it is useful to review the channels through which having natural resources might be detrimental for growth, and what the policy implications might be. There are two interesting hypotheses. The first, formalized by Lane and Tornell (1996), is rooted in political economy and invokes a ‘voracity effect’. Different interest groups, it is argued, fight to capture the rents from natural resources, and this induces a poor allocation of resources, tilted towards rent-seeking and inducing inefficient taxation. The other hypothesis is that an economy might have a limited endowment of some key factor – for example, the capital stock. In such a situation, the exploitation of natural resources, which provides high rents but probably makes a small contribution to overall productivity growth, may also shift the allocation of capital away from growth-enhancing activities. If the problems with natural resources development stem from a voracity effect, the solution is not to forbid their exploitation but rather to strengthen institutions. After all, there will always be some source of rents, but a healthy institutional framework, one that leads to sound fiscal policy, taxation, regulation and decision-making at the government level, may prevent the capture of economic policy by interest groups that do not properly represent the interests of society. In contrast, if the problem is that the exploitation of natural resources uses

192 Role of FDI and Natural Resources in Economic Development

scarce resources, such as physical capital, then the best solution may be to open the capital account and allow FDI in the natural resources sector. In a world of free capital mobility, the lack of domestic capital could be overcome by allowing the inflows of foreign capital. Strengthening institutions to avoid voracity effects, although necessary, may not be enough. The abundance of natural resources may create a ‘Dutch disease’, by which the exchange rate might appreciate and reduce profitability in other sectors (Sachs and Warner, 1995). In addition, the existence of a booming natural resource sector might create windfall gains in public finances, which may be used to subsidize current consumption rather than for investment. On the other hand, opening the capital account to increase the availability of capital may be not enough to increase growth if the scarce resource in developing countries is human capital and not necessarily physical capital. As argued by Bravo-Ortega and De Gregorio (2002), if the scarce factor is human capital, opening the capital account does not solve the problem, because human capital is essentially not mobile across countries. This point is made using a two-sector model. The model assumes that production in one sector, natural resources, is subject to decreasing returns to scale, whereas the other sector, manufacturing, is subject to decreasing returns to scale at the firm level, but there is an externality that leads to aggregate constant returns to scale and, thus to permanent growth. The rate of growth of the economy is the weighted average of the rates of growth of the two sectors. If human capital is scarce, the exploitation of natural resources, which is the sector that enjoys rents but not dynamism to increase the rate of growth, could ultimately reduce growth because it demands a limited amount of human capital. In contrast, if human capital is abundant, there is no such crowding out. Thus the model captures the idea that natural resources limit growth as long as the level of human capital is low, leaving insufficient resources to devote to growth-enhancing activities. The implication of the model is that countries with low levels of human capital may reduce their rate of growth by exploiting natural resources. Indeed, the empirical evidence presented in the Bravo-Ortega and De Gregorio paper shows that there is an interaction effect between human capital and the abundance of natural resources, and that there is a threshold level of human capital beyond which natural resources are good for growth. Nevertheless, this discussion misses the most important point: what is good for an economy’s welfare is not necessarily its rate of growth of output, but rather the level of output. It is easy to imagine an economy where the discovery of natural resources may lead to a decline in growth

José De Gregorio 193

but an increase in income that ultimately raises welfare. On the other hand, it is difficult to imagine that a country could be better off by giving away its natural resources, as could be wrongly implied from the result of Sachs and Warner (1995). The empirical evidence regarding the level of income, rather than the rate of growth, shows that the richer a country is in natural resources, the greater its income. Maloney (2002) has made a similar point regarding the innovative capacity of a country, and the benefits it can extract from the exploitation of natural resources. Reviewing the historical evidence, he argues that high investment in human capital and scientific infrastructure helps countries both to take greater advantage of technological advances abroad and to increase productivity growth in the natural resources sector. Maloney also argues that there is no reason to think that natural resource industries have inherently slower productivity growth than other sectors, such as industry. The high starting levels of human capital in the Scandinavian countries and in other high-income countries with a broad base of natural resources help to explain why the exploitation of these resources was not detrimental for the countries’ growth. Moreover, even the argument that natural-resource industries display slow productivity growth compared to manufacturing has recently been questioned by Martin and Mitra (2001). They found, in a large sample of developing and developed countries, that technical progress has been faster in agriculture than in manufacturing, weakening further the argument that governments should use discriminatory policy to shift the economy away from natural resources to industry. As in the case of FDI, there is strong evidence that having a well-educated labour force is the key to growth and, specifically, the key to taking full advantage of both FDI and natural resources.

5

Concluding remarks

The majority of studies show that foreign direct investment is beneficial for economic growth, or at least that good macroeconomic performance and FDI move together. The most recent micro evidence also shows the presence of significant spillovers into local supplier sectors other than that in which multinationals produce. But, there is no solid basis for arguing that any one sector should be promoted over any other, and therefore, as a general guideline, inducements to FDI should not be made on a discriminatory sectoral basis without a clear rationale. The issue of industrial policy goes beyond the purpose of this chapter, but there are no strong reasons to argue that developing natural resources rather than

194 Role of FDI and Natural Resources in Economic Development

manufacturing may be detrimental to growth. However, as this chapter has emphasized, taking full advantage of the benefits of FDI requires a well-educated labour force to promote technological diffusion and the adoption of better technologies. The same applies to the development of natural resources: here too it is beneficial to have a high level of human capital. This prevents crowding out among different activities, and it allows for innovation to take place, starting in the natural resources sector and spreading downstream or to other sectors. The experience of the Nordic countries shows how prosperity is in no way inconsistent with the good fortune of being well-endowed with natural resources. Also important is the principle of national treatment. There is no reason for countries to discriminate between local and foreign investors, and perhaps one of the most important rules for all countries is that foreigners should be treated in the same way as locals. However, in countries making the transition towards being a reputable recipient of FDI, it may be necessary to provide some guarantees to foreigners, in particular those that apply uniquely to foreign investors, such as access to foreign exchange to remit profits and capital. As countries develop, it is to be expected that capital flows will take different forms, and perhaps the share, although not the magnitude, of FDI may decline. For this reason, policy-makers must look at FDI in the broad context of capital flows, and realize that trying to force these flows to take one form or another may create artificial distortions. Of course, and for good reasons, countries may worry about the maturity of capital inflows and FDI, but this does not imply that, for flows of a given maturity, there should be some preference for one form over another. Each type of capital inflow serves a different purpose from the point of view of investors, and the choice depends to a large extent on the institutional framework of the economy. In addition, if the concerns are with the stability of the economy and capital flows, the best instrument is a strong system of prudential regulation of the financial system. An economy with strong institutions is the best incentive for FDI, as this is an effective way to guarantee the protection of property rights. A recent World Bank survey of 191 MNEs, including thirty of the world’s 100 biggest, found that almost 80 per cent planned to expand abroad in order to create new operations or to buy existing companies (see Table 9.2): these companies use FDI mainly to gain access to markets. This would put small developing countries at a disadvantage in attracting FDI. But with greater openness, and progress in transportation and communications, this disadvantage could be overcome, since the second most important reason for locating abroad is to reduce costs.

José De Gregorio 195 Table 9.2

World Bank FDI survey: expansion strategies

Locating new facilities outside home country in next 3 years? Yes No

79% 21%

Expansion strategy in foreign country? Expand an existing company facility Build or lease a facility M&A

15% 43% 42%

Source:

World Bank (2002).

Table 9.3 World Bank FDI survey: factors influencing location (percentages) 1st

2nd

First and second most important objectives for investing abroad Improved market access 55 15 Reduce operating costs 17 31 Other factors 8 11 Source raw materials 6 0 Consolidate operations 4 16 Developed new product lines 4 11 Improved productivity 2 0 Developed new technologies 2 0 Improved labour force access 1 11 Reduce risk 1 5 Source:

World Bank (2002).

As the survey also shows (see Table 9.3), what matters for the choice of location is not special incentives (for example, lower taxes is not among the most important factors), but rather strength of institutions. Access to customers is naturally the most important factor of all. But the stability of the social and political environment and ease of doing business follow close behind. Countries can benefit greatly from FDI. To attract it, and to take full advantage of it when it comes, growth-promoting institutions are the essential ingredient. Notes 1 For further analysis of FDI among developed countries, see Lipsey (2000). 2 Goldberg (2004) discusses the evidence of FDI in the financial sector, arguing that it has important benefits regarding allocative efficiency through increased

196 Role of FDI and Natural Resources in Economic Development competition. She also compares the findings for the financial sectors with those in other sectors. 3 The evidence regarding international integration is more conclusive for the case of trade openness; for a recent discussion, see Dollar and Kray (2004). 4 For further details, see Chile Foreign Investment Committee (2002).

References Aitken, B. and A. Harrison (1999) ‘Do Domestic Firms Benefit from Foreign Investment? Evidence from Venezuela’, American Economic Review, vol. 89(3), pp. 605–18. Balasubramanyam, V. N., M. Salisu and D. Sapsford (1999) ‘Foreign Direct Investment as an Engine of Growth’, Journal of International Trade and Economic Development, vol. 8(1), pp. 27–40. Blalock, G. and P. Gertler (2004) ‘Welfare Gains from Foreign Direct Investment through Technology Transfer to Local Suppliers’, mimeo, Cornell University. Blomstrom, M., R. Lipsey and M. Zejan (1992) ‘What Explains Growth in Developing Countries?’, NBER Working Paper no. 4132. Borensztein, E., J. De Gregorio and J.-W. Lee (1998) ‘How Does Foreign Direct Investment Affect Economic Growth?’, Journal of International Economics, vol. 45(1), pp. 115–35. Bravo-Ortega, C. and José De Gregorio (2002) ‘The Relative Richness of the Poor? Natural Resources, Human Capital and Economic Growth’, Working Paper no. 139, Central Bank of Chile. Carkovic, M. and R. Levine (2005) ‘Does Foreign Direct Investment Accelerate Economic Growth?’, in T. H. Moran, M. Blomstrom and E. Graham (eds), Does FDI Promote Development (Washington DC: Institute for International Economics), pp. 195–220. Chile Foreign Investment Committee (2002) ‘Chile’s FDI Policy: Past Experience and Future Challenges’, DAFFE/IME/RD (2002), vol. 4 (Paris: OECD). De Gregorio, J. (1992) ‘Economic Growth in Latin America’, Journal of Development Economics, vol. 39(1), pp. 58–84. Desai, M. A., C. F. Foley and J. R. Hines, Jr. (2002) ‘Capital Controls, Liberalizations, and Foreign Direct Investment’, mimeo, Harvard University. Dollar, D. and A. Kray (2004) ‘Trade, Growth and Poverty’, Economic Journal, vol. 114, pp. F22–F49. Goldberg, L. (2004) ‘Financial-Sector FDI and Host Countries: New and Old Lessons’, NBER Working Paper no. 10441. Hausmann, R. and E. Fernandez-Arias (2000) ‘Foreign Direct Investment: Good Cholesterol?’, Working Paper no. 417, Research Department, Inter-American Development Bank. Lane, P. and A. Tornell (1996) ‘Power, Growth and Voracity Effect’, Journal of Economic Growth, vol. 1(2), pp. 213–41. Lederman, D. and W. Maloney (2002) ‘Open Questions about the Link between Natural Resources and Economic Growth: Sachs and Warner Revisited’, mimeo, World Bank. Lipsey, R. E. (2000) ‘Interpreting Developed Countries’ Foreign Direct Investment’, NBER Working Paper no. 7810.

José De Gregorio 197 Maloney, W. (2002) ‘Missed Opportunities: Innovation and Resource-Based Growth in Latin America’, mimeo, World Bank. Martin, W. and D. Mitra (2001) ‘Productivity Growth and Convergence in Agriculture versus Manufacturing’, Economic Development and Cultural Change, vol. 49(2), pp. 403–22. Prasad, E. S., K. Rogoff, S. J. Wei and A. Kose (2003) ‘Effects of Financial Globalization on Developing Countries: Some Empirical Evidence’, IMF Occasional Paper no. 220 (Washington, DC: International Monetary Fund). Sachs, J. and A. Warner (1995) ‘Natural Resource Abundance and Economic Growth’, NBER Working Paper no. 5398. UNCTAD (1999) World Investment Report (New York: United Nations). UNCTAD (2001) World Investment Report (New York: United Nations). World Bank (2002) Foreign Direct Investment Survey (Washington, DC: The World Bank).

10 Multinationals and Foreign Direct Investment in India and China∗ Subrata Gupta Jogamaya Devi College, Calcutta University, India

1

Introduction

This chapter seeks to focus on the role of foreign direct investment (FDI) in the economy since economic liberalization in 1991, and contends that the inflows of FDI so far in India have been disappointingly low. After considering a short historical perspective of India’s policy towards FDI, this chapter considers the structure and industrial breakdown of the inflows of FDI. An important element in the published discussion of the impact of FDI on the Indian economy is a comparison with China. China’s success in absorbing and utilizing FDI inflows in the post-reform period since 1978 has been contrasted with the corresponding failure of India: given its potential, India has attracted much less FDI than it should have attracted. While the Indian record can be explained in part as the outcome of certain mistaken policy prescriptions, it is amazing how a communist country such as China could free itself from its dogmatic barriers and attract FDI to such benefit.

2

The evolution of Indian policy towards FDI

India’s policy towards FDI was initially announced in the Industrial Policy Statement of 1948, soon after attaining independence – to invite ∗

I am grateful to Professor Edward Graham of the Institute for International Economics, Washington, DC, for his comments on an earlier version of this chapter. 198

Subrata Gupta 199

foreign capital subject to certain provisions. The period 1948–67 was one of import substitution and a receptive attitude to FDI if on mutually advantageous terms, and with majority local ownership. Foreign investors were assured of remittance of profits and dividends. But despite the abundance of cheap labour in the country and a growing market for consumer goods, few foreign investors were tempted, some for fear that foreign firms might ultimately be nationalized. During its Second Five-Year Plan, India benefited from the installation of three steel plants – at Durgapur, Bhilai and Rourkela, with assistance from the governments of the United Kingdom, the USSR and the Federal Republic of Germany, respectively. But, generally, production from foreign investments was small. The first Export Processing Zone (EPZ), in which foreign investors were accorded tax concessions, was set up at Kandla in 1965. From 1968 to 1979, India had a restrictive attitude to FDI on the grounds of protecting the domestic base of ‘created assets’. The government nationalized some major oil-producing and retailing multinational corporations in the early 1970s, and government listed industries in which FDI was not desirable, to protect local companies. In 1973, the Foreign Exchange Regulation Act (FERA) came into force and all foreign companies operating in India were required to register under Indian corporate legislation, with up to 40 per cent equity. Exemptions from the general 40 per cent were made only for companies operating in hightech sectors or sectors with the potential for good export performance. The second EPZ where the foreign investors were given some incentives with regard to taxation was set up in 1972 at Santacruz. A process of cautious deregulation started during the 1980s. It was realized that the international competitiveness of Indian goods had suffered from growing technological obsolescence and inferior product quality, a limited range, and high cost because of a highly protected local market. But the period of cautious deregulation of foreign investment could not on its own succeed in attracting much FDI. However, the liberalization of industrial and trade policies initiated in the 1980s was accompanied by an increasingly receptive attitude towards FDI and foreign licensing collaborations. On 24 July 1991, with the announcement of the New Industrial Policy, India fully liberalized its economy and became completely open to FDI. In terms of this policy, the industrial approval system in all industries had been abolished, apart from eighteen strategic or environmentally sensitive industries. In thirty-four high-priority industries, FDI of up to 51 per cent was approved automatically. As a part of economic reforms initiated in 1991, tariffs were reduced sharply on more products, the

200 Multinationals and FDI in India and China

average weighted tariff rate on imports being reduced from 87 per cent in 1990–1 to 25 per cent by 1994–5, and 20 per cent by 1997–8. The foreign technology requirements for inviting FDI were discontinued. The FERA was amended, and restrictions imposed on foreign companies by FERA were also withdrawn. Four more EPZs were created in addition to those at Kandla and Santacruz. New EPZs were created at Kochin, Chennai, Noida (in Uttar Pradesh) and Falta (in West Bengal). In 2001, two Special Economic Zones were created. In 1994, India became a member of the Multilateral Investment Guarantee Agency (MIGA) following which all investments approved by the Government of India are insured against expropriation and nationalization. Up to 100 per cent equity was allowed on export-orientated and high-tech industries. The Reserve Bank of India (RBI) can now extend automatic approval to proposals of high priority and high-tech areas where foreign equity does not exceed 51 per cent, and in the mining sector where the foreign equity does not exceed 50 per cent. A Foreign Investment Promotion Board (FIPB) has been set up to look after the inflows of FDI in cases where foreign equity exceeds 51 per cent and the industry is not on the list of the high priority sectors, or where it does not cover import of capital goods. The Secretariat of Industrial Approvals (SIA) of the Ministry of Industry also gives approval to foreign investment in certain cases. The overview of the evolution of FDI policy pursued in India indicates that there exists some confusion about the economic impact of foreign direct investment on the Indian economy.

3

FDI inflows into India

Table 10.1 displays the total number of foreign collaboration approvals involving foreign investment, and Table 10.2 shows the aggregate value of inflows, each from the economic liberalization in August 1991 to March 2003. Throughout the 1990s, the actual inflow of FDI has been much less than the approvals in each year.

4

Inflows of FDI

FDI inflows in India are recorded under five broad headings: (i) RBI’s automatic approval route for equity holdings up to 51 per cent; (ii) FIPB’s discretionary approval route for larger projects with more than 51 per cent equity holding; (iii) the acquisition of shares route (since 1996); (iv) RBI’s non-resident Indian (NRI) schemes; and (v) external

Subrata Gupta 201 Table 10.1 Foreign collaboration approvals, August 1991–March 2003

No. of foreign collaborations No. of approvals Foreign investment involved (billions of rupees)

SIA

RBI

FIBP

Total

4 475

8 570

10 832a

23 877

1 583 58.5

4 142 214.8

10 620 2 582.8

16 345 2 856.1b

Notes: a The approvals by FIPB include 98 proposals for American Depository Receipts (ADRs), Global Depository Receipts (GDRs) or Foreign Currency Convertible Bonds (FCCBs) involving investment of 484.71 billion rupees. b US$73.70 billion. Sources: SIA Newsletter April 2003; Economic and Political Weekly (Mumbai), 25 October 2003, p. 4199.

Table 10.2 Actual inflow of foreign direct investment (FDI) and non-resident Indians (NRI), August 1991–March 2003 (Rs crore) Governmental approval RBI approval NRI scheme ADRs/GDRs/FCCBs Amount of inflows on acquisition of shares

61 273 13 545 8 359 25 629 18 433

Sub-total Closing balance of advance Stock swapped

127 239 6 910 84

Total

134 233a

Note:

a

US$ 33327.60 million.

Sources: SIA Newsletter April 2003; Economic and Political Weekly (Mumbai) 25 October 2003, p. 4194.

commercial borrowings (ADR/GDR/FCCB route); the Indian definition of FDI differs from that of the IMF and the World Investment Report of the UN. The IMF definition includes external commercial borrowings, reinvested earnings and subordinated debt, while the World Investment Report excludes external commercial borrowing. The World Investment Report, 2001 (UNCTAD, 2001) defines the flow of FDI as comprising capital provided (either directly or through other enterprises) by a foreign direct investor to an FDI enterprise, or capital received from an FDI enterprise by a foreign direct investor. There are three components of FDI: equity capital, reinvested earnings, and intra-company loans.

202 Multinationals and FDI in India and China

Total inflows of FDI in India appear to have been undervalued, since FDI inflows are entirely evaluated on equity investments while ignoring other components. In China, reinvested earnings and intra-company loans together accounted for about 30 per cent of total FDI inflows during 1997, and about 51 per cent of total FDI inflows in 1998 (UNCTAD, 2001, p. 113). It has been argued that the official FDI figures for China may be somewhat inflated compared to those in India, and need to be interpreted with caution (Srivastava, 2003). In China, while Hong Kong has been a major direct investor, part of the investments may be a result of ‘round-tripping’ from the mainland as domestic (Chinese) investors try to take advantage of tax and tariff benefits extended to foreign investors (Graham and Wada, 2001). It is very difficult to explain the causes of the major gaps between intended flow and actual flow of FDI into India. The actual inflow of FDI increased gradually after 1993 and continued until 1997. After a decline in 1998, it rose again in 1999 and 2000. In terms of the percentage of GDP (see Table 10.3), FDI net inflows in India indicate a poor picture at only 2.60 per cent, even less than those in Sri Lanka, Pakistan and Bangladesh. It should be noted that FDI as percentage of GDP rose from 0.03 per cent of GDP in Bangladesh in 1995 to 3.25 per cent in 1998; it also rose in Sri Lanka and China, where the share increased to 11.91 per cent of GDP in 1998. UNCTAD (2001, p. 23) provided data on FDI inflows into thirty developing countries during the period 1993–8: of the total FDI inflows to these countries, 25.7 per cent went to China,

Table 10.3 Net FDI inflows as percentage of GDP in selected Asian countries, 1990–8 Countries

1990

1995

1998

Malaysia Thailand Indonesia China Philippines Vietnam Sri Lanka Pakistan Bangladesh India

16.25 6.92 3.10 2.83 4.95 1.91 2.41 3.22 0.05 0.20

10.87 2.96 6.74 12.54 8.88 25.48 1.67 6.53 0.03 2.30

25.85 24.69 2.71 11.91 12.81 15.38 4.84 4.61 3.25 2.60

Source:

World Development Indicators (World Bank).

Subrata Gupta 203 Table 10.4

FDI inflows to India by country, August 1991–March 2003

Country by rank

USA Mauritius UK Japan South Korea Germany Netherlands Australia France Malaysia Singapore Italy Belgium Israel Cayman Islands Switzerland Canada Thailand Hong Kong Sweden NRIs Euro Issues/ADRs/ GDRs/FCCBs All others Total

Investment (millions of rupees)

Percentage of total

574 340 343 500 232 240 114 240 98 270 92 420 89 510 67 810 65 580 60 570 52 740 48 050 45 110 42 450 38 870 31 290 28 760 24 610 23 020 20 600 106 510 484 710

20.1 12.0 8.1 4.0 3.4 3.2 3.1 2.4 2.3 2.1 1.8 1.7 1.6 1.5 1.4 1.1 1.0 0.9 0.8 0.7 3.7 17.0

171 000 2 856 200

6.1 100.0

Source: Ministry of Commerce and Industry, Government of India, SIA Newsletter, April 2003.

which ranked number one. India ranked 17th, with only 1.4 per cent of the FDI inflows. The USA, the largest investing country, accounting for 20.1 per cent (in the period covered by Table 10.4), is followed by Mauritius, a conduit for investors from many states (including the USA) under a treaty that allows Mauritius-based companies exemption from taxes. The United Kingdom is the next largest investor, followed by Japan, South Korea (which has emerged as a major investor in automobiles and consumer durables since 1995), and Germany. Malaysia has also been a significant investor since 1995; other significant investors include the Netherlands, France, Australia, Singapore and Italy.

204 Multinationals and FDI in India and China Table 10.5 Foreign collaboration approvals by industry, August 1991–March 2003 Industrial group

Number of approvals Technical Financial

Amount of FDI Percentage of approved total (billions of rupees)

Basic goods Ferrous metals, non-ferrous metals, mining, fuel and power, fertilisers, chemicals, oil refining, etc.

1 559

2 113

1084.18

38.0

Capital goods Electrical equipment, electronics, boilers and steam plants, industrial machinery, transportation equipment, industrial instruments, misc. engineering, medical appliances, etc.

3 352

3 621

261.84

9.2

Intermediate goods Rubber goods, glass, ceramics, leather and leather goods, etc.

262

608

53.29

1.8

1 443

3 297

298.47

10.5

42

136

94.36

3.3

635

6485

1 063.10

37.2

7 293

16 260

2 855.24

100.00

Consumer non-durables Textiles, paper, foodstuffs, agro-industrial goods, etc. Consumer durables Passenger cars, commercial office and household equipment, refrigerators, televisions, etc. Services Computer software, telecommunications, air/sea transport, financial services, hotels and tourism, hospitals and diagnostic centres, etc. Totals

Sources: Ministry of Commerce and Industry, Government of India, SIA Newsletter, April 2003; Economic and Political Weekly (Mumbai), 25 October 2003.

As Table 10.5 shows, the bulk of FDI went into basic goods and into services – these accounted for 38.0 per cent and 37.2 per cent, respectively, of the total. Power generation, oil refineries, fertilizers and chemicals, telecommunications, computer software and financial services experienced substantial inflows, as did food products and agro-industries.

Subrata Gupta 205

5

The economic impact of FDI

FDI inflows constitute a source of resources for investment and help to bridge the savings–investment gap in the host country; it also provides foreign exchange to help close the trade gap, as well as contributing to the raising of technological standards, efficiency in business management and export competitiveness. All these should exert a favourable impact on the recipient country’s GDP and on its openness to the world economy, as reflected in its ratio of India’s foreign trade turnover to GDP. That indicator rose from 19.53 in 1992 to 24.78 per cent in 1998, the ratio of exports to GDP increasing from 8.94 in 1992 to 11.03 per cent, and that of imports to GDP from 10.59 to 13.75 per cent (World Development Indicators database; Nayar, 2001). The impact of FDI inflows may not be wholly favourable. An improvement in the terms of trade, under the impact of the trans-nationalization of production, may promote investment by increasing real income, making capital goods (mostly imported in developing countries) cheaper to the domestic industries. But it could also decrease investment by decreasing the demand for domestic goods compared to imported goods (Cardoso, 1993). In India, despite access to large, long-term resources, foreign firms’ share in fixed asset formation in the corporate sector remained a meagre 10 per cent in the 1990s and, compared to domestic firms, foreign firms used a smaller share of their investible resources in physical investment during the period 1992–7, the ratio of gross fixed assets to total uses of fund for the foreign private sector being lower than that for the Indian private sector by about 13 per cent (CMIE, 1997). FDI is hence not always positively associated with the growth of investment. If capital inflows in the form of FDI are large, the real exchange rate may appreciate, leaving the nominal exchange rate constant. As a result of the absorption of the inflows into reserves, there may be an expansion of money supply leading to a rise in prices without a corresponding increase in output, thus leading to an appreciation of the real exchange rate (Rangarajan, 2000). An increase in the real exchange rate would increase the price of imported capital and intermediate goods, and result in a contraction of investment (Serven and Solimano, 1992; Fry, 1995). But in that case the contraction in investment might take place in domestic goods only, and investment in tradable goods might actually increase (Van Wijnbergen, 1982). One of the most important determinants of domestic investment is the availability of institutional credit in developing countries

206 Multinationals and FDI in India and China

(see Blinder and Stiglitz, 1983; Fry, 1995). The cheap money policy pursued by the banking system in India aims at raising domestic investment, but there is no guarantee that inflows of FDI would always have a positive impact on domestic investment. It is possible that excessive concessions on taxes and inflows of FDI to some of the more vulnerable or noncompetitive sectors of the economy might have a negative effect on the economy in future (Agrawal, Chapter 5 in this volume). It is a fact that in a labour-surplus, capital-scarce economy, FDI could promote GDP growth by increasing domestic investment and by generating employment opportunities for surplus labour, particularly when there is a migration of surplus labour to foreign enclaves (that is, EPZs and SEZs) and by improving technological skills and human capital. Again, inflows of FDI may also lead to immiserizing growth when such inflows of FDI result in excessive profits in the domestic economy if the economy is subject to trade and financial distortion (Brecher and Diaz Alejandro, 1977). With the inflow of FDI, multinational enterprises (MNEs) are expected to develop substantial R&D in the host country, which can upgrade production standards and managerial skills, but the spillovers found in advanced countries may not be found in India (Kumar and Siddhartan, 1997). Multinationals as a group, as demonstrated by several studies, behave differently from non-affiliated local firms (Kumar, 1990; Dunning, 1993; Caves, 1996). Compared to local unaffiliated firms, MNEs enjoy a number of advantages with respect to brand names, goodwill, access to global technology, and market information. But so far as India is concerned, R&D activities have yet to produce a favourable impact on domestic investment. Agrawal (Chapter 5 in this volume) has estimated that the impact of FDI on GDP growth was positive in India during 1980–96, and particularly significant over 1990–6. He also found that, since 1980, FDI inflows have contributed more to investment and also to GDP growth in South Asia (which includes India) than an equal amount of foreign borrowing. This is an advantage that FDI demonstrates over foreign borrowing. MNEs create employment opportunities through the outsourcing of work to reduce costs and improve the return on investment–producing labour-intensive products where the labour cost is lower. The emergence of China as a major economic power is largely dependent on such outsourcing, and India has benefited from outsourcing by MNEs, particularly in the sectors of communication and information technology, and to some extent in textiles. When outsourcing involves the sending of labour-intensive work, whether in manufacturing or in services,

Subrata Gupta 207

to another country, it is termed ‘offshoring’. But it might not benefit India permanently if MNEs keep wages low and might arouse popular discontent in the home country.1

6

Comparing India and China in FDI inflows

Prior to its 1978 Law on Foreign Joint Ventures, China was, for all practical purposes, closed to foreign direct investment. That law permitted foreigners to enter into contracts for a joint venture (JV) with a Chinese firm (at that time generally a state-owned enterprise) subject to approval by the central government (a requirement eased in 1986), in order to bring into China technologies that were not to be found in the state enterprise. Foreign investment was allowed only in selected branches and was slow to enter until the government’s administration of JVs became more transparent in 1983 and the law was liberalized further in 1985. From 1995, China moved into the first rank among developing economies as recipient of FDI, as Table 10.6 shows, the actual inflow in 1998 more than quadrupling that of 1992. Four Special Economic Zones (SEZs) were created in China in 1980, three in Guangdong province (in the cities of Shanton, Shenzhen and Zuhai) and one in Fujian Province (in Xiamen); a fifth SEZ was set up in Hainan Island in 1988. In addition, fourteen coastal cities were designated in 1984 as Economic and Technical Development Zones (ETDZs), which in practice differed very little from SEZs. All such zones are in coastal areas, with a well-developed infrastructure (including access to deep-water ports and airports) and a supply of good human capital. Foreign investment enterprises (FIEs) operating through the SEZs have also played a significant part in expanding Chinese foreign trade, as Table 10.7 indicates. Economic reform in China is continuing at the time of writing and the expansion of foreign trade as a whole and of the performance of FIEs in China has continued since 1998. Most of the increase in FDI may be attributed to the successful operation of SEZs, which Table 10.6 Inflow of FDI into China, 1988–98 (US$ billions)

Actual transfers

1988

1992

1993

1994

1995

1996

1997

1998

3.2

11.2

27.3

33.8

35.9

40.2

44.3

45.6

Sources: Statistical Year Book of China 2000; Economic and Political Weekly (Mumbai), 6 November 1999; Chandra (1999).

208 Multinationals and FDI in India and China Table 10.7 Chinese foreign trade and FIEs, 1985–98 (US$ billions) Year

1985 1990 1995 1996 1997 1998 Sources:

Total

Of which, through FIEs

Exports

Imports

Exports

Imports

27.4 62.1 148.8 151.1 182.7 183.8

42.3 53.4 132.1 138.8 142.4 140.2

0.3 7.8 46.9 61.5 74.9 81.0

2.1 12.3 62.9 75.6 77.7 76.7

Up to 1996, Sun (1998); 1997 and 1998, Chandra (1999).

facilitated export-orientated ventures. It may be noted that China created SEZs long after India had created its EPZs in 1965 and 1972. In India, the inflow of FDI was 6.75 billion rupees in 1992, rising to 164.25 billion rupees in 1997, 193.42 billion rupees in 2000 and 192.85 billion rupees in 2001 – that is, less than $US5 billion (Economic Survey, 2003), whereas in China FDI inflows exceeded $US45 billion in 2001 (UNCTAD, 2001). Many reasons may be cited for India’s failure to utilize its potential to attract FDI. The first difference with China is that the creation of EPZs in India provided no incentive for the governments of the various states to take measures to develop the zones in their regions because the EPSs were under the full control of central government. A second difference is that all foreign direct investments in India are routed through the central government or the FIPB and the RBI, thus alienating state governments because of centralized bureaucratic control. More recently, central government has allowed state governments the power to negotiate directly with foreign investors, to the benefit of states such as Karnataka, Andhrapradesh and Maharashtra. A third factor has been lack of adequate economic infrastructure, leading to bottlenecks in the transport network and power supply. Fourth, militant trade unionism and labour indiscipline (particularly in West Bengal) and the lack of a developed work culture have in some cases dissuaded foreign investors. Fifth, breakdowns of law and order in some areas (as in Bihar) have not been conducive to foreign direct investment. The constraints cited above are responsible for India failing to realize fully its potential to attract foreign direct investment, in contrast with the situation in China. Whereas in China the population at large have

Subrata Gupta 209

co-operated with foreign-owned firms in the interests of industrialization and export promotion, a large section of the Indian population has always been sceptical about foreign investment, believing that it would substitute capital for jobs. However, the increasing flow of FDI in India shows that there has been an erosion of such ideas. The SEZs along China’s coastline were designed to give foreign investors and domestic enterprises favourable conditions for export promotion, such as the duty-free import of intermediate and capital goods, tax holidays, and the provision of land, power, physical security, transport and specially created industrial parks. In contrast, India’s approach to export processing zones has been one of relative neglect rather than support. One major reason for the failure of the Indian EPZs to perform as well as China’s SEZs is their inefficient logistical links with ports, poor infrastructure in the areas surrounding the zones, physical insecurity, and non-transparency in the design, setting-up and functioning of the zones (Bajpai and Sachs, 2000). In China, although new sources of employment were generated and exports rapidly promoted in the SEZs and SEZs-like zones, the benefits were to some extent limited by the scarcity of spillovers, especially during the 1980s, from foreign-invested firms.2 China has an abundant supply of disciplined cheap labour and, as such, most of the operations of the foreign-invested firms in these zones were concentrated in relatively labour-intensive and technology non-intensive activities (Chadha, 2000). As noted above, state governments in India have sometimes been averse to the EPZs regulated wholly by the central government, unlike the case in China, where the local and provincial governments are given responsibility. A Trade Policy Statement issued by the Government of India on 31 March 2002 proposed a relaxation of such central regulation: ‘a Special Economic Zone is a specifically delineated duty-free enclave and shall be deemed to be foreign territory for the purposes of trade organisations and duties and tariffs’. The difference between the earlier EPZs and the later SEZs is that, while in an EPZ customs permission is necessary to take raw materials from one place to another, this is not needed in an SEZ. Various restrictions were relaxed and infrastructural inadequacies are being rectified. In considering the impact of foreign investment on a country’s overall growth, Chinese development may be contrasted with that of Brazil. Brazil has attracted massive foreign investment since 1994, but much of it has been portfolio acquisition of privatized assets: neither Brazil’s growth nor its export performance has improved in subsequent years (Nagaraj, 2003).

210 Multinationals and FDI in India and China

7

Concluding observations

Between 1968 and 1979, the government’s policy of import substitution, its restrictive attitude to protect the domestic capital base, and exchange control hindered the inflow of FDI into India. A process of cautious deregulation of foreign investment, started in 1980s, failed to attract even a moderate inflow. It was only after economic liberalization in 1991 that FDI became substantial, albeit lagging far behind the contemporary inflow into China. China opened its economy to FDI after 1978, establishing Special Economic Zones and SEZ-like areas, with dynamic results. It appears that India has undertaken insufficient measures to attract foreign investment. Financial incentives are not lacking, but an inadequate infrastructure, bureaucratic delays, rigid industrial labour laws and in some cases militant trade unionism have been obstacles. India’s predominantly agrarian economy, with land productivity only a third of China’s and hence with consumer purchasing power too low for the high-cost products of foreign firms is a major reason why India’s potentialities in attracting FDI have not been fully realized. Indian government policy to improve the economic infrastructure for investment generally, and for FDI in particular, has not yet yielded substantial results (one contributory factor being that the rupee cost of power supply by foreign firms is too high for Indian consumers). On the other hand, FDI in the consumer-goods industries has increased domestic competition, and resulted in greater choice and quality improvement. The lesson is that India should liberalize further, introduce labour market reforms in the industrial sector, and invest more in infrastructure. Such a package would greatly enhance the inflow of FDI into Indian industry. Notes 1 The US Senate has voted to ban Federal contractors from using taxpayers’ money to move American jobs offshore. 2 Graham (2004), who used the term ‘foreign-invested firm’ because local affiliates of foreign investors are so called in China, even if they are majorityor wholly-owned by the foreign investor; the term reflects the fact that, at the outset, operations with foreign investment had to be in the form of joint ventures with locally-controlled firms, a requirement that was eased in 1986.

References Bajpai, Nirupam and Jeffrey D. Sachs (2000) ‘India’s Decade of Development’, Economic and Political Weekly (Mumbai), 15 April. Blinder, Alan S. and J. E. Stiglitz (1983) ‘Money, Credit Constraints and Economic Activity’, American Economic Review, vol. 73, pp. 277–302.

Subrata Gupta 211 Brecher, R. A. and C. F. Diaz-Alejandro (1977) ‘Tariffs, Foreign Capital and Immiserizing Growth’, Journal of International Economics, vol. 7, pp. 317–22. Cardoso, Eliana (1993) ‘Private Investment in Latin America’, Economic Development and Cultural Change, vol. 41, pp. 833–48. Caves, Richard E. (1996) Multinational Enterprise and Economic Analysis (Cambridge: Cambridge University Press). Chadha, Rajesh (2000) ‘ExIm Policy Changes’, Economic and Political Weekly (Mumbai), vol. 35(16), (15 April) pp. 1343–6. Chandra, Nirmal Kumar (1999) ‘FDI and the Domestic Economy: Neoliberalism in China’, Economic and Political Weekly (Mumbai), 6 November. CMIE (1997) Report (New Delhi/Mumbai: Centre for Monitoring the Indian Economy). Dunning, J. H. (1993) Multinational Enterprises and the Global Economy (Boston, Mass: Addison-Wesley). Economic Survey (2003) (New Delhi: Government of India). Fry, Maxwell J. (1995) Money, Interest and Banking in Economic Development (Baltimore, Md.: Johns Hopkins University Press). Graham, E. M. (2004) ‘Do Export Processing Zones Attract FDI and Its Benefits – The Experience from China’, Journal of International Economics and Economic Policy, vol. 1(1), pp. 87–103. Graham, E. M. and E. Wada (2001) ‘Foreign Direct Investment in China: Effects on Growth and Economic Performance’, Institute for International Economics Working Paper no. 01-3, Washington, DC. Kumar, Nagesh (1990) Multinational Enterprises in India (London/New York: Routledge). Kumar, N. and N. S. Siddhartan (1997) Technology, Market Structure and Internationalisation (London/New York: Routledge). Nagaraj, R. (2003) ‘Foreign Direct Investment in India’, Economic and Political Weekly (Mumbai), 26 April. Nayar, Baldev Raj (2001) ‘Opening Up and Openness of the Indian Economy’, Economic and Political Weekly (Mumbai), 15 September. Rangarajan, C. (2000) ‘Capital Flows: Another Look’, Economic and Political Weekly (Mumbai), 9 December. Serven, Luis and Andres Solimano (1992) ‘Private Investment and Macroeconomic Adjustment – A Survey’, World Bank Research Observer, no. 7. Srivastava, Sadhana (2003) ‘What Is the True Level of FDI Flows to India?’, Economic and Political Weekly (Mumbai), 28 June. Sun, H. (1998) ‘The Macroeconomic Impact of Direct Foreign Investment in China, 1979–96’, World Economy, vol. 21, 5 July. Van Wijnbergen, S. (1982) ‘Stagflationary Effects of Monetary Stabilization Policies: A Quantitative Analysis of South Korea’, Journal of Development Economics, vol. 37, pp. 133–69. UNCTAD (2001) World Investment Report 2001 (Geneva: United Nations).

11 Portuguese Investment in Brazil – the Challenges of an Iberian Logic Cézar Miranda Guedes Federal Rural University of Rio de Janeiro, Brazil

and Mario Gomez Olivares Technical University of Lisbon, Portugal

1

Introduction

Since the 1980s, the world economy has undergone a process of institutional change characterized by systematic financial deregulation and a wave of privatizations. The globalization of trade, finance and production has thus gathered momentum, side by side with the expansion of transnational and multinational enterprises, which have reached a tremendous financial weight, increasing international liquidity and intensifying the search for short-term profits (Scherer, 1999). With regard to Latin America, the fact that – unlike the situation in the past – trade and investment have ‘drifted apart’ and become increasingly independent of one another is also worthy of notice. In a context of generalized liberalization among the domestic economies, the emergence of new national markets has brought out the need to adapt to the strong neo-liberal bias of the new development model, in which additional financial resources are being channelled so as to foster the growth of those markets that have opened themselves up to the world economy and to foreign investment. In this chapter, we focus on the repercussions this has had for the Portuguese, Spanish and Latin-American economies, in terms of new problems and potential opportunities. Throughout the 1990s, and particularly from 1995 onwards, investments made by Portuguese corporations in Brazil increased substantially. In 1996, Brazil overtook Spain as the prime destination of Portuguese 212

Cézar Miranda Guedes and Mario Gomez Olivares 213

direct investment abroad, and since than, Portugal has ranked as one of the major sources of foreign direct investment in Brazil. In this chapter, we maintain that, rather than following a pattern of its own, the trend displayed by Portuguese investment in recent times follows the general logic of Iberian investment, as initiated by the Spanish in the 1980s by channelling large financial resources into Latin America (particularly its Southern Cone – that is, the subcontinent that includes Argentina, Brazil, Chile, Paraguay and Uruguay). Accordingly, we demonstrate three main characteristics peculiar to this Iberian logic, and discuss them in detail. The first of these characteristics relates to the fact that, unlike the usual case with investments originating in other ‘Triad’ countries (that is, the highly developed countries of North America and Europe plus Japan), Portugal and Spain direct most of their investments abroad to countries that are not a part of the Triad, or even of the OECD.1 The second specific characteristic of Iberian investment concerns the profile of the investments – which, in the case of both countries, are predominantly concentrated in the service sector, comprising mainly non-tradable services aimed at the domestic market. Indeed, examples of partnerships between Portuguese and Spanish investors can be found in both the telecommunications and electricity supply sectors, in a context of increasing international expansion of Iberian corporations aimed at both European and global markets. Finally, the third specific characteristic of the Iberian logic is its predominant direction towards countries that share a language and culture with either Portugal or Spain, allowing for a substantial reduction in transaction costs, as well as other considerable, albeit less visible, advantages. The remainder of this chapter is as follows: Section 2 analyses Portugal’s place in the European and world economies; and Section 3 discusses foreign direct investment in Latin America and expands upon the subject of the logic of Iberian investment, as briefly outlined in this introduction. Section 4 concludes.

2 The place of the Portuguese economy in the European and world economies The socioeconomic disparities between Portugal and the rest of Europe have been decreasing since the 1960s, as Portugal assumed an increasingly outward-orientated approach to commercial and financial flows that ultimately led to it joining the European Economic

214 Portuguese Investment in Brazil

Community (EEC) in 1986. The issue at stake at the time of writing is somewhat more ambitious: turning the European Union into a federation of national states following the abolition of national borders and currencies, and the adoption of the euro as a symbolic and economic expression of the European Union (EU). The process currently in place will eventually bring to an end those national trade barriers that still remain in place, deprive national governments of the possibility of pursuing independent monetary and exchange-rate policies, and transfer most of the responsibility for macroeconomic policy-making to various new supra-national entities.2 The remarkable nature of this process lies in the fact that there are no ready-made models to be followed; rather, entirely new political concepts and supra-national institutions have to be designed. The adoption of the euro by the various EU member countries was conditional upon meeting a number of convergence criteria, as stated in the Maastricht Treaty: ●

● ●

●

an inflation rate within 1.5 per cent of the average rate of the three EU countries with the lowest inflation; a budget deficit under 3 per cent of GDP; a global public debt of less than 60 per cent of GDP, or clearly showing a tendency towards reaching that reference level; and a stable exchange rate for at least two years within the European Monetary System (EMS).

States wishing to join the EMU had to prove their ability to ensure sustained nominal convergence built on budget and price stability, so as not to jeopardize the viability of the common currency and of the supra-national pursuit of macroeconomic policies. In the end, Portugal managed to meet all the demands of the nominal convergence criteria. However, it is worthwhile mentioning the importance of ensuring that the costs of nominal convergence do not themselves jeopardize the real, structural convergence between the member states of the European Union. Various economic, political and historical issues are at stake, since, with regard to the Continent’s internal spatial logic, Portugal falls outside the core. In fact, Portugal’s position can adequately be described as one of ‘integrated periphery’ with regard to the London–Milan axis – that is, the area that stretches from Southern England to Northern Italy (Durand et al., 1992). Portugal’s industrialization is notably restricted to areas where firms of industrialized European countries invested to

Cézar Miranda Guedes and Mario Gomez Olivares 215

take advantage of the country’s lower wages. Most Portuguese firms are small, although they participate in exports connected to multinational collaboration in textiles, footwear, wood and cellulose (see Lança, 2000). The fact that disparities are gradually decreasing – notwithstanding the Portuguese economic structure, business orientation and quality of human resources –leave it in a rather vulnerable and fragile position that, compared with European standards, limits its competitiveness and potential for internationalization. One must bear in mind that Portugal did not take full part in the innovation cycles that characterized the first and second industrial revolutions. Hence the development of industrial capital took place belatedly, having occurred at a time in which substantial technological and financial barriers were already in place in the world economy. Up to the mid-twentieth century, Portugal’s industrial structure essentially comprised labour-intensive, low-value-added goods with little potential for export (apart from a few primary goods) and aimed mainly at the domestic market. But in the latter half of the twentieth century, by taking advantage of the post-war international expansion, Portugal’s manufacturing and service sectors gradually became more dynamic. New processes, products and technologies were introduced, new labour and production management techniques were adopted, and new circuits for the dissemination of scientific knowledge and technological know-how were put into place. However, no efforts were made to foster endogenous innovation, which is the ultimate deciding factor for sustained long-term competitiveness. In our analysis of Portuguese competitiveness, we draw on the work of Lança (2000) on Portuguese exports, and on a comparative study of European industry (European Communities, 1999). In the ‘core’ of Portuguese industrial exports, which in 1996 accounted for 80 per cent of the total, competitiveness arises from three factors: (i) lower costs for textiles and tanning (31 per cent), timber, woodworking and glass (8 per cent); (ii) natural resources and/or economies of scale for pulp and paper, oil refining and non-metallic minerals (8 per cent), although wine-making, based on agricultural resources, is in decline (3 per cent); and (iii) human resources with higher qualifications than is generally the case in manufacturing industries for electrical machinery (12 per cent) and transport equipment (17 per cent). The author of this analysis concludes that the specialization pattern of Portuguese industry shows some serious vulnerabilities. These vulnerabilities can be more clearly perceived by means of a series of commonly used indicators, namely: the evolution

216 Portuguese Investment in Brazil

of international demand, both in the present and in the foreseeable future; the development of the national capacity to gradually ... move upstream on the production chain so as to manufacture competitive equipment; the vulnerability of the industries already in place to the foreseeable removal of all national and international barriers to trade; and, finally, its vulnerability to the worldwide reorganization of industry. (Lança, 2000, p. 33) In a comparative analysis of the trends during 1988–98 in EU (European Communities, 1999) manufacturing, Portugal (with Austria), ranks as one of the countries in which per capita value-added has increased the most: its annual increment of 6.7 per cent is exceeded only by Ireland at 7.9 per cent. This improvement derives chiefly from the expansion of the motor vehicle and electrical appliance branches. Portugal is the only EU country to have gone counter to increased specialization in the sense that the relative weight, in both exports and production, of those sectors that have national comparative advantage (textiles, food processing and woodworking) has in fact decreased in the period. Such a trend may nevertheless be quite positive, since it indicates that Portuguese industry is undergoing structural changes and moving towards the production of goods generating more value-added. This, however, arises less from domestic than from multinational investment, made by transnational corporations as part of their global strategy. But if that capital is not backed by a qualified labour force and adequate technological and innovation infrastructures, negative externalities may prevail and induce disinvestment, as has already been shown in textiles and footwear. This risk is by no means negligible, especially because of the eastward expansion of the EU, since the countries that joined in May 2004 offer lower direct and indirect costs, and access to a higher-qualified labour force (Guedes, 2001). With this in mind, it should be stressed that Portuguese manufacturing companies do not have enough strength to compete in international markets, and in the 1990s the expansion of Portuguese firms was concentrated in the service sector. Furthermore, the domestic market is too small for some of the largest Portuguese firms to justify investment in the expectation of gaining global markets. Portuguese corporations are in no position to become ‘global players’ and therefore implement more selective or subordinate strategies. This is precisely why Brazil, also because of the characteristics of its internal market, presented itself as an attractive alternative destination for Portuguese investment, as had been the case in the 1980s with Spanish investment in Latin America, particularly in

Cézar Miranda Guedes and Mario Gomez Olivares 217

the Southern Cone. In this connection, a brief discussion of the structure of the Latin-American Southern Cone is relevant, since most analyses of that continent treat it as a whole. The Common Market of the South (Mercosur), comprising Argentina, Brazil, Paraguay and Uruguay was created in 1991, drawing its inspiration from the EU. Chile and Bolivia gained associate status in, respectively, 1996 and 1997. The analysis in this chapter dealing with Mercosur also relates to these two countries. The four full members have a combined GDP of nearly 1 trillion dollars, accounting for 60 per cent of Latin-American GDP, and a potential market of 210 million people.3 Although the six countries share a common historical and cultural background, as well as a number of common economic and geopolitical characteristics and constraints, they are substantially disparate, and cannot be treated as an integrated whole. Within the commercial and financial flows between the Latin-American countries and the Triad, as well as within Latin America, those of the Southern Cone are largely with the EU rather than with the USA. This was reflected in 1995 by the Interregional Cooperation Agreement between Mercosur and the EU, aimed at gradually liberalizing trade between the two regions. After five further years of negotiation, however, the October 2004 target for its replacement by a bilateral trade agreement was not met. In addition to the EU, Brazil and Chile have significant commercial ties with the Asia-Pacific region. During the 1990s, trade among the Mercosur countries increased from US$4.1 billion to US$20 billion, a relative increase that was more substantial than that taking place in the world as a whole, and was accompanied by a substantial, albeit a lesser, rise in Mercosur trade with the rest of the world. Since the increase in intra-regional flows did not take place to the detriment of its trade with other countries, Mercosur can be considered an open regional economy. In sum, Mercosur countries are not as dependent on the US economy, and are more complementary to one another, than the rest of Latin America. With regard to North America, the closer Latin-American countries are to the territory of the USA, the larger the relative weight of their trade and investment with the USA.4 Mexico provides a clear example of this, having recently overtaken Japan as the second most important US trade partner. Mexican–US trade accounts for 85 per cent of the Mexican total. The two largest export earners are the oil industry and the Indústrias Maquiladoras de Exportación (IMEs), export-processing zones, in which virtually all the investment comes from the USA. IMEs account for nearly half of the Mexican active population employed in industry.

218 Portuguese Investment in Brazil

Hence, it is only in the Southern Cone, rather than in the whole of Latin America, that the EU is the dominant trade partner. It was against this general background that the logic of Iberian investment became evident in the 1990s. Just how connected the Iberian economies are with this part of the American continent can be shown by the dependence of the largest Iberian investors (such as Telefónica and Portugal Telecom) on the performance of the Southern Cone economies. This was clear when these economies were affected severely by the adoption of a fixed parity between the Argentinean peso and the US dollar from 1991 onwards, then by the dissemination of that exchange rate policy throughout the region, and by the devaluation of the peso. This succession of events upset not only the financial markets, but also the whole of the economy of the region – including its trade relations, since internal demand was affected and the large foreign corporations present in the region saw the market value of their assets drop considerably.5 The problems faced by this region over the 1990s demonstrate that policies such as the adoption of a fixed parity with regard to the US dollar are by no means viable in these countries.

3 Foreign direct investment in Latin America and the logic of Iberian investment The international crisis that took place in the early 1980s gave way to a period of expansion characterized by the intensification of the competition among businesses at the international level. The effects of this process of competitive globalization were especially visible in increasing numbers of mergers and acquisitions (M&A), particularly in the USA, Europe and Japan. The technological changes brought about by the development and rapid dissemination of microelectronics and its applications to telecommunications and information technologies have had a substantial impact on the organization of industry, while fostering globalization by making it possible to convey large amounts of information in real time and at very low cost. The increasing availability of information with respect to the international supply of goods and services has prompted the globalization of demand, which in turn has promoted the global expansion of supply and the intensification of international competition. However, it should be noted that this process is not related solely to the emergence of new economic spaces, in which corporations, by taking advantage of further economies of scale, compete to create and enlarge their market power by means of the traditional competition strategies (price, quality, distribution, costumer support), but rather that the entire

Cézar Miranda Guedes and Mario Gomez Olivares 219

productive system, as well as the factors that determine competitiveness, have themselves undergone profound changes. The engines of this process have been of two different types: technological change on the one hand, and innovations in the organization of production on the other. The business relations within corporations, as well as between different corporations, have changed, since the new logic of production favours flexibility, quality and co-operation above all other characteristics, and establishes a close link between the adoption of flexible technologies in manufacture and new ways of managing the production process. Competitive globalization has prompted an increasing number of M&A and has fostered the liberalization of certain sectors even in industrialized countries. This has been the case with certain specific services, such as banking and finance, which have rendered possible, and benefited from, the international expansion of direct investment in this sector. Mercosur represented a reaction to this process on the part of the emerging countries and has constituted a focus of attraction for foreign investment and trade – in an early phase, particularly from the leading economies – that is, the USA, Germany and Japan. In the 1990s, by following an Iberian rather than a European logic, Spain and Portugal assumed increasing relevance as foreign investors and trade partners, because they have firm specific advantages as earlier colonial powers in the zone that allowed the maintenance of cultural, linguistic and family links with Latin-American countries, and because Spain and Portugal supported the democratization process in the 1980s, and hence also gained priority in political communication. Analysis of the internationalization process of Portuguese and Spanish corporations shows that most Spanish investments have been directed to Latin America, whereas the chief destination of Portuguese investments has been Brazil. We stressed above that Mercosur is an Iberian, rather than European, investment ground, going beyond the mere relative weight of investment flows – a unique logic that is peculiar to Iberian investment. The first characteristic of the Iberian investments of the 1990s is that both Portugal and Spain have directed most of their investment flows to non-industrialized countries. The second is that most investments have been in the service and commerce sectors, aimed at controlling the internal market. The third characteristic is that, for the most part, Iberian investments have been carried out in countries that share a common language and culture with one of the Iberian countries – that is, in Latin America. There is an extensive literature on the beneficial effects of

220 Portuguese Investment in Brazil

sharing a common currency in reducing transaction costs and, similarly, on the beneficial effects of sharing a common language in minimizing communication costs (Ramos Silva, 2000). Before characterizing the nature and significance of those Portuguese and Spanish investments, we briefly discuss foreign investment in Latin America as a whole, as it was remarkably large. Total foreign investment there increased from US$47,694 million in 1980 to an astonishing US$415,614 million in 1998, a tenfold increase (which amounts to roughly four times the Portuguese GDP for the year 2001). Brazil was the prime destination of this investment, accounting for a third of the total (US$156,798 million), followed by Mexico (US$60,783 million), Argentina (US$45,466 million), Bermuda (US$30,905 million), Chile (US$30,481 million), Colombia (US$14,320 million) and Venezuela (US$13,304 million). Considering that Bermuda is an offshore tax haven for international corporations and, as noted above, Mexico is a special case because of its proximity to the USA, we conclude that the lion’s share of this FDI, apart from these two special cases, was concentrated in the Southern Cone countries. In 1999, investments continued on a similar path, with Brazil ranking first (US$26,500 million), followed by Argentina (US$20,000 million) and Chile (US$3,500 million), as shown in Table 11.1. The relevance of these investment flows, which were much larger relative to historical flows and to their weight in these economies, can be measured by their contribution to gross fixed capital formation (GFCF) in the region (see Table 11.2). In 1997, FDI accounted for an average contribution of 16.1 per cent of GFCF for the whole of the region: for 11.9 per cent in Brazil; 12.7 per cent in Argentina; 16.3 per cent in Mexico; 27.9 per cent in Chile; 34.4 per cent in Venezuela; and 38.2 per cent in Colombia. The figures are even more impressive in the case of the smaller countries such as Costa Rica, Peru, Bolivia and Ecuador. The magnitude of FDI into this region demonstrates its importance at this stage of the globalization process in laying the foundation for the export competitiveness of these countries, based on comparative advantage in mining, agriculture and manufacturing. These FDI flows were heavily concentrated in M&A rather than in greenfields investment. The M&A occurred particularly in four sectors: financial services (banks, insurance companies, pension funds); telecommunications; energy supply and transport (production, distribution, airlines); and other services (commerce, urban waste disposal), as Table 11.3 demonstrates. The leading actors in this process – which took place at a time of extensive privatization and take-overs in most Southern Cone countries – have

Table 11.1

Foreign investment flows in Latin America, 1990–9, by country (millions of US dollars)

Latin America and the Caribbean Argentina Bolivia Brazil Chile Colombia Mexico Uruguay Paraguay Remaining countries Source:

1990

1991

1992

1993

1994

1995

1996

1997

1998

1999

6 758

11 065

12 506

10 359

23 706

24 878

39 329

55 222

59 934

70 275

1 836 66 324 654 484 2 549 – 76 770

2 439 50 89 697 433 4 742 – 84 2 531

3 218 91 1 924 538 679 4 393 – 118 1 545

2 059 121 801 600 719 4 389 102 75 1 493

2 480 147 2 035 1 672 1 297 10 973 155 137 4 811

3 756 391 3 475 2 204 712 9 526 157 155 4 502

4 937 472 11 666 3 445 2 795 9 186 137 246 6 445

4 924 728 18 608 3 353 4 394 12 831 113 270 9 501

4 175 870 29 192 1 842 2 509 10 238 164 235 10 710

20 000 745 26 500 3 500 250 11 000 145 350 7 885

CEPAL, Balance preliminar de las economias latinoamericanas 1999, Santiago de Chile, 1999.

221

222 Portuguese Investment in Brazil Table 11.2 FDI flows as a percentage of GFCF in selected Latin-American and Caribbean countries, 1987–92 to 1997 (millions of US dollars) Countries Latin America and the Caribbean Argentina Bolivia Brazil Chile Colombia Paraguay Venezuela Remaining countries Note:

a

1987 – 92a

1993

1994

1995

1996

1997

5.4

6.0

9.1

9.4

12.5

16.1

7.6 7.9 1.8 14.4 6.8 4.1 5.5 5.1

5.8 12.9 1.5 9.3 10.0 5.0 3.1 14.2

6.1 14.7 2.3 21.8 10.3 7.8 7.9 19.01

10.5 35.9 3.8 19.1 6.1 7.5 7.9 15.2

12.4 39.8 7.1 27.5 21.0 11.3 21.0 16.3

12.7 53.8 11.9 27.9 38.2 12.0 34.4 27.2

Annual average.

Source: UNCTAD, World Investment Report 1999.

Table 11.3 Latin America and the Caribbean: privatizations exceeding US$1m involving foreign investors in selected sectors, 1998–9 Sector

Millions of dollars

Primary Manufacturing Services Telecommunications Electricity supply Transport Basic sewage infrastructure Gas supply Financial services

2 813 5 675 43 946 21 092 12 138 7 670 1 754 988 304

Total

96 380

Source: CEPAL, Centro de información de la Unidad de Inversiones y estrategias empresariales, 1999.

been European and American corporations. Table 11.4 illustrates the relationship between FDI flows and the principal sector activity of the private companies that have been subject to take-overs. In this period (1998–9), investments in Brazil accounted for more than half the total investment in the region, coinciding with the acceleration of the privatization process by the government, headed by then-President F. H. Cardoso.

Cézar Miranda Guedes and Mario Gomez Olivares 223 Table 11.4 Latin America and the Caribbean: private company take-overs by foreign investors exceeding US$1 m, by sector and amounts involved, 1998–9 Activity sector

Millions of dollars

Oil and gas Manufacturing Services Banking and finance Electricity supply Commerce Telecommunications Other services

13 375 5 655 21 169 9 220 5 378 3 308 3 053 210

Total

40 199

Source: CEPAL, Centro de información de la Unidad de Inversiones y estrategias empresariales, 1999.

Table 11.5 Latin America and the Caribbean: FDI flows originating in Europe, the USA and Japan, 1995–7 (net flows, in million of dollars) Europe

Argentina Brazil Mercosur Chile Andean countries Mexico Central America Latin America and the Caribbean Source:

USA

Japan

1995 – 7

%

1995 – 7

%

1995 – 7

%

5 700 10 420 16 651 3 195 7 257 4 611 79 32 285

17.7 32.3 51.6 9.9 22.5 14.3 0.2 100.0

3 825 17 311 21 291 3 199 4 861 11 629 1 124 43 015

8.9 40.2 49.5 7.4 11.3 27.0 2.8 100.0

103 1 815 1 917 50 201 194 0 2 362

4.4 76.8 81.2 2.1 8.5 8.2 0.0 100.0

IADB/IRELA (1998).

Table 11.5 shows investments originating in Europe, Japan and the USA. By themselves, they account for nearly all the FDI in the region – indeed, investment flows originating within the region or in other parts of the world can be considered negligible. Half the investments made by European and North American investors in Latin America and the Caribbean in the period shown were directed to Mercosur countries

224 Portuguese Investment in Brazil

(80 per cent in the case of Japan, whose investment was smaller and largely concentrated in Brazil, Argentina, Venezuela and the Andean countries). If investment destined for Chile, Mexico and the Andean countries is included, there is a considerable concentration in a handful of countries. This is because these are the most developed countries in the region and have the capacity to adjust rapidly to the new international division of labour. With regard to European investment, it is important to note that a very significant part of this process has been initiated by Iberian corporations – mainly Spanish, but also Portuguese. Even though Portuguese investments are smaller in volume, they represent a remarkable internationalization effort, in view of the size of the Portuguese economy. Portuguese FDI in 2000 amounted to approximately US$5 billion (some 5 per cent of GDP), which is equivalent to the FDI entering Portugal. Table 11.5, on FDI flows destined for Latin America and the Caribbean by national origin, shows an increase in European FDI throughout the 1990s. That originating in Spain stands out from 1994 onwards – exceeding FDI originating in Japan, Germany, the United Kingdom and France – and has contributed to the approximation of the share of European investment to that of North American investment. From 1986 onwards, the effects of the international co-ordination of macroeconomic policy, of the fall in the price of oil, and of the somewhat more optimistic growth expectations, became manifest. MNEs have engaged in fierce international competition with one another in an effort to consolidate and expand their influence over the most attractive markets. This has led to a higher concentration of FDI flows, relegating most developing countries to a subordinate position with regard to FDI originating in the industrialized countries. In this process, Latin America has stood out as a major destination for investment flows, ranking as a close third behind the Triad and the Asian countries. Spain’s aggressive investment stance has helped to explain this development, while the contribution of Portuguese investment has been by no means negligible. Spanish investment in Latin America has been on the rise since the 1980s, seeking to take advantage of the favourable international circumstances – in particular, the Latin-American increasing openness during the 1990s. Following a period when these countries were hostile to FDI, they then sought it eagerly to finance expansion of their export sectors as well as for the privatization and restructuring of domestic industry. These countries also sought to balance their foreign trade in accordance with the liberal development model, which evoked

Cézar Miranda Guedes and Mario Gomez Olivares 225

the primacy of the market and the reduction of the role of the state (Zapata, 1999). In global terms, according to the ECLAC, FDI in Latin America and the Caribbean during the 1990s had three main characteristics (CEPAL, 2000). First, its rapid growth, from US$6,758 million in 1990 to US$70,725 million in 1999; second, its concentration in a few countries; third, its stability in the face of the international financial crisis, since it continued to rise at a time when it was falling in other parts of the world; and fourth, the acquisition of assets already in place, accounting for two-thirds of those FDI flows. The executive in charge of promoting privatization at USAID, Henriqueta Holsman, has claimed that ‘Industries such as telecommunications, finance and energy supply are being restructured so as to meet the challenges of an integrated global economy. The globalisation of the international economy calls for these industries to engage in the privatisation process’ (Martin, 1993, p. 9). These are precisely the sectors in which most Iberian investments have been carried out. The slow recovery from the foreign debt crisis, the financial vulnerability of governments and a general economic stagnation gave way, from the 1980s onwards, to a new development model in which the national economies have opened their borders to international capital, notably by disposing of state-owned assets in strategic sectors of the domestic market. As shown in Table 11.6, Spanish investors have taken advantage of these opportunities. The investment strategy that had been initiated in the 1980s6 was pursued with renewed vigour, bringing total investment flows up from US$1,037 million 1995 to US$5,653 million in 1997.

Table 11.6 Latin America and the Caribbean: FDI flows (excluding offshore centres), net flows, 1990–7, 1995, 1996 and 1997 (millions of dollars and percentages)

1 2 3

Europe Spain USA Japan

Total 1 + 2 + 3 Source:

Total 1990–7

%

1995

44 354 11 007 79 779 4 291

34.5 8.6 61.8 3.4

6 496 1 037 15 283 492

8 278 1 577 9 918 780

17 510 5 653 17 814 1 091

100

22 271

18 976

36 415

127 424

IADB/IRELA, 1998.

1996

1997

226 Portuguese Investment in Brazil

Beginning in the 1980s, and continuing into the 1990s, Spanish investors invested strongly in the service sector, namely in transport and communications, urban infrastructure, oil and energy supply, banking and insurance (see Table 11.7). Even though investments have been highly concentrated in these sub-sectors, there have been sporadic considerable investments in sectors such as agriculture and tourism, which, albeit not all too meaningful in terms of global Spanish FDI, assumed considerable relevance for the Latin-American countries, by fostering non-traditional exports in the tourism, food processing and fishing sectors. Spanish investments have typically been carried out by a small number of corporations, such as Banco Bilbao Vizcaya Argentaria (BVB); Banco Santander Central Hispano; Endesa; Repsol; and Telefónica. Hence these investments have been important not only for Latin America, but also for the profitability and market value of these corporations. Spanish corporations have led the M&A process throughout Latin America and have gained strategic control over various sectors, such as telecommunications and the financial system. Table 11.7 shows the

Table 11.7 Spanish FDI in Latin America, 1993–6 (millions of dollars and percentages) 1993 Primary sector Agriculture, cattle breeding and fishing Oil and fuel processing Services Construction Commerce Transport and communications Electricity, gas and water supply Financial services (banking and insurance) Hotels Holdings Total Source:

1994

1995

1996

1997

1998

1993–8 (%)

0 0

18 18

37 37

438 9

19 19

36 19

1.9 0.3

0

0

0

429

0

17

1.5

429 5 0 106

3 217 52 0 2 211

1 597 64 16 179

3 438 45 26 176

6 836 200 12 407

12 206 170 127 103

9.4 1.8 0.6 11.1

9

51

116

217

810

2 352

12.0

69

148

195

1 366

2 449

1 626

19.8

0 303

0 755

36 991

37 1 571

55 2 813

32 7 796

0.5 48.1

521

3 388

1 743

4 024

7 238

12 638

100.0

ECLAC, CIUIEE-DDPE, Ministério de Hacienda e Economia de Espan˜ a, 2000.

Cézar Miranda Guedes and Mario Gomez Olivares 227 Table 11.8 Portuguese FDI (¤ thousands and percentages) Destination

1998

%

United Kingdom France Germany Spain Switzerland USA Brazil PALOP Other

58 093

0.707

36 942 5 664 353 683 17 912 81 950 3 692 256 104 326 3 810 184

Total

8 211 988 100.0

Source:

1999

%

2000

%

138 909

1.685

250 471

2.2

0.449 35 253 0.689 5 823.3 4.306 497 852 0.218 13 460 0.997 68 117 44.96 1 579 532 1.27 185 163 46.39 5 664 216

0.427 0.706 6.041 0.163 0.826 19.167 2.246 68.73

39 441 125 233 564 980 15 804 411 873 2 942 278 262 748 6 504 234

8 240 735

100.0

11 117 062

0.35 1.12 5.08 0.142 3.7 26.46 2.363 58.5 100.0

Bank of Portugal.

relative importance of the various sectors (associated with corporations such as Telefónica, Endesa, Repsol, Banco Santander or BVB). Since the late 1990s, Portuguese investment has begun to take part in the global economy in terms of the formation of international capital (see Table 11.8). Several large investments were made in Brazil in that period, and some Portuguese corporations participated in a number of investments led by Endesa and Telefónica. Generally speaking, Portuguese investments have shared similar characteristics to Spanish investments, although the former are mainly concentrated in Brazil and in a number of specific sectors, such as commerce, urban infrastructures and, above all, telecommunications – that is, services aimed at the internal market. Like Spain, Portugal has chosen Latin America, particularly Brazil, as the natural and preferred destination for its investments abroad – compared to the pattern of Portuguese foreign trade, which for the most part is carried out within Europe. Table 11.8 illustrates the importance of Brazil as the chief destination for Portuguese FDI in the period 1998–2000, confirming the trend initiated in the mid-1990s. Portuguese investment abroad experienced a twentyfold increase in just five years, denoting a daring attempt on the part of a number of large Portuguese corporations (which had themselves been privatized in the early 1990s) to compete in markets characterized by a significant internal demand. Figure 11.1 illustrates this increase and corroborates the argument that Portuguese investment in Brazil – albeit fostered haphazardly at an early stage by the privatization and modernization processes taking

228 Portuguese Investment in Brazil

3000 2409

US$ millions

2500 2000

1755

1500 1000 500

1290 695 202 1996 1997 1998 1999 2000

0

Figure 11.1 Portuguese investment in Brazil Source:

Central Bank of Brazil.

place in the Brazilian economy – has assumed a sustained, structural nature. Typically, Portuguese investments abroad have been carried out in the communications, banking and finance, and service sectors (see Table 11.9). The fluctuations in the relative importance of each sector from one year to the next are caused by the fact that these investments are correlated with the privatizations taking place at the same time in Latin America, particularly in Brazil. That is, since most investments take the form of a take-over of foreign companies, investment figures are closely related to the pace of the privatization processes and to the restructuring needs of the target companies. In the year 2000, according to the Portuguese Central Bank (Bank of Portugal, 2000), total Portuguese investment abroad amounted to US$6 billion; that is, roughly equivalent to 5 per cent of Portuguese GDP. A major part of these investments corresponds to the take-over of Telesp Celular by Portugal Telecom. Since Telesp had signed a business agreement with Telefónica (the largest Spanish telecommunications operator), a joint venture has thus been created that provides an example of an Iberian joint investment strategy7 (the concerted action on the part of Endesa and EDP together building new electric plants for Brazil, provide another example). The remaining 25 per cent was in banking (CGD), supermarkets and the wood processing industry (Grupo Sonae). In 2000, Portuguese investment in Brazil was carried out by a number of major

Cézar Miranda Guedes and Mario Gomez Olivares 229 Table 11.9 Portuguese FDI (¤ thousands and percentages) Activity sectors

1998

8 136 Agriculture, hunting, forestry and fishing Mining 29 340 Manufacture 129 939 647 973 Electricity, gas and water supply Construction 45 657 Commerce 57 967 Transport, storage 3 495 067 and communications Banking and 572 544 finance Real estate, 3 167 406 rentals and B2B services Other activities 84 366 Total Note: Source:

8 211 989

%

1999

%

2000

%

0.1

1 996

0.0

9 042

0.1

0.0 1.6 7.9

2 944 225 513 7 005 280

0.0 2.7 8.5

2 539 634 932 384 567

0.0 5.7 3.5

0.6 0.7 42.6

87 676 88 478 3 925 909

1.1 1.1 47.6

103 015 124 377 144 734

0.9 1.1 1.3

7.0

676 578

8.2

2 500 711

22.5

38.6

2 124 624

25.8

6 957 252

62.6

1.0

406 489

4.9

255 892

2.3

100.0

8 240 735

100.0

11 117 061

100.0

B2B = business-to-business. Bank of Portugal.

Portuguese corporations – for example, Banco Mello, BES, CGD, Grupo Pestana, EDP, Cimpor, Somague, Grupo CGM and Brisa. However, one should bear in mind that the biggest share of these investments was been made by Portugal Telecom and the Sonae distribution group. This trend contrasts with the rather insignificant investment in Central and Eastern Europe (even though it did rise) and to the diminutive investment in the African countries (which nevertheless were of great relevance for the African Portuguese-speaking countries).

4

Concluding remarks

One of the chief characteristics of foreign investment in Latin America in the 1990s was its direction, primarily the acquisition or take-over of already existing assets and thus did not increase the domestic productive capacity already in place. Foreign investors first sought to take

230 Portuguese Investment in Brazil

over state-owned companies, and then do the same with large private companies. However, in the case of certain sectors, such as mining and automobiles, foreign investment has in fact resulted in an increase in productive capacity. Similarly, FDI has allowed the modernization of the financial sector, even though its impact in this sector has been mainly to fund new facilities. The shift in the economic development paradigm prevalent in Latin America, by placing renewed emphasis on the creation of open, outward-orientated competitive economies based on private property, fostered a unique and unrepeatable privatization that created extraordinary opportunities for FDI – and for Iberian investment in particular. As a result of that shift, companies in every main activity sector were privatized, and countries opened up to foreign direct and portfolio investment, seeking modernization and growth. In a context of increasing mobility of international capital, foreign investors in Latin America engaged in a policy of acquiring assets aimed at modernizing the LatinAmerican business structure in a strategy of national consolidation based on the international accumulation of capital. This strategy was visible in many sectors in Latin-American countries: Spanish corporations have taken over the Latin-American markets for energy supply, telecommunications, banking and finance. Spanish investments in Latin America have also served as a platform for Portuguese investment in Brazil – a rather wider market that was opened up somewhat more belatedly to foreign capital. The competitive weaknesses of traditional Portuguese industry, such as textiles and footwear, prevent it from exporting successfully to Latin America. The more recent and dynamic export sectors, such as automobiles and electronics, are aimed primarily at intra-industrial trade within the global logic of multinational corporations, but can nevertheless give rise to new export opportunities (as is the case with car parts and mouldings). All this should be taken into account in analysing the pattern of Portuguese foreign trade and international investment, and in comparing it with that of Europe. The logic of Portuguese investment – and, in a way, that of Spanish investment as well – consists of taking advantage of the potential of the emergent markets of Latin America – seeking to take over privatized sectors to take control of the internal markets for non-tradable goods and services. The future of FDI in LatinAmerican countries depends on the recovery of world trade and the economic dynamics of investment in the emerging economies. The privatization process will no longer have a great incentive to invest, except in the case of Brazil, which has an important industrial private

Cézar Miranda Guedes and Mario Gomez Olivares 231

sector, but there is always the possibility of exploring new concepts of FDI, such as the participation in new projects involving services – for example, transport or tourism; expanding the traditional branches of mining, agriculture and fishing; and in major projects involving, for example, gas and oil. There is no clear and reliable scenario for the medium term, partly because of political changes in the area: while there is no strong opposition among political leaderships to FDI, there is a social restlessness and a growing opposition to accepting the free movement of capital, in which the burden and risk of external debt is a feature, as seen in the case of Argentina. FDI was an important instrument for the growth of the region in the last decade of the twentieth century, much more than free trade. Much depends for the future of FDI in the region on negotiations between the USA and the Latin-American countries in the context of a free trade area. Accordingly, Portuguese (and Spanish) investments have tended to concentrate on the production of goods destined for the internal market, in which demand has been steadily rising – such as energy supply, urban infrastructure, communications, commerce and other services. Often, Portuguese investors have forged alliances with Spanish firms, aimed at sharing Latin-American markets (in a similar way to the Tordesillas Treaty of 1494). Latin-American countries are in need of further and faster capital growth – the engine of accelerated economic growth – which is why their own interests in many ways match those of Iberian investors, who in turn seek to pursue their expansion into markets that share a similar cultural background and allow larger profits at lesser risk.

Notes 1 Much of the growth in international business is among developing markets, ones that are outside the traditional economic powerhouses such as the USA, the EU and Japan. See, for example, Sweeney and McFarlin (2001, ch. 2). 2 Fiscal policy is formally independent, but the Stabilization Pact, which is a political consequence of the need for monetary policy directed towards a stabilization of the euro conducted by the independent ECB, restricts the autonomy of national governments. 3 The nominal GDP of Mercosur has dropped substantially since 1999, largely because of the devaluation in Brazil – a country that accounts for more than double the combined economic weight of the three other countries in Mercosur. For a more detailed coverage of Mercosur, see the website of the Brazilian Institute of Geography and Statistics (IBGE) at http://www.ibge.gov.br, entitled the European Community–Mercosur Statistical Cooperation Project. This reflects the commitment of the EC to giving Mercosur countries the

232 Portuguese Investment in Brazil

4

5

6

7

benefit of Eurostat and, by including Chile, making the whole Southern Cone an essential partner in its economic diplomacy. For a statistical analysis of the composition of the goods trade in all the LatinAmerican and Caribbean visible trade by commodity and by destination, see CEPAL (2001) and for an analysis of national competitiveness by industry and company, see Mortimore and Perez (2001). See Albuquerque and Romão (2000) for a global analysis of the economic relations between Portugal and Brazil; and Ramos Silva (2001) for a cross-sector analysis of investment and the balance of trade in the 1990s. By 1988, Spanish FDI in Latin America amounted to US$1,967.7 million, reaching US$3,288.5 million in 1990 and US$6,511.7 million in 1999 (Arahuetes and Argüelles, 1996) The significance of this is that São Paulo has more customers than Telefónica’s share in the telecommunications firms of the Southern Cone, Argentina excluded.

References Albuquerque, R. and A. Romão (eds) (2000) O Diálogo dos 500 Anos Brasil-Portugal Desenvolvimento e Cooperação (Rio de Janeiro: EMC Editores). Arahuetes, García A. and Alvarez J. Argüelles (1996) ‘Las Inversiones Directas de España en América Latina, 1981–1992’, Afers Internacionals, 31 January (Barcelona: Fundació Cidob). Bank of Portugal (2000) Relatório anual de 2000 (Lisbon: Bank of Portugal). CEPAL (2000) Panorama de la Inserción Internacional de América Latina y El Caribe, 1999–2000 (Santiago, Chile: United Nations). Durand, M. F. et al. (1992) ‘Europe’, in Le Monde – Espaces et Systèmes (Paris: Presses de la Fondation Nationale des Sciences Politiques), p. 32. European Communities (1999) The Competitiveness of European Industry (Luxembourg: Office for Official Publications of the European Community). Guedes, C. (2001) ‘Competitividade e Inserção Internacional: Algumas Observações Sobre a Economia Portuguesa’, Actas do IV Encontro Internacional de Economia Europeia e Internacional (Lisbon: Instituto Superior de Economia, Universidade Técnica de Lisboa). IADB/IRELA (1998) Inversión extranjera directa en América Latina: la perspectiva de los principales inversores (Madrid: IADB). Lança, I. S. L. (ed.) (2000) A indústria portuguesa–especialização internacional e competitividade (Oeiras: Celta). Martin, B. (1993) In the Public Interest? (London, Zed Books). Mortimore, M. and W. Peres (2001) ‘La competitividad empresarial en América latina y el Caribe, Revista de la Cepal, no. 74, August, pp. 37–59. Ramos Silva, J. (2000) ‘O factor linguístico na economia global’, Revista Gestão e Economia (Universidade da Beira); Special Issue ‘500 anos de Descobrimentos do Brasil’, pp. 182–204. Ramos Silva, J. (2001) ‘ “As Relações Económicas Entre o Brasil e Portugal na Década de 1990”, Os Laços Linguísticos e Históricos à Prova da Globalização’, Revista Internacional de Estudos Políticos, vol. 3, no. 2, August (Rio de Janeiro; UERJ) pp. 46–67.

Cézar Miranda Guedes and Mario Gomez Olivares 233 Scherer, André Forti (1999) ‘Mundialização financeira e investimento directo estrangeiro: notas sobre a experiência brasileira recente’ (Porto Alegre: IV Encontro Nacional de Economia Política, UFRGS). Sweeney, Paul D. and Dean B. McFarlin (2001) Organizational Behavior: Solutions for Management (New York: McGraw-Hill/Irwin). Zapata, F. (1999) ‘Estado Nação, livre comércio e integração económica na América Latina’, in Emprego e Desenvolvimento Tecnológico – Processo de Integração Regional (São Paulo: Dieese).