_s WORLD BANK LATIN AMERICAN AND CARIBBEAN STUDIES THE WORLD BANK 23440 2002 From Natural Resources to the Knowledge Economy -T 1. . . - ~~~~< 7 M$ 1 * ..>-S'w- r /- David de Ferranti, Guillermo E. Perry, Daniel Lederman, and William F. Maloney WORLD BANK LATIN AMERICAN AND CARIBBEAN STUDIES Vzieapoints FROM NATURAL RESOURCES TO THE KNOWLEDGE ECONOMY Trade and Job Quality by David de Ferranti Guillermo E. Perry Daniel Lederman tVilliam E Maloney THE WORLD BANK Washington, D.C. 3 2002 The International Bank for Reconstruction and Development / THE WORLD BANK 1818 H Street, NW Washington, DC 20433 All rights reserved. 1 2 34 014)4 02 01 The findings, interpretations, and concLusions expressed here are those of the author(s) and do not necessarily reflect the views of the Board of Executive Directors of the World Bank or the governments l-hey represent. The World Bank cannot guarantee the accuracy of the data included in this work. The boundaries, colors, denominations, and other information shown on any map in this work do not imply on the part of the World Bank any judgment of the legal status of any territor) or the endorsement or acceptance of such boundaries. Rights and Permissions The material in this work is copyrighted. No part of this work may be reproduced or transmitted in any form or by any means, electronic or mechanical, including photocopying, recording, or inclusion in any information storage and retrieval system, withouc the prior written permission of the World Bank. The World Bank encourages disseminLation of its work and will normally grant permission promptly. For permission to photocopy or reprint, please send a request with complete information to the Copyright Clearance Center Inc, 222 Rosewood Drive, Danvers, MA 01923, USA, telephone 978-750-8400, fax 978-750-4470, wwwv.copyright.com All other queries on rights and licenses, including subsidiary rights, should be addressed to the Office of the Publisher, World Bank, 1818 H Street NW, X i.V..l . DC 20433, USA, fax 20295D222422 e-mail ., .1.I ..r it .- About This Book David de Ferranti is vice president andi Guillermo E. Perry is chief economist in the Worlcl Bank's Latin America and the Caribbean Regional Office. Daniel Lederman is an economist and Willliam E Maloney is senior economist in the World Bank's Latin America and the Caribbean Regional Office. For more information on publications from the World Bank Latin America and the Caribbean Region, please visit us at www.worldbank.org/lacpublications (www.BancoMundial.org/publicaciones). Cover design by Jeffrey Kibler, The Magazine Group. Cover art: Dia/sg.sv uitl the lnieerse, 1999, by Manuel Cunjama. Mexico. Manuel Cunjami was born in 1971 in Tuxtla Gutierrez, Chiapas. Ile began exhibiting his work in Chiapas in 1992. Cunjam&i's work is included in the collection of the World Bank Art Program. The World Bank Art Program makes particular efforts to identify artists from developing nations and make their work available to a wider asidience. The art program organizes exhibits, educational and cultural partnerships, competitions, artists' projects, and site-specific installations. ISBN 0-821 3-5009-9 Library of Congress Cataloging-in-Publication Data From natural resoLurces to the knowledge economy : trade and job quality / David de Ferranti ... [et al.]. p. cm. - (World Bank Latin American and Caribbean studies) Includes bibliographical references. ISBN 0-8213-5009-9 1. Latin America-Commercial policy. 2. Latin America-Economic policy. 3. Labor supply-Latin America. 4. Occutpations-Latin America. I. De Ferranti, David M. II. Series. HF32;0.5.Z5 F76 2001 38.98 -dc21 20010)146929 Contents Acknowledgments ........................................................ ix Acronyms and Abbreviations . ........................................................ xi Chapter 1: Introduction and Summary ....................................................... 1 What Should Countries Rich in Natural Resources Do to Grow Faster and Improve the Quality of Their Jobs? .... . 2 Deepen Openness to Trade and FDI Flows ... and Secure Market Access ... ....... ........... 2 Build New Endowments .................... I................... . I ........................... 3 Play to Your Strengths: Don't Turn Your Back on Your Natural Advantages ............................... 4 Answering Old and New Concerns about Trade ......... . ............................................ 5 What Determines What We Trade: Traditional and "New" Endowments-Natural Resources Are Not Destiny ..... 5 Natural Resources Are Not a Curse but an Asset for Development ....................................... 5 Lessons from Successful Natural Resource-Rich Countries: Play to Your Strengths ................... ...... 6 Prebisch Revisited .......................................................................... 6 Lessons from LAC History: Don't Turn Your Back on Your Natural Advantages ...............8.......... 8 Building New Areas of Comparative Advantage ........................................ ... . . ...... 8 Are the New Jobs Good Jobs? . ..................... .......................... 9 C o n clu sion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . 10 N o te s . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . II Chapter 2: Comparative Advantage, Diversification, and Intra-industry Trade: Determinants and Consequences . 13 New Trade Theories, New Endowments, and New Patterns of Trade .......... .......... 13 New Factor Endowments: Industrial Clusters, Logistics Costs, and ICT and Knowledge .15 Clusters, Geography, and Knowledge .......................................................... 15 Logistics Costs ................ .......................................................... 16 ICT and Knowledge: Measurement and Correlation with Development .................................. 18 Comparative Advantage: Facts from LAC ...................................................... 20 The Regional Average since the Early 1980s ......................... ............................... 23 Specific Country Experiences since the Early 1980s .............................................. 23 Are Traditional Endowm ents Destiny? ........................... ............................ .... 30 Determinants of Trade Structure: International Evidence ............................................. 30 Trade and E ndow m ents ................ ....................................................... 3 1 Econom etric Evidence ........................................................................ 33 Export Diversification and IIT: Where Are We and Does It Matterv ................3...................... 38 Export Diversification and IIT in LAC ............................................................ 38 Export Diversification, IIT, and Growth ....................................................... 40 Export Diversification and IIT: Can Anything Be Done? ............................................. 42 Export Concentration, IIT, and Comparative Advantage: Can Sectoral Policies Help Reduce Concentration and Raise IIT? .................. .. ............... ............ ........... 44 Summary of Findings about Comparative Advantage, Diversification, and IIT ................................ 46 N o te s . . . . . . . . . . . .... . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4 7 Chapter 3: It's Not Just What You Produce, But How: Lessons from Comparative History .49 The Lesson from History: Play to Your Strengths .................................................... 49 FROM NATULRAL RESOU[R(. 1:0 THE KNOWLEDGE ECONOMYY TRADE AND ) JOB QUAL11Y Total Factor Productivity Growth and Devclopment ................................................. 52 Knowledge and Natural Resources: A Proven Recipe for Growtrh.5...................................3..... S Two Final Concerns .................................................. 59 Why Was Latin America's Experience So Disappointing2 ................................................ 61 Initial Condirions ................................................... 1 I Distribution .........................................6........... ....... ...4.... 6A Education and Technical Capacity .......................................... 5 Disruptive Industrialization in the Postwar Period ............ ............................. 66 The Siren's Call of Industrialization ......................................... 6(8 A Turning Away from Natural Resource Strengths .6..............................9........... 6 Are the Successful Models Replicable? The Death of Distan(e, Fragmentation, and First-Mover Advantages .... ....... 7) New Opportunities Appear ................................................................... 73 N o te s . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7 3 Chapter 4: Recent LAC Experiences: The Role of Knowledge and Institutions .............................. 75 Chile's Agricultural Performance: The Case of Fresh Fruits Exports ....................................... 76 Economic Reforms and the Performance of the Agricultural Sector in LAC ............................. . 76 Econometric Results ......................................................................... 76 Reforms and Changes in Incentives in Chilean Agriculture ........ . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 77 Net Rates of Protection . ........................................................................ 7 Research and Development in Chilean Agriculture ................................................... 8 Argentina's and Uruguay's Export Diversification after Liberalization ...................................... 82 A Unilateral atid Rcgional Liberalization in Argentina anlI Uruguay ......................... ........... 82 Diversification Indicators ....................................................................... ', Decomposition of Concentration by Sectors and Regions .............................................84I '. 4 Decomposition of Export Concentration by Sectors ........... ..................................... 8,4 Diversification Indicators by Region: Mercosur and the R,-st of the World .................................. '4 Regional Integration, Economies of Scale, and Transport Costs ......................................... 85 Regional Tariff Preferences: Diversification through Trade Diversion .. Y7 Empirical Evidence ....................................................................... :',7 13razil's Reforms, Manufacturing Productivity, and EMBRAER ...............................89........... 9 Productivity Growth in Brazilian Manufacturing: Review of the Evidence .8...........9....... S9 Brazil's EMBRAER ............................................................I()........... 1)( Exports and Foreign Investment in Central American Countries: Tax Incentives, Institutions, or Human Capital? ......__ {) Exports and Growth during the 1990s in Costa Rica and El Salvador ... 1 Tax Incentives in EPZs: Inevitable Reforms in the Future ............................................. ') FDI and High-Tech Firms in Costa Rica: The Roles of ClINDE and Humian Capital . .... ,7 Human Capital in Costa Rica and El Salvador ............ . .................... ..............1(3 Summing Up: Different Paths of Development in the Future ......................I.................. 1(.3 The Impact of NAFTA on Mexico's Trade Structure ................................................. 11 4 The Structure of NAFTA Trade .. . ................................................. 1 FDI in Mexico: The Role of Geography and Human Capital .......................................... 106 The Mexican Electronics Industry .................................................. . 1S Background and Facts ................................... ............... ()8 Factors Determining Location ................................................... Il) The Dynamics of Agglomeration .................................................. 111 The Role of Education and Knowledge Policies ................................................. 112 Summing Up ............................................ ......................... 112 Tourism and Development in the Caribbean ............................................... ... 3i Tourism and Development ...1............................................... II Possibilities for Growth ..................................15............... l Is Tourism Fickle? ................... I ..... .... ........................116............. I Notes .................................................................. .7............... ]17 Chapter 5: How We Work: Job Quality in Emerging Sectors ........................................ 121 Employment ......................................1......... 21 Wages and Distribution. ........................................1....... 2 Inequality ................................................ 127 lV CONTEN'T'S Labor and Human Capital Supply: The Missing Link .............................................. 131 Sectoral Reallocation ......................... .........1....... ..........................2 132 Informality ................ 132 The Nature of the Informal Sector: A Comprehensive Approach ........... .......................... 134 Does Trade Liberalization Impact Informality? ... .......................... . . .I I. 138 Microenterprises: Improving How They Work in the Global Environment ............................... 141 The Real Problem: Isolation from Domestic and Global Markets ...................................... 142 New Jobs from Trade: Opportunity or Exploitation?R ................................................. 145 EPZ Work .148 Who Works in EPZs? .150 Teleservice Industries. 152 Tourism .152 Nontraditional Agricultural Exports .153 Women in NTAEs .155 Indigenous Groups: How Do They Integrate .156 Conclusion: How Can NWe Continue to Raise Job Quality? .......................... I................... 159 Notes .................................................................6.................. 16 Bibliography: .165 Annex: Selected Figures from Chapters 2, 3, and 4 .179 Boxes Box 2.1 Net Exports as an Indicator of Comparative Advantage Driven by Endowments. . 4 Box 2.2 The Structure of Trade in Large Developing Countries: China, India, and Korea . 29 Box 2.3 Data: Traditional and New Endowments ................................................. .2 Box 2.4 Econometric Methodology: Heckman Selection Models .36 Box 3.1 The Swedish Growth Secret: Innovation Networks and Aggressive Labor .54 Box 3.2 And More Specifically, Knowledge Clusters: The Swedish Pulp and Paper Industry .................. 55 Box 3.3 Finland: From Trees to Telecommunications .............................................. 57 Box 3.4 How Did Australia Come to Run Chilean Mines?. 58 Box 3.5 Why Didn't Chileans Find La Escondida? ................................................ 60 Box 3.6 Prebisch Redux: Are LAC's Terms of Trade Declining? ............................. .... 62 Box 3.7 Canada and Mexico: So Far from God, So Close to a Huge Market .63 Box 3.8 Obstacles to Technical Development in Nueva Granada 67 Box 3.9 Where Do You Go at the End of the Frontier? .......... ................. ...... .... 69 Box 3. 1 0 The Costs of Chilean Industrialization: Frustrated Development in Fishing, Fruit, and Forestry .7 1 Box 4.1 Effective and Net Rates of Protection; Some Algebra ....................................... 78 Box 4.2 R&D Policies and the Emergence of the Fruit Sector in Chilean Agriculture ...................... 80 Box 4.3 Incentives for R&D: Some Policy Issues ..... ............................................ 83 Box 4.4 Mercosur as a Platform for World Exports: New Empirical Evidence ... ........................ 88 Box 4.5 Information Services in Barbados and Jamaica .................... ........................ 96 Box 4.6 FDI and Human Capital in LAC .......................... .. .............. 99 Box 5.1 Losers and Winners in the Original NAFTA ............................................ 123 Box 5.2 Stolper-Samuelson: Why We Thought Unskilled Workers Should Benefit from Liberalization ......... 130 Box 5.3 Will Technological Progress Usher in a Golden Age of Equality? .............................. 130 Box 5.4 Migration, Remittances, and Their Implications ..... .................................... 133 Box 5.5 Are Most Informal Sector Workers Voluntary? ..................................5......... 1 5 Box 5.6 Home-Based Work: Exploitation or Flcxiblc Work Arrangcmcnt? .... .............. I....... 144 Box 5.7 Promoting International Linkages for Artisans .. . . . . 146 Box 5.8 Alfred Marshall Meets Juan Valdez (and Burro) ........................................... 149 Box 5.9 Voices from the Maquila: Is This Exploitation or a Choice? ............ ...................... 15() Box 5.10 What Do Employers Look For? What Do They See?........................................ 154 Box 5.11 We Donot Want Those Bad Jobs. Or Do We? ............................................ 155 Box 5.12 New Jobs, New Gender Roles .......................................5................ 157 Box 5.13 Mexico Village-Town CGE Model .................................................... 159 Box 5.14 Labor Standards in EPZs ....................................................... ... 160 V R(OM NA-'FSRAL RESOURCES TO THf KNOWLUDGE ECONOMY TRADF AND JOB QUALITY Tables Table 2.1 Transport Costs and Infrastructure by Coontry Groups .17 Table 2.2 Transport Costs and Infrastructure: Selected Latin American Countries .17 Table 2.3 Effect of a 1 Percent Transport Cost Reduction on Value Added by Industry: Argentina, Brazil, and Mexico .18 Table 2.4 Indicators of ICT and Knowledge as a Percentage of the United States Levels .19 Table 2.5 Components of Leamrer's 10 Commodity Clusters. 22 Table 2.6 Growth of W orld Machine Exports, 1990-99........................ .................... 23 Table 2.7 Variance of Variables of Interest across and wirhin Countries .3 Table 2.8 The Role of Conventional and New Factors in Accounting for Patterns of Comparative Advantage across Countries during 1976-99 .................... 37 Table 2.9 Does Export-Revenue Concentration Affect E(onomic Growth? Econometric Results .13. 3 Table 2.10 What Can Be Done to Raise IIT? Econometric Results .'. . Table 3.1 Rates of Growth of GDP per Capita, 1820-1 989 .52 Table 3.2 Participants in the Knowledge and Skills Clusters in the Swedish Paper and Pulp Industry ... 56 Table 4.1 Date of Economic Reform, Growth of Labor Productivity and Exports per Worker in Agriculture, Selected Latin American Countries, 1980-99. 76 Table 4.2 Decomposition of Producer Price Changes in Chilean Agriculture .79 Table 4.3 Indicators of Trade Concentration, Argentina and Uruguay, 1986-99 .84 Table 4.4 Issues in R&D Incentive Policies ..4 Table 4.5 Costa Rica and El Salvador: Incentives for FDI .)7 Table 4.6 FDI as a Percentage of GDP and Human Capiral Indicators in LAC Countries: Econometric Results ............................................................ 110l Table 4.7 Perception of High-Technology Firms on the lEducation/Training in Costa Rica .... ............... 101 Table 4.8 Costa Rica and El Salvador: Human Capital Indicators ...................................... I)4 Table 4.9 Tobit Estimates .1............................................................8. Table 4.1 0 Results from a Limited Input-Output Model for Barbados ................................... l 14 Table 4.11 Export Revenue Volatility ............1........1.................................... 1l6 Table 5.1 Average Unemployment Rates in LAC .14................................................. i. Table 5.2 Labor Force Participation Rates in the 1990s .......... .................................. 125 Table 5.3 Home-Based Workforce ............................................................ 1 14 Table 5.4 Business Registration in Selected Countries and by Region ....I......................... I i5 Table 5.5 Characteristics of Maquila/Free Trade Zones . ...........................1............... I i8 Table 5.6 Perceptions of Male and Female Employees in i he Tourism Sector ............................. l z)4 Table 5.7 Estimated Percentage Effects of a 10 Percent Decrease in Grain Prices and Compensating Income Transfer ......... .............................................. I -19 Figures Figure 2.1 Development and Knowledge Index ..................0.................................... ( Figure 2.2 Development and ICT Index ........................................................ : Figure 2.3 Shares of World Merchandise Exports, 1976-99 ................I............................. 7- Figure 2.4 Average LAC Net Exports per Worker by Commodity Groups, 1982-97 ... 5 Figure 2.5 Argentina: Structure of Net Exports, 1980-99 ............................................ 25 Figure 2.6 Brazil: Structure of Net Exports, 1981-99 ........................... .................... .6 Figure 2.7 Chile: Structure of Net Exports, 1981-99 .. . '7 Figure 2.8 Costa Rica: Structure of Net Exports, 1981-99 .. . 27 Figure 2.9 Dominican Republic: Structure of Net Exports, 1981-99 . ' Figure 2.10 Mexico: Structure of Net Exports, 1981-99 .8 Figure 2.11 China: Net Exports per Worker, 1984-99 ....0 Figure 2.12 India: Net Exports per Worker, 1978-98 .0.) Figure 2.13 Republic of Korea: Net Exports per Worker, 1S76-99 . I Figure 2.14 Characteristics of Countries with a Comparative Advantage in Tropical Agriculture, Raw Materials, Petroleum, and Labor-Intensive Manufactures . i Figure 2.15 Characteristics of Countries with a Comparative Advantage in Forestry, Cereals, and Animals ... . 53 Figure 2.16 Characteristics of Countries with a Comparative Advantage in Capital-Intensive Manufactures, Machines, and Chemicals .. . . . 31 Figure 2.17 Marginal (Average) Effects of Country Characteristics on Comparative Advantage .... 35 Figure 2.18 Median Herfindahl Index of Concentration of Exports by Regions .3.. . 39 vi CONTI NT S Figure 2.19 Herfindahl Index of Merchandise Export Revenue Concentration ............................... 39 Figure 2.20 Grubel-Lloyd Index of lIT by Regions, 1980 and 1998 ...................................... 41 Figure 2.21 IIT in LAC Countries Since the Early 1980s ........... ................................... 41 Figure 2.22 Export Concentration and Subsequent Growth .......... .................................. 42 Figure 2.23a Export Concentration and IIT by Sector . ................................................ 45 Figure 2.23b Export Concentration and IIT by Sector . ................................................ 45 Figure 2.23c Export Concentration and IIT by Sector . ................................................ 46 Figure 3.1 Australia: Structure of Net Exports in Selected Induscrialized Countries .50 Figure 3.2 Canada: Structure of Net Exports in Selected Industrialized Countries. 50 Figure 3.3 Finland: Structure of Net Exports in Selected Industrialized Countries . 51 Figure 3.4 Sweden: StrLicture of Net Exports in Selected Industrialized Countries ........................... 5 1 Figure 3.5 Percent Annual TFP Growth (1967-92) ................................................. 58 Figure 3.6 Segmented Trends in Real Commodity Prices, 1900-2()00 .62 Figure 3.7 Literacy Rates in Selected Countries (1870-1925) .65 Figure 3.8 Transportation and Communication Cost (1920-90) .72 Figure 4.1 After Reforms, LAC Countries Caught up with Chile: An Illustration of Econometric Resuls .77 Figure 4.2 Concentration Index Decomposition by Main Aggregates, Argentina Exports, SITC .85 Figure 4.3 Concentration Index Decomposition by Main Aggregates, Uruguay Exports, SITC .85 Figure 4.4 Theil Concentration Indexes for Exports; Argentina, Rest of the World, and Mercosur .86 Figure 4.5 Herfindahl Index for Exports, Uruguay ............................................... 86 Figure 4.6a El Salvador: Composition of Exports, 1995-96 .93 Figure 4.6b El Salvador: Composition of Exports, 1999-2000 .93 Figure 4.7a Costa Rica: Composition of Exports, 1995-96 .94 Figure 4.7b Costa Rica: Composition of Exports, 1999-2000 .94 Figure 4.8 Costa Rica and El Salvador: Exports and Imports from EPZs .95 Figure 4.9 Costa Rica and El Salvador: Share of Exports from EPZs .95 Figure 4.10 Costa Rica and El Salvador: Net FDI ...... ............................................. 98 Figure 4.11 Perception of International Investors on Labor Force Quality ................................. 101 Figure 4.12 Costa Rica: Fraction of Firms Using Training Programs .................................... 1(2 Figure 4.13 Costa Rica: Composition of the Workers in the High-Technology Firms ........................ 102 Figure 4.14a Costa Rica: Composition of Engineers in the High-Technology Firms .......................... 103 Figure 4.14b Composition of Technicians in the High-Technology Firms ............. .. ................... 103 Figure 4.15 Costa Rica: Supply of New Engineers in the Areas in Most Demand by New High-Technology Firms . . . 104 Figure 4.16 Costa Rica and El Salvador: Public Expenditures in Education ................................ 105 Figure 4.17 Mexico: Structure of NAFTA Net Exports .............................................. 106 Figure 4.18 Mexico's IIT with North America: NAFTA, IIT, and the MIIT Index .......................... 106 Figure 4.19 Mexico: Geographic Distribution of Variables ............................................ 107 Figure 4.20 Tourism's Share of Exports and GDP . .................................................. 113 Figure 4.21 LAC's Share and Rates of Growth . ..................................................... 114 Figure 4.22 Real Tourism Receipts per Tourist Day .................. ............................... 115 Figure 5.1 Composition of Labor Force .122 Figure 5.2 Unemployment and Openness, 1998 .123 Figure 5.3 Unemployment and Resource Intensity, 1998 .124 Figure 5.4 Chile: Steady Manufacturing Employment Growth after Reforms .125 Figure 5.5 Change in Women's Labor Force Participation Rates by Household Structure .126 Figure 5.6 Change in the Ratio of Women's Wage to Married Men's Wage .126 Figure 5.7 Current Raw Wages by Leamer Category .127 Figure 5.8 Wages by Leamer Category Adjusted for Human Capital .127 Figure 5.9 Current Raw Wages by Trade Exposure .128 Figure 5.10 Wages by Trade Exposure Adjusted for Human Capital ..................................... 128 Figure 5.11 Wage Inequality Spread ........................................................... 129 Figure 5.12 Relative Labor SupplY and Wage of Workers with No Schooling Compared to Workers with Lower Secondary Schooling in Brazil ................................................. 131 Figure 5.13 Relative Labor Supply and Wage of Workers with Tertiary Schooling Compared to Workers with Upper Secondary Schooling in Brazil ...... ................... ...................... 132 Figure 5.14 Remittances as a Share of GDP and Exports, 1999 ........................................ 133 Figure 5.15 Earnings Inequality Decomposition for Salaried versus Self-Employed Workers, 1995 ............... 134 I SROM NATIJRAL RESOUR(.FS TO tiHE ;KNOtWLEDGE ELCONOMY: S RADi AND JO)B QLUALI Tf Figure 5.16 Mexico-Reasons to Start a Microenterprise ............................................. 1 35 Figure 5.17 Self-Employment versus Indusrrial Productivir-y .......................................... 116 Figure 5.18 Relative Formal/Informal Sector Sizes and Incomes and the Real Exchange Rate . . .................. 117 Figure 5.19 Relative Formal/Informal Sector Sizes and Income and the Real Exchange Rate, Colombia .... ....... 1 i8 Figure 5.20 Impact of Minimum Wages on the Distribution of rhe Wages of Informal and Formal Salaried Workers ......................................................... I Uq Figure 5.21 Formality Rates by Trade Exposure .................0................................. I i( Figure 5.22 Formality Rates by Leamer Categories ................................................. I i Figure 5.23 Domestic and Foreign Businesses as Microenterprise Clients, Mexico and the Dominican Republic. 1 1I Figure 5.2-4 Informality Rate for Salaried Workers in Greater Buenos Aires, 1980-99 .............. ......... I i2 FigtIre 5.25 Share of Informal Workers in Greater Buenos Aires ........................................ I i Figure 5.26 Premium on Reported Profits of Mexican Microfirms with Firms as Clients Compared to Other Microenterprises ........ ................................................. I i 3 Figure 5.27 Hourly Wage Premium to Working outside the Home Compared to Home-Based Work .... ........ I i l Figure 5.28 Microenterprise Access to Finance .......... .. . .. . . . .. . .. . . .. . . .. . .. . . .. . . .. . . .. . . I i7 Figure 5.29 Dominican Republic: Earnings Distribution by Sector; Selected Sectors .................... 1 i9 Figure 5.30) Total Maquila Employment and Women's Shaire in Mexico ........... ....................... 1) I Figure 5.31 Mexico: Change in Wages and Women's Share of Employment in the Maquila Sector, 1980 98 ....... I U1 2 Figure 5.32 Bolivian Highlands: Exports Determinants .............................................. 1`18 Figure A. I Average LAC Net Exports per Worker by Commodity Groups, 1982-97 .. I81 Figure A.2 Argentina: Structure of Net Exports, 1980-99 I.8 Figure A.3 Brazil: Structure of Net Exports, 1981-99 .'2 Figure A.4 Chile: Structure of Net Exports, 1981-98 .12 Figure A.5 Costa Rica: Structure of Net Exports, 1981-99. .1S Figure A.6 Dominican Republic: Structure of Net Exports, 1981-99.. I 1 Figure A.7 Mexico: Structure of Net Exports, 1981-99 i.I Figure A.8 Australia: Structure of Net Exports iii Selected Industrialized Countries. I i Figure A.9 Canada: Structure of Net Exports in Selected Industrialized Cotntries.. 1 5 Figure A. 10 Finland: Structure of Net Exports in Selected Industrialized Countries ..........................8 I,'5 Figure A. I I Sweden: Structure of Net Exports in Selected Irndustrialized Countries .......................... 1$1'6 Figure A.12 Mexico: Structure of NAFTA Net Exports .. .......................................... 186 Acknowledgments T H HIS REPORT IS THE PRODUCT OF A COLLECTIVE EFFORT BY A WORLD BANK TEAM COORDI- nated by the Office of the Chief Economist for the Latin America and the Caribbean Region and led by Guillermo E. Perry. The principal authors by chapter were Guillermo E. Perry (Chapter 1), Daniel Lederman (Chapters 2 and 4), and William F Maloney (Chapters 3 and 5). We were honored by the knowledge and talents of specialists (and their numerous coauthors) who wrote background papers: Patricio Aroca, Magnus Blomstrom (Stockholm School of Economics), Eric Bond (Penn State University), John Cuddington (Georgetown University), Pablo Fajnzylber (Federal University of Minas Gerais, Brazil), William Foster (Catholic University of Chile), Rolando Guzman, Arn Kokko (Stockholm School of Economics), Magdalena Lizardo, Catherine Mann (Institute for International Economics), Patricio Meller (University of Chile), Alex Monge-Naranjo (Northwestern University), Pablo Sanguinetti (Torcuato di Tella University, Argentina), Rodrigo Soares (University of Chicago), J. Edward Taylor (University of California, Davis), Alberto Valdes, Anthony Venables (London School of Economics), and Gavin Wright (Stanford University). Friends and colleagues from the World Bank made ef- appreciation for the helpful input received from Luther forts that went well beyond the call of duty. We received Miller of the Caribbean Tourism Organization (CTO). important contributions from: Andreas Blom, Wendy The report benefited greatly from discussions with a Cunningham, Norbert Fiess, Jose R. L6pez-Calix, Ana wide range of civil society organizations and individuals. Maria Menendez, Gabriel Montes, Marcelo Olarreaga, More than 100 representatives from think tanks, trade Susana Sanchez, Norbert Schady, Claudia Sepulveda, unions, community-based organizations, women's move- Quentin Wodon, and Lixin Colin Xu. In addition, valuable ments, Afro-Latino communities, advocacy groups, and comments were received at various stages of the project development nongovernmental organizations (NGOs)- from Ian Bannon, Mauricio Carrizosa, Paul Collier, Shelton among others-helped us build a bridge between theory Davis, Indermit Gill, Marcelo Giugale, Ali Khadr, Paul and practice. We conducted a series of consultation meet- Levy, Will Martin, Ernesto May, Lee Morrison, Anne Pillay, ings in Brazil, Colombia, the Dominican Republic, and David Rosenblatt, Carolina Sanchez, Maurice Schiff, Luis Mexico, which provided us with useful information to both Serven, Will Shaw, David Yuravlivker, and Cara Zappala. formulate and test our hypotheses. In addition, more than Important contributions and guidance were received 30 participants from several countries in the region were from Nina Pavcnik at Dartmouth College, and at the ini- engaged in a productive electronic dialogue, which guided tial stages from Ken Sokoloff at the University of Cali- the design of the study. Special recognition should be fornia, Los Angeles. We would also like to express our given to Dr. Ignacio Roman, who arranged further consul- ix UROM NATL RAL RES()UER FS TO THE KNOWLEDGE ECONOMY TRADE AND JOB QUALITY tation and field visits in Jalisco, Mexico. These consulta- Any errors or omissions in the report arc the solc rc- tions could not have taken place without the support of our sponsibility of the authors and should not be attribute d Civil Society team, especially Kathy Bain and Roberto to any of the above individuals or the institutions th,;-y Senderowitsch. represent. Acronyms and Abbreviations ANOVA Analysis-of-variance CBI Caribbean Basin Initiative CEO Chief executive officer CEPAL Centro Estudios para America Latina CET Common External Tariff CGE Computable general equilibrium CINDE Coalici6n Costarricense de Iniciativas de Desarrollo CONASUPO Compafifa Nacional de Subsistencias Populares CORFO Chilean Development Corporation CTA Centro Tecnico da Aeronautica CTO Caribbean Tourism Organization ECLAC Economic Commission for Latin America and the Caribbean EMBRAER Empresa Brasileira de Aeronautica EMS Electronics manufacturing services EPZ Export processing zone ERP Effective rate of protection FDI Foreign direct investment FIA Ministry of Agriculture's research fund FONDECYT Fondo Nacional de Desarrollo Tecnol6gico y Productivo, Chile FONDEF Fondo de Fomento al Desarrollo Cientifico y Tecnol6gico, Chile GATT General Agreement on Tariff and Trade GDP Gross domestic product GMM Generalized Method of Moments ICM Sales tax (Brazil) ICT Information and communications technology IIT Intra-industry trade IMF International Monetary Fund INIA Instituto Nacional de Investigaci6n Agraria (National Institute of Agricultural Research) IP Internet Protocol IPI Industrial products tax (Brazil) ITA Instituto Tecnol6gico da Aeronautica LAC Latin America and the Caribbean LDC Less-developed country Mercosur Mercado Comuin del Sur MIIT Marginal intra-industry trade xi FIRI)bM NA Ft RAL RFSOtLR( LS IO TIlE KN(\WLED(IF ECONOMY TR ADE AND jOB Ql ALI TY NAFTA North American Free Trade Agreement NIC Newly industrialized country NRP Net rate of protection NTAE Nontraditional agricultural exports OECD Organisation for Economic Co-operation and Development OEMs Original Equipment Manufacturers PCB Printed circuit board PROCAMPO Programa de Apoyos Direcros al Campo PROCOMER Promotora del Comercio Exterior de Costa Rica (The Costa Rican Foreign Trade Corporation) R&D Research and development RIA Regional integration agreements SITC Standard International Trade Classification TFP Total factor productivity UNESCO United Nations Educational, Scientific and Cultural Organization USAID United States Agency for International Development WTO World Trade Organization xii CHAPTER 1 Introduction and Summary FTER OVER A DECADE OF REMARKABLE PROGRESS IN TRADE REFORM, OBSERVERS IN A Latin America and the Caribbean (LAC) are again asking a question of long standing in the region: Does the way in which the region integrates into the global market promise rapid growth and good jobs for its workers? Indeed, an old and central concern has re-emerged. In most LAC countries their rich natural resource endowments are still determining what they export. This is happening despite earlier predictions that LAC would follow the Asian manufacturing suc- cesses, and despite the widely heralded arrival of what we loosely call the "knowledge economy." The worry is that continued specialization in natural resources will leave LAC behind in the slower "old" economy. More fundamentally, what determines what LAC will export in the 21st century? Traditional concepts of comparative advantage stress endowments of labor, land, capital, and natural resources. However, "new" trade theory stresses that at least as important are other "endowments" such as geography (distance to large markets), technical knowledge, human capital, public infrastructure, quality of institutions, and more gen- erally the ability of firms to provide the right products to the right markets at the right time. Further, the spectacular decline in transport and communications costs over the last century has led to the "fragmenta- tion" of production processes once performed in a single country into multiple stages of production dis- tributed around the world. This fragmentation may imply tures are over; that is, are these emerging jobs-in Export that some successful development trajectories are now fore- Processing Zones (EPZs), or myaquilas, long-distance teleser- closed, and may rekindle the concerns of the old depen- vices, nontraditional agriculture, and tourism-good jobs? dency school that LAC will be stuck with the nondynamic These themes are the subject of this, the annual World parts of the global production process. Could natural re- Bank publication for Latin America and the Caribbean. The sources not only be a curse, as some have argued, but now report contributes to the debate in two ways. First, it takes also LAC's inescapable destiny? Or by making increasingly stock of where LAC economies are as participants in the unpredictable the ultimate location of industries, do the global economy by reviewing the trends in the region's new facts and trade theories suggest opportunities for the exports, drawing lessons from close examination of emerg- region that were previously not imaginable? ing sectors, and locating the present reality in a historical The parallel concern is that recent decades have seen a and international context. Second, it looks at the bottom diminishing quality of jobs measured in terms of wages, line of any trade policy: how workers and families have benefits, and working conditions. The most notable fact is fared, and in particular, how much of the movements in the untold personal hardship caused by job loss in indus- income distribution and informality that took place in the tries that could not compete internationally. The longer- last decade can be attributed to trade, what kinds of jobs are term questions are about what can be expected when the being created, and who benefits from them. The central dislocations of the transition to the new economic struc- question we ask is, then, what strategies and policies should I RUNt NATU RAEL RESOURCLs TO 'I'HE KNO)WLEDCE Ei ()NOMY TR DOFE ANi) JOB Q lALATY a natural resource-rich country follow in order to develop should they bridge their natural resource wealth with the and to improve the quality of jobs? new knowledge economy? The report attempts to provide answers to a series of both positive and normative questions related to this fundamen- What Should Countries Rich in Natural tal concern. Specifically, we address a series of questions Resources Do to Grow Faster and Improve related to trade, natural resources, and growth in Chapters the Quality of Their Jobs? 2 to 4: Will LAC countries continue to specialize in com- The policy recommendations that stem from this report modities? Is their natural resource wealth their destiny? Can can be conveniently grouped into three categories: they build comparative advantages in technologically sophisticated economic activities related or unrelated to their * Foster openness to trade, market access, and FDI flows. natural resource wealth? Is their natural resource wealth a * Build new endowments in human capital, knovl- curse for development? Or, on the contrary, are natural re- edge, better institutions, and public infrastructure. sources assets for development? Is there or is there not an * Play to your strengths: don't turn your back on your inescapable dilemma between the old economy based on natural advantages. natural resources and the new knowledge-intensive econ- omy of the 21st century? Can natural resource-based activ- Deepen Openness to) Trade and FDI Flons ... ities be knowledge-intensive industries that lead develop- and Secure Market Access ment for a long time? In Chapter 5 we also address two We find in this report that opening to trade and FDI has led questions related to trade and jobs: Are the new jobs emerg- to export diversification (both within and outside natural ing from free trade "poor-quality" jobs? Has free trade resource-based activities) and to higher intra-industry trade increased unemployment, informality, and wage inequality? (IIT) in the region, and we present evidence showing that We attempted to answer these questions through a vari- such developments have been causally related to highe r ery of approaches. We reviewed what new trade theories and future growth elsewhere. We also find, as many previous recent empirical studies have to say on these subjects. We studies have, that greater openness and FDI promote inno- looked at available quantitative evidence and conducted vation and skills, and permit realizing the potential of nat- cross-country panel data and econometric estimations to ural advantages such as natural resource wealth, but also of validate or contradict predictions from theory. We commis- such factors as location, size, natural beauty, and cultur:'. sioned papers to draw lessons from comparative history Finally, we find that new jobs created by trade and FIF (Australia, Finland, Sweden, and the United States com- flows are "good jobs." Thus, Latin American countries pared with LAC), and case studies of recent LAC experiences would do well to deepen their efforts toward greater trade (for Chile, new agricultural exports; for Mexico, the North integration and attraction of FDI flows that have characte-- American Free Trade Agreement [NAFTA] and the elec- ized the latter decade. tronic complex in Jalisco; for Brazil, industrial productivity At the same time, however, we find that unilateral open- growth and aircraft exports; for Mercosur, export diversi- ing is often not enough. Protectionism in developed coun- fication in Argentina and Uruguay; for Costa Rica and El tries and developing countries alike limits the potential of Salvador, foreign direct investment (FDI) and EPZs; for the export expansion in several activities in which LAC couI-I- Caribbean, tourism) by leading experts in the field and well- tries possess comparative advantages (agriculture, process- known local consultants. ing of raw materials, textiles and apparel, and services:'.' We summarize our findings and answers to these ques- Uncertainty about continued market access often inhibits tions in the following sections. But before doing so, and strong FDI and trade flows. In other words, secure market because we are a policy-oriented institution, we offer up access is crucial for realizing the potential of trade arid front a summary of policy recommendations from our investment flows. Indeed, this report shows that NAFTA answers to the main normative questions related to these has contributed decisively to a large increase in Mexican concerns: What should countries rich in natural resources trade and to a fundamental change in its production and do to develop and to improve the quality of their jobs? trade structure, taking advantage of its locational advan- Should they turn away from their natural resources in tages and the opportunities for fragmentation of produc- attempting to develop new comparative advantages? How tion opened by the sharp reduction in transport and tele- INTRODUCTION AND SUMMARY communication costs. Also, the Caribbean Basin Initiative ments, such as human capital, knowledge, and good insti- (CBI) has permitted some Central American and Caribbean tutions and public infrastructure, explain as much, if not countries (such as Costa Rica, the Dominican Republic, more, of the comparative advantages of countries and their and El Salvador) to better exploit their locational advan- evolution over time, as traditional factor endowments such tages, though not to the same degree as Mexico. as land, labor, and physical capital. Public policy has a large Thus, LAC countries could obtain significant gains from role in building up such endowments.2 a new development round centered in agriculture and ser- * General Education and Lifelong Learning. Policies vices, and from further free trade agreements with Organi- aimed at developing educational systems that pro- sation for Economic Co-operation and Development (OECD) vide quality education in general, but that are also countries. The advancement of current negotiations with focused on lifelong learning and training, are critical the United States toward bilateral trade agreements and to building human capital. This is, in fact, a require- the Free Trade Zone of the Americas, and with the Euro- ment to ensure that our workers can continuously pean Union and Japan, acquire significant importance in upgrade how they produce in whatever sector they this context. However, it should be said that other LAC might be employed. It is also a prerequisite for the countries may not benefit as much as Mexico from free emergence of strong innovation systems. trade agreements with the United States, because they do * Research and Development Incentives and Innova- not possess the same locational advantages. tion Systems. Knowledge and technical progress are We also find that trade agreements among LAC coun- the main force behind productivity growth and the tries can foster export diversification. Indeed, our study on emergence of new comparative advantages. This is an Mercosur reveals that it has served as a platform for the area where the public sector should be involved, due development of new exports to third markets. It is thus to the high spillover effects of knowledge that are useful to deepen such agreements, though because they unlikely to be promoted to their socially optimal may also entail some trade diversion costs, it is important level by private firms alone. This report argues that to keep external tariffs low and relatively uniform. Open countries should experiment with various types of tax regionalism continues to be a good policy stance (see Burki, incentives and public subsidies to promote both pri- Perry, and Calvo 1997). vate and public investments in research and develop- Establishing an enabling environment for FDI is an ment (R&D). However, the precise design of such important complement to open trade policies. We find that policies will depend on the existing institutional FDI flows respond to better institutions, human capital, pub- capacity of the governments to enforce tax laws and lic infrastructure, and knowledge clusters (see below). But we monitor the quality of those investments. Adequate also find that special regimes (like EPZs) can serve a useful intellectual property rights protection is also essen- purpose in countries with weak institutional infrastructure, tial for the development of innovations. In addition, because they help reduce transaction costs and enhance the public policy should promote the establishment of credibility and predictability of policies to attract FDI. Nat- knowledge clusters and networks, encompassing pri- urally, such regimes must conform with World Trade Orga- vate firms, independent research institutions and nization (WTO) agreements, and thus some of the tax incen- universities, and the public sector. Such develop- tive schemes that have been used by some countries will have ments are critically important not just for the so- to be phased out by 2003. Finally, we find some evidence in called high-tech manufacturing sectors, but also for favor of active promotion programs, such as those developed many natural resource-based activities. by the Coalici6n Costarricense de Iniciativas de Desarrollo * The Special Role of Information and Communica- (CINDE), a public-private venture in Costa Rica. tions Technology. The evidence presented in this report suggests that information and communica- Build New Endowments tions technology (ICT) can help reduce coordination A crucial finding of this report is that countries can and do costs, which might enable firms to be more effectively create new comparative advantages through policies that plugged into international industrial clusters, ensure build new endowments. Indeed-validating the predic- the dynamism of the tourism industry, and facilitate tions of new trade theories-we find that "new" endow- access to markets for nontraditional agricultural 3 FROM NATURAL RESOURCES TO THE KNOWLEDGE ECONOMY TRADE AND JOB QUALITY products. The role of the public sector in this area is Natural resource-based activities can be knowledge in- to provide the appropriate enabling environment to dustries. Mining was the "national learning experience" in allow the development of telecommunications infra- the United States that led to building a strong technolc,gi- structure (for example, telephone lines and cellular cal system from which modern manufacturing developed. telephony). That is, regulations should aim at raising Forestry and forest products are highly knowledge intensive the level of competition and access in domestic in Finland and Sweden, which explains why they remain economies for the provision of these types of services. competitive compared to countries with much lower wages. * Public Infrastructure. Good public infrastructure is Fresh-fruit production and marketing in Chile has a high- essential to reduce transportation costs and "economic tech content. Biotechnology is as high tech as chips and distance" to markets. We find that it is especially semiconductors, if not more. important for the development of natural resource- Even more, the stock of natural resources themselves c:an based exports. In this context, improving the regula- be increased by the application of knowledge. This is es- tion of private investment in infrastructure appears as pecially evident in the case of oil. Just 30 years ago most a major policy priority. analysts predicted that supplies from economically viable * Good Institutions. Last but nor least, good instiru- potential oil reserves would he limiting world growth by tions (rule of law, security, effective property rights, the end of the century. Thanks to advances in exploration transparency, and doing away with excessive regula- and extraction techniques, proven oil reserves are today tory burdens, and efficiency of public service delivery) much larger than they were then. Venezuela significantly are critical for the efficiency of investment and pro- increased its economic oil reserves through the develop- ductivity growth in all sectors of the economy, but ment of oriemulsion, allowing the commercial exploitation especially for developing comparative advantages in of large heavy crude reserves. Today Finland and Sweden technologically sophisticated industries. have more forests than they did a hundred years ago. The evidence presented in this report indicates that pro- Play to Your Strengths: Don't Turn Your Back on Your ductivity growth in agriculture has outpaced that of man- Natural Advantages ufacturing in both developed and developing countries. This report finds that a major mistake in LACs economic Productivity growth in agriculture in Latin America has history has often been to turn its back to its natural advan- increased significantly after trade reform, though in many tages, whatever they may be: natural resources wealth, lo- countries there was an initial period of decrease. In short, cation, natural beauty, culture. Such attitudes have been natural resource-based activities can have high productiv- largely driven by prejudices about the alleged superiority of ity growth, technical spillovers, and forward and backward manufacturing activities compared to natural resource- linkages, as much as modern manufacturing. Such activi- based activities. The evidence presented in this report should ties can become knowLedge industries. They are in no way dispel such prejudices and other concerns about natural incompatible with the new knowledge economy. resource-based activities. What we found is that the key to success is to complc- Indeed, natural resources-based activities can lead growth ment natural resource wealth with good institutions, human for long periods of time. This is patently evident in the capital, and knowledge. Natural resources and knowledge are development history of natural resource-rich developed a proven growth recipe. Developing dynamic natural countries, such as Australia, Finland, Sweden, and the resource-based sectors is not incompatible, either, with United States. Mining was the main driver of growth and building new comparative advantages in footloose and high- industrialization in Australia and the United States over tech manufacturing. We find both sectors coexisting not only more than a century, as forestry has been in Finland and Swe- in natural resource-rich developed economies, but also in the den. These countries continue to be significant net exporters already highly diversified export structure of Brazil and Mex- of natural resource-based products, along with high-tech ico, and in small Costa Rica and the Dominican Republic. To products. The recent success of Chile, with the highest what extent successful development will be based more in growth rate in the region in the last 15 years, has been almost natural resource-based activities or in footloose manufactur- fully led by exports of natural resource-based products. ing or service activities will depend largely on other trade 4 INTRODUCTION AND SUMMARY advantages such as location or size. Countries close to the that, especially for manufactures but also for certain agri- United States will more likely develop strong manufacturing cultural products, the "new" endowments explain a larger links with that market through opportunities for fragmenta- share of the international differences in comparative advan- tion of production, especially because they have been able to tage than traditional endowments. secure market access through NAFTA and CBI-NAFTA par- Falling transport and communication costs have spawned ity. Small economies will not likely evolve as diversified pro- new industries such as specialty vegetables in Guatemala, duction and trade structures as large Brazil or Mexico. But all flowers in Colombia, fresh fruits in Chile, teleservices in of them will integrate successfully into the world economy if Jamaica, and tourism in previously isolated locales. They they play to their strengths and complement their natural have also enabled the fragmentation of production processes advantages with human capital, knowledge systems, public that were previously largely contained in the industrialized infrastructure, and good institutions. countries. Indeed, the statistical evidence shows that com- parative advantage in labor-intensive manufactures exhibits Answering Old and New Concerns about Trade a higher over-time variation during the last quarter century than all other commodity groups examined in this report. What Determines What We Trade: Traditional This fact can be interpreted as indicating that developing a and "New" Endowments-Natural Resources comparative advantage in these types of goods is more likely Are Not Destiny to occur than for other commodities. Not only do countries At a very aggregate level, the export profiles of most of the like El Salvador and Mexico provide the labor-intensive countries of the region have, with few exceptions, changed subprocesses of many globally assembled products, but remarkably little since the early 1980s. The finding that Guadalajara has become a printer design center for Hewlett- LAC remains an exporter of goods intensive in natural Packard, and Costa Rica now hosts Intel's most recent chip- resources is undeniable, and therefore merits special atten- manufacturing plant, with spillovers in design and software. tion. But it is also true that countries like Costa Rica, the However, pronouncements of the "death of distance" have Dominican Republic, and Mexico have experienced dra- been overstated. It is not coincidental that these countries matic changes that are as informative about the possible are close to the largest global market. future trade patterns of the region. Trade openness, together with the accumulation of The cross-country evidence presented in Chapter 2 in- "modern" endowments, has also helped diversify exports dicates that the structure of trade is, indeed, determined and increase IIT (that is, the share of total trade accounted by traditional notions of factor endowments including la- for by exports and imports of similar products). Indeed, bor, capital, natural resources, and land. However, the data there was significant export diversification and an increase also show that modern concepts of national endowments in IIT in LAC during the 1990s, the latter encouraged by emphasized by modern trade theory also have substantial higher levels of education, openness, better institutions, explanatory power. It is not surprising that the comparative and infrastructure. Regional integration efforts also con- advantage in resource-based activities is mostly determined tributed to the transformation of the patterns of trade and by natural resource endowments, and that labor-intensive to increasing export diversification and IIT. This conclu- manufactures appear where labor is plentiful. In contrast, sion is supported by the evidence discussed in our case net exporters of capital-intensive goods, machinery, and study of the influence of NAFTA on the Mexican economy, chemicals have much higher levels of "modern" endow- and to a lesser extent by our study of the effects of Merco- ments, especially knowledge in the case of capital-intensive sur on Argentina. activities, and knowledge, education, ICT, and quality public institutions in the cases of chemicals and machines. Natural Resources Are Not a Curse but an Asset But what is striking is that net exporters of forestry prod- for Development ucts and cereals also have high levels of these nontraditional However, most countries find themselves again concentrat- factors. Hence, the endowments of the "knowledge econ- ing in natural resource-based products, and this has raised omy" also seem to be important for some natural resource- concerns dating back centuries. No less an economist than based activities. Overall, the statistical evidence shows Adam Smith wrote in 1776 that "projects of mining, 5 FROM NATLJRAL RESOLFRCES TO TIHfE KNOWLEDGE ECONOMY TRADE AND JOB QUALITY instead of replacing the capital employed in them, together ative industries of both, made Australia one of the richest with the ordinary profits of stock, commonly absorb both economies in the world by the early 20th century. The dis- capital and stock. They are the projects, therefore, to which coveries of new deposits and the generation and export of of all others a prudent law-giver, who desired to increase mining-related knowledge-in mineral detection, environ- the capital of his nation, would least choose to give any mentally sound mining practices, and processing, all based extraordinary encouragement ...... The best-known cri- on a massive educational and research infrastructure-riay tiques of natural resource-based development in LAC came put it near the top of the list again. One of Australia's most from Prebisch (1959) and Singer (1950). They were con- famous mining firms, Broken Hill Propriety, Ltd., estab- cerned about an apparent long-term deterioration of the lished long after Chile began copper mining, discovered prices of these commodities relative to the price of manu- and manages Chile's largest mine, La Escondida. factures imported by the region. They also believed that The Scandinavian countries that produce aircraft, luxory technical progress opportunities were limited relative to cars, designer furniture and, most recently, advanced tele- those in manufactures. In modern times, Jeffrey Sachs and communications products, also built slowly on their Andrew Warner (1995a) from the Harvard Center for In- strengths in natural resources. Sweden's Volvo and Saab ternational Development found that during two decades of emerged partially as backward linkages of the forestry in- the late 20th century, 1970 to 1989, countries rich in nat- dustry, but perhaps more interesting is how Nokia, origi- ural resources grew more slowly than their counterparts. nally a wood-pulp producer, became a major player in the This case is not strong. First, the sample period is not global cellular phone industry. The key elements were orga- historically representative. As Maddison (1994) shows, nizational structures, knowledge networks, and aggressive from 1913 to 1950, resource-rich countries, including LAC human capital policies that, though developed to pursue countries, grew faster than the then-industrialized coun- resource processing, were transferable to high-tech indos- tries, and Asia experienced negative growth. Further, Sachs tries. In short, the historical record is clear: when managed and Warner's period includes the "lost decade" of the well and located in the proper institutional framework, nat- 1980s, which resulted from abuse of easy foreign indebted- ural resources can be vital for development. ness in the 1970s, the traumatic demise of the protection- ist model of development, and the wrenching transitions Prebisch Revisited to more open economies. Though some of these factors are What happened to Prebisch's twin concerns in light of this correlated with natural resource abundance, none are in- historical experience? If anything, the modern literature (n trinsic to it. decreasing transport costs and agglomeration externalities (Puga and Venables 1999) offers a formalization of depea- Lessons from Successful Natural Resource-Rich dency theory and possible confirmation of its diagnosis. Countries: Play to Your Strengths If manufactures are a more dynamic sector than natural The most convincing evidence is offered by history: it is resource-based sectors, either, to paraphrase Prebisch, be- impossible to argue that Australia, Canada, Finland, Swe- cause of decreasing relative prices or lower possibilities for den, and the United States did not base their development technological progress, then decreasing transport costs may on their natural resources. In fact, even today they are net lead to a "deindustrialization" of the South that reduces its exporters of natural resource-based products. long-term dynamism. The evidence is now strong that the U.S. industrial suc- But the emerging literature and several papers commi s- cess resulted from a gradual transition to resource-intensive sioned for this study suggest that on both counts Prebisth manufacturing industries, and only quite late in its devel- was too pessimistic. First, the worsening terms-of-trade opment to more knowledge-intensive industries. Canada effect was probably a false alarm. Economic historians have was the country that inspired "staples theory"-where pri- documented increasing relative commodity prices through- mary goods exports drive development over an extended out the 19th century. Most recently, Harvard historian Jeff period of time through either demand or supply linkages. Williamson has argued that Prebisch was probably wrong Although wool is Australia's most famous staple, extra- about their decline at the beginning of the 20th century. ordinary and continuing success in mining, and the deriv- Rapidly decreasing transport costs made LAC's commodi- 6 INTRODLUCTION AND SUMMARY ties appear relatively cheaper in London, where Prebisch sectors where the country has a clear comparative advantage. measured them, but the reverse would have been true in R&D policies should not underestimate the innovative the Port of Buenos Aires. Finally, a paper by Cuddington, potential that can be unleashed in agricultural activities. Ludema, and Jayasuriya (2001), commissioned for this proj- Mining or forestry can be a low-innovation sector with ect, concludes that from the beginning of the century to few possibilities for long-term growth, but in the cases 1973, there was no trend in relative commodity prices. mentioned above, this was emphatically not the case, nor is Second and more important, the view that manufactur- it the case today. Tree harvesting, pulp and paper process- ing has something special (in terms of backward and for- ing, and other downstream industries continue to show ward linkages, technological innovation, and other poten- productivity gains in Finland and Sweden. As Wright tial externalities), must be called into question-as many (1999) argues, "mining, in the U.S. case, was fundamen- contemporaries of Prebisch did (such as Viner 1952 and tally a collective learning phenomenon," and this applies to North 1955). More than what is produced, the important the Australian experience as well. Natural wealth itself is issue is whether there is an enabling environment for not fully exogenous, and the stock of economically useful adopting technologies. Modern growth literature increas- oil or mineral reserves (or productive land) is a function of ingly stresses the development of "national innovative knowledge and improvements in exploration and extrac- capacity," or the ability of countries to produce and com- tion techniques. Brazil's discovery of vast new deposits of mercialize knowledge over the long term. This capacity is iron ore, bauxite, tin, and copper using satellite technolo- determined by interrelationships among a variety of social gies confirms this insight. institutions (universities, research centers) and actors (pri- Tourism is a natural beauty-based growth industry vate firms, the public sector). worldwide, with expenditures growing by 3.8 percent Evidence is offered by Martin and Mitra (2001), who annually in the last two decades. The case study included examine total factor productivity (TFP) growth-the part here suggests that in several Caribbean countries it supports of growth unaccounted for by the increase in measurable high living standards and has been an important substitute factors such as land, labor, or capital-in agriculture and for the declining sugar and banana industries. LAC has manufacturing from 1967 to 1992 for a large sample of slipped in market share over the last decades, but could countries. Not only was TFP growth 50 percent faster in increase the sector's dynamism with an aggressive program agriculture than in manufactures, but the industrialized of product differentiation (ecotourism, historical and cul- countries experienced rates substantially above those of tural tourism, adventure tourism) and infrastructure, en- less developed countries (LDCs). In fact, several of the big hancing skills, ICT, and marketing research to compete natural resource success stories-Denmark, France, and with new locations. Again, a successful strategy requires not Sweden-continue to show the highest TFP growth rates only the natural resources, but also complementary inputs in agriculture. in human capital, technical knowledge, and infrastructure. Work conducted for this book on LAC agriculture con- In a nutshell, both the historical and recent LAC experi- firmed that agricultural productivity increased in all coun- ences reviewed in this report reveal that what is important tries after trade reforms. 't'he most successful case in terms is not uhat countries produce, but how. In all these cases, a of output per worker and export growth during 1980-99 in large knowledge network (or cluster) that generates innova- agriculture was Chile. Part of this successful experience was tion and facilitates the adoption of foreign technologies, due to adequate macroeconomic management and, espe- stands out as a critical ingredient in sectoral dynamism. cially, a competitive and relatively stable real exchange rate. Finally, the concern that natural resource-intensive As important was Chile's long history of experimenting exports will crowd out other promising industries seems with various types of sector-specific programs that provided less problematic than often thought. Although Norway incentives for private investment, for expanding productive and the Republica Bolivariana de Venezuela do show evi- capacity, and for promoting research and development in dence that their massive development of petroleum or nat- the sector. The main lesson we derive from this case study is ural gas has precluded other industries, Chile and Sweden that policymakers should be encouraged to experiment with do not, and they are arguably more similar to the majority various mechanisms for promoting innovation, especially in of LAC's economies. 7 FROM NAT URAL RESOURCES TO TEIE KNOWLEDGE ECONOMY TRADE AND JOB QUTALITY Lessons from LAC History: Don't Turn Your Back their advantage in developing forward linkages from nat- on Your Natuiral Advantages ural resources. If true, authors of the emerging literature Why, then, were LAC countries not as successful in using ought to be cautious about urging the promotion of indtis- their natural wealth as a basis for development? Our his- trial "clusters," and readers should be cautious about torical review indicates that several factors prevented such believing what they read. The fragmentation process an outcome during most of the 19th and early 20th cen- implies large efficiency gains, and forcing uneconomical turies: political insrability, barriers to trade, weak property productive chains may be little more than inefficient indus- rights, deficient infrastructure, volatile public finances, and, trial policy. especially, poor general and technical education and other The overall conclusion of both our empirical findings on barriers to innovation explicitly related to knowledge man- the determinants of trade specialization and the new theory agement and generation. is that in the current context of globalization, it is difficult After World War II there was an additional factor: LAC to predict where industries will be located in the future. countries turned their backs on their natural wealth. They The main reason for this is that international flows of inter- opted for promoting an import-substitution industriali- mediate goods, knowledge, capital, and labor, facilitated .oy zation strategy that did not play to their endowment low transport costs and improvements in communications strengths, discouraged innovation in manufacturing, and technology, open up opportunities for industries of various taxed the natural resource sectors in myriad ways. While degrees of technological sophistication to be established we do observe a continuum of inefficient interventions in almost anywhere around the globe. favor of the manufacturing sector to the detriment of the Thus, some Central American and Caribbean economies resource-exporting sector in other natural resource-rich have been highly successful in attracting FDI and in devel- countries, neither Australia, Canada, nor Scandinavia had oping labor-intensive exports to the United States through the extreme turn toward inward-looking policies, or the EPZs. These regimes reduce transaction costs, and provide a dramatic macrodisequilibria that gave Latin America a spe- stable and reliable regime for FDI (that is, they are efficient cial fame. At the extreme end of the continuum, we find and stable institutions) and corporate income tax incentives. Swedish labor unions across all sectors militantly insisting Costa Rica has been able to attract more high-tech FDI than on an open trade stance and energetic pursuit of produc- other countries in the subregion, and indeed has built an tivity-enhancing technologies. impressive and dynamic cluster in electronics, the well- The two phenomena were self-reinforcing: recurring bal- known Intel plant being the crown jewel. We find that this ance-of-payment problems led to a greater distrust of the achievement was possible mainly due to its higher level of global market and to political instability, and discouraged human capital, but also to the general quality of its instita- entrepreneurship and barriers to innovation. Arguably, the tions and active promotion policies undertaken by CINDE. measures to coax rapid manufacturing growth out of coun- Compared to the rest of LAC, since the mid- 1 990s MeK- tries with substantial structural barriers to innovation and ico experienced the most impressive transformation of its without a scientific tradition led to heavy rates of taxation on trade and productive structure. The critical factor is that traditional sectors. In the end, the protected manufacturing Mexico finally decided to play to its strength-proximiry industries lacked long-term dynamism, and the potential of to the largest and most technologically advanced economy the natural resource-based sectors was squandered. in the world. The case study on the electronics industiy, which has been one of the most dynamic in the Mexican Building New Areas of Comparative Advantage economy during recent years, indicates the importance of Many skeptics of natural resource-led development will incipient knowledge clusters and throws further light on accept these historical arguments, but will deny that pre- the role of public policy. Particularly after the implementa- viously successful experiences of natural resource-based tion of NAFTA, the computer and telecommunications development are replicable under today's conditions. In equipment industry in Jalisco has attracted the providers particular, the sharp reduction in transport costs that took of electronics manufacturing services and the suppliers of place in the last century eliminated the "natural protec- parts and components used in other locations throughout tion" that some natural resource-rich countries used to the world. These investments appear to be leading to the 8 INTRODUCTION AND SUMMARY creation of integrated supply chains, which could lead to tively slowly, due either to overall slower growth or to pro- important gains in efficiency, based on both lower costs and ductivity gains that reduce labor absorption. better logistics. Overall, this case highlights the possibility These dislocations, though extremely costly in human of evolving from pure miaquila-type assembly operations to terms, are transitional, and do not imply permanently higher skill-intensive manufacturing and R&D activities, success- rates of unemployment, as shown by the experiences of fully integrated into global networks. Chile and Mexico, the two economies in the region where A case study on Brazil's Empresa Brasileira de Aeronau- trade exposure has increased the most. Overall, unemploy- tica (EMBRAER), a major exporter of aircraft, sheds addi- ment rates were not especially higher in the 1990s com- tional light on the possibility of developing industries pared to the 1970s, except in countries with restrictive unrelated to traditional endowments, and on the impor- labor policies, such as Argentina and Colombia. tance of an outward orientation and knowledge infrastruc- There is no obvious channel through which greater open- ture. Early on, EMBRAER combined high R&D invest- ness would lead to a degradation of wages. We find that ments with strategic partnerships with foreign companies classification of manufactures by increasing sophistica- in order to achieve its present technological capacity, while tion-labor intensive, capital intensive, machinery, and always keeping an eye on foreign markets in which it now chemicals-is reflected in increasing wages, even after competes very successfully. It is also clear, however, that the adjusting for human capital. What is not obviously true is necessary government support to get EMBRAER to this that these sectors offer "better" jobs than petroleum, min- point was substantial, and it is not clear that the net pre- ing, forestry products, and even tropical agriculture in sent value of these public subsidies is positive, or that there terms of returns to human capital. Further, it appears that are major spillovers to other industries. Thus this case in countries as diverse as Argentina, Brazil, Costa Rica, the raises the issue of how far policy should go in supporting Dominican Republic, and Mexico, those industries most high-tech industries, even when they are commercially suc- exposed to international competition pay the highest wages. cessful after many years of government subsidies. Much of the literature, in fact, attributes the observed What these experiences, and the teleservices industries worsening of wage distribution in many countries to an in Jamaica, the smelting industry in Trinidad that increased demand for skills as a result of trade liberaliza- processes Brazilian ore, and myriad others tell us is that tion. In fact, the evidence suggests that very little of the today it is possible for developing countries to build com- observed movements appear to be attributable directly to parative advantages in technologically advanced activities, trade, and supply-side factors, such as long-term move- even if these appear unrelated to their original endowments ments in education and the dramatic entry of women into of natural resources. They need not, and indeed may not be the labor force, are easily as important. Nonetheless, it able to, replicate the paths followed by developed coun- raises the important point that, over the short term, the tries. But the recipe for success depends more than ever on creation of "better" jobs, in the sense of requiring high efficient institutions and investment in human capital, skills and better education-the jobs we want-may imply technical knowledge, and infrastructure. a tradeoff with wage distribution. Strikingly, the increased inequality in northern Mexico is attributed to the fact that Are the New Jobs Good Jobs? the jobs brought through the maquilas and NAFTA, while The region has seen some disturbing trends in job quality low skilled by U.S. standards, are above average by Mexi- over the last decade. But it is not clear how much these are can standards, and hence have increased the skill premium. related to trade liberalization and, if they are, whether they Public policy must take advantage of the opportunity that are permanent. higher skill premiums present (enhanced incentives for In terms of unemployment, the region has seen some families and workers to improve educational levels and very substantial reallocations of its workforce after trade skills) to improve the supply of quality education and opening. Large numbers of workers lost their jobs in previ- training. If they do, the end result may be a higher overall ously protected industries, and some experienced either level of real wages, with lower (or at least no more) wage very long periods of unemployment, large wage losses, or disparities than before. In fact, in the two countries with both. Further, formal sector employment has grown rela- the longest history of liberalization, Chile and Mexico, the FROMNI NATt RAL RESOITRCES TO I HE KNOWLE DGE ECONOMY TRADE AND JOB Qi LALI TY trend toward widening wage disparities reversed itself sev- ditional agriculture, and tourism. We find the criticism eral years after liberalization. to be overstated. Though wages and benefits are lower The growth of the "unprotected" or informal sector in than in similar jobs in industrialized countries (that is one the region has been especially noted: the Economic Com- of the reasons why some of these jobs are relocating to mission for Latin America and the Caribbean (ECLAC) developing countries), they are higher than for average ir suggests that 60 percent of new jobs in the 1990s were cre- comparable jobs in the country in question. Interviews ated in the informal sector. What drives this, and how we with workers suggest that this is precisely why they take should think about it, is not obvious, however. In some them. cases, such as Colombia after 1997, the traditional dualis- Who ends up employed or benefiting from these new tic view of formal sector rigidities forcing unwilling work- jobs depends on the characteristics of the individual ers into marginial informiial jobs has merit, although it is not worker. People, like nations, trade based on their compar;,- clear that there was any link to trade liberalization. But tive advantages. A striking fact that emerges in case after what is striking is that in Argentina, Brazil, and Mexico, case is that women have dominated the workforce itn informal self-employment expanded, along with an in- emerging industries. This is a pattern that was observed in crease in relative informal incomes in the beginning of the the industrialized countries a century ago as well, and to l990s, at the same time that the exchange rate appreciated. some degree it arose because most men were already work- This suggests that the dramatic rise in informality during ing, and hence women were available to fill the new this period was driven in part by increased opportunities in employment opportunities. But in most sectors a range the nontradeables sector that boomed in many countries of of gender-linked characteristics ranging from dexterity the region following the liberalization of capital accounts to reliability and lower incidence of alcoholism appear to and stabilization programs. motivate employers' choices. For the most part, femalh- A little reflection suggests that there probably should dominated jobs tend to represent opportunities, and may not be a relationship between trade and informality. have social repercussions beyond the purely economic. Aggressive exporters such as Belgium, the Republic of Significant improvements in job quality over the short Korea, and Singapore, or resource-intensive economies such term are circumscribed by basic economic constraints. as Australia, Canada, Finland, Sweden, and the United Most of these new jobs tend to be in industries where wage States, have low levels of informality. We find no secular increases not matched by productivity gains might erode trend either in total share of informal salaried workers, or competitive advantages and force firms to fail or relocate the number of informal microfirms linked to large or elsewhere. Over the longer term, better-paying jobs in bet- exporting firms in Mexico after liberalization or NAFTA. ter conditions can only be achieved by increasing the skills In fact, the share of firms that declare any linkage to large of the workers, and encouraging some degree of differenti- or exporting firms is so small as to suggest that arguments ation in products, and most fundamentally, by raising pro- that informality is primarily a way for exporting firms to ductivity. In the end, the process of raising job quality is avoid paying benefits through subcontracts seems without the process of development. any basis. The macroeconomic fluctuations driven by attempting to stabilize the peso had a far more obvious Conclusion effect on short-term informality levels than any long-term We must reiterate that rich endowments of natural resources, trend that might be attributed to NAFTA. We argue, in combined with the aggressive pursuit and adoption of new fact, that the problem is not excessive or exploitive inte- technologies, are a proven growth recipe. Further, the evi- gration of microenterprises in the global economy, but dence strongly indicates that their development does not rather barriers to more extensive integration posed by oner- preclude the development of manufacturing or other activi- ous registration procedures, deficient credit markets, and ties in the "knowledge" economy. In short, countries that poor availability of information. have "played to their strengths" have done well. We also address the concern that the jobs being created An equally important lesson is that what is important are somehow inferior by looking at four new types of jobs is not so much uhat is produced, but hown it is produced. in the maquilas or EPZs, the teleservice industry, nontra- Taking advantage of global technological progress is essen- i( INTRODIJCTION AND SIUMMARY tial in every field, and it cannot be done cheaply. The recur- road toward the "new" knowledge economy can and should rent lesson of the successful natural resource developers, be paved by public policies. and of contemporary theory, is the necessity of engendering a high level of human capital and developing a capacity for "national" learning and innovation. The present literature Notes 1. See a description of the use of protectionist instruments in on how nations learn and what types of policies are neces- place (tariff peaks, tariff escalation, subsidies, quotas, antidumping) and their limiting effect on LDC exports in World Bank 2001. viding a road map in this area. However, since the knowl- 2. Though they are also to some extent endogenous to the process edge produced by one firm can be applied by others, the of growth and development. CHAPTER 2 Comparative Advantage, Diversification, and Intra-Industry Trade: Determinants and Consequences OST COUNTRIES OF THE LATIN AMERICA AND CARIBBEAN REGION LIBERAL- ized their economies during the late 1980s and early 1990s. This process of liberaliza- tion included trade reforms, whereby import tariffs were slashed and nontariff barriers were reduced and simplified, and the welcoming of FDI. Much progress has also been made in other areas of reform, especially in macroeconomic stabilization and privatiza- tion (Burki and Perry 1997). When the reforms were first debated, it was broadly expected that the region would specialize in labor- intensive production of goods and services. However, the evidence presented in this chapter shows that the region's revealed comparative advantage has in fact changed little since the early 1980s. On average, the region has maintained a comparative advantage in the production of commodities that have a relatively high content of natural resources, and it has an advantage in the various types of land used for the harvest- ing of agricultural goods. It remains a net importer of capital- and labor-intensive manufactures. Despite the lack of structural change in the region's trade patterns, however, there have been significant productiv- ity gains, and consumers also now enjoy greater varieties of high-quality products. Given the concerns discussed in Chapter 1 about the con- New Trade Theories, New Endowments, sequences of different patterns of trade, the first step in the and New Patterns of Trade analysis should be to briefly review new trade theories that The theory of international trade has been transformed in seek to evaluate the factors that help determine what coun- recent decades, with the focus of study shifting away from tries produce. This is done in the next section. The subse- the stylized world of perfect markets, identical technolo- quent discussion identifies the revealed comparative advan- gies across countries, and immobility of factors and inputs tage of LAC economies during the last two decades of the of production across borders. Recent literature instead has 20th century, and empirically examines how traditional and tended to emphasize the following: modern notions of national endowments affect the structure of international trade and how trade structure in turn affects * The role of input trade (Jones 2000; Jones and Find- the rate of economic growth. Special emphasis is placed lay 2001) on the role of public policies. The conclusion is that trade * Frictions in international trade and investment flows liberalization and the development of modern factors of due to geography, institutions, transport, and infor- production can help accelerate the pace of development mation costs (Venables 2001; Bond 2001) in the region. What matters is how we produce, not what * The transmission of knowledge across borders (Gross- we export. man and Helpman 1991) 13 FRONM NATULRAL RESOULRCES TO T111: KNOWLEDGE ECONOMY TRA DE AND JOB QIUAI-ITY * Technological differences across borders (Trefler nally analyzed by Hirschman (1958), which is often cited 1995; Hakura 2001) in the literature on industrial clusters (see Bergman and * Monopolistic competition in differentiated prod- Feser 2001). Perhaps the clearest examples of this pattern ucts with increasing returns to scale (Krugman of trade have been observed in Finland and Sweden, two 1979; Ruffin 1999). economies that have rich forestland and that export forest products (see Chapter 3). Both countries have developed These new considerations do not mean that the tradi- sophisticated, export-oriented, high-tech industries frorn tional notions of comparative advantage and trade driven industries that originally provided inputs for the pro- by cross-country differences in relative factor endowments duction of forest products. Similar backward linkages are are no longer operational. Rather, traditional notions of present in various LAC economies, where mining or agri- endowments can be combined with the modern considera- cultural products buy inputs and hire services from other tions to provide a rich conceptual framework for under- sectors; in Chile, for example, the mining industry hircs standing how and why LAC developed the patterns of trade local engineering services (Meller 2001). These backward we observe today. linkages do not always result in the development of a sel- Perhaps the main lesson from the new theories is that arate export industry, however. anything is possible. A common pattern of integration in Hirschman was also the first to discuss forward linkages today's global economy is characterized by the fragmenta- from natural resources, such as those observed in the devel- tion of the manufacturing of labor-intensive goods (Jones opment of the steel industry in the United States during the 2000). This structure of trade is observed, for example, 19th century (Jones and Findlay 2001). Currently, forward in the Dominican Republic, El Salvador, and, to a lesser linkages can be seen in the metallurgical industry in Brazil extent, in Costa Rica and Mexico, The term "fragmenta- and other developing countries, but low transport costs in tion" tries to capture the idea that the capital-intensive the 20th century have reduced the profitability of these stage of production is conducted mainly in the industrial- types of linkages by fragmenting the stages of production ized, capital-rich economies, while the labor-intensive of steel and of other metallurgical industries, with ores now stage of the production process is conducted in the devel- exported for manufacturing to countries that have the know- oping countries, where labor costs are lower. how, the capital, and other factors needed to most efficiently The main difference between this and the traditional pat- produce the final products. An additional constraint on LAC tern of inter-industry trade is that the transportation of countries hoping to develop forward linkages from their intermediate goods from capital-rich economies to develop- agricultural or mining activities is the trade policies of ing countries has now become cheap enough for it to be major export markets that favor the export of basic com- profitable to export the intermediate good for processing in modities by discriminating against imports of processed raw the developing country. The costs of shipping the finished materials-policies dcsigned to protect domestic industries goods back to the rich consumer markets also must be rela- that process agricultural or natural resources. tively low for this pattern of trade to be profitable. In sum, The rise of intra-industry trade (IJT)-that is, exports the costs of sending the capital input to the developing and imports of similar goods-is perhaps the most notewor-- economy plus the costs of sending finished manufactures thy feature of the post-World War II global economy. This back must be lower than the profit gains realized through type of trade is clearly observed in the case of Mexico, which-i locating the labor-intensive stage of the manufacturing exports electronics and automobiles to the United States, process in the developing economy. Consequently, it is not but which also imports similar products. A key ingredient surprising that this pattern of trade has emerged in several in Mexico's case is FDI and the importation of capital and of the LAC economies located close to the U.S. market. The other inputs for the production of these goods. The litera- manufacturing firms in these economies have become part of ture on the determinants of IIT (see Krugman 1979) empha- international industrial clusters, a concept analyzed further sizes that the emergence of such trade depends on product below. differentiation, in which firms across borders produce simi- A third pattern of trade features linkages between nat- lar goods that are distinguished from one another by brand ural resource sectors and domestically manufactured inter- or by other subtle differences. The main driving force of IIT mediate goods This is a rype of backward linkage, origi- in this context, however, is increasing returns to scale-that 1 4 CONMPARATIVE ADVANTAGE. DIVERSIMICATI()N, ANT) 1NTRA-INDUlSTRY TRAI)E DFTFRMINANTS ANI) CONSEQIUENCES is, unit costs of production that decline as the quantity pro- erally, are replete with examples of firms that are linked to duced increases. In other words, productivity gains are at the each other either as competitors or as parts in a single sup- center of the competitive process that yields IIT. This pat- ply chain (Porter 1990). Sometimes these firms are geo- tern of trade can thus emerge when industries benefit from graphically nearby. Bergman and Feser (2001) state that an either large initial investments or fixed costs (Ruffin 1999), "industry cluster may be defined very generally as a group or when producers become more efficient through experi- of business enterprises and nonbusiness organizations for ence or through the introduction of technological innova- whom membership within the group is an important ele- tions from abroad or from nearby industries (see Venables ment of each member firm's individual competitiveness." It 2001, for a literature review.) The next section explores the cannot be overstated that an industrial cluster is held to- empirical determinants of IIT. gether by the profit objectives of the individual firms- Agglomeration, or clustering, effects have also received that is, each firm benefits from its relationship or proxim- substantial attention (Puga and Venables 1999). The dis- icy to the other firms in the cluster. tinguishing feature of such effects is that the production of Business administration literature defines a plethora of one type of final good leads to the development of new man- different types of clusters, of which vertical and horizontal ufacturing processes as a consequence of learning spillovers; clusters are perhaps the two main categories. Vertical in- for example, firms in certain knowledge-intensive sectors dustrial clusters comprise groups of firms that are part of a can learn from the experience of other firms and thus single supply chain. The textile-apparel cluster, for exam- develop new products. The emergence of this type of hori- ple, includes firms that supply textiles to other firms, zontal cluster has been observed in geographic areas includ- which manufacture clothing and apparel to be sold to con- ing Silicon Valley in California, and around certain firms in sumers. The profitability of firms belonging to such clus- Costa Rica and Mexico. The rapid decline in the costs ters depends to a large extent on their capacity to supply of information dissemination since the late 20th century the inputs and the final goods on time. Given that trans- suggests that the geographic dimension of industrial clus- port costs can help determine the profitability of a firm, ters might become less important, but the process nonethe- geography can thus play an important role in the estab- less demonstrates that a vibrant manufacturing sector can lishment of these clusters. Coordination and transaction emerge in formerly agricultural or mining economies as a costs are also important, together with transport costs com- consequence of foreign investment, input trade, and learn- prising what some analysts call logistics costs (Guasch and ing spillovers across firms. As discussed by Puga and Ven- Kogan 2001; Bond 2001). It is in large part because trans- ables, trade liberalization, by reducing the costs of imported port and coordination costs are likely to be negatively cor- inputs, can lead to agglomeration effects that result in sig- related with distance that vertical industrial clusters often nificant welfare and productivity gains for developing coun- appear as regional clusters or as a regional agglomeration of tries. In contrast, protectionism might lead to much less industries. dynamic industrialization Horizontal industrial clusters comprise firms that pro- The following sections investigate in more detail the duce similar products, and that learn techniques and man- formation of industrial clusters and the role of geography agement practices from each other. In economic terms, if and knowledge in fomenting clusters, and include a dis- one firm's activities benefit others, the positive effects of cussion of the impact of logistics costs in determining trade those activities that go beyond the firms' confines are called structure. They also address some issues of measurement "externalities." Venables (2001) writes that the "externali- and the correlation between ICT and knowledge and the ties generated in one sector will typically only affect firms level of development. in a set of industries, perhaps the same industry or others that are in some way linked to it." The main sort of tech- New Factor Endowments: Industrial Clusters, nological externalities are knowledge spillovers, such as the Logistics Costs, and ICT and Knowledge spillover of research and development activities, technical or managerial know-how, or the knowledge accumulated Clusters, Geography, and Knowledge with experience (that is, learning by doing). There is a popular specialized literature that emphasizes Horizontal industrial clusters sustained by positive that national economies, and the world economy more gen- knowledge spillovers are often also geographically agglom- 15 FROM NATURAL RESOURCES TO TILE KNOWLEDGE ECONOMY: TR ADE AND JOB QUALITY erated. This is due less to transport costs, which play an national Monetary Fund (IMF) statistics indicate that for insignificant role in the transmission of knowledge, than to the countries in the bottom quartile of income distribution the interaction and information exchange between firm (based on per capita gross domestic product [GDP] in managers. Perhaps more important, agglomeration effects 1990), the average transport cost for importable goods was can be the result of the existence of trained workers who 20 percent of the value of imports. For countries in the top "job-hop" across firms, taking with them the knowledge quartile, the average transport cost rate on imports was and training accumulated while working for different firms only 6.3 percent. (Cooper 2001). This difference in transport cost rates presents a signifi- An interesting issue to consider in reference to the cant barrier for firms in developing countries to entering agglomeration of firms in a horizontal cluster is whether foreign markets, both in terms of the cost of getting goods the recent ICT advances might dilute the advantage of geo- to market and in terms of acquiring imported inputs. It graphic proximity for the sharing of knowledge across also provides protection to relatively inefficient import- firms. The following subsections focus on the role of logis- competing firms in developing countries. In this sense, the tics costs, knowledge, and ICT in determining both eco- effects of logistics costs are analogous to those of the "effec- nomic structure and aggregate performance, with a special tive rate of protection" from the tariff literature: the value emphasis on measurement issues. added in a particular production activity is changed f;lr given prices of the primary factors of production. The way Logistics Costs this change occurs depends on the characteristics of the Logistics costs can be thought of as the costs of getting the activity, related to the importance of transportation costs right goods to the right location at the right time.1 These for the final product and for the intermediate goods. costs include freight, storage, warehousing, insurance, and The main determinants of logistics costs are the locaticn administrative costs. They relate both to the selling of final of a country and the quality of the infrastructure. Table 2.1 goods and to the purchase of intermediate goods. presents several of these measures of logistics costs for dif- There are several factors that affect logistics costs in a ferent regions of the world. These measures are, from country. Geography, the quality of infrastructure, and the columns (1) to (8): average transport costs (CIF/FOB-] ); management techniques of firms are the most obvious ones. average air distance to the closest major market (Asia, Countries closer to major markets can bring their goods to Europe, and North America); port efficiency index (based these markets at lower costs simply because smaller dis- on a shipper survey, from 1 [lowest quality] to 7 [highest tances have to be traveled to deliver the goods. Also, infra- quality]); average number of days it takes for goods to clear structure increases the efficiency of the transport system customs (based on the median response from a survey of and favors competition on transport networks. Thus, ship- importers); kilometers of paved roads per square kilometer; ping costs may be high because of the inefficient transport kilometers of railroad lines per square kilometer; telephone system, such as in Mercosur, wherc three quarters of the main lines per capita; and air freight per capita (logarithml Is overland trade between Brazil and its partners travels over for these last four variables). a single bridge. Costs can be reduced by the adoption of As can be seen from Table 2.1, the various measures of more effective management of the supply chain, and higher infrastructure are quite highly correlated. Indeed, high- coordination among firms, as in the just-in-time inventory income countries have values of all infrastructure measures management techniques that led to substantial productiv- that are from 1 to 1.35 standard deviations above the world ity improvements in the United States. average, while Sub-Saharan Africa lags in all measures with These logistics costs are particularly important in ana- quantities from 0.6 to 1 standard deviation below the lyzing international trade, because a firm dealing in foreign world average. markets faces higher costs than local producers. This dif- The evidence also suggests that logistics costs faced by ference arises not only because of the transport costs of LAC countries are significantly higher than those of devel- shipping goods between countries, but also because of the oped countries (see Table 2.2 for values for specific coun- communication and coordination costs of organizing the tries). Nevertheless, there seems to be nothing special supply chain at a distance and in a foreign country. Inter- about LAC: once one controls for the income level, popula- 16 COMPARATIVE ADVANTAGE. DIVERSIFICATION, AND INTRA-INDUSTRY TRADE DETERMINANTS AND CONSEQUENCES TABLE 2.1 Transport Costs and Infrastructure by Country Groups AVERAGE AVERAGE DISTANCE MEDIAN KM OF KM OF TELEPHONE TRANSPORT TO CLOSEST PORT CLEARING PAVED RAILROAD MAIN COSTS MAJOR EFFICIENCY TIME ROADS PER LINES PER LINES PER AIRFREIGHT GDP PER (CIF/FOB-I) MARKET INDEX AT PORT KM2 KM2 CAPITA PER CAPITAL CAPITA COUNTRY GROUP (1) (2) (3) (4) (5) (6) (7) (8) (9) Developed Countries 4.24 1,561 5.76 3.50 1.15 1.02 1.35 1.16 2.11 South America 9.83 6,528 2.84 8.81 -0.65 -0.65 0.16 -0.12 -0.12 Central America 10.96 3,086 3.59 6.56 0.09 -0.04 0.24 -0.33 -0.37 Middle East and North Africa 10.12 3,512 3.77 5.50 -0.09 -(0.32 0.20 0.11 -0.08 Sub-Saharan Africa 19.54 6,237 4.55 12.00 -0.73 -0.60 -1.07 -0.61) -0.70 South Asia 9.74 5.121 2.79 n.a. 0.20 -0.01 -1.(0 -0.53 -0.67 Transition Economies n.a. 2,248 3.3( 4.13 1.14 0.97 0.46 -1.90 -0.9( East and Southeast Asia 8.99 5,651 4.54 5.67 0.70 0.04 0.28 0.55 0.55 Middle-Income Europe 8.57 1,525 4.48 n.a. 1.10 11.76 1.11 0.69 0.69 n.a. = Nor available. Souroes: CIF/FOB is from IMF ( 1995) except far Sourh America, wIhich is obtained from Hummels (1999); distance from capital city to closest of Rotterdam, New York, or Tokyo (km), obtained from Gallup Sachs. and Mellinger (1998); port efficiency and clearing time data from Mi,co and Perez (2(101); iiifrastrrucrtr data are froiii Caiiiiing (1999) foe 1990 and are expressed as standard normal devlations from the world mean. World Bank measure (1990) from Gallup. Sachs, and Mrllinger (1998). The gfoup definitions given in each column heading are provided in the Appendix of Bond (2001). tion, and size of the country (by regressing the respective costs. The potential impact of these improvements on the infrastructure measure on these variables for a cross-section value added in the different sectors within a country will of countries in 1990), the comparison of the predicted val- depend on the trade pattern of the final goods and the impor- ues with the actual values shows no systematic bias associ- tance of traded intermediate products in production. To ated with LAC. Worse infrastructure and higher logistics illustrate the importance of these factors, Table 2.3 presents costs in Latin America are mainly associated with the char- a simulation with the effect of a 1 percent transport cost acteristics of the countries in the region. reduction on the value added by industry, for Argentina, Still, these higher logistics costs are reflected in higher Brazil, and Mexico. The simulation is based on a model of freight rate markups, longer waiting times for goods to clear production with logistics costs, assuming constant prices ports, and higher inventory holdings of raw materials by of inputs and nontraded goods, and constant demands for firms (see Guasch and Kogan 2001). Improvements in the inputs, intermediary goods, and the final good. Thus, the quality of infrastructure have the potential to reduce these simulation can be seen as an approximation of the immedi- TABLE 2.2 Transport Costs and Infrastructure: Selected Latin American Countries AVERAGE AVERAGE DISTANCE MEDIAN KM OF KM OF TRANSPORT TO CLOSEST PORT CLEARING PAVED RAILROAD TELEPHONE COSTS MAJOR EFFICIENCY TIME ROADS PER LINES PER MAINLINES AIRFREIGHT GDP PER (CIFiFOB-1) MARKET INDEX AT PORT KM2 KM2 PER CAPITA PER CAPITA CAPITA COUNTRY (1) (2) (3) (4) (5) (6) (7) (8) (9) Argentina 7.50 8,570 3.81 7.00 -o1.36 0.26 0.50 0.14 0.21 Brazil 7.30 7.700 2 92 10.00 -0.50 -0.59 0.24 0.19 -0.09 Chile 8.80 8.290 3 76 3.00 -0.56 0.05 0.31 0.96 0.17 Mexico 4.50 3,360 3.34 4.00 -0.03 11.33 0.31 -0.54 -0.06 Paraguay 13.30 7,580 * n.a. -1.38 -0.17 -0.81 -0.46 Uruguay 4.60 8,560 5.00 n.a. n.a. 0.48 0.68 -0.93 -0.04 ' Landlocked country. n.a. = Not available. Sourcr.,: ClF/FOB from Hummels (1999) except for Mexico, which is obrained from the IMF; other variables as in Table 2.1. 17 FRa)M N ATURAL RES00 RCES 1O TilE KNOWLEDUE ECONOMY TRADE AND JOB QUALITY TABLE 2.3 ICT and Knowledge: Measurement and Correlation Effect of a 1 Percent Transport Cost Reduction on Value Added ivith Developnzent by Industry: Argentina, Brazil, and Mexico Indicators for knowledge and ICT measure the relative INDUJSTRY ARGENTINA BRAZIL MEXICO position of each country with respect to others in these Ley Agriculture. Forescry, areas of development.2 Eight indicators were selected. Four and Fishing 1.11 -1I33 -1.1( reflect innovation activity and R&D in each economy, and Food, Beverages, Tobacco 1.29 5.64 -1.4 'Textiles -l.22 -1.39 2. i3 four reveal the level of ICT development. The indicatrrs Apparel -1.08 -1.47 1.69 are: Leather Products 1.9Y 3.0(0 2.70 Footwear -1.72 2.56 1.56 WiVod ProdLICts -1.64 2. 3 -1.39 1. Research and development as share of gross national FLirniiture -1.)5 -5.61 2.53 income IGNI) Paper -1.47 2. 35 -1. I Prinrting and Publishing -1.116 -2.28 -1.38 2. Scientists in R&D per million people Chemicals -1.44 -1.85 -1.47 3. Patent applications by residents and nonresidents per Petroleum 1. 2 -1.55 -5.32 RLubber Products 2.05 4.22 -1. *3 1,000 people Nonmetallic Minerals -1.47 2.49 -1.54 4. Patent applications in the United States by country (lass 1.2i -1.47 1.80 Jr(in and Steel - 5.( 4.74 -2.02 per 1,000 people NonferrouIs Metals -1.-( 4.25 1.68 5. Telephone mainlines per 1,000 people Metal Products - n() 1.95 -1.59 Niinelectrical Machinery 1.(lI -2.45 -1.67 p , p Electrical Machinery -1. ( -2.34 1.92 7. Personal computers per 10,000 people Transpirrtation Equipment -1.10 -3.25 2.0)9 8. Internet hosts with active Internet Protocol (I P) Mfiscellaneous Manufactures -1.05 -1.61 -1.67 Ntining 1.i16 -1.48 1.22 addresses per 1,000 people. Swine: Bond 2ii0i i. The definitions and relevance are self-explanatory for ate impact of the change in the logistics costs given the cur- most of these variables. Mobile phones and telephones mea- rent allocations in the economy. sure the depth of connectivity in a country. The other two The model predicts that, for Argentina, the major bene- ICT variables are personal computers and Internet hosts.' ] n ficiaries of the infrastructure improvements would be the the knowledge area we included R&D as share of gross leather, agriculture, food, and petroleum sectors; for Brazil, national income and scientists working in R&D. We also they would be metal, food products, and rubber; and for included patent applications filed by nationals and nonna- Mexico, they would be the leather, textile, transport equip- tionals as an indicator of both innovation activity and as a ment, electrical machinery, and furniture sectors. It is note- measure of the need and ability of the state to protect intel- worthy that in these cases the main beneficiaries are agricul- lectual property. Patent applications in the United States by ture, natural-resource intensive, and labor-intensive sectors. country of origin of the inventor was also included.4 This Overall, there seems to be substantial need for improve- variable helps control for the variability in the previous ment in LAC infrastructure. These improvements will variable caused by differences in the institutional develop- require both investments in new infrastructure and im- ment of each country. However, as discussed in Grossman provements in the regulatory environment. The potential and Helpman (1991), patent applications are also deter- gains from these improvements could be substantial. As an mined by the size of the domestic market where the example, a recent World Bank report points to the fact that patented products are sold. Consequently, neither indicatoir the costs of handling a container in Brazil could be cut is a perfect proxy for innovation output. All the variable, almost in half (from US$4,775 to US$2,614) if multimodal except patent applications in the United States, come from transport reforms were implemented. Cost reductions of this the World Developnent Indicators. Patent applications in the magnitude, amounting to 4.5 percent of the values of ship- United States come from the U.S. Patent Office. ments for imports, would reduce transport costs on Brazil- based on averages for 1995 to 2000 or 1990 to 1999, ian goods as reported in Table 2.3 to a level comparable to Table 2.4 shows the country variables expressed as a per- the rates reported for U.S. trade. centage of the U.S. levels. The most striking feature is that 18 COMPARATIVE ADVANTAGE, DIVERSIFICATION, ANI) INTRA-INDUSTRY TRADE DETERMINANTS AND CONSEQLJENCES TABLE 2 4 Indicators of ICT and Knowledge as a Percentage of the United States Levels INFORMATION AND COMMIUNICATIONS TECHNOLOGY KNOWVLEDGE PATENT R&D RESIDENTS PATENT TELEPHONE MOBILE INTERNET PERSONAL SCIENTISTS & NON- APPLICATION MAINLINES PHONES HOST COMPUTERS R&D AS (PER RESIDENTS IN U.S (PER 1,00110 (PER 1,000 (PER 1,0()0 (PER 10,00C SHARE OF MILLION (PER 1,0()(J (PER 1,000 PEOPLE) PEOPLE) PEOPLE) PEOPLE) GNI PEOPLE) PEOPLE) PEOPLE) LAC Countries Argentina 30.04 21.56 1.41 9.12 14.25 17.79 19.34 0.52 Bolivia 8.94 6.12 0.09 1.94 18.75 4.79 1.78 0.(4 Brazil 20.38 13.00 1.04 6.12 30.88 4.53 24.36 ().19 Chile 1.0(9 17.05 2.109 10.58 28.88 11.69 14.24 11.24 Colombia 24.23 14.08 0.36 6.76 4.50 n.a. 3.98 0.07 Costa Rica 29.96 8.65 1.40 9.45 7.13 n.a. n.a. 0.47 Ecuador 13.00 6.36 0.13 -1.06 0.53 3.42 2.99 0.06 Mexico 16.30 9.42 0.91 9.i)4 10.88 5.09 40.70 0.27 Peru 9.78 7.84 0.27 2.90 9.38 9.85 3.29 0.04 Venezuela, R.B. de 17.06 23.47 0.32 8.90) 19.00 5.70 11.22 0.35 Average 16.59 18.20 1.59 . 23.41 n.a. 3(1.7 0.23' Asian Countries China 11.73 5.02 0.02 1.34 24.85 10.33 5.50 0.()7 India 3.66 0.34 (.(11 0.50 28.*5 3.89 1.08 0.04 Korea, Rep. of 65.74 73.7( 4.04 35.11 87.76 56.18 285.36 211.88 Thailand 12.77 16.08 0.38 4.61 5.63 3.01 9.61 (.()9 OECD Countries Germany 83.29 21.08 70.72 61.74 87.76 76.32 232.99 34.86 Japan 79.03 18.40 151.66 51.83 1()8.()2 138.S 2 371.80 72.87 United States 100.00 1110.00 100.00 10(.()() 1()().00 100.00 10().0( 10().()() AVERAGES BY LEVEL OF INCOME Low income 3.05 ().67 0.03 0.86 n.a. n.a. 30.80 n.a. Middle income 14.78 12.51 0.85 4.89 33.62 18.02 35.83 n.a. High income 84.04 111.29 65.60 69.52 89.35 86.13 334.78 ( 9.24a v.a. = Not available. a Average of the corresponding countrnes included in this table. for all developing countries except Korea, development in On the other hand, the other three Asian countries knowledge and ICT is very shallow. Korea spends almost as (China, India, and Thailand) show very low levels of knowl- much as the United States on R&D, but the number of sci- edge and ICT development. In fact, their figures are closer entists in R&D in Korea is about half that in the United to those of the selected LAC economies. In China and India, States. The indicators of connectivity (phones and mobile knowledge and ICT are almost exclusively led by R&D, phones) are about 65 percent of those in the United States. which in both cases is about 25 percent of that in the The penetration of personal computers is about 35 percent United States. China has about 10 percent of scientists in of that in the United States, and the level of Internet hosts R&D, which for a big country amounts to a large contin- (4 percent) and the number of patent applications by Kore- gent of engineers. ans in the United States (20 percent) are much lower. On Even though the LAC countries have a much lower level the other hand, the number of patent applications in Korea of knowledge and ICT development than Korea and the when controlled by population is three times that of the United States, there are special characteristics worth men- United States, suggesting a relevant system of property tioning. Brazil and Chile have relatively more R&D than rights and innovational activity in the country. China and India, which in their own structure of "knowl- 19 FROM NATuJRAL REsOURCES TO THE KNOWLEDGE ECONOMNY TRADE AND JOb QUALITY edge" depended strongly on R&D. LAC countries also have and access to rhe Internet are, according to rhis basic result, higher connectivity levels (phones and mobiles), especially highly correlated with income per capita. The relationship Argentina, Chile, Costa Rica, and Repdblica Bolivariana de for the knowledge index and development is also high. Venezuela. Nonetheless, these variables are 30 percent of GDP per capita explains about 60 percent of the variance the level achieved in the United States. The penetration of in the knowledge index (R2=0.6). personal computers is similar for most LAC countries, but Thus far, the evidence is clear concerning the impact of Argentina, Chile, Costa Rica, and Repdblica Bolivariana de logistics on the structure of trade, and ICT and knowledge Venezuela are the leaders (around 10 percent of the United are key correlates of the level of development. The follow- States). Argentina has a relative strength in the number of ing section explores more broadly the determinants of scientists per 1,000 people, even though it is only 19 per- trade structure around the world, with a special focus on cent of the level in the United States. Patent applications the role of "new" endowments, including ICT and knowl- in the United States are minimal for all the countries, but edge. However, a brief discussion of measurement issues is patent applications in Mexico are 40 percent of those in the necessary prior to analyzing the patterns of trade in LAC United States, and this is the only area of "knowledge" countries. where Mexico has a relative strength. On average. LAC countries are above middle-income countries in ICT, but Comparative Advantage: Facts from LAC are lagging behind on the components of knowledge. The concept of "comparative advantage" is one of the best In the following section we use knowledge and ICT understood by academic economists, and the least accepted composite indexes to analyze the determinants of trade pat- by society. When economists mention 'comparative advan- terns. These indexes are generated using the principal com- tage" they are referring to economic activities that an econ- ponents of the variables described herein. Figures 2.1 and omy can produce at lower relative costs than others. [n 2.2 show the relationship of each one of the composite other words, it means that countries have certain strength-is, indexes and GDP per capita. It is evident that there is a and the resulting patterns of trade and production should positive and nonlinear relationship between knowledge and reflect those strengths, rather than policy distortions. ICT and the level of development across countries during This report relies on indicators proposed by Edward 1976-99. The fit of the regression is lhigh for the ICT index Leamer of the University of California at Los Angeles, who (R2=0.8). Thus, communications, computer penetration, is one of the leading world experts on international tra(le FIGURF 1 I Development and Knowledge Index Development and Knowledge [9 - y = 0.42957x -- 5.5257x + 16.436 10 - R' - 0.6246 X 6 - -o- VI I II. -2 6- - - -- -- 4.5 5.5 6.5 7.5 8.5 9.5 10.5 Logarithm of Gross Do0mestic Product per Capita S~~~~~~~~~~~~~~~~2 COMPARATIVE ADVANTAGE, DIVERSIFICATION. AND INTRA-INDUSTRY TRADE DETERMINANTS AND CONSEQUENCES FIGURE 2.2 Development and ICT Index Development and ICT y 0(.2882x - 6.2-365x2 + 44.969x - 108.89 8- 6 - () - _, .__ -o 4 0- O- 2 - -----*- .., Oi.S .,-fiL = - -- -2 _- A -- l l l l.. - I _ 5 6 7 8 9 10 Logarithm of Gross Domestic Product per Capita analysis. In several important professional publications, The other three clusters of manufactured products (labor Professor Leamer proposed using a set of 10 industrial clus- intensive, capital intensive, and chemicals) maintained a ters, which are aggregates of over 2,200 products traded relatively stable share of the world market during the internationally (see Leamer 1984, 1987, and 1995). The 1990s. In contrast, the agricultural and mining commodi- distinguishing feature of each commodity group is that ties experienced a noticeable decline in their corresponding empirical analyses conducted by Leamer (1984, 1995) show shares. However, these data should not be interpreted as that the products included in each cluster tend to be indications that agricultural and mining commodities are exported by similar countries in terms of their endowments zero-growth industries. In fact, world exports of cereals of different types of labor, land, and natural resources. For grew, on average, by 2.6 percent per year during the 1990s, reference, Table 2.5 shows the components of Leamer's 10 exports of animal products by 3.2 percent, raw materials commodity aggregates, based on the widely used Standard by 3.6 percent, petroleum by 4.1 percent, agriculture by International Trade Classification (SITC). 4.4 percent, capital-intensive manufactures by 4.6 percent, As a precursor for further analysis of the evolution of pat- labor-intensive products by 6.7 percent, and chemicals by terns of trade around the world, Figure 2.3 shows the shares 7 percent. The rising share of world exports of machines of world merchandise exports of the four clusters of manu- reflects the fact that machine exports grew even faster than factures, plus the sum of the shares of petroleum and raw the others-its overall average growth rate was about 8.2 materials, and the sum of the four agricultural sectors dur- percent per year. ing 1976-99. The cluster of machines has had the lion's The theoretically rigorous measure of comparative ad- share of the world market, and also experienced a surge dur- vantage is net exports, the difference between the value of ing the 1990s. In the companion Table 2.6, it is evident exports and the value of imports of each of these product that the growth of machines was not even for all products aggregates. The mathematical derivation of this conclusion during the last decade, although all products in this cluster comes from Leamer (1984) and is presented in Box 2.1. The had positive growth rates. The fastest-growing products intuition, however, is quite clear: countries will export were electrical equipment, office and automatic data proces- products that are produced with the strong aspects of their sors, and telecommunications equipment. There are no sur- economies, which could be large endowments of natural prises in these numbers. resources or highly skilled labor, or any other type of 21 FROM NAITLTRAL RESOURCES TO THE KNOWLEDGE ECONOMY TRADE AND JOB QUALITY TABLE 2.5 Components of Leameres 10 Commodity Clusters AGGREGATES SITC AGGREGATES SITE Petroleum Labor Intensive Petroleum and derivatives 3 Nonmetal minerals 66 Furniture 82 Raw Materials Travel goods and handbags 83 Crude fertilizers and minerals 27 Arr apparel 84 Metalliferrous ores 28 Footwear 85 Coal and cole 32 Miscellaneous manufactured articles 89 Gas, natural, and manufactured 34 Postal packaging, nor classified 91 Electrical current 35 Special transactions, not classified 93 Nonferrous metal 68 Coins (nongold) 96 Forest Products Capital Intensive Lumber, wood, and cork 24 Leather 61 Pulp and waste paper 25 Rubber 62 Cork and wood manufacturers 63 Textile yarn and fabric 65 Paper 64 Iron and steel 67 Manufactured metal n. e. s. 6' Tropical Agriculture Sanitary fixtures and fittings 81 Vegetables 5 Sugar 6 Machinery Coffee 7 Power generating 71 Beverages 11 Specialized 72 Crude rubber 23 Metalworking 73 General industrial 74 Animal Products Office and data processing 75 Live animals 0 Telecommunications and sound 76 Meat 1 Electrical 77 Dairy products 2 Road vehicles 78 Fish A Other transportation vehicles 79 Hides and skins 21 Professional and scientific instruments 87 Crude animals and vegetables 29 Photographic apparatus 8S Processed animal and vegetable oils 43 Firearms and ammunition 95 Animal products n. e. s. 94 Chemical Cereals, etc. Organic 51 Cereals 4 Inorganic 52 Feeds 8 Dyeing and tanning 53 Miscellaneous 9 Medical and pharmaceutical products 54 Tobacco 12 Essences and perfumes 55 Oil seeds 22 Fertilizers 56 Textile fibers 26 Explosives and pyrotechnics 57 Animal oil and fat 41 Artificial resins and plastics 58 Fixed vegetahle oils 42 Chemical materials n. e. s. 59 n. a s. = Not elsewhere specified. SITC = Standard Ineernatonal Trade Classification. Sasece: Leamer 1995. endowment. The precise definitions of the types of endow- those discussed in Balassa (1989), is that theoretically ments that could be considered as sources of strengths for countries should be net exporters of products that inten- any given economy at a particular point in time were dis- sively use the resources and economic strengths they have cussed in the previous section of this chapter. at their disposal. This is so because an exclusive focus on Another advantage of using net exports as an indicator exports will ignore the possibility that countries import a of revealed comparative advantage, rather than other mea- substantial amount of goods that they also export, which sures that focus only on the structure of exports, such as could indicate that the hypothetical country is a net 2) (OMPARATIVE ADVANTAGE, DIVERSIFICATION, AND INTRA-INDUSTRY TRADE. DETERMINANTS AND CONSEQUENCES FIGURE 2.3 Shares of World Merchandise Exports, 1976-99 0.50 - 0.45 - 0.40 - 0.35 - # * < -o- Petroleum and o 30- Natural Resources 0.30 - -E- Foresc, Agriculture, 0.25 - , and Animals A Labor Intensive 0.20 - capital Intensive 0 15 .'_ '._ = ._ _ -.... .u _ _ ' = _ Machines 0.15 - - -. C Chemical 0.10 -L 0.05 0 0.00- I l l l I I I I I l lo lI TaBLE 2.6 move to trade liberalization was meant to bring about a Growth of World Machine Exports, 1990-99 change in the structure of trade in favor of labor-intensive ( Percent) ANNUAL sectors. However, the region's average pattern of international PRODUCT GROWTH RATE trade, as reflected in the net exports per worker across Arms 0.3 Leamer's (1984 and 1995) clusters of commodities, has Metalworking 3.7 remained dependent on and even increased the net exports of Specialized 3.9 Photography 4.4 land- and natural resource-intensive commodities, as shown General Industrial 6.8 in Figure 2.4. By 1997, the regional average of net exports of Road Vehicles 7.0 Other Vehicles 7.6 raw materials per worker was about US$250; the number for Power Generating 8.( petroleum and tropical agriculture was about US$210; while Professional and Scientific Instruments 8.8 Troelecomuniations and S oientificInstrumend snet exports of animal products per worker was almost US$ 50. Telecommunications and Sound 9.7 Office and Data Processing 10.7 In contrast, in 1997, net imports per worker of machines were Electrical 11.9 almost US$ 800. The hopes of structural change were perhaps TOTAL 8.2 misguided, because neoclassical trade theory predicts that countries with abundant land and natural resources will spe- importer of the factors of production used in the elabora- cialize even more in these sectors after liberalization. Never- tion of those products. In other words, the indicator of net theless, the increased dependence on suchl products hlas againi exports provides adequate information concerning the raised eyebrows and instigated recent calls for reevaluating products in which an economy has strengths relative to the the structural effects of development strategies based on lib- rest of the world. Hence, the rest of the analysis in this eral trade policies (see, for example, Katz 2000; and Ramos chapter focuses on the pattern of net exports for the 1998). regional average, and for a set of LAC countries. Specific Country Experiences since the Early 1980s The Regional Average since the Early 1980s The regional averages discussed above, however, are mis- As mentioned earlier, in Latin American countries that exper- leading in that they do not show the rather large differcnces imented with protectionist regimes roughly since the inter- that exist within the region. Figures 2.3 through 2.10 show war period (the 1920s and 1930s) until the early 1980s, the the evolution of the composition of net exports for our core 23 RI)fM NATUR WI RFSO[JRCES TO THE KNDWLED(,E ECONOMY TRADE AND DOB QL ALI1 i BOX 2 1 Met Exports as an Imlicator of Comprative Advantae Driven by Endowments Assuming Leontieff technology, the framework in Leamer C = (5) (1984) be' ins with the system of equations that relate fac- tor supplies to factor demands as follows where iw S the wotld's output vector and s IS the pro- portion consumed by each country K = aK Y + aK2Y2, and (1) The proportional-consumption assumption is rather L = I YI - aL2Y, (2) implausible and is used for the sake of simplicity, but it is not a frundamental part of the argument Deviations from K and L are the amounts of two factors of production, t assumptod not of the analy- call them capital and labor, available in a given country These amounts are country specific and are assumed to be sis (see Leamer 1984) Hence the vector of net exports internationally immobile The Y's denote the quantity is simply the product of the inverse of the vector of fac- produced in the given country of two commodities (labeled tor intensities across product clusters and the difference and ,) The a's are the traditional factor intensities deter- between each country's vector of endowments and the mined by the available production technologies in each world's vector of endowments An often forgotten step sector, and y the represet te units of each factor required in the derivation of testable hypotheses is that the key to produce a unit of output Equations (I) and (2) repre- dependent variable is net exports, not gross exports or sent a system that can be solved for outputs Y as a func- gross imports This iS clear after considerig the fact that tion of the inputs K and L and the factor intensities net exports are the difference between domestic produc- In matrix notation, this setup can be generalized to a tion and consumption model with multiple products and multiple factors of NX = Y - C = A-' (V - sV,) (6) production as long as the latter do not exceed the num- ber of products, or as long as the model is just identified In principle, empirical models of neoclassical trade the- or underidentified Then ory should be estimated with net exports as the dependent variable, and excess factor endowments as the explanatory Y = A-`V (3) variables In turn, the signs of the estimated coefficients on where Y is the vector of product outputs and V is the vec- the endowment variables, or the values inside the inverted tor of endowments The A is the vector of factor intensi- A matrx, reflect the factor intensities of production ties, which is invertable as long as the production tech- At this point, it is important to note that the inverted nologies are different across sectors so that the ratios of vector A contains factor intensities across product clus- factor intensities across sectors are not identical ters, not relative factor abundances across countries How- Still following Leamer (1984), the production of the ever, each country's consumption share (relative to the world economy as a whole can also be written in the samte world) is a weighted average of its factor shares (also rela- format tive to the world's endowments), so that s is Yw = A-'VW (4) K/Kg, > s > LIL,, or, KlK < s < L/LW (7) Assuming that countries consume commodities in the That is, a capital-abundant country will have KIKW > same proportions, the country consumption levels can be s > L/Lw, while a labor-abundant country will have expressed as KlKW, < s < L/LW set of six countries Of these, only Costa Rua and Mlexzio right after the implementation of NAFTA (Janiary 19941, (Figures 2 8 and 2 10, respectively) show a significant and after the currency devaluation of December 199, change in their trade structure in the sense that at some Beginning in 1995, Mexico became a net exporter of point between 1980 and 1999 they became net exportcrs of machines More disaggregated data reveals that, within the new products In the case of Mexico, this occurred in 1995, cluster of machines, Mexico is an important exporter of 24 COMPARATIVE ADVANTAGE, DIVERSIFICATION, AND INTRA-INDLISTRY TRADE DETERMINANTS AND CONSEQUENCES FIGURE 2.4 Average LAC Net Exports per Worker by Commodity Groups, 1982-97 0~~~ -; 0.8 - D X__ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ __ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ : 0.6--- 0.4- Chemicals Machines 4< 0.2 1 t~~ t--~--k~--f-= -C -I- L Intensive ~~ ~~ - -~~~~_3Capita InlteDsive >,Ei 0 1 - i ---Cereals o -0.2 ~~~~~~~~~~~~~~~~~~~~~~~~~-0-'> -Tropical Agricultuire ----Forest - U- Raw Materials S- 06 1:1~~~~~~~~~~~~~~~~~~~~~~ Petroleum o.6 Z-0.8 -I - I I l I I l I l I I I I l I 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 Note: Sample of 22 countries: Argentina, Barbados, Belize, Bolivia, Brazil, Chile, CoLombia, Costa Rica, Ecuador, El Salvador, Guatemala, Honduras, Jamaica, Mexico, Nicaragua, Panama, Paraguay, Peru, the Repfiblica Bolivariana de Venezuela, Suriname, Trinidad and Tobago, and Uruguay. To see this figure in color, refer to Figure A. 1 in the Annex. FIGURE 2.5 Argentina: Structure of Net Exports, 1980-99 100 -- 60 - EM- - - 0 Chemicals ~~~~~~ 40 ~~~~~~~~~~~~~~~~~~~~~ Machines 40 - _ Capital InieH C In' sive 29 l Labor Intensive 20 - Fl ~~~~~~~~~~~~~~~~~~~~~~~CerealS 4, U~~~~~~~~~~~~~~~~~~~~~~~~~~ Animals Z 0 - - - - - - - -El Tropical Agriculture -20 , ERaw Materials -40 _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ -6() -S O - l l l l l l l l l l l l l l l l l l I °o p\ SW 9b qf 4,5 o'a N$; 'N "2

qbtX N <> >n Cereals ol Tropical Agriculture o Forest * Raw Marerials -20 *~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~E PetroleLoM N,Ntsc To see this figure in color, refer to Figure A.3 in the Annex. 26 COMPARATIVE ADVANTAGE, DIVERSIFICATION, AND INTRA-INDUSTRY TRADFT DETERMINANTS AND CONSEQUENCES FIGURE 2.7 Chile: Structure of Net Exports, 1981-98 80- - 60- o E~~~~~~~~~~~~~~~~~~~~~~~~~~~~~0Chemicals M~~~~~~~~~~~~~~~~~~~~~~~~~~~ ~Machines i Capital Intensive c 5~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~- Labor Intensive C | g n ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~-1 lwr Cerc-als * Animals 5 _ _0;1 ; I S S _ u! i El T ro p ical A g ricu lture 71 L2 Z 0~~~~~~~~~~~~~~~~~~ Forest i Raw Materials m E] ~~~~~~~~~~~~~~~~Perroleum -8()0 1 \ t 'b5' 4NLt r E t '5 '5 'y 'z^ >5 '5 ' '5\ '5'5 Note: To see this figure in color, refer to Figure A.4 in the Annex. illustrates the fact that economic development and poverty the United States, in order to evaluate the factors and poli- reduction need not be hampered by the existence of natural cies that aided the development of these economies. wealth. This report will cover not only the experience of The Dominican Repuiblic (Figure 2.9) is an example of an Chile in the agricultural and mining sectors, but also the economy that, on the merchandise side, has maintained a historical development experiences of now-industrialized clear comparative advantage in tropical agriculture and economies such as Australia, Canada, Finland, Sweden, and labor-intensive manufactures. Its comparative advantage in FIGURE 2.8 Costa Rica: Structure of Net Exports, 1981-99 8()- 60 40 -_ _ __ _ _ ___Chemicals___ 0 Machines 20 -U Capital Intensive El1 Lahor Intensive -80 5 Cereals 1 1 1 1 1 1 1 1 1 1 1 1 1 F C r Note:0 Tropicals Agricutore 2 Raw Materials -40 - - - - - - - - - El~~~~~~~~~~~~~~~~~~~ Petroleum -6o - ~ ~ ~ ~ ~ N ~ N ' ~ < ' 5 5 0 Nore: To see this figore in color, refer to Figure A.5 in the Annex. 27 FROM NATUIRAL RESOURICES TO TIIE KNOW LEDGE ECONO(MY TRADE AND JOB QUALITY FIGIIURE 2'. Dominican Republic: Structure of Net Exports, 1981-99 "40 - so _______ ____ __ _. C . * Chemnicals c - |i - - . [1_____ For ________ - ______-- - M achic es t:-6()- * _ _ _ _* ___ _ E3 Capital Intensive C- ~~~0 __________________________________ O~~~~~~~~~~~[ Labor Intensive -() El Cereals l l l l l l i E Animals - _- 4 - -___- ___ - 0 Tropical Agriculture z Fl~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ Foresr 2 URaw Materials 6(o --_ _ _ _ _ - - _ _ _ _ _ _ _ _ - Petroleum Na;e: To see this figure in color, rcfer to Figure A.6 in the Annex. Data were not available for 1984, 1986. 1987, and 1990(-99. FIGtURE 2.(1( Mexico: Structure of Net Exports, 1981-99 __ 011 U~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ Chiemicals -* Capirat Intensive Il Labor Intensive b 20 F eel - 0 Tropical Agriculture Ul Furest C, __20 M~~~~~~~~~~~~~~~~~~~~~~~~~~ Raw Materials 0 PetroleuIM Note: Tio see this figLire in color, refer tn Figure AT7 in the Annex. agricultural products has persisted for a long time, clearly stimulated by policies establishing EPZs. Like northern due to its relativcly high endowment of arable land relative Mexico, the manufacturing sector of this country has beDl- to its labor force and capital stock. As discussed below, efited from its proximity to the large consumer market of exports of labor-intensive manufactures from the Domini- the United States. Later in this report, the role of commu- can Republic and several Central American countries were nications technology, including the Internet, and interna- 28 COMPARATIVE ADVANTAGE, DIVERSIFICATION, AND INT'RA-INDUSTRY TRADE DET'ERMINANTS AND CONSEQUENCES tional transport costs will be analyzed in terms of their An interesting issue explored below emerges from the importance for attracting FDI in manufacturing activities comparison of Costa Rica and the Dominican Republic. that are vertically linked with a broader manufacturing Costa Rica succeeded in attracting major investment in the production process. That is, countries like the Dominican relatively more "sophisticated" industrial cluster of machin- Republic, which are close to large consumer markets and to ery (that is, computer chips), while the Dominican Republic the source of more capital-intensive inputs for production, has maintained a comparative advantage in labor-intensive will tend to attract foreign investment to take advantage of manufactures. It is likely that this qualitative difference is both their geographic location and their competitive labor mainly due to the difference in the quantity and quality of force. the countries' human capital. The case of the Dominican Observers of the global economy, especially of flows of Republic is also interesting due to its dynamic tourism sec- manufactured products, recognize that firms that export tor. Hence we will cover this sector in Chapter 4. certain goods are facing the pressure of competition from The main issue raised by the recent evolution of the pat- other large developing countries, especially countries such terns of trade of LAC countries is the concern about the as China, India, and Korea. In particular, there are fears remaining, and in some cases rising, dependence on export that the opening of China and India, with their large pools revenues provided by natural wealth, such as mining and of competitive labor, pose significant challenges for exports petroleum reserves and arable land. The following section of labor-intensive manufactures such as apparel and foot- tries to clarify the potential determinants of the observed pat- wear. Box 2.2 reviews the structure of trade in these large terns of trade, and then discusses statistical evidence based on developing countries. international comparisons from throughout the globe. BOX 2.2 The Structure of krade in Large Developing Countbes: China, India, and Korea Figures 2.11, 2.12, and 2.13 show the evolution of net illustrates how an economy integrated into the global exports per worker of manufactures for three large devel- matketplace can naturally change its comparative advan- oping countries: China, India, and Korea. China shows a tage as development progresses. continuous inctease in its net exports of labor-intensive What do these comparisons imply for Latin America goods since the late 1980s, when the process of economic and the Caribbean? First, the emerging patterns of trade in liberalization was launched. India also has a comparative these economies indicate that labor-intensive manufactures advantage in labor-intensive goods, but remains rela- will continue to be a very competitive sector of the world rively closed. The emergence of these international pow- economy, especially as the process of liberalization contin- erhouses in the export of labor-intensive manufactures ues in China and India. Consequently, the survival of labor- poses a challenge for other developing economies wishing intensive industries in LAC will depend on the extent to to expand their labor-intensive exports. which they are able to compete with Chinese and Indian Korea is the most open economy of the three. During exports by playing to their strengths, mainly their close the entire period from 1976 to 1999 Korea maintained a proximity to the United States. Moreover, as discussed clear comparative advantage in manufactures, especially later in this chapter, the adoption of ICT can also help labor- and capital-intensive manufactures. However, since enhance the performance of labor-intensive firms. This 1992, Korea's net exports of machinery increased dra- consideration is particularly important for the Dominican matically, while since 1989 irs net exports of labor-inren- Republic and some Central American countries that have sive manufactures declined steadily. This structural managed to develop labor-intensive industries oriented change in its pattern of trade is probably a consequence toward foreign markets. Koreas pattern of trade is similar of Korea's success in terms of raising real wages and mov- to the one emerging in Mexico. Consequently, we cat ex- ing up the ladder of manufacturing processes. This case pect stiff competition in machines by Korean firms. 29 FROM NATURAL RESOURCES TO TIlE KNOWLEDGE ECONOMY TRADE AND JOB QUALITY FIGURE 2.11 China: Net Exports per Worker, 1984-99 0.10 a (.08 _ L' Liabl Intensive | Cap ne v:,n _ Machines -- Chemicals o.o6- 0 0.01- 0. -00 2 C-00 -0.04 - Z o.o6 1984 1985 1986 1987 1988 1989 199( 1991 1992 1993 1994 1995 1996 1997 1998 1999 FIGURE 2.12 India: Net Exports per Worker, 1978-98 0.02- 0.02 - Labor Intensive {} Capital Intensive] 0.0_ es _ ChSw emicals| | , 0.01 - 0.0 O.J- 0 (.()2 1978 1979 1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 Nokte Data were not available for 1981, 1982, and 1999. Are Traditional Endowments Destiny? quite sensitive to social and economic policies. The follow.- The short answer to this question is "no," having natural ing analysis discusses statistical evidence that shows the riches does not mean that a country will only produce and importance of modern notions of factor endowments. export commodities based on the exploitation of natural resources, land, or both. It turns out that other types of Determinants of Trade Structure: "endowments" also matter a great deal in the determination International Evidence of trade patterns. Furthermore, several of the factors that Given that the previous casual discussion of various pat- help explain why some countries are net exporters of manu- terns of trade simply established that all patterns are theo- factures, while others export agricultural commodities, are retically possible and that natural wealth (that is, in terms 30 COMPARATIVE ADVANTAGE, DIVERSIFICATION, AND INTRA-INDUSTRY TRADE DETERMINANTS AND CONSEQUENCES FIGURE 2.13 Republic of Korea: Net Exports per Worker, 1976-99 1.60- a |4-- Labor Intensive Capital Intensive a 1.40 - | ____ Machines ---- Chemicals v1.20- 1.00- a 0.80- - b 0.60- t 0.40- t _ 2 X0(.20 - v 0.00- C O__<, -g , W -0.20 -__ Z -o>0..0_ I l l l l l l l l l l l l l l l l l lI I 1976 1977 1978 1979 1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 of land and mining reserves) are not necessarily destiny in seems that for any given country, developing a compara- terms of trade patterns, what do data actually show? tive advantage over time in this type of manufactured goods is more likely than for other commodity groups. It Trade and Endowments is likely that this finding is due to the possibility of In this section we discuss empirical evidence based on data attracting FDI in search of competitive labor. More gener- for over 55 countries during 1976-99. Box 2.3 explains the ally, most indicators related to net exports of manufactured data. Greater detail can be found in the background paper goods tend to vary more over time than indicators related by Lederman and Xu (2001). Our objective here is to com- to agricultural commodities. The only exceptions to this pare the influence of traditional concepts of factor endow- rule are the probabilities of being a positive exporter of ments (that is, land, labor, and capital) to the effects of animal products and cereals, which have a relatively low modern notions such as education, knowledge, ICT, insti- ratio in column (3). tutions, infrastructure, and volatility. The fact that changing comparative advantage over time Table 2.7 shows in dicators of the patterns of trade and seems to be relatively difficult (as shown by the relatively their potential determinants for our sample of countries. low variation of the trade variables within countries) does Column (1) shows the standard deviation of the variables not mean that nothing can be done. Indeed, the data dis- across countries; column (2) shows the average standard cussed so far show that it is possible to experience struc- deviation within countries (that is, over time); and column tural changes over time even if this is relatively rare when (3) shows the ratio of the first divided by the second. This compared to the variation that exists across countries. The last set of numbers is interesting because it provides a sense real issue is, therefore, what can be done either to enhance of how much each indicator varies across countries relative the performance of the industries in which a developing to its variation within countries (that is, over time). country has a comparative advantage or to change its struc- An interesting finding is that the value of net exports of ture of trade. Figures 2.14, 2.15, and 2.16 shed some light labor-intensive manufactures and the probability of being on this question. a positive exporter of these goods are the trade indicators Each graph contains the difference between the average with the lowest relative variance across countries, when value of each endowment indicator for countries (and years) compared to the variation over time. For example, the that have a comparative advantage (that is, positive net probability of being an exporter of these goods varies only exports) minus the average for those countries (years) that 2.2 times more across countries than over time. Hence it do not have a comparative advantage. Each one corresponds 31 FROM NATURAL RESOLURC ES TO TIIE KNOWLEDGE ECONOMY: TRADE AND JOB QUALITY BOX 2.3 Dab; Tradiional and New Endowments The traditional explanatory variables of comparative advan- data, namely the value of net exports and the condition of tage are the endowments of crop- and forestland (hectares being a positive net exporter, both change much less over per worker) and capital (per worker). The "new" variables time than many of the explanatory variables. The cross- considered here include our index of institutional quality (a country variance of the value of net exports is two to nine composite index based on data from the International Cozen- times greater than the within-country variance. The con- try Risk Guide), ICT, domestic land transport infrastructure dition of being a net exporter has an even higher relative (a combination of the length of paved roads and radIwdys), cross-country variance than the value of net exports for the index of knowledge (R&D expenditures, technical work- most sectors, except animal products, and to a lesser extent ers as a share of the labor force), the .a e race years of educa- labor-intensive manufactures and cereals. tion of the adult population, and the volatility of the real All the variables chosen as explanatory variables in the effective exchange rate. The data cover 1976 to 1999 for comparative advantage equation-that is, those listed in over 55 countries. The regressions were conducted after bold letters-are relatively more stable over time than dropping the top and bottom 1 percent. the value of net exports. The only exception is the index As mentioned, the explanatory variables were divided of ICT, whose over-time variance is surprisingly high. into two categories; those that determine comparative This is due primarily to the very fast growth of the num- advantage and those that determine the value of net ber of registered Internet hosts in the late 1990s, when exports per worker. Table 2.7 shows the standard devia- many countries went from having zero to having positive tions for the dependent (the sign and value of net exports) numbers. Therefore, we opted to include it in the com- and explanatory variables. Column (1) shows the standard parative advantage together with the other endowment deviation of the variables across countries; column (2) variables that change little over time. All characteristics shows the average standard deviation within countries of neighbors were put in the trade intensity function, (that is, over time); and column (3) shows the ratio of the thus limiting the variables in the comparative advantage first divided by the second. The first obvious observation equation to relatively stable home-country characteris- is that the cross-country variance for all variables is greater tics. (For more detailed information on the data and across countries than within countries. However, the trade methodology, see Lederman and Xu 2001.) to a given agricultural or manufacturing industrial cluster, ments of all types, except for land. But these countries have again based on Leamer's commodity aggregates. Some inter- particularly high levels of the "modern" endowments. This esting generalizations can be drawn from these pictures. is especially the case of knowledge for capital-intensive An interesting similarity exists between countries with exports; knowledge, education, ICT, and institutions fitr a comparative advantage in tropical agriculture and those both machines and chemicals. Hence it is clear that havit-ig with a comparative advantage in labor-intensive manufac- land resources does not automatically mean that a count y tures. The countries that export these goods seem to have cannot develop a comparative advantage in manufacturts. relatively low levels of virtually every endowment, except More important, countries that export certain agricultural land in the case of tropical agriculture. In contrast, coun- commodities tend to be very sophisticated in terms of their tries with a comparative advantage in forestry and cereals level of technological and institutional development. have higher endowments of all types than countries that The main weakness of the analyses presented so far is import those goods. Countries with a comparative advan- that they ignore the fact that all these endowments inter- tage in forestry have particularly high levels of infrastruc- act to produce particular patterns of international trade. In ture; those with a comparative advantage in cereals have other words, the simple correlations presented here should particularly high levels of education and infrastructure. not be interpreted as evidence about the precise determi- Finally, countries that export capital-intensive exports, nants of comparative advantage. The following economet- machines, and chemicals have higher than average endow- ric analysis is required to ascertain the extent to which the 32 (COMPARATIVE ADVANTAGE, DIVERSIFICATION, AND INTRA-INDUSTRY TRADE DETEkRMINANTS AND) (ONSEQUENCES TABLE 2.7 Variance of Variables of Interest across and within Countries VARIABLES (I): BETWEEN S.D. (2): MEAN WITHIN S.D. (3): (1)1(2) Net exports of: Raw materials 0.40 0.0)7 5.71 Forestry 0.54 0.(8 6.75 Tropical agriculture 0.17 0.05 3.4(o Animals 0.48 (.()8 6.00 Cereals 0.19 (1.05 3.80 Labor intensive 0.31 0.13 2.38 Capital intensive 0.21 0.07 3.00 Machines 0.85 0.27 3.15 Chemicals 0.31 0.10 3.10 Probability of being a positive net exporter of: Raw materials greater than () 0.46 0.06 7.67 Forestry greater than 0 0.44 0(.05 8.80 Tropical agriculture greater than 0 o.46 0.05 9.20 Animals greater than ( 0.43 0.15 2.87 Cereals greater than 0 0.43 (.12 3.58 Labor intensive greater than 0 o.42 0.19 2.21 Capital intensive greater than ) 0.40 0.1() 4.()0 Machines greater than 0 0.3() 0.06 5.()0 Chemicals greater than 0 0.34 0.05 6.80 Explanatory variables: Annual rate of change of the real exchange rate 0.33 (1.25 1.32 ICT Index 1.79 1.27 1.41 Institurions Index 1.55 0.44 3.52 Capital per worker 1.29 0.30 4.31) Income per capita 1.03 0.22 4.68 Adjusted Openness Index 47.09 8.39 5.61 Years of schooling 2.67 0.46 5.80 Knowledge 2.49 0.38 6.55 Cropland per worker 11.72 0.07 10.29 Forestland per worker 1.43 0.10 14.30 Land Transport Index 1.48 0.10 14.80 iNote: See Box 2.3 and Lederman and Xu (20(1) for details on the data. Variables in bold were used as determinants of comparative advanrage In the econometiric exercises. endowments affect the likelihood that any given country at emergence of sophisticated manufacturing activities, which any given time will have a comparative advantage in a cer- in the future might become completely detached from the tain industry. aforementioned sectors. The main question posed in this section was whether Econometric Evidence natural endowments are destiny in terms of determining Perhaps the most important message from the discussion of the pattern of comparative advantage across countries and the various patterns of integration is that, in the current over time. The empirical evidence presented above indi- context of globalization, it is difficult to predict where cates that land and capital per worker do play an important industries will be located in the future. The main reason for role in determining comparative advantage. However, they this is that international flows of intermediate goods, are not destiny, because other country-specific characteris- knowledge, capital, and labor, facilitated by low transport tics, which are arguably more sensitive to public policies, costs and improvements in communications technology, also play an important role. This general conclusion comes open up opportunities for industries of various degrees of from the descriptive evidence. However, the previous technological sophistication to be established almost any- analysis was based on an examination of the characteristics where in the world. Furthermore, agricultural and eco- of countries that have positive net exports of the various nomic activities based on the discovery, extraction, and types of goods, which cannot establish the precise impact processing of natural resources offer opportunities for the of the various country endowments on comparative advan- 33 FROM NATURAL RESOURCES TD THE KNOWLEDGE ECONOMY TRADE AND JOB QUTALITY FIGURE 2 14 Characteristics of Countries with a Comparative Advantage in Tropical Agriculture, Raw Materials, Petroleum, and Labor-intensive Manufactures (Compared to the Rest) 4- _ Tropical Agriculture * Petroleum * Raw Materials U2 Labor Intensive 3- 0 2- o2 E I- 0- -e _____rf- r --2 2 ~- I I I I Institurions Infrastructure ICT Cropland Forestland Knowledge Education Capital Volatility EIGUJRE 2.15 Characteristics of Countries with a Comparative Advantage in Forestry, Cereals, and Animals (Compared to the Rest) 4-- F Forest U Cereals * Animals o2 0 L u: ~~ -H -H= 1- L I 4-- _L AL - -12- -2 - I I I I IIII Institutions Infrastruccure ICT Cropland Forestland Knowledge Education Capital Volatility FIGLIRE 2.16 Characteristics of Countries with a Comparative Advantage in Capital-intensive Manufactures, Machines, and Chemicals (Compared to the Rest) -2 Institutions Infrastructure ICT Cropland Forestland Knowledge Education Capital Vrolatiliry 34 COMPARATIVE ADVANTAGE, DIVERSIFICATION, AND INTRA-INDlTSTRY TRADE DETERMINANTS AND CONSEQUENCES FIGURE 2.17 Marginal (Averagel Effects of Country Characteristics on Comparative Advantage (Comparative Advantage = Probability of Positive Net Exports; Only Statistically Significant Effects Shown) 0.5 0.4 - Land per worker' * Capital per worker 2 Knowledge E] Education * Infrastructure Institutions 2 ICT * Volatility[ 0.3 - l 0.2- -0.1 2 0. _045 -0.6 Forescry Tropical Animals Cereals labor Capital Machines Chemicals Agric-ulture Intensive Intensive Agriculture Sectors Manufacturing a, Except knowfedge, infrastructure, institutions, ICT; impact of ont standard deviation, which are indexes. b. Either foresttand or cropland; the stat. Significant one used. Note: Education is general-, knowledge includes technicians/patents/R&D. Sample: Over 55 countries, 1976-99. Source: Lederman and Xu 2001. tage because many of these characteristics move together. On average, small improvements in education (besides Hence further econometric analysis that examines the cropland) can have significant positive consequences for net "1marginal" effects of each of the country eDdowmneDts can exports of tropical agricultural products. In interpreting be a useful complement. this result, readers should be aware that this product cate- The main econometric results concerning the determi- gory includes commodities, such as fresh fruits, which are nants of comparative advantage are summarized graphically exported by relatively (compared to developing countries) in Figure 2.1t7; the methodology is discussed in Box 2.4. educated countries, such as Chile and New Zealand. Like- Each bar shown in the graph indicates the impact of the cor- wise, net exports of cereals benefit from small increases in responding country endowment on the probabilitv of com- cropland, general education, and infrastructure. These parative advantage in the agricultural and manufactured- results for cereals clearly reflect the influence of countries goods sectors. such as Australia, Canada, and the United States, which For given endowments, small increments in forestland, have highly educated populations and extensive roads and knowledge, and infrastructure have significant effects on railways. Also, on average, countries that raise their general the likleihood of having a comparative advantage in education and transport infrastructure on the margin will forestry exports. These results are driven by several Scandi- increase their chances of developing a comparative advan- navian countries, such as Finland and Sweden (see Chapter tage in animal products. 3), which invest in R&D and have developed important All else remaining constant, countries that raise their technologies in a variety of areas, including forestry. Like- endowments of knowledge and ICT will increase their wise, these countries also have advanced domestic infra- chances of developing a comparative advantage in labor- structure. Furthermore, sorme developing countries that intensive manufactures. This result is strikingly consistent export forestry products, such as Chile, also have relatively with the trade theories that emphasize the role of coordi- more advanced infrastructure than other developing coun- nation costs, and logistics costs more generally, in stimu- tries. Finally, comparative advantage in forestry is impeded lating the fragmentation of production across borders. That by macroeconomic volatility. is, the labor-intensive stages of production are made more 3- FROM NATURAL RESOURCES TO THE KNOWLEDGE FE ONONMY TRADE AND JOB QlALITY BOX 2.4 Eeonometric Metdhodolo. Heckman Selection Models In using the Heckman selection model (Heckman 1976, unobservables for the selection equation and the main 1979), we simultaneously estimate a model of the deter- equations are alLowed to be correlated. The equation of the minants of comparative advantage with another of export determinants of export (import) status is the "comparative (and separately, import) intensity. Factor endowments, advantage equation," The equation of the determinants of broadly defined, determine whether the country is a net net export values is the "trade intensity equation." In esti- exporter (or a net importer); then trade policy (that is, mating the import and the export segments separately, we "adjusted openness"), the institutional environment, infra- have allowed the coefficients for the net export equation to structure, and economic size (of the domestic economy differ across the net-export and the net-import subsamples, and of its principal neighboring countries) determine the which provides an intuitive way to allow for nonlinearities value of net exports (imports). Formally, in our estimates of the determinants of trade intensity. Due to data limitations, the analysis does not include 1= X, f3 + Si deposits of minerals and petroleum. To correct this omis- i = 1 if I* > o sion, we conducted additional statistical exercises, using NXit = Xp± . + i the condition of being a net exporter of raw materials and where cov (Ea,, /,,) 5f 0 petroleum as a proxy. The results discussed in the main text for agricultural and manufacturing products did not where 1* represents the index function for exports (im- change significantly. Another caveat is that the "endow- ports); a positive value of the function leads to the status of ments" might be "caused" by trade. Hence the econo- net exporter (importer). The N)X,t function represents the metric models used the lagged values of the endowments value of net exports (imports) per worker. Note that the in the analysis discussed here. profitable when countries raise their levels of ICT develop- tions, especially those involving the enforcement of intellec- ment, which helps producers coordinate their activities tual property rights, and even those regulating the inceni- better with their input suppliers located in the developed tives for R&D expenditures. The good news is that small (capital-rich) countries. improvements in institutional quality might have large pos- Small improvements in knowledge hlave even greater itive effects on these sectors. More generally, the results show effects on comparative advantage in capital-intensive manu- that since the mid-1970s, new endowments have had signif- factures-more so than for labor-intensive goods. Also, mar- icant effects on the patterns of trade around the world. ginal increments in knowledge and in the amount of capital To finalize the discussion about whether traditional per worker have significant impacts on the comparative endowments are destiny, Table 2.8 reports the possible advantage in machines. For a given level of knowledge, tiny ranges of the variance of comparative advantage attribur- improvements in institutional quality favors comparative able to traditional notions of endowments (land and capital advantage in chemicals, which are also exported by countries per worker) and to the modern notions of endowments with high endowments of capital per worker and land-trans- (knowledge, schooling, infrastructure, institutions, ICT, port infrastructure. As discussed earlier in this chapter, the and so forth). The table shows the plausible ranges esti- knowledge index used for this analysis is composed largely mated with various analysis-of-variance (ANOVA) exer- of outcome indicators of R&D expenditures and the number cises, which had the predicted probabilities from the previ- of patent applications. The fact that the quality of institu- ous models as the dependent variable. tions has a large (marginal) impact on comparative advan- The evidence shows that only the variance of comparativti tage in chemicals indicates that a proper institutional frame- advantage in raw materials and forest products explained by work for the creation of knowledge can be greatly improved the conventional factor endowments are greater than the with (statistically) small efforts to reform domestic institu- variance explained by the modern concepts, which are more COMPARATIVE ADVANTAGE, DIVERSIFICATION, AND INTRA-INDUSTRY TRADE. DETERMINANTS AND CONSEQUENCES TABLE 2.8 The Role of Conventional and New Factors in Accounting for Patterns of Comparative Advantage across Countries during 1976-99 SHARE OF VARIANCE EXPLAINED BY "CONVENTIONAL" FACTORS: "NEW" FACTORS: INSTITUTIONS, ICT, CROPLAND, FORESTLAND, AND INFRASTRUCTURE, VOLAT1LITY, COMPARATIVE ADVANTAGE IN: CAPITAL PER WORKER EDUCATION, AND KNOWLEDGE Raw Materials (.35-0.78 0.19-0.62 Forestry Products 0.52-0.76 0.19io.40 Tropical Agricultural Products 0.09-0.55 0.40-0.87 Animal Products 0.11-0.35 0.65-0.89 Cereals 0.25-0.54 0.42-().72 Labor-Intensive Manufacturing 0.15-0.35 11.62-0.83 Capital-Intensive Manufacturing 0.02-0.12 0.86-0.96 Machinery 0.06-0.62 0.38-0.94 Chemicals 0.22-0.61 0.39-0.78 Note: The dependent variable in the ANOVA exercises for each row is the predicted probability of exsporting the specific product: the underlying model is the selection equ,ation of the Heckman model. The reported number in the cells are the share of the variance attributable to each group of variables. Since the shatr depends on the 'ordering" of eth explanatory variables in tho ANOVA equation. we report the range of shares nuder diffecent cu-binat-ons. malleable through public policies. More precisely, for raw this is especially strong in labor- and capital-intensive materials, the minimum and maximum shares of the con- manufactures. In the latter, the new factors explain be- ventional factors are higher than the corresponding shares of tween 86 and 96 percent of the variance. In the cases of the new factors. In the case of forest products, the minimum machinery and chemicals, the maximum share reaches 94 share of the conventional factors is greater than the maxi- percent for machines and 78 percent for chemicals. mum of the modern factors. However, as mentioned earlier, From a public policy standpoint, the overall results are we acknowledge that the explanatory power of traditional encouraging. It cannot be overemphasized that land endow- factors would probably be slightly higher if we included ments and other natural riches are not destiny. Developing mining and petroleum reserves, although our additional countries can develop comparative advantage in manufactur- estimations using rather imperfect proxies for these vari- ing activities through sustained efforts to raise the quality of ables yielded very similar results. domestic institutions, improve the public transport infra- There are some surprising results concerning comparative structure, and help the private sector adopt new information advantage in agricultural and animal products. The share of and communications technologies. Perhaps more important, the global pattern of comparative advantage in tropical agri- policymakers should not underestimate the role played by cultural products and cereals explained by the new factors general education and knowledge in determining the global could be as much as 87 to 89 percent of this variance. The patterns of trade. Although we have emphasized that for any conventional factors could explain no more than 54 to 55 given level of educational attainment of the adult popula- percent. In the case of animal products, the evidence is even tion, knowledge creation is a lkey factor in the performance of stronger in favor of the new factors, because the minimum most manufacturing activities, you cannot have knowledge possible share of the variance explained by these factors is without general education; engineers and innovation are not almost twice as high as the maximum variance possibly created in a vacuum. Yet knowledge by itself seems to have explained by the more traditional notions of factor endow- strong and independent effects on the capacity of countries ments. In our view, these results are surprising because agri- throughout the globe to export manufactured goods. cultural products obviously require land, but we have already The following sections explore the causes and conse- discussed the influence of domestic infrastructure and gen- quences of other structural features of international trade, eral education on agricultural comparative advantage. namely the degree of export diversification and IIT. These It is perhaps less surprising that the new factors explain featutes have important consequences for the pace of eco- relatively high shares of the global variation in comparative nomic development, and their determinants are somewhat advantage in manufactured goods. The evidence shows that different from the factors discussed thus far. 37 ,ROM NATURAL RESRDiR ES TO THE -KNOWLEDGE ECONOMY- TRADE AND JOB QUALITY Export Diversification and lIT: Where Are We, insritutions and infrastructure to ameliorate the impart of and Does It Matter? geographic distance from the major world markets. A recurrent preoccupation of LAC policymakers is that This report links sectoral comparative advantage with their natural riches produce a highly concentrated structure the issues of export concentration and IIT. It looks at the of export revenues, which then leads to economic volatility evidence concerning the relationship between comparative and lower growth. This section examines the evidence con- advantage in the 10 industrial clusters discussed previously cerning export revenue concentration. It first compares the and the degree of export concentration and IIT. We attempt region's degree of export revenue concentration with those to answer the question of whether sectoral policies can help of other parts of the world. The evidence indicates that the reduce export concentration and raise IIT. The data show region enjoyed declining export concentration during the that there is no clear relationship between the degree of con- 1990s, after the trade reforms. This declining trend is also centration and IIT across sectors. Consequently, we con- evident in the set of LAC countries. After presenting these clude that sectoral policies that aim at addressing export facts, we explore the relationship between export revenue concentration and IIT by "changing" a country's compara- concentration and economic growth. The evidence is clear: tive advantage might have unexpected results. Indeed, concentration reduces subsequent economic growth, and attempts to increase exports of manufactures might actually only a fraction of this effect is explained by the correlation reduce export diversification and IIT. between export concentration and macroeconomic volatil- ity. It is likely that a large portion of the remaining nega- Export Diversification and IIT in LAC tive effect of concentration on growth is due to the nega- There are various ways to measure the degree of concent-ra- tive relationship (but not necessarily causation) that exists tion of any economic variable. For example, the menu of between concentration and the incidence of ITT (that is, the indicators for measuring the degree of concentration of in- share of total trade that is composed of imports and exports come across households is quite large, and includes :he in the same product class). Gini index, the Theil index, and the Herfindahl index. F or IIT is correlated with the pace of economic development the moment we rely on the Herfindahl index to examine because it is driven by productivity gains. Much of the recent trends in export concentration in LAC. The case of recent literature on trade has focused on scale economies in Argentina is examined in Chapter 4 and alternative miea- production, which can produce patterns of trade that are sures of concentration are used therein. The advantage of unrelated to factor endowments. Thus, economies of scale the Herfindahl index is that it is not sensitive to the incor- provide an explanation for trade flows among countries poration or elimination of product lines.' with similar endowments and that export similar products Figure 2.18 shows the evolution of the median (or typi- to each other. Regarding the relationship between IIT and cal) index of concentration of export revenues for the region economic growth, IIT is more likely to be driven by prod- compared with the rest of the world. The export concen- uct differentiation, productivity gains, and/ or the life cy- tration of the typical LAC country experienced a secular cle of product development (see, for example, Krugman decline since 1981. Nonetheless, the degree of concentra- 1979).' That is, IIT is driven by economies of scale and tion is still higher than in Asia, Eastern and Central productivity gains, which in turn should be reflected in a Europe, and the industrialized countries of the OECD. country's growth rate. Another interesting aspect of these indicators is that the To determine the stylized facts about the recent evolu- higher they are the more volatile they are. This is revealed tion of the incidence of IIT in the region as a whole, and for in Figure 2.18, where the corresponding indexes for Sub- our core set of countries, we studied the determinants of Saharan Africa and the Middle East and North Africa are IIT across the globe, using annual data from 1970 to 1997. quite volatile. This is expected because a high concentra- We found that public policies can affect the incidence of tion of export revenues in a few products will also result in IIT. Among the most important policies to consider are a high sensitivity of the index to particular variations in the increasing the rate of education of the population so that international prices of those commodities. Thus we cannot the region can catch up with the rest of the world, main- say much about the relative position of Sub-Saharan Africa taining an open trade regime, and improving domestic to Latin America and the Caribbean in the last three obser- COMPARATIVE ADVANTAGE, DIVERSIFICATION, AND INTRA-INDUSTRY TRADE DETERMINANTS AND CONSEQUENCES FIGURE 2.18 Median Herfindahi Index of Concentration of Exports by Regions (Minimum of 11 Country Observations per Year; Calculations Based on over 2,000 Product Lines, Four-Digit Standard International Trade Classification [SITC]) Or_~~~~~~~~~~~~~'K 0.25- - "_ 0.05 - g ~~~~~~~ ~ ~~~~~~~ ~~I I I I I , I, , I s I , _ I - Asia >K- Europe and Central Asia Latin America and the Catibbean - Middle East and North Africa 3- Sub-Saharan Africa -4- OECD Note; Data were not available in all yearc for all countries. FIGURE 2.19 Heifindahi Index of Merchandise Export Revenue Concentration (Calculations Based on 2,208 Product Lines, Four-Digit SITC) 0.6 - - ". Argentina 0.5 - A Brazil '>K' Chile Costa Rica 0.4 - If Dominican Republic Ecuador A -.~~~~~~~~~~~~~~~~~-- ~~~~~~~~~Mexico 0~1.3 - 0).2 _ / . - 4 0.1 - 0. ____-a---'-x--- - -D o~~~ - l')~ l l l- - l l l l l l l -l l- Note: Data were not available in all years for all countries. vations in the charr. That is, we do not know whether this whether this declining trend is also present in our core set was due to, for example, a temporary decline in the price of of countries. Africa's major commodity exports. In contrast, as men- Figure 2.19 shows the indexes of export concentration tioned, LAC did experience a secular decline in its concen- for a set of LAC economies. Of the sample, Brazil has the tration of export revenues. The question that remains is most diversified export structure, which is reflected in the 39 FROM NATURAL RESOIJRCES TO THE KNOWLEDGE ECONOMY. TRADE AND JOB QUALITY relatively low values of its concentration index. It has been export revenue concentration, except for Costa Rica in 1999. diversified for some time, especially when compared to the This is to be expected from a small economy that is highly rest of the sample. Argentina also has a diversified export dependent on oil exports. But it also experienced a process structure, and has experienced a decline in concentration of diversification or falling export concentrarion. In sum, since 1991. This experience is examined in detail in Chap- the region as a whole and most individual countries expe- ter 3. Mexico experienced the most dramatic decline in rienced an increase in export diversification after trade lib- export concentration. In fact, since 1990 (which was not eralization.7 The discussion in the following section sug- the year of Mexico's highest level of concentration) the gests that these trends are good news for the futare Mexican index declined from about 0.13 to 0.02 in 1999. development of the region. By the end of the 1990s Mexico had reached the Brazilian Regarding the share of IIT, Figure 2.20 shows that level of diversification of exports. LAC's IIT rose between 1980 and 1998, but it is still Chile also experienced an important decline of export behind other regions. Nevertheless, the regional upward revenue concentration, but it remains relatively concen- trend of IIT also appears in the individual LAC countries trated, especially when compared with the aforementioned examined in this report (Figure 2.21). Overall, the upward countries of South America and Mexico. Costa Rica's ex- trends in export diversification and IIT in LAC countries perience is also interesting. Like the rest of the sample, coincided with the implementation of trade liberalization it experienced a steady process of diversification during throughout the region. Whether these trends represent the 1990s, until the very end. In 1999, after Intel began good news for the region is an empirical issue addressed in exporting computer chips from Costa Rica, the country's the following section. export structure experienced a dramatic increase in the level of concentration. It remains to be seen whether this Export Diversification, IIT, and Growth degree of concentration will prove to be transitory as the Why should the concentration of export revenues hamper economy begins to develop new potential export products. economic growth? There are various plausible channels For now, however, this case raises the issue that strong through which the concentration of export revenues in a foreign investment and the emergence of a new manufac- handful of commodities can affect growth performance. turing industry will not automatically lead to greater di- First, the concentration of exports could be associated with versification of export revenues. Indeed, Costa Rica will macroeconomic volatility. This might happen because coun- probably face macroeconomic challenges in attempting to tries that rely on a few commodity exports can face more manage swings in the international price of computer severe and sudden variations in their terms of trade. In turn, chips. However, the volatility of high-tech product prices economic uncertainty can hamper domestic investment might be lower than that of agricultural commodities, and (Pyndick 1988; Serven 1998; Lederman and others 200 a) it is possible that the country will quickly develop new and stifle other business decisions such as whether to sell export industries, in part due to potential dynamic effects goods abroad or for the home market, thus negatively affe t- of having Intel. At the moment, it is clear that Costa Rica's ing trade (Maloney and Acevedo 1995). Second, export con- high-tech sector is composed not only of Intel, but also centration can be a symptom, rather than a cause, of poor includes a variety of domestic firms that are providing economic growth due to low incidences of lIT. As discussed, inputs to Intel, plus other foreign firms that have expansion IIT is driven by product differentiation and competition plans in the country partly as a result of the signal sent to among firms with monopoly power in their respective prod- foreign investors by Intel's decision. This experience will be uct niches of the international market. An important conse- examined more closely in Chapter 4 (see also Rodriguez- quence of IIT is that it is driven by productivity gains, which Clare 2001). could be reflected in greater aggregate growth rates. Finally, Quality data for the Dominican Republic were not high degrees of export concentration might be related either available for every year. However, the scattered evidence to civil conflicts (Collier and Hoeffler 1998) or to other forms shown in Figure 2.19 seems to indicate that this country of institutional failures due to the existence of large supplies also experienced a process of export diversification since the of natural riches. This has been coined the "voracity effect" early 1980s. Finally, Ecuador clearly has the highest level of by Lane and Tornell (1999).i 411 COMPARATIVE ADVANTAGE, DIVERSIFICATION, AND INTRA-INDUSTRY TRADE DETERMINANTS AND CONSEQUENCES FIGURE 2.20 Grubel-Lloyd Index of IIT by Regions, 1980 and 1998 0.6 - 0.5 - ~ ~ I11980 U1998 0.4- 0.2- 0.I 0- Africa Asia Europe and Latin America Middle East and OECD Other High Central Asia and the North Africa Caribbean Regions FIGURE 2.21 IIT in LAC Countries Since the Early 1980s 0o6 _ 1~~~~-0 Argentina 0(-Costa Rica 0.5 - _ Brazil 0;S Dominican Republic -Ah Chile -49 Mexico/ C, v 0.2 @ 02 -__s _/ > ~~ ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~_-\ 0.1 - --- ,. -'-- 0~ I I I I I I I I I I~~~~~~~~11) Iq I~ 1) I I I I o <° , NOS q¢ oO r b g # O °9 9 9 gW959( 9\ 90 9 Note: Daca were not available in all years for all countries. So, what do the data say? As a preliminary incursion because this simple correlation could be spurious. For into the relationship between export concentration and instance, it could be the result of other variables that pro- economic growth, Figure 2.22 shows clearly that export duce both low growth and high export concentration. The concentration is negatively correlated with subsequent simple correlation is also not convincing because Figure growth. The graph contains the fitted line from a linear 2.22 itself reveals that although the negative correlation is regression. At first sight, the data show that countries with statistically significant, by itself it explains a very small higher levels of export concentration tend to grow slowly, portion of the cross-country and over-time variation of However, this evidence is not enough to be convincing growth rates, thus indicating that there are many other fac- 41 fRODM NATU.RAL RESOU1RCES TO THE KNOWLEDGE ECONOMY. TRADE AND JOB QUALITY FtGU-RE 2 22 Export Concentration and Subsequent Growth 0.2 y = -0.(301x + 0.063 * . * std err: (.00)89) (.o)026) 0.15 - + , , ' = R - 0.0167 B * *J e i * * * 0.05 - -4. *, '*,' -1 .1 - -1.15 - I I . I () ().1 0.2 '(.3 0.4 (.5 (1.6 (.7 (0.8 (.9 Hefindahl Index tors that affect growth. This issue, unfortunately, cannot be in column (2) of Table 2.9. The regression results correspon-d addressed without more sophisticated statistical analysis. A to an exercise similar to the one discussed above, but with discussion on multivariate regression results follows. explanatory variables of interest that might represent chan- Table 2.9 shows the results of two econometric models, nels through which export concentration affects growth. where the dependent variable is the average real-GDP The results indicate that roughly 50 percent of the prevy- growth rate during three-year periods corresponding to ously estimated impact of export concentration on growth is 1976 to 1999 (the same data used in Figure 2.22). These due to the volatility of the exchange rate, the incidence of exercises include several explanatory variables suggested by IIT, and the quality of domestic institutions.'"' When the the existing literature on economic growth (see, for exam- volatility of the exchange rate, the incidence of IIT, and the ple, Barro and Sala-J-Martin 1995). In addition to the stan- quality of domestic institutions are included in the regres- dard controls, we include our indicator of export-revenue sion, the direct effect of export concentration on growth is concentration, and the results appear in column (1). The reduced to half of the previous estimate. In other words, results suggest that a 1 percent increase in the concentra- about half of such estimated effect is due either to the effec-s tion of exports is associated with a 0.5 percent decline in of export concentration on the volatility of the exchange the growth rate of real-GDP per capita. rate, the incidence of lIT, and the quality of domestic inst The studies by Sachs and Warner (1995a, 1997) find tutions, or to exogenous differences in these variables." that the exports of raw materials as a share of GDP are neg- atively correlated with growth. In additional econometric Export Diversification and IM Can zAu I/hin.k Be Done.? estimations (not presented here) we reproduced the esti- We have already seen that since 1976 the degree of export mates by Sachs and Warner. The results suggest that once diversification and IIT has been positively correlated with we controlled for the degree of export concentration, the subsequent economic growth across countries.'2 Again, to statistical significance of raw materials exports vanished.9 It inform policy it is important to understand the determi- is likely, therefore, the Sachs-Warner result operates through nants of these phenomena. We now shed some light o01 the export concentration channel. what can be done by LAC policymakers to raise export Still, a policy-relevant question is what are the channels diversification and the share of lIT. through which export-revenue concentration affects the pace To understand export diversification in LAC countries of economic growth? This question is addressed empirically during the 1990s, we conducted additional empirical analy- -42 COMPARATIVE ADVANTAGE, DIVERSIFICATION. AND INTRA-INDUSTRY TRAI)E DETERMINANTS ANI) CONSFQUJENCES TABLE 2 9 liberalization was associated with reductions in the concen- Does Export-Revenue Concentration Affect Economic Growth? tration of export revenues. This was due primarily to the Econometrc Results reduction of the anti-export bias of protectionist policies. ANNUAL ANNUAL Regional integration also helped to diversify exports. This GROWTH GROWTH particular hypothesis is examined in detail in Chapter 4 RATE RATE... DEPENDENTAVARIABLEE(11 (1 with reference to the case of Argentina's participation in DEPENDENT VARIABLE (I) (2) Initial GDP per Capita -0.343 -0.304 Mercosur. Additional econometric exercises indicated that (11.15)9 (12.63) factors that reduce transaction costs, such as the quality of E1ducation o.o4i 0.025 domestic institutions, also are negatively correlated with (2.699' (2.27)1 Government Expenditures (%9 of GDP) -0.001 0.0(1 export concentration. (0.24) (0.27) Since the early 1970s specialists have devoted much Investment Rate 0.011 0.006 (455)a (3.3 69 effort to trying to understand the determinants of IIT. On Population Growth -0.03 1 -0.016 one hand, studies that use data from industries within (2d or G(1.57) countries have emphasized the role of economies of scale, Trade over GDP (.001 -0.000 (0.70) (0.64) product differentiation, and imperfect competition. In this Exporr Concentration -0.5()8 -0.254 context, industries produce a variety of similar goods with Import Concentration (2.24)b (1.90)' similar factor intensities. Cross-country studies, on the Import Concentration ~~~~-1.1)28 (2.46)b other hand, have focused almost exclusively on bilateral Real Exchange Rate Volatility -0.251 trade flows among industrialized countries. (3.1 i)' IIT Index 0.389 For this study, we conducted an empirical analysis using (2.27)b cross-country annual data covering 1970 to 1997. These Institutions Index 0.004 (0.32) trade data were then used to construct the widely used Constant Grubel-Lloyd (1975) index of the incidence of lIT, which is Oberaton 4the indicator used in the previous sections of this report." Observrations 475 475 a. significant at I percent. Following the existing literature, the analysis focuscs on b. Significant at 5 percent. five types of explanatery variables. First, an important set c. Significant at 10 percent. Nate. Econometric methodology: GMM T-staristics in parenthrses. The variables are in of variables concerns the degree of similarity between the first-differences to get rid of country-specific effects. The instrumental variables are the lagged levels of the endogenous variables, and the differences of out-of-system instru- c mental variables (that is, those included in the determinants of the export Herfindahl world, and consequently the indicators of similarity used index but not in The growth eqoacions). The overidenrifying restriction tests suggest sh.t ac insiLraunens a-e largely valid. Re-Is f-ni fiaed- aid andi-elfeci nidel herein are constructed with reference to world characteris- produced resalts that are consistent svith those presented in this table tics. More specifically, this analysis studies the effect of the "stocks" of education (total years of schooling of the popu- lation), capital, and labor in each country on the incidence ses in two stages. The first stage entailed the estimation of of IIT. The log of the country-level stocks are differenced statistical models where the index of export concentration with respect to the world's total stocks of these variables, was explained by GDP per capita and its squared term for thus producing an indicator of the similarity of each coun- LAC countries in the 1990s. The second stage entailed the try relative to the rest of the world. Second, we study the same analysis but with data for the 1980s. The results from impact of the size of the domestic economy (that is, GDP) these two exercises revealed that the relationship between on IIT, in order to capture the potential effects of so-called development and export concentration was different during economies of scale. Third, the analysis considers the poten- the two periods. This counterfactual analysis clearly shows tial effects of the level of development on IIT, because it is that if LAC countries had not liberalized, and thus would thought that consumers from rich societies tend to prefer have behaved like they did in the 1980s (after controlling product varieties, more so than consumers in the least- for their per capita income levels), the degree of concentra- developed countries. Finally, we consider indicators of the tion of export revenues would have been higher. Hence the extent of international transport and transaction costs and first clear policy implication concerning export concentra- the level of overall international trade (as a share of GDP) tion is that for the vast majority of LAC countries, trade as potential determinants of IIT. 43 FROM NATURAL RESOURCES TO THE KNOWLEDGE ECONOMY, TRADE AND JOB QUALITY Table 2.10 contains the results of the econometric analy- increase in product variety (Krugman 1979). The othier sis. The three indicators of similarity are significant deter- variables are meant to broadly capture the potential effe(:ts minants of IIT, as are the size of the domestic market and of international transaction and transport costs. The corre- the level of development. This set of findings indicates that sponding results show that high ratios of trade to GDP tend large and rich economies with high levels of education tend to be associated with higher IIT; geographic distance (which to have higher IIT indexes. The policy implication is is related to transport costs) tends to reduce IIT; and cur- clearly related to the education variables, which is the most rency unions (which eliminate international transaction policy sensitive of the lot. Countries should emphasize costs related to currency exchanges) tend to increase lIT 14 efforts to increase the provision of general education, so The index of domestic institutions has the expected effect of that poor LAC countries can at some point reach the levels increasing IIT, but it is not statistically significant. of education of the industrialized countries. In other words, This analysis provides a handful of policy propositions, the pace of education improvement should be faster in the and the results are consistent with those found in the liter- poor countries than in the rich countries for the poor coun- ature (see, for example, Stone and Lee 1995; Clark and tries to increase their IIT. Stanley 1998). Perhaps the most important propositions are The rest of the explanatory variables are related to trade related to trade orientation, education, and international orientation, geographic distance from major markets (Brus- transport and transaction costs. The policy implications of sels, Tokyo, and Washington), the quality of domestic insti- the first two are quite clear: trade liberalization will lead to tutions, and the existence of currency unions. The trade ori- higher IIT, and we have already mentioned the need for entation is meant to capture the theoretical proposition that poor countries to "converge" to the levels of education of more trade (relative to domestic production) leads to higher the industrialized world. The latter proposition can take a IIT, where one of the gains for domestic consumers is the long time to come to fruition, but it cannot be overempha- sized that general education plays an important role in determining trade patterns, as is mentioned throughout TABLE 2.10 WhAtL Can Be Done to Raise lIT? Econometnc Resultsthis report. The policy implications related to transport and What Can Be Done to Raise IIT? Econometnc Results transaction costs are also interesting. Perhaps more impor- EXPLANATORY VARIABLES (1) (2) (3) tant, policymakers wishing to raise their country's level of FACTOR ENDOWMENT IIT should consider improving the institutions related to SIMILARITY Capital 0.02Y 0.0313 0.034' customs administration and the infrastructure related to in- Labor 0.020' 0.023' 0.024' ternational commerce (ports, railways, roads), which play Education 0.002a 0.001' 0.003' an important role in reducing the effects of geographic dis- SCALE ECONOMIES tance from the major export markets of the world. GDP 0.047a 0.048" 0.049' The following section conducts further policy analysis CONSUMER PREFERENCES about the relationship between sectoral comparative advari- GDP per Capita 0.058' 0.054' 0.055' tage and export concentration and IIT. The objective is to assess the potential of reducing export concentration, rais- TRADE IMBALANCE 0.130 ing the incidence of IIT by focusing on sector-specific poli- TRADE ORIENTATION cies, or both. (X+M)/GDP o.0181 0.017' 0.019' TRANSPORT AND Export Concentration, IIT, and Comparative TRANSACTION COSTS Advantage: Can Sectoral Policies Help Reduce Geographic Distance -0.014" -0.012' -0.013' Institucions 0.032 0.025 Concentration and Raise IIT? Currency Union 0.002b 0.001' 0.001' Another policy-oriented analysis is to examine the correla- NO. OF OBSERVATIONS 1,873 1,873 1,873 tion between comparative advantage in the 10 sectors dis- a. Significant at 5 percent level. cussed in this chapter and the indexes of export diversifica- b. Signficant at 1o percent level. tion and IIT, both of which are correlated with economic Note: Methodology: Tobit regressions with panel data. Dependent variable is the Grubel-Lloyd index Regional and time dummies vere incuded, as was a constant. growth. Many LAC analysts and policymakers believe that 44 COMPARATIVE ADVANTAGE, DIVERSIFICATION, AND INTRA-INDUSTRY TRADE DETERMINANTS AND CONSEQUENCES countries that export manufactured products automatically a comparative advantage in a particular sector will tend to have higher diversification and IIT. If we find that certain have higher (positive difference) or lower (negative differ- patterns of comparative advantage are associated with ence) levels of export concentration and IIT than countries higher levels of export diversification and IIT, then the pol- without a comparative advantage in the specified sectors. icy recommendations discussed will also apply to the issues Figure 2.23a contains the estimates of the differences of of export diversification and IIT. the averages for countries with a comparative advantage in Figures 2.23a, 2.23b, and 2.23c provide this analytical petroleum and raw materials. Net exporters of these prod- perspective. They show the differences among averages of ucts tend to have higher indexes of export concentration country characteristics, and tell us whether countries with and lower indexes of IIT than other countries. FIGURE 2.23a Export Concentration and IIT by Sector Raw Materials and Petroleum Export Concentration I Raw Materials _ Petroleum IIT -0.15 -0.10 -0.05 0 0.05 0.10 0.15 0.20 Difference of Group Means FIGURE 2.23b Export Concentration and IIT by Sector Chemicals, Machines, and Animals Exporc Concencration *Chemicals jMachines *Animals IIT -0.15 -0.10 -0.05 0 0.05 0.10 0.15 0.20 Difference of Group Means 45 FROM NATURAL RESOUIRCES 10 THE KNOWLEDGE ECONOMY: TRADE AND JOB QUALITY FIGIJRE 2.23c Export Concentration and IIT by Sector Capital and Labor Intensive, Cer eals, Agriculture, and Forestry Export Concentration CapitalInesv Labor Intensive A griclture TIT orest -0.15 -0.10 -0.05 0 0.05 0.10 0.15 0.20 Difference of Gr oup Means Figure 2.23b shows the estimates for countries with a tures, both of which seem to have the desirable qualities of comparative advantage in chemicals, machines, and ani- low export concentration and high IIT? The answer must be mals. These groups share the characteristic that their bars a cautious one: yes, unless the particular policies have other for export concentration and IIT point in the same direc- undesired distortionary effects. For instance, a public pro- tion: Countries with a comparative advantage in chemicals gram designed to improve the infrastructure and regulatory and machines have higher export concentration than other framework to aid the adoption of new information and com- countries, but they also have higher IIT indexes. Net munications technology will help develop a comparative exporters of animals have the opposite characteristics- advantage in labor-intensive manufactures. The goal of thlis lower export concentration, but also lower IIT. type of program should be to increase the adoption of new Finally, Figure 2.23c contains the calculations for capi- information and communications technology for the econ- tal-intensive and labor-intensive manufactures, and for omy as a whole, including reviewing the regulatory frame- cereals and agriculture and forestry products. Countries work for communication and helping to develop the neces- with a comparative advantage in these products share the sary infrastructure (see Mann 2001). These policies, in turn, characteristic that their averages of export concentration are likely to enhance the competitiveness of labor-intensive and IIT point in different directions. Those that are net industries that belong to a broader manufacturing process exporters of capital- and labor-intensive manufactures tend that has been fragmented across borders. This impact is to have both low export concentration and high IIT, while likely to be stronger for this type of industry than for oth- the reverse is true for the agricultural exporters. ers, but this does not mean that the rest of the economy will Overall, the evidence presented in Figures 2.23a, 2.23b, not benefit from these policies. On the contrary, there are and 2.23c indicates that comparative advantage in some likely to be beneficial economywide effects. sectors is associated simultaneously with high export con- centration and high IIT (machines and chemicals), while Summary of Findings about Comparative others are associated with low export concentration and low Advantage, Diversification, and IIT IIT (capital- and labor-intensive manufactures). Moreover, The analyses in this chapter provide several conclusiols some agricultural sectors behave similarly to some manu- that might be useful for policymakers and analysts. facturing sectors. On average, the LAC region has maintained a compara- Should a country implement policies to develop a com- tive advantage in petroleum, raw materials, tropical agri- parative advantage in capital- or labor-intensive manufac- culture (including fruits and vegetables), and animal prod- 46 COMPARATIVE ADVANTAGE, DIVERSIFICATION, AND INTRA-INDUSTRY TRADE DETERMINANTS AND CONSEQUENCES ucts. While there has been a lack of structural change, this 3. A caveat regarding the data for Internet hosts is that all hosts has nonetheless been accompanied by productivity and wel- without a country code are assumed to be located in the United fare gains for consumers due to trade reforms. Perhaps more States, inflating the figure for that country and reducing it for all the other countries. interesting, the regional average masks important differ- 4. The U.S. Pacent Office files the nationality of the first inventor ences across countries. Some countries, especially Costa Rica in the patent application. and Mexico, have moved into the fast-growing global mar- 5. This is especially true when IIT is driven by horizontal rather ket in machines, while Brazil has maintained a diversified than vertical 1IT (Thom and McDowell 1999). Horizontal IIT is structure of net exports by exporting agricultural com- trade in products coming from the same stage in the production modities and several types of manufactures. process, whereas vertical IIT is trade among products, some of which New trade theory indicates that anything is possible in can be inputs in the production of the other products. For example, trade in automobiles is horizontal, while trade in automobiles and the world economy of the 21st century. The evidence shows auto parts is vertical. I that "new" endowments, such as knowledge, ICT, quality 6. The Herfendahl index is defined as H = x,/x, xj , where, domestic institutions, and domestic infrastructure, plusI domesticity, institutions,la and redomesti irastructure,splus stands for a particular product and n is the total number of products. volatility, explain a large share of the trade patterns oh- When a single export product produces all the revenues, H = 1; when served around the globe. Hence natural resources are not export revenues are evenly distributed over a large number of prod- destiny. ucts, H approaches 0. LAC countries experienced a significant diversification 7. Econometric exercises not presented in this report also showed of merchandise export revenues in the 1990s, and a rise in that export diversification was higher in the 1990s in LAC, after con- the share of lIT. Both were partly due to trade reforms, trolling for the level of development. these sren ofaIIT.Bothwere encour tause the statistil ref . 8. Sachs and Warner (1995a, 1997, 1999) argue that dependency These trends are encouraging because the statistical evi- on exports of natural resources and agriculture hurt economic growth dence shows that export diversification and IIT are posi- through its negative Dutch Disease effect on manufacturing industries, tively correlated with subsequent economic growth. The which they argue theoretically can have unexploited economies of region's dependence on natural resources has clearly not scale. Hence when manufacturing falls, aggregate productivity growth prevented these trends, thus indicating that natural re- falls as well. This is an argument distinct from the ones being discussed here about export-revenue concentration on any type of commodity. sources a nt c a. 9. In fact, when using the Generalized Method of Moments Chapter 3 further explores this last point by reviewing (GMM) estimates with instrumental variables, the Sachs-Warner the evidence concerning how natural resources can be used variables sometimes have a positive and significant sign. to promote fast productivity growth. Certain industrialized 10. The incidence of IIT is the Grubel-Lloyd index, which is a countries developed successfully by tapping the potential measure of the share of imports and exports of identical products (at for productivity growth offered by their abundant natural the four-digit level of the SITC. Revision 2) with the world. r l. The concentration of imports appears with a significant sign in Table 2.9. However, this variable was not significant on its own, in the long run had rich endowments of natural resources, and was not robust to the choice of econometric technique. Thus we but they accumulated knowledge while maintaining an out- downplay its potential role. ward orientation. Chapter 4 looks at the specific experiences 12. Claudia Sepulveda (World Bank) conducted the econometric of LAC countries, showing in detail how trade reforms, analysis of IIT. regional integration, the promotion of FDI, and the use of 1dx, -MAe, ' ' ~~~~~~~~~~13. The index is defined as IIT = 1- ' , where indi- modern endowments have improved the economic perfor- (X, +4,) mance of both the emerging natural resource and manufac- cates a product category and n is the total number of products This turing sectors. index varies between 0 and 1, and it shows the share of total trade that is conducted among identical products (that is, imports and Notes exports of the same product category). 1. This subsection is based on Bond (2001). 14. The data on currency unions come from Glick and Rose 2. Ana Maria Menendez (World Bank) wrote an initial draft of (2(0(01). These authors also provide an empirical analysis of the this section. impact of currency unions on the magnitude of trade flows. 47 CHAPTER 3 It's Not Just What You Produce, But How: Lessons from Comparative History C A HAPTER 2 MADE TWO KEY OBSERVATIONS. FIRST, PATTERNS OF TRADE ARE DETERMINED NOT only by traditional factors of production-land, labor, capital, and natural resources-but equally or more by "new" factors-knowledge, education, ICT, and the quality of institutions. The ability to create these factors, combined with the historical decrease in transport and com- munications costs, offers the opportunity for the region to move into industries that were heretofore unimaginable. The second observation is that, while some countries-Costa Rica and Mexico, for example-have experienced dramatic changes in their production structures, most countries remain largely exporters of natural resource-based goods. This second finding could be a concern for two reasons. First, some may interpret it as evidence that there have been few of the promised efficiency gains of trade liberalization. Second, recent literature, in particu- lar Sachs and Warner (1997), appears to confirm a traditional Latin American suspicion that resource-rich countries underperform. This chapter answers both concerns with the same finding: It is not so much what is produced, as houw it is produced. The estimated static efficiency gains of eliminating trade distortions-eliminating the famous Harberger triangles-have never exceeded more than a few percentage points of GDP. It is the dynamic process of productivity growth arising from in- of, their resource base. In fact, as Figures 3.1 through 3.4 novation and adoption of new technologies that propels show, their net exports are heavily resource intensive. development. This, as much as any reallocation of factors The evidence is now strong that the U.S. industrial success among industries, is the promise of trade liberalization. must be considered that of a natural resource-rich nation that Further, the evidence reviewed in this chapter shows that made a gradual transition to resource-rich manufacturing natural resource sectors have as much or more potential for industries (see Irwin 2000). The United States was the rich- gains in productivity as the manufacturing sectors. est country in terms of natural resources, and its transforma- tion into the global manufacturing leader, while somewhat The Lesson from History: Play to Your Strengths technology driven, was sparked by the discovery of iron ore History offers numerous examples of how sustained and reserves in Minnesota that reduced the cost of iron-intensive dynamic growth is possible based on any number of patterns manufactures below those of the competition. Even today, it of export specialization. The manufacturing success of Japan remains first and foremost a net exporter of foodstuffs. and the Asian newly industrialized countries (NICs) is one Innis (1933) and Watkins (1963) had Canada in mind route. However, many of the world's most developed coun- when they developed their "staples theory." They saw pri- tries-Australia, Canada, Scandinavia, and the United States- mary good exports, beginning with fur and fisheries and have successfully developed on the basis of, and not in spite then progressing to forestry and wheat-through either 49 FROM NATURAL RESO(UR('ES TO THE KNOW'LEDGE EBONOMY TRADE AND JOB QUALITY FIGURE A. I Australia: Structure of Net Exports in Selected Industiialized Countries 60- 40- * Chemicals o | | El Mach|ines -sr,- U Capital l l l l l l l l l l l l l l l l laIIntensive Ul Labor Intensive Ul Cereals U Animals Nore: El Trupical AgriCfigureIn color refer toltgll A.8 in UTpaArox c -2 Forest U Raw Materials C Petroleum -60 -Su) cyb q9 Q° C> 4b 4t ,sy , and Ericsson's break- decided to transform Nokia from largely a producer of throughs-have been essential for success in both sectors. paper, tires, and cable to a leading high-tech company. They argue that while it is not possible to systematically The foundation had been laid in the electronics division, create innovations, it is possible to systematically prepare which had expanded with the public radiotelephone sys- for those commercial and technological opportunities and tem in the early 1970s, and had been marketing foreign challenges that will occur. computers (Bull and Honeywell Elliot), and had found a market adapting foreign computer equipment into pack- Sorce: Blnomsrrb and Kokko 2001. 57 FR()M NAT UR AL R ESrISIR( LS IT( T HE K NO'X LEDGE E (ION (MY TRAD F AN H JOB QUALITY FIGURE A 5 Percent Annual TFP Growth (1967-92) 8 - | Developing Counrriesl Fra,- 6 - Developed Countries 1 l)e~~~~~~~~~~~~~~~~~~~~~Initrk Avtr.gc _Ar 4 I-~~~~~~~ Inmmm.m (,rme i - we.lell 'r .l I y 1 liltrle1- , G" 1-1 - V LIemmmmI]UIS i s..jrm L. l k. l sl . c(* I(ImnrI t * °I ~ ,eIWI1~ S(.il. ) _ * Ind . * +~~~~~~~~~~ ~~ PhIImpplmL Uni d Nttr sp erhelmmJ <~~~~~~~~~~~I Smlm. er \j \ Au rmI. h (I - S. mplm Ame,,,, Tkl....l. ..b,c,, .. *(,.R-t _A1_ I l l ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ml l Ie -6 -i 0 ( _ 4 6 MarLufacturlng 5 li: BaCsed on NMartin and Mlira (20(11). BOX 3.4 How Did Australia Come to Run Chilean Mines? Chile's largest producer of copper is run by Broken Hill and a School of Mines was established in Ballarat in 1870 Proprietary Company (BHP), hailing from Australia, a and in Bendigo in 1873. The University of New South country roughly the size of Chile and with a shorter tradi- Wales (ITTNSW) was founded in 1949 on the campus of the tion of copper mining. Though ALu.r.jiaiY nillir, piJiLv Technical College, with the Massachusetts Institute of has fluctuated dramatically across tirri. -nc n\plntnriun TzJihn,i.I's and the Berlin University of Technology as for its ability to export mining tepr:rtn,c lict in hc [-lid- mill-,!. and a core ['m -'AIn on research and teaching in sci- up of human capital in this area over time. Fi L!e ri r, and ence and technology. The UNSW School of Mining Engi- metallurgists were recruited to Australia in 1886 and this neering now ranks as one of the largest educators of min- firmly linked the country to the innovations generated in ing engineers in the world. These institutions rest on a the United States (Wright 1999). Duncan and Fogarty commitment to universal education beginning with the (1984) argue that "geological knowledge and mining Victoria Education Act of 1872 that established free, secu- expertise became part of the Australian ht ris Le, enriched lar, and compulsory education in that state-this in a by schools of mines of world class and the industry has country that until the 1840s had been a penal colony of the been in the forefront in the development and application United Kingdom. of mining and treatment trhn,lo^t." In fact, Australia Out of this context emerged one of Australia's most lagged behind the United States until after P 1412 l-mh InAtlucnil niinini: companies and industrial conglomer- 47 engineers per 100,000 people compared to 128 per ates. In 1883, a boundary rider on a sheep station in New 100,000-but would reach 163 by 1955. Further, several South Wales discovered lead and silver and formed a syn- important universities could offer local beachheads for for- dicate with other local workers, leased the mine, and in eign research. The Sydney Mechanics Institute was estab- 1885 -1. 'atd BHP. Called by those of the region "the cra- lished in 1843 and the Sydney Technical College in 1878, dle of Australian industrialization," mines and smelters both with the goal of the JIm1. i.i -n s, i,n rric l.- ind I i Vi I ld * ).nrrd and i%tlen easy-to-access oxide zones were knowledge; Sydney University tu1X tCraFl hhtd Ir I' t\aiMl'trtd. BHP rnccall urUists and engineers introduced 58 ITrS NOT JUST WIIAT YJOU PROI)UCE, BLUT HOW LESSONS FROM COMPARATIVE HISTORY the flotation process that, as a residual, allowed the in Chile. Australians now lead the world in mineral expansion of zinc production by new firms. BHP diversi- detection technology embodied in their "transparent fied into manufactured steel products, coal and iron ore, earth" initiative, and in mine-closing techniques and petroleum, and shipping and shipbuilding, and eventu- mining-related environmental knowledge, and export ally severed its ties completely with its birthplace. BHP more mining know-how than their famous wines. and similar conglomerates became modern corporations, The Australian case suggests several lessons. First, with vertical control from mining to blast furnaces to mineral wealth was a blessing. Second, literate and free wire rope factories to shipping lines with links to foreign everymen can be an important source of entrepreneur- capital through joint ventures. BHP would gain a global ship and innovation. Third, the ready adoption of min- reach, acquiring mines in Canada and in the state of Utah ing and metallurgical expertise allowed local exploita- in the United States, and would eventually discover and tion of the mines and, eventually, the export of mining undertake the development of La Escondida copper mine knowledge. ered mineral deposits-with having deficient capacity in Box 3.5.) We argue that much of the reason may have to do these areas. with both explicit barriers to adoption, and to a lack of an The case of Australia, discussed in Box 3.4, suggests environment that facilitates innovation. that a similar process of collective learning could be repli- cated later in a small peripheral economy. Australia's BHP Two Final Concerns company found La Escondida in Chile, one of the world's Two concerns that are found in much of the literature on largest deposits of copper. Something similar appears to natural resource-rich countries merit comment. have occurred in Brazil. Baer (2001) argues that modern First, another of Prebisch's concerns was that a declining satellite surveying techniques radically increased Brazil's share of world income accounted for by commodities would known endowments over those known in the late 1960s. lead to worsening terms of trade: Latin America's exports Previously, most mineral reserves were thought to lie in the would buy progressively fewer manufactured imports. Re- mountain ranges of Central Brazil, especially in the state of cent literature has called this diagnosis into question. As Minas Gerais. However, massive deposits of iron ore were part of a larger challenge to what he depicts as a weak and discovered in 1967 in the Serra dos Carrajas in the Amazon, scientifically unsound dependency literature, Stanford histo- which was also found to contain large deposits of bauxite. rian Stephen Haber (1997) reviews the work of numerous Tin reserves now appear to exceed those in Bolivia, copper scholars and finds that during the 19th century the region's was found in Bahia, and offshore oil exploration vastly terms of trade actually rose. He concludes that "the weight expanded available reserves of oil. In Peru, mining exports of the evidence points to the conclusion that there has been doubled between 1992 and 1999, making it the world's no secular deterioration in Latin America's terms of trade, second-largest silver, bismuth, and tin producer, the but rather that there have been cyclical swings with no dis- world's sixth-largest copper producer, and eighth-largest cernable long-term trend." Hadass and Williamson (2001) gold producer. Still, Wright argues, this is far below Peru's reexamine evidence on the Prebisch-Singer hypothesis, con- potential. structing estimates of the terms of trade for 19 developing All three sectors suggest that Latin America's challenge and industrialized countries, and find that the terms of trade is not necessarily getting out of what it presently produces actually imiproved for 11 regions. This is due largely to well, but taking advantage of the possible productivity rapidly declining transport costs during the sample period: gains. The question emerges of why it is foreign technology the decline in the relative price of Latin America's exports that is exploiting Chilean ores, and why it is Australia that that Prebisch's 1870 to 1950 U.K. data showed would have exports more mining know-how than wine, and not Chile, been reversed in any port in Latin America. In a sense, every- which started copper mining over a century earlier. (See body's terms of trade improved, and up to World War I, the 59 FRoM NATU RAL RFSOURC ES TO THF KNOWLEDGE FCONOMY TRADE ANI) JOB QUIALITY BOX 3.5 Why Didn't Chileans Find La Escondida? Chile's largest mine was discovered and managed by Aus- 1775, was established, and sought to improve the coun- tralia's BHP, as discussed in Box 3.4. Why didn't Chile try's agriculture and industry by the diffusion of indus- follow a similar trajectory? Chile's mining sector became trial arts and the promotion and adoption of useful inven- a classic foreign-owned enclave with relatively few link- tions. But even a century later, in 1915, only 19 of the ages. In Chile, A Case of Frustrated Development. Pinto 970 students graduating from the National University Santa Cruz is only the best read of Chile's historians who would complete work in engineering (Will 1957). Chile's point to the willingness to live passively off natural rich culture would produce two Nobel Prize winners in wealth, a culture of conspicuous consumption, and defi- literature, a major surrealist painter, and first-class musi- cient entrepreneurship among the Chilean elite. But in cians. But it lagged behind in engineering sciences. other areas, Chileans were very entrepreneurial. Exports Arguably this weakened Chile's capacity for innova- boomed after the elimination of Spanish restrictions on tion and adoption of new technology. Pinto (1959) spec- trade. Copper production by Chileans increased four-fold tacularly underlines the foregone power of a gradual from 1844 to 1860. In response to increased demand accumulation of know-how to the maintenance of com- arising from the gold rush in California in the United petitive position and generation of new areas of compar- States, Chilean wheat exports rose ten-fold in value from ative advantage: 1848 to 1850, and area under wheat cultivation more than tripled between 1850 and 1870 (Conning 2001). te chnological deands of th darlie An alternative cause may lie in the historical lack of an period in contrs o whatris ocrring today some areas of mnining or industries, were relatively infrastructure for scientific learning and innovation. Until the 20th century, Chilean governments remained arnbiva- mode an thu nottoosl What could and ha lent about universal primary education, and higher stud- to be done i the national mining companies and in agriculture ... was perfectly compatible with the ies required going to Lima until the establishment of the University of San Felipe in 1758. Most students studied resources accumulated in the long periods of theology and law; medicine was considered unsuitable bonanza. If the process had been initiated and main- for the uppe class,andmathematicswasparticularl tained adequately, without doubt, it would have cre- for the upper class, and mathematics was particularly ae h en ocnrn oecalnigtss neglected. A major break occurred in 1797 with the estab- such as those posed by copper mining when it was lishment of the Academy of San Luis, which attempted to r ~~~necessary to exploit less rich veins. However, faced introduced more practical courses in languages, arith- metic, geometry, and drafting, all of which its founder with the technological revolution, the local mining ' ' . ~~~~~~~~~~~~companies had behind them neither sufficient accu- saw as essential to the development of the colony's agri- mpated r es,no the nitionalior adcu- cukur indutry nd comerce Burrhis emaind an mulated resources, nor the organizational or admin- culture, industry, and commerce. But this remained an istrative capacity that were indispensable. In these uphill battle in a society where traditionally the Inquisi- tion enforced, albeit incompletely, religious orthodoxy, cinrouctinc f frei cap an o expetise. and where the largest collection of books at the time was introduction of foreign capital and expertise. the 5,000 at the University of San Felipe, which focused Chilean copper mines were seen as inefficient, poorly on theology, canon law, and the lives of the saints, none of managed, and lagging behind technically. The big and which had been printed within the previous half-century. visible advances were in the Guggenheim mines at El In 1819, revolutionary leaders would establish the Insti- Teniente and Chuquicamata. In the end, rather than any tuto Nacional, an amalgam of the existing institutions that lack of entrepreneurial spirit, there may have been an would focus on, in addition to -ht'loLoN, engineering, nat- implicit acknowledgment that technically, Chileans ural sciences, and medicine. In 1813, the first Sociedad could not managc their resources as well as the British Economica, patterned on those that took root in Spain after and later the Americans could. But the path chosen 60 IT'S NOT JUST WHAT YOU PRODUCE. Bt T HOW LESSONS FROM COMPARATIVE IHISTORY meant that no indigenous mining universities or institu- in a country that began exporting copper long before the tions would emerge, and there would be limited indige- United States or Australian firms that would dominate nous capacity developed for innovation or analysis. the industry in Chile. Even today, there is relatively lit- Meller (2001) argues that "in the 1950s one could have tle interaction between the copper companies and uni- learned more about Chilean copper in foreign libraries versities or other think tanks. than in Chilean ones. Neither was there training of Lagos (1997) argues that one way to continue devel- Chilean engineers and technicians specializing in cop- opment of the North after the inevitable decline in min- per." The fact that in 1952 the ( ,,nroIller General admit- ing production is for the region to transform itself into a ted that he had no idea of what went on in the companies regional service center for the mining industry of neigh- suggests that part of the feeling of vulnerability and boring countries Argentina, Bolivia, and Peru. A crucial dependency must be attributed to the lack of technical aspect of this is the development of advanced science capacity to monitor and confidently critique the actions in the fields of the environment, production, and explo- of the Gran Mineria. It was not until 1955 that the Cop- ration technologies, markets, applications, and other key per Department was created to oversee U.S. firms' copper issues. Lagos suggests that a possible starting point operations, and a bureaucracy established of Chilean pro- would be for the state to create a research fund directed fessionals, engineers, and economists, as the basis for to strategic research projects, not only in Chile, with the local Chilean expertise. "In short," Meller argues, "it took target of positioning copper globally in all these fields. about 40 years, from 1925 to 1965, to develop a domes- There may also still be a chance to be a player in the tic capacity to analyze the role of copper and to educate emerging lithium industry that is central to developing Chilean professionals and technicians in the management batteries for electric cars. In both cases, it may be neces- of the [large copper firms]." This is a striking statement sary to cultivate more of a knowledge cluster. periphery improved more than the center. Finally, for this ing" (cited in Lagos 1997), and Lagos concludes that a report, Cuddington, Ludema, and Jayasuriya (2001) use Dutch Disease scenario is not likely to occur in Chile. Gyl- modern time series econometric techniques and find no fason (1999) also suggests that though there is some evi- downward trend across the century before 1973 (see Box 3.6 dence of Dutch Disease emerging in Iceland and Norway, and Figure 3.6). Even if the post 1973 downward trend were the same is not true for Denmark, Finland, and Sweden. permanent, Meller argues that, measured in terms of the Given the strength in automobile, machinery, and telecom- quality of manufactures, purchasing power may have risen: munications exports, it seems difficult to argue that a high a Chilean can buy four times the number of computers in share of natural resource-initensive exports was an insuper- 2001 than he could in 1981. able obstacle in these latter cases. Second, there is no overwhelming evidence of Dutch Disease effects that would crowd out other desirable activ- Why Was Latin America's Experience ities. Competition for capital or labor or, alternatively, ap- So Disappointing? preciation of the exchange rate due to large inflows of Latin America clearly did not follow the lead of these coun- foreign capital, are probably not as severe as sometimes tries, arguably for two reasons: (a) the poor foundation for arguied. The literature is not convincing on this point, technological progress laid in the colonial period, and (b) the however. A 1996 study by Ilades-Georgetown University cultivation of inefficient manufacturing sectors to the detri- estimated that the impact of the copper mining boom ment of the resource sectors in the post-World War II period. amounted to about 2 to 4 percent of crowding out of in- vestments and some labor from other sectors. Overall, the Initial Conditions negligible crowding-out effects and increases in both ag- During the initial phase of export growth, the foundations gregate production and new resources show that "the cop- were not laid for dynamic innovative economies, as they per boom is closer to a full blessing than to a mixed bless- were elsewhere. Economic historian David Landes (1998) 61 FROM NATllRAl. RESOUARCES 10 TIlE KNOW,LFDGE ECONOMY TR kl)E ANI) JOB QUALITY BOX 3.6 While these more cultural themes emerge in mLich of Prebisch RedUx Are LACs Tens of Trade Decliningd the Latin American historical literature, we choose to foc tis on more structural conditions, in particular those that In a paper written for this report, Cuddington, impeded the generation of a dynamic, innovative society. Ludema, and Jayasuriya (2001) take a more rigorous Many of these are familiar: barriers to trade, imperfrct econometric approach. They find breaks in the under- property rights, deficient infrastructure, unsound public lying data-generation process in 1921 and 1974, and finance, and deficient market size. We focus particularly n,n until the latter date, they cannot reject a random walk issues of education, technical capacity, and other barriers ro with no trend in relative commodity prices. After 1974 innovation explicitly related to knowledge managemen-t they cannot reject a secular decline of roughly 1 percent and generation. per year (see Figure 3.6). The first were barriers to trade and explicit barriers to F)GURE 3.6 diversification. The Spanish and Portuguese both forbade Segmented Trends in Real Commnodity Prices, 1900-2000 trade with other nations and enforced trading monopohl,s (Baer 2001). The vast expansion of trade in Chile after [ 2.3 independence testifies to the lost dynamism of the previous 2.2 decades. Further, fear of competition for manufacturo.d 2.1 . ,l products led Portugal to order the destruction of all smel t- c a ^ ' S ! ^fi 2.0 t ing furnaces to make iron implements throughout the mid 0.10- 1.9 18th-century, thus short-circuiting a potentially important - t 0.50 SO- x1.S backvard linkage (Baer 1969). E we ooo h I s I A M W /iA1 - l .7 Second, Latin America's birth as independent states w.1s violent, and the transition to stable, consensual states difil- -005 cult. This implies that, while Australia, Canada, Scancli- -0.10 - 9 is - OLItlier' navia, the United Kingdom, and the United States were able __.15 _ _ ____l_____,_,_,to quickly establish stable, legitimate governments and 1900 '10 '20 '30 '40 '50 '60 '70 '80 '90 2000 could forge ahead in promoting-at the very least-stabil- Real Commodity_Prices - Structural Trend ity, education, infrastructure networks, and often direct nui- -Fitted with Two Breaks Residual Break turance for emerging industries, the principal job of Latin states in the 19th century was nation building. Mexico offiers a particularly striking example of the roots of the lost halJ century after independence (see Box 3.7). Mining, which suggests that the slow adoption of technological advances had powered Mexican development for most of the 18tIh in the southern European countries relative to their rapid century, fell 60 percent and remained depressed for decade-. adoption in Scandinavia was a legacy of a period of being Mines flooded, most expertisc fled, emerging downstream intellectually closed following the recon qulista in Spain. industries and transport networks stalled, and national J-f Numerous authors refer to a r'rntier mentality of living off nances were thrown into chaos (Cardenas [997). In Colom- the income from the new world mines that discouraged the bia, independence and then numerous civil wars extending emergence of entrepreneurship and the pursuit of scientific into the present constantly intcrrupted construction of rail- knowledge. Though trading mining riches for manufac- roads and the consolidation of the national economy or loc,1l tured goods is consistent with comparative advantage, Lan- finances. In few ex-colonies of the 19th century was there a des argues that it did not encourage a tinkering or entre- government capable of implementing a sustained and coher- preneurial culture at the core of the industrializing nations, ent vision of economic development. which is essential to developing new areas of comparative These setbacks were critical factors in the delay of advantage. These traits were inevitably communicated to a growth. Recent literature supports Coatsworth's early argu- greater or lesser extent to the Spanish colonies, along with ments that the lack of good roads and canals, and the delay the stifling institution of the Inquisition. in the construction of railroads, were two of the most impor- 62 IT S NOT JUST WHAT YOtl PRODI)UCE. BET H(OW LESSONS KROM f-OMPARATIVE HISTORY BOX 3.7 Canada and Mexico: So Far from Cod, So Close to a Huge Market Canada and Mexico both share a border with a vast mar- of Canada's important industries were tariff-jumping affil- ket and the principal source of innovation in the 20th iates of foreign firms as the key factor inhibiting growth. century. For both countries, this implied both depen- In fact, many successful countries-Australia, Canada, dency and opportunity. Yet Canada's per capita income Finland, and Norway-were dominated by large econo- has expanded roughly three times more than Mexico's mies in such a way that Latin America's dependency lit- since the 1870s, and its industries are thought to be as erature could apply there with force as well (Maloney efficient as those in the United States. Why would these 2001a). But as Cortes Conde (1985) notes, compared to outcomes be so different? Latin American countries, Canada took advantage of the Mexico's beginning was hampered by the factors that large market to the south, beginning with exports of news- would delay the development of much of Latin America. print, and was aggressive in adapting foreign technologies The first factor was a violent search for independence that (Wylie 1990). would cost it 50 years. Further, poor roads and few rail- Technological progress was hampered in Mexico by roads delayed integration of its market with the exterior, several factors. First, there were distributional issues. and weak educational and technological infrastructure Haber (1992), for example, suggests that the Flndidora de impeded increasing the human capital of its people. The Fierro y Acero de Monterrey adopted modern technologies, Canadian government early on made a commitment to but the plant never operated at more than 30 percent railroads and farming infrastructure, and to universal capacity due both to the thinness of the market and to, public education. perhaps, a workforce unsocialized as modern salaried Industrialization in Mexico in the late 19th century workers. Canada inherited the equal distribution of lands would be undertaken almost entirely by the resident and an educated populace that allowed the emergence of Americans, English, Germans, Spanish and, particularly, indigenous industry that would grow to form large, ver- the French (Hansen 1971), and would drive the growth tically integrated corporations. boom of the Porfiriato. Using machinery from the home- Second, due to concentration in the credit markets and land, the French started the textile industries in Veracruz other forms of economic and noneconomic pressure, new- and Puebla (Buffington and French 1999). Foreigners also corners to the market tended to avoid established indus- started Mexico's first iron and steel plant, the Fundidora de tries and hence did not put on the pressure to further Fierro y Acero de Monterrey, in 1903, that would build on innovate. Unlike in Brazil and the United States, most the region's ore deposits and anchor its industrial devel- investment capital came from a tight clique of national opment (Hansen 1971). Hansen argues that there were financiers who grew to control the most important man- entrepreneurial spillover effects that drew many Mexicans ufacturing companies. Haber's study of the paper manu- into the capitalist ranks. However, this would also lead to facturing industry in Mexico and the United States finds substantial foreign ownership of production. that both the San Rafael y Anexas paper company and the Canada also struggled with foreign control. The per- International Paper Company (IPC) came to control their centage of the value of production in minerals controlled respective markets by the 1890s through mergers and by U.S.-controlled and affiliated companies in 1932 consolidations to capture economies of speed, and cut off ranged from 39 percent in iron to 63 percent in nonfer- the supply of raw materials to would-be competitors. rous metals, including electrical apparatus (Wylie 1990). However, monopoly rents attracted competitors financed Canada maintained "defensive" tariffs since the turn of the by the stock and bond markets in rhe United States, and century, although they would never be of the magnitude drove IPC's market share from 64 percent in 1900 to 48 of those in Mexico. Dehem (1962) cited the fact that most percent by 1905. No such challenge would occur in Mex- (continnzes on nextpage) 6i FROM NATURAL RESOURCES TO THE KNOWLEDGE EC ONOMY TRADE AND JOB QUALITY BOX 3.2 Continued ico, and San Rafael would not lose its monopoly position pal talents lay in making deals to maintain their monop- until 1936, when it was nationalized by the state as a oly positions and in manipulating the economic apparatus strategic industry. of the state to provide them with protection from foreign Third, few captains of industry had any technical expe- and domestic competition" (Haber 1992:24). rience. Of the nine men on the board of directors of Fun- Finally, beginning at the end of the Porfiriato, Mexico didor Monterrey, only one had any relevant technical expe- increasingly insulated itself from foreign competition. rience. "These were not the tinkerers of the English and not until the mid-1980s did it embrace the degree of Industrial Revolution or the production-orientated engi- openness to U.S. technology, markets, and innovation neers of U.S. industry. They were financiers whose princi- that Canada did. tant factors in retarding Mexico's economic growth (Carde- country as responsible for the far greater success of the nas 1997; Haber 1997). Their eventual appearance inte- merino sheep industry in Australia compared to Argentina, grated the market, making it easier for local industry to introduced in the same year in both countries. reach scale economies, and also made possible the new ex- ports of copper, zinc, lead, cotton, coffee, tobacco, and sisal. Distribution Summerhill (1997) estimates that by 1900 the static social Another major difference with the successful natural savings in terms of greater specialization and freight costs resource exporters is found in the unequal distribution of was as much as one-quarter of Brazilian GDP, and 20 per- wealth, land, financial capital, and education. The Scandi- cent of Mexico's, rising to 40 percent by 1920. navian countries did not start with an egalitarian tabula Weak financial sectors nor only prohibited financing of rasa. In the 18th century, Danish land was in the hands ot'a railroads, but also, as Haber notes in the case of Mexico, few thousand families on large estates tilled by serfs, and prevented the entry of competing firms that might have only 23 percent of rural households owned land in Finland. stimulated innovation. The small markets, the politicized What laid the foundation for the Scandinavian transforn.i- nature of defending property rights and enforcing con- tion to modern wealthy societies was the agrarian reforms tracts, and the lack of appropriate legislation delayed the that ranged in timing from Denmark's precocious begin- emergence of a banking system in Mexico (Cardenas 1997), nings in 1788 to Norway's and Sweden's reforms in the and frequent government defaults on the debt impeded the 1850s and later Finland's in the 1920s, which creatcd development of equity markets. Haber attributes the vastly small- and medium-size privately owned farms or plots of more rapid expansion and technological sophistication of forestland (Blomstr6m and Meller 1991). In fact Blom- the Brazilian compared to the Mexican textile industry to strom and Kokko (2001) argue that "it is hardly possible to the presence of a more developed, less "insider"-based cap- overemphasize the importance of the improvement in agri- ital market in Brazil. cultural productivity for Swedish industrialization," which Weak public finances partially explain the lower com- facilitated transfer of labor and savings, and made possible mitment to both education and infrastructure in the post- exports that generated capital for investment in forestry and colonial states, extending into the 20th century. Cortes manufacturing, providing a market for local craft-based Conde (1985) argues that Argentina's loss of market in production. Similar importance has been attached to th-ie wheat had to do with a lack of commitment to infrastruc- homestead-style settlement of the plains of Canada and the ture, agricultural extension, and quality control that were United States for the emergence of local industries. A Latin critical in Canada and were recognized as such by Argen- American analogue can be found in the even landholdings tines at the time. Australians see similar policies in their of Antioquia, Colombia, that numerous authors find pro- IT S NOT JUST WHAT YOU PRODUCE, BUT HOW LESSONS FROM COMPARATIVE hIISTORY vided the initial demand that jump-started Medallin (Pala- led to dramatically smaller efforts toward universal education cios 1979; Urrutia 1979). than the successful natural resource exporters made. By Recent thinking suggests that Latin America's persistent 1870, more than 80 percent of the population age 10 or inequality may have had even further-reaching effects. In above in Canada and the United States was literate, over "Inequality, Institution, and Differential Paths of Growth three times the percentage in Argentina, Chile, Costa Rica, Among New World Economies," Engerman and Haber and Cuba, and four times the percentage in Brazil and Mex- (1997) argue that the period of sustained economic growth ico. By 1925, Argentina, Chile, Costa Rica, and Uruguay during the 18th and early 19th centuries that distinguished attained literacy rates of over 66 percent, while the literacy Canada and the United States from the other New World rates in Bolivia, Brazil, Colombia, Guatemala, Honduras, economies was fundamentally due to the patterns of settle- Mexico, Peru, and Venezuela hovered at 30 percent until ment and crops that led to a relatively unequal distribution much later (Mariscal and Sokoloff 2000) (see Figure 3.7). of income in the slower-growing areas. This concentration As Engerman and Haber (1997) note, this proves par- preserved the political influence of the advantaged elites and ticularly important given that early industrialization marginalized much of the population, thus losing their reflected the cumulative impact of incremenral advances potential contributions, particularly through lower access to made by individuals throughout the economy, rather than the franchise, restricted access to primary schooling and nat- being driven by progress in a single industry or the actions ural resources, use of financial institutions, and property of a narrow elite. As one manifestation critical to the devel- rights.' opment of innovation, they note that the greater equality in human capital accounted partially for the high rates of Education and Technical Capacity invention in the United States overall, but also that the This marginalization is particularly striking in education. more general concern with the-opportunities for extracting The Colonial period enforced a negative intellectual bias in the returns from invention contributed to a patent system many ways. Most countries had a local franchise of the Inqui- which was probably, at the time, the most favorable in the sition and, largely for reasons of political control, the icon of world to common people. This stands in stark contrast to intellectual discourse, the printing press, was banned in Mexico and Brazil, "where patents were restricted by costs Brazil until the head of the Portuguese empire was moved to and procedures to the wealthy or influential, and where the Rio deJaneiro in 1809 (Baer 2001).9 The same concerns with rights to organize corporations and financial institutions control, extreme inequality of income, weak public finance, were granted sparingly, largely to protect the value of and perhaps an intellectual commitment to a small state, all rights already held by powerful interests." FIGURE 3.7 Literacy Rates in Selected Countries (1870-1925) 9' - 6o - Argenina Bazil Chll Costa Rica Jam la Mexio Australia Canada Sweden Unitd Sowier Mara an oooff 2(0, and Merdih 1995 65 FRONI NATLUKAL RFS(ORFRCFS fl) TIIE KNOWLED(GE ECONOMY TRADF AND JOB QUALIIT In addition to the poor distribution of education, the forestry sectors and the Australian, Finnish, and North nature of education in Latin America was less technical than American mining sectors. Hence the kind of dynamic expan- that found elsewhere, and this may have been central to how sions of those sectors, and knowledge spillovers that occurred the resource sectors developed in Latin America. Spanish there, were less likely to appear. For example, despite a prim- higher education was largely religiously based and focused itive tradition of iron smelting dating from the mid-l(-th on law, philosophy, and theology, and this pattern was repli- century, Brazil's capacity was hampered by limited technical cated in the colonies. The Spanish enlightenment after 1750 expertise and, arguably as a result, low adoption of new tet-h- saw the establishment of groups of autonomous sociedades nologies that impeded competition with British goods'1" ecnnamicas that sought to diffuse technology from abroad and Baer (1969) sees as a critical event for the development of the establish libraries throughout the country, and of some steel industry the foundation in 1879 of the School of Min-es Royal Societies emphasizing applied science. These had at Ouro Preto, Minas Gerais that led to the establishment of spillovers for the colonies: in both Chile and Colombia spe- the first new blast furnace since the failures of the beginni ng cific royal initiatives gave the initial impetus to scientific of the century. Much like the Antioquia School of Minin-g, inquiry in the last decades of colonization (see Will 1957; graduates of the Escola de Engenharia do Exercito, established Safford 1976). But Spain itself lagged behind; it began in 1930, led the steel industry as it developed through the training engineers seriously only in the 185(Os, and by 1867 1960s. Again, this must be compared to Sweden, which was had only one Esc-icela tie Inigenieros Indastria/es, located in exporting engineers by the turn of the century. Barcelona (Riera i Tuebols 1993). When Latin Americans Second, the lack of technical capacity established a began to look for innovation, they followed the Spanish in dependency not only on foreign know-how, but on forei,gn looking to the British, the French, the Germans and, even- adaptations of know-how to local circumstances. Chileain tually, the Americans. This needs to be compared to Swe- historian Anibal Pinto Santa Cruz is very clear that one rea- dlen, which had technical schools in the 1820s, a modern son that Chileans left copper development to foreigners (and engineering industry by 1850, and was exporting engineers this may also be the case with nitrates earlier) was simply and innovations to the United States by 1900. By that year, that the incremental learning process that would have serious research in chemistry was being carried out at the enabled them to adapt or develop techniques to mine lower- University of Oslo that laid the foundation for the dominant grade ores never took place. Chile did not develop the cap;.:_c- fertilizer, electrochemical, and electrometallurgical indus- ity to even monitor the gran inineria until the 1950s. tries in Norway (Hveem 1991). In sum, a variety of structural problems, ranging from Latin America lagged behind Spain. Chile's first sociedad poor distribution of income, weak property rights, volatile eono;inia was established in 1813, and schools teaching public finances, political instability, and the very late engineering and the natural sciences were established in development of a structure that fomented the adoption of 1819, but even a century later they were not producing foreign technologies and stimulated domestic innovation serious numbers of engineers. The birth of Colombia's tech- impeded the full development of Latin America's econo- nical corps was also long and difficult, and highlights an mies and, in particular, the natural resource sectors in additional dynamic of development that has perhaps been which they were strongest. Some of these, particularly the overlooked. As Box 3.8 suggests, not only was it the recur- knowledge-related factors, may have been less crirical in ring violence and political instability that silenced promi- the 19th century when Latin America was still growing nent scientists and undermined fledgling universities, or extensively toward the "frontier." However, as the next sec- fiscal instability that made financing of the sciences unreli- tion argues, they may well have become so when growth able, but a lack of demand for engineers prevented the became more dependent on appropriating the quasi-rents career from being feasible. The building of railroads pro- of technological progress. vided a vital source of employment for the scientific profes- sion; however, frequent work stoppages due to war or pub- Disruptive Industrialization in the lic finance problems made such employment unreliable. Postwar Period But the lack of technical capacity arguably had two major The second major determinant of Latin America's poor per- impacts. First, it prevented the development of the kinds formance as a natural resource-rich region lay in the very of knowledge clusters that characterize the Scandinavian self-conscious and deliberate process of fomenting manu- 66 IT S NOT JUSI WHAT YOU PR(DUCE, BUT IIOW: LESSONS FROM COMPARATIVE HISTORY BOX 3.8 Obstacles to Technical Development In Nueva Granada Colombia offers insights into the difficulties of establish- ment at the School of Natural Science and Medicine ing the type of technical education central to the success ranged from 144 to 197, and at the School of Law enroll- of Australia, Scandinavia, and the United States. Safford ment never fell below 100. This may have been due to the (1976) documents the barriers due to political instability traditional biases in Spanish culture, but it was probably and the vagaries of public finances that plagued much of also due to the multiple civil wars that, among other the region. As SLLg"ivLit t . however, is the feedback from things, slowed railway construction and hampered general economic expansion to the growth and specialization of growth that would have provided employment. Further, the technical base. the Sociedad Colomibian de Ingenieros complained that changes An early and seminal vector of the Spanish Enlighten- in political parties and personalities so often changed the ment in Colombia, Jose Celestino Mutis arrived in 1761 focus of the school, from being the "brains of industry" to to serve as personal physician to the Viceroy. An accom- simply designing military ordinance, that the teaching of plished botanist, Mutis also taught mathematics and engineering had failed to develop permanence or solidity. astronomy from 1762 to 1776 at El Colegio del Rosario, Further hampered by the lack- of a permanent location, one of the two principal universities for the Colombian labs, or equipment, the school would finally close during elites. He ran afoul of the Inquisition over issues sur- the War of the Thousand Days (1899-1903). Nonetheless, rounding the teaching of the Copernican system, but later by the 1880s, Colombia would develop a class of engineers viceroys promoted his idea of a public university with up on the latest in Western engineering and capable of more science and 'useful knowledge" in the curriculum. some innovation. In 1783, they also financed astronomical, meteorological, During this period agriculture would gain more and geographical research that included the collection of attention. The first Sociedadde ,A1 I .!.,', fColombianos was potential export crops, indigo, and nutmeg. In 1805 founded in 1878, and the Trujillo government soon after Mutis turned over the Bogota astronomical observatory to embarked on a broad program of agricultural research another accomplished scientist, Francisco Jose de Caldas, and development, the establishment of experiment sta- who also took over the professorship of mathematics at El tions, and extensive data collection, all patterned on the Rosario upon Mutis' death. From 1808 to 1810 Caldas U.S. Department of Agriculture. Agricultural programs published articles of an emerging group of science enthu- existed at the National University since the 1 870s, and siasts in the Semanario del Nuevo Reino de Granada, a jour- Trujillo sought to extend education by offering subsidies nal dedicated to science and the promotion of material for agricultural instruction in primary and secondary progress. These beginnings of this fledgling scientific schools, and created the National Institute of Agricul- community floundered on the recurring obstacles of ture. Public finance problems would undermine the whole Colombian history. A shortage of immediately practical program, however. The Institute lacked a chemistry lab- employment possibilities kept attendance of Caldas's oratory, suffered from deficient professorial ability, and mathematics class extremely low. More tragically, most of lasred for only five years. Professional instruction in the contributors to El Semanario were also active in the agronomy was not attempted again for 20 years, and was independence movement and were shot in the Spanish not firmly established until 1916. During the fiscal emer- Pacification of 1815. gencies of the 1890s, subsidies to primary education were A series of transitory military schools from 1848 to suspended, plans to send samples of local plants to inter- 1885 contributed to the science focus of Colombian uni- national expositions abandoned, and funding of the versities and to developing a source of engineers, partially experimental stations cut. by FLIranrcc ine' a source of employment. The technically The legendary Antioquia School of Mining, modeled oriented National University was established in 1868. after the University of California, Berkeley, possessed in However, in the 1890s, the Faculty of Mathematics and Engineering had enrollments of 22 to 52, while enroll- (continues onext page) 67 FROM NAI ULRAL RESOURCES TO THiE KNOWLEDGE ECONOMY TRAIDF AND JOB QUAlITY BOX 3.8 Continued the 1880s a French chemist, a Belgian metallurgist, and variety of markets and specializations, and hence innova- an Antioqueflo mining engineer trained at Columbia, tion in any narrow field was impossible. but it was terminated in the civil war of 1895. When Technical schools and public works projects in the last revived after 1904, along with a mining degree it offered third of the 19th century generated a proficient technical separate programs in civil engineering, industrial chem- class in Colombia. However, the problems of political, istry, and electrical and mechanical engineering. Alumni economic, and fiscal instability endemic to much of Latin would play a critical role not only in both large rnining America delayed its emergence and made less secure its enterprises and railways, but also in the development of roots, necessarily reducing its contribution to the overall Medallin's renowned manufacturing sector in the early growth of the country. years of the 20th century. However, it was also the case that deficient demand would force engineers to serve a Sourme: Safford 1976. facturing in the mid-20th century that dealt a double blow tures that played to the strengths of their natural resources to possibilities for innovation and growth. Not only did it and can be seen as pursuing productivity growth in both, create a sector with limited potential for long-term pro- and they too are near the TFP growth frontier. As a dra- ductivity gains, but to finance it the traditional or poten- matic contrast, Blomstrbm and Meller (1991) argue tial sources of growth in the resource sector were under- When Latin America decided to force industrializa- mined."t In effect, it was Latin America's reluctance to play tion by import substitution, it was not an industrial- to its natural resource strengths, and its embarking on an ization based on the countries' endowments that was inefficient industrialization strategy, that led to low growth supported. While the Scandinavian countries slowly in both sectors where countries like Sweden and the United and gradually filled in the empty slots in their input- States saw rapid gains in both. output tables, the Latin American countries filled in The Siren's Ca/i of Indiistrialization 1 2 all the numbers at the same time; and even worse, Latin America saw import substitution of manufactured they tried to fill in the U.S. numbers! Suddenly there were several small Latin American economies with goods as an alternative to its traditional resource-intensive sectors. But, again, arguably, it is not just what is produced production structures similar to that of the United but how it is produced that is critical. In the long term, States. the sector must not only be a source of jobs, but also of Chile and Ecuador had as many car producers in the dynamic gains in productivity. The Argentine economist 1960s as the United States. Not only did Latin American Guido Di Tella (1985) sees the critical development junc- countries want to accelerate the growth process, but thcy ture to be whether countries moved beyond a state of chose paths unrelated to their comparative advantage. And exploiting the pure rents of a frontier or extraction of min- they did it behind high tariff walls that institutionalized the eral riches into exploiting the quasi-rents offered by inno- barriers to innovation discussed by Hirschman, Prescott, vation, or whether their entrepreneurs became hooked on Parente, and others. "collusive rents" offered by state-sanctioned or otherwise To varying degrees, many of the late-developing, imposed monopoly (see Box 3.9). As the most extreme case, resource-based economies faced similar junctures, and m we can think of the Asian NICS as examples of the former. fact Australian observers see far more similarities than dif- Whatever dynamism emerging in the economy came from ferences in the goals of mid-century technocrats of the two manufactures, as the TFP figures show, they were among regions. In Opening Late-Industrializing Economzies: Lessons from the greatest adopters of technology in the postwar period. Argentina and Australia, Wynia (1990) sees both regions as Scandinavia and the United States evolved industrial struc- attempting "more merciful and less costly industrial revolu- 6'3 IT'S NOT JUST WHAT YOtU PRODUCE, BUT HOW: LESSONS FROM COMPARATIVE HISTORY BOX 3.9 Where Do You Go at the End of the FrOW? Referring to the closing of the Argentine frontier, Di rents. To the extent that development was based on Tella (1985) argues: innovation, these countries were switching to an alternative and unlimited source of g row -h. To the This kind of area of new settlement was bound to seetsrtesfgrwthattr afr i a c a extent that it was based on collusion, it opened up see its rates of growth falter after initial coloniza- a limited, alternative path."> tion. Argentina behaved, to some extent, in this fairly predictable fashion. But the same was not The distinction between industrialization driven by true for the other countries. It must be acknowl- the appropriation of quasi-rents arising from adopting edged that the ability of the United States, Canada innovations abroad vetsus exploiting artificially created and Australia to continue a process of vigorous rents, is critical. Again, it is not that you have a manu- growth even at the end of the expansion of the fron- facturing sector, but the way in which you have a sector, tier has been a most extraordinary feat, and one that that defines whether the country has created a source of could not be taken for granted. At that point the growth, or a drag on growth. The Asian NICs could not successful cases were able to move to a quasi-rent afford the latter for long. Korea and Taiwan (China) built based stage-early for the most successful of all, the sunset clauses into their infant industry protection to United States, less so for Canada and Australia, and ensure competitiveness, and established export targets to rather later for Argentina; further development for ensure efficient scale and technology. Brazil's informatics the United States and Canada was more clearly industry not only became addicted to protection and based on innovation and less so in Australia. For proved undynamic, but it dragged downstream indus- Argentina it arose exclusively from collusive quasi- tries such as the automobile industry with it. tions, guided by sentiments of nationalism and the need to Latin America is thus not sui generis in adopting the reduce dependency on large economies. Governments in policies it did, and it should be seen as lying at the extreme both countries sought to do it by relying heavily on gov- end of a continuum that extends through Canada and Aus- ernment regulations and controls, contrived economic rents tralia to the United States and Sweden at the opposite and and applied the best minds to managing this 'rent seeking most successful end. Acknowledging the similarities is approach to economic modernization,' `3 vital since it prevents us from isolating the region as some But prolonged rent-seeking policies had similar effects sort of rare and unredeemable case operating under separate everywhere. Observers in Australia and Canada cite the sti- economic laws. Indeed, the persistent Australian interest in fling and dependency-increasing effects of protection on Argentina stems precisely from its perceived kinship and a their industrial sectors. Australia, for example, had an desire to avoid its fate. indigenous automobile industry of some promise, and BHP- type conglomerates with solid roots, but it also nurtured A Turning Away from Natural Resource Strengths import-substituting industries that were neither of efficient What this implies is that, far from being a dynamic growth scale nor appropriate, given comparative advantages. At the sector with strong incentives to raise TFP through tapping furthest extreme from this pattern is Sweden which, despite the growing world knowledge base, the manufacturing sec- equivalent levels of unemployment during the depression, tor requires transfers from the traditional sectors to be viable. maintained low tariffs and an aggressive outward orientation The U.S. and Swedish economies found organically emerg- throughout the postwar period, largely at the insistence of ing industrial sectors that added value to their abundant organized labor. The significance of this case is that it shows natural resources. Latin America was highly suspicious of that it was not preordained that resource-rich economies had its natural resources sectors, both because of the external to turn inward. dependency implied, and because of Prebisch's argument of 69 I ,R(M NATURAL RSI. RE SOJRES TI THE KNOWLEDC E ECONOMY TlZ AI)E AND JOLF QUALITY declining terms of trade. In the end, however, the hothouse Scandinavia had the extreme turn toward inward-looking industrial sectors were more dependent than ever on natural behavior, nor the dramatic macrodisequilibria that gave resource-based sectors that, at the same time, were being Latin America a special fame. The two phenomena are self- dramatically undermined. reinforcing: recurring balance-of-payment problems led to a Australian observers see the degree of this turning away greater distrust of the global market, to political instability, from the natural resource base as critical to explaining their to discouraged entrepreneurship, and to barriers to innova- country's relatively better performance compared to Argen- tion. Arguably, the measures to coax rapid manufacturing tina's. As Australia encouraged investment in petroleum and growth out of countries without a scientific tradition and refining and electrical equipment in the postwar period, it with substantial strLctural barriers to innovation led to initially neglected the rural sector, which expanded at only heavy rates of taxation on traditional sectors. In the end, half the rate of population growth. But this policy was re- manufacturing lacked long-term dynamism, and the nat- vcrsed in 1952 with the granting of investment subsidies, ural resource sector lost the potential it had. Sweden stands extension of credit, price stabilization programs, and expan- as the road not taken. The openness to foreign trade, the sion of research and extension programs. This led to a dou- insistence on technological upgrading by its labor union-s, bling of production over the next decade due to the "clover and playing to irs natural resource strengths led tO a gradoU- revolution," improved crop and wool yields, and the intro- ally accumulated capacity to innovate in both the natural duction of myxomytosis in 1950 as a way of eliminating the resource and manufacturing sectors. plague of rabbits. AtistralikL continued to stiffer from mild cycles of boom and balance-of-payments crises, and required Are the Successful Models Replicable? IMF assistance in 1952. But a rebirth of interest in tradi- The Death of Distance, Fragmentation, tional mining sectors in the 1960s led to increased and First-Mover Advantages clynamism in the resources sector that might in the near Can the experiences of Australia, Canada, and Scandinavia future lead Australia to have the fifth-highest per capita be replicated in the early 21st century? The answer is not income (Wright 2001). It also diverted attention away from obvioosly yes. In particular, transport costs offered a mt a- the inefficiencies of the import substitution industrialization sure of protection to fledgling downstream industries, ex ,;-n strategy. In a sense, the manufacturing sector continued to to industries with low tariffs, which have since fallen to a be a drag on the natural resource sector, but the increasing fraction of their torn-of-the-century levels. As Chapter 2 productivity and dynamism of the natLral resource sector discussed, Jones and Kierzkowski (1990) and Jones (2()000) was such that it was able to overcome it. show that decreasing transport costs can lead to the frag- During the same initial inward-looking period of the mentation of production structures that in the past wocld 1940s and 1 950s, Argentina inflicted deep damage on its have constituted what some term a "vertical cluster." They traditional leading sector, driving output growth to 0.2 per- argue that various prodIbIction blocks in a vertically integrated cent per year and Leaving the country perilously close to ceas- process are connected by srvliee links. As the scale of the ing to export foodstuffs. In Chile, continuing tensions with activity expands, the firm may find it profitable to out- foreign copper companies led to, among other things, a source some blocks to other countries in which differen(es steady decline in copper-generated foreign exchange reserves, in factor prices or productivities lower the marginal cost of the lifeline of the manufacturing sector. Massively negative production of the fragments. Transport, communication, rates of protection prevented the emergence of exports in and coordination costs between blocks are the cost otf such natural resource-based areas that, after liberalization, surged fragmentation, and these are precisely the costs that have as dominant sectors (see Box 3.10). The combination of inef- decreased so dramatically over time. ficient industrialization and the demise of its traditional Another way of phrasing this is that the increasingly export sectors left the countries exceptionally vulnerable to globalized world may be leading to fewer forward and the cycles of boom and bhst that characterized the region. backward linkages. The traditional view of primary exports In sum, while there is a continuum of inefficient inter- driving development in other sectors, originally formulated ventions in the manufacturing sector, to the detriment of in staples theory to explain the Canadian case, can be short- the resource-exporting sector, neither Australia, Canada, nor circuited. As Puga and Venables (1999) show, this is not IT S NOT JlST WllAT YO I PROI)UCF, BU1' HOW- LFSSO)NS lIROM COMPARATIVE II OTORY BOX 3.10 The Costs of Chilean Industrialization: Frustrated Development in Fishing, Fruit, and Forestry While mining remains Chile's primary export, the recent nity in 1954 entitled Negative Aspects of Economic Interven- success stories that resonate most with the United States tion: Failures of an Experiment, praised CORFO's irreplace- and the Scandinavian experience are in the fishing, able role in creating the electricity and fishing industries, forestry, and agriculture sectors. The Promethean efforts but derided the gross inefficiency of Huachipato and the of the Chilean Development Corporation (CORFO), National Petroleum company, and saw the capriciousness founded in 1939, were financed by a 15 percent tax on of exchange controls as the overriding disincentive to copper, and by the 1950s grew to control 30 percenr of needed foreign investment. The halving of export volume investment in plants and equipment. CORFO's efforts over the previous decade, the stagnation of agriculture, would lay the foundation for the dynamic export indus- and the frustration of Chile's tremendous potential in tries of the next half-century. To develop the fishing in- vegetable and fruit exports were laid at the feet of irra- dustry, CORFO contracted technical assistance missions, tional intervention in the price mechanisms and the per- established a marine biology station near Valparaiso in sistently overvalued exchange rate. 1945, granted sizable tax exemptions in 1952, and joined In the 1960s, recurrent balance-of-payments crises the army and the University of Chile in surveying rhe would lead the government of Eduardo Frei to seek coastal waters in 1954. It took the first inventories of for- to promote nontraditional and traditional exports. Yet, est stocks, and contracted the 1944 Haig technical assis- Chile's areas of natural comparative advantage were tance mission, which revealed "the indisputable truth stymied by the gross protection and inefficiencies that tliat an adequate management of our forests could be- were the logical culmination of a system of protection and come the basis for a . . . great industry of forest prod- incentives that had mutated into literally incomprehensi- acts." In 1953 it financed processing plants for cellulose ble degrees of distortion. A researcher at the Catholic and newsprinr. In the fruit industry CORFO financed University noted that in 1965, "the multiplicity of technical assistance missions, extended credit for cultiva- instruments used, and the frequency with which they tion and experimental plots, and invested in supporting were modified, had arrived at such extremes that it was infrastructure. In 1941 it financed efforts to promote ex- humanly impossible to have a clear vision of their final ports of wood products and wine. During the 1950s and impact by sector or for the economy as a whole" (Jean- early 1960s CORFO established an experimental fishing neret 1972). She found effective rates of protection ex- station in Arauco, financed construction of modern boats treme by global standards, ranging from -100 to 650 and dock facilities in Tarapaca and Valdivia, and founded compared to -50 to 500 for Brazil, -25 to 200 for fish canneries and fishmeal mills. The World Bank- Malaysia, and -17 to 106 for Norway. These heavy nega- financed Paper and Carton Manufacturing Company in tive rates of protection implied that 10 of 21 industries Bfo Bio stimulated paper- and cellulose-related forestry studied could export only at a loss, and that "some of activities after 1957. these sectors, principally wood, paper, paper products, CORFO may have been correct in boasting, on its fish and other minerals, would have become, perhaps, sig- 20th birthday, of Chilean history being divided into two nificant exporters. A contemporary observer, Marko eras-one before the construction of the Huachipato iron Mamalakis, also wondered at the inability of the agro- works near Concepci6n in 1947, and one after, which export industry to grow, given that export dernand for transformed the region into an important center of man- raw or processed Chilean fruit, seafood, oils, wine and so ufacturing in Latin America, hosting peripheral indus- forth [was] almost unlimited" (Mamalakis 1976). tries in steel wires, ferrous alloys, zinc and tin recovery That these disincentives to invest and innovate were plants, electrical equipment, and tools and machinery. critical is borne out by subsequent history. As is well But early on, local observers wondered at the costs. A compilation of seminars given in the business commu- (rutinmer on next Page 71 FROM NATURAL RESOURCES TO THE KNOWLEDGE ECONOMY TItADE AND JOB QUALITY BOX 3.1() Continued known, the history since 1975 was one of relentless pur- to 45 percent, and exports of fruit, fishmeal, lumber, and suit of integration with the world economy and a correc- wood furniture rose from 2 percent to peaks of 9 percent, tion of the distortions accumulated in the previous 8 percent, and 6 percent, respectively; paper and cellulose decades. In the next 20 years, noncopper exports in- rose from 4 percent to 11 percent; and chemicals rose creased by a factor of 10, essentially eliminating the tra- from 1 percent to 4 percent. ditional foreign exchange bottleneck to industry. The tra- ditional dependence on copper fell as its export share fell Soarc: Maloney 1997. problematic unless there are scale economies, market 1800 levels (O'Rourke and Williamson 1999), and as Fig- imperfections, or other externalities implicit in the down- ure 3.8. shows, the trend has continued dramatically since stream process that a country would benefit from were then, and even more dramatically in communications. lHis- these processes retained in-country. In particular, if indus- tory suggests that fragmentation was a fact before the turn try can benefit from agglomerating in one area, then as of the 20th century, and that it had the potential to break transaction costs fall we may see the "south" deindustrial- apart downstream linkages. As one example, Findlay and izing. As industry is less constrained by location near mar- Jones (2001) argue that, compared to Lancashire, United ket or inputs, it agglomerates in the "north." If it is true Kingdom, New England textile mills required both more that the greatest productivity gains will be realized in man- labor and higher raw material per unit of output. However, ufacturing, (we have argued that it probably is not), then lower prices of U.S.-produced cotton would have given the the traditional bifurcation of dynamic north and stagnant United States an advantage in the presence of barriers to south common in the dependency literature will emerge. trade in cotton. Unfortunately, the reverse was true. Cotton The phenomena of decreasing transport and communi- transport costs were very low, and as Bils (1 984) argues, the cation costs and fragmentation are historical facts. By American cotton textile industry would probably not hLve 1920, transport costs had already fallen to a quarter of their been viable without the protection it received. FIGURE , 8 Transportation and Communication Cost (1920-90) 100 9 0 - 780- 7() - _\ 41)-X 3(0 - 20 - - * () II I I '_ _ _ _ _ 192() 193() 1940 1950 196(0 197(0 1980 199() |.K~.. Ocean Freight 6 Airfreight _ _ _ _ Satellite Trans-Atlantic Phone Swirce: Branson. 72 IT'S NOT JtJST WHIAT YOU PRODUCE, BUT HOW LESSONS FROM COtMPARATIVE HISTORY The same is likely true in the mineral-based industries some manufactures from Japan until wages became too discussed earlier. High ore transport costs effectively gave high, and then they were relocated to China or Malaysia. downstream U.S. industries a price advantage over foreign This pattern was not observed in the early part of the processors. But these same industries also maintained a car- 20th century, and as the next chapter shows, it offers new tel prohibiting exports to ensure that this was the case. As possibilities for Latin America: plummeting Internet con- early as 1910, Australian minerals were being processed nectivity and communications costs make Jamaica a center across the world in Germany, and the coming of World for telemarketing and design consulting. Guadalajara and War I made it far easier for BHP's processing capacity in Costa Rica became important parts of cutting-edge infor- Newcastle to become established as a forward linkage in mation technology industries. that country. The potential efficiency gains implicit in the fragmenta- Pressures toward fragmentation of the processes under- tion process also offer an important caveat to the emerging lying "staples" models of development were thus strong a literature promoting "clusters," envisioned as sequences of full century ago, and are only stronger today. This can be related products. If the blocks were more efficiently broken seen, for example, in the few linkages to industry arising up, keeping them together may create an inefficient indus- from the expansion of the Chilean pulping capacity from try. As an example, BHP originally calculated that smelt- 1987 to 1992. The Nacimiento plant was built by a ing copper in Chile was not cost-effective, partly due to Finnish company specializing in pulp-mill engineering, higher energy costs, and now ships the ore elsewhere for Jaako Poyry, and the Mininco plant by the Canadian firm processing. There may be externalities emerging from keep- Simons and the Swedish firm IPK. The machinery was ing smelting near the point of extraction, but if not, the imported from Canada, Finland, Sweden, and the United country will lose overall. Focusing on production chains States. Saab and Volvo may have had their start in provid- where comparative advantage does not dictate can begin to ing transport to the forest industry, but Chile's path would look like inefficient industrial policy. be different. Arguably, a more efficient policy might use the income from present natural resources to create comparative advan- New Opportunities Appear tage in new areas, for instance, by creating a well-educated However, though certain paths are made difficult, as Chap- workforce, or investing in communication or physical ter 4 will show, the situation is not at all bleak. infrastructure. Chile may never produce forestry or mining First, as discussed above, natural-resource-based indus- equipment, but lower communication costs, a well-trained tries can be very knowledge intensive and have the poten- workforce, and a network of universities and research insti- tial for very high growth rates. Further, falling transport tutes could make it an exporter of mining and forestry costs have allowed the emergence of natural-resource- expertise, or computer software. intensive industries such as flowers in Colombia and Ecua- The next chapter examines in more detail the successful dor, specialty vegetables in Guatemala, and fresh fruits in development and necessary supporting conditions of indus- Chile that were not possible before. Far-flung locales can tries playing to the region's traditional strength in natural now take advantage of the expanding market in tourism. resources, and those in new areas. Second, this ability to fragment production allows global efficiency gains, but does not imply that all industries move 1 It may be argued that natural resource collateral may have per- to the industrialized world. Krugman (1979) and Venables mitted excessive indebtedness and ill use of readily available lending (2001) argue that as transport costs continue to fall, benefits funds. However, the Asian NICs had che same response to higher oil from agglomeration lessen. Puga and Venables (1999) show prices, and similar debt-to-GDP ratios. The big difference for LAC that as this happens, wage differentials between countries was that the more open and dynamic export economies of the Asian may cause many manufacturing or other fragments to again NICs were able to manage debt service in ways that Latin America be relocated to poorer countries. With the Asian develop- could not. 2. Parente and Prescott's (2000) simulations suggest that a TFP ment experience in mind, they envision a process where level one-third that of the United States can explain GDP differences intermediate processes move in a cascading fashion from one of 1:27, or roughly the difference between the incomes of the highest- country to another as wages rise. Hence, Korea took over and lowest-income countries in the world. Colombia, they simulate, 773 FROM NATURAL RE5OLIRCES TO T HE KNOLEDI)GE ECONOMY: TRADE AND JOB QUALITY his TFP levels of 64 percent of the United States, and 59 percent of 8. These argoments also apply to Brazil. As Baer (1969) notes, Paraguay (Dollar and Wolff 1997). education remained virtually nonexistent until 1776, and even atter 3. We use "innovation" not only to refer to the process of gener- that "the few schools that functioned had little impact on the cultural ating new knowledge, but to making the necessary adaptations to level of the population." externally developed techniques. 9. This section based on Maloney (2001b). 4. Hall and Jones (1999), after controlling for differing levels of 10. Baer (1969) reports that the techniques used at the end of the education, still find i, r-, levels of log TFP highly correlated with 19th century were primitive. Of the 30 ironworks in the headwalter log of output per worker with a correlation coefficient of .89. region of the Rio Doce in 1879, seven used Italian forge methods and 5. See Stern, Porter, and Furman (2000); Romer (1990); Nelson the rest used the old African cadinho technique. (1993); Wright (1999). All seek to explain the factors determining 11. This section based on Maloney (20)01). the flow of innovation in terms of the interrelationships of a variety 12. With apologies to Duncan and Fogarty (1984). of social institutions and actors. 13. "None of this is confined to Latin America. Rent-seeking emo- 6. Most recently Matusayama (2)00), Sachs and Warner (1995a, nomics is not derived from that region's patrimonial political tradi- 1997, 1999), Rodriguez and Sachs (1999), and Rodrik (1996, 1997) tions or Hispanic affection for corporatist ways of conducting poli- have argued that agriculture has few prospects, and as Wright (2001) tics. Rather, it was a strategy chosen by authorities in nations that notes, there is a bias toward seeing mineral sectors as pure extraction, were, at the time that uconomic modernization was acceleratd, with few possible gains from technology. already too activated socially and politically to permit less politically 7 . Looking across the 1 970s and 1 98(0s, Bernard and Jones (1 996) self-conscious approaches to economic renovation... . The Aluts- find TFP growth of 2.6 percent in agriculture compared to 1.2 per- tralians were not radically different from the Argentines in their cent in industry, and in only one of their 14 sample countries was approach to the protection of industry and labor ... . They wcre TFP growth in industry. Lewis, Martin, and Savage (1988) find pro- guided by sentiments of nationalism and nativism, stressing the ductivity growth higher in agriculture in the Australian economy nation's defense against competition from cheaper labor and/or more than for the rest, and Martin and Warr (1993) find similar evidence powerful foreign economies" (Wynia 1990: 187-8). for Indonesia. 7'i CHAPTER 4 Recent LAC Experiences: The Role of Knowledge and Institutions HIS CHAPTER FOCUSES ON COUNTRY EXPERIENCES, MOST OF THEM DEALING WITH SPE- cific sectors. The role of nontraditional endowments, such as geography, knowledge, institu- tions, ICT, and foreign direct investment (FDI), seems to be key for the performance of emerg- J f it ing economic activities in LAC. This is true for a variety of production processes, ranging from fr uits in Chile to computer electronics in Costa Rica and Mexico, and tourism in the Caribbean. Moreover, it is also clear that trade liberalization and regional integration also helped diversify the structure of exports in most countries. NAFTA was particularly important for Mexico, while Mercosur was important for Argentina. The EPZs in Central America and the Caribbean have helped these countries take advantage of their close proximity to the United States. It seems that such arrangements can provide an institutional safe haven in terms of relatively stable regulations for FDI, although corporate tax incentives need to be reviewed in the near future. In both Costa Rica and Mexico, however, human capital and FDI have jointly stimulated the emergence of knowledge-intensive manufacturing activities. Hence the role of knowledge and human capital cannot be overestimated in the promotion of dynamic FDI. The experience of Costa Rica's CINDE also shows that proactive FDI promotion policies undertaken jointly between the private and public sectors can yield handsome rewards for development. endowments discussed in the recent scientific literature In the future, ICT development might enhance the per- (see Chapter 2). The dynamic potential of combining formance of industries ranging from the maquila sector in national strengths, such as geography, abundant natural Mexico and Central America to the tourism business in the resources, natural beauty, and even cultural heritage, with Caribbean. other modern factors of production such as sound institu- In addition, support from the public sector for research tions, ICT, and knowledge is a good recipe for growth and and development was a key ingredient in Chile's agricul- development. tural success story, and in the recent commercial success of There is thus a clear similarity between the recent LAC Brazil's EMBRAER. The case of agriculture in Chile also experiences discussed in this chapter and the historical indicates that macroeconomic management was a key in- experience of the industrialized countries reviewed in gredient for the success of the sector after the mid-1980s. Chapter 3: in both instances the emergence of new indus- The case studies that follow provide specific examples tries depended on how countries played to their strengths, of how policies can affect economic structure with de- ranging from abundant natural resources to geographic sirable consequences. In the end, the path from the use factors, by progressively applying new technologies and of natural resources to the knowledge economy is aided knowledge to their production processes, while strengthen- by intelligent policies, combined with the types of new ing ties with the global economy. 75 FROM NATURAL RESOU RCtES TO THE KNOWLEDGE ECONOMY TR.ADE AND JOB QUALITY Chile's Agricultural Perfonnance: A quick glance at Table 4.1 reveals that Chile was the top The Case of Fresh Fruits Exports agricultural performer during 1980-99. Furthermore, in Economic reforms in Latin America have been followed the vast majority of cases, the performances of productivity by concerns regarding the impact of trade reforms on the and exports per worker were better in the period after the performance of the agricultural sector. These doubts en- reforms. Nevertheless, for individual countries, the diffi.r- compass several issues, ranging from the dynamism of ence in performances is statistically significant only for a the sector to its effects on poverty (see Foster and Valdes handful of cases, probably due to the limited number of 2001). observations in the different periods. Table 4.1 suggests that In this section, we focus mainly on the case of Chile, but with respect to efficiency, the reforms tended to benefit agri- we begin by analyzing the productive performance of the culture. More detailed empirical analyses of the impact of agricultural sector in LAC countries before and after the the economic reforms on the performance of the agriculture reforms. We investigate whether economic reforms brought sector is presented in Lederman and Soares (2001). In geu- improvements in the productivity and export performance eral, the evidence indicates that reforms tended to have an of agriculture. But even after the reforms, Chilean agricul- immediate effect of reducing the productivity growth in tural performance was still outstanding compared to the rest agriculture relative to Chile, and this reductioni was recov- of the region. We find that part of the explanation is the ered after approximately three-and-a-half years. This result long history of applying knowledge to the development of is illustrated in Figure 4.1. On average, at the time of the new exportable agricultural products. implementation of reforms, LAC countries were underpter- forming relative to Chile by about 3 percent. After the ini- Economic Reforms and the Performance of the Agricil- tial decline, the gap between LAC agricultural productivity tural Sector in LAC and Chile became less severe with time. Also, on average, Table 4.1 presents some basic information for several LAC reforms tended to have an immediate positive impact on the countries.1 It indicates the year in which economic reforms growth of exports, and this impact tended to be progres- were adopted in the different countries, according to the sively intensified over time. independent analysis of Sachs and Warner (1995a). The table also summarizes the behavior (average annual growth Econometric Results rates) of agricultural labor productivity and exports per The country-by-country analysis discussed in Lederman ar.d worker, before and after the reforms. Soares (2001) shows that the general pattern observed in the TABLE 4.1 Date of Economic Reform, Growth of Labor Productivity and Exports per Worker in Agriculture, Selected Latin American Countres, 1980-99 YEAR OF PREREFORM POSTREFORM COUNTRY REFORM PRODUCTIVITY EXPORT'S N PRODUCTIVITY EXPORTS N Argentina 1991 0.15 % -3.320% 11 0.84Yc% 0.015c 8 Bolivia 1985 -1.28%7 31.61% 5 -0.30% Il.143a 1-i Brazil 1991 -0.63% -8.67% 11 1.59% 1.3 6 S Chile 1976 - - 0 2.58%> 4.39% 19 Colombia 1986 -2.28% -5.15% 6 4).88% -7.90%7 13 Costa Rica 1986 -1.69G -4.71% 6 0.97% 0.64y, 1 Ecuador 1991 0.81%k -3.15% 11 -1.475c -1.05%7 8 El Salvador 1989 -4.37%, -19.53% 9 -1.48%a 10.75%y 10 Guatemala 1988 -1.87% -8.985c 8 -0.21761a 2.119%. 11 Honduras 1991 -0.45% -8.295'. 11 -2.73e% -9.72%. 8 Mexico 1986 -1.87% -1.535' 6 -2.00%ci 0.98% I' Peru 1990J -1. 1% -8.74 10( 2.16% 4.41% 9 Uruguay 1990 -1.67%, -1.40%' 10 1.68%7. -1.07% 9 Total -1.16% -8.20%. 123 0.16%' 0 47%7 h 143 a. Postreform average is significantly larger than prereform average at 10 percent. b. Postreform average is signficantly larger than prereform average at 5 percent. ,Nont: Variables are growth rates (difference in natural logarirhms) of value added m agricult:.re, forest. and fishlng per worker (labor force); and total exports (FOB) of agricultural products per worker (labor force), both in 1990 U.S. dollars. Souae. Reform dates from Sachs and Warner (1I995b). 76 RECENT LAC EXPERIENCES: TILE ROLE OF KNOWLEDGE AND INSTITUTIONS FIGURE 4.1 pects of the reform process. Perhaps more important for the After Refonns, LAC Countries Caught up with Chile: farm sector were the indirect effects of exchange rates and An Illustration of Econometric Results interest rates, two key prices to which the sector is partic- 0.0 - ularly sensitive. By now it is well recognized that the ex- -05 - change rate is the most important "price" affecting the -1.0 - agricultural economy (Valdes 1986). We now examine the 15 - /effects of reform on sector prices. Net Rates of Protection u 25 To examine the long-term trends in the evolution of incen- 3-0. tives in the farm sector, the most general measure is the . < -3.5 - value added of agricultural activities relative to nonagri- -4.( 4- cultural activities. Most trade and price policies are re- 4.5 - stricted to tradable goods in the case of agriculture, and most agricultural goods in LAC countries are tradables. 50 - 2 4 I 6 7 8 9 ] Box 4.1 reviews the algebra of a relevant indicator of rela- Years since Reforms tive price incentives-the net rate of protection (NRP). The rest of this section reviews the evidence concerning NRPs for various agricultural commodities in Chile. aggregate for productivity could also be seen in each indi- As the NRPs in Table 4.2 show, prior to 1974 Chile vidual country. Overall, the results suggest that, with respect fAvored inTe rp sh and graps an most favored export-oriented crops (apples and grapes) and most to efficiency, the reforms tended to bring long-term benefits import-competing commodities (beef, maize, milk, sugar to the agricultural sector. In this respect, concerns about the beets, and wheat). Immediately following 1974, the NRPs impacts of economic liberalization on the performance of for exportables fell to very low rates, and since the initiation agriculture seem to be misplaced. Yet Chile still outper- of the second phase of reforms in the early 1980s, NRPs formed the other LAC economies due to its long history of have effectively been zero. For import-competing crops the reforms, its adequate macroeconomic management, and its hisorial nvetmets n kowldgeandinnvaton n ari- story is notably different. The NRPs for milk and wheat were relatively high in the decade of the 1960s. Measured culture. In the following paragraphs we look more closely at theChieanexpriece,wit a pecal ocu ontherol of NRPs during the early 1970s are not very meaningful, given the Chilean experience, with a special focus on the role of tergm fpiecnrl,hprnlto,sotgs n the regime of price controls, hyperinflation, shortages, and macroeconomic management and knowledge creation.'' rampant black markets. NRPs were relatively high during 1984-89 due to the government's response to the strain on Reforms and Changes in Incentives in the traditional farmer sector arising from low international Chilean Agriculture prices and a strong appreciation of the currency between The trade reforms most immediately affected the incentives 1979 and 1982. Chile instituted price bands for wheat, facing producers through changes in the prices of tradable sugar, and oilseed, and minimum import prices for milk goods. 2 This yielded incentives to move resources from during that time. Although there was a currency deprecia- import-competing goods toward the export-oriented and tion in the late 1980s, protections remained. During the nontraded sectors. A central goal of reform in general, and 1990s, the currency again appreciated, and the already insti- especially with respect to agriculture, was to reduce the tuted price bands cushioned traditional producers. explicit and implicit anti-export bias that existed previ- Table 4.2 shows the decomposition of real domestic ously. Imported input prices also fell as tariffs decreased prices for selected Chilean agricultural products for the with liberalization, which was a significant element in period from 1960 to 1993 or 1995. For example, during determining the effects of reform for some countries. 1975-83 the average 0.18 percent decrease in the real price In addition to the changes in direct price incentives of milk is the result of a 2.39 percent decrease in the real brought about by freer trade, there were other price effects border price, a 46.47 percent increase in the real exchange induced by deregulation and privatization and other as- rate, and a 44.26 percent decrease in price supports.3 As the 77 FROM %AI U ORAL RESOURC,FS 10) THE KNOWLEDGE ECONO(fY TIR IE AND) JOB (QL ALlTY BOX 4.1 Effeelive and Net Rates of Prmtection: Some Algebra The algebraic expression that captures the returns to pri- adjusted for internal transport costs, quality difference, mary factors may be written as the value added of agri- and other factors, because traditionally, especially in agri- culture relative to an average value added in nonagricul- culture, there were several forms of intervention (nontar- ture (tradables and home goods): iff barriers) and not simply tariffs. With a simple tariff (export taxes) as the only instrument restricting impor- VAA = VAA tables (exportables), these measures should reflect the VA!V4 aVA~ +(1-a)VAH.. explicit tax charged. What these measures do not capture is the effect of Although this has been used occasionally for the case h * d of specific cornnodities (for example, Hurtado, Valdes, what is called indirect protection (Krueger, Schiff, and ofd spechifi commodities (ore exampnle rtado Valdes 1988), which includes possible misalignment of anid Muchnik 1v99), more commonly relative prices sub- the exchange rate and the effects of industrial protection stitute for value added measures: on domestic relative prices. The net and effective rates of PA PA _ PIpHA protection (positive or negative) can be adjusted accord- P1,, oP, +(I-a)P, aPIPR +(1-a) ingly. The total adjusted NRP (incorporating direct and indirect interventions) may be expressed for a specific The effects of trade and price policies are captured by agricultural good i: the effective and net rates of protection (ERP and NRP), which compare prevailing domestic price ratios with NRpJ I,'PHA those that would exist with free trade. The ERP is the Pi' PI. most relevant to capture the impact on incentives; the NRP is better suited to measure the effects on prices paid where the P* is the counterfactual border price adjusted by consumers. It should be noted that most studies on for exchange rate misalignments, and PN is the counter- agricultural protection measure ERP and NRP by direct factual price of nonagricultural goods adjusted for both comparisons between border and domestic prices, industrial protection and exchange rate misalignment. table demonstrates, there were increases in real domestic Research and Development in prices for fruits and wheat during the initial phase of Chilean Agriculture reforms, 1975 to 1983. While many reformers havc tendcd to retreat from deve.l- During the second phase of the reform, 1984 to 1989, opment-oriented institutions concerned with the farm sec:- real domestic prices declined for all products, except beef tor, it has been broadly accepted that the unsuccessful and apples.' The declines in real domestic prices for wheat, delivery of supportive services of such institutions can maize, and fruits, in spite of a steady depreciation of the impede the achievement of reforms. The ability of R&D real exchange rate, were due to a decline in border prices. services to contribute to productivity growth, especially in After 1990 there was a cumulative decline in all real the case of poor farmers, can ameliorate the negative andi domestic prices, except for grapes, due to an appreciation of enhance the positive effects of trade liberalization (Tabor the Chilean currency and, for four of the selected products, 1995). This in turn can make the implementation of agri - due to a decrease in real border prices. To the extent that cultural reforms more palatable. In addition, strengthening the changes in the real domestic prices of these selected the growth of the agricultural sector through technical products crudely approximate the changes in the returns to change and productivity growth can improve overall farming, one can say that in the case of Chile the main employment, income growth, and food price levels, thus forces behind agriculture's price incentives were beyond the easing the costs of adjustment more generally. In particu- control of policy. The main factors were the border prices. lar, strerngthening export agriculture can directly earn for- 78 RECENT LAC EXPERIENE ES TIlE ROLE OF KNO(WLEDGE AND IN>TITUTIONS TABLE 4.2 Decomposition of Producer Price Changes in Chilean Agricuture CUMULATIVE PERCENTAGE CHANGE IN REAL DOMESTIC REAL BORDER REAL EXCHANGE (I PLUS THE) PRODUCT PERIOD PRICES PRICES RATE TARIFF RATE OTHERS Exportables Apples, Red 196(1-70 1 *3.77 4t.1)2 3(0.94 17.1 i 44.67 1971-74 -2A -21.22 44.77 -10.38 -15.58 1975-83 6.4 -8.4 46.47 -6.77 -24.9 1984-89 30.48 -4.1 36.11 0 -1.53 1990-93 -429.1 -64.88 -14.79 0 -349.43 Grapes, Thompson 1960-70 99.29 21.53 3)0.94 17.14 29.67 1971-74 54. 5 4.37 44.77 -0.59 5.8 1975-83 41.51 2(0.91 46.47 16.55 -9.31 1984-89 -4.7 -3(1.47 33.67 0 -7.91 1990 -94 22.24 21.37 -16.8 0 17.67 Importables Beef 196(0-70) 412 9 57 3(o.94 ) (1.69 1971-74 87.27 42.14 .i-177 0 0.36 1975 83 51.72 -76.v6 46.47 0 -22.12 1984-89 136.69 64.93 33.67 13.98 24.12 1990-95 30).05 16.88 26.84 -3.54 -16.54 Maize 1960-70 n.a. n.a. n.a. n.a. n.a. 1971-74 n.a. n.a. n.a. n.a. n.a. 1975-83 n.a. n.a. n.a. n.a. n.a. 1984-89 -4.24 -45.12 25.27 -7.84 23.46 199)-9 3 -22.18 2 83 1 ]6.25 -3.54 21.44 Milk 1960-70 3.29 -5.37 30.94 0 -22.28 1971-74 77.44 14.27 44.77 I) 18.41 1975-83 I0.18 -2.39 46.47 0 - 4.26 1984-89 -6.25 9.46 33.67 13.9S -63.35 1990-9- -16.51 -19.66 16.25 -3.54 22.94 Sugarbeers 196()-70) n.a. n.a. n.a. n.a. n.a. 1971-74 n.a. n.a. n.a. n.a. n.a. 1975-83 n.a. n.a. n.a. n.a. n.a. 1984-89 -1.35 39.16 215.27 -7.84 -57.93 1990)-93 -7.57 18.72 16.25 -3.54 -6.5 Wheat 1960-70 6.33 -13.15 26.76 0 -7.28 1971-74 53.81 80.13 9.89 0 -36.21 1975-83 35.51 -88.39 81.35 0 42.55 1984-89 -1.42 -32.83 37.84 -7.84 2.41 1990-95 -27.8 -1t284 -26.84 7.84 4.04 n.a. Not apphcrbl. Sausrc: Foster and Va)des 20It1. eign exchange and indirectly finance imports. Yet, to some In the case of Chile, the role of R&D in agricultural per- extent, the private sector usually replaces public support by formance is most clearly observed in the fruit sector. This opening new avenues for the agricultural sector to achieve experience is narrated in Box 4.2. The rest of this section technological and managerial advances (Umali 1991). In reviews the role of R&D policies more generally. terms of who its "clients" are, however, private support is Prior to 1973, agricultural research in Chile was con- more likely to be oriented toward larger, commercial farm- centrated in the Agriculture Ministry's parastatal Instituto ers than the public sector, especially toward those in Nacional de Investigacion Agraria (INIA). The agency was growth sectors, which are often the export sectors in the responsible for all scientific specialties, crops, and regions. case of LAC. In terms of both funding and personnel, other institutions, 79 FROM NATURAL RESOURC ES TO THE KNOWLEDGE ECONOMY TRADE AND JOB QUALITY BOX 4.2 R&D Pelicies and the Emegec of the Fnmt ecW In Cie Agitculhure The most dramatic story occurs in the fruit sector where transfer, identified and began to plant new varieties suit- exports grew at a rate of 20 percent annually in the first able for foreign markets, improved orchard and posthar- 20 years since the reforms of 1974. Areas planted to com- vest manzagemncn, upgraded fruit research and teaching, mercial orchards almost tripled and fruit production and developed the infrastructure necessary to export fruit quadrupled. Jarvis (1992., attributes this success to the to foreign mnarkets. Jarvis notes that the bulk of these speed with which Chileans were able to :ransrcr. adapt, developments were carried out by the public sector. and extend fruit technologies i nr l ly developed for Cal- Exports rose slowly across this period and several export ifornia and other fruit-growing regions to Chile He companies emerged that gained experience with foreign argues that private initiative in these areas was driven by markets. Chilean firms achieved the volume needed to changes in price relationships and industry structure that charter special fruit cargo ships. Fruit handling, better increased returns to private R&D. The Corporaci6n de cold-storage managemenr. and reduced transit times al- Fomento (CORPO) played an important role in the early lowed improvements in fruit quality at destination. 1960s in surveying ecxisting fruit orchards, analysis of Significant barriers rermained. Uncertainties surround- potential demand in foreign markets, elaboration of pro- ing land reform and xnacropolicy of the late 1960s and duction goals, introduction and screening of new vari- early 19'(Js deterred private investment. Import quotas eties, establishment of nurseries to propagate disease-free and high tariffs, slow and inefficient transport and port plants, construction of cold-storage facilities at srrategic handling, and bureaucratic red tape slowed progress. locations to promote postharvest care, ph;rosarniiarN With the policy reforms of the early 1970s, quota restric- inspection of exported fruit, establishment of favorable tions on imported inputs were reduced, and in 1976 credit lines and working capital, and "drawback" pay- import tariffs were cut to a uniform 10 percent from the ments for fruit exports. In 1965 a 10-year program of previously high average level of 96 percent. Export proce- cooperation between the University of California and the dures were streamlined and labor unions were proscribed. University of Chile was established to permit technical Strong world prices for fruit and a competitive peso raised cooperation and improve graduate training. This helped returns to capital in the range of 25 to 50 percent. the University of Chile develop a first-rate faculty in Chile improved technology at all levels of the produc- fruit-related sciences and to begin modern fruit research. tion chain-domestic transport, port operation, intema- Spillover effects strengthened government agencies and tional shippings baing, and telecommunications-and other universities. As one crude measure, Jarvis docu- in all aspects of fruit production, packing, and cold stor- ments that the number of theses on fruit issues submitted age. Jarvis examines in detail particular gains in planted for the Agricultural Engineering degree increased by 2.5 varieties, mdnagement, and transport, and the private percent and as a share from 13 to 31 percent from channels of technolopg dissemination. The number of 1_9'6-so to 1986-199(. In 1 96- Chile established the fruit entrepreneurs increased four-fold. Jarvis further con- National Institute of Agr ultrural Research (INIA), cludes that there were s[rong spillovers to other sectors which paid relatively high salaries and attracted skilled that saw the possibili tnus of exporting. Most of the inno- researchers. The agency initiated a fruit research program vation was carried out by the private sector, although from the start. By these mneans, Chile developed the sc- Jarvis wonders about the need for public provision as entific personnel and knowledge to achLieve rechnolog cal fruit markers have become tighter. including universities, contributed a small proportion to (Venezian and Muchnik 1994). With the introduction of the overall system of agricultural research. Private sector reforms, while direct public funds remained at a level of efforts were relatively slight; research investments in 1973 less than a half percent of agricultural GDP, the develop- represented only 2 percent of the annual budget of INIA ment of other funding sources led to at least a doubling of 80 RECENT LAC EXPERIENCES THE ROLE OF KNOWLEDGE AND INSTITUTIONS total support for agricultural research-from 0.4 percent agriculture. While the previous emphasis on traditional of sector GDP to 0.9 percent between 1973 and 1992. crops by INIA was continued by that institution, the pri- In the spirit of the economic reforms after 1975, several vate sector and universities adapted quickly to the new institutions were created to promote private sector partici- incentives generated by reforms. Magnifying the effect of pation and competition in research and development. The domestic research activities related to commercial agricul- operation of these institutions required the use of either ture generally, and to export agriculture especially, was the collaborative funding or research (or both) originating in easy availability of proven crops and technologies from the private sector. Moreover, the support of these institu- abroad, most notably from California. For-profit research tions was for the most part open to rivalry across regions activities proliferated, especially where the returns to the and between researchers in any economic sector. Except for identification and adaptation of new varieties and methods the Ministry of Agriculture's research fund (Fondo de In- were more easily internalized. vestigaci6n Agrfcola, FIA), research directed toward the The human capital and experience officially located in farm sector had to compete for resources available to all INIA, but put to use by private enterprises, were important types of investigations, elements in the initial identification and evaluation of vari- A national fund for science and technology (Fondo eties and technologies accessible from abroad. As R&D in Nacional de Desarrollo Tecnologico y Productivo, the service of private interests gained in stature and funding, FONDECYT) was established in 1982, and the national INIA expanded its area of responsibility and replaced in part development corporation created a research fund previous government services that had been unable to sur- (FONTEC) in 1984. More important for university-based vive in the political environment following the initiation of research, a 1989 law introduced tax incentives for research reforms. Moreover, the political adherence to market-based donations to institutions of higher education, and 1992 solutions, and the obvious successes of commercial farmers, saw the implementation of the development fund Fondo de exporters, and processors, called into question public sup- Fomento al Desarrollo Cientffico y Tecnologico (FONDEF), port for activities that otherwise might be sustained by pri- in support of R&D, which was underwritten with an Inter- vate interests. As a result, INIA's role evolved toward a focus American Development Bank loan. By 1990 private expen- on smaller-scale agriculture. Nevertheless, one consequential diture on agricultural research had increased 19 times its benefit to all agricultural interests that followed the institu- 1973 level, and represented approximately 13 percent of tion's adoption of a development role was the enhanced coor- total spending on research in the sector. Within a few years dination of research with the demands of farmers. following the introduction of FONDEF, private spending The Chilean experience related to the contributions of nearly doubled to about 20 percent of total farm sector agricultural research and development reinforces the research expenditures. The greater diversification of insti- importance, within the context of overall reform, of dereg- tutions and sources of financial support for farm sector ularion and privatization in the provision of more-reliable R&D was accompanied by a shift in the nature of funding and lower-cost services. The climate that permitted these for the largest research institution, INIA. Prior to the contributions, however, has also tended to de-emphasize reforms, INIA relied on taxpayer monies for 90 percent of longer-term research questions and the institutional infra- its budget, the remainder coming from sales of services and structure that supports scientific research, development, seeds and other farm products. Following 1975, govern- and training leading to benefits not easily internalized by ment policy pushed INIA toward self-financing. By 1985 private concerns. The net cost or benefit of such a shift in the institution was earning 40 percent of its income from the balance of agricultural research and development policy sales and another 20 percent from grants, loans, and other away from long-term issues is not obvious.5 It is also worth nongovernment sources. noting that the successes of the Chilean system in respond- The increased availability, diversification, and private ing to the incentives brought about by the reforms were direction of funding sources was associated with a greater made possible not merely by the similarity of Chile's nat- weight giveln to research on exportable crops and the inves- ural resources to those of regions in more economically tigation of postproduction technologies, product character- advanced countries. Regardless of its deficiencies, the pre- istics and quality, and other topics important to commercial reform agricultural research system had sustained a pool of 81 UROM1 NAT[IRAL RESOUR( FS TO THE KNOWLEDGE ECONOMY TRAIDE ANI) JOtB QIPALI1TY human capital that, however inefficiently employed, was By 1996, the member countries were, on average, pretty available for use when the reforms shifted the emphasis close to the liberalization objectives. While the external tar- onto private efforts. Box 4.3 discusses more general issues iff in Argentina and Brazil converged from above toward the concerning the design of R&D incentive programs. CET, those of Paraguay and Uruguay approached the CET from below, thus reflecting the relatively lower level of pro- Argentina's and Uruguay's Export Diversification tection that those countries had initially, compared to their after Liberalization big neighbors. Regarding the tariff levels observed for the exempted items, the small countries set very high tariffs tor A UJnilateral and Regional Liberalization in Argentina the excluded items. On the other hand, the big countries, in a ad Urulguiay particular, Brazil, have high rariffs on rhe external excludcd Argentina's and Uruguay's trade liberalization was accom- iterns (21. 39 percent). This fact could potentially play an plished by policies applied unilaterally and regionally, and important role in encouraging exports from Argentina and within the multilateral negotiations under the auspices of Uruguay to the Brazilian market. This hypothesis is ex- the General Agreement on Tariff and Trade/World Trade plored fLrther below. Organization (GATT/WTO)." The process of trade liberal- ization in Argentina started as a unilateral policy in 1988, Diversification Indicators with the so-called Canitrot Reform. At the beginning of In this section we investigate the precise nature of the 1991, trade liberalization was pushed even further; the changes outlined above. Particularly, we focus on the average tariff fell to an unprecedented level of 1 2.2 percent degree of export diversification. Following Sapir (1996) we in mid-1991. Overall, this unilateral liberalization reduced construct the Gini coefficient and the Theil Coefficient to the average tariff in Argentina from 45 percent in 1987 to measure the extent of diversification of exports and imports arotund 13 percent in 1994 (Berlinski 1998). in both countries. The Gini Coefficient ranges between 0 The process of trade liberalization in Uruguay also and 1. For example, a high Gini Coefficient is associatt'd rccived a strong push in the early 1990s, but trade liber- in the income distribution literature with high-incorm-e alization had been pursued without major inrerrtiptions inequality. In our context, it will be an indicator ofa highly since the end of the 1970s. The tariff reduction in Uruguay concentrated trade structure. Different indicators weight was accelerated after 1990. The average nominal protection differently changes in the distribution of export and import declined from 30 percent in 1990 to 17 percent in 1993, shares, so it is a good practice to check the robustness of and reached 13 percent in 1995. results to different indicators.8 Table 4.3 presents the Gini From 1995 onward, trade policies in Argentina and and Theil Coefficients for Argentina and Uruguay. tTruLguay were set within Mfercosur. This integration initia- Several observations can be made. First, the evolution of tive was established among Argentina, Brazil, Paraguay, diversification does not depend on the inclicator we uSe." and UJruguay in 1991 with the signing of the Asunci6n Second, as one would have expected for a small country lil:e Treaty. In its first article this treaty states that the aim of Argentina, exports are more concentrated than imports the agreement is to achieve "the free circulation of goods, throughout the period. Third, imports and exports show a services ancd productive factors among the member coun- consistent decline toward greater diversification. Hoxvevcr, tries, through the elimination of the tariff and nontariff export concentration declined steadily and continuously restrictions to the circulation of merchandises and of any while the import concentration indexes show a structural other equivalent measure." It also established a Common change around 1991 and 1992, when trade liberalization External Tariff (CET) and a common commercial policy was implemented. toxward the rest of the world. The full implementation of For Uruguay, exports are also more concentrated relatixe free trade within the region and the establishment of the to imports, and both indicators decline in the period CET were scheduled for 1995. These objectives were par- (greater diversification). Again, as in the case of Argentina, tiallv achieved with significant exemptions (for details, see the concentration of imports declined sharply at the begin- BoLizas 1996; Terra 2(0;0 and Sanguinetti, Pantano, and ning of the 1990s, reflecting the deepening of trade liber- Posadas 2001). alization policies taking place in those years. Indeed, this RECENT LA(< EXPERIEN(ES TIHE ROLE OF KNOWULDDGE AND INSTITUf lONS bOX 4.5 Incentives for R&D: Some Policy Issues R&D has long been recognized as an important factor two systems. Table 4.4 summarizes some of the issues to determining the pace of economic development. They be considered when discussing incentives for R&D. reduce production costs of existing products and create The cost-sharing scheme allows the government to new ones, thus contributing to sustained growth and partially control the composition of investments in R&D, enhancing economic welfare. Recent microevidence for which in principle can be used to enhance welfare. It also Ireland also shows that, at the plant level, R&D-active assigns the monitoring of programs to special public firms are usually associated with plants with higher sur- institutions (such as development agencies), which are vival rates and higher job quality (Kearns and Ruane possibly more capable of checking the appropriate use of 2001). But there are externalities and public good fea- resources than a tax agency restricted to looking at the tures of investments in R&D that tend to make decen- accounting statements, as in the tax credit system. But tralized solutions-market outcomes-inefficient. These the cost-sharing mechanism also has its drawbacks. In are related to the generation of knowledge that can he principle, it is questionable whether the public sector has publicly used, and to externalities typical of human cap- a better ability than the private sector to decide what are ital investments. For these reasons, governments usually the investments with higher social return. The decision use incentive devices to increase R&D investments and process inside the government is certainly affected by to try to attain a socially optimal allocation of funds for rent-seeking and political lobbying, thus increasing the R&D. uncertainty in relation to the optimality of the outcome. The most common incentive takes the form of tax Also, although development agencies already exist in credits for investment in R&D. The rather scarce evi- most countries, a system of public subsidies would possi- dence suggests that these incentives increase investments bly overload the existing institutions and increase opera- in research, and that each additional dollar in tax credit tional costs. All these factors have to be weighed against for R&D stimulates roughly one additional dollar of the problems of the tax credit system. These include the R&D (Hall and Van Reenen 2000). The impact is rela- fact that, in this case, the whole decision process remains tively small, and does not seem to have a multiplier effect delegated to the private sector. The effectiveness of tax on private investment decisions. Considering that firms credits is limited by the possibility of its use for tax- usually adapt their accounting practices to tax laws and evasion purposes and the increased complexity of the re- incentives, this figure implies that each dollar given away sulting tax structure. through tax exemptions is probably associated with less Therefore, the choice between the two systems, or the than one additional dollar really spent on R&D. The nat- ideal design of any of them, is far from clear. Which of ural question, therefore, is whether these resources would the factors mentioned above is the most important one be better allocated if directly managed and invested by probably depends on the specific productive and political the government in specific R&D programs. structure of each country. In this sense, there seems to be One altertnative is a subsidy in the form of cost shar- space for experimencation, using pilot programs and the ing of R&D investments between the private and public institutional structures already in place in the different sectors. In this case, the public sector would add a de- countries. This would reveal the differential impacts of fined amount for each dollar invested by private firms. In the alternative structures in different environments, giv- terms of instruments, and of the impacts on government ing a more solid guide for policy decisions. In any case, accounts and R&D expenditures, a scheme like this the issues metntioned here should be kept in mind, would probably be equivalent ro a tax credit system, since because they are factors that will determine the costs and the tax incentive can be set at any percentage of the total benefits of R&D incentive programs, including their fis- R&D investment by private firrns. Hence it is always pos- cal impact, the response of private R&D expenditures, sible to reproduce the fiscal and R&D impacts of a given the composition of R&D expenditures across sectors, the cost-sharing system with some specific tax credit struc- monitoring mechanisms, and the potential political and ture. But there are operational differences between these institutional constraints. FROJM NATUTRAL RESOURCES TO THE KNOWLEDGE ECONOMY TRADE AND JOB QUALITY TABLE -13 , Indicators of Trade Concentration, Argentina and Uruguay, 1986-99 ARGENTINA URUGUAY IMfPORTS EXPORTS IMPORTS EXPORTS GINI THEIL GINI THEIL GINI THEIL GINI THIIL 1986 0.76 1.2 0.89 2.0 0.76 1.4 0.92 2 i [992 o.64 0.8 0.86 1.7 0.72 1.1 0.9( 2 l 1 999 (1.64 ().8 0.83 1.A 0.65 0.8 (1.85 I i S,,rUe: Sangoinetci. Pantano. and Posadas 20U1. increase in the varieties of imports was great news for con- diversification, and this proportion remained stable during sumer welfare in both countries. the entire period. On the import side (not shown), we observe an increase in the participation of between-group Decomposition of Concentration by Sectors and Regions diversification since 1991 (around 40 percent), suggesting a In this subsection we present a more detailed analysis of the significant effect of trade liberalization on the composition driving forces behind the evolution of concentration of of imports at a very aggregate level of products. exports. The analysis first uses variants of the Theil Coeffi- Figure 4.3 shows the calculations for Uruguay. The cient to determine the extent to which trade diversification results are similar to those obtained for Argentina in the in Argentina and Uruguay occurred either u'ithin product sense that most of the diversification is explained by groups or between product groups. In turn, the analysis within-group diversification. The main difference is that explores the impact of Mercosur on trade concentration in we find a stable behavior of the participation of within- Argentina. group (and between-group) diversification during the e-n- tire period for both exports and imports (not shown). Decomposition of Export Concentration by Sectors The Theil decomposition analysis allows us to check Diversification Indicators by Region: Mercosur whether diversification occurred either within or between and the Rest of the World product aggregates. Figure 4.2 presents the results for A final important aspect of the behavior of the diversification Argentina."' The main conclusion is that a large share of the indicators is to check whether diversification was different diversification of exports of Argentina is accounted for by across trade flows within Mercosur and with the rest of the within-group diversification. Almost 80 percent of the de- world. If this is so, part of the overall diversification proce ss cline in the Theil Coefficient is explained by within-group could be attributed to the increase in regional trade. Figure 4.4 shows Theil concentration indexes for Argentina's exports. TABLE 1A Exports to Mercosur are less concentrated than those to Issues in R&D Incentive Policies the rest of the world. On the import side (not shown), imports coming from the rest of the world are more diver- TAX COST- sified. Regarding the dynamics of the concentration indica- ISSUES CREDII SHARING SUBSIDY tors, the export indicator for Mercosur behaved more errat- ically compared to the export indicator corresponding to t-he 2. R&D expenditures Possibly tquivalent Possibly equivalent rest of the world. More important, export diversification 3. Composition of 1()1) percent privately Shaeed decision in~~estnn,nt determined ~~within Mercosur Increased over time, though some diverii- investment deteemined 1. Monitoring and Tax collection agency Special public fication occurred even before 1991. This may reflect that on administrative institutions the export side considerable mark-et access was obtained responsibilitv prior to 1991 through the partial agreements signed among 5. Challenges Tax evasion, complex Renr-seeking for Argentina, Brazil, and Uruguay. On the other hand, import tax structure subsidies, public sec- concentration declined significantly after 1991, showing the toe's decisionmaking ability, institutional effect of across-the-board regional (and unilateral) trade lib- overload eralization that has taken place since that year. 84 RECENT LAC EXPERIENCES: THE ROLE OF KNOWLEDGE AND INSTITLITIONS FIGURE 4 2 Concentration Index Decomposition by Main Aggregates, Argentina Exports, SITC * Theil within Group E Theil between Group 8(1 60 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 Soaurrc: Saoiguiiietti, Pantano, a[Id Posadas 2001. FIGURE 4. i Concentration Index Decomposition by Main Aggregates, Uruguay Exports, SITC Theil within Group 17 Theil between Group ()f~~ I I I I III I 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 Sowire: Sanguinetti, Pantano, and Posadas 2001. The evidence for Uruguay (see Sanguinetti, Pantano, where those geared to Mercosur are less concentrated than and Posadas 2001) shows that on the export side we also those going to third markets. The remaining issue is how observe a greater level of diversification of those going to Mercosur aided trade diversification. Mercosur. The differences in the level of the concentration indicators across destinations seem to be less significant Regional Integration, Economies of Scale, (and have declined over time) than those found for and Transport Costs Argentina. Still, when we calculate the Herfindahl concen- In relatively small countries it will nor be profitable to pro- tration index (see Figure 4.5) the data show a significant duce goods subject to large economies of scale. This could and relatively stable difference in export concentration be because of lack of inputs or because the local market is 85 FROM NATURAL RESOUIRCES TO THE KNOWLEDGE ECONOMY TRADE AND JOB QLALITY FfG[JRE 4.4 Theil Concentration Indexes for Exports; Argentina, Rest of the World, and Mercosur 2.2 - _ Mercosur _S Rest of World 2.0 - \_ 1.8 - 1. 6 1.4 - 1.2 1.() - l l l l l l l l l l 1 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 Sw/ne: Sanguinetti, Panrano, and Posadas 2001. F[GURE 4 5 Herfindahl Index for Exports, Umguay (1.16 - 0.1 I ().12- 0.1(1- 0.(8 - 0.(16- 0.04 - 0(.02 - - Herfindahl Mercosur Herfindahl RestofWorld (.1 )- I II I I I I I I I I 1986 1987 1988 1989 199() 1991 1992 1993 1994 1995 1996 1997 1998 1999 Sos,re: Sanguinetti, Pantano, and Posadas 2001. too small to achieve the required amount of sales to cover regional integration compared to unilateral trade liberaliza- costs. Regional integration agreements (RIA) can help tion? Three types of arguments can be made. First, unilat- overcome this scale problem as local markets are pooled eral liberalization does not ensure access to export markets, into a single larger market (see World Bank 2000 and Cor- though it stimulates efficiency in domestic production, and den 1972). Thus, once countries form an RIA and the in this respect it encourages exports. Reciprocity, on the domestic market expands, new products subject to econo- other hand, is a key aspect of regional liberalization, and mies of scale will be produced. because of this it could be an effective tool to gain access to The question that arises here is whether this inconve- other markets. nience of smallness can also be overcome with general, non- A second related argument in favor of an RIA is that preferential trade liberalization. Under what circumstances regional integration may enhance and protect market is this argument about economies of scale stronger for access. The multilateral trading system is still far from 81; RECENT LAC EXPERIENCES: THE ROLE OF KNOWLEDGE AND INSTITUTIONS assuring market access. Contingent protection is pervasive authors assessed the impact of Mercosur on export concen- throughout the world, and neither protection in the form tration. In this case the dependent variable is the change in of antidumping safeguards nor other border frictions can be Argentina's export shares to Brazil relative to Argentina's completely avoided. Regional integration could provide a export shares to the rest of the world. This analysis focuses WTO-plus environment where free trade within the on whether Mercosur helped Argentina diversify by allow- regional markets is more sustainable and less subject to this ing exports of new products to Mercosur but not to the rest contingent form of protection. The case of the European of the world, either through trade diversion or economies Union with the Single Market Program of 1989 is a clear of scale. Thus it is qualitatively different from the analysis example of a regional integration scheme where deep inte- in Box 4.4, which focuses on whether Mercosur allowed its gration has been pursued. Even Mercosur has established members to export new products to the world after export- a WTO-plus environment in some areas of trade policy. ing them to Mercosur. Additional exercises presented in For example, safeguard actions are not allowed within the Sanguinetti, Pantano, and Posadas (2001) considered alter- region. " native periods. Finally, another aspect in which an RIA can facilitate All the econometric evidence discussed by Sanguinetti, trade within the area, relative to that with the rest of the Pantano, and Posadas (2001) shows that tariff preferences world, is by encouraging cooperation among countries in had a significant effect on the share of Argentine exports to the area of physical integration. This could produce a sig- the region (Brazil) compared with the rest of the world. nificant reduction in transport costs, especially in the case The net effect of tariff preferences on Argentine export where the countries that formed an RIA share common shares to Brazil is positive and significant, while the same borders. For some products, this reduction may imply that preferences have had a negative and significant effect on these goods are traded within the region but not with third export share of the same items to the rest of the world. markets. Examples of these products are electricity, natural Thus, as expected, tariff preferences have encouraged ex- gas, and cement, all of which are typically subject to ports of certain items to Mercosur markets relative to extra- economies of scale in production.'2 Mercosur destinations. The evidence concerning the role of economies of scale Regional Tariff Preferences: Diversification through in Mercosur is more ambiguous, suggesting that, contrary Trade Diversion to what was predicted by theoretical arguments, sectors Regional integration will change relative prices in member subject to larger economies of scale have expanded their countries. Imports from partners become cheaper due to the exports to the rest of the world relative to Mercosur. This elimination of tariffs. This in turn affects import demand result is nevertheless not robust to small changes in the and as a consequence affects trade flows and production. As period of analysis (1992 to 1995 and 1992 to 1996). Thus, indicated previously, the presence of tariff preferences may we should interpret this result with caution. foster local production and exports of products that could Overall, we conclude that tariff preferences played a pos- not have been exported under a nonpreferential liberaliza- itive role in encouraging Argentine exports to the region. tion. Thus part of the diversification of exports we observe However, the empirical models, and in particular the tariff may be a consequence of trade diversion. However, it is also preference margins, explain a small proportion of the vari- possible that RIAs might become platforms for world ation across exports of the share going to Mercosur relative exports under certain circumstances. This possibility and to the rest of the world. Thus other factors not contem- new empirical evidence is discussed in Box 4.4. The follow- plated in the aforementioned empirical analysis may have ing paragraphs review empirical evidence concerning the played a key role. Within those, we suspect that transport question of whether Mercosur helps explain the export costs are relevant.3 In addition, most of the diversification diversification trend experienced by Argentina in the 1990s. of Argentina's and Uruguay's exports occurred within sec- tors. Hence the removal of the anti-export bias brought by Empirical Evidence the unilateral trade liberalization is probably the main Sanguinetti, Pantano, and Posadas (2001) present the sta- driving force of trade diversification in these economies. tistical results for Argentina covering 1991 to 1995. The Mercosur helped a bit, but the action was elsewhere. 87 FROM NATURAL RESOURC ES TO THIE KNOWLEDGE EC ONOMY TRADE AND JOB QUALITY BOX 4.4 Mercosur as a Pbtlorm for Wowld Exports: New Emhlcal Evkdnce Supporters of regional integration often argue that the Paraguay and Uruguay benefited from export information formation of a larger market may serve as a platform for spillovers, regardless of the creation of Mercosur. exporters to world markets (Devlin and French-Davis On the other hand, at a more disaggregated level 1999). The region can serve as a "classroom" for potential (SITC I digit), there is evidence of platform effects for exporters, where they can learn "how to export" and create all Mercosur member exporters. Argentina's exporters of a reputation as reliable suppliers. The knovd edge acquired chemicals (SITC 5) and machinery (SITC 7) have taken in the regional market can then be used to penetrate more advantage of their increase in exports to the regional mar- distant markets outside the regional agreement. The re- ket, after the creation of Mercosur in 1991, to improve gional market not only allows for the exploitation of econ- their performance in rest-of-the-world markets. The same omies of scale and learning by doing, but information is true for Brazilian and Paraguayan exporters of basic about customs procedures, required design of export prod- manufacturing (SITC 6: textiles, wood, paper, and leather) ucts, foreign consumer tastes, and firm reputation are gen- and for Brazilian and Uruguayan exporters of machinery. erated rhrougsh exports to the regional market. Moreover, there is also evidence for the small members Nicita, Olarreaga, and Soloaga (2001) explore the of Mercosur (Paraguay and Uruguay), both at the aggre- effect that regional exports within Mercosur had on the gate and disaggregated level, that they have benefited ability of members to export manufacturing products to from other Mercosur member export performance in rest- the rest of the world. From a policy perspective, the pres- of-the world markets. Since the creation of Mercosur, ence of platform effects may create sufficiently large gains irnprovements of other Mercosur member export perfor- to compensate for potential trade diversion associated mance in rest-of-the-world markets has helped exporters with preferential market access. in Paraguay and Uruguay increase their exports to the rest Information flows to other countries on rtc ional ex- of the world in all industries, except for basic manufac- porter performance in other Mercosur markets are cap- turing in Uruguay and other manufacturing (SITC 8: tured by weighting the evolution of exporter market share apparel, footwear, instruments, and furniture) in Paraguay. in the regional market by the bilateral share of trade in For example, exporters of Uruguayan machinery to the newspapers and periodicals (for example, journal of Corn- rest of the world have therefore benefited from better merce, Export Channe, Made for Export, Gazeta Mercantile) export performance of Brazilian exporters of machinery in between Mercosur members and each rest-of-the-world those markets. potential market. In sum, although platform effects are not generally After controlling for potential information flows present across all industries for Mercosur exporters, there among rest-of-the-world countries on the export perfor- is evidence that exporters in different Mercosur countries mance of Mercosur members in their respective markets, have benefited from platform effects associated with the there is no evidence of a platform effect associated with creation of Mercosur in at least some industries. Since the creation of Mercosur in 1991 at the aggregate level. 1991 exporters to the rest of the world in Paraguay and However, in the case of the small members of Mercosur Uruguay have also benefited from large Mercosur mem- (Uruguay and Paraguaiy), there is evidence of export in- ber export performance in those markets. However, note formation spillovers in the regional markets, but also in that whether these platform effects dominate the trade- the rest of the world, before and after the creation of Mer- diverting effects associated with preferential access behind cosur in 1991. This suggests that the tariff preferences (high) external tariffs remains an open question. granted ftom 1991 onward had little effect on their avail- ability to penetrate other f re in markets. Exporters in Simrrcs:Nicita,OlaTeaga,and: S .I.... 2' iri 88 RECENT LAC EXPERIENCES. THE ROLE OF KNOWLEDGE AND INSTITUTIONS Brazil's Reforms, Manufacturing Productivity, became a necessity for the survival of Brazilian manufac- and EMBRAER turing firms. However, it is very important to remember As shown, the economic reforms undertaken in the past that the real appreciation of the real was also associated decades in Latin America had significant impacts on the with extremely high interest rates and with low aggregate productivity and export performance of the agricultural sec- economic growth rates. Thus, the real exchange rate appre- tor. This was also the case in manufacturing. In this section, ciation was not good for the economy as a whole. we first review the evidence on the turnaround in Brazilian The opening of the economy also provided new means manufacturing productivity during the 1990s. We then for efficiency improvement. After decades of limited access look in detail at the case of EMBRAER, the Brazilian air- to state-of-the-art equipment, components, and technolo- craft manufacturer that in recent years has become a com- gies in general-whenever a product had a "national simi- mercial success. lar" its imports were automatically prohibited-Brazilian firms gained access to the same suppliers used by their for- Productivity Growth in Brazilian Manufacturing: eign competitors. Thus, effective rates of protection dimin- Revieu of the Evidence ished at an even faster rate than nominal tariffs, and Although estimates of productivity vary considerably with imports of intermediate and especially capital goods the specific methodologies adopted, and with the source and reacted to the opening of the economy more rapidly than level of aggregation of the data, all the available studies indi- those of final products. cate a considerable improvement in Brazilian productivity Although the timing of the productivity turnaround during the 1990s. At the aggregate level, studies performed coincided with the adoption of trade liberalization, its at Brazil's Central Bank and Planning Ministry find increases underlying causes have still been the subject of some of annual productivity growth from the 1980s to the 1990s debate. Indeed, it has been argued that part of the produc- of, respectively, 2.7 percent and 1.1 percent. World Bank tivity growth recovery could be associated with cyclical fac- estimates suggest an increase of 1.9 percent (see Teixeira da tors, particularly the deep recession that followed Brazil's Silva 2001, Bonelli and Fonseca 1998, and Loayza 2001). failed attempts at stabilization during 1990 and 1991. As expected, the improvements were even larger in the This argument, however, has been overtaken by the fact manufacturing sector, where exposure to import competition that productivity increases have persisted during the peri- is greatest. Although the estimates also vary considerably ods of output growth after 1992. Moreover, the basic result depending on the level of aggregation of the data, studies of a significant turnaround in productivity growth appears performed on the basis of industry-level figures suggest to be robust to the use of different data sources and the par- increases of around 4.5 percent from the 1980s to ticular approach adopted to measure the services of labor 1990-97.i4 Also, the efficiency gains of Brazilian manufac- and capital-including therein different measures of capac- turing firms were reflected in considerable reductions of ity use, data on electricity consumption, corrections for their markups during the 1990s. From 1990 to 1995, the unemployment, hours worked, and so forth. difference between real prices and costs fell by 21.1 percent Finally, econometric estimates confirm a positive impact in the Brazilian manufacturing sector, while another 5.3 per- of trade liberalization on the level and growth of manufac- cent reduction occurred during 1995-98 (see Moreira 2000). turing productivity. For instance, studies performed using The pressure to increase efficiency came from the un- firm-level data suggest that both the level and the rate of precedented access of Brazilian consumers to imported growth of TFP were significantly associated with reduc- goods, as nontariff trade barriers were mostly eliminated tions in tariffs and with real exchange rate appreciations and tariffs were reduced to almost one fourth of their pre- (see Hay 1997 and Muendler, Sepulveda, and Serven vious average. Also, as part of the successful inflation stabi- 2001).t5 Similar conclusions are obtained with data tabu- lization policy implemented in 1994, the real exchange lated at the industry level, which also suggests a positive rate experienced a significant appreciation during most association between the degree of import penetration and of the decade, which further stimulated imports. In this the rate of industry TFP growth (see Rossi Junior and Fer- context, cost-reducing and quality-increasing strategies reira 1999). 89 FROM NATURAL RESOURCES TO THE KNOWLEDGE ECONOMY TB ADE AND JOB QUALITY Brazil's EMBRAER were absorbed, at no cost, from other Air Force-related insti- Chapter 2 discussed the recent evolution of the structure of tutions. The government also designated to EMBRAER a Brazil's net exports.'6 It is clear that the country maintains a fraction of the income tax paid by all companies established comparative advantage in a variety of products, ranging from in the country, a transfer which amounted to US$500 rmil- tropical agriculture to capital-intensive manufactures. How- lion between 1969 and 1985.11 Other benefits include the ever, the behavior of net exports of machines since the early exemption of import, export, sales (ICM), and industrial 1990s indicates that the only sector in which Brazil main- products (IPI) taxes, import protection, large orders from tains a notable comparative advantage is in the export of Ministries of Aeronautics and Agriculture, government loans transport equipment (excluding road vehicles), especially and grants, and the support of Brazilian diplomacy in inter- aircraft. The Brazilian company EMBRAER, the relatively national sales, especially in the military market. successful firm that produces small airplanes, is an example. It must be noted that in an oligopolistic market such as In June 1999, at the Paris Le Bourget air show, the one for commuter aircraft, governments may have EMBRAER announced sales of 200 commuter jets, with con- incentives to pursue strategic trade policies, in order to tracts totaling US$6.6 billion. During 1999, EMBRAER's shift profits from foreign to domestic firms. However, sales of US$1.9 billion made it the fourth-largest aircraft although there is evidence that this has in fact been a com- manufacturer and the second-largest in the market for mon practice in that market, welfare gains from govern- regional jets, with net revenue of US$230 million. The excel- ment interventions of this type are not guaranteed, espie- lent recent performance of EMBRAER is even more impres- cially if possible retaliation is taken into consideration (see sive if one considers the fact that at the beginning of the Baldwin and Flamm 1989). 1990s the formerly state-owned Brazilian company was on As for EMBRAER's particular market strategies, one of the brink of bankruptcy. the most important is the company's early focus on reaching In December 1994 the company was bought by a con- foreign markets as opposed to concentrating on Brazil's pro- sortium led by a local financial conglomerate. The govern- tected local market, as many other manufacturing compa- ment assumed the company's debt and reduced its owner- nies in that country did. For instance, EMBRAER's first ship to 6.8 percent. The new management implemented a important product, the 19-passenger turboprop EMB-110 thorough reengineering process, and gave full priority to or Bandeirante, was first delivered in 1973 to the Brazilian the project of the ERJ-145, a 50-passenger jet that was Air Force, and in 1979-82 it had already attained a 32 per- already at the end of the development process, and the sales cent share of the U.S. market for 15- to 19-passenger air- of which would later be responsible for the company's craft. The Brasilia, a 30-passenger, twin-engine turboprop reversal of fortune. However, a complete explanation of the first delivered in 1985, reached market shares of 25 percent company's recent performance has to contain at least two worldwide and 29 percent in the United States. Thus, in the other important dimensions. early 1980s, EMBRAER's ratio of exports to production was First, EMBRAER benefited from substantial government already close to 50 percent, increasing to more than 60 support, including but going far beyond the export subsi- percent at the end of that decade. After its recent conme- dies that are now the focus of an ongoing dispute between back, EMBRAER's exports reportedly represent 95 percent Brazil and Canada at the WTO. Second, the company's com- of total sales. mercial success in the global market for regional jets cannot EMBRAER also differed from the typical import-substi- be dissociated from the knowledge and skills accumulated tution-driven company in that it avoided any attempt to by EMBRAER over the years, related to the company's early reach high degrees of local content in its products. This foreign market exposure, associated with its export orienta- option, it must be noted, was not available to most Braz2l- tion and the coupling of high R&D investments with strate- ian industries, which were subject to domestic content laws gic partnerships with foreign manufacturers. that forced them to use domestic suppliers for inputs."' As for the government policies used to support During the 1980s, the ratio of imports to production was EMBRAER, it must be emphasized that the company bene- close to 50 percent, and it has increased in recent years, fited from much more than the current subsidized credit reaching 62.4 percent in 1999. In fact, almost all the key lines. EMBRAER's initial projects, staff, and equipment components and systems in EMBRAER's planes, such as 9o RECENT LAC EXPERIENCES. THE ROLE OF KNOWLEDGE AND INSTITUTIONS engines and avionics, are purchased from foreign-based sup- its products, and a softening of the international aircraft pliers: 95 percent in the case of the ERJ-145 program and market, led to delays in the project. 85 percent for the ERJ-170/190 family of 70-to-108-pas- Production of the EMB-145 began during EMBRAER's senger jets, whose initial deliveries are expected for 2002.19 third phase, after its privatization. Among the changes intro- As a consequence of EMBRAER's early focus on inter- duced by the new management must be emphasized the shift national sales, the company has been an important contrib- from an engineering-driven to a market-driven strategy. utor to Brazilian exports. However, because of its intense Indeed, among the planes developed by EMBRAER, the use of foreign inputs, the company's net export perfor- three main commercial failures were characterized by being mance has been less impressive, especially if one takes into overpriced for their market segment, even though all were consideration that its local suppliers also tend to have a technical successes with good operational records. Moreover, large import-to-production ratio. For instance, it has been especially during the late 1980s, the company opted for an estimated that between 1975 and 1988, EMBRAER's net expansion of its product line by engaging in technically exports averaged US$123 million per year. But average net sophisticated expensive development projects at a time when exports in that period shrink to less than US$20 million its financial structure was already fragile. This type of deci- when an adjustment is made to consider imports by sion would probably not be made by the new private con- EMBRAER's suppliers (Dagnino 1993:54). trollers, committed to the profitability and financial sustain- As for EMBRAER's technological strategies, one could ability of their investments. distinguish three phases in its history of product develop- The new phase has also been characterized by the use of ment and manufacturing (Frischtak 1993, 1994). The first risk-sharing partnerships, four in the case of the EMB-145, was characterized by the manufacturing of products devel- and 10 in the case of the new ERJ-170/190 family.20 oped outside EMBRAER, either by the Air Force's Centro Another important partnership was created in November T6cnico da Aeronautica (CTA) or by foreign companies. It 1999, when a French consortium including Aerospatiale- is a period of intense learning in production technologies, Matra, Dassault Aviation, Thompson-CSF, and Snecma and in the areas of marketing and customer technical assis- bought a 20 percent stake in EMBRAER. Among other tance. A crucial role was played by the knowledge absorbed gains, this new partnership is expected to improve the from foreign companies that partnered with EMBRAER company's access to international financial markets, and to during this period. Those important agreements, with Aer- contribute to its expansion in the military and in foreign macchi, Northtrop, and Piper, were made possible by the markets in general. For the French companies, the partner- intervention of the Brazilian government, which in practice ship with EMBRAER could probably increase their odds of used them as conditions to gain access to the Brazilian civil being selected as suppliers of the Brazilian Air Force in its and military markets. forthcoming new round of equipment procurement (see A second phase started in the late 1970s and lasted until Goldstein 2000 and Bernardes 2001). the company's privatization in 1994. It was characterized by It is important to note that a rigorous cost-benefit the development of in-house formal R&D activities, which analysis of Brazil's commitment to the development of generated a series of increasingly sophisticated planes. EMBRAER is hampered by the difficulties in quantifying While EMBRAER still counted on government finance for the numerous benefits that the company received from the the military planes, three of the five main projects of this Brazilian government. However, it is clear that the total period were targeted primarily at the civil market, and were amounts involved are very large, because they cover a variety financed with the company's own resources and commercial of government policies: the provision of specialized human loans. Although all projects were the source of important resources, preference in government procurement to the internal spillovers, generating knowledge and skills on company or its partners, direct transfers of tax revenues and which successive planes would be based, only two of them capital injections in general, a variety of tax exemptions, were a clear commercial success-the Tucano and the Brasilia. subsidized credit, and import protection, among others. The development of a sixth plane was started in the late The total opportunity cost of the above resources should 1980s-the EMB-145-but the company's excessive in- be weighted against the social benefits derived from debtedness, in addition to the commercial failure of some of EMBRAER's establishment and growth, including the 91 FROM NATUlRAL RESOURCES TO THE KNOWLEDGE ECONOMY: TRADE AND JOB QUALITY externalities or spillovers reaped by other companies, and Exports and Foreign Investment in Central EMBRAER's contribution to Brazil's balance of payments, Amercan Countries: Tax Incentives, Institutions, given its focus on international sales. Although these ben- or Human Capital? efits are difficult to quantify, the fact that EMBRAER has Over the last decade, Costa Rica and El Salvador experi- concentrated on system-integration activities and main- enced exceptional economic growth.21 In both cases, the tains a low degree of indigenous content in its products, expansion of exports-in particular those from EPZs-- suggests that at least in terms of net exports, and the devel- suggests that foreign investment and exports played a opment of specialized local suppliers, the company's contri- major role in fostering growth. Nevertheless, the profile of butions have been limited. exports and dynamic sectors differs substantially across the More indirect spillover effects were probably significant two countries, suggesting that their future development at the local level because EMBRAER, together with CTA paths will be quite different. Most notably, the types of and the Instituto Tecnol6gico da Aeronautica (ITA), report- technologies being adopted are considerably distinct. edly contributed to the transformation of Sao Jose dos This case study examines the policies implemented in Campos, from a relatively small city in the 1950s, to a these countries, and the challenges that must be faced. The prosperous industrial center with a population of a half- study starts by describing the recent economic performance million today. It must be noted, however, that significant of the two countries. It then investigates the determinants of local investments by multinationals in the automobile and the different composition of exports and foreign investments consumer goods industries had already taken place before between Costa Rica and El Salvador during the 1990s. The the establishment of EMBRAER in 1969. much higher stock of human capital in Costa Rica seems In sum, much of the current market success of to explain why the more technologically advanced foreign EMBRAER is related to its early exposure to foreign mar- investments chose Costa Rica as opposed to other countries kets, both through its export orientation, and in terms of in the region. To explore this view, the study looks at the its strategy of combining high R&D investments with explanations provided by firms about their decisions to imports of foreign inputs and know-how, and strategic locate in Costa Rica. partnerships with foreign companies. However, although EMBRAER successfully developed into a world-class man- Exports and Growth dutring the 1 990s in Costa Rica ufacturer of technologically sophisticated products, the and El Salvador costs involved in its development were enormous, and we Like all Central American economies, Costa Rica and LEl do not know whether the company could have succeeded Salvador experienced dramatic events during the 1980s. without substantial government support. Moreover, it is Costa Rica initiated the decade with a balance-of-payments not clear that the net present value of the public invest- crisis triggered by large internal and external macroeco- ments made in EMBRAER are positive, even considering nomic imbalances. The stabilization required a large deval- its recent commercial success. For instance, due to the uation, as well as fiscal and monetary tightening. The fixed company's focus on system inregration and its limited use exchange rate regime was abandoned, and a more flexib]e of local suppliers of inputs and components-both impor- crawling peg was adopted, together with other structural tant aspects of its market and technology strategies- adjustment measures. EMBRAER's interaction with local industries has not been In the 1990s, all the economies in Central America, significant, so that knowledge and technology spillovers starting at different times and at different rates, embarked have probably been limited. Even if the company has been on a process of liberalization of markets, reduction of in'- a significant contributor to Brazilian exports, its net ex- port protection, reduction of the size of the public sector, ports have been much more limited due to the fact that and incentives to exports. With macroeconomic stability in EMBRAER's imports have also been considerable. Finally, place, the scenario was remarkably different, and long-term it must be remembered that the effectiveness of credit or development objectives were again a top priority. Like in other types of subsidies directed at shifting profits from for- the 1960s and 1970s, policy focused on how to speed up eign firms to EMBRAER is limited by the possible retali- growth. The goal was now sought through exports of new ations that those measures may trigger. goods to new markets and through foreign investment. 92 RECENT LAC EXPERIENCES: THE ROtE OF KNOWLEDGE AND INSTITUTIONS Both Costa Rica and El Salvador grew remarkably fast FIGURE4.6a during this period, with per capita GDP average annual El Salvador: Composition of Exports, 1995-96 growth rates of 6 percent and 7.7 percent, respectively. To put these numbers in perspective, Costa Rica expanded its 20.5% GDP by 78 percent, and El Salvador doubled its GDP dur- / ing the period. The yearly aggregate exports of both coun- /09 -, Coffee tries during the decade grew even more dramatically, dou- . 1 Shrimp bling for Costa Rica and more than tripling for El Salvador. ro. EPZ rs The ratio of exports to GDP for both countries grew sig- nificantly over time, most notably in the second half of the 1990s. The expansion of exports was not concentrated in tradi- tional goods, nor was it the result of favorable movements Sourwe Monge 2001. in world prices. On the contrary, the expansion of exports was concentrated in new sectors, which changed the con- figuration of the two economies. To illustrate how fast FIGURE 4.6b the change took place, Figures 4.6a, 4.6b, 4.7a, and 4.7b El Salvador. Composition of Exports, 1999-2000 display the average composition of the exports during 99% 1995-96 and 1999-2000 for both countries. Clearly, the 1 4% importance of the traditional goods-coffee for El Salvador, and coffee and bananas for Costa Rica-decreased. 5 i.8% / * ffe Though part of the decline in the share of traditional / Shrim goods was due to new producers in the international markets O tLhers 5 EPZ and increased protection from developed countries, mainly 34.1% the European Union, part was also the result of the develop- ment of EPZs. For El Salvador, the share of exports out of EPZs, mostly maq&ilas, increased from 41 percent to 54 per- cent in less than five years. For Costa Rica, this same share went from 27 percent to 58 percent.22 For Costa Rica, such Swirce Monge 2001. an impressive growth was in great part due to the initiation of operations of Intel and other high-tech firms in the coun- try (Remec, Sawtec, Cinair, and Sensortronics, for example). tion in the region, the worldwide movement toward free The rising importance of EPZs in both countries is illus- trade, and the formation of trade blocks-would have trated in Figure 4.8, which displays total exports and increased the flows of foreign investment to the region, and imports of firms in EPZs, and Figure 4.9, which shows the the exports of maqiila and other products. It should be fraction of the total gross and net exports generated by noted that in several countries, notably in the Caribbean, EPZs. Here, the net series is constructed by removing the rapid advances in information and communications tech- imports made by firms in EPZs, that is, (ExportsEPZ - nology have also added impetus to the development of new ImportsEPZ)/(Toral Exports - ImportsEPZ). The importance export activities in EPZs (see Box 4.5). However, govern- of EPZs in gross export and import flows accelerated at the ments in both Costa Rica and El Salvador actively pro- end of the 1990s. Furthermore, EPZs were not simply moted the development of EPZs. increasing both gross exports and imports, since the ratio of net exports more than quintupled for Costa Rica and Tax Incentives in EPZs: Inevitable Reforms almost tripled for El Salvador, in less than five years. in the Future There is little doubt that forces working throughout the Both Costa Rica and El Salvador provide significant bene- period-such as the political and macroeconomic stabiliza- fits to the establishment of export-oriented firms in their 93 FROM NATURAL RESOURCES TO THE KNOWLEDGE ECONOMY TRADE AND JOB QUALITY FIGURE 4.7a training of the labor force. Also, in Costa Rica the benefits Costa Rica: Composition of Exports, 1995-96 seem to target employment in underdeveloped regions, but 11 1% those are additional benefits, not requirements. 26.7% O While it is difficult to assess the actual effectiveness of these benefits, it would be extreme to think that they have /t' 18.2% U Coffee been redundant. Figure 4.10 displays the FDI received by 18--.2% . 3 1:fBanana 0 Meat the two countries during the 1990s. It is obvious that they r S' E AgrSculture have grown dramatically. After a slow start, especially for El M Manufacturing Salvador, the flows of FDI accelerated during the second half 1.3% of the decade. Figure 4.10 also shows the large difference between Costa Rica and El Salvador in the volume of invest- 28.4%Fo - - - | . I r ments. Until 1996, FDI in El Salvador was negligible. Costa Rica, on the other hand, has consistently received large Susr,-e: Monge 2001. inflows and, apart from an expected slowdown during FIGUIRE .7b 1998-2000, the FDI flow increased steadily. Only in 1998 Costa Rica: Composition of Exports, 1999-2000 did El Salvador receive more FDI than Costa Rica, but even that cannot be taken as a signal of reversion because it vvas 58.4%, _ .; . . mostly due to the large privatizations taking place in El Sal- / e -f, 0.s6Wvador. In 1998 and 1999, El Salvador successfully privatized four electricity distribution coEnpanies, three thermal gener- / BnCoffee ation plants, and the national telephone company (which == Meat was split in two). The sum of these operations amounted to El Sugar * Agriculture almost US$1 billion. * Manufacturing C EPZ A common feature of the EPZs in these countries is t4e 17.9%( use of tax incentives to promote foreign investment. This is also true for EPZs in the Caribbean. In the case of the Dominican Republic, Law 8-90 permits a maximum of 20 Source: Monge 200)1. percent of production to be sold in the local market. These types of export requirements will become illegal under the territories. And, despite the fact that most of the estab- WTO framework by January 2003. Hence the aforemers- lished firms are foreign owned, this specific set of incentives tioned countries, as well as others in the LAC region, w:.ll is directed to exports, independently of nationality. Never- need to make reforms to bring their incentive systems theless, both countries also have incentives directly aimed within the WTO's legal framework. Efforts are under way to at foreign investors. Table 4.5 compares the incentives pro- raise the income-per-capita criteria used by the WTO to vided by Costa Rica and El Salvador as of 2001. It is clear provide exemptions to this rule. But if these efforts faI, that the incentives are biased toward export-oriented mar- these countries will need to change their incentive structure. kets, in sharp contrast with the incentives from the import- More specifically, the relevant authorities should consider substitution schemes of the 1960s and 1970s. reducing corporate income tax rates to low levels for all C(osta Rica and El Salvador offer similar incentives. firms, which should continue to attract FDI. This approach There are no restrictions on the repatriation of profits or is now favored by Costa Rica. The main advantage of this remittances, no import or exports taxes, and no sales taxes solution is its simplicity, because the corresponding tax sy5- as long as the goods are sold abroad. However, there are tem is significantly simplified, thus reducing the costs (eva- important differences. On one hand, El Salvador is more sion, etc.) associated with complicated tax structures. On generous with fiscal incenrives. In every tax dimension the other hand, this approach requires a substantial amount explicitly mentioned, El Salvador offers either a larger ben- of regional coordination among Central American and Ca- efit or a longer duration. Costa Rica, on the other hand, ribbean countries to prevent a race to the "bottom" in terms seems more proactive in terms of providing assistance with of tax rates, which could cause fiscal difficulties. 9-[ RECENT LAC EXPERIENCES: THE ROLE OF KNOWLFD(GE AND INSTITUTIONS FIGURE 4.8 Costa Rica and El Salvador: Exports and Imports from EPZs | Exports El Salvador - Imporcs El Salvador _ Exports Costa Rica - - Imports Costa Rica 4,000 3,500 -- m3.00() -/ 2,500- v:. ,000 - - v. 1,500 --- . 1,000 - B_7 o-_-.-_-o------>---- 500 - 0 1993 1994 1995 1996 1997 1998 1999 2000 Soarce: Monge 2001. FIGURE 4.9 Costa Rica and El Salvador: Share of Exports from EPZs Ratio of Gross Exports El Salvador -0- Ratio of Net Exports El Salvador Ratio of Gross Exports Costa Rica -- Ratio of Net Exports Costa Rica 60 - 50 - 40 - io ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~~~~- ,.,i ---,-------- 20- 10- l l ll 1993 1994 1995 1996 1997 1998 1999 2(000 Sorarce: Monge 20)01. Another alternative is to remove the export require- look for locations where there are other similar plants to ments and replace them with criteria permitted by the take advantage of any knowledge externalities. WTO, such as objectives for developing marginal geo- In the meantime, efforts at the regional level should coor- graphic areas, or employment generation could also help dinate the implementation of such changes to prevent in- stimulate future investments. The disadvantage of this set vestment diversion and a race to the bottom (in terms of incentives is that dynamic firms might not be attracted of providing more and more onerous fiscal incentives for to those areas precisely because high-tech firms usually firms) among LAC countries. We believe that making such a 95 FROM NATURAL RESOLRCES TO THE KNOWLEDGE ECONOMY TRADE AND JOB QIJALITY BOX 4.5 tfonnaion Seices in Barbados and Jamaka For the countries that manage to successfully participate arrangemen[s between universities and private companies in the new knowledge-based economy, the development are also found in Barbados, where the local University of of the ICT sector is the source of dynamic economic the West Indies has had fruitful collaborative partnerships opportunities. With the development of the Internet and with information services firms, the dramatic reduction in telecommunications costs, the In the field of telecommunications infrastructure, local diffusion of ICT has accelerated. governments have recognized the fact that the activities of Confronted with the persistent decline of their tra- businesses in the information technology sector increas- ditional primary exports, countries such as Bdrbados, ingly rely on the movement of large amounts of data, so Jamaica, and Trinidad and Tobago have visualized the that the provision of modern and competitive telecom- [CT sector as the source of significart opportunities munication services is crucial for their development. for export diversification, and for the attraction of FDI. Thus, both Barbados and Jamaica have embarked on tele- Local authorities have made developing an export-ori- communications reform programs intended to assure the ented information services sector one of their top priori- provision of competitive, state-of-the-art services such as ties, and have implemented a series of policies to that end. low-cost broadband capacity at special rates for informa- Because comparative advantage in information services tion services companies. is crucially related to the availability of a labor force with As a result of the attractive conditions offered to infor- the appropriate skills, and to a good telecommunications mation services companies, some Caribbean countries infrastructure, important efforts have been devoted to im- have been able to attract a considerable number of in- proving local conditions in those two areas. In addition, vestors in areas such as customized software development, new policies have been implemented to advance the rele- industrial design, telemarketing, data entry, and multi- vant institutional setup, including the creation of new media. These firms comprise both foreign, but also investment promotion agencies, EPMZs and industrial increasingly domestic, firms. The number of such compa- parks; strengthening intellectual properts rights; and nies entering the offshore information services sector has establishment of venture capital funds directed at pro- grown exponentially, particularly during the 1990s. In moting new technology-based companies. 1997 they employed 3,000 workers in Barbados, or 2.5 In education, measures have been taken at different percent of the work force. In Jamaica's Mon ego Bay Free- levels to adjust the supply of skiled workers to the needs zone, for example, in 2000 those companies employed of the newly established sector. The measures include 4,500 employees, sutpassing employment in the declin- strengthening the information technology component of ing manufit.ruring industry. primary and secondary education, and making changes It is important to note that some characteristics of Bar- in the curriculum and degrees offered by graduate and bados and Jamaica provide additional incentives for FDI undergraduate university programs, notably in Jamaica's that are not necessarily met in other Latin American or University of the West Indies, which is considering estab- Caribbean countries. For example, these countries have lishing a science park. some of the highest teledensities in the world; labor forces Interestingly, some of the main initiatives in the area are literate and English speaking; labor costs are compar- have been led by the private sector. In Jamaica, for in- atively low, including benefits; and these countries have stance, a successful in-house training program developed geugraplhic proximity to the large North American mar- by a software firm, that was intended to prepare software ket. However, in many respects the experience of these programmers for the company, was evenrually franchised Caribbean countries is illustrative of the challenges and to the island's second main university, the University of opportunities involved in the development of the infor- Technology. With this collaborative framework, supported mation services sector. by the government educational and export promotion agencies, the originally private training program will be applied in other institutions across Jamaica. Cooperative SeAnrce Broane 2001. 96 RECENT LAC EXPERIENCES: THE ROLE OF KNOWLEDGE AND INSTITUTIONS TABLE 4.5 Costa Rica and El Salvador: Incentives for FDI TYPE OF INCENTIVE COSTA RICA EL SALVADOR Income Tax Exemption depending on the location. Ranges from 100I% for 100% exemption svhile the firm operates in the country. 8 years and 50% for 4 other years to 100%'o for 12 years and 50%,, for 6 other years. Impu-rt Taxes and Tariffs 100% exemption on materials, machinery and most 100% exemption on materials, equipment and machinery. equipment. Sales Taxes 100%7 exemption for the entire period. 100%' exemption for the entire period. Municipal Taxes 100% exemption on estate taxes and taxes on net capital. 100% exemption from taxes on assets. Reinvestment Exemption of 75% of income tax for 4 years. Not mentioned. Training and Research Assistance from the National Institute of Learning (INA). Not mentioned. Employment Tax exemptions assigned for firms based on underdeveloped Not mentioned. regions. The exemptions are for 5 years, starting ar 1(01% the first year eiidinig in 2%c the fifth year. Other Exemption from taxes to rcsale of estate. Exemption of taxcs to resale of estate. Exemption from export taxes. No taxes or restrictions on repatriation of profits. No taxes or restrictions on repatriation of profits. No taxes on remittances abroad. No taxes on remittances abroad. Processing firms and service firms can sell up to 25%. and 50%, Free sale in the local market as long as import, income, sales in local market and other taxes are paid. Developers and managers ot the industrial parks have che same Managers and developers of industrial parks are exempted from incentives. 100% of the income tax (for 10 years), taxes over net worth (fir 15 years) and taxes on resale of assets (except the manager of the park). Local governments can grant additional benefits. iSi,ur. Monge 2001. transition should not be costly from a developmental view- sidered a serious candidate. The usual argument is that the point, because the main benefits of these zones for foreign final decision was based on the combination of geographi- investors are the establishment of transparent institutions cal proximity to the United States and internal labor mar- safeguarding the property rights, and the geographic location ket conditions-with political stability and the quality of and skills of the local labor force. Yet human capital played a the labor force being highlighted. While Intel is the most key role in determining the structure of FDI in Costa Rica. commonly mentioned high-tech firm installed in Costa Rica, it is not the only one. Recently, other technology FDI and High-Tech Firms in Costa Rica: firms, such as Procter and Gamble, Roche, and Abbott Lab- The Roles of CINDE and Human Capital oratories, also started operations in the country. In 1996, Intel decided to locate a plant in Costa Rica, and It is true that Costa Rica has been more politically sta- the plant was built in approximately two years. Why did ble than most Latin American countries. However, it is not Intel choose Costa Rica? The Costa Rican government clear that high-tech firms are particularly vulnerable to offered an attractive set of incentives, such as subsidies on expropriation or other forms of political risks. On the con- electricity and several tax exemptions. But other countries, trary, the high-tech sector seems to be less sensitive to the such as Brazil, Chile, El Salvador, Indonesia, Mexico, the risk of expropriation, since the main source of value is the Philippines, and Thailand, competed with similar incen- know-how and human capital of the managers. tive packages. Indeed, at the beginning of the selection An important factor that helped attract FDI to Costa process by Intel executives, Costa Rica was not even con- Rica was the role of the CINDE, a private nonprofit orga- 97 FROM NATURAL RESOURCES TO TIH KNOWLEDGE ECONOMY TR SDE AND JOB QUALITY FIG URE 4.10 Costa Rica and El Salvador Net FDI 1.200 - _| - Costa Rica * El Salvador 1,000- .2 800- ? 600- c 400- 200 0 -200 -IllllllllI 199( 1991 1992 1993 1994 1995 1996 1997 1998 1999 20 0 Sur-ce: Monge 2001. nization. It was founded in 1983 by prominent business- health services, nightlife and cultural amenities, and natural people, supported by the Costa Rican government, and resources (for which the country was increasingly better financed by grants from the United States Agency for Inter- known, given the ecotourism boom). By 1995, several high- national Development (USAID). Its broad mission was to tech multinational corporations (DSC Communications Ccr- help in the development of the economy, but the attraction poration; Sawtek, Inc.; Merrimac Industries; and Remec) had of FDI was always one of its top priorities. In the early set up plants in Costa Rica. l 990s CINDE realized that the country was losing com- When word got out in 1996 that Intel was shopping for petitiveness in unskilled-labor-intensive industries to other a new site, CINDE played a key role in attracting attention members of the CBI and also due to the prospects for to Costa Rica, and essentially managed to get the country NAFTA, which would give Mexico better access to the U.S. on the list of candidates. Not only was CINDE important market than beneficiaries of the CBI. At the same time, in convincing Intel to consider Costa Rica as a possible CINDE was losing the USAID funding it had enjoyed since location for its plant, but it was also extremely helpful to its creation. Given these circumstances, it decided to focus Intel in conducting its research and obtaining the credible its FDI attraction efforts on fewer sectors, choosing the ones and consistent information it demanded. Moreover, its links that were a better match for Costa Rica's relatively high to and credibility with the government allowed it to play a education levels (that is, skilled-labor-intensive industries). key role in arranging successful meetings between Intel For the strategic plan of 1993, CINDE focused on sectors executives and government authorities. associated with the electrical, electronic, and telecommuni- Besides its support of CINDE, the government was very cations industries. These sectors not only required more diligent in responding to Intel's concerns in areas such as skilled workers, but were also experiencing fast growth in education, electricity, and taxes. The concessions made the United States, and strong competitive pressures were were not specific to Intel, but were generally applicable ro forcing companies to search for low-cost locations around other companies as long as they met the required condi- the world. It was thought that these sectors were a particu- tions. In this sense, it could be argued that these were not larly good match for Costa Rica, not only because of its concessions, but rather Intel-inspired reforms to improve high-quality public institutions, but also because of its sup- the country's performance. For example, the "concessions' ply of technicians and engineers at relatively low cost, and included the addition of a one-year certificate program of also because of the widespread knowledge of English. More- education focused on technical skills and physics or chem- over, there was a high quality of life, with good access to istry competency, and a one-year associate degree program 98 RECENT LAC EXPERIENCES THE ROLE OF KNOWIEDGE ANTD INSTITUTIONS focused on semiconductor manufacturing. These reforms ity of the labor force asked firms located in various coun- came from recommendations made by a special committee tries in Latin America to rate, on a scale from 0 to 10, sev- that was created for this purpose and was composed of, eral aspects of the labor force in each country. Figure 4.11 among others, the Minister of Education and the Minister shows the results of the survey in terms of two variables: of Science and Technology. 23 productivity and speed of learning. The actual numbers While the activities of the government and CINDE should be interpreted with caution, because different were key, in what follows we explore further the role of the industries may be located in the different countries, and qualified labor force. Box 4.6 conducts a further statistical sample sizes may vary significantly. Yet, these responses analysis that shows that the levels of FDI across LAC coun- make a strong case in favor of the hypothesis that the labor tries can be explained by differences in the levels of human force in Costa Rica compares favorably with other coun- capital. tries, even with the most advanced ones, such as Argentina The results of a survey from the Association of Ameri- and Chile. The results should not surprise those who are can Chambers of Commerce in Latin America on the qual- aware of the fact that Costa Rica consistently ranks at the BOX 4.6 FM and Human CapilS in LAC In our discussion of the determinants of the rniritudt these ptriods onthe dittfrent indicators of human capital, and composition of exports and FDI in Costa Rica and El and the second set includes additional controls in these Salvador, we argued that human capital played a major reg rcLijonS (the controls are output per worker and a CBI role. Our view is that the qualirx of the labor force in the dummy variable). Results for the coefficients on the different countries determines the attractiveness of the human capital indicators are summarized in Table 4.6. country to foreign investors. A higl,ti sk;ilit labor force The results show a consistent positive relation be- attracts capital in more dytnamic, knno ledazu-inrcnsive tween initial levels of human capital and subsequent sectors, thus sustaininL' export d namnissm and crnhanrimni Fl , Most of the time, this correlation survives the inclu- the prospects for economic growth. sion of additional controls. The results imply, for exam- Here we systematically investigate the relation be- ple, that an increase of five years in the iverage level of tween FDI and human capital for a groAlp of 21 Latin education of the population above age 1i5 is usually asso- Ametican countries. We look at the relation between the ciated with an increase in FDI of 3 percent of G)DP. The 1990 human capital level of a country and the amount of results also suggest that the ~ n i ti%i ir of FIP to the ini- net FDI received by that country (as a share of GDP) in tial level of human capital is even higher at higher levels the subsequent years. To check the robustness of the of education (secondmar-), and that it increased signifi- results, we use different human capital indicators and dif- cantly in the second half of the I q.-K9k I- The results for the ferent time frames for the FDI. We also analyze the deter- United Srares share similar p-atrerns, and FDI from this minants of the net FDI received spe cifi(ally from the ouLntrrv seems to be even more sensitive to human capi- United States, to check whether it is in any way dififrent tat levels. Overall, the evidence supports the point that from the general pattern observed. The human capital human capital determines the amount ot FD I received by indicators come from the Barro and Lee dataser, and they a country. The role of education in spurrinp: economic are the fullow in,,. all for the population above age 15: growth seems to go beyond the tradirional view of percentage of secondary school attained, :-aer;lve school- enhancing productivity within given economic activities. ing years, average years of pr i ma ry school ng, and a% erage Education can determine the prospects of the country in years of secondary schooling. The periods of PD! analyzed terms of access to technologie, and development of new are 1985-98, 1990-98, and I 98 1-98. The first set of dynamic sectors. results refers to simnlip repressions of the average P1) for 99 FROM NATURAL RESOURCES TO THE KNOWLEDGE ECONOMY TRADE AND JOB QLTALITY TABLE 4.6 FDI as a Percentage of GDP and Human Capital Indicators in LAC Countries: Econometric Results UNIVARIATE REGRESSIONS FLOWS FROM UNITED STATES. 1985-98 1990-98 1995-98 1994-98 ALTERNATIVE INDICATORS COEFFICIENT COEFFICIENT COEFFICIENT COEFFICIENT OF HUNIAN CAPITAL T-STATISTIC R2 T-STATI'TIC R2 T-STATISTIC R2 T-STATISTIC R Percent Secondary Attainment 0.071 0.282i 0.0938 0.2857 0.121 (0.2039 0.1634 0.2.948 2.8737 2.8984 2.3191 2.9632 Average Schooling 11.4284 0.2367 (0.5633 11.2318 0.7747 0.1877 0.921)8 11.2145 2.5522 2.517 2.2029 2. -X95 Average Primary Schooling (0.5699 ().1732 ().7291 0.1474 0.953 0.11)79 0.9858 0.0132 2.0973 1.9057 1.5936 1.4694 Average Secondary Schooling 1.3716 0.3208 1.7829 0.3495 2.5182 0.2986 2.9672 1.3i853 3.1491 3.3592 2.9903 3.6281 N obs= 21 N obs = 2 1 N obs= 21 N obs= 11: MULTIPLE REGRESSIONS: INCLUDE OUTPUT FER WORKER AND CARIBBEAN BASIN DUMMY FLOWS FROM UNITED STATES: 1985-98 990-98 1995-98 1994-i98 ALTERNATIVE INDICATORS COEFFICIENT COEFFICIENT COEFFICIENT COEFFICIENT OF HUNfAN CAPITAL T-STATISTIC R2 T STATISTIC R2 T-STATISTIC R2 T-STATISTIC R.? Percent Secondary Attainment (0.0)667 0.4581 0.0715 0.376 0.0921 (0.2623 (0.1765 (o.414 2.5909 2.045'9 1.5852 3.1636 Average Schooling 0.4755 0.434 0.4757 0.351 0.6794 0.2608 1.2491 01.4ii95 2.3521 1.7927 1.57 3.4762 Average Primary Schooling 0.5041 0.3577 0.47556 0.2894 0.6104 (.2006 1.5193 (t. .31)7 1.5428 1.055S 0.8362 2.2854 Average Secondary Schooling 1.5495 0.5241 1.5821 0.4369 2.3459 0.3485 3.4523 0.5`94 i.2485 2.628 2.3712 4.5531 Nobs= 21 Nobs= 21 Nobs= 21 Nobs= - I N,,i: Dependent variables are average level of total net FDI/GDP f(br the periods indicated and average level iif net FDI/GDP from the United States (1994-98). Independentr aet ables are indicators of human capital for 19911. Controls included in the multiple regressiois are- dummy for countries in the CBI and output per worker. Countries incladed in the tortal FDI/GDP sample are Argentina. Bihivia, Beaail, Chile. Colombia, Costa Rica, the Dominican Republic, Ecuadoe, El Salvador, Guatemala, Haiti, Honduras,Jamaica, Menico. Nicaragu.a, Pana ar, Parga, Pereu, rhe Rrpdbli-a Bolivaeiana dc Venezuela, Trioidad and Tobago, and Uruguay. Countries included in the U.S. FDI/GDP D amp-p are Argentina, Barbados, Brazil, Chile, Colombia. Costa Rica. the Dominican Republic. Ecuador, Guatemala. Honduras, Jam.ica, Mexici, Panama. Peru, the Repiblica Boltvariana de VenezulIa and Trinidad and Tobago. Se,ine: Estimates by A. Monge (Northwestern Univrsity). top in the United Nations Development Programme's Hu- The scores are relatively lower, but still not bad (above 8 man Development Index. out of 10), in terms of specific knowledge and skills for A survey on the labor market conditions faced by high- engineers. In this last dimension, technicians and skilled tech firms was conducted in 2000 by Promotora del Co- and unskilled workers have a more modest score, just above mercio Exterior de Costa Rica (PROCOMER), and asked 28 6. This is not surprising, because this new industry requires firms questions about the quality of managers, technicians, activities to be performed by workers who did not have any engineers, and skilled and unskilled labor, and questions on previous experience. Another aspect worth mentioning is the quality of the training programs and curricula of exist- the gap in terms of English proficiency. Consistently, firms ing educational institutions (technical and higher educa- indicate that employees at all levels are not quite as fluent tion). In general, the perception is favorable. Managers, in English as would be desirable. English proficiency, how- engineers, technicians, and skilled and unskilled workers ever, does not really seem to pose a long-term problem receive high marks in terms of productivity, speed of learn- because the industry of English as a second language has ing, and disposition to work variable shifts. Most notably, bloomed in Costa Rica. English has also been introduced in managers and engineers have the highest scores in terms of the national curriculum of primary schools in the whole productivity, creativity, initiative, and speed of learning. country. 1(10 RECENT LAC EXPERIENCES: THE ROLE OF KNOWLEDGE AND INSTITUTIONS FIGURE 4.11 Perception of Intemational Investors on Labor Force Quality 1 0 -- _ - _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 5 3 2 Costa Rica Chile Argentina Mexico Brazil Sosarce: Monge 2001. The PROCOMER survey also suggests that wages of perception of the firms in relation to quality and appropri- high-skilled labor are attractive to foreign firms. It is well ateness of the programs for their needs. While the adequacy known that labor costs in Costa Rica are higher than in of the curriculum for their specific needs was questioned, other neighboring countries. But, for any economic deci- the quality of the two institutions received very good sion, the relevant consideration is the price of effective reviews. Table 4.7 reports the scores, in ascending scale labor, that is, the cost of labor per unit of time versus its from 0 to 10. Private universities and the public national productivity. Interestingly, the firms in the survey gener- university obtained significantly lower marks. Moreover, ally do not find the costs of different workers in Costa Rica few enterprises in the survey indicated that they had hired to be high: only 4 percent make that claim for engineers professionals from the public university. and managers, 12 percent for administrative staff and tech- This group of firms has also needed to train its workers nicians, and 20 percent for skilled and unskilled workers. because of both the specific tasks required for each job and The general pattern is that the relation of productivity to the newness of the sector in the economy. Most of the firms cost improves with the level of human capital of the use the local plant or the headquarters of the company for employee. Indeed, the responses of the firms in the sector training, and only 46 percent and 39 percent, respectively, suggest that there are significant cost reductions by operat- use public and private local institutions (see Figure 4.12). Yet ing in Costa Rica as opposed to the United States. the survey also indicates that there is room for improvement. The two main educational institutions, the University On one dimension, the high-tech firms indicate that workers of Costa Rica and the Costa Rican Institute of Technology, need to improve their discipline, especially toward following also receive good evaluations. The survey asked about the the safety requirements on the job. Also, as Table 4.7 makes TABLE 4.7 Perception of High-Technology Firms on the Education/Training in Costa Rica (On a Scale of 1 to 10) INSTITUTION QUALITY OF EDUCATION QUALITY OF TRAINING PROGRAMS Technological Inscitute of Costa Rica 9.33 9.00 University of Costa Rica 9.02 8.75 National University 7.75 6.80 Private Universities 6.67 6.40 National Institute of Learning 7.90 7.75 Technical Secondary Schools 8.30 8.00 Other 8.50 8.30 Source: Monge (2001) based on a PROCOMER Sursey. 101 FR(OM NAT L RAL RESOUIRCES TO ITHE KNOWLED GE ECONO.MY TRADE AND JOB QIJALJTY FIGURE 4 12 Costa Rica: Fraction of Finms Using Training Programs 7H1 - 60 50- ~. 3 0 ') 11 - 320- 10 Headqlarters Plant in Public Costa Ruan Private Costa Rican Private Foreign Other Costa Rica InstitLItions Institutions Institutions Sa,zos: Monge 2001. c lear, although the quality of the two main universities and The challenge for Costa Rica is to provide enough quan- technical schools is satisfactory, at least 70 percent of the tity and quality of professionals in those areas. Interesting ly, respondents express concerns about the appropriateness of the educational market seems to be catching up with the the short-term classes and training programs offered. demands of the new sector. There is a significant increment The operation of this new sector provides Costa Rica in the areas of computation and information technology and with an interesting opportunity, but also with serious chal- in industrial engineering (Figure 4.15). At a lower level, lenges. If the sector is to grow, the country has to produce the government and the market are also actively reacting to workers with the required skills. Figures 4.13, 4.14a, and the new demand for skilled workers and technicians: in less 4.14b decompose the employment of the firms in the sur- than four years, enrollment increased by almost 10 percent vev into different skill groups. In most firms, the largest in the technical colleges, reaching almost 44,000 students share of employment is composed of production workers. in 1999; the National Institute of Learning saw its studenllt But it is clear that, relative to any other sector, these firms body increase from around 47,000 in 1994 to 124,191 in are high-skill-intensive, since 25 percent of their labor force is composed of engineers and technicians, and 20 per- FIGURE . 1 3 cent are skilled workers. Figures 4.14a and 4.14b decom- Costa Rica: Composition of the Workers in the pose engineers and technicians into areas of expertise. High-Technology Firms According to the survey, most of the firms indicated that they planned to increase their levels of operation. About 60 2 _ percent indicated that the growth will originate mainly from the expansion of existing lines, while 40 percent indi- * Unskilled O4~, Technicians cated that expansion will arise from new lines of produc- o-1 Engineers tion. The planned expansion is not small. The firms expect (0% 0 MSanger to hire approximately 2,100 new employees in 2001; 2,3 30 . ,k*xeudve in 2002; and more than 2,600 in 2003. From their answers, it seems that the new demand will have a larger share of 7 skilled than unskilled workers. Still, their projections in- clucde 500 engineers and technicians hired in 2001 alone. 18% For the other two years the estimates are slightly lower. Soairce: Monge 2001. 11(2 RECENT LAC EXPERIENCES THE ROLE OF KNOWLEDGE AND INST-ITUTIONS FIGURE 4 14 (with the exception of pupils per tutor in secondary educa- Costa Rica: Composition of Engineers in the tion). This observation becomes even more relevant once High-Technology Finins one realizes that El Salvador compares favorably with the other Central American countries (for example, Honduras 23% - 44K7 and Nicaragua). Furthermore, these differences are the re- / Industrial Engineers sult of government policies. Figure 4.16 compares the ex- / *~~~~~~~~~~ Industrial Engineers o Computing penditures on education in Costa Rica and El Salvador as [ MeLhanical Engineers fractions of GDP and government expenditures. Costa Rica 4% 0 Maintenance and , _ s v Quality Control spends much more in education, both because it is richer AX \t .- * , *~~~~~1 Chemical Engineers 3%_P U ChemicaOther and because it spends a higher fraction of its income. The 62 *OEertriral Engineers - fractions declined during the 1980s, partly because of the fiscal imbalances and, in the case of El Salvador, increased Io ------------ military spending. 1091 These differences show up, for example, in the degree of Solece: Monge 20101. illiteracy in the population and its composition. El Salvador has a much higher illiteracy rate (28 percent compared to 5 percent in 1997), which is concentrated in the younger FIGURE /t.Iit, population. While the composition of the illiterate popula- Composition of Technicians in the High-Technology Firms tion in Costa Rica is becoming more concentrated in the 14% elderly over time, the fraction of young individuals in the total illiterate population in El Salvador is high and 57_'_/ 1C,. increasing, reflecting the poor coverage of primary educa- U Industrial Engineers tion and probably the impact of the war. This implies that, .. -- 0.9 1: Computing for long periods of time, El Salvador will have an active - 0~~~~~~F Merhanical Engineers ln ie o Maintenance and labor force with a high percentage of low-skilled workers. Quality Control v Chemical Engineers *Othe r 132 E ElectronicsiElectrrricity Sum2ming Up: Different Paths of Development in the Futuire Costa Rica and El Salvador are similar countries in terms of natural and geographic conditions and cultural aspects. Srm,n*-: Mu nge 2001. Also, from a macroeconomic perspective, they had similar performances during the 1990s, driven by exports and for- eign investment. Nevertheless, the Costa Rican economy, 1997; and the applications for diplomzados and technicians supported by a highly educated labor force, experienced an (one- and two-year programs) in the Costa Rican Institute important transformation during the period. Several tech- of Technology increased by almost 300 percent. nology-intensive firms started operating in the country, thus changing the profile of its employment and exports. El Sal- Human Capital in Costa Rica and El Salvador vador, on the other hand, concentrated its expansion mainly Combined with stable institutions, the skills of the Costa in traditional labor-intensive sectors, mainly apparel. Rican labor force made it the leading candidate in LAC for The operation of the high-tech firms in Costa Rica has a the development of high-tech industries. To effectively cre- two-way relation with the quality of the labor force. First, ate a basis of skilled workers, technicians, engineers, man- its introduction was determined precisely by the relatively agers, and other professionals, any country needs to have a high quality of the local labor force. Second, at the same solid primary and secondary education. Table 4.8 compares time, its successful operation demands increasing amounts the statistics for Costa Rica and El Salvador. Costa Rica is and quality of labor at different levels. The great challenge far ahead of El Salvador, not only in coverage for all levels, for the Costa Rican economy is to respond adequately to but also in terms of the efficiency of the educational system the increased demand, by improving the overall quality of 103 FROM NATURAL RESOURCES TO THE KNOWLEDGE ECONOMY TRADE AND JOB QUALITY FIGURE 4.15 Costa Rica: Supply of New Engineers in the Areas in Most Demand by New High-Technology Finms 700 l0.0 Computer/lnformation Technology -_- Industrial Electric/Electronic -.- Mechanical | 600 C: ° 500 B Z 400 300 ~ E 300 ,,_,~~~~~~------------_____--' 200 100 o - - - - - - - - _ _ -_ _ _ _ - - ---- ---- ---->Ho<-- --- --- 1995 1996 1997 1998 19''9 Graduation Year Source: Monge 20(01. universities and high schools, and education in rural areas. gains obtained from the development of the traditional Besides, infrastructure reforms that have been lagging behind industrial sectors should be partly directed to an aggressive (such as telecommunications) can become a limiting factor if policy of investments in human capital to improve the cov- not dealt with in a timely fashion. If Costa Rica responds to erage and efficiency of basic education. Recent innovations these challenges, the possibility of attaining a sustained in the education sector of El Salvador, such as increasing growth path-where investments in human capital reinforce parental involvement in the management of schools, could growth, which feeds back into human capital accumula- provide avenues for future improvements in this area. The tion-is an attractive and plausible development path. recent adjustments in terms of infrastructure and telecom- El Salvador has a difficult task ahead: long-term growth munications are important advances. But with low human could be limited by the quality of its labor force. A transi- capital, those investments might not lead to dynamic tion to more dynamic sectors is unlikely in such an envi- growth in labor productivity. ronment. The demographic profile of the labor force in terms of the relation between ages and skills raises even The Impact of NAFTA on Mexico's more serious concerns about the long-term prospects. The Trade Structure In Chapter 2 we analyzed the evolution of Mexico's net ex- port structure since the early 1980s. We noted that begin- TABLE 48 ning in 1995, the country developed a clear comparative Costa Rica and El Salvadon Human Capital Indicators advantage in machine exports, especially in road vehicles, COSTA EL and office (data-processing) and telecommunications equip- RICA SALVADOR ment. In the following subsections we take another look at Literacy Rare (1995) 95.0 72.10 Gross Enrollment Primary (1997) 103.5 973 Mexico's trade structure by comparing its structure within Gross Enrollment Secondary (1997) 48.4 36.8 NAFTA to that observed with the whole world. After the Gross Enrollment Tertiary (1997) 30.3 17.8 analysis of the structure of we turn our attention to Pupils per Teacher Primary (1997) 30.0 33.0 trade, Pupils per Teacher Secondary (1997) 18.0 16.0 . the electronics industry, which includes the office equip- Coeff. of Efficiency Primary UNESCO)0 (1995) 80.(1 63.4 ment sector, including personal computers. The objective Years Input per Graduate Primary (19951 7.3 14.2 is to gain some understanding of whether Mexico's com- a. United Nations Educational, Sciertific, and Cultural Organization. Siao-: Monge 20(11. parative advantage in these products has been transformed 10)4 RECENT LAC EXPERIENCES- THE ROLE OF KNOWLEDGE AND INSTITUTIONS FIGURE 4.16 Costa Rica and El Salvador. Public Expenditures in Education 4o -~~~~~~~~~ Costa Rica: GDP *o*Costa Rica: Government Experiditures| 35 - +t " 30_..___--,_ S20rfe: Monge2001. from production processes driven by competitive labor that the exchange rate adjustment that took place in late costs or by the establishment of new industties driven by 1994 might also have played an important role in stimu- innovation and knowledge creation. lating exports of manufactures. But the fact that the struc- tural change within NAFTA occurred earlier indicates that The StruBctuere of NAFTA Trade the negotiation of the agreement itself was an important Figure 4.17 shows the structure of Mexico's net exports to force for change. Canadian and U.S. markets during 1986-98. As with the Figure 4.18 provides further descriptive evidence about structure of its net exports to the whole world, this chart the impact of NAFTA on Mexico's trade structure. The shows a structural change toward the end of the period, graph shows the evolution of the Grubel-Lloyd index of when net exports of machines turned positive. This change LIT introduced in Chapter 2, as well as Brulhart's (1994) was driven by exports of road vehicles, telecommunications index of marginal intra-industry trade (MIIT). The MIIT equipment, and office and automated data-processing index measures the extent to which changes in trade struc- equipment. However, this change occurred in 1993 for ture are driven by IIT. exports to Mexico's NAFTA partners, while it appeared a The graph contains the indexes constructed with data year later in the structure of overall net exports. Hence it from Mexico's trade with the world and trade with its seems that change occurred in anticipation of the actual NAFTA neighbors. The data clearly show that all four implementation of the agreement in January 1994. It is indexes of IIT jumped after 1991, when the formal negoti- worth recalling that the United States had approved fast- ations were launched. Indeed, the intra-NAFTA indexes are track legislation for the negotiation of NAFTA in 1991. virtually indistinguishahle from the overall indexes for most Thus it is quite likely that foreign and domestic firms made of the period. Consequently, we can safely conclude that investment decisions to expand their operations in Mexico Mexico's LIT is mainly driven by its North American trade. prior to the formal implementation of the agreement. This This is nor surprising when one considers the fact that interpretation of the evidence is consistent with economer- NAFTA merchandise exports rose from about 80 percent of ric evidence provided by Freund and McLaren (1999), total exports in 1991 to over 90 percent by early 2001. The who find anticipated structural effects of trade agreements share of merchandise imports originating in North America throughout the world. However, we must acknowledge fluctuated around 75 percent during the last decade. 105 FRO)M NAtIRAL RESOURCES TO THE KNOWLEDGE ECONOMY Tb ADE AND JOB QUALITY FIGURE i 17 Mexico: Stnrcture of NAFTA Net Exports 8() - 60 - -_ __ _ _ __ _ _ __ _ _ 40 - _ __ _ _ __ _ _ X~~ ~ ~~~~~~~~~~~~~~~~~~~~~~~~~ CCEhemical[ 1dsS 20 El Machinery U Capital Intensive Labor Intensive D 0 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~Cereals 80( A nrepical sgr t ue inrt 0.6 Forest I Raw Materials _____ ~~~PJPetroleum --6(1- _ - ~~~~MIIT l) 5 _~O ~- NQNAFTA IIT r > =8 >~~ N N A AH N- ; , Wi'th Tosethis evidnene in col ande,e ito FisrefAlcuin the Anenytax. eio TeJeo egah onMexic'snI ithae NoThAneca Nfo TAowing auscind thek M at Ithex Teeaesvrlipratqusin eae oterl NAFTA MIIT ~ ~ FstTrckNAT 0.3 - 1981 1982 1983 1984 1985 1986 1987 1988 19891 1991) 1991 1992 1993 1994 1995 1996 1997 1998 1999) With this evidence in hand, it is difficult to deny that FDI inz Mexico: The Role of Geography th-e negotiation of NAFTA had substantial structural effects anzd Human Capital on Mexican trade. The following subsection looks at the There are several important questions related to the role internal geographic distribution of FDI in Mexico. We at- of geography and human capital in the determination of tempt to understand how geography and some of the other foreign direct investment in Mexico.2' First, though the 'new" endowments such as human capital help explain the agglomeration of FDI on the border with the United States location of foreign firms within Mexico. is well known, is distance the dominant factor, and hence I O(1 RECENT LAC EXPERIENCES: THE ROLE OF KNOWLEDGE AND INSTITETIONS are states or countries farther away at an insuperable disad- ently, is Chiapas marginalized because it is distant, poorly vantage? Second, since Lucas (1979) showed that a human educated, rural, and heavily indigenous? capital-augmented production function can generate radi- Figure 4.19 shows that most FDI per capita is concen- cally different returns to capital than might be calculated trated on the northern border. However, there are impor- on the basis of capital-labor ratios alone, how important is tant exceptions. First, the highest concentration is, in fact, a well-educated work force to attracting FDI? Third, given in Mexico City, although this appears to result from the the recent literature on agglomeration effects or infrastruc- fact that companies with headquarters there might not reg- ture, how important is the concentration of economic ac- ister the actual location of their investments, but rather the tivity in urban areas? Fourth, how much of a barrier to par- location of headquarters. On the other hand, the states ticipating in FDI are language or cultural characteristics immediately around the capital also show high per capita related to being members of indigenous groups? Put differ- rates, and the central region has, traditionally, been the FIGURE 4.19 Mexico: Geographic Distribution of Variables Per Capita FDI (From 1994 to 20(10) Schooling (in Years) s . > -s1 r, ~~~~~~~~~~~MiLlions of DolLars \ 6|,,lYears 600 06001200Mis122-52 568 04502-9725 .07)8 972-08 - 5.01-17.00 N I S N IL a a 600 0 600 1200 Mil62 600 0 600 1200 Mies Degree of Urbanization Percent Speak^ing Indigenous Languages _ ~~~~~~~~~~Urban Population Share ~ -_Percentage Sr: Arlca andMalo .0.405 0.001.() -0.5- 6 _1.1-510 ^^ , {XJT| o ~~~~~~~~~~~~06-()- ^ 5.01-17.00 5: - . _ ~~~~~~ ~ ~ ~~~~0. 7-0.8 5,.17.10-25.00( > +l-t ' ~~ . _ 0-1.0 2 5.10)-37.00 N X N-_) , ; A We E ~ :W*XE 600 0 600~~~~~~~ D Md.~~~~ 6tO ° 600 1200 Mib5 Source: Aroca and Maloney 200 1 107 FROM NATURAL RESOURCES TO THE KNOWLEDGE ECONOMY TRADE AND JOB QUAlITY center of industrial activity. Second, the region around are very distant from the United States by road, they are very Guadalajara in central Mexico is not particularly close to close to Miami by boat, and hence our distance measure is the border, yet receives substantial FDI, perhaps due to the probably incorrect. The relatively high density of indige- presence of Hewlett-Packard and other transitional firms. nous population appears to be correlated with relatively Finally, Quintana Roo, in the Yucatan peninsula, also high FDI, given the road distance from the United States. shows relatively high FDI per capita. In sum, proximity to the United States, higher educa- Schooling is more evenly distributed, although it shows, tion, and more urbanized areas capturing perhaps agglom- again, the highest levels along the border, the central region, eration effects or infrastructure are positively correlated and Quintana Roo, suggesting some correlation with FDI as with FDI. However, areas with indigenous populations well. The difficulty of isolating the influence of any particu- appear to receive less FDI only because they tend to also lar social variable is highlighted by the fact that the degree of have lower education, are more rural, and are located farther urbanization also seems highly correlated with both educa- from the United States. The following section explores in tion and FDI. This complicates attempts to identify the influ- detail how geography and knowledge have affected the evo- ence of ethnicity since the highest concentration of indige- lution of the electronics industry in Mexico. nous-language speakers is in the south, which also tends to have lower levels of both urbanization and education. The Mexican Electronics Industry Again, the high degree of correlation of these variables Can labor-intensive offshore manufacturing activities means that multivariate analytical techniques are necessary, evolve into skill-intensive, high-value-added operations? and these are presented in Table 4.9. Columns A, B, and C What is the role of government promotion policies in the suggest that distance, education, and urban population context of an open trade regime? Although there are no share enter with the expected signs, although it is difficult general responses to these queries, the recent experience of to identify the impact of the last two. The share of indige- the Mexican electronics industry does provide some clues nous population does not enter significantly, suggesting that to what the possible answers may be. there is no effect independent of the fact that these commu- nities tend to have lower education levels, live in rural areas Background and Facts with less infrastructure, and are concentrated in areas far Between 1995 and 1998, Mexico's GDP grew at an average from the United States. Column D generates a positive impact rate of 5.9 percent per year. This impressive performance, of the indigenous variable if we drop a dummy included to after a quick recovery from the effects of the 1994 devalua- capture the fact that, although Yucatan and Quintana Roo tion, was partly due to the dynamism of the country's man- IABLE 4.9 Tobit Estimates DEPENDENT VARIABLE: FDI PER CAPITA (TOTAL 1994-2000) VARIABLES COEF. A T COEF. B T COEF. C T COEF D T COEF. E T COEF. F T Education t80.4 1.13 388.5 3.35 496.4 4.94 Urban Population Share 1343.3 1.67 2047.1 3.8 Percent Indigenous Population 364.6 0.28 -88.2 -0.07 230.5 0.17 1887 2.51 -5885.5 -2.7 -2698.8 -2.32 Distance to the United States -47.1 -5.65 -49.5 -6.01 -47.9 -5.44 -42.4 -5.25 Dummy 468.2 1.09 545.5 1.25 656.4 1.48 1447.7 2.06 Constant -1086.1 -1.14 -178.9 -0.35 -1644.1 -1.64 -2629.5 -3.15 509.4 3.71 413.2 3. I Sample Size 31 31 31 31 31 31 LR chi2(df) 51.4 50.1 48.8 46.4 10.7 5.6 Pseudo R2 14.30%o 13.90% 13.60% 12.90% 3.00% 1.60% Log Likelihood -154.3 -154.9 -155.6 -156.8 -174.7 -177.2 Note: EDI iower than US$50 per period of 1995-2000 set to zero leading to 9 left-censored observations at FDDI<0, 22 uncensored observations. The Federal District (Mexico City) was dropped from the sample. Education is measured in years. Percent indigenous variable is calculated as the percentage of the population of age live years or older or more that speaks a native langag,e. Dummy equals I for Qoinrana Roe and Rucaran. and 0 otherwise. Sou-rc: Aroca and Maloney 2001. 108 RECENT LAC EXPERIENCES THE ROLE OF KNOWLEDGE AND INSTITUTIONS ufacturing sector, the annual rate of growth of which was 9.4 ropolitan area (see Casaneuva and Brown 2000). Among the percent during the same period (see Lederman and others main manufacturers of final products with plants in Jalisco 2001a, 2001b). Manufacturing exports evolved from 43 per- are IBM and Hewlett-Packard, which established them- cent of total exports in 1990 to 85 percent in 1999. Within selves in the 1970s, at the time when the Mexican policy manufacturing, no industry matched the growth of the Mex- for the informatics industry required foreign firms to pro- ican electronics manufacturers. Between 1995 and 1998, for duce locally in order to supply the domestic market. Other instance, output and exports in that industry grew at aver- manufacturers with more recent investments include NEC, age annual rates of 25 percent and 31 percent, respectively. Unisys, Siemens, AT&T, Alcatel, Ericsson, and Motorola. Exports of electronic products totaled US$29 billion in More than half of the manufacturers of electronic prod- 1998, representing 27.6 percent of manufacturing exports. ucts operating in Mexico are maquilas. These are mostly Although imports have also increased significantly, net foreign-owned companies that assemble imported com- exports of electronics products have been positive since ponents for re-export, benefiting from fiscal incentives 1995, attaining almost US$5.6 billion in 1998. As in other and simplified administrative procedures. Although most manufacturing industries, the United States is Mexico's maquilas are located near the U.S. border, the fraction of main trading partner in electronics, accounting for 93.8 per- maquila employment located in nonborder states has in- cent of the country's exports and 83.5 percent of its imports. creased from around 5 percent in 1986 to more than 22 Although NAFTA led to the elimination of tariffs in the percent in 2000. bilateral trade between Mexico and the United States, it did The share of the maquila operations in total electronics not pose a threat to their respective electronics industries. industry employment increased from 38 percent in 1992 to Indeed, the United States already had a vcry low tariff on 54 percent by 1998. Almost all the consumer electronics electronics products (below 5 percent), and the Mexican manufacturers operate as maquilas, and so do most makers exports of electronics are complementary to those of its of telecommunications equipment. The non-maquila seg- neighbor. Indeed, the main Mexican export products are ment of the industry operates under the same conditions as consumer electronics and data-processing equipment, the rest of the manufacturing industry, and has stronger mainly TV sets and computers, whose shares in total elec- local ties, both in terms of markets and suppliers. In fact, tronics exports are 17 percent and 15 percent, respectively.25 until the opening of the economy to foreign trade during As for the Mexican imports, they consist largely of parts and the late 1980s, the non-maquila electronics industries components, notably semiconductors (20 percent) and cine- established in Mexico were oriented mainly toward the scopes (11 percent). domestic market, in which they enjoyed considerable pro- In fact, the consumer electronics industry is even geo- tection from import competition. graphically integrated with the United States, with most With trade liberalization, however, and particularly manufacturers located in Mexico's northern states, especially after the implementation of NAFTA, the outward orienta- near the border. The state of Baja California, in particular, tion of the non-maquilta sector has increased considerably. has attracted 81.6 percent of the investments made in the For instance, the ratio of exports to production in the non- electronics industry during 1995-98, which amounted to maquila sector increased from 35.1 percent in 1989 to 60 more than US$2.3 billion, and is the host of the main percent in 1994 to 83.5 percent in 1996. In the specific Korean and Japanese manufacturers of consumer electronic case of the computer industry, it must be noted that the products. As for the computer and telecommunications liberalization of trade only occurred in 1990, at least two equipment industries, they are mostly located in Jalisco, years after the other subsectors of the electronics industry. which ranked second in terms of investments in electronics, Moreover, during three years computer manufacturers ben- totaling US$394 million during 1995-98. Concentrated in efited from fiscal incentives that included tariff exemptions, the city of Guadalajara, the electronics industry established in the context of a special regime that rewarded local tech- in Jalisco grew 52.6 percent between 1993 and 1997, nological development and the use of domestic parts and increasing its share of the narional production of computer components. and telecommunication equipment from 31.9 percent to With 637,000 workers in 1998, the electronics industry 71.4 percent, mostly at the expense of the Mexico City met- is a relatively large employer, responsible for 9 percent of 109 FROM NATtJRAL RESOURCFS TO THE KNOW'LEDGE ECONOMY' TRADE AND JOB QUTALITY the jobs in manufacturing. Within the electronics industry, As already mentioned, a very important factor in the the maquilas have shown the strongest performance in decision to locate electronics manufacturing operations in terms of employment generation, and they have been Mexico is the possibility of exporting with zero tariffs to responsible for most of the new jobs created in the indus- Canada and the United States, provided the NAFTA ruaes try. Between 1992 and 1998 employment grew 94 percent of origin are met. The possibility of the maquilas importiog in the miiaquiila sector, as opposed to a meager 3 percent in parts and components with zero tariffs has recently becorne non-maquzila companies. This is due both to the compara- restricted to imports from NAFTA partner countries, on- tively larger growth in the number of maqysilas, and to the der a clause of the agreement signed in 1994. However, fact that these types of companies perform assembly opera- FDI has not been significantly affected, in part because the tions that are relatively more labor intensive. share of parts and components purchased within NAF'I'A has increased significantly, and because of the proliferation Factors Determining Location of free-trade agreements dropping the duties on imports As for the factors that explain the decision of foreign com- from non-NAFTA countries. This is notably the case of t he panies to locate production facilities in Mexico, it must be goods imported from the European Union, which will he remembered that before the 1990s, and outside the maquila tariff exempt under a free-trade agreement signed in 1999. sector, the main motivation for setting up local plants was Moreover, for some Asian countries, a further incenti ve to gain access to the protected Mexican market. This moti- to establish maquila operations in Mexico is the fact that vation is no longer relevant, so the new foreign investors they face quantitative restrictions in the U.S. market. These are driven, at least in part, by favorable local conditions rel- aspects appear to compensate for the fact that manufactur- ative to locations in other countries, in order to sell not ing wages in Mexico are still higher than in several locations only to the Mexican but also to the U.S. and other markets. in Asia-China and Malaysia-and Central and Eastern Moreover, given the dramatic reductions in transport costs Europe-Bulgaria, Hungary, and the Russian Federation. and the possibiliry of 'fragmenting" the stages of the elec- For instance, it has been estimated that the total landed cost tronics manufacturing process, location decisions may now (including inventory costs, tariffs, and shipping) of produc- involve the fragmentation of those stages across different ing electronics products for the U.S. market is up to 5 per- cities and countries. cent lower in Mexico than in Malaysia (Merrill Lynch 2000). The main motivation for setting up electronics manufac- Besides traditional cost considerations related to relative turing plants in Mexico has been the combination of low labor and capital abundance, a factor that has become cru- labor costs, geographic proximity to the United States and, cial in the investment decisions of electronic companies is particularly after NAFTA, trade preferences and certainty of the feasibility of implementing cfficient logistics structures access to the U.S. market. The role of geography and human for the relationships with both suppliers and customers. capital in attracting foreign investment to Mexico has been Indeed, because of the very fast rate of technological change analyzed in a previous section. The point here is that geog- in the electronics industry, with important technological raphy has been crucial in all segments of the electronics "leaps" occurring with a frequency of less than two years, industry. However, among maquiilas it has been coupled to a companies cannot afford to maintain sizable stocks of either decision of locating only the labor-intensive assembly oper- final products or parts and components. For instance, it Was ations in Mexico, while using mostly imported parts and indicated in interviews that a Pentium processor loses 10 components. Although labor-intensive operations have also percent of its value every month, which explains the necd been the focus of non-maquila businesses, their range of to maintain a strong integration among companies in the products and operations has widened considerably toward computer supply chain, mainly through the use of infor- more skill-intensive activities, and they have made a grow- mation technologies. In fact, the semiconductor industry ing use of inputs produced locally by companies that have is the one in which that type of integration appears to be also been attracted to the country. This option has been most crucial (Mann 2001). favored among manufacturers of products with more Although the need to produce "to order" derives to a dynamic technologies, computers for instance, who tend to great extent from the swift rate of technological obsoles- supply the American market in a "made-to-order" regime. cence of electronics products, the high importance attrib- [ Io RECENT LA( EXPERIENCES THE ROLE OF KNOWLEDGE AND INSTITUTIONS uted to logistics is also related to the high degree of inter- be emphasized, however, that only after the advent of trade nationalization of the electronics industry. Particularly in liberalization did those companies increase their invest- recent years, due to widespread trade liberalization and ments in local R&D activities. plummeting transport costs, the production activities of The requirements for the second-stage manufacturing electronics manufacturers have become increasingly frag- activities vary according to the specific type of component. mented across plants and countries. Thus, while total inter- Although some components are relatively simple and their national trade grew 6.8 percent per year during the 1990s, manufacturing processes are mostly labor intensive, the pro- trade in electronics products increased at an average annual duction of most of the inputs of the electronics industry is rate of 15 percent. particularly capital- and knowledge-intensive, as is notably The renewed importance of logistics implies a crucial the case for semiconductors. Thus, in Mexico, most of the role in location decisions for the availability of an appro- components used by the electronics industry are imported priate transport infrastructure. In this respect, compared to from the United States, and to a lesser extent, from Asia. other countries, Mexico benefits from the fact that products The two final stages of electronics manufacturing are the can be shipped by truck to the United States, reaching the most labor intensive, and their location is determined border in less than 24 hours from many locations-18 mostly by cost and logistics considerations. Although the hours from Jalisco to Texas, for example. Moreover, with degrees of vertical integration vary by product and by cor- the increased flow of airfreight between Guadalajara and poration, most OEMs tend to focus on R&D and on final the United States, the number of daily flights to and from system-integration activities. They thus tend to rely in- several major cities in the United States has increased sig- creasingly on providers of electronics manufacturing ser- nificantly-there was only one flight per day a few years vices (EMS), which perform most of the third and fourth ago. There are also a number of overnight airfreight com- phases of the manufacturing process, and purchase their panies such as DHL, Federal Express, and UPS that offer components from specialized suppliers. With the diffusion their services in Guadalajara (Merrill Lynch 2000). of "lean production systems" and "just-in-time" relation- ships between OEMs (or EMS providers) and their suppli- The Dynamics of Agglomneration ers, a trend has developed in the direction of an increasing Good logistics are also easier to implement when the com- geographic concentration of integrated supply chains. panies that perform different parts of the manufacturing At least in Mexico, the regional agglomeration of elec- process decide to set up plants at the same location. At least tronics manufacturing activities has been led by the invest- in Jalisco this has been increasingly the case in recent years, ments of OEMs, which, after being established in a partic- and this has further contributed to the attractiveness of the ular location have, in some cases, attracted some of their region as a potential location for electronics manufacturers. suppliers of parts and components. This process has been To better understand these recent developments, it is worth- relatively limited in the case of the mnaquila operations, while to sketch the stages of a typical production process in which tend to purchase most of their inputs abroad. In the that industry. There are four main stages: R&D, manufac- non-maquila sector, however, notably in the case of Guadala- ture of electronic parts and components, assembly of printed jara, the OEMs have been followed by manufacturers of circuit boards (PCBs), and final assembly. some of their main inputs, including producers of compo- The first stage, R&D, is performed by so-called original nents, PCBs, plastics, packing materials, and sheet metal, equipment manufacturers (OEMs), which in turn distri- and electronics distributors. It is worth noting, however, bute these activities across various R&D centers, which are that most of these companies are foreign, while domestic often specialized in certain types of products. Larger OEMs manufacturers are limited to activities with relatively low usually have R&D centers in different countries, with each capital and knowledge intensity, such as packing materials, specific location determined by corporate strategies and the labels, and cleaning and maintenance services (Casaneuva local availability of trained scientists and engineers. In and Brown 2000). the case of Jalisco, this availability appears to have played a Recently, some of the leading EMS providers have also role in the decision of several companies to implement been attracted to the region by the OEMs to which they locally the development of some of their products. It must provide services elsewhere in the world. After the establish- FROM NATURAL RESOURCES TO THE KNOWLEDGE ECONOMY TRADE AND JOB QUALITY ment in 1987 of the first major EMS provider in Guadala- role this played in the location decisions of the first OEMs jara, at least seven other leading companies have set up local established inJalisco in the 1970s, interviews with industry plants in recent years, mostly after 1997. Some of the main leaders suggest that there are significant externalities de- EMS companies already in the region are SCI systems, Flex- rived from the availability of a skilled workforce in the tronics, Jabil Circuit, NatSteel Electronics, and Soletron. region. Although some of these companies, as well as other EMS In particular, the presence of trained scientists and engi- providers, have plants in other Mexican cities-Hermosillo, neers partially explains the establishment of R&D centers Monterrey, Puebla, and Tijuana-it is clearly in Guadalajara by the main OEMs, which have developed locally some that the largest concentration of EMS providers is found products-printers, for instance-that are now manufac- (Merrill Lynch 2000). tured in other countries.26 Moreover, one of the leading Interestingly, the establishment of EMS providers in the companies in the industry has decided to transfer from the region is attracting more manufacturers of parts and com- United States to Guadalajara part of its strategic financial ponents to Guadalajara, which could in turn increase the operations, a decision that also reveals the high quality of competitiveness of manufacturers of final products, poten- the local telecommunications infrastructure. There is also tially attracting further OEMs and EMS providers. More- evidence of increasing collaboration between local universi- over, some recent modifications in the scope of the product ties and the electronics manufacturers established in Jalisco, lines manufactured locally by these companies suggest that although it has so far been mostly limited to the training of they have been upgrading the technological content of personnel and support for the establishment of new local their manufacturing activities. Thus, at least some of the R&D centers by OEMs (Casaneuva and Brown 2000). The EMS providers have been widening their local product growth of the development activities performed in these lines in the direction of increasingly sophisticated prod- centers could in principle lead to spinoffs in the form of the ucts, for instance from personal computers and peripherals appearance of new, technology-based companies that would to notebooks, servers, and network equipment. Comple- take advantage of the availability of trained personnel, with mentarily, within each product line, those companies have work experience in the OEMs' R&D centers. As indicated evolved from performing locally only the assembly of in interviews, these types of incentives are behind the ongo- PCBs, to also internalizing the final assembly of full sys- ing development of a local software industry. tems (Merrill Lynch 2000). Some EMS providers have offered incentives for the Summing Up establishment of parts and component suppliers, for in- In recent years the electronics industry has been one of the stance in the form of instant access to necessary manufac- most dynamic in the Mexican economy, both in terms of turing infrastructure in their industrial parks. However, it output and export growth. Particularly after the imple- is worth noting that government incentives have played no mentation of NAFTA, the differences between maquila and role in the Jalisco agglomeration, except, as mentioned, in non-maq/ila companies in terms of outward orientation the early years after trade liberalization. This fact has led have been reduced, as the latter have considerably increased industry leaders to lobby for fiscal exemptions based on both their exports and imports. However, unlike the con- claims that Jalisco would be losing important investment sumer electronics manufacturers, which operate mostly as projects to countries that do offer fiscal advantages to in- maquilas in the northern border states, importing almost all vestors, particularly in Asia. of their inputs, the companies that manufacture computer and telecommunications equipment in Guadalajara have The Role of Education and Knowledge Policies created strong ties to other companies established in the As for the upgrading of the activities performed locally by region. Indeed, they have been able to attract the providers EMS providers, in terms of their technological content, it of electronics manufacturing services and the suppliers of has been stimulated in part by the local availability of parts and components that they use in other locations skilled personnel. Indeed, the metropolitan area of Guadala- throughout the world. These investments appear to be jara has six universities and is the second most important leading to the creation of integrated supply chains, which educational center in Mexico. Although it is not clear what could lead to important gains in competitiveness, based on t 12 RECENT LAC EXPERIENCES THE ROLE OF KNOWLEDGE AND INSTITUTIONS both lower costs and better logistics. These gains should Tourism and Development in the Caribbean attract new manufacturers, generating a virtuous circle in Tourism is a critical export sector for many countries in the which the expansion of the industrial agglomeration gen- region.27 For many economies in the Caribbean, tourist erates externalities that contribute to further cost reduc- receipts are the majority of exports, and their share in GDP tions, particularly through the improvement of logistics, can rise above 40 percent (1998 data) (see Figure 4.20). and thus leads to the attraction of new investments. Further, tourism is a growth industry globally. The Overall, the case of the electronics manufacturing com- arguments that suggest that primary products will enjoy a panies in Jalisco highlights both the possibility of evolv- decreasing share of world income apply in reverse as in- ing from pure maquila-type assembly operations to skill- creasing incomes and more leisure time raise tourist expen- intensive manufacturing and R&D activities, successfully ditures proportionally more than income. This is suggested integrated into global networks. This case also illustrates by Figure 4.21, which shows annual growth rates in real the crucial role played by the externalities derived from a global tourist expenditures of over 3.8 percent for the last good education and transportation infrastructure, and the 20 years. Some economies in the region have been able to increasing returns associated with the establishment of take advantage of this growth. In the Dominican Republic, integrated supply chains. Some of the first manufacturers for example, tourism has been a booming sector, rising were established in Jalisco at the time when protectionist from 9.1 percent of total exports during 1971-80 to 29 policies required local production to access the Mexican percent by 1991-2000. However, similar to commodities, market. However, it must be noted that such a chain of the supply of alternatives also affects revenues. Tourist events would be improbable in smaller countries where the expenditures in Latin America have risen 0.51 percent over size of local market does not justify tariff jumping. Also, the comparable period, and the region has dramatically lost the development of integrated supply chains and the tech- market share. This suggests growth opportunities for many nological upgrading of the activities performed locally took countries of the region. place only during the period that followed trade liberaliza- tion, particularly after NAFTA. Finally, it is worth noting Tourism and Development that the role of government policies has recently been very The impact of tourism can be significant on the local econ- limited, with no fiscal incentives being granted to new omy, and both input-output and computable general equi- investors outside the maquila sector. librium (CGE) analysis can reveal the magnitude and nature FIGURE 4.20 Tourism's Share of Exports and GDP Venezuela, R.B. de Mexico _ Tourism Receipts/Total Exports * Tourism Receipts/GDP Ecuador Brazil Trinidad and Tobago Argentina _ Peru Uruguay _ J : Dominican Republic 1 . , I_i= Jamaica . I Grenadar na2 St. Vincent and the Grenadines . - = St. Kitts and Nevis ml _ i Barbado s _ _ _ _ __ _ __ _ ' 2 _____ Antigua and Barbuda- _ & _ 2 - v 1_ _ ! ' The Bahamas j j - _ j ______. St. Lucia 0 10 20 30 40 50 60 70 80 Percent Source: Maloney and Montes Rojas 2001. 113 FRONI NAT URAL RESOURCES TO THE KNOWLEDGE ECONOMY T RADE ANI) JOB QUALITY FIGLIRE 4 21 LAC's Share and Rates of Growth ;o~ ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~1 20 Rates of Growth: World and LAC I 1(0 _ l9X _ i - 10~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~1 6 -31) -40 -~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ -A l _ I I I I I I I II I I I I I I A 1981 1982 1983 I984 1985 1986 1987 1988 1989 1990 1991 1992 1993 19i4 199S 1996 1997 1998 1999 Swine: Maloney and Montes Rojas 20(11. of linkages. As Table 4.10 shows, the GDP multiplier for Although it is difficult to generalize based on the avail- Barbados is 1.4, close to the average calculated for other able data, for Barbados the impact on government finances countries and at the high end of industry multipliers.2' seems favorable. The state receives over 10 percent of tourist Though more difficult to measure, the contribution to spending and the ratio of revenue to spending on infra- Iabor creation can be large. For the Windward Islands and structure and other services exceeds 1.5. Finally, the impact Leeward Islands tourism directly and indirectly accounts of cruise ship arrivals is smaller along most dimensions. TI-is for up to 24 percent of total employment, with direct is to be expected given the lower spending on housing and employment accounting for 49 percent of the workforce in food that would have linkages to the rest of the economy. Anguilla. Employment multipliers range widely from 1.4 In Barbados, this may be partially accounted for by the for Barbados to estimated magnitudes of 3 in the Wind- high foreign content of "shopping," which is estimated at ward Islands and the Dominican Republic. 75 percent. Still, the contribution of cruise ship spending is TABlE 11) Results from a Limited Input-Output Model for Barbados OVERNIGhfT CRUISE SHIP TOTAL Cwntjinetion t, GDP Direct Contribution to GDP (at factor cost) 1 ..9% 1 7% I 4.29, Direct and Indirecr to GDP (at factor cost) 18.4'. 1.5 1'i, l'9.9% GDP Multiplier 1.42 1.15 1.4 ToLurism-Related Imports/Tourism Spendlng -,1.2 39.8 i1.9 Impasit on Emiployment Direct Employment 17,631 1,'27 19,()61 Indirect Employment 6.133 2(02 6.345 Direct + Indirect/Total Employment - 2.5 1.5 2 4.1 Employment Multiplier 1 348 1.142 1.332 Ioin3ca on Governozent Financee Tourism-Related Government Revenue 121,2'28,325 1(,75',247 131,985,572 ToUrIsm-Related Government Spending 7S,94/2,I11 8,771,368 87.713.679 Re tcnue/Spending 1.536 1. 26 1.505 Tourism Related Government Revenue/Tourism Spending 11%'l 12.2% 1 1 '.f S,,ncie: Caribbean Development Bank 1996. 114 RECENiT LAC EXPERIENCES THE ROLE 0l KNOWLEDGE AND INSTITUTIONS not negligible. In the Galapagos, cruise ships owned by Increasing "productivity" of the value generated, given entities on the mainland generate 46 percent of the total the tourist base, can occur along all three dimensions. value added on the islands, channeling income through crew 1. Expand Capacity: For long-established tourist destina- wages and inputs supplied by the islanders. Locally based tions such as Barbados, the expansion of accommodation cruise ships accounted for 17 percent of gross island prod- capacity is limited to replacing existing hotels with larger uct, followed by fishing (8 percent), commerce (7.5), and hotels or, alternatively, floating capacity offshore in the farming (5 percent). In short, a small economy can be largely form of cruise ships. The second type of tourist, the cruise- supported on the multiplier effects of cruise ship visitors. ship tourist, tends to spend less overall, but given the expansion in the number of total tourists, they are desir- Possibilities for Growth able, and hence there is strong competition among islands As Figure 4.22 suggests, revenues from tourism constitute to attract them. a large fraction of both exports and GDP in many econo- 2. Reduice Vacancy Rates: Room-occupancy rates in the mies in the Caribbean. The natural question arises as to Caribbean in 1999 averaged 65 percent, ranging from over what the long-term potential for growth is. If at some point 80 percent in Anguilla to 30 percent in Belize. Hence, there the number of tourists reaches the physical capacity of the is room for using existing capacity more completely.29 Econ- environment, then how does income continue to rise- omists at the CTO have suggested that local establishments We can think of the income per tourist resource (for adopt a system similar to that used by the airlines to ensure example, beachfront, Mayan ruins) as full booking.)0 These gains, although potentially large, are L S"one-off" and cannot be extended indefinitely. Incomze = L * V 8 S 3. Increasing Value per Tourist: Expanded capacity and B L V reducing vacancy rates are fundamentally "extensive," and where L is the capacity for lodging over the potential for once exhausted, only increasing spending per tourist offers total lodging dictated by the environment (beachfront, for the potential for continued growth. Figure 4.22 suggests example), and hence the first term measures the degree to that high value added tourism can play a large role in a which the natural tourist potential has been exploited development strategy. For example, 50 percent of Bermuda's physically. The second term is the "occupancy rate" of lodg- GDP comes from tourism, and the mean spending per ing potential, or the number of visitors over total potential tourist is very high-roughly US$200 per visitor day. The visitors. The final term is the spending per visitor. relationship with GDP could imply that raising value per FIGLRE 4.22 Real Tourism Receipts per Tourist Day 0.25 0.20 Anguilla AO.S- Antigua and Barbuda -~-Aruba ~~~ 0.15 ""'EJ'- ~~~~~~~~~~~~~~~~~~~~Bahamas, The . . ~~~~~~~~~~~~Barbados --X - Bermuda 0.10 ---Cura,ao Martinique ( .00 I I I I I I I I I I 1984 1985 1986 1987 19S8 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 Soli,nr: Maloney and Montes Rojas 2001. 115 FROM NATURAL RESOURCES TO THE KNOWLEDGE ECONOMY: TRADE AND JOB QUALITY tourist has helped lead to higher income levels, or it may The results are slightly more positive if we look at imply that in high-GDP countries, higher labor and land tourist receipts not in terms of how much one visitor boys costs imply a higher-cost tourism product and hence higher in terms of a basket of U.S. goods, but in terms of the man- spending per tourist. But it must also be the case that the ufacturing unit value index used to compare the evolution product is proportionately better, and this highlights the of commodity prices. However, overall, we do not find need for continued innovation and differentiation of the tremendous gains in the value per tourist over time. This tourism product. Bermuda and St. Martin can be more ex- suggests that islands are overall not very mobile in the pensive only if a similar service cannot be easily found else- quality of their product. Further, it suggests that, without where. The question of substitutability is discussed later. initiatives to find other ways for tourists to spend their Has the Caribbean been increasing revenues by expand- money, the growth benefits will not be realized. ing the number of tourists, or by increasing the value of their product? Figure 4.22 suggests a very mixed answer. Is Tourism Fickle? For Bermuda, Curacao, and Barbados inflation-adjusted re- Table 4.11 presents the standard deviation of growth rates ceipts per tourist have been stable or increasing, but in of exports for several commodities that often complement Antigua and Barbuda, the Bahamas, Aruba, and Martinique, tourism in island economies, particularly. What drives the tendency over the last decade has been downward. volatility separates tourism from the other commodities. This may be partially due to the fact that "tourists" Where price fluctuations are perhaps the larger concern for include both cruise passengers and overnight visitors. An commodities, it is the sudden decline in tourist arrivals due increase in the relative proportion of cruise passengers could either to recession in the country of origin or increased lead to an apparent decline in spending per visitor when, in attractiveness of alternative destinations that causes more fact, all we see is a desirable increase in lower-spending concern for tourism. But what is clear is that in virtually tourists. Table 4.10 suggests that it is highly likely that every case, tourism revenues are the least volatile. Only in cruise ship receipts per person have decreased signifi- Trinidad and Tobago do tourism revenues vary by more cantly-by about 5.5 percent per year-and that overnight than the other commodities and, even then, by very little. stays may, optimistically, have increased by 1 percent. De- The central concern is that either a perceived similarity cline or small gains could be due to two factors.3' First, the among islands or simply a fad will lead to a rapid redirection rise of "all-inclusive" packages, while not necessarily de- of tourists to other locales. Put more rechnically, the price creasing the demand for goods and services of the host, does elasticity of demand for tourism is very high. The only pu;D- give large tour operators greater bargaining power with lished study to date (Rosenzweig 1988) finds very high inter- local providers.32 Competition from other destinations, in- Caribbean elasticities of over 2-a 1 percent rise in one cluding within the United States, has also put downward island's prices will lead to a loss of 2 percent of its tourists, and pressure on prices. an elasticity of 1 with respect to Mexico. Our findings, TABLE 4 li Export Revenue Volatility (Standard Deviation of Growth Rate) COUNTRY TOURISM StJGAR FRUITS COFFEE PETROLEUM AND G AS Barbados 0.08 0.17 0.33 0.96 3.61 Belize 0.36 0.11 0.21 0.69 Dominica 0.11 2.55 0.91 0.48 Dominican Republic (.11 0.13 0.08 0.32 o.69 Grenada 0.17 2.39 0.49 0.27 Haiti 0.13 2.79 1.09 0.71 1.65 Jamaica 0.09 0.23 (.15 0.13 2.66 St. Kitts and Nevis 0.11 0.37 ().91 1.S2 0.94 St. Lucia (.09 1.13 0.29 0.37 St. Vincent and Grenadinies 0.17 0.30 0.47 Trinidad and Tobago 0.29 0.23 0.22 0.28 0.24 Arsragea 0.4315 1.84 0.99 1.522 2.34 a. Includes all the conn.r.es 116 RECENT LAC EXPERIENCES THE ROLE OF KNOWLEDGE AND INSTITUTIONS although not strictly comparable, are much lower-about destinations. Argentina and Chile have seen increases in 0.4-and agree broadly with results from Spain and the tourism to their Patagonian extremities. Areas rich in nat- Mediterranean.33 This seems a bit more plausible for two ural beauty or simply beachfront, such as Brazil, could reasons. expand. Peru could follow Mexico with a more intensive First, what determines the price responsiveness of tour- marketing of its historical and cultural past. ism is, as with manufactures, product differentiation. If all 2. Increasing Value per Tourist. For countries or islands destinations are identical beaches, then price elasticities with mature tourist industries, or in areas where environ- should be very high. Similarly, as wages rise in such an econ- mental concerns prohibit further extensive development, omy over time, we would expect the tourism sector to con- improved "productivity" gains will come from providing tract over time and move to other locales. But we do not more targeted and specialized tourist services. This can observe this in the case where a destination has something range from adventure tourism, such as shark or wall diving, very particular to offer. France has very high wages and a to greater emphasis on cultural or historical tourism. In booming tourist industry due to the fact that demand to see these cases, as with other natural resource-rich exports, the Eiffel Tower is virtually vertical-double or triple the there is a great possibility for applying human capital and price and people will still go. However, the Dominican know-how to add value to the basic resource. At the same Republic, which has developed largely on mass-based time, relentless pursuit of product differentiation-for tourism with little "product differentiation," is closer to the Trinidad and Tobago, perhaps a global steel pan competi- identical beach story. To date, the Caribbean islands ave tion; for Chile, the extremeness of the northern Atacama t dr . B h Desert; for Argentina, the Train of the Clouds-are the way tended to offer a fairly differentiated product. British to maintain market share and value as wages rise in the rest tourists charmed by the Barbados's air of "little England" are not likely to visit Bob Marley's stomping grounds in Jamaica of the economy, and to reduce the fickleness of tourists. 3. Going Online. To use Mann's (2001) terminology, bhecaus ofd smaies movdemoupent partinic he, FrencS partrni tourism is a very information technology-sensitive indus- their old colonies Guadeloupe, Martiniqlue, and Srt Martin, the Dutch Curac,ao and St. Maartin, and they are not likely try. Getting the message out, defining the image, facilitat- ing arrangements, and engaging in differentiated pricing to b drwn o Cncunbecuseof odet chnge inpries, through the Internet will be the base level of competition Second, accommodation prices are keyed to the external mark-et, and over the medium term changes inexchange over the next decades. Further, arrangements are increas- maret o vwa erte mikedi eterm hanges in exhage ingly being made electronically. Travel booking will gener- rates or wages are likely to reflect more in profits than in ate US$3.1 billion by 2002 (Holder 2001). prices. Over the longer term, we may expect more hotel 4. Increase Research. Tourism represents another case where building in areas with higher profit levels, such as Cuba. 4 ) ~~~~substantial R&D is essential to take advantage of market This is probably not the case with large economies such as opportunities. The global market is evolving rapidly, and Brazil and Mexico, where a large share of what the tourist niches open for specialized products. Yet, at present, tourism consumes is "interior," and hence would be more affected statistics are not ideally suited to the types of analysis re- by exchange rate changes. quired; the Caribbean Hotel Association, the regional pri- However, decreasing transport costs and the develop- vate sector tourism organization, has no research program; ment of venues in Asia or even in the United States do offer and the universities are not heavily invested in this area alternatives that will increase this elasticity over time. (Holder 2001). Somewhat counterintuitively, Cuba's coming on line may have a salutary effect by calling more attention to the Caribbean Basin relative to other regions. Notes In sum, tourism remains a desirable sector with growth 1. This section is based on Lederman and Soares (2001). potential in the region. However, the same lessons apply to 2. This section borrows heavily from Foster and Valdes (2001). the tourism sector as to any other resource sector. 3. The last term in the decomposition of price movements is com- 1. Playing to Resource Strength. Many of the econiomies puted as a residual. combining changes in trade and price interven- tions and other possible changes in variables affecting the margins of the region have the potential to rap into the tourism between domestic and border prices. The reader should also consider market: most obviously Costa Rica, Ecuador, and Trindad that, although the real exchange rate is equal for all products at any and Tobago increasingly find their niches as ecotourism point in time, for each product the relevanr domestic and border 117 FRONI NATURAL RES)OURCES TO THE KNOWLEDGE ECONOMY: TRADE AND JOB QUALITY prices and the exchange rate were computed at the harvesc month by calculations performed on the basis of the available microlevel corresponding to the product in question. For milk and beef, sold data, which are representative of medium and large but not of rhe throughout the year, the exchange rates represent an annual average. supposedly more dynamic small firms. For instance, Muendler, Se- 4. The data for domestic prices of apples are suspect. pcilveda, and Serven (2001) find an increase of 1.8 percent in manu- 5. In general, while the immediate reaction might be to consider facturing productivity growth from 1 987-90 to 1 991-98. this outcome a cost of adopting a certain market-oriented economic 15. It is worth noting that the problems of endogeneity relate,_l to policy (to be balanced against corresponding benefits), upon reflec- reverse causality from productivity to trade openness are minimized ton, it is not obvious that it has been or will be in the medium term in these studies because of the use of firm-level data. a serious detriment to Chile, To the degree that Chilean agriculture 16. This section is based on Fajnzylber (2001a). can incorporate varieties and techniques developed elsewhere, there 17. See Dagnino (1993) for an account of the government bense- would be less to gain by replicating a sysrem of taxpayer stupport fits received by EMBRAFR. for universities and other institutions specializing in agricultural 18. See Luzio and Greenstein (1995), for example, for an analysis research leading to beneficial but hard-to-internalize results. Even of the price and performance practices of the Brazilian microcom- with respect to the human capital necessary to make use of the latest puter industry, which until 1990 was subject to severe local cont,nt agricultural production and processing technologies, the benefits of requirements and lagged international standards by at least three and absorbing the fixed costs of sustaining large agricultural research fac- as much as five years. ulties are diminished in light of the access to foreign graduate pro- 19. In this new family of jets, the number of suppliers is also grams, consultants, and scholars. The marginal dollars spent on a expecred to decline, from more than 400 to around 40, with "f,rst publicly funded, long-term, public good-oriented research project or line" suppliers becoming system integrators themselves. This should educational institution specializing in agriculture might offer higher lead to a further reduction in the manufacturing and integration returns if used to give a potential Chilean agronomist or engineer the activities carried out within EMBRAER (see Bernardes 2011). basic training and English proficiency necessary to take advantage of 20. The risk-partnership arrangement was initiated when technical services or educational opportunities available elsewhere. EMRAER was still a state-owned company, in order to pursue che This would be especially so if similar projects-leading to the same EMB-145 project in the context of a severe shortage of resources publicly accessible results-are already being conducted at the faced by the company at the time. EMBRAER's share in the EMB- expense of foreign taxpayers who have been conveniently persuaded 145 and EMB-170/190 projects is, respectively, 60 percent and 45 of rheir duty to support public research efforts and sophisticated percent. In the case of the EMB-145, the support of Banco Nacional graduate training programs. de Desenvolvimento Econnmico e Social (BNDES) was also instru- G This section is based on Sanguinetti, Pantano, and Posadas mental, because it financed US.$ 100 million of the development costs (2001a). (Bernardes 2001). 7. The Asuncion Treaty also stipulates far-reaching objectives in 21. This section is based on Monge-Naranjo (2001) and RodrfgueLz- terms of coordination of policies in other areas, particularly, policies Clare (2001). applied in the following sectors: agriculture, industry, public taxes 22. EPZs in Costa Rica include firms from the Zonas Francas and and expenditures, monetary rules, exchange rates, capital market, Perfeccionamiento Activo. services, transport, and communications. 23. Briefly, the other reforms were the following: In rhe area of 8. The 1Herfindahl concentration index was discussed in Chapter 2. electricity, the rates were very high, so the government asked the reg- 9. The magnitude of the fall in the concentration indicator (the ulator to establish a new lower rate for energy-intensive industiial gain in diversification) depends on which index we use. For example, facilities. In the area of taxes, there was an ambiguity in the appli,a- using the Gini Coefficient, export concentration falls about 10 per- bility of a new tax to companies in the EPZ system, so the govern- cent, while using the Theil Coefficient, the fall is nearly 20 percent. ment requested a formal interpretation from the Attorney General, 10. The commodity grotups are the one-digit Harmoniized System which turned ouit favorable to Intel and other companies rhen enrer- of Classification aggregates. Results do not change when using alter- ing the EPZ system. native groups. 24. This section is based on Aroca and Maloney (2001). 1 1. On the other hand, antidumping actions are still permitted 25. In fact, Mexico is the largest supplier of TV sets for the U.S. (as of mid-2001), although there was a compromise to eliminate market, and the second-largest supplier of electronics products in them by the end of 2000. See Sanguinetti and Salustro (2000) for general, surpassed only by Japan. details about tariff and nontariff barriers in Mercosur. 26. On the role of human capital. see Box 4.5 and the section on 12. This argument applies not only to goods subject to economies the experiences of Costa Rica and El Salvador. of scale. 27. This section is based on Maloney and Montes Rojas (20)011). 13. Garriga and Sanguinetti (1995) present an early assessment of 28. Aroca (2001) calculates multipliers for Chile of real estate the effect of transport costs on the composition of total trade by des- 1.02, agriculture 1.14, construction 1.21, transportation and com- tination in Mercosur. municarion 1.27, mining 1.28, manufacturing 1.28, retail 1.31, fish- 14. See Bonelli and Fonseca (1998) and Rossi Jinior and Ferreira ing 1.35, business services 1.41, and utilities 1.66. (14999). The recovery in manufacturing productivity is also supported 29. Caribbean Tourism Statistical Report 1999-2000. 1 [8 RECENT LAC EXPERIENCES: THE ROLE OF KNOWLEDGE AND INSTITUTIONS 30. Interview with Luther Miller, CTO (2001). Spending through one supplier, however, provides a stronger position 31. We regress total receipts on the number of overnight tourists with subcontractors. Hence these packages might have the same (or overnight adjusted for reported average length of stay, or hotel overall impact but different distributional implications. occupancy rates times installed room capacity) and number of cruise 33. We follow Arellano and Bond (1991) in running a dynamic ship visitors. Though a long-term cointegrative relationship might panel specification in differences of number of tourists from five geo- be expected, the results were much more stable in differences. graphic areas to each island on the calculated purchasing power par- 32. A study commissioned for the Caribbean Development Bank ity in each country relative to the principal competitor countries, and (1996) concluded that there was no evidence that all-inclusives con- several measures of total tourist demand. sume fewer local goods and services than any other stay-over visitors. 34. We are grateful to Mr. Luther Miller of the CTO for this point. 119 CHAPTER 5 How We Work: Job Quality in Emerging Sectors T HE GOAL OF ALL TRADE POLICY IS TO IMPROVE THE QUALITY OF LIFE OF A COUNTRY'S citizens, and as one key dimension, the overall quality of jobs. To paraphrase the theme of this report, it's not what you do, but how you work. Yet over the last decade, concern has emerged that trade liberalization and the emerging patterns of trade have caused workers to lose ground. This chapter argues that this is not the case. There have been major job dislocations of workers as firms adjust to more competition, but probably over the medium term, unemployment issues will not be the central problem. The quality of the jobs created will be. But here the evidence is generally encouraging. Wages in more open or resource-dependent sectors hold their own, or do better than others, perhaps due to more aggressive adoption of technologies and demand for high skill levels- Technological progress is also held largely responsible for the widening wage differentials observed throughout the region. This trend, though probably not permanent, is far preferable to the reverse. It implies that the new jobs being created are "good" in the sense of demanding more of workers in terms of human capital. This chapter examines more closely one dimension of job quality that has drawn much attention: the increase in informality over the 1990s. It offers an alternative framework that argues that this trend has more to do with the macroevolution of the region's economies, and in particular the nontradeables booms of the early 1990s. Again, increased informality does not appear chemicals and capital- and labor-intensive manufactures. to be a permanent feature of the region's labor market. Brazil lost much of its cereals industry to Argentina under Finally, this chapter looks at several types of jobs that Mercosur, and its manufacturing industry suffered more have emerged over the period of greater integration, in- generally. Mexico also experienced a large decline in cereal cluding jobs in EPZs, tourism, teleservice industries, and and animal production and substantial increases in all types nonagricultural exports. of manufactures. Costa Rica lost much of its labor-intensive manufacturing processes to Mexico after NAFTA, but also Employment saw a substantial increase in manufactures of computer Several LAC countries experienced high unemployment in chips. (See Box 5.1 for a review of the "original" NAFTA.) the 1990s. What is far less clear is how much unemploy- In each case, substantial numbers of workers lost their jobs, ment can be attributed to greater openness, and how much and some experienced either very long periods of unemploy- is permanent. ment, large wage losses, or both. The 2000 LAC regional To be sure, part of this unemployment must be laid at publication, Securing Our Future in a Global Economy, made the door of increased international competition. As Figure the case for why an improved safety net is necessary to 5.1 shows, the region has seen very substantial reallocations moderate the high costs of these adjustments. of its workforce. Argentina lost much of its automobile However, these dislocations, though extremely costly in industry while seeing an expansion in more sophisticated human terms, are transitional, and do not imply perma- 121 FROM NATURAL RESOURCES TO THE KNOWLEDGE ECONOMY TRADE AND JOB QUALITY FIGURE 5. 1 Composition of Labor Force (AII Workers) Argentinaa Brazil Chemicals Chemicals Machines * 1 8 Machines _999 Capital Incensive * . Capital Intensive Labor Incensive = Labor Intensive Cereals Cereals * l Tropical Tropical Forest Forest Materials Materials Petroleum PetroleLim Animals Animats II 2 s 6 8 ( S 1 5 Percent Percent Costa Rica Mexico Chemicals Chemicals l Machines j 1992 Machines 1992 Capital Intensive Capital Intensive Libx)r Intensive : - Labor Intensive Cereals Cereals Tropical Tropical Forest Forest Materials Materials PerroleLim Petroleum Animals _ _ Animals 2 4 6 8 101 0 1() 21 Percent Percent a. Urban workers only Sw,rie: National Labor Force Surveys. nently higher rates of unemployment. In the long term, omy, experienced double-digit rates of unemployment for equilibrating forces bring demand for workers in line with several years following liberalization, but from 1986 i-o supply. Cross-country comparisons can be seen as approxi- 1997 had among the lowest rates in the region until the mating the "long term," and statistical analysis suggests Asian crisis severely depressed its export market in 1998. In that, globally, there is no relationship between openness and fact, Figure 5.4 suggests that after a period of "shaking out" unemployment (see Figure 5.2).l Nor is there any relation- from the recession of the early 1980s, employment grew ship between resource intensity and unemployment (see steadily from 1983 to 1995. Similarly, Mexico's present rare Figure 5.3). Even within the region, as Table 5.1 shows, of unemployment is roughly at its traditional level, despire unemployment rates in many liberalizing economies are not a dramatic integration with the United States. high by historical standards (with the notable exception of Further, the observed high rates of unemployment in Argentina). More specifically, Chile, the LAC country that some countries partially reflect the unfinished business of is the most advanced and most integrated in the world econ- macrostabilization. No one following the debate over I'' HOW 'WE WORK JOB QtTALITY IN EMERGING SECTORS whether Argentina should attempt a depreciation of 80X 5.1 Losers and Winners in the Orignal NAM its currency can attribute the present economic situation entirely or even primarily to liberalization. The challenge of The first North American Free Trade Area was estab- bringing inflation down from triple-digit levels and estab- lished by the U.S. Constitution, signed in 1789, which lishing responsible and credible fiscal and monetary policies forbade barriers to trade among states. Many of its has led to multiple stabilization plans and striking volatil- signers were farmers who effectively unemployed their ity in wages and employment over the last 20 years in grandchildren. As an example, the town of Maynard, Argentina, Brazil, Mexico, and Peru. Nor can Colombia's Massachusetts, was primarily agrarian at the turn of the present unemployment rates be attributed primarily to lib- 19th century, but its relatively poor and rocky soil eralization. The financial crisis, combined with ill-advised made it no match for the imports from the new terri- minimum-wage policies, has created unnecessarily high lev- tories in the Midwest. Its economic base would be els of unemployment, exacerbated by rapid migration from reborn in the form of textile manufacturing, which violence-torn rural areas. Again, the two countries that have would again be subsequently overcome by cheaper gone the farthest in opening, and that have re-established imports from the southern states. Maynard is now bet- solid macrofundamentals, Chile and Mexico, have shown ter known as the birthplace of Digital equipment, normal rates of unemployment by historical standards. which was one of the founders of the Rt. 128 technol- ogy oom nd wichmadeits eadqartrs i theoldFinally, unemployment is a function of both demand for mill buil andings. h made its headquarters in theold workers and supply. As Table 5.2 suggests, supply has risen minl eachiincarntiogfamerswen.bakrut, ex- substantially in Argentina, Bolivia, and Chile-by around 3 tIeaworkersnlost atheirjobs fandmth subseqent diskrupt - percent-and if figures are to believed, in Peru-by almost tileworkcuse unstol thumrjobs,and thardshiButint rrosect, 10 percent. Such increases require substantial increases in tions caulsed untold human hardship. But in retrospect, jbceto omiti osatuepomn ae.Fr would it have been good for Maynard workers to pro- job creation to maintain constant unemployment rates Fur- tect New England agriculture from "foreign" competi- ther, the increase in women's participation was especially *on from other states of the Union? large, ranging from 5 to 10 percent across the 1990s, and across most household structures (Figure 5.5). The concomi- FIGURE 5. 2 Unemployment and Openness, 1998 3.o - 0O LAC Countries Cl Rest of the World 2.5- - 2.0 - 1.5- 0 a Nicaragua 1.0 - Honduras Costa Rica, Barbados - n Tbg W ~ ~ ~~~~~~~~~m Trirnidad and Tobago Oom-s,s 0Ec*ador a Panama 0.5 - D- Chic loa * Urog-asP Ven-zuola, R B d Argn- n om I 0.0 - I lI I I I I 0 2 4 6 8 10 12 14 16 1S 20 Unemployment Rate (Percent) Note: Openness is measuredI as X + MIGDP. Swine: WSr/,/ Dere/oprcnt lndlzflrr,tors ?(I . t23 FROM NAIURAL RESO[I'CES TO THE KNOWLEDGE ECONOMY: TRADE AND JOB QUALITY FIGURE 5 3 Unemployment and Resource Intensity, 1998 8 _ | O0 LAC Countries U Rest of the World 6- -t - Venezuela, R B. dr c/ _ Mexico - adura Costa Rica pChlle Brazil tTrugy Ecuador Argentina Trin:da,g and Tobago W () _ ~ ~ ~ {11 *m - U3 - - Nicarogua bPanama -a~ :4 -2~~~~~~~- i 2 - egs B Ei -tU - Barbadosl 6 8 - ~~~~~~I I I IIIIII 0 2 4 6 8 10 12 14 16 18 .0 Unemploymenr Rate (Percent) Note: Resource intensity is measured by net exports per worker of nonmanufactrued goods. Source: World Developnment Indsicato,r 2(001. tant 5 percent average increase in women's wages relative to Wages and Distribution men's wages may reflect that their rising rates of participa- As important as the number of jobs created is the quality tion have led them to be younger and their human capital of those jobs. The most straightforward measure is tli-e on average to be of higher quality than that of men, and that wage a worker receives, both what it can buy and how it there are fewer who became obsolescent during restructur- compares to other wages in the economy; that is, the wage ing (Figure 5.6). However, as will be discussed in the last distribution. section, it may also be due to the booms in the industries As Stallings and Peres (2000) and Weller (2000) note, that have emerged in the region that have tended to reward real manufacturing wages grew strongly from 1991 to 1998 the characteristics associated with being female. in Bolivia (37 percent), Brazil (24 percent), Chile (29 per- cent), Costa Rica (23 percent), and Jamaica (70 percenti. TAbLE 5.1 The stagnation in Argentina (-2 percent) and Mexico (-2.5 Average Unemployment Rates in LAC percent) again must be largely blamed on difficulties in (Percent) macrostabilization in the Argentine case, and the Tequila Crisis in Mexico. Overall, they see the reforms as having a ENTIRb COUNTRY 197'(s 1980s 1990s PERIOD positive impact on wages. Argentina 4.1 4.8 12.1 7.0 Looking at a sectoral level reveals two important find- Bolivia 6.3 7.7 4.9 6.3 2 Brazil 6.6 ¶.o 5.6 ¶9 ~~~~~~ings . First, though survey data on rural and natural resoun:e Brazil 6.6 5.4 5.6 5.9 Chile 10.5 14.4 7.1 10.7 sectors, especially mining, are often weak or unrepresentai- Colombia 9.6 11.3 11.5 10.8 tive, Figures 5.7 to 5.10 show both raw wages and wages Costa Rica 5.0 6.9 5.3 5.7 Guyania 0.5 (1.3 - adjusted for human capital and by Leamer categories. Some Honduras - 4.5 5.8 5.1 striking patterns emerge. Leamer's classification of manu- Mexico 7.0 4.7 3.7 5.1 Paraguay 7, 1 1.8 6.3 8.5 factures by increasing sophistication-labor intensive, cap>- Peru :/4 7.4 8.5 7.8 ital intensive, machinery, and chemicals-is reflected in Uruguay 9.2 I11.6 9.9 9.9 VenezuLela, R.B. de 5.7 8.8 11.3 8.3 increasing levels of wages, and wages adjusted for human- LAG--Mean 6.6 7.6 7.6 7.3 capital. But equally important, it is not obviously true that ILAC--A4ednan 6.S 7.4 6.7 7.() these sectors offer higher wages than petroleum, mining, - Nor avala-ble. Soone: Dr F-reaurti arid odh-es 2000 forestry products, or even tropical agriculture. This is prob- 12 4 HOW WE WORK JOB QUALITY IN EMERGING SECTORS FIGURE 5.4 Chile: Steady Manufacturing Employment Growth after Refonns 18()- 170- otal Output/Total Employment -_.Output Index, 16() ~~~- * ^ Employment Index,- 156- 10- -' 120- 90 90 - 1979 1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 Note: 1,501 Chilean firms, 1979-95. For 1995, only employment index daca were available. source: Chitean manufacturing survey. TABLE 5.2 ranked by their degree of exposure to international compe- Labor Force Participation Rates in the 1990s tition. For each country, the bar on the extreme left repre- (Percent) sents the third of workers in import-competing industries that are most exposed, measured as the dollar value of COUNTRY PERIOD FIRST YEAR LAST YEAR CHANGE Argentina 1991-97 59.8 62.6 2.8 imports per worker. The next bar captures the third of Bolivia 1989-96 52.8 56.4 3.6 workers progressively less exposed, and so on. The right Brazil 1992-97 61.5 60.1 -1.4 three bars capture increasing exports per worker moving Chile 1990-96 51.6 54.4 2.8 right. The middle bar corresponds to workers in nontraded Colomnbia 1988-95 56.8 58.1 1.3 Costa Rica 1990-96 53.5 52.2 -1.3 industries such as services or commerce. What jumps out Jamaica 1989-96 69.8 68.4 -1.4 imdaeyi rjptenta mre nAgnia Mexico 1991-96 56.6 554 1.8 immediately is a U orJ pattern that emerges in Argentina, Peru 1991-97 58.1) 67.8 9.8 Brazil, Costa Rica, the Dominican Republic, and Mexico, S(,ura Weller 2)00. where those industries most exposed to competition pay higher wages. Figure 5.10 shows that, even adjusting for ably not too surprising. In the long term, labor mobility levels of human capital, education, and experience, this will equilibrate total remuneration among sectors, and pattern remains and in fact becomes more striking. those that cannot pay as well will shrink and lose workers. One explanation is consistent with the theme of this And, as Chapter 3 notes, natural resource-rich sectors have report: it is not only what is produced, but how. As Chap- the potential to show productivity gains as high as those in ter 3 discussed, external competition stimulates the adop- manufacturing, and hence there is no reason for these sec- tion of new technologies and these, in turn, require train- tors to be in decline or to pay less. Again, Australia and the ing and skill upgrading. As only one example, in Mexico Scandinavian countries still have large fractions of their the tripling of manufacturing exports observed during the workforces in primary product-intensive industries and 1990s has been associated with increased rates of adoption enjoy extremely high standards of living. In sum, as a first of modern production technologies, an acceleration of pro- cut, reorientation toward products with a higher natural ductivity growth, and a relative increase in the demand for resource content does not obviously imply worse jobs. skilled labor. Lopez-Acevedo (2001a) shows that between Further, exposure to global competition more generally 1992 and 1999 the rates of increase in the adoption of incrcases wages. Figure 5.9 looks at wages in industries modcrn manufacturing technologies were not only higher 125 FROM NATURAL RESOURCES TO THE KNOWLEDGE ECONOMY TRADE AND JOB QUALITY FIGUtRE 5.) Change in Women's Labor Force Participation Rates by Household Structure 40 * Sinigle, No Children 3( - E Single, with Children F-GI REI 536 Chil 1() - _ -20 I ArgenLina Chile Colombia Costa Rica El SalvaduE Honduras Mexico Peru Venezucla LesocL: Cunningham 21101. FIGIt RE 5.6 Change in the Ratio of Women's Wage to Married Men's Wage N -311 ~-20 -11) 0 20) 31 40 50 6(0 Percenti SOw~,cc: CLinningham 20)0 1 amon-g exporters than nonexporters, but also increased atso find higher rates of both human capital and skill build- more rapidly. Audretsch and Lopez-Acevedo (2001) also ing prior to entering export markets, and find a clear assco- show that the modernization of production processes dur- ciation between plant-level efficiency (TFP) and exporting. ing the 1990s was associated with an increase in employ- To the degree that these additional skills are nor captured ment, labor productivity and, at least during the second by the fairly standard and imprecise human capital vani- half of the decade, higher wages. Hallberg and Tan (1998) ables, these more open industries will appear to pay higher 1216 HOW WE WORK JOB QUALIITY IN EMERGING SECTORS FIGURE 5.7 Current Raw Wages by Learner Category 1,200 U Argentina' JBrazil 1,000 ____ Costa Rica 1 Mexico J Dominican Republic k 800 _E_ r -1l Venezuela, R.B. de 6oo- 400- u 2001 - -L Animals Petroleum Materials Forest Troptcal Cereals Labor Capital Machines Chemicals Nontradeablc Intensive Intensive a. Urban workers only. Source: National Labor Force Surveys. FIGU.RE 5.8 Wages by Leamer Category Adjusted for Human Capital 200 - Argentina" 150 __ Z 1 00 - - _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _C _ _ __ C o sta R ica z . | ~~~~~~~~~~~~~~~~~~~~~~~Mexico <, 5 ~~~~~~~~~~~~~~~~~~~~~~~~~Dominican Republic : lz)o -_ | _ N _ 2 ~~~~~~~~~~~~~~~~~~~~Venezuela, R.B. de -50 Animals Petroleum Materials Forest Tropical Cereals Labor Capital Machines Chemicals Intensive Intensive Leamer Categories a Urban workers only. S'cssrce: National Labor Force Surveys. wages. Further, these industries may also pay higher than nologies. In the long term this is the only possible path to market-clearing or "efficiency" wages to hold onto workers increasing wages, and job quality more generally. who they have trained.3 Again, these results are entirely consistent with the Inequality insistence of Swedish trade unions on maintaining an open The fact that more exposed sectors may be demanding trade regime and on constantly upgrading production tech- higher skill levels is frequently cited as a possible explana- 127 FROM NATURAL RESOURCES TO THE KNOWLEDGE ECONOMY TRADE AND JOB QUALITY FIGURE 5.9 Current Raw Wages by Trade Exposure 700 - 600 -_ High Import l- 600 - N \i. - [jL ,,500 _ - * N..-.r,I. !-. 1-l . 1 c 111~~~~~~~~~~~~~ 1. 1 .;,,1 400- 300 - 200 C 70() _ _ * 1- = S i-- 100 0II Argentina Brazil Costa Rica Mexico Dominican Republic Venezuela, RB. de Note: Nonagriculcural workers. Means by trade balance per worker centiles. Sourc,: National Labor Force Surveys FIGURE 5 10 Wages by Trade Exposure Adjusted for Human Capital 90 '.' 8() - _ = * ~~~~~~~~~~~~High Import 7I) Medium Import ' 41t)-* 10~~~~~~~~~~~~T1 Arguntina Brazil Costa Rica Mexico Dominican Republic Venezuela, RB. de N')te: Nonagricultural workers. Conditional on human capital. Swosrce: National Labor Force Surveys. tion for the apparent trend toward greater wage inequality counter to what many expected. The Stolper-Samuelson observed after liberalization in Argentina, Chile, Colombia, theorem, one of the four central results of the Hekscher- Costa Rica, Mexico, and Uruguay.4 Ohlin-Samuelson framework, suggests that trade liberal- These movements were not necessarily permanent. ization would stimulate the production of goods that used While the ratio of the wages of the highest- to the lowest- Latin America's presumed abundant factor of production, paid workers increased in the 1980s in Chile and the mid- unskilled labor, and raise its relative return. As Box 5.2 1990s in Mexico, as Figure 5.11 shows, it had largely shows, the logic behind the Stolper-Samuelson theorem is reverted by the early and late 1990s, respectively. Nonethe- very simple, and therein lies both its appeal and its many less, these movements immediately following reform run limitations. 128 HOW WE WORK- JOB QUALITY IN EMERGING SECTORS FIGURE 5.11 Wage Inequality Spread (90th over the 10th percentile) Mexico (1987-99) 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 6 - I I I I I I I I I I I I I I I i I I 3 Ue 3- \ ~~~~~~~~~~~~~~~=g_~~~N, 2.5 4 3r 422- l 1- O.1.A\-C 5 -5- ~~~~~~~~~~~~~~~~.- ~~~~~~~~~~ - . ~~~~~~~~~~1.5 3 +~~~~~~~~~~ Chl - 3- Meic 1- ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~0.5 - Chile Mxc 0 - I I I I I II I 0 1970 1972 1974 1976 1978 1981) 1982 1984 1986 1988 199( 1992 1994 1996 Chile (1970-96) Note: Chilean upeness began in 197i but increased during the 1980s. Mexicani liberalization began in 1985 anid deepened with NAFTA (1994). In particular, two assumptions do not correspond to LAC's and Lopez-Acevedo (2001b) find that Chilean and Mexican reality. First, there are no intermediate factors or mobile fac- firms with greater access to technology have, overall, a tors of production in the Hekscher-Ohlin-Samuelson frame- higher demand for skilled workers. Dramatic increases in work, yet one of the motivations of the trade reform was skilled and unskilled labor in the industrialized world have exactly to give firms access to better capital and intermediate been almost entirely attributed to technological progress goods. This could potentially have two important effects. and very little to trade. Sanguinetti and others (2001a, First, as Jones (2000) argues, if factors of production are 2001b) find that import penetration increases inequality in mobile, the traditional Stolper-Samuelson results can be Argentina, although Galiani and Sanguinetti (2000) find overturned under reasonable conditions. But as important that this can account for only a small share of the increase. from the point of view of this report is that mobile capital, or Santamaria (2000) find a similarly small effect for Colombia. the technology it embodies, may be skill biased, implying Pavcnik, Blom, and Schady (2001) show that in Brazil, after that it must be combined with higher skills, and this could controlling for workers' personal characteristics, wages rose also reverse the expected effect. Feenstra and Hanson (1995, in the industries where tariffs experienced larger reductions, 1996), for example, show that, though the United States but they do not find any correlation with inequality. Simi- moved their least-skill-intensive part of the production larly, Sanguinetri and others find no connection in Uruguay process to the Mexican mnaquila sector, nonetheless, these Most recently, Behrman, Szekely, and Birdsall (2001) find processes were skill intensive hy local standards. that for 18 Latin American countries during 1977-98, tech- Mfore generally, technological change, however induced, nological progress, rather than trade flows, appears to be a appears to increase the demand for skilled workers. Looking channel through which reforms affect wages (see Box 5.3). at Colombia, Mexico, and Taiwan (China), Tan and Batra These effects are, in reality, hard to separate. First, since (1997) find that firms that invest in research and develop- Latin America is not a major source of technological inno- inieut or traininig of workers tenid to have wider spreads vationi, mllost is "imported." Second, most authors acknowl- between skilled and unskilled workers than firms that do edge that any sector adjusting to increased trade exposure not. Consistent with studies of the OECD, Pavcnik (2000) will displace workers to all other industries, not just those 129 FR(ON[ NATITRAL RESOURC ES TO THE KNW)LEVDGE ECONOMY TRADE AND JOB QUALITY BOX 5.2 SkoerSanelsoe : Why We Thon ske Wettems Shoud Beneftm Jbakaw The logic behind the Stolper-Samuelson theorem is sim- umented in this report, Latin A-merica does not export ple. Suppose we have two goods-wine, where we use labor-intensive manuactures in general. Compared to two units of labor for each unit of capital, and automo- the OECD, the region is unskilled-labor abundant, but biles, where we use one unit of labor for each unit of cap- compared to Asia, it is not, and hence the Stolper- ital. And suppose that before liberalization, cars were Samuelson effect may simply be reflecting that the prod- heavily protected and wine suffered negative rates of uct whose price is rising is not a labor-intensive good, protection. This was more or less the case in Chile in but a resource-intensive good. 1967. Second, Stolper-Samuelson is fundaamentally a long- Now we liberalize, which lowers the price of cars and term story where factors are fully mobile among indus- raises the price vintners get, reducing the productivity of tries. This leads to the counterintuitive finding that labor cars and raising the production of wine. If we produce and capital should never be on the same side of a protec- one less car, this frees up one unit of labor and one unit tion dispute: if capital is intensive in the industry, then of capital. To produce one more bottle of wine takes that more protection will help capitalists and hurt labor. Yet one unit of capital, but two units of labor. This means in real life we see both owners and unions resistant to that there is excess demand for labor, and the wage must lowering protection. This makes much more sense when rise relative to capital. The logic is the same when the we realize that in the short to medium term, both groups two factors are skilled and unskilled labor. are tied to the sector, either by invested capital or specific There are several reasons why the predicted result may skills, and hence are not mobile. Decreasing protection not have occurred in LAC beyond the two discussed in on existing unskilled-labor-intensive products will lead the text. First, as Leamer has noted, and as has been doc- to a decline in their wages relative to other factors. BOX 5.3 W bi TechmeIgkai Pmgress Usher In a Golen Ae of Equa.? From a global perspective it is not necessarily obvious of the early 19th-century weavers, whose incomes soared that, if -:chnutlo i i.al progress tends to reward more after the mechanization of spinning, only to crash when skilled workers, then relatively unskilled latin American the technological revolution reached their own craft. workers continue to fall behind. Krugm.in (1999) argues I suspect, then, that the current era of growing in- (as does Landes [1998J) that, in other historical episodes, equality and the devaluation of ordinary work will turn technology has displaced the relatively skill endowed as out to be only a temporary phase. In some sufficiently well, and he sees no reason why it should not be 'college long term the tables will be turned. Those uncommon i:radu.itet saving in the future. skills that are rare because they are so unnatural will be "So here is a speculation. The time may come when largely taken over or made easy by computers, while most tax lawyers are replaced by expert systems software, machi nus will still be unable to do what every person can. but human beings are still needed-and well paid for In other words, I predict that the current age of inequal- such truly difficult occupations as gardvning. house ity will give way to a golden age of equality. In the very cleanine, and the thousands of other services that will long n, of course, the machines will be able to do every- receive an ever growing share of our expenditure as mere thing we can. By that time, however, it will be their consumer goods become sread iy cheaper. The high-skill respuinsibi1t } to rake care of the problem." professions whose members have done so well during the last 20 years may turn out to be the modern counterpart swrc. Kaarn.r. d 19cl: Z03). HiOW WE WORK JOB QUALITY IN EMERGING SECTORS affected by trade, thereby muddying any statistical effect of behind. This asymmetric push in education altered remuner- trade. ation of education in the labor market in two ways. First, as What we can say, however, is that technological progress illustrated in Figure 5.12, the wage of unskilled workers is desirable and hence, in the short term, there may be a (defined as workers with no school diploma) experienced a short-term tradeoff between good jobs that are generated in relative wage increase compared to medium-skilled workers modernizing industries, and wage distribution in econo- (defined as workers with complete primary or secondary edu- mies with poorly educated work forces. Countries want to cation). The relative wage increase coincides with a stable create jobs that require higher levels of schooling and hence decline in the relative supply of unskilled labor, Second, as the likelihood of training, continual career development, depicted in Figure 5.13, the wage of highly skilled workers, and worker satisfaction. As discussed above, even in the defined as workers with tertiary education, increased. Mexican mia quilas, the newly created jobs demand relatively Blom and Velez (2001) estimate that around 60 percent higher levels of skill than that possessed by the workforce. of the increase of the skill premium to tertiary education Though no one would want to further "dumb down" can be attributed to the asymmetric expansion of the edu- n'aquila employment, this implies that relatively skilled cation system, and only 40 percent is due to a shift in labor workers will see a rise in their relative earnings. In fact, as demand toward highly skilled labor. Since they find no Venables (2001) notes, the widening differentials appear- structural breaks in the behavior of demand over the last ing over the last decade offer the right market signal to two decades, even this 40 percent appears unrelated to young people to accumulate higher levels of human capital. trade liberalization per se. Overall, the reduction of returns to secondary education drove the overall fall in inequality Labor and Human Capital Supply: The Missing Link despite the offsetting effect of rising returns to a college In fact, the deficient supply of skilled workers in the work- education.5 force is a critical part of the distribution story and poses a Similarly, V6lez and others (2001) argue that during the major policy challenge to the region's governments. In 1990s, the increased participation by women was partially Brazil, large shifts in educational composition have taken responsible for the worsening wage distribution in Colom- place in the last two decades and this has had an ambiguous, bia. Generally, women's skill levels tend to be lower than although large, impact on relative wages. The primary and average and hence they depress earnings to lower-skilled secondary education system expanded substantially, increas- workers. Given the large increases in female participation ing mean years of schooling from 4.8 in 1981 to 6.9 in 1999 across the region, this is likely to be a more general phe- and decreasing the Gini Coefficient of years of schooling from nomenon. However, it may also offer an explanation for the 0.49 to 0.37. However, the teruiary education system lagged fact that there does not seem to be a corresponding increase PIGURE 5 12 Relative Labor Supply and Wage of Workers with No Schooling Compared to Workers with Lower Secondary Schooling in Brazil 3.5- .41 > 5 | _B-- Relative Supply (Left scale) -- Relative Wage (eight scale) | 5 - ~~~~~~~~~~~~~~~~~~~~~~~~~~~39 seC c . N. ~ ~ ~ ~ ~ ~ ~ ~ ~~~~~~~~~~~~~~~-37 12.5 5 c-u U ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~~~~~~~~~~~~~~~~~~~-33 - -~~~~~~~~~~~~~~~~~~ 2.0 ----w-31 S ~-.. ~ 1.5 X.~~~~~~~~~~~~~~~~~~~~~~~ 27 1.0 25. 198 1985 1988 1992 1996 1999 Nate: Surveys were unavailable for 1984, 199 1, and 1994. Source: Velez and others 2001. 131 FRO)M NATURAL RESOURCES TO TILE KNOWLEDGE ECONOMY TR ADE AND JOB QUALITY FIGURE 5.13 Relative Labor Supply and Wage of Workers with Tertiary Schooling Compared to Workers with Upper Secondary Schooling in Brazil 6.0- -280 be - | - -U- Relative Supply (left scale) ~-- Relative Wage (righ[ scale) S -2 S c2 u 5.8 270c .~a 5.6 4. 2601 - 4.6 '1() co~~~~~~~~~~4 ; 5l" - 250 X a a. 5~ -2. 240 4.8 220~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~4 4.6- I 210 1981 1983 1986 1988 1990 1995 19)97 1999 Ns)t: Surveys svere unavailable for 19S4, 1991, and 1994. S,/urt: Velez and others 200L. in per capita houisehold inequality over the 1990s at the coun- earnings show roughly twice the inequality as those of try or regional levet (Wodon and Yitzhaki 2001; Stallings salaried workers. This is due to the riskiness of self-employ- and Peres 2000). Though the rising participation of women ment (see Wodon, Maloney, and Barenstein 2000), and the helped worsen the wage distribution of individuals in Colom- higher measured returns to both education and experience bia, this represented an additional earner in poor households, among the self-employed>' This means that the increase in and hence reduced overall inequality. self-employment could also have an impact on measureid The issue of migration spans both international move- inequality. ments in factors and human capital supplies. Though the In sum, the emerging picture of the evolution of wages report does not examine the issue in any depth, Box 5.4 and distribution is not bleak. The sectors more exposed to suggests some of the channels through which flows of trade appear to pay more than those in the less-exposed stc- remittances, which amount to large fractions of export rev- tors. Further, the observed worsening distribution of wages enues, affect both distribution and poverty. seems largely due to the fact that firms are adopting te(e- nologies and more efficient modes of production and, in t.e Sectoral Reallocation process, are creating jobs that demand more from, but are A final consideration that leads us into the next section is also more satisfying to, workers. that the expansion of nontradeables sectors, and the con- comitant expansion in self-employment in the 1990s, may Informality also have had adverse impacts on measured inequality An emerging preoccupation in the region is that trade lii- among income earners, although not necessarily wage earn- eralization has contributed to the rising number of "infor- ers. This could occur through two channels. First, the trade- mal" workers-those unprotected by social security and ables and nontradeables sectors may have differing demands health benefits. As Stallings and Peres (201)0) of the Centro for skills such that the shift toward nontradeables observed Estudios para America Latina (CEPAL) noted, in the 1991)s across the period would affect wage distribution. San- six of every 10 jobs in the seven countries surveyed7 were guinetti and others (2001a) find that controlling for this created in the informal sector, suggesting a substantial ero- shift lessens the direct impact of import penetration on the sion of job quality." skill premium, but directly contributes to it in Argentina. 'I'he channel through which the increased openness may They find no statistically significant effects for Uruguay. have had this impact is not clear. At one extreme, those Second, the nontradeables sectors have a higher fraction of opposing greater integration into the global economy argue the self-employed, and as Figure 5.15 shows, self-employed that global manufacturing networks seek to control costs 1.32 HOW WE WORK JOB QUALITY IN EMERGING SECTORS BOX 5.4 Migration, Remittances, and Their limpications In many cuuntrries. labor has becorne an important export, the reverse is tme in the urban areas (Gonzalez, Koning, and its associated remittances amount to an important and Wodon 200la). share of poor household income. In El Salvador, 23 percent An often-overlooked source of investment capital for of the work force is estimated to migrate, and in the microenterprises are remittances from relatives working Dominican Rcpublic and Mexico the rates are above 12 abroad. In Mexico, Woodruff and Zenteno (2001) find percent. Remnittances, which have been growing at double- that within the 10-state migration region, more than digit rates, are equivalent to nearly one third of the a third of invested capital is sourced from flmih re- region's foreign direct investment and exceed othfcial de- mittances, and that in total, remittances add to about a velopment inflows. As shown in Figure 5.14, remittances fifth of aggregate investment capital in urban Mexican account for 10 percent of GDP or 20 percent of exports microfirms. In El Salvador one-quarter of remittance in m .n% countries. They exceed total coffee exports in receivers have established small businesses as a result of Colombia and tourism in Mexico (Lpez-C,61ix 2001). the availability of remittances (Lopez and Seligson 1 irI-I. Remittances have profound effects on all elements of Moreover, migration networks may act as a source of the receiving village. They constitute an important social information, support, and cooperation that reduces risks safety net in natural disasters (Meyers it )'Y; and have and increases rewards. In this case, those with greater important impacts on household budgets. In the late access to remittances are also likely to have greater access 1980s, a special survey by CEPAL found that in El Sal- to formal credit and to significantly higher chances of vador, Guatemala, and Nic.ara,gua. 4 to 8 percent was spent firm entry, survival, and expansion. on children's health and education, 5 to 6 percent on Studies on the Dominican Republic and El Salvador investment, and the remainder on c:urnumpr ion (CEPAL also suggest that migration also affects consumption 2001). A 1997 study by FondoMicro in the Dominican habits, and may create negative attitudes toward work. Republic reports that 84.9 percent of the received funds In both countries, most household remittance receivers went toward txpenur- of the family, while 7.1 percent went (mainly composed of women, children, and parents) de- toward the business, and 8 percent went toward other velop a dependent subsistence culture: they receive just expenses. The impact on inequality can differ by region: enough money to provide for basic needs, but not enough Since those migrating from rural areas are often those with for investment (Georges 1990; Lopez and Seligson 1990; the highest human capital, inequality may increase, and and Funkhouser 1"'Q2 FIGURE 5.14 Remittances as a Slae of GDP and Exports, 199 Brazil Colombia Domninican Republic Ecuador El Salvador Guatemala Haiti Honduras Jamaica _ __ Mexico Nicaragua - Peru _/ 0 20 40 60 140 160 Percent Source: L6pez-Clix 2001, World Bank 2001, and Meyers 2001. 133 I R()M NATU RAL RES)IR tRCLS 1O) TIIF K \NOWLED CF E (()N)NIY TR ADF AND] JOB Qt A LITX Fi-IUPRE S . l Eamings Inequality Decomposition for Salaried versus Self-Employed Workers, 1995 Arguntina _ | S,ilarit l F] Self-EmphleJvt Bolivi.l Chile CtA,inbia ______ _-- Uru~g,.ayI \Vt tzLueli. R.B. JCe _ _1 I.2 l.-) II 02).8 I I Theil Index W(id.Dn: nd a, Malonuv,.and Barensrelit ()()( by reducing worker benefits: large- and medium-size firms three emerging stylized facts about the sector. First, as Box facing international competition, or multinationals them- 5.5 suggests, the informal self-employed in Argentina, Mex- selves, outsource production to informal small firms, in- ico, and Paraguay report being largely voluntary. To say vol- cluding home-based and self-employed microentrepreneurs untary is not to say well-off or even happy, only that in a (Portes, Castells, and Benton (1989); Borrus and Zysman market without segmentation, this option was deemed b(et- 1997). In a sense, this view takes the fragmentation of the ter than any alternative job in the formal sector. Second, the production process down to the microfirm level, but with a vast majority of informal jobs, generally arouncd 70 to 8() very negative twist: informal firms get their comparative percent, operate in nontradeables sectors, such as comn- advantage by avoiding worker protections and taking on merce, transport, and services. Third, the sector is heetro- the risk associated with demand volatility. geneous, including both voluntary members who would not The evidence offers little support for this view This sec- be made better off in formal sector jobs, and involunta-y tion argues that the increasing informalization of the region members who are effectively "rationed" by higher th.in is not primarily related to trade liberalization per se, but market-clearing wages due to minimum-wage laws or oth,;r rather to the increasing attractiveness of nontradeables em- rigidities." This view encompasses both traditional dtualis- ployment, which itself was partially related to macrostabi- tic views of segmented labor markets and newer views )f lization and capital inflows. In some cases it appears to integrated markets where the share of voluntary members is resolt from increased rigidities in labor markets and tax determined by how binding formal sector rigidities may be. regtIlations. Further, those few microfirms linked to larger These features suggest that, as a first approximation, the firms either through subcontracting or other market trans- informal self-employed could be considered to be unregL- actions tend to be both better paid and more formal. tated entrepreneurs in the nontradeables sector. This, in Arguably, the larger challenge is how to reduce the isola- turn, has implications for how we view the trends over thie tion of informal microfirms and help them to take advan- last decade. First, as Balassa (1964) noted, the size of the tage of the emerging opportunities. nontradeables sector is largely determined by its produ(,- tivity compared with the tradeables sector. Productivity The Nature of the Izformal Sector: gains in the formal sector both increase the attractiveness of A Comprehensive Approach formal sector jobs relative to being self-employed, and raise A more general approach is necessary to understand the the relative cost of nontradeable goods, reducing their dynamics of the informal sector, which can incorporate demand. This relationship is borne out in Figure 5.17, 134 1lHOW WE WORK JOB QUAL11Y IN FMERUING sE(CTOItS BOX 5.5 Are Most lnfonnal Sector Workers Voluntary? Evidence from Argentina, Mexico, and Paraguay suggests Why Would Workers Prefer to Be "Unprotected"? that most people in the informal sector are there volun- There are several reasons why workers may be willing to tarily. In Mexico, 70 percent report entering the sector voluntarily become "unprotected" and rely more on in- either because of more flexibility, higher wages, or sim- formal safety nets. First, developing-country microentre- ply the desire to be their own boss. Panel data suggest preneurs may not be fundamentally different from their that of those workers who started in the formal salaried counterparts in the industrialized world who also take sector but moved into self-employed 15 months later, on responsibility for medical insurance or saving for re- two-thirds report moving voluntarily, citing a desire for tirement that was previously covered by their employers. greater independence or higher pay as the principal Second, in a market with flexible wages, the cost of motives (see Figure 5.16). These findings are consistent employer-provided benefits is partially passed down to with the sociologists Balan, Browning, and Jelin's (1973) workers in the form of lower wages. If workers do not value extensive interviews with Monterrey workers who state the benefits as much as the decline in wages, they will seek that being one's own boss was well regarded and that out jobs in the unregulated sector where remuneration is movements into self-employment from salaried positions entirely monetary. This might happen if social security or often represented an improvement in job status. This is health benefits provision is very inefficient, if there are very broadly consistent with interview data from Argentina: a weak linkages between benefits and contributions, or if one small survey in the province of Jujuy revealed that 80 member of the farnily is already formal and hence the rest percent of the self-employed had no desire to change of the family already receives benefits. Third, the interviews jobs, and under 18 percent viewed self-employment as a of Balan, Browning, and Jelin (1973) suggest that the very temporary activity before they found a "real" job. In legislation that is thought to induce rigidities into the Greater Buenos Aires, another survey found that while labor market in fact stimulates such turnover and encour- 36 percent would have preferred to work more hours, ages workers to leave saLaried employment. The paucity of only 26 percent were looking for other work. In openings for promotion on the rigid escalafon and the ceil- Paraguay, only 28 percent of those in the informal self- ing on mobility opportunities for manual workers makes employed sector stated a desire to change occupations. self-employment the remaining outlet for further advance- Among those often thought to be the worst off, informal ment. These last two issues suggest that, in contrast to the salaried workers, the percent rose only to 32 percent. usual view, the extant labor protections may make formal sector work less desirable rather than less attainable. FIGURE 5.16 Mexico-Reasons to Start a Mkcroenterpnse Other Laid off or fired fromwork _ r|-1996 [] 1994 Flexible hours of work_ k_ _ Did not find a salaried job __ - To obtain higher income than as a I J salaried worker To complement family incorne _ Family tradicion _ _l Want to be independent I I 1 1t 0 5 10 15 20 25 30 35 40 Percent MEs Sourne: Maloney 2000 and Sanchez, Joo, and Zappala 2001. )35 FROM NATURAL RESOURCES TO THE KNOOWLEDGE ECONOMY: Ti.ADE AND JOB QUALITY FIGURE ). 1 Self-Employment versus Industrial Productivity 50 _ 415- 40- Bolivia H z El SaI-ador Venezuela, R B. de 35n - -- 0 a' 'a I inndneas Gua-et ala - Uruguay, 3. 0- Aaaa Panama Mtexico a_ CTurke * * Chile treece -El - K-Csta Rica *Koee, Rep de Italy _ o 0-UnitB-il Uned Kingdom CEt Portugal v SpZdainteln rmy aC New Zealand iclnGemy Poland C.:ech Republic II i lgnd * S ,e -t- I (I - *> Ireland Finland Sweden / A S-.ngary Netherlands Aseimar- 1 , L'-b Norway Uo.red St.wes () - l l l l l l l lI 6.() 6.5 7.0 7.5 8.0 8.5 9.1) 9.5 10.(1 1(.5 1 1.( Log of Induscrial Value Added per Worker Scar.e: Maloney 201a. which shows that the large size of the self-employed sector rate is appreciating. This suggests that the dramatic rise in reflects low formal sector productivity, and hence lower informality during this period was driven by increased wages of the region. As the relative attractiveness of formal opportunities in the nontradeables sectors that boomed in sector work increases with the development process and many countries of the region following the liberalization of productivity rises, the share of self-employment will fall. the trade and capital accounts. The case is most suggestlve Second, this characterization of the informal sector also in Brazil, where statistically the real exchange rate, the size implies that the short-term and long-term overall macro- of the informal self-employed sector, and the relative earn- behavior as captured in standard small economy or "Aus- ings can be shown to move together, suggesting that much tralian" models of the economy may offer important expla- of tde large increase in informaliry is volintary.1) nations for the increase in informality we observe in the Third, in the wake of these booms, there does seem to be region (Maloney 1997; Fiess, Fugazza, and Maloney 2001). evidence of an increase in informality concomitant wit!) a Figure 5.18 plots the ratio of the number of formal sector decline in relative informal income in Mexico from 1992 to workers relative to the number of informal workers, their 1994, in Argentina post-1993, and in Brazil immediately relative incomes, and the real exchange rate for Argentina, before the 1998 depreciation, which suggests segmentation Brazil, and Mexico, which together account for 60 percent arising from nominal rigidities presenting difficulties in of the regional population. Much of the regional trend is adjusting to further shocks. As an extreme case, Figure 5.1 9 driven by Brazil, which showed a steep rise in the number shows a massive increase in relative informal self-employ- of workers seni carteira, or without the signed work card that ment in Colombia after the decline of the boom and onset would give them access to benefits. In a traditional dualis- of the crisis in 1998. Unfortunately, the data do not permit tic or segmented view of the informal sector, we would us to separate informal and formal salaried employees for expect to see the two labor market variables move against long periods of time. However, the estimated wage distri- each other: a rise in the formal sector wage, perhaps due to butions in Figure 5.20 suggest that in 1998 Colombia's an increase in minimum wages, forces workers into the minimum wage was far more binding in both the formal informal sector and drives down the informal wage relative and informal salaried sectors than in the other three coun- to the formal. But what is striking is that in Argentina, tries. Further, indexing it to overly pessimistic forecasts of Brazil, and Mexico, these series move together in the begin- inflation led to it rising another 6 percent in real terms in ning of the 1990s, and at the same time that the exchange 1999, despite the worst crisis in the postwar period. 1I6 P ~~~~~~~~Relative Wages Relative Wages Relative Wages C 0~C -1 --I C .Z ' C Ny Z a) ~~~~~~~4 ~ ~ ~ ~ ~ - 'ta- 4~~~~~~~~~~~ A I~~~~~~~~~~~~~~~~~~cC / F~~~~~~~~~~~~~~~ a) B ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~~~~, I ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ t -~~~~~z - . 0,-n .*~~~~~~~~~~~~ -1~~ C .0 [.i-~ ~ ~ ~ ~ ~ ~~ ~~~~~~~~~~~~ ~~~~~*-~~~~~~~ ~~~Z C 0)~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~C - 004 (~~~~~~~~~~~~~~C -1 0 01 1~~ora/nom l aaidFrmlIfra SaaidFrm lIfralSlre FROMf NATU RAL RESOilRCFES 'rO THE KNOIWLEDGE ECION(MY TRADE AND JOB QUJALITlY FIGURE 5.19) sourcing or "gray" area workers. Most compelling is the Relative Formal/Informal Sector Sizes and Income and the case of Mexico. If any country were to be thought of as Real Exchange Rate, Colombia (Informal Selt-Employed/ the low-skilled manufacturing appendage to global capital- All Salaried Workers) ism, it would have to be the United States' southern neigh- 1.4 - Colombia 4. bor. Yet Figure 5.18 does not suggest a permanent rise in informality, either of self-employment or informal salaried l1. - workers. In the first case, we do see a sharp rise in informal 1.2 A 4: I l. ; __ salaried workers across the 1990s, but this is sharply re- 1.1 A ___NAIL--- C- 1.1) J U 81 3.11 E versed after the 1995 depreciation and crisis. Since many Y 15.9> _ . - _ _ |. - 2.5 informal salaried workers are hired by informal microfirms, -c ().9 _ -- | t - -2.5 the boom at the beginning of the 1990s, and the resulting .( 8 X A rationing during the crisis, can explain much of the in- 0.7 - -^+' 4: < E crease. But the level of informality is clearly on trend to (1.6 - '; ,' \~g_ }' _1.5 ' return to 1988 levels in the next year or two, and we see rno () 5 - , - secular trend either in total share of informal salaried work- 0. _ 11 .1.0 ers, or their share in large-firm employment. The macro- .J05 cfrcf,+' .45,4etA '.44' ,<->,«n We9X>>,o1 ,>9Ub7 , 9 seconomic fluctuations driven by attempting to stabilize t he ..2 RA cTh-nge rat- Wage/FKSI -UP rmal/SE peso are far more obvious than any long-term trend that F: Formal might be attributed to NAFTA. SE: Self-employed When we look more closely, we do not see strong evidence Sowine: Fies, Fugazza, and Matoney 2(11)1. of linkages between informal microfirms and multinatioral and large firms. The fact is that the vast majority In sum, a sizeable component of the observed rise in of microenterprise activity remains oriented toward the informality in the beginning of the 1990s may be due to domestic consumer market. In the Dominican Republic arid the booms in the nontradeables sector and the necessary Mexico, almost 90 percent of microenterprises sell their macroeconomic adjustments afterward. goods and services primarily to individuals, with the remain- ing 10 percent selling to firms. Only a small fraction sell Does Trade Liberalization Impact Informnality? their goods or services to international markets, ranging What is more difficult to establish is a direct link to greater from 1 percent in Mexico to 1.7 percent in the Dominican integration in the global economy. It is worth noting that Republic to 4 percent in Ecuador. As Figure 5.23 illustrates, relatively open economies such as Belgium and Singapore, in the Dominican Republic the percentage of firms with and resource-intensive economies such as Australia, Canada, business clientele that might possibly have linkages with for- Finland, Sweden, and the United States, are not especially eign firms has increased but remains low, and in Mexico the informal. Nor is it obvious from Figures 5.21 and 5.22 that number of microfirms claiming to be linked to such firms there is a tight correlation between either openness or par- actually fell from 1994 to 1996. While this may reflect the ticular types of exported goods. For Argentina, Brazil, and temporary effects of the crisis, this nonetheless makes it even Mlexico, the familiar U-pattern appears with more exposed more difficult to argue that there has been an increase in our- industries having relatively more formal work forces. Fig- sourcing since NAFTA was passed. In fact, the higher rat's ure 5.22 suggests that most of Leamer's export categories of informality among microenterprises may be more due t-o have higher rates of formality than nontradeables sectors, a shift toward the nontradeables service areas, which usually although animal and cereal production appears in general to show lower degrees of formality. be among the least formal. Certainly it would not be possi- The difficulty in pinning the rise in informality on trade ble to argue that manufactured products, particularly labor- liberalization also appears in Argentina. In Figure 5.24 intensive goods, are uniformly more formal than the more Gasparini plots the share of the informal work force relative resource-oriented sectors. to the formal. What is striking is not only the dramatic Nor is there much evidence that competitive pressures secular trend, but that the trend begins in the beginning of during liberalization have increased the amount of out- the 1980s, long before Mercosur or the opening of the HOW WE W 0ORK JOI QUIALITY IN F MER(iING SECTORS FIGURE 5 20 Impact of Minimum Wages on the Distribution of the Wages of Informal and Fonnal Salaried Workers Minimum Wage Minimum Wage II ....... --. nat2 4 6 8 102 4 6 8 1 Logwages Logwages Argentina, 1998:3 Brazil, 1998 Miniimum Wage 1,,~~~~~~~~~~~~~~~~~~~~~~~~~I I g ............ .~~~~~~~~~~~~~~~~~~~~~~~~~. - - - - - - - - - - - - - - - - - - - I - - - _. 9 G 8 1 15 4 6 8 10 12 Logwages Logwages Colombia, 1998 Mexico, 1999:1 Saiirce: Maloney and NoOnez 200(0. economy. Decomposing the changes observed, Gasparini is tion to the large increase in self-employment, which is unable to attribute to increased openness the source of responsible for much of the rise in informal salaried em- increased informality. Given Mexico's relative stability of ployment, Blom, Pavcnik, and Schady (2001) find a secular informal share, it seems that Argentina's informalization increase in informality in every major sector. However, may be more related to extraordinary macrovolatiliry com- employing the same technique discussed in analyzing the bined with fairly rigid labor markets, and the rise in pay- movements in income distribution, they are unable to find roll taxes across the period. Figure 5.25 suggests that the a direct link between a decline in tariffs by industry and the increased informality has occurred across firms of all sizes. increase in informality. What is happening is general, and It is nor, for example, a subcontracting out of processes to not specific to firms more open to trade. It, again, m-ay be microfirms, but increased evasion at all levels, partly related to issues of stabilization or labor taxes. Brazil is a central case, both because of the dramatic This does not imply that insertion into the global econ- increase in informality and the size of the economy'. In addi- omy does nor somehow contribute, but the impact may nor 1 3 9 FROM NATURAL RESOURCES TO THE KNOWLEDGE ECONOMY: TRADE AND JOB QUALITY FIGURE 5.21 Formality Rates by Trade Exposure 10( - 9()- ______ s0 60- ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~~~~~~~UHigh Import 60gentinl Medium Import Nb,te Nonagricultural workers. 4Means by Lowde Low Import U Nonrradeable Fon 40 te Low Export - H Medium Export 50t) _ ] ] aro,Znrinqa %1 - H -High Export 2(1 10 0 Argentina Brazil Mexico Venezuela, RIB. de Ni.e Nonagricultural workers. Means by trade balance per worklr cenriles. Soarce: National Labor Force Surveys. FICGURE 5 2) Formality Rates by Leanmer Categories 7 0~ 4- 20 Animats Petroleum Materials Forest Tropical Cereals Labor Capital Machines Chemicals Nontradeab Intensive Intensive Learner Categories a. Urban workers only. Soarcee: National Labor Force Surveys. be as direct as sometimes thought. Helfner and Castro de fraction of which would have migrated to cites and perhaps Rezende (2001) show that the increased exposure to compe- contributed to swelling the size of the informal sector. tition, particularly from Argentina through Mercosur, had a In sum, an observer in 1995 would have seen a dramatic devastating impact on wheat and other cereal production in rise in informality in Argentina, Brazil, Colombia, anid the south of Brazil. This, combined with the appreciated Mexico, four countries accounting for 75 percent of thie exchange rate and cheaper imported fertilizer and machin- region's population. But arguably much of this was volun- ery that raised the relative cost of labor 70 percent, meant a tary and driven by demand pull to the nontradeables sector. decrease in the rural workforce of around 4 percent, a large And to an observer in 200 t, much of the increase observed 140 HOW WE WORK JOB QUALITY IN EMERGING SECTORS FIGURE 5.23 Domestic and Foreign Businesses as Microenterprise Clients, Mexico and the Dominican Republic Manufacturing Services Commerce 40 ES Stes with Low FDI E States with High FDI Mexico ~10 S 0 4f 1994 1996 1998 1994 1996 1998 1994 1996 1998 20 - l)tImnli1lItan IWlpLihli... = 15- 10 - ,: °___-1 All sectors Manufacturing Commerce Services Source: Sanchez, Joo, and Zappala 2001. in the second half of the 1990s could be attributed to stan- ported selling their products to foreign individuals or busi- dard difficulties in macroadjustment in the face of nominal nesses, 16 percent of owners said that trade liberalization rigidities. Neither looking across countries nor at one had left them better off than before, compared to 24 per- country across time suggests a tight link to greater inte- cent who said they were worse off. The impact was even grarion in the world economy. greater within individual sectors: 60 percent of those in commerce reported being better off, while half of those in Microenterprises: Improving Hou They Work in the textiles reported being hurt. Global Environment On the other hand, for firms that have been able to con- This is not to say that greater openness has nor had large nect with larger firms, the overall impact has probably been impacts on the microfirm sector. In the Dominican Repub- positive. Male-headed microfirms whose clients are small or lic, for example, despite the fact that only 2 percent re- large firms, or those located in areas with high FDI, earned 141 I ROM NATtJRAL RESOU RCES TO THE KNO)W'LEDGE l. ONOMY TIIADE AND JOB QUALITY IIGURE 5921 Infonnality Rate for Salaried Workers in Greater Buenos Aires, 1980-99 35- 25 - ____i Trade Liberalization _) -,_ 15- 1 5 - I l l l l l l lI I I I I I I I I I I W? ibN ~a. 9>;\ -7 7 6) 4 N > 0~~~~~~~~NN0> "" a. Simprle OLS regression of change in women's share on wage changes. attractive to men now than in the past. Increasingly, men applied to work in other industries. Second, wages are with spouses are entering this sector, perhaps because it higher in this sector than in other typically female jobs, su(ch offers more earnings stability, which breadwinners must offer as textile manufacturing, and comparable to manufacturing tlheir families. On the demand side, Kopinak (01995) suggests wages. Since the work is visibly located in offices with a sizet- that employers are shifting their perception of the ideal able work force, the workers generally receive all the bene- worker away from young, single women (although their fits that accrue to formal sector work. employment in the sector continues to grow) toward men Processing firms, which compete for contracts on pric- and women with families, who are thought to take their jobs ing, prefer to hire young women, who constitute 90 per- more seriously due to their family responsibilities (UNIFEM cent of the workforce. They tend to have low levels of edu- 2000). This would be consistent with a greater need for cation, to be young (18 to 19 years old), able to type, and employee stability to train for more sophisticated processes. have an aptitude for dealing with technology and the stress of high workloads. Again, Pearson (1997) finds that em- Teleservice Industries ployers appear to assume that women are more patient, Data on the emerging long-distance service industries are dexterous, and flexible in learning new skills, and hence are also sparse, although Pearson (1997) offers a window into more productive at secretarial-type work, and this has led the Jamaican data-processing firms. Similar to inaquila to a high demand for female labor. Furthermore, the high industries, the jobs are low skilled with few options for addi- levels of concentration and short turnaround times require tional training, a flat promotion structure, and tedious work, young women who do not have household constraints that thus limiting the jobs as potential careers for all but the may limit their time and productivity. However, as small portion who do advance up the narrow job ladder. demand grows, employers are starting to hire older women. However, data processing does offer advantages that the more traditional assembly jobs or agriculture do not. First, Tourism workers receive some training in technology-related areas, Data on job quality in tourism are not easily accessible. lasting from two weeks to three months, which can be Lizardo and Guzman (2001) find that in the Dominican 152 140W W'E WORK: JOB QUALIITY IN EMERGING SECTORS Republic far fewer workers are found in the bottom quin- Nonetheless, in the short term, tourism offers a particularly tiles, most are solidly in the middle of the distribution, and visible reminder of the intrinsic randomness and injustice of overall the skill level of the industry is high: 32 percent of the global distribution of wealth and opportunity. workers have completed secondary school and 21 percent have complctcd college, compared to 25 pcrcent and 16.4 Nontraditional Agricultural Exports percent, respectively, for the economy as a whole. The study As discussed in the previous section, concerns about low of Barbados in Chapter 4 suggests comparable wages in productivity growth in the nontraditional agricultural ex- tourism and nontourism sectors. Further, the collected num- ports (NTAE) sector are probably unfounded. However, it bers may not include tips, which further push up the rela- is in this area that among the most serious concerns are tive wage. In addition, it seems possible that over the raised about workers' rights, labor standards, and discrim- longer term, wages and job quality can improve. Industri- ination. Rural workers have traditionally been among the alized countries and wealthy islands, such as Bermuda, worst paid, they have lived in the most severe poverty, and have large tourism sectors paying industrialized-country social relations retained a feudal flavor well into the 20th wages. The key, again, however, is to have a sector that century. Employment progress has not always been as fast offers a differentiated product that attracts travelers, even as would be desired, and work conditions are difficult, as is as the wage, and hence, the cost of the product, rises. often the case in the rural areas.22 As found in the inaquila and agricultural sectors, women But again, the comparison must be to alternative jobs in are often overrepresented in the industry. Chant (1991) finds the local areas, and here the evidence seems to suggest that, for Mexico that women's participation in tourism jobs overall, the growing NTAE sector offers opportunity. Wages exceeds their overall participation rate by 10 percentage in NTAEs tend to be higher than other wages in rural areas. points. Again, as Box 5.10 suggests, jobs may be allocated Women in the flower export sector in Ecuador earn 40 per- according to the perceived comparative advantage, which cent more than other rural women, while men earn one- may affect long-term career progress. The relative domi- third more than their counterparts in other agricultural nance of women also appears in the Dominican Republic, work. Further, two-thirds of men and women cite the low although men's employment in tourism has been increasing, risk in terms of wages and employment as an attraction of from 64.4 percent in 1994 to 70 percent by 1999. The roots the sector, and 13 percent of women highlight the benefit of the decreasing feminization of the industry are not clear, of having their own wage (Newman 2000). Similarly, in but the story may be similar to that for the other sectors: Chile, both men and women in the fruit-exporting indus- where the industry starts with above-average rates of female tries earn approximately 50 percent more per day than do participation, perhaps due to the availability of untapped workers in non-NTAE jobs in the same locality (Jarvis and female labor, a rapidly expanding industry may almost by Vera-Toscano 2001). These differentials are smaller during defiinitioi eveintually progressively employ more men.c2 the off-season, but the premium to NTAE work largely per- Tourism employment is also affected by continuing con- sists across the production cycle. cerns over the environmental sustainability of the industry. This picture seems hard to square with, for instance, the The Galapagos Computable General Equilibrium (CGE) situation of migrant workers in Guadalajara, Mexico, de- model discussed in Chapter 4 also highlights the hard scribed in Box 5.11, where working conditions are indeed choices implicit in environmental preservation: Reducing harsh. But again, what is clear is that workers do consider tourism by 10 percent would reduce the income of those these jobs to be better than their best alternative. Further, presently on the island by 7 to 9 percent, with a potentially as Box 5.11 suggests, it is not clear how much latitude large increase in poverty. there is for large increases in worker protection or quality. Finally, unlike other products, tourism is consumed The tomato market is highly competitive and Marshall's within the host country, and this can generate a microcosm of law appears strongly binding. global income disparities and a sense of being a second-class What is missing from such analysis is a total economy citizen in one's own country.2' In the long term, this contra- view of the impact on the rural poor. Carter, Barham, and diction is not necessary: Parisians may feel overwhelmed by Mesbah (1997) study three agricultural export booms and summer tourists, but they do not feel like aliens in France. find very different impacts dependent on three factors: (a) 153 FRO(M NATt1RAL RESOURCES TO THE KNO\\ LEDGY ECONOMY TRADE AND JOB QUALITY BOX 5.10 Wbat Do Empoyers Lak For? Wbat Do They See? A window into the nature of the demand for labor in the Though this allocation to jobs according to perceived tourism industry is offered by sociologist Sylvia Chant comparative advantage" has a certain logic, Chant argues (1991), who studied the tourism industry in Puerto Val- that it works against women in two ways. First, the types larta, Mexico. She found employers very clear on their of jobs that women are in are those without opportunities perceptions of men's and women's skills and how they fit for vertical or horizontal movement. A chambermaid may into their industry (Table 5.6). The hotel industry, virtu- become promoted to head maid, but that is the end of her ally by definition, has a high demand for skills tradition- job ladder, while a bartender may become a waiter, a head ally found in the household; hence women tended to waiter, and eventually a maitre d'. Second, men are in work as chambermaids, laundrywomen, and cashiers in more visible jobs, so that although the legal minimum small food establishments, and as clerks in souvenir wages for men's activities such as bellboy, general cleaner, shops, while men tended to work as clerks in more groundsman, or bar attendant were roughly 75 percent ot upscale shops, and bellhops, bartenders, doormen, chefs, that earned by cooks, chambermaids, or laundryworkers, and waiters in more upscale establishments where tips and commissions could dramatically reverse this. strength was more valued and which were more in the traditional domain. Stserce: Chant 1991: 103. TABLE 5.6 Pemeptons of Male and Female Employees In the Tounsm Sector lin Ofder of Most Cited) MEN WOMEN POSITIVE Greater work experience Docility Greater pbysical strength Reliability Foreign language skills/experience Punctuality :,!I.r, r t ,I.Ir, toworkovertinme Flexibility (additional tasks) Willingness to take orders Domestic skills Diligence Patience cooperative disposition NEGATIVE Drunkenness on the job Neurotic/temperamental Greater rates of absenteeism (due to idleness/heavy drinkiag) Greater races of absenteeism (due to family problems/pregnancy) Troublemuaking (union activity) Unwillingnessfiuability tu work overtime Dishonesty Lack of education/experience inparience Physical weakness Inflexibility (extra tasks) Lack of authority (as maanager) Unwillingness to do detmeaning/female work Demanding (higher wages) Resistant to authority (especially female supervisors) Soaex: Chant 1991. whether small-scale Lnits participate in producing the the latest boom in winter-vegetable exports followed an export crop and can enjoy higher incomes, (b) whether the inclusive path. Smaliholders took on the production of crop induces a pattern of structural change that systemati- broccoli and snow peas, and primary export crops were not cally improves or worsens the poor's access to land, and obviously less labor absorbing than displaced traditional (c) whether agricultural exports absorb more or less labor of crops. Further, they observe land transferred from larger to landless and part-time farming households. In Guatemala smaller farmers that adopted new agricultural products. At Is-s HOW WE WORK JOB QUALITY IN EMERGING SECTORS OX 5. 11 We Don't Want Those Bad Jobs. Or Do We? Sayula, Jalisco, Mexico lies roughly an hour from Guada- places; and 3 percent had no, bad, or contested land. But, lajara, Mexico's Silicon Valley, and it is a major center importantly, only 1 percent said they were deceived by for the growing tomato export industry. It also offers an the contractor who brought them to Jalisco. This, com- important example of the difficulties in raising job qual- bined with the recurrent seasonal nature of the jobs and ity in an industry where products lack differentiation, well-developed information networks, suggests that mi- capital is easy to move, and hence small increases in la- grants knew about the conditions they would find and bor costs may lead to major declines in employment (see were so poor as to choose to take the jobs anyway. Losing Box 5.8), these jobs arguably worsened their poverty. In 1999, the Jalisco human rights commission sued The relationships between human rights workers and Chajoma Industries for human rights violations and won a the remaining plants are cooperative and progress has large settlement against them. Migrant workers were liter- been made as a result. Firms like Bonanza 2000 have bet- ally housed in old chicken coops and sanitation and health ter, although still very basic, facilities, and they forbid conditions were extreme. However, rather than raising the families to put their children in the field. Wages are worker conditions, within months Chajoma had left Jalisco. still low and educational facilities largely absent. But the It is reasonable to think, as one human rights worker margin for maneuver is small. Bonanza has extreme com- remarked, that "We don't want those bad jobs anyway." petition, not only from other states, but from countries But the issue is complex. The System for the Integral where wages are lower. And the migrants have voted Development of the Family, jointly with the National with their feet. They want these jobs. In the long term, Autonomous University of Mexico, interviewed 728 mi- the best hope is to raise the ok I II level of the migrants and grant families in the Sayula area as to their reasons for make them qualified for b:etter jobs, such as those emerg- leaving their home areas in Veracruz, Guerrero, and Ga- ing in the technology industry in Guadalajara. In the xaca, three of the poorest states in Mexico. Fifty-three per- short term, the ethical tradeoff is harsh. cent said they lacked good work; 27 percent said that they needed money; 15 percent wanted to travel to other Samrce, Jalisco 2000. the other extreme, small farmers faced severe credit con- Women in NTAEs straints that prohibited them from participating as owners The agribusiness labor force began as highly feminized and in Paraguay's wheat and soy boom, and the large farms continues to be today. Women make up more than 80 per- absorbed relatively little Labor. The net effect has been cent of the NTAE work force worldwide (Carr 2000), and largely exclusionary, Finally, participating in Chiles fruit Latin America is no exception. Women constitute over two- boom also required large capital investments with long ges- thirds of the flower-sector workers in Ecuador (Newman, tation periods, factors that precluded small farmers with Larreamendy, and Maldonado 2000), more than 50 percent limited access from participating in the boom. Many small of the temporary workers in the Chilean fruit business, over parcel holders, faced with high prices for land and squeezed 80 percent of the Colombian flower sector, 70 percent of in traditional crops, simply sold out. On the other hand, the Brazilian grape production, and over half of the NTAEs in overall employment impact of the boom was large and Costa Rica, Guatemala, and Honduras (Barrientos and oth- appears to have largely offset the first impact. Carter, Bar- ers 1999). In Chile, as in the Mexican ;maqulila sectors, ham, and Mesbah's (1997) conclusion is twofold. First, women were perceived as possessing the dexterity suited there is nothing intrinsically beneficial or harmful to the to specialized agriculture, and as having experience with poor about expanding export-oriented crops. It depends household agriculture. They were also perceived as less mil- substantially on the crop and the local context. Second, itant and thus less likely to oppose the new organization of addressing the barriers to credit and information could work (assembly line) developed by agribusiness. As a result, ensure a more inclusive development of these sectors. women tend to work in caring for plants, harvesting, and 155 FROM NATURAL RESOURCES TO THIE KNOWLEI)GE ECONOMY- TRADE AND JOB QUALITY' preparation for shipping, while men tend to specialize in Indigenous Groups: Hou Do They Integrate? heavy field work, fumigation, and physically transporting The impact of trade liberalization on indigenous peoples the agricultural products throughout the growth, prepara- and racial minorities is important, because they are among tion, and export process (Newman, Larreamendy, and Mal- the poorest in Latin America.25 The Sayula case highlights, donado 2000; Barrientos and others 1999).2 But perhaps as in particular, the relationship between indigenous commu- important as this comparative advantage, as Jarvis (2001) nities and jobs in the new economy. From the time of the argues, the growth of the sector was so rapid, and more conquest, indigenous groups of the region have suffered employment alternatives existed for men in the rural labor discrimination and exploitation. But even if this were nct markets than for women, so women were the source of the case, it is doubtful that these communities would be available labor for the emerging industries. well positioned to take advantage of opportunities that a Perspectives differ on whether to view often backbreak- more open economy might offer. On one hand, there is ing and repetitive jobs as providing opportunity for women, some room for exploiting a comparative advantage in "cul- or simply a new modality of exploitation. But again, inter- tural" goods. There can be no doubt that access to internm- views with workers in the Ecuadorian flower industry sug- tional markets has vastly expanded demand for traditional gest reasons for working there that are similar to those heard products, and almost certainly reduced their product elas- in the maquila sector in many ways. Thirty-six percent of ticity of demand. Nonetheless, it is also clear that exploit- interviewed workers felt that the industry offered a stable, ing these global niche markets has not been sufficient to lift dependable income, while 30 percent cited ease of entry these communities out of poverty, and this is partly due to (low job-search costs) (Newman, Larreamendy, and Maldo- factors which give them a comparative disadvantage in the nado 2000) in locations where jobs were few and the alter- emerging industries. At the most basic level, only 55 per- natives for women were local commerce or migration to the cent of the migrants to Jalisco discussed earlier spoke ciry to work as domestic servants and in construction, trans- Castilian; the vast majority speak Nahuatl or Huastelo, and port, or small-scale agriculture for men (Newman, Larrea- this makes coordination difficult even in the tomato busi- mendy, and Maldonado 2000). Additional benefits men- ness. As with other indigenous groups, they have among tioned were medical services and transportation to the work the lowest level of literacy, and the need to put their chil- site, and women cited the benefit of the development of new dren in the fields perpetuates low educational attainment. skills in nondomestic-type activities. There is simply no possibility that they would have suff- CGender wage differentials are smaller in the NTAEs than cient human capital to work in the Hewlett-Packard or in the general rural labor market, but some persist, perhaps IBM maquilas only an hour away from the Jitomate fields. due to gender segregation of jobs within the NTAE indus- Nor is it likely that firms will relocate to their home try. For example, in the Ecuadorian flower industry, women areas, which are often distant, poorly served by infrastruc- earn 2 percent less than men per hour, but in the general ture and schooling facilities, and distant from markets. labor market, they earn only half of men's wages. The exist- Viewing indigenous communities as any other "countrY' ing differential in the NTAEs is attributed to men's segre- seeking to trade, we know from Chapter 2 that these char- gation in more skilled jobs, such as operating tractors. On acteristics are unlikely to generate a dynamic modern econl- the other hand, in the Chilean fruit industry women earn omy. But further, theory tells us that when factors of pro- approximately 10 percent more than men, while women duction can migrate, it is possible for regions to have "no working in non-NTAEs earn approximately 10 percent less comparative advantage" and not produce anything for than men. The higher daily earnings of women is attributed 'trade." A study done by the Bolivian Vice-Ministry for to the high value placed on preparation of the product for Indigenous Affairs of the Ministry of Sustainable Develop- market, which is usually done by women, who are assumed ment and the World Bank (1999) applied Porter's "dia- to be more cautious with the fruit (Barrientos and others mond" evaluation strategy to identify the competitiveness 1999) (see Box 5.12). Finally, asJarvis (2001) shows, where of indigenous communities. Despite the possibility of piecework is the primary form of payment, women actually building on biodiversity products, as Figure 5.32 shows, earn more than men, perhaps because they have longer expe- measured along standard dimensions, no pueblo was very rience in the sector (seven years compared to four years for competitive, 23 percent were competitive, and 77 percent men), and they work longer hours.24 were either not competitive at all or were only slightly 156 HOW WE WORK JOB QUALITY IN EMERGING SECTORS B3OX 5.12 New Jobs, New Gender Roles Interviews with men and women working in full-time However, the self-esteem gained through working is (Ecuadorian flower NTAEs), seasonal (Chilean fruit partially offset by the feeling of failure to fulfill the role NTAEs), and home-based (Guatemalan specialty agricul- of a woman (N% mrnan 2000). In particular, women and ture) industries suggest that access to new jobs gives men still feel that women's role is in the household, and women increased bargaining power in the household and some women are unhappy and uncomfortable delegating alters relations vis-a-vis men, but the degree of change is childcare responsibilities to a relative or a paid daycare subject to the permanence of the job. center. In fact, 35 percent of women in NTAEs who Interviews with 123 workers in three Ecuadorian choose not to work cite 'care of the household and chil- flower NTAE zones showed that the steady, well-paid dren" as the rationale behind their decision. jobs were changing the roles and expectations between In contrast, interviews with women and men in the men and women, but the changes were at times difficult. seasonal Chilean fruit-packing NTAE industries reveal In particular, focus groups discussed that at the house- that responsibilities in the household and women's deci- hold level, women with a steady source of income have sionmaking power changed less than amnong full-time more decisionmaking power, ranging from how to spend Ecuadorian NTAE workers. Although women did make money to the feeling that they have exit options should decisions about how to spend the money they earned, their spouse or partner treat them badly. Domestic vio- often on labor-saving appliances that allow them to work lence is lower in communities where a high percentage of more efficiently in the household, Barrientos (1999) sug- the workforce is in the NTAE sector, and responsibility gests that the home sphere has not evolved in terms for reproductive health is higher. However, it should be of women's new roles and responsibilities as quickly as noted that the propensity to .o ruailh leave a spouse or the market has, and that women's empowerment in the partner is not higher in NTAE regions than non. N TAE home is questionable because women still work the dou- regions. In addition, the long work shifts, particularly of ble shift (although Newman [2000J does not find evi- women who can work late into the night due to electric- dence of a double shift) and identify with a traditional ity in the packing plants, unlike men in the field who female identity. quit when the sun goes down, has shifted some of the Finally, among women working in contract farming, domestic burden to men. Those men who also work in since most work with their husbands (Katz 1 995), there the NTAEs share more of the homecare duties with their is not a notable change in gender roles. Instead, this orga- wives than do men who work elsewhere. They were more nization of farming is an extension of the unpaid farming likely to see women as equals (though this transforma- that they did when the farms were not contracted, such tion is incomplete due to the high segregation of jobs by that work time, decisionmaking power, and even the gender in the sector). At the individual level, women decision to work is still decided by the husband. The pri- who work can better protect their interests. The money mary gender impact of this form of farming was that earned by women is allocated more toward women's older girls had more household rrsl l,n.,ib!hlrics so their needs and priorities-both expenditures for the house- mothers could work in the fields, but this may be a gen- hold, such as food, clothing, children's needs, and also for eral trend when there are work opporruniries rather than herself t,Nrve man 2Uu1 i(. specific to contract farming. competitive, with scores of under 1.5 on a scale of 3. The tradeoff between development and retaining the anchor of challenges are not only of bringing up the level of educa- traditional society, land. tion and infrastructure. The fact that 75 percent of pueblos There are cases where indigenous communities have were only slightly or not competitive in the use of their pri- been able to extensively take advantage of the global mar- mary factor of production, land, also implies a long-term ket. Indigenous communities in the south of Mexico have 1 57 FROM NATURAL RESOURCES TO THE KNOWLEDGE ECONOMY TRADE AND JOB QUALITY FIGURE 5' 2 are not to the comparative advantage of Mexico, a reduc- Bolivian Highlands: Exports Determinants tion in the government-supported price of corn would decrease rural incomes, employment, and wages, and stirm- Biodiversity ulate a sharp increase in rural outmigration. However, Aymara these negative impacts of agricultural policy reforms have -' Qouechua not materialized. Despite an 18 percent real drop in the /-/-' Amazonia support price for white corn between 1994 and 1997, corn / '\Oriental output declined only slightly, from 18.13 million metric tons to 18.02 million metric tons (according to the Mex- Internatlonal Infrastrticture Demand ,' can Ministry of Agriculture). There is no evidence that rural workers suffered disproportionately from structural reforms in Mexico, or that Mexico-to-U.S. migration i- creased above its (rising) trend of recent years. Taylor (2001) argues that the reason for the low impact of the falling prices was precisely that rural camipesino corn- Market Sociability munities are much more diversified than previous.y Mte: Indigenous people of Bolivia. Average of communities index. thought, and face very high transaction costs getting their Solrr: World Bank 1999. crops to market, making prices locally and not internation- ally determined. In a sample of 196 households in ttie central Mexican state of Michoacan, a number of surplus- producing farmers sold their excess production locally, made progress in developing environmentally sound means rather than selling to the government at the higher guarani- of both cultivating and selling coffee on the international teed price. When asked why, most responded that it was too market that offers a medium-term horizon for development costly to transport their harvest to the nearest government based precisely on their traditional factor of production. (Compafifa Nacional de Subsistencias Populares, CONA- Nonetheless the obstacles are daunting. As has been SUPO) purchase point, approximately 40 miles away. Otr- shown, despite the vast inflow of FDI in Mexico, virtually ers cited cases in which neighbors had paid to transport none of it has gone to the states with the highest indigenous their maize, only to have it rejected on the grounds that the population, and there is a statistically negative relationship quality was not sufficiently high or the silo was already full. between the percent of the population that speaks an in- This added marketing uncertainty, discouraging farmers digenous tongue and FDI. However, if we also compensate from selling unless there was a sufficient price spread to for shipping distance to the United States, level of educa- cover not only transportation, but also an implicit market- tion, and level of urbanization, the indigenous character of ing risk premium. Some farmers complained that, lackin-g the region proves unimportant. It is not being indigenous their own vehicles, they would have had to sell to interm,- per se, but the locale and level of education that are impor- diaries at a price below the guaranteed price (reflecting the tant-again, the community's comparative advantage. In high transaction costs and marketing risks in this region). fact, Yucatan and Quintana Roo, which are on the Yucatan Both diversification and transaction cost effects can be Peninsula and have ready access by sea to U.S. markets, have analyzed in a villagc-levcl CGE model that, when adjusted higher-than-predicted FDI, suggesting that indigenous lan- to reflect the characteristics of his sample, explains what guage or culture need not be a barrier to participating in happened (see Box 5.13). When perfect product integra- emerging industries. tion with the national, and hence international, market w.as Ironically, these same barriers may have buffered com- assumed, a massive decline in output of 26 percent ap- munities somewhat from the heat of international compe- peared, and migration to the United States increased 4 per- tition. For example, NAFTA and related initiatives call for cent. But when prices were determined locally, production a gradual (1 5-year) phaseout of price supports for corn sta- actually increased, due to the impacts on consumption of ples. Studies using aggregate computable general equilib- the transfer from the Programa de Apoyos Directos al rium (CGE) models concluded that, since these products Campo (PROCAMPO). 158 HOW WE WORK JOB QUALITY IN EMERGING SECTORS BOX 5.13 Medcac Village-Town CGE Model The village-town CGE model for Michaocan, Mexico, is village-town prices equal to outside market prices for really two models-one for the town and the other for the tradeables. For nontradeables (that is, goods or factors for three villages. These two models are linked to each other, which high transaction costs isolate the village-town from especially through trade. Each model has five compo- regional or national markets), local prices are endogenous nents: (a) household-farm production, (b) household-farm "equilibrium" prices, at which local supply and demand income, (c) expenditures, (d) a set of general equilibrium are in balance. The component behavioral equations are closure equations, which ensure that either local markets estimated from the village data and the simulations done clear or the village-town is a net buyer ("importer") from under two assumptions: that corn prices are determined in or seller ("exporter" of marketed surplus) to outside mar- national (international) markets, or that transaction costs kets, and (e) a price-determination component, which sets imply that they are determined locally. TABIE 5.7 Estimated Percentage Effects of a 10 Perment Decrease in Grain Prices and Compensatitng Income Transfer STAPLE PRICE NATIONALLY STAPLE PRICE LOCALLY DETERMINED VARIABLE DETERMINED (TRANSACTION COSTS) Basic Grain Prices -10.0 0.2 Output Stapte -25.7 0.7 Other Crop 3.5 -0.1 Livestock 4.3 -0.1 Mexico-U.S. Migration 0.4 0.l Labor Factor Demand -1.3 0.3 Wage -0.1 0.0 Village Reat GDP -2.5 0.2 Household Real Income Comnmercial 0.5 1.6 Subsistence 1.1 0.8 Capital Investment -0.5 1.1 Marketed Surplus 247.7 - Nt aailable. Sou,re: Taylor 2001 and Yuwez-Naade and Dyer 1999. The lesson is, then, that the poor infrastructure, geo- of the debate and what the available evidence suggests about graphical distance, and even cultural norms that often iso- their impact on trade patterns. As has been stressed late indigenous communities prevent ready access to the throughout, the ability to raise job-quality wages, benefits, new opportunities offered by the global marketplace, but and job environment is circumscribed by the productivity of also gives them time to adjust to competition on the workers and Marshall's law of labor demand. The labor stan- import side. This time can be used to both raise the qual- dards debate overlaps greatly with the debate examining the ity of education and to think through appropriate strate- impact of other labor market distortions or rigidities. gies of growth that maintain, to whatever degree possible, On the second point, the evidence is highly ambiguous. cultural integrity. Mah (1997) shows that developing countries that ratify International Labour Organisation conventions have lower Conclusion: How Can We Continue to exports as a share of GDP than those that do not. He then Raise Job Quality? interprets this as evidence that these developing countries Over the short term many observers have focused on the lose part of their comparative advantage as they become issue of labor standards. This is a complex issue that this more similar to their industrialized trading partners. Van report can only touch on by noting some of the parameters Beers (1998) somewhat counterintuitively concludes that 159 FROM NATtURAL RESOLURCES T(O THIE KNOWLEDGE ECONOMY TR,DE AND JOB QUALITY strict labor standards, such as those prevalent in OECD problem. Below some total packagc of remuneration, firns countries, disproportionately affect skilled labor by reduc- cannot attract workcrs. It may be more of an issue where ing skilled-labor-intensive exports. Rodrick (1996) shows labor protections force total remuneration above market that more rigid regulation increases the cost of labor and clearing and the transnational bargaining undermines therefore diminishes a country's comparative advantage in insider benefits. More generally, this concern has manifested labor-intensive goods. As numerous authors have noted, itself through the idea that the large informal sectors are, de such results are hard to interpret anyway since there is often facto, workers stripped of these benefits. As this chapter sug- a correlation between high labor standards and other inter- gests, this very commonly cited scenario is not, in fa(.t, ventions, such as industrial policies or other distortions. prevalent in the region. (See Rama and Tabellini 1992 and Kuruvilla 1996.) Of greater concern is who gets to set the labor standard3s. 1Hlaving said this, however, it is important to point out Latin American governments have been suspicious of labor that, with the exception of Rodrick (1996), the studies dis- side clauses to trade agreements. This is because higher leg- cussed above use exports as a share of GDP or some other islated workplace quality or benefits that do not cau,e macroeconomic measure as their dependent variable. 26 This unemployment necessarily diminish the fraction of total implies that, while we can conclude that stricter labor stan- remuneration paid in wages. There is no guarantee that dards have little effect on the overall trading performance regionally or globally set standards about the composition of developing countries, this tells us little about the effect of benefits will reflect local worker or local preferences. that they may have on other variables of interest. More fundamentally, however, the concern is the potent,.al Probably of more relevance is the emerging literature on for abuse as trade barriers in the industrialized countries. the employment impacts of various labor market interven- In any case these proposed measures can make only mar- tions. Numerous studies in the region suggest that labor ginal changes in worker welfare. Historically, the only w,iy market interventions have costs-those benefiting do bet- to raise living standards in a sustained fashion is through rer, but jobs may be lost. As only one example, recent work on minimum wages by Fajnzylber (2001a) in Brazil and AMaloney and Nunez (2001) in Colombia, does find impor- BOX 5.14 .. . . ~~~~Labor Standards in IEPZs tant negative employment effects of raising minimum wages. Pages and Heckman (2000), for example, also find adverse It is cornmonly argued that labor standards are most employment effects of firing restrictions. In a world where likely to be ignored in EPZs to increase competitiveness. labor-cost differences across competing developing coun- However, Romero (1995) points out that EPZs gener- tries are fairly small (for example, the difference in labor ally pay higher wages and offer better working condi- costs between Malaysia and Mexico is a mere 3 percent), a tions than the rest of the economy because (a) they pay mocderate increase in labor costs may lead firms to substi- productivity incentive bonuses (for example, piece rates) tute labor for capital, or to substitute across different types and overtime; (b) they tend to be larger, and usually pay of labor in order to maintain their competitive edge. scales and working conditions are correlated with firm Where the interest in labor standards goes beyond simple size, perhaps because large firms are easier to monitor analysis of labor market interventions is in the debate about and regulate than smaller ones; (c) most foreign-owned whether they should be set globally. Some of the sentiment firms apply their domestic "best-practice codes in all in favor of global standards arises from highly publicized their branches; and (d) the government often regulates cases of treatment of workers in export processing plants. On higher minimum wages for firms in this sector. average, this appears to be overstated. As this chapter sug- Exceptions, nevertheless, can occur in those cases gests, for example, EPZs do not pay below comparable where monitoring procedures are lax or union rights wages in the country and tend to pay benefits. As Box 5.14 restricted. This opens the case for government moni- suggests, this appears to be more the rule than the exception toring and international coordination. in much of the world. The related concern, however, is that transnational firms can play one country off another to Sore: Romero 1995. reduce wage and benefits costs. At some level, this is not a 160 HOW Wr WCORK JOB QUAIATY IN EMERGING SECTORS increasing labor productivity-work-ers can be paid more if jas and Broners (1989). In fact, Rees and Shah in the United King- they produce more. As this report has argued, the challenge dom find no significant relationship between self-employed earnings of raising productivity is a multidimensional one involving and human capital variables. Their regressions included far more ofetaising producctivity capistamultiimeiona inontinvoling a than the basic human capital variables, which prevents a straightfor- measutes to aceeaecaia omaward comparison of their explanatory power in the various sectors. the adoption of foreign technologies, and those to increase 7. The countries included were Argentina, Brazil, Chile, Colom- the human capital of the workforce. Swedish labor unions bia, Costa Rica, Mexico, and Peru. understand the fundamental fact that Swedish forestry 8. Conceptual and statistical definitions of the informal sector are workers earn more than those in Brazil and Chile not not as clear-cut as one might expect because of its characteristically because of labor legislation, but because of a higher degree diverse and complex nature. In this section, the informal sector includes three groups of workers: (a) employers who hire at most 15, or in most cases less than five, paid workers, with or without appren- ters that encourages constant innovation and increasing rices; (b) self-employed workers who own and operate one-person labor productivity. The process of raising job quality is thus businesses alone or with the help of unpaid workers, generally family the process of development. members, and apprentices; and (c) employees in these microfirms regardless of their degree of protection. Alternative definitions of the informal sector focus on the issue of "protectionism," and thus Notes include owners or workers in firms of fewer than 15 employees who 1. In LAC, we can find a slightly positive correlation between do not have social security or medical benefits. unemployment and openness, but it is highly dependent on just a 9. Cunningham and Maloney 2001 employ cluster and factor analy- few small, open economies such as Barbados, Trinidad and Tobago, sis to identify a typology of informal microentrpreneurs in Mexico. and Nicaragua, and is sensitive to specification. For example, if we 10. Fiess, Fugazza, and Maloney (2000) use a multivariate Jo- adjust openness for the fact that small countries are inherently more hansen approach Uohansen 1988) to explore cointegration relation- open, the relationship reverses. ships in relative earnings, relative sector size, and the real exchange 2. This section is based on Maloney and Montes Rojas (2001). rate, and formulate hypotheses to identify the degree of segmentation 3. See Abuhadba and Romaguera (t993) for Brazil, Chile, and the in the labor market in Brazil, Colombia, and Mexico. Positive United States; Marquez (1990) for the Republica Bolivariana de comovements between relative wages and sector sizes is attributed to Venezuela; and Krebs and Maloney (1999) for Mexico. segmentation, while negative comovements of relative wages and sec- 4. See Galiani, Arim, and Pantano (2001) for Argentina; Santa- tor sizes is classified as integration. Different labor market regimes are maria (2000) for Colombia; Robbins (1994) for Chile; Robbins and empirically identified with the LR-test on the coefficients of the coin- Gindling (1999) for Costa Rica; Cragg and Epelbaum (1996) for regration vectors. The period under consideration runs from 1985:QI Mexico; Acosta and Montes Rojas (21)01), and Sanguinetti and oth- to 1999:Q2 for Colombia, from 1987:Ql to 1999:Ql for Mexico, and ers (2001b) for Uruguay. from 1992:Ql to 1998:Q2 for Brazil. For Mexico the test of inte- 5. The reduction of wage differentials of medium-skilled to com- grated markets cannot be rejected prior to 1995. For Brazil, the pletely unskilled workers dominates the unequalizing effect of ter- hypothesis of integration cannot be rejected and the hypothesis of tiary education and produced a net reduction of 2.5 points of the segmentation is rejected at the I percent level. For Colombia, the Gini between 1988 and 1996. If the returns to tertiary education had hypothesis of segmentation cannot be rejected, while the hypothesis remained at the 1988 level (that is, 19 percent instead of 23.9 per- of integration can be. In all three cases, however, tests of the stability cent), the total reduction of wage inequality due to changes in ofthe cointegration space suggest the possibility ofdifferent relation- returns to education would have been 2 more percentage points ofthe ships in different subperiods. This finding is consistent with the argu- Gini. Moreover, if returns to education had fallen to U.S. levels (I 3 ment that rigidities in the formal sector may bind in some periods and percent). the additional reduction in wage inequality due to change not in others. depending on the macroeconomic environment. in returns would have been twice as large. 11. These strategies include augmentation of the diversity of 6. The arguments from the small-firm-dynamics literature sug- products and innovation; reduction of the costs of inputs, labor, raw gest a higher variance or volatility of earnings among the self- materials, and other supply goods; networking and forming linkages employed (businesses can be driven out of the market). The selection horizontally or vertically; switching product lines; training employ- process for survival in the seLf-employed sector typically leads to a ees to be more efficient and managers to be more effective; market- broader distribution of earnings for any given level of human capital, ing; and increasing the quality and customization of services (prod- compared to what would be obtained if workers were all salaried with uct differentiation). non-stochastic and smoothly increasing wages. This leads to higher 12. Barton and Bear (1999) advocate business models that pro- levels of inequality in the self-employment sector. It also implies that vide basic sets of standardized services to a broadly targeted market standard wage regressions will have less explanatory power in the that can be augmented at relatively little additional cost; such mod- self-employment sector, as found by Rees and Shah (1986) and Bor- els can be commercially viable and cost-effective, and increasingly 161 FROM NATURAL RESOURCES TO THF KNOWLEDGE ECONOMY: TRADE AND JOB QUALITY profitahle as more consurmers hecome aware of the gains of telecoom- holds econometrically They also point ou)r rhat areas wirh a relari'.ely muomcations and information services. strong presence of EPZs and tourism exhibit lower unemployn-lent 13. Lack of detailed microdata on firms is probably a major reason and illiteracy rates. why banks have not developed large portfolios of small-firm loans. 17. Literacy rates are higher-only 3 percent of naquita workers First, developing credit information registries, which provide rapid in the Dominican Republic are illiterate, compared to 14.6 per,cent access to standard information on borrower behavior from a variety of of workers in the rest of the economy. On average, workers have so,nme sources, such as lenders and utility companies, and making borrower secondary education in both countries, and approximately one more credit history available to all lenders, allows sorting potential borrow- year of education than workers in the rest of the economy. Most of the ers at low cost and reduces adverse selection problems (Galindo and difference is in the lower incidence of workers with no education- Miller 2001). Credit registries may provide additional incentives for only 1.3 percent compared to 9.4 percent for the economy as a timely repayment because they allow borrowers to create a "reputation whole-and lower rates of college graduation-7.6 percent cC m- collateral" in credit markets as good credit risk. Using credit-scoring pared to 16 percent for the economy overall. Primary and seconclary technologies, banks can take advantage of such information to assess education are correspondingly higher. the risk and profitability of microfirm loans. In the United States, 18. In their interviews, Sklair (1989) and Tiano (1994) found t nat credit scoring is already allowing large banks to expand into small- employers perceived women as "docile and dexterous," and that these business lending (Mester 1997). Second, a potentially important step traits were more suitable to a highly competitive, repetitive assembly wvould be reforming aspects of the legal and institutional framework. process. Movable assets (equipment, machinery, inventories, livestock, and 19. The industries that women dominate are not necessarily the accounts receivable), which constitute the bulk of microfirm capital, lowest paying. Assembly of tools is the highest-paying industry, had very limited capacity to carry debt. These assets represented most especially for men, who are overrepresented in the sector, especially of the aggregate value of assets of Mexican microfirms. for example. in the nonborder areas. However, machinery, transportation, and However, present legislation unduly restricts the use of movable assets electronics also pay well, all industries where women dominate. Flur- to secure credit transactions (Holden 1996; Fleisig 1995). Further, in thermore, 65 percent of working women are in these indusrry some cases, microfirm owners cannot pledge land and buildings as col- groups. The lowest-paying industries are food and textiles, where lateral because of unclear property rights. women dominate in the nonborder region, with over 50 percent of 14. Using bank data, Clarke and others (2001) find that during women in these categories, but it is mixed on the border, where only the 1990s in Argentina, Chile, Colombia, and Peru, lending by all 8 percent of women work in these industries. Regardless of the types of banks to small businesses generally shrank. However, after industry, though, men earn more than women (Figure 5.31). The dif- controlling for other factors that might affect lending, large foreign ferential is greatest in the tools industry, with an average of 92 per- banks actually appear to lend more to small businesses (as a share of cent for the whole meaquila industry. total lending) than large domestic banks in two of the four case study 20. The decreasing share of women in tourism is a double-edge-d countries, Chile and Colombia. Finally, one of the most interesting sword, though. On one hand, these jobs reinforce women's stereo- findings of the authors is that public banks do not appear to surpass types, since they are highly domestic and provide fewer labor market private banks in the extent to which they lend to small businesses, opportunities for women than do production jobs, where there are suggesting that privatization of public banks would not obviously promotion ladders. On the other hand, the seasonality of the jobs and hurt small businesses. the lack of discrimination based on household structure may giv,e 15. Employment in the msaquila sector is growing, but less women flexibility to work in both the market and the home. Fur- rapidly than employment in non-maquila manufacturing. Between thermore, any income earning by women (in any industry) will 1980 and 1990, when the msaquila industry experienced its highest increase their decisionmaking powers (Chant 1991). employment growth in both countries, mesaquila employment grew by 21. See, for example, "Alien" lyrics by Rohan Seon, sung by the 273 percent in Mexico and 692 percent in the Dominican Republic. Mighty Pep (1994 Saint Lucia Carnival King Winner). This compares to non-maquita manufacturing employment growth of 22. In the Ecuadorian flower industry, for example, nearly 70 pe-- 780 percent in Mexico. Employment growth slowed to 101 percent cent of male workers complain of the danger of the insecticides that by 1990-97 in Mexico, and 40.2 percent in the Dominican Repub- are heavily used on the crops, both during the planting season and for lic, but non-maqiitla manufacturing growth in Mexico was 237 per- shipping. Women in the same industry recognize the danger of men's cent over the same period. work with chemicals (64 percent), but they themselves primarily 16. Only 4.4 percent of all workers in EPZs and tourism come complain of repetitive tasks and remaining on their feet throughout from poor households, while this number is 18 percent for those in the work shift, which can last up to 16 hours. Similar complaints are other sectors. Due to the lack of information on household income heard in the Chilean fruit industry. However, some workers say they net of the worker's wage, or on whether the worker is the primary prefer the cleanliness and orderliness of the packing plants to alter- earner in the family, we need to be cautious in interpreting this dif- native job arrangements (Barrientos and others 1999). ference in poverty rates as a result of employment in EPZs and 23. The assignation of tasks by gender is a natural extension of tourism. Lizardo and Guzman (2001) show that this relationship the work done on the home farms and in the domestic sphere. For 162 SIOW WE WORK JOB QUALITY IN EMERGING SECTORz example, Katz (1995) likens the training of vines to weaving, thus pation in these jobs difficult for women, even as men are taking over making this task a feminine task. some of the homecare burdens. 24. It is not clear whether men or women have a comparative 25. Using household-level data for several countries, Wodon and advantage in NTAEs. Although the sector started out being heavily others (2001a) show that belonging to an indigenous population leads represented by women, the change in land policies and the fewer to a reduction in per capita income, even after controlling for other options available for rural employment (except migration) have led to household characteristics. In Guatemala, the negative impact of being an increase of men in this employment. Since men and women are indigenous represents about 15 percent of per capita income. In sorted into tasks in NTAEs based on physical characteristics (such as Bolivia, households not speaking Spanish or a foreign language also strength), assumed characteristics (such as dexterity), and gender tend to be poorer. This is the case for Quechua and Aymara speakers, biases (such as employing men only in chemicals), only the first, but the impact is not significant for rural households speaking which is a small group of NTAE jobs, points toward a clear compar- Guarani. In Brazil, black (Preta) and ethnically mixed (Parda) groups ative advantage. It is hard to document that, in fact, women are face a reduction in income of about 11) to 20 percent when compared indeed more careful and thus becter able to prepare fruit for export, to otherwise similar white households. These results suggest that there for example. However, as long as this bias exists, women will have a may be some level of discrimination in labor markets against indige- comparative advantage in the well-paid Chilean fruit industry. How- nous populations or specific ethnic groups depending on the country, ever, the organization of production is highly constraining for women so that assessing the impact of trade liberalization on these groups is (long, exhausting shifts, nowhere to care for children, unable to go all the more important since they are already at a disadvantage. home for lunch) who still have home duties, making further partici- 26. Rodrick (1997) uses labor cost per worker in manufacturing. 163 Bibliography The word processed describes informally reproduced works Desarrollo #21. Economic Commission for Latin Amer- that may not be commonly available through libraries. ica and the Caribbean, Chile. Baer, Werner. 1969. The Development of the Brazilian Steel Abuhadba, Mario, and Pilar Romaguera. 1993. "Inter- Indiustry. Nashville, Tenn.: Vanderbilt University Press. industrial Wage Differentials: Evidence from Latin . 2001. The Brazilian Economy. Grouth and Develop- American Countries." Journal of Development Studies 30 ment. 5th Edition. Westport, Conn.: Praeger Publishers. (1): 190-206. Balan, Jorge, Harley L. Browning, and Elizabeth Jelin. Acosta, Pablo, and Gabriel V. Montes Rojas. 2001. "Trade 1973. Men in a Developing Society. Geographic and Social Reform, Technical Change and Inequality: The Case of Mobility in Monterry. Mexico. The Institute of Latin Mexico and Argentina in the 90's." World Bank, Wash- American Studies. Austin and London: University of ington, D.C. Texas Press. Amsden, Alice H., and Takashi Hikino. 1994. "Staying Balassa, Bela. 1964. "The Purchasing Power Parity Doc- Behind, Stumbling Back, Sneaking Up, Soaring Ahead: trine: A Reappraisal." Journal of Political Economy 72 Late Industrialization in Historical Perspective." In (December): 584-96. William J. Baumol, Richard R. Nelson and Edward N. . 1989. "''Revealed' Comparative Advantage in Ja- Wolff, eds., The Convergence of Productivity. Its Significance, pan and the United States." Journal of International Eco- and Its Varied Connotations. Oxford and New York: Ox- nomic Integration 4: 8-22. ford University Press. Balassa, Bela, and Luc Bauwens. 1988. Changing Trade Pat- Arellano, Manuel, and Stephen Bond. 1991. "Some Tests terns in Manufactured Goods: An Econometric Investigation. of Specification for Panel Data: Monte Carlo Evidence Amsterdam: North-Holland. and an Application to Employment Equations." Review Baldwin, Richard, and Harry Flam. 1989. "Strategic Trade of Economic Studies 5: 277-97. Policies in the Market for 30-40 Seat Commuter Air- Armington, Paul S. 1969. "A Theory of Demand for Prod- craft." Weltwirtschaftliches Archi2, (125): 484-500. ucts Distinguished by Place of Production." Interna- Barham, B., M. Clark, E. Katz, and R. Schurman. 1992. tional Monetary Fund Staff Papers 16 (1): 159-78. "Nontraditional Agricultural Exports in Latin Amer- Aroca, Patricio. 2001. "Impacts and Development in Local ica." Latin America Research Review 11 (26): 43-82. Economies Based on Mining: The Case of the Chilean II Barrientos, Stephanie, Anna Bee, Ann Matear, and Isabel Region." Resource Policy 27: 119-34. Vogel. 1999. Women and Agribusiness: Working Miracles in Aroca, Patricio, and William F. Maloney. 2001. "A Note the Chilean Fruit Export Sector. London: Macmillan Press. on the Determinants of FDI in Mexico: Distance, Edu- Barro, Robert J., and Xavier Sala-i-Martin. 1995. Economic cation and Ethnicity." World Bank, Washington, D.C. Growth. New York: McGraw Hill, Inc. Processed. Barton, Clifton, and Marshall Bear. 1999. "Information Arriagada, Irma. 1998. The Urban Female Labour Market in and Communications Technologies: Are They the Key Latin America: The Myth and the Reality. Serie Mujer y to Viable Business Development Services for Micro and 165 FRO(M NATtIRAL RESOIIRCEL 1T0 THE KNOWLEDGE ECONONMY TRAI)E AND JOB QUALITY Small Enterprises?" Microenterprise Best Practices Proj- Blom, Andreas, Pinelopi Koujianou Goldberg, Nina ect, USAID, wwwmip.org/pubs/mbp/ict.htm. Washing- Pavcnik, and Norbert Schady. 2000. "Trade Liberaliza- ton, D.C. tions and Industry Wage Differentials in Brazil." Da rt- Basu, K. 1998. "Child Labor: Cause, Consequence and mouth College and World Bank, Washington, D.C. Cure, with Remarks on International Labor Standards." Processed. World Bank Policy Research Working Paper No. 2027. Blomstr6m, Magnus, and Arn Kokko. 2001. "From Nat- Washington, D.C. ural Resources to High-Tech Production: The Evolution Baumol, William J., Richard R. Nelson, and Edward N. of Industrial Competitiveness in Sweden and Finland." Wolff, eds. 1994. The Convergence of Productivity, Its Sig- Stockholm School of Economics. Stockholm, Sweden. nifcance. and Its Varied Connzotations. Oxford and New Processed. York: Oxford University Press. Blomstr6m, Magnus, and Patricio Meller. 1991a. Diverging Becker, Gary S., and Kevin M. Murphy. 1992. "The Divi- Paths: Comparing a Ceiztury of Scandinavian and Latinz Am,cr- sion of Labor, Coordination Costs, and Knowledge." The ican Developmient. Washington, D.C.: Inter-American De- QuarterlyJournal of Economics 106 (4): 1137-60. velopment Bank. Behrman,Jere, Miguel Szekely, and Nancy Birdsall. 2001. . 1991b. "Issues for Development: Lessons from "Economic Reforms and Wage Differentials in Latin Scandinavia-Latin Comparisons." In Magnus Blorn- America." Paper presented at the Poverty and Ap- strom and Patricio Meller, eds., Diierging Paths: Co,- plied Micro Seminar Series. World Bank, Washington, paring a Centuzry of Scandinav,ian and Latin American D.C. Development. Washington, D.C.: Inter-America Develop- Bergman, Edward M., and Edward J. Feser. 2001. Indies- ment Bank. trial and Regional Cl/zsters: Concepts and Comparatil eAppli- Bond, Eric. 2001. "Trade Structure and Development: The cations. Regional Research Institute, West Virginia Uni- Role of Logistics Costs in Latin American Countries." versity, West Virginia. Department of Economics, Penn State University. Uri- Berlinski, J. 1998. El Sisterna de Incentivos en Argentinza. versity Park, Penn. Processed. Instituto Torcuato Di Tella, Serie Documentos de Tra- Bonelli, Regis, and Renato Fonseca. 1998. "Ganhos de pro- bajo. Buenos Aires, Argentina. dutividade e de efici&ncia: novos resultados para a econo- Bernard, A., and C. I. Jones. 1996. "Productivity Across mia brasileira." Pesqlisa e Planejamento Economico (Brazil), Industries and Countries: Time Series Theory and Evi- 28 (2): 273-314. dence." Review of Econotnics and Statistics 78 (1): 135-46. Borjas, George J., and Stephen G. Bronars. 1989. "Con- Bernardes, Roberto. 2001. "Redes de Inovacao e Cadeias sumer Discrimination and Self-Employment."Jozurnal of produtivas Globais: Impactos da Estrategia de Com- Political Economry 97 (3): 581-605. peticao da Embrear no Arranjo Aeronautico da Regiao Borrus, Michael, and John Zysman. 1997. "You Don't de Sao Jose dos Campos." Fundac,ao SEADE, Sao Paulo. Have to Be A Giant: How the Changing Terms of Corn- Processed. petition in Global Markets are Creating New Possibili- Bils, Mark. 1984. "Tariff Protection and Production in the ties for Danish Companies." Danish Research Unit fi)r Early U.S. Cotton Textile Industry."Journal of Econonmic Industrial Dynamics Working Paper No. 97-5. Aal- History 44 (4): 1033-45. borg, Denmark. Blom, Andreas, and Carlos Eduardo Velez. 2001. "The Dy- Bouzas, Roberto. 1996. "Mercosur y Liberalizaci6n Comrer- namics of the Skill-Premium in Brazil; Growing De- cial Preferencial en America del Sur: Resultados, Temas mand and Insufficient Supply?" Background paper to y Proyecciones." In Richard G. Lipsey and Patricto World Bank report on Inequality in Brazil. Washington, Meller, eds., Westernz Hemisphere Trade Integration: A D.C. Canadian-Latin American Dialogue. International Politi- Blom, Andreas, Dorte Verner, and Holm-Nielsen. 2001. cal Economy Series. New York: St. Martin's Press. "Education, Earnings and Inequality in Brazil 1982- Brome, Pearson. 2001. "Information Services: Integrating 1998; Implications for Education Policy." World Bank into the World Economy." University of the West Indies. Policy Research Working Paper. Washington, D.C. Processed. 166 B I B LI OG RAP PHY Brown, Cynthia J. 2000. "The Impact of Foreign Direct Chant, Sylvia. 1991. Women and Surtival in Mexican Cities. Investment on Small Business and Employment Forma- Manchester: Manchester University Press. tion in Mexico." Unpublished dissertation. University Chen, Martha, Jenifer Sebstad, and Leslie O'Connell. 1999. of Texas Pan American, Edinburg, Tex. "Counting the Invisible Workforce: The Case of Home- . 2001. "Gender in Mexico's Maquila Industry." Based Workers." World Development 27 (3): 603-10. World Bank, Washington, D.C. Processed. Christen, Robert P. 2000. "Commercialization and Mission Brown, Drusilla K. 2000. "International Trade and Core Drift: The Transformation of Microfinance in Latin Labor Standards. A Survey of the Recent Literature." America." Consultative Group to Assist the Poorest. Medford, Mass.: Tufts University. Processed. Occasional Paper. World Bank, Washington, D.C. Brown, Drusilla K., Alan V Deardorff, and Robert M. Clark, Don P., and Denise Stanley. 1999. "Determinants of Stern. 2000. "U.S. Trade and Other Policy Options and Intra-Industry Trade Between Developing Countries Programs to Deter Foreign Exploitation of Child Labor." and the United States. " Journal of Economnic Development Tufts University, Working Paper. Medford, Mass. 24 (2): 79-94. Brulhart, Marius. 1994. "Marginal Intra-Industry Trade: Clarke, George, Robert Cull, Maria Soledad Martinez Measurement and Relevance for the Pattern of Indus- Peria, and Susana M. Sanchez. 2001. "Bank Lending to trial Adjustment." Weltwirtschaftliches Archizc 130 (3): Small Businesses in Latin America: Does Bank Origin 600-13. Matter?" Paper to be presented at the Latin America Buffington, Robert M., and William French. 1999. "The and the Caribbean Economic Association 2001. Culture of Modernity." In Michael C. Meyer and Clert, C., and Quentin Wodon. 2001. "The Targeting of William Beezley, eds., The Oxford History of Mexico. Ox- Government Programs in Chile: A Quantitative and Qual- ford: Oxford University Press. itative Assessment." Cuadernos de Economia. Burki, S. Javed, and Guillermo E. Perry. 1997. The Long Collier, Paul, and Anke Hoeffler. 1998. "Greed and Griev- Mllarch: A Reform Agenda for Latin Amtierica and the Ca- ance in Civil War." World Bank Policy Research Paper ribbean in the Next Decade. Washington, D.C.: World No. 2355. Washington, D.C. Bank. Conning, Jonathan. 2001. "Latifundia Economics." De- Burki, S. Javed, Guillermo E. Perry, and Sara Calvo, eds. partment of Economics, Williams College. Williams- 1997. "Towards Open Regionalism." World Bank, town, Mass. Processed. Washington, D.C. Cooper, David P. 2001. "Innovation and Reciprocal Exter- Cardenas, Enrique. 1997. "A Macroeconomic Interpreta- nalities: Information Transmission via Job Mobility." tion of Nineteenth-Century Mexico." In David Can- Journal of Economic Behavior & Organization 45: 403-25. ning, 1999, "Infrastructure's Contribution to Aggregate Cord, C., and Quentin Wodon. 2001. "Do Mexico's Agri- Output." Policy Research Working Paper No. 2246. cultural Programs Alleviate Poverty? Evidence from the World Bank, Washington, D.C. E jido Sector." Cuiadernos de Economia, 114: 239-56. Carr, Marilyn, Martha Chen, and Jane Tate. 2000. "Glob- Corden, W. Max. 1972. "Economies of Scale and Customs alization and Home-Based Workers." Feminist Economics Union Theory."Journal of Political Economy 80 (3): 465- 6 (3): 123-42. 75. Carter, Michael R., Bradford L. Barham, and Dina Mesbah. Cortes Conde, Roberto. 1985. "Some Thoughts on the 1997. "Agricultural Export Booms and the Rural Poor Industrial Development of Argentina and Canada in the in Chile, Guatemala and Paraguay." Latin American 1920s." In D. C. M. Platt and Guildo di Tella, eds., Research Review 31 (1): 33-65. Argentina. Auistralia and Canada: Stzudies in Compara- Casanueva Reguart, Cristina, and Flor Brown Grossman. tive Development, 1870-1965. London: MacMillan Press, 2000. "Globalization and Industrial Restructuring in 149-60. Mexico: The Cases of the Automobile and Electronic Cragg, Michael Ian, and Mario Epelbaum. 1996. "Why Industries." Mexico: Instituto Tecnol6gico y de Estu- Has Wage Dispersion Grown in Mexico? Is It the Inci- dios Superiores de Monterrey. Mexico City Campus, dence of Reforms or the Growing Demand for Skills?" Mexico. Processed. Journal of Development Economnics 51 (1): 99-116. 167 FROM NATURAL RESOLTRCES TO THE KNOWLEDGE ECONOMY TRADE AND JOB QUALITY Cuddington, John T., Rodney Ludema, and Shamila A. Development, 1870-1965. London: MacMillan Press, pp. Jayasuriya. 2001. "Prebisch-Singer Redux." George- 19-37. town University/World Bank. Washington, D.C. Djankov, Simeon, Rafael La Porta, Florencio Lopez-cle- Processed. Silanes, and Andrei Shleifer. 2000. "The Regulation of Cunningham, Wendy. 2000. "Latin America and the Ca- Entry." National Bureau of Economic Research Work- ribbean Gender Database." Forthcoming. ing Paper No. 7892. Cambridge, Mass. Cunningham, Wendy, and William E Maloney. 2001. Dollar, David, and Edward N. Wolff. 1997. "Convergence "Heterogeneity Among Mexico's Micro-Enterprises: An of Industry Labor Productivity Among Advanced Application of Factor and Cluster Analysis." Policy Re- Economies, 1963-1982." In Edward N. Wolff, ed., lThe search Working Paper No. 1999. World Bank, Wash- Economnics of Productivity 2: 39-48. ington, D.C. Duncan, Tim, and John Fogarty. 1984. Auistralia and Cunningham, Wendy, and Carlos Ramos. 2001. "The Argentina: On Parallel Paths. Carlton, Victoria, Aus- Home as the Factory Floor." Draft. World Bank, Wash- tralia: Melbourne University Press. ington, D.C. Engerman, Stanley L., Stephen H. Haber, and Kenneth L. Dade, Carlo. 2001. "Approaches to Increasing the Prodtic- Sokoloff. 1999. "Inequality, Institutions, and Different tive Value of Remittances." Inter-American Founda- Paths of Growth Among New World Economies. tion, Washington D.C. Cambridge, Mass.: National Bureau of Economic Re- Dagnino, Renato. 1993. "Competitividade da Induistria search. Processed. AeronAutica." ECIB-Estudo da Competitividade da Engerman, Stanley L., and Kenneth L. Sokoloff. 1997. Industria Brasileira." Nota T6cnica Setorial. Available "Factor Endowments, Institutions, and Differential at: http://www.mct.gov.br/publi/Compet/Default.htm. Paths of Growth Among New World Economies: A De Ferranti, David, Guillermo Perry, Indermit S. Gill, and ' ' ~~~~~~~~~View from Economic Historians of the United Stares." Luis Serv6n, with Francisco H. G. Ferreira, Nadeem ' ~~~~~~~~~~~~In Stephen H. Haber, ed., i-ou' Latin Amneri-a f:ell B5e- Ilahi, William F. Maloney, and Martin Rama. 2000. hind: Essays on the Economic Histories of Brazil and Mexico. SecuBring Our Future in a Global Economny. Washington, 1800-1914. Stanford, Cal.: Stanford University Press. D.C.: World Bank. de Soto, Hernando. 1989. The Other Path: The Invisible Rev- Fajnzylber, Pablo. 2001a. "Aircraft Manufacturing in Latin olution in the Third World. New York: Harper and Row America? Notes on Brazil's EMBRAER." World Bank, Publishers. Washington, D.C.: Processed. Deardorf, Alan V. 1994. "Exploring the Limits of Compar- . 2001b. "Minimum Wage Effects Throughout th-te ative Advantage." Weltwirtschaftliches Archiv 130 (1): Wage Distribution: Evidence from Brazil's Formal and 1-18. Informal Sectors." Centro de Planejamento Regional - . 1984. "Testing Trade Theories and Predicting Universidade Federal de Minas Gerais (CEDEPLAR - Trade Flows." Chapter 10 in R. W. Jones and P. B. UFMG), Working Paper No. 151. Belo Horizonte, Kenen, eds., Handbook of International Economics, Volume Brazil. 1, International Trade. Amsterdam: North Holland. Fajnzylber, Pablo, and William E Maloney. 2000. "Labor Dehem, Robert. 1962. "The Economics of Stunted Demand and Trade Reform in Latin America." Policy Growth.' Canadian Journal of Economnics and Political Sci- Research Working Paper No. 2491. World Bank, Wash- ence 28: 502-510. ington, D.C. Devlin, Robert, and Ricardo French-Davis. 1999."Towards Feenstra, Robert C., and Gordon H. Hanson. 1995. "For- an Evaluation of Regional Integration in Latin America eign Investment, Outsourcing and Relative Wages." in the 1990s." WUorld Economny 22: 261-90. Working Paper No. 5121. National Bureau of Econ- Di Tella, Guildo. 1985. "Rents, Quasi Rents, Normal omic Research, Cambridge, Mass. Profits and Growth: Argentina and the Areas of Recent _ _. 1996. "Globalization, Outsourcing, and Wage Settlement." In D. C. M. Platt and Guildo di Tella, Inequality." Working Paper No. 5424, National Bureau Argentina, Australia and Canada: Stiudies in Comparative of Economic Research, Cambridge, Mass. 16 BIBLIOGRAPHY Feldstein, Martin, and Charles Horioka. 1980. "Domestic versidad Torcuato Di Tella, Buenos Aires, Argentina. Saving and International Capital Flows." The Economic Processed. Journal 90: 314-29. Galindo, Arturo, and Margaret Miller. 2001. 'Can Credit Fiess, Norbert M., Marco Fugazza, and William F. Mal- Registries Reduce Credit Constraints? Empirical Evi- oney. 2001. "Real Exchange Rates, Labor Market dence on the Role of Credit Registries in Firm Invest- Rigidities, and Informality." World Bank, Washington, ment Decisions." Paper prepared for seminar, Towards D.C. Processed. Competitiveness: The Institutional Path. Inter-American Findlay, Ronald, and Ronald W. Jones. 2001. "Input Trade Development Bank, Washington, D.C. and the Location of Production." American Economic Re- Gallup, John Luke, Jeffrey D. Sachs, and Andrew D. view 91: 29-33. Mellinger. 1998. 'Geography and Economic Develop- Fleck, Susan. 2001. "A Gender Perspective on Maquila ment." Working Paper Series No. 1856. Harvard Insti- Employment and Wages in Mexico." In Maria Correia tute of Economic Research, Cambridge, Mass., pp. 1-57. and Elizabeth Katz, The Economics of Gender in Mexico. Garces-Diaz, Daniel. 2001. "Was NAFTA Behind the Washington, D.C.: World Bank. Mexican Export Boom (1994-2000)." Banco de Mexico. Fleisig, Heywood. 1995. "The Power of Collateral." Public Processed. Policy for the Private Sector, Note No. 43. World Bank, Garriga, Marcelo, and Pablo Sanguinetti. 1995. "Es el Washington, D.C. Mercosur un Bloque Natural?: Efectos de la politica Foster, William, and Alberto Valdes. 2001. "Has Reform Comercial y la Geograffa sobre el Comercio Regional." Failed Latin American Agriculture? A Review of Argen- Estudios Econ6micos. Revisra de la Fundaci6n Medite- tina, Chile and Colombia." World Bank, Washington, rranea, Abril/Junio. Buenos Aires, Argentina. D.C. Processed. Gasparini, Leonardo C. 2000. "La Informalidad Laboral Freund, Caroline L., and John McLaren. 1999. "On the en la Argentina: Evoluci6n y Caracterizaci6n." La Dynamics of Trade Diversion: Evidence from Four Economfa Oculta en la Argentina. Fundaci6n de Inves- Trade Blocs." U.S. Board of Governors of the Federal tigaciones Econ6micas Latinoamericanas, Argentina, Reserve System. International Finance Discussion pp. 162-223. Papers, 637: 1-51. Gaytan Hernandez, and Ana Isabel. 2000. "Tirar los pesti- Frischtak, Claudio R. 1993. "Development of the Brazilian cidas a for de piel. Percepciones del riesgo, cuerpo y Electronics Industry: A Study of the Competitiveness of salud en fumigadores del jitomate." Master's Thesis in Four Subsectors." In Bjorn Wellenius, Arnold Miller Social Anthropology. Centro de Investigaciones y Es- and Carl J. Dahlman, eds., Developing the Electronics tudios Superiores en Antropologia Social. Guadalajara, Induistry. World Bank Symposium, Washington, D.C. Jalisco. Processed. pp. 181-207. Georges, Eugenia. 1990. The Making of a Transnational ______. 1994. "Learning and Technical Progress in the Communnity: Migration, Development and Cultural Change Commuter Aircraft Industry: An Analysis of Embraer's in the Dominican Republic. New York: Columbia Univer- Experience." Research Policy (Netherlands) 23: 601-12. sity Press. Funkhouser, Edward. 1992. "Mass Emigration, Remit- Glick, Reuven, and Andrew K. Rose. 2001. "Does a Cur- tances and Economic Adjustment: The Case of El Sal- rency Union Affect Trade? The Time Series Evidence." vador in the 1980s." In George Borjas and Richard National Bureau of Economic Research Working Paper Freeman, eds., Immigration and the Work Force: Economic No. 8396. Cambridge, Mass. Consequences for the United States and Source Areas. Chi- Goldberg, Pinelopi Koujianou, and Nina Pavcnik. 2001. cago: University of Chicago Press. "Trade Protection and Wages: Evidence from the Fussell, Elizabeth. 2000. "Making Labor Flexible: The Colombian Trade Reforms." National Bureau of Eco- Recomposirion of Tijuana's Maquiladora Female Labor nomic Research, Cambridge, Mass. Processed. Force." Feminist Economics 6 (3): 59-79. Goldstein, Andrea. 2000. "The Political Economy of Galiani, S. and P. Sanguinetti. 2000. "Wage Inequality and High-Tech Industries in Developing Countries: Aero- Trade Liberalization: Evidence from Argentina." Uni- space in Brazil, Indonesia and South Africa." Organisa- 169 FRO(M NATURAL RESOURCES TO TIlE KNOWLEDGE ECONOMY TRADE AND JOB QUIALITY tion for Economic Co-operation and Development, of Economic Research, Working Paper No. 8188. Cami- Development Centre, Paris. Processed. bridge, Mass. Gonzalez Koning, G., and Quentin Wodon. 2001a. "Remit- Hakura, Dalia S. 2001. "Why Does HOV Fail? The Role tances and Inequality." World Bank, Washington, D.C. of Technological Differences within the EC." Journal of Processed. International Economics 54: 361-82. . 200lb. "Remittances, Education, and Child Hall, Bronwyn, and John Van Reenen. 2000. "How Effec- Labor." World Bank, Washington, D.C. Processed. tive Are Fiscal Incentives for R&D: A Review of the Griliches, Zvi. 1986. "Economic Data Issues." In Handbook Evidence." Research Policy 29: 449-69. of Econometrics, Volume 3, pp. 1466-1514. New York: Hall, Robert E., and Charles I. Jones. 1999. "Why Do Elsevier Science. Some Countries Produce So Much More Output Per Grossman, Gene M., and Elhanan Helpman. 1991. Innova- Worker Than Others?" Quarterly Jouirnal of Economzics tion and Grouwth in the Global Economy. Cambridge, Mass.: 114 (1): 83-116. MIT Press. Hallberg, Kristin, Hong Tan, and Leonid Koryukin. 2000. Grubel, Herbert G., and P. J. Lloyd. 1975. Intra-Indiustry "Export Dynamics and Productivity: Analysis of Mexi- Trade: The Theory and Measurement of International Trade can Manufacturing in the 1990s." World Bank, Was:h- in Differentiated Products. New York: John Wiley & ington, D.C. Processed. Sons. Hansen, Roger D. 1971. The Politics of Mexican Development. Guasch, J. Luis, and Joseph Kogan. 2001. "Inventories in Baltimore: The Johns Hopkins University Press. Developing Countries: Levels and Determinants: A Red Harrigan, James, and Egon Zakrajsek. 2000. "Factor Sup- Flag for Competitiveness and Growth." Policy Research plies and Specialization in the World Economy." Inter- Working Paper No. 2552. World Bank, Washington, national Research Department. Federal Reserve Bank of D.C. New York. Processed. Gylfason, Thorvaldur. 1999. "Natural Resources and Eco- Hay, Donald. 1997. "The Post-1990 Brazilian Trade Lil- nomic Growth: A Nordic Perspective on the Dutch Dis- eralization and the Performance of Large Manufacturing ease." Working Papers No. 167, the United Nations Firms." Instituto de Pesquisa Econ6mica Aplicada. UJniversity, World Institute for Development Economic Texto para Discussao 523. Research, Finland. Heckman, James. 1976. "The Common Structure of Sta- Haber, Stephen H. 1992. "Assessing the Obstacles to tistical Models of Truncation, Sample Selection, and Industrialization: The Mexican Economy, 1830-1940." Limited Dependent Variables and a Simple Estimator Jouirnal of Latin Americ.an StudcAies 24: 1-32. for Such Models." The Annals of Economic and Social Mea- ___ .. 1997a. "Financial Markets and Industrial Devel- surement 5: 475-92. opmenr: A Comparative Study of Governmental Regu- _ _. 1979. "Sample Selection Bias as a Specification lation, Financial Innovation, and Industrial Structure in Error." Econometrica 47: 153-61. Brazil and Mexico, 1840-1930." In Stephen H. Haber Helfner, Steven M., and Gervasio Castro de Rezende. 200]. ed., Hou' Latin America Fell Behind: Essays on the Ec-onomic "Brazilian Agriculture in the 1990s: Impact of the Pol- H-listories of Brazil and Mexico 1800-1914. Stanford, Cal.: icy Reforms." Paper prepared for delivery at the XXIV Stanford UTniversity Press. International Conference of Agricultural Economists, _____. 1997. "Introduction: Economic Growth and Latin August 13-18, 2000, Berlin. American Economic Historiography." In Stephen H. Hirschman, Albert 0. 1958. The Strategy of Economic Devel- Flaber, ed., Hou' Latin America Fell Behind: Essays on the opment. New Haven, Conn.: Yale University Press. Economic Histories of Brazil and Alexico I 800-9 l 14. Stan- Hjalmarsson, Lennart. 1991. "The Scandinavian Model of ford, Cal.: Stanford University Press. Industrial Policy." In Magnus Blomstrom and Patricio Hadass, Yael S., and Jeffrey G. Williamson. 2001. "Terms Meller, eds., Diverging Paths: Comiparinzg a Centuiry o, of Trade Shocks and Economic Performance 1870- Scandinavian ancd Lalin American Dewelopment. Washing- 1940: Prebisch and Singer Revisited." National Bureau ton, D.C.: Inter-American Development Bank. 17 C BIBLIOGRAPIHY Holden, Paul. 1996. "Collateral without Consequence: Jelin, Elizabeth, Matilde Mercado, and Gabriela Wyczykier. Some Causes and Effects of Financial Underdevelop- 2001. "Home Work in Argentina." Boosting Employ- ment in Latin America." The Enterprise Research Insti- ment Through Small Enterprise Development (SEED) tute of Latin America, Washington, D.C. Working Paper No. 6. Series on Homeworkers in the Global Holder, John S. 2001. "Some Ideals for Tourism Change Economiy. International Labour Organization, Geneva. in a Changing World." Secretary General. Caribbean Jones, Ronald W. 2000. Globalization and the Theory of Input Tourism Organization, Barbados. Trade. Cambridge, Mass.: MIT Press. Hummels, David. 1999. "Toward a Geography of Trade Jones and Kierzkowski. 1990. "The Role of Services in Pro- Costs." Processed. Purdue University. duction and International Trade: A Theoretical Frame- Hveem, Helge. 1991. "Developing an Open Economy: work." In R. Jones and A. Krueger, eds., The Political Norway's Transformation 1845-1975." In Magnus Econoimzy of International Tradle. Oxford: Basil Blackwell, Blomstr6m, and Patricio Meller, eds., Diverging Paths: pp. 31-48. C.- . a.; aCentury of Scandina-vian and Latin Anericai Katz, Elizabeth. 1995. "Gender and Trade Within the Development. Washington, D.C.: Inter-American Devel- Household: Observations from Rural Guatemala." opment Bank. World Development 23 (2): 327-42. Imbs, Jean, and Romain Wacziarg. 2000. "Stages of Katz, Jorge. 2000. "Cambios estructurales y productividad Diversification." London Business School, London, Eng- en la industria latinoamericana, 1970-1996." Revista de land, and Stanford University, Stanford, Cal. Processed. la CEPAL 71 (agosto): 65-84. Innis, Harold. 1933. Problems of Staple Produiction in Canada. Katz, Lawrence F., and Kevin M. Murphy. 1992. " Changes Toronto: University of Toronto Press. in Relative Wages, 1963-1987: Supply and Demand International Labour Organization. 1995. World Employ- Factors. QuiarterlyjJournal of Economics 107 (1): 3 5-78. ment. Geneva: ILO. Kearns, Allan, and Frances Ruane. 2001. "The Tangible Irwin, Douglas. 1996. Against the Tide: An Intellectual His- Contribution of R&D-Spending Foreign-Owned Plants to a Host Region: A Plant Level Study of the Irish Manu- tory of Free Trrade. Washington, D.C.: The American En- facturing Sector (1980-1996)." Research Policy 30: 227-44. etprinse Institute. * 2000. "How Did the United States Become a Net Keller, Wolfgang. 2001. "The Geography and Channels of E r n er - Diffusion at the World's Technology Frontier." National Exporter of Manufactured Goods?" National Bureau of BueuoEcnmcRsahWrkgPprN. EconmicResarchWoringPapr No 768. am- Bureau of Economic Research Working Paper No. Economic Research Working Paper No. 7638. Cam- 85.Cmrde as bridge, Mass. ~~~~~~~~8150. Cambridge, Mass . bridgeTr,Mass Key, Nigel. 1999. "Contract Farming, Smallholders, and Jansson, Tot, and Geoffrey Chalmers. 2001. "The Case for RuaDelomninLtnA ri:ThOgnz- Business Registration Reform in Latin America." Inter- tion ofvAgopoeng ir an therScaleTof Orgaower American Dvelopment ank, Washigton, D.C tion of Agropocessing Firms and the Scale of Outgrower American Development Bank, Washington, D.C. Production." World Development 27 (2): 381-401. Processed. Kopinak, Kathryn. 1995. "Gender as a Vehicle for the Sub- Jarvis, Lovell S. 1992. "Cambios en los roles de los sectores ordination of Women Maquiladora Workers in Mexico. publico y privado en el desarrollo tecnologico: lecciones Latin America Perspectives 22: 30-8. a partir del sector fruticola chileno." Colecci6n Estudios Kraay, Aart, and Jaume Ventura. 2001. "Comparative CIEPLAN 36: 5-39. Advantage and the Cross-Section of Business Cycles." Jarvis, Lovell, and Esperanza Vera-Toscano. 2001. "Sea- National Bureau of Economic Research Working Paper sonal Adjustment in a Market for Temporary Agricul- No. 8104. Cambridge, Mass. tural Workers: Fruit Workers in Chile." Draft. Wash- Kraay, Aart, Norman Loayza, Luis Serven, and Jaume Ven- ington D.C., World Bank. tura. 2000. "Country Portfolios." National Bureau of Jeanneret, Teresa. 1972. "El sisrema de protecci6n a la Economic Research Working Paper No. 7795. Cam- industria chilena." In Proceso a la industrializacion chilena. bridge, Mass. Centro de Estudios de Planificaci6n Nacional. Santiago, Krebs, Tom, and William F. Maloney. 1999. "Quitting and Chile: Ediciones Nueva Universidad. Labor Turnover. Microeconomic Evidence and Macro- 171 FROCM NATtJRAL RESOUTRCES TO THE KNOWLEDGE ECONOMY TRADE AND JOB QUALITY economic Consequences." Policy Research Working the Tequila Crisis: Basic Economics, 'Confidence Ef- Paper No. 2068. World Bank, Washington, D.C. fects' or Market Imperfections?" Journal of International Krueger, A. 1996. "Observations on International Labor Money and Finance. Standards and Trade." National Bureau of Economic . 2001b. "Mexico: Five Years after the Crisis." In Research Working Paper No. 5632. Cambridge, Mass. Annual Bank Conference on Development Economics, 2000. Krugman, Paul R. 1979. "Increasing Returns, Monopolis- Washington, D.C.: World Bank. tic Competition and International Trade." Journal of Lederman, Daniel, and Lixin Colin Xu. 2001. "Compara- 7nternational Economics 9: 469-479. tive Advantage and Trade Intensity: Are Traditiornal _____. 1999. Pop Internationalism. Cambridge, Mass.: MIT Endowments Destiny?" World Bank, Washington, Press. D.C. Processed. Kuruvilla, S. 1996. "Linkages Between Industrialization Lederman, Daniel, and Rodrigo Reis Soares. 2001. "A Strategies and Industrial Relations-Human Resources Note on the Impact of Economic Reforms on the Per- Policies: Singapore, Malaysia, the Philippines and formance of the Agriculture Sector in Latin America." India." Industrical and Labor Relations Review 49 (4). World Bank, Washington, D.C. Processed. Lagos, Gustavo. 1997. "Developing National Mining Poli- Levenson, Alec R., and William F. Maloney. 1998. "The cies in Chile: 1974-96." Resources Policy 23 (1/2): Informal Sector, Firm Dynamics, and Institutional Par- 51-69. ticipation." Policy Research Paper No. 1988. World Landes, David S. 1998. The Wealth and Poverty of Nations: Bank, Washington, D.C. Why Some Are So Rich and Some So Poor. New York and Levy, Santiago, and Sweder Van Wijnbergen. 1995. "Tran- London: Norton. sition Problems in Economic Reform: Agriculture in Lane, Philip R., and Aaron Tornell. 1999. "The Voracity the North American Free Trade Agreement." The Amer- Effect." American Economic Review 89: 22-46. ican Economic Review 85 (4): 738-54. Leamer, Edward E. 1984. Sources of International Comparative Lewis, Karen. 1999. "Trying to Explain Home Bias :n Advantage: Theory and Evidence. Cambridge, Mass.: MIT Equities and Consumption." Journal of Economiiic Litera- Press. ture 37: 571-608. ______ 1987. "Paths of Development in the Three- Lewis, P., W. Martin, and C. Savage. "Capital and Invest- Factor, n-Good General Equilibrium Model."Journal of ment in the Agricultural Economy." Quarterly Revieu' oj Political Economy 95 (5): 961-99. the Rural Economy 10 (1): 48-53. _____. 1995. "The Heckscher-Ohlin Model in Theory Li, Hongyi, Lixin Colin Xu, and Heng-Fu Zou. 2000. and Practice." Princeton Studies in International "Corruption, Income Distribution, and Growth." Ec_- Finance No. 77. Department of Economics, Princeton nomics and Politics 12 (2): 155-82. University, Princeton, N.J. Linbeck, Assar. 1974. Swedish Economic Policy. Berkeley and Leamer, Edward E., and Jerome Levinsohn. 1995. "Inter- Los Angeles: University of California Press. national Trade Theory: The Evidence." Chapter 3 in Lizardo, Magdalena, and Rolando M. Guzman. 2001. G. M. Grossman and K. S. Rogoff, eds., Handbook of Inter- "Patrones de integracion a la economfa global: ,Qu6 national Economics, Volume 3. Amsterdam: North Holland. comercializa America Latina? ,Qu6 hacen sus traba- Leamer, Edward E., Hugo Maul, Sergio Rodriguez, and jadores? El Caso de la Republica Dominicana." World Peter K. Schott. 1999. "Does Natural Resource Abun- Bank, Washington, D.C. Processed. dance Increase Latin American Income Inequality?" Loayza, Norman. 2001. "Growth in Brazil." World Bank, Journal of Development Economics 59: 3-42. Washington, D.C. Lederman, Daniel. 2001. "The Political Economy of Pro- Lopez, Jose R., and Mitchell Seligson, 1990. Small Businers tection: Theory and Chilean Experience." Ph.D. Disser- Development in El Salvador: The Impact of Remittance5. tation. The Johns Hopkins University, Baltimore. Report of the Commission for the Study of International Lederman, Daniel, Ana Maria Men6ndez, Guillermo Perry, Migration and Cooperative Economic Development. and Joseph Sriglitz. 200 la. "Mexican Investment after Washington D.C.: Government Printing Office. 172 BIBLIOGRAPHY Lopez-Acevedo, Gladys. 2001a. "Technology and Firm . 2001b. "Dealing with Dependency: A Compara- Performance in Mexico." World Bank, Washington, tive View." Washington, D.C.: World Bank. Processed. D.C. Processed. . "Technical Capacity and Development: Australia, . 2001b. "Technology and Skill Demand in Mex- Latin America and Scandinavia." World Bank, Wash- ico." World Bank, Washington, D.C. Processed. ington, D.C. Processed. .2001c. "Determinants of Technology Adoption in Maloney, William F., and Rodrigo Acevedo. 1995. "Trade Mexico." World Bank, Washington, D.C. Processed. Reform, Uncertainty and Export Promotion: Mexico L6pez-Calix, Jose R. 2001. "LAC Migration and the New 1982-88." Journal of Development Economics 48: 67-89. Patterns of Integration." World Bank, Washington, D.C. Maloney, William E, and Gabriel V. Montes Rojas. 2001. Processed. "Demand for Tourism." World Bank, Washington, D.C. Lucas, Robert E. B. 1979. "Sharing, Monitoring, and In- Processed. ccntives: Marshallian Misallocation Reasscsscd. "Journal Maloney, William F., and Jairo Nuniez, with Wendy Cun- of Political Economy 87 (3): 501-21. ningham, Norbert Fiess, Claudio Montenegro, Ed- Lumenga-Neso, Olivier, Marcelo Olarreaga, and Maurice mundo Murrugarra, Mauricio Santamaria, Corinne Schiff. 2001. "On 'Indirect' Trade-Related R&D Spill- Siaens, and Claudia Sepulveda. 2001. "Measuring the overs." World Bank Development Economics Research Impact of Minimum Wages: Evidence from Latin Group, Washington, D.C. Processed. America." World Bank Policy Research Working Paper Luzio, Eduardo, and Shane Greenstein. 1995. "Measuring No. 2597. Washington, D.C. the Performance of a Protected Infant Industry: The Mamalakis, Markos. 1976. The Growth and Structure of the Case of Brazilian MIicrocomputers." Review of Economics Chilean Economy: From Independence to Allende. New Ha- and Statistics 77: 622-23. ven: Economic Growth Center, Yale University Press. Maddison, Angus. 1994. "Explaining the Economic Per- Mann, Catherine L. 2001. "Implications of Global Internet formance of Nations, 1820-1989." In William J. Bau- Commerce for Trade Competitiveness: A Consideration mol, Richard R. Nelson and Edward N. Wolff, eds., for Selected Latin and Asian Countries. Chile, Mexico, Convergence of Productivity. New York: Oxford University Peru; Korea, Thailand, Vietnam." Institute for Interna- Press, pp. 20-60. tional Economics/World Bank, Washington, D.C. Mah, J. 1997. "Core Labour Standards and Export Perfor- Processed. mance in Developing Countries." World Economy 20 (6). Mariscal, Elisa, and Kenneth L. Sokoloff. 2000. "School- Mahler, Sarah. 2000. "Migration and Transnational Is- ing, Suffrage, and the Persistence of Inequality in the sues." Working Paper No. 4. Hamburg: Institut fur Americas, 1800-1945." In Stephen Haber, ed., Political Iberoamerika-Kunde. Germany. Institutions and Economic Growth in Latin America. Essays Maloney, William F. 1997. "Chile." In Laura Randall, ed. in Policy, History, and Political Economy. Stanford Univer- The Political Economy of Latin America in the Postwar sity, Stanford, Cal.: Hoover Institution Press. Period. Austin: University of Texas Press. Marquez, Gustavo. 1990. "Wage Differentials and Labor _____. 1999. "Does Informality Imply Segmentation in Market Equilibrium in Venezuela." Ph.D. dissertation. Urban Labor Markets? Evidence from Sectoral Transi- Boston University, Department of Economics, Boston, tions in Mexico." The World Bank Economic Review 14 (2): Mass. 275-302. Martin, Will, and Devashish Mitra. 2001. "Productivity _- 2000. "Informality Reconsidered." World Bank, Growth and Convergence in Agriculture and Manufac- Washington, D.C. Processed. turing." Economic Detelopnlent and Cultural Chage 49 (2): . 2001a. "Self-Employment and Labor Turnover in 403-22. Developing Countries: Cross-Country Evidence." In Martin, Will, and Peter G. Warr. 1993. "Explaining the Shantayanan Devarajan, F. Halsey Rogers, and Lyn Relative Decline of Agriculture: A Supply-Side Analy- Squire, eds., World Bank Econoniiji' Forum, Volume 1. sis for Indonesia." World Bank Economii- Revieu' (Interna- Washington, D.C: World Bank. tional) 7: 381-401. 173 FRONt NATUJRAL RES(t RCFS TO THE KNOWLEDGE ECONOMY TRADE AND JOB QUALITY Matsuyama, Kiminori. 2000. "A Ricardian Model with a Nelson, R., ed. 1993. National Innovation Systems: A Cmo),,- Continuum of Goods under Nonhomothetic Prefer- parative Analysis. New York: Oxford University Press. ences: Demand Complementarities, Income Distribu- Newman, Constance, Pilar Larreamendy, and Ana Mat la nion, and North-South Trade." Jouirnal of Political Econ- Maldonado. 2000. "Mujeres y Floricultura: Cambios y omy 108 (61): 1093-1120. Consecuencias en el Hogar." Draft. World Bank. Meller, Patricio. 2001. "Chilean Copper: Facts, Role and Nicita, Alessandro, Marcelo Olarreaga, and Isidro Soloag;l. Issues." Universidad De Chile Facultad De Ciencias 2001. "The Region as a Platform to the World." World Fisicas y Matemkticas Deparramento de Ingenieria Bank, Washington, D.C. Processed. Industrial/World Bank, Washington, D.C. Processed. North, Douglass C. 1955. "Location Theory and Regional Meredith, David. 1995. "The Role of Education and Economic Growth." Journal of Political Economny: 24'- Health Services in the Economic Development of 258. Australia and Argentina 1880-1940." School of Eco- O'Rourke, Kevin H., and Jeffrey G. Williamson. 199.). nomics Discussion Paper, the University of New South Globalization and History. The Evolution of a Nineteent,- Wales. Centuiry Atlantic Economny. Cambridge, Mass.: MIT Press. Merrill Lvnch 2000. "Emerging as an Electronics Man- Pages-Serra, Carmen, and James Heckman. 2000. 'Tle uifacturing Hub for North America." New York. Cost of Job Security Regulation: Evidence from Latin l'rocessed. American Labor Markets." Economnia. Journal of the Lat.n Mester, Loretta J. 1997. "What's the Point of Credit Scor- American and Caribbean Economics Association I (1). ing?" Business Revieu. Philadelphia: Federal Reserve Palacios, Marco. 1979. El c en Colombia (1850-1970: Bank of Philadelphia. Una historia Econ(;mnica. Social. y Pol/tica. Bogota, Colom - Meyers, Deborah. 1998. "Migrant Remittances to Latin bia: Editorial Presencia Ltda. America: Reviewing the Literature." Inter-American Dia- Par Ste nL. d Prescott. Ltda. logue, Washington D.C. ~~~Parente, Stephen L., and Edward C. Prescott. 2000. Barri- .l2001. Washington Regional Map:FlowsandImpacto ers to Riches. Cambridge, Mass.: MIT Press. _ 2001. "The Regional Map: Flows and lmpact of Pavcnik, Nina. 2000a. "Trade Liberalization, Exit, and Remittances in LAC." Washington, D.C.: World Bank. Pavcnikina Impao"rae Liberlaon Exit,an Processed. Productivity Improvements: Evidence fom Chilean Micco, Alejandro, and Natalia Perez. 2001. "Maritime Plants." National Bureau of Economic Research Work- Transport Costs and Port Efficiency." Inter-American ing Paper No. 7852. Cambridge, Mass. Development Bank, Washington, D.C. Processed. . 2000b. "What Explains Skill Upgrading in Less MT,F/IDB. 2001. "Remittances to Latin America and the Developed Countries? National Bureau of Economic Caribbean: Comparative Statistics." Washington, D.C. Research Working Paper No. 7846. Cambridge, Mass. Processed. Pavcnik, Nina, Andreas Blom, and Norbert Schady. 200]. Moilge, Alexatider. 2001. "Exports and Foreign Invest- "Trade Liberalization and Industry Wage Differentials ment in Costa Rica and El Salvador During the 1990's." in Brazil." Dartmouth College and World Bank, Wash- Department of Economics, Northwestern University/ ington, D.C. Processed. World Bank, Washington, D.C. Processed. Pearson, Ruth. 1997. "Gender and New Technology in the Moreira, Maurfcio Mesquita. 2000. "A Industria Brasileira Caribbean: New Work for Women?" In Janet Momseri, nos Anos Noventa: o que ja se pode dizer?" Sao Paulo: ed., Womnen and Change in the Caribbean: A Pan-Caribbean Banco Nacional de Desenvolv'imento Econ^nzico e Social Perspective. Kingston: Ian Randle Press. (BNDES). Processed. Pinto Santa Cruz, Anfhal. 1959. Chile. un caso de desarroh's Muendler, Marc, Claudia Sepulveda, and Luis Serven. frustrado. Editorial Universitaria. Santiago, Chile. 2001. "Productivity Growth in Brazilian Industry." Porter, Michael E. 1990. The Competitive Advantage of Na- World Bank, Washington, D.C. Processed. tions. New York: Free Press. Murphy, Kevin. M., Andrea Shleifer, and Robert. W. Portes, Alejandro, Manuel Castells, and Lauren A. Bentor, Vishny. 1993. "Why is Rent-Seeking So Costly to eds. 1989. The Informal Economiy. Baltimore: The Johns Growth?" American Economnic Review May: 409-14. Hopkins University Press. 17 B I BL IO GRAP HY Prebisch, Raul. 1959. "Commercial Policy in the Under- Guadalajara." In Alejandro Portes, Manuel Castells and developed Countries." The American Economic Revieuw, Lauren A. Benton, eds., The Informal Economy. Balti- Papers and Proceedings 49 (2): 25 1-73. more: The Johns Hopkins University Press. Pritchett, Lant. 1996. "Measuring Outward Orientation in Rodriguez, Francisco, and Jeffrey D. Sachs. 1999. "Why LDCs: Can It Be Done?" Journal of Development Economics Do Resource-Abundant Economies Grow More 49: 307-35. Slowly?"Journal of Economzic Growth 4: 277-303. Prugl, Elisabeth. 1997. "Microentrepreneurs and Home- Rodriguez-Clare, Andres. 2001. "Costa Rica's Develop- workers: Convergent Categories." World Development 25 ment Strategy Based on Human Capital and Technol- (9): 1471-82. ogy: How It Got There, the Impact of INTEL, and ______. 1999. The Global Construction of Gendler: Home- Lessons for other Countries." Paper written for the based, Work in the Political Economzy. New York: Colum- Human Development Report 2001. San Jose, Costa bia University Press. Rica. Processed. Puga, D., and A. J. Venables. 1999. "Agglomeration and Rodrik, D. 1996. "Labor Standards and International Economic Development; Import Substitution Versus Trade: Do they Matter and What Do We Do About Trade Liberalization." EconomicJournal 109: 292-31 1. Them." In R. Z. Lawrence, D. Rodrik and J. Whalley, Pindyck, Robert. 1988. "Irreversible Investment, Capacity eds., Emerging Agenda for Global Trade: High States for Choice and the Value of the Firm." American Economic Developing Countries. Baltimore: The Johns Hopkins Revieu 78: 969-985. University Press. Rama, M., and G. Tabellini. 1997. "Endogenous Distor- _ _. 1997. "Has Globalization Gone Too Far?" Insti- tions in Product and Labor Markets." Centre for Eco- tute for International Economics, Washington, D.C. nomic Policy Research, Discussion Paper No. 1143. Romer, P. 1990. "Endogenous Technological Change." London: Cambridge University Press. Jouirnal of Political Economy 98: 71-102. Ramos, Joseph. 1998. "A Development Strategy Founded Romero, T. 1995. "Labor Standards and Export Processing on Natural Resource-Based Production Clusters." Zones: Situation and Pressure for Change." Development CEPAL Review 66 (December): 105-27. Policy Review 13. Reardon, Thomas, Julio Berdegue, and German Escobar. Rosenzweig, Jeffrey A. 1988. "Elasticities of Substitution "Rural Nonfarm Employment and Incomes in Latin in Caribbean Tourism." Journal of Development Economics America: Overview and Policy Implications." World De- 29: 89-100. velopment 29 (3): 395-409. Rossi Junior, Jose Luiz, and Pedro Cavalcanti Ferreira. Rees, Hedley, and Anup Shah. 1986. "An Empirical 1999. "Evoluc,o daprodutividade industrial brasileira e Analysis of Self-employment in the U.K." Journal of abertura comerical." Pesquisa e Planejamento Economico 29 Applied Econometrics 1 (1): 95-108. (1): 1-35. Riera i Tuebols, Santiago. 1993. "Industrialization and Ruffin, Roy J. 1999. "The Nature and Significance of Technical Education in Spain, 1850-1914." In Robert Intra-Industry Trade." Economic and Financial Review. Fox and Anna Guagnini, eds., Education. Technology and Federal Reserve Bank of Dallas Q4: 2-9. Indiestrial Perfornmance in Eierope, 1850-1939. Cam- Sachs, Jeffrey, and Andrew M. Warner. 1995a. "Natural bridge: Cambridge University Press. Resources and Economic Growth." National Bureau of Robbins, Donald J. 1994. "Relative Wage Structure in Economic Research Working Paper. Cambridge, Mass. Chile, 1957-1992: Changes in the Structure of Demand . 1995b. "Economic Reform and the Process of for Schooling." Estidios de Economica 21: 49-78. Global Integration-" In William Brainard and George Robbins, Donald, and T. H. Gindling. 1999. "Trade Lib- Perry, eds., Brookings Papers on Economic Activity - VI eralization and the Relative Wages for More Skilled 1995 - 25th Anniversary Issue. Washington, D.C.: The Workers in Costa Rica." Review of Development Economics Brookings Institution. 3 (2): 140-54 . 1997. "Natural Resources and Economic Roberts, Bryan R. 1989. "Employment Structure, Life Growth." Revised version. Harvard Institute for Inter- Cycle, and Life Chances: Formal and Informal Sectors in national Development, Cambridge, Mass. 175 FROM NATURAL RESOURCES TO THE KNOWLEDGE ECONOMY TRADE AND JOB QUALITY ______. 1999. "The Big Push, Natural Rcsourcc Booms Sklair, Lcslie. 1989. Assemblingfor Development: The Maquifa and Growth."Journal of Development Economics 59: 43-76. Industry in Mexico and the United States. Winchester, Safford, Frank. 1976. The Ideal of the Practical: Colombia's Mass.: Unwin Hyman. Struiggle To Form a Technical Elite. Austin and London: Smith, Adam. 1776. An Inquiry into the Nature and Cau.Ies University of Texas Press. of the Wealth of Nations. Oxford: Clarendon Press. Sanchez, Susana M., Anna Joo, and Cara Zappala. 2001. Stallings, Barbara, and Wilson Peres. 2000. Grouth, "Constrained Global Integration: A Note on Micro- Employment and Equity. The Impact of the Economic Reforms enterprises in Latin America." World Bank, Washing- in Latin America and the Caribbean. Economic Commi s- ton, D.C. Processed. sion for Latin America and Caribbean. Washington, Sanguinetti, Pablo, and Miguel Salustro. 2000. "El Merco- D.C.: The Brookings Institution Press. sur y el sesgo regional de la politica comercial: aranceles Stern, Scott, Michael E. Porter, and Jeffrey L. Furman. y barreras no tarifarias." CEDI, Buenos Aires, Argen- 2000. "The Determinants of National Innovative Ca- tina. Processed. pacity." National Bureau of Economic Research Work- Sanguinetti, Pablo, Juan Pantano, and Josefiina Posadas. ing Paper No. 7876. Cambridge, Mass. 2001. "Trade Liberalization and the Dynamics of the Srone,Joe A., and Hyun-hoon Lee. 1995. "Determinants of Trade Structure in Argentina and Uruguay." World Intra-Industry Trade: A Longitudinal, Cross-Country Bank, Washington, D.C. Processed. Analysis." Weltwirtschaftliches Architv 131 (1): 67-85. Sanguinetti, Pablo, Rodrigo Arim, and Juan Pantano. Summerhill, William. 1997. "Transport Improvements 2001. "Changes in Production and Employment Struc- and Economic Growth in Brazil and Mexico." In ture and Relative Wages in Argentina and Uruguay." Stephen H. Haber, ed., How Latin America Fell Behind: Universidad Torcuato di Tella/World Bank, Washing- Essays on the Economic Histories of Brazil and Mexico 180(,- ton, D.C. Processed. 1914. Stanford, Cal.: Stanford University Press. Santamaria, M. 2000. "External Trade, Skill, Technology Tabor, S. R. 1995. "Structural Adjustment and InstitLo- and the Recent Increase of Income Inequality in Co- tional Change." In S. R. Tabor, ed., Agricultiiral Research lombia." Dissertation. Georgetown University, Wash- in an Era ofAdjustment. Washington, D.C.: World Bank. ington, D.C. Tan, Hong, and Geeta Batra. 1997. "Technology and Firn Sapir, A. 1996. "The Effects of Europe's Internal Market Size-Wage Differentials in Colombia, Mexico, and Program on Production and Trade: A First Assessment." Taiwan (China)." World Bank Economic Review 11 (I): Weltuwirtschafliches Archiv. 59-83. Sargent, John, and Linda Matthews. 1999. "Exploitation Taylor, J. Edward. 2001. "Microeconomics of Globaliza- or Choice? Exploring the Relative Attractiveness of tion: Evidence from Mexico, China, El Salvador, and the Employment in the Maquiladoras." Journal of Business Galapagos Islands." World Bank, Washington, D.C. Ethics 18: 213-27. Processed. Schott, Peter K. 2000. "Do Countries Specialize?" Yale Teixeira da Silva Filho, Tito Nicias. 2001. "Estimando o School of Management, New Haven, Conn. Processed. produto potencial brasileiro." Banco Central do Brasil. Serven, Luis. 1998. "Macroeconomic Uncertainty and Pri- Trabalhos para Discussao 17. vate Investment in LDCs: An Empirical Investigation." Terra, I. 1998. "Uruguay en el MERCOSUR: Perspectivas World Bank Policy Research Working Paper No. 2035. del Comercio Intraregional. " Centro Estudios para Amer- World Bank, Washington, D.C. ica Latina (CEPAL). Singer, Hans W. 1950. "The Distribution of Gains Be- Thom, Rodney, and Moore McDowell. 1999. "Measuring tween Investing and Borrowing Countries." The Ameri- Marginal Intra-Industry Trade." Weltuirtschaftliches Archh, can Economic Review 40 (2): 473-485. 135 (1): 48-61. Sistema para el Desarrollo Integral de la Familia. 2000. Tiano, Susan. 1994. Patriarchy on the Line: Labor, Gender, "Perfil Socieconomico de Jornaleros Migrantes en and Ideology in the Mexican Maquila Industry. Philadel- Sayula, Jalisco Mexico." Jalisco, Mexico. Processed. phia: Temple University Press. 176 BIBLIOGRAPHY Tomei, Manuela. 2000. "Home Work in Selected Latin Viner, Jacob. 1952. International Trade and Economic Devel- American Countries: A Comparative Overview." Interna- opment. Glencoe, Ill.: Free Press. tional Labour Organization, Geneva. Unpublished. Watkins, Melville. 1963. "A Staple 'T'heory of Economic Torres, Gabriel. 1997. "La Fuerza de la Ironfa." El Colegio Growth." The Canadian Journal of Economics and Political de Jalisco y Ciesas, Mexico. Science 29: 141-158. Trefler, Daniel. 1995. "The Case of the Missing Trade and Weller, Juirgen. 2000. Reformas econdmicas, crecimiento y em- Other HOV Mysteries." American Economic Review 85 pleo: los mercados de trabajo en America Latina y el Carihe. (5): 1029-46. Santiago, Chile: Fondo de Cultura Econ6mica/Comisi6n Umali, D. 1991. Public and Private Sector Roles in Agricul- Economica para America Latina y el Caribe. tulral Research: Theory and Evidence. Washington, D.C: Wheatly, Jeffrey J. 1999. World Telecommunication Economics. World Bank. Exeter, England: Short Run Press Ltd. UNIFEM (United Nations Development Fund for WIEGO. 2000. "Women in Informal Employment: Glob- Women). 2000. El Imipacto del TLC en la Mano de Obra alizing and Organizing." Women in the Informal Economwy. Femenina en AIe'xico. Ncw York: United Nations. www.wiego.org. Urrutia, Miguel. 1979. Cincuenta a&os de desarrollo econ6mico Will, Robert Milton. 1957. "Some Aspects of the De- colombiano. Bogota: La Carreta. velopment of Economic Thought in Chile (circa Vaillant, M. 2000. "Limits to Trade Liberalization: A 1775-1878)." Unpublished doctoral dissertation, Duke Political Economy Approach." Ph.D. Thesis. University University. of Pompeu Fabra. Wodon, Quentin. 1999. "Migration and Poverty in Latin Valdes, Alberto. 1986. "Impact of Trade and Macroeco- America and the Caribbean." World Bank, Washington, nomic Policies on Agricultural Growth: The South D.C. Processed. American Experience." In Economi- and Social Progress in . "Income Mobility and Risk During the Business Latin America, 1986 Report. Washington, D.C: Inter- Cycle." Economics of Transition 9 (2): 449-61. American Development Bank. Wodon, Quentin, with contributions from Robert Ayres, Van Beers, C. 1998. "Labor Standards and Trade Flows of M. Barenstein, Norman Hicks, K. Lee, William Mal- OECD Countries." World Economiiy 221 (1). oney, Pia Peeters, Corrine Siaens, and S. Yltzhaki. 2000. Vanek, Jaroslav. 1968. "The Factor Proportions Theory: Poverty and Policy in Latin America and the Caribbean. The N-factor Case." Kyklos 21 (4): 749-56. World Bank Technical Paper No. 467. World Bank, Velez, Carlos Eduardo, Jose Leibovich, Adriana Kugler, Washington, D.C. Cesar Boulill6n, and Jairo Nnfiez. 2001. "The Reversal Wodon, Quentin, and G. Gonzalez Koning. 2001. "Do of Inequality Trends in Colombia, 1978-1995: A Com- Cash Transfers to Farmers Reduce Migration? Procampo bination of Persistent and Fluctuating Forces." Inter- in Mexico." World Bank, Washington, D.C. Processed. American Development Bank and World Bank, Wash- Wodon, Quentin, G. Gonzalez Koning, and Corinne Siaens. ington, D.C. Processed. 2001. "Remittances and Housing, World Bank." World Venables, Anthony J. 2001. "Trade, Location, and Devel- Bank, Washington, D.C. Processed. opment: An Overview of Theory." London School of Wodon, Quentin, and M. Minowa. 2001. "Training for the Economics and CEPR/World Bank, Washington, D.C. Urban Unemployed: A Reevaluation of Mexico's Probe- Processed. cat." In S. Devaradjan, E Hasley and L. Squire, eds., Venezian, E., and E. Muchnik. 1994. "Structural Adjust- World Bank Economists' Forum. World Bank, Washington ment and Agricultural Research in Chile." Briefing D.C. Paper No. 9, International Service for National Agri- Wodon, Quentin, and S. Yitzhaki. 2001. "Evaluating the cultural Research. The Hague, Netherlands. Impact of Government Programs on Social Welfare: The Villalobos, Sergio, ed. 1990. "Historia de la Ingenieria Role of Targeting and the Allocation Rules Among Pro- en Chile." Insituto de Ingeneiros de Chile, Ediciones gram Beneficiaries." World Bank, Washington, D.C. Hachette. Santiago, Chile. Processed. 177 I ROM NATI RAL KESOUR(CES TO THE KNOWLEDGE ECONOMY TR SDE AND JO)B QUALITY WXodon, Quentin, R. Castro-Fernandez, G. Lopez-Acevedo, Wright, Gavin. 1999. "Can a Nation Learn? American Corrine Siaens, Carlos Sobrado, and Jean-Philippe Tre. Technology as a Network Phenomenon." In Naomi R. 2001a. "Poverty in Latin America: Trends (1986-1998) Lamoreaux, Daniel M. G. Raff and Peter Temin, eds., and Determinants." Caadernos de Economia 114: 127-154. Learning by Doing in Markets. Firrmls and ColJntries. Clii- Wlodon, Quentin, William Maloney, and Matias Baren- cago: National Bureau of Economic Research and Uni- stein. 2000. "Self-Employment as an Explanation for versity of Chicago Press. High Inequality in Latin America." World Bank, . 2001. "Resource Based Growth, Then and Now." Washington, D.C. Processed. Stanford University/World Bank, Washington, D.C. Wood, Adrian. 1994. "Give Heckscher and Ohlin a Processed. Chance!" Weltuirtschaftliches Archie, 130 (1): 20-48. NWylie, Peter J. 1990. "Indigenous Technological Adapta- Woodruff, Christopher. 2001. "Remittances and Micro- tion in Canadian Manufacturing, 1900-1929." Cana- enterprises in Mexico." Monterrey: Monterrey Institute dian Journal of Ec-onomics 2 3: 856-72. of Technology and Graduate Studies. Processed. Wynia, Gary W. 1990. "Opening Late-Industrializirlg Woodruff, Christopher, and Rene Zentano. 2001. "Remit- Economies: Lessons from Argentina and Australia." P1l- tances and Microenterprises in Mexico." University of icy Sciencies 23: 185-204. California, La Jolla, Cal., and Monterrey Institute of Yeats, Alexander J. 1998. "Does Mercosur's Trade Perfor- Technology, Guadalajara, Mexico. Processed. mance Raise Concerns About the Effects of Regional 'World Bank. 1995. Ii _. Developmient Report. Washington, Trade Arrangements?" XVorld Bank Economic Revieu 12 D.C. (1): 1-28. . 2000. Trade Blocs. New York: Oxford University Yusuf, Shahid. 2001. "Globalization and the Challenge Press. for Developing Countries." World Bank, Washington, ___. 2001. Global Economic Prospects and the Developing D.C. Processed. Countries 2002. Washington, D.C. 1 7 ANNEX Selected Figures from Chapters 2, 3, and 4 ANNEX SELEFiTLD FIGU1ILS FROM CHAPTERS 2, i. AND 4 FIGURE A. I Average LAC Net Exports per Worker by Commodity Groups, 1982-97 1 - ^ 1)8- 0.6 - GA -.4 _ _ * Chemicals 4 - Machines - 0.2 - Capital Intensive - 0 --L o - n Ln _-_ __ _ [_ rs_ _ r* - j -.- Labor Intensive _- 0 - Cereals -i~~~* - nimals -(.-- Tropical Agriculture -O. i s ~~~~~~~~~~~~~~~~~~~~~~~~~~~Forest -0.4 7 m-_ Raw Materials r Petroleum z -o.6 1982 L98S 198- 1985 1986 18)87 1988 1989 19)0 1991 1992 1093 19 ' S5 19)9 [Q97 Nore: Sample of 22 countries: Argentina, Barbados, Belize, Bolivia, Brazil, Chile, Colombia, Costa Rica, Ecuador, El Salvador, Guatemala. Honduras, Jamaica, Mexico, Nicaragua, Panama. Paraguay, Peru. the Repiblica Bolivariana de VenezLtela, Suriname, Trinidad and Tobago, and Uruguay. FIGURE A 2 Argentina: Structure of Net Exports, 1980-99 100 81() El Chiemicals u U~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ Machines L40 - U l Capital Intensive * Libsir Intensive 20 - - - Cereals * Animals 0 - M~~~~~~~~~~~~~~~~~~~~~~~~~~~~ Tropical Agriculture 0 Foresr -jo Raw NMaterials ED Petriileum -6o0 181 FROM NATURAL RESOURCES ro THE KNOWLEDGE ECONOMY TRADE AND JOB QUALITY FIGURE A.3 Brazil: Structure of Net Exports, 1981-99 1(( 0 8(0- 6t) - _ Chemicals *- 1 Machines 40 - L X Capital Intensive * Labor Incensive 20 - n Cereals E A4Animals tS 0 L L ; # # S H a W W # H W W | Tropical Agriculture U1 Forest --20 U Raw Materials lo --60 --80 FIGULRE Al4 Chile: Structure of Net Exports, 1981-98 .5 U Chemicals 13 Machines Z 20 - F- Caia IncensiveI I I I ELabor Intensive UFlCereals * Animals > 0~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ Tropical Agriculture t5 ~ ~ ~~~~~~~~~~~~~~~~~~~~~~~~~~~UForest * Raw Materials U] Pettoleum -4- - 0 ', N '- ' 010 1.,2 ANNEX: SELECTED FIGURES FROM CHAPTERS 2, 3, AND 4 FIGURE A.5 Costa Rica: Structure of Net Exports, 1981-99 so - 60 - 40m- 0 Chemicals * Machines 20 -0 - - -F Capital Intensive * Labor Intensive 0 Cereals o0- -- - E Anima,s * Tropical Agriculture FIGURE A.6 Dominican Republic: Structure of Net Exports, 1981-99 6~~~~d Mi For s°, -20 -t -9o -9 -i -92 -ot 9etO <>P 0 9 98 , t >\ 9 -40 - Domi- Da werepnotblic: Structure 1984, Net')8 , 198 1 aao t9te-99 1 Animals ~~~~~~~-4o *~~~~~~~~~~~~~~~~~~~~~~~~~~~- Tropical Agrenicutue 0l Forest i1' U~~~~~~~~~~~~~~~~~~~~~ Raw Materials -60-- 0 Petroleum -80 - _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ - I~~~~~~~~'li 11-11l91 ~ 1~ Note: Data were not avaiLable for 1984, 1986, 1987. and 1990-99. 18 3 FROM NATURAL RFSOLJ RCES TO T IF KNOt\S LED(;E E(CONONIY TRADE ANI) JO( Ql ALITY FIGURE A.- Mexico: Structure of Net Exports, 1981-99 10 - Fl