Making Global Value Chains Work for Development Making Global Value Chains Work for Development Daria Taglioni Deborah Winkler © 2016 International Bank for Reconstruction and Development / The World Bank 1818 H Street NW, Washington, DC 20433 Telephone: 202-473-1000; Internet: www.worldbank.org Some rights reserved 1 2 3 4 19 18 17 16 This work is a product of the staff of The World Bank with external contributions. The findings, interpretations, and conclusions expressed in this work do not necessarily reflect the views of The World Bank, its Board of Executive Directors, or the governments they represent. The World Bank does not 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 any judg- ment on the part of The World Bank concerning the legal status of any territory or the endorsement or acceptance of such boundaries. Nothing herein shall constitute or be considered to be a limitation upon or waiver of the privileges and immunities of The World Bank, all of which are specifically reserved. Rights and Permissions This work is available under the Creative Commons Attribution 3.0 IGO license (CC BY 3.0 IGO) http://creativecommons. org/licenses/by/3.0/igo. Under the Creative Commons Attribution license, you are free to copy, distribute, transmit, and adapt this work, including for commercial purposes, under the following conditions: Attribution—Please cite the work as follows: Taglioni, Daria, and Deborah Winkler. 2016. Making Global Value Chains Work for Development. Trade and Development series. Washington, DC: World Bank. doi:10.1596/978-1-4648-0157-0. License: Creative Commons Attribution CC BY 3.0 IGO Translations—If you create a translation of this work, please add the following disclaimer along with the attribution: This translation was not created by The World Bank and should not be considered an official World Bank translation. The World Bank shall not be liable for any content or error in this translation. Adaptations—If you create an adaptation of this work, please add the following disclaimer along with the attribution: This is an adaptation of an original work by The World Bank. Views and opinions expressed in the adaptation are the sole responsi- bility of the author or authors of the adaptation and are not endorsed by The World Bank. Third-party content—The World Bank does not necessarily own each component of the content contained within the work. The World Bank therefore does not warrant that the use of any third-party-owned individual component or part contained in the work will not infringe on the rights of those third parties. The risk of claims resulting from such infringe- ment rests solely with you. If you wish to re-use a component of the work, it is your responsibility to determine whether permission is needed for that re-use and to obtain permission from the copyright owner. Examples of components can include, but are not limited to, tables, figures, or images. All queries on rights and licenses should be addressed to the Publishing and Knowledge Division, The World Bank, 1818 H Street NW, Washington, DC 20433, USA; fax: 202-522-2625; e-mail: pubrights@worldbank.org. ISBN (paper): 978-1-4648-0157-0 ISBN (electronic): 978-1-4648-0162-4 DOI: 10.1596/978-1-4648-0157-0 Cover image: © Patrick Ibay. Used with permission. Further permission required for reuse. Cover design: Bill Pragluski, Critical Stages LLC Library of Congress Cataloging-in-Publication Data has been requested. CONTENTS Foreword xiii Preface xv Acknowledgments xvii About the Authors xix Abbreviations xxi Overview Making GVCs Work for Development 1 Introduction 1 Part I: Why GVCs Require Fresh Thinking 1 Part II: Quantifying a Country’s Position in GVCs 2 Part III: Strategic Questions and Policy Options 4 Part IV: Country Engagement 7 Note 7 References 7 PART I Why GVCs Require Fresh Thinking 9 Chapter 1 Here’s Why 11 Introduction 11 Firm and Policy Perspective 13 Evolution of GVC Trade 16 Assessment of a Country’s Potential in GVCs 20 Policy Dimension: Entering GVCs, Expanding Participation, and Ensuring Sustainable Development 25 Notes 30 References 32 Chapter 2 Consider Bulgaria 35 Introduction 35 Bulgaria’s Domestic Value Added in Exports 35 Channels for Increasing Domestic Value Added in Exports 36 Bulgaria’s GVC Participation and Firm-Level Productivity 45 What Must Be Done? 47 Notes 50 References 50 Annex 2A 51 v vi Contents PART II Quantifying a Country’s Position in GVCs 53 Chapter 3 What Do Imports and Exports Say about GVC Participation? 55 GVC Participation Using Gross Trade Data 55 Informed Classifications 58 GVC Participation Using Data on Trade in Value Added 62 Buying and Selling Sides 62 Notes 63 References 70 Chapter 4 Buyer-Related Measures 71 Introduction 71 Intermediates in Gross Imports 71 Imported Inputs Embodied in Gross Exports 71 Value Added in Gross Exports 73 Length of Sourcing Chains 80 Buyer Dimension: Summary 85 Notes 86 References 86 Chapter 5 Seller-Related Measures 87 Introduction 87 Intermediates in Output or Gross Exports 87 I2E Trade in Gross Exports 88 Domestic Value Added in Gross Exports of Third Countries 88 Who Are the Ultimate Consumers of a Country’s Value Added? Value Added in Final Domestic Demand 91 Length of Selling Chains: Distance to Final Demand 94 Domestic Gap between Buying and Selling Chains 97 Seller Dimension: Summary 98 Notes 99 References 100 Chapter 6 Other Measures of GVC Participation: From Macro to Micro 101 Introduction 101 GVC Participation Index 101 Network Metrics and Visualizations 104 Role of Services in Value Added 108 Main Actors and Their Links in GVCs Using Firm-Level Measures 112 Notes 115 References 116 Chapter 7 Use of GVC Measures to Assess the Drivers and Impacts of GVC Participation 117 Introduction 117 What Are the Determinants of GVC Links? 117 Do GVC Links Matter for Economic Upgrading? 121 Quantifying the Labor Market Dimension of GVCs 125 Summary 130 Annex 7A. Regression Results 132 Annex 7B. Factors Mediating Productivity Spillovers from Foreign Direct Investment 137 Annex 7C. Factors Mediating Productivity Spillovers from GVC Integration in Bulgaria 139 Notes 143 References 143 Contents vii PART III Strategic Questions and Policy Options 145 Chapter 8 Entering GVCs 147 Introduction 147 Attracting Foreign Investors and Facilitating Domestic Firms’ Entry into GVCs: Strategic Questions 147 Policy Options 161 Notes 174 References 175 Chapter 9 Expanding and Strengthening GVC Participation 179 Introduction 179 Promoting Economic Upgrading and Densification in GVCs: Strategic Questions 179 Strengthening Absorptive Capacity: Which Domestic Firm Characteristics Help Internalize Spillovers? 185 Policy Options 186 Notes 193 References 194 Chapter 10 Turning GVC Participation into Sustainable Development 199 Introduction 199 Promoting Social Upgrading and Cohesion: Strategic Questions 199 Promoting Environmental Sustainability: What Benefits from Environmental Regulation? 205 Policy Options 205 Notes 211 References 211 PART IV Country Engagement 213 Chapter 11 Designing a Country Engagement Strategy Based on Sound Analytics 215 What Is the Goal of This Guide? 215 Who Is This Guide For? 217 Steps in Component 1 219 Annex 11A Interview Guide for Fieldwork 223 Annex 11B Checklist of Topics from Combined Desk Research and Fieldwork 228 Appendixes 231 Appendix A Dimensions of GVC Participation: A Tentative Checklist 233 Appendix B Broad Economic Categories Classification 237 Appendix C Customized versus Generic Intermediates 239 Appendix D Parts and Components 241 Appendix E Value Chain Categories 243 Appendix F Sector and Product Clusters 245 Appendix G Main Types of Data Used to Measure GVC Participation 251 Appendix H The World Bank Export of Value Added Database 255 Appendix I Survey Year and Number of Domestic and Foreign Manufacturing Firms, by Country 259 Notes 261 References 262 viii Contents Boxes 1.1 Defining GVCs 12 1.2 The Disruptive Effects of Computer-Aided Technologies and Digital Innovation 19 1.3 What Is Special about Network Analysis? Finding Structure in Economic Problems 24 2.1 Methodology for the Identification of Key GVCs and Peer Countries 36 4.1 Choice of Comparator Countries 72 6.1 Why Firm-Level Analysis? 112 6.2 GVC Measures Based on World Bank Enterprise Surveys 114 7A.1 Data: GVC Indicators and Policy Variables 132 8.1 Network Analysis of a Product Value Chain Using I-O Tables 150 8.2 The Moroccan Aerospace Industry 154 8.3 Examples of Strategic Analysis and the Dynamics of Change Management: The Ventilation Industry and the Truck Cluster in Sweden 155 8.4 The Impact of Thailand’s 2011 Flooding 157 8.5 Why the Form of Governance Matters 158 8.6 Four Strategies to Widen Power Asymmetries in GVCs 160 8.7 Lessons from Failed Industrial Policies 163 8.8 Chile: ProChile Internationalization Plan 165 8.9 Case Study: Regulatory Reform and Infrastructure Building in Greek Logistics 173 9.1 The Czech Republic’s Supplier Development Program 188 9.2 Case Study: Renault-Dacia Regional Design and Development Activities in Romania 189 9.3 Own Design and Branding in Turkey 193 10.1 Succeeding in New Knowledge-Intensive Niche Sectors 206 10.2 Bangladesh’s Minimum Wage in the Apparel Industry 206 11.1 World Bank Group Approach to Diagnostic Work and Formulation of Action Plans to Strengthen a Country’s Position within Specific GVCs 216 G.1 Major Input-Output Databases 252 H.1 Value Added in Exports 258 Figures O.1 Strategic Policy Framework 5 1.1 Supplier-Buyer Links between China and Japan in the Automotive Industry 14 1.2 Stylized Facts about GVCs: A Multipolar World with Diverging Performances 17 1.3 Services Trade 18 1.4 Services Forward Links, 2007 19 1.5 Two Perspectives When Measuring GVC Participation 20 1.6 Two Perspectives When Measuring GVC Seller and Buyer Functions 21 1.7 GVC-Driven Development 22 1.8 From Sector to Functional Upgrading 23 1.9 Achieving Functional, Product, and Inter-sector Upgrading in All of a Country’s Agricultural, Manufacturing, and Services Production through Skills, Capital, and Process Upgrading 23 1.10 Network Representation of Value-Added Trade, 2011 25 1.11 GVC Transmission Channels 28 2.1 Growth of Domestic Value Added Embodied in Gross Exports, 1995–2011 36 2.2 Decomposition of Domestic Value Added Generated through Exports 37 2.3 Share of Firms Exporting Directly or Indirectly 37 2.4 Share of Total Inputs Sourced Locally 38 2.5 Share of Total Sales Exported Directly or Indirectly 38 2.6 Share of Firms Using Material Inputs/Supplies of Foreign Origin 38 2.7 Domestic and Foreign Value Added Embodied in Gross Exports: Bulgaria and Selected Countries, 1995 and 2008 39 Contents ix 2.8 Coefficients from Regression Results for Value-Added Components of Gross Exports, Overall Country Sample and Bulgaria 39 2.9 Coefficients from Regression Results for Value-Added Components of Gross Exports, by Sector, Overall Country Sample 40 2.10 Bulgaria’s Buying and Selling Patterns, 1995 and 2009 41 2.11 Final Demand, by Destination, Bulgaria and Peer Countries, 1995 and 2008 42 2.12 Domestic Value Added in Third Countries’ Gross Exports, Bulgaria and Peer Countries, 1995 and 2008 43 2.13 Import and Export Upstreamness and Gap, Bulgaria and Peer Countries, 2000 and 2012 44 2.14 Strategic Policy Framework Applied to Bulgaria 47 2.15 Doing Business Indicator: Overall and Protecting Investors, 2014 48 2.16 Logistics Performance Index: Overall and Customs Efficiency, 2014 49 2.17 Innovation Capacity and Skills, 2012 50 3.1 Malaysia’s Top 50 Exports and Imports, 2012 58 3.2 Most Relevant Buyers of Computer Storage Devices 61 3.3 Most Relevant Suppliers for Computer Storage Devices 61 3.4 Buying and Selling Patterns: Japan, 1995 and 2011 64 3.5 Buying and Selling Patterns: China, 1995 and 2011 65 3.6 Buying and Selling Patterns: Poland, 1995 and 2011 66 3.7 Buying and Selling Patterns: Mexico, 1995 and 2011 67 3.8 Buying and Selling Patterns: Germany, 1995 and 2011 68 3.9 Buying and Selling Patterns: United States, 1995 and 2011 69 4.1 Intermediate Imports, 1996–2012 72 B4.1.1 Country Positioning in the Global Economic Space 73 4.2 Countries’ Integration in GVCs: Share of Intermediate Imports in Gross Imports and Electrical and Electronics, 2009 and 2012 73 4.3 Intermediate Imports Embodied in Exports and Electrical and Optical Equipment, Selected Countries, 2009 and 2011 74 4.4 Re-imports and Re-exports in Supply Chain Trade 75 4.5 Decomposition of Gross Exports 76 4.6 Quantifying the Value Added of Gross Exports 76 4.7 Decomposition of Gross Exports, Selected Countries, 2011 77 4.8 Foreign Value Added in a Country’s Gross Exports, 1995 and 2011 77 4.9 Comparison of Four Buyer-Related Measures of GVC Participation, Selected Countries, 2009 78 4.10 Value-Added Trade: US$10 Million in Mexican Car Exports to the United States 79 4.11 Foreign Value Added in Thailand’s Transport Equipment Sector Exports, by Source Region, 1995–11 80 4.12 Foreign Value Added in Gross Exports and Electrical and Optical Equipment, by Source Region, Selected Countries, 2011 80 4.13 Foreign Value Added in Gross Exports and Electrical and Optical Equipment, Selected Countries, 2011 81 4.14 Foreign Value Added in Gross Exports in the U.S. Electrical and Optical Equipment Sector, by Source Industry, 2011 81 4.15 Foreign Value Added in Gross Exports, Electrical and Optical Equipment, by Source Sector, Selected Countries, 2011 82 4.16 Foreign Value Added in Gross Exports, Chemicals and Chemical Products, by Source Sector, Selected Countries, 2011 83 4.17 Length of Sourcing Chains, by Industry, 2008 84 5.1 Intermediate Exports, 1996–2012 88 5.2 Domestic Value Added Embodied in Third Countries’ Exports, 1995 and 2011 89 5.3 Domestic Value Added Embodied in Third Countries’ Exports, Excluding Mining and Quarrying and Coke, Refined Petroleum Products, and Nuclear Fuel, 2011 90 x Contents 5.4 Domestic Value Added Embodied in Third Countries’ Exports, Global Market Share, Selected Countries, 1995 and 2011 91 5.5 Domestic Value Added Embodied in Third Countries’ Exports, Global Market Share, Chemicals and Chemical Products and Electrical and Optical Equipment, Selected Countries, 1995 and 2011 91 5.6 RCA in Chemicals and Chemical Products and Electrical and Optical Equipment, 1995 and 2011 92 5.7 Domestic Value Added in Gross Exports, by Destination Region, 1995 and 2011 92 5.8 Domestic Value Added in Gross Exports in Electrical and Optical Equipment, by Destination Region, 1995 and 2011 93 5.9 Domestic Value Added in Gross Exports in Chemicals and Chemical Products, by Destination Region, 1995 and 2011 93 5.10 The Ultimate Consumers of a Country’s Export Value Added 94 5.11 Domestic Value Added Embodied in Foreign Final Demand, 2011 95 5.12 Foreign Value Added Embodied in Domestic Final Demand, 2011 95 5.13 Domestic Value Added in Foreign Final Demand, Top Five Partner Shares, Selected Exporters, 2011 96 5.14 Upstreamness of Industries in Malaysia 97 5.15 Upstreamness in Malaysia and Comparators, 2012 and Progression Since 2000 97 5.16 Import Upstreamness, Export Upstreamness, and Domestic Gap, Malaysia, 2000–13 98 6.1 GVC Participation Index, 2011 102 6.2 GVC Participation Index, Excluding Mining and Quarrying and Coke, Refined Petroleum Products, and Nuclear Fuel in the Numerator, 2011 102 6.3 GVC Participation Index, Malaysia and Peer Countries, Chemicals and Chemical Products, 2011 103 6.4 GVC Participation Index, Selected Countries, Electrical and Optical Equipment, 2011 103 6.5 Breakdown of Malaysia’s GVC Participation Index, 2011 103 6.6 Evolution of the Network of Value-Added Trade, 1995 and 2011 104 6.7 Buyer and Seller Perspectives, 2011 105 6.8 World Gross Trade Network for Apparel, 2013 108 6.9 Bangladesh’s Gross Trade Network: Main Buyers of Bangladeshi Apparel (Cotton) Consumption Products, 2013 109 6.10 Domestic Value Added of Services Sectors Embodied in Manufacturing Gross Exports, All Countries, 2009 110 6.11 Contribution of Services Sectors to Export Value Added of Goods Sectors in Morocco, 2007 111 6.12 Forward and Backward Links in Export Value Added, Trade and Transport Services, 2001 and 2007 111 6.13 Forward and Backward Links in Export Value Added, Other Private Services, 2001 and 2007 112 6.14 Input Sources of Multinationals in Agribusiness, 2012 113 6.15 Domestic Suppliers’ Output Sold to Multinationals in Agribusiness, 2012 115 6.16 Sales Channels of Domestic Suppliers in Agribusiness, 2012 115 7.1 GVC Integration and Overall Logistics Performance Indicator, 2008 119 7.2 GVC Integration and Skill Levels, 2008 119 7.3 GVC Integration and Geographical Distance to the Closest Knowledge Center (Germany, Japan, and the United States), 2008 120 7.4 Labor Value Added in Chinese Machinery and Equipment Exports, 1995–2011 126 7.5 Labor Value Added in Indian Other Private Services Exports, 1995–2011 127 7.6 GVC Participation and the Labor Component of Domestic Value Added in Exports 128 7.7 Growth in the Labor Component of Domestic Value Added in Exports by Level of GVC Participation Growth and Foreign Value Added 128 Contents xi 7.8 Jobs in the Business Sector Sustained by Foreign Final Demand, 1995 and 2008 129 8.1 Malaysia: RCA, Gross Exports, and Domestic Value Added Embodied in the Country’s Gross Exports, 2009 149 B8.1.1 Manufacturing Inter-Sector Links, NAICS 31–33, 2007 151 B8.1.2 Most Relevant Buyers of Computer Storage Devices 152 B8.1.3 Most Relevant Suppliers for Computer Storage Devices 152 B8.1.4 Computer Storage Devices Network for Malaysia 153 B8.1.5 Malaysia as an Importer of Downstream Products 153 B8.1.6 Malaysia as an Exporter of Downstream Products 153 B8.2.1 Upward Mobility: Approximate Employment in the Moroccan Aerospace Industry 154 8.2 Five GVC Governance Structures 159 8.3 Logistics Services in a Typical Supply Chain 166 8.4 Reducing Supply Chain Barriers: Impact on GDP and Trade Growth 167 8.5 Services Involved in the Internationalization of Production (at Sandvik Tooling) 172 9.1 Example of Possible Inter-Sector Upgrading in Nicaragua 181 9.2 Standards in Agrifood GVCs 190 9.3 Diffusion of Standards and Other Codes of Conduct in GVCs 191 9.4 Tasks Performed by Apparel Industries in Torreon, Mexico 191 10.1 Social Cohesion as an End of, and a Means for, Development 200 10.2 Social “Grading” of Jobs 201 10.3 Upgrading and Downgrading 202 10.4. Economic and Social Upgrading and Downgrading in Apparel, 1990s to 2000s 203 B10.2.1 Minimum Wage per Month for Selected Countries 206 Tables O.1 Selected Policy Objectives and Performance Indicators by Focus Area 6 2A.1 Bulgaria’s Position in the Global Network of Trade in Value Added, 2008 51 2A.2 Logistics Performance, Domestic Component, 2014 52 3.1 Malaysia’s Top 50 Exports, 2012 56 3.2 Malaysia’s Top 50 Imports, 2012 57 3.3 Turkey’s Share of Exports and Value Added, 2003 and 2010 59 3.4 Auto Cluster 60 4.1 Indicators of Value Added Embodied in Gross Exports 79 4.2 Summary of the Main Buyer-Related Measures 85 5.1 Indicators of Value Added Embodied in Final Demand 96 5.2 Summary of the Main Seller-Related Measures 99 6.1 Network Measures 106 6.2 Network Measures, All Sectors, E&E, and Chemicals, 2009 107 6.3 Indicators of Services Value Added 110 7.1 Jobs Generated by Five Components of Foreign Trade, 2009 130 7.2 Manufacturing GVC Workers, 1995 and 2008 131 7A.1 GVC Integration as a Buyer and Domestic Value Added, National Characteristics, 1995–2011 133 7A.2 GVC Integration as a Buyer and Domestic Value Added, National Characteristics, Selected Years, 1995–2011 134 7A.3 GVC Integration as a Seller and Domestic Value Added and the Role of National Characteristics, 1995–2011 135 7A.4 GVC Integration as a Seller and Domestic Value Added, National Characteristics, Selected Years, 1995–2011 136 7C.1 Structural Integration in GVCs from a Buyer’s Perspective and Its Impact on Productivity, the Role of Absorptive Capacity, Manufacturing Firms, OLS 139 xii Contents 7C.2 Structural Integration in GVCs from a Seller’s Perspective and Its Impact on Productivity, the Role of Absorptive Capacity, Manufacturing Firms, OLS 139 7C.3 Structural Integration in GVCs and Its Impact on Productivity, the Role of Absorptive Capacity, Manufacturing Firms, OLS 140 7C.4 Structural Integration in GVCs from a Buyer’s Perspective and Its Impact on Productivity, the Role of National Characteristics, Manufacturing Firms, OLS 141 7C.5 Structural Integration in GVCs from a Seller’s Perspective and Its Impact on Productivity, the Role of National Characteristics, Manufacturing Firms, OLS 142 8.1 Addressing Obstacles at the Border: Policy Objectives and Performance Indicators 168 8.2 Increasing the Connectivity of Domestic Markets: Policy Objectives and Performance Indicators 169 8.3 Improving Cost Competitiveness While Avoiding the Trap of Low-Cost Tasks: Policy Objectives and Performance Indicators 170 8.4 Improving Drivers of Investment: Policy Objectives and Performance Indicators 171 8.5 Encouraging and Protecting Foreign Investment: Policy Objectives and Performance Indicators 172 8.6 Improving Domestic Services Infrastructure and Market Structure: Policy Objectives and Performance Indicators 173 9.1 Fostering Innovation and Building Capacity: Policy Objectives and Performance Indicators 189 9.2 Improving Standards: Policy Objectives and Performance Indicators 191 10.1 Promoting Social Upgrading: Policy Objectives and Performance Indicators 208 10.2 Engineering Equitable Distribution of Opportunities and Outcomes: Policy Objectives and Performance Indicators 210 11.1 Desk-Based Analysis 218 11.2 Stakeholders to Target during Fieldwork 220 A.1 A Multidimensional Checklist of a Country’s Participation in GVCs 234 B.1 Broad Economic Categories Classification 237 C.1 Customized Intermediates in the Apparel and Footwear Sector 239 D.1 List of Manufacturing Parts and Components 241 E.1 Assignment of Products to Five Value Chain Categories in Five Main GVC Sectors 243 F.1 GVC Clusters 245 BG.1.1 International Input-Output Databases 252 H.1 Turkey’s Exports, Gross and Value-Added Measures, by Sector, 2007 258 I.1 Survey Year and Number of Domestic and Foreign Manufacturing Firms, by Country 259 I.2 Survey Year and Number of Domestic and Foreign Manufacturing Firms, Selected Countries 261 FOREWORD The global value chain (GVC) revolution has trans- are struggling with the challenge of making GVCs formed trade, leading to changes in trade-growth- work better for their national development strate- development links, trade-competitiveness links, and gies. Other low- and middle-income nations—espe- trade-governance options. In my view, twentieth cially in Africa and South America—still view GVCs century globalization is about made-here-sold-there as some sort of trap, creating a new core-periphery goods crossing borders: the trade system helped pattern with “good” jobs in the North and “bad” jobs nations sell things. But twenty-first century global- in the South. Yet even the most reluctant are com- ization is also about factories crossing borders, so ing around to the idea that the success of nations intra-factory flows of goods, know-how, investment, such as China in the GVC competition means that training, ideas, and people are now international all other low- and middle-income nations have to commerce. The trade system helps nations make face the sort of competition that comes when GVCs things, not just sell things. combine high-tech with low wages. In essence, GVCs GVCs also denationalized comparative advan- killed import substitution as a viable industrializa- tage, and that changed the options facing all nations. tion strategy, so that pursuing strategies that nations Instead of building the whole chain domestically to such as the Republic of Korea and the United States become competitive internationally (the twentieth pursued in the past became almost unthinkable. In century way), in the twenty-first century, low- and that domain, the book is extremely welcome. The middle-income nations join GVCs to become com- GVC revolution requires fresh thinking; twentieth petitive and then industrialize by densifying their century paradigms are insufficient or misleading participation. The flip side is that the competitive- when applied to twenty-first century challenges. This ness options of high-income nations have changed. book is a solid step in that direction. Much research Globally competitive firms knit together national remains to be done, but the book will help govern- comparative advantages to make components in ments—and policy scholars—understand the issues. the most cost-effective location. Firms and nations The basic structure of the book is well thought that eschew GVCs must struggle to compete. In through. short, GVCs killed import substitution for low- and Part I introduces key concepts to provide an middle-income countries and naively nationalistic accessible and highly logical framework for think- industrial policies for high-income countries. ing about GVCs and—importantly—for why GVCs Making Global Value Chains Work for Development require new thinking. That is a key element, because is very timely in that those facts are now coming I find that many policy makers in low- and middle- into focus in the global discussion on development. income countries (and many academics in high- Some low- and middle-income nations—for exam- income countries) view GVCs as just a new buzzword ple, most East and Southeast Asian economies— for rationalizing old policy ideas. It is essential to get have fully embraced the GVC revolution, but they this message out, so that governments will stop using xiii xiv Foreword old analytics to think through new challenges. Firms Part III is less well developed simply because in all nations are much further along in view of the the research does not exist to support a diagnostic changes, but they do not really have a way to con- approach to policy. In the economic literature, a ceptualize them simply. The first chapter will help on great deal of storytelling and macro data purport to both scores. show that nations participating in GVCs are seeing Part II provides a review of the many concepts faster growth and expanding exports on the intensive and measurement tools that have been discussed and extensive margins, but we do not really know over the 20 years or so since GVCs really took off. In enough to guide policy makers’ decisions on exactly the past three or four years, the range of GVC mea- what to do. surements exploded with new data sets, including the Overall, this book is an excellent product. It is too Trade in Value Added data set of the Organisation for early to write a definitive work on GVCs and devel- Economic Co-operation and Development and the opment. My guess is that at least a decade of research World Input-Output Database. The critical concepts will be necessary to reach that point. But govern- used in those data sets are a bit tricky, because they ments face challenges that must be met today. This are so far from the standard, black box/production- book is an excellent contribution to making such function approach to trade. Again, this book pro- decisions on a more solid, evidence-based founda- vides a good, accessible introduction to the measures tion. I wholeheartedly commend it. and how they compare. Richard Baldwin Professor, International Economics, Graduate Institute, Geneva, Switzerland Director, Centre for Economic Policy Research, London, England Founder and Editor-in-Chief, VoxEU.org PREFACE From banana chips to computer chips, the way the and better wages? For which type of countries are global economy produces and exchanges goods export-processing zones a viable tool of industrial- has never been more dynamic or more intercon- ization? Will firms in those zones actually generate nected. The fragmentation of production across more spillovers than those outside the zones? global chains and the importance of foreign inputs This book presents a crucial starting point for in virtually all sectors affect everybody: participants, applying fresh thinking to the GVC revolution and nonparticipants, and countries at all income and its implications for policy and development. It does development levels. Increasingly complex global so by providing three main contributions to the cur- value chains (GVCs) are a dominant economic reality rent debate on GVCs. First, it provides a framework in the twenty-first century. They present critical new for more easily conceptualizing GVCs and thus, for challenges to the ways of evaluating and improving a more structured discussions and debates on GVCs country’s trade and competitiveness. and their implications for development. Second, it This book comes at the perfect moment for low- serves as a repository of analytical tools—on which and middle-income countries seeking to join or the World Bank Group will work to expand as new upgrade in GVCs. Until now, the development com- tools become available. Third, it is a collection of munity has had a very emulative, unidirectional dis- best-practice policies illustrated through case stud- course. A narrow focus on the success stories among ies, which will also be expanded to include evidence- GVCs has resulted in policy prescriptions that too based data. All this is accomplished through an often seek to make each country the next Singapore; innovative mix of methodologies from the economic that simply will not suffice. Over the past few years, and business school literature, embracing top-down as some of the initial success stories—such as Ireland and bottom-up approaches. or even my home country, Costa Rica—have come I see this work as the spearhead of the World Bank to face challenges in the sustainability of their posi- Group’s newly established Trade and Competitive- tion in GVCs, questions and concerns rightly have ness Global Practice effort to lead the intellectual and been raised. policy agenda on GVCs. It is a promising first step for In light of the new reality of GVCs, a thorough better understanding the role of GVCs in economic review of tools and policies is in order. The time has development in the twenty-first century—especially come to reevaluate conventional wisdom. How can the impact of GVCs on increasing the prosperity of the risk of investment attraction policies be more the bottom 40 percent of global citizens. I strongly accurately assessed? What might their impact be on believe that continuing to develop innovative tools is domestic investors? What are the inherent tensions not only necessary but essential. Now is the time for between GVC attraction strategies—often based on questions, reflections, and nuances—and that is what low wages—and achieving higher labor productivity this work brings. Anabel González Senior Director, Trade and Competitiveness Global Practice World Bank Group xv ACKNOWLEDGMENTS Making Global Value Chains Work for Development consultation on the overall narrative of the book, was prepared by Daria Taglioni (Task Team and to Gary Gereffi on specific concepts. Leader) and Deborah Winkler, under the guidance Special thanks go to our peer reviewers, Paulo of Anabel González, Senior Director of the World Correa (Lead Economist, Trade and Competitive- Bank Group’s Trade and Competitiveness Global ness Global Practice), Emiliano Duch (Lead Finance Practice. The book is part of a broader, multiyear and Private Sector Development Specialist, Trade work program of the Trade and Competitiveness and Competitiveness Global Practice), Frederico Gil Global Practice, which is aimed at offering a com- Sander (Senior Country Economist, Macroeconom- prehensive framework and analytical instruments ics and Fiscal Management Global Practice), and that can be used to undertake a systematic assess- Loraine Ronchi (Lead Economist and Global Agri- ment of a country’s competitiveness and trade business Lead, Trade and Competitiveness Global performance. The book draws on contributions, Practice), as well as peer reviewers from the World background notes, and discussions from a variety Trade Organization and the Organisation for Eco- of experts. These include Olivier Cattaneo (poli- nomic Co-operation and Development (OECD) cies for entering and strengthening participation in (Development Centre; Statistics Directorate; Direc- global value chains), Massimiliano Calì and Alen torate of Science, Technology, and Industry; and Mulabdic (labor content of exports), Claire Hollweg Trade and Agriculture Directorate), and the OECD (measures of social upgrading and computations Initiative for Policy Dialogue on Global Value of some of the measures in part II), Asier Mariscal Chains, Production Transformation, and Develop- (skills and transmission of knowledge), Miles ment, which has been provided without explicitly McKenna (environmental sustainability), Anasuya identifying the names of the authors. Raj (methodology for the selection of compara- The authors would also like to thank other tor countries), Gianluca Santoni (measures of net- experts and consultants who contributed valuable work analysis), as well as Emiliano Duch, Thomas discussions, comments, and other inputs, includ- Farole, Sumit Manchanda, Syed Akhtar Mahmood ing Guillermo Arenas, Jean François Arvis, Bertram (specific aspects of country engagement strategies), Boie, Christina Busch, Ana Paula Cusolito, Roberto Laura Alfaro, Pol Antras, Paola Conconi, Ana Paula Echandi, Thomas Farole, Michael Ferrantino, Mona Cusolito, Thomas Farole, Ana Margarida Fernandez, Haddad, Eric van der Marel, William Milberg, Jose Jan de Locker, Asier Mariscal, Timothy Sturgeon, Daniel Reyes, Frank Sader, and Sebastian Saez, Ezequiel Zylberberg, and the International Finance as well as participants in the CompNet Research Corporation/International Labour Organization Network of the Eurosystem, OECD. Better Work Programme (key issues to cover in firm During the development and piloting of parts of interviews and surveys). The authors are grateful the book in different countries and regions across to Richard Baldwin for his invaluable expertise and the world, including Bulgaria, Cambodia, China, xvii xviii Acknowledgments Malaysia, Morocco, Poland, South Africa, Turkey, World Economic Forum for permission to use their and Vietnam, the authors benefited from the sup- material. port, comments, and suggestions of colleagues at The authors are grateful to Communications the World Bank Group, including Enrique Aldaz- Development Incorporated, led by Bruce Ross- Carroll, Fabio Artuso, Jean-Pierre Chauffour, Julian Larson, and Sandra Gain for editing the book; Paola Latimer Clarke, Doerte Doemeland, Thomas Farole, Scalabrin, Susan Graham, and Denise Bergeron Frederico Gil Sanders, Mariem Malouche, Sandeep from the World Bank’s publishing and knowledge Mahajan, Kamer Karakurum Ozdemir, Catriona unit for overseeing the publication and dissemina- Mary Purfield, Richard Record, Jose Guillherme Reis, tion process; Patrick Ibay for providing the overall Martin Reiser, Emilia Skrok, and Chunlin Zhang. layout and graphical design of the figures and the Thanks also go to Premachandra Athukorala, graphical concept of the cover. Amir Alexander Xiao Jiang, Sebastien Miroudot, Timothy Fouad and Patrick Ibay provided support during the Sturgeon, and Marcel Timmer, as well as the Asian book’s preparation and Patrick also coordinated the Development Bank; the Center on Globalization, publication process. Governance, & Competitiveness; Japan Automobile This project was supported in part by the govern- Manufacturers’ Association; Kommerskollegium; ments of Finland, Norway, Sweden, and the United OECD; Proceedings of the National Academy of Kingdom through the Multi-Donor Trust Fund for Sciences of the United States of America; and the Trade and Development. ABOUT THE AUTHORS Daria Taglioni Daria Taglioni is a lead economist and the Global Solutions Lead on Global Value Chains in the Trade and Competitiveness Global Practice of the World Bank Group. Ms. Taglioni’s published work in economic pol- icy analysis covers topics in international trade and finance, including countries’ competitiveness in the global economy and the relationship between financial markets and performance. Ms. Taglioni is the author of the book Valuing Services in Trade: A Toolkit for Competitiveness Diagnostics (with Sebastian Saez, Erik van der Marel, Claire Hollweg, and Veronika Zavacka). Her articles have appeared in peer-reviewed journals, including the Journal of International Economics, Economic Policy, Journal of Banking and Financial Economics, Journal of Economic Integration, World Economy, Emerging Markets Review, European Economy, OECD Journal, and Journal of Financial Transformations, as well as in edited volumes by the World Bank Group, Centre for Economic Policy Research, European Central Bank, and Organisation for Economic Co-operation and Development (OECD). Before joining the World Bank Group, Ms. Taglioni worked at the European Central Bank and the OECD. She holds a PhD in international economics from the Graduate Institute Geneva. Deborah Winkler Deborah Winkler is a senior consultant in the World Bank Group’s Trade and Competitiveness Global Practice. Deborah has worked on issues of global value chains, export competitiveness, foreign direct investment, trade in services, and the welfare effects of international trade and offshoring. She is the author of Outsourcing Economics: Global Value Chains in Capitalist Development (with William Milberg) and Services Offshoring and Its Impact on the Labor Market. Ms. Winkler is the editor of Making Foreign Direct Investment Work for Sub-Saharan Africa: Local Spillovers and Competitiveness in Global Value Chains (with Thomas Farole). Her recent articles have appeared in World Development, Journal of Economic Geography, and World Economy, as well as edited volumes by the World Bank Group, Centre for Economic Policy Research, International Labour Office and World Trade Organization, and Oxford Handbook Series. She received her PhD in economics from Hohenheim University in Germany. xix ABBREVIATIONS ASEAN Association of Southeast Asian Nations BEA Bureau of Economic Analysis (United States) BEC Broad Economic Categories BONwin inflows of value added BONwout outflows of value added BPO business process outsourcing CCw clustering index CI CzechInvest cm centimeter CRI competitiveness reinforcement initiative E&E electrical and electronics E&O electrical and optical EPZ export processing zone EU European Union EXIOPOL Environmental Accounting Framework Using Externality Data and Input–Output Tools for Policy Analysis FAO Food and Agriculture Organization (of the UN) FDI foreign direct investment g gram GDP gross domestic product GTAP Global Trade Analysis Project GVC global value chain HS Harmonized System HSE health, safety, and environmental I2E import to export ICIO inter-country input-output ICT information and communications technology IDE Institute of Developing Economies IFC International Finance Corporation IKTIB Istanbul Textile and Apparel Exporter Association xxi xxii Abbreviations ILO International Labour Organization I-O input-output IPR intellectual property rights ISIC International Standard Industrial Classification IT information technology JETRO Japan External Trade Organization JV joint venture KPI key performance indicator LACEX Labor Content of Exports LMIC low- and middle-income country LP logistics performance LPI Logistics Performance Index m meter M&A merger and acquisition M&E monitoring and evaluation MC-GVC Measuring Competitiveness in Global Value Chains MF mediating factor MNC multinational corporation MNE multinational enterprise MRIO multi-regional input-output tables NACE General Industrial Classification of Economic Activities within the European Communities NAFTA North American Free Trade Agreement NAICS North American Industry Classification System nec not elsewhere classified NEM nonequity mode of investment nesoi not elsewhere specified or included OECD Organisation for Economic Co-operation and Development OEM original equipment manufacturer R&D research and development RCA revealed comparative advantage RMG ready-made garment RTR Renault Technologie Roumanie SEZ special economic zone SITC Standard International Trade Classification SMEs small and medium enterprises TiVA Trade in Value Added UN United Nations UNCTAD United Nations Conference on Trade and Development WDI World Development Indicators WIOD World Input-Output Database WTO World Trade Organization Overview MAKING GVCS WORK FOR DEVELOPMENT Introduction unprecedented skill and knowledge transfer, but also with important distributional consequences, as Making Global Value Chains Work for Development world income shifts toward countries involved in provides a framework, analytical tools, and policy GVCs and a “superstar effect” is generated in many options. The book shows why global value chains countries, with the risk of growing inequalities (GVCs) require fresh thinking. It presents a meth- domestically. odology for quantifying the extent of a country’s Internationally fragmented production is not participation in GVCs, based on available data. It new. For decades, low- and middle-income coun- also proposes a strategic framework to guide policy tries (LMICs) have imported parts from countries makers in identifying the key objectives of GVC par- with more advanced technology. But generally these ticipation and development and in selecting suitable imports were only for the assembly of locally sold economic strategies to achieve them. goods. Because the goods produced were not part of a global network, flows of know-how and the rate of Part I: Why GVCs Require Fresh Thinking technology transfer were less intense. And because there were fewer opportunities to buy and sell on Part I begins by asserting that the economic implica- global markets, the push to improve productivity tions of GVCs must be rethought for the twenty-first was also less strong. century. GVCs entail four key features that set them The new characteristic of GVCs from a develop- apart from traditional production and trade. These ment perspective is that factories in LMICs have are customization of production, sequential produc- become full-fledged participants in international tion decisions going from the buyer to the suppliers, production networks, and this fact can present high contracting costs, and global matching not only important development prospects. of goods and services, but also production teams and ideas (Antràs 2015). • LMICs no longer are just importing parts for Goods and services produced in GVCs are fre- assembly for local sales; they are absorbing valu- quently customized to the needs of their intended able foreign technology and know-how and buyers. Customization, in turn, entails sequential importing inputs that they process and export production and sales decisions that go from the final in the form of goods, parts, components, and buyer backward to the producers of upstream inputs. services used in some of the most sophisticated Global production of customized goods and services products today. Baldwin and Lopez-Gonzalez also entails intensive contracting between parties, (2013) call this process import to export, or I2E. often subject to distinct legal systems. Because of • LMICs no longer have to master the entire pro- the fragmented contracting environment, a signifi- duction process of a good. They can specialize in cant share of GVC trade is intra-firm.1 GVCs also only a segment of the international production lead to matching production teams globally, with process while reaching sufficient production scale 1 2 Making Global Value Chains Work for Development to meet their bottom line, thanks to the access to well-targeted policies. From a policy perspective, the global markets. critical issue is how GVCs integrate into the economy • As firms from different countries no longer just as a whole. If GVCs remain de-linked from the local trade goods and services, but work together in context, lead firms will keep driving most decisions vertically integrated systems of production, shar- and governments may have limited influence and ing blueprints, technicians, managerial practices, ability to leverage these decisions for domestic eco- and productivity-enhancing tools and techniques, nomic development. The policy challenge therefore GVCs provide access to “accelerated learning” and extends to creating and strengthening links with transfer of tacit knowledge, at a rate unthinkable domestic firms and ensuring that the host country in a traditional trade setting. benefits from technology transfers, knowledge spill- • With GVCs, the range of actors in international overs, and increased value addition in the country. trade and production has expanded. South-South It is equally important to ensure that GVC partici- trade flows and investment are rapidly gain- pation benefits domestic society through more and ing importance. Outward foreign direct invest- better paid jobs, better living conditions, and social ment (FDI) by the BRICS (Brazil, the Russian cohesion. In a nutshell, the key question is, how can Federation, India, China, and South Africa) rose LMICs make GVCs work for development? from US$7 billion in 2000 to US$145 billion in To exemplify how this book can help policy mak- 2012 and US$200 billion in 2013, that is, almost ers find answers to that challenge, part I ends with one-third of global FDI (Gómez-Mera and oth- a case study of Bulgaria. The study shows how ana- ers 2015). It is not just manufacturing. In coun- lysts can make use of the quantitative tools described tries such as India, the Philippines, and Vietnam, in part II of the book, as well as the strategic policy to name just a few, dynamic knowledge intensive framework developed in part III, to identify a coun- software and business services have emerged and try’s position in GVCs, its scope for upgrading, and are showing strong growth. And it is not just the policies that can help achieve that goal. Finally, part story of large multinationals anymore. The frag- IV closes the book, offering guidelines on how to mentation of production together with advances design and implement a national strategy to achieve in information and communications technology GVC-led development. The guidelines are based on are creating new entrepreneurial possibilities for experience on the ground by World Bank Group small and medium enterprises to access markets teams. abroad, giving rise to a new category of so-called micro-multinationals, which are small firms that Part II: Quantifying a Country’s Position develop global activities from their inception in GVCs (Mettler and Williams 2011). • Participation is not a given but raises new chal- Integrating a country’s domestic suppliers into lenges. Competition is fierce and LMICs face a set GVCs increases the possibility for GVC spillovers of challenges to enter international production, through exporting to a buyer abroad or supplying upgrade to higher value-added products, tasks, to a multinational in the country. But countries and sectors, and ensure social upgrading and should also consider from a buyer’s perspective cohesion from participation. the opportunities that GVC participation can pro- • Countries are also faced with new policy trade- vide. Firms can join existing global and regional offs. High growth and development potential are value chains through importing parts and compo- associated with exposure to the increasing com- nents that are used in production at home without plexity and uncertainty that is associated with the need to build a complete array of value chains organizing production across several locations. at home. In the past, for a country to become an apparel exporter, for example, it would need design Opening borders and attracting offshore facto- capabilities and textile mills; to export in the auto- ries is important, and those steps help jump-start motive sector, it would need to produce engines and entry into GVCs. But retaining GVCs, maximizing all the subcomponents, and be able to produce on their benefit to the domestic economy, and ensur- the scale necessary to compete with foreign produc- ing their sustainability require well-designed and ers. Under the new trade dynamics, a country can Overview: Making GVCs Work for Development 3 specialize in certain activities (for example, sewing, • Which are the source countries of foreign value specific components, or subassemblies) and import added that are used as an input in the exports by the balance of manufacturing needs. Although such the country of interest? a situation does not guarantee significant value cap- • Which countries are the final consumers of ture and upgrading from inception, it does provide domestic value added embodied in the exports by a vital first step toward producing world-class, high- the country of interest? quality goods and services. Nowhere is that more • What is a country’s performance and what are evident than in China, and more widely across East its main functions in GVCs (buyer or seller; pre- Asia, where GVCs are at the heart of the open-econ- dominantly headquarters, factory, or assembly omy growth model that has been responsible for the location)? growth and poverty reduction success story of the • Are domestically owned firms well integrated in region in recent decades. GVCs? Quantitative measures of GVC participation and • Through which channels and in which sectors do guidelines for analysis make it possible to deliver domestic value-added products and services con- informed policy suggestions. Correct identification tribute most to the country’s gross exports? of constraints and remedial actions, and assessment • Does the position (upstream/downstream) of of the efficacy of new policy measures cut across the the country’s and sector’s participation in GVCs gamut of the statistical information system, includ- matter for domestic value added and growth of ing macro and, crucially, micro (firm-level) data. The domestic value added? organizing framework and indicators in part II make • What is the impact of GVC participation on task it possible to answer questions related to a country’s trade (goods and services) and the factors of pro- GVC participation and the economic and social gains duction (such as workers, ideas, and investments)? from such participation. The key takeaway from part • Is GVC participation creating positive effects and II is that a sound analysis of countries’ participa- spillovers to the domestic economy? tion in GVCs requires assessing performance across a wide range of indicators and concepts. No single Part II assesses a country’s GVC participation measure or concept can be used to determine suc- through three types of measures: cess or failure in GVC integration. The development of value-added trade data represents a fundamental 1. GVC participation measures, by country and sec- step forward in understanding GVC trade. tor (including trade in value added) Nevertheless, currently, measures of GVC par- 2. Network analysis of international trade ticipation and domestic value added in trade are 3. Firm-level measures of direct links in GVCs not widely available. Existing GVC databases are presented at a fairly aggregated level of goods and GVC participation measures differentiate between services, and do not always cover LMICs. No single buyer- and seller-related measures and combine database in isolation provides a complete picture those measures to assess countries’ overall GVC par- of GVC participation and how much value added ticipation. Growing GVC participation on the buying in trade is being generated domestically. By shift- side indicates that a country’s exports increasingly ing emphasis from the broad country level to an rely on intermediate imports. Growing GVC partici- increased focus on firms and narrowly defined sec- pation on the selling side indicates a country’s growth tors, part II suggests moving further in the direction in domestic value added caused by own or third- of better measures of GVC participation and domes- country exports. So various measures drawing on tic value added embodied in trade. trade in value-added data estimate the source of value Questions that can be addressed with the tools (domestic or foreign, by country and industry) that described in part II include the following: is added in goods and services produced for export or final demand. Emphasis is placed on how value • How extensive is a country’s or sector’s GVC addition from gross exports has changed over time participation? What is a country’s, sector’s, or and how it is linked to the country’s participation product’s degree of centrality and structural inte- in GVCs. Meanwhile, the narrow view of whether gration in GVCs? a country captures a growing share of the value of 4 Making Global Value Chains Work for Development exports is dismissed, as it misses the key benefit of Part III: Strategic Questions and growth of domestic value added (in levels) originat- Policy Options ing from GVCs: positive changes in foreign sourcing are associated with positive changes in the per capita GVC participation does not automatically gener- domestic value added in exports, which suggests that ate development. Part III shows that development greater use of foreign inputs is complementary to requires getting more value added from a country’s growing per capita domestic value added in exports. productive factors (economic upgrading and den- Network metrics typically focus on a country’s sification), improving the quality and quantity of gross trade, trade in value added, trade in parts those factors (especially labor skills and technologi- and components, or other groupings of trade flows cal capabilities), redressing market failures, and engi- that proxy for GVC trade. The metrics can be com- neering equitable distribution of opportunities and puted for overall trade, individual sectors, or indi- outcomes—which all add up to social cohesion. All vidual products in three ways. First, there are several this must occur while reinforcing living standards, indicators that examine a country’s centrality and including employment, wages, working conditions, structural integration in GVCs. Second, the net- economic rights, gender equality, economic secu- work trade index is an improved measure of assess- rity, and protection of the environment—altogether ing a country’s trade openness overall (Santoni and known as social upgrading. The internationalization Taglioni 2015), its openness in a sector, or its trade of production processes helps with very few of those of an individual product relative to peers. The index development challenges, but it provides the policy accounts not only for direct trade relationships with space to address them. partner countries, but also for the interactions of The book offers policy makers analytical tools and the countries with their partners, in an iterative pro- policy options to formulate a country’s GVC partici- cess that covers the entire network. Third, bilateral pation strategy—how the country can enter a GVC network relations can be visualized as a world map and then leverage its position to expand GVC par- of proxies for vertical trade networks. In this con- ticipation by shifting and improving resources in a text, concepts such as minimal spanning trees visu- way that advances development goals. Formulating ally identify the trade partners with the strongest or a country’s GVC participation strategy includes most relevant links, according to a chosen param- determining whether a GVC delivers labor market– eter. Overall, network analysis helps to capture enhancing outcomes for workers at home. Thinking heterogeneity between the individual nodes in the at the country level brings to the fore constraints networks (for example, countries or combinations such as the supply of various types of labor, skills, of countries and sectors, or countries and products) and absorptive capacity. GVCs can create new and in the links between the nodes, to understand opportunities on the labor demand side, but supply the complex, multidimensional phenomena that and demand cannot meet if the supply is missing. characterize GVCs. That fact emphasizes the importance of embedding Firm-level measures focusing on direct links in national GVC policies in a broader portfolio of poli- GVCs add more granularity to the analysis. They cies aimed at upgrading skills, improving physical can be aggregated up to the sector and country levels, and regulatory infrastructure, and enhancing social but—where data are sufficiently available—they cohesion. can also be used to look at the dispersion around The strategic policy framework in part III focuses the average for any given measure of interest. Firm- on strategies to help LMICs maximize their gains level survey data directly capture the main actors in a from participation in GVCs. To develop an effective value chain—buyers and suppliers—and allow com- and sustainable strategy of GVC participation, gov- parisons of GVC links across industries or between ernments must identify key binding constraints and types of actors (for example, foreign-owned versus design the necessary policy and regulatory interven- domestically owned firms, or firms of different sizes) tions—as well as infrastructure and capacity build- in a country or a single industry across countries. ing—which allow them to achieve distinct objectives The links between buyers and suppliers include mul- and address specific challenges (figure O.1): tinational corporations and domestic suppliers in a country, domestic final producers and suppliers 1. Entering GVCs: attracting foreign investors and abroad, and domestic suppliers and buyers abroad. facilitating domestic firms’ entry into GVCs Overview: Making GVCs Work for Development 5 Figure O.1. Strategic Policy Framework Focus area Objectives Strategic questions Policy options Which tasks? Creating world-class GVC links – Which form of GVC participation? – Jump-starting GVC entry through EPZs – How can tasks be identified? and other competitive spaces Attracting foreign – Which risks? – Attracting the "right" foreign investors investors and Which form of governance? – Helping domestic firms find the Entering GVCs “right” trade partner and technology facilitating domestic – Which form of governance between firms’ entry into GVCs lead firms and suppliers? abroad – Buyer- or producer-driven value – Improving connectivity to chains? international markets – Which power relations in GVCs? Creating a world-class climate for foreign tangible and intangible assets – Ensuring cost competitiveness – Improving drivers of investment and protecting foreign assets – Improving domestic value chains and quality of infrastructure and services Which transmission channels? Strengthening GVC local-economy Which type of economic upgrading? links on the buyer’s and seller’s Promoting economic sides upgrading and Which type of densification? densification Which foreign firm and country Expanding and characteristics influence Strengthening absorptive capacity strengthening spillovers? – Maximizing the absorption potential GVC participation of local actors to benefit from GVC Strengthening Which domestic firm characteristics spillovers domestic firms’ help internalize spillovers? – Fostering innovation and building absorptive capacity capacity – Complying with process and product standards – Bundling tasks Which relationship between economic and social upgrading? Creating a world-class workforce Promoting social Which type of social upgrading? – Developing skills upgrading and Is downgrading a possibility? – Promoting social upgrading Turning GVC cohesion – Engineering equitable distribution of participation into Which links between social upgrading and cohesion? opportunities and outcomes sustainable development Promoting environmental What benefits from environmental Implementing climate-smart policies sustainability regulation? and infrastructure Note: EPZs = export processing zones; GVCs = global value chains. 2. Expanding and strengthening GVC participation: institutions can, in addition, move along the value promoting economic upgrading and densifica- chain, strengthen participation, and achieve higher tion, and strengthening domestic firms’ absorp- added value in a sustainable way. Therefore, the book tive capacity raises strategic questions in each of the three focus 3. Turning GVC participation into sustainable devel- areas, offers a range of possible answers, and points opment: ensuring skill upgrading, social upgrad- to critical issues that must be considered. ing, and equitable distribution of opportunities and outcomes while promoting environmental • Which form of GVC participation can a country sustainability. pursue? • How can GVC tasks be identified? The goal is to enable policy makers to make • What are the possible risks of GVC participation? informed choices. All in all, GVCs offer a role to play • Which forms of governance exist between lead for economies at different levels of development firms and suppliers? at any point in time. Economies that have in place • Which power relations characterize specific a supporting environment and well-functioning GVCs? 6 Making Global Value Chains Work for Development • What are the GVC transmission channels? Countries also need to be aware of the power rela- • Which type of economic upgrading, densifica- tions in GVCs between the lead firm and other firms, tion, and social upgrading can countries pursue? and the scope for diversifying specific supply chain • Which foreign firm, domestic firm, and country risk. Because a large part of GVC integration hap- characteristics influence GVC spillovers? pens through FDI, countries must examine whether • What is the relationship between economic FDI leads to positive spillovers for local actors (espe- and social upgrading, and is downgrading a cially domestic firms and workers), and they must possibility? know about possible factors at the foreign investor, • What are the links between social upgrading and domestic firm, and national and institutional levels cohesion? that could—positively or negatively—mediate such • What benefits to sustainable GVC participation benefits. Countries also have to decide which type can originate from environmental regulation? of economic upgrading (product, functional, or inter-sector), densification, and social upgrading A country that seeks to participate in GVCs (employment, wages, or labor standards) they want must ask which tasks it should focus on and which to pursue, and assess the relevance of implement- types of GVC governance are possible. A country ing climate-smart policies and infrastructure in their that is already integrated in GVCs must evaluate GVC integration strategies. the risks that could threaten its survival in the value Finally, policy options are proposed for each of chain, such as becoming more vulnerable to exter- the three focus areas: nal shocks. By locating various stages of produc- tion in countries where production costs are lower, 1. Which policies support GVC entry? firms decrease the marginal cost of production, but 2. How can policies influence the expansion and raise other costs by increasing the complexity and strengthening of GVC participation? uncertainty associated with organizing production 3. Which policies help turn GVC participation into across several locations. Changes in this “trade-off ” sustainable development? affect outsourcing and offshoring decisions, and can be heavily influenced by national policy choices. In To guide policy makers in prioritizing policies, recent years, some evidence has started to emerge part III lists performance indicators that can be on “back-shoring” activities because of rising costs, used to identify the most important challenges that intellectual property rights concerns, digitalization a country must address. Table O.1 presents selected of the economy, and changing perceptions about the performance indicators, of those described in part stability and reliability of GVCs. III, for illustration. Table O.1 Selected Policy Objectives and Performance Indicators by Focus Area Focus area Policy options Selected performance indicators Entering GVCs Improving connectivity to international LPI (international)—overall and components; efficiency of markets customs (WDI) Ensuring cost competitiveness Unit labor costs Improving drivers of investment Ease of doing business index—overall (WDI) Protecting assets Ease of doing business index—protecting investors (WDI) Improving domestic value chains and LPI (domestic)—quality of infrastructure, quality and quality of infrastructure and services competence of services (WDI) Expanding and Fostering innovation and building capacity R&D intensity strengthening Complying with process and product Diffusion of voluntary standards and ISO certification GVC participation standards ownership (WDI, national statistics); surveys/field assessments in country Turning GVC Developing skills Education statistics participation into Promoting social upgrading Wage statistics; employment statistics; labor standards sustainable Engineering equitable distribution of Indicators on access to information; antidiscrimination laws development opportunities and outcomes and rights; social insurance and assistance Note: GVCs = global value chains; ISO = International Organization for Standardization; LPI = Logistics Performance Index; R&D = research and development; WDI = World Development Indicators. Overview: Making GVCs Work for Development 7 Governments that seek to join GVCs have to • The creation of synergies on the ground requires create (1) world-class GVC links and (2) a world- multiple interventions (advisory, analytics, financ- class climate for foreign tangible and intangible ing, advocacy) and long-lasting engagement. assets. The first item requires attracting the right for- • Policy advice supporting GVC-based growth eign investors and improving connectivity to interna- models requires sound analytics, evidence, and tional markets; the second requires assets protection data. It also requires 360-degree assessment of and high-quality infrastructure and services. The the competitiveness of a country’s economy, in ease of doing business indicator, for example, can its entirety, and drilling down to specific sectors, provide an overview of how attractive a country is to GVCs, tasks, and activities, to identify, prepare, foreign investors, especially in its protection of assets. and inform all interventions. The logistics performance index can help countries • Interventions need to build on analytical founda- assess how well they fare on connectivity to interna- tions and follow well-targeted and action-bound tional markets and border efficiency. The logistics action plans, but they do not need to follow a performance index can also help in examining the standard sequence or timeline abstracting from quality of a country’s infrastructure and services. country-specific and context-specific conditions. To expand and strengthen a country’s GVC par- Coordination, information sharing, and lever- ticipation, policy makers must focus on strength- aging synergies between different interventions ening existing GVC–local economy links, as well are important. Coordination demands are high as the absorptive capacity of local actors, to help within government agencies, GVC stakeholders, them maximize the benefits from GVC spillovers. and donor partners. Absorptive capacity includes innovation capacity, • A participative approach, with alignment on and which, for example, could be measured by research ownership of the agenda by all stakeholders, is and development intensity. critical. Effective stakeholder engagement mecha- nisms are a central anchor for continued, long- Turning GVC participation into sustainable devel- lasting results (but are often the least funded). opment also means creating a world-class workforce • Network effects and positive spillovers from GVC with policies that promote skill development, social participation across sectors, based on integrated upgrading, and equitable distribution of opportuni- solution packages, are achievable over time. ties and outcomes, and implementing climate-smart Dynamic learning, replication, and scale-up can policies and infrastructure. Performance indicators be fostered through global and cross-country include, but are not limited to, education, wage, and platforms. employment statistics, as well as indicators on labor • A shared vision and a common understanding of standards, access to information, antidiscrimination the project goals and objectives between imple- laws and rights, and social insurance and assistance. menting teams, local and international stake- holders, and other development partners are Part IV: Country Engagement important for success. To complement the content of parts I to III, part The rest of part IV shares World Bank Group IV offers guidelines to engage with country stake- experience in leveraging the combination of quanti- holders for implementing a national strategy to tative, desk-based analysis, fieldwork, and in-country achieve GVC-led development. The odds of success capacity building to produce sound, country-specific in GVCs are affected by policy and its implementa- diagnostics. tion in a wide range of influencing areas. Part IV is intended to bring attention to the synergies between Note the various areas of policy and help support coun- tries’ efforts to identify the necessary reforms to 1. About 50 percent of total U.S. exports, for example, are intra-firm, according to Antràs (2015). trigger a virtuous cycle of “reform-GVC entry and upgrading-development.” The following are important recommendations References and lessons learned for interventions at the country Antràs, Pol. 2015. Global Production: Firms, Contracts, and level that the World Bank Group experience of work- Trade Structure. Princeton, NJ: Princeton University ing with countries across the globe suggests: Press. 8 Making Global Value Chains Work for Development Baldwin, Richard, and Javier Lopez-Gonzalez. 2013. Mettler, Ann, and Anthony Williams. 2011. The Rise “Supply-Chain Trade: A Portrait of Global Patterns of the Micro-Multinational: How Freelancers and and Several Testable Hypotheses.” NBER Working Technology-Savvy Start-Ups Are Driving Growth, Jobs Paper 18957, National Bureau of Economic Research, and Innovation. Policy Brief, Lisbon Council, Brussels, Cambridge, MA. Belgium. Gómez-Mera, Laura, Thomas Kenyon, Yotam Margalit, Santoni, Gianluca, and Daria Taglioni. 2015. “Networks Josó Guilherme Reis, and Gonzalo Varela. 2015. New and Structural Integration in Global Value Chains.” In Voices in Investment: A Survey of Investors from Emerging The Age of Global Value Chains, edited by João Amador Countries. Washington, DC: World Bank. and Filippo di Mauro. Washington, DC: Center for Economic and Policy Research. PART I WHY GVCs REQUIRE FRESH THINKING Part I begins by asserting that global value chains (GVCs) must be rethought for the twenty-first century. Chapter 1—“Here’s Why”—shows that the new GVC-enabled flow of know-how from high-income countries to low- and middle-income countries is a key factor in determining the role of GVCs in industrialization and development. From a policy perspective, the critical issue is how GVCs integrate into the economy as a whole and how to maximize the benefits from technology transfers, knowledge spillovers, and increased value addition. But it is equally important to ensure that participation in GVCs benefits domestic society through more and better-paid jobs, better living conditions, and social cohesion. To exemplify how this book can help policy makers find answers to that challenge, chapter 2— “Consider Bulgaria”—provides a case study of Bulgaria. The study shows how analysts can make use of the quantitative tools described in part II of this book, as well as the strategic policy framework developed in part III to identify a country’s position in GVCs, its scope for upgrading, and policies that can help achieve that goal. Chapter 1 HERE’S WHY Introduction multinational corporations (MNCs) seek to improve local innovation, knowledge-based capital, and com- Global value chains (GVCs) can be thought of as petencies. The Samsung Group—which employs factories that cross international borders (box 1.1).1 369,000 people in 510 offices worldwide—worries Producing high-quality goods and services in GVCs about shortages of technical and engineering skills in involves more than simply trading goods and ser- Africa and how those shortages affect its efforts to vices internationally; it also entails the cross-border embed its African workforce in Samsung’s global movement of know-how, investments, and human production networks. In 2011, to address such capital. When Toyota makes car parts in Thailand, it shortages, Samsung launched Samsung Electronics does not rely on local know-how. Instead, it imports Engineering Academies in Kenya, Nigeria, and South Toyota technology, management, logistics, and any Africa. Outstanding performers are sent to annual other bits of know-how not available in Thailand, Learnership Programs in Seoul as part of Samsung’s because Thai-made parts have to fit seamlessly with program for young leaders. The initiative serves the parts made in Japan and elsewhere. GVCs, in effect, company’s broader goal to develop 10,000 electron- “unbundle” factories by offshoring firm-specific ics engineers across the continent by 2015.2 Other know-how along the stages of production, and those corporations are investing in building the skill base international flows of know-how are a key reason in LMICs, too.3 Lucent Technologies supports educa- why GVCs offer unprecedented development oppor- tion and learning programs in 16 countries through- tunities to participating countries. out Africa, Asia, Europe, and Latin America; Nike Internationally fragmented production is not and the United Kingdom’s Department for Interna- new. For decades, low- and middle-income countries tional Development run a program to support access (LMICs) have imported parts from countries with to economic assets for adolescent girls; Microsoft more advanced technology, although generally only provides support to incorporate information tech- for the assembly of locally sold goods. Because the nology (IT) into the daily lives of young people in goods produced were not part of a global network, the Philippines, Poland, the Russian Federation, and flows of know-how were less intense. The new char- South Africa; Cisco provides funds, expertise, and acteristic of GVCs from a development perspective equipment to create national networks of IT training is that factories in LMICs have become full-fledged centers in India, Mexico, South Africa, and the West participants in international production networks. Bank and Gaza, in addition to the work of the Cisco They are no longer just importing parts for assem- Networking Academy, which has 10,000 academies bly for local sales. They are exporting goods, parts, in 165 countries; finally, Nokia enhances life skills components, and services customized to the needs of and leadership skills of young people in several coun- the intended buyers and used in some of the most tries, including Brazil, China, and Mexico. sophisticated products on the planet. The new GVC-enabled flow of know-how from Given the need for customization and integra- high-income countries to LMICs is a key fac- tion of production facilities internationally, large tor in determining the role of GVCs in industrial 11 12 Making Global Value Chains Work for Development Box 1.1. Defining GVCs From a business organization perspective, value chains describe multinational enterprises (MNEs), such as IBM, Siemens, and the sequence of productive (value-added) activities that capital Toyota, nowadays rely on a complex web of suppliers, vendors, and labor (or firms and workers) perform to bring a good or ser- and service providers of all kinds and in multiple locations. At the vice from its conception to end use and beyond.a “Value chain same time, a set of highly influential global buyers gained scale analysis” is intended as the science of identifying bottlenecks and and influence in the 1990s, including retailers such as Walmart opportunities between different stages of production and tasks. and Tesco and branded merchandisers such as Nike, Zara, and Value chains are said to be “global” when they include steps, Uniqlo.e Building on successful experiments in the 1970s and processes, and actors from at least two countries;b they can be 1980s by a handful of pioneering retailers, such as J. C. Pen- regional if the scope of production takes place within the same ney and Sears, global buyers nowadays place huge orders with geographic region. From an economic perspective, the phenom- suppliers around the world without establishing any factories or enon of global value chains (GVCs) identifies a production struc- farms of their own.f Unlike traditional MNEs, where equity ties ture in which tasks and business functions are distributed among link headquarters with foreign affiliates, global buyers link to several companies, globally, or regionally.c The key features of their suppliers through non-equity external sourcing ties. Often, GVCs are therefore the international dimension of the production intermediaries (for example, trading companies such as Hong process and the “contractualization” of buyer and seller relation- Kong SAR, China’s Li & Fung) are used to link buyers to producers ships, often across international borders. in multiple countries. GVCs, in effect, integrate the know-how of lead firms and sup- To highlight the complexity of the interactions among global pliers of key components along all the stages of production and in producers, recent literature makes reference to the concept of multiple companies and offshore locations. Typically coordinated global production “networks” rather than “chains.”g Accordingly, by lead firms, GVCs involve international trade flows within their in the more realistic metaphor of networks, links can be seen as networks of foreign affiliates (foreign direct investment), contrac- connecting nodes, some more central and some more peripheral. tual partners (non-equity modes of investment), and arm’s-length Given the predominance of the term GVCs in the literature, this external suppliers. d Well-functioning supply chains—which report uses it to refer generically to chains, networks, or both. define the physical movement of goods all along the value chain, When more specific references are needed, they will be explicitly including domestic and international segments—are a key con- mentioned in the text. cern in GVCs. This is the case because good logistics, which Capital and labor are not the only factors of production. defines the art of managing the supply chain and includes good “Ideas” can be singled out as a third factor of production, although connectivity, streamlined procedures for imports and exports, they could also be understood as high-skilled labor input. In a and low cost of logistics services, is an important determinant of global context, the value-added activity performed in one country countries’ ability to join and strengthen participation in GVCs and crosses international borders in goods or services tasks. Differ- a key factor in determining the costs of sourcing from and sup- ent tasks of the value chain contain a different amount of such plying to global markets. Getting to the border is one of the most factors of production. For example, specialized workers tend to pervasive constraints for exports of firms in low- and middle- be necessary in higher value-added tasks of the GVC. In the auto- income countries (LMICs), while the costs of logistics services motive, electronics, and electrical appliance industries, ideas are can be disproportionately high for smaller and younger firms or for more strongly embedded in the early preproduction stages, such more remote locations. Improving logistics is also where LMICs as research and development and design, or in postproduction have the most potential to reduce trade costs, according to recent (logistics, marketing, and branding), thus requiring such special- surveys. Finally, well-functioning trade facilitation measures ized workers in those tasks. In other industries, notably the craft- enable GVC trade by reducing the time, cost, and uncertainty based ones (such as furniture making), innovation development is involved in importing and exporting. maximized when ideas (product design) and manufacturing opera- But most production processes do not happen in a sequence tions are joint,h because innovation in those sectors often stems of dependent activities. Instead, they take place in more complex from a bottom-up approach.i networks of production, in which participating firms are special- ists in one activity and external international sourcing arrange- a. Porter (1985); Sturgeon (2001). b. Gereffi and others (2001); Gereffi, Humphrey, and Sturgeon (2005). ments imbue inter-firm trade with characteristics similar to c. Grossman and Rossi-Hansberg (2012). intra-group trade: better control from the center, higher levels of d. UNCTAD (2013). bilateral information flow, tolerance of asset specificity, and har- e. Feenstra and Hamilton (2006). monization and immediate integration of business processes that f. Gereffi (1999); Ponte and Gibbon (2005). g. Henderson and others (2002). increase the potential for foreign activities to integrate seam- h. Buciuni, Coro, and Micelli (2013); Pisano and Shih (2009). lessly with activities performed at home. Large brand-carrying i. Breznitz and Murphree (2011). Here’s Why 13 development. LMICs can now industrialize by local factories with the relevant international pro- joining GVCs without the need to build their own duction network and protecting proprietary assets. value chain from scratch, as Japan and the Republic The predictability, reliability, and time sensitiv- of Korea had to do in the twentieth century.4 That ity of trade flows are important factors behind firms’ enables LMICs to focus on specific tasks in the decision about a location, according to major trade value chain rather than producing the entire prod- and competitiveness indexes and case studies.6 In uct, thereby lowering the threshold and costs for many cases, countries cannot participate in certain industrial development. LMICs can benefit from parts of GVCs because of requirements for timely foreign-originated intellectual property; trademarks; production and delivery. In effect, time is money in operational, managerial, and business practices; GVCs. A day of delay in exporting has a tariff equiv- marketing expertise; and organizational models. alent of 1 percent or more for time-sensitive prod- The result is that a new policy framework has ucts.7 Slow, unpredictable land transport keeps most emerged in which imports matter as much as, if not of Sub-Saharan Africa out of the electronics value more than, exports and in which the flows of goods, chain.8 Lead firms and intermediate producers in services, people, ideas, and capital are interdependent GVCs need reliable, predictable, and timely access to and must be assessed jointly (box 1.1). Countries inputs and final products to satisfy demand on time. that understand the opportunities that GVCs offer Hence, good infrastructure and efficient borders are and adopt the appropriate policies to mitigate the critical, as they relate to the predictability, reliability, risks associated with them have the opportunity— and time sensitivity of trade flows. through GVCs—to boost employment and produc- Strong, well-enforced property rights are the tivity in all their agriculture, manufacturing, and other element essential to attracting and keeping services production. Job creation and labor pro- foreign investors. 9 Firms export valuable, firm- ductivity growth are sometimes viewed as compet- specific technology and know-how, only part of ing goals, as higher labor productivity enables firms which can be protected through patents, trade- to produce a larger amount of value added without marks, and other forms of intellectual property necessarily increasing the number of workers at the regulations (IPRs). The know-how embodied in same rate (static productivity effects). business and organizational models, operational Research shows that GVC integration leads to and managerial practices, production processes, higher net jobs but lower job intensity5 and has and export processes cannot be patented or trade- strong potential for productivity gains via sev- marked; and even when intellectual property can be eral transmission channels (dynamic productivity patented or trademarked, IPR treaties and domes- effects), as discussed later in this chapter, which go in tic regulations aimed at promoting fair competition hand with increased labor demand caused by more only imperfectly protect such know-how. Because vertical specialization and higher output in GVCs. GVCs necessarily involve contracting relationships between agents located in countries with hetero- geneous legal systems and contracting institutions, Firm and Policy Perspective “contracts are often neither explicit nor implicit; Connecting Factories and Protecting Assets When they simply remain incomplete.”10 The way in which different national systems deal with contractual fric- Doing Business Abroad: The Firm Perspective tions and incomplete contracts and the way host The international location of new production facili- countries enforce contracts between private par- ties is ultimately in the hands of GVC lead firms. ties are additional elements driving firms’ choice of Conceptually, the new possibilities created by location, and those elements also factor into firms’ globalization and the information and communica- boundaries in global sourcing.11 tions technology revolution create two distinct sets of The connectivity of factories and the nature necessities for firms, which countries must address: of contracting across countries are therefore key (1) connecting factories and (2) protecting assets. determinants—along with capital intensity—of a Because cross-border factories must work as a unit, firm’s decision to make or buy and whether to do so lead firms in GVCs care about efficiently connecting domestically or internationally. Figure 1.1 illustrates 14 Making Global Value Chains Work for Development Figure 1.1. Supplier-Buyer Links between China and Japan in the Automotive Industry Japan China Production FAW outsourcing of GAC Toyota Cars Co., Mazda 6 and Mazda Motor Co., Ltd. Ltd. Mazda 8 30.5% equity stake; production of Production of Tianjin FAW 47.7% equity FAW 25% equity Camry, Camry Vitz, Belta, Xiali stake Group stake Hybrid, Yaris, and Platz Automobile Corp. E-Z, and Highlander Co., Ltd. 30% equity stake 40% equity stake; Changang Changan Tianjin FAW 50% production of Vios, 20% equity Mazda Ford Toyota Motor equity Corolla, Crown, Reiz, stake Automobile Mazda Engine Co., Ltd. stake Corolla EX, and Rav4 Co., Ltd. Co., Ltd. 50% equity stake 25% equity stake 50% equity stake; production Sichuan Chongqin of Coaster, FAW Toyota 50% equity Changan Toyota Land Cruiser, stake Ford Motor Co., Automobile Prado, Corolla, Ltd. Co., Ltd. and Prius Source: Japanese Automobile Manufacturers Association 2013, 55, adapted as of March 2013. Note: Japanese companies are yellow, Chinese counterparts are green. The arrows indicate ownership or other forms of control. The Japan Automobile Manufacturers Association states: “In principle, the tie-ups shown above cover only technical cooperation related to motor vehicle production and exclude sales tie-ups.” these concepts using actual ownership relationships instead just handed over through licensing agree- between some of the key suppliers and buyers in the ments or other non-equity investments. Technical Sino-Japanese auto industry. Those relationships cooperation and arm’s-length trade signal looser move from Japan to China—that is, from the higher- forms of collaboration. With the dramatic growth of income to the lower-income country and from the outsourcing practices, competition between compa- technological hub to the production site. The good nies has shifted from horizontal (with firms compet- connectivity between China and Japan and the prox- ing in the same sector for the same customer base) imity of the two countries satisfy the first concern to vertical (with firms in the same value chain com- of lead firms: connecting factories. Meanwhile, the peting to perform specific and specialized tasks). correspondence between the type of control and the Lead firms compete with first-tier and lower-tier strategic importance of assets in the Sino-Japanese suppliers.12 automotive sector accurately illustrates the second The links between Mazda, the fifth largest Japanese key concern of global investors: protecting assets. car manufacturer in production volume, and China’s Control of the subsidiary takes place in a variety FAW Car Group (FAW) illustrate the complexity of of ways. The most strategic assets are tied to the lead vertical competition (figure 1.1). Whereas Mazda firm through forms of direct capital control over the outsources the production of the Mazda 6 and the supplier (such as majority equity stakes). Assets of Mazda 8 to FAW, the latter also competes with the lower importance (such as older technologies) are former. FAW produces other models, under different Here’s Why 15 brands, using technology from Mazda’s competitors, be connected with a wide-ranging domestic reform including Daihatsu, Toyota, and Volkswagen. Mazda agenda aimed at helping countries enhance firms’ also has its own line of luxury cars that directly com- productivity by building internal capacities and pro- pete with models from the lead firms.13 viding access to capital and connectivity, and ensur- ing a responsive and effective governance structure for identifying opportunities and addressing chal- Creating Links to the Local Economy: lenges from GVC participation. The Policy Maker Perspective GVCs require targeted policies and analysis across In the same way that import substitution industrial- a wide range of areas, which may not always be easy ization gave way to export-oriented industrialization, for a country’s policy makers to formulate and con- the latter is now being replaced by efforts to identify nect to each other and to GVCs. Governments may an entry point into vertically specialized industries not necessarily be aware of the effects of domestic and upgrade within GVCs. Attracting offshore fac- policies on integration and upgrading in GVCs. The tories and ensuring domestic firm participation in odds of success in GVCs are affected by policy and its international GVCs has become a major priority for implementation in areas as different as trade (tariff many policy makers in LMICs. and nontariff barriers), domestic services regulations, From a policy perspective, however, the critical investment regulations and incentives, compliance issue is how GVCs integrate into the economy as a with process and product standards, innovation, whole. Attracting and keeping offshore factories is not industry, entrepreneurship, labor markets, educa- enough. Opening borders and attracting investment tion, and infrastructure and connectivity. Countries are important and help jump-start entry in GVCs. But may not appreciate fully the importance of the syn- to retain GVCs, maximize their benefit to the domes- ergies between the core areas of trade and investment tic economy, and ensure their sustainability, countries regulation and well-tailored complementary mea- must integrate the domestic productive sector. The sures. Countries also may not be able to identify the policy challenge extends, therefore, to creating and appropriate investment in education and vocational strengthening links with domestic firms and ensuring training, infrastructure, and connectivity; the best that the host country benefits from technology trans- setting for labor market policies; which interna- fers, knowledge spillovers, and increased value addi- tional standards to adopt; how to design and develop tion in the country. If GVCs remain de-linked from adequate supplier programs; effective cluster devel- the local context, lead firms drive many decisions, and opment programs and competitive spaces (special governments may have limited influence and abil- economic zones, growth poles, growth corridors, and ity to leverage such decisions for domestic economic so forth); or services regulations conducive to busi- development. It is equally important to ensure that ness efficiency. Finally, countries may not be able GVC participation benefits domestic society through to identify and implement sustainable and effective more and better-paid jobs, better living conditions, financing and incentive schemes. and social cohesion. Even when governments are aware of these issues, The right strategies can help LMICs increase and putting in place regulations that do not unnecessar- strengthen their participation in GVCs and foster ily restrict effectiveness in GVC participation may development. Those strategies will be discussed at be difficult. In most countries, many agencies have length in part III. Nevertheless, a point to remember a role in setting and enforcing regulation that may is that to create an effective and sustainable strategy affect value chains and the efficiency of their sup- of GVC participation, governments must focus on ply chain.14 Those agencies also often legislate and identifying key binding constraints and designing implement regulation in an uncoordinated man- the necessary policy and regulatory interventions— ner because regulators set policies with domestic as well as infrastructure and capacity building—with regulatory objectives in mind. As a consequence, a “whole of value chain approach.” Such an approach international coordination is not necessarily able to is needed to achieve development objectives through foster GVCs’ production and trade along the cor- GVC participation and address specific challenges in responding supply chain. International coordina- entering GVCs, expanding and strengthening par- tion conflicts with domestic regulatory objectives ticipation, and ensuring sustainability and inclu- may explain why existing trade agreements, invest- sive growth. Trade and investment policies need to ment agreements, and similar forms of international 16 Making Global Value Chains Work for Development cooperation are rarely designed to foster GVC par- Europe, Mexico, and parts of the Middle East, such ticipation (Hoekman 2014). as Morocco and Turkey, are increasingly integrated Given this background, the policy maker’s prior- in GVCs, other parts of the world remain marginal. ity should be to identify and lift binding constraints, That is also the case for most of Africa, Mercosur (the unlock productivity growth, and improve the overall trade bloc comprising Argentina, Brazil, Paraguay, competitiveness of the country. Many governments Uruguay and Venezuela as full members, and Bolivia, are willing to invest significant time and effort to Chile, Colombia, Ecuador, and Peru and Suriname adopt policy that influences the cost of production as associate countries), and South Asia. Another and trade within a GVC. The appropriate analysis key difference between the group of countries and and policy strategies can help trigger a virtuous cycle regions that is more integrated into GVCs and the of “reform—GVC entry and upgrading—develop- group with low integration is that whereas the latter ment,” whereby the private sector is encouraged to remain resource-based economies, the former have keep investing retained earnings in the continued shifted their specialization to manufacturing. improvement of existing activities, new activities, and An initial trigger has been the integration of comparative advantage tasks in countries’ agriculture, China, India, and Russia, which added new massive manufacturing, and services sectors, thereby generat- product and labor markets that had been marginal to ing a process of inclusive growth for the host country. the multilateral trading system before 1989. The inte- gration of these countries into the world economy nearly doubled the scope of play for globalization.16 Evolution of GVC Trade Faced with slow growth at home, large enterprises Once concentrated among a few large economies, from high-income countries set up operations in the global flows of goods, services, and capital now newly opened markets, especially in China. Although reach an ever larger number of economies world- relocation was partly to carve out brand recognition wide. Global gross exports of goods increased ten- and a market share in rapidly expanding consumer fold over 1980–2013, and that of services, 9.8-fold. markets, the firms also saw an opportunity, through Foreign direct investment (FDI) net inflows were 34 GVCs, to cut costs on goods produced for export to times higher in 2013 than in 1980. By 2013, as many international and home markets. Moreover, under as 3,000 bilateral investment treaties had been signed pressure from financial markets, large American to create the framework of deep agreements neces- and European enterprises embarked on a “second sary to connect factories and protect the assets of unbundling” of corporate functions during the foreign firms, and the sales of foreign-owned firms 1990s.17 In an effort to focus on “core competencies,” amounted to US$26 trillion.15 nearly every business function that was considered All these flows have grown over time, creating “noncore” became subject to possible external sourc- increasingly dense and complex networks. The value ing from more specialized, more competitive, and of most bilateral goods flows between major world often less unionized suppliers.18 regions (Asia, Eastern Europe, Latin America, the According to Hoekman (2014), the heterogene- United States, and Western Europe) is now greater ity in GVC participation is largely caused by per- than the gross domestic product (GDP) of the par- sistent heterogeneity in trade costs. In addition to ticipants. In 1980, by contrast, the only flows of goods trade costs, which are determined by a country’s exceeding the value of GDP were those connecting connectivity among domestic markets and with the United States with Western Europe and Western international markets, a country’s drivers of invest- Europe with the Middle East, North Africa, and Sub- ment—including its skills and technological capacity Saharan Africa. The globe has grown into a multi- and the protection of foreign assets—further deter- polar world economy with diverging performances mine its extent of GVC participation. Improvements (figure 1.2). The triad formed by China, the European in industrial capabilities in many LMICs have cre- Union, and the United States accounts for 53.6 per- ated many more opportunities for their firms in the cent of world goods and services exports and 53.9 past 20 years. What previously had to be done within percent of world goods and services imports. India’s the boundaries of multinational enterprises (MNEs) trade, by comparison, is very small, accounting only can now be externally sourced from newly compe- for one-fifth of China’s goods and services exports. tent suppliers and service providers with offices and GVC participation and trade costs remain het- factories around the world.19 While the twin trends erogeneous. Although East Asia, Central and Eastern of external and international sourcing meant that Here’s Why 17 Figure 1.2. Stylized Facts about GVCs: A Multipolar World with Diverging Performances Triad: China, European Union, and the United States Triad’s global share, % India-China exports ratio Imports 53.9 53.6 Exports India 1:5 China Heterogeneous trade costs Heterogeneous GVC participation Role of FDI Sales of foreign-owned firms, US$ Resource based Shifted to manufacturing 26 trillion Africa, South Asia, East Asia, Central and MERCOSUR and Eastern Europe, Despite barriers to trade Mexico, and Turkey A five-fold increase having fallen everywhere since 1990 GVC trade Higher Know-how, Mostly Longer chains, value-added intraindustry GVC R&D, and higher import champions services embedded technological and intraregional content into imported trade development value added Sources: Adapted from UN Comtrade Database; UNCTAD 2013; Manyka and others 2014; Hoekman 2014. Note: FDI = foreign direct investment; GVC = global value chain; Mercosur = Southern Common Market (Bolivia, Chile, Colombia, Ecuador, and Peru); R&D = research and development. existing suppliers received vast quantities of new most countries have increased their dependence on work and were pressured to follow their customers foreign inputs, measured by the share of foreign to offshore locations,20 at the same time and for the value added as a percentage of their gross exports, same reasons, the most efficient suppliers that were as they increasingly rely on imported inputs that based in LMICs grew rapidly from being small com- are processed and subsequently exported. But that panies to becoming MNEs in their own right.21 is not important. What matters is that those addi- Although production systems today are very tional imports are helping countries to grow faster complex, with multi-layered international sourcing the domestic value added that is exported, and that networks and fast-evolving, technology-enabled busi- the imports enrich the skill set available in the coun- ness models that constantly change the geography of try. And indeed, not only GVC champions, such as GVC trade, the bulk of it is intra-industry and intra- China; Poland; Taiwan, China; Turkey; and Vietnam, regional. Many goods require shorter supply lines, to name a few, but also late adopters of GVCs, such which has allowed countries near large consumer as Bulgaria and Cambodia, have seen their domestic markets to attract export processing activity. Eastern value added embodied in gross exports increase sig- Europe, in particular, joined traditional “export pro- nificantly and at par with increasing foreign sourcing. cessing” locations such as Mexico and North Africa. Manufacturing functions were among the first to Yet, the fragmentation of production implies that be externally sourced, but services soon followed. By in most manufacturing processes value chains have the 2000s, the computerization of work and the emer- become longer. A mechanical consequence is that gence of low-cost international communications 18 Making Global Value Chains Work for Development enabled a surprisingly wide range of service tasks to exemptions for maritime liner transport, privatiza- be standardized, fragmented, codified, modularized, tion of ports and port services, and divestiture and and more readily sourced externally and cheaply breakup of telecom monopolies are, according to transferred across long distances. Even aspects of Hoekman (2014), the main examples of regulatory research and development (R&D) are now sourced measures reducing the cost of delivering services from foreign suppliers. As in goods production, the across borders.24 In addition, services have increased application of IT to the provision of services has in importance as a determinant of competitiveness. allowed some degree of so-called mass customiza- Countries with a higher content of services in the tion, which is the association of customization to downstream economy are also those producing more increasing process automation and high-volume complex goods (figure 1.4). production.22 Services trade and the role of services The agriculture sector has also evolved. It now in boosting the economy as a whole have increased: represents just 2 percent of global trade (down from more than 60 percent of the current stock of global 9 percent in the 1960s) and—just as with services— FDI is in services. The composition of services has the composition of trade in agriculture has changed also changed, with modern services, such as busi- from the dominance of traditional commodities to ness services, gaining in relative importance at the increasing trade in higher-value processed products. expense of traditional services, such as travel.23 FDI The shift is also tightly linked to GVCs. In addition is also a main engine of growth for services trade. to the barriers to connecting to the agrifood GVC in Mode 3 (delivery through foreign affiliates) covers the first place,25 the efficiency and functioning of the about 50 percent of overall services trade (figure 1.3). agrifood value chain depends on the availability and The explosion of services trade is a result of fall- quality of a variety of embedded services, including ing trade and investment barriers as well as new quality control, logistics, storage facilities, packaging, digital technology, which have reduced the costs of insurance, and distribution. delivering services across borders and transformed For example, avocados are portrayed in a case study many goods into services (box 1.2). The deregula- by the U.S. Agency for International Development for tion in air and road transport, abolition of antitrust Chile.26 The fruit can be sold locally or internation- ally—at very different stages of processing. At the most basic “ingredient” level, the fruit is grown with Figure 1.3. Services Trade little control over its quality, harvested, and sold to Mode 1 35% intermediaries at low profit margins. The same pro- ducers of avocados can instead achieve better bargain- ing power and profit margins by entering or setting up more complex and sophisticated value chains, and by focusing on producing higher-quality primary prod- ucts (production tasks) that can be sold in faraway and demanding markets. They can do so by embed- Mode 2 ~10-14% ding the range of services just mentioned (quality control and so on) and by adding to the production $7 Mode 3 50% technology that enhances the quality of the fruit and BILLION Global commercial better controls the ripening of the fruit to ensure that services trade, 2010 it happens at the point of destination—whether that be next door or on the other side of the globe. To achieve the standards demanded in global markets, the producers of the primary good (the fruit) need a quality management system that grants higher quality standards by controlling harvest and postharvest procedures. By doing so, the producer improves the tasks of comparative advantage (agri- Mode 4 ~1-5% cultural production) with the assistance of more technology and services. Source: Adapted from Saez and others 2015. Note: Values are U.S. dollars. Mode 1 = cross-border trade; mode 2 = consumption In conclusion, what matters is the value addi- abroad; mode 3 = commercial presence; mode 4 = presence of natural persons. tion generated domestically and the longer-term Here’s Why 19 Box 1.2. The Disruptive Effects of Computer-Aided Technologies and Digital Innovation Value chains are rapidly changing under the pressure of digital edge technologies transforming value chains and processes. Such innovation. Cloud computing, the “Internet of Things” and “Big computer-based technologies can be disruptive—particularly for Data” are transforming business models, power relations, and companies and countries specializing in standardized production sources of value added in entire industries as diverse as the and assembly activities and not investing in human capital and health industry, distribution services, and the automotive sec- technological empowerment—because they have the potential tor. The ubiquity of e-mails, sensors, electronic data collection, to change the conventional upgrading patterns. The technologies social networks, tools for virtual collaboration (Dropbox or Google do so by transforming goods into online transfers of data, which Docs), online labor marketplaces, platforms such as eBay and Ali- allow production at the consumer’s location. For instance, 3D baba, and other cost-convenient sites for sales and professional printing is a process by which individual machines build products collaborations by small and medium enterprises are all produc- by depositing layer upon layer of material. Model-based definition tivity-enhancing instruments—albeit some possibly disruptive— instead uses a fully annotated 3D digital model as a master and grounded in digital technology. provides a seamless flow of digital thread through the product life As companies develop more sophisticated ways to leverage cycle. Copy-exact techniques allow for duplicating entire produc- digital technology, they are also shifting many processes that tion processes in remote locations and on larger or smaller scales. used to be labor intensive to computer-aided machinery. The digi- The technique was used, for example, by Intel to match its manu- tization of manufacturing may soon allow customized production facturing site to its development site at all levels, from equipment at no incremental cost and in smaller quantities (which means to process, and data collected at several levels were compared lower overall costs) than with assembly lines. The result is not with data from research and development (R&D) sites to get an only that the advantages of standardized mass production may be exact match. fading away, but also the distinction between preproduction, pro- Although these methods are still used mainly for R&D, proto- duction, and postproduction may become less and less relevant. types, and the construction of very complex components, there Analogously, the distinction between goods and services produc- are increasing examples of the methods being used for manufac- tion may become more and more a statistical artifact. turing consumer products, from toys to bicycles and even housing. Model-based definition, additive manufacturing (such as 3D printing), and copy-exact techniques are only three of the cutting- Sources: WEF 2012; World Bank 2016. development prospects that inflows of world-class or indirectly by the rest of a country’s productive technology and a richer and more sophisticated system—in other words, by producers in upstream skill set allow. The value added embodied in a coun- sectors that supply inputs to the export sector. try’s exports can be generated directly by exporters Particularly important is whether that value addi- tion—which should be measured in absolute levels and not as a share of exports—increases over time. Figure 1.4. Services Forward Links, 2007 The combination of traditional drivers of inter- 0.45 nationalization (arm’s-length trade and intra-firm trade related to FDI) combined with extensive and 0.40 complex models of external international sourc- 0.35 ing leads to cross-country, inter-firm relationships Services forward links 0.30 increasingly similar to intra-group characteristics. Patterns of cross-border investment and trade based 0.25 on product cycles—where producers in less devel- 0.20 oped countries receive older, outmoded products 0.15 from more advanced economies27—are rapidly giv- ing way to more unified global production systems 0.10 and markets, with different countries specializing in 0.05 specific aspects, or stages, of the development and 0 production of leading-edge goods and services. –2.0 –1.0 0 1.0 2.0 GVCs offer countries that embrace them the Index of economic complexity chance to grow faster, import skills and technology, Complexity index and services forward links and boost employment and productivity in all the Linear (complexity index and services forward links) country’s agriculture, manufacturing, and services Source: Saez and others 2015. activities. 20 Making Global Value Chains Work for Development Assessment of a Country’s Potential of value that was added in another country. The basic in GVCs concept is “importing to export,” or I2E, as Baldwin and Lopez-Gonzalez (2013) call it. As figure 1.5 illus- To guide policy makers in achieving development trates, one country (Japan in this example) exports through GVC integration requires investigating the parts that are incorporated in the exports of another key concepts and metrics of a country’s GVC par- country (China here). That single flow of intermedi- ticipation. Understanding how countries fare in such ate goods is the basis of two key measures of GVC key concepts and metrics allows a better identifica- integration: tion and analysis of specific value chains, activities, and business segments, which are the object of case • On the sales side, it indicates that Japanese export- studies, such as those based on Michael Porter’s five ers are selling to a GVC. forces.28 • On the sourcing side, it indicates that China is Assessments of country GVC participation focus buying from a GVC. on three concepts: The term GVC trade typically refers to I2E manu- 1. Function in GVCs: the buyer’s and seller’s factured goods and related services, but more gener- perspectives ally it also includes imported raw materials used in 2. Specialization and domestic value-added contri- exports.29 The relevance of I2E on the seller’s and bution: specialization in low or high value-added buyer’s sides is illustrated in detail in part II. activities, and patterns of upgrading and develop- To put the I2E concept in an operational context, ment through GVCs the book introduces a distinction between the 3. Position in GVC network and type of GVC node: seller’s and buyer’s sides of GVC participation. In hub, incoming spoke, or outgoing spoke; clus- many cases, countries are GVC buyers and GVC sell- tering properties; and centrality in the global ers, but that distinction reflects the difference in eco- network nomic mechanisms and determinants that lead to a country’s successful performance in absorbing valu- The multidimensional nature of GVCs can be able foreign value added compared with growing captured by looking at the relationships between domestic value embodied in GVC trade flows. We goods, services, workers, ideas, and investments consider three types of buyer roles in GVCs: input (box 1.1)—going beyond value added to look at the purchases (1) for production of final exports, (2) for actors in GVCs and assess the effects of GVCs on jobs production of intermediate inputs in the value chain, and wages. and (3) for assembly. There are also three main seller functions: supply of (1) turnkey components, Function in GVCs: Buyer’s and Seller’s Perspectives (2) primary inputs, and (3) other inputs (figure 1.6). The types of flows (goods, services, people, ideas, Classic trade involves goods made 100 percent in one and capital) predominantly associated with the buy- country and sold in another. Measures of GVC trade er’s or seller’s role are more easily discussed by first quantify deviations from that classic trade concept— focusing on the buyer’s or seller’s functions separately essentially, how much of a country’s exports consist and then considering them jointly. That evaluation Figure 1.5. Two Perspectives When Measuring GVC Participation Japan China SELLER I2E BUYER Import to Export Source: Adapted from Baldwin and Lopez-Gonzalez 2013. Here’s Why 21 Figure 1.6. Two Perspectives When Measuring GVC Seller and Buyer Functions FUNCTIONS Supply of For turnkey assembly components Supply of For primary production of inputs intermediates Supply of For other SELLER BUYER production of inputs final exports is more easily actionable from the policy angle. If, technological and know-how requirements needed for example, the domestic value chain is found to be to perform the task increase. short, or little transformation is taking place domes- In many value chains, the highest value added tically, the supply-side bottlenecks and opportunities lies with intangible activities, which are intensive in for expansion on the buying side could be more read- human capital and technology.30 In some industries, ily identified than those on the selling side. such as electronics and apparel, the latter tend to be located either at the beginning of the value chain (preproduction activities, such as basic and applied Specialization and Value Addition R&D and design) or at the end (postproduction Ultimately, what matters is the value addition gen- activities driven by marketing knowledge, such as erated in the country and whether it increases over marketing, commercialization, advertising, brand time, which is not a new question for economics. management, specialized logistics, and after-sale Value addition is a function of productivity, but it customer services). In other industries, such as fur- is associated with the breadth, variety, and sophis- niture, the intangible, high value-added activities tication of tasks and activities in which a country (such as design) are likely to take place jointly with specializes. production.31 Finally, in sectors such as chemicals, The range of activities in a value chain is very high value-added activities tend to be concentrated broad. The activities range from manufacturing upstream. inputs, outputs, and assembly operations to inbound The value added in different industries can be in and outbound logistics, marketing, sales, and a range different segments of the value chain, but invariably, of other service activities. And there are activities as higher-income countries have a stronger specializa- diverse as the production of other inputs, machin- tion in higher value-added activities within value ery, and equipment, as well as R&D, technological chains. This fact reflects the greater use of technology development, and functions aimed at organizing the and service inputs—whether in agriculture, manu- firm’s infrastructure, human resource management, facturing, or services, and whether in preproduction, and procurement. Broadly, the value-added con- production, or postproduction. Greater use of tech- tent of such activities and tasks tends to grow as the nology and service inputs is, in turn, the outcome 22 Making Global Value Chains Work for Development Figure 1.7. GVC-Driven Development industries as diverse as the health industry, distribu- tion services, and the automotive sector. The ubiquity S OF THE ECONOMY SECTOR of e-mails, sensors, electronic data collection, social networks, tools for virtual collaboration (Dropbox Manufacturing or Google Docs), online labor marketplaces, eBay, and similar platforms as cost-convenient sites for sales by small and medium enterprises (SMEs) are all productivity-enhancing instruments grounded in Agriculture Services digital technology. In a world dominated by complex and frag- mented production processes, economic develop- ment can occur through economic upgrading and Global Global densification. Economic upgrading is largely about Chains Value chains value gaining competitiveness in higher value-added prod- ucts, tasks, and sectors. Densification involves engag- ing more local actors (firms and workers) in the GVC network. In some cases, this could mean that performing lower value-added activities on a larger scale can generate large value addition for the coun- Technology Know-how Richer skill-set try. Raising domestic labor productivity and skills contributes to the overall goal to increase a country’s value added as a result of GVC participation. Upgrading by empowering production sectors through The proponents of the “new paradigm” empha- size the role of functional upgrading (figure 1.8), moving to higher value-added tasks. But other forms of economic upgrading are equally relevant. Upgrading does not necessarily mean transitioning of more complex knowledge- and capital-intensive from an agricultural to a services economy, as tra- activities (figure 1.7). ditional development views suggest (development The ability of a small country such as Denmark in broad sectors, or the “old paradigm,” as dubbed to establish and maintain its position among the top by the GVC literature). It can instead mean increas- eight world exporters of food products exemplifies ingly embracing higher value-added production the relationship. Denmark achieved that position by with the contribution of better skills and know-how, massively applying information and communica- capital and technology, and processes (figure 1.9). In tions technologies and support services (R&D, logis- that sense, economic upgrading in GVCs via skills, tics, commercialization, advertising, and after-sale capital, and process upgrading overcomes the old services) to the production and processing of food. paradigm and extends the new paradigm focused Moreover, and linked to the first item, Denmark solely on functional upgrading (figure 1.8); it allows has made continuing efforts to upgrade products, achieving higher value-added production in the form processes, and functions, by introducing capital- of product, functional, and inter-sector upgrad- intensive inputs, thereby increasing value addition. ing. Denmark’s strength in global food production Digitization also makes every step of the produc- and Chile’s production of high-quality avocados tion process more productive and, in some cases, for export provide a clear case for improvement. changes the nature of production. Digitization trans- Improvement can occur by identifying (1) the tasks forms some goods into services (e-books, digital or activities of initial comparative advantage within news, and entertainment), and 3D printing trans- sectors, and then (2) policies to empower such activi- forms goods into online transfers of data that locate ties of comparative advantage, and (3) policies to the production process next to the consumer (box empower the underlying existing skills with technol- 1.2). Cloud computing, the “Internet of Things,” ogy and better human capital inputs. Following these and “Big Data” are transforming business models, steps may help countries to produce better qual- power relations, and sources of value added in entire ity products, establish more efficient processes, and Here’s Why 23 Figure 1.8. From Sector to Functional Upgrading Figure 1.9. Achieving Functional, Product, and Inter-sector Upgrading in All of a Country’s Old paradigm New paradigm Focus on final goods Focus on intermediates Agricultural, Manufacturing, and Services Production through Skills, Capital, and Process Upgrading Specialized services High Functional upgrading P re - pro d D Services value ology m en t added R& ct i High hn lo p on T ec u e dev value-added I n t e r-g production up r d in g grad u c t FROM LOW TO HIGH VALUE ADDED FROM LOW TO HIGH VALUE ADDED s e cdi n o a g ra l P p ro Design t u Medium- servic e s C o m m e r cial skilled Manufacturing support and sales ale za ti o r- s i n A ft e Capital upgrading Value- c productio asi added B n growth gr oc e s s u p g ra ng S ki di adi Commodities Primary lls ng Pp r u A ss e bly m Source: Adapted from Cattaneo and Mirodout 2013. ls and cap Note: R&D = research and development. skil ab nt Curre Tasks ilitie jump to higher value-added functions or develop tage s with specialization in more profitable industries (figure an mp v co 1.9). Part III outlines strategies that countries pursue a r a ti v e a d to do so. Position in GVC Networks and Type of GVC Node for example, as measured by the Bonancich power In the complex, multidimensional space of GVCs, index (Bonancich 1987) how do countries fare overall? Network analysis and 4. Clustering: the transitivity of the network—how metrics shed some light on this topic by capturing much the neighbors of country c are connected to the complexity and heterogeneity of actors and trade each other links (box 1.3). Large and dense networks may be 5. Visualization through a minimal spanning tree: a assessed more easily by creating a network topology, process that illustrates the network reporting the consisting of a set of centrality measures that capture strongest flow for each node (box 1.3). The most different aspects of the network. Stylized representa- connected countries—the central nodes, as they tions of the network make it easier to visualize some are the main trade partner for several countries— dominant aspects of the network and the actors. are the “roots” of the tree, distinguished from the The following are relevant measures: peripheral countries, the “leaves.” The size of the node reflects a country’s strength or centrality in 1. Strength: average flow for country c the network (figure 1.10). 2. Closeness: mean distance from country c to all other countries Network indexes and tree representations are use- 3. Centrality: the location of country c relative to the ful in many ways. They can be constructed to account overall structure of the network through mea- for the heterogeneity of trade links and, accordingly, sures of “structural integration” in the network— to visualize trade flows relevant to GVCs, such as 24 Making Global Value Chains Work for Development Box 1.3. What Is Special about Network Analysis? Finding Structure in Economic Problems Network analysis and metrics are primarily about finding structure necessarily cancel out.d Instead, the shocks affect aggregate in the data describing the link between the nodes (agents, coun- fluctuations through general equilibrium channels. The size of tries, and firms). This approach differs from traditional economet- the impact depends on the network structure, that is, the prop- rics in many ways. erties of the matrix corresponding to the underlying production The first difference is that network analysis accounts for het- relations, which can play a key role in determining the depth and erogeneity in the links between individual observations. That is frequency of large economic shocks, such as downturns.e not the case for traditional econometrics, which assumes one A third difference is that in econometrics, indicators need to of two corner cases, both assuming uniform links: a fully con- be related to independent variables. The network toolbox is more nected network or random connections. The difference is key, as eclectic and flexible, and it allows the analysis to accommodate it has conceptual and computational implications. It also clearly nonlinearity and topology, and, generally speaking, the full set of underscores that the usefulness of network analysis goes beyond relationships between variables, including those traditionally cov- visualizations of phenomena. In trade, for example, measures of ered by econometrics. Gravity modeling, the workhorse of empiri- eigenvector centrality are to be preferred to openness measures, cal international and spatial economics, for example, can also be as the former account not only for direct links, but also for indi- addressed through network analysis. The resulting concept of rect links.a And the links suggest new insights, for example, that connectivity is associated with economic benefits for the more a relatively central position in a production network (a new input connected nodes on transport networks. is used by already central technologies) makes an input’s wide A previous generation of dimensionality reduction tools— diffusion across the network more likely.b principal component analysis, multidimensional scaling, and clus- The second difference lies in the assessment of the structure tering—also looked for structure in the data, but network analy- of the network. Network analysis allows for several metrics to sis allows more visual identification of the relevant dimensions. synthesize a node’s complex and multidimensional set of char- For example, a network representation of a proximity matrix acteristics in one indicator, such as centrality metrics, with the will be more visual than the traditional dendrogram of a cluster- nodes retaining their full set of characteristics (or complexity). ing analysis. However, some more recent tools from complexity Standard econometrics would proceed otherwise to explore the and computer science can be superior—strictly speaking—to a structure of the network. It would, for instance, regress the values network toolbox for certain applications and retain some of its of the adjacency matrix against independent variables, working advantages in representing the same data. Two examples include de facto with averages. And the structure of the network mat- Kohonen’sf self-organizing map and nonlinear (exponential) com- ters. It determines the nature and size of impacts. For example, in ponent analysis. GVCs, input-output links can generate a cascade effect induced a. Santoni and Taglioni (2015). by the propagation of micro-shocks through the production b. See the results from the input diffusion model by Carvalho and Voigtländer network.c In traditional statistical concepts, the law of large num- (2015). bers would suggest that micro-shocks cancel out, as the distri- c. Acemoglu and others (2012); Acemoglu, Akcigit, and Kerr (2015); Acemoglu, Ozdaglar, and Tahbaz-Salehi (2015a, 2015b); Carvalho (2014). bution of a large universe of firms would tend to have a normal d. Gabaix (2011). distribution. But the “granular” hypothesis of Gabaix shows that e. Acemoglu and others (2015). with asymmetric distributions of firms, firm-level shocks do not f. Kohonen (1982). value-added trade. They can also illustrate other Multidimensional Nature of GVCs types of flows (for example, parts, components, ser- vices, or FDI) or flows in individual sectors or of spe- A multifaceted, multidirectional approach examines cific products. Network indexes allow observers to the nexus of goods, services, investments, workers, identify the position of individual countries in GVC and ideas in GVCs (box 1.1). Specifically, the frame- networks, their centrality, and the nature of the trade work covers tasks to produce goods and services and flows. factors of production—that is, capital flows, includ- Moreover, network measures, such as central- ing FDI, as well as workers, ideas, information, and ity and clustering, reveal the indirect links between intellectual property (patents, trademarks, and copy- countries.32 For example, trade in intermediates for rights). Connecting tasks with factors of production many Central American countries is connected to has become increasingly important because the qual- the United States through Panama, the region’s main ity and availability of production factors in a country logistics hub. Looking at various network measures in affect downstream activities in the GVC. combination allows analysts to detect that aspect of Looking at the relationships between some of Panama’s participation in trade and GVC networks. those components is not new. Economists have Here’s Why 25 Figure 1.10. Network Representation of Value-Added Trade, 2011 Iceland Denmark Norway United Kingdom Malta Mexico Turkey Hungary Brunei Darussalam South Africa Lithuania Tunisia New Zealand Chile Japan Belgium France Malaysia Australia Russian Federation Vietnam Sweden Saudi Arabia Canada Slovak Republic Germany Estonia Latvia United Indonesia Romania States Argentina Israel China Czech Republic Italy Bulgaria Netherlands Brazil Thailand Austria Portugal Finland Croatia Croatia Slovenia Cambodia Poland Spain Korea, Rep. Greece India Colombia Philippines Taiwan, China Switzerland Cyprus Ireland Singapore Luxembourg Hong Kong SAR, China Source: Santoni and Taglioni 2015. Note: In 2011, China, Germany, and the United States formed the three main roots of the value-added trade network. That is, those three countries were the most relevant sources and destina- tions for the value added embodied in other countries’ exports, especially neighboring economies. examined the relationship between trade and invest- market–enhancing outcomes for workers at home as ment for quite some time. The economics profession well as social upgrading and cohesion. has traditionally tended to view trade and investment GVCs can lead to development. But at the country as separate phenomena; the standard question was level, such constraints as inadequate skills, labor, and whether they were complements or substitutes.33 The absorptive capacity remain.34 GVCs can create new emphasis now is to look at them jointly. Similarly, opportunities on the labor demand side, but sup- economists and policy makers should analyze the ply and demand cannot meet if the supply is miss- tasks and production factors of a GVC jointly. ing. That potential gap illustrates the importance of To capture this concept of “jointness,” table A.1 in embedding national GVC policies in a broader port- appendix A describes examples of patterns expected folio of policies aimed at upgrading skills, improving in goods and services tasks as well as in the factors physical and regulatory infrastructure, and enhanc- of production, including workers, ideas, and invest- ing social cohesion.35 ments, depending on their role in GVCs. This addi- To address all these policy dimensions of GVC- tional information is rarely available in the form of led development, part III proposes a framework hard data and must be gathered primarily through that identifies three focus areas: (1) entering GVCs; surveys or field assessments. (2) expanding and strengthening participation in GVCs, through upgrading to higher value-added Policy Dimension: Entering GVCs, activities and densifying economic participation; and (3) turning GVC participation into sustainable Expanding Participation, and Ensuring development. The text links these focus areas with Sustainable Development specific objectives, strategic questions, and ensuing GVCs represent a new path for development. They policy options (figure O.1 in the Overview). can help LMICs accelerate industrialization and the “servicifying” of the economy. For policy mak- Joining GVCs: Policy Options to Facilitate GVC Entry ers, the focus is on shifting and improving access to resources while also advancing development goals— Integrating domestic firms (suppliers and final pro- and on whether GVC participation delivers labor ducers) into GVCs can help LMICs accelerate their 26 Making Global Value Chains Work for Development industrialization. Facilitating GVC entry requires reasons. Most studies of the backward links of firms in creating world-class GVC links and a world-class EPZs find the links to be minimal, with domestic trade climate for foreign tangible and intangible assets. remaining very low and technology spillovers rare.36 However, GVC participation is a necessary but not Ultimately, however, a sustainable and inclusive sufficient condition for development. Although GVCs policy of GVC participation and upgrading requires open doors, they are not magical. Most of the hard establishing a broadly competitive national environ- work still has to be done at home, with domestic ment for offshore FDI and domestic firms. Overall pro-investment, pro-skills, pro-jobs, and pro-growth investment attraction policies matter greatly. In reforms. Creating demand for high-productivity designing investment promotion measures, various workers must be matched with a supply of capable factors are important for policy makers to consider, workers who have the relevant skills. In other words, particularly measures that explicitly target foreign when thinking about the first step in facilitating GVC investors. However, policy makers should ensure that entry, policy makers must have a clear road map of the measures do not discriminate against domestic how entry will lead to strengthened and broader par- investors. Policy makers also must identify and attract ticipation and economic and social upgrading. Policy “the right” foreign investors. That endeavor includes makers must keep a keen eye on the workforce’s assessing the nature of investment and the motiva- competencies and how they match up with foreign tions of potential foreign investment (efficiency- investment. seeking, resource-seeking, or a market-seeking export platform), as well as its technology contribution and Creating World-Class GVC Links the technology gap with domestic firms. Investment Countries can join GVCs by facilitating the entry of promotion should not only focus on lead firms in domestic firms or by attracting foreign investors. The GVCs, but also target turnkey global suppliers, and FDI option includes more direct access to foreign possibly, important lower-tier suppliers.37 know-how and technology. Costa Rica and Thailand Moreover, countries can facilitate the participa- have managed to attract FDI and turn it into sus- tion of domestic firms through arm’s-length trade by tainable GVC participation in very different ways. In helping them find the right trade partners and tech- all cases, however, ensuring a set of conditions that nology abroad. That help can include setting up firm includes excellent infrastructure, streamlined export directories, offering practical advice, and promoting procedures, and a tariff-friendly environment is exports and imports more generally. Government necessary. assistance can also include e-tools to help domestic For LMICs that face significant infrastructure and companies commercialize their intellectual prop- regulatory gaps, establishing a broadly competitive erty, identify and take advantage of freely available national environment for offshore FDI is difficult. technologies, or assist them to establish licensing As a result, many of those countries seek to establish agreements. “competitive spaces”—enclave locations such as spe- GVC entry also requires the improvement of a cial economic zones (SEZs) and export processing country’s connectivity with international markets. zones (EPZs), where the rules of business are differ- Bad connectivity means high costs, low speed, and ent from those that prevail in the national territory high uncertainty. Thus, successful participation in and the costs of factors of production are lower. The GVCs requires policy makers not just to ensure effi- zones usually are rapidly built sites equipped with cient cross-border connections, but also to increase excellent infrastructure; streamlined customs, regu- the connectivity of domestic markets and enhance latory, and administrative procedures; and favorable the resilience and efficiency of the domestic segment tax conditions (such as tariff drawbacks on imports of the supply chain. of intermediates). Competitive spaces have played a central role in the development of GVCs in many sec- Creating a World-Class Climate for Firms’ Assets tors in “Factory Asia” and “Factory North America.” Low wages may be a way for countries to enter And in many lower-income countries, exports come GVCs, and low-wage industrial jobs can be a big overwhelmingly from such spaces. For SEZs, EPZs, pro ductivity step-up from subsistence agricul- and other competitive spaces to contribute to sus- ture, underemployment, and low-skill service jobs. tained economic development, however, they have to However, the goal should be higher labor produc- be linked to the rest of the economy. The problem is tivity, so the country can remain cost-competitive that, by their nature, they resist such links for several despite rising wages and living standards. Although Here’s Why 27 static labor productivity effects may be negative for reforming storage fees, which improved the country’s employment creation (if the same amount of value Logistics Performance Index score. added is created with fewer workers), GVC integra- Meanwhile, a light-handed industrial policy can tion has strong potential for dynamic productivity foster entry into GVCs by overcoming market failures gains via several transmission channels, as will be or capturing coordination externalities. An analogy is discussed in the next section. urban policy. If individual initiatives are completely What matters are unit labor costs, not wages. uncoordinated, the result can be over-congested cit- Chinese labor, for example, remains cost-effective ies that fail in their basic goal of improving the lives despite rising wages, because labor productivity is also of residents. At the other extreme, government con- rising. Moreover, lower unit labor costs alone are not trol of every investment decision can stifle growth sufficient; the capacity to meet production require- and innovation and fail to improve lives. A key dif- ments also must be taken into consideration.38 Put ference between GVC-led development and classical simply, lower labor costs will not attract GVC-linked models of development—through structural trans- foreign investors without the right infrastructure and formation from agriculture to manufacturing to ser- capacity building. Labor policies aimed at attracting vices—is that government coordination in GVC-led foreign investment should therefore be matched by models must take place at the micro level. Countries other initiatives, including packages of infrastructure do not have to pick a sector as the “winner”; they and public-private vocational training. must assess the existing skills and capabilities in the Removing restrictions and barriers to foreign country, domestic and international demand condi- investment and increasing the protection of foreign tions, and competition from other countries. Based assets are keys to attracting foreign investors. Those on such information, the government may then set efforts imply policies such as (1) allowing more for- in place the appropriate incentive framework on the eign equity into domestic companies,39 (2) facilitat- supply and demand sides. In this way, governments ing the movement and employment of key personnel, can help plan and encourage entry into the appropri- (3) relaxing domestic content rules when their role ate tasks and, consequently, densification of the GVC and purpose are not clearly defined, (4) relaxing rules participation that has already begun, while also fos- on foreign exchange and repatriation of benefits, and tering domestic demand for goods and services pro- (5) strengthening investor protection and the right to duced domestically—all on a market-driven basis. challenge domestic regulations and decisions. The sophistication and competitiveness of domestic firms are essential conditions in the host Completing the Firms’ Ecosystem: Policy Options to economy. Countries that are home to large and com- Expand Development beyond the Initial GVC Enclave petitive firms have an advantage in attracting foreign investors and fostering the participation of domes- Strengthening GVC–Local Economy Links tic firms through arm’s-length trade, because the After a country enters GVCs, the next set of policy domestic firms can act as turnkey suppliers. Some considerations must ensure that the GVCs are as inte- domestic firms also have the potential to become grated as possible into the domestic economy. The lead firms. Countries with predominantly SMEs find logic of this effort is that strong links with the domes- entering GVCs more difficult unless the SMEs are tic economy should result in greater diffusion of part of a well-established and integrated industrial knowledge, technology, and know-how from foreign cluster, such as the Italian industrial districts.40 investors. The problem is that foreign investors do not The benefits of efficient transportation and logis- actively pursue—and sometimes resist—such integra- tics at the border could be undermined by inefficient tion, for reasons that range from economic constraints, domestic links (such as the unreliability or high cost to technological and quality gaps with domestic sup- of domestic transportation or the lack of cold chains pliers, to shortages of specialized workers and skills. for fresh products), as well as regulatory bottlenecks. For policy makers, the goal is to turn economic Foreign investors evaluate the ease of access to effi- upgrading and densification through GVC participa- cient services and infrastructure in the host coun- tion into sustainable development. Part of that effort try, including access to cheap and reliable energy, should include understanding how the potential finance and trade support, telecommunications (for for FDI spillovers differs across firms, sectors, and e-commerce or electronic transfers), and transport. tasks—and designing investment attraction policies For example, Indonesia reduced vessel dwell time by that do not discriminate against domestic players.41 28 Making Global Value Chains Work for Development It is also important to ask what economic upgrading These transmission channels enable GVCs to sup- through GVCs means for average living standards— port development and industrialization efforts in such as employment, wages, work conditions, and four ways (figure 1.11).42 economic security—and for wider social upgrad- First, GVCs—through backward value chain ing, distributional concerns, and nonmaterial factors links—generate a demand effect and an assistance such as democracy, labor rights, human rights, gen- effect in the host country: der equality, environment, cultural issues, respect for minority rights, and more. • Demand effect. Lead firms tend to require more or The main transmission channels for economic better inputs from local suppliers. and social upgrading include the following: • Assistance effect. Lead firms can assist local suppliers through knowledge and technology • Forward links: sales of GVC-linked intermediates sharing, advance payments, and other types of to the local economy, spurring production and/or assistance. productivity in various downstream sectors • Backward links: GVC-linked purchases of local In addition, the forward and backward links gen- inputs, spurring production and/or productivity erate technology spillovers, improving the produc- in various upstream sectors tivity of local firms through two mechanisms: • Technology spillovers: improved productivity of local firms in the same or related downstream or • Diffusion effect. The assistance effect leads to dif- upstream sectors as a result of GVC production fusion of knowledge and technology in the sup- • Skills upgrading: similar to technology spillovers, pliers’ industry. but transferred through the training of and • Availability and quality effects. GVC participation demand for skilled labor increases the availability and quality of inputs in • Minimum scale: for example, GVC participation the buyer’s industry. may stimulate investments in infrastructure that would otherwise not be profitable and that may Second, GVC participation can translate into spur local production in other sectors pro-competitive, market-restructuring effects that Figure 1.11. GVC Transmission Channels Technology Diffusion spillovers effect Demonstration Availability effect & quality effect Market Backward/ restructuring forward links Pro-competition Demand effect effect Demonstration Assistance effect effect Amplification of Demand effect Minimum pro-competition Domestic Training effect Labor scale effect impact of GVC Labor turnover markets achievements Sustainability participation effect effect Here’s Why 29 are not limited to GVC participants but extend to • Labor turnover effect. Knowledge embodied in the nonparticipants: workforce of participating firms (such as MNCs or their local suppliers) moves to other local firms. • Pro-competition effect. GVC participation increases competition for the limited resources in Strengthening Absorptive Capacity the country (between foreign investors and local The degree to which local firms and workers benefit firms, but also between participants and nonpar- from knowledge and technology spillovers ultimately ticipants in GVCs), increasing overall average pro- depends on the absorptive capacity of domestic ductivity in the medium run.43 actors. This is the most important area of GVC spill- • Demonstration effect. Knowledge and technology over policy, particularly in helping local firms and spillovers arise from direct imitation or reverse workers access opportunities. Building the absorp- engineering by local firms (GVC and non-GVC tive capacity of local firms requires general and participants) of GVC products, business models, industry-specific investments to upgrade technical marketing strategies, production processes, and capacity and, most important, to achieve quality and export processes. efficiency standards for production and export. Industry-specific and general policies—for educa- Third, minimum scale achievements have a two- tion, standard-setting, and innovation—and comple- fold impact: mentary policies—such as development of adequate supplier programs or effective clusters—are critical • Amplification effect. Minimum scale achievements for sustaining long-term spillovers. Bolstering pro- amplify pro-competition effects. The achieve- ductivity and production and innovation capacities, ments stimulate investment in infrastructure and including human capital and other resources, can be backbone services, which would not be realized achieved through coordinated efforts over a range without the scale of activity generated by GVCs. of initiatives. The initiatives may include developing Once the infrastructure is in place, it is likely to public-private partnerships aimed at research and spur local production in other sectors and the development collaboration, increasing the supply of non-GVC economy. sufficiently qualified researchers in local universities, • Sustainability effect. Minimum scale achieve- and aligning higher education curricula and training ments also strengthen the ability of the country specializations with local economic activities. Policy to sustain GVC participation over time. The GVC makers should also help domestic firms comply with literature is rife with examples of the key role of process and product standards. Such public, private, improvements in backbone infrastructure and or voluntary standards must be respected throughout services, such as logistics, to improve timeliness the entire value chain, because every stage of pro- and reliability in transporting goods, parts, and duction can affect the quality of the final product or components, and therefore enable countries to service, which could affect the lead firm’s reputation. integrate vertically into GVCs.44 Over the long term, a country cannot be competitive in GVCs by offering a single task (for example, assem- Fourth, GVCs benefit labor markets through the bly); it must offer a bundle of tasks. Diversifying into following mechanisms: service tasks and promoting service exports offer a largely untapped income potential for many LMICs.45 • Demand effect. GVC participation is characterized by higher demand for skilled labor from MNCs or Creating a World-Class Workforce other GVC participants. Multinationals may tem- Developing skills is a key element of competitiveness, porarily bid away human capital by paying higher and it affects the ability to participate in GVCs and wages or offering enhanced employment benefits. achieve economic and social upgrading within GVCs. The effect tends to dim, however, as soon as the Economic upgrading requires the availability of new productivity of domestic firms is raised or the skills and knowledge by increasing the skill content market adjusts to the tighter labor supply. of a country’s activities (and thus its workforce) or • Training effect. Local firms participating in GVCs developing competencies in niche market segments.46 are more likely to receive training (say, from Economic upgrading and social upgrading are thus MNCs or their international buyers). linked to and dependent on each other. Indeed, lead 30 Making Global Value Chains Work for Development firms have strong incentives to train their workforces Implementing Climate-Smart Policies and to comply with the firms’ standards. Beyond private Infrastructure initiatives, there is a strong case for public investment Firms today are more vulnerable than ever to shifts in skills development to meet the needs of interna- in the economy and exogenous disruptions, such tional trade and participation in GVCs.47 as climate change. Climatic disruption can impair Economic upgrading may drive social upgrad- firms’ ability to access inputs and deliver final prod- ing, but that is not automatic. Complementary policy ucts, making countries’ preparedness an increasingly could promote social upgrading and maximize the critical factor in firms’ location decisions. Climate sustainable development impact of GVC activities. change is a multi-sector, uncertain phenomenon, Social policies are needed to create an equitable dis- which makes evaluating economic impacts and tribution of opportunities and outcomes. Without designing robust and appropriately prioritized adap- social cohesion and policies ensuring that all segments tation strategies difficult for countries. of society benefit from GVC participation, develop- The global trade landscape is trending toward ment would be unsustainable. Social upgrading can more climate-friendly international standards and derive from labor regulation and monitoring, such mandatory sustainability reporting regimes, including as occupational safety, health, and environmental on issues of (1) wildlife trafficking, (2) illegal logging, standards in GVC production sites. Well-functioning (3) sustainable management of ocean and coastal labor markets are also important, because integrating resources, (4) energy efficiency, (5) infrastructure into GVCs requires reallocating resources. for electric vehicles, (6) responsible mining practices, For social upgrading to translate into social cohe- (7) chemical health and safety cooperation, (8) trade sion through better living standards, countries must in environmental goods, and (9) aviation emissions. ensure equal opportunities to strengthen social But for countries to comply with such standards, cohesion by (1) creating a sense of belonging and long-term strategic policy responses will be neces- active participation, (2) promoting trust, (3) offer- sary. This will require the mainstreaming of a triple ing upward social mobility, and (4) fighting inequal- bottom-line approach to planning, one that accounts ity and exclusion. Equal access to jobs (including for financial, social, and environmental policy impli- for women and minorities) is the most important cations. Increasing the scale of global production opportunity in GVCs. Access to widely advertised requires carefully planned investments in infrastruc- information about job vacancies and practical advice ture. With an effective strategic vision, countries can on how to get those jobs (through job search assis- strengthen the ability of their firms to sustain GVC tance) is a precondition for equality of access to jobs. participation over time. In addition, workers need to be informed about their rights. Despite their important roles in the Notes labor market, farmers, the self-employed, and infor- mal workers in particular often are unaware of their 1. The phenomenon has been called “vertical special- rights in relation to landowners, traders, or employ- ization” by Balassa (1967) and Findlay (1978), “slicing up of the value chain” by Krugman (1995), and many other ers. Cooperatives, associations, and trade unions can names by other economists, including “international frag- be effective channels of information. mentation of production” (Arndt and Kierzkowski 2001), But those information channels require that free- “transnational production” (Feenstra 1998), and “global dom of association and collective bargaining rights production networks” (Ernst and Kim 2002; Henderson already exist in the country. Such provisions encour- and others 2002). Vertical specialization identifies a pro- age proactive social dialogue that can address ten- duction structure in which tasks and business functions are spread over several companies that are globally or sions before they lead to conflict. And facilitating regionally dispersed. Tasks, rather than sectors, define the access to jobs for excluded or disadvantaged groups specialization of countries in the value chains, as indicated helps economies tap a largely idle segment of the by Grossman and Rossi-Hansberg (2008). workforce with productive potential, and increases 2. ACET (2014). social cohesion. Antidiscrimination laws and manda- 3. Dunbar (2013). 4. Baldwin (2012). tory or voluntary affirmative action programs—such 5. Cali and Hollweg (2015). as proactive measures for hiring women, minori- 6. WEF (2013). ties, or other groups—are important measures for 7. Hummels and others (2007). achieving greater equality of opportunity.48 8. Christ and Ferrantino (2011). Here’s Why 31 9. Feenstra and others (2013). 25. WTO-OECD (2013). 10. Rodrik (2000, 179). 26. Sagrario, Pierrestegui, and Mas (2009). 11. Antràs and Yeaple (2014). Antràs and Chor (2013) 27. Vernon (1966, 1979). lists a range of reasons for incomplete contracting in inter- 28. Porter (2008). national settings, including the limited amount of repeated 29. Importing to export on the sales and sourcing sides interactions; lack of collective punishment mechanisms is related to the bilateral concepts of backward and forward associated with international transactions; and natural vertical specialization introduced by Lopez-Gonzalez and difficulties in contract disputes involving international Holmes (2011), in which “backward” refers to sourcing transactions, such as determining which country’s laws and “forward” to sales. That usage contradicts the standard are applicable to the specific contract. Finally, even when usage of backward and forward links from economic geog- the relevance of the law to the contract in question is clear, raphy, in which “backward links” refer to sales and “for- local courts may be reluctant to enforce a contract involv- ward links” refer to sourcing (Ottaviano and Puga 2003). ing residents of foreign countries. 30. Gereffi and Frederick (2010). 12. The extent of vertical competition varies, depending 31. Buciuni, Coro, and Micelli (2013). on the power relations within the specific value chain (see, 32. Santoni and Taglioni (2015). for example, Milberg 2004). Interestingly, horizontal and 33. Brainard (1993); Helpman (1984); Horstmann and vertical competition are driven by similar forces: the inter- Markusen (1992). play between traditional cost advantages, institutional fac- 34. Kummritz, Taglioni, and Winkler (2016). tors, and proximity to the final consumer, which together 35. Kummritz, Taglioni, and Winkler (2016). determine which tasks are more profitable in given loca- 36. Milberg and Winkler (2013). tions (Cattaneo and others 2013). 37. Farole and Winkler (2014b). 13. Daihatsu licenses the Terios SUV technology (an 38. Cattaneo and others (2013). older technology that has been phased out in the Japanese 39. China has been effective in attracting FDI, even with domestic market) to FAW (Paultan.org). The latter engages restrictions on joint ventures. However, that success is in the manufacture and sale of passenger cars and related largely a result of China-specific conditions: a large domes- accessories. FAW offers its products under three brand tic market and a large pool of low-cost but well-trained names: Benteng, Mazda6/Atenza, and Hongqi. Some of the workers. Countries that do not have specific factors to Benteng cars are produced using old models of the Mazda attract investors or to use as leverage will have less space for sedan, and others using the second-generation Volkswagen maneuvering when dictating joint venture conditions with Jetta. The company also produces the Mazda6/Atenza for foreign investors. the Chinese and Japanese markets. The production and 40. Becattini (1990); Porter (1990). commercialization of this model is outsourced by Mazda 41. Understanding the spillover potential of differ- Japan, a competitor of Daihatsu. The advantage for Mazda ent FDI at the micro level is likely to become an impor- is that it can focus on models that are more strategic from tant policy priority in the coming years. This is not the a corporate point of view, such as Premacy and Familia. case only for small and lower-income countries that rely Finally, FAW has its own brand: the Hongqi luxury car increasingly on FDI and have a limited pool of resources (FAW corporate website: http://www.faw.com/). Hongqi to devote to attracting foreign investors, but also for large cars have been manufactured since 1958; the original countries. Another important priority in designing FDI- models were reserved for the high-ranking party elite of related policy should be ensuring that the incentives used the communist party. They remained in production until to attract foreign investors do not create a bias against local 1981 (The Economist, “The Home Team,” November 13, integration. Moreover, policy makers must leverage invest- 2008, www.economist.com/node/12544893). The current ment incentives to promote spillovers actively, including Hongqi fleet includes the H7, which is an executive car local supplier development, provision of technical assis- based on the Toyota Crown platform. This intricate system tance, and training of workers, joint research, and more. of collaboration and business relationships is an excellent The spotlight should be on value addition rather than example of the degree of vertical competition in the auto- in-country ownership. Instead of establishing rigid local motive sector. content requirements, the focus should be on collabora- 14. For the distinction between value chains and supply tive development of flexible localization plans, in which chains, refer to box 1.1. investors come up with their own proposals on how they 15. UNCTAD (2013). will deliver spillovers to the local economy. Incentivizing 16. Freeman (2006). foreign investors to collaborate with local universities, 17. Baldwin (2011). research institutes, and training institutes is also important 18. See Sturgeon (2002) for a detailed case study on the (Farole and Winkler 2014a). trend toward external sourcing in the electronics industry. 42. The discussion on mechanisms triggered by GVC 19. Sturgeon and Lester (2004). participation evolves partially from the taxonomy intro- 20. Humphrey (2003). duced by Farole, Staritz, and Winkler (2014). 21. Kawakami (2011). 43. In the short run, average productivity may decrease 22. Pine (1999). and local firms may lose market shares because of intensi- 23. Saez and others (2015). fied competition. 24. Hoekman (2014). 44. WEF (2013). 32 Making Global Value Chains Work for Development 45. Cattaneo and others (2013). Brainard, S. Lael. 1993a. “An Empirical Assessment of 46. Humphrey and Schmitz (2002). the Factor Proportions Explanation of Multinational 47. Cattaneo and others (2013). Sales.” NBER Working Paper 4583, National Bureau of 48. OECD (2011); World Bank (2013). Economic Research, Cambridge, MA. Breznitz, Dan, and Michael Murphree. 2011. Run of the Red Queen. Government, Innovation, Globalization, References and Economic Growth in China. New Haven, CT: Yale University Press. Acemoglu, Daron, Ufuk Akcigit, and William Kerr. 2015. Buciuni, Giulio, Giancarlo Coro, and Stefano Micelli. 2013. “Networks and the Macroeconomy: An Empirical “Rethinking the Role of Manufacturing in Global Value Exploration.” NBER Macroeconomics Annual 2015, Chains: An International Comparative Study in the volume 30. National Bureau of Economic Research, Furniture Industry.” Industrial and Corporate Change Cambridge, MA. (4): 1–30. Acemoglu, Daron, Vasco M. Carvalho, Asuman Ozdaglar, Cali, M., and C. Hollweg. 2015. “The Labor Content of and Alirez Tahbaz-Salehi. 2012. “The Network Origins Exports in South Africa: A Preliminary Exploration.” of Aggregate Fluctuations.” Econometrica 80 (5): World Bank, Washington, DC. 1977–2016. Carvalho, Vasco M. 2014. “From Micro to Macro via Acemoglu, Daron, Asuman Ozdaglar, and Alireza Production Networks.” Journal of Economic Perspectives Tahbaz-Salehi. 2015a. “Microeconomic Origins of 28 (4): 23–48. Macroeconomic Tail Risks.” NBER Working Paper Carvalho, Vasco M., and Nico Voigtländer. 2015. “Input 20865, National Bureau of Economic Research, Cambridge, MA. Diffusion and the Evolution of Production Networks.” ———. 2015b. “Networks, Shocks, and Systemic Risk.” Economics Working Papers 1418, Department of NBER Working Paper 20931, National Bureau of Economics and Business, Universitat Pompeu Fabra, Economic Research, Cambridge, MA. Barcelona, Spain. ACET (African Center for Economic Transformation). Cattaneo, Olivier, Gary Gereffi, Sébastien Miroudot, and 2014. 2014 African Transformation Report: Growth with Daria Taglioni. 2013. “Joining, Upgrading and Being Depth. Accra, Ghana: ACET. Competitive in Global Value Chains: A Strategic Antràs, Pol, and Davin Chor. 2013. “Organizing the Global Framework.” Vol. 1. Policy Research Working Paper Value Chain.” Econometrica 81 6: 2127–204. 6406, World Bank, Washington, DC. Antràs, Pol, and Stephen R. Yeaple. 2014. “Multinational Cattaneo, O., and S. Miroudot. 2013. “From Global Value Firms and the Structure of International Trade.” Chains to Global Development Chains: An Analysis of Handbook of International Economics, 4: 55–130. Recent Changes in Trade Patterns and Development Arndt, Sven W., and Henryk Kierzkowski. 2001. Paradigms.” In 21st Century Trade Policy: Back to the Fragmentation: New Production and Trade Patterns in Past? Volume in Honor of Professor Patrick Messerlin, the World Economy. Oxford: Oxford University Press. edited by E. Zedillo and B. Hoekman. New Haven, CT: Balassa, B. 1967. Trade Liberalization among Industrial Yale University Press. Countries. New York: McGraw-Hill. Christ, Nannette, and Michael J. Ferrantino. 2011. “Land Baldwin, Richard. 2011. “Trade and Industrialisation after Transport for Exports: The Effects of Cost, Time, and Globalisation’s Second Unbundling: How Building Uncertainty in Sub-Saharan Africa.” World Development and Joining a Supply Chain Are Different and Why It 39 (10): 1749–59. Matters.” NBER Working Paper 17716, National Bureau Dunbar, Muriel. 2013. “Engaging the Private Sector in of Economic Research, Cambridge, MA. Skills Development.” Health and Education Advice and ———. 2012. “Global Supply Chains: Why They Emerged, Resource Team, Oxford, United Kingdom. Why They Matter, and Where They Are Going.” Ernst, Dieter, and Linsu Kim. 2002. “Global Production Discussion Paper 9103, Centre for Economic Policy Networks, Knowledge Diffusion, and Local Capability Research, London. Formation.” Research Policy 31: 1417–29. Baldwin, Richard, and Javier Lopez-Gonzalez. 2013. Farole, T., C. Staritz, and D. Winkler. 2014. “Conceptual “Supply-Chain Trade: A Portrait of Global Patterns Framework.” In Making Foreign Direct Investment and Several Testable Hypotheses.” NBER Working Work for Sub-Saharan Africa: Local Spillovers and Paper 18957, National Bureau of Economic Research, Competitiveness in Global Value Chains, edited by Cambridge, MA. T. Farole and D. Winkler, 23–55. Washington, DC: Becattini, Giacomo. 1990. “The Marshallian Industrial World Bank. Districts as a Socio-Economic Notion.” In Industrial Farole, Thomas and Deborah Winkler. 2014a. “Policy Districts and Interfirm Co-Operation in Italy, edited Implications.” In Making Foreign Direct Investment by Frank Pyke, Giacomo Becattini, and Werner Work for Sub-Saharan Africa: Local Spillovers and Sengenberger. Geneva: International Institute for Labor Competitiveness in Global Value Chains, edited by Studies. Thomas Farole and Deborah Winkler, 263–79. Bonancich, Phillip. 1987. “Power and Centrality: A Family Washington, DC: World Bank. of Measures.” American Journal of Sociology 92 (5): ———. 2014b. “The Role of Mediating Factors for FDI 1170–82. Spillovers in Developing Countries: Evidence from a Here’s Why 33 Global Dataset.” In Making Foreign Direct Investment Trade (Natura Facit Sal-tum).” Journal of International Work for Sub-Saharan Africa: Local Spillovers and Economics 32 (1–2): 109–29. Competitiveness in Global Value Chains, edited by Hummels, David, Peter Minor, Matthew Reisman, and T. Farole and D. Winkler, 59–86. Washington, DC: Erin Endean. 2007. “Calculating Tariff Equivalents World Bank. for Time in Trade.” Technical Report submitted to the Feenstra, Robert C. 1998. “Integration of Trade and U.S. Agency for International Development, Nathan Disintegration of Production in the Global Economy.” Associates, Arlington, VA. Journal of Economic Perspectives 12 (4): 31–50. Humphrey, John. 2003. “Globalization and Supply Chain Feenstra, Robert, and Gary Hamilton. 2006. Emergent Networks: The Auto Industry in Brazil and India.” Economies, Divergent Paths: Economic Organization and Global Networks 3 (2): 121–41. International Trade in South Korea and Taiwan, China. Humphrey, J., and H. Schmitz. 2002. “How Does Insertion Cambridge, UK: Cambridge University Press. in Global Value Chains Affect Upgrading in Industrial Feenstra, Robert C., Chang Hong, Hong Ma, and Barbara Clusters?” Regional Studies 36 (9): 1017–27. J. Spencer. 2013. “Contractual versus Non-Contractual Japan Automobile Manufacturers Association, Inc. 2013. Trade: The Role of Institutions in China.” Journal of The Motor Industry of Japan 2013. Tokyo, Japan: Japan Economic Behavior & Organization 94 (C): 281–94. Automobile Manufacturers Association, Inc. Findlay, Ronald. 1978. “An ‘Austrian’ Model of International Kawakami, Momoko. 2011. “Inter-firm Dynamics of Trade and Interest Rate Equalization.” Journal of Notebook PC Value Chains and the Rise of Taiwanese Political Economy 86: 989–1008. Original Design Manufacturing Firms.” In The Freeman, Richard 2006. “The Great Doubling: America Dynamics of Local Learning in Global Value Chains; in the New Global Economy.” Usery Lecture, April 8, Experiences from East Asia , edited by Momoko Georgia State University. http://emlab.berkeley.edu Kawakami and Timothy Sturgeon, chapter 1. Basing- /users/webfac/eichengreen/e183_sp07/great_doub.pd. stoke, UK: Palgrave Macmillan. Gabaix, Xavier. 2011. “The Granular Origins of Aggregate Kohonen, Teuvo 1982. “Self-Organized Formation of Fluctuations.” Econometrica 79 (3): 733–72. Topologically Correct Feature Maps”. Biological Gereffi, Gary. 1999. “International Trade and Industrial Cybernetics 43 (1): 59–69. Upgrading in the Apparel Commodity Chain.” Journal Krugman, Paul. 1995. “Growing World Trade: Causes and of International Economics 48: 37–70. Consequences.” Brookings Papers on Economic Activity Gereffi, Gary, and Stacey Frederick. 2010. “The Global 26 (1): 327–77. Apparel Value Chain, Trade, and the Crisis: Challenges Kummritz, Victor, Daria Taglioni, and Deborah Winkler. and Opportunities for Developing Countries.” In 2016. “Economic Upgrading through Global Value Global Value Chains in a Postcrisis World, edited by Chain Participation: Which Policies Increase the Value O. Cattaneo, G. Gereffi, and C. Staritz, 157–208. Added Gains?” World Bank, Washington, DC. Washington, DC: World Bank. Lopez-Gonzalez, Javier, and Peter Holmes. 2011. “The Gereffi, Gary, John Humphrey, Raphael Kaplinsky, and Timothy J. Sturgeon. 2001. “Introduction: Nature and Evolution of Vertical Specialisation: What Globalisation, Value Chains and Development.” IDS Is the Role of Preferential Trade Agreements?” Trade Bulletin 32 (3): 1–8. Working Paper No. 2011/41, Swiss National Centre of Gereffi, Gary, John Humphrey, and Timothy J. Sturgeon. Competence in Research, Bern, Switzerland. 2005. “The Governance of Global Value Chains.” Review Manyka, James, Jacques Bughin, Susan Lund, Olivia of International Political Economy 12 (1): 78–104. Nottebohm, David Poulter, Sebastian Jauch, and Sree Grossman, Gene M., and Esteban Rossi-Hansberg. 2008. Ramaswamy. 2014. “Global Flows in a Digital Age: “Trading Tasks: A Simple Theory of Offshoring.” How Trade, Finance, People, and Data Connect the American Economic Review 98 (5): 1978–97. World Economy.” McKinsey Global Institute Report, ———. 2012. “Task Trade between Similar Countries.” McKinsey & Company, Washington, DC. Econometrica 80 (2): 593–629. Milberg, William. 2004. “The Changing Structure of Trade Helpman, Elhanan. 1984. “A Simple Theory of Linked to Global Production Systems: What Are the International Trade with Multinational Corporations.” Policy Implications?” International Labour Review 143 Journal of Political Economy 92 (3): 451–71. (1–2): 45–90. Henderson, Jeffrey, Peter Dicken, Martin Hess, Neil Coe, Milberg, William, and Deborah Winkler. 2013. Outsourcing and Henry Wai-Chung Yeung. 2002. “Global Production Economics: Global Value Chains in Capitalist Networks and the Analysis of Economic Development.” Development. New York: Cambridge University Press. Review of International Political Economy 9 (3): 436–64. OECD (Organisation for Economic Co-operation Hoekman, Bernard. 2014. “Supply Chains, Mega-Regionals and Development). 2011. “Perspectives on Global and Multilateralism: A Road Map for the WTO.” Development 2012: Social Cohesion in a Shifting Research Paper RSCAS 2014/27, Robert Schuman World.” OECD, Paris. Centre for Advanced Studies, Centre for Economic Ottaviano, Gianmarco I. P., and Diego Puga. 2003. Policy Research, London. “Agglomeration in the Global Economy: A Survey of Horstmann, Ignatius J., and James R. Markusen. 1992. the ‘New Economic Geography.’” World Economy 21 “Endogenous Market Structures in International (6): 707–31. 34 Making Global Value Chains Work for Development Pine, B. Joseph, II. 1999. Mass Customization: The New ———. 2002. “Modular Production Networks: A New Frontier in Business Competition. Cambridge, MA: American Model of Industrial Organization.” Industrial Harvard Business School Press. and Corporate Change 11 (3): 451–96. Pisano, Gary P., and Willy C. Shih. 2009. “Restoring Sturgeon, Timothy, and Richard K. Lester. 2004. “The American Competitiveness.” Harvard Business Review New Global Supply-Base: New Challenges for Global 87 (7–8). Suppliers in East Asia.” In Global Production Networking Ponte, Stefano, and Peter Gibbon. 2005. “Quality and Technological Change in East Asia, edited by Standards, Conventions and the Governance of Global Shahid Yusuf, M. Anjum Altaf, and Kaoru Nabeshima. Value Chains.” Economy and Society 34 (1): 1–31. Washington, DC: World Bank. Porter, Michael E. 1985. Competitive Advantage: Creating UNCTAD (United Nations Conference on Trade and and Sustaining Superior Performance. New York: Free Development). 2013. “World Investment Report Press. 2013—Global Value Chains: Investment and Trade for ———. 1990. “The Competitive Advantage of Nations.” Development.” UNCTAD, Geneva. Harvard Business Review 78 (2): 73–93. Vernon, Raymond. 1966. “International Investment and ———. 2008. “The Five Competitive Forces That Shape International Trade in the Product Cycle.” Quarterly Strategy.” Harvard Business Review 86 (1): 78–93. Journal of Economics 80 (2): 190–207. Rodrik, Dani. 2000. “How Far Will Economic Integration ———. 1979. “The Product Cycle Hypothesis in a Go?” Journal of Economic Perspectives 14 (1): 177–86. New International Environment.” Oxford Bulletin of Saez, Sebastian, Daria Taglioni, Erik van der Marel, Economics and Statistics 41 (4): 255–67. Claire H. Hollweg, and Veronika Zavacka. 2015. WEF (World Economic Forum). 2012. “The Future of Valuing Services in Trade: A Toolkit for Competitiveness Manufacturing.” Global Agenda Council on the Future Diagnostics. Washington, DC: World Bank. of Manufacturing, WEF, Cologny, Switzerland. Sagrario, Ins, Mara Pierrestegui, and Gabriel Mas. ———. 2013. “Enabling Trade: Valuing Growth 2009. “Combining Strategic Analysis and Change Opportunities.” WEF, Geneva. Management: A Tool for Improving the Competitive- World Bank. 2013. World Development Report: Jobs. ness of Firms.” Micro Report No. 154, U.S. Agency for Washington, DC: World Bank. International Development, Washington, DC. ———. 2016. World Development Report 2016: Digital Santoni, Gianluca, and Daria Taglioni. 2015. “Networks Dividends. Washington, DC: World Bank. and Structural Integration in Global Value Chains.” In WTO-OECD (World Trade Organization and Organisa- The Age of Global Value Chains, edited by Joao Amador tion for Economic Co-operation and Development). and Filippo di Mauro. Washington, DC: Center for 2013. Aid for Trade and Value Chain in Agrifood . Economic Policy Research. Geneva: WTO. Sturgeon, Timothy J. 2001. “How Do We Define Value Chains and Production Networks?” IDS Bulletin 32 (3): 9–18. Chapter 2 CONSIDER BULGARIA Introduction Bulgaria’s Domestic Value This chapter examines the case of Bulgaria, which Added in Exports we use to illustrate how the tools described in this In a world of GVCs, what matters is the growth of book exemplify a country’s integration and eco- a country’s domestic value added embodied in gross nomic upgrading in global value chains (GVCs). exports, because a significant share of gross exports The chapter shows how analysts can make use of the may consist of foreign value added (via imported quantitative tools described in part II of the book— inputs). Domestic value added consists of value as well as the strategic policy framework developed added that is created in an industry (direct domestic in part III—to identify a country’s position in GVCs, value added), value added that is created in other sec- its scope for economic upgrading, and policy sugges- tors supplying the industry (indirect domestic value tions to achieve that goal. In particular, the chapter added), and re-imported intermediates (see chap- explores how value addition from gross exports has ter 4). Bulgaria’s growth of domestic value added changed over time in Bulgaria and how it is linked to embodied in gross exports shows moderate over- the country’s participation in GVCs. For easy naviga- all performance compared with its peer countries. tion, the measures and concepts in this chapter refer- Bulgaria’s compound annual growth rate between ence the relevant chapters or charts in the book.1 1995 and 2008 reached 10.7 percent and, thus, was Identifying the effect of GVC integration on higher than the rate in established GVC countries country performance is exacerbated by the fact that such as Germany (6.9 percent), the United States (5.7 the exogeneity of GVC participation is often difficult percent), and Japan (3.3 percent). However, Bulgaria to establish. If GVC integration is endogenous to trailed China (17.5 percent) and the regional peers, changes in the overall economic environment or to in particular, Romania (14.8 percent), Poland (13.8 developments in politically influential parts of the percent), and the EU-102 (13.1 percent). Among the economy, the causality between GVC participation comparator countries, only Turkey (10.4 percent) and country performance could run in both direc- and Portugal (6.7 percent) performed worse than tions. This makes Bulgaria a good candidate to eval- Bulgaria (figure 2.1). uate, as it became a member of the European Union Bulgaria’s growth in domestic value added (EU) in 2007. The close supervision of reform embodied in gross exports varies significantly across progress by the European Commission represents a sectors (for the methodology to identify key GVCs rare example of a mostly exogenous policy reform, in a country, see box 2.1 and chapter 3). Agriculture, including the country’s openness to flows of capital, textiles, leather and footwear, and basic metals and labor, trade, and ideas. This suggests that the open- fabricated metal products performed well com- ing of sectors in Bulgaria was an exogenous develop- pared with the EU-10 average over 1995–2008, but ment rather than a response to domestic lobbying or the performance of other product groups was lack- other unobservable factors. luster. Growth in agriculture was 9 percent, placing 35 36 Making Global Value Chains Work for Development Figure 2.1. Growth of Domestic Value Added Embodied and optical equipment, and in particular food and in Gross Exports, 1995–2011 beverages. Overall, Bulgaria’s growth seems to have China been largely driven by the traditional sectors and has lagged in the medium- and high-tech sectors, show- Romania ing a structure similar to that of Portugal. Poland Channels for Increasing Domestic Value EU-10 Added in Exports Bulgaria Between 1995 and 2008, the main source of growth Turkey in domestic value added embodied in exports was the increase in direct value added generated by export- Germany ers, which was 62.1 percent of the total (figure 2.2). The value-added growth contribution generated Portugal through backward links—that is, by the domestic United States supply sectors—was 37.8 percent. Growth of re- imported intermediates was negligible. The growth Japan of the direct value added by Bulgarian exporters is likely to be associated with improvements in produc- 0 4 8 12 16 20 tivity and export competitiveness via participation in (CAGR, %) GVCs. Chapter 1 discusses the transmission channels Source: Adapted from the Organisation for Economic Co-operation and Development–World Trade Organization Trade in Value Added database. for economic and social upgrading in GVCs (see fig- Note: See chapter 4 for more information on the decomposition. CAGR = ure 1.11 in chapter 1). compound annual growth rate; EU = European Union. International GVC Links the country third after Romania (13 percent) and Portugal (10 percent) and above the EU-10 average Macro Perspective: Bulgaria’s Structural (8.3 percent). Bulgaria’s growth also exceeded that of Integration in GVCs the EU-10 in textiles, leather and footwear, and basic Network analysis and metrics help capture the com- metals but fell significantly short of the EU-10 aver- plexity and heterogeneity of actors and trade links age in machinery, transport equipment, electrical in GVCs (see chapter 6). Bulgaria exhibits a low to Box 2.1. Methodology for the Identification of Key GVCs and Peer Countries The desk-based selection of potentially important sectors try. The GVCs include (1) agriculture; (2) food and beverages; for global value chains (GVCs) is based on three quantitative (3) textiles, leather, and footwear; (4) chemicals and nonmetallic analyses—based on the concept of a sector’s revealed compara- mineral products; (5) basic metals and fabricated metal products; tive advantage, assessment of total forward and backward GVC (6) machinery and equipment not elsewhere classified; (7) electri- participation measures, and the composition of the top 50 to 100 cal and optical equipment; and (8) transport equipment. Although export and import products (which usually account for roughly Bulgaria’s involvement in the automotive GVC is still small, the 50 to 70 percent of countries’ total exports and imports). The last country intends to expand its GVC participation in that area. This are identified at the most disaggregated statistical level avail- enables us to contrast Bulgaria with other countries—in par- able and categorized by final use (capital, consumption, and ticular, Romania—that successfully integrated into the transport intermediate goods) and by chain category (final products, main equipment value chain. inputs/parts, standard inputs, raw materials, and machinery Peer countries include Bulgaria’s regional neighbors, which and equipment [see also chapter 3]). Crossing gross trade data are Poland, Portugal, Romania, and Turkey, and the European with informed classifications (see chapter 3) provides additional Union-10 average (whenever possible), as well as key GVC bench- insights. mark countries, which are China, Germany, Japan, and the United The analysis for Bulgaria focuses on eight specific GVCs, States (for the methodology for selecting the peer countries, see which play a significant role (actual or potential) in the coun- box 4.1 in chapter 4). Consider Bulgaria 37 medium level of structural integration in GVCs, Figure 2.2. Decomposition of Domestic Value Added Generated which is slightly worse on the selling side. Table through Exports 2A.1 in annex 2A shows buyer-related and seller- Direct domestic Indirect domestic Reimported domestic related measures of structural integration in GVCs, value added value added value added as measured by the Bonacich power index (Bonacich 1987). They are denoted with the terms BONwin and BONwout, respectively. Bulgaria’s overall extent of structural integration in GVCs as a buyer (BONwin) lies in the medium spectrum and is weakest in trans- port equipment. The country’s overall structural integration in GVCs as a seller (BONwout) lies in the 62.1% 37.8% 0.1% low to medium range across the country sample, and Source: Adapted from the Organisation for Economic Co-operation and Development–World Trade is lowest in agriculture, food and beverages, electri- Organization Trade in Value Added database. cal and optical equipment, and transport equipment. Note: See chapter 4 for more information on the decomposition. The countries with the highest levels of structural integration in GVCs are China, Germany, and the United States. which is lower than for most comparator countries. Minimal spanning trees help in visualizing the Foreign manufacturing firms in Bulgaria source complexity and heterogeneity of actors and trade about 52.6 percent of their inputs locally, which sug- links in GVCs (for the concept of network visu- gests that backward links of foreign direct invest- alizations, see chapter 6). Figure 1.10 in chapter 1 ment (FDI) are in the medium range (figure 2.4). illustrates a minimal spanning tree based on trade Shifting the focus to Bulgarian exporters’ GVC links, in value-added data (as computed by Santoni and the picture looks slightly better. Domestic exporters Taglioni 2015). The main root of the tree is Germany, in Bulgaria export the second highest share of their making it the most relevant source of the “imported” sales (8.9 percent), but they still have a long way to go value added in other countries’ exports, including to reach the share of sales exported by Turkey (16.7 Bulgaria, but also in peers such as Poland, Portugal, percent) (figure 2.5). and Turkey, as well as China. Bulgaria’s structural On the buying side, Bulgaria’s domestic firms are integration in GVCs is less peripheral compared with much better integrated into GVCs compared with Portugal, Romania, and Turkey. From the regional firms in the peer countries. GVC integration on the perspective, value-added trade flows between Germany and Poland are the most relevant. Figure 2.3. Share of Firms Exporting Directly or Indirectly Micro Perspective: Firms’ Integration in GVCs A country’s integration in GVCs is the outcome of Bulgaria firms’ GVC links. Domestic firms can become sellers China in GVCs by supplying to multinational firms in the country or by exporting inputs or final products to Germany international buyers. Domestic firms can also act as Poland buyers by sourcing intermediates from abroad (for a description of GVC measures based on World Bank Portugal Enterprise Surveys, see box 6.2 in chapter 6). Bulgaria’s domestic firms are less well integrated Romania in GVCs on the selling side than are local subsidiaries Turkey of multinationals. Whereas 66 percent of the foreign firms sampled in Bulgaria export at least 1 percent, the 0 20 40 60 80 100 share drops to 18 percent for domestic firms (figure Percent 2.3). The share of domestic firms that export is higher Domestic firms Foreign firms in Poland (23.1 percent) and Romania (21.1 percent), and almost twice as high in Turkey (35.9 percent). In Source: Adapted from Enterprise Surveys (World Bank). Note: The relevant data for each country are as follows: Bulgaria (2013), China Bulgaria, domestically owned manufacturing firms (2012), Germany (2005), Poland (2013), Portugal (2005), Romania (2013), and Turkey tend to source only 65.3 percent of their inputs locally, (2008). 38 Making Global Value Chains Work for Development Figure 2.4. Share of Total Inputs Sourced Locally contrast, only 12.9 percent of domestic manufactur- ing firms import intermediates (figure 2.6). The extent Bulgaria of GVC links on the buying side is also greater for China domestic manufacturing firms in Bulgaria, reaching 34.7 percent, which is the second highest share after Germany Romania (46.6 percent) (100 percent minus the share Poland of total inputs sourced locally, as shown in figure 2.4). Portugal Domestic versus Foreign Content of Gross Exports Some countries, interest groups within countries, Romania and even international organizations (see, for exam- Turkey ple, UNCTAD 2013) still view GVCs as a trap that creates a new core-periphery pattern with “good” 0 20 40 60 80 100 jobs in the North and “bad” jobs in the South. The Percent key piece of evidence that GVC skeptics present Domestic firms Foreign firms to support their view is that the share of domestic value added embodied in exports as a percentage of Source: Adapted from Enterprise Surveys (World Bank). Note: The relevant data for each country are as follows: Bulgaria (2013), China gross exports tends to shrink for emerging countries. (2012), Germany (2005), Poland (2013), Portugal (2005), Romania (2013), and Turkey Indeed, the share of domestic value added embodied (2008). The data include manufacturing firms only. See box 6.2 in chapter 6 for the definition of this measure. in exports as a percentage of gross exports has fallen in Bulgaria and all its peer countries. Bulgaria’s share declined from 62.7 percent in 1995 to 60 percent in Figure 2.5. Share of Total Sales Exported Directly or 2008, which is less than the share in any of the peer Indirectly countries (figure 2.7). Evidence of decreasing domestic value added in Bulgaria total exports is a reflection of the increased sophis- China tication and length of value chains. All countries Germany Figure 2.6. Share of Firms Using Material Poland Inputs/Supplies of Foreign Origin Portugal Bulgaria Romania China Turkey Germany 0 20 40 60 80 100 Poland Percent Portugal Domestic firms Foreign firms Romania Source: Adapted from Enterprise Surveys (World Bank). Note: The relevant data for each country are as follows: Bulgaria (2013), China (2012), Germany (2005), Poland (2013), Portugal (2005), Romania (2013), and Turkey Turkey (2008). See box 6.2 in chapter 6 for the definition of this measure. 0 20 40 60 80 100 Percent buying side provides information on the sources of technology transfer and the types of GVCs a coun- Domestic firms Foreign firms try is likely to join. In Bulgaria, more than 60 percent Source: Adapted from Enterprise Surveys (World Bank). of domestically owned manufacturing firms source Note: The relevant data for each country are as follows: Bulgaria (2013), China (2012), Germany (2005), Poland (2013), Portugal (2005), Romania (2013), and Turkey inputs of foreign origin, which is the second high- (2008). The data include manufacturing firms only. See box 6.2 in chapter 6 for the est share in the sample after Romania. In China, by definition of this measure. Consider Bulgaria 39 Figure 2.7. Domestic and Foreign Value Added Embodied in Gross Exports: Bulgaria and Selected Countries, Total, 1995 and 2008 100 90 80 70 % of gross exports 60 50 40 30 20 10 0 1995 2008 1995 2008 1995 2008 1995 2008 1995 2008 1995 2008 1995 2008 1995 2008 1995 2008 1995 2008 Bulgaria Portugal EU-10 Romania Germany Poland China Turkey United Japan States Domestic value added Foreign value added Source: Adapted from the Organisation for Economic Co-operation and Development–World Trade Organization Trade in Value Added database. increasingly rely on foreign value added (imported Figure 2.8. Coefficients from Regression Results for Value-Added inputs) in GVCs. One way to demonstrate this Components of Gross Exports, Overall Country Sample and Bulgaria increase is by measuring the correlation between the Services All growth rate of gross exports and developments in the Bulgaria direct (intrasector) domestic value added embodied Manufacturing All in gross exports, the indirect (upstream) domes- Bulgaria tic value added embodied in gross exports, and the Primary All foreign value added embodied in gross exports (for Total All definitions of these measures, see chapter 4; for the Bulgaria econometric model, see chapter 7). The econometric results indicate that domestic 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 value added gains substantially from increases in Coefficient gross exports. More specifically, all the value-added Direct domestic value added Foreign content components of gross exports are positively correlated Indirect domestic value added with growth of gross exports. In Bulgaria, growth of Source: Adapted from the Organisation for Economic Co-operation and Development–World Trade the direct domestic value added embodied in gross Organization Trade in Value Added database. Note: Computations are based on econometric results (not reported). exports shows the highest correlation with growth of gross exports (figure 2.8). In the overall sample of countries, growth of gross exports shows the highest in gross exports increases most in the services sec- correlation with growth of the indirect (upstream) tor, in particular for the overall sample of countries domestic value added embodied in gross exports. For (figure 2.8). Figure 2.9 reports the coefficients for the overall sample and for Bulgaria, growth of for- individual industries in the overall country sample eign value added embodied in gross exports shows (the number of observations was too low to estimate the weakest correlation with growth of gross exports. the model for Bulgaria alone). The main observa- The results by sector indicate that the growth of tion is that overall gains in the domestic content indirect (upstream) domestic value added embodied of gross exports are always greater than those in 40 Making Global Value Chains Work for Development Figure 2.9. Coefficients from Regression Results for Value-Added Components of Gross Exports, by Sector, Overall Country Sample Agriculture Food products Textiles & apparel Chemicals Basic metals Machinery Electrical equipment Transport equipment 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 Coefficient Direct domestic value added Indirect domestic value added Foreign value added Source: Adapted from the Organisation for Economic Co-operation and Development–World Trade Organization Trade in Value Added database. Note: Computations are based on econometric results (not reported). foreign content. At the same time, the results for spe- Final Demand for Value Added in Gross Exports cific industries point to substantial heterogeneity in Despite Bulgaria’s EU integration, the final demand composition. for Bulgarian value added embodied in gross exports (for the measure, see chapter 5) from the EU-27 has Imports-to-Exports Patterns increased only slightly and remains significantly The basic concept of GVC trade is “importing to lower than the EU average. Between 1995 and 2008, export,” or I2E, as Baldwin and Lopez-Gonzalez Bulgaria’s share increased from 52.5 to 54.7 percent (2013) call it. Essentially, the measure of GVC trade (figure 2.11). Bulgaria’s share in 2008 was thus sub- captures how much of a country’s exports consist of stantially less than the EU-10 average of 69.1 percent value that was added in another country. The mea- and the shares for other EU peers, such as Poland sure includes the sourcing side in GVCs (that is, how (71.3 percent) and Romania (64.7 percent). For much a country is buying) and the selling side in GVCs Bulgaria, final demand from the EU-27 market grew (how much a country is selling) (for more details, only in transport equipment and, to a lesser extent, see chapter 3). Bulgaria buys its I2E mainly from food and beverages and machinery and equipment. Germany, Italy, the Russian Federation, and Turkey The final demand for Bulgarian value added embod- (figure 2.10). Between 1995 and 2009, Bulgaria’s reli- ied in its gross exports from other countries increased ance on sourcing I2E from Russia dropped sharply from 27.8 to 32.5 percent over the same period, and, to a lesser extent, also from Germany, whereas whereas final demand from East Asia declined from the I2E sourcing share from Romania and Turkey 10.9 to 5.3 percent. showed a relatively strong expansion over the period. On the selling side, Bulgaria sold its I2E products International GVC Links and Growth in Bulgaria’s mainly to Belgium, Germany, Italy, Romania, and Domestic Value Added That Is Exported Turkey. The country’s reliance on those I2E destina- Does the intensity and nature of GVC links matter tions—especially Belgium and Germany—increased for growth in domestic value added that is exported? between 1995 and 2009, with the exception of Italy. This study explores this question through economet- Other EU countries, such as France, the Netherlands, ric and statistical analysis from several angles. First, and the United Kingdom, also lost in relative market the analysis explores whether the degree of struc- share as destinations of I2E products over the period. tural integration in global value-added trade matters. Consider Bulgaria 41 Figure 2.10. Bulgaria’s Buying and Selling Patterns, 1995 and 2009 1 Russian Federation Belgium 1 2 Germany Germany 2 3 Turkey Turkey 3 4 Italy Italy 4 5 Romania Romania 5 6 Greece Greece 6 7 China Russian Federation 7 8 France Netherlands 8 9 Austria Austria 9 10 Poland France 10 Import partners ranked from highest to lowest percentage, 2009 Export partners ranked from highest to lowest percentage, 2009 11 Netherlands Spain 11 12 Czech Republic Ireland 12 13 United Kingdom Poland 13 14 United States Canada 14 15 Hungary United Kingdom 15 16 Spain Hungary 16 17 Belgium Czech Republic 17 18 Slovak Republic Denmark 18 19 Slovenia Sweden 19 20 Brazil Slovenia 20 21 Denmark Slovak Republic 21 22 Sweden China 22 23 Canada United States 23 24 Korea, Rep. Korea, Rep. 24 25 Japan Portugal 25 26 India Finland 26 27 Ireland Cyprus 27 28 Finland Taiwan, China 28 29 Taiwan, China Lithuania 29 30 Portugal Luxembourg 30 31 Malta Mexico 31 32 Cyprus India 32 33 Luxembourg Latvia 33 34 Indonesia Indonesia 34 35 Australia Brazil 35 36 Lithuania Japan 36 37 Mexico Estonia 37 38 Latvia Malta 38 39 Estonia Australia 39 16 14 12 10 8 6 4 2 0 –2 –2 0 2 4 6 8 10 12 14 16 Import partners’ growth, 1995–2009 (%) Export partners’ growth, 1995–2009 (%) Import share, 2009 Export share, 2009 Trade share, 2009 (%) 13 10 8 6 4 2 1 0.0002 Source: Adapted from Baldwin and Lopez-Gonzalez 2013. Second, the analysis asks whether the greater inte- domestic value-added component. The underlying gration of Bulgaria in GVCs as a buyer (relative to econometric and statistical frameworks can be found the weaker integration as a seller) negatively affects in chapter 7. Bulgaria’s domestic value-added growth from gross Bulgaria’s increasing GVC integration is cor- exports. Third, the analysis looks more closely at the related with value-added gains. We find a positive relation between the growth of foreign value added correlation between growth of GVC participation embodied in gross exports and the growth of the and growth of domestic value added embodied in 42 Making Global Value Chains Work for Development Figure 2.11. Final Demand, by Destination, Bulgaria and Peer rather tends to be associated with a gain in domes- Countries, Total, 1995 and 2008 tic value-added exports. The correlation is highest for the primary sectors, followed by manufacturing Bulgaria 1995 2008 and services. The premium, which is highest in tex- China 1995 tiles and apparel, agriculture, and machinery, may be 2008 driven by the beneficial effects of importing foreign EU-10 1995 know-how and technology. 2008 Finally, higher growth of foreign value added in Germany 1995 gross exports is associated with higher growth of 2008 domestic value added embodied in gross exports. In Japan 1995 other words, higher rates of imported inputs enable 2008 Bulgaria and all the other countries in the sample to Poland 1995 increase their production, productivity, and value- 2008 added content in exports. The elasticity is slightly Portugal 1995 lower for Bulgaria compared with the overall sample. 2008 Romania 1995 The results make a clear case for integrating in GVCs 2008 and encouraging imports by demonstrating that Turkey 1995 more imports are linked to more value addition from 2008 exports. United States 1995 2008 Scope for Further Value-Added Growth 0 15 30 45 60 75 Low Domestic Value Added in Percent Third Countries’ Exports ASEAN-8 East Asia EU-27 NAFTA Other There remains large scope for further value-added growth. Bulgaria’s sales of inputs to third countries— Source: Adapted from the Organisation for Economic Co-operation and Development–World Trade which are used in those countries’ exports—tend to Organization Trade in Value Added database. Note: Destinations are only available at the aggregated level. ASEAN = Association of Southeast be of low value added (for the measure, see chapter Asian Nations; EU = European Union; NAFTA = North American Free Trade Agreement. 5). Although all countries in the sample increased their domestic value added in third-country exports as a percentage of gross exports, Bulgaria’s expan- gross exports. Although that correlation suffers from sion is among the lowest (figure 2.12). Its share is problems of endogeneity, the correlation is higher among the lowest in the sample for chemical and in Bulgaria compared with the full country sample nonmetallic mineral products, metals, and machin- in the Trade in Value Added database developed by ery and equipment; however, the country managed a the Organisation for Economic Co-operation and strong increase in its share for electrical and optical Development and the World Trade Organization (for equipment. more information on this database refer to Appendix G). The value-added premium for Bulgaria’s integra- Large Distance to Final Demand and Lower tion in GVCs is even higher on the selling side than Opportunities for Increasing Domestic Value on the buying side. The highest positive correlation Added along the Chain stems from the manufacturing sector, whereas it is A final useful metric is to combine import and export weakest for the services sector. A breakdown by sec- “upstreamness”—the distance of a country’s import tors indicates that the correlation is highest for tex- and export baskets to the final consumer—to com- tiles and apparel and metals, sectors in which the pute the gap between the buying and selling chains, premium for Bulgaria over the full country sample which is defined as the difference between import is also highest. These results may be associated with and export upstreamness (for more on upstreamness Bulgaria’s comparative advantage in those sectors. and the domestic gap between the buying and selling Interestingly, the unbalanced extent of integration chains, see chapter 5). A positive gap indicates that in GVCs on the buying side (higher) versus the sell- exports are relatively more downstream compared ing side (lower) does not seem to harm Bulgaria, but with the import mix, or that exports are closer to Consider Bulgaria 43 Figure 2.12. Domestic Value Added in Third Countries’ Gross Exports, Bulgaria and Peer Countries, Total, 1995 and 2008 100 90 80 70 % of gross exports 60 50 40 30 20 10 0 1995 2008 1995 2008 1995 2008 1995 2008 1995 2008 1995 2008 1995 2008 1995 2008 1995 2008 Japan United Germany Romania Poland Portugal Turkey Bulgaria China States Source: Adapted from the Organisation for Economic Co-operation and Development–World Trade Organization Trade in Value Added database. final demand than are imports. Conversely, a nega- implies a larger average number of production steps tive gap indicates that a country’s export profile is performed in the country. Bulgaria’s small gap could more upstream than its import profile. Alternatively, be an indicator of lower opportunities to expand a negative gap may indicate that the country has a domestic value added along the chain. sophisticated consumer market and therefore is an In Bulgaria, textiles and transport equipment and, intensive importer of finished consumer goods, such to a lesser extent, food and beverages and machinery as the United States. and electronics are closest to final demand. Although Bulgaria’s exports and imports are in a relatively food and beverages moved slightly upstream, upstream position, far from the final consumer, and machinery and electronics moved even closer to final have moved up further over time. Bulgaria’s import demand, as did chemicals and transport equipment. upstreamness in 2012 was the third largest after By contrast, metals and agriculture—the two sectors China and Japan, whereas its export upstreamness that are the most upstream—moved further from was the largest of the sample. As for most countries, final demand. both measures increased over 2000–12, reflecting Does the position in the value chain (upstream- that GVCs have become longer with the increased ness) and the length of the domestic segment fragmentation of production. Only Germany and of GVCs matter for the growth rate of domestic Poland managed to move their exports closer to the value added that is exported? This question can be final consumer (figure 2.13). addressed econometrically by examining the growth Bulgaria’s gap is relatively short, which implies rate of the upstreamness and the domestic length of that the average number of production steps and, sourcing chains—that is, the distance between a sec- thus, opportunities to increase domestic value added tor of interest and the first supplier in the value chain along the chain are fewer than elsewhere. This situ- (for an econometric model, see chapter 6). ation will not change unless the country manages Going upstream has no impact on the growth to enter tasks that are not carried out domestically. of domestic value added embodied in manufactur- Bulgaria showed the second smallest gap after the ing exports for the overall sample of countries. The United States, which resulted from similarly high correlation between growth of upstreamness and import and export upstreamness. The gap is slightly growth in the domestic value-added content of positive, which suggests that exports are relatively exports is statistically insignificant in manufacturing, closer to final demand than are imports. A larger gap although a statistically significant correlation exists 44 Making Global Value Chains Work for Development Figure 2.13. Import and Export Upstreamness and Gap, Bulgaria and Peer Countries, 2000 and 2012 Bulgaria 2000 2012 China 2000 2012 EU-11 2000 2012 Germany 2000 2012 Japan 2000 2012 Poland 2000 2012 Portugal 2000 2012 Romania 2000 2012 Turkey 2000 2012 United States 2000 2012 1.75 2.00 2.25 2.50 2.75 3.00 Upstreamness Export upstreamness Import upstreamness Positive domestic gap Negative domestic gap Sources: Adapted from Chor 2014; United Nations Comtrade database. for agriculture and services. By contrast, for Bulgaria, odds with the results for the overall country sam- the relationship is mostly statistically significant, ple—that is, significantly negative. In other words, even in manufacturing industries. A breakdown by being closer to final demand in the services sector in industry suggests that Bulgaria’s moving upstream in Bulgaria pays off. the value chain is associated with greater growth of Finally, expanding the length of domestic sourc- domestic value added that is exported in agriculture, ing chains is associated with greater domestic value- food and beverages, and textiles. added content in Bulgarian gross exports of food and By contrast, in chemicals, metals, and electron- beverages, textiles and apparel, metals, and transport. ics, moving downstream creates a positive associa- By contrast, the association is negative for services, tion. The results for the full country sample indicate agriculture, and chemicals, which suggests that in instead that the relationship is statistically insignifi- these sectors, specialization in core activities and cant for all the individual manufacturing sectors. In higher import intensity may be associated with bet- agriculture, however, we find that moving down- ter quality, higher value-added output. stream is associated with growth in the domestic value added embodied in gross exports, a result that Summary is at odds with the case of Bulgaria, where moving upstream generates greater value addition. Finally, The analysis so far suggests that a key priority for in the results for the full country sample, upstream- Bulgaria, after having successfully integrated into ness in the services sector is associated with growth regional and GVCs, is to target economic upgrading of the value added embodied in exports. The results and densification. The econometric results suggest a for the services sector may be driven by the fact positive correlation between Bulgaria’s integration that research and development (R&D) and design in GVCs and its ability to boost the domestic value- activities are located at the beginning of many value added content of exports. However, the indicators chains. Interestingly, however, the correlation coef- for Bulgaria’s position in GVCs point to a still low to ficient for the services sector in Bulgaria is again at medium level of integration and positioning in the Consider Bulgaria 45 low value-added segment. The results suggest ample the variables and the regression results). A lower scope for upgrading. technology gap of a firm (relative to the median pro- ductivity level of foreign firms in the same sector) positively mediates productivity gains from GVC Bulgaria’s GVC Participation and participation on the buying and selling sides in the Firm-Level Productivity full country sample, and the positive effect is even What is the potential of GVCs to enhance the produc- larger for Bulgaria. In other words, firms that lag fur- tivity of firms located in Bulgaria (domestic and for- ther behind foreign firms in their median productiv- eign)? This question may be addressed by modifying ity also benefit less from GVC integration. the analysis by Farole and Winkler (2014) and apply- Exporters and foreign firms in Bulgaria gain less ing it to Bulgaria. The model assesses how a firm’s from structural integration in GVCs than do the absorptive capacity and a country’s institutional same types of firms in the overall sample of 22 coun- variables influence firm productivity from structural tries. Other factors that positively mediate the impact integration in GVCs in manufacturing industries of structural integration in GVCs from a buyer’s and (see annex 7B in chapter 7; for a literature review of seller’s perspective are the firm’s technology level, the mediating factors, see chapter 9). The data con- size, export share, and FDI status; these findings hold sist of a cross-section of domestic manufacturing for the average firm in the full country sample. In firms in low- and middle-income countries (LMICs) Bulgaria, the positive effects from export share and from the World Bank’s Enterprise Surveys. Because FDI status are smaller. These results may be related the effect of GVC participation is the topic of inter- to infrastructural challenges or barriers at the border, est, we merged the data set with two sector measures which do not support a firm’s openness to trade and of structural integration in GVCs: BONwin (buyer’s FDI as much in Bulgaria as in other countries. perspective) and BONwout (seller’s perspective), as Although the firm’s location in an agglomera- computed by Santoni and Taglioni (2015). The data tion reduces the productivity gains from the sector’s include measures for more than 14,000 manufactur- structural integration in GVCs as a buyer and as a ing firms in 22 LMICs (see table I.2 in appendix I). seller in the full country sample, the effect in Bulgaria A sector’s structural integration in GVCs—as is more positive for buyers and more negative for buyer and seller—has a positive impact on firm pro- sellers. In other words, agglomerations entail positive ductivity. The estimations for the full country sample urbanization economies, large consumer markets, clearly confirm that GVC participation increases the deep labor pools, links to international markets, and productivity of firms in a country—be they domes- clusters of diverse suppliers and institutions—when tic or foreign firms (see tables 7C.1 to 7C.3 in annex firms rely on imported inputs in GVCs, lowering 7C in chapter 7). The various transmission chan- production costs and increasing firm productivity. nels through which GVC participation can increase These benefits outweigh the potential negative con- productivity at home are depicted in figure 1.11 in gestion costs that occur because of increased demand chapter 1. The positive impact in the full sample is for resources in agglomerations (such as power out- stronger for the seller-side measure than for the ages and waiting times). By contrast, firms that are buyer-side measure. In other words, GVC integration selling in GVCs may face higher negative congestion as a buyer (via importing intermediates) in the short costs, which seem to be higher than the potential term leads to higher productivity gains than does benefits of agglomeration. GVC participation as a seller (via exporting). Why are productivity spillovers for sellers in The following subsections discuss the role of Bulgarian agglomerations lower in the short run? absorptive capacity and national characteristics for One reason could lie in the short-term nature of the GVC integration-productivity nexus. the regressions: buyers in Bulgaria may benefit more quickly from GVC integration (for example, through the availability of high-tech inputs)—espe- Absorptive Capacity cially when located in agglomerations—whereas the Several characteristics at the firm level—in particu- productivity gains for sellers may take more time lar, a firm’s lower technology gap—can increase the to materialize. In the short run, sellers in Bulgaria productivity spillovers from a sector’s structural may face negative competition and rivalry effects— integration in GVCs (see annex 7B and tables 7C.1 which could be even greater in agglomerations— to 7C.3 in annex 7C in chapter 7 for definitions of which hamper productivity. The weaker GVC links 46 Making Global Value Chains Work for Development for sellers in GVCs might magnify those potential host country residents significantly affect spillovers mechanisms. In the long run, it could be that mul- in GVCs. This relationship takes a U-shaped form, tiple positive effects offset the negative competition that is, the extent of spillovers increases only below and rivalry effects for sellers. or above certain threshold levels of human capital (Meyer and Sinani 2009). It is possible that the coun- National Characteristics and Institutions tries in the sample (which covers only LMICs) are in the medium level of the U-curve for skills and the The productivity spillovers from structural integra- low level of the U-curve for R&D intensity. tion in GVCs are lower in more open trade regimes National and institutional characteristics in and more developed countries in the full country Bulgaria positively mediate the productivity spill- sample (see annex 7B and tables 7C.4, and 7C.5 in overs from structural integration in GVCs as a annex 7C in chapter 7 for definitions of the vari- buyer—that is, on the sourcing side. The results show ables and the regression results). A country’s share a positive association between integration in GVCs of exports (as a percentage of gross domestic prod- and measures of financial freedom (such as bank- uct [GDP]), less trade protectionism, and a higher ing efficiency and independence from government GDP reduce the positive productivity spillovers on control and interference in the financial sector). The firms from GVC participation, as a buyer and as a same holds true for more government spending on seller. GVC participants in an outward oriented trade education (as a percentage of GDP), higher share setting may focus more strongly on international of people who have completed secondary and ter- distribution and marketing, whereas in an inward tiary education, higher R&D intensity, higher share oriented policy regime, they might bring newer tech- of exports (as a percentage of GDP), less trade pro- nologies to the host countries (Crespo and Fontoura tectionism, and higher per capita income, which all 2007). Local firms in an open trade regime may also show a positive and significant mediating relation- be more exposed to competitive pressures through ship with firm productivity in Bulgaria. Although international trade competition, which could lead to many of these characteristics reduce productivity negative spillovers in the short term. spillovers in the full country sample, in Bulgaria they The more advanced a country is in income level, can help increase the productivity gains from GVC the lower are the productivity spillovers from GVC integration as a buyer. participation. If income is accepted as a broad mea- By contrast, many national and institutional sure of national competition (and other institutional characteristics in Bulgaria are negatively associ- factors), the findings suggest that more developed ated with the productivity spillovers from struc- countries with higher levels of competition benefit tural integration in GVCs as a seller. Less restricted less from GVC integration. In such contexts, firms labor or financial markets, more government spend- may have lower incentives to improve (Barrios and ing on education, higher share of people who have Strobl 2002). completed secondary and tertiary education, higher Productivity spillovers from structural integra- share of R&D intensity, more freedom to invest, tion in GVCs are lower in countries with higher higher share of exports in GDP, less trade protec- education, whereas they are higher in countries with tionism, and higher GDP have a negative and sta- high innovation capacity, according to the full coun- tistically significant correlation with firm-level try sample. Government spending on education (as productivity in Bulgaria. Compared with the full a percentage of GDP) and the share of people who country sample, the negative mediating effects of have completed secondary and tertiary education national characteristics are weaker for Bulgaria from reduce the positive productivity spillovers on firms a buyer’s perspective but stronger from a seller’s per- from GVC participation, as a buyer and as a seller. By spective. This was also the case for firm location in contrast, higher R&D intensity shows a positive and agglomerations. significant impact on productivity spillovers for both What explains the negative mediating relation- types of GVC integration. Meyer and Sinani (2009), ship with GVC participation as a seller but the posi- for instance, include three measures of a country’s tive mediating relationship with GVC integration as availability of human capital and show evidence a buyer in Bulgaria? Buyers in Bulgaria might ben- that the share of workers with tertiary education, the efit more quickly from GVC integration (for exam- R&D intensity in the private sector, and the num- ple, through the availability of high-tech inputs), ber of patents (per billion U.S. dollars) granted to which is further supported by a business-friendly Consider Bulgaria 47 environment—in particular, trade openness, skilled strengthening Bulgaria’s GVC integration—particu- workforce, and R&D intensity. Sellers, by contrast, larly on the selling side—by fostering better links may face negative competition and rivalry effects in with domestic producers and enhancing absorptive the short run, and a business-friendly environment capacities. The GVC analysis described in this chap- may lead only to greater competition and, thus, more ter, combined with selected performance indicators negative effects. The weaker GVC links for sellers in of Bulgaria’s regulatory and institutional framework GVCs might magnify the effect. In the long run, how- related to GVCs (see table O.1 in the Overview), ever, multiple positive effects can be expected to offset reveals that to integrate further into GVCs, Bulgaria the negative competition and rivalry effects for sellers. must improve (1) connectivity to international mar- kets as well as to the domestic segment of the sup- ply chain and (2) the quality of infrastructure. The What Must Be Done? country also must focus on strengthening its links, Applying the strategic policy framework to Bulgaria maximizing the absorption potential of local actors reveals two priority areas for policy interven- to benefit from GVC spillovers and fostering its inno- tion, as discussed in World Bank (2015): (1) facili- vation capacity to expand and strengthen its GVC tating domestic firms’ entry into GVCs and (2) participation (figure 2.14). Figure 2.14. Strategic Policy Framework Applied to Bulgaria Focus area Objectives Strategic questions Policy options Which tasks? Creating world-class GVC links – Which form of GVC participation? – Jump-starting GVC entry through EPZs – How can tasks be identified? and other competitive spaces Attracting foreign – Which risks? – Attracting the "right" foreign investors investors and Which form of governance? – Helping domestic firms find the Entering GVCs “right” trade partner and technology facilitating domestic – Which form of governance between firms’ entry into GVCs lead firms and suppliers? abroad – Buyer- or producer-driven value – Improving connectivity to chains? international markets – Which power relations in GVCs? Creating a world-class climate for foreign tangible and intangible assets – Ensuring cost competitiveness – Improving drivers of investment and protecting foreign assets – Improving domestic value chains and quality of infrastructure and services Which transmission channels? Strengthening GVC-local economy Which type of economic upgrading? links on the buyer’s and seller’s Promoting economic sides upgrading and Which type of densification? densification Which foreign firm and country Expanding and characteristics influence Strengthening absorptive capacity strengthening spillovers? – Maximizing the absorption potential GVC participation of local actors to benefit from GVC Strengthening Which domestic firm characteristics spillovers domestic firms’ help internalize spillovers? – Fostering innovation and building absorptive capacity capacity – Complying with process and product standards – Bundling tasks Which relationship between economic and social upgrading? Creating a world-class workforce Promoting social Which type of social upgrading? – Developing skills upgrading and Is downgrading a possibility? – Promoting social upgrading Turning GVC cohesion – Engineering equitable distribution of participation into Which links between social upgrading and cohesion? opportunities and outcomes sustainable development Promoting environmental What benefits from environmental Implementing climate-smart policies sustainability regulation? and infrastructure Note: The orange lines highlight priority areas for policy intervention in Bulgaria. EPZs = export processing zones; GVCs = global value chains. 48 Making Global Value Chains Work for Development Facilitate Domestic Firms’ Entry into GVCs Figure 2.15. Doing Business Indicator: Overall and Protecting Investors, 2014 Macro- and micro-level indicators reveal that there is scope for Bulgarian firms to enter GVCs, espe- 1 United 4 cially on the selling side. The analysis points out that States 6 10 Bulgaria exhibits a low to medium level of struc- 16 tural integration in GVCs—in particular in agri- 20 Germany 21 culture, food and beverages, electrical and optical Japan 27 equipment, and transport equipment, the last being 30 Portugal 31 34 slightly worse on the selling side. At the micro level, Bulgaria’s domestic firms are less well integrated 40 in GVCs on the selling side. More scope exists for Poland 45 Rank domestic firms to become exporters. In the follow- 50 52 ing, we analyze some policy options that can help Bulgaria 58 achieve this goal. 60 Countries can join GVCs either by facilitating Turkey 69 70 domestic firms’ entry or by attracting foreign inves- Romania 73 tors. Bulgaria’s relatively favorable business environ- 80 ment, in conjunction with the country’s accession to the EU in 2007, led to large inflows of FDI, especially 90 during the years around EU membership. Between China 96 98 2005 and 2009, average net foreign inflows as a per- 98 centage of GDP reached 20 percent, compared with Overall ease of Protecting doing business investors 6.4 percent in Romania and 4.4 percent in China and Poland. Source: Adapted from World Bank Logistics Performance Index and Doing Business data. Bulgaria’s business climate exhibits solid perfor- mance relative to its peers, also with regard to inves- tor protection. The country’s overall ease of doing arranging shipments, quality of logistics services, business, which captures a country’s regulatory envi- ability to track and trace consignments, and delivery ronment when starting and operating a local firm, times (figure 2.16). Bulgaria also lags behind its peers ranges in the medium spectrum compared with its in customs efficiency (64th), whereas its peers and peers and global GVC players (figure 2.15). Business global GVC players, including China, perform much is more difficult to do in China, Romania, and Turkey better in both regards. than in Bulgaria, but much easier in Germany, Japan, Although Bulgaria’s quality of services performs and the United States. In the area of protecting inves- relatively well in most aspects, the quality of its tors, Bulgaria is at the same level as Poland, Portugal, infrastructure lags behind that of most of its peers. and Romania, but is ranked better than China and, According to the domestic Logistics Performance surprisingly, Germany. Index,3 Bulgaria is trailing its peers in the quality of However, successful GVC integration requires trade and transport-related infrastructure, especially Bulgaria to improve its connectivity to international ports and airports. The country also lacks quality in markets. Poor connectivity means high costs, low roads (ranking as the second worst performer after speed, and high uncertainty. Successful participa- Poland) and rail (alongside Poland, Turkey, and the tion in GVCs thus requires policy makers not just United States) (see table 2A.2 in annex 2A). Bulgaria’s to address barriers at the border, but also to increase competence and quality of trade and logistics-related the connectivity of domestic markets and enhance services, by contrast, are relatively high in most the resilience and efficiency of the domestic segment aspects (especially freight forwarders, warehous- of the supply chain. Bulgaria ranks lowest (47th) ing and distribution services, customs agencies and among its peer countries in the overall international brokers, and road-haulers and consignees/shippers). Logistics Performance Index, a measure that takes By contrast, health/sanitary and phytosanitary and into account a country’s customs efficiency, qual- quality/standards inspection agencies perform rela- ity of trade and transport infrastructure, ease of tively poorly. Consider Bulgaria 49 Figure 2.16. Logistics Performance Index: Overall and GVCs and the local economy target foreign investors Customs Efficiency, 2014 primarily, but those policies can also include other 1 international buyers outside the country. Several pol- 1 2 Germany icy options are available (see chapter 9). 9 10 10 Strengthen Bulgaria’s Absorptive Capacity 14 Japan 16 United GVCs can lead to benefits, but at the country level con- States 20 straints remain, such as inadequate capacity to absorb such potential gains. Attracting foreign investors and 26 28 other international buyers and linking them to the 30 30 domestic economy should create optimal conditions 31 31 Portugal Rank 32 Poland for local firms and workers to benefit from spillovers 34 Turkey 38 China of knowledge and technology. The degree to which 40 40 they ultimately benefit, however, depends crucially on the absorptive capacity of domestic actors. That is 47 the area of spillover policy in which government has 50 the most important role to play, in particular by the absorptive capacity of firms and workers and helping 59 Romania local firms and workers access opportunities. 60 64 Bulgaria The analysis indicates that innovation capacity— 64 at the firm and country levels—matters for positive Overall LPI Customs LPI productivity spillovers from GVC integration. The Source: Adapted from World Bank Logistics Performance Index and Doing econometric analysis shows that a higher level of Business data. technology positively mediates spillovers in GVCs. In particular, a smaller technology gap relative to for- eign firms is beneficial for productivity spillovers in Strengthen GVC–Local Economy Links on the GVCs, and that effect is even stronger for Bulgaria Selling Side compared with the average firm in the full country sample. Similarly, higher R&D intensity at the coun- Firm-level analysis indicates there is scope to try level increases the productivity gains from GVC strengthen GVC-local economy links, especially on integration in the full country sample. In this area, the selling side. Although the share of FDI inflows the gains are even higher in Bulgaria for buyers in as a percentage of GDP is very high in Bulgaria, the GVCs, but lower for sellers. challenge lies in linking FDI with the local economy. Although the share of skilled workers in Bulgaria The analysis reveals that the share of local inputs is high, the country’s innovation capacity has room to sourced by multinational manufacturing firms expand further. More than one-quarter of Bulgaria’s in Bulgaria lies only in the middle spectrum; and workforce has tertiary education, putting the country although Bulgarian exporters’ GVC links are stron- at a level similar to Germany and Poland. The pro- ger, scope for growth remains. portion of educated workers in Portugal, Romania, Weaker links on the selling side in Bulgaria may and Turkey is less than 20 percent (figure 2.17). Public have magnified the negative mediating effect on pro- and private expenditures for basic research, applied ductivity spillovers for sellers in GVCs—in particu- research, and experimental development, however, lar, in the short run. Strong links with the domestic are the second lowest after Romania (0.64 percent of economy should offer greater benefits of GVC par- GDP)—at only one-quarter the levels of Germany, ticipation at home via several transmission channels Japan, and the United States and only one-third the (figure 1.11 in chapter 1). Linkage development can level of China (figure 2.17). Reduction of the skills focus on the breadth of links (variety of local inputs) gap will require the active engagement of universities and their depth (degree of local value added), so and research institutes. making that distinction is key (Morris, Kaplinsky, Building absorptive capacity goes beyond increas- and Kaplan 2011). Policies promoting links between ing R&D intensity. Measures to build the absorp- 50 Making Global Value Chains Work for Development Figure 2.17. Innovation Capacity and Skills, 2012 References % of GDP Baldwin, Richard, and Javier Lopez-Gonzalez. 2013. 0.5 1.0 1.5 2.0 2.5 3.0 3.5 “Supply-Chain Trade: A Portrait of Global Patterns and Several Testable Hypotheses.” NBER Working Bulgaria Paper 18957, National Bureau of Economic Research, Cambridge, MA. China Barrios, Salvador, and Eric Strobl. 2002. “Foreign Direct Investment and Productivity Spillovers: Evidence from Germany the Spanish Experience.” Review of World Economics 138 (3): 459–81. Japan Bonacich, Phillip. 1987. “Power and Centrality: A Family Poland of Measures.” American Journal of Sociology 92 (5): 1170–82. Portugal Chor, Davine. 2014. “Where Are Countries Positioned along Global Production Lines?” Macroeconomic Romania Review: 94–99. Monetary Authority of Singapore. Crespo, N., and M. P. Fontoura. 2007. “Determinant Turkey Factors of FDI Spillovers—What Do We Really Know?” World Development 35 (3): 410–25. United States Farole, Thomas, and Deborah Winkler. 2014. “The Role 0 5 10 15 20 25 30 of Mediating Factors for FDI Spillovers in Developing % of total labor force Countries: Evidence from a Global Dataset.” In Making R&D expenditure (% of GDP) Foreign Direct Investment Work for Sub-Saharan Africa: Local Spillovers and Competitiveness in Global Value Labor force with tertiary education (% of total labor force) Chains, edited by Thomas Farole and Deborah Winkler, Source: Adapted from World Bank 2012. 59–86. Washington, DC: World Bank. Note: No labor force data are available for China, Japan, or the United States. Meyer, Klaus E., and Evis Sinani. 2009. “When and Where Does Foreign Direct Investment Generate Positive Spillovers? A Meta-Analysis.” Journal of International tive capacity of local firms require general and Business Studies 40 (7): 1075–94. industry-specific investments to upgrade technical Morris, M., R. Kaplinsky, and D. Kaplan. 2011. capacity and, most important, achieve quality stan- “Commodities and Linkages: Meeting the Policy dards. Because licensing of technology from foreign Challenge.” MMCP Discussion Paper 14, Making investors and other international buyers is a sig- the Most of Commodities Programme, The Open nificant source of technical spillovers, governments University, U.K. Santoni, Gianluca, and Daria Taglioni. 2015. “Networks should provide incentives for that licensing. and Structural Integration in Global Value Chains.” In The Age of Global Value Chains, edited by Joao Amador and Filippo di Mauro. Washington, DC: Center for Notes Economic and Policy Research and VOXEU. 1. A significant part of this chapter is based on a tech- UNCTAD (United Nations Conference on Trade and nical background paper written by the authors for World Development). 2013. “World Investment Report Bank (2015). It shows how to analyze a country’s participa- 2013—Global Value Chains: Investment and Trade for tion in GVCs, using Bulgaria as an example. The chapter Development.” UNCTAD, Geneva. uses only publicly available data. World Bank. 2012. World Development Indicators . 2. EU-10 comprises Bulgaria, the Czech Republic, Washington, DC: World Bank. Estonia, Hungary, Lithuania, Latvia, Poland, Romania, ———. 2015. Productivity Growth in Bulgaria: Trends and Slovak Republic, and Slovenia. Options. Washington, DC: World Bank. 3. http://data.worldbank.org/data-catalog/logistics -performance-index. Consider Bulgaria 51 Annex 2A Table 2A.1 Bulgaria’s Position in the Global Network of Trade in Value Added, 2008 Chemicals and Basic metals Machinery Electrical Sector Food and Textiles, apparel, non-metal and fabricated and equipment, and optical Transport Total Agriculture Beverages and footwear minerals metals nec equipment equipment out out out out out out out out out in in in in in in in in in Nw Nw Nw Nw Nw Nw Nw Nw Nw Nw Nw Nw Nw Nw Nw Nw Nw Nw CCw CCw CCw CCw CCw CCw CCw CCw CCw Country BO BO BO BO BO BO BO BO BO BO BO BO BO BO BO BO BO BO Argentina 0.63 0.12 0.13 0.52 0.14 0.13 0.58 0.14 0.14 0.52 0.13 0.13 0.54 0.12 0.12 0.51 0.11 0.12 0.48 0.10 0.11 0.47 0.10 0.11 0.51 0.13 0.12 Australia 0.68 0.13 0.14 0.54 0.14 0.14 0.59 0.14 0.14 0.50 0.11 0.14 0.58 0.12 0.14 0.60 0.15 0.14 0.56 0.13 0.14 0.52 0.11 0.14 0.53 0.12 0.14 Austria 0.68 0.14 0.14 0.50 0.12 0.14 0.59 0.15 0.14 0.55 0.14 0.14 0.60 0.14 0.14 0.60 0.15 0.14 0.59 0.16 0.14 0.57 0.14 0.14 0.58 0.16 0.14 Belgium 0.70 0.15 0.14 0.55 0.15 0.14 0.61 0.16 0.14 0.57 0.15 0.14 0.64 0.16 0.14 0.61 0.16 0.14 0.59 0.15 0.14 0.56 0.14 0.14 0.58 0.16 0.14 Brazil 0.67 0.13 0.14 0.56 0.15 0.15 0.61 0.14 0.15 0.53 0.12 0.14 0.59 0.13 0.14 0.58 0.14 0.14 0.55 0.13 0.14 0.53 0.12 0.13 0.55 0.14 0.14 Brunei Darussalam 0.54 0.09 0.10 0.36 0.07 0.09 0.36 0.05 0.09 0.42 0.11 0.09 0.37 0.04 0.10 0.35 0.02 0.10 0.43 0.00 0.09 0.34 0.04 0.10 0.34 0.02 0.09 Bulgaria 0.60 0.12 0.11 0.47 0.14 0.10 0.50 0.11 0.10 0.47 0.12 0.11 0.53 0.12 0.11 0.52 0.13 0.11 0.50 0.12 0.11 0.48 0.11 0.10 0.45 0.10 0.10 Cambodia 0.50 0.10 0.08 0.37 0.12 0.05 0.39 0.10 0.06 0.43 0.12 0.08 0.37 0.07 0.07 0.36 0.07 0.06 0.34 0.08 0.06 0.35 0.08 0.06 0.33 0.06 0.06 Canada 0.69 0.14 0.14 0.57 0.16 0.14 0.59 0.14 0.14 0.53 0.13 0.14 0.61 0.14 0.14 0.60 0.15 0.14 0.58 0.14 0.14 0.56 0.14 0.14 0.58 0.16 0.14 Chile 0.64 0.12 0.13 0.52 0.15 0.12 0.56 0.14 0.12 0.46 0.10 0.12 0.54 0.12 0.13 0.51 0.10 0.14 0.50 0.10 0.13 0.39 0.03 0.13 0.48 0.10 0.13 China 0.75 0.16 0.15 0.56 0.14 0.16 0.64 0.16 0.16 0.65 0.19 0.17 0.66 0.16 0.16 0.65 0.17 0.16 0.66 0.18 0.16 0.66 0.18 0.16 0.62 0.17 0.16 Czech Republic 0.67 0.14 0.13 0.49 0.13 0.12 0.54 0.13 0.13 0.53 0.14 0.13 0.57 0.14 0.13 0.58 0.14 0.13 0.58 0.15 0.13 0.57 0.15 0.13 0.57 0.15 0.13 Denmark 0.68 0.14 0.13 0.54 0.14 0.14 0.60 0.16 0.14 0.53 0.13 0.13 0.59 0.14 0.14 0.56 0.13 0.13 0.58 0.15 0.13 0.55 0.14 0.13 0.53 0.13 0.13 Estonia 0.59 0.11 0.11 0.44 0.12 0.09 0.49 0.12 0.10 0.45 0.12 0.09 0.50 0.11 0.10 0.49 0.11 0.10 0.46 0.10 0.10 0.48 0.11 0.10 0.44 0.10 0.10 Finland 0.67 0.14 0.13 0.46 0.10 0.13 0.54 0.12 0.13 0.50 0.12 0.13 0.58 0.13 0.13 0.58 0.14 0.13 0.58 0.15 0.13 0.57 0.15 0.13 0.54 0.13 0.13 France 0.73 0.15 0.15 0.61 0.18 0.16 0.65 0.17 0.16 0.60 0.16 0.16 0.65 0.16 0.15 0.63 0.16 0.15 0.64 0.17 0.16 0.61 0.16 0.15 0.63 0.18 0.16 Germany 0.76 0.16 0.16 0.60 0.16 0.17 0.67 0.17 0.17 0.62 0.16 0.17 0.68 0.17 0.16 0.67 0.18 0.17 0.68 0.19 0.17 0.64 0.17 0.17 0.66 0.19 0.17 Greece 0.65 0.13 0.13 0.50 0.13 0.12 0.54 0.13 0.13 0.50 0.12 0.13 0.55 0.12 0.13 0.55 0.13 0.12 0.49 0.10 0.12 0.50 0.11 0.12 0.45 0.08 0.12 Hong Kong SAR, China 0.67 0.13 0.13 0.42 0.07 0.13 0.47 0.08 0.13 0.42 0.06 0.14 0.56 0.12 0.13 0.50 0.09 0.13 0.52 0.11 0.13 0.54 0.12 0.14 0.49 0.10 0.13 Hungary 0.66 0.14 0.13 0.51 0.15 0.12 0.55 0.13 0.12 0.50 0.13 0.12 0.56 0.13 0.12 0.56 0.14 0.12 0.55 0.14 0.13 0.56 0.16 0.12 0.56 0.16 0.13 Iceland 0.55 0.10 0.10 0.38 0.10 0.07 0.44 0.10 0.08 0.37 0.08 0.08 0.46 0.10 0.09 0.42 0.08 0.09 0.40 0.08 0.09 0.45 0.11 0.09 0.38 0.07 0.09 India 0.69 0.14 0.14 0.53 0.13 0.15 0.59 0.14 0.14 0.59 0.16 0.15 0.61 0.14 0.14 0.59 0.14 0.14 0.57 0.14 0.14 0.56 0.13 0.14 0.55 0.14 0.14 Indonesia 0.66 0.13 0.13 0.50 0.12 0.14 0.59 0.14 0.14 0.56 0.15 0.14 0.58 0.13 0.13 0.55 0.12 0.13 0.55 0.14 0.13 0.54 0.13 0.13 0.51 0.12 0.13 Ireland 0.68 0.14 0.13 0.51 0.13 0.13 0.60 0.16 0.13 0.50 0.11 0.13 0.61 0.15 0.14 0.53 0.11 0.13 0.54 0.13 0.13 0.57 0.15 0.13 0.50 0.11 0.13 Israel 0.65 0.13 0.12 0.51 0.15 0.12 0.51 0.11 0.12 0.50 0.13 0.12 0.57 0.14 0.12 0.53 0.12 0.12 0.52 0.12 0.12 0.54 0.14 0.12 0.51 0.12 0.12 Italy 0.72 0.15 0.15 0.57 0.15 0.16 0.63 0.16 0.15 0.62 0.17 0.16 0.64 0.16 0.15 0.63 0.16 0.15 0.64 0.17 0.16 0.60 0.15 0.15 0.61 0.16 0.16 Japan 0.73 0.15 0.15 0.51 0.11 0.16 0.57 0.12 0.15 0.57 0.14 0.16 0.64 0.15 0.15 0.63 0.16 0.16 0.64 0.16 0.16 0.63 0.16 0.16 0.63 0.17 0.16 Korea, Rep. 0.71 0.15 0.14 0.48 0.10 0.14 0.57 0.13 0.14 0.58 0.15 0.15 0.63 0.15 0.14 0.61 0.15 0.14 0.61 0.15 0.15 0.62 0.17 0.15 0.60 0.17 0.15 Latvia 0.58 0.11 0.11 0.44 0.12 0.10 0.49 0.11 0.10 0.45 0.11 0.10 0.48 0.09 0.11 0.48 0.11 0.10 0.44 0.09 0.10 0.44 0.09 0.10 0.43 0.09 0.10 Lithuania 0.59 0.11 0.11 0.47 0.13 0.10 0.51 0.12 0.11 0.47 0.12 0.11 0.53 0.12 0.11 0.48 0.11 0.10 0.45 0.10 0.10 0.46 0.10 0.10 0.45 0.10 0.10 Luxembourg 0.64 0.13 0.12 0.46 0.11 0.11 0.52 0.12 0.12 0.50 0.13 0.12 0.54 0.12 0.12 0.55 0.14 0.12 0.51 0.12 0.12 0.49 0.11 0.12 0.44 0.08 0.12 Malaysia 0.69 0.14 0.14 0.57 0.17 0.14 0.59 0.14 0.14 0.54 0.14 0.13 0.60 0.15 0.13 0.56 0.13 0.13 0.59 0.16 0.13 0.60 0.16 0.14 0.52 0.12 0.14 Malta 0.54 0.10 0.09 0.36 0.09 0.07 0.41 0.09 0.08 0.34 0.07 0.08 0.43 0.09 0.08 0.36 0.06 0.08 0.37 0.07 0.08 0.46 0.11 0.09 0.38 0.08 0.08 Mexico 0.67 0.14 0.13 0.51 0.13 0.13 0.55 0.13 0.13 0.52 0.14 0.13 0.57 0.13 0.13 0.57 0.14 0.13 0.55 0.13 0.13 0.58 0.16 0.13 0.57 0.16 0.13 Netherlands 0.71 0.15 0.14 0.60 0.18 0.15 0.65 0.18 0.15 0.56 0.14 0.15 0.64 0.16 0.15 0.61 0.15 0.14 0.61 0.16 0.15 0.58 0.15 0.15 0.58 0.15 0.15 New Zealand 0.60 0.12 0.11 0.49 0.14 0.11 0.54 0.14 0.12 0.48 0.12 0.11 0.50 0.11 0.11 0.50 0.11 0.11 0.46 0.10 0.11 0.46 0.10 0.11 0.45 0.10 0.11 Norway 0.68 0.13 0.14 0.54 0.14 0.14 0.59 0.14 0.14 0.48 0.09 0.14 0.59 0.13 0.15 0.59 0.14 0.14 0.58 0.14 0.14 0.54 0.12 0.14 0.53 0.13 0.14 Philippines 0.63 0.13 0.12 0.43 0.10 0.11 0.46 0.09 0.11 0.49 0.13 0.11 0.48 0.09 0.11 0.47 0.09 0.11 0.48 0.10 0.12 0.56 0.15 0.13 0.49 0.11 0.12 Poland 0.68 0.14 0.13 0.52 0.14 0.13 0.58 0.15 0.13 0.55 0.15 0.13 0.59 0.14 0.13 0.60 0.15 0.14 0.58 0.15 0.14 0.57 0.14 0.13 0.59 0.16 0.14 Portugal 0.64 0.13 0.12 0.47 0.12 0.12 0.54 0.13 0.12 0.54 0.15 0.13 0.56 0.13 0.12 0.54 0.13 0.12 0.52 0.12 0.12 0.53 0.13 0.12 0.53 0.14 0.12 Romania 0.63 0.12 0.12 0.48 0.13 0.11 0.49 0.10 0.12 0.52 0.14 0.12 0.54 0.13 0.12 0.55 0.14 0.12 0.53 0.13 0.12 0.51 0.12 0.12 0.52 0.13 0.12 Russian Federation 0.71 0.14 0.16 0.59 0.15 0.17 0.59 0.12 0.16 0.53 0.11 0.16 0.64 0.14 0.17 0.64 0.15 0.17 0.61 0.14 0.16 0.55 0.12 0.15 0.57 0.14 0.16 Saudi Arabia 0.66 0.12 0.14 0.49 0.10 0.15 0.56 0.12 0.14 0.50 0.10 0.15 0.61 0.13 0.15 0.56 0.12 0.15 0.48 0.08 0.14 0.49 0.09 0.14 0.47 0.08 0.14 Singapore 0.71 0.15 0.14 0.47 0.10 0.14 0.58 0.14 0.14 0.51 0.12 0.14 0.63 0.16 0.14 0.57 0.14 0.13 0.59 0.16 0.14 0.60 0.16 0.14 0.56 0.14 0.14 Slovak Republic 0.64 0.13 0.12 0.47 0.13 0.11 0.52 0.13 0.11 0.50 0.14 0.11 0.55 0.13 0.12 0.55 0.14 0.12 0.54 0.14 0.12 0.54 0.14 0.12 0.55 0.15 0.12 Slovenia 0.61 0.12 0.11 0.43 0.11 0.10 0.48 0.11 0.10 0.48 0.13 0.10 0.53 0.13 0.11 0.53 0.13 0.11 0.52 0.13 0.11 0.50 0.12 0.11 0.50 0.13 0.11 South Africa 0.63 0.13 0.12 0.50 0.15 0.11 0.52 0.13 0.11 0.46 0.11 0.11 0.54 0.12 0.12 0.56 0.14 0.13 0.53 0.00 0.12 0.48 0.10 0.12 0.53 0.14 0.12 Spain 0.70 0.14 0.14 0.56 0.16 0.14 0.61 0.15 0.14 0.58 0.15 0.15 0.63 0.15 0.14 0.60 0.15 0.14 0.59 0.15 0.14 0.58 0.14 0.14 0.60 0.16 0.15 Sweden 0.70 0.14 0.14 0.51 0.12 0.14 0.59 0.14 0.14 0.53 0.13 0.14 0.61 0.15 0.14 0.60 0.15 0.14 0.61 0.16 0.15 0.58 0.15 0.14 0.59 0.16 0.14 Switzerland 0.69 0.14 0.14 0.50 0.12 0.14 0.59 0.15 0.14 0.54 0.14 0.14 0.62 0.15 0.14 0.59 0.14 0.14 0.60 0.16 0.15 0.59 0.15 0.14 0.55 0.13 0.14 Taiwan, China 0.69 0.14 0.14 0.49 0.12 0.13 0.52 0.11 0.13 0.55 0.15 0.14 0.61 0.15 0.13 0.59 0.15 0.14 0.59 0.15 0.14 0.61 0.16 0.15 0.55 0.13 0.14 Thailand 0.68 0.14 0.13 0.51 0.13 0.13 0.59 0.15 0.14 0.57 0.16 0.13 0.58 0.13 0.13 0.55 0.13 0.13 0.55 0.13 0.13 0.59 0.16 0.14 0.54 0.13 0.13 Turkey 0.67 0.13 0.13 0.51 0.13 0.13 0.56 0.13 0.13 0.58 0.16 0.14 0.58 0.14 0.13 0.60 0.15 0.13 0.56 0.14 0.13 0.53 0.13 0.13 0.56 0.15 0.13 United Kingdom 0.72 0.15 0.15 0.55 0.13 0.16 0.63 0.15 0.16 0.59 0.15 0.16 0.65 0.16 0.15 0.63 0.15 0.16 0.62 0.16 0.16 0.61 0.15 0.15 0.62 0.17 0.16 United States 0.75 0.15 0.16 0.64 0.18 0.18 0.67 0.17 0.17 0.61 0.15 0.17 0.68 0.16 0.17 0.64 0.15 0.17 0.66 0.17 0.17 0.63 0.16 0.17 0.65 0.18 0.17 Vietnam 0.63 0.13 0.12 0.53 0.17 0.11 0.56 0.15 0.12 0.54 0.16 0.12 0.51 0.11 0.11 0.50 0.11 0.11 0.48 0.11 0.11 0.50 0.12 0.11 0.46 0.10 0.11 Source: Calculations using data from Santoni and Taglioni 2015. Note: The cells are colored according to the strength of the metrics—a green cell indicates a strong measure, an orange cell indicates a weak measure, and a white cell indicates an average measure. BONwin = eigenvector centrality based on inflows of value added; BONwout = eigenvector centrality based on outflows of value added; CCw = clustering index; nec = not elsewhere classified. 52 Making Global Value Chains Work for Development Table 2A.2 Logistics Performance, Domestic Component, 2014 Quality of infrastructure: Evaluate the quality of trade- and transport-related infrastructure in your country of work (% of respondents answering low/very low) United Bulgaria China Germany Japan Poland Portugal Romania Turkey States Ports 66.7 5.4 0.0 50.0 0.0 33.3 0.0 12.5 11.1 Airports 33.3 0.0 0.0 0.0 0.0 0.0 0.0 16.1 5.6 Roads 44.4 0.0 0.0 0.0 50.0 0.0 33.3 12.5 16.7 Rail 55.6 5.3 0.0 0.0 50.0 33.3 0.0 61.3 50.0 Warehousing/transloading 22.2 0.0 0.0 0.0 0.0 0.0 0.0 3.1 5.6 facilities Telecommunications and IT 0.0 5.3 0.0 0.0 0.0 0.0 0.0 6.3 0.0 Competence and quality of services: Evaluate the competence and quality of service delivered by the following in your country of work (% of respondents answering high/very high) United Bulgaria China Germany Japan Poland Portugal Romania Turkey States Road 66.7 28.2 90.0 50.0 50.0 100.0 33.3 80.7 50.0 Rail 44.4 15.8 68.4 50.0 0.0 0.0 0.0 20.0 33.3 Air transport 55.6 43.6 94.4 50.0 0.0 66.7 33.3 70.0 66.7 Maritime transport 44.4 50.0 84.2 50.0 50.0 33.3 100.0 83.3 55.6 Warehousing/transloading 77.8 38.5 90.0 50.0 50.0 66.7 66.7 77.4 50.0 and distribution Freight forwarders 88.9 44.7 85.0 50.0 100.0 66.7 66.7 80.7 66.7 Customs agencies 77.8 33.3 85.0 50.0 100.0 33.3 33.3 54.8 33.3 Quality/standards inspection 33.3 18.4 85.0 50.0 50.0 33.3 33.3 46.7 29.4 agencies Health/SPS agencies 22.2 23.7 76.5 50.0 0.0 33.3 33.3 33.3 22.2 Customs brokers 77.8 29.0 80.0 50.0 100.0 33.3 66.7 54.8 55.6 Trade and transport 44.4 34.2 77.8 50.0 50.0 0.0 66.7 67.7 50.0 associations Consignees or shippers 66.7 35.9 84.2 50.0 0.0 0.0 66.7 61.3 23.5 Source: Calculations using 2014 World Bank Logistics Performance Index data. Note: The cells are colored according to the performance of the indicator—a green cell indicates strong performance an orange cell indicates weak performance, and a white cell indicates average peformance. IT = information technology; SPS = sanitary and phytosanitary. PART II QUANTIFYING A COUNTRY’S POSITION IN GVCS To guide policy makers in achieving development through integration in global value chains (GVCs), the key concepts and metrics of the country’s GVC participation must be investigated. Understanding how countries fare in such key concepts and metrics allows a better identification of specific value chains, activities, and business segments, which are the object of the case studies in this part. The assessment of a country’s GVC participation focuses on three concepts: • Role in GVCs: the buyer’s perspective versus the seller’s perspective • Specialization and domestic value-added contribution: specialization in low or high value added, preproduction, assembly, postproduction, or support activities • Position in GVC network and type of GVC node: incoming spoke, hub, or outgoing spoke, clustering properties, or centrality in the global network Chapter 3—“What Do Imports and Exports Say about GVC Participation?”—discusses how gross import and export flows can be used to gather some initial insights into a country’s participation in GVCs. The chapter also delves into how much of the gross flows represent value addition in the country of interest and how to quantify the domestic value added embodied in a country’s exports. Chapter 4—“Buyer-Related Measures”—covers more indirect measures, such as the share of intermediates in gross imports based on combining gross trade data with informed classifications, and more direct quantifications of a country’s position in GVCs, such as the foreign value added embodied in the country’s gross exports or the length of sourcing chains. Adding information on factors of production (labor and wages, ideas, and investment) enables further characterization of the buyer function in GVCs for a country. Similarly, in chapter 5—“Seller-Related Measures”—the measures can be more indirect, such as the share of intermediates in output or gross exports, or more direct, including the domestic value added embodied in gross exports of third countries and the length of selling chains. Like the analysis of the buyer dimension, information on factors of production further helps characterize a country’s participation as a seller in GVCs. Chapter 6—“Other Measures of GVC Participation: From Macro to Micro”—complements the buyer- and seller-related measures with additional measures of GVC participation at the macro and micro levels. The chapter includes the GVC participation index, network metrics and their visualizations, the role of services in value added, and firm-level links in GVCs. Chapter 7—“Use of GVC Measures to Assess the Drivers and Impacts of GVC Participation”—presents selected topics for a research agenda on ways to test for the drivers and impacts of GVC participation using statistical methods or econometrics or by quantifying direct relationships in international input- output tables. The multidimensional nature of GVCs can be captured by examining the relationships between goods, services, workers, ideas, and investments, going beyond value added to identify the actors in GVCs and how to assess the impacts of GVCs on jobs and wages. This part uses country-specific examples to illustrate the concepts and suggested analysis. Chapter 3 WHAT DO IMPORTS AND EXPORTS SAY ABOUT GVC PARTICIPATION? GVC Participation Using Gross Trade Data The usefulness of eyeballing the top imported and exported products as a starting point in GVC Gross export data indicate what products a country analysis can be seen in the example in table 3.1, exports, whereas gross import data indicate what it which reports Malaysia’s 50 top exports, and table imports. These data do not provide any indication of 3.2, which reports the country’s 50 top imports. The the domestic or foreign source of the inputs or the most important export product, other monolithic value addition generated in the country. To gather integrated circuits (HS code 854230), accounts for information on the latter requires more sophisticated 10.5 percent of overall exports. The importance of data (discussed herein). Even so, a first assessment of individual items rapidly decreases. Two items, lique- a country’s global value chain (GVC) participation fied natural gas and petroleum and oils (not crude), can be based on gross exports and imports data. cover more than 6 percent of exports each, and 11 The first consideration when investigating a additional products cover a share of 1 to 5 percent of country’s potential in GVCs is what the country total exports each. All other items, individually, rep- exports and imports. Looking at the top 50 to 100 resent less than 1 percent of total exports. Imports export and import products of a country, classified follow a similar distribution. at the most disaggregated level (at least Harmonized Refinements of such a first-cut analysis increase System [HS] 6-digit or Standard International Trade the relevance of GVC analysis. Four types of refine- Classification 5-digit products) is a good starting ments should be considered: point. For most countries, the top 50 exports and imports are likely to cover at least 50 percent of the • Consider raw commodities separately from other total trade value in each direction, and the top 100 products. Although raw commodities are impor- cover at least 75 percent. tant import and export items for most countries A country’s distribution of exports tends to follow worldwide, their relevance in a GVC analysis a lognormal, power, or Zipf ’s law distribution. Zipf ’s is limited. Therefore, the various analyses sug- law, originally applied to language, states that given gested in this chapter and elsewhere in part II of some universe of items, the frequency of any item is the book should be run in two ways—including inversely proportional to its rank in the frequency raw commodities and excluding them—and the table. That is, a few items account for the bulk of results compared. the given universe; the contribution of most items is • Compare product-level imports with export val- marginal. Exports and imports loosely follow such ues, volumes, and prices of the top traded products. asymmetric distribution laws. Therefore, the mar- If exports and imports follow a similar distri- ginal additional information that can be gathered bution and the values or volumes traded have a from import and export products beyond the top 50 similar growth or level, this may suggest that rela- to 100 is generally small. tively little transformation may be taking place 55 56 Making Global Value Chains Work for Development Table 3.1. Malaysia’s Top 50 Exports, 2012 HS-6 Rank Code Description Value (US$) Share (%) 1 854230 Monolithic integrated circuits 23,846,665 10.5 2 271111 Liquefied natural gas 17,974,365 7.9 3 271000 Petroleum oils/oils from bituminous minerals (not crude) 15,419,273 6.8 4 151190 Palm oil or fractions simply refined 10,935,036 4.8 5 270900 Petroleum oils/oils from bituminous minerals (crude) 10,440,086 4.6 6 847170 Storage units 5,881,013 2.6 7 847330 Parts & accessories of the machines of heading No. 84.71 5,331,637 2.3 8 151110 Crude oil 4,504,723 2.0 9 852812 Color television receiver 3,214,780 1.4 10 401519 Other gloves 3,072,135 1.4 11 854140 Photosensitive/photovoltaic/LED semiconductor devices 2,491,291 1.1 12 400122 Technically specified natural rubber (TSNR) 2,382,671 1.0 13 151620 Vegetable fats & oils & their fractions 2,348,416 1.0 14 854290 Parts of electronic integrated circuits, etc. 2,248,361 1.0 15 847180 Other units of automatic data processing machines 2,057,050 0.9 16 853710 Electrical control & distribution boards, <1kV 1,981,275 0.9 17 711319 Jewelry & parts of precious metal except silver 1,876,044 0.8 18 854129 Transistors, other than photosensitive transistors 1,743,263 0.8 19 400599 Compounded unvulcanized rubber in primary forms, nes 1,666,362 0.7 20 853400 Printed circuits 1,417,904 0.6 21 853690 Electrical switch, protector, connector for <1kV, nes 1,361,028 0.6 22 903090 Parts & accessories, electrical measuring instruments 1,352,642 0.6 23 441213 Plywood, outer ply of tropical wood 1,305,017 0.6 24 852520 Transmission apparatus incorporating reception apparatus 1,262,680 0.6 25 999999 Commodities, nes 1,241,447 0.5 26 852990 Parts of radio/TV transmit/receive equipment, nes 1,126,296 0.5 27 854190 Parts of semiconductor devices & similar devices 1,049,935 0.5 28 851780 Electronic apparatus for telephone line 926,542 0.4 29 271129 Petroleum gases & gaseous hydrocarbons, nes, liquefied 925,991 0.4 30 844359 Other printing machinery 910,630 0.4 31 841510 Window or wall types, self-contained 900,525 0.4 32 382319 Other industrial monocarboxylic fatty acids 840,678 0.4 33 940360 Other wooden furniture 793,101 0.3 34 800110 Tin, not alloyed 780,070 0.3 35 854110 Diodes, other than photosensitive or light emitting diodes 778,835 0.3 36 903040 Gain/distortion & crosstalk meters, etc. 772,792 0.3 37 852320 Magnetic discs 764,608 0.3 38 382490 Chemical prep, allied in 712,140 0.3 39 852540 Still image video cameras & other video camera recorders 691,899 0.3 40 850910 Vacuum cleaners 654,134 0.3 41 847990 Parts of machines & mechanical appliances, nes 599,320 0.3 42 851790 Parts of telephone line/telegraph equipment, nes 596,025 0.3 43 390110 Polyethylene having a specific gravity of less than 0.94 587,103 0.3 44 392690 Plastic articles, nes 585,215 0.3 45 940350 Wooden furniture of a kind used in the bedroom 583,510 0.3 46 382370 Industrial fatty alcohols 581,726 0.3 47 880330 Other parts of airplanes or helicopters 572,487 0.3 48 903082 Instruments for measuring or checking semiconductors 569,050 0.3 49 151329 Palm kernel or babassu oil & fractions thereof 564,780 0.2 50 844390 Parts of printing machinery & ancillary equipment 559,765 0.2 Source: Adapted from the United Nations Comtrade database. Note: Products from the electrical and electronics sector are color-coded in green, whereas those from the oil, gas, and petrochemical sector are color-coded in yellow, and products from all other sectors in gray. HS = Harmonized System; kV = kilovolts; LED = light-emitting diode; nes = not elsewhere specified. What Do Imports and Exports Say about GVC Participation? 57 Table 3.2. Malaysia’s Top 50 Imports, 2012 HS-6 Rank Code Description Value (US$) Share (%) 1 271000 Petroleum oils/oils from bituminous minerals (not crude) 15,596,099 7.9 2 854230 Monolithic integrated circuits 15,193,085 7.7 3 854290 Parts of electronic integrated circuits, etc. 10,704,458 5.4 4 270900 Petroleum oils/oils from bituminous minerals (crude) 8,963,271 4.6 5 847330 Parts & accessories of the machines of heading No. 84.71 3,672,195 1.9 6 880240 Airplanes & other aircraft, of an unladen weight <15,000kg 3,246,371 1.7 7 852520 Transmission apparatus incorporating reception apparatus 2,168,024 1.1 8 270119 Other coal, whether or not pulverized, but not agglomerated 2,103,913 1.1 9 852990 Parts of radio/TV transmit/receive equipment, nes 2,061,850 1.0 10 853400 Printed circuits 1,961,253 1.0 11 870323 Automobiles, spark ignition of 1500-3000cc 1,880,437 1.0 12 740311 Refined copper: Cathodes & sections of cathodes 1,674,219 0.9 13 710813 Non-monetary: Other semi-manufactured forms 1,534,985 0.8 14 853120 Indicator panels incorporating LCD or LED 1,438,878 0.7 15 854190 Parts of semiconductor devices & similar devices 1,267,170 0.6 16 847130 Portable digital automatic data processing machines, >10kg 1,156,743 0.6 17 400110 Natural rubber latex, whether or not pre-vulcanized 1,134,514 0.6 18 710812 Non-monetary: Other unwrought forms 1,116,523 0.6 19 711590 Articles of, or clad with, precious metal, nes 1,041,547 0.5 20 999999 Commodities nes 1,017,398 0.5 21 847170 Storage units 1,010,919 0.5 22 170111 Cane sugar w/o flavoring or coloring matter 973,042 0.5 23 847989 Other machines & mechanical appliances 940,269 0.5 24 853690 Electrical switch, protector, connector for <1kV, nes 921,485 0.5 25 151190 Palm oil or fractions simply refined 908,447 0.5 26 400122 Technically specified natural rubber (TSNR) 898,192 0.5 27 180100 Cocoa beans, whole or broken, raw or roasted 877,533 0.4 28 151110 Crude oil 796,517 0.4 29 870421 Diesel powered trucks, >5tonnes 782,424 0.4 30 300490 Medicaments nes, in dosage 775,336 0.4 31 880330 Other parts of airplanes or helicopters 717,258 0.4 32 392690 Plastic articles, nes 707,399 0.4 33 847990 Parts of machines & mechanical appliances, nes 703,534 0.4 34 851780 Electronic apparatus for telephone line 689,394 0.4 35 853710 Electrical control & distribution boards, <1kV 657,043 0.3 36 870829 Other parts & accessories of bodies (including cabs) 653,313 0.3 37 901380 Other devices, appliances, & instruments 624,489 0.3 38 310420 Potassium chloride 624,451 0.3 39 903090 Parts & accessories, electrical measuring instruments 616,256 0.3 40 230400 Soybean oil: Oil-cake & other solid residues 611,933 0.3 41 760110 Aluminum, not alloyed 609,133 0.3 42 100630 Semi-milled or wholly milled rice 599,986 0.3 43 210690 Food preparations, nes 596,375 0.3 44 730511 Pipeline submerged arc welded steel, diameter >406mm 587,927 0.3 45 870322 Automobiles, spark ignition of 1000-1500cc 572,274 0.3 46 850440 Static converters 561,011 0.3 47 520100 Cotton, not carded or combed 546,293 0.3 48 844390 Parts of printing machinery & ancillary equipment 542,525 0.3 49 100590 Maize except seed corn 535,904 0.3 50 854129 Transistors, other than photosensitive transistors 481,679 0.2 Source: Adapted from the United Nations Comtrade database. Note: Products from the electrical and electronics sector are color-coded in green, whereas those from the oil, gas, and petrochemical sector are color-coded in yellow, and products from all other sectors in gray. cc = cubic centimeters; HS = Harmonized System; kV = kilovolts; LCD = liquid crystal display; LED = light-emitting diode; mm = millimeters; nes = not elsewhere specified; w/o = without. 58 Making Global Value Chains Work for Development domestically, with the domestic segment of the analysis of a country’s imports and exports. We use country’s major GVCs being relatively short. the case of Malaysia for the illustrations. • Use informed classifications to extract as much infor- mation as possible from gross trade data. Regrouping Illustration #1: Final Use data in meaningful clusters or categorized by informed classifications is also very helpful. To start, the analysis focuses on the most basic clas- • Document trade flows at the subnational level. sification, the UN Broad Economic Categories. The Acknowledging the fact that subnational differ- majority of Malaysia’s top 50 exports and imports, ences may exist, data that take into account the measured at the HS 6-digit level, are intermediate subnational perspective should be used when products from two sectors: oil, gas, and petrochemi- available. cals (color-coded in yellow in the figures) and elec- trical and electronics industry (in green) (tables 3.1 and 3.2). Informed Classifications Comparison of the top imports and exports of a country provides further preliminary insights on Informed classifications are useful for identifying fea- their participation in GVCs. Figure 3.1 provides such tures and investigating specific aspects of GVC trade, including parts and components, technical func- tions, and so forth. Multiple classifications exist, with Figure 3.1. Malaysia’s Top 50 Exports and Imports, 2012 concordance tables that allow matching them with standard trade data, and researchers are constantly 1 854230 1 1 2 271000 3 3 developing new ones. Some examples of useful classifi- 151190 4 4 270900 5 5 cations that are widely available include the following: 5 847170 6 847330 7 7 151110 8 • The United Nations (UN 2002) Broad Economic 9 10 10 Categories focus on the final use and distinguish 400122 12 between consumer goods, capital goods, and 854290 14 15 intermediates.1 15 853710 16 • Athukorala (2010) identifies parts and compo- nents at a very detailed level of aggregation for 853400 20 20 20 853690 21 21 East Asia. 903090 22 • Sturgeon and Memedovic (2011) show how to 852520 24 24 Rank 999999 25 25 identify final goods and intermediate goods—the 25 852990 26 26 854190 27 latter further categorized in standard or custom- 851780 28 28 ized intermediates—in specific GVCs (electronics, 30 31 vehicles and parts, and textiles/apparel/footwear, 32 33 and more recently raw and processed food and 34 35 35 chemicals and related products). • Taymaz, Voylvoda, and Yilmaz (2011), based 39 on engineering considerations—and in a paper 40 847990 41 applied to Turkey—similarly assign products and activities to five value chain categories (final 392690 44 45 products, main inputs/parts, standard inputs, raw 880330 47 material, and machinery and equipment). The 48 system focuses on five typical GVCs: televisions, 50 844390 50 motor vehicles, food, machinery, and textiles and HS-6 Export Import rank rank apparel, the last distinguishing cotton, wool, syn- thetic, and other. Products from the electrical and electronics sector Products from the oil, gas, and petrochemical sector A discussion of these classification systems can Products from all other sectors be found in appendixes B to E. The following exam- Source: Adapted from the United Nations Comtrade database. ples illustrate how the classifications can enrich the Note: HS = Harmonized System. What Do Imports and Exports Say about GVC Participation? 59 a comparison for Malaysia. The figure shows that not stylized facts that emerge from the analysis of pro- only are they intermediate goods, parts, and compo- duction and exports in Turkey are the following: nents that feed into the production of other prod- ucts, but the majority of the top products also appear • For motor vehicles in 2010, more than 70 percent as top imports and top exports. This fact raises the of the sector’s exports were final products. The question of how much value added is provided second most important production stage in the within segments of GVCs located in the country. Turkish automotive value chain, by export value, This type of preliminary test therefore flags whether is standard input production. Exports of standard participation in GVCs remains marginal, with little inputs accounted for nearly 25 percent of total domestic transformation. exports. Although the share of main parts and components exports increased from 2003 to 2010, Illustration #2: Value Chain Category those exports remained marginal (less than 3 per- cent in 2010). The second example uses the classification of Taymaz, • Similar to motor vehicles, Turkish textile export- Voylvoda, and Yilmaz (2011) to categorize firm-level ers tend to concentrate in the final stage of textile export and production data (measured as firm value production. About 70 percent of the export value added) for the motor vehicles, textiles and apparel, and more than 50 percent of value addition is and agrifood sectors in Turkey (table 3.3).2 This clas- generated by final goods exports. sification assigns exports to one of five value chain • The agrifood sector stands apart from the previ- categories, namely: final products, main inputs/ ous two. Although the majority of production is parts, standard inputs, raw materials, and machinery in final goods, the majority of export growth has and equipment. Although this information is only been concentrated in fairly unsophisticated prod- available for five industries—motor vehicles, televi- ucts, such as grains, nuts, and lentils, a fact that is sions, food, machinery, and textiles and apparel—it reflected in the high export share of raw materials. is nevertheless useful, as these five industries repre- • Overall, the numbers in the table reveal that sent important GVCs in many countries. The key Turkish participation in the agrifood value chain Table 3.3. Turkey’s Share of Exports and Value Added, 2003 and 2010 (% of total) Main Standard Raw Machinery and Final inputs inputs materials equipment Total Motor vehicles Export share, 2003 73.5 0.4 24.8 0.2 1.1 100 Value-added share, 2003 48.0 2.7 46.9 0.2 2.1 100 Export share, 2010 72.3 2.4 23.9 0.3 1.1 100 Value-added share, 2009 38.4 1.1 53.3 0.7 6.5 100 Textiles and apparel Export share, 2003 78.8 11.9 7.3 1.7 0.3 100 Value-added share, 2003 55.5 19.1 13.5 9.2 2.7 100 Export share, 2010 70.0 15.0 8.6 5.7 0.7 100 Value-added share, 2009 56.0 17.0 13.2 11.8 2.0 100 Agrifood Export share, 2003 34.3 19.2 n.a. 44.6 1.8 100 Value-added share, 2003 48.9 18.5 n.a. 17.4 15.2 100 Export share, 2010 37.4 17.9 n.a. 42.0 2.7 100 Value-added share, 2009 56.0 16.8 n.a. 13.3 13.9 100 Source: World Bank 2014. Note: n.a. = not available. 60 Making Global Value Chains Work for Development is less advanced than in the other two sectors. The by Sturgeon and Memedovic (2011) as customized low specialization in machinery and equipment is or standard parts and components. A global view of also a symptom that Turkish GVC participation international trade through these lenses suggests that: tends to concentrate in low value-added segments of GVCs. • High-income countries specialize in final assem- bly—but assembly is also an important activ- Illustration #3: Customized Trade and the World ity in upper-middle-income countries (such as Bank MC-GVC Database and Country Dashboards Argentina, Hungary, Mexico, South Africa, and Turkey), as well as in some poorer countries (such Distinguishing customized from standard interme- as Cambodia and India). diates can be used to assess participation and links • The supply chain for customized parts and com- at the industry and product levels. The World Bank ponents extends widely into the middle-income Group for example has constructed a database for level, but not to low-income countries. A drill- measuring competitiveness in GVCs, the Measuring down shows that ignition wiring sets for autos, Competitiveness across Global Value Chains for example, are widely exported from middle- (MC-GVC) database, which allows tracking of six cat- income countries. egories of goods trade in three archetypal GVC indus- • Japan keeps more of its customized parts and tries: electronics, apparel and footwear, and autos and components at home than North America or motorcycles. For each of these industries, the data- Western Europe does (unlike in electronics, where base identifies whether goods are for intermediate or the supply chain is more widely distributed). final use and, for intermediates, if they are classified Information at the country level can be used to Table 3.4. Auto Cluster leverage firm-level hypotheses on participation Nomenclature of Economic Activity rates and trends, export and import links, specific Revision 2 code Revision 2 description products of interest, two-way trade in differentiated 1392 Manufacture of made-up textile articles, except apparel products, and competition in destination markets. 2219 Manufacture of other rubber products An example is the production of color television 2222 Manufacture of plastic packing goods receivers. Analysis of gross trade flows at the prod- 2229 Manufacture of other plastic products uct level matched with information on customized 2433 Cold forming or folding trade shows that Tunisian color televisions have an 2511 Manufacture of metal structures and parts of structures 2550 Forging, pressing, stamping and roll-forming of metal; increasing market share in France, but there is ruth- powder metallurgy less competition at the global level, as the same mar- 2572 Manufacture of locks and hinges ket segment is populated by important global actors, 2573 Manufacture of tools such as LG from the Republic of Korea. 2593 Manufacture of wire products, chain, and springs 2732 Manufacture of other electronic and electric wires and cables Illustration #4: Product Clusters 2740 Manufacture of electric lighting equipment 2790 Manufacture of other electrical equipment Another useful method to enrich the data is to com- 2822 Manufacture of lifting and handling equipment bine production or export data that belong to dif- 2841 Manufacture of metal forming machinery ferent sectors but to the same GVC (or cluster of 2849 Manufacture of other machine tools 2892 Manufacture of machinery for mining, quarrying, and products). An example of a cluster for the auto sec- construction tor is illustrated in table 3.4, which reports the clus- 2899 Manufacture of other special-purpose machinery, nec tering of sectors by economic activity carried out 2910 Manufacture of motor vehicles by a regional development agency in Romania, the 2920 Manufacture of bodies (coachwork) for motor vehicles; Romania-West Development Agency (following the manufacture of trailers and semi-trailers 2931 Manufacture of electrical and electronic equipment for Nomenclature of Economic Activities, revision 2). motor vehicles The table shows that the auto GVC includes not only 2932 Manufacture of other parts and accessories for motor vehicles and their parts and components, but also vehicles molding and other metallurgy activity, production 3299 Other manufacturing, nec Source: World Bank 2013. of specialized textiles, rubber products, packaging Note: NACE = Nomenclature of Economic Activities; nec = not elsewhere classified. goods, plastics, and a variety of additional products What Do Imports and Exports Say about GVC Participation? 61 from the electric and electronics industry, machin- Figure 3.2. Most Relevant Buyers of Computer Storage Devices ery, and equipment. Appendix F provides examples of clusters for agrifood, construction, energy, health, 336390 information and communications technology, tex- tiles, and tourism. Although such cluster classifi- cations tend to be similar across countries, some 332710 334610 differences at the very granular level can be expected. 326110 Different market segments and business models 334510 affect the way activities are clustered together. 33441A Input-output (I-O) tables are a useful first step in identifying clusters of activities across sectors. Figure 334112 334418 3.2 shows the main sectors buying computer stor- 332800 age devices (North American Industry Classification System [NAICS] product 334112), and figure 3.3 335999 identifies the main inputs of this product, based on 3259A0 334511 U.S. I-O tables. The green nodes in figure 3.2 are 334413 334111 downstream industries that use NAICS 334112 as 33451A inputs in production and for which computer stor- age devices (red) represent at least 1 percent of the total input requirements for their production (nodes 33411A and links are built using network analysis metrics, Most relevant buyers NAICS product 334112 Other illustrated in chapter 6; for more details, see box 8.1 in chapter 8). The buyer industries are as follows: Source: Santoni and Taglioni forthcoming. Adapted from Benchmark Input-Output Data, Bureau of Economic Analysis, U.S. Department of Commerce. Note: Red lines designate flows of computer storage devices (red circle) to main buying sectors (green • Sector 334510: electromedical and electrothera- circles). NAICS = North American Industry Classification System. peutic apparatus • Sector 334111: electronic computer manufactur- Figure 3.3. Most Relevant Suppliers for Computer Storage Devices ing with small business administration standards, which includes manufacturing and/or assembling 336390 electronic computers, such as mainframes, per- sonal computers, workstations, laptops, and com- puter servers 332710 334610 • Sector 33411A: other computer manufacturing 326110 334510 • Sector 334511: search, detection, navigation, guidance, aeronautical, and nautical system and 33441A instrument manufacturing • Sector 33451A: other measuring and controlling 334112 334418 device manufacturing. 332800 What sectors are the most important suppliers for 335999 this sector? The green nodes in figure 3.3 show the 3259A0 334511 following suppliers of inputs to 334112: 334413 334111 33451A • Sector 334610: software reproduction • Sector 33411A: other computer manufacturing • Sector 334418: printed circuit assembly 33411A • Sector 335999: all other miscellaneous electrical equipment and component manufacturing Most relevant suppliers NAICS product 334112 Other • Sector 33441A: other electronic component Source: Santoni and Taglioni forthcoming. Adapted from Benchmark Input-Output Data, Bureau of Economic Analysis, U.S. Department of Commerce. manufacturing Note: Green lines designate main input flows from supplying sectors (green circles) to the computer • Sector 332800: metal treating storage sector (red circle). NAICS = North American Industry Classification System. 62 Making Global Value Chains Work for Development • Sector 3259A0: other chemical product and prep- this part of the book introduces GVC participa- aration manufacturing tion measures. Most of the measures illustrated in • Sector 326110: plastics packaging materials and chapters 4 to 7 require the use of these databases. unlaminated film and sheet manufacturing Furthermore, some of the measures can only be com- • Sector 334413: semiconductor and related device puted with some of the databases mentioned; when manufacturing that is the case, it is explicitly mentioned in the text. • Sector 332710: machine shops To facilitate the illustration of some of the under- • Sector 336390: other motor vehicle parts lying economic concepts, key measures of GVC par- manufacturing. ticipation that draw on these databases are discussed by differentiating between buyer-related (chapter 4) The upstream products that computer storage and seller-related (chapter 5) measures and combin- devices use as inputs are more numerous than the ing the measures to assess the overall GVC partici- downstream products. Such mappings enable the pation of countries (chapter 6). These databases can definition of clusters of activities or products for fur- also be used for network analysis—for example, to ther analysis. construct measures of centrality and structural inte- Clustering can also include information on gration in GVCs (chapter 6). Finally, the data can be the geographic location of activities in coun- used in econometric and statistical methods (chapter tries. Examples include the U.S. Cluster Mapping 7) that go beyond the illustration of countries’ par- project and the European Union Cluster Portal.3 ticipation in GVCs. These proposed methods allow Acknowledging that clusters and their impacts can for testing the economic relevance of specific mea- vary across regions in a country, the projects also sures and examining the drivers and effects of GVC take into account the subnational perspective.4 participation. The quantifications in this part of the book can be analyzed at the aggregate and sector levels. The GVC Participation Using Data on Trade in level of detail depends on data availability. Some Value Added of the most sophisticated measures are available at The previous section discussed how to use gross fairly aggregate sector levels, whereas most of the less import and export flows, as well as production data, sophisticated measures are available for narrowly to gather some initial insights into a country’s par- defined industries. Because aggregate trends may ticipation in GVCs. Gross trade flows can be decom- hide important developments in underlying indus- posed in various ways. The most obvious distinction tries, it is suggested to use a range of tools of differ- is between domestic and foreign value added (see ent sophistication levels, which combined reveal a chapter 4 for more details). As stated in the intro- general overview of how countries fare in GVCs and duction, because in GVCs countries import inputs provide the ability to zoom into specific issues. to export them after processing, what matters is the value addition generated in the country. This section Buying and Selling Sides delves into how much of the gross flows represent value addition in the country of interest (chapter 2 A key role of GVCs in industrial and economic illustrates the relevance of this exercise in the exam- development is boosting the competitiveness of the ple of Bulgaria). exports of low- and middle-income countries by Addressing this question requires moving facilitating the combination of foreign technology beyond traditional trade data. New databases have with the countries’ labor, capital, and technology. greatly facilitated this task—particularly the World Imports are important for competitiveness. A coun- Input-Output Database (WIOD), created by a con- try’s ability to participate in GVCs depends as much sortium of 11 institutions; the Organisation for on its capacity to import world-class inputs effi- Economic Co-operation and Development–World ciently as the country’s capacity to export. A country Trade Organization’s Trade in Value Added data- cannot become a major exporter in GVCs without base;5 the United Nations Conference on Trade and first becoming a successful importer of intermediate Development–EORA GVC database; and the World imports, because imported intermediate inputs con- Bank Export of Value Added database, which is tain foreign technology. based on the Global Trade Analysis Project database This section suggests ways to identify the (appendixes G and H). Drawing on these databases, extent to which countries source—domestically What Do Imports and Exports Say about GVC Participation? 63 or internationally—the intermediates they use in variety of countries (right side), indicating China’s exporting, which will provide a first indication of specialization in final goods exports. their participation in GVCs. The section then shows Compared with China, the selling patterns are ways to quantify the domestic value added embodied different for Poland (figure 3.6). Poland’s participa- in countries’ exports. tion is on par on the buying side (left side). It buys Import to export (I2E) patterns are a useful start- from a variety of countries, mainly Germany, Russia, ing point. Figures 3.4 through 3.9 illustrate I2E on and China (in decreasing order) and other European the sales and buying sides for six important actors in partners, but sells a higher I2E share primarily to GVCs: Japan, China, Poland, Germany, Mexico, and Germany and other European markets. This finding the United States, respectively. The sales and buying underscores the importance of the regional dimen- patterns for I2E for each country are normalized by sion, particularly for the European GVCs. Mexico’s the country’s exports (on the sales side) or imports dependence on inputs from the United States (on the buying side). The result shows where the emerges clearly (figure 3.7): almost 12 percent of country sources the intermediates it uses to export, Mexico’s I2E originates from the United States. The as well as where it sells the intermediates used in its other important suppliers (in decreasing order) are partners’ exports. China, Japan, Germany, Canada, and Korea. In the figures, each graph has two matched sides. Germany’s I2E (figure 3.8) and that of the United The left side shows the country’s I2E buying pat- States (figure 3.9) are more similar to that of Japan. tern—that is, the share of its exports made up of imported intermediates from the partners in the list. In both cases, the countries sell domestic parts and The right side shows the country’s bilateral exports components that are then embodied in many other of I2E trade as a share of its total exports. For each countries’ exports. The most important buyers for partner, the shares are shown for 2011 (size of circle), the United States are (in decreasing order) China, while the position on the x-axis illustrates the evolu- Canada, and Mexico (figure 3.9, right side). German tion of buying and selling patterns (growth between intermediates, by contrast, feed into Chinese and 1995 and 2011 in percent), respectively. Countries most of its regional partners’ exports (figure 3.8, right are ranked in decreasing order for 2011; tiny partners side). On the buying side, the United States seems to have been removed to improve readability. rely very little on foreign countries. I2E inputs are The distinction between the buying and selling mainly sourced from Canada, China, and Mexico (in sides of I2E is clear in Japan’s I2E with China (fig- decreasing order) (figure 3.9, left side). Germany’s ure 3.4). China is a very important destination for most important sources of I2E (in decreasing order) Japanese parts and components that are embodied are the United States, France, United Kingdom, in other countries’ exports. In 2011, more than 10 Russian Federation, and Italy (figure 3.8, left side). percent of Japan’s exports consisted of intermediate goods sold to China and subsequently embodied in Chinese exports—shown on Japan’s sales side of I2E Notes (right side). In figure 3.5, which shows China’s I2E, 1. Note that the OECD STAN Bilateral Trade Database the exact same flow of GVC intermediates is shown by industry and end-use category (BTDxE), which is based as China’s I2E on the buying side as almost 5 per- on the Broad Economic Categories, distinguishes the fol- cent (it is normalized by China’s imports instead of lowing end-use categories: intermediate goods, household consumption, capital goods, and mixed end-use (personal Japan’s exports). computers, passenger cars, personal phones, precious Japan and China have noticeably different par- goods, packed medicines, and miscellaneous). ticipation in GVCs. According to the I2E measure, 2. World Bank (2014). Japan imports I2E goods from few countries, mostly 3. For the U.S. Cluster Mapping project, see http:// China, the United States, Saudi Arabia, and Australia www.clustermapping.us; for the EU Cluster Portal, see (in decreasing order of importance) (figure 3.4, left http://ec.europa.eu/growth/smes/cluster/index_en.htm. side); but a large fraction of Japan’s exports are of 4. The U.S. Cluster Mapping project has the infor- mation broken down to the county level; the EU Cluster parts embodied in other countries’ exports (right Portal shows the data at the Nomenclature of Territorial side). China sources its I2E inputs from many more Units for Statistics (NUTS-2) level. countries, mainly from Japan; United States; Korea; 5. Access the World Bank’s Trade in Value Added Taiwan, China; and Germany (figure 3.5, left side), database at http://data.worldbank.org/data-catalog but sells a much lower share of I2E goods to a wide /export-value-added. 64 Making Global Value Chains Work for Development Figure 3.4. Buying and Selling Patterns: Japan, 1995 and 2011 1 China China 1 2 United States Korea, Rep. 2 3 Saudi Arabia Taiwan, China 3 4 Australia United States 4 5 Indonesia Thailand 5 6 Korea, Rep. Malaysia 6 7 Russian Federation Germany 7 8 Germany Singapore 8 9 Malaysia Mexico 9 10 Taiwan, China Canada 10 11 United Kingdom United Kingdom 11 12 France Russian Federation 12 13 Canada France 13 14 Thailand Vietnam 14 15 Brazil India 15 16 South Africa Italy 16 17 Switzerland Australia 17 18 Italy Hong Kong SAR, China 18 19 Singapore Indonesia 19 20 India Czech Republic 20 Import partners, ranked from highest to lowest percentage, 2011 Export partners, ranked from highest to lowest percentage, 2011 21 Chile Spain 21 22 Philippines Philippines 22 23 Spain Poland 23 24 Vietnam Switzerland 24 25 Brunei Darussalam Hungary 25 26 Mexico Ireland 26 27 Hong Kong SAR, China Turkey 27 28 Netherlands Belgium 28 29 Sweden Sweden 29 30 Norway Netherlands 30 31 Belgium Brazil 31 32 Ireland Slovak Republic 32 33 Finland Austria 33 34 Austria Norway 34 35 Denmark Denmark 35 36 New Zealand Israel 36 37 Israel South Africa 37 38 Poland Luxembourg 38 39 Argentina Finland 39 40 Colombia Saudi Arabia 40 41 Turkey Chile 41 42 Hungary Argentina 42 43 Czech Republic Portugal 43 44 Portugal New Zealand 44 45 Luxembourg Colombia 45 46 Romania Costa Rica 46 47 Greece Romania 47 48 Slovak Republic Slovenia 48 49 Costa Rica Greece 49 50 Bulgaria Tunisia 50 51 Slovenia Cambodia 51 52 Cambodia Estonia 52 53 Iceland Bulgaria 53 54 Tunisia Brunei Darussalam 54 55 Estonia Latvia 55 56 Lithuania Iceland 56 57 Latvia Lithuania 57 58 Croatia Malta 58 59 Cyprus Croatia 59 60 Malta Cyprus 60 11 10 9 8 7 6 5 4 3 2 1 0 –1 –2 –2 –1 0 1 2 3 4 5 6 7 8 9 10 11 Import partners’ growth, 1995–2011 (%) Export partners’ growth, 1995–2011 (%) Import share, 2011 Export share, 2011 Trade share, 2011 (%) 13 10 8 6 4 2 1 Source: Adapted from Baldwin and Lopez-Gonzalez 2013. 0.0002 What Do Imports and Exports Say about GVC Participation? 65 Figure 3.5. Buying and Selling Patterns: China, 1995 and 2011 Japan United States 1 1 United States Korea, Rep. 2 2 Korea, Rep. Japan 3 3 Taiwan, China Germany 4 4 Germany Taiwan, China 5 5 Australia Mexico 6 6 Saudi Arabia Malaysia 7 7 Russian Federation France 8 8 Malaysia Thailand 9 9 Brazil United Kingdom 10 10 Indonesia India 11 11 India Italy 12 12 France Singapore 13 13 United Kingdom Canada 14 14 Singapore Russian Federation 15 15 Italy Hong Kong SAR, China 16 16 Thailand Vietnam 17 17 Canada Czech Republic 18 18 Hong Kong SAR, China Spain 19 19 Switzerland Poland 20 20 Chile Australia 21 21 Import partners, ranked from highest to lowest percentage, 2011 Export partners, ranked from highest to lowest percentage, 2011 Philippines Hungary 22 22 South Africa Indonesia 23 23 Spain Turkey 24 24 Sweden Sweden 25 25 Netherlands Brazil 26 26 Mexico Belgium 27 27 Austria Switzerland 28 28 Belgium Norway 29 29 Vietnam Netherlands 30 30 Norway Ireland 31 31 Ireland Denmark 32 32 Israel Philippines 33 33 Finland Austria 34 34 Turkey South Africa 35 35 Denmark Slovak Republic 36 36 Poland Israel 37 37 Argentina Finland 38 38 Colombia Chile 39 39 Czech Republic Cambodia 40 40 New Zealand Argentina 41 41 Hungary Luxembourg 42 42 Portugal Saudi Arabia 43 43 Slovak Republic Portugal 44 44 Romania Romania 45 45 Costa Rica New Zealand 46 46 Greece Colombia 47 47 Brunei Darussalam Slovenia 48 48 Luxembourg Greece 49 49 Bulgaria Tunisia 50 50 Slovenia Bulgaria 51 51 Tunisia Estonia 52 52 Cambodia Costa Rica 53 53 Lithuania Croatia 54 54 Estonia Lithuania 55 55 Latvia Latvia 56 56 Croatia Iceland 57 57 Iceland Cyprus 58 58 Malta Brunei Darussalam 59 59 Cyprus Malta 60 60 12 10 8 6 4 2 0 –2 –2 0 2 4 6 8 10 12 Import partners’ growth, 1995–2011 (%) Export partners’ growth, 1995–2011 (%) Import share, 2011 Export share, 2011 Trade share, 2011 (%) 13 10 8 6 4 2 1 0.0002 Source: Adapted from Baldwin and Lopez-Gonzalez 2013. 66 Making Global Value Chains Work for Development Figure 3.6. Buying and Selling Patterns: Poland, 1995 and 2011 1 Germany Germany 1 2 Russian Federation Czech Republic 2 3 China United Kingdom 3 4 Italy France 4 5 France Italy 5 6 United States Hungary 6 7 United Kingdom China 7 8 Czech Republic Russian Federation 8 9 Sweden Sweden 9 10 Japan Spain 10 11 Spain Slovak Republic 11 12 Norway United States 12 13 Netherlands Austria 13 14 Austria Denmark 14 15 Korea, Rep. Switzerland 15 16 Belgium Belgium 16 17 Switzerland Canada 17 18 Slovak Republic Norway 18 19 Hungary Netherlands 19 20 Denmark Ireland 20 Import partners, ranked from highest to lowest percentage, 2011 Export partners, ranked from highest to lowest percentage, 2011 21 Finland Turkey 21 22 India Finland 22 23 Turkey Romania 23 24 Ireland Korea, Rep. 24 25 Brazil Luxembourg 25 26 Taiwan, China Mexico 26 27 Romania Lithuania 27 28 Saudi Arabia Taiwan, China 28 29 Canada Latvia 29 30 Lithuania Japan 30 31 Indonesia India 31 32 Australia Singapore 32 33 Israel Malaysia 33 34 Greece Estonia 34 35 Thailand Slovenia 35 36 Malaysia Thailand 36 37 South Africa Bulgaria 37 38 Portugal Portugal 38 39 Singapore Israel 39 40 Slovenia South Africa 40 41 Luxembourg Greece 41 42 Mexico Australia 42 43 Chile Brazil 43 44 Bulgaria Vietnam 44 45 Argentina Croatia 45 46 Vietnam Iceland 46 47 Hong Kong SAR, China Indonesia 47 48 Latvia Argentina 48 49 Colombia Tunisia 49 50 Philippines Hong Kong SAR, China 50 51 Iceland Saudi Arabia 51 52 Tunisia Chile 52 53 Estonia Cyprus 53 54 Croatia Philippines 54 55 Cyprus New Zealand 55 56 Malta Malta 56 57 New Zealand Colombia 57 58 Cambodia Costa Rica 58 59 Costa Rica Cambodia 59 60 Brunei Darussalam Brunei Darussalam 60 22 20 18 16 14 12 10 8 6 4 2 0 –2 –2 0 2 4 6 8 10 12 14 16 18 20 22 Import partners’ growth, 1995–2011 (%) Export partners’ growth, 1995–2011 (%) Import share, 2011 Export share, 2011 Trade share, 2011 (%) 13 10 8 6 4 2 1 0.0002 Source: Adapted from Baldwin and Lopez-Gonzalez 2013. What Do Imports and Exports Say about GVC Participation? 67 Figure 3.7. Buying and Selling Patterns: Mexico, 1995 and 2011 1 United States United States 1 2 China Canada 2 3 Japan China 3 4 Germany Spain 4 5 Canada Korea, Rep. 5 6 Korea, Rep. Germany 6 7 Italy India 7 8 Taiwan, China Japan 8 9 Brazil United Kingdom 9 10 France France 10 11 United Kingdom Italy 11 12 Spain Taiwan, China 12 13 Russian Federation Thailand 13 14 Malaysia Brazil 14 15 Saudi Arabia Singapore 15 16 India Malaysia 16 17 Switzerland Colombia 17 18 Australia Australia 18 19 Chile Chile 19 20 Indonesia Belgium 20 Import partners, ranked from highest to lowest percentage, 2011 Export partners, ranked from highest to lowest percentage, 2011 21 Thailand Russian Federation 21 22 Singapore Portugal 22 23 Sweden Ireland 23 24 Philippines Switzerland 24 25 Netherlands Argentina 25 26 Colombia Costa Rica 26 27 Austria Turkey 27 28 South Africa Czech Republic 28 29 Argentina Hungary 29 30 Poland Poland 30 31 Ireland Netherlands 31 32 Belgium Sweden 32 33 Norway Norway 33 34 Israel Vietnam 34 35 Czech Republic Austria 35 36 Hong Kong SAR, China Israel 36 37 Denmark Denmark 37 38 Costa Rica South Africa 38 39 Finland Indonesia 39 40 Turkey Luxembourg 40 41 Vietnam Hong Kong SAR, China 41 42 Portugal Slovak Republic 42 43 Hungary Finland 43 44 Romania Saudi Arabia 44 45 Greece Philippines 45 46 Slovak Republic Tunisia 46 47 New Zealand Romania 47 48 Slovenia Greece 48 49 Luxembourg New Zealand 49 50 Bulgaria Slovenia 50 51 Brunei Darussalam Bulgaria 51 52 Tunisia Iceland 52 53 Lithuania Estonia 53 54 Estonia Croatia 54 55 Croatia Latvia 55 56 Latvia Lithuania 56 57 Cambodia Cambodia 57 58 Iceland Malta 58 59 Malta Cyprus 59 60 Cyprus Brunei Darussalam 60 24 22 20 18 16 14 12 10 8 6 4 2 0 –2 –2 0 2 4 6 8 10 12 14 16 18 20 22 24 Import partners’ growth, 1995–2011 (%) Export partners’ growth, 1995–2011 (%) Import share, 2011 Export share, 2011 Trade share, 2011 (%) 13 10 8 6 4 2 1 0.0002 Source: Adapted from Baldwin and Lopez-Gonzalez 2013. 68 Making Global Value Chains Work for Development Figure 3.8. Buying and Selling Patterns: Germany, 1995 and 2011 1 United States China 1 2 France France 2 3 United Kingdom Italy 3 4 Russian Federation United Kingdom 4 5 Italy Switzerland 5 6 China United States 6 7 Switzerland Austria 7 8 Netherlands Czech Republic 8 9 Austria Poland 9 10 Spain Spain 10 11 Poland Belgium 11 12 Japan Hungary 12 13 Belgium Sweden 13 14 Czech Republic Korea, Rep. 14 15 Norway Russian Federation 15 16 Sweden Luxembourg 16 17 India Ireland 17 18 Brazil Netherlands 18 19 Korea, Rep. Mexico 19 20 Ireland Denmark 20 Import partners, ranked from highest to lowest percentage, 2011 Export partners, ranked from highest to lowest percentage, 2011 21 Denmark Canada 21 22 Hungary Japan 22 23 Canada Slovak Republic 23 24 Finland Taiwan, China 24 25 Turkey Malaysia 25 26 Slovak Republic Turkey 26 27 Romania Finland 27 28 South Africa Singapore 28 29 Saudi Arabia India 29 30 Singapore Thailand 30 31 Australia Norway 31 32 Taiwan, China Portugal 32 33 Luxembourg Romania 33 34 Portugal Brazil 34 35 Mexico Slovenia 35 36 Indonesia Australia 36 37 Slovenia Israel 37 38 Malaysia South Africa 38 39 Greece Bulgaria 39 40 Thailand Hong Kong SAR, China 40 41 Chile Greece 41 42 Hong Kong SAR, China Vietnam 42 43 Colombia Indonesia 43 44 Israel Saudi Arabia 44 45 Argentina Chile 45 46 Philippines Argentina 46 47 Bulgaria Tunisia 47 48 Iceland Estonia 48 49 Vietnam Croatia 49 50 Lithuania Latvia 50 51 Tunisia Lithuania 51 52 Croatia Philippines 52 53 Latvia New Zealand 53 54 Cyprus Malta 54 55 New Zealand Colombia 55 56 Estonia Iceland 56 57 Malta Cyprus 57 58 Costa Rica Costa Rica 58 59 Brunei Darussalam Cambodia 59 60 Cambodia Brunei Darussalam 60 10 8 6 4 2 0 –2 –2 0 2 4 6 8 10 Import partners’ growth, 1995–2011 (%) Export partners’ growth, 1995–2011 (%) Import share, 2011 Export share, 2011 Trade share, 2011 (%) 13 10 8 6 4 2 1 0.0002 Source: Adapted from Baldwin and Lopez-Gonzalez 2013. What Do Imports and Exports Say about GVC Participation? 69 Figure 3.9. Buying and Selling Patterns: United States, 1995 and 2011 1 Canada China 1 2 China Canada 2 3 Mexico Mexico 3 4 Japan Germany 4 5 Germany Ireland 5 6 Saudi Arabia Korea, Rep. 6 7 United Kingdom United Kingdom 7 8 Russian Federation France 8 9 Korea, Rep. Japan 9 10 France Singapore 10 11 Brazil Taiwan, China 11 12 Italy Malaysia 12 13 India India 13 14 Colombia Italy 14 15 Switzerland Netherlands 15 16 Taiwan, China Spain 16 17 Ireland Switzerland 17 18 Australia Belgium 18 19 Norway Thailand 19 20 Singapore Luxembourg 20 Import partners, ranked from highest to lowest percentage, 2011 Export partners, ranked from highest to lowest percentage, 2011 21 Spain Russian Federation 21 22 Indonesia Brazil 22 23 Israel Sweden 23 24 Netherlands Australia 24 25 Malaysia Denmark 25 26 Chile Norway 26 27 Sweden Israel 27 28 Belgium Hungary 28 29 Thailand Turkey 29 30 South Africa Chile 30 31 Austria Poland 31 32 Hong Kong SAR, China Czech Republic 32 33 Poland Hong Kong SAR, China 33 34 Denmark Austria 34 35 Argentina Finland 35 36 Turkey Vietnam 36 37 Philippines Argentina 37 38 Vietnam South Africa 38 39 Finland Philippines 39 40 Czech Republic Costa Rica 40 41 Greece Indonesia 41 42 Hungary Greece 42 43 Portugal Colombia 43 44 Costa Rica Saudi Arabia 44 45 New Zealand Portugal 45 46 Romania Slovak Republic 46 47 Luxembourg New Zealand 47 48 Slovak Republic Romania 48 49 Slovenia Slovenia 49 50 Bulgaria Iceland 50 51 Lithuania Tunisia 51 52 Brunei Darussalam Bulgaria 52 53 Cambodia Estonia 53 54 Croatia Cambodia 54 55 Tunisia Croatia 55 56 Estonia Cyprus 56 57 Latvia Lithuania 57 58 Iceland Latvia 58 59 Cyprus Brunei Darussalam 59 60 Malta Malta 60 10 8 6 4 2 0 –2 –2 0 2 4 6 8 10 Import partners’ growth, 1995–2011 (%) Export partners’ growth, 1995–2011 (%) Import share, 2011 Export share, 2011 Trade share, 2011 (%) 13 10 8 6 4 2 1 0.0002 Source: Adapted from Baldwin and Lopez-Gonzalez 2013. 70 Making Global Value Chains Work for Development References Türkiye’nin Konumu.” TUSIAD-Koc Universitesi Ekonomik Arastirma Forumu Calisma Raporlari Serisi. Athukorala, Premachandra. 2010. “Production Networks “The Transformation of International Production and Trade Patterns in East Asia: Regionalization or Linkages and Turkey.” TUSIAD-Koc University, Globalization?” Working Paper Series on Regional Economic Research Forum Research Reports. http:// Economic Integration 56, Asian Development Bank, eaf.ku.edu.tr/sites/eaf.ku.edu.tr/files/eaf_rp_1101.pdf. Manila. UN (United Nations). 2002. “Classification by Broad Baldwin, Richard, and Javier Lopez-Gonzalez. 2013. Economic Categories.” United Nations, Department of “Supply-Chain Trade: A Portrait of Global Patterns Economic and Social Affairs, Statistical Papers Series M and Several Testable Hypotheses.” NBER Working No. 53, Rev. 4. Paper 18957, National Bureau of Economic Research, World Bank. 2013. “Romania Western Region Com- Cambridge, MA. petitiveness Enhancement and Smart Specialization Santoni, Gianluca, and Daria Taglioni. Forthcoming. Final Report.” Paper presented at the World Bank Network Analysis Using U.S. Benchmark Input-Output and the Agency for Regional Development–Region Data. Washington, DC: The World Bank. West, Romania closing-project conference, “Advisory Sturgeon, Timothy J., and Olga Memedovic. 2011. Services on Competitiveness Enhancement and Smart “Mapping Global Value Chains: Intermediate Goods Specialization Policies in the West Region,” Timișoara, Trade and Structural Change in the World Economy.” Romania. Working Paper 10/2010, United Nations Industrial ———. 2014. “Turkey: Trading up to High Income: Development Organization, Vienna. Country Economic Memorandum.” Report 82307, Taymaz, Erol, Ebru Voylvoda, and Kamil Yilmaz. 2011. World Bank, Washington, DC. “Uluslararasi Üretim Zincirlerinde Dönüș üm ve Chapter 4 BUYER-RELATED MEASURES Introduction High shares of intermediate imports in total imports are common for these countries, reflect- The global value chain (GVC) participation mea- ing the importance of primary commodities in sures in this chapter focus on the sourcing side in the import basket of dynamic and rapidly industrial- GVCs. The key questions are, where are a country’s izing economies (figure 4.2). The far lower shares of exports made, and where is their value created? This intermediates of imports in electrical and electronic is the buyer’s perspective, as shown on the left side in equipment support this finding. The difference also figures 3.4 to 3.9 in chapter 3. suggests that sector analysis of this indicator is very important. The indicator can be computed using the Intermediates in Gross Imports concepts and classifications discussed in chapter 3 (BEC, value chain category, customized trade, and The first indicator is the share of intermediates by other classifications). Broad Economic Categories (BEC) in gross imports, The measure does not reveal whether the inputs as shown for high-income countries (HICs) and low- are used domestically or exported. The following and middle-income countries (LMICs) between 1996 measures address this topic, focusing on imported and 2012 (figure 4.1).1 Two patterns stand out. First, inputs or foreign value added embodied in gross the share of intermediates in gross imports rose most exports. in HICs, which reflects the global fragmentation of production and the offshoring of tasks from HICs to LMICs. Second, the share is substantially higher Imported Inputs Embodied in Gross Exports for LMICs, reaching 71 percent in 2012 against 61 I2E in Intermediate or Total Imports percent in HICs. That phenomenon probably occurs because LMICs specialize in processing intermedi- The indicator importing to exports (I2E) in inter- ates for subsequent export (which could be shown by mediates measures the buyer’s intermediate imports further separating processed from primary interme- embodied in its gross exports as a percentage of the diates), whereas final goods had a larger share of the buyer’s total intermediate imports.2 A very similar imports of HICs. measure was presented in chapter 3. The only dif- In this chapter, the measure is illustrated for ference in this chapter, compared with chapter 3, Malaysia and selected peer countries, including Chile, is that the denominator here includes intermediate China, Mexico, the Philippines, Poland, the Republic rather than total imports, resulting in smaller shares. of Korea, Singapore, South Africa, Thailand, and Figure 4.3 reports the measure for the selected group Vietnam (box 4.1). The countries are all middle to of countries, as well as for Germany, Japan, and the high performers in GVCs. United States. 71 72 Making Global Value Chains Work for Development Figure 4.1. Intermediate Imports, 1996–2012 trade paradigm, by contrast, would reward opening import access as a key ingredient of strategies that 75 reinforce a country’s competitiveness. 70 I2E by Source Country % of gross imports Imported intermediate inputs contain foreign tech- 65 nology. A key role of GVCs in industrial develop- ment is to boost the competitiveness of the exports 60 of LMICs by facilitating the combination of foreign technology with their labor, capital, and technology. 55 A very useful indicator of GVC participation, then, is the origin of the imported inputs embodied in a country’s gross exports, as was illustrated in the left- 50 hand side of figures 3.4 to 3.9 in chapter 3. 1996 1998 2000 2002 2004 2006 2008 2010 2012 Low- and middle-income countries Distinguishing between Domestic and Foreign Value High-income countries Added in Imports Source: Adapted from the United Nations Comtrade database. Note: Calculations are based on the Broad Economic Categories classification Although imported inputs embodied in gross exports (appendix B). is a very useful indicator for a country’s participation in GVCs, it does not distinguish between the foreign and domestic contents of the value of the imported Survey data confirm that, in the aggregate, a large inputs. Imported inputs may contain domestic value share of the imports of goods and services is used as added that is exported to a foreign location, pro- inputs for exports. According to a recent business cessed, and re-imported. An example is U.S. inputs survey of 250 lead firms and suppliers in the agrifood imported by Canada and used to produce Canadian sector, more than 80 percent of businesses in GVCs exports.4 The imported inputs from the United perceive imports of goods and services as being States may already contain Canadian value added important or critical for their exports.3 This finding from upstream processes in which the United States challenges the mercantilist approach to trade and has imported inputs from Canada. Those products trade negotiations, in which the focus is on market are called “re-imports” from a Canadian perspective access and the reciprocity of concessions. The new and “re-exports” from a U.S. perspective (figure 4.4). BOX 4.1. Choice of Comparator Countries International comparisons set countries’ competitiveness in con- ment, competitors with export baskets of similar composition, or text; therefore, several examples in this part of the book discuss neighboring countries. The World Bank Group has constructed a results for a subset of countries. The challenge is to select compa- world database of peer countries to inform export competitive- rable countries. The following methodology was adopted. ness analyses. The matching of countries with their peers is avail- Malaysia was selected as an example. Malaysia was an early able at https://mec.worldbank.org/buildercompare#comparator adopter of global value chains, but is a middle performer, with countries. Benchmark countries for each of the 121 countries in areas of excellence and areas in which it needs to catch up (see the data set are determined based on the following five indica- World Bank 2014). To put the country’s performance in context, tors: population, human capital, physical capital, gross domestic the most appropriate peer countries were identified (listed in the product per capita, and export basket composition. Countries are main text). ranked by similarity in decreasing order, enabling the analyst Selection of the peer countries was based on a data-driven to select the most similar ones as benchmark countries. Figure method, informed by the judgments of country experts. The B4.1.1 shows the global network of countries. method identifies countries similar in size or economic develop- (Box continues next page) Buyer-Related Measures 73 BOX 4.1. (continued) Figure B4.1.1. Country Positioning in the Global Economic Space United States Japan Canada United Kingdom France Germany Netherlands Italy Switzerland Spain Austria Norway Belgium Singapore Qatar Ireland Sweden Korea, Rep. United Arab Emirates Kuwait Australia Israel Russian Federation Denmark Portugal Finland Czech Republic Bahrain Saudi Arabia Luxembourg Hungary Greece Poland Turkey China Slovak Republic Libya Philippines Malaysia Indonesia Slovenia India Iceland Cyprus Croatia Kazakhstan Algeria Estonia Romania Gabon Vietnam Nepal Yemen, Rep. Sudan Lithuania Ukraine Mozambique New Zealand Côte d’Ivoire Uganda Bulgaria Egypt, Arab Rep. Burundi Malta Thailand Pakistan Cambodia Latvia Venezuela, RB Senegal Albania Syrian Arab Republic Rwanda Mexico Botswana Ghana Tanzania South Africa Kenya Malawi Tunisia Congo, Rep. Armenia Kyrgyz Republic Morocco Gambia, The Jordan Zambia Benin Niger Cameroon Togo Sri Lanka Mauritania Moldova Brazil Namibia Colombia Lesotho Zimbabwe Central African Republic Argentina Mauritius Dominican Republic Peru Fiji Ecuador Guatemala Barbados Chile El Salvador Costa Rica Panama Honduras Uruguay Paraguay Maldives Belize Jamaica Bolivia Guyana Source: Adapted from Raj (2014). Re-importing and re-exporting can be quite impor- Figure 4.2. Countries’ Integration in GVCs: Share of Intermediate tant for some industries and countries. Imports in Gross Imports and Electrical and Electronics, 2009 and 2012 Chile 2009 Value Added in Gross Exports 2012 China 2009 Value-added trade statistics can be used to single out 2012 the domestic or foreign value added embodied in Korea, Rep. 2009 2012 exports. The advantage of these data is that they help Malaysia 2009 determine where things are actually made. For exam- 2012 ple, the data can be very important for quantifying Mexico 2009 2012 the impact of GVC participation on jobs. Because Philippines 2009 about three-quarters of domestic value added comes 2012 from labor, value-added trade statistics roughly show Poland 2009 2012 where the export-linked jobs are located, by country Singapore 2009 and sector. 2012 Gross exports are decomposed in various ways, South Africa 2009 2012 the most obvious being between domestic and for- Thailand 2009 eign value added (see the bars in figure 4.5). A first 2012 set of indicators looks at the value added embod- Vietnam 2009 2012 ied in gross exports. The first-pass indicator simply distinguishes between domestic and foreign value 0 10 20 30 40 50 60 70 80 90 100 added, usually expressed as a share of gross exports. Percent The second pass digs deeper into where the domestic Total Electrical and electronics value added is actually created. This indicator breaks Source: Adapted from the United Nations Comtrade database. 74 Making Global Value Chains Work for Development Figure 4.3. Intermediate Imports Embodied in Exports and Electrical and clarifies differences between the following key and Optical Equipment, Selected Countries, 2009 and 2011 concepts: Chile 2009 2011 • Gross exports. The total value of exports as shown China 2009 in traditional trade and balance-of-payments sta- 2011 Germany 2009 tistics (for goods and services) captures the value 2011 added embodied in the production of the good Japan 2009 or service exported, as well as all domestically 2011 Korea, Rep. 2009 sourced and imported inputs embodied in the 2011 good or service. Malaysia 2009 • Direct domestic value added embodied in exports. 2011 Mexico 2009 Gross exports minus domestically sourced 2011 (EXGR_IDC), re-imported (EXGR_RIM), and Philippines 2009 foreign inputs (EXGR_FVA) capture the true 2011 Poland 2009 sector-specific domestic value added of exports 2011 (EXGR_DDC). This information is important Singapore 2009 in an environment in which global production 2011 South Africa 2009 is fragmented across countries. For example, a 2011 business process outsourcing (BPO) service in Thailand 2009 India contains telecommunications services from 2011 United States 2009 local providers and foreign owners of satellites. 2011 The measure EXGR_DDC nets out domestic and Vietnam 2009 foreign inputs and captures the true value added 2011 generated in India’s BPO sector. 0 10 20 30 40 50 60 70 80 90 100 • Total domestic value added of exports. For the % of total intermediate imports total, the direct domestic value added of exports is added to the value added of the inputs sourced Total Electrical and optical equipment domestically (indirect domestic content of gross Source: Adapted from the Organisation for Economic Co-operation and Development–World Trade Organization Trade in Value Added database. exports [EXGR_IDC]) and the value added of re- imported inputs (EXGR_RIM). In the BPO exam- out the total domestic value added into (1) domes- ple, the measure captures the value added of the tic value added in the particular sector (autos, in the BPO service plus the value of the domestic satel- example), (2) domestic value added in upstream sec- lites used as input in the underlying telecommuni- tors supplying the sector with parts, and (3) domes- cations service, but the measure does not include tic value added in intermediates first shipped abroad the value of the foreign-owned satellite input. The for further processing and then re-imported. measure captures the full domestic value added of Combining information from input-output tables an exported service or good. Quantitatively, how- with information from trade flows makes such com- ever, in most countries re-imported inputs tend to putations possible. Four data sets for trade in value be very small, so the sum of the direct and indi- added are discussed in chapter 3 and appendixes G rect value added contribution is highly correlated and H. The data sets make it possible to assess the with the total domestic value added embodied in domestic or foreign value-added content of coun- exports. tries’ gross exports. Model-based computations enable quantifying To exemplify the relevance of these concepts, gross the value added of a given export (good or service). exports (of goods and services) are decomposed in If information is available on the geographic ori- the four measures, for the same countries (figure gin of the inputs, the value added specific to inputs 4.7). The share of foreign value added is lowest in produced domestically can be quantified, follow- commodity exporters and HICs. The lower the ratio ing an iterative process (figure 4.6). The schematic is, the higher is the domestic value-added content in in figure 4.6 distinguishes between domestic value gross exports, and thus the lower is the importance added (green) and foreign value added (yellow), of I2E. Commodity goods often are exported in their Buyer-Related Measures 75 Figure 4.4. Re-imports and Re-exports in Supply Chain Trade CANADA U.S. re-exports 1 2 2 1 U.S. re-imports UNITED STATES U.S. car industry Canadian car industry U.S. exports intermediates U.S. imports intermediates from 1 to Canada for processing. 1 Canada for processing. U.S. imports intermediates back U.S. exports intermediates back to 2 embodied in goods and services; 2 Canada embodied in goods and bilateral flow is normalized by all services; bilateral flow is normalized U.S. exports to Canada. by total U.S. imports from Canada. Source: Adapted from Baldwin and Lopez-Gonzalez 2013, figure 5. Note: U.S. re-exports are Canadian re-imports but with different normalizations. raw form or embody only low additional value from for most countries, in the sample in figure 4.7, that other sectors or from abroad, which explains the amount adds up to 1 percent of total gross exports high share of direct domestic value added. By con- only for China and Germany. trast, HICs tend to have a more diversified domestic supply base that requires them to rely less on foreign Foreign Value Added in Gross Exports imports. China’s gross exports rely more on indirect value added (37 percent) in supplying sectors. Subtracting the total domestic value-added of By contrast, most other emerging countries and exports measure as a percentage of gross exports Germany, Japan, and the United States (44, 48, and from 1 yields the foreign value added embodied in 52 percent, respectively) strongly depend on direct gross exports as a percentage of gross exports. This domestic value added. That countries relatively mar- figure captures the country’s GVC position as a user ginal to GVCs (such as Chile and South Africa) and of foreign value added in its exports. For the world as countries at their core (Germany, Japan, and the a whole, only 20 percent of gross exports constitutes United States) post a low I2E ratio suggests that the value that was added in a foreign country.5 measure—in isolation—is not indicative of failure or By plotting the foreign value added in a country’s success. Its relevance needs to be assessed in combi- gross exports as a percentage of gross exports, three nation with the wider range of measures presented patterns stand out (figure 4.8): in this book. Reflecting the fact that the domestic value added • Almost all countries saw their numbers increase in re-imported intermediates is generally very low between 1995 and 2011. The expansion was 76 Making Global Value Chains Work for Development Figure 4.5. Decomposition of Gross Exports particularly strong in emerging countries, such as Cambodia, Hungary, Korea, Poland, and Turkey. Domestic Directly in the auto sector value added • The share is lower for large countries—especially the manufacturing giants—but Germany and even China are almost twice as integrated interna- tionally as the United States. • The share rises to very high levels in the small- est countries, such as Ireland, Luxembourg, and Singapore. In sectors supplying the auto sector The difference between gross exports and value- added exports can be extremely stark for some CAR products. China’s exports of iPhones, for example, INDUSTRY EXPORTS include less than 10 percent of Chinese value added, In re-imported intermediates whereas Norway’s exports of oil contain almost 100 percent of Norwegian value added. At the national Foreign In various industries value added level, however, the difference is moderate for most countries, with some standouts: Korea has a remark- ably high foreign content for a country of its size and level of industrialization. Australia is a standout for its low number, but that surely reflects its reliance on primary product exports, which are naturally high in local content. How do the four key measures compare for individual countries? Figure 4.9 shows the share Source: Adapted from Baldwin and Lopez-Gonzalez 2013. of imported intermediate inputs in total imports, Figure 4.6. Quantifying the Value Added of Gross Exports Indirect domestic Direct value added value added of EXGR_IDC EXGR_DDC of exports intermediate inputs Re-imported Value added EXGR_RIM Gross intermediate of intermediate inputs inputs exports Foreign value added of intermediate EXGR_FVA inputs Note: Foreign value added is highlighted in yellow and domestic value added is highlighted in green. Buyer-Related Measures 77 Figure 4.7. Decomposition of Gross Exports, Selected Figure 4.8. Foreign Value Added in a Country’s Gross Exports, 1995 Countries, 2011 and 2011 54.31 25.24 0.03 20.42 Luxembourg Chile Hungary 29.88 36.95 0.96 32.21 Slovak Republic China Czech Republic 43.71 29.69 1.00 25.60 Ireland Germany Taiwan, China 48.29 36.62 0.36 14.72 Singapore Japan Korea, Rep. 33.29 24.61 0.36 41.75 Malaysia Korea, Rep. Bulgaria Thailand 36.06 22.97 0.27 40.70 Malaysia Malta Cambodia 43.97 23.97 0.32 31.74 Mexico Vietnam Slovenia 49.14 26.97 0.04 23.84 Estonia Philippines Finland 37.37 29.88 0.21 32.54 Belgium Poland Iceland 38.68 19.01 0.15 42.16 Portugal Singapore Denmark 43.91 36.56 0.03 19.50 Tunisia South Africa Poland 37.90 22.90 0.13 39.07 China Thailand Mexico 51.71 32.51 0.72 15.06 Sweden United States Latvia Austria 39.63 24.03 0.07 36.28 Vietnam Costa Rica Spain 0 10 20 30 40 50 60 70 80 90 100 Italy % of gross exports Turkey Germany Direct domestic value added content Israel Indirect domestic value added content from France domestic intermediates Greece Romania Re-imported domestic value added content India Foreign value added content Lithuania Philippines Source: Adapted from the Organisation for Economic Co-operation and Canada Development–World Trade Organization Trade in Value Added database. United Kingdom Switzerland Cyprus the share of I2E in total imports, the foreign value- Hong Kong SAR, China added content in gross exports as a percentage of Chile Croatia gross exports, and the domestic value added in gross Netherlands exports as a percentage of gross exports for 2009. South Africa Norway New Zealand Full Decomposition of Value Added by Sector and United States Japan Source Country Australia Argentina The decomposition of gross exports illustrated in Russian Federation figures 4.5 and 4.6—and applied to specific countries Indonesia Brazil in figures 4.7 to 4.9—can be used further to iden- Colombia tify the source of value added by sector and country. Brunei Darussalam Saudi Arabia The decomposition addresses where and in which 0 10 20 30 40 50 60 industries the value added that makes up a coun- % of gross exports try’s gross exports is produced. Figure 4.10 breaks 1995 2011 down US$10 million of Mexican auto exports. The Source: Adapted from the Organisation for Economic Co-operation and Development–World Trade first column shows that US$10 million in car exports Organization Trade in Value Added database. 78 Making Global Value Chains Work for Development Figure 4.9. Comparison of Four Buyer-Related Measures of GVC added). The value added in the rubber and plastic Participation, Selected Countries, 2009 intermediates comes from Mexico and the United Total States, because the Mexican rubber and plastics sec- Chile E&E tor imports some inputs from the United States. The China Total only figure that requires no further calculation is the E&E Mexican value added in the car sector. Germany Total Although the measures that have been discussed E&E Japan Total so far can be computed using any of the data- E&E bases illustrated in chapter 3, the Organisation for Korea, Rep. Total Economic Co-operation and Development–World E&E Trade Organization Trade in Value Added (TiVA) Malaysia Total database and WIOD are the only two international E&E Mexico Total data sets that provide decompositions that drill down E&E to the individual “country of origin by sector” level. Philippines Total With the full value-added breakdown of a country’s E&E gross exports to the world or to specific destinations, Poland Total E&E the information can be arranged in two basic ways: Singapore Total by source country or by source industry. The nec- E&E essary indicators in the TiVA database are listed in South Africa Total table 4.1; the left column shows the TiVA indicator E&E names; the right column provides their definitions. Thailand Total E&E Similar measures are available in WIOD. United States Total E&E Value Added in Gross Exports by Source Country Vietnam Total E&E The breakdown of foreign value added into source 0 10 20 30 40 50 60 70 80 90 100 countries or industries is useful from a buyer’s per- spective because it identifies which foreign sources Domestic value added in gross exports (EXGR_DVASH) add the most value to its exports. Figure 4.11 shows Foreign value added in gross exports (EXGR_FVASH) Imported intermediate inputs in total imports (IIM) the shares of foreign value added in gross exports as I2E in total imports (I2EB_IM) a percentage of gross exports for Thailand’s transport equipment sector by source region. The share of total Source: Adapted from the Organisation for Economic Co-operation and Development–World Trade Organization Trade in Value Added database. foreign value added in gross exports (sum across all Note: E&E = electrical and electronic equipment; GVC = global value chain; I2E = importing to export. source regions) rose from 48 percent in 1995 to 55 per- cent in 2011. Among the source countries, East Asian economies (China; Hong Kong SAR, China; Japan; from Mexico to the United States, in this example, Korea; and Taiwan, China) contributed approximately contains intermediates of iron and steel sourced 23 percent to Thailand’s gross exports of transport abroad worth US$3 million, intermediates of rubber equipment. East Asia’s contribution remained con- and plastics sourced in Mexico worth US$2.5 mil- stant over the period, whereas other source regions lion, and US$4.5 million of Mexican value added in expanded their shares, in particular, other regions, the car industry. The US$4.5 million consists of pay- the Association of Southeast Asian Nations without ments to productive factors in Mexico (wages, inter- Thailand, and South and Central America. est, dividends, and so on) and the Mexican firm’s For the subsample of countries discussed earlier, profit margin on the export sale. the source of value added is shown for total exports The iron and steel inputs embodied in the Mexican and electrical and optical (E&O) equipment (fig- cars exported to the United States come from the ure 4.12). Germany (from European Union–28, iron and steel sectors in Australia, Mexico, and the EU-28), Mexico (from North American Free Trade United States (US$1 million each). The imported Agreement, NAFTA), Malaysia (from East Asia), iron and steel has Mexican value added because the Poland (also from EU-28), Thailand (from East U.S. iron and steel industry uses Mexican inputs in Asia), and Vietnam (also from East Asia) source the its exports to Mexico (US$1 million of Mexican value highest shares of foreign value added from a single Buyer-Related Measures 79 Figure 4.10. Value-Added Trade: US$10 Million in Mexican Car Exports to the United States TiVA DECOMPOSITION VARIABLES BY SECTOR Mexico Domestic value EXGR_DDC $4.5M Car sector added in car sector $10M Domestic value EXGR_RIM $2M Rubber and MEXICAN added (all plastics GROSS EXPORTS intermediates) intermediates EXGR_IDC $1M United Foreign value EXGR_FVA $0.5M Iron and steel States added (all intermediates intermediates) $1M Australia $1M Source: Baldwin and Lopez-Gonzalez 2013. Note: TiVA = Trade in Value Added. Table 4.1. Indicators of Value Added Embodied in Gross Exports Indicator Definition EXGR Gross exports, by sector (US$ millions) EXGR_DVA Total domestic value added embodied in gross exports, by sector (US$ millions) EXGR_DVASH EXGR_DVA in EXGR, by sector (%) EXGR_DDC Direct (intrasector) domestic value added embodied in gross exports, by sector (US$ millions) EXGR_IDC Indirect (upstream) domestic value added embodied in gross exports, by sector (US$ millions) EXGR_RIM Re-imported domestic value added embodied in gross exports, by sector (US$ millions) EXGR_FVA Foreign value added embodied in gross exports, by sector and source country (US$ millions) EXGR_FVASH EXGR_FVA in EXGR, by sector and source country (%) Source: Adapted from the Organisation for Economic Co-operation and Development–World Trade Organization Trade in Value Added database. region—in the range of 12 to 17 percent. In the E&O is higher, representing about 26 percent of gross industry, reliance on a single source region is higher. exports (25 percent in E&O), mainly sourced from More than 35 percent of Vietnam’s gross exports con- the EU-28. tain value added originating from East Asian loca- tions, whereas the share ranges from 12 to 29 percent Value Added in Gross Exports by Source Industry in China, Malaysia, the Philippines, Singapore, and Thailand. The regional dimension of GVCs, in the Similarly, foreign value added can be disaggregated aggregate and in the E&O sector, emerges clearly. by source industries. For the U.S. E&O equipment Advanced countries (Germany, Japan, and the sector in 2011, the total share of foreign value added United States) post lower shares of foreign value embodied in exports as a percentage of gross exports added from any source region (figure 4.13). U.S. was 14.8 percent (figure 4.14). As might be expected, gross exports show the lowest foreign value-added intra-industry foreign value added contributed one content, a little over 15 percent (15 percent in E&O). of the largest shares (2.8 percent), followed by for- Japan’s foreign value-added share in gross exports is eign value added from mining and quarrying (1.9 also low, making up less than 15 percent (17 percent percent) and real estate, renting, and business activi- in E&O). Germany’s share of foreign value added ties (1.5 percent). Wholesale and retail trade; hotels 80 Making Global Value Chains Work for Development Figure 4.11. Foreign Value Added in Thailand’s Transport Equipment Sector Exports, by Source Region, 1995–11 30 28 26 24 22 20 % of gross exports 18 16 14 12 10 8 6 4 2 0 1995 2011 1995 2011 1995 2011 1995 2011 1995 2011 1995 2011 ASEAN w/o East Asia EU-28 NAFTA South and Other regions Thailand Central America Source: Adapted from the Organisation for Economic Co-operation and Development–World Trade Organization Trade in Value Added database. Note: ASEAN = Association of Southeast Asian Nations; EU = European Union; NAFTA = North American Free Trade Agreement. Figure 4.12. Foreign Value Added in Gross Exports, Total and and restaurants (2.8 percent) was the largest contrib- Electrical and Optical Equipment, by Source Region, Selected utor of foreign value added share in gross exports, Countries, 2011 pointing to the importance of services in GVCs. Chile Total Here we shift the focus to cross-country compari- E&O sons, to investigate the share of foreign value added China Total E&O embodied in gross exports as a percentage of gross Germany Total exports, by source sectors, in the E&O equipment E&O and chemicals industries. Figures 4.15 and 4.16 con- Japan Total E&O firm the importance of business services and whole- Korea, Rep. Total sale and retail trade, in particular for middle-income E&O countries and HICs. The most important source sec- Malaysia Total E&O tors are highlighted in yellow. Most foreign inputs are Mexico Total sourced from the same industry in the E&O equip- E&O ment sector (panel j in figure 4.15). For chemicals, Philippines Total E&O the main foreign source industry is upstream mining Poland Total and quarrying (panel c in figure 4.16). E&O Singapore Total E&O Length of Sourcing Chains South Africa Total E&O The I2E concept focuses on bilateral relations— Thailand Total E&O essentially, who a country is involved with in GVCs. United States Total Although it is informative for some issues, I2E misses E&O the “chain” aspect of GVCs. The Japanese compo- Vietnam Total E&O nents used in Chinese exports, for example, are likely to contain imported components from, say, Korea or 0 5 10 15 20 25 30 35 40 the United States. % of gross exports One measure that reflects such multi-country con- ASEAN EU-28 South and Central America siderations is the length of value chain sourcing. The East Asia NAFTA Other measure developed by Fally (2011) and applied to the Source: Adapted from the Organisation for Economic Co-operation and Development–World Trade TiVA data by DeBacker and Miroudot (2013) cap- Organization Trade in Value Added database. Note: ASEAN = Association of Southeast Asian Nations; E&O = electrical and optical; EU = European tures this attribute by looking at a recursive measure Union; NAFTA = North American Free Trade Agreement. of I2E on the sourcing side. To illustrate the concept, Buyer-Related Measures 81 it is useful to consider a simple value chain, in which Figure 4.13. Foreign Value Added in Gross Exports, Total and the automotive industry in Germany requires 40 Electrical and Optical Equipment, Selected Countries, 2011 cents of auto parts from Poland for each dollar’s Vietnam worth of automotives produced (the other 60 cents Malaysia being value added by the German-based automotive industry). Assuming the Polish parts are 100 percent Thailand Polish value added, the length of the sourcing chain is Mexico 1.4; that is, the German stage always counts as 1 stage China and the 40 percent value added in Poland counts as Poland 0.4 stage. The term “sourcing chain” can be somewhat misleading, in that it also takes into account the direct Korea, Rep. domestic value added contribution of the sector of Singapore interest, and does not only consider the upstream sec- South Africa tors from which inputs are sourced. Philippines The concept is recursive in more complex exam- ples. For instance, if the 40 percent Polish value added Germany included parts made in France, the sourcing chain Chile would be longer. If the French parts were 100 percent Japan made in France and made up 30 percent of the value of United States the Polish parts exports to Germany, the Polish parts chain would be 1.3 (1 for Poland and 0.3 for France). 10 20 30 40 50 60 70 Thus, the German auto industry’s chain length would % of gross exports be 1 + 0.4(1.3) = 1.52.6 The TiVA data provide a handy Total Electrical and optical equipment means for comparing the average number of produc- Source: Adapted from the Organisation for Economic Co-operation and Development–World Trade tion stages in a given industry and country. Organization Trade in Value Added database. Figure 4.14. Foreign Value Added in Gross Exports in the U.S. Electrical and Optical Equipment Sector, by Source Industry, 2011 Percent 2.76 C50T55: Wholesale and retail trade; hotels and restaurant 2.75 C30T33: Electrical and optical equipment 1.87 C10T14: Mining and quarrying 1.54 C70T74: Real estate, renting, and business activities 1.51 C27T28: Basic metals and fabricated metal products 0.98 C23T26: Chemicals and non-metallic mineral products Total foreign value 0.90 C60T64: Transport and storage, post, and telecommunication added in U.S. 0.68 C65T67: Financial intermediation electrical and optical equipment 0.41 C29: Machinery and equipment, nec exports 0.30 C40T41: Electricity, gas, and water supply 0.28 C75T95: Community, social, and personal services 0.20 C20T22: Wood, paper, paper products, printing, and publishing 0.16 C01T05: Agriculture, hunting, forestry, and fishing 0.14 C34T35: Transport equipment 0.10 C36T37: Manufacturing nec; recycling 0.08 C45: Construction 0.06 C17T19: Textiles, textile products, leather, and footwear 0.06 C15T16: Food products, beverages, and tobacco Source: Adapted from the Organisation for Economic Co-operation and Development–World Trade Organization Trade in Value Added database. Note: nec =not elsewhere classified. 82 Figure 4.15. Foreign Value Added in Gross Exports, Electrical and Optical Equipment, by Source Sector, Selected Countries, 2011 b. Agriculture, d. Food e. Textiles, textile f. Wood, paper, hunting, products, products, paper products, g. Chemicals h. Basic metals j. Electrical and forestry, and c. Mining and beverages, and leather, and printing, and and nonmetallic and fabricated i. Machinery and optical a.Total fishing quarrying tobacco footwear publishing mineral products metal products equipment, nec equipment Chile China Germany Japan Korea, Rep. Malaysia Mexico Philippines Poland Singapore South Africa Thailand United States Vietnam 0 20 40 60 80 0 1 2 3 0 5 10 15 20 0 1 2 3 0 1 2 3 0 1 2 3 0 5 10 15 20 0 5 10 15 20 0 1 2 3 0 5 10 15 20 Making Global Value Chains Work for Development % of gross exports o. Wholesale p.Transport and r. Real estate, s. Community, m. Electricity, and retail trade; storage, post renting, and social, and k. Transport l. Manufacturing gas, and water hotels and and telecommu- q. Financial business personal equipment nec; recycling supply n. Construction restaurants nication intermediation activities services Chile China Germany Japan Korea, Rep. Malaysia Mexico Philippines Poland Singapore South Africa Thailand United States Vietnam 0 1 2 3 0 1 2 3 0 1 2 3 0 1 2 3 0 5 10 15 20 0 5 10 15 20 0 5 10 15 20 0 5 10 15 20 0 1 2 3 % of gross exports Source: Adapted from the Organisation for Economic Co-operation and Development–World Trade Organization Trade in Value Added database. Note: The most important source sectors are in yellow and use a scale ranging from 0 to 20. Less relevant source sectors are in green and use a scale ranging from 0 to 3. nec = not elsewhere classified. Figure 4.16. Foreign Value Added in Gross Exports, Chemicals and Chemical Products, by Source Sector, Selected Countries, 2011 b. Agriculture, d. Food e. Textiles, textile f. Wood, paper, hunting, products, products, paper products, g. Chemicals h. Basic metals j. Electrical and forestry, and c. Mining and beverages, and leather, and printing, and and nonmetallic and fabricated i. Machinery and optical a.Total fishing quarrying tobacco footwear publishing mineral products metal products equipment, nec equipment Chile China Germany Japan Korea, Rep. Malaysia Mexico Philippines Poland Singapore South Africa Thailand United States Vietnam 0 20 40 60 80 0 1 2 3 0 5 10 15 20 25 0 1 2 3 0 1 2 3 0 1 2 3 0 5 10 15 20 25 0 1 2 3 0 1 2 3 0 1 2 3 % of gross exports o. Wholesale p.Transport and r. Real estate, s. Community, m. Electricity, and retail trade; storage, post renting, and social, and k. Transport l. Manufacturing gas, and water hotels and and telecommu- q. Financial business personal equipment nec; recycling supply n. Construction restaurants nication intermediation activities services Chile China Germany Japan Korea, Rep. Malaysia Mexico Philippines Poland Singapore South Africa Thailand United States Vietnam 0 1 2 3 0 1 2 3 0 1 2 3 0 1 2 3 0 5 10 15 20 25 0 5 10 15 20 25 0 1 2 3 0 5 10 15 20 25 0 1 2 3 Buyer-Related Measures % of gross exports Source: Adapted from the Organisation for Economic Co-operation and Development–World Trade Organization Trade in Value Added database. Note: The most important source sectors are in yellow and use a scale ranging from 0 to 25. Less relevant source sectors are in green and use a scale ranging from 0 to 3. The minimum value of the index is 1 when no intermediate inputs are used to produce a final good or service. nec = not elsewhere classified. 83 84 Making Global Value Chains Work for Development Figure 4.17. Length of Sourcing Chains, by Industry, 2008 TV and communication equipment Motor vehicles Basic metals Electrical machinery Other transport equipment Textiles, leather, and footwear Machinery and equipment, nec Rubber and plastics products Chemicals and chemical products Fabricated metal products Precision and optical instruments Food products and beverages Wood and products of wood Manufacturing nec; recycling Other non-metallic mineral products Office and computing machinery Construction Paper, printing, and publishing Electricity, gas, and water supply Petroleum products and nuclear fuel Transport and storage Hotels and restaurants Research and development Agriculture Post and telecommunications Other social and personal services Finance and insurance Computer and related activities Public administration Other business activities Health and social work Wholesale and retail trade Renting of machinery and equipment Mining and quarrying Education Real estate activities 0 0.50 1.00 1.50 2.00 2.50 3.00 3.25 Length International Domestic Total Source: Adapted from OECD Global Value Chains Indicators, May 2013. Note: The minimum value of the index is 1 when no intermediate inputs are used to produce a final good or service. nec = not elsewhere classified. Figure 4.17 shows the length of sourcing chains other transport equipment; and textiles, leather, and averaged across all countries in the TiVA data set, dif- footwear. Services, on average, have shorter value ferentiating between the domestic and international chains, with notable exceptions, such as transport and production stages. The six industries with the longest storage. chains are television and communication equipment; The measure can be useful in several ways. For motor vehicles; basic metals; electrical machinery; instance, an increase in GVC length over time Buyer-Related Measures 85 suggests that the value chain has become more com- refers to the respective sections of the book, in which plex, with stages done in more countries. The mea- the measures are introduced. For more information sure also can show the relative complexity of the on the underlying data, see appendixes G and H. GVCs of which a country is part—associating long The measures presented so far cover only trade GVCs with complex ones. in tasks (goods and services). Chapter 1 shows that GVCs are a multidimensional phenomenon that also Buyer Dimension: Summary involves the flow of factors of production. From a buyer’s perspective, several measures can be analyzed Table 4.2 summarizes the main measures of the to complement the task-based assessment. Although buyer dimension. More indirect measures—such as a wider set of examples can be found in table A.1 the share of intermediates in gross imports based in appendix A, the following are some of the key on informed classifications—are easily available, indicators: whereas more direct quantifications of a country’s performance in GVCs—such as the full decomposi- • Wages. High wages are likely to be associated with tion of a country’s gross exports—are available for buyers that are also final producers, close to final only a limited set of countries. The table indicates demand, and able to generate high value added. whether the measures are available for goods tasks, Low wages are likely to be associated with buy- services tasks, or both, as well as the level of analysis ers that are mainly assemblers or are involved in (country, sector, or firm). For easy reference, the table activities with little transformation. Table 4.2. Summary of the Main Buyer-Related Measures Level of Where to Measure Task analysis Data sources find in book? Share of intermediates in Goods Country, sector Gross import data (UN Comtrade, BACI, WITS), Chapter 4, gross imports categorized using informed classifications (BEC, p. 71 parts and components, technical classifications) Share of intermediates in Goods Country, sector, Gross import data (customs firm-level), categorized Chapter 4, gross imports, range of firm using informed classifications (BEC, parts and p. 71 imports, bundle of imported components, technical classifications) products, and countries Imported inputs embodied Goods and/or Country, sector International I-O data (WIOD, World Bank Export Chapter 4, in exports, as percentage of services Value Added database) p. 71 gross imports Imported inputs embodied Goods and/or Country, sector, International I-O data (WIOD) Chapter 4, in exports, as percentage of services source country p. 72 gross imports and by source country Share of foreign value added Goods and/or Country, sector International I-O data (WIOD, TiVA, World Bank Chapter 4, in gross exports services Export Value Added database) p. 75 Share of foreign value added Goods and/or Country, sector, International I-O data (WIOD, TiVA) Chapter 4, in gross exports by source services source country p. 78 country Share of domestic value Goods and/or Country, sector International I-O data (WIOD, TiVA, World Bank Chapter 4, added in gross exports services Export Value Added database) p. 74 Multinational’s share of inputs Goods and/or Country, sector, Enterprise surveys or other firm-level surveys Chapter 6, from domestic suppliers in services firm p. 113 total inputs Domestic producer’s share Goods and/or Country, sector, Enterprise surveys or other firm-level surveys Chapter 6, of imported inputs in total services firm p. 115 imported inputs Length of sourcing chains Total Country, sector International I-O data (WIOD, TiVA); U.S. I-O table Chapter 4, and gross trade data (Comtrade, BACI, WITS) p. 80 Note: BACI = International Trade database by CEPII; BEC = Broad Economic Categories; I-O = input-output; TiVA = Trade in Value Added; WIOD = World Input-Output Database; WITS = World Integrated Trade Solution. 86 Making Global Value Chains Work for Development • Ideas. Buyers that aim to create value addition 3. OECD-WTO 2013. from GVC participation are likely to be active 4. Baldwin and Lopez-Gonzalez 2013. 5. Baldwin and Lopez-Gonzalez 2013. buyers of international patents and foreign tech- 6. Technically, the upstream-length index—which nology licenses. The existence or absence of a Fally (2011) denotes as “N” because it can be inter- sound framework for intellectual property is preted as the number of upstream stages—is defined as likely to be associated with buyers specializing in N cs = 1 + ∑µbr N br . Here, Ncs is the GVC weighted num- higher value-added tasks—unless GVCs happen br ber of stages for country “c” in sector “s.” It is 1 (for the predominantly or exclusively within the boundar- processing done in country c) plus a weighted sum of ies of multinational corporations. the length of the GVC for inputs that country c’s sector • Investments. Buyers are likely to observe foreign s uses. The weights are the value shares of inputs from direct investment (FDI) inflows and high FDI partner nation b’s sector r (that total is defined for each stocks in sectors and products of GVC specializa- nation, but the national subscript is omitted to reduce clutter in the formula). The formula focuses on inter- tion, sometimes also associated with FDI in down- national stages, but the same formula can be applied to stream or upstream sectors. If a country is a strong parts sourced domestically by sector s, which allows the buyer overall, support for upstream services is also analysis to quantify the length of the domestic chain too. likely to be of interest to foreign investors, includ- ing banks and companies in distribution, trans- port, telecommunications, and so on. References Baldwin, Richard, and Javier Lopez-Gonzalez. 2013. “Supply-Chain Trade: A Portrait of Global Patterns Notes and Several Testable Hypotheses.” NBER Working Paper 18957, National Bureau of Economic Research, 1. The share of imported intermediates in gross Cambridge, MA. IMINPcs DeBacker, Koen, and Sébastien Miroudot. 2013. “Mapping imports, IIM, is defined as IIM cs = where IMINP Global Value Chains.” OECD Trade Policy Papers No. IMGRcs 159, Organisation for Economic Co-operation and denotes intermediate imports and IMGR gross imports. Development Publishing, Paris. The measure is easily available; intermediates can, for Fally, Thibault. 2011. “On the Fragmentation of Production instance, be detected using the BEC classification in in the U.S.” University of Colorado, Boulder. http:// appendix B. It enables analysis of the extent to which a sciie.ucsc.edu/14AIEC/Fragmentation_Fally.pdf. buyer imports intermediate inputs. Categorizations of the OECD and WTO (Organisation for Economic Co- numerator informed by technical considerations, such as operation and Development and World Trade those discussed previously and reported in appendixes C to Organization). 2013. Aid for Trade at a Glance 2013: E, are also possible. Intermediates include energy imports. Linking to Value Chains. Paris: Organisation for 2. I2EB_IMINP: Economic Co-operation and Development. I 2EBcs Raj, Anasuja. 2014. “Peer Countries.” World Bank, I 2EB _ IMINPcs = where I2EB denotes the buy- IMINPcs Washington, DC. er’s intermediate imports embodied in its gross exports World Bank. 2014. “Boosting Trade Competitiveness.” and IMINP the buyer’s total intermediate imports in a Malaysia Economic Monitor, June 2014, Washington, sector. DC. Chapter 5 SELLER-RELATED MEASURES Introduction according to one of the classifications. Second, the The global value chain (GVC) participation indica- measure does not indicate whether intermediate out- tors discussed in chapter 4 focus on the sourcing side put is used domestically or exported. in GVCs. The touchstone question was: where are a The share of intermediates in gross exports takes country’s exports made, and where is the value cre- the exporting perspective into account. Again, the ated? Key questions for the selling dimension are: intermediates can be identified according to one of who are the ultimate customers for a country’s value the informed classifications discussed in chapter 3 and added, and to what countries is the country export- appendixes B–F. In some countries, such as Malaysia, ing its value added? Australia, for example, exports intermediates dominate the export basket (see figure iron ore to China, but part of that product ends up 3.1 in chapter 3). This measure quantifies GVC partic- in the United States and Germany rather than China. ipation, measured by the importance of intermediates That is the seller’s perspective, as shown on the right- in the export basket of a country or, put differently, hand side in figures 3.4 to 3.9 in chapter 3. whether the country supplies products used as inter- mediates for further processing in other countries. Over time, the measure for specific countries and sec- Intermediates in Output or Gross Exports tors—and relative to peers—can provide a first-pass A first basic measure of the seller’s involvement in indication of whether a country has become a more the production of inputs, as opposed to final goods, important supplier in GVCs. The same limitations is the share of intermediates in gross output.1 This plague this measure as for the share of intermediates measure is akin to the share of intermediates in in gross imports, which is discussed in chapter 4. The gross imports presented in chapter 4, but focuses on measures do not reveal the use of the inputs in the domestic production rather than trade. The mea- export destination—whether they are used domes- sure quantifies participation in GVCs, measured tically or processed and exported to third countries. by the importance of intermediates in a country’s The measure that addresses this caveat focuses on the overall production.2 Interesting variations of this role of foreign inputs or value added for third coun- measure can be constructed by focusing on different tries’ exports. types of intermediate inputs, using one or more of Before turning to that topic, however, there are the informed classifications illustrated in chapter 3 two stylized facts about intermediate exports world- and reported in appendixes B to E or by identifying wide that are worth highlighting. First, the share of production by clusters, as discussed in chapter 3 and intermediates in gross exports has increased over appendix F.3 However, the measure suffers from two the past one and a half decades (figure 5.1), which main limitations. First, it is useful if it is quantified reflects greater global fragmentation of production, at a very disaggregated level. Therefore, to be useful, but may also be the result of rising commodity prices, the measure requires survey, industry, or firm data especially in the 2000s. Second, the share is slightly sufficiently disaggregated to disentangle production higher for low- and middle-income countries, which 87 88 Making Global Value Chains Work for Development Figure 5.1. Intermediate Exports, 1996–2012 difference is that this indicator accounts only for the seller’s intermediates that are domestically produced, 66 whereas in the previous case the intermediates could 64 also contain some foreign value added. The portion of a country’s exports used as imported inputs in the 62 buyers’ exports is formalized as DVA3EX.5 DVA3EX % of gross exports can be related to the total gross exports in a country 60 to obtain DVA3EX_EX. From 1995 to 2011, almost all countries increased 58 their GVC participation; the few exceptions were 56 Cambodia, Croatia, and Luxembourg (figure 5.2). The increase was partly the mechanical effect of 54 longer value chains and increasing specialization in GVCs worldwide. Nevertheless, cross-country dif- 52 ferences exist. The leading countries are all natural 1996 1998 2000 2002 2004 2006 2008 2010 2012 resource exporters, including Brunei Darussalam, Low- and middle-income countries Norway, the Russian Federation, and Saudi Arabia, High-income countries whose gross exports consisted of 38 to 43 percent Source: Adapted from the United Nations Comtrade database. domestic value added embodied in third countries’ Note: Calculations are based on the Broad Economic Categories classification (see appendix B). exports. Japan and the United States—two large, non-natural resource–intensive countries—did not come very far (at approximately 33 and 25 percent, suggests that high-income countries export more respectively). On the other side of the spectrum are final goods, although the difference has become nar- countries with very low export shares of domestic rower since the late 1990s. value added embodied in third countries’ exports, ranging from only 12 percent in Cambodia and Luxembourg, to 14 to 16 percent in China, Croatia, I2E Trade in Gross Exports Ireland, Mexico, Thailand, Turkey, and Vietnam. The indicator importing to exports (I2E) in gross The aggregate measure tends to give a somewhat exports measures intermediates sold by a country biased picture, because natural resource–intensive to a buyer for use in the buyer’s exports (I2E from countries, especially oil and gas exporters, unsurpris- the buyer’s perspective) as a percentage of the seller’s ingly show the largest shares. It therefore makes sense gross exports.4 The measure can be computed for the to show this indicator excluding mining and quarry- economy as a whole and for individual sectors. ing and coke, refined petroleum products, and nuclear Figures 3.4 to 3.9 in chapter 3 show the measure fuel in the numerator (figure 5.3). The indicator (the selling side) for Japan, China, Poland, Mexico, for total gross exports remains in the denominator. Germany, and the United States, respectively. The The picture is somewhat different from the previ- figures show, for example, that in 2011 over 10 per- ous figure, with natural resource–intensive export- cent of Japan’s gross exports consisted of intermedi- ers ranked behind Japan and closer to non-natural ate goods sold to China and subsequently embodied resource–intensive countries. Despite the correc- in Chinese exports (figure 3.4, right side). Similarly, tion, Brunei Darussalam, Chile, Norway, Russia, and almost 6 percent of Poland’s gross exports con- Saudi Arabia still exhibit one of the largest shares: 29 tained I2E goods sold to Germany and embodied in to 30 percent of their total gross exports is made up German exports (figure 3.8, right side). of domestic value added subsequently embodied in third countries’ exports. Among high-income coun- tries, Austria; Iceland; Japan; the Republic of Korea; Domestic Value Added in Gross Exports of the Netherlands; Switzerland; Taiwan, China; and Third Countries the United States show relatively large shares of 24 to This indicator is very similar to the previous one, as it 32 percent. indicates the contribution of domestically produced Another useful way to look at the data is through intermediates to exports in third countries. The only changes in global market share. To the extent that such Seller-Related Measures 89 Figure 5.2. Domestic Value Added Embodied in Third changes can be considered an indicator of competi- Countries’ Exports, 1995 and 2011 tiveness,6 measuring a country’s global market share changes in domestic value added embodied in third Brunei Darussalam Saudi Arabia countries’ exports can be considered a measure of Norway increasing comparative advantage in GVCs. Figure 5.4 Russia Federation Japan illustrates this measure for the selection of emerging Chile countries used in the previous examples: Chile, China, Indonesia Colombia Korea, Malaysia, Mexico, the Philippines, Poland, Australia Singapore, South Africa, Thailand, and Vietnam. All Netherlands market shares are expressed in percent of global value Philippines South Africa added embodied in third countries’ exports. All coun- Iceland tries in the sample—except South Africa—saw their Switzerland United States market share increase. The most spectacular increase Austria was for China, which jumped from 1.3 percent to 6.6 United Kingdom Sweden percent in global value-added market share. Romania Similar indicators can be constructed at the sec- Brazil Denmark tor level. Figure 5.5 shows examples of the chemicals Germany and chemical products and the electrical and optical Taiwan, China (E&O) equipment sectors. In chemicals and chemi- Latvia Belgium cal products (highlighted in green), global market Poland shares are broadly unchanged over the period— Hong Kong SAR, China Finland except for China, which had a market share of less Slovenia than 1 percent in 1995 and almost 5 percent in 2011; Lithuania France Japan, whose value-added market share diminished Italy from 7.3 to 4.3 percent; and Germany, whose value- Slovak Republic added market share shrunk by almost 40 percent Korea, Rep. Estonia during this period. In E&O (highlighted in yellow), Singapore the increase was spectacular for the Asian countries Malaysia Spain (except Japan and Singapore), including China, Czech Republic Korea, Malaysia, and Vietnam, which managed to Israel India gain a small but valuable market share (0.35 per- Canada cent), given that it was starting from close to nothing. Greece Tunisia Figure 5.6 illustrates another indicator of com- Portugal petitiveness and specialization patterns—the revealed Malta comparative advantage (RCA) indicator for the E&O Cyprus Costa Rica and chemical sectors. The figure shows the way in Bulgaria which countries’ comparative positions changed Hungary New Zealand between 1995 and 2011. Instead of using gross Argentina exports, the measure is constructed using domestic Vietnam Ireland value added embodied in gross exports. This method China is a more accurate representation of comparative Thailand advantage, because it nets out foreign value added Turkey Mexico imported into the country (see chapter 4). As in tradi- Croatia tional RCA measures, a country is said to have a com- Luxembourg Cambodia parative advantage in the sector if the RCA measure is greater than 1 (to the right of the RCA line). All Asian 0 5 10 15 20 25 30 35 40 45 countries in the sample—Vietnam aside—reveal a % of gross exports comparative advantage in E&O goods. Chile, Poland, 1995 2011 South Africa, the United States, and Vietnam, by con- Source: Adapted from the Organisation for Economic Co-operation and trast, post a relatively weak specialization in E&O. Development–World Trade Organization Trade in Value Added database. From a dynamic point of view, the Philippines shows 90 Making Global Value Chains Work for Development Figure 5.3. Domestic Value Added Embodied in Third Countries’ Exports Excluding Mining and Quarrying and Coke, Refined Petroleum Products, and Nuclear Fuel, 2011 lam russa ion Cambo Luxe % of abia rat ei Da ede Japan gross exports r mbo di A Brun nF dia Me 35 urg Sau sia Cro xico Rus Vie atia ile es ay tna 30 Ch pin rw Tu Th m rke ip No ail an hil ia y Ch d 25 P es on ds Irel ina Ind e rlan New and 20 Neth Zea and land Icel Arge 15 ralia ntina Aust Cypru erland s 10 Switz Bulgaria Austria 5 Canada Romania Hungary United States Taiwan, Chi Tunisia na South ica Africa Costa R Germ a Malt any l Swe uga den Port Uni ece ted Gre Be Kin ia lgi gdo Ind ia Co u m m ys lom ala bia M l ae Po Isr lan Kong lic Ho ub d ng ain Lat ep a hR Sp Slo via oni Den re ec ublic ven Est ep. Finlan apo SA Cz Lithuania Brazil Italy France mar R, ia Korea, R Sing k Rep Ch k d ina Slova Source: Adapted from the Organisation for Economic Co-operation and Development–World Trade Organization Trade in Value Added database. Note: Gross exports include the country’s total exports. a sizable reorientation toward E&O, with an RCA that in gross exports in total, as well as in E&O and chem- went from 2 in 1995 to almost 3 in 2011. In chemi- icals by destination region. The European Union cals, only five countries report an RCA greater than 1: (EU), East Asia, and North American Free Trade Germany, Korea, Singapore, Thailand, and the United Agreement (NAFTA) countries differ in importance States. These results mesh with results found through- for those sectors. out the indicators—that Southeast Asia as a region is In the aggregate, as well as for E&O, the most important for value-added trade in E&O goods, as important trend has been the shift of Asian countries opposed to chemicals, for which value added is con- from NAFTA countries toward intraregional demand. centrated in other world regions. For example, in 2011, Thailand’s overall exports in From a seller’s perspective, identifying which for- value-added terms went mostly to destinations in eign sources most demand its exported value added East Asia (32 percent), other destinations (28.4 per- is useful. Domestic value added can be separated by cent), and ASEAN countries (16.8 percent), and only destination countries. Figures 5.7 to 5.9 focus on the marginally to the EU and NAFTA (10.5 and 10.4 per- geographic side and show, for the countries in the cent, respectively). The Asian comparator countries preceding sample, the shares of domestic value added display similar patterns. Seller-Related Measures 91 Figure 5.4. Domestic Value Added Embodied in Third Figure 5.5. Domestic Value Added Embodied in Third Countries’ Exports, Global Market Share, Selected Countries’ Exports, Global Market Share, Chemicals Countries, 1995 and 2011 and Chemical Products and Electrical and Optical Equipment, Selected Countries, 1995 and 2011 China Chile 1995 Korea, Rep. 2011 China 1995 Singapore 2011 Germany 1995 Malaysia 2011 Japan 1995 Mexico 2011 Korea, Rep. 1995 Poland 2011 Thailand Malaysia 1995 2011 South Africa Mexico 1995 2011 Chile Philippines 1995 2011 Philippines Poland 1995 2011 Vietnam Singapore 1995 2011 0 1 2 3 4 5 6 7 South Africa 1995 2011 Percent Thailand 1995 1995 2011 2011 United States 1995 Source: Adapted from the Organisation for Economic Co-operation and 2011 Development–World Trade Organization Trade in Value Added database. Vietnam 1995 2011 0 4 8 12 16 20 24 Mexico’s total domestic value added in gross Percent exports in 2011 went, in large part, to NAFTA and Chemicals and chemical products Poland’s mostly to the EU. Finally, Germany’s value Electrical and optical equipment added was exported to the EU first (52 percent), then Source: Adapted from the Organisation for Economic Co-operation and to other destinations (22 percent), and only margin- Development–World Trade Organization Trade in Value Added database. ally to countries in East Asia and NAFTA (11.7 and 10.4 percent, respectively). U.S. exports of domestic value added were more spread out: EU (24 percent), added in final demand by country of origin. The con- NAFTA (23 percent), other destinations (23 per- cepts can be understood by revisiting the simplified cent), East Asia (19 percent), and South and Central Mexican car industry example, which assumes that America and ASEAN (6 and 5 percent, respectively). car production uses only two types of intermediate inputs: (1) iron and steel, and (2) rubber and plastics. Who Are the Ultimate Consumers of a Figure 5.10, panel a, shows the observed trade flow Country’s Value Added? Value Added in between three countries (Australia, Mexico, and the United States). Australia and the United States export Final Domestic Demand iron and steel to Mexico, and the United States also The previous discussion focused on direct links exports rubber and plastics to Mexico. Mexico exports between a country and the buyers of its exports. final cars to the United States. The implicit trade The Trade in Value Added (TiVA) and World Input- flows—which keep track of the ultimate consum- Output databases can also provide an understanding ers of value added—are shown in figure 5.10, panel of the final consumers of a country’s value-added b. Australia’s iron and steel sector is really exporting activities. The TiVA database contains a set of readily to U.S. consumers, not Mexican ones, and the United available indicators that focus on the share of value States is “exporting” iron and steel to itself. 92 Making Global Value Chains Work for Development Figure 5.6. RCA in Chemicals and Chemical Products Figure 5.7. Domestic Value Added in Gross Exports, by and Electrical and Optical Equipment, 1995 and 2011 Destination Region, 1995 and 2011 Chile 1995 Chile 1995 2011 2011 China 1995 China 1995 2011 2011 Germany 1995 Germany 1995 2011 2011 Japan 1995 Japan 1995 2011 2011 Korea, Rep. 1995 Korea, Rep. 1995 2011 2011 Malaysia 1995 Malaysia 1995 2011 2011 Mexico 1995 Mexico 1995 2011 2011 Philippines 1995 Philippines 1995 2011 2011 Poland 1995 Poland 1995 2011 2011 Singapore 1995 Singapore 1995 2011 2011 South Africa 1995 South Africa 1995 2011 2011 Thailand 1995 Thailand 1995 2011 2011 United States 1995 United States 1995 2011 2011 Vietnam 1995 Vietnam 1995 2011 2011 0 0.5 1.0 1.5 2.0 2.5 3.0 0 20 40 60 80 RCA Percent Chemicals and chemical products ASEAN East Asia EU-28 NAFTA Electrical and optical equipment South and Central America Other RCA Line Source: Adapted from the Organisation for Economic Co-operation and Source: Adapted from the Organisation for Economic Co-operation and Development–World Trade Organization Trade in Value Added database. Development–World Trade Organization Trade in Value Added database. Note: ASEAN = Association of Southeast Asian Nations; EU = European Union; Note: RCA line = 1. RCA = revealed comparative advantage. NAFTA = North American Free Trade Agreement. This sort of information is useful in several ways. The FDDVA measure looks at the sales side of For example, Australian steel exports are affected by this “final consumer” measure. A corresponding any U.S. trade barriers against Mexican autos. Thus, measure on the buying side also exists—the foreign U.S.–Australian trade relations are not solely depen- value added embodied in domestic final demand dent on their bilateral trade. Moreover, the informa- (FDFVA)—which looks at where the value is added tion may be helpful in understanding how changes in in a particular country’s final demand. This measure demand patterns—say, caused by an economic crisis can be interpreted as “imports in value added.”7 or a simple recession—can affect a country’s exports. The value added in final demand is a func- The domestic value added embodied in foreign tion of trade openness and GVC integration. As final demand—FDDVA in the TiVA data set—shows expected, the domestic value added embodied in how sectors export value through direct final exports foreign final demand as a share of a country’s gross and indirect exports of intermediates by way of other domestic product (GDP), FDDVA_GDP, is largest in countries to foreign final consumers (households, small, open economies, such as Brunei Darussalam, charities, government, and investment). The measure Luxembourg, and Singapore. The indicator is also illustrates the full upstream impact of final demand large in countries whose inputs are exported to in foreign markets on domestic output. other countries, including countries in Eastern Seller-Related Measures 93 Figure 5.8. Domestic Value Added in Gross Exports Figure 5.9. Domestic Value Added in Gross Exports in Electrical and Optical Equipment, by Destination in Chemicals and Chemical Products, by Destination Region, 1995 and 2011 Region, 1995 and 2011 Chile 1995 Chile 1995 2011 2011 China 1995 China 1995 2011 2011 Germany 1995 Germany 1995 2011 2011 Japan 1995 Japan 1995 2011 2011 Korea, Rep. 1995 Korea, Rep. 1995 2011 2011 Malaysia 1995 Malaysia 1995 2011 2011 Mexico 1995 Mexico 1995 2011 2011 Philippines 1995 Philippines 1995 2011 2011 Poland 1995 Poland 1995 2011 2011 Singapore 1995 Singapore 1995 2011 2011 South Africa 1995 South Africa 1995 2011 2011 Thailand 1995 Thailand 1995 2011 2011 United States 1995 United States 1995 2011 2011 Vietnam 1995 Vietnam 1995 2011 2011 0 20 40 60 80 100 0 20 40 60 80 Percent Percent ASEAN East Asia EU-28 NAFTA ASEAN East Asia EU-28 NAFTA South and Central America Other South and Central America Other Source: Adapted from the Organisation for Economic Co-operation and Source: Adapted from the Organisation for Economic Co-operation and Development–World Trade Organization Trade in Value Added database. Development–World Trade Organization Trade in Value Added database. Note: ASEAN = Association of Southeast Asian Nations; EU = European Union; Note: ASEAN = Association of Southeast Asian Nations; EU = European Union; NAFTA = North American Free Trade Agreement. NAFTA = North American Free Trade Agreement. Europe and Southeast Asia. The indicator is small in Germany, the United Kingdom, France, Italy, and the large countries, such as Japan and the United States United States, in that order. The markets of China, (figure 5.11). Germany, Japan, Mexico, the Philippines, Singapore, Foreign value added embodied in domestic final and Vietnam rely most on the United States for final demand as a share of GDP, FDFVA_GDP, is highest demand of its value-added exports, with Mexico a in countries well integrated in GVCs and dependent clear outlier (64 percent). China is the most impor- on imported inputs from other countries, especially tant final consumer for some countries’ value added, in Eastern Europe and Southeast Asia (figure 5.12). including Chile, Korea, Malaysia, South Africa, and The indicator also allows for identification of the Thailand, while Japan also matters strongly as final ultimate buyer of a country’s value added, which consumer for all countries in the sample except figure 5.13 shows for the set of selected countries Germany and Poland. used in the previous examples. The main final buy- This trend is different from that observed when ers of the value added of all countries in the data set considering gross exports, in which the importance are China and the United States. The exception is of the United States for many of the countries in Poland, whose value added is ultimately consumed in the sample was less pronounced for overall exports. 94 Making Global Value Chains Work for Development Figure 5.10. The Ultimate Consumers of a Country’s Export Value Added a. Explicit trade flows b. Implicit trade flows Observed trade flows Value added trade flows (solid green arrows) (solid yellow arrows) Australia Mexico United States Australia Mexico United States $1 $1 I&S I&S I&S I&S I&S I&S 5 $0. R&P R&P R&P R&P R&P R&P $3 $2. 5 CAR CAR CAR CAR CAR CAR $1 $1 $10 $2 $4.5 United States $1 $0.5 Source: Adapted from Baldwin and Lopez-Gonzalez 2013, 10. Note: The car exports from Mexico to the United States (worth US$10 million) contain intermediates of iron and steel worth US$3 million and intermediates of rubber and plastics worth US$2.5 million (dashed green arrows). For more detail, see figure 4.10 (chapter 4). I&S = iron and steel; R&P = rubber and plastics. Using the example of Malaysia, China is responsible For example, countries specialized in very for 16 percent of Malaysia’s domestic value added upstream activities produce raw materials—say, iron embodied in foreign final demand. Japan and the ore—or the intangibles at the start of the production United States (12 percent each) are also important process—say, research and design. Countries that final consumers of Malaysia’s domestic value added specialize in, for example, final assembly or customer embodied in foreign final demand. Using gross services will be very close to final demand. Countries exports, the United States accounts, instead, for 9 in activities at the center of the value chain focus percent of total Malaysian exports, only the fourth on the standardized, labor-intensive manufacturing largest export partner behind Singapore (14 per- jobs. However, these assumptions do not hold for cent), China (13 percent), and Japan (12 percent). every type of production and GVC. They are true for Table 5.1 summarizes the indicators of value some sectors and value chains (such as electronics, added embodied in final demand available in the particularly in East Asia), but not for others (such as TiVA database. high-end furniture manufacturing, where acquisi- tion of raw materials and design usually take place at the same stage of production).8 Length of Selling Chains: Keeping in mind the caveats, the analysis that Distance to Final Demand follows asks how to assign a specific country to a GVC length on the buying side of importing to category of upstreamness. The analysis draws on export measures the number of upstream stages for the work of Chor (2014), and applies it to measure a specific sector in a specific country. Another use- where a country (Malaysia in our example) is posi- ful measure looks at a similar concept on the sales tioned along the global production line. The analy- side. That measure gauges the “upstreamness” of a sis assesses whether Malaysia’s exports tend to be in country’s exports—roughly, the number of down- relatively upstream industries, near the start of the stream stages between the country’s producers and production process and far from final demand, or in final consumers. Antràs and others (2012) call it the downstream industries, closer to the final consumer. “distance to final demand.” Chor (2014) calculates a measure of the production Seller-Related Measures 95 Figure 5.11. Domestic Value Added Embodied in Figure 5.12. Foreign Value Added Embodied in Foreign Final Demand, 2011 Domestic Final Demand, 2011 Brunei Darussalam Vietnam Luxembourg Bulgaria Singapore Slovak Republic Ireland Cambodia Malaysia Tunisia Saudi Arabia Latvia Vietnam Malta Hungary Cyprus Estonia Estonia Iceland Slovenia Malta Luxembourg Slovak Republic Thailand Thailand Hungary Czech Republic Malaysia Bulgaria Czech Republic Taiwan, China Hong Kong SAR, China Slovenia Lithuania Hong Kong SAR, China Ireland Switzerland Taiwan, China Norway Iceland Cambodia Singapore Cyprus Costa Rica Latvia Belgium Lithuania Romania Austria Poland Korea, Rep. Korea, Rep. Belgium Portugal Tunisia Austria Sweden Finland Denmark Israel Chile Switzerland Germany Croatia Poland Turkey Costa Rica Chile Netherlands Sweden Russian Federation Philippines Croatia Greece Finland United Kingdom Israel Saudi Arabia New Zealand South Africa Romania Denmark Portugal Germany South Africa New Zealand United Kingdom Spain Philippines Canada Spain Italy Canada France Indonesia India Italy Brunei Darussalam Mexico Indonesia France Norway Argentina Netherlands Australia Mexico Greece Australia India Russian Federation Turkey Colombia Colombia Argentina China United States Japan China Brazil Japan United States Brazil 0 15 30 45 60 75 0 10 20 30 40 50 % of GDP % of GDP Source: Adapted from the Organisation for Economic Co-operation and Source: Adapted from the Organisation for Economic Co-operation and Development–World Trade Organization Trade in Value Added database. Development–World Trade Organization Trade in Value Added database. 96 Making Global Value Chains Work for Development Figure 5.13. Domestic Value Added in Foreign Final Demand, Top The petrochemical industry remains the furthest Five Partner Shares, Selected Exporters, 2011 from final demand, significantly more so than total Chile exports or E&E. The electrical industry is the clos- China est, followed by manufacturing and electronics. E&E Germany and manufacturing exports have moved upstream Japan since 2000—or further away from final demand— Korea, Rep. unlike petrochemicals, which has moved down- Malaysia stream (despite remaining an upstream industry in Mexico general). Philippines Poland Compared with its peers, Malaysia has one of Singapore the highest upstreamness measures, behind only South Africa Australia, Chile, Indonesia, and South Africa (fig- Thailand ure 5.15). This finding is no surprise, given that United States the most upstream industries tend to be related to Vietnam the extraction and processing of raw materials and 0 5 10 15 20 25 30 35 40 45 50 55 60 65 resources, and those comparator countries are all % of domestic value added in foreign final demand natural resource exporters. The results show that in the GVCs in which Malaysia participates, the coun- Australia France Italy Taiwan, China try maintains a position relatively further from final Brazil Germany Japan United Kingdom consumption. Canada India Korea, Rep. United States China Indonesia Mexico Among Malaysia’s peers, only two countries have managed to move downstream (New Zealand and Source: Adapted from the Organisation for Economic Co-operation and Development–World Trade Organization Trade in Value Added database Vietnam), and two other countries (the Philippines and Poland) have moved marginally downstream. line position, or upstreamness, for 426 industries The largest upstream movers are Australia; Chile; (279 of which are manufacturing) using 2002 data Hong Kong SAR, China; Indonesia; Malaysia; and from U.S. input-output (I-O) tables. The measure South Africa. Most countries have increased their is based on I-O relationships, or how much sector u upstreamness because the overall lengths of the value purchases from each sector s as inputs.9 chains have increased with the fragmentation of The average position of Malaysia’s exports from production, and the countries herein are no excep- final demand can then be calculated as the average tion. Moreover, the offshoring-outsourcing process upstreamness measure for each industry, weighted that lengthens GVCs primarily tends to affect the by the importance of that industry in Malaysia’s early stages of production, so that the countries that export basket. Figure 5.14 provides that calculation are most upstream are likely to be relatively more for each year from 2000 to 2013 for total exports, as affected. That fact notwithstanding, a new wave of well as electrical and electronic (E&E) equipment service offshoring and outsourcing has been taking (combined and separately), petrochemicals, and place in recent years, which may affect such conclu- manufacturing exports. sions in the future. Table 5.1 Indicators of Value Added Embodied in Final Demand Indicator Definition FDDVA Domestic value added embodied in foreign final demand, by importing country and exporting sector (US$) FDDVASH FDDVA, by importing country and exporting sector (% of total FDDVA) FDDVA_GDP FDDVA, by importing country and exporting sector (% of GDP) FDFVA Foreign value added embodied in domestic final demand, by origin country and origin sector (US$) FDFVASH FDFVA, by origin country and origin sector (% of total FDFVA) FDFVA_GDP FDFVA, by origin country and origin sector (% of GDP) TSVAFD Bilateral trade balances in value added, by partner country, FDDVA minus FDFVA (US$) TSVAFD_GDP Bilateral trade balances in value added, by partner country (% of GDP) TSVAFD_TSGR Difference in trade surpluses, value added in final demand minus gross trade Source: Adapted from the Organisation for Economic Co-operation and Development–World Trade Organization Trade in Value Added database. Note: These indicators are in the Trade in Value Added database, which covers 61 economies. GDP = gross domestic product. Seller-Related Measures 97 Domestic Gap between Buying Figure 5.14. Upstreamness of Industries in Malaysia and Selling Chains 3.60 A final useful metric is to combine import and 3.42 export upstreamness to compute the domestic gap 3.24 between the buying and selling chains of individual 3.06 sectors (for the concept of upstreamness or dis- Upstreamness 2.88 tance to final demand, see the previous section). We 2.70 calculate the export and import upstreamness for 2.52 each year from 2000 to 2013 for total trade as well 2.34 as trade in the E&E and petrochemical industries 2.16 of Malaysia. Figure 5.16 plots the export or import upstreamness on the left axis, and their difference 1.98 (import upstreamness minus export upstreamness) 1.80 on the right axis, which is an indicator of the domes- 02 03 04 05 06 07 08 09 12 13 00 10 01 11 20 20 20 20 20 20 20 20 20 20 20 20 20 20 tic gap between the buying and selling chains. A pos- E&E Electrical Electronics itive gap indicates that exports are relatively more Manufacturing Petrochemicals Total downstream compared with the import mix, or that Sources: Adapted from Chor 2014; United Nations Comtrade. exports are closer to final demand than are imports. Note: E&E = electrical and electronic. This is the case in economies in which the manu- facturing sector has been a key source of export- led growth, such as China, Japan, and Thailand. Figure 5.15. Upstreamness in Malaysia and Comparators, 2012 and Progression Since 2000 Conversely, a negative gap indicates that a country’s export profile is more upstream than its import pro- Chile file. This is the case in economies whose exports are Australia concentrated in agricultural products and primary South Africa Indonesia commodities, such as Australia and New Zealand. Malaysia Another scenario is that the negative gap indicates Singapore that the country is a large importer of finished con- Philippines sumer goods, rather than being a reflection of the United States composition of its exports; the United States is one Korea, Rep. Thailand example. India The domestic gap between the buying and sell- New Zealand ing chains in Malaysia has been changing since 2000. Japan Whereas it was positive before 2010, it has become Poland negative since then. The shift is being driven by Mexico Hong Kong SAR, China exports that have become more upstream or further Vietnam from final demand; import upstreamness has not China changed significantly over the past decade. Malaysia Cambodia is apparently becoming less plugged into global pro- duction lines as an importer of upstream interme- 1.0 1.5 2.0 2.5 3.0 3.5 diate inputs that are subsequently processed and Upstreamness assembled. Moved upstream Export upstreamness, 2012 Neither E&E nor petrochemicals seems to be Export upstreamness, 2000 Moved downstream behind the change. Although the domestic gap of Sources: Adapted from Chor 2014; United Nations Comtrade. the E&E industry has become smaller since 2008, it remains positive but low—as the preceding analysis of limited GVC participation in E&E that Malaysia imports more complex (downstream) shows—which points to the weak selling side. The petrochemicals for domestic use than the less com- domestic gap of the petrochemicals industry has plex (upstream) petro chemicals that the country been mainly negative since 2002, which indicates sells in GVCs. 98 Making Global Value Chains Work for Development Figure 5.16. Import Upstreamness, Export Upstreamness, and Seller Dimension: Summary Domestic Gap, Malaysia, 2000–13 a. Total Table 5.2 summarizes the main measures of the seller 3.6 0.5 dimension. The table indicates whether the mea- sures are available for goods tasks, services tasks, or 3.4 0.4 both, as well as the level of analysis (country, sector, 3.2 or firm). For easy reference, the table refers to the 0.3 respective sections in the book where the measures 3.0 are introduced. For more information on the under- Upstreamness Domestic gap 0.2 2.8 lying data sets, see appendixes G and H. 0.1 GVCs are a multidimensional phenomenon that 2.6 involves the flow of factors of production, as shown 0.0 2.4 in chapter 1. From the seller’s perspective, several measures can be analyzed to complement the task- 2.2 -0.1 based assessment that has been discussed. Although 2.0 -0.2 a wider set of examples can be found in table A.1 in appendix A, the following are some of the main indi- 00 01 02 03 04 05 06 07 08 09 10 11 12 13 20 20 20 20 20 20 20 20 20 20 20 20 20 20 cators to consider: b. E&E 3.6 0.5 • Wages. Low or high wages (relative to GVC part- 3.4 0.4 ner countries) are a priori indeterminate. High 3.2 wages are likely to be associated with sellers who 0.3 are also owners of GVC assets, technology, and 3.0 know-how. Upstreamness Domestic gap 0.2 • Ideas. Countries specializing in high value-added 2.8 0.1 tasks are likely to export a lot of royalties and 2.6 fee services. The existence or absence of a sound 0.0 framework for intellectual property is likely to 2.4 be associated with sellers specializing in higher 2.2 -0.1 value-added tasks, unless GVCs are predomi- nantly or exclusively within the boundaries of 2.0 -0.2 multinational corporations. 00 01 02 03 04 05 06 07 08 09 10 11 12 13 • Investment. Outward direct investment is more 20 20 20 20 20 20 20 20 20 20 20 20 20 20 c. Petrochemicals prevalent for sellers than for buyers, predomi- 3.6 0.5 nantly in sectors of high GVC intensity, aiming at 3.4 0.4 fostering large-scale production abroad. Foreign direct investment tends to be preferred to non- 3.2 0.3 equity modes in transactions that embody more 3.0 sophisticated technology and know-how. Upstreamness Domestic gap 0.2 2.8 0.1 2.6 0.0 2.4 2.2 -0.1 2.0 -0.2 00 01 02 03 04 05 06 07 08 09 10 11 12 13 20 20 20 20 20 20 20 20 20 20 20 20 20 20 Export upstreamness Import upstreamness Domestic gap Sources: Based on the methodology in Chor 2014 and data from United Nations Comtrade. Note: E&E = electrical and electronic equipment. Seller-Related Measures 99 Table 5.2 Summary of the Main Seller-Related Measures Level of Where to find Measure Task analysis Data sources in book? Share of intermediates in total output Goods and/or Country, sector, Production data (national statistics, Chapter 5, services firm UN-Stat manufacturing data set, p. 87 firm-level data) Share of intermediates in total exports Goods Country, sector Gross export data (Comtrade, BACI, Chapter 5, WITS), categorized using informed p. 87 classifications (broad economic category, parts and components, technical classifications) Share of intermediates in total exports, Goods Country, sector, Gross export data (customs firm-level), Chapter 5, range of exports, bundle of exported firm categorized using informed classifications p. 87 products, and countries (broad economic category, parts and components, technical classifications) I2E trade (% of gross exports) Goods and/or Country, sector International I-O data (WIOD, TiVA, World Chapter 5, services Bank Export of Value Added database ) p. 88 Domestic value added (% of gross value Goods and/or Country, sector International I-O data (WIOD, World Bank Chapter 6, of output) services Export of Value Added database) p. 110 Domestic value added in gross exports Goods and/or Country, sector International I-O data (WIOD, TiVA, World Chapter 5, of third countries services Bank Export of Value Added database) p. 88 Domestic value added in gross exports Goods and/or Country, sector World Bank Export of Value Added Chapter 6, (direct and forward or backward links) services database p. 110 Domestic value added embodied Goods and/or Country, sector International I-O data (WIOD, TiVA) Chapter 5, in foreign final demand (% of GDP) services p. 92 Foreign value added embodied in Goods and/or Country, sector International I-O data (WIOD, TiVA) Chapter 5, domestic final demand (% of GDP) services p. 92 Domestic supplier’s share of output to Goods and/or Country, sector, Enterprise surveys or other firm-level Chapter 6, multinationals in total output services firm surveys p. 113 Domestic supplier’s share of exports Goods and/or Country, sector, Enterprise surveys or other firm-level Chapter 6, in output services firm surveys p. 113 Length of selling chains Total Country, International I-O data (WIOD, TiVA) or Chapter 5, sector national I-O data (for example, U.S I-O p. 94 tables), and trade data Domestic gap between buying and Goods and/or Country, sector, International I-O data (WIOD, TiVA) or Chapter 5, selling chains services firm national I-O data (for example, U.S I-O p. 97 tables), and trade data Note: BACI = international trade database by CEPII; GDP = gross domestic product; I2E = importing to export; I-O = input-output; TiVA = Trade in Value Added database; WIOD = World Input- Output Database; WITS = World Integrated Trade Solution. Notes C and E), at parts and components (appendix D), or at OUTINPcs specific clusters of activity (appendix F). 1. IIOCS = where OUTINP denotes inter- I 2ES cs OUTGRcs 4. I 2ES _ EX CS = where I2ES denotes inter- mediate output and OUTGR denotes gross output in EXGRcs country c and sector s. mediates sold by a country that are embodied in gross EXINPcs exports of third countries and EXGR denotes the seller’s 2. IIE CS = where EXINP denotes intermediate EXGRcs gross exports in a sector. exports and EXGR denotes gross exports in country c and DVA 3EX cs 5. DVA 3EX _ EX CS = where DVA3EX de- sector s. Note that intermediates include energy exports. EXGRcs 3. For example, following the Broad Economic notes domestic value added embodied in third countries’ Categories (BEC) classification (appendix B), or by look- exports and EXGR denotes the seller’s gross exports in a ing at technical or customized classifications (appendixes sector. 100 Making Global Value Chains Work for Development 6. Gaulier and others (2013). Review Papers and Proceedings 102 (3): 412–16. 7. OECD (2013). DOI:10.1257/aer.102.3.412 8. Taglioni and Winkler (2014) discuss this issue. Baldwin, Richard, and Javier Lopez-Gonzalez. 2013. 9. Formally, the measure of upstreamness of industry s “Supply-Chain Trade: A Portrait of Global Patterns is computed as and Several Testable Hypotheses.” NBER Working ∑ = 1d su Fu + 3 ∗ ∑u ∑ N N N d s νd νu Fu Paper 18957, National Bureau of Economic Research, F U IS = 1 ∗ s + 2 ∗ u ν=1 + Cambridge, MA. Ys Ys Ys Chor, Davine. 2014. “Where Are Countries Positioned where Ys is the total output of the industry, Fs is the value along Global Production Lines?” Monetary Authority of of that output that goes to final uses (final consumption or Singapore Macroeconomic Review, April: 94–99. investment), and dsu is the value of inputs from industry s Gaulier, Guillaume, Gianluca Santoni, Daria Taglioni, and that are required by industry u to produce $1 of the latter’s Soledad Zignago. 2013. “In the Wake of the Global output. With this definition, an industry that has its entire Crisis: Evidence from a New Quarterly Database of output channeled to final uses, namely with Fs = Ys, will Export Competitiveness, Volume 1.” Policy Research have Us = 1. Working Paper 6733, World Bank, Washington, DC. OECD (Organisation for Economic Co-operation and Development). 2013. Science, Technology, and Industry References Scoreboard 2013: Innovation for Growth. Paris: OECD. Taglioni, Daria, and Deborah Winkler. 2014. “Making Antràs, Pol, Davin Chor, Thibault Fally, and Russell Global Value Chains Work for Development.” Economic Hillberry. 2012. “Measuring the Upstreamness of Premise No. 144, Poverty Reduction and Economic Production and Trade Flows.” American Economic Management Network, World Bank, Washington, DC. Chapter 6 OTHER MEASURES OF GVC PARTICIPATION: FROM MACRO TO MICRO Introduction The index is measured as the percentage of the coun- try’s gross exports. This chapter goes beyond the buying and selling Figure 6.1 shows the GVC participation index for sides in global value chains (GVCs) and comple- 61 economies worldwide (highlighted in blue).2 The ments the measures illustrated in chapters 4 and 5 results suggest that geographic size—particularly with additional useful measures of participation. relative to regional peers—seems to matter. Smaller First, the chapter illustrates how the buyer and seller economies—such as Luxembourg; Singapore; dimensions can be combined to quantify an overall Slovak Republic; and Taiwan, China—have partici- indicator (the GVC participation index). Second, the pation rates of 60 to 70 percent. The participation chapter focuses on network metrics. It shows how a of larger countries is lower. The participation of country’s position overall, in a sector, in a specific middle-size countries—such as Brazil, Mexico, and GVC, and with respect to individual products can Turkey—is 35 to 50 percent. The index for China be measured and visualized using network metrics. (48 percent) is relatively low, comparable to those Third, the chapter pays special attention to the role of Germany and Japan. China’s participation index of services in value added. Fourth, the chapter intro- reflects very low Chinese value added in third coun- duces measures of direct links in GVCs using firm- tries’ exports, as well as medium foreign value added level data—the micro perspective. in China’s gross exports, as commonly perceived. In addition to country size, the distance to consumer GVC Participation Index markets is another determinant of the participation in GVCs, giving New Zealand, for example, one of Combining two measures that were previously intro- the lowest indexes (less than 33 percent). duced on the buying and sales sides provides a single Figure 6.1 provides a clear indication of whether measure of a country’s involvement in vertically frag- countries tend to be specialized in buying activities mented production, as a user of foreign value added (foreign value added embodied in gross exports, for its own exports and as a supplier of domestic or backward links, highlighted in green) or selling value added embodied in intermediate goods or ser- activities (domestic value added embodied in gross vices used in other countries’ exports. This measure exports of third countries, or forward links, high- is captured by the GVC participation index.1 lighted in yellow). The buying side plays a larger Conceptually, the index can be broadly consid- role in most countries, except for resource-intensive ered a GVC-specific measure of trade openness. The countries, especially Brunei Darussalam, Colombia, higher is the foreign value added embodied in gross Norway, the Russian Federation, and Saudi Arabia. exports (see chapter 4) and the higher is the value Figure 6.2 shows the GVC participation measure, of domestic inputs exported to third countries and excluding mining and quarrying, and coke, refined used in their exports (see chapter 5), the higher is the petroleum products, and nuclear fuel in the numer- participation of a given country in the value chain. ator. To assess the impact of mining and quarrying 101 102 Making Global Value Chains Work for Development Figure 6.1. GVC Participation Index, 2011 Figure 6.2. GVC Participation Index Excluding Mining and Quarrying and Coke, Refined Petroleum Products, and Nuclear Fuel in the Numerator, 2011 Luxembourg Luxembourg Taiwan, China Taiwan, China Slovak Republic Czech Republic Hungary Hungary Czech Republic Slovak Republic Korea, Rep. Iceland Singapore Slovenia Malaysia Ireland Iceland Malaysia Ireland Singapore Slovenia Korea, Rep. Belgium Denmark Norway Malta Finland Estonia Denmark Poland Bulgaria Thailand Poland Finland Estonia Latvia Thailand Austria Malta Belgium Sweden Bulgaria Latvia Philippines Austria Sweden Vietnam Vietnam Chile Cambodia Russian Federation Germany Philippines Tunisia Tunisia Romania Portugal Switzerland Germany Chile Romania China Cambodia Japan China Portugal United Kingdom France Italy Italy Netherlands Costa Rica Japan Netherlands Switzerland Mexico Brunei Darussalam United Kingdom France Hong Kong SAR, China Mexico Spain Spain Norway Lithuania Israel South Africa Lithuania Saudi Arabia Russian Federation Costa Rica Turkey Israel South Africa Australia Canada Hong Kong SAR, China Cyprus Indonesia Indonesia Greece Greece India United States Canada India Turkey Australia United States Saudi Arabia Cyprus Brunei Darussalam Colombia Croatia Brazil New Zealand Croatia Brazil New Zealand Colombia Argentina Argentina 0 12 24 36 48 60 72 0 12 24 36 48 60 72 % of total gross exports % of total gross exports Backward links Forward links Participation index Backward links Forward links Participation index Source: Adapted from the Organisation for Economic Co-operation and Source: Adapted from the Organisation for Economic Co-operation and Development–World Trade Organization Trade in Value Added database. Development–World Trade Organization Trade in Value Added database. Other Measures of GVC Participation: From Macro to Micro 103 Figure 6.3. GVC Participation Index, Malaysia and Peer Figure 6.4. GVC Participation Index, Selected Countries, Chemicals and Chemical Products, 2011 Countries, Electrical and Optical Equipment, 2011 Singaporee Malaysia Korea, Rep, China Germany Philippines United States Korea, Rep, Thailand Japan Malaysia Singapore Poland Mexico Japan Thailand China Vietnam South Africa Poland Vietnam Chile Philippines Germany Chile United States Mexico South Africa 0 0.8 1.6 2.4 3.2 4.0 4.8 5.6 6.4 0 6 12 18 24 30 Percent of total gross exports Percent of total gross exports Backward links Participation index Backward links Participation index Forward links Forward links Source: Adapted from the Organisation for Economic Co-operation and Source: Adapted from the Organisation for Economic Co-operation and Development–World Trade Organization Trade in Value Added database. Development–World Trade Organization Trade in Value Added database. and petroleum and coke on the overall measure, we Figure 6.5. Breakdown of Malaysia’s GVC Participation keep total gross exports in the denominator. Natural Index, 2011 resource–intensive exporters are ranked at the lower Electrical and optical equipment 40.99% end of the spectrum. To illustrate how these concepts apply to specific countries, figures 6.3 to 6.5 focus on the GVC par- ticipation index for Malaysia and peer countries in the chemicals and chemical products sector (figure 6.3) and electrical and optical (E&O) equipment sector (figure 6.4). The sector participation index is computed using total exports as the denominator. Malaysia’s GVC participation in the chemicals and chemical products sector is higher than the level in Chemicals and chemical products 15.37% most of its peer countries, at 4 percent, and in the E&O sector Malaysia’s GVC participation is the high- est compared with its peers, at 25 percent. Moreover, chemicals reflect the general findings that the par- Basic metals and fabricated metal products 6.28% ticipation in GVCs is stronger on the buying side and weaker on the selling side. For E&O, Malaysia’s perfor- Food products, beverages, and tobacco 5.95% mance on the buying side is considered stronger than Transport and storage, post, and telecommunication 5.45% on the selling side, following the patterns of China, Wholesale and retail trade; hotels and restaurants 5.12% Mexico, Thailand, and Vietnam. Together, the two sectors contribute more than one-half of Malaysia’s Other 20.83% overall participation in GVCs (figure 6.5). The contri- bution of E&O to overall participation is 41 percent (participation index of 25 percent relative to an over- all participation index of 60 percent in 2011), whereas the chemicals sector contributes 15 percent (partici- Source: Adapted from the Organisation for Economic Co-operation and pation index of 4 percent relative to 60 percent). Development–World Trade Organization Trade in Value Added database. 104 Making Global Value Chains Work for Development Network Metrics and Visualizations network space shows that China has moved to the center of the global trade network in the past 17 years, Metrics and representations of trade networks can attracting most Asian countries with its gravitational help identify the important suppliers and sellers of a pull (figure 6.6). The visualization is based on the country’s value added (see box 1.3 in chapter 1). undirected trade network (that is, without differenti- Plotting the network of value-added trade on a ating imports from exports) between 1995 (figure 6.6, Figure 6.6. Evolution of the Network of Value-Added Trade, 1995 and 2011 a. In 1995, Germany and the United States were at the core of the network Bulgaria Malta Romania Croatia Argentina Hungary Italy Poland Iceland Brunei Darussalam Brazil Greece Slovenia Taiwan, China Philippines Spain Denmark Canada Tunisia Malaysia Saudi Arabia Austria France Vietnam Netherlands Turkey Cyprus Korea, Rep. Singapore Germany United States Lithuania Chile United Kingdom Switzerland Russian Federation Thailand Ireland Israel Mexico Cambodia China Croatia Latvia Slovak Republic Portugal Czech Republic Colombia South Africa Hong Kong SAR, China India Japan Indonesia Belgium Sweden Estonia Finland Australia Norway New Zealand Luxembourg b. In 2011, Germany, China, and the United States were at the core of the network Norway Denmark Iceland Hungary United Kingdom Brunei Darussalam Malta Mexico New Zealand Turkey South Africa Tunisia Chile Malaysia Belgium Lithuania France Japan Sweden Saudi Arabia Canada Slovak Republic Latvia Australia Russian Federation Vietnam Germany Argentina Romania Estonia United States Indonesia Italy China Netherlands Israel Czech Republic Bulgaria Brazil Thailand Finland Austria Portugal Croatia Croatia Poland Slovenia Cambodia Korea, Rep. Spain Greece India Colombia Philippines Taiwan, China Switzerland Cyprus Ireland Singapore Luxembourg Hong Kong SAR, China Source: Santoni and Taglioni 2015. Other Measures of GVC Participation: From Macro to Micro 105 panel a) and 2011 (figure 6.6, panel b), and con- roots of the tree (links are in darker colors and big- structed using the measure of domestic value added ger), whereas links to peripheral countries (leaves) embodied in gross exports. are in milder colors and smaller. The size of the In figure 6.6, the graphs visualize a minimal span- nodes reflects a country’s centrality. ning tree—a reduced network that reports for each China’s move to the core, bringing it from the country only the strongest relation in value-added periphery to one of the three central nodes of the flows, considering imports and exports of value global trade and production networks, primarily added. The most connected countries represent the results from the buying side (figure 6.7, panel b). On Figure 6.7. Buyer and Seller Perspectives, 2011 a. Seller perspective, largest suppliers:a Most countries buy value added from the United States and Germany Bulgaria Brunei Darussalam Latvia Philippines Cyprus Hong Kong SAR, China Lithuania Russian Federation Singapore Slovenia United Kingdom Malaysia Brazil Argentina Cambodia Norway Greece Malta Estonia China Vietnam Poland Israel India United States Sweden Japan Chile Denmark Thailand Hungary Finland Taiwan, China Korea, Rep. Ireland Slovak Republic Colombia Mexico Turkey Netherlands Indonesia Germany Iceland Canada Italy Romania Australia Saudi Arabia Croatia South Africa Belgium Croatia Czech Republic New Zealand Spain Switzerland France Portugal Australia Luxembourg Tunisia b. Buyer perspective, largest buyers:a Most countries sell value added to Germany and China Hong Kong SAR, China Cambodia India Saudi Arabia New Zealand Vietnam Indonesia Singapore Brunei Darussalam Brazil France Tunisia Philippines Croatia Australia Russian Federation Belgium Portugal Thailand Spain Malta Turkey Colombia Chile China Slovenia South Africa Australia Croatia Malaysia Finland Korea, Rep. Germany Italy Bulgaria Ireland Romania Israel Greece Netherlands Taiwan, China Poland Czech Republic United Kingdom Mexico United States Japan Hungary Iceland Denmark Canada Cyprus Norway Luxembourg Estonia Slovak Sweden Switzerland Republic Argentina Latvia Lithuania Source: Santoni and Taglioni 2015. a. Reduced networks. 106 Making Global Value Chains Work for Development the selling side, Germany and the United States have closeness or proximity. The eigenvector centrality can remained the main suppliers of value added (figure be computed from the buyer’s (BONwin) or seller’s 6.7, panel a). Thus, most Asian countries are in the (BONwout) perspective. The CCw is a measure of the supply system of the United States and in the buy- transitivity of the network, measuring how much the ing system of China. For example, in 2011, Malaysia neighbors of country c are connected to each other. It sourced more than 11.7 percent of overall imported captures whether country c is strong because it trades value added from the United States, whereas China a lot with other countries that are also strong. absorbed more than 29.8 percent of the value added For total trade, many of the countries central to Malaysia exported. the network from the buyer’s perspective are also Network analysis allows much more than visu- central from the seller’s perspective—not only many alizations of dominant patterns. It can quantify high-income countries (HICs) in Europe and North full patterns of made-here-sold-there trade, as well America but also China. Still, China and Germany as its evolution, using network metrics such as the are more important from the buyer’s perspective— eigenvector centrality (BONwin for inflows of value demanding value added—whereas the United States added, BONwout for outflows) and the clustering is more important from the seller’s perspective— index (CCw)—see table 6.1.3 supplying value added. To illustrate how network metrics apply to trade in Again, Malaysia is used to illustrate the results in value added, table 6.2 reports the indexes for the two more detail. Of 56 economies, Malaysia ranks 13th main concepts that underpin the preceding figures— from the buyer’s perspective and 22nd from the sell- eigenvector centrality and clustering—which are also er’s perspective, indicating that it is well integrated constructed using domestic value added in gross into GVCs from the buyer’s and seller’s perspectives exports. These measures are reported for total trade but is stronger on the buyer’s side. Malaysia is well as well as for electrical and electronic and chemicals. clustered in the network (ranking 18th), suggesting The measure of eigenvector centrality—or structural that it trades with other countries that have strong integration—is a measure of the centrality of coun- links. Unsurprisingly, China, Germany, Japan, and try c relative to the overall structure of the network. the United States are the strongest economies. It is the most representative measure of the net- Bangladesh’s ready-made garments (RMG) work and captures the strength of the links and their industry is used to illustrate the use of network Table 6.1 Network Measures Measure Description In-strength and out-strength Sum of values of inflows or outflows. The use of normalized link weights implies that the values for in-strength and out-strength report the market share of country c. The values show that usual market shares are a particular case of network centrality measures, when considering only first-order connectedness. Eigenvector centrality Expresses the idea that the influence of a node is proportional to the influence of its neighbors: (BONwin, BONwout) the node’s eigenvector centrality is largely determined by the eigenvector centrality of its neighbors (multiplied by a constant). Closeness A measure of how close (topological distance) a node is to all other nodes. In general terms, the concept of distance in network analysis is related to the number of steps needed for some node to “reach” another network node. In the case of weighted networks, not the number of steps but the value of the links (the inverse of link value) is considered; the strongest flows result from shorter distance. Clustering (CCw) Expresses network transitivity, that is, how much the neighbors of country c are connected to each other. Hubs and authorities A drawback of eigenvector centrality with disconnected networks (with more than one strongly connected component) is that all nodes report a zero centrality. Hubs and authorities—also eigenvector-based centralities—represent a way to circumvent this problem. That is not the case with the network built from TiVA; only a strongly connected component results. So the out-eigenvector centrality is basically equal to the hub (99.9 percent correlation), whereas the in-eigenvector is equal to the authorities. The out-eigenvector then identifies hubs: nodes that point (sell) to highly connected nodes; in-eigenvector identifies authorities: nodes that are pointed to (buy from) highly connected nodes. Note: TiVA = Trade in Value Added. Other Measures of GVC Participation: From Macro to Micro 107 Table 6.2. Network Measures, All Sectors, E&E, and Chemicals, 2009 Total E&E Chemicals Country Cluster BONwin BONwout Cluster BONwin BONwout Cluster BONwin BONwout Argentina 0.621 0.117 0.127 0.461 0.097 0.116 0.562 0.117 0.126 Australia 0.668 0.132 0.140 0.517 0.112 0.139 0.608 0.125 0.146 Austria 0.677 0.139 0.138 0.559 0.141 0.139 0.618 0.141 0.137 Belgium 0.698 0.147 0.143 0.561 0.136 0.145 0.664 0.160 0.148 Brazil 0.660 0.127 0.140 0.523 0.119 0.135 0.610 0.128 0.144 Brunei Darussalam 0.528 0.093 0.100 0.338 0.040 0.091 0.381 0.044 0.099 Bulgaria 0.597 0.118 0.113 0.473 0.106 0.108 0.538 0.123 0.106 Cambodia 0.505 0.099 0.084 0.360 0.083 0.064 0.388 0.071 0.072 Canada 0.677 0.139 0.139 0.552 0.137 0.138 0.630 0.143 0.142 Chile 0.631 0.122 0.129 0.386 0.034 0.128 0.569 0.119 0.128 China 0.748 0.163 0.157 0.660 0.187 0.168 0.693 0.166 0.159 Czech Republic 0.666 0.139 0.132 0.565 0.153 0.132 0.599 0.137 0.129 Denmark 0.678 0.141 0.136 0.549 0.137 0.135 0.615 0.137 0.138 Estonia 0.578 0.112 0.108 0.472 0.111 0.102 0.502 0.104 0.103 Finland 0.661 0.134 0.133 0.562 0.152 0.131 0.594 0.130 0.132 France 0.725 0.152 0.154 0.606 0.155 0.158 0.683 0.162 0.157 Germany 0.755 0.162 0.162 0.639 0.168 0.170 0.714 0.174 0.167 Greece 0.640 0.126 0.129 0.486 0.106 0.123 0.574 0.123 0.126 Hong Kong SAR, China 0.668 0.132 0.139 0.537 0.124 0.140 0.593 0.123 0.136 Hungary 0.653 0.136 0.127 0.557 0.155 0.127 0.587 0.136 0.123 Iceland 0.542 0.103 0.098 0.448 0.109 0.091 0.482 0.109 0.085 India 0.692 0.143 0.143 0.569 0.142 0.144 0.631 0.138 0.147 Indonesia 0.652 0.128 0.135 0.539 0.132 0.133 0.600 0.128 0.137 Ireland 0.680 0.142 0.137 0.567 0.150 0.137 0.642 0.151 0.143 Israel 0.644 0.131 0.127 0.535 0.136 0.127 0.589 0.134 0.125 Italy 0.715 0.148 0.152 0.595 0.150 0.155 0.669 0.158 0.153 Japan 0.718 0.146 0.156 0.625 0.162 0.165 0.667 0.149 0.158 Korea, Rep. 0.712 0.151 0.147 0.622 0.170 0.155 0.657 0.155 0.146 Latvia 0.571 0.107 0.109 0.431 0.089 0.102 0.488 0.092 0.107 Lithuania 0.580 0.112 0.109 0.448 0.098 0.100 0.533 0.118 0.108 Luxembourg 0.636 0.132 0.122 0.476 0.102 0.118 0.557 0.123 0.117 Malaysia 0.686 0.144 0.139 0.591 0.158 0.145 0.634 0.149 0.138 Malta 0.546 0.104 0.099 0.456 0.109 0.096 0.463 0.092 0.089 Mexico 0.660 0.138 0.130 0.572 0.160 0.131 0.586 0.127 0.131 Netherlands 0.709 0.151 0.146 0.581 0.147 0.149 0.675 0.165 0.150 New Zealand 0.593 0.115 0.114 0.446 0.098 0.106 0.517 0.106 0.109 Norway 0.672 0.134 0.141 0.531 0.121 0.138 0.612 0.127 0.147 Philippines 0.629 0.126 0.123 0.558 0.152 0.130 0.488 0.086 0.114 Poland 0.676 0.140 0.137 0.561 0.144 0.137 0.617 0.142 0.134 Portugal 0.638 0.128 0.126 0.518 0.131 0.121 0.577 0.129 0.123 Romania 0.626 0.124 0.124 0.510 0.125 0.123 0.560 0.126 0.116 Russian Federation 0.695 0.134 0.155 0.543 0.116 0.153 0.660 0.138 0.167 Saudi Arabia 0.648 0.118 0.143 0.470 0.080 0.140 0.627 0.131 0.154 Singapore 0.706 0.150 0.143 0.599 0.163 0.146 0.664 0.161 0.144 Slovak Republic 0.635 0.132 0.121 0.536 0.145 0.121 0.563 0.130 0.115 Slovenia 0.605 0.122 0.114 0.498 0.120 0.110 0.544 0.124 0.108 South Africa 0.626 0.124 0.124 0.472 0.100 0.121 0.554 0.115 0.122 Spain 0.698 0.144 0.145 0.572 0.141 0.147 0.652 0.152 0.147 Sweden 0.690 0.143 0.142 0.576 0.147 0.144 0.636 0.146 0.142 Switzerland 0.689 0.142 0.143 0.584 0.153 0.145 0.650 0.153 0.145 Taiwan, China 0.687 0.141 0.142 0.607 0.164 0.151 0.634 0.148 0.138 Thailand 0.673 0.139 0.136 0.584 0.161 0.139 0.604 0.133 0.134 Turkey 0.663 0.135 0.134 0.532 0.129 0.131 0.605 0.137 0.132 United Kingdom 0.717 0.148 0.153 0.603 0.154 0.157 0.674 0.157 0.158 United States 0.746 0.153 0.166 0.634 0.161 0.173 0.706 0.162 0.173 Vietnam 0.627 0.129 0.119 0.488 0.116 0.115 0.531 0.106 0.117 Source: Based on data from Santoni and Taglioni 2015. Note: The colors of the cells are according to the strength of the metrics: a green cell indicates a strong measure, a yellow cell indicates a weak measure, and a white cell indicates an average measure. BONwin = eigenvector centrality based on inflows of value added; BONwout = eigenvector centrality based on outflows of value added; CCw = clustering index; E&E = electrical and electronic. 108 Making Global Value Chains Work for Development Figure 6.8. World Gross Trade Network for Apparel, 2013 a. Seller perspective b. Buyer perspective Portugal Portugal Slovak Republic Slovak Republic Spain Spain Morocco Morocco Denmark Turkey Cambodia Denmark Turkey Cambodia Sweden Sweden Czech Republic Norway Norway France Czech Republic France Austria Austria Canada Bangladesh Bangladesh Uzbekistan Uzbekistan United Kingdom Canada Unted Kingdom Switzerland Switzerland Russian Federation Sri Lanka Russian Federation Australia Pakistan India Macedonia, FYR Germany Macedonia, FYR Australia India Sri Lanka New Zealand Pakistan Germany New Zealand Hong Kong SAR, China Tunisia Tunisia Netherlands Hong Kong SAR, China Netherlands Italy South Africa South Africa Chile Italy Chile Belgium China Belgium Egypt China Egypt Brazil Brazil United States United Arab Emirates United States Colombia United Arab Emirates Colombia Romania Singapore Romania Singapore Honduras Japan Honduras Peru Kazakhstan Panama Korea, Rep. Japan Ireland Panama Korea, Rep. Peru Ireland Slovenia Thailand Thailand Kazakhstan Vietnam Vietnam Malaysia Malaysia Israel Israel Nicaragua Philippines Nicaragua Slovenia Philippines Indonesia Indonesia Guatemala Poland Mexico Guatemala Saudi Arabia Poland Mexico Saudi Arabia Bulgaria Bulgaria Finland Finland Haiti Haiti Circle Color scale From low to high import share Source: BACI data from CEPII data. Note: Both panels show apparel consumption product flows across economies; links are proportional to world market share. In the exporter graph (panel a), Bangladesh and China are highlighted as the two main exporters by using different colors: yellow for Bangladesh and green for China; the rest of the countries are gray; the node size is proportional to the export market share of each economy. For the buyer graph (panel b), a gradient color scheme is used to distinguish the largest importers (green) from the small importers (white); the node size is propor- tional to the import market share of each economy. The position of the circles in each graph reflects the number of links and relative weight (optimized using a force-directed algorithm, Gephi software). To improve readability, only trade flows (links) accounting for at least 0.001 percent of world trade are shown. The 62 economies in the network cover 91 percent of world trade in the selected apparel products. BACI = international trade database by CEPII. visualization for GVC analysis. The Bangladeshi internationalization of production.4 A positive cor- RMG industry is the second largest in the world in relation, for example, exists between the specializa- exports, after China (see figure 6.8, panel a). The tion of a country in services in 1985 and its per capita total value of Bangladesh’s RMG exports reached gross domestic product in 2010; a positive correlation US$24.1 billion in 2013. The RMG industry accounts also exists between the overall economic complexity for more than 80 percent of Bangladesh’s total of an economy and the importance of the domestic exports and is the country’s largest source of foreign services sector as a contributor of value added for currency. Around 60 percent of RMG exports go to downstream exports (figure 1.4 in chapter 1).5 the European Union, and an additional 19.6 percent Recent data on trade in value added suggest that go to the United States (figure 6.9). services represent about 30 percent of the value added in manufacturing exports. Figure 6.10 illus- trates the services share of value added embodied in Role of Services in Value Added gross exports as a percentage of gross exports in five The focus of the measures in chapter 4 is on quan- manufacturing sectors by type of service input. tifying domestic and foreign value added in gross Measures quantifying the services dimension in exports. A special case that deserves attention is the GVCs are available in the Organisation for Economic role of services value added in gross exports (see Co-operation and Development—World Trade table 6.3, as explained later in this section). That con- Organization Trade in Value Added database and the cept is particularly important in the context of the World Bank’s Export of Value Added database. The Other Measures of GVC Participation: From Macro to Micro 109 Figure 6.9. Bangladesh’s Gross Trade Network: Main Buyers of Bangladeshi Apparel (Cotton) Consumption Products, 2013 Poland Sweden Denmark Russian Federation Germany Australia Switzerland Netherlands Bangladesh Austria Canada United Kingdom Belgium Italy Turkey United States France Japan Spain Circle Color scale From low to high import share Source: BACI data from CEPII data. Note: The graph shows Bangladesh’s export relations for the selected apparel goods. The position of each circle in the graph reflects the number of links and relative weight (optimized using a force-directed algorithm, Gephi software). To improve readability, only trade flows (links) accounting for at least 1 percent of each country’s exports are shown. BACI = international trade database by CEPII. latter is based on data from the Global Trade Analysis Focusing on forward links in Morocco, figure 6.11 Project database and work by Francois, Manchin, suggests that services represent a large share in most and Tomberger (2013), who constructed two matri- goods export sectors—in particular, in manufactur- ces that allow identifying the value-added contribu- ing exports. The share of total services (highlighted in tion of specific sectors to other sectors that either dark gray) exceeds 30 percent in the metals (37 per- sell the final good to the domestic market (domestic cent), paper and publishing (34.3 percent), and min- value-added Sector_GMatrix) or export it (export erals and manufactures not elsewhere classified (both value-added Sector_HMatrix) (see table 6.3). Rows 30 percent) export industries. Services also contrib- indicate the supply sectors, and columns represent ute an important share to Morocco’s most important the demand sectors—that is, sectors that use a spe- manufacturing export sectors: chemicals (21.7 per- cific input. Forward links, VXsharefwd, designate cent), machinery and equipment (23.7 percent), and the total share of a specific input being used across transport equipment (13.6 percent). Among these all sectors and are calculated as the sum over a row. export sectors, trade and transport services (high- Analogously, backward links, VXsharebwd, denote lighted in white) play the largest role, followed by the share of different inputs used in a specific sec- other private services (highlighted in light gray). tor and are calculated as the sum over a column. Forward links of services are stronger than back- The most recent matrix is available for 2011 (for ward links in most countries, including Morocco. more information on the database, see table 6.3 and The following analysis focuses on a cross-country appendix H). comparison of forward and backward links in export 110 Making Global Value Chains Work for Development Table 6.3 Indicators of Services Value Added TiVA database, covering 61 economies Indicator Description EXGR_SERV_DVASH Domestic value added of the services sectors (ISIC Rev. 3 45-95), by sector (% of total gross exports). EXGR_SERV_FVASH Foreign value added of the services sectors (ISIC Rev. 3 45-95), by sector (% of total gross exports). Other indicators See table 4.1 (in chapter 4) and table 5.1 (in chapter 5) for other indicators that can be computed focusing on services value added only. World Bank Export of Value Added database, based on GTAP database (105 countries worldwide) Indicator Description Sector_GMatrix This matrix contains the total domestic value added, based on links. Depending on whether rows or columns are considered, the sum corresponds to forward (row) or backward (column) links. Thus, reading a row for a given sector (sectors are presented on the y-axis) provides information about the value of the sector’s inputs into each sector (on the x-axis). The matrix corresponds to matrix G, as described in the explanatory note in appendix H and in Francois, Manchin, and Tomberger (2013). Sector_HMatrix This matrix contains the total export value added, based on links. Depending on whether rows or columns are considered, the sum corresponds to forward (row) or backward (column) links. Thus, reading a row for a given sector (sectors are presented on the y-axis) provides information about the value of the sector’s inputs into each sector (on the x-axis). The matrix corresponds to matrix H, as described in the explanatory note in appendix H and in Francois, Manchin, and Tomberger (2013). DomVAshare This vector denotes the domestic share of value added of gross value of output per sector. The diagonal matrix, B, as described in the explanatory note and Francois, Manchin, and Tomberger (2013), contains the shares on its diagonal. GXshare Denotes the share of each sector in total exports per country, based on the gross value of exports. See table 3 in Francois, Manchin, and Tomberger (2013). DXshare Denotes the share of each sector’s exports of total exports per country, based on direct value added, ignoring links. See table 4 in Francois, Manchin, and Tomberger (2013). VXsharefwd Denotes the total value added in exports, based on forward links per sector and country. This vector corresponds to the row-sums of matrix H in the explanatory note in appendix H. See table 5 in Francois, Manchin, and Tomberger (2013). VXsharebwd Denotes the total value added in exports, based on backward links. It is obtained by taking the column-sums of matrix H. See table 6 in Francois, Manchin, and Tomberger (2013). Sources: Export of Value Added database, Francois, Manchin, and Tomberger 2013; Organisation for Economic Co-operation and Development–World Trade Organization TiVA database; GTAP database. Note: GTAP = Global Trade Analysis Project; ISIC = International Standard Industrial Classification; TiVA = Trade in Value Added. Figure 6.10. Domestic Value Added of Services Sectors Embodied forward links are stronger than backward links—that in Manufacturing Gross Exports, All Countries, 2009 is, services contribute to export value added more Chemicals strongly than they make use of export value-added Food products contributions from other sectors, especially other Machinery and equipment private services. Textiles Morocco’s forward links of trade and transporta- Transport equipment tion services were lower in 2007 compared with 2001, but higher than in most countries worldwide—that 2 4 6 8 10 12 is, trade and transportation services were less impor- Percent tant for export sectors (forward links) in 2007 than in Finance Other services Transport and storage Distribution and repair 2001. Forward links of trade and transportation ser- Business services vices declined from 26 to 22 percent of export value Source: Adapted from the Organisation for Economic Co-operation and Development–World Trade added, while backward links fell from 21 to 17 per- Organization database. cent. Compared with its peers, Morocco shows lower Note: The share of distribution does not include distribution services for final goods. forward and backward links of trade and transport services than the Arab Republic of Egypt, Tunisia, value added for trade and transport, as well as other and Turkey; nevertheless, Morocco’s links remain private services, in 2007 (figures 6.12 and 6.13). higher than those of most countries worldwide. The fact that most countries, including Morocco, Strong forward (and also, to a lesser extent, back- are located below the 45-degree line indicates that ward) links characterize other private services in Other Measures of GVC Participation: From Macro to Micro 111 HICs, but not in Morocco. The cross-country com- Figure 6.11. Contribution of Services Sectors to Export Value Added parison reveals that other private services contribute of Goods Sectors in Morocco, 2007 more strongly to exports in HICs, whereas backward Agriculture, forestry, & fisheries links play an important but less relevant role. Other Beverages & tobacco products private services in Morocco showed some progress Chemicals, rubber, & plastic products between 2001 and 2007, increasing backward links Energy extraction from 6.0 to 8.8 percent and forward links from 7.9 Ferrous metals Leather products to 9.8 percent; but other countries, such as Egypt and Machinery & equipment Romania, still show stronger links. Manufactures, nec All this evidence emphasizes that a country’s Metal products competitiveness, even in manufacturing, seems to Metals, nec be related to an efficient domestic services sector or Minerals, nec to the degree of the country’s openness to import- Mineral products Paper products, publishing ing such services. In GVCs, the services sector is Processed foods particularly important. Managing the complexity of Textiles the chain and preserving the production standards Transport equipment throughout the chain require strong coordination Wearing apparel that relies on efficient services (such as business, Wood products technical, financial, transportation, and distribution 0 5 10 15 20 25 30 35 40 services) and the movement of key personnel across % of export value added borders (such as engineers, auditors, lawyers, and Construction Trade & transport services Total managers). Electricity, gas, & water Other private services Part 2 of Saez and others (2015) discusses in detail Sources: Adapted from Francois, Manchin, and Tomberger 2013; World Bank Export Value-Added the full range of measures for assessing the impor- database (World Bank); Global Trade Analysis Project (Purdue University). tance of services as a source of competitiveness in the Note: nec = not elsewhere classified. economy.6 Figure 6.12. Forward and Backward Links in Export Value Added, Trade and Transport Services, 2001 and 2007 Trade and transport services a. 2001 b. 2007 0.7 0.7 0.6 0.6 0.5 0.5 Backward links (2001) Backward links (2007) 0.4 0.4 0.3 0.3 EGY TUR TUN TUN 0.2 0.2 TUR Morocco DEU Morocco ROM DEU IND ROM 0.1 0.1 JPN JPN USA IND RUS USA CHN MEX CHN RUS BRA BRA MEX 0 0 0 0.2 0.4 0.6 0 0.2 0.4 0.6 Forward links (2001) Forward links (2007) High Income Upper Middle Lower Middle Low income 45-degree line Sources: Adapted from Francois, Manchin, and Tomberger 2013; Export Value-Added database (World Bank); Global Trade Analysis Project (Purdue University). 112 Making Global Value Chains Work for Development Figure 6.13. Forward and Backward Links in Export Value Added, Other Private Services, 2001 and 2007 Other private services a. 2001 b. 2007 0.5 0.5 0.4 0.4 Backward links (2007) Backward links (2001) 0.3 0.3 IND Morocco 0.2 0.2 Morocco USA TUN EGY ROM IND MEX BRA 0.1 TUR USA BRA 0.1 DEU RUS ROM RUS DEU JPN TUN TUR JPN CHN TUN CHN MEX 0 0 0 0.1 0.2 0.3 0.4 0.5 0 0.1 0.2 0.3 0.4 0.5 Forward links (2001) Forward links (2007) High Income Upper Middle Lower Middle Low income 45-degree line Sources: Adapted from Francois, Manchin, and Tomberger 2013; Export Value-Added database (World Bank); Global Trade Analysis Project (Purdue University). Main Actors and Their Links in GVCs Using industries in a country (if data are available for only Firm-Level Measures one country) or in a single industry across countries (if data are available for several countries). Such Firm-level data have the advantage of capturing comparisons allow countries to assess their level of directly the main actors in a value chain—final pro- integration in GVCs in a specific industry and iden- ducers and their suppliers (other advantages are tify possible areas for policy changes. discussed in box 6.1). Depending on data availabil- This section distinguishes between four types of ity, the extent of GVC links can be compared across firms that characteristically take part in GVCs: Box 6.1. Why Firm-Level Analysis? Aside from the fact that firms are heterogeneous in character- export? Do they rely on imported inputs? If so, from which source? istics and performance, production models are seeing major Do they conduct research and development? Are they paying changes worldwide, which deeply affect economies’ transmission higher wages than non-GVC companies? Are they more intensive mechanisms—domestic and international. in technology and services? But macro aggregations miss the critical features and effects These are just a few sample questions (for a more extensive of firm heterogeneity on the wider economy, because the aggre- list of questions, see annex 11A in chapter 11). Going deeper into gations are not adaptable to changes and innovations in the busi- firm-level dynamics can help policy makers and others not only ness landscape—either within countries or internationally. For improve their aggregate assessments of competitiveness and policy initiatives to deliver the expected results in jobs, domestic GVC participation, but also identify the drivers and the reaction growth, and exports requires identifying the typologies of actors of the real economy to policy interventions (an issue beyond the in global value chains (GVCs) by answering a raft of questions. scope of this book). Are the companies that participate in GVCs domestic or foreign In short, rounding out more aggregate assessments with firm- owned? Are they large multiproduct firms or small and medium level data can improve policies to raise a country’s competitive- enterprises? What is their function in the production process? ness and secure the benefits of being in GVCs. Do they mainly sell inputs to domestic multinationals, or do they Other Measures of GVC Participation: From Macro to Micro 113 1. Multinationals relying on inputs from domestic Figure 6.14. Input Sources of Multinationals in suppliers Agribusiness, 2012 2. Domestic suppliers to multinationals in the country 100 12 15 10 5 13 3. Domestic suppliers that export 18 14 80 21 6 4. Domestic producers relying on imported inputs 26 76 36 % of total inputs Another type of firm is the hybrid case of non- 60 17 equity modes of investment, whereby a multina- 50 47 tional has a contractual relationship with a domestic 40 firm in the host country and maintains some degree 34 of control over the operation and conduct of busi- 20 ness but has no ownership stake. This type of firm includes contract manufacturers that produce fully 0 assembled goods for large retailers, as well as con- Ghana Kenya Mozambique Vietnam tract farming, business process outsourcing, fran- Domestically-owned firms chising, contract management, strategic alliances, Foreign-owned firms abroad and joint ventures.7 Such contractual relationships Foreign-owned firms operating in country Parent firm/other susidiaries of parent firm can fall into all four categories, which are discussed Source: Adapted from a survey by the World Bank International Trade Department in the following subsections. (simple averages). Multinationals’ Share of Inputs from Domestic Suppliers to higher FDI spillover potential in Vietnam (figure 6.14). Box 6.2 describes how this measure (and oth- Most countries devote great attention and resources ers) can be obtained from World Bank Enterprise to attracting foreign direct investment (FDI). They Surveys. hope not only to generate such benefits as jobs, for- eign exchange, and tax revenues but also, perhaps Domestic Suppliers’ Share of Output to more important, to realize dynamic benefits to the Multinationals domestic economy through “spillovers.” That term generally refers to productivity improvements result- The next measure is closely related to the previous ing from knowledge diffusion from multinational one, as it also measures the percentage of a supplier’s affiliates to domestic firms. Although evidence on domestic output sold to a multinational.11 intra-industry spillovers is mixed, some studies have According to a survey by the World Bank shifted the focus to vertical spillovers in upstream International Trade Department of 88 supplier firms and downstream sectors.8 Studies support the idea of in agribusiness, the output share of suppliers to mul- positive backward spillovers from multinationals to tinationals ranges from 21.5 percent in Ghana to local suppliers; the evidence on forward spillovers is 51.8 percent in Mozambique (figure 6.15). The per- more mixed.9 centage increased in all countries except Ghana after More links between multinationals for domestic the domestic supplier started doing business with the inputs thus promise higher FDI spillover potential for multinational. the local economy, captured by its domestic sourc- ing intensity—the percentage of domestic inputs in a Domestic Suppliers’ Share of Exports multinational’s total intermediate inputs.10 The results of a World Bank International Trade The next measure for detecting a country’s exposure Department survey of 30 multinationals in agribusi- to GVCs is the percentage of sales being exported.12 ness in Ghana, Kenya, Mozambique, and Vietnam are Depending on data availability, a supplier’s only indicative, but they seem to suggest that input exports can be further decomposed into direct links between multinationals in agribusiness and exports, indirect exports, and sales to firms that use suppliers are much higher in Vietnam (76 percent) inputs for export products. The same survey of 88 than in African countries (50 percent or less), leading suppliers in agribusiness shows that exports in the 114 Making Global Value Chains Work for Development Box 6.2. GVC Measures Based on World Bank Enterprise Surveys Many of the indicators described herein can be obtained from the Based on this subsample, the share of exports in output can be World Bank Enterprise Surveys. One major advantage of these calculated from the answer to one question: surveys is that their questions are the same across all countries. • Question: In the past fiscal year, what percentage of this Moreover, the data represent a random sample of firms using establishment’s sales fell in each of the following categories? three levels of stratification: sector, firm size, and region. • Possible answers include (1) national sales, (2) indirect exports (1) A multinational’s share of inputs from domestic suppliers can (sold domestically to a third party that exports products), and be calculated from the answers to two questions: (3) direct exports. • Question: What percentage of this firm is owned by each of A supplier’s export share can be computed based on answer 3  the following? only or on the sum of answers 2 and 3. • Answer: ___% foreign private individuals, companies, or (3) Finally, a domestic producer’s share of imported inputs can be organizations. derived from the answers to three questions. Generally, foreign ownership of at least 10 percent qualifies a  • Question: What percentage of this firm is owned by each of firm to be considered a multinational. the following? Based on foreign firms only, the multinational’s share of domestic • Answer: ___% domestic private individuals, companies, or inputs can be obtained. organizations. • Question: As a proportion of all material inputs or supplies Domestic ownership of at least 90 percent qualifies a firm to  purchased [that year], what percentage of this establishment’s be considered domestic. material inputs or supplies were of domestic origin? Domestic producers can be singled out based on the following • Answer: ___% material inputs or supplies of domestic origin. question: (2) A domestic supplier’s share of exports can be calculated from • Question: In the past fiscal year, this establishment’s produc- the answers to three questions: tion falls into which category? • Question: What percentage of this firm is owned by each of • Possible answers include (1) only goods for sale to final con- the following? sumers, (2) semi-finished goods used as inputs by other firms, • Answer: ___% domestic private individuals, companies, or (3) mostly finished goods but also some semi-finished goods, organizations. and (4) mostly semi-finished goods but also some finished Domestic ownership of at least 90 percent qualifies a firm to  goods. be considered domestic.  Typical producer firms should focus mainly on the production of final goods, so categories 1 and 3 are the most appropriate Domestic suppliers can be singled out based on the following to take into account. This question is not covered in all Enter- question: prise Surveys, so the measure in some cases reflects a domes- • Question: In the past fiscal year, this establishment’s produc- tic firm’s (rather than a producer’s) share of imported inputs. tion falls into which category? Based on domestic producers only, the firm’s share of imported • Possible answers include (1) only goods for sale to final con- inputs can be obtained. sumers, (2) semi-finished goods used as inputs by other firms, (3) mostly finished goods but also some semi-finished goods, • Question: As a proportion of all material inputs or supplies and (4) mostly semi-finished goods but also some finished purchased that year, what percentage of this establishment’s goods. material inputs or supplies were of foreign origin? Typical supplier firms should mainly focus on the production of  • Answer: ___% material inputs or supplies of foreign origin. intermediate goods, so categories 2 and 4 are the most appro- priate to take into account. This question is not covered in all Enterprise Surveys, so the measure in some cases reflects a domestic firm’s (rather than a supplier’s) share of exports. Source: Based on World Bank Enterprise Surveys. broader sense made up 15 percent of total sales in (dominated by direct exports), Ghanaian suppliers Mozambique and 36 percent in Vietnam (figure rely more on indirect exports or sell to firms that use 6.16). The survey results also reveal different under- their inputs for export products (box 6.2 describes lying export structures: whereas the types of export how this measure can be obtained from World Bank channels in Kenya and Vietnam are very similar Enterprise Surveys). Other Measures of GVC Participation: From Macro to Micro 115 Figure 6.15. Domestic Suppliers’ Output Sold to Figure 6.16. Sales Channels of Domestic Suppliers in Multinationals in Agribusiness, 2012 Agribusiness, 2012 60 100 11.9 5.7 11.4 8.8 4.3 2.8 12.5 21.6 1.4 2.3 50 80 24.7 9.7 % of total output 68.4 % of total sales 40 60 65.9 63.6 30 40 85 20 20 10 0 Ghana Kenya Mozambique Vietnam Ghana Kenya Mozambique Vietnam Previous year Domestic sales When firm started supplying them Direct export Indirect export Source: Adapted from a survey by the World Bank International Trade Department Sales to firms that use inputs for exports (simple averages). Source: Adapted from a survey by the World Bank International Trade Department (simple averages). Domestic Producers’ Share of Imported Inputs includes total domestic value added of the services sector embodied in gross exports and EXGR is gross exports. The focus should be not only on domestic suppliers but also on domestic producers of final goods that The total decomposition of domestic value added of the can be part of GVCs if they rely on imported inputs. services sector is formalized as EXGR_SVACcs=EXGR_ DDC_SVcs+EXGR_IDC_SVcs+ EXGR_RIM_SVcs, where A domestic producer’s percentage of imported inputs EXGR_DDC_SV is the direct domestic services value in total intermediate inputs is a widely accepted mea- added, EXGR_IDC_SV the indirect domestic services sure of offshoring.13 value added, and EXGR_RIM_SV re-imported domestic services value added. 5. Saez and others (2015, 48). Notes 6. Saez and others (2015). 7. UNCTAD (2011). 1. Developed by Koopman and others (2011), the 8. Javorcik (2004); Blalock and Gertler (2008). index can be formalized as 9. See, for instance, the meta-analysis by Havranek and (EXGR _ FVAcs + DVA 3EX cs ) Irsova (2011). GVC _ PARTcs = = EXGRcs 10. The measure can be formalized as EXGR _ FVASH cs + DVA 3EX _ EX cs INP _ DOM csi MULT _ DOM csi = where INP_DOM re- where EXGR_FVA is the foreign value added embodied in INP _TOTALcsi gross exports in sector s and country c, as defined in chapter fers to the value of inputs that multinational i in sector 4, and refers to the sourcing side in GVCs. DVA3EX is the s and country c purchases from domestic suppliers, and domestic value added of sector s and country c embodied INP_TOTAL refers to the total value of inputs bought. in gross exports of third countries, which was defined in 11. The measure is defined as chapter 5 and relates to the selling side in GVCs. Both are OUTP _ MULTcsi SUP _ MULTcsi = where OUTP_MULT measured as a percentage of gross exports, EXGR. OUTP _TOTALcsi 2. OECD (2012). refers to the value of output that supplier i in sector s, 3. Those concepts, which underpin the graphical rep- country c, sells to multinationals, and OUTP_TOTAL resentations of figures 6.6 and 6.7, are described in chap- refers to the total value of output produced. ter 1. They differ from econometric assessments in ways 12. The measure is defined as described in box 1.3 in chapter 1. EX csi SUP _ EX csi = where EX denotes a sup- 4. The concept can be formalized as OUTP _TOTALcsi EXGR _ SVAC cs plier i’s exports and OUTP_TOTAL its total output in SERV _VAGRcs = where the EXGR_SVAC EXGRcs country c and sector s. 116 Making Global Value Chains Work for Development 13. The measure can be formalized as Koopman, Robert, William Powers, Zhi Wang, and Shang- INP _ IM csi Jin Wei. 2011. “Giving Credit Where Credit Is Due: PROD _ IM csi = where INP_IM denotes a Tracing Value Added in Global Production Chains.” INP _TOTALcsi NBER Working Paper 16426, National Bureau of domestic producer i’s value of imported inputs in sector Economic Research, Cambridge, MA. s and country c, and INP_TOTAL represents the value of OECD (Organisation for Economic Co-operation and total inputs. Development). 2012. “Managing Aid to Achieve Trade and Development Results: An Analysis of Trade-related Targets.” COM/DCD/TAD(2012)12/FINAL, OECD, References Paris. Blalock, G., and P. Gertler. 2008. “Welfare Gains from Saez, Sebastian, Daria Taglioni, Erik van der Marel, Foreign Direct Investment through Technology Claire H. Hollweg, and Veronika Zavacka. 2015. Transfer to Local Suppliers.” Journal of International Valuing Services in Trade: A Toolkit for Competitive- Economics 74 (2): 402–21. ness Diagnostics, Volume 1. Washington, DC: World Francois, J., M. Manchin, and P. Tomberger. 2013. Bank. “Services Links and the Value Added Content of Trade.” Santoni, Gianluca, and Daria Taglioni. 2015. “Networks Policy Research Working Paper 6432, World Bank, and Structural Integration in Global Value Chains.” In Washington, DC. João Amador and Filippo di Mauro (eds) The Age of Havranek, Tomas, and Zuzana Irsova. 2011. “Estimating Global Value Chains: Maps and Policy Issues, London: Vertical Spillovers from FDI: Why Results Vary and CEPR. What the True Effect Is.” Journal of International UNCTAD (United Nations Conference on Trade and Economics 85 (2): 234–44. Development). 2011. WIR11. World Investment Report Javorcik, Beata Smarzynska. 2004. “Does Foreign Direct 2011: Non-Equity Modes of International Production and Investment Increase the Productivity of Domestic Development. New York and Geneva: United Nations. Firms? In Search of Spillovers through Backward Linkages.” American Economic Review 94 (3): 605–27. Chapter 7 USE OF GVC MEASURES TO ASSESS THE DRIVERS AND IMPACTS OF GVC PARTICIPATION Introduction Decomposition of Gross Export Growth What are the determinants of global value chain What are the growth contributions of the differ- (GVC) links? Do GVC links matter for economic ent components of gross exports—as introduced in upgrading? And what is the link between GVC par- chapter 4—for export growth? We can decompose ticipation and labor market outcomes? Although the the growth rate of gross exports (EXGR) into the GVC participation measures defined in the previous direct (intra-sector) domestic value added embodied chapters are not suited to answer these questions, in gross exports (EXGR_DDC), indirect (upstream) they can be used in combination with each other or domestic value added embodied in gross exports with other measures to shed further light on two key (EXGR_IDC), and foreign value added embod- questions policy makers need to ask: which policies ied in gross exports (EXGR_FVA) (all in natural help a country enter GVCs, and, more important, logarithms): does GVC participation lead to development? D.lnEXGRcst = α + βD.lnEXGR_DDCcst Although providing answers to these questions + γD.lnEXGR_IDCcst + δD.lnEXGR_FVAcst would go beyond the scope of this book, this chapter + Dcs + Dt + εcst presents a research agenda and examples of possible estimation strategies for ways to test for the drivers D. denotes first differences, while subscripts c, s, and and impacts of GVC participation using statistical t designate country, sector, and time, respectively. All methods or econometrics or quantifying direct rela- regressions control for country-sector and year fixed tionships in international input-output tables. effects, denoted by Dcs and Dt, respectively. This statistical relationship does not include re-imported domestic value added, EXGR_RIM, What Are the Determinants of GVC Links? because in many cases that value is 0, and taking nat- The first section of this chapter focuses on the deter- ural logarithms would yield many missing observa- minants of GVC links. The first step decomposes tions. Although all components are expected to show gross export growth into its components. If gross positive coefficient signs, comparing differences export growth is accepted as a measure of GVC links across countries or sectors within a country is never- on the selling side, the decomposition allows for theless interesting. detecting how much of the value added is generated This assessment can be based on data from the Trade at home and abroad. in Value Added (TiVA) data set of the Organisation for The second step adopts two measures of GVC Economic Co-operation and Development (OECD) links—GVC integration at the country or sector level and World Trade Organization (WTO), which covers and a GVC participation dummy at the firm level. 61 OECD and non-OECD countries and 34 sectors The text focuses on different determinants of GVC (two primary sectors, 17 manufacturing sectors, 10 links at the country, sector, and firm levels. commercial services sectors, and five other services 117 118 Making Global Value Chains Work for Development sectors), for 1995, 2000, 2005, and 2008 to 2011. GVC Integration and Logistics Performance Because all the measures are reported in nominal val- Analysis of high- and middle-income countries based ues, applying appropriate deflators (preferably at the on information available in the OECD-WTO TiVA country or sector level) is important. database in 2008 confirms that higher GVC integra- tion, as a buyer and seller, hinges on better overall logistics performance (figure 7.1). The LPI takes into Correlations of GVC Integration with Country-Level account a country’s customs efficiency, quality of Characteristics trade and transport infrastructure, ease of arrang- What are the key country characteristics associ- ing shipments, quality of logistics services, ability ated with GVC integration? Looking at the deter- to track and trace consignments, and delivery times. minants of GVC participation requires developing For high-income countries (HICs), the positive cor- a sound theoretical model. However, initial insights relation is even stronger than for middle-income can be gathered by assessing the statistical correla- countries (not shown here), which suggests that the tion between measures of GVC integration with importance of efficient logistics may rise with more selected indicators at the country level. This analy- complex tasks. sis uses the measure of structural integration in GVCs—BONwin (buyer’s perspective) and BONwout GVC Integration and Skill Levels (seller’s perspective), as computed by Santoni and Analysis of high- and middle-income countries with Taglioni (2015) and introduced in chapter 6. Here information available in the OECD-WTO TiVA we focus on the following three country character- database for 2008 suggests that a higher share of istics, which, according to the economic literature, workers with tertiary education is positively corre- are important determinants of GVC participation: lated with GVC participation, from the buyer’s and (1) logistics performance, (2) share of people with a seller’s perspectives (figure 7.2). The positive corre- tertiary education in the workforce, and (3) geograph- lation seems to be stronger for GVC integration as a ical distance to the closest global knowledge center. seller, which implies that becoming an exporter in Good logistics performance is important because GVCs depends more crucially on skills than import- key components of GVC production are time sensi- ing GVC inputs. tive, and reliable connectivity allows firms to connect factories across borders more efficiently. A skilled GVC Integration and Geographical Distance to workforce is recognized as an important determi- Knowledge Centers nant of countries’ success in GVCs because it allows Shorter geographical distance to major knowledge producing at the high standards of productivity, centers is positively correlated with higher GVC efficiency, sophistication, and timeliness required to integration. Figure 7.3 plots the correlation between serve global markets. Countries closer to the hubs GVC integration and geographical distance to the in GVCs and to the global centers of knowledge are closest (in kilometers) of three knowledge centers for favored by easier access to tacit knowledge. Unlike a range of high- and middle-income countries with knowledge embodied in technology, tacit knowl- information available in the OECD-WTO TiVA data- edge requires frequent and continued face-to-face base in 2008. The analysis takes into account the fol- interaction between the staff and managers of lead lowing knowledge centers: Germany, Japan, and the firms or turnkey suppliers and those of other firms United States. Figure 7.3, panel a, suggests that lon- in the GVC, and the importance of tacit knowledge ger distance to the closest knowledge center is neg- increases for more complex tasks. The analysis uses atively correlated with GVC integration as a buyer. the overall Logistics Performance Index (LPI) to This finding might support the view that countries quantify logistics performance; the share of workers that are geographically closer to technology cen- with tertiary education to quantify the skill level; and ters are able to import more knowledge-intensive the geographical distance from Germany, Japan, and inputs, which increases their GVC integration as the United States as a proxy for distance from knowl- a buyer directly and perhaps also indirectly (in the edge centers. Those three countries are identified as sense that greater access to technology enables them global knowledge centers through the network anal- to participate in more GVCs, which requires them to ysis illustrated in chapter 6 (eigenvector centrality of import more GVC inputs). Geographical distance to outflows of value added). major knowledge centers is also negatively correlated Use of GVC Measures to Assess the Drivers and Impacts of GVC Participation 119 Figure 7.1. GVC Integration and Overall Logistics Performance Indicator, 2008 a. GVC integration as a buyer b. GVC integration as a seller 0.17 0.17 CHN DEU USA GBR DEU FRA RUS KOR USA JPN NLD CHN JPN FRA SWE IND CZE ITA ITA 0.15 ESP 0.15 BEL GBR POL SGP SAU IRL BEL CAN HUN MYS SWE POL KOR NOR NLD MEX DNK IND ESP SGP RUS VNM THA AUT BRA IDN AUS SVK FIN THA MYS CHE TUR ARG TUR LUXCHE NOR FIN 0.13 BRA ISR PRT CAN 0.13 MEX CZE GRC AUT BONwout PHL IDN SVN ZAF AUS HKG BONwin ROU ROU PRT DNK HKG GRC ISR LUX CHL PHL HUN IRL BGR SAU NZL VNM SVK LTU CHL NZL ARG LTU LVA SVN 0.11 EST 0.11 ZAF LVA EST BGR KHM 0.09 0.09 KHM 0.07 0.07 2.25 2.50 2.75 3.00 3.25 3.50 3.75 4.00 4.25 4.50 2.25 2.50 2.75 3.00 3.25 3.50 3.75 4.00 4.25 4.50 Overall LPI Overall LPI Country 95% Confidence interval Fitted values Sources: Adapted from Santoni and Taglioni 2015; Organisation for Economic Co-operation and Development–World Trade Organization Trade in Value Added database; World Bank World Development Indicators 2007. Note: BONwin = buyer-related measure of structural integration in GVCs; BONwout = seller-related measure of structural integration in GVCs; GVC = global value chain; LPI = Logistics Performance Index. Figure 7.2. GVC Integration and Skill Levels, 2008 a. GVC integration as a buyer b. GVC integration as a seller 0.17 0.17 DEU DEU 0.16 NLD 0.16 RUS SWE FRA GBR JPN ITA ESP GBR NLD ITA ESP 0.15 SGP BEL JPN 0.15 FRA CHE THA AUT MYS SAU SWE NOR MEX DNK IRL CAN GRC SGP 0.14 CZE POL CHE 0.14 THA AUT MYS BEL CAN HUN TUR AUS TUR RUS IDN DNK BONwout BONwin IDN SVK ISR POL 0.13 PRT HKG FIN 0.13 ZAF GRC PHL LUX NOR HKG FIN AUS HUN CHL IRL ISR ROU SVN MEX LUX ZAF 0.12 SAU CHL 0.12 CZE PHL BGR NZL NZL LTU ROU PRT SVN LTU EST 0.11 LVA 0.11 SVK LVA BGR MLT EST 0.10 ISL 0.10 MLT ISL 0.09 0.09 0 10 20 30 40 50 60 0 10 20 30 40 50 60 Workers with tertiary education (%) Workers with tertiary education (%) Country 95% Confidence interval Fitted values Sources: Adapted from Santoni and Taglioni 2015; Organisation for Economic Co-operation and Development–World Trade Organization Trade in Value Added database; World Bank World Development Indicators 2007. Note: Education data are not available for many countries. BONwin = buyer-related measure of structural integration in GVCs; BONwout = seller-related measure of structural integration in GVCs; GVC = global value chain. 120 Making Global Value Chains Work for Development Figure 7.3. GVC Integration and Geographical Distance to the Closest Knowledge Center (Germany, Japan, and the United States), 2008 a. GVC integration as a buyer b. GVC integration as a seller 0.17 0.17 DEU CHN USA DEU 0.16 GBR FRA USA ITA 0.16 RUS NLD KOR SGP JPN FRA CHN JPN IRL SWE RUS GBR ITA BEL TUR MYS 0.15 0.15 DNK AUT ESP TWN NLD NOR KOR SAU CAN POL FIN IND CHE ESP SWE SGP CHE MEX THA BEL TWN IND AUS AUS CAN MYS CZE HKG ISR FIN THA BRA 0.14 LUX HUN PRT 0.14 AUT DNK HKG IDN IDN BRA ZAF IRL TUR MEX CHL SVK NOR GRC VNM POL GRC BONwin BONwout ROU PHL CHL CZE HUN ISR ZAF PRT 0.13 SVN BGR SAU 0.13 LUX ARG LTU ARG SVK ROU PHL VNM EST NZL LTU SVN BGR NZL LVA 0.12 MLT 0.12 LVA EST BRN ISL KHM ISL 0.11 0.11 MLT BRN 0.10 0.10 KHM 0.09 0.09 –1 0 2 4 6 8 10 11 –1 0 2 4 6 8 10 11 Minimum distance to major knowledge center Minimum distance to major knowledge center (km, in thousands) (km, in thousands) Country 95% Confidence interval Fitted values Sources: Adapted from Santoni and Taglioni 2015; Organisation for Economic Co-operation and Development–World Trade Organization Trade in Value Added database; CEPII 2012. Note: BONwin = buyer-related measure of structural integration in GVCs; BONwout = seller-related measure of structural integration in GVCs; GVC = global value chain; km = kilometers.. with GVC integration as a seller, but the correlation Pr(gvcict = 1) = Pr(πict = βfirmict seems to be weaker (figure 7.3, panel b). We find that + γpolicycst + S(1 – gvcit–1) > 0), resource-rich countries are located more remotely where subscript c denotes country, s is sector, firm is from knowledge centers. firm-level determinants of GVC participation, and policy is policy determinants. Determinants of Firm-Level GVC Entry The analysis focuses on the following estimation equation: This subsection amends the Roberts and Tybout (1997) theoretical model on the determinants of gvcict = α0 + βfirmict + γpolicycst + Di + Dcs + Dt + εict exporting. A firm i’s propensity to participate in a where α0 denotes the constant; Di, firm fixed effects; GVC at time t depends on the firm’s expected prof- Dcs, country-sector fixed effects; Dt, year fixed effects; its, π, which, in turn, are influenced by expected rev- and εict, the idiosyncratic error term. enues, R, and costs, c, plus sunk GVC entry costs, S: Following the literature on the firm-level determi- Pr(gvcit = 1) = Pr(Rit > cit + S(1 – gvcit–1), nants of exporting, the model includes firm size, firm age, foreign ownership status, as well as measures of where gvc denotes a GVC dummy at the firm level. workers’ skills and productivity as determinants of The term S(1 – gvcit – 1) is 0 if the firm participated GVC participation. This leads to the following ver- in GVCs in period t – 1, and 1 otherwise. In other sion of the preceding equation: words, the firm participates in GVCs if expected gvcict = α0 + β1ln empict + β2 ln ageict profits π > 0. + β3fdiict + β4 ln wageict + β5 ln tfpict The firm’s expected profits π are affected by firm- + γpolicycst + Di + Dcs + Dt + εict level characteristics and the policy environment, which can generate or lower revenues R or costs c. where emp denotes the total number of employees The equation then translates into the following: (in logarithms); age denotes the years of operation Use of GVC Measures to Assess the Drivers and Impacts of GVC Participation 121 (in logarithms); fdi is a dummy that equals 1 if the Do GVC Links Matter for Economic foreign private ownership ≥ 10 percent, and 0 oth- Upgrading? erwise; wage is the average real wage per worker (in logarithms) as a proxy for worker skills; and tfp is The second part of this chapter focuses on the total factor productivity (in logarithms). role of GVC links for economic upgrading. Three Next, the focus is on the policy determinants main measures of economic upgrading are adopted: of GVC entry, as discussed in part III of this book, (1) growth of domestic value added embodied in including the following policy determinants at the gross exports at the sector level in the first section, country-sector level (depending on data availabil- (2) level of domestic value added at the sector level ity, the determinants can also be measured at the in the second section, and (3) firm-level labor pro- country-region or country level only): ductivity in the third section. Different measures of GVC links are also explored, including GVC mea- gvcict = α0 + β1ln empict + β2 ln ageict sures of structural integration as buyers and sellers + β3fdiict + β4 ln wageict + β5 ln tfpict in networks, foreign value added embodied in gross + γ1epzcst + γ2opencst + γ3 connectcst exports, domestic value added embodied in exports + γ4competitivecst + γ5investcst of third countries, GVC participation index, posi- + γ6domestcst + Di + Dcs + Dt + εict tion in GVCs (upstreamness), domestic length of sourcing chains, and share of foreign output in a where epz is a dummy that equals 1 if the country has sector. an export-processing zone in a sector, and 0 if not; open denotes measures of openness to international trade and foreign direct investment (FDI); connect Growth of GVC Links and Domestic denotes measures of connectivity to international Value Added in Exports markets (for example, logistics performance, cus- Do the intensity and nature of GVC links matter for toms, and infrastructure); competitive covers mea- growth in domestic value added that is exported? sures of competitiveness in unit labor costs and labor This question can be explored through econometric productivity; invest captures drivers of investment analysis from several angles. First, the most obvious (for example, intellectual property protection, level question to explore is whether the degree of struc- of competition, administrative burden, and corrup- tural integration in global value-added trade matters. tion); and domest denotes measures of the quality of Second, econometric analysis can be used to investi- domestic value chains and the services infrastructure gate how greater integration of a country in GVCs as (for example, competence and quality of services). a buyer—as opposed to weaker integration as a seller Depending on the type of variables chosen, it might (that is, more unbalanced GVC integration)—affects be necessary to use logarithms. domestic value-added growth from gross exports. The equation can be estimated using the probit/ Third, the analysis can examine more closely the logit type of regression model where the dependent relation between the growth of foreign value added variable is a binary variable taking the value 1 or 0. embodied in gross exports and the domestic value- added component. Fourth, it can look at the role of a Determinants of Sector GVC Participation country’s position in the value chain (upstreamness or distance to final demand). Finally, econometrics If firm-level data—in particular, information on can be used to investigate the role of the domestic GVC participation—are not available, an alternative length of the sourcing chains. could be to estimate the impact of the policy deter- minants just discussed on sector GVC participation in a country, gvcpart, where GVC participation is Growth of GVC Participation and Domestic Value not a dummy variable, but enters the equation in the Added Embodied in Exports form of values (in logarithms). The estimation equa- What is the relationship between the growth rate of tion then looks as follows: GVC participation and a country’s growth of domes- tic value added embodied in exports? To address this ln gvcpartict = α0 + γ1epzcst + γ2opencst question, we can correlate BONwin and BONwout + γ3connectcst + γ4competitivecst + γ5investcst growth measures of “eigenvector centrality” or struc- + γ6domestcst + Dcs + Dt + εcst tural integration in GVCs (as introduced in chapter 122 Making Global Value Chains Work for Development 6) with the growth of domestic value added embod- D.lnEXGR_DVAcst = α + βD.lnBONbalcst ied in exports, EXGR_DVA: + γD.lnBONbalcst* Dummycountry of interest + Dcs + Dt + εcst D.lnEXGR_DVAcst = α + βD.lnBONcst + γD.lnBONcst *Dummycountry of interest The analysis can use an inverse measure of bal- + Dcs + Dt + εcst anced GVC integration, defined as ln(BONwin– BONwout). Because EXGR_DVA is reported in The two measures of structural integration in nominal values, the use of appropriate deflators GVCs, BONwin and BONwout, are suggested as (preferably at the country-sector level) is important. measures of eigenvector centrality. These measures A larger difference designates less balanced GVC are preferred over other measures of trade open- participation because GVC integration as a buyer is ness or GVC participation, as the former account not larger than as a seller. If the sign of the coefficient only for direct links, but also for indirect links. For is positive, the results could point to the beneficial example, in accounting for the trade links between, effects of importing foreign know-how and technol- say, Argentina and Mexico, BONwin and BONwout ogy. Alternatively, the analysis can use a measure of also take into consideration the full network of trade, balanced GVC participation in levels (rather than including for example the trade links between China growth rates). Alternative measures of balanced GVC and the Republic of Korea. Alternative measures of participation—for example, based on the forward GVC participation, such as the GVC participation and backward components of the GVC participation index, can also be used. index—can also be used. The domestic value added embodied in exports is measured in logarithms. Because EXGR_DVA is Growth of Foreign and Domestic Value Added reported in nominal values, appropriate deflators Embodied in Gross Exports (preferably at the country-sector level) must be used. A more direct way to look at the role of foreign To detect the correlation specific to the coun- inputs is to run correlations between the growth rate try of interest, the interaction term γlnBONcst* of EXGR_FVA and the growth rate of EXGR_DVA Dummycountry of interest—the dummy for the country (in logarithms): of interest—is added, where the dummy takes the value 1 if that country is the one under analysis, and D.lnEXGR_DVAcst = α + βD.lnEXGR_FVAcst 0 otherwise. We also control for country-sector fixed + γD.lnEXGR_FVAcst*Dummycountry of interest effects, Dcs, and fixed year effects, Dt. + Dcs + Dt + εcst The signs of the coefficients of the correlations Because EXGR_DVA and EXGR_FVA are reported provide a first indication of whether a growing inte- in nominal values, the use of appropriate deflators gration in GVCs may have the potential to increase (preferably at the country-sector level) is important. the growth of domestic value added embodied in Although a positive sign on the coefficient would gross exports. Although the analysis does not allow for be expected, because importing enables exporting for establishing causality, growing GVC integration can most countries, the size of the coefficient by sector lead to higher output, productivity, and value added could indicate which sectors benefit more strongly at home via several transmission channels (see figure from foreign know-how or technology. 1.11 in chapter 1). The correlation can also be run at the sector level to detect differences across GVCs. Growth of Upstreamness and Domestic Value Added Embodied in Gross Exports Growth of Balanced GVC Participation and Does the position of a firm in the value chain Domestic Value Added Embodied in Gross Exports (upstreamness or distance to final demand) mat- To detect whether balanced GVC integration—as a ter for the domestic value added that is exported? buyer and as a seller—matters for economic upgrad- The following computation illustrates correla- ing, correlations can be run between the growth rate tions between the growth rate of upstreamness of balanced integration in GVCs (BONbal) and the (UPSTREAM) and the growth rate of domestic value growth rate of domestic value added embodied in added embodied in gross exports (EXGR_DVA) (in gross exports (EXGR_DVA) (in logarithms): logarithms): Use of GVC Measures to Assess the Drivers and Impacts of GVC Participation 123 D.lnEXGR_DVAcst = α + βD.lnUPSTREAMcst We estimate a standard fixed effects model for + γD.lnUPSTREAMcst*Dummycountry of interest sector s in country c at time t of the following form: + Dcs + Dt + εcst lnDVAcst = α + βlnGVCcst + γlntrade Alternatively, the level rather than the growth of + δ1lncapitalcst + δ2lnempcst + Dcs + Dst + Dct + εcst upstreamness can be included as the right-hand vari- where DVA denotes domestic value added, capital able. Because EXGR_DVA is reported in nominal is capital stock, and emp is the number of employ- values, the use of appropriate deflators (preferably at ees. GVC captures our measures of GVC integration, the country-sector level) is important. which enter the function as part of the technology It is important to bear in mind that the results shifter. The first GVC indicator is the amount of are clearly sector specific and depend on the cur- EXGR_FVA; the second indicator is the amount of rent position of the country in a specific value chain, domestic value added re-exported by third countries as well as the value-added contribution of the tasks (DVA3EX). EXGR_FVA quantifies a country’s back- upstream and downstream of the current ones. More ward links into GVCs, or GVC integration as a buyer, upstreamness in the electronics sector, for exam- while DVA3EX quantifies a country’s forward links ple, can be beneficial if the country is specialized into GVCs, or GVC integration as a seller (see chap- in assembly or production activities; higher value- ters 4 and 5 for more details). added activities, such as research and development In addition to GVC integration, we also include and design, are located at the beginning of the value a measure of final goods trade (trade), to separate chain. A move toward downstream postproduction the potential positive GVC effect from the simple activities (marketing, logistics, after-sales, and so positive effect of trade openness. For this, we calcu- forth) can also be beneficial. late the amount of foreign value added processed or consumed domestically. This covers imports of final Growth of Domestic Length of Sourcing Chains and goods and intermediate goods assembled and con- Domestic Value Added Embodied in Gross Exports sumed domestically. Because the latter part might The relationship between the growth rate of the overlap with GVC trade, we might have a down- domestic length of sourcing chains (LENGTH) and ward bias in our estimates. However, not controlling the growth rate of EXGR_DVA (in logarithms) is for openness would prevent us from separating the illustrated as follows: effects of GVC trade and final goods trade. D.lnEXGR_DVAcst = α + βD.lnLENGTHcst All variables are measured in logarithms at the + γD.lnLENGTHcst*Dummycountry of interest sector level in a country. We also include country- + Dcs + Dt + εcst sector fixed effects (Dcs), country-time fixed effects (Dct), and sector-time fixed effects (Dst). The last are Because EXGR_DVA is reported in nominal val- included to capture innovation that is part of the ues, the use of appropriate deflators (preferably at technology shifter. the country-sector level) is important. Again, levels As we are interested in the contribution of can be used instead of growth rates for the length country- specific policy variables to economic variable. upgrading through GVCs, we include an interaction term between a set of national characteristics and GVC Links and Domestic Value Added1 our GVC indicators: lnDVAcst = α + β1lnGVCcst + β2lnGVCcst*policyc This step focuses on the effect of GVC integra- + γlntrade + δ1lncapitalcst + δ2lnempcst + Dcs + Dst tion—as a buyer and a seller—on domestic value + Dct + εcst added, also taking into account the mediating role of national policy. Domestic value added is gener- The sign of the coefficient β 2 indicates whether ated by combining labor with capital stock, and is a certain policy helps increase (β2 > 0) or reduce dependent on a country’s technology shifter. The (β2 < 0) the impact of GVC integration on domestic technology shifter is assumed to be a function of value added. international trade and innovation, which is consis- As measures for the policy variables (policy) we tent with the trade literature. employ variables capturing a country’s infrastructure, 124 Making Global Value Chains Work for Development foreign presence, legal institutions, and innovation result from the willingness of foreign investors to capabilities. Hence, we include variables that ana- provide technology and know-how, and the effect of lyze a country’s ability to join GVCs (infrastructure, increased competition with local firms for limited foreign presence, legal system, and institutions) and resources in the country (for a discussion of how for- its ability to upgrade (innovation). Because of the eign firm characteristics can influence spillovers, see sometimes incomplete data, we use the average over chapter 9). the period. Measures of absorptive capacity, by contrast, The regression results for a set of 40 countries allow for testing the preparedness of domestic firms and 35 industries for 1995–2011 are shown in annex to absorb new technology and know-how, using 7A. Many other aspects of a country’s institutional various channels. Measures of domestic institu- characteristics (including education, trade openness, tional preparedness and other national characteris- financial regulation, labor market regulation, busi- tics indicate the ability of the domestic economy to ness environment, competition, development, and so facilitate technology spillovers, enhance the produc- on) can also be assessed. tivity of domestic firms, and generate positive effects If we are interested in a country’s performance on wages and skilled labor (for a discussion of how in a specific role, we can add two more interaction absorptive capacity and national institutions can terms to the equation: influence spillovers, see chapter 9). The econometric estimation uses a cross-section lnDVAcst = α + β1lnGVCcst + β2lnGVCcst*policyc of more than 25,000 domestic manufacturing firms + β3lnGVCcst* Dummycountry of interest in 76 low- and middle-income countries (LMICs) + β4lnGVCcst*policy* Dummycountry of interest from the World Bank’s Enterprise Surveys (table I.1 + γlntrade + δ1lncapitalcst + δ2lnempcst in appendix I). The measure of intra-industry FDI + Dummycountry of interest + Dcs + Dst + Dct + εcst spillovers in the strict sense captures only horizontal where the sign of the coefficient β3 indicates whether spillovers, but because sectors are defined at a broad the impact of GVC integration is higher or lower in level, FDI spillovers are likely to capture some verti- the country of interest than for the rest of the coun- cal spillovers. For example, the rubric “auto and auto try sample (β1). The sign of the coefficient β4 reveals components” includes manufacturers of final auto- whether the mediating effect of policy in the country motive products and suppliers of automotive com- of interest is more positive or negative than for the ponents, so FDI in this sector could affect domestic rest of the country sample (β2). final producers of cars as well as domestic suppliers of auto components. Similar situations are likely in GVC Participation and Firm-Level Productivity: such sectors as food, electronics, and chemicals and Mediating Factors pharmaceuticals. Integrating a country’s domestic firms into GVCs Within-Industry Impact of FDI not only increases the possibility for productivity Data constraints make it difficult to perform direct test- gains through supplying to a multinational firm in ing of the channels for FDI spillovers to the wider host the country, but also through exporting to a buyer economy. Farole and Winkler (2014) have developed abroad. In addition, countries should not neglect the an econometric analysis to assess how foreign inves- opportunities for productivity gains that GVC par- tor characteristics, domestic firms’ absorptive capacity, ticipation can provide from importing inputs that and a country’s institutional variables influence intra- contain knowledge and technology. Alternatively, industry productivity spillovers to domestic firms from the model could also include firm-level measures of FDI (annex 7B). The method focuses on the within- GVC integration as right-hand side variables, such as industry impact of foreign output share on domestic the share of exports in a firm’s total output or the firm productivity and the role of mediating factors. share of imported inputs in total inputs. Specifically, Farole and Winkler ask, what is the poten- tial of global production networks to enhance the pro- Within-Industry Impact of Structural ductivity of domestic firms? Integration in GVCs Foreign investor characteristics are likely to cap- Similarly, the analysis can be used to examine the ture the effect of the effectiveness of intra-industry effect of GVC participation of an industry on a firm’s demand and assistance, technology spillovers that productivity by merging the Farole and Winkler Use of GVC Measures to Assess the Drivers and Impacts of GVC Participation 125 (2014) data set with two sector measures of struc- household income. Using the newly developed tural integration in GVCs: BONwin (buyer’s per- LACEX database (table BG.1.1 in box G.1 in appen- spective) and BONwout (seller’s perspective), as dix G), economists at the World Bank found that in described in chapter 6 and computed by Santoni South Africa, GVC integration has led to higher net and Taglioni (2015). Because the measures of struc- jobs but lower job intensity (Calì and Hollweg 2015; tural integration are based on the OECD-WTO TiVA Hollweg 2015). Using a social matrix accounting database, fewer observations are available, leading to approach, it appears that jobs per exports (in U.S. a total of more than 14,000 manufacturing firms in dollars) decline in GVC integration. Jobs growth 22 LMICs (table I.2 in appendix I). Alternative mea- comes through indirect links, and mainly services sures of GVC participation, such as the GVC partici- inputs (with implications for skill bias). pation index available in the TiVA database, might This section provides an overview of measures also be used. that can be used to identify the impact on labor and The baseline equation follows the estimation wages. The measures are categorized in two groups: equation in the previous section (annex 7B): indirect measures of social upgrading, and direct measures of social upgrading. lnlpirst = α + βBONcst + γ(BONcst*MF) + δ(BONcst*MF*Dummycountry of interest) + ζlncapintirst Indirect Measures of Social Upgrading + Dummycountry of interest + Dr + Ds + Dt + εirst This subsection presents indirect measurements of lnlpirst denotes the log labor productivity for domes- the link between GVC participation and labor mar- tic firm i in region r, country c, sector s at time t; ket outcomes. The specific sectors that are relevant lncapintirst denotes capital intensity in natural loga- for participation in GVCs can be analyzed using the rithms. The key variable of interest is the interaction following methods. effect between the measure of structural integration in GVCs (BON) in country c and sector s at time t Descriptive Statistics and the “mediating factors” (MF) that are specific Descriptive statistics may be used to assess which to the country of interest (see annex 7B for the list sectors are associated with better labor market out- of mediating factors and their definitions). This comes. The researcher would examine countries’ sec- term is indicated in the equation as δ(BONcst*MF tor averages of the number of employees, wages and *Dummycountry of interest). salaries, wage rate (wages and salaries divided by the The total impact of GVC integration—taking into number of employees), or labor share (wages and account the mediating role of absorptive capacity salaries as a percentage of value added). Such statis- and host country characteristics—on firms located tics, for example, can be obtained from the United in the country can be obtained as follows: Nations Industrial Development Organization’s Industrial Statistics database. β + γMF + δ(MF*Dummycountry of interest) A topic of interest is how the impact of mediat- Analysis of Employment-Generating Industries and ing factors in the country of interest differs from the Their Level of GVC Integration impact of the whole sample of countries in the data Analysis of employment-generating industries and set—that is, the effects have to be interpreted relative their level of GVC integration may be carried out to the full sample. Annex 7C shows an application of by running cross-country “controlled correlations” this model to Bulgaria. The results are described in at the sector level, whereby the labor market indica- chapter 2. tors discussed in the previous section are regressed on indicators of GVC involvement while controlling for other factors, such as region and gross domes- Quantifying the Labor Market tic product. The analyst can also run pooled regres- Dimension of GVCs sions controlling for industry fixed effects to see The last part of this chapter addresses which GVC- which industries have more labor-market-enhanc- oriented industries have a higher demand for labor, ing outcomes conditional on GVC involvement. such that integrating into GVCs in those sectors The sector-level GVC indicators may include the has a greater potential to create jobs and increase following: 126 Making Global Value Chains Work for Development • Network measure of structural integration in can be applied across countries and industries if the GVCs, to assess if centrality, clustering, closeness, data are available. or strength in a network has an effect on labor market indicators Labor Content of Gross Exports • Domestic value added in gross exports, to con- The first direct measure of social upgrading is the sider whether value-added generation in exports labor content of gross exports. The newly devel- can be associated with positive labor market oped World Bank data set on LACEX can be used outcomes to explore the social upgrading linked to GVC par- • GVC participation index or its individual com- ticipation.2 The data set is computed on the basis of ponents, to see whether countries that participate the social accounting matrix data available in the more as buyers and sellers generally have better Global Trade Analysis Project for intermittent years labor market results between 1995 and 2011. The matrix includes data • Length of the sourcing chain—either domestic for more than 100 countries and 24 or 57 sectors and international segments taken individually or (see appendix G). both combined—to see how a greater number Two cases illustrate successful GVC insertion in of production stages is related to labor market the past two decades: (1) Chinese machinery and outcomes equipment and (2) Indian private services. The • Upstreamness or distance to final demand, to see former contains non-transport machinery, includ- whether a country’s position in the value chain ing the electronics sector; the latter contains mainly matters for labor market results. information technology (IT)–enabled services, including back office and IT services exports. China’s labor value added in the machinery and Direct Measures of Social Upgrading equipment sector has expanded dramatically over This subsection presents more direct measurements time, particularly its backward link component of the link between GVC participation and labor (figure 7.4, panel a). This finding is confirmed by market outcomes by drawing on various indicators the ratio of backward to direct labor value added already developed in the literature. The indicators in exports, which has increased rapidly since 1997 Figure 7.4. Labor Value Added in Chinese Machinery and Equipment Exports, 1995–2011 a. Labor value added of exports b. Total over direct labor value c. Labor value added per US$1 d. Labor intensity of exports (US$, thousands) added of exports of exports (US$) compared to the rest of the world 350,000 3.5 0.40 1.2 300,000 0.35 1.0 3.0 250,000 0.30 0.8 200,000 0.25 2.5 150,000 0.20 0.6 100,000 0.15 2.0 0.4 50,000 0.10 0 1.5 0.05 0.2 19 5 97 01 04 07 11 19 5 97 01 04 07 97 01 04 07 11 19 5 97 19 5 11 01 04 07 11 9 9 9 9 19 20 20 20 20 19 20 20 20 20 19 19 20 20 20 20 20 20 20 20 Backward links Forward links Direct labor value added Source: Adapted from Calì and others forthcoming. Use of GVC Measures to Assess the Drivers and Impacts of GVC Participation 127 Figure 7.5. Labor Value Added in Indian Other Private Services Exports, 1995–2011 a. Labor value added of exports b. Total over direct labor value c. Labor value added per US$1 d. Labor intensity of exports (US$, thousands) added of exports of exports (US$) compared to the rest of the world 30,000 3.0 1.2 1.5 25,000 1.0 2.5 1.2 20,000 0.8 2.0 0.9 15,000 0.6 1.5 0.6 10,000 0.4 5,000 1.0 0.2 0.3 19 5 97 01 04 07 11 19 5 97 01 04 07 11 19 5 11 97 01 04 07 11 19 5 97 01 04 07 9 9 9 9 19 20 20 20 20 19 20 20 20 20 19 20 20 20 20 20 19 20 20 20 Backward links Forward links Direct labor value added Source: Adapted from Calì and others forthcoming. (figure 7.4, panel b)—unlike the forward links-direct declined for each of the three measures of labor value ratio, which has remained constant. This suggests added since 1995 (figure 7.5, panel c). In particular, that China has increased its domestic production the total labor content of exports on the basis of for- in the sectors providing inputs for final exports of ward links almost halved, from US$1.1 per US$1 of machinery. The increase has also translated into an exports in 1995 to US$0.6 in 2011. But the direct and increase in the share of domestic labor value added in total labor content of exports on the basis of back- exports. The total labor content of machinery exports ward links have increased since 1997, also relative to in backward links increased from US$0.23 per US$1 the rest of the world (figure 7.5, panel d). of exports in 1995 to almost US$0.4 in 2011 (figure 7.4, panel c). In other words, each $100 of machin- Labor Component of Domestic Value Added ery exports generated $40 of wages in the economy in Exports (in green), only $11 of which is a result of the direct A second direct measure of social upgrading, which labor in final production (in black). The increase has was developed by the United Nations Conference been much milder for direct and forward links. The on Trade and Development (UNCTAD 2013), is the increase in the labor intensity of China’s machinery labor cost component of domestic value added in exports also has been more marked relative to the exports, which acts as a proxy for the employment- rest of the world (figure 7.4, panel d). generating potential of exports. Using the United For India’s other private services exports, the Nations Conference on Trade and Development direct labor value added and total labor value added (UNCTAD) EORA GVC database for 187 coun- on the basis of forward links are more relevant than tries, countries are ranked according to their 2010 the value added generated through backward links GVC participation rates in decreasing order (figure (figure 7.5, panel a). Over time, the direct labor con- 7.6). The labor component of domestic value added tent of exports has grown more rapidly than the in exports increases with higher GVC participa- labor content of forward links (figure 7.5, panel b), tion: it reaches 43 percent of value added in exports but neither has grown relative to the value of exports. for countries with the highest GVC participation The labor content for each $100 of exports has rate (first quartile), compared with 28 percent for 128 Making Global Value Chains Work for Development Figure 7.6. GVC Participation and the Labor Jobs Sustained by Foreign Final Demand Component of Domestic Value Added in Exports The third indicator, jobs sustained by foreign final Median share of the labor cost component demand, is being developed by OECD-WTO as part in the domestic value added of the TiVA database for 40 countries.3 The indicator 1st quartile calculates the number of jobs in the total economy sustained by foreign final demand, which captures GVC participation From low to high 2nd quartile the full upstream impact of final demand in foreign markets on domestic employment. Rather than con- 3rd quartile sider the domestic value added in total exports (as 4th quartile was the basis of the previous indicator), which could be used as intermediates in third countries and be 20 25 30 35 40 45 50 exported as final goods, the indicator considers the Percent domestic value added in foreign final demand. Sources: UNCTAD 2013; UNCTAD-EORA GVC database. Note: The data are for 187 countries ranked according to the 2010 GVC participa- Between 1995 and 2008, a higher share of tion rate in decreasing order and grouped in quartiles. Median values of the employment consisted of jobs sustained by foreign quartiles are reported. GVC = global value chain; UNCTAD = United Nations Conference on Trade and Development. final demand (figure 7.8), yet that percentage varies according to countries’ size and specialization. For example, based on preliminary estimates, the share countries with the lowest GVC participation rate for Germany almost doubled between 1995 and 2008, (fourth quartile). with about 10 million jobs sustained by foreign final In addition, countries with faster growth in GVC demand. For China, the number increased by about participation have faster growth in the labor compo- two-thirds, from 89 million to 146 million. These fig- nent of domestic value added in exports (figure 7.7). ures are averages for the whole economy, including From 2000 to 2010, the countries that experienced services sectors with lower exposure to international fast growth in GVC participation saw the labor com- trade, but they can also be disaggregated by industry. ponent of exports rise faster (14 percent) than did For example, about one-third of U.S. jobs in elec- countries with slow growth (9 percent). The relation- tronics and almost two-fifths of Japanese jobs are ship holds even when country participation in GVCs derived from foreign final demand. depends on higher foreign value added share, which reduces the share of domestic value added of exports. Jobs Generated by Foreign Trade in GVCs The fourth indicator is the number of jobs generated Figure 7.7. Growth in the Labor Component of Domestic Value Added in Exports by Level of GVC Participation by a country’s trade in GVCs—jobs generated domes- Growth and Foreign Value Added tically and abroad—using the World Input-Output Database (WIOD) for 39 countries over the period Countries with 1995–2009. The sources of employment creation rapidly growing from international trade can be decomposed into GVC participation five components: (1) exports, (2) imports, (3) import content of exports, (4) export content of imports, Countries with and (5) intermediates contained in imports.4 The slow growth of first two components are labor demand from final GVC participation goods trade; the last three are from trade in interme- 0 4 8 12 16 20 diates, or the result of a country’s GVC participation. Percent A country’s participation in GVCs can lead to Countries using Countries using domestic or foreign labor demand. Because of the more foreign less foreign import content of exports, a country’s exports gener- value added value added ate jobs and incomes in foreign countries. Likewise, Sources: UNCTAD 2013; UNCTAD-EORA GVC database. a country’s imports from foreign countries might Note: The data are for 187 countries. “Countries with rapidly growing GVC participation” refers to the 50 percent of countries with the highest 2000–10 GVC contain its own exports to those foreign countries as participation growth rate. “Countries using more foreign value added” refers to intermediate inputs. Because of the export content the 50 percent of countries with the highest foreign value-added share in exports in 2010. GVC = global value chain; UNCTAD = United Nations Conference on Trade of imports, a country’s imports generate jobs domes- and Development. tically. Given third-party intermediates contained in Use of GVC Measures to Assess the Drivers and Impacts of GVC Participation 129 imports, trade between two countries will, in turn, Figure 7.8. Jobs in the Business Sector Sustained by Foreign Final create jobs in the third country. Therefore, the total Demand, 1995 and 2008 domestic labor demand can be viewed for each Luxembourg country as the sum of labor demand by domestic Estonia exports and domestic content of imports. The sum Ireland of the remaining components yields the total foreign Slovak Republic labor demand resulting from each country’s trade Hungary position. Belgium The jobs generated by each component can be Slovenia computed for various industries. The sector-level Czech Republic Switzerland information has been aggregated to a single employ- Sweden ment figure for each country (table 7.1).5 Large Austria HICs tend to be the most responsible for GVC- Denmark based labor demand, with China, France, Germany, Netherlands the Netherlands, and the United States having the Norway greatest labor demand resulting from GVC partici- Finland pation. In 2009, most of the countries in the sample Germany demanded more foreign labor than domestic labor Poland through exports, with the exceptions including Korea, Rep. China, India, and Indonesia. Italy China Jobs in GVC Manufacturing Portugal Canada The fifth indicator is for selected countries between France 1995 and 2008, using the WIOD. The jobs in the Bulgaria GVC manufacturing indicator presents a broader Russian Federation picture of the structure of employment in GVCs Turkey within a country. It is the most direct measure in the Spain literature of the domestic employment impacts of Greece manufacturing GVC participation (table 7.2).6 The India indicator measures—directly and indirectly—the Mexico number of GVC jobs involved in the production of Indonesia final manufacturing goods (also known as manu- Japan factures), as well as their sector of employment in Australia a country. United States Apart from China and Turkey, the share of manu- Brazil facturing GVC jobs in overall employment declined, 0 10 20 30 40 50 60 70 80 driven by manufacturing GVC job losses in agricul- % of total employment ture and manufacturing. Only about one-half of the 1995 2008 workers in manufacturing GVCs are employed in Source: OECD 2013, figure 7.8.1. manufacturing; the other half are employed in non- Note: The business sector is defined according to International Standard Industrial Classification manufacturing industries that deliver intermediates. Rev. 3, Divisions 10 to 74, that is, total economy excluding agriculture, forestry, and fishing (Divisions 01-05), public administration (75), education (80), health (85), and other community, social, and At the same time, employment in manufactur- personal services (90–95). ing GVCs increased in the services sector. For some European countries, such as Germany, Italy, and Spain, GVC job increases in services were higher by educational attainment.7 HICs in the European than job losses in manufacturing and agriculture, but Union experienced a strong shift toward specializa- that trend was not apparent in other countries. tion in GVC activities performed by high-skilled Changes in the skill structure of GVC manufac- workers. Relative to the overall labor force, the share turing workers and their average wages have been of high-skilled workers in total GVC employment analyzed—between 1995 and 2008—and include increased much faster than the share of medium- low-, medium-, and high-skilled workers, proxied skilled workers. 130 Making Global Value Chains Work for Development Table 7.1. Jobs Generated by Five Components of Foreign Trade, 2009 (thousands) Domestic Foreign labor Export Third-party Differences content of Import content imports in (domestic minus Country Exports imports Imports of exports imports foreign labor) China 140,249.1 3,270.9 17,462.8 4,221.9 2,238.0 119,597.4 India 34,914.8 89.6 8,064.4 1,291.5 496.6 25,151.9 Indonesia 10,236.6 24.0 3,891.8 448.4 289.0 5,631.4 Brazil 7,143.3 21.9 3,210.6 168.8 486.7 3,299.0 Bulgaria 882.3 1.4 465.3 97.9 98.2 222.4 Romania 1,597.0 6.0 1,097.3 186.6 293.7 25.4 Latvia 162.2 0.7 161.0 23.1 51.4 –72.5 Estonia 160.1 0.3 155.0 50.5 39.2 –84.3 Malta 45.1 0.0 119.0 33.9 23.8 –131.5 Cyprus 34.8 0.0 143.4 14.1 35.4 –158.1 Lithuania 250.5 1.0 383.8 102.7 68.5 –303.5 Slovenia 223.8 0.4 345.2 113.5 106.5 –340.9 Mexico 6,054.1 46.7 4,317.6 1,590.4 848.1 –655.2 Portugal 797.8 4.2 1,122.8 218.7 353.3 –892.8 Slovakia 738.4 4.9 977.2 458.0 264.7 –956.6 Poland 3,592.6 26.9 3,149.1 911.0 747.0 –1,187.6 Hungary 1,129.2 5.8 1,349.1 713.2 417.8 –1,345.1 Finland 433.5 2.0 1,644.0 449.7 323.2 –1,981.4 Czech Republic 1,674.7 15.9 2,176.4 993.2 544.1 –2,023.2 Turkey 2,056.6 6.2 3,146.6 456.5 506.2 –2,046.5 Greece 204.9 0.8 1,807.2 83.4 386.6 –2,071.5 Denmark 529.4 3.4 1,974.9 463.1 542.5 –2,447.7 Taiwan, China 3,119.7 23.2 3,807.2 1,681.9 517.2 –2,863.4 Russian Federation 6,532.3 47.3 8,398.5 225.3 897.5 –2,941.7 Ireland 578.8 2.4 2,278.2 897.9 440.0 –3,034.9 Sweden 828.5 6.7 2,520.9 697.5 694.6 –3,077.8 Austria 942.3 8.9 2,575.1 734.4 739.2 –3,097.4 Belgium 1,325.9 17.3 4,281.9 1,793.5 1,326.9 –6,059.2 Australia 1,081.5 5.4 7,268.1 470.9 563.1 –7,215.2 Spain 2,300.8 30.6 7,774.1 1,050.5 1,385.3 –7,878.4 Italy 3,427.0 45.6 9,109.3 1,437.0 1,891.9 –8,965.6 Canada 2,718.2 34.0 10,140.8 1,489.8 1,421.4 –10,299.8 Korea, Rep. 3,812.6 35.9 11,020.0 2,521.8 841.1 –10,534.4 France 3,114.5 70.5 11,471.2 1,898.5 2,674.1 –12,858.8 Netherlands 2,397.5 31.2 10,891.6 3,845.3 1,189.4 –13,497.7 England 3,897.1 80.0 15,583.6 1,746.0 2,499.5 –15,852.0 Japan 3,871.4 65.6 20,451.8 1,483.2 1,495.4 –19,493.2 Germany 8,473.3 366.8 22,449.3 5,591.3 4,619.4 –23,819.8 United States 6,851.7 510.9 61,198.0 3,101.0 6,484.2 –63,420.6 Total 268,383.9 4,915.2 268,383.9 43,755.9 38,840.7 –77,681.4 Sources: Jiang and Milberg 2013, 5, based on data from the World Input-Output Database. Summary analysis to establish causality and dig into the driv- ers and impacts of GVC links at the micro level. For This chapter focused on two basic questions in many LMICs, however, firm-level data are not avail- analyzing the relationship between GVC links and able. Acknowledging this challenge, this chapter pro- the domestic economy: what are the determinants posed several methodologies and data sources that of GVC integration, and what are its impacts? The can be used and combined, depending on the specific quality of the assessment depends on the method- country context and data availability. ology that is applied, which, in turn, depends heav- Regarding the determinants, the chapter began ily on data availability. A fully developed regression with the decomposition of gross export growth, model using firm- or sector-level data, for example, which serves as a first assessment of where the is preferable to bivariate correlations using country- growth of the value added embodied in gross exports level data, because the regression model allows the is generated in terms of the country of origin (that is, Use of GVC Measures to Assess the Drivers and Impacts of GVC Participation 131 Table 7.2 Manufacturing GVC Workers, 1995 and 2008 Manufacturing GVC workers Change in manufacturing GVC (% of all workers Manufacturing GVC workers in workers between 1995 and 2008 in the economy) 2008 (thousands) employed in (thousands) employed in All All Country 1995 2008 Agr. Mfg. Serv. sectors Agr. Mfg. Serv. sectors United States 16 11.1 1,143 8,837 6,892 16,872 –331 –3,144 –1,138 –4,612 Japan 22.6 19.4 1,298 6,491 4,417 12,207 –794 –2,225 148 –2,871 Germany 26.8 26.4 400 5,481 4,766 10,647 –161 –666 1,388 561 France 22 18.7 303 2,195 2,355 4,853 –96 –423 368 –151 United Kingdom 20.1 12.6 115 1,946 1,931 3,992 –128 –1,148 –347 –1,624 Italy 29.1 25.5 333 3,553 2,559 6,444 –192 –234 517 91 Spain 23.2 17.5 271 1,827 1,494 3,592 –97 185 353 440 Canada 20.8 16 157 1,138 1,482 2,777 –102 –136 193 –45 Australia 18.2 14.5 165 641 855 1,661 –48 3 196 150 Korea, Rep. 29.7 22.8 655 2,646 2,077 5,378 –468 –735 524 –679 Netherlands 22.8 19 89 643 929 1,661 –42 –87 158 29 China 31.7 33.3 121,342 87,568 49,468 258,378 9,963 20,508 11,965 42,436 Russia 24.7 21.9 4,259 6,749 6,228 17,237 –1,403 –2,120 2,198 –1,325 Brazil 29.6 28.7 8,347 9,490 9,823 27,660 –705 2,450 4,118 5,863 India 27.9 27.3 57,926 41,933 26,483 126,343 2,118 10,896 7,025 20,039 Mexico 30.3 24.4 2,817 6,128 3,205 12,150 –400 1,403 1,121 2,124 Turkey 27.1 30.4 1,778 3,115 1,554 6,446 –341 620 584 863 Indonesia 32.1 25.6 13,921 7,427 5,725 27,073 –1,899 –425 1,380 –944 Sources: Timmer and others 2014, appendix table 5; based on data from the World Input-Output Database. Note: GVC = global value chain. foreign versus domestic) and sector (within a sector integration at the sector level, including measures of or in upstream sectors). The section then looked in structural integration in GVCs, foreign value added more detail at the determinants of GVC integration embodied in gross exports, domestic value added at the country, sector, and firm levels. At the country embodied in exports of third countries, upstream- level, logistics performance, skill levels, and access ness, length of sourcing chains, and the share of to knowledge (measured as geographical distance to foreign output in total output. If firm-level data are knowledge centers) were shown to be important fac- available, the analysis could also include firm-level tors for a country’s extent of GVC links. Sector-level measures of GVC integration as right-hand-side determinants can include the existence of an export variables, such as the share of exports in a firm’s processing zone, openness to international trade and total output or the share of imported inputs in total FDI, connectivity to international markets, competi- inputs. tiveness in unit labor costs and labor productivity, In addition to economic upgrading, the chapter drivers of investment, and the quality of domestic also looked at the impact of GVC integration on value chains and the services infrastructure. Firm- social upgrading. Again, the analysis can assess the level determinants include size, age, foreign owner- impact of the GVC-related measures listed previ- ship status, workers’ skills, and productivity. ously on employment, wages and salaries, or the As for the impacts of GVC integration, the chapter labor share, using regression analysis. The next step differentiated between economic and social upgrad- was to present a more direct way to measure the ing. Economic upgrading is captured by measur- link between GVC participation and labor market ing the growth of domestic value added embodied outcomes, by drawing on various indicators already in gross exports, the level of domestic value added developed in the literature, which are based on inter- or productivity (for example, labor productivity or national input-output data. total factor productivity), which serve as dependent Finally, other impacts that were not explicitly cov- variables in the analyses presented. Although the first ered in this chapter but warrant further investigation is only available at the sector level, the second and include the impact of GVCs on the macro economy, third can also be studied using firm-level data. Using as well as the links with working conditions, edu- regression analysis, measures of economic upgrad- cation and skills, other aspects of the society (for ing were then related to various measures of GVC instance, poverty), and the environment. 132 Making Global Value Chains Work for Development Annex 7A. Regression Results middle-income countries (but cannot control for capital stock and employment), the interaction term Focusing on global value chain (GVC) integration is positive and significant (table 7A.2). This result as a buyer, table 7A.1 shows the results for the full may be indicative of a lower policy threshold effect sample of 40 countries in the World Input-Output in emerging countries; that is, better infrastructure Database (WIOD) covering the period 1995–2011. matters more. For a description of the data, see box 7A.1. Column Higher general infrastructure investment (table 1 in table 7A.1 shows the results based on the esti- 7A.1, column 3), by contrast, shows the expected mation equation. The production factors labor and positive mediating effect. Better rail coverage and capital significantly increase domestic value added. more investment in rails (columns 4 and 5) also pos- The same applies to the trade-related variables for- itively mediate the effect. A higher value of air cargo eign value added for domestic processing (trade) (column 6) has no effect, while more investment and, most important, our GVC measure of foreign in airports (column 7) shows a negative mediating value added embodied in exports (EXGR_FVA). A effect. The latter effect is surprising, but could indi- 10 percent increase in GVC integration as a buyer cate that better airports may act as a driver to source is expected to increase domestic value added by 0.7 more inputs internationally (or to offshore more percent. In other words, GVC integration as a buyer inputs), which could reduce domestic value added helps increase a country’s domestic value added and if foreign production factors substitute for domestic thus fosters economic upgrading. ones. A higher value of air cargo positively mediates Next, we assess the role of national character- the impact of GVC integration as a buyer using the istics. We first focus on a country’s infrastructure OECD data set (table 7A.2), which again may indi- (table 7A.1, columns 2 to 8). Surprisingly, better cate a lower policy threshold effect. Finally, more logistics performance negatively mediates the impact investment in roads shows a positive effect (table of GVC integration as a buyer on value added (col- 7A.1, column 8). umn 2). With the use of the OECD data set instead, Second, we look at the role of foreign direct which includes more emerging and low- and investment (FDI), which shows the expected positive Box 7A.1. Data: GVC Indicators and Policy Variables GVC Indicators transport-related infrastructure) and total infrastructure invest- To calculate the global value chain (GVC) indicators, we rely on ment (percentage of gross domestic product [GDP]) capture the two databases that provide inter-country input-output (ICIO) general quality of infrastructure. Kilometers of rail lines (per per- tables, which allows us to track value-added flows across coun- son) and investment in rail infrastructure (percentage of GDP), as tries and industries. The first database is the World Input-Output well as the value of air cargo (percentage of GDP), investment in Database (WIOD). It covers 40 countries, 35 industries, and the airport infrastructure (percentage of GDP), and investment in road years 1995–2011. A major advantage of WIOD is that it also pro- infrastructure (percentage of GDP) capture specific aspects of a vides data on price levels, employment, and capital stocks at the country’s infrastructure. industry level. This allows us to include the controls we need and The investment data are taken from the OECD transport to deflate the level variables (using 1995 as the base year). database, but are only available for 35 countries in the sample. The second database is the Organisation for Economic Co- Therefore, we only include them in the WIOD regressions. The operation and Development (OECD) ICIO database, which covers remaining variables are taken from the World Development Indi- 61 countries, 34 industries, and the years 1995, 2000, 2005, and cators (WDI) and are available for all countries except Taiwan, 2008–11. The advantage here is the extended country coverage, China.. especially of low- and middle-income countries. However, since To proxy for foreign presence and innovative capabilities, we no additional variables are provided, we have to rely on our trade use WDI data on foreign direct investment inflows (percentage measure and fixed effects to get clean estimates. Therefore, the of GDP), research and development (R&D) intensity, and patent WIOD estimates represent our benchmark. applications (per person). In addition, we use OECD data on busi- ness sector R&D spending (percentage of GDP) for the WIOD Policy Variables sample. Finally, we rely on rule-of-law data from the World Gover- We use seven proxies to measure a country’s infrastructure: the nance Indicators to measure legal institutions. infrastructure Logistics Performance Index (quality of trade- and Table 7A.1. GVC Integration as a Buyer and Domestic Value Added, National Characteristics, 1995–2011 Variable (1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (11) (12) (13) EXGR_FVA 0.0713*** 0.443*** 0.0182 –0.0033 0.0196 0.0742*** 0.178*** 0.0409 0.0582*** 0.0624*** 0.121*** 0.111*** 0.0884*** (0.0178) (0.0886) (0.0363) (0.0150) (0.0373) (0.0186) (0.0339) (0.0364) (0.0199) (0.0170) (0.0268) (0.0312) (0.0202) Employment 0.346*** 0.339*** 0.362*** 0.289*** 0.337*** 0.342*** 0.332*** 0.360*** 0.338*** 0.342*** 0.347*** 0.357*** 0.341*** (0.0585) (0.0560) (0.0622) (0.0570) (0.0684) (0.0579) (0.0680) (0.0615) (0.0582) (0.0580) (0.0584) (0.0898) (0.0574) Capital stock 0.287*** 0.290*** 0.302*** 0.292*** 0.310*** 0.302*** 0.322*** 0.307*** 0.301*** 0.287*** 0.283*** 0.312*** 0.298*** (0.0393) (0.0376) (0.0474) (0.0386) (0.0481) (0.0389) (0.0496) (0.0473) (0.0384) (0.0390) (0.0387) (0.0570) (0.0384) Trade 0.164*** 0.173*** 0.168*** 0.173*** 0.178*** 0.175*** 0.177*** 0.168*** 0.173*** 0.162*** 0.166*** 0.159*** 0.179*** (0.0221) (0.0214) (0.0221) (0.0234) (0.0261) (0.0225) (0.0246) (0.0221) (0.0226) (0.0218) (0.0219) (0.0294) (0.0224) EXGR_FVA*infrastructure LPI –0.116*** (0.0261) EXGR_FVA*infrastructure investment 0.0984** (0.0403) EXGR_FVA*Rail line coverage 0.0003*** (4.36e–05) EXGR_FVA*rail investment 42.81*** (15.88) EXGR_FVA*air cargo 520,633 (538,886) EXGR_FVA*airport investment –189.6*** (68.40) EXGR_FVA*road investment 12.32* (6.815) EXGR_FVA*FDI inflows 0.0038*** (0.0014) EXGR_FVA*contract enforcement 0.0216 (0.0178) EXGR_FVA*total R&D intensity –0.0436*** (0.0165) EXGR_FVA*business R&D intensity –4.52e+06*** (1.75e+06) EXGR_FVA*patent applications –7.99e-05*** (2.10e-05) Constant 3.554*** 1.700*** –3.013*** 2.516*** –4.138*** 3.615*** –4.261*** 5.535*** 0.171 0.234 1.415*** –0.557 3.947*** (0.462) (0.473) (0.691) (0.523) (0.739) (0.470) (0.789) (0.407) (0.511) (0.525) (0.527) (0.816) (0.473) Observations 11,253 10,953 9,794 10,410 9,518 10,953 8,958 9,794 10,953 11,253 11,253 8,649 10,953 R–squared 0.869 0.876 0.872 0.880 0.874 0.874 0.874 0.872 0.875 0.869 0.870 0.853 0.875 Source: Data are from the World Input-Output Database. Note: The dependent variable is domestic value added. EXGR_FVA is lagged. Robust standard errors are clustered at the industry-country level. All level variables are in natural logarithms. Country-year, industry-year, and industry-country fixed effects are included. EXGR_FVA = foreign value added embodied in exports; FDI = foreign direct investment; GVC = global value chain; LPI = Logistics Performance Index; R&D = research and development. *** p < 0.01, ** p < 0.05, * p < 0.1. Use of GVC Measures to Assess the Drivers and Impacts of GVC Participation 133 134 Making Global Value Chains Work for Development Table 7A.2. GVC Integration as a Buyer and Domestic Value Added, National Characteristics, Selected Years, 1995–2011 Variable (1) (2) (3) (4) (5) (6) (7) (8) EXGR_FVA 0.0874*** –0.0136 0.0586*** 0.0751*** 0.0701*** 0.0709*** 0.0404* 0.0888*** (0.0128) (0.0582) (0.0181) (0.0136) (0.0141) (0.0146) (0.0206) (0.0143) Trade 0.286*** 0.287*** 0.297*** 0.284*** 0.284*** 0.283*** 0.273*** 0.271*** (0.0201) (0.0212) (0.0254) (0.0202) (0.0203) (0.0202) (0.0242) (0.0221) EXGR_FVA*infrastructure LPI 0.0333* (0.0178) EXGR_FVA*rail line coverage 5.14e-05 (4.39e-05) EXGR_FVA*air cargo 2.08e+06*** (563,500) EXGR_FVA*FDI inflows 0.0026*** (0.000761) EXGR_FVA*contract enforcement 0.0309** (0.0124) EXGR_FVA*total R&D intensity 0.0369*** (0.0126) EXGR_FVA*patent applications 1.03e-05 (2.13e-05) Constant 5.339*** 5.476*** 5.552*** 5.444*** 5.419*** 5.462*** 5.584*** 5.912*** (0.168) (0.164) (0.187) (0.152) (0.138) (0.149) (0.170) (0.165) Observations 8,488 8,235 7,408 8,348 8,348 8,488 8,098 8,096 R-squared 0.852 0.857 0.869 0.855 0.855 0.853 0.848 0.855 Source: Data are from the Organisation for Economic Co-operation and Development database. Note: The dependent variable is domestic value added. EXGR_FVA is lagged. Robust standard errors are clustered at the industry-country level. All level variables are in natural logarithms. Country-year, industry-year, and industry-country fixed effects are included. EXGR_FVA = foreign value added embodied in exports; FDI = foreign direct investment; GVC = global value chain; LPI = Logistics Performance Index; R&D = research and development. *** p < 0.01, ** p < 0.05, * p < 0.1. sign (column 9). More foreign presence increases the our GVC indicator. Table 7A.3 focuses on the effects gains from importing foreign value added for domes- using the full WIOD country sample. GVC integra- tic value added, possibly because of more efficient tion as a seller substantially increases domestic value distribution channels and/or investment put in place added (column 1). The elasticity is higher than those by foreign firms. Better contract enforcement has no of all other control variables, while it was smaller mediating effect (column 10), while the mediating when using the amount of foreign value added in impact is positive in the OECD country sample. This exports (see table 7A.1), indicating that being a seller could again suggest that in emerging countries better in GVCs contributes more strongly to boost eco- contract enforcement matters more. nomic upgrading than being a buyer only. Third, we assess the role of a country’s innovation In the next step, we assess whether certain coun- capacity. Surprisingly, the interaction terms with all try characteristics influence the results. First, we look three measures of innovation are negative and signifi- at the moderating role of infrastructure (columns cant (table 7A.1, columns 11 to 13). One reason could 2 to 8). Surprisingly, only airport-related indicators be that GVC integration has a smaller positive impact matter. A higher value of air cargo shows a positive on domestic value added in more innovative coun- impact (column 6), while more investment in air- tries, possibly because domestic production factors ports negatively mediates the effect (column 7). The contribute relatively more to value added. This expla- latter effect is surprising, but confirms the findings nation seems to be supported by the findings using from GVC integration on the buying side (see table the OECD country sample that includes more emerg- 7A.1). These results could indicate that better air- ing countries (table 7A.2), where a higher research ports may act as a driver to source more inputs inter- and development intensity positively mediates the nationally (or to offshore more inputs) to be used relationship. More patents also show a positive coef- in a country’s export products, which could reduce ficient sign, but are not statistically significant. domestic value added if foreign production factors We now focus on GVC integration from a seller’s substitute for domestic ones. A higher value of the perspective, using the amount of domestic value Logistics Performance Index and rail line coverage added re-exported by third countries (DVA3EX) as positively mediates the impact of GVC integration as Table 7A.3. GVC Integration as a Seller and Domestic Value Added and the Role of National Characteristics, 1995–2011 Variable (1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (11) (12) (13) DVA3EX 0.289*** 0.286*** 0.298*** 0.248*** 0.274*** 0.286*** 0.258*** 0.296*** 0.284*** 0.293*** 0.287*** 0.275*** 0.285*** (0.0516) (0.0510) (0.0557) (0.0533) (0.0590) (0.0512) (0.0585) (0.0542) (0.0512) (0.0521) (0.0509) (0.0746) (0.0508) Employment 0.243*** 0.257*** 0.249*** 0.242*** 0.251*** 0.259*** 0.261*** 0.252*** 0.260*** 0.240*** 0.242*** 0.235*** 0.256*** (0.0378) (0.0376) (0.0451) (0.0396) (0.0460) (0.0371) (0.0456) (0.0447) (0.0369) (0.0381) (0.0377) (0.0513) (0.0374) Capital stock 0.145*** 0.159*** 0.151*** 0.151*** 0.155*** 0.158*** 0.148*** 0.151*** 0.159*** 0.144*** 0.141*** 0.135*** 0.161*** (0.0210) (0.0211) (0.0208) (0.0223) (0.0230) (0.0212) (0.0220) (0.0207) (0.0213) (0.0207) (0.0200) (0.0249) (0.0210) Trade 0.193*** 0.328 0.143 0.239*** 0.206** 0.186*** 0.402*** 0.168* 0.161*** 0.215*** 0.163*** 0.287*** 0.199*** (0.0424) (0.202) (0.0907) (0.0690) (0.0801) (0.0429) (0.0606) (0.0988) (0.0458) (0.0525) (0.0627) (0.0688) (0.0456) DVA3EX*infrastructure LPI –0.0466 (0.0595) DVA3EX*infrastructure investment 0.0866 (0.0677) DVA3EX*rail line coverage –1.9e–05 (0.0002) DVA3EX*rail investment 13.23 (30.03) DVA3EX*air cargo 2.0e+06** (789,801) DVA3EX*airport investment –341.7** (153.6) DVA3EX*road investment 10.76 (12.17) DVA3EX*FDI inflows 0.0055*** (0.00191) DVA3EX*contract enforcement –0.0318 (0.0364) DVA3EX*total R&D intensity 0.0360 (0.0391) DVA3EX*business R&D intensity –1.1e+06 (6.3e+06) DVA3EX*patent applications –6.2e–05 (5.0e–05) Constant 3.554*** –2.701*** 1.801*** 5.862*** 4.166*** 3.836*** 3.323*** 8.067*** 0.305 1.358*** 2.601*** –0.852 –0.771 (0.462) (0.653) (0.575) (0.559) (0.508) (0.550) (0.505) (0.622) (0.482) (0.499) (0.479) (0.744) (0.521) Observations 11,253 11,012 9,845 10,460 9,568 11,012 9,008 9,845 11,012 11,312 11,312 8,680 11,012 R–squared 0.869 0.885 0.885 0.890 0.887 0.885 0.891 0.885 0.886 0.881 0.881 0.872 0.885 Source: Data are from the World Input-Output Database. Note: The dependent variable is domestic value added. DVA3EX is lagged. Robust standard errors are clustered at the industry-country level. All level variables are in natural logarithms. Country-year, industry-year, and industry-country fixed effects are included. DVA3EX = domestic value added re-exported by third countries; FDI = foreign direct investment; GVC = global value chain; LPI = Logistics Performance Index; R&D = research and development. *** p < 0.01, ** p < 0.05, * p < 0.1. Use of GVC Measures to Assess the Drivers and Impacts of GVC Participation 135 136 Making Global Value Chains Work for Development Table 7A.4. GVC Integration as a Seller and Domestic Value Added, National Characteristics, Selected Years, 1995–2011 Variable (1) (2) (3) (4) (5) (6) (7) (8) DVA3EX 0.512*** 0.270** 0.452*** 0.511*** 0.504*** 0.478*** 0.536*** 0.564*** (0.0245) (0.109) (0.0371) (0.0263) (0.0264) (0.0288) (0.0287) (0.0216) Trade 0.225*** 0.223*** 0.232*** 0.225*** 0.225*** 0.225*** 0.206*** 0.205*** (0.0142) (0.0139) (0.0165) (0.0143) (0.0143) (0.0137) (0.0130) (0.0119) DVA3EX*infrastructure LPI 0.0850*** (0.0314) DVA3EX*rail line coverage 0.0003*** (5.11e-05) DVA3EX*air cargo 700,776 (451,674) DVA3EX*FDI inflows 0.0017*** (0.0005) DVA3EX*contract enforcement 0.0564** (0.0229) DVA3EX*total R&D intensity 0.0171 (0.0178) DVA3EX*patent applications –4.94e-05 (5.45e-05) Constant 3.517*** 3.610*** 3.047*** 3.628*** 3.774*** 3.565*** 3.458*** 3.600*** (0.132) (0.128) (0.134) (0.142) (0.136) (0.144) (0.172) (0.167) Observations 8,502 8,240 7,409 8,362 8,362 8,502 8,103 8,101 R–squared 0.929 0.935 0.945 0.930 0.930 0.930 0.934 0.937 Source: Data are from the Organisation for Economic Co-operation and Development database. Note: The dependent variable is domestic value added. DVA3EX is lagged. Robust standard errors are clustered at the industry-country level. All level variables are in natural logarithms. Country-year, industry-year, and industry-country fixed effects are included. DVA3EX = domestic value added re-exported by third countries; FDI = foreign direct investment; GVC = global value chain; LPI = Logistics Performance Index; R&D = research and development. *** p < 0.01, ** p < 0.05, * p < 0.1. a seller using the OECD sample of 61 industrialized integration as a buyer. Innovation does not matter in and emerging countries (table 7A.4), which again either data set (table 7A.3, columns 11 to 13). may indicate a lower policy threshold effect. The results suggest that GVC integration as a seller Second, we find that FDI inflows clearly show leads to higher domestic value-added gains than GVC a positive mediating impact on the relationship integration as a buyer. However, national character- between GVC integration as a seller and domestic istics seem to matter less for the effect on economic value added (table 7A.3, column 9). This result con- upgrading in GVCs for sellers than for buyers, in par- firms the positive results for GVC integration on the ticular in high-income countries. This finding could buying side (see table 7A.1). The effect can also be indicate that for more advanced economies, firm- confirmed using the OECD data set (table 7A.4), level characteristics or absorptive capacities, such as underlying the positive effect of FDI in GVCs on productivity and skill intensity, are more important economic upgrading in industrialized and emerging for becoming a seller in GVCs (which would confirm countries. studies on the determinants of exporting that empha- Contract enforcement, by contrast, does not size the role of productivity). When more emerging matter in the WIOD country sample (table 7A.3, countries are included in the data set, the results sug- column 10), although contract enforcement has a gest that national policies matter more strongly for significant positive impact using the OECD country economic upgrading from both buying and selling in sample (table 7A.4). These results suggests that con- GVCs. Policies also seem to have a generally positive tract enforcement appears to be more important in impact in the enlarged country sample, suggesting a emerging countries, supporting the findings for GVC lower policy threshold effect. Use of GVC Measures to Assess the Drivers and Impacts of GVC Participation 137 Annex 7B. Factors Mediating Productivity website to communicate with clients or suppli- Spillovers from Foreign Direct Investment ers and 0 otherwise; and email = 1 if the firm uses email to communicate with clients or sup- The baseline equation, estimated by ordinary least pliers and 0 otherwise. The technology indica- squares, takes the following form: tor serves as a proxy for a sector’s average FDI lnlpirst = α + βFDIcst + γ(FDIcst*MF) technology intensity in a country. + δ(FDIcst*MF*Dummycountry of interest) + lncapintirst + Dummycountry of interest + Dr 2. Measures of absorptive capacity in the host + Ds + Dt + εirst economy: lnlpirst denotes the log labor productivity for domes- a. gap = domestic firm’s labor productivity (LP) tic firm i in region r, sector s, at time t. Foreign direct relative to median LP of multinational firms in investment (FDI) is defined as the share of foreign the sector in natural logarithms; a higher num- output as a percentage of total output at the sector ber indicates a lower gap. level in a country. b. tech = domestic firm’s technology indica- The key variable of interest is the interaction effect tor as defined in the previous section, where between the FDI variable in country c and sector s at tech ∈ {0, 1, 2, 3, 4}. The technology indicator time t and the “mediating factors” (MF), which are serves as a proxy for research and development specific to the country of interest. That term is indi- intensity, which is unavailable. cated in the equation as δ(FDIcst*MF*Dummycountry c. skills = domestic firm’s share of high-skilled of interest). To avoid spurious results for the correlation labor in the firm’s total labor force. of interest, the equation controls for a constant α, the d. size = domestic firm’s total number of per- FDI spillovers measure in country c and sector s at manent and temporary employees, in natural time t, its interaction with MF across all countries, logarithms. the level in logs of capital intensity of domestic firm e. aggl = region’s total number of manufactur- i—lncapintirst—a dummy that takes the value 1 if the ing and services firms as a percentage of the host country is the country of interest and 0 other- country’s total number of manufacturing and wise; as well as sector, region, and time fixed effects. services firms. This measure is a proxy for Standard errors are robust to heteroskedasticity and urbanization economies (locational advan- clustered at the country-sector level. tages) and covers domestic and foreign firms. The mediating factors tested are as follows: f. exp = domestic firm’s share of direct or indi- rect exports in firm sales. 1. Measures of spillover potential by the foreign firm: 3. Measures of national characteristics and a. own = a sector’s average percentage of foreign institutions: ownership in a country. b. market = a sector’s average percentage of FDI a) labor = measure of labor freedom, in natural sales to the domestic market in a country. This logarithms, from the Heritage Foundation; it measure serves as a proxy for a sector’s aver- captures labor market institutions. The vari- age FDI motive in a country, whereby a higher able ranges from 0 to 100 (highest labor free- share is associated with market-oriented FDI. dom) and includes various aspects of the legal c. inp = a sector’s average percentage of domestic and regulatory framework of a country’s labor input purchases of FDI firms in a country. This market, such as minimum wages; laws inhibit- measure captures a sector’s average sourcing ing layoffs; severance requirements; and mea- strategy of foreign firms in a country, whereby surable regulatory burdens on hiring, hours, a higher share is associated with more local and so forth. The measure is mainly based on sourcing. data from the World Bank’s Doing Business d. tech = iso + tech_for + website + email with annual studies. 0 ≤ tech ≤ 4, where iso = 1 if the firm owns b) finance = measure of financial freedom, in nat- internationally recognized quality certification ural logarithms, from the Heritage Foundation. and 0 otherwise; tech_for = 1 if the firm uses The variable measures banking efficiency, as technology licensed from foreign firms and 0 well as independence from government control otherwise; website = 1 if the firm uses its own and interference in the financial sector, with 138 Making Global Value Chains Work for Development scores ranging from 0 to 100 (highest finan- h. trade2 = measure of trade freedom, in loga- cial freedom). This measure relies on various rithms, from the Heritage Foundation; a com- underlying data sources, including (in order posite measure of the trade-weighted average of priority) the Economist Intelligence Unit, applied tariff rate and nontariff barriers, with International Monetary Fund, Organisation scores ranging from 0 to 100 (highest trade for Economic Co-operation and Development, freedom), reflecting the absence of trade pro- and official government publications of each tectionism. The measure is based on various country. underlying sources, including data from the c. educ1 = government spending on educa- World Bank, World Trade Organization, and tion, as a percentage of gross domestic prod- Economist Intelligence Unit. uct (GDP), from the World Development i. business = measure of business freedom, Indicators (WDI) database. in natural logarithms, from the Heritage d. educ2 = people who have completed second- Foundation; it is an outcome-based indicator ary and tertiary education, as a percentage of of a country’s institutional development. It is a population ages 15 years and older, from Barro measure that reflects the ability to start, oper- and Lee (2010). ate, and close a business, with scores ranging e. rd = country’s expenditures on research and from 0 to 100 (highest business freedom). The development, as a percentage of GDP, from the measure mainly relies on the World Bank’s WDI database. Doing Business annual studies. f. investment = measure of investment free- j. hhi = measure of sector concentration, to cap- dom, in natural logarithms, from the Heritage ture competition in a domestic firm’s sector. Foundation; it serves as a proxy for investment The hhi of sector concentration is defined as promotion. The score ranges from 0 to 100 the sum of squares of a firm’s output share by (highest investment freedom) and measures sector. If only one firm operates in a sector, the the ability of individuals and firms to move hhi would be 1. A lower hhi reflects higher sec- their resources in and out of specific activities tor diversity. This measure includes domestic internally and across the country’s borders. and foreign firms. This variable is mainly based on official gov- k. income = a country’s per capita GDP (US$ at ernment publications of each country on capi- 2000 prices), in natural logarithms, from the tal flows and foreign investment. WDI database. It captures national competi- g. trade1 = country’s share of exports of goods tion, but also other aspects of the national and and services as a percentage of GDP, from the institutional environments. WDI database. Use of GVC Measures to Assess the Drivers and Impacts of GVC Participation 139 Annex 7C. Factors Mediating Productivity Spillovers from GVC Integration in Bulgaria Table 7C.1. Structural Integration in GVCs from a Buyer’s Perspective and Its Impact on Productivity, the Role of Absorptive Capacity, Manufacturing Firms, OLS (1) (2) (3) (4) (5) (6) (7) Variable gapirst techirst skillsirst sizeirst agglrct expirst fdiirst BONwincst 2.3316 –2.6228** 0.7005 –4.2015*** 0.8967 –1.3804 –0.5884 (0.434) (0.026) (0.600) (0.001) (0.697) (0.252) (0.637) BONwincst*MF 6.3302*** 1.7829*** 0.2092 1.0560*** –0.0492 3.3622*** 3.8519*** (0.000) (0.000) (0.637) (0.000) (0.993) (0.000) (0.000) BONwincst*MF*bulgariac 1.7456** 0.4047 –0.4018 0.2034 17.8798 –1.8881** –3.6582** (0.014) (0.373) (0.892) (0.395) (0.112) (0.031) (0.021) lncapintirst 0.0484*** 0.2433*** 0.2631*** 0.2570*** 0.2647*** 0.2631*** 0.2501*** (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) bulgariac –1.1657*** –1.5192*** –1.1953*** –1.4954*** –1.5890*** –0.9818*** –1.1845*** (0.000) (0.000) (0.001) (0.001) (0.000) (0.001) (0.001) constant 7.8520*** 6.2189*** 4.3364*** 5.8411*** 5.8362*** 4.6493*** 4.6640*** (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) Observations 8,178 8,734 8,619 8,734 8,734 8,672 8,415 R-squared 0.91 0.50 0.47 0.49 0.47 0.48 0.49 Sources: Based on Farole and Winkler 2014, Santoni and Taglioni 2015. Note: The dependent variable is log labor productivity (lnlpirst). All regressions include sector, subnational region, and year fixed effects. Standard errors are clustered at the country-sector level. GVC = global value chain; OLS = ordinary least squares. *p < 0.1, **p < 0.05, ***p < 0.01. Summary • Other factors that positively mediate the impact of structural integration in GVCs from a buyer’s • Structural integration in global value chains perspective are a firm’s technology level (tech), (GVCs) has a positive effect from a buyer’s per- size (size), export share (exp), and foreign direct spective if all the mediating factors are taken into investment (FDI) (fdi) status. However, that only account (see table 7C.3, first column). holds for the full country sample. In Bulgaria, the • A lower technology gap positively mediates pro- positive effects from export share and FDI status ductivity gains from GVC participation on the are smaller (table 7C.1, columns 6 and 7). buying side in the full country sample, and the • Although agglomeration (aggl) has a negative positive effect is even larger for Bulgaria (table influence in the overall sample, the effect is posi- 7C.1, column 1). tive for Bulgaria (table 7C.1, column 5; and table 7C.3, column 1). Table 7C.2. Structural Integration in GVCs from a Seller’s Perspective and Its Impact on Productivity, the Role of Absorptive Capacity, Manufacturing Firms, OLS (1) (2) (3) (4) (5) (6) (7) Variable gapirst techirst skillsirst sizeirst agglrct expirst fdiirst BONwoutcst 21.5631*** 2.8876 7.0853* 0.3527 12.1879* 3.0935 6.0138* (0.003) (0.392) (0.058) (0.922) (0.076) (0.336) (0.076) BONwoutcst*MF 6.4980*** 1.8536*** 0.2234 1.1055*** –16.0688 3.6661*** 4.2005*** (0.000) (0.000) (0.618) (0.000) (0.250) (0.000) (0.000) BONwoutcst*MF*bulgariac 2.6355*** 0.6088 –1.5955 0.2116 –352.6137*** –2.2371** –4.0507** (0.000) (0.246) (0.603) (0.472) (0.001) (0.017) (0.020) lncapintirst 0.0401*** 0.2420*** 0.2628*** 0.2563*** 0.2642*** 0.2626*** 0.2491*** (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) bulgariac –0.4180 –1.3305*** –0.8988** –1.3411*** 1.0230 –0.8387*** –0.9766** (0.200) (0.000) (0.016) (0.004) (0.164) (0.008) (0.013) constant 5.1944*** 5.4665*** 3.4426*** 5.2289*** 4.4005*** 3.9991*** 3.7278*** (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) Observations 8,178 8,734 8,619 8,734 8,734 8,672 8,415 R–squared 0.92 0.50 0.48 0.49 0.48 0.48 0.49 Sources: Based on Farole and Winkler 2014; Santoni and Taglioni 2015. Note: The dependent variable is log labor productivity (lnlpirst). All regressions include sector, subnational region, and year fixed effects. Standard errors are clustered at the country-sector level. GVC = global value chain; OLS = ordinary least squares. *p < 0.1, **p < 0.05, ***p < 0.01. 140 Making Global Value Chains Work for Development Summary Table 7C.3. Structural Integration in GVCs and Its Impact on Productivity, the Role of Absorptive • Structural integration in GVCs has a positive Capacity, Manufacturing Firms, OLS effect from a seller’s perspective if all mediating (1) (2) factors are taken into account (see table 7C.3, col- Variable BONwincst BONwoutcst umn 2). The positive impact is stronger for the BONcst 9.2658* 32.3833** seller-side measure compared with the buyer-side (0.068) (0.012) measure. BONcst*gapirst 6.2972*** 6.4560*** (0.000) (0.000) • A lower technology gap (gap) positively medi- BONcst*gapirst*bulgariac 2.0540*** 2.8115*** ates productivity gains from GVC participation (0.000) (0.000) on the selling side in the full country sample, and BONcst*techirst 0.2685*** 0.2791*** (0.004) (0.004) the positive effect is even larger for Bulgaria (table BONcst*techirst*bulgariac –0.2954 –0.0963 7C.2, column 1). (0.161) (0.698) • As was the case for the buyer-related GVC mea- BONcst*skillsirst 0.0810 0.0767 sure, other factors that positively mediate the (0.514) (0.568) BONcst*skillsirst*bulgariac –1.2359 –0.9956 impact of structural integration in GVCs from a (0.314) (0.327) seller’s perspective are a firm’s technology level BONcst*sizeirst 0.1264*** 0.1131** (tech), size (size), export share (exp), and FDI (0.008) (0.018) BONcst*sizeirst*bulgariac 0.1545 –0.0556 (fdi) status. However, that only holds for the full (0.474) (0.712) country sample. In Bulgaria, the positive effects BONcst*agglrct –30.9621*** –37.9136 from export share and FDI status are smaller (0.009) (0.115) BONcst*agglrct*bulgariac 75.6443*** –943.2283*** (table 7C.2, columns 6 and 7). (0.001) (0.002) • Interestingly, the mediating impact of agglomera- BONcst*expirst –0.0089 –0.0331 tion (aggl) turns negative from a seller’s perspec- (0.976) (0.920) tive, whereas the effect is positive from a buyer’s BONcst*expirst*bulgariac –0.7764 –0.8087 (0.571) (0.556) perspective (table 7C.2, column 5; and table 7C.3, BONcst*fdiirst 0.3081 0.3633* column 2). The interpretation could be that (0.125) (0.053) agglomerations entail positive urbanization econ- BONcst*fdiirst*bulgariac –0.5836 –0.5684 (0.119) (0.201) omies when firms rely on external inputs in GVCs, lncapintirst 0.0435*** 0.0375*** which lowers production costs and increases firm (0.000) (0.000) productivity, and those benefits outweigh poten- bulgariac –1.2247*** 5.6863*** (0.002) (0.000) tial negative congestion costs. Firms that are sell- constant 6.9362*** 3.6770** ing within GVCs, by contrast, may face higher (0.000) (0.030) negative congestion costs (for example, related to Observations 7,751 7,751 transportation), which seem to be higher than the R-squared 0.92 0.92 potential benefits in agglomerations. Sources: Based on Farole and Winkler 2014; Santoni and Taglioni 2015. Note: The dependent variable is log labor productivity (lnlpirst). All regressions include sector, subnational region, and year fixed effects. Standard errors are clustered at the country-sector level. GVC = global value chain; OLS = ordinary least squares. *p < 0.1, **p < 0.05, ***p < 0.01. Table 7C.4 Structural Integration in GVCs from a Buyer’s Perspective and Its Impact on Productivity, the Role of National Characteristics, Manufacturing Firms, OLS (1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (11) Variable laborct financect educ1ct educ2ct rdct investmct trade1ct trade2ct businessct hhisct incomect BONwincst 18.4738 15.1480 15.4471* 5.8874 –9.0676** –18.5153 5.4447* 46.2275*** 50.0742 0.9948 14.6179 (0.519) (0.291) (0.063) (0.114) (0.045) (0.153) (0.070) (0.007) (0.114) (0.456) (0.130) BONwincst*MF –4.2677 –3.7896 –83.5696* –18.3914 1545.9797** 5.0911 –11.7106* –10.9109*** –11.9946 –0.9732 –1.7905 (0.537) (0.309) (0.058) (0.100) (0.029) (0.148) (0.052) (0.007) (0.118) (0.470) (0.134) BONwincst*MF*bulgariac 1.8937* 2.2559** 72.6620* 30.3471*** 1948.5219** 1.4553 18.7248** 3.2866*** 2.3850*** 0.0000 1.0696** (0.074) (0.031) (0.071) (0.004) (0.039) (0.183) (0.025) (0.000) (0.010) (.) (0.033) lncapintirst 0.2711*** 0.2708*** 0.2816*** 0.2707*** 0.2594*** 0.2712*** 0.2712*** 0.2699*** 0.2709*** 0.2722*** 0.2708*** (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) bulgariac –2.1101*** –2.0583*** –2.8288*** –2.6351*** –2.0001*** –2.1376*** –3.3296*** –2.8066*** –2.2941*** –2.3292*** (0.000) (0.000) (0.000) (0.000) (0.001) (0.000) (0.000) (0.000) (0.000) (0.000) constant 5.9350*** 5.8794*** 7.5136*** 5.5893*** 6.8063*** 6.1145*** 6.4155*** 5.9171*** 6.1632*** 4.2449*** 5.8306*** (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) Observations 8,515 8,515 6,579 8,515 6,673 8,515 8,515 8,515 8,515 8,263 8,515 R-squared 0.48 0.48 0.47 0.48 0.44 0.48 0.48 0.48 0.48 0.48 0.48 Sources: Based on Farole and Winkler 2014; Santoni and Taglioni 2015. Note: The dependent variable is log labor productivity (lnlpirst). All regressions include sector, subnational region, and year fixed effects. Standard errors are clustered at the country-sector level. GVC = global value chain; OLS = ordinary least squares. *p < 0.1, **p < 0.05, ***p < 0.01. Summary • National and institutional characteristics in Bulgaria positively mediate the • By contrast, some of those national characteristics have a negative effect in effect of structural integration in GVCs on firm productivity across the board the full country sample—for example, government spending on education when examining the buying side (table 7C.4). (educ1), share of exports (trade1), and the absence of trade protectionism • Higher financial freedom (finance), government spending on education (trade2). Therefore, for those variables, the net effect is less negative or even (educ1), share of people with completed secondary and tertiary education positive for Bulgaria. (educ2), share of research and development (R&D) expenditures in gross domestic product (GDP) (rd), share of exports (trade1), absence of trade protectionism (trade2), and GDP (income) all show positive and significant mediating effects in Bulgaria. Use of GVC Measures to Assess the Drivers and Impacts of GVC Participation 141 142 Table 7C.5. Structural Integration in GVCs from a Seller’s Perspective and Its Impact on Productivity, the Role of National Characteristics, Manufacturing Firms, OLS (1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (11) Variable laborct financect educ1ct educ2ct rdct investmct trade1ct trade2ct businessct hhisct incomect BONwoutcst –18.9168 80.4004 45.1467** 31.4952*** 17.1380 –45.0303 11.8399* 188.8899*** 271.4564** 6.8400* 99.3937*** (0.803) (0.177) (0.030) (0.006) (0.201) (0.500) (0.092) (0.000) (0.022) (0.056) (0.008) BONwoutcst*MF 6.3794 –19.2451 –256.3010** –89.1525** 3533.6598* 13.3403 –14.8726 –43.2515*** –64.0242** –0.7015 –11.2453** (0.736) (0.203) (0.032) (0.016) (0.077) (0.443) (0.339) (0.000) (0.024) (0.616) (0.011) BONwoutcst*MF*bulgariac –29.8746*** –34.3717*** –1.87e+03*** –426.7757*** –2.31e+04*** –32.1103*** –276.8090*** –38.9288*** –31.0791*** 0.0000 –19.4073*** (0.001) (0.000) (0.000) (0.000) (0.009) (0.001) (0.001) (0.000) (0.000) (.) (0.000) lncapintirst 0.2705*** 0.2704*** 0.2814*** 0.2704*** 0.2592*** 0.2705*** 0.2707*** 0.2693*** 0.2703*** 0.2719*** 0.2703*** (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) bulgariac 12.2958*** 14.2412*** 20.6887*** 14.5512*** 11.0190** 11.7757*** 11.0847*** 16.9562*** 12.9836*** 13.6429*** (0.002) (0.000) (0.000) (0.000) (0.012) (0.004) (0.008) (0.000) (0.001) (0.000) constant 8,515 8,515 6,579 8,515 6,673 5.8825*** 5.8061*** 5.1181*** 6.4553*** 3.4222*** 4.3802*** 0.48 0.48 0.47 0.48 0.44 (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) Making Global Value Chains Work for Development Observations 8,515 8,515 6,579 8,515 6,673 8,515 8,515 8,515 8,515 8,263 8,515 R-squared 0.48 0.48 0.47 0.48 0.44 0.48 0.48 0.48 0.48 0.48 0.48 Sources: Based on Farole and Winkler 2014; Santoni and Taglioni 2015. Note: The dependent variable is log labor productivity (lnlpirst). All regressions include sector, subnational region, and year fixed effects. Standard errors are clustered at the country-sector level. GVC = global value chain; OLS = ordinary least squares. *p < 0.1, **p < 0.05, ***p < 0.01. Summary • National and institutional characteristics in Bulgaria negatively mediate the • The full country sample also shows the negative effect of government spend- effect of structural integration in GVCs on firm productivity across the board ing on education (educ1), the share of people with secondary and tertiary when examining the selling side (table 7C.5). education (educ2), the share of exports in GDP (trade1), the absence of trade • Less restricted labor (labor) or financial markets (finance), more government protectionism (trade2), and per capita GDP (income). Therefore, the negative spending on education (educ1), a higher share of people with completed sec- influence is even more pronounced for Bulgaria in those areas. ondary and tertiary education (educ2), a higher share of R&D expenditures • The only variable in the overall sample with a positive impact is a country’s in GDP (rd), more freedom to invest (invest), a higher share of exports in R&D intensity. GDP (trade1), more absence of trade protectionism (trade2), and higher GDP (income) all show a negative and significant mediating impact on productiv- ity in Bulgaria. Use of GVC Measures to Assess the Drivers and Impacts of GVC Participation 143 Notes Chains, edited by Thomas Farole and Deborah Winkler, 59–86. Washington, DC: World Bank. 1. This section draws on Kummritz, Taglioni, and Hollweg, C. 2015. “The Labor Content of Exports in Belize: Winkler (forthcoming), which is part of ongoing work A Preliminary Exploration.” World Bank, Washington, at the World Bank that aims to develop a taxonomy of DC. GVC participation and economic upgrading for a set of Jiang, Xiao, and William Milberg. 2013. “Capturing the countries. Jobs from Globalization: Trade and Employment in 2. Calì and others (2016). Global Value Chains.” Working Paper 30, Capturing 3. The World Bank’s Trade in Value Added database the Gains: Economic and Social Upgrading in Global can be accessed at http://data.worldbank.org/data-catalog Production Networks, University of Manchester, /export-value-added. Manchester, U.K. 4. Jiang and Milberg (2013). Kummritz, Victor, Daria Taglioni, and Deborah Winkler. 5. Jiang and Milberg (2013). Forthcoming. “Economic Upgrading through Global 6. Timmer and others (2014). Value Chain Participation: Which Policies Increase the 7. Timmer and others (2014). Value Added Gains?” World Bank, Washington, DC. OECD (Organisation for Economic Co-operation and Development). 2013. Science, Technology and Industry References Scoreboard 2013, Innovation for Growth. Paris: OECD. Barro, Robert J., and Jong-Wha Lee. 2010. “A New Data Set Roberts, M. J., and J. R. Tybout. 1997. “The Decision to of Educational Attainment in the World, 1950–2010.” Export in Colombia: An Empirical Model of Entry NBER Working Paper No. 15902, National Bureau of with Sunk Costs.” American Economic Review 87 (4): Economic Research, Cambridge, MA. 545–64. Calì, M., and C. Hollweg. 2015. “The Labor Content of Santoni, Gianluca, and Daria Taglioni. 2015. “Networks Exports in South Africa: A Preliminary Exploration.” and Structural Integration in GVCs.” In The Age of World Bank, Washington, DC. Global Value Chains, edited by João Amador and Filippo Calì, M., J. Francois, C. Hollweg, M. Manchin, D.A. di Mauro. Washington, DC: Center for Economic and Oberdabernig, H. Rojas-Romagosa, S. Rubinova, and Policy Research. P. Tomberger. 2016. “The Labor Content of Exports Timmer, Marcel P., Abdul Azeez Erumban, Bart Los, Robert Database.” Policy Research Working Paper No. 7615, Stehrer, and Gaaitzen J. de Vries. 2013. “Slicing Up World Bank, Washington, DC. Global Value Chains.” Journal of Economic Perspectives Farole, Thomas, and Deborah Winkler. 2014. “The Role 28(2). of Mediating Factors for FDI Spillovers in Developing UNCTAD (United Nations Conference on Trade and Countries: Evidence from a Global Dataset.” In Making Development). 2013. “World Investment Report Foreign Direct Investment Work for Sub-Saharan Africa: 2013—Global Value Chains: Investment and Trade for Local Spillovers and Competitiveness in Global Value Development.” UNCTAD, Geneva. PART III STRATEGIC QUESTIONS AND POLICY OPTIONS By integrating their domestic firms (suppliers and final producers) into global value chains (GVCs), low- and middle-income countries can help their economies industrialize, become services oriented faster, and move closer to their development goals. Part II suggests how to measure various aspects of GVC participation and, thus, how to identify key policy needs. This part of the book builds on those findings, suggesting “strategic questions” and approaches to addressing them—“policy options.” Including real-world examples, the text proposes a diagnostics exercise to identify three focus areas. Chapter 8—“Entering GVCs”—discusses ways for countries to enter global production networks. Those avenues include ways to attract foreign investors, as well as strategies to enhance the participation of domestic firms in GVCs. Suggestions for entering GVCs encompass measures to ensure that the country can offer world-class links to the global economy and create a friendly business climate for foreign tangible and intangible assets. Chapter 9—“Expanding and Strengthening GVC Participation”—discusses ways for countries to lever their position in GVCs to achieve higher value addition through economic upgrading and densification. The concept of economic upgrading is largely about gaining competitiveness in higher-value-added products, tasks, and sectors. Densification involves engaging more local actors (firms and workers) in the GVC network. Strengthening GVC–local economy links, absorptive capacity, and skills contributes to the overall goal to increase a country’s value added that results from GVC participation. Chapter 10 tackles the challenge of “Turning GVC Participation into Sustainable Development.” The chapter focuses on social and environmental sustainability of GVCs. Labor market–enhancing outcomes for workers at home and more equitable distribution of opportunities and outcomes create social support for a reform agenda aimed at strengthening a country’s GVC participation. Climate-smart policy prescriptions and infrastructure can mitigate the challenges for firms from climatic disruptions, ensuring the long-term predictability, reliability, and time-sensitive delivery of goods necessary to participate in GVCs. Focus area Objectives Strategic questions Policy options Which tasks? Creating world-class GVC links – Which form of GVC participation? – Jump-starting GVC entry through EPZs – How can tasks be identified? and other competitive spaces Attracting foreign – Which risks? – Attracting the "right" foreign investors investors and Which form of governance? – Helping domestic firms find the Entering GVCs “right” trade partner and technology facilitating domestic – Which form of governance between firms’ entry into GVCs lead firms and suppliers? abroad – Buyer- or producer-driven value – Improving connectivity to chains? international markets – Which power relations in GVCs? Creating a world-class climate for foreign tangible and intangible assets – Ensuring cost competitiveness – Improving drivers of investment and protecting foreign assets – Improving domestic value chains and quality of infrastructure and services Chapter 8 ENTERING GVCs Introduction tasks be identified? (3) Which risks? Before coun- try analysts consider these questions, they should This chapter focuses on the strategic questions, pos- be aware of the pitfalls of basing their strategies sible answers, and critical issues that policy makers on sector-based conceptual frameworks. Chapter must consider when seeking to enter global value 1 shows that reasoning along broad sector lines chains (GVCs). A country that seeks to participate in assumes that countries sell final goods to each other GVCs must ask which tasks it should focus on and and that, as countries grow richer, they transition which types of GVC governance are possible. The from specializing in the primary sector to manufac- chapter suggests that governments that seek to join turing and ultimately to services. GVCs have to create (1) world-class GVC links and In contrast to this sector-based vision, a “new (2) a world-class climate for foreign tangible and paradigm” centered on tasks has recently gained intangible assets. The first item requires attracting popularity. Its premise is that in the world of the right foreign investors and improving connec- GVCs—dominated by complex and fragmented tivity to international markets; the second requires production processes—development is best achieved high-quality infrastructure and services. Countries by specializing in the tasks and activities of com- also need to be aware of the different power relations parative advantage among the broad range available. in GVCs between the lead firm and other firms, and After all, a firm’s location decisions are task specific. the scope for diversifying specific supply chain risk. Yet that approach, too, is partial, as it captures only functional upgrading efforts and strategies. Product Attracting Foreign Investors and and inter-sector upgrading—defined in chapter 9— Facilitating Domestic Firms’ Entry into are also necessary and can be achieved through the GVCs: Strategic Questions upgrading of skills, capital, and processes (see figure 1.9 in chapter 1). That higher-income countries have Entering a GVC requires answering two strategic a stronger specialization in high-value-added manu- questions: (1) What tasks are performed in a GVC? facturing and services than lower-income countries (2) What form of governance does the GVC follow? indeed reflects the former’s greater use of skills and The first is a more country-level question; the sec- know-how, capital and technology, and improved ond emphasizes that entry into GVCs is ultimately a processes in its production, whether in agriculture, firm’s decision. industry, or services—hence the term task-based development strategies. Therefore, this part discusses Which Tasks? all three major forms of upgrading in GVCs—prod- The first strategic question has three sub-questions: uct, functional, and inter-sector—and three ways to (1) Which form of GVC participation? (2) How can achieve them—skills, capital, and process upgrading. 147 148 Making Global Value Chains Work for Development Which Form of GVC Participation? foreign investors is the potential they create to help Before identifying and focusing on the tasks and internationalize domestic firms, particularly their risks in GVCs, countries need to be aware of the two suppliers. They do this in two main ways: indirectly, sets of approaches for entering GVCs: (1) attracting by requiring domestic firms to meet international foreign investors and (2) facilitating domestic firms’ standards (as in quality and timely delivery) and by access to GVCs (“internationalizing” those firms). contributing to building the scale and productivity Regarding the first approach, why do countries go of their domestic suppliers; and directly, by provid- to great lengths to attract foreign direct investment ing access to their international marketing, supply, (FDI)? One simple answer is that many countries and distribution networks.5 have built up too little domestic capital to stimulate Still, linking to foreign-owned subsidiaries of for- growth. FDI thus represents an important source of eign firms is not the only way for domestic firms to private capital. And given the relatively long-term join GVCs. They can consider other approaches that outlook of direct (versus portfolio) investors, FDI involve arm’s-length trade: generally is less risky than other financial flows, because it tends to be less vulnerable to rapid out- • Exporting inputs to international buyers flows caused by exogenous shocks. Moreover, perva- • Becoming domestic final producers that import sive information asymmetries—with powerful lead intermediates firms able to maintain and increase markups and with competitive suppliers subject to pressure from Another approach to consider is the hybrid case of buyers on supply price, delivery time, quality, and contract manufacturers that produce fully assembled payment schedule at the bottom—may lead to a sub- goods for large retailers (such as Walmart or Gap) or optimal level of cross-border investment, justifying lead firms that focus on design, development, and public intervention.1 marketing and that outsource the actual produc- But the more important answer is that FDI has tion of their products, such as Nike, Calvin Klein, the potential to deliver far greater “dynamic” benefits or Fisher-Price.6 Contract manufacturers therefore to host economies through the spillovers they deliver fall into the latter two categories. They are part of (mainly through technological and other advantages non-equity modes of investment (NEMs), which that stimulate higher productivity). Spillovers, in this also include business arrangements such as contract context, generally refer to the diffusion of knowl- farming, business process outsourcing, franchising, edge—unintentional or intentional, if sharing that contract management, strategic alliances, and joint knowledge is not compensated in some way—from ventures. In those cases, a multinational has a con- multinational affiliates to local firms. Thus, spillovers tractual relationship with a domestic firm in the host encompass technology and all forms of codified and country and maintains some degree of control over tacit knowledge related to production, including the operation and conduct of business (more so than management and organizational practices. It also in the case of arm’s-length trade), but has no owner- includes the benefits that can accrue to local par- ship stake.7 GVC participation through arm’s-length ticipants when they link into the global networks of trade and NEMs can also lead to spillovers. multinational investors.2 This chapter clarifies that the form of GVC par- Not all FDI is the same, however; its development ticipation matters for development. The chapter also impact varies depending on the extent of foreign discusses how the form of governance in GVCs is ownership. Fully foreign-owned FDI, for example, not a prerogative of public policy, but endogenous may induce the lead firm to transfer more knowl- to lead firms, although countries may adopt comple- edge—through technology, say—to the host country.3 mentary policies to meet lead firms’ needs to lever Partly foreign-owned FDI could also be beneficial for GVC opportunities. local firms; the lead firm’s interests are less well pro- tected, which makes technology leakages more likely. How Can Tasks Be Identified? Larger domestic participation might also increase the It is often difficult for policy makers and analysts to chances of relying on domestic suppliers.4 identify the tasks in which a country has a compara- Regarding the second approach—international- tive advantage, partly because full production- and izing domestic firms—one important spillover from trade-related statistics are rarely available at the task Entering GVCs 149 level in low- and middle-income countries (LMICs). not yet active. In that case, countries can focus on the Combining different approaches—complementary third step—identifying the optimal export sectors but different in data requirements—allows inves- and value chains—and devote less attention to the tigators to identify broad sectors, value chains, and starting sector or product specialization. Concepts specific activities, thereby enabling the country to of economic proximity between products may help determine its GVC entry strategy. identify the difficulties inherent in “jumping” to new One strategy encourages entry into tasks—in sec- sectors and activities.) tors or value chains—in which the country already has expertise. The strategy internationalizes the Step 1. Identify Sectors with the Highest RCA, existing production of goods or services, or that of Based on Value-Added Export Data new tasks—in a more aggregated sector or in a value Identification of the export sectors in which a coun- chain in which the country already specializes. For try has an RCA should be based on value-added example, Kenya—already an important producer of rather than gross export data. Malaysia,8 for example, fruits and vegetables—later joined the horticulture has an RCA greater than one in four of nine manu- GVC within the same industry. facturing sectors—electrical and optical equipment Tasks can be identified in three steps. Step 1 iden- (the most important GVC sector); machinery and tifies the broad export sectors in which a country has equipment (not elsewhere classified); chemicals and a revealed comparative advantage (RCA), which can non-metallic mineral products; and wood, paper, be based on value-added export data. Step 2 analyzes paper products, printing, and publishing—on both the upstream and downstream output of a GVC measures (figure 8.1). But for electrical and optical product. Step 3 identifies differences in economic equipment, the value added–based RCA is about 15 characteristics of tasks within those export sectors percent lower—a key distinction. and value chains, such as tasks that may create the largest domestic value added or have important Step 2. Analyze Upstream and Downstream Output of potential for diversification. a GVC Product 9 (Another strategy identifies a country’s potential Network analysis applied to input-output (I-O) for entry into tasks in sectors in which the country is tables can help in assessing the features of the value Figure 8.1. Malaysia: RCA, Gross Exports, and Domestic Value Added Embodied in the Country’s Gross Exports, 2009 Food products, beverages, and tobacco Textiles, textile products, leather, and footwear Wood, paper, paper products, printing, and publishing Chemicals and nonmetallic mineral products Basic metals and fabricated metal products Machinery and equipment, nec Electrical and optical equipment Transport equipment Manufacturing nec; recycling 0 0.25 0.50 0.75 1.00 1.25 1.50 1.75 2.00 RCA RCA_EXGR_DVA RCA_EXGR RCA Line Source: Adapted from the Organisation for Economic Co-operation and Development–World Trade Organization Trade in Value Added database. Note: RCA = revealed comparative advantage; RCA_EXGR = revealed comparative advantage based on gross exports; RCA_EXGR_DVA = revealed comparative advantage based on domestic value added in gross exports; nec = not elsewhere classified. 150 Making Global Value Chains Work for Development chains in which a country specializes. Using the U.S. which are all likely to require higher technological I-O tables has the advantage of documenting I-O and skill content. Matching these findings to trade relationships at the finest level of disaggregation. But data, the analysis shows that although Malaysia’s using those tables for assessing tasks in third coun- position as an exporter of downstream products is tries has one important caveat: the analysis may be relatively marginal, its most important competitor in biased because of differences in technology across producing computer storage devices is China, which countries. Still, the richness from the very detailed is also the largest buyer of Malaysian exports of the documentation of the production structure in the product, as well as a leading exporter of downstream U.S. I-O tables and the absence of comparable data products—factors that may help shape GVC entry for almost all countries worldwide justify their use. strategies. The method of analysis has the following steps: Step 3. Identify Which Tasks within a Broad Sector or 1. Identify the positioning of the export product of Value Chain Create the Largest Domestic Value Added interest in the wider network of inter-sector pro- or Promise for Growth and Development duction links. In the absence of market failures (monopolistic 2. Identify sectors that are the main buyers of the rent or exclusive or controlled access to resources), product and sectors that are the main suppliers tasks tend to depend on the know-how (quantity and their relative economic contribution (mea- and quality of workers) and capital stock (including sured in value added or exports). technology) available to perform them. If only a frac- 3. Assess the relative position of countries of inter- tion of the workforce is highly skilled, launching into est as suppliers of the product, as well as in the tasks that depend primarily on skilled workers does production of upstream and downstream prod- not make sense for a country. The goal is to choose ucts, and the relative value added or export tasks that create the largest domestic value added, contribution. given the labor and capital endowments in the home 4. Repeat steps 1 to 3 for upstream and downstream country. sectors to map out a wider portion of the value That is indeed what Morocco did to develop its chain for the product of interest. aerospace industry (box 8.2), based on its predomi- nantly low-skilled workforce. Good performance Box 8.1 applies this concept to computer storage allowed the country to transition to higher value- devices, Malaysia’s main export product. The analysis added segments. reveals that the product is small and peripheral to the Information about the value added of tasks can- manufacturing production network (based on U.S. not be easily obtained using available statistical data. I-O tables) and that the product’s main buyers are Using I-O tables and gathering firm-level data are relatively concentrated in more sophisticated sectors, two ways to address some of the data constraints Box 8.1. Network Analysis of a Product Value Chain Using I-O Tables Figure B8.1.1 shows the inter-sector links for the sectors in manu- requirements in the production of j are shown. North American facturing, using the highest available disaggregation provided by Industry Classification System (NAICS) product 331110, iron and the U.S. input-output (I-O) tables (388 products) for 2007. a The steel mills and ferroalloys, is the most structurally integrated node size is proportional to the so-called OUT-degree: bigger/ into the manufacturing production network, supplying inputs to darker nodes are those that supply intermediates to a larger num- a large number of manufacturing industries. The network visual- ber of industries (the color is correlated to the size of the node— ization also puts into perspective the position of product 334112, that is, they deliver the same information). Links, from sector i computer storage devices, which is an important export product to sector j, are proportional to the share of i in the overall input of middle-income countries—such as Malaysia—in the electron- demand of j, excluding j’s inputs sourced from j. The network is ics global value chain (GVC). The figure shows that the product is built considering all intermediate flows from i to j, using all the relatively small and peripheral to the manufacturing production information available in the network structure. For visual clarity, network (in red). only flows above the 1 percent threshold of total intermediates (Box continues next page) Box 8.1. (continued) Figure B8.1.1. Manufacturing Inter-Sector Links, NAICS 31–33, 2007 334111 33411A 334112 334515 334516 334220 325310 334418 324122 334210 333314 33641A 327992 33451A 336414 335999 335991 311990 334413 327999 334511 327400 33441A 327310 3118A0 339113 334513 31151A 311700 327320 311810 311300 334514 324190 3259A0 334510 323120 327100 334290 334300 334512 311513 311514 311230 333315 333313 325620 339114 339116 334517 311910 325120 33331A 311210 335110 31122A 323110 327200 327330 339115 339910 335313 311520 312120 339112 311420 312130 311225 322210 331411 335930 331419 336411 322220 339940 335920 311940 322230 335311 335314 33299A 326160 332800 332913 311221 327993 334610 311410 311119 326110 311930 325180 327390 322130 312140 331420 336413 336412 326120 335911 311615 335120 331490 325413 322120 325110 335912 322299 335210 332320 336320 325411 312110 325190 333415 333295 325320 311920 326190 335222 332710 333994 326130 333993 325414 312200 325211 332410 33299B 311111 325910 335224 325412 325520 325610 326140 339950 335221 336992 327991 33211A 33351B 335228 322110 332430 33211B 332720 31161A 339990 335312 33341A 339930 325130 333414 33391A 333991 333611 322291 327910 33712A 33291A 325510 332996 333514 324121 314110 326150 33131B 331110 333130 3252A0 332600 336611 331200 33131A 332420 332310 333220 337900 313100 337127 336991 339920 332114 332991 313200 33351A 313300 332200 33329A 316000 326220 326290 333912 332500 331520 314900 333511 315000 337121 33721A 33399A 336500 33399B 333612 314120 337215 336370 337122 333920 333613 321100 331510 326210 333111 336612 333120 3219A0 337110 336360 336999 321200 336390 333112 333618 321910 336214 336212 336350 336112 3363A0 336111 336310 Node color scale From low to high import share 336120 336211 Entering GVCs Source: Santoni and Taglioni forthcoming. Note: NAICS = North American Industry Classification System. (Box continues next page) 151 152 Making Global Value Chains Work for Development BOX 8.1. (continued) Sector Buyers of This Product Sector Suppliers of This Product Figure B8.1.2 reports the outflows of sector 334112 (red node). Figure B8.1.3 displays the inflows to 334112 (red node)—that The green nodes represent industries that use 334112 as inputs is, the most important suppliers of intermediates for this sector in production and for which computer storage devices constitute (green nodes): sectors 334610 (software reproduction), 33411A at least 1 percent of the total input requirements for their pro- (other computer manufacturing), 334418 (printed circuit assembly), duction (for visual clarity, we highlight only the links associated 335999 (all other miscellaneous electrical equipment and compo- with sector 334112). These are sector 334510 (electromedical nent manufacturing), 33441A (other electronic component manu- and electrotherapeutic apparatus), sector 334111 (electronic facturing), 332800 (metal treating), 3259A0 (other chemical product computer manufacturing with the U.S. Small Business Associa- and preparation manufacturing), 326110 (plastics packaging mate- tion small business standard, which includes manufacturing and rials and unlaminated film and sheet manufacturing), 334413 (semi- assembling electronic computers, such as mainframes, personal conductor and related device manufacturing), 332710 (machine computers, workstations, laptops, and computer servers), sector shops), and 336390 (other motor vehicle parts manufacturing).c 33411A (other computer manufacturing), sector 334511 (search, detection, navigation, guidance, aeronautical, and nautical sys- Figure B8.1.3. Most Relevant Suppliers for Computer tem and instrument manufacturing), and sector 33451A (other Storage Devices measuring and controlling device manufacturing).b These sectors are more sophisticated than computer storage devices, which suggests that entering the downstream stages of production may 336390 imply for Malaysia a need to upgrade its technology and skills. A detailed analysis of the production structure and relative value added of the downstream products to the item of interest—such 332710 334610 as the one suggested in Step 3 of the method—would further 326110 334510 make it possible to assess how easy jumping to the next step in the downstream value chain would be. 33441A 334112 Figure B8.1.2. Most Relevant Buyers of Computer Storage 334418 Devices 332800 335999 336390 3259A0 334511 334111 334413 33451A 332710 334610 326110 334510 33411A 33441A Most relevant suppliers NAICS product 334112 Other 334112 334418 Source: Santoni and Taglioni forthcoming. 332800 Note: The green lines designate main input flows from supplying sectors (green circles) to the computer storage devices sector (red circle). NAICS = North American Industry 335999 Classification System. 3259A0 334511 334413 Relative Positions of Countries as Suppliers of This Product 334111 33451A Figure B8.1.4 depicts the relative position of Malaysia (red node) as a supplier of computer storage devices in 2012 (NAICS 2007 code 334112, computer storage devices). The links between the 33411A other nodes show the exports of downstream products—products that use computer storage devices as major inputs—using U.S. Most relevant buyers NAICS product 334112 Other I-O tables for 2007.d The node size is proportional to a country’s market share in world exports. For Malaysia, the market share for Source: Santoni and Taglioni forthcoming. computer storage devices exports in 2012 was 5.6 percent. The Note: The red lines designate flows of computer storage devices (red circle) to main most important competitor was China. buying sectors (green circles). NAICS = North American Industry Classification System. (Box continues next page) Entering GVCs 153 BOX 8.1. (continued) Figure B8.1.4. Computer Storage Devices Network for Figure B8.1.5. Malaysia as an Importer of Downstream Malaysia Products Taiwan, China Taiwan, China Philippines Philippines Malaysia_f Malaysia_f Belgium Belgium Thailand Malaysia Thailand Malaysia China China Indonesia Indonesia Sweden Sweden Netherlands Netherlands Singapore Singapore Hong Kong SAR, Hong Kong SAR, China United China United United States Czech United States Czech Kingdom Republic Kingdom Republic Germany Germany Mexico Mexico Japan Japan Switzerland Switzerland Italy France Italy France Korea, Rep. Slovak Korea, Rep. Slovak Canada Canada Republic Republic Malaysian exports of NAICS product 334112 Other Malaysian imports of downstream products Other Most relevant buyers Most relevant sellers of downstream products Source: Santoni and Taglioni forthcoming; BACI World Trade Database, CEPII. Source: Santoni and Taglioni forthcoming; BACI World Trade Database, CEPII. Note: The red lines designate export flows of Malaysian computer storage devices Note: The green lines designate import flows of downstream products that use computer (red circle) to most relevant buyers (green circles). The green lines designate export storage devices as inputs to Malaysia (red circle) from most relevant sellers (green flows of downstream products that use computer storage devices as inputs. NAICS = circles). NAICS = North American Industry Classification System. North American Industry Classification System. The nine largest buyers (green nodes) of Malaysian exports Figure B8.1.6. Malaysia as an Exporter of Downstream of computer storage devices absorbed 50 percent of the country’s Products exports in this sector. For the other countries, the size of the node reflects the market share in exports of downstream products: Taiwan, China China (green links) is the most important exporter of downstream Malaysia_f Philippines products, with an export market share of 37.3 percent. Exports Belgium Malaysia Thailand from China to the United States are 10 percent of world flows, China and flows from China to Hong Kong SAR, China, are 8.2 percent of Indonesia world flows. (In the other direction, exports from Hong Kong SAR, China, to China represent 6.3 percent of overall world flows.) Sweden Netherlands Figure B8.1.5 visualizes the position of Malaysia as a buyer of Hong Kong SAR, Singapore downstream products (with respect to computer storage devices) United China from other countries. The node MYS_f considers the position of Czech United States Republic Kingdom the country as an importer of downstream products (the largest Germany Mexico Japan exporters are the green nodes). Figure B8.1.6 reports the position of Malaysia as a seller of downstream products (with respect to Switzerland computer storage devices) to other countries. The node MYS_f Italy France Korea, Rep. Slovak Canada Republic considers the position of the country as an exporter of down- stream products (the largest importers are the green nodes). Malaysian exports of downstream products Other Most relevant buyers of downstream products Sources: Santoni and Taglioni forthcoming; BACI World Trade Database, CEPII. Note: The red lines designate Malaysian export flows of downstream products that use computer storage devices as inputs (red circle) to most relevant buyers (green circles). NAICS = North American Industry Classification System. (Box continues next page) 154 Making Global Value Chains Work for Development BOX 8.1. (continued) a. The network representation is built on the 2007 Commodity-by-Commodity c. Sector 33441A corresponds to the following NAICS (2007) codes: 334411 Direct Requirements table from the U.S. Bureau of Economic Analysis, which (electronic computer manufacturing); 334412 (bare printed circuit board provides only the total requirement table (TOT); the direct requirement was manufacturing), 334414 (electronic capacitor manufacturing), 334415 (electronic derived as DR = (TOT-I)xTOT -1, following Acemoglu and others (2012). The data resistor manufacturing), 334416 (electronic coil, transformer, and other inductor are in the form of a square matrix, where the (i, j ) entry represents the share of manufacturing), 334417 (electronic connector manufacturing), and 334419 commodity i (row) used in the production of commodity j (column). Column sums (other electronic component manufacturing). Sector 3529A0 corresponds to provide the total share of intermediate inputs in each commodity. 325920 (explosives manufacturing) and 32599 (all other chemical product b. Sector 33451A from the I-O tables groups the following NAICS (2007) codes: manufacturing). 334518 (watch, clock, and part manufacturing) and 334519 (other measuring d. Downstream industries are those that use 334112 as input in production and for and controlling device manufacturing). Sector 33411A corresponds to 334413 which computer storage devices represent at least 1 percent of the total input (semiconductor and related device manufacturing) and 334419 (other electronic requirements for their production. Those industries are depicted as green nodes component manufacturing). in figure B8.1.2. Box 8.2. The Moroccan Aerospace Industry Over the past decade, leading aviation companies, such as Boeing Figure B8.2.1. Upward Mobility: Approximate Employment of the United States and Bombardier of Canada, have invested in in the Moroccan Aerospace Industry increasingly sophisticated factories in Morocco. That investment is part of the government’s strategy to expand into more advanced manufacturing, including aerospace and electronics—a move that is expected to attract more basic industries in its wake. Approximate employment in In 2001, Boeing and French electrical wiring company Labinal Moroccan aerospace industry opened a small operation, Matis, to prepare cables for Boeing 737 jetliners. Workers assembled wire bundles and shipped them to 10,000 Boeing plants in the United States for installation. Initially, that 8,000 work did not require any technical background, but workers hit 70 percent efficiency of industry norms within two years. As the 6,000 company expanded, job openings attracted many highly educated 4,000 applicants, more than 80 percent of them with few job opportuni- ties in traditional industries. Today, Matis workers prepare wires 2,000 not just for Boeing, but also for General Electric, Dassault Avia- tion, and Airbus. 0 Matis’s parent company—now called Safran—then invested 2004 ’06 ’08 ’10 ’12 in more advanced manufacturing. In 2006, its Aircelle division Sources: Gimas; Moroccan Ministry opened a plant that produces jet engine housings. of industry Morocco’s aviation industry recently employed almost 9,000 The Wall Street Journal people (figure B8.2.1), who are paid approximately 15 percent Source: Reprinted with permission from Wall Street Journal. more than the country’s average monthly wage of about US$320. Note: GIMAS = Moroccan Aerospace Industries Association. Sources: Michaels 2012; interviews with people from the private sector in Morocco. and quantify value added and manufacturing links; they often have to be based on ad hoc surveys. Firm- nevertheless, those data sources have limits. With level official information, such as census and bal- I-O tables, in most cases, relationships cannot be ance sheet data, seldom has the detailed information documented at the fine level of disaggregation neces- required, and when it does, access to the data is very sary. GVC frameworks that use firm-level analysis to frequently restricted. Improving such data sets would determine the different stages of production of a sec- enable analysts to apply new theoretical GVC mod- tor and the value of each task are very costly, because els to firm-level data. One such attempt has been Entering GVCs 155 made by Del Prete and Rungi (2015), who apply the Strategic analysis as applied to GVCs requires property rights model by Antràs and Chor (2013) to the utmost attention to the international (regional Italian firm-level data.10 and global) dimension of production and demand. For these reasons, the assessment must be based on Specifically, it requires a market analysis that includes different sources and methodologies. Methodologies a benchmarking exercise that (1) assesses key attri- for identifying tasks within sectors include gathering butes of firms (that is, quality, price, reliability of qualitative information from one or more of the fol- supply, flexibility, and time from order to delivery) lowing: industry associations, chambers of commerce against local, regional, and global competitors; and and industry, ministries of trade and industry, com- (2) keeps an eye on how technology and increas- panies, experienced technical experts in companies ing global integration of goods, services, capital, and academic centers, and existing value chain case and knowledge flows change the boundaries of the studies. For a detailed assessment of suitable tasks, industry and the competitive space. Such an analysis countries can follow methodologies that combine also should segment end-markets as much as pos- strategic analysis with cluster change-management sible, because generally multiple actual and potential tools (Christensen and Kempinsky 2004; Conejos end-markets exist, each with different demand char- and others 2000; USAID 2006).11 These tools are not acteristics and returns, as well as different opportu- a substitute but a complement to the analyses sug- nities and challenges. Finally, identification of the gested in this chapter. policy area of intervention should be multidimen- The methodology used for the strategic analy- sional, assessing the subnational, national, regional, sis usually is based on the concepts initially devel- and multilateral levels. Such an approach allows an oped by Michael E. Porter, a professor at Harvard assessment of the strategic positioning of firms and University.12 It includes evaluating the sources of recommendations for attainable strategic options in a cluster’s competitive advantage; detailed and the global context. forward-looking industry analysis, with emphasis on Once strategic analysis has identified suitable future trends; evaluation of the strategic positioning tasks, methodologies focusing on the process of for the producers and firms in a country and rec- change can help specify actions to generate short- ommendations for attainable strategic options; and term results and engage an industry or a cluster in highlights of the value chain in which firms operate the dynamics of change (USAID 2006). Box 8.3 pro- and main areas of improvement. vides examples of how such methodologies allow Box 8.3. Examples of Strategic Analysis and the Dynamics of Change Management: The Ventilation Industry and the Truck Cluster in Sweden Companies of all sizes are globalizing production, often through (5) generic strategic options for the future, (6) key success fac- value chain clustering. Low-cost countries may create satellite tors for the options, (7) ideal value chain and cluster diamond for clusters of companies to a lead firm, as seen with Bangalore’s chosen option/s, (8) benchmarking of the cluster against the refer- hi-tech cluster or Timisoara’s footwear and auto clusters. High- ence/ideal cluster, (9) feasible options for firms in the cluster, and income countries have responded by moving jobs and business (10) areas of improvement. models in an entire industry or cluster to higher-productivity That approach allowed the country’s ventilation industry to tasks. Strategic analysis methodologies have helped companies understand the need to shift from selling (heating, ventilating, and redefine their business model and identify tasks of comparative air conditioning equipment) to selling clean air services—to stay advantage. Greater value added can be created either through in business. The approach also encouraged the truck industry to incorporating new technologies originating from strong research refocus from selling trucks to offering full transport solutions. In and development capabilities, usually upstream from the produc- both cases, the shift entailed moving from selling products to sell- tion process, or through inserting (or expanding) value-added ser- ing concepts and services, such as fleet management systems. vices, which originate from a deep and sophisticated knowledge Although this approach seems most useful to the private sec- of customers with technology (such as using “Big Data”). tor, it is important for public policy, too. The approach can align Duch (2000) proposed an analysis of two clusters in Swe- private initiatives and public interventions. den—the ventilation and truck industries—based on 10 steps: (1) mapping of the cluster, (2) strategic segmentation, (3) evolution of the segment’s attractiveness, (4) advanced demand analysis, Sources: Conejos and others 2000; Christensen and Kempinsky 2004. 156 Making Global Value Chains Work for Development identifying tasks in high-income countries that are goods—faster than if they had the same information challenged by the loss of jobs and business to lower- as final goods producers.15 cost countries.13 The next subsections look more closely at sell- ers’ end-market and downstream risks and at buyers’ Which Risks? upstream risks. GVC integration entails not only economic benefits, but also risks on the sourcing and selling sides, and Sellers’ End-Market and Downstream Risks countries must be aware of those risks. Yet govern- A seller’s end-market risk has been discussed for quite ments cannot control the risks directly, because GVC some time. Sector, firm, or geographic concentration participation is the endogenous result of a choice is a potential source of high volatility in value added made by firms. and a likely determinant of sharp readjustments in a The seller’s risk refers to demand shocks, includ- country’s gross domestic product (GDP) during a cri- ing end-market risks, and to a wide range of other sis. By contrast, a diversified portfolio generally helps downstream risks along the value chain. Similarly, dampen price fluctuations, as having more products, the buyer’s upstream risks refer to supply shocks on firms, or production facilities in diverse geographic the sourcing side that result from unforeseen events areas is likely to lead to independent price dynamics, or bottlenecks taking place along the value chain of with smoothing effects on total earnings. Put differ- upstream suppliers. ently, a more diversified production portfolio should Downstream and upstream risks are larger in lead to a more stable stream of export revenues. GVCs than in non-GVC trade or exports based on The export diversification discussion applies well purely domestic value chains. The risks also are to a world of final goods exporters (rather than to larger for more complex goods, such as automobiles, a world characterized by importing-to-export in for which parts and components are produced in dif- which countries export intermediates to the lead ferent countries and assembled in one location. The firm or final goods producer). Suppliers in GVCs, by higher is the number of countries involved in key contrast, do not have that option because they often tasks of production and the higher is the customiza- produce specialized (customized) inputs for only tion of the task to the downstream output, the higher one or a few buyers (see, for example, figure 1.6 in is the exposure of participants to potential risks. chapter 1). The suppliers may also depend on the Conversely, exports of unprocessed consumer goods, technology and know-how provided by the lead firm. goods produced by purely domestic value chains— Or, in an effort to become a supplier in GVCs, they which are organized in a single country—or final might incur specific sunk cost investments, which goods produced in shorter and less sophisticated make finding alternative buyers more difficult. That GVCs are likely to be more resilient. risk also applies to contract manufacturers that pro- Downstream and upstream risks in GVCs can duce final goods for large buyers. more generally be related to operational risks because The risks are also greater for suppliers in GVCs of the supplier’s dependence on a monopsony for its than for lead firms. GVCs adjust quickly to demand product; multiple border crossings, modes of trans- changes in end markets, as lead firms seek to shift port, hand-offs, and countries; and disparate tech- the burden of risks (associated with declines in nology issues and security concerns.14 Risks can also demand) to supplier firms, especially when supply be caused by shifts in a firm’s strategies, such as GVC chains are well coordinated.16 Such burden shifting consolidation or task bundling. came through strongly in the economic crisis of 2008 A final risk is the uncertainty of firms in an eco- and importers’ ensuing inventory changes, revealing nomic downturn. Such uncertainty is greater for GVC countries’ vulnerability.17 For the apparel GVC, more peripheral firms and occurs more frequently declining demand from leading apparel-importing among upstream firms. When demand for final goods countries led to a fall in apparel volumes and values slows, exporters can continue for a while on inven- for suppliers in LMICs and to higher unemployment tory rather than order new intermediates. Having and more factory closures.18 less information about any fall in demand for final A “pecking order” of risk exists among suppliers. goods, suppliers of inputs may start avoiding risk— First-tier and second-tier suppliers tend to face less by cutting production and trade in intermediate risk than marginal suppliers. In a financial crisis or Entering GVCs 157 another unexpected shock, buyers tend to transfer Buyers’ Upstream Risks business from marginal outfits to their core opera- From a buyer’s perspective, the novel risk element tions. During the 2008–09 global trade collapse, relates to upstream supply shocks, because import- foreign-owned Polish firms were more resilient ing goods (or services) to export increases a buyer’s than average, partly as a result of intra-group lend- dependence on upstream inputs. Two such upstream ing mechanisms that supported affiliates facing risks are natural disasters and changes in—this external credit constraints.19 Many foreign-owned time—suppliers’ strategies. firms in Poland were turnkey suppliers for foreign The 2011 flooding in Thailand (box 8.4) and the multinationals. triple Tohoku disaster in Japan—earthquake, tsu- From a seller’s perspective, the major novel risk nami, and nuclear—starkly revealed the vulnerabil- elements in many value chains are changes in the ity of GVCs to natural events. Tohoku was especially strategies and management of lead firms. The asym- pernicious in automotive products, computers, and metric power relations between suppliers (compet- consumer electronics, where downstream producers ing with each other) and the lead firm (frequently, rely heavily on Japanese suppliers of specialized parts a buyer that is far downstream in the GVC with oli- and components.21 In addition to the severe effects it gopoly power) enable strategic changes. had on Japan’s economy, the Tohoku disaster took a Thailand’s high-technology and small and toll especially on other Asian countries, which have medium corporate sectors, for example, are highly higher shares of intermediate goods imports than dependent on the decisions of Japanese companies other parts of the world.22 in Thailand. Some of them, such as Nikon (cam- Changes in upstream supplier strategies may eras) and Yazaki (car parts), are shifting produc- also pose a risk for intermediate buyers in GVCs. tion of lower-value manufacturing to lower-income Suppliers that—because of the underlying GVC neighbors, such as Cambodia and the Lao People’s governance structure—have more market power or Democratic Republic.20 Improvement of regional target economic upgrading within the GVC could transport links is therefore increasing the opportuni- perform new tasks to supplement and build on exist- ties—and risks—for the region’s economies overall, ing ones. That type of change poses a threat to exist- as the lower-income countries continue their moves ing downstream suppliers of those tasks, particularly to attract foreign investment. Box 8.4. The Impact of Thailand’s 2011 Flooding Thailand’s 2011 flooding—combined with the government’s inef- The flooding had a ripple effect on final production in other ficiency in managing the recovery—led to price hikes and produc- countries. Shortages of auto parts from an inundated plant in tion cuts in third countries. Ayutthaya forced Honda to cut production around the world.a It The flooding hit many industrial clusters in central areas. also caused price hikes for hard disk drives, because of the direct According to a business survey by the Bank of Thailand in impact of production stoppage and the indirect impacts of defen- 2012, 43 percent of businesses reported that usual operations sive purchases by consumers and inventory hoarding by resellers could be restored within only three months, 46 percent in four and wholesalers.b to six months, and the remaining 11 percent in more than six The flooding and the government’s inefficiency in managing months. flood recovery have raised investor concerns about rising pro- Manufacturing was hit hardest. Whereas 56 percent of manu- duction costs stemming from higher insurance premiums and facturing firms reported that the impact on their businesses was firms building their own flood defenses. Those concerns could “severe” or “very severe,” only 41 percent of nonmanufacturing undermine Thailand’s longer-term investment attractiveness. Of firms made that claim. In contrast, 31 percent of nonmanufactur- 50 multinational firms directly affected by the floods, 38 percent ing firms reported “no or a small impact,” but only 14 percent of intend to “scale back” activities. manufacturing firms reported that level of impact. The stronger impact on manufacturing stemmed largely from disruptions of Source: JETRO 2012, reported in Ye and Abe 2012. intermediate input supplies in the automotive and electronics sec- a. Chongvilaivan (2012). tors, and in computers and optical instruments. b. Ye and Abe (2012). 158 Making Global Value Chains Work for Development if the upstream supplier manages to offer the bun- between the lead firm and its suppliers usually occur dled tasks at a competitive cost. at arm’s length, requiring little or no formal coop- eration, and the costs of switching to new partners Which Form of Governance?23 is low on both sides. The central governance mecha- nism is price, rather than a powerful lead firm. As GVCs have developed, and suppliers have increased their technological sophistication and Modular Governance scale of operations, the dichotomy between in-house The modular governance structure exists when com- (“make”) and arm’s-length (“buy”) global supply plex transactions are fairly easy to codify. Suppliers relations has given way to a multiplicity of lead firm– in modular chains typically make products to a supplier relations. Those relations involve various customer’s specifications and take full responsibil- degrees of investment, technical support, and long- ity for process technology, using generic machinery term contracting and monitoring, as reflected in the that spreads investments across a wide customer growing importance of NEMs for internationaliza- base. Such governance often appears in industries tion. Largely for that reason, the form of governance dominated by transactions between a lead firm and matters (box 8.5). turnkey, full-package suppliers—especially in the domains of autos, apparel, footwear, electronics, and Which form of Governance between Lead Firms business services. This structure keeps the switching and Suppliers? costs low and transaction-specific investments few, GVCs can be organized in one of five governance although buyer-supplier interactions can be very structures: market, modular, relational, captive, and complex. The links (or relationships) are more sub- hierarchy (figure 8.2).24 They can be measured by stantial than in simple market structures, because of three variables: complexity of information between the high volume of information flowing across the actors in the chain, how the information for produc- inter-firm link. Information technology and stan- tion can be codified, and supplier competence.25 dards for exchanging information are key to how this structure functions. Market Governance The market governance structure involves fairly sim- Relational Governance ple transactions. Information on product specifica- With the relational governance structure, buyers and tions is easily transmitted, and suppliers can make sellers rely on complex information that is not eas- products with minimal input from buyers. Exchanges ily transmitted or learned, which leads to frequent Box 8.5. Why the Form of Governance Matters The scope for entering global value chains (GVCs) and deter- flexibility, diversification of location to reduce risk, and lower pro- mining the value of exports in GVCs is not fully in the hands of duction costs.b Comparing vertical foreign direct investment ver- countries. Most lead firms decide strategically where to produce sus arm’s-length outsourcing in a North-South framework, Grover (domestically or offshore) and whether to make some levels of (2011) postulates that outsourcing is more welfare enhancing in the value chain abroad (foreign direct investment) or buy them the South if the domestic absorptive capacity, defined as the ratio from an external firm either at arm’s length (domestic or offshore of skilled relative to unskilled labor, is above a certain threshold outsourcing) or through non-equity modes of investment, such as level. Bernard and others (2010) show that the choice between contract manufacturing. these two forms of governance is sector-specific. The firm’s governance decisions go beyond mere transactions Similar to firms, countries should think strategically about the costs and core competencies. A theoretical model may be con- forms of GVC participation that will best advance their develop- sidered, in which firms, on the basis of productivity and sector ment goals. The firms may not be able to choose the governance characteristics, decide whether to integrate production of inter- structure, but they should be aware of how governance character- mediate inputs or outsource it.a Firms with different productivity istics can mediate the impacts of GVC participation—and there- levels choose different ownership structures and supplier loca- fore condition firms’ decisions. tions, and those choices affect the relative prevalence of different organizational forms. But the motives for offshoring and outsourc- a. Antràs and Helpman (2004). ing for the strategic firm also encompass the pursuit of greater b. Milberg and Winkler (2013). Entering GVCs 159 Figure 8.2. Five GVC Governance Structures a. Market b. Modular c. Relational d. Captive e. Hierarchy END USE Customers Lead Lead Lead Integrated firm firm firm firm Full-package Relational Price supplier supplier MATERIALS Suppliers Component & Component & Captive material material suppliers suppliers suppliers From low to high degree of explicit coordination From low to high degree of power assymetry Source: Adapted from Gereffi, Humphrey, and Sturgeon 2005, 89. Note: GVC = global value chain. interactions and knowledge sharing between the par- production capabilities does not encroach on this ties. Such links require trust and generate mutual reli- attribute but benefits the lead firm by increasing ance, which are regulated through reputation, social the efficiency of its value chain. Ethical leadership is and spatial proximity, family and ethnic ties, and the important to ensure that suppliers receive fair treat- like. Despite mutual dependence, however, lead firms ment and an equitable share of the market price. still specify what is needed and thus have the ability to exert some level of control over suppliers. Producers Hierarchical Governance in relational chains are more likely to supply differen- With the hierarchical governance type of struc- tiated products based on quality, geographic origin, ture, chains are characterized by vertical integra- or other unique characteristics. Relational links take tion, and managerial control exists within lead firms time to build, so the costs and difficulties in switching that develop and manufacture products in-house to a new partner usually are high. (make). This structure usually exists when product specifications cannot be codified, products are com- Captive Governance plex, or highly competent suppliers cannot be found. With captive governance structures, small suppli- Although it is less common than in the past, this sort ers depend on one or a few buyers that often wield of vertical integration is still important in the global a great deal of power. Such networks feature a high economy. degree of monitoring and control by the lead firm. The power asymmetry forces suppliers to link to Buyer- or Producer-Driven Value Chains? their buyer under conditions set by—and often spe- Over time, any of these forms of governance can cific to—that buyer, leading to “thick” ties and high change as an industry evolves; similarly, governance switching costs for both parties. The core compe- patterns within an industry can vary from one link tence of the lead firms tends to be in areas outside of the chain to the next. Depending on the nature production, so helping the suppliers upgrade their of the lead firm in the chain, the analysis can also 160 Making Global Value Chains Work for Development distinguish between buyer-driven and producer- consider a firm’s location. Firms in GVCs have two driven value chains.26 The former occur mainly in basic locational decisions: stay in their home country consumer products, such as apparel, footwear, and or open an affiliate or NEM abroad. These options toys. The GVC is driven by large retailers that do generally apply to lead firms and large, first-tier sup- not manufacture but focus instead on design and pliers with market power in GVCs. Lower-tier sup- marketing and subcontract the production. The lat- pliers do not have the capacity to carry the sunk costs ter are typical in industries such as automobiles and of foreign investment. Although governments can- aeronautics, which require mid- to high-technology not directly influence that decision (it is the firm’s), production as well as substantial scale economies. they can adopt policies to attract FDI or NEMs. They are driven by multinational producing firms A major determinant of country policies to attract that may subcontract some aspects of production FDI is the potential to deliver substantial knowledge but that keep research and development and final or productivity spillovers for local firms and work- goods production at the firm. However, there are ers. A vast set of empirical evidence has been amassed major cases of buyer-driven relationships, in which over the past decade on the existence and direction the buyer focuses solely on the postproduction seg- of FDI-generated horizontal and vertical produc- ment—for example, the BMW (buyer)–Aprilia (pro- tivity spillovers. In a comprehensive meta-analysis, ducer) relationship in motorcycles. Havranek and Irsova (2011) take into account 3,626 estimates from 55 studies on FDI spillovers and find Which Power Relations in GVCs? 27 evidence for positive and economically important The governance structure of GVCs is particularly backward spillovers from multinationals on local important, because it defines the GVC’s power rela- suppliers in upstream sectors and smaller positive tions, which determine how financial, material, and effects on local customers in downstream sectors. human resources are allocated within the chain.28 However, the authors reject the existence of horizon- GVCs with different governance structures have dif- tal FDI spillovers. ferent degrees of power asymmetries (box 8.6 and Local firms, including NEMs, can similarly ben- figure 8.2), including GVCs in agrifood, consumer efit from international trade within GVCs, particu- electronics, textiles, and apparel. larly when exporting inputs to international buyers Although extremely useful from an analytical abroad, but also when importing intermediates from perspective, the five governance structures do not international suppliers. The extent of spillovers to Box 8.6. Four Strategies to Widen Power Asymmetries in GVCs Asymmetric power often is endogenous to the formation and gov- segments in-house. Such risks include end-market and down- ernance of some global value chains (GVCs), as oligopolistic lead stream risks. firms follow a cost-cutting strategy managed through offshore • Branding is a textbook example of constructing an entry bar- sourcing in GVCs. Such endogenous asymmetry can take a variety rier. Despite considerable theoretical analysis of entry barri- of forms, depending on the lead firm’s strategic focus. Four strate- ers, study of the economics of pure branding within GVCs has gies stand out: been limited. Branding tilts the bargaining power in produc- tion to the firm that holds the brand. In industries with stan- • Inducing competition is the process of diversifying among dardized production technology—including apparel, footwear, suppliers to spur competition among them. Playing one sup- airlines, computing (at times), consumer electronics, and auto- plier off another, working with multiple suppliers, and even mobiles—branding is a key part of a lead firm’s strategy. creating new supplier firms have become standard strategies • Minimizing technology sharing is a strategy in which lead of lead firms in GVCs, to keep input prices low. Such diversi- firms protect their proprietary assets through patents, trade- fication also reduces risk after, say, a political, economic, or marks, and other forms of intellectual property regulations, natural disaster in a country or a unionization effort or work especially when investing abroad, to reduce the amount of protest at a given plant. Inducing competition is easiest where potential technology leakages. For example, Boeing carefully global capacity is already excessive. controls technology in its sourcing with Japanese, U.K., and • Offloading risk to suppliers has been documented in a vari- U.S. parts producers.a ety of industries, including apparel and electronics. The surge of offshoring and outsourcing practices also helps lead firms Source: Milberg and Winkler 2013. offload risks that they previously faced when producing those a. Nolan, Sutherland, and Zhang (2002). Entering GVCs 161 domestic suppliers depends on the type of gover- “competitive spaces”—enclave locations where the nance structure between the lead firm and its local rules of business are different from those that pre- suppliers. vail in the national territory, and the costs of factors In addition to the governance structure in GVCs, are lower. An example is export processing zones international buyer characteristics can mediate (EPZs), which are rapidly built sites equipped with potential spillovers from belonging to a GVC. The excellent infrastructure, streamlined procedures, and buyer’s motives (whether market, cost, resource, or favorable tax conditions (such as tariff drawbacks on asset seeking), global production and sourcing strat- imports of intermediates). egies (which could also involve co-sourcing and co- location), technology intensity, home country, and Jumpstarting GVC Entry through the Creation of the duration of supplier relations can all—through EPZs and Other Competitive Spaces 29 international trade—influence potential spillovers in In many lower-income countries, exports come over- a way similar to how foreign investor characteristics whelmingly from EPZs, which—along with the other mediate FDI spillover potential. types of competitive spaces—can provide a way for Likewise, some host country characteristics and the country not only to attract foreign capital, but institutions that are important for FDI spillovers can also to connect the local labor force to established lead to spillovers through domestic firms’ involve- GVCs. The critical second step is then to connect the ment in international trade. Host country character- competitive spaces to the rest of the economy.30 So, istics and institutions that affect the availability and within the framework of GVCs, competitive spaces quality of labor (a country’s learning and innovation have a clear rationale, but empirical research also infrastructure) and the international movement of shows that their ability to generate development goods and services (a country’s trade policy) are of yields mixed results. The case of EPZs illustrates the major importance to spillovers. These mediating fac- complex issues that converge to determine the ability tors for spillovers in GVCs will be discussed in more of competitive spaces to deliver development. detail in chapter 9. EPZs are spaces in a country that are intended to attract export-oriented companies by offering those companies special concessions on taxes, tariffs, and Policy Options regulations. Some of the typical special incentives for Lead firms think strategically when making deci- EPZs include the following: sions, so governments should too, when reviewing two sets of policies: (1) creating world-class links in • Exemption from some or all export taxes GVCs to optimize international flows of inputs and • Exemption from some or all duties on imports of outputs among production facilities and create effi- raw materials or intermediate goods cient links with global markets and (2) creating a • Exemption from direct taxes, such as profit, world-class business climate for foreign tangible and municipal, and property taxes intangible assets (see the example of Bulgaria in fig- • Exemption from indirect taxes, such as value- ure 2.14 in chapter 2). added tax on domestic purchases • Exemption from national foreign exchange Creating World-Class GVC Links controls • Free profit repatriation for foreign companies Countries can join GVCs either by facilitating • Provision of streamlined administrative services, domestic firms’ entry or by attracting foreign inves- especially to facilitate import and export tors. The foreign investment option includes more • Free provision of enhanced physical infrastruc- direct access to foreign know-how and technology. ture for production, transport, and logistics Countries such as Costa Rica and Thailand have managed to attract FDI and turn it into sustain- Other, less transparent features of EPZs some- able GVC participation in very different ways. In all times provide further incentives for firm investment cases, however, providing excellent infrastructure, and export. One such feature is a relaxed regulatory streamlined export procedures, and a tariff-friendly environment, including labor rights and standards environment is necessary. One way to jumpstart that (notably the right to unionize), foreign ownership, process, particularly for countries with poor national and leasing or purchasing of land. Another fea- infrastructure and high import tariffs, is to create ture (although clearly not available to all countries 162 Making Global Value Chains Work for Development simultaneously) is an undervalued currency that Some notable exceptions include the Republic of renders costs lower (in foreign currency terms) and Korea, where the share of inputs purchased from the raises export competitiveness. domestic economy rose from 13 percent in 1972 to EPZs continue to contribute an important share 32 percent in 1978 and remained that high through of national gross exports in many LMICs, particu- the 1980s.33 The country’s EPZs were set up to attract larly lower-income economies. During the 1990s, foreign investment and promote the electronics sec- many countries vastly expanded their EPZ exports: tor. The level of integration is particularly impres- Costa Rica’s EPZs, for example, shot up from 10 sive, given that about 80 percent of investment in the percent of manufactured gross exports in 1990 to EPZs was foreign. The state played an important role 50–52 percent in the early 2000s; Bangladesh saw its in fostering the link by providing duty drawbacks to gross EPZ exports rise from 3.4 percent in 1990 to non-EPZ firms in its “equal footing policy.”34 21.3 percent in 2003.31 In some smaller LMICs, EPZ Technology spillovers also are limited, as the exports accounted for 80 percent or more of gross low-skill, assembly-type production so common in exports in 2006. EPZs is simply not conducive to technology transfer. For EPZs to contribute to sustained economic And the higher skill-intensive EPZs, such as those development, however, they have to be linked to the involving software or other business services, often rest of the economy. The problem is that, by their are enclaves, de-linked from the rest of the economy nature, they resist such links for several reasons. For except for its high-skills labor force. The technology one, EPZs are generally created to attract foreign is embodied in imported capital, and the knowledge firms to promote jobs and exports precisely because is embodied in management. Evidence shows—for domestic firms are uncompetitive internationally example, in the case again of Korea in the mid- and cannot generate foreign exchange. So, from the 1980s—that knowledge transfers increase when the start, domestic firms are behind in their capacity to skill intensity of production rises.35 provide low-cost, high-quality inputs to production At least two other characteristics of EPZs restrain in EPZs. their potential to advance development. First, EPZs Another reason for resistance may be that EPZs may indeed create employment and pay average are dominated by foreign firms that have well- wages slightly higher than those of similar jobs out- established relations with foreign input producers. side EPZs, but they generally have not been asso- Many foreign firms may follow a co-sourcing strat- ciated with notable improvement in wages and egy, relying on imported inputs from established labor standards. Second, EPZs raise an issue of the suppliers abroad, or they may follow co-location compatibility of some incentives with World Trade strategies that require established foreign input sup- Organization (WTO) agreements—notably, offshore pliers to enter EPZs. Most studies find that the back- production creates obstacles to aligning domestic ward links from firms in EPZs are minimal, with onshore rules with best international practices.36 domestic orders remaining very low and technology spillovers rare. That finding underpins the terms- Attracting the “Right” Foreign Investors of-trade weakness for many LMIC manufacturing EPZs and other competitive spaces are a special exports. case. A sustainable longer-term strategy of invest- Moreover, most EPZs allow duty-free imports ment attraction requires that governments target of material inputs. Non-EPZ domestic firms may more general, nationwide measures. In designing not import inputs duty-free, putting them at a cost investment promotion measures, various factors are disadvantage in input production. The share of important for policy makers to consider, particularly inputs purchased from domestic suppliers com- factors that explicitly target FDI. Attracting foreign monly ranges from 3 to 9 percent, as reported for El investors is the first of two sets of approaches for Salvador, Guatemala, the Philippines, and Sri Lanka countries entering GVCs. The other is facilitating in the mid- to late 1990s. In the Dominican Republic domestic firms’ access to GVCs, which is the focus of in 2004, after 30 years of EPZ presence and robust the next section. growth in EPZ exports and employment, EPZs pur- Foreign investors vary in their potential to deliver chased 0.0001 percent of material inputs from the spillovers.37 Governments therefore must iden- domestic market.32 tify and attract the “right” foreign investors, taking Entering GVCs 163 steps that include assessing the nature of the invest- A key difference between GVC-led and other ment and the motivations of potential FDI or NEM avenues of development is that GVCs require gov- (for example, efficiency seeking/export platform, ernment coordination at the micro level. Still, gov- resource seeking, or market seeking), as well as their ernments should not aim to pick a sector as the technology contribution and the technology gap “winner” (box 8.7). They should instead help firms with domestic firms. Investment promotion should plan and encourage entry into the appropriate tasks not only focus on lead firms in GVCs, but also tar- and, consequently, densification of already-begun get turnkey global suppliers and, possibly, important GVC participation (see, for example USAID (2006), lower-tier suppliers.38 as discussed in Step 3). A light-handed industrial policy can foster par- The following are recommendations for designing ticipation in GVCs and links with the domestic public policy to attract FDI and NEMs with potential economy by overcoming market failures or captur- for spillovers.39 ing coordination externalities. Urban policy provides an analogy: if individual initiatives are completely • Keep the most important policies focused on uncoordinated, the result can be over-congested cit- ensuring an attractive general investment climate ies that fail in the basic goal of improving citizens’ and a trade-conducive policy environment. lives. At the other extreme, government control of • Ensure that investment policy explicitly considers every investment decision can stifle growth and the nature of investment and the motivations of innovation—and so also fail to improve everyone’s potential FDI and NEMs, as their degree of spill- lives—in cities, towns, and rural areas. over is likely to vary. Box 8.7. Lessons from Failed Industrial Policies Many countries have designed and run industrial policies to pro- essential to get the private sector to commit to investing in mote production transformation, reconversion, or upgrading. Some innovation and production. policies have achieved their objectives, but many others have • Capture by incumbents. Consultations with the private sector failed. Even the success stories include elements of failure over often end up being led by incumbents, but innovation and pro- time, as countries learn through trial and error. Focusing on the les- duction diversification also depend on the creation and expan- sons from success is common, but failure can be just as instructive. sion of new firms. Targeted mechanisms to encourage the cre- ation of startups are needed to avoid the risks of policies that • Indiscriminate subsidies. Granting subsidies without condi- will only help to maintain the status quo instead of catalyzing tions increases the risk of adverse selection of beneficia- dynamic change. ries and the development of assistance-dependent behavior • Low critical mass for investments. If the government’s contri- among firms. Such a policy rarely translates into productivity bution is too small, the government will not be able to mobi- improvements. lize the matching funds from the private sector. • Never-ending support. The absence of sunset clauses (a pro- • Short-term horizon and annual budgeting. The creation and vision that if a contract is canceled, neither buyer nor seller strengthening of domestic scientific, technological, and pro- shall be subject to penalty) in support programs to companies duction capabilities take time, so industrial policies with discourages efforts to increase productivity. short-term horizons and based on annual budgets tend not • Cathedrals in the desert. Building factories or research labo- to be credible. Multiyear plans and budgets are necessary to ratories in remote locations works only when it is part of a achieve results, but they require robust monitoring and evalu- broader plan for creating backward and forward links, or when ation (M&E) to correct failures during implementation. the policy is matched with programs to foster local infrastruc- • Lack of M&E mechanisms. The limited capacity to generate ture development. feedback between policy design and implementation reduces • Prevention of competition. Although the creation of new activ- the effectiveness of policies that evolve through trial and ities and industries may require support in the early stages error. That lack also narrows the scope for regularly revising (the traditional “infant industry” argument), gradual exposure the policy to reduce the risks of capture and adverse selection. to internal and external competition can ensure that those activities grow in a productive way. • Closed-door, bureaucracy-led priorities. This type of policy cuts the chances of generating the information flows and trust Source: OECD-WTO 2013b. 164 Making Global Value Chains Work for Development • Assess the appropriate technology contribution • Use industrial policy light-handedly. Weaknesses explicitly during FDI evaluation. The assessment in institutions, private sector capacity and orga- could include ascertaining how much the tech- nization, and skills and absorptive capacity are nologies that investors may bring are likely to the norm in low-income countries, which raises be absorbed in the economy, given the current an array of challenges to fostering links. The trick capacity. is to fashion a light-handed industrial policy (in • Target promotion efforts beyond original equip- chosen sectors that conform to reasonable pro- ment manufacturers and lead firms to tier-one jections of comparative advantage) that focuses global suppliers and beyond. This means that the on overcoming market failures or capturing requirements and incentives to promote spillovers coordination externalities, including packages of should be pushed down below the lead firms to infrastructure expenditures and public-private include first-tier—or even second-tier—suppli- vocational training. ers and the investors to whom they contract out operations. Helping Domestic Firms Find the “Right” Trade • Avoid bidding away the benefits of spillovers Partners and Technology Abroad by offering excessive firm-specific incentives to Governments can help potential buyers and sup- attract FDI and NEMs. Incentives tend to be most pliers—domestic and international—by making commonly associated with attracting export plat- the right connections, say, by setting up an online form investment, given its more footloose nature, firm directory that includes the sector, expertise, although realizing spillovers from exactly that and firm profile. Such directories should include type of investment may be the most challenging. information on certificates that local suppliers have • Recognize that the “right” investment to deliver obtained. Becoming a supplier to lead firms requires spillovers requires foreign and domestic inves- meeting specific quality, legal, labor, health, safety, tors, so ensure that investment policies are not environmental, and other standards. Walmart, for biased against domestic investors and that they example, provides a manual that includes “respon- support mutual interaction. EPZs are one exam- sible sourcing” requirements with which potential ple of bias: they often are established primarily suppliers need to comply.41 And the International for foreign investors and may have explicit or de Trade Centre (United Nations Conference on Trade facto barriers to domestic investors. Countries and Development/WTO) has launched a tool called that are home to large and competitive compa- Standard Maps, which provides comprehensive, nies have an advantage in attracting FDI because verified, and transparent information on voluntary the domestic firms can act as turnkey suppliers. standards and other initiatives, covering issues such Countries in which firms are predominantly small as food quality and safety. This tool also includes and medium enterprises (SMEs) find attracting self-assessments for producers to rate their business FDI more difficult and so become inclined to pro- against standard requirements.42 vide overly generous incentives. Devoting some of Government assistance can also include e-tools those resources to helping SMEs become part of a to help domestic companies (1) commercialize their well-established and integrated industrial cluster intellectual property, (2) identify and take advan- brings greater “bang for the buck.”40 tage of freely available technologies, or (3) assist • Facilitate joint ventures (JVs) where they can them to establish licensing agreements, as Morocco add value, but avoid coercion. JVs seem to be does through the Office Marocain de la Propriété effective for facilitating spillovers, particularly Industrielle in the framework of its Horizon 2015 of older technologies and know-how (which, for program.43 Other practical advice that governments low-income countries, are likely to be most rel- can provide to potential local suppliers includes the evant). However, this admonition should not be requirements they must meet to become exporters misread as encouraging attempts to force inves- of intermediates. Effective forms of matchmaking tors to engage in JVs with local partners. The cor- include holding buyer-supplier fairs or meetings. relation depends on the FDI or NEM motive, and The government’s role also covers the promo- demand-led JVs are more likely to share knowl- tion and marketing of exports and imports. Export edge openly than are forced partnerships. promotion ranges from country image building, to export support services (such as trade fairs), to Entering GVCs 165 market research and publications. Japan’s External regulatory matters (licensing, implementation, Trade Organization (JETRO), for example, has been enforcement, or trade facilitation at the border). In successful in promoting exports partly because of short, policy is key for creating an overall conducive its emphasis on researching foreign markets and environment for logistics services (figure 8.3). providing information to Japanese firms.44 Chile’s Rarely does a single “magic bullet” of policy reform export promotion agency—the Chilean Trade exist, and improving the international connectivity Commission, or ProChile, for short—has helped of a country touches on many dimensions: tighten- promote Chilean salmon in the U.S. market, work- ing forward and backward links within GVCs; secur- ing with Canadian producers.45 Chile’s 2001 Interna- ing the flow of inputs and outputs; creating efficient tionalization Plan has helped improve the exporting links with global markets; reducing “the thickness of skills of smaller exporters and encouraged new SME borders;”47 lowering traditional barriers to trade; and exporters (box 8.8). promoting trade facilitation. Improved connectivity In a world of GVCs, however, importing to export also serves the goals of GVC participation by lower- also requires public efforts to focus on import pro- ing costs, increasing speed, and reducing uncertainty. motion, because a country’s ability to participate in Regarding cost reduction, GVCs have changed GVCs depends on its capacity to import world-class the perspective on traditional barriers to trade, such inputs. JETRO, for instance, established import pro- as tariffs. Some recent studies suggest that reducing motion facilities as early as the 1990s to adapt to the supply chain barriers to trade (border administra- increasing openness of Japan’s trade.46 tion, transport and communications infrastructure, and related services) would have greater impact on Improving Connectivity to International Markets the growth of GDP and trade than the complete How effectively does a country’s logistics infrastruc- elimination of tariffs. Cutting supply chain barriers ture operate and connect to its neighbors and to to trade could increase GDP by nearly 5 percent and global markets? Geography plays a role, with coun- trade by 5 percent, against less than 1 and 10 percent, tries in remote locations (such as Chile, Kazakhstan, respectively, for complete tariff removal.48 LMICs and Mongolia) or with large archipelagos (Greece would be the main benefactors of trade facilitation and Indonesia) at a disadvantage. However, policy (figure 8.4). Transport costs, according to LMIC sup- also matters for logistics performance, whether pliers, remain the main obstacle to entering, estab- for infrastructure investment and operation or lishing, or upgrading in GVCs.49 Box 8.8. Chile: ProChile Internationalization Plan Chile is a middle-income country that relies heavily on mining and Chile co-financing programs if they have promising export plans. metals but has substantial agricultural export capacity. In the past The programs take about one year to complete. ProChile covers up two decades, Chile has become a major export success in agricul- to 90 percent of the cost if participants have an exportable prod- ture and agroprocessing, including products such as salmon and uct for which international demand exists and if they use labor- wine, and horticulture. ProChile is widely acknowledged as hav- intensive production methods. ing played a critical role in the country’s export growth. Since the early 1990s, the number of exporters in Chile has To improve the export skills of smaller existing exporters and doubled. The diversification of sectors, products, and markets has encourage new small and medium enterprise (SME) exporters, been dramatic, with the number of new products doubling, the ProChile developed its Internationalization Plan in 2001. One com- number of markets growing by more than 50 percent, and the rela- ponent, Interpac, is for agricultural SMEs; the other, Interpyme, is tive concentration of the mining sector falling sharply. Between for industrial SMEs. Those programs provide Chilean companies 1996 and 2006, Chile’s nontraditional exports (which account for with systematic training in the exporting issues faced by SMEs. 90 percent of its SME exports) increased from US$6 billion to The programs include training parts on production capabilities, US$15 billion. Several impact evaluation studies have shown that market research, logistics, marketing plans, banking, international ProChile has had a positive and significant impact on export par- law, searching for partners, and the export process. Interpac and ticipation, new product introduction, and firm-level technological Interpyme are operated by a team of private sector consultants and management improvements. hired by ProChile, and participants receive individualized, one-on- one counseling. They complete one part at a time, and when they have completed the full program, they become eligible for Pro- Source: Partly derived from Nathan Associates 2004. 166 Making Global Value Chains Work for Development Figure 8.3. Logistics Services in a Typical Supply Chain Producer Domestic Export International transport gateway transport Consolidate, load, Intermodal transfer, Load, transport, deliver, unload storage, clearance unload Distributor/ Domestic Import wholesaler transport gateway Consolidate, load, Intermodal transfer, deliver, unload storage, clearance Domestic Distribution Domestic transport center transport Consolidate, load, Deconsolidation, storage, Consolidate, load, deliver, unload inventory management, deliver, unload packaging, labeling Retailer Although drivers for offshore outsourcing often chains. Logistics performance captures the different have been linked to a desire to cut labor costs, other dimensions of supply chain efficiency, including how drivers include predictability, reliability, and timeli- supply chains connect globally and regionally and ness—that is, increased speed and reduced uncer- how each is influenced by national endowments and tainty.50 Many countries cannot join certain stages of policies. The three pillars of logistics performance are GVCs because of their inability to meet requirements for timely production and delivery; time really is • Availability and quality of trade-related infra- money. A day of delay in exporting has a tariff equiv- structure: ports, airports, roads, and railroads alent of 1 percent or more for time-sensitive prod- • Friendliness and transparency of trade proce- ucts.51 Slow and unpredictable land transport keeps dures implemented by customs and other border most of Sub-Saharan Africa out of the electronics control agencies value chain.52 Sellers often are willing to pay more for • Development and quality of logistics services, such air freight. Delays in GVCs also create uncertainty, as trucking, warehousing, freight forwarding, ship- inhibiting countries from participating in GVCs for ping and customs clearing, and value-added logis- goods such as electronics or fruits and vegetables.53 tics services (third- and fourth-party logistics) To guide policy makers in enacting reforms of the logistics sector, the World Bank launched the now Logistics performance and the ability of countries widely accepted concept of logistics performance to connect to international markets therefore depend in 2007. The World Bank also introduced a frame- on a range of policy interventions that can be imple- work—now a standard—to analyze national supply mented at the national or, increasingly, regional level. Entering GVCs 167 Figure 8.4. Reducing Supply Chain Barriers: Impact on GDP and Trade Growth 71% 33% 2% 6% 11% Russian Federation 9% 9% 7% Europe, except FSU and other FSU 26% Korea, Rep. United States and 10% Canada Other oil producers 46% 11% 34% 3% 65% 49% Japan Non-Oil Middle East 26% and North Africa 11% 31% 34% South and Central Asia Mexico 74% China; Hong Kong SAR, China; 18% 12% Taiwan, China 30% 63% 55% Southeast Asia Brazil Sub-Saharan Africa 38% 39% 1% 2% Oceania Rest of Americas GDP increase Trade increase 2-3% 4-5% 6-8% ≥9% Exports Imports Source: Reprinted with permission from WEF 2013. Note: GDP = gross domestic product; FSU = former Soviet Union. Priority areas for logistics performance improve- country’s competitiveness and ability to participate ment in most countries include the following: in GVCs depends as much on its capacity to import world-class inputs efficiently as on its capacity to • Regional integration and development of trade export processed or final goods. corridors: border crossings and transit regimes Second, trade within GVCs magnifies the costs of • Customs reform and trade facilitation tariff protection when intermediate inputs are traded • Border management extending beyond customs across borders multiple times, and the efficiency of • Port reform the value chain could be challenged if a country at an • Regulation and development of logistics services intermediate stage of production has high tariffs.55 (such as trucking, third-party logistics, freight Tariff escalation is a further, direct obstacle to the forwarding, and warehousing) offshore outsourcing of key stages of production; it • Development of performance metrics reduces the length of a GVC and the upgrading pros- • Public-private coalitions for reforms pects of LMICs in the chain. Customs efficiency can be another obstacle at the Addressing Obstacles at the Border 54 border, often in LMICs. One approach to simplify Policies on obstacles at the border (table 8.1) should border processing and clearance is a national single- address traditional barriers to trade, as well as cus- window system, in which buyers and sellers submit toms matters, notably efficiency and procedures, all information through a single electronic gateway. including rules of origin. When production is within But establishing such a system requires a strong gov- GVCs, addressing the traditional barriers to trade is ernment mandate supported by political will and much more critical, for two main reasons. stakeholder engagement, as well as the cooperation First, GVCs broaden the scope of traditional of multiple government agencies, many of which export barriers to include barriers to imports: a have to undergo substantial institutional reform.56 168 Making Global Value Chains Work for Development Table 8.1. Addressing Obstacles at the Border: Policy Objectives and Performance Indicators Policy objectives • Address obstacles to trade at the border, including trade • Simplify customs procedures, including sanitary and facilitation phytosanitary; technical barriers to trade; and other • Suppress quotas and other quantitative restrictions on certifications, rules of origin, valuation, and so forth, to imports and exports conform with agreements or international best practices • Reduce tariffs, suppress tariff peaks and tariff escalation, or • Implement WTO or regional/bilateral commitments (for simplify tariff schedules example, common external tariff) • Modernize (reform) customs, and harmonize procedures and cooperation across borders Performance indicators • Trade restrictiveness indexes: OTRI, TTRI (WTI 1.1) • Export restrictions (WTI 1.13) • Binding coverage and bound rates (WDI) • Logistics Performance Index and its indicators—efficiency of • Share of tariff lines with peaks/specific rates (WDI, WTI 1.6) customs and other border procedures (LPI, WTI 4.1) • MFN applied tariffs: AV+AVE or AV only (WDI; WTI 1.2, 1.3) • Trading across Borders—Doing Business (IFC, WTI 4.2) • Applied tariffs, including preferences (WDI, WTI 1.4) • Trade-enabling and global competitiveness indexes—goods • Tariff escalation (WTI 1.5) market efficiency: burden of customs procedures, prevalence of trade barriers, trade tariffs, efficiency of customs • MFN 0 tariff lines/import value (WTI 1.7) administration, efficiency of import–export procedures, • Tariff bounds/overhang (WTI 1.8) transparency of border administration (WEF GCI 6.10, 6.11, • Non-AV tariffs (WTI 1.9) 6.13; ETI 1.01–4.02) • Nontariff measures (WTI 1.10) • Average time to clear exports through customs/time to export/ • Customs duties (WTI 1.11) import (WDI) • Documents to export/import (WDI) Sources: Cattaneo and others 2013, based on OECD 2012a. Note: AV = ad valorem; AVE = ad valorem equivalent; ETI = Enabling Trade Index; GCI = Global Competitiveness Index; IFC = International Finance Corporation; LPI = Logistics Performance Index; MFN = Most Favored Nation; OTRI = Overall Trade Restrictiveness Index; TTRI = Trade Tariff Restrictiveness Index; WDI = World Development Indicators; WEF = World Economic Forum; WTI = World Trade Indicators; WTO = World Trade Organization. Increasing the Connectivity of Domestic Markets prior dominance of producing and consuming a ser- The policy objectives and measures in table 8.2 aim vice onsite. to increase the connectivity of domestic markets LMICs have caught up on ICT penetration through improvements in logistics and transport and interregional Internet bandwidth, which has and telecommunications, with a greater focus on increased their ability to produce and export ser- transport for goods and telecommunications for off- vices. But the poorest among those countries still shoring services. have a long way to go. The progress in LMICs has Importer logistics performance is associated with been accompanied by liberalization of services sec- higher components and parts trade.57 Its influence is tors, which has been fostered by constant privatiza- much higher for trade in parts and components than tion, competition, and independent regulation over for trade in final goods. The quality of logistics in the the past two decades. Most LMICs that are now importing country is thus an important determinant attracting large amounts of FDI in the services sector in a lead firm’s location decisions, but the relationship were characterized by protectionist policies before between logistics performance in the exporting coun- opening to foreign ownership of companies. try and trade in parts and components is less clear. In addition to logistics performance, the develop- Creating a World-Class Business Climate for ment of GVCs—particularly offshoring services—to Foreign Tangible and Intangible Assets58 a large extent has been fostered by information and communications technologies (ICTs). ICTs trans- Ensuring Cost Competitiveness While Avoiding the mit codified design specifications between actors in Trap of Low-Cost Tasks product-based chains and are the main medium for Low wages may be a way for countries to enter GVCs. participation in cross-border services exports. ICTs According to firm surveys, costs (production, labor, have enabled information to be uncoupled from transport, and investment) and tax incentives are physical storage, which renders the transfer of huge major drivers of lead firms’ decisions to invest or amounts of data possible in seconds, eroding the source production in LMICs. Wage differentials have Entering GVCs 169 Table 8.2. Increasing the Connectivity of Domestic Markets: Policy Objectives and Performance Indicators Policy objectives Increasing the accessibility and connectivity of the • Reform transport, logistics, and ancillary services, including domestic market and the security, predictability, infrastructure, regulation, competition for land (road and rail), reliability, and efficiency of transports/logistics, maritime/water, and air telecommunications, and ICT: • Harmonize regional infrastructure for trade corridors, and • Reform the telecommunications sector, including ensure other forms of regulatory cooperation infrastructure, regulation, competition, and access for all • Improve vertical governance in infrastructure, including through segments; to include fixed lines and mobiles fast-tracking and streamlining the regulatory environment, • Develop the ICT sector and the Internet (infrastructure, private–public dialogue on regulatory changes needed, and regulation, competition, access) enhancement of budget capital execution Performance indicators • Logistics Performance Index and its indicators—quality • Number of seats available, airlines, international routes, and of transport and IT infrastructure, international transport airport passenger statistics (IATA, WDI) costs, logistics competence, traceability and timeliness of • World telecommunication/ICT indicators database and ICT shipments, and domestic transport costs (WDI, LPI, WTI 4.1) development index (ITU) • Trading across Borders—Doing Business (IFC, WTI 4.2) • Foreign participation/ownership in telecoms (ITU, WTI 1.14) • Trade enabling and global competitiveness indexes— • Competition index in telecoms (ITU, WTI 1.14) infrastructure: quality of infrastructure overall, roads, • Number of international gateways, landing stations, licenses railroads, ports, air transport, available seats, fixed telephone for fixed and mobile phones, and Internet providers (national lines/100, mobile phone subscriptions/100, availability data, WB and OECD STRI) and quality of transport infrastructure and services, and availability and use of ICTs (WEF GCI 2.01–2.09, WEF ETI • Mobile and fixed-line telephone subscribers/population 4.01–7.05, WDI) covered by mobile cellular network (WDI, WTI 4.4) • Technological readiness (WEF GCI 9.01–9.06) • Average cost of 3-minute call to the United States (WTI 4.4) • Africa Infrastructure Country Diagnostic (AICD) • Personal computers (WTI 4.4) • Liner Shipping Connectivity Index (UNCTAD, WTI 4.3) • Internet or broadband users/subscribers (WDI, WTI 4.4) • Baltic Exchange Dry Index (WTI 4.3) • Internet bandwidth and secured servers (ITU, WDI) • Lead time to export/import (WDI) • Port container traffic (WDI, WTI 4.3) • Total/air freight and costs (WTI, 4.3) Sources: Cattaneo and others 2013, based on OECD 2012a. Note: AICD = Africa Infrastructure Country Diagnostic; ETI = Enabling Trade Index; GCI = Global Competitiveness Index; IATA = International Air Transport Association; ICT = information and communications technology; IFC = International Finance Corporation; IT = information technology; ITU = International Telecommunication Union; LPI = Logistics Performance Index; OECD = Organisation for Economic Co-operation and Development; STRI = Services Trade Restrictiveness Index; UNCTAD = United Nations Conference on Trade and Development; WB = World Bank; WDI = World Development Indicators; WEF = World Economic Forum; WTI = World Trade Indicators. been primary drivers of the globalization of produc- labor productivity growth are sometimes viewed as tion. But costs encompass a wide range of drivers, competing goals, if more value added is created with and high costs could, for example, stem from a lack the same amount of workers (static productivity of infrastructure or competition in basic services. effects), this book argues that GVC integration not High costs could also result from excessive adminis- only has a strong potential for productivity gains via trative burdens (including those at the border), strict several transmission channels (dynamic productivity labor laws (a weak business environment), or wide- effects), but also creates jobs because of the increased spread political and social insecurity or corruption. vertical specialization and value added in GVCs However, the goal should be higher labor pro- (densification), as discussed in the next chapter. ductivity and higher wages, allowing the country to Productivity and the capacity to meet production remain cost competitive despite rising living stan- requirements must also be considered when assessing dards. Unit labor costs in themselves are irrelevant. costs. If cost savings because of relocation go hand in For example, China remains competitive even with hand with productivity losses, lead firms might end rising labor costs. Recent research finds, for exam- up facing higher overall costs. Moreover, value chain ple, that minimum wage growth in China allows tasks based exclusively on labor cost advantages tend more productive firms to replace the least produc- to be easy to relocate. A strategy based on low wages tive ones and forces incumbent firms to strengthen exclusively is therefore risky and unsustainable over their competitiveness.59 Although job creation and the long term. Investment or tax incentives should 170 Making Global Value Chains Work for Development Table 8.3. Improving Cost Competitiveness While Competitiveness (World Economic Forum) reports Avoiding the Trap of Low-Cost Tasks: Policy Objectives provide lists of key measures for business operations, and Performance Indicators as well as indications of a country’s performance Policy objectives based on selected criteria. The range of measures is • Ensure cost competitiveness related to production, labor, very large, from the regulatory environment to the transport, and investment. functioning of markets (such as state trading enter- • Foster productivity gains, skills development, and prises and government procurement). Protection of technological empowerment. intellectual property is a decision tipping point for Performance indicators many lead firms. The cost of administrative burdens • Unit labor costs and wage data (ILO ILOSTAT and KILM, OECD) also becomes larger in GVCs, as management needs • Labor productivity (ILO KILM, OECD) to coordinate a wider cast of actors. • For skills development and technological empowerment A country’s political stability, governance, and indicators, see table 9.1. degree of corruption are other factors to consider Note: ILO = International Labour Organization; ILOSTAT = ILO database of labor in the decision to join a GVC. Those metrics (with statistics; KILM = Key Indicators of the Labour Market; OECD = Organisation for Economic Co-operation and Development. others, summarized in table 8.4) relate to security (including assets and personnel) and predictabil- ity, the key drivers of intra-firm GVC trade (FDI) be carefully used to foster productivity gains, skills and on-time delivery to consumers. Within GVCs, development, and technological empowerment (for suppliers often are expected to meet the lead firm’s a list of performance indicators, see table 8.3). corporate social responsibility codes, which raise challenges for audit and execution in small, LMIC Improving Drivers of Investment and Protecting firms.63 Foreign Assets To prevent a “race to the bottom” on incentives, Drivers of investment, particularly the protection of however, policy makers can seek to promote invest- foreign assets, have a large influence on a country’s ment through regional integration. That process location attractiveness for foreign investors. Those includes four steps: (1) identifying regional invest- drivers affect a country’s participation in GVCs, ment barriers (such as through intensive private regardless of their governance structure. sector consultations and interviews), (2) defining Protecting assets is mainly about protecting firm- the reform agenda, (3) implementing reforms, and specific technology and know-how, but only some (4) benchmarking reform progress against the of those elements can be defended through patents, defined reform agenda. Throughout the process, an trademarks, and other forms of intellectual property important measure is to engage the private sector laws. Many other elements cannot be protected this with the national public sector and regional institu- way, including business and organizational models, tions (such as through private-public dialogues) as a managerial practices, production processes, and export feedback mechanism and reform engine.64 procedures. As global production networks necessarily GVC entry through foreign investment requires involve contracting relationships between agents in maximum fluidity in the mobility of production fac- countries with differing legal systems and contract- tors. Barriers to FDI are likely to exclude a country ing institutions, contracts often are incomplete.60 The from major GVCs, or confine the country to cer- reasons for incomplete contracting in international tain forms of GVC governance. Stability clauses in settings include a limited amount of repeated interac- contracts and participation in major international tions; lack of collective punishment mechanisms; and (including regional) arbitration and dispute settle- natural difficulties in contract disputes, such as deter- ment mechanisms are also important (table 8.5). mining which country’s laws apply—and even when that is known, local courts may be reluctant to enforce Organizing Domestic Value Chains and Improving a contract involving residents of foreign countries.61 the Quality of Infrastructure and Services The way in which different national systems deal How well the domestic segment of the value chain with contractual frictions and incomplete contracts is is organized is as important as that for the inter- therefore important in driving firms’ choices of loca- national segment (see figure 8.3). The benefits of tion and boundaries in global sourcing.62 efficient transport and logistics at the border, for That statement is also proven empirically. The example, can be undermined by inefficient domes- annual Doing Business (World Bank) and Global tic links, including the unreliability or high cost of Entering GVCs 171 Table 8.4. Improving Drivers of Investment: Policy Objectives and Performance Indicators Policy objectives Intellectual property protection: Administrative burden: • Improve the intellectual property regime and administration • Adopt administrative reforms to simplify and reduce to comply with trade agreements, to include patents, authors’ administrative procedures (as an example, guillotine reform); rights, geographic indications, and so forth increase transparency, predictability, timeliness, and security • Improve enforcement mechanisms and practices of administrative decisions (for example, suppression of • Promote the intellectual property regime and related training authorizations) or technical assistance Other constraint resolution: Competition, including privatizations and concessions: • Create EPZs, business clusters, technology centers, and the like • Privatize, offer concessions, and open sectors to competition • Revise labor regulations for greater labor market efficiency • Elaborate and implement a competition framework, including • Revise regulations regarding the form of business operations competition law, competition authority (for example, and partnerships (for instance, franchises, multi-sector independence, resources), competition law enforcement (for partnerships) instance, investigations, sanctions), and related training or • Increase security for operations and staff against crime and technical assistance violence Government procurement: Promote investment through regional integration: • Adjust laws pertaining to public procurement, including • Eliminate barriers to expansion of cross-border investments transparency, selection criteria, national preference, and so within the region forth • Converge levels of investment protection within the region Corruption: and increase transparency to prevent “race to the bottom” on incentives • Reform to fight corruption in the public (for instance, customs) and private sectors • Promote and adopt international instruments for corruption reform Performance indicators • Ease of Doing Business Index (IFC, WTI 3.1, WDI) • Time spent in meetings with tax officials/expected gifts/ • World Governance Indicators—corruption, rule of law, informal payments to public officials (WDI) government effectiveness, regulatory quality, political • Firms using banks to finance investment (WDI) stability (WTI 3.2) • Strength of legal rights index (WDI) • Enabling Trade and Global Competitiveness Indexes • Time required to enforce a contract (WDI) • Regulatory environment (WEF ETI, 8.01-08) • Time required to obtain an operating license/register property/ • Institutions: property rights, ethics and corruption, undue start a business (WDI) influence, government inefficiency, security (WEF GCI • Value of seized counterfeited goods (national statistics) 1.01-1.16) • Number of registered trademarks, patents, and the like (WIPO, • Labor market efficiency (WEF GCI 7.01-7.09) WDI) • Goods market efficiency (WEF GCI 6.01-6.16) • Number of competition investigations and sanctions (national • Business sophistication: state of cluster development (WEF statistics) GCI 11.03) • Public procurement penetration ratio—public imports/public • Enterprise ownership (government, private foreign, private demand percentage (national statistics) domestic) (ADI) • Security costs (ADI) • Cost of business startup procedure/procedures to register a business (WDI) Sources: Cattaneo and others 2013; also based on OECD 2012a and World Bank 2014. Note: ADI = Africa Development Indicators; EPZ = export processing zone; ETI = Enabling Trade Index; GCI = Global Competitiveness Index; IFC = International Finance Corporation; WDI = World Development Indicators; WEF = World Economic Forum; WIPO = World Intellectual Property Organization; WTI = World Trade Indicators. domestic transport, the fresh product cool chain, and infrastructure (for 39 percent) were the two most low-quality storage. Regional markets and stocks are serious national supply-side constraints identified critical for agriculture’s inclusion in GVCs. by LMIC GVC suppliers as affecting their ability to Locational attractiveness to foreign investors is enter, establish, or upgrade in a GVC.65 also determined by the ease of access to efficient ser- The “servicification” of manufacturing is par- vices and infrastructure, including access to energy ticularly important as production international- (cheap and reliable), financial and trade support, izes because as many as 40 services may be involved telecommunications, and transport. Access to finance when a manufacturing firm internationalizes (figure (for 52 percent of the firms surveyed) and transport 8.5). Recent trade in value-added data suggest that 172 Making Global Value Chains Work for Development Table 8.5. Encouraging and Protecting Foreign Investment: Policy Objectives and Performance Indicators Policy objectives • Remove barriers to foreign investment • Allow more foreign equity/ownership/partnership • Facilitate the free movement and employment of key personnel across borders • Remove discriminatory policies (including licensing, taxes, subsidies, and so forth) • Increase the protection of foreign assets • Strengthen investor protection, including rights to challenge domestic regulations/decisions • Develop alternative dispute resolution mechanisms for foreign investors (for example, recognition of international arbitration, bolstering of domestic arbitration capacities) • Adjust the laws on nationalization, expropriation, foreign ownership, stability clauses, and the like Performance indicators • GATS commitments (WTO), regional commitments, and domestic laws • Services trade restrictiveness indexes (WB, OECD) • Arbitration awards (ICSID and other arbitration bodies’ statistics) • Protecting investors (ADI) Sources: Cattaneo and others 2013, based on OECD 2012a. Note: ADI = African Development Indicators; GATS = General Agreement on Trade in Services; ICSID = International Centre for Settlement of Investment Disputes; OECD = Organisation for Economic Co-operation and Development; WB = World Bank; WTO = World Trade Organization. services represent at least 30 percent of the share of efficient services (auditors, lawyers, and managers) value added in manufacturing trade (see also chap- and the movement of key personnel across borders ter 6).66 Thus, a country cannot be competitive and (table 8.6). join GVCs—even in manufacturing—unless it has Improving a country’s domestic logistics envi- efficient domestic services or is open to import- ronment—a key services sector in GVCs—requires ing them. Managing the complexity of the value infrastructural interventions and regulatory changes chain and preserving production standards along spanning many different sectors, as seen in the exam- it require strong coordination efforts that rely on ple of Greece (box 8.9). Figure 8.5. Services Involved in the Internationalization of Production (at Sandvik Tooling) Legal services Security services Insurances Accounting, bookkeeping etc. Packaging Health related services Taxation services Printing, publishing Hotel and restaurants Medical services Design Travel agency services Computer services Building-cleaning services Maritime transport - freight Research and development Photographic services Inland waterways - freight Rental/Leasing Courier services Air transport - freight/passenger Advertising Logistic services Road transport - freight/passenger Market research Postal services Cargo-handling services Services incidental to Telecommunications Storage and warehouse services manufacturing Audio-Visual services Freight transport agency service Placement of personnel Educational services Feeder services Maintenance and repair Environmental services Energy services Convention services Banking services Source: Reprinted with permission from the National Board of Trade 2010. Entering GVCs 173 Table 8.6. Improving Domestic Services Infrastructure and Market Structure: Policy Objectives and Performance Indicators Policy objectives Performance indicators Improving access to finance: • Banking GATS commitment index (USITC, WTI 1.14) • Reform the financial sector, including microfinance, to • Export credit—insured exposures (WTI 4.5) increase the affordability and availability of financial • Indicators of financial structure, development, and soundness services (IMF) • Ensure export credit and trade finance • Access to finance (WDI) • Enabling trade and global competitiveness indexes—financial market development (WEF GCI 8.01-8.08) Improving other domestic infrastructure, including storage • Procedures and time to build a warehouse (WDI) and energy: • Time required to get electricity (WDI) • Upgrade storage infrastructure • Energy statistics/access to electricity (IEA, WDI) • Reform access, regulation, and competition in energy • Quality of electricity supply (WEF 2.07) (production and distribution) and other natural resources • Power outages in firms/value lost in power outages (WDI) essential to certain activities (for instance, water in • Electricity cost (WTI 4.6) agriculture) • Pump price for fuel (WTI 4.6) Improving business support and the organization, • Logistics Performance Index and its indicators—quality of connectivity, and performance of markets, including transport and IT infrastructure, international transport costs, e-commerce: logistics competence, trackability and timeliness of shipments, • Adopt export and investment promotion and incentives and domestic transport costs (WB, WTI 4.1) • Global Competitiveness Index—business sophistication: • Give analyses and information on markets, opportunities, threats, and so forth extent of marketing, state of cluster development, value chain breadth, control of international distribution production process • Undertake marketing, branding, international presence, and sophistication, delegation of authority (WEF GCI 11.05-11.09) promotion efforts • Value of e-commerce, number of ICT firms, number of secured • Form sector, professional, or other forms of associations servers (WDI, ITU, national statistics) (such as chambers of commerce) and consultations • Postharvest losses (African Postharvest Losses Information • Develop trade corridors and other regional forms of hard System) and soft networks (for example, regional regulatory agency, regional distribution network) • Develop regional markets and stocks, boards of trade, and price regulation mechanisms • Organize value chains and sectors, including storage and distribution channels • Develop e-commerce (including infrastructure, legal framework, protection of data, security of payments) Sources: Cattaneo and others 2013, based on OECD 2012b. Note: GATS = General Agreement on Trade in Services; GCI = Global Competitiveness Index; ICT = information and communications technology; IEA = International Energy Agency; IMF = International Monetary Fund; IT = information technology; ITU = International Telecommunication Union; USITC = United States International Trade Commission; WB = World Bank; WDI = World Development Indicators; WEF = World Economic Forum; WTI = World Trade Indicators. Box 8.9. Case Study: Regulatory Reform and Infrastructure Building in Greek Logistics More than 95 percent of goods traded between Europe and Asia railway connections—requiring extensive reforms and strate- are transported over deep seas, through two primary routes. Large gic investments. To facilitate that goal, the Greek government, container ships leave ports in Asia and go to Rotterdam, the Neth- advised by the World Bank, is taking steps to remove regulatory erlands. Many go through the Suez Canal, entering the Mediter- bottlenecks and improve the country’s international connectivity. ranean, usually bypassing Greece. However, Greece’s economic Those steps include reforms in transformational sectors such as crisis has helped focus domestic policy makers’ attention on the trucking, rail, and ports; in the regulatory environment; and in potential benefits of being a regional transport hub in the way the smaller micro initiatives, such as improving the enforcement of Netherlands is in Northern Europe. regulations, promoting coordination between authorities, enhanc- But becoming a regional gateway requires competitive logis- ing transparency vis-à-vis the private sector, and better monitor- tics along the whole supply chain—beyond efficient ports and ing the performance of the sector and evaluating the impact of (Box continues next page) 174 Making Global Value Chains Work for Development BOX 8.9. (continued) reforms using modern methods. Key actions enacted since 2010 • Improving the regulatory environment. Reforms should include the following: improve the viability of Greek logistics companies, improve logistics efficiency, and encourage competition along all the • Privatizing port operations at the Port of Piraeus (Greece’s segments of the logistics value chain. Key actions include main port) of the Piraeus Port Authority and the national rail- drafting a logistics strategy and a logistics master plan, pass- way company, Trainose. Piraeus is the focal point of a logistics ing a new law on the logistics industry, and establishing a push by the government. Part operated by the China-based strong institutional framework by which the private sector has Cosco Pacific Ltd., Piraeus is the 11th largest container-ship- the power to hold the public sector accountable. ping port in the European Union, and it is the fastest growing • Adopting a trade facilitation strategy. The strategy has estab- port in the European Union (by number of containers) since lished a single window for trade facilitation and additional Cosco started operations. initiatives, such as setting up business process analysis to • Investing in infrastructure. In 2013, the government completed map export procedures, improving customs procedures, and a long-delayed 17-kilometer link from the port to the national introducing risk management methods. rail network following Cosco Pacific’s arrival. This is now attracting international investors such as Hewlett Packard to Greece. Source: Taglioni and others 2013. between the costs of inputs and outputs. If reliable infor- Notes mation on the value added of tasks is unavailable, which 1. Farole and Winkler (2014a). often is the case in services sectors, then the skill intensity 2. Farole and Winkler (2014a). Knowledge spillovers of a performed task—that is, the employee’s educational can diffuse from foreign firms to local producers within level or work experience—can serve as a good proxy for the same industry (intra-industry, or horizontal spillovers) the task’s value added (Gereffi and Fernandez-Stark 2010). or to another industry (inter-industry, or vertical spill- An example is offshore services GVCs. Call centers or rou- overs). In the latter case, spillovers can affect local input tine business process outsourcing tasks can be performed or services suppliers in upstream sectors (backward spill- by workers with a high school diploma. Market research, overs) and local customers in downstream sectors (for- however, generally requires workers with a minimum of a ward spillovers). bachelor’s degree, and the highest value-added tasks often 3. Dimelis and Louri (2002); Takii (2005). are carried out by workers with master’s degrees or PhDs. 4. Crespo and Fontoura (2007); Toth and Semjen That classification helps policy makers identify which tasks (1999). may be entered based on the skill levels of their workers 5. Farole and Winkler (2014a). (Gereffi and Fernandez-Stark 2010). 6. The value added generated by the lead firm comes 11. A separate project of the World Bank Group, Trade from preproduction activities, such as design, and postpro- and Competitiveness Global Practice, is creating a frame- duction activities, including marketing and retailing. work for such detailed analysis and identification of tasks. 7. UNCTAD (2011). 12. Porter (1980, 1985, 1990, 1998). 8. The RCA can be computed based on the domestic 13. Based on the preceding concepts, a separate World value added embodied in a country’s gross exports (see, for Bank Group project is producing a framework for system- example, the World Bank’s Export of Value Added database atically applying strategic analysis and cluster change man- and the OECD-WTO’s TiVA database in Appendixes G agement tools to identify GVC tasks in World Bank client and H). For countries for which trade in value-added data countries. are unavailable, or for customized aggregations of prod- 14. MacDonald (2006). ucts and sectors (for example, a specific cluster of activities 15. Ferrantino and Taglioni (2014). spanning different broad sectors, such as the automotive 16. Milberg and Winkler (2010). cluster or the textile cluster), RCA indexes can be con- 17. Alessandria, Kaboski, and Midrigan (2010, 2013). structed based on intermediates, parts, and components, 18. Gereffi and Frederick (2010). which can be identified using the informed classifications 19. Kolasa, Rubaszek, and Taglioni (2010). discussed in part 1 and appendixes B through E. 20. Peel (2014). 9. This step draws on ongoing research at the World 21. Lohr (2011); Escaith and Gonguet (2011). Bank by Jean Francois Arvis, Daria Taglioni, and Gianluca 22. IMF (2011); Cattaneo and others (2013). Santoni. 23. This section draws on information from Cattaneo 10. Gereffi and others (2001). For the goods sectors, the and others (2013, box 2). value added of a task can be determined as the difference 24. Gereffi, Humphrey, and Sturgeon (2005). Entering GVCs 175 25. Frederick and Gereffi (2009); Gereffi, Humphrey, 2008–09: An Inventory Adjustment?” NBER Working and Sturgeon (2005). Paper 16059, National Bureau of Economic Research, 26. Gereffi (1994). Cambridge, MA. 27. This section draws on information from Milberg ———. 2013. “Trade Wedges, Inventories, and and Winkler (2013). International Business Cycles.” Journal of Monetary 28. Gereffi (1994, 97). Economics 60 (1): 1–20. 29. This section draws on information from Milberg Antràs, Pol. 2014. “Grossman-Hart (1986) Goes Global: and Winkler (2013). Incomplete Contracts, Property Rights, and the 30. Milberg and Winkler (2013). International Organization of Production.” Journal of 31. Engman, Onodera, and Pinali (2007); Aggarwal Law Economics and Organization 30 (suppl 1): i118–i17. (2005). Antràs, Pol, and Davin Chor. 2013. “Organizing the Global 32. Engman, Onodera, and Pinali (2007, 34–35). Value Chain.” Econometrica 81 (6): 2127–204. 33. Kusago and Tzannatos (1998). Antràs, Pol, and Elhanan Helpman. 2004. “Global 34. Engman, Onodera, and Pinali (2007, 39). Sourcing.” Journal of Political Economy 112 (3): 552–80. 35. Engman, Onodera, and Pinali (2007). Antràs, P., and Stephen R. Yeaple. 2014. “Multinational 36. Farole and Akinci (2011). Firms and the Structure of International Trade.” In 37. Farole and Winkler (2014c). Handbook of International Economics, vol. 4, edited by 38. Farole and Winkler (2014b). Gita Gopinath, Elhanan Helpman, and Kenneth Rogoff, 39. Farole and Winkler (2014b). 55–130. Amsterdam: Elsevier. 40. Becattini 1990; Porter (1990). Arvis, Jean-François, Gaël Raballand, and Jean-François 41. Walmart (2014). Marteau. 2010. The Cost of Being Landlocked: Logistics 42. www.standardsmap.org. Costs and Supply Chain Reliability. Washington, DC: 43. www.directinfo.ma. World Bank. 44. Beltramello and others (2011). Becattini, G. 1990. “The Marshallian Industrial District as 45. Pietrobelli (2008). a Socio-Economic Notion.” In Industrial Districts and 46. Beltramello and others (2011). Inter-Firm Cooperation in Italy, edited by F. Pyke, G. 47. OECD (2012b). 48. WEF (2013). Becattini, and W. Sengenberger. International Institute 49. OECD-WTO (2013a). for Labor Studies, Geneva. 50. WEF (2013). Beltramello, A., K. De Backer, V. Mercader, and L. Moussiegt. 51. Hummels and others (2007). 2011. “Opening Japan: Comparisons with Other G20 52. Christ and Ferrantino (2011). Countries and Lessons Learned from International 53. Arvis, Raballand, and Marteau (2010); Christ and Experience.” OECD Science, Technology, and Industry Ferrantino (2011). Working Papers, 2011/02, Organisation for Economic 54. This section draws on Cattaneo and others (2013). Co-operation and Development Publishing, Paris. 55. OECD (2012a). http://dx.doi.org/10.1787/5kg6nk6w3v7c-en. 56. McLinden (2013). Bernard, Andrew B., J. Bradford Jensen, Stephen J. Redding, 57. Saslavsky and Shepherd (2012). and Peter K. Schott. 2010. “Intra-Firm Trade and 58. This section draws on Cattaneo and others (2013). Product Contractibility.” American Economic Review, 59. Mayneris, Poncet and Zhang (2014). Papers and Proceedings, 100 (5): 444–48. 60. Rodrik (2000). Cattaneo, Olivier, Gary Gereffi, Sébastien Miroudot, and 61. Antràs (2014). Daria Taglioni. 2013. “Joining, Upgrading and Being 62. Antràs and Yeaple (2014). Competitive in Global Value Chains: A Strategic 63. UNCTAD (2012). Framework.” Policy Research Working Paper 6406, 64. World Bank (2014). World Bank, Washington, DC. 65. OECD-WTO (2013a). Chongvilaivan, Aekapol. 2012. “Thailand’s 2011 Flooding: 66. OECD (2014). Its Impact on Direct Exports and Global Supply Chains.” Asia-Pacific Research and Training Network on Trade (ARTNeT) Working Paper Series 113, References ARTNeT, Bangkok. Acemoglu, Daron, Vasco M. Carvalho, Asuman Ozdaglar, Christ, Nannette, and Michael J. Ferrantino. 2011. “Land and Alireza Tahbaz-Saleh. 2012. “The Network Origins Transport for Exports: The Effects of Cost, Time, and of Aggregate Fluctuations.” Econometrica 80 (5). Uncertainty in Sub-Saharan Africa.” World Development Aggarwal, Aradhna. 2005. “Performance of Export 39 (10): 1749–59. Processing Zones: A Comparative Analysis of India, Christensen, Lars, and Peter Kempinsky. 2004. “Reinforcing Sri Lanka and Bangladesh.” Working paper 155, Indian the Competitiveness of Clusters.” Translated by Council for Research on International Economic Emiliano Duch. Originally published as “Att stärka klus- Relations, New Delhi. ters konkurrenskraft,” chapter 15 of Att molilisera for Alessandria, George, Joseph P. Kaboski, and Virgiliu regional tillväxt: Regionala Utvecklingsprocesser, Kluster Midrigan. 2010. “The Great Trade Collapse of och Innovationssystem. Lund, Sweden: Studentlitteratur. 176 Making Global Value Chains Work for Development Conejos, J., E. Duch, J. Fontrodona, J. M. Hernández, A. Gereffi, Gary, and Karina Fernandez-Stark. 2010. “The Luzárraga, and E. Terré. 2000. Cambio Estratégico y Offshore Services Value Chain: Developing Countries Clusters en Cataluña. Barcelona: Gestion. and the Crisis.” In Global Value Chains in a Postcrisis Crespo, Nuno, and Maria Paula Fontoura. 2007. World. A Development Perspective, edited by Olivier “Determinant Factors of FDI Spillovers—What Do We Cattaneo, Gary Gereffi, and Cornelia Staritz, 335–72. Really Know?” World Development 35 (3): 410–25. Washington, DC: World Bank. Del Prete, Davide, and Armando Rungi. 2015. “Organizing Gereffi, Gary, and Stacey Frederick. 2010. “The Global the Global Value Chain: A Firm Level Test.” IMT Lucca Apparel Value Chain, Trade, and the Crisis: Challenges EIC Working Paper Series 04, Institute for Advanced and Opportunities for Developing Countries.” In Global Studies, Lucca, Italy. Value Chains in a Postcrisis World: A Development Dimelis, Sophia, and Helen Louri. 2002. “Foreign Perspective, edited by Olivier Cattaneo, Gary Gereffi, Ownership and Production Efficiency: A Quantile and Cornelia Staritz, 157–208. Washington, DC: The Regression Analysis.” Oxford Economic Papers 54 (3): World Bank. 449–69. Gereffi, Gary, John Humphrey, Raphael Kaplinsky, Engman, Michael, Osamu Onodera, and Enrico Pinali. and Timothy J. Sturgeon. 2001. “Introduction: 2007. “Export Processing Zones: Past and Future Role in Globalisation, Value Chains and Development.” IDS Trade and Development.” OECD Trade Policy Working Bulletin 32 (3): 1–8. Paper 53, Organisation for Economic Co-operation Gereffi, Gary, John Humphrey, and Timothy J. Sturgeon. and Development, Paris. 2005. “The Governance of Global Value Chains.” Review Escaith, Hubert, and Fabien Gonguet. 2011. “International of International Political Economy 12 (1): 78–104. Supply Chains as Real Transmission Channels of Grover, Arti. 2011. “Vertical FDI versus Outsourcing: A Financial Shocks,” Journal of Financial Transformation, Welfare Comparison from the Perspective of the Host Capco Institute 31: 83–97. Country.” Review of Development Economics 15 (2): Farole, Thomas, and Gokhan Akinci. 2011. Special 293–306. Economic Zones: Progress, Emerging Challenges, and Havranek, Tomas, and Zuzana Irsova. 2011. “Estimating Future Directions. Washington, DC: World Bank. Vertical Spillovers from FDI: Why Results Vary and Farole, Thomas, and Deborah Winkler. 2014a. What the True Effect Is.” Journal of International “Introduction.” In Making Foreign Direct Investment Economics 85 (2): 234–44. Work for Sub-Saharan Africa: Local Spillovers and Hummels, David, Peter Minor, Matthew Reisman, and Competitiveness in Global Value Chains , edited Erin Endean. 2007. “Calculating Tariff Equivalents by Thomas Farole and Deborah Winkler, 5–22. for Time in Trade.” Technical Report submitted to the Washington, DC: World Bank. U.S. Agency for International Development, Nathan ———. 2014b. “Policy Implications.” In Making Foreign Associates, Arlington, VA. Direct Investment Work for Sub-Saharan Africa: Local IMF (International Monetary Fund). 2011. “Tensions Spillovers and Competitiveness in Global Value Chains, from the Two-Speed Recovery: Unemployment, edited by Thomas Farole and Deborah Winkler, Commodities, and Capital Flows.” World Economic 263–79. Washington, DC: World Bank. Outlook, IMF, Washington, DC. ———. 2014c. “The Role of Mediating Factors for FDI JETRO (Japan External Trade Organization). 2012. “Survey Spillovers in Developing Countries: Evidence from a of Business Sentiment on Japanese Corporations Global Dataset.” In Making Foreign Direct Investment in Thailand for the 2nd Half of 2011.” JETRO, Work for Sub-Saharan Africa: Local Spillovers and Tokyo. http://www.jetro.go.jp/thailand/e_survey/pdf Competitiveness in Global Value Chains , edited /jccaut11eng.pdf. by Thomas Farole and Deborah Winkler, 59–86. Kolasa, Marcin, Michal Rubaszek, and Daria Taglioni. Washington, DC: World Bank. 2010. “Firms in the Great Global Recession: The Role Ferrantino, Michael J., and Daria Taglioni. 2014. “Global of Foreign Ownership and Financial Dependence.” Value Chains in the Current Trade Slowdown.” Emerging Markets Review 11 (4): 341–57. Economic Premise No. 137, Washington, DC, World Kusago, Takayoshi, and Zafiris Tzannatos. 1998. “Export Bank. Processing Zones: A Review in Need of an Update.” Frederick, Stacey, and Gary Gereffi. 2009. “Value Chain Social Protection Discussion Paper 9802, World Bank, Governance.” USAID Briefing Paper. U.S. Agency Washington, DC. for International Development, Washington, DC. Lohr, Steve. 2011. “Stress Test for the Global Supply http://www.cggc.duke.edu/pdfs/Frederick_Gereffi Chain.” The New York Times, March 19. http:// _ValueChainGovernance_USAID_BriefingPaper www.nytimes.com/2011/03/20/business/20supply _Feb2009.pdf. .html?pagewanted=all&_r=0. Gereffi, Gary. 1994. “The Organization of Buyer-Driven MacDonald, Andrea. 2006. “Managing a Global vs. Global Commodity Chains: How U.S. Retailers Shape Domestic Supply Chain.” World Trade Magazine, July Overseas Production Networks.” In Commodity 6. http://fyketrading.homestead.com/World_Trade Chains and Global Capitalism, edited by Gary Gereffi _Article_reprint.pdf. and Miguel Korzeniewicz, 95–122. Westport, CT: Mayneris, Florian, Sandra Poncet, and Tao Zhang. 2014. Greenwood Press. “The Cleansing Effect of Minimum Wage: Firm-Level Entering GVCs 177 and Aggregate Effects of the 2004 Reform of Minimum ———. 1985. Competitive Advantage: Creating and Wage Rules in China.” Working Paper 2014-21, CEPII, Sustaining Superior Performance. New York: Free Press. Paris. ———. 1990. “The Competitive Advantage of Nations.” McLinden, Gerard. 2013. “Single-Window Systems: What Harvard Business Review 78 (2): 73–93. We Have Learned.” The Trade Post (blog), World Bank ———. 1998. On Competition. Boston: Harvard Business Group, April 30. http://blogs.worldbank.org/trade School. /single-window-systems-what-we-have-learned. Rodrik, Dani. 2000. “How Far Will Economic Integration Michaels, Daniel. 2012. “Morocco’s Aviation Industry Go?” Journal of Economic Perspectives 14 (1): 177–86. Takes Off.” Wall Street Journal March 20. http://online Santoni, Gianluca, and Daria Taglioni. Forthcoming. .wsj.com/news/articles/SB100014240529702040598045 “Benchmark Input-Output Data.” Bureau of Economic 77226763868263758. Analysis, U.S. Department of Commerce, Washington, Milberg, William, and Deborah Winkler. 2010. “Trade, DC. Crisis, and Recovery: Restructuring of Global Value Saslavsky, Daniel, and Ben Shepherd. 2012. “Facilitating Chains.” In Global Value Chains in a Postcrisis World: International Production Networks: The Role of Trade A Development Perspective, edited by Olivier Cattaneo, Logistics.” Policy Research Working Paper 6224, World Gary Gereffi, and Cornelia Staritz, 23–72. Washington, Bank, Washington, DC. DC: World Bank. Taglioni, Daria, Baher El-Hifnawi, Jean-Francois Arvis, ———. 2013. Outsourcing Economics: Global Value Chains and Lauri Ojala. 2013. Greek Logistics: Unlocking in Capitalist Development. New York: Cambridge Growth Potential through Regulatory Reform and University Press. Complementary Measures. Washington, DC: World Nathan Associates. 2004. “Best Practices in Export Bank. Promotion.” Technical report submitted to USAID, Takii, Sadayuki. 2005. “Productivity Spillovers and Nathan Associates, Arlington, VA. Characteristics of Foreign Multinational Plants in National Board of Trade. 2010. “At Your Service. Indonesian Manufacturing 1990–1995.” Journal of The Importance of Services for Manufacturing Development Economics 76: 521–42. Companies and Possible Trade Policy Implications.” Toth, Istvan Janos, and Andras Semjen. 1999. “Market Links Kommerskollegium 2010:2. and Growth Capacity of Enterprises in a Transforming Nolan, Peter, Dylan Sutherland, and Jin Zhang. 2002. Economy: The Case of Hungary.” In Market Links, Tax “The Challenge of the Global Business Revolution.” Environment and Financial Discipline of Hungarian Contributions to Political Economy 21 (1): 91–110. Enterprises, edited by Istvan Janos Toth and Andras OECD (Organisation for Economic Co-operation and Semjen. Budapest: Institute of Economics, Hungarian Development). 2012a. “Managing Aid to Achieve Trade Academy of Sciences. and Development Results: An Analysis of Trade-related UNCTAD (United Nations Conference on Trade and Targets.” COM/DCD/TAD (2012) 12/FINAL. OECD, Development). 2011. WIR11. World Investment Paris. ———. 2012b. “Reducing the Thickness of Borders to Report 2011: Non-Equity Modes of International Promote Trade and Participation to Global Value Production and Development. New York and Geneva: Chains: An Issues Paper.” COM/DCD/TAD/RD (2012) United Nations] 2/RD4. OECD, Paris. ———. 2012. Corporate Social Responsibility in Global OECD-WTO (Organisation for Economic Co-operation Value Chains: Evaluation and Monitoring Challenges and Development and World Trade Organization). for Small and Medium Sized Suppliers in Developing 2013a. “Aid for Trade at a Glance 2013: Linking to Value Countries. Geneva: UNCTAD. Chains.” OECD, Paris. USAID (United States Agency for International ———. 2013. Perspectives on Global Development 2013: Development). 2006. “Implementation of Best Prac- Industrial Policies in a Changing World. Paris: OECD. tices for Value Chain Development Projects.” Draft for ———. OECD. 2014. “Global Value Chains and Africa’s discussion. USAID, Washington, DC. Industrialisation.” African Economic Outlook 2014. Walmart. 2014. Standards for Suppliers Manual, available OECD, Paris. online at: http://cdn.corporate.walmart.com/d1/7e Peel, Michael. 2014. “Thailand Political Turmoil Imperils /ee6f5c8942f69ad4183bc0683771/standards-for Foreign and Domestic Investment.” Financial Times, -suppliers-manual.pdf March 9. http://www.ft.com/cms/s/0/6a2c75f4-a5e7- WEF (World Economic Forum). 2013. Enabling Trade: 11e3-9818-00144feab7de.html#axzz48Z7q1AoR. Valuing Growth Opportunities. Geneva, Switzerland: Pietrobelli, Carlo. 2008. “Global Value Chains in the WEF. Least Developed Countries of the World: Threats and World Bank. 2014. “Investment Policy Product.” Investment Opportunities for Local Producers.” International Climate, Trade and Competitiveness Global Practice . Journal of Technological Learning, Innovation, and Washington, DC: World Bank. Development 1 (4): 459–81. Ye, Linghe, and Masato Abe. 2012. “The Impacts of Natural Porter, Michael E. 1980. Competitive Strategy: Techniques Disasters on Global Supply Chains.” Asia-Pacific for Analyzing Industries and Competitors. New York: Research and Training Network on Trade (ARTNeT) Free Press. Working Paper 115, ARTNeT, Bangkok. Focus area Objectives Strategic questions Policy options Promoting economic Which transmission channels? Strengthening GVC-local economy upgrading and Which type of economic upgrading? links on the buyer’s and seller’s densification Which type of densification? sides Expanding and Which foreign firm and country strengthening characteristics influence GVC participation spillovers? Creating a world-class workforce Strengthening – Developing skills Which domestic firm characteristics domestic firms’ help internalize spillovers? absorptive capacity Strengthening absorptive capacity – Maximizing the absorption potential of local actors to benefit from GVC spillovers – Fostering innovation and building capacity – Complying with process and product standards – Bundling tasks Chapter 9 EXPANDING AND STRENGTHENING GVC PARTICIPATION Introduction of that effort is that strong links with the domestic economy result in greater diffusion of knowledge, This chapter focuses on strengthening a coun- technology, and know-how from foreign investors or try’s participation in global value chains (GVCs) trade partners abroad. Unfortunately, foreign inves- and upgrading to higher value-added activities. tors and trade partners do not actively pursue—and Policy makers must focus on strengthening existing sometimes resist—such integration for reasons rang- GVC–local economy links, as well as the absorptive ing from economic constraints to technological and capacity of local actors, to help them maximize the quality gaps with domestic suppliers, and to short- benefits from GVC spillovers. Policy makers need to ages of specialized workers and skills. decide which type of economic upgrading (product, For policy makers, economic upgrading and den- functional, or inter-sector) they want to pursue. By sification are keys to turning GVC participation into locating various stages of production in countries sustainable development. The concept of economic where production costs are lower, firms decrease the upgrading is largely about gaining competitiveness marginal cost of production but raise other costs by in higher value-added products, tasks, and sectors. increasing the complexity and uncertainty associ- GVC densification involves engaging more local ated with organizing production across several loca- actors (firms and workers) in the GVC network. In tions. Changes in this “trade-off ” affect outsourcing some cases, this could mean that performing lower and offshoring decisions and can be heavily influ- value-added activities on a large scale can generate enced by national policy choices. In recent years, large value addition for the country. Raising domestic some evidence of “back-shored” activities caused by labor productivity and skills contributes to the over- rising costs, intellectual property rights concerns, all goal to increase a country’s value added as a result and digitalization of the economy, and changing of GVC participation. Although static labor produc- perceptions about the stability and reliability of tivity effects are sometimes viewed to be negative for GVCs has started to emerge, pointing to the criti- employment creation (if more value added is created cal role of domestic policies to retain GVC-related with the same amount of workers), GVC integra- investment and ensure that it leads to positive spill- tion has strong potential for productivity gains via overs for local actors. several transmission channels (dynamic productivity effects), as discussed in this chapter, which go hand- Promoting Economic Upgrading and in-hand with the increased labor demand caused by more vertical specialization and higher value added Densification in GVCs: Strategic Questions in GVCs. How can countries complete the ecosystem of firms This chapter concentrates on two options to beyond the initial GVC enclave and ensure that GVCs expand development beyond the initial enclave: are integrated into the domestic economy? The logic (1) promoting economic upgrading and densification 179 180 Making Global Value Chains Work for Development in GVCs and (2) strengthening domestic firms’ in the supplier’s industry) as well as the availability absorptive capacity to benefit from spillovers in and quality effects (GVC participation increases the GVCs (see figure O.1 in the Overview). availability and quality of inputs). Policy can help move a country’s resources into Second, GVC participation can translate into pro- higher value-added activities. Value added is defined competitive market restructuring effects that extend as the sum of wage income, profit income, and tax to nonparticipants through the pro-competition revenue. All factors that influence these three ele- effect. GVC participation increases competition for ments can be considered determinants of value limited resources in the country—between multi- added, including the ability to produce goods at a national corporations (MNCs) and local firms, and higher level of quality and sophistication, as well as between participants and nonparticipants in GVCs— access to skills, knowledge, innovation, and technol- raising overall average productivity in the medium ogy. But before discussing such policy options, the run.2 GVC participation also increases competition chapter focuses on four basic strategic questions fac- through the demonstration effect of GVC prod- ing low- and middle-income countries (LMICs). ucts, business models, marketing strategies, produc- tion processes, or export processes. Knowledge and technology spillovers arise from direct imitation or Which Transmission Channels? reverse engineering by local firms, whether or not To target policy efforts efficiently, countries should they are GVC participants. identify the main transmission channels for economic Third, minimum scale achievements have a twin and social upgrading (see figure 1.11 in chapter 1): impact. In the amplification effect, they amplify pro-competition effects, stimulating investment in • Forward links are sales of GVC-linked intermedi- infrastructure and backbone services, which would ates to the local economy, spurring production not be realized without the scale of activity gener- and/or productivity in downstream sectors. ated by GVCs. The infrastructure, once in place, is • Backward links are GVC-linked purchases of local likely to spur local production and/or productivity inputs, spurring production and/or productivity in other sectors and in the non-GVC economy. With in various upstream sectors. the sustainability effect, minimum scale achievements • Technology spillovers are improved productivity of strengthen the country’s ability to sustain GVC par- local firms in the same or related downstream or ticipation over time. The GVC literature is rife with upstream sectors as a result of GVC production. examples of the key role of improvements in back- • Skills demand and upgrading are similar to tech- bone infrastructure and services, such as logistics, nology spillovers but transferred through the to improve timeliness and reliability in transporting training of and demand for skilled labor. goods, parts, and components, which enables coun- • Minimum scale achievements occur, for example, tries to integrate vertically into GVCs.3 when GVC participation stimulates investments Fourth, GVCs benefit labor markets through the in infrastructure that would otherwise not be following: profitable and that may spur local production in other sectors. • Demand effect. GVC participation is characterized by higher demand for skilled labor from MNCs or These transmission channels enable GVCs to other GVC participants. Multinationals may tem- support development and industrialization in four porarily bid away human capital by paying higher ways.1 First, GVCs—through forward and backward wages or offering enhanced employment benefits, links—generate a demand effect (lead firms tend to but that effect tends to dim as soon as the pro- require more or better inputs from local suppliers) ductivity of domestic firms is raised or the market and an assistance effect (lead firms can assist local adjusts to the tightening labor supply. suppliers through, for example, sharing knowledge • Training effect. Local firms in GVCs are more likely and technology, and advance payments) in the host to receive training (for example, from MNCs or country. The forward and backward links generate their international buyers). technology spillovers, which improve the produc- • Labor turnover effect. Knowledge embodied in the tivity of local firms through the diffusion effect (the workforce of participating firms (such as MNCs assistance effect diffuses knowledge and technology or their local suppliers) moves to other local firms. Expanding and Strengthening GVC Participation 181 Which Type of Economic Upgrading? economic upgrading, it should involve tasks with a higher unit value (value added per unit of out- Economic upgrading does not necessarily mean to put). For example, knowledge acquired in the “move up the value chain”—in other words, to per- television GVC may be used in the monitor/com- form or integrate downstream activities—but offers puter GVC. Taiwan, China, has been successful in a wider range of possibilities. Depending on the type such inter-sector upgrading.7 Another possible of economic upgrading that a country pursues, three example could be the move from sewing products objectives can be defined: in Nicaragua’s apparel and footwear industries to sewing covers for car seats (figure 9.1). • Moving into more sophisticated products in the existing value chain (product upgrading).4 Product How can countries upgrade inter-sectorally? sophistication can be measured in increased unit Once countries have singled out the tasks in which values (value added per unit of output). they have a comparative advantage (as described in • Increasing value-added shares (in output of final chapter 8), they need to identify sectors that require product) in existing GVC tasks (functional upgrad- similar tasks but add more value. The following mea- ing). Functional upgrading is defined as the move sures can be used: into more technologically sophisticated or more integrated tasks of a production process and • Labor’s share in value added can be used to get a relates to the overall skill content of activities.5 first indication of a sector’s labor intensity. It is usually measured as a higher share of value • The skill intensity of sectors can be calculated if added in the output of the final product. sector data by type of labor input are available • Moving into new value chains with higher value- (say, by using firm-level data). added shares (inter-sector upgrading). Firms can • Technology intensity is a more sophisticated mea- pursue inter-sector upgrading, moving horizon- sure to identify similar sectors. The classification tally into new value chains that require similar by Lall (2000) has high-, medium-, and low-tech; knowledge and skills.6 To qualify such a move as resource-based; and primary sectors/products. Figure 9.1. Example of Possible Inter-Sector Upgrading in Nicaragua Car seats Seat Foam and Cover Final frames parts sewing assembly Fabric and Apparela Cut Sew Finish leather Fabric and Footweara Soles Uppers Assembly leather Source: Adapted from Sturgeon and Zylberberg 2012. a. Industry value chains currently active in Nicaragua. 182 Making Global Value Chains Work for Development The measures are most meaningful if they are upgrading—contributes to the overall goal to undertaken at a highly disaggregated sector level. increase a country’s value added that results from To detect similar tasks with higher value added in GVC participation. other industries, analysts should also use qualita- The concept of economic upgrading is based on tive information from technical sources, companies, the premise that existing local GVC participants and other field experts. That information can help become more competitive, which enables them to identify which sectors are similar in their processes advance into new products, tasks, and sectors with and required tasks, and which type of inter-sector higher value added. Densification, by contrast, aims upgrading has been successful in other countries. at facilitating the participation of more local firms Ideally, that information should be backed up by and workers in already existing GVC-related prod- evidence of past success (with firms that have moved ucts, tasks, and sectors in a country. Value addition into higher value-added products in other sectors). in the latter case happens through scale effects, as Other measures of economic upgrading include densification creates higher profit income, wage profit growth, export growth, growth in export mar- income, and tax revenue. In some cases, this could ket shares (especially if associated with unit value/ mean that performing lower value-added activities quality growth), reduced relative incidence of unit on a large scale can generate large value addition for labor costs, and increase in capital intensity.8 the country. The three objectives can be achieved by upgrad- The key is to enable local firms and workers to ing the production factors of labor and capital, and participate in GVCs that are already present in the also by increasing total factor productivity (see figure country by strengthening domestic firms’ absorptive 1.9 in chapter 1). Policy options and indicators are capacity and developing worker skills. Policy makers discussed in more detail in this chapter. The policy should assess which of those is the priority for a options should aim for the following: country. • Improving the skills and know-how of the workforce Which Foreign Firm and Country Characteristics (skills upgrading). Developing skills is a key ele- Influence Spillovers? ment of competitiveness, and it affects the abil- ity to participate in GVCs and achieve economic A major determinant of country policies to attract and social upgrading within GVCs. This can hap- foreign direct investment (FDI) and non-equity pen by increasing the skill content of a country’s modes of investment (NEMs) is the potential to workforce. deliver substantial spillovers of knowledge—and • Improving the absorptive capacity and technology ultimately productivity—to local firms and workers. of firms (capital upgrading). Strengthening the A vast set of empirical evidence has been amassed absorptive capacity of local firms requires gen- over the past decade on the presence and direction of eral and industry-specific investments to upgrade FDI-generated horizontal and vertical productivity technical capacity and achieve quality standards spillovers. Overall, the results are mixed and suggest and innovation capabilities. that the theoretical postulated spillover effects do not • Increasing productivity in existing GVC tasks (pro- automatically materialize with FDI, which indicates cess upgrading). Domestic firms performing GVC the need for still more research.10 tasks can pursue process upgrading by better Three groups of mediating factors determine organizing their production or introducing new potential spillovers to domestic firm productivity: technologies to capture efficiency gains.9 In other (1) the spillover potential of the foreign firm, (2) the words, process upgrading refers to total factor absorptive capacity of firms in the host economy, and productivity growth in existing activities in the (3) the national characteristics and institutions of the value chain that cannot be directly attributed to host country. GVC entry via FDI attraction is mainly the production factors of labor or capital. determined by the first and third factors, which are discussed following the conceptual framework devel- oped in Farole, Staritz, and Winkler (2014). The sec- Which Type of Densification? ond group is discussed in the next section. Densification is about engaging more local Given the increasing importance of GVCs and actors (firms and workers) in the GVC net- export platform FDI, understanding how spillover work. Densification—in addition to economic potential differs is likely to become an important Expanding and Strengthening GVC Participation 183 policy priority, particularly for small and low-income clothing, footwear, electronics, and automotive sec- countries that rely on that type of investment. tors.16 Moreover, the share of intermediates sourced locally by multinationals is likely to increase with the Which Foreign Firm Characteristics Influence distance between the host and the source economy. Spillover Potential? 11 The share is also likely to be larger for multinationals The degree of foreign ownership affects local firms’ originating in countries outside the country’s prefer- potential to absorb FDI spillovers. A higher share of ential trade agreement, because such a trade agree- foreign ownership—thus, larger control over man- ment makes imports from the home country less agement and lower potential for knowledge leak- attractive.17 ages—correlates positively with the parent firm’s Spillovers also depend on the technology intensity incentive to transfer knowledge, such as in the form of the multinational’s goods produced in the host of technology.12 A larger domestic ownership share, country.18 Products characterized by greater technol- however, could also be beneficial for local firms, ogy—or products that are more intensive in research because the foreign investor’s interests are less well and development (R&D)—generally contain a protected, which makes technology leakages more greater element of knowledge and a broader set of likely. Larger domestic participation might further skills. However, the production of high-tech prod- increase the likelihood of relying on domestic sup- ucts might also involve low-tech processes, which pliers.13 Empirical studies controlling for different could offset that effect.19 structures of foreign ownership tend to support the Related to technology intensity is the foreign more positive spillover effects of joint ventures.14 investor’s home country, which may have an effect Explanations include the possibility of more verti- on the production strategy and technologies in host cal links, as well as stronger technology leakages for countries, but also may have other effects on spillover partly owned foreign firms.15 potential. The home country more generally influ- Various motivations for FDI and NEMs are likely ences managerial practices and cultures on the use to mediate spillover potential. The conventional of expatriate workers, attitudes toward training local wisdom is that resource-seeking investment has less workers, and skills development. Further, end-market potential for spillovers because of its capital and tech- segmentation—closely linked to foreign investors’ nology intensity and limited time horizons. By con- home countries through historical, cultural, and lan- trast, manufacturing investment often is considered guage ties, as well as trade policies—is common. All to have higher spillover potential because it is driven these patterns affect spillover potential.20 The foreign largely by efficiency-seeking motives. Indeed, manu- investor’s home country also positively influences facturing investment that is more labor-intensive, domestic firms’ absorptive capacity, because workers has greater requirements for a broad range of goods in domestic firms observe and imitate technologies, and services inputs, and has lower barriers to domes- management practices, and cultural values.21 And a tic links (relative to resource-seeking investment) is a foreign affiliate’s distance to its parent firm affects its strong candidate for contributing spillovers. Market- spillover potential, particularly for efficiency-seeking seeking investment, particularly in retail, is also con- investment. Several studies find that foreign investors sidered to provide higher spillover potential, because are more likely to purchase local inputs from domes- retailers tend to source from local producers—in tic suppliers if the home country is further away.22 particular, food and other perishables. The evidence A multinational firm’s entry mode may influence remains ambiguous, however, which suggests context the pace or extent of investment-induced benefits specificity. for local firms. For example, a greenfield invest- Analogously, a multinational firm’s sourcing strat- ment is more likely to be accompanied by technol- egy may affect spillover potential. If an MNC sources ogy, whereas with mergers and acquisitions (M&As), on a global scale, it may follow a co-sourcing strat- the multinational firm is more likely to adopt egy, which increases its reliance on imported inputs the host country’s technology and only gradu- from established suppliers abroad. An MNC might ally improve it.23 Whereas greenfield investments follow co-location strategies that require an estab- self-evidently increase investment, capacity, and lished foreign input supplier to enter the host coun- employment, M&As and other types of brownfield try. Both circumstances could make the entrance of investments may not do so, because the new foreign new local suppliers more difficult, which is a par- owners may rationalize and even reduce capacity and ticularly common situation for multinationals in the employment. 184 Making Global Value Chains Work for Development The pace and irregularity of foreign entry can and locating in countries with few serious competi- also affect spillovers, by constraining multination- tors.31 Policies that mandate technology transfer to als from building stable relationships with local local firms may increase the transmission of knowl- suppliers, which results in multinationals being less edge and technology between the affiliate and the likely to rely on domestic inputs. Further, local firms local market, but such policies may result in the for- might not have enough time to observe and imitate eign investor limiting the level and nature of knowl- good practices and for local workers to acquire skills, edge transfers to the affiliate. which results in negative competition effects.24 Financial markets in LMICs may also be a factor Finally, the length of foreign presence may affect in the absorption of spillovers.32 Multinationals can spillovers. Foreign firms with a longer presence in the have an ambiguous impact on access to finance for country may have a more positive effect on produc- local firms: multinationals may ease such access by tivity spillovers, largely resulting from longer supplier bringing in scarce capital to LMICs; but if MNCs relationships, although that effect may taper off.25 borrow locally, they may increase local firms’ financ- ing constraints.33 That, in turn, can influence a Which Host Country Characteristics and local firm’s absorptive capacity, and well-developed Institutions Influence Spillovers? 26 markets may facilitate a domestic firm’s absorptive Host country and institutional factors can influence capacity links.34 foreign and domestic firm characteristics, as well A country’s trade policy shapes the amount and as the transmission channels for knowledge diffu- type of foreign investment. Spillovers are larger in sion from multinationals to local firms. Although countries that are more open to trade. A country’s the focus here is on spillovers from FDI, many host trade policy regime is related to its capacity to attract country characteristics can also be expected to lead foreign firms, because foreign investors are less con- to spillovers from GVC participation through NEMs strained by the size and efficiency of the local mar- of investment. ket.35 Moreover, foreign investors in an open trade Labor market regulations may influence the effect setting are more integrated globally and thus tend of foreign investment on domestic firms through to adopt the newest technologies. However, it can various channels. Higher absolute and relative labor also be argued that foreign investors in an outward- market flexibility than in the foreign investor’s home oriented trade policy regime tend to focus more on country seems to have a positive effect on the chances international distribution and marketing and less on of securing initial foreign investment.27 Labor mar- new technologies.36 ket regulations in general, and wage constraints in Trade policy also affects domestic firms. In an particular, can affect the skills in a firm, and hence open trade regime, domestic firms are more exposed their absorptive capacity.28 Overly rigid labor mar- to international competitive pressures, which will kets can reduce the likelihood of labor turnover and prepare them to absorb spillovers.37 Moreover, a FDI spillovers.29 Conversely, overly flexible labor country’s trade policy also affects the likelihood of markets may generate frequent labor turnover, which domestic firms’ becoming exporters and learning reduces the time for domestic workers to acquire by exporting. Although the effect of exporting on skills and knowledge from foreign firms. domestic firms’ absorptive capacities is ambiguous, The strength of intellectual property rights in a exporting clearly moderates the direction and extent host country may help attract high-quality foreign of FDI spillovers. FDI spillovers are larger in coun- investment initially and, therefore, create the poten- tries that are more open to trade.38 For example, for tial for FDI spillovers. But some people argue that China, horizontal and vertical spillover effects from although strong intellectual property rights may FDI are negative when final goods and input tariffs help attract such investment and allow knowledge are higher.39 and technology to be transferred to the affiliate, they Investment policy and promotion mediate spill- may also hinder the transmission of those advances overs by helping to attract foreign investment in to the local market.30 Multinational firms use several general (the focus of most export promotion efforts) instruments—in addition to ensuring strong prop- and by encouraging policies to promote spillovers erty rights—to protect technology spillovers to local (much less common). Investment promotion con- competitors in the same sector, such as paying higher tributes to bringing in firms that should have higher wages to avoid labor turnover, ensuring trade secrecy, spillover potential, given their quality and technology Expanding and Strengthening GVC Participation 185 position.40 For example, positive FDI spillovers in studies find that FDI spillovers increase with aver- Chinese manufacturing are higher from foreign age education and innovation) and is particularly firms enjoying investment subsidies and exemptions important for expanding GVC participation.45 from value-added taxes relative to spillovers from foreign firms that do not reap those benefits.41 Strengthening Absorptive Capacity: Special economic zones (SEZs) may affect Which Domestic Firm Characteristics Help spillovers. Local Chinese manufacturing firms in Internalize Spillovers? 46 SEZs have smaller productivity spillovers from FDI than do non-SEZ domestic firms.42 That may At the domestic firm level, R&D, human capital, occur because most SEZs focus on export process- firm size, firm location, export behavior, technology ing combined with a high percentage of imported gap, type of ownership, and sector competition are inputs, which limits the potential for FDI spillovers mediating factors that allow countries to adopt com- because demand for local suppliers is constrained. plementary policies for leveraging the opportunities Moreover, the spatial and legal structures that gov- of GVC participation. These factors determine the ern SEZs often inhibit their integration with the local firm’s absorptive capacity. Although the focus local economy. here is on spillovers from FDI, many firm character- Industrial policies, particularly programs to sup- istics can also be expected to lead to spillovers from port the development of local small and medium GVC participation through international trade and enterprises, can mediate FDI spillovers, especially NEMs, especially in modular or relational gover- where the technology and productivity gaps between nance forms in which the degree of knowledge shar- foreign and local firms are large, or where few local ing is relatively high (see the section “Which Form firms exist. Collaboration with foreign firms and of Governance between Lead Firms and Suppliers?” support to develop local supplier networks through in chapter 8). supplier development programs run by foreign affili- The technology gap between foreign and domes- ates but supported by governments have done much tic firms has been identified as one the most impor- to facilitate spillovers in, for example, the automotive tant mediating factors for FDI spillovers.47 A large and electronics sectors. Local content provisions that gap can be beneficial for local firms because their require a certain share of inputs to be sourced locally catching-up potential increases,48 but local firms have also gained prominence, as in China, but the might not be able to absorb positive FDI spillovers if track record of those provisions is mixed, and they the gap is too big or too small.49 Some studies recon- depend on domestic absorption capacity and sup- cile the two views and find a nonlinear relationship plier development. between a domestic firm’s technology gap and FDI- Weak institutions—including corruption, red induced productivity benefits.50 tape, and intellectual property rights—are linked to The supportive role of R&D in local firms is solid protection for local firms, network-driven business in high-income countries, such as Ireland, Spain,51 practices, and inefficient markets, which possibly Sweden, and the United States.52 It is also strong in constrains foreign investors from fully exploiting LMICs or emerging countries, including the Czech their competitive advantages. That drawback may Republic, Hungary, India, Indonesia, and the Slovak influence the types of FDI and NEM that are initially Republic, and a large cross-section of 78 LMICs.53 attracted, as well as domestic firms’ absorptive capac- A domestic firm’s ability to absorb foreign tech- ity. The empirical evidence is mixed. Firm-level data nology can be positively related to its share of skilled for 17 emerging countries during 2002–05 reveal no labor,54 but that benefit may apply only to smaller evidence that the extent of FDI spillovers is affected firms.55 In that case, FDI does not affect large domes- by the degree of corruption or red tape.43 The evi- tic firms with a high proportion of human capital, dence also shows that a country’s transparency has a because those firms are probably the most similar U-shaped effect on FDI spillovers: countries with a to multinationals in technology and market share. medium level of transparency benefit the least from No evidence exists for the positive effect of skilled FDI, whereas countries with low and high levels workers.56 In contrast, the competition effect might show stronger FDI spillovers.44 enable larger domestic firms to keep skilled workers The local innovation and learning infrastructure more readily, compared with smaller firms, which influences the share of human capital in firms (most may lead to negative spillovers for the latter. Smaller 186 Making Global Value Chains Work for Development firms have fewer means to attract skilled workers by versus state-owned firms, which can be studied best paying higher wages or offering additional benefits.57 in the context of China or the transition economies Firm size has been positively related to a domes- in Central and Eastern Europe. Private firms may be tic firm’s capacity to absorb FDI spillovers.58 Larger more likely to benefit from FDI spillovers because of firms may (1) be better positioned to compete with their willingness to restructure and imitate (demon- multinationals and imitate their tools;59 (2) pay stration effect), and because of their larger export better wages and therefore find attracting workers orientation, which enables those firms to access employed by multinational firms easier; and (3) be knowledge internationally. 66 By contrast, state- more visible, perhaps organized in associations, and owned enterprises typically are larger, are technically thus more likely to be selected as local suppliers by competitive, and may have easier access to finance, foreign firms. which increases their absorptive capacity; but they Several aspects of domestic firm location are tend to be less market oriented, which may lower important in FDI productivity spillovers. Foreign their absorptive capacity.67 firms co-locating (agglomeration) in the same sector Finally, the level of competition influences the and region, for example, can significantly increase extent of FDI spillovers. Competitive pressures from the productivity and employment of local firms.60 multinational firms might be lower if the local firm However, firm location in SEZs can have a nega- already faces stiff competition at the sector level. As tive impact on FDI spillovers if the zone focuses on with exports, local firms in competitive sectors may export processing and has a high share of imported have less incentive to improve, which results in lower inputs. More regional development and a domestic benefits from FDI spillovers. Still, local firms may be firm’s geographic proximity to multinational firms better equipped to benefit from positive demonstra- seem to have a positive effect.61 tion effects.68 Exporting has been linked to a domestic firm’s absorptive capacity for at least two reasons. First, Policy Options local exporting firms generally are characterized by higher productivity—whether through learning by Expanding GVC participation requires three sets exporting or self-selecting into exporting—which of policies: (1) to strengthen existing links in GVCs, makes them more competitive against negative (2) to strengthen a country’s absorptive capacity to rivalry effects created by multinationals.62 Second, benefit from intensified GVC integration, and (3) to the more a local firm exports, the less the competitive create a world-class workforce (see figure O.1 in the pressures from multinational firms are felt (assum- Overview). Although some policies in the third set ing that the multinational firm does not enter the aim to strengthen a country’s absorptive capacity—for same export market); hence, there is an incentive example, by promoting skills development—a broad to improve, which lowers the extent of positive FDI range of policies target other aspects of upgrading. spillovers. However, empirical studies show no clear evidence of whether exporting increases or lowers Strengthening GVC–Local Economy Links on the the productivity gains from FDI.63 Buyer’s and Seller’s Sides69 Spillovers can also depend on the sectors in which domestic firms operate.64 FDI-enhanced productiv- Strong links with the domestic economy—through ity spillovers in food processing, for example, seem forward and backward links, technology spillovers, to be driven by efficiency improvements, whereas skills demand and upgrading, minimum-scale technological progress seems to be the main driver in achievements (see figure 1.11 and the subsection electrical machinery. FDI spillovers may be smaller “Which Transmission Channels?” in this chapter), for domestic firms in services sectors because of and other forms of collaboration and interaction— the lower absorptive capacity of firms in those sec- should offer greater benefits of GVC participation tors. A foreign presence in technology-intensive or at home. The development of links can focus on the high-tech industries tends to lead to larger positive breadth of links (variety of local inputs) and on their spillovers compared with foreign presence in labor- depth (degree of local value added), so making a dis- intensive or low-tech industries.65 tinction is key.70 Policies that promote links between Type of ownership is another factor. Some stud- GVCs and the local economy primarily target foreign ies have focused on the difference between private investors, but can also include other international Expanding and Strengthening GVC Participation 187 buyers outside the country. The policies include the as establishing skill requirements. The framework following: should also include addressing gaps in domestic contract enforcement and other barriers to for- • Ensure that the incentives used to attract foreign mal contracting with local suppliers. investors do not create a bias against local integra- • Establish incentives for foreign investors and other tion. The most important issue is to ensure that international buyers to work with local universities, foreign-owned companies do not have privileged research institutes, and training institutes. Such access to instruments such as import tax and duty incentives include research funds, matching grant concessions or duty drawbacks. Similarly, reserv- programs, and fiscal incentives for R&D in the host ing EPZs for foreign-owned companies can create country, as well as internships, outplacements, and barriers to supply by domestic firms. joint training and curriculum development. • Leverage investment and other incentives to pro- mote actions that support spillovers. If generat- Although these policy options target international ing spillovers is among the principal rationales firms—particularly foreign investors—in GVCs, for offering incentives to foreign investors or policies for developing links should emphasize other international buyers, if those incentives are (1) the absorptive capacity of domestic firms to ben- not predicated on spillover outcomes (which are efit from GVC participation and (2) the develop- difficult to measure), the incentives should be ment or improvement of worker skills. The next two predicated at least on foreign investors or other subsections address these areas. international buyers engaging in activities to sup- port spillovers. Strengthening Absorptive Capacity • Ensure that local content regulations operate under the right conditions and are clearly defined. Maximizing the Absorption Potential of Local (For example, what is “local” and what is “con- Actors to Benefit from GVC Spillovers 71 tent”?) The focus should be on value addition Attracting foreign investors and other international rather than in-country ownership. Regulations buyers and linking them to the domestic economy can be effective, but only when the domestic sup- should create the conditions for local firms and ply side is up to the task of being a competitive workers to benefit from spillovers of knowledge and supplier. Otherwise such regulations are likely to technology. The degree to which they ultimately ben- weaken the competitiveness of investors, thereby efit, however, depends on the absorptive capacity undermining the overall outcomes. In any case, of the domestic actors. This is the area of spillover setting strict local content targets can be coun- policy in which government has the most important terproductive and difficult to enforce. Instead of role, particularly by building the absorptive capac- establishing rigid local content requirements, the ity of firms and workers and by helping local firms aim should be the collaborative development of and workers access opportunities. For example, the flexible localization plans, in which investors come Czech Republic has policies to help create a competi- up with their own proposals for delivering spill- tive local supplier network, as described in box 9.1. overs to the local economy. That approach allows The policies should include the following: for sufficient flexibility across sectors and firms. • Have a clear and comprehensive framework to • Support supply-side capacity building, taking support the upgrading of domestic firms. This step into account the heterogeneity of domestic firms. is important to facilitate supplier development The potential of domestic firms to supply foreign programs initiated by foreign investors or other investors and other international buyers and to international buyers. Traditional linkage pro- upgrade in higher value-added activities varies grams merely scratch the surface; they are likely to enormously across domestic firms. Supplying be effective only in the context of a more compre- foreign investors and other international buyers hensive set of policies on links. A comprehensive should be an activity for the most productive, framework should include bridging information high-potential domestic firms. Government pro- gaps by facilitating exchanges of information on grams focused on upgrading technical capacity foreign investors’ and other international buy- should focus primarily on those firms, setting out ers’ needs and local supplier capabilities, as well clear requirements for firm participation. 188 Making Global Value Chains Work for Development Box 9.1. The Czech Republic’s Supplier Development Program After the country’s emergence from communism and entry into foreign manufacturers to shortlist and contact potential the European Union, CzechInvest (the investment promotion suppliers. agency—CI) learned from surveying investors that multinationals • Matchmaking by identifying the components and services considered the local supplier network a key determinant in their foreign investors were considering subcontracting (Meet the investment decisions, second only to labor availability. Yet multi- Buyer), arranging seminars and exhibitions with Czech sup- national investors imported 90 to 95 percent of their components pliers and foreign affiliates, and taking proposals to potential to meet production requirements. foreign investors. CI’s top management saw an opportunity: to address inves- • Upgrading selected Czech suppliers. CI managers selected tors’ demand for inputs and willingness to source locally by suppliers according to predefined criteria in high-technology strengthening the capabilities of Czech suppliers. From CI’s per- industries, and then produced an upgrading plan. In an elec- spective, creating a robust, competitive Czech supplier base for tronics pilot, CI identified 45 companies as potential can- key prominent sectors was a way to embed foreign direct invest- didates, trained them, and after seven months reevaluated ment (FDI) into the economy and channel its benefits, thereby them; CI subsequently offered tailored assistance to the 20 helping to retain and attract investors while supporting domestic most promising firms. suppliers. CI launched the Pilot Supplier Development Program (also called the Twinning Program) in electronics, the country’s An evaluation of the pilot 18 months after it ended in July 2002 fastest-growing and second largest FDI sector after automotives. showed that 15 suppliers had landed new, renewable contracts, The program’s orientation was demand driven and practical. worth more than US$46 million for 2000–03. Based on those Its overall objective was to equip suppliers with the informa- results, CI rolled out Twinning II, which extended the program to tion and skills to meet investor requirements and win more (and aeronautics, automotives, pharmaceuticals, and engineering. higher) value-added contracts. The program had three elements: • Collecting and distributing information on the products and capabilities of potential Czech component suppliers to enable Source: Potter 2001. • Build the absorptive capacity of local firms. This in the long run by improving the sophistication requires general and industry-specific invest- and competitiveness of local firms. ments to upgrade technical capacity and, most important, achieve quality standards. Because Fostering Innovation and Building Capacity 72 licensing of technology from foreign investors GVCs ease capacity constraints because a country and other international buyers is a significant does not need to develop a fully integrated industry source of spillovers, governments should provide to participate in GVCs. Still, capacities and produc- incentives for it. The biggest gap in support, how- tivity (as much as cost) are important drivers for for- ever, is likely to be outside the technical arena—in eign investors and lead firms that search for global basic business and financial management. Flexible offshore locations. Given the significance of flows delivery and financing models are necessary to in the new trade paradigm (as opposed to stocks), a allow for sector-specific approaches and collabo- location’s responsiveness, capacity to innovate, and ration with foreign investors. adaptability to the lead firm’s requests are also key • Narrow the technical and managerial skills gap factors.73 with foreign investors and other international With the shift in demand to emerging markets, buyers. This includes actively engaging universi- lead firms have to define strategies in which innova- ties and research institutes to embed spillovers. tion “centers” are in fact decentralized. According to • Adopt open policies to promote imports and the concept of reverse innovation, lead firms need to skilled immigration. This step may be critical to innovate in LMICs—often in clusters—and even- promote localization in the long term. A policy of tually bring the results back home.74 That requires openness—not only for access to imported goods the host country to develop innovation capabili- and services, but, more controversially, for access ties, based on education and skills, often involv- to (imported) skilled workers—is likely to pay off ing public-private partnerships for R&D (box 9.2), Expanding and Strengthening GVC Participation 189 increasing the supply of qualified researchers at local universities, and aligning higher education curricula Box 9.2. Case Study: Renault-Dacia Regional Design and and training with the local economy’s needs. Development Activities in Romania Economic upgrading is often about “creating the knowledge behind the product,” but a country In 2007, Renault-Dacia moved part of its regional design and devel- opment activities to Renault Technologie Roumanie (RTR) in Roma- might not be able to upgrade because of barriers nia, the largest Renault engineering center outside France, with in other stages of production, such as services. The some 2,500 engineers. diversification into services tasks and the promotion RTR mainly accommodates engineering functions, along with pur- of services exports offer largely untapped potential chasing, design, and support. With three locations in Romania, RTR for many LMICs, but also require them to be well brings together all the activities needed in an automotive project. prepared. For example, moving out of production The relocation of the design and development activities was and into R&D, engineering, or marketing services driven by Dacia’s entry-level car and the idea that designing cars in an emerging market would help the company respond better to new requires flexibility in trading those services, includ- consumer markets in Eastern Europe and Asia. The center now over- ing the temporary movement of service providers. It sees the development of all entry-level vehicles (about 35 percent of may also require establishing and enforcing intellec- all Renault vehicles worldwide). tual property rights. Table 9.1 summarizes possible policy objectives Source: Based on interviews with private sector stakeholders. to foster innovation and capacity building, as well as available performance indicators. GVCs—so much so that “failure to comply with these Complying with Process and Product Standards standards can result in exclusion from the GVC.”75 Although respect for standards might vary depend- According to a recent business survey in the agri- ing on the maturity of the GVC’s lead firm and the food sector of 250 lead firms and suppliers in LMICs, final market, it is a key element for the functioning of about 60 percent of the firms named the ability to Table 9.1. Fostering Innovation and Building Capacity: Policy Objectives and Performance Indicators Policy objectives Bolstering productivity, production, and innovation capacities, including human capital and other resources: • Adopt innovation policies and incentives (for example, R&D, innovation centers) and adapt/diffuse technologies in trade-oriented sectors. • Provide education and training to match domestic skills with international standards and demand in trade-oriented sectors; upgrade skills. • Develop production capacities in trade-oriented sectors, both hard (storage, conditioning, cooling chains, and so forth) and soft (value chain management, for instance). • Create clusters and other task-bundling efforts. • Change production (methods and equipment) to more efficient and sustainable use of natural resources and energy. Performance indicators • Computer, communications, and other services; ICT goods and services imports/exports (WDI) • Investment in telecoms with private participation (WDI) • Firms offering formal training (WDI) • Number of patent applications filed by residents and nonresidents, domestically and abroad (WDI, WIPO) • Education statistics: secondary and tertiary education, specialties, male/female, and so forth (UNESCO, ILO, WDI) • Global competitiveness index–business sophistication (WEF GCI 11.01–11.09) • Innovation (WEF GCI 12.01–12.07) • Extent of staff training (WEF GCI 5.08) • Labor statistics—activity rates, unemployment, male/female, and so forth (ILO, WDI) • Innovation indicators and surveys—public and private R&D expenditure, high and medium-high technology manufacturing, knowledge intensive services (OECD) • Production capacities—sector output—and productivity statistics (national statistics, WIOD) Sources: Cattaneo and others 2013, based on OECD 2012. Note: GCI = Global Competitiveness Index; ICT = information and communications technology; ILO = International Labour Organization; OECD = Organisation for Economic Co-operation and Development; R&D = research and development; UNESCO = United Nations Educational, Scientific, and Cultural Organization; WDI = World Development Indicators; WEF = World Economic Forum; WIOD = World Input-Output Database; WIPO = World Intellectual Property Organization. 190 Making Global Value Chains Work for Development meet quality and safety standards as the main fac- input providers and assemblers along the chain.77 tor influencing sourcing and investment decisions Despite the role of private standards in GVCs, pub- in GVCs.76 Similarly, 40 percent of the firms pointed lic standards, public infrastructure for certification to noncompliance with mandatory import require- and accreditation, and the enforcement by public ments as a typical trade problem with LMIC suppli- authorities of health, safety, and environment rules ers. About 37 percent suggested that improving the are essential to attract GVC production segments. standards infrastructure and certification capacity Inadequate public standards can raise the cost of would be the most effective way to integrate new local production or create unnecessary obstacles to suppliers in LMICs into GVCs; almost 50 percent trade—or both (figure 9.2). of the firms providing trade-related technical and Excessively low or badly enforced local standards capacity building focused on compliance with safety minimize the backward links and positive spillovers and quality standards. of FDI and offshore production in a country: inputs Standards relate to processes (such as labor, will have to be imported to meet the lead firm’s stan- social, and environmental standards, often in cor- dards, and the local tasks will be confined to basic porate social responsibility or code of conduct) and transformation and manufacturing. Analysis of products (such as quality). The standards must be the retail sector suggests three phases: a first phase respected along the entire value chain, because every in which no local products meet the retailer’s stan- stage of production could affect the quality of the dards and most products are imported; a second final product or service. In agrifoods, for example, in which local producers adjust to the standards of such standards translate into traceability require- the retailer (often with its help), and local products ments aimed at protecting consumer health and replace imported ones; and a third in which the best increasing product information for consumers. local products that meet international standards are Standards in GVCs are public and private, with exported and distributed by the retailer abroad. an increasing prevalence of “voluntary” standards Conversely, excessively high local standards are imposed by lead firms (buyers or producers) on all equally disturbing and could constitute unnecessary Figure 9.2. Standards in Agrifood GVCs Concentrated food retail (supermarkets and fast-food chains) Bilateral oligopolies Buyer-driven chains Private/most Public and private/ comprehensive safety-focused standards standards Concentrated food production Fragmented food production (farmers and manufacturers) (farmers and manufacturers) Producer-driven Traditional markets chains Limited public Public and private/ standards/least quality-focused comprehensive standards standards Fragmented food retail (supermarkets and fast-food chains) Source: Adapted from Lee, Gereffi, and Beauvais 2012. Note: GVC = global value chain. Expanding and Strengthening GVC Participation 191 Figure 9.3. Diffusion of Standards and Other Codes of Table 9.2. Improving Standards: Policy Objectives and Conduct in GVCs Performance Indicators First-tier or not specified Policy objectives Technical and sanitary and phytosanitary standards: Subcontractors • Build capacity for certification and accreditation (labs, personnel, resources, and so forth). Second-tier or beyond • Adopt or reform domestic norms and standards to comply with international best practices. Joint ventures • Promote standards—including voluntary standards—and Licensees related training. • Ensure private sector support to comply with standards. 0 20 40 60 80 100 Performance indicators Percent • Diffusion of voluntary standards and ISO certification Source: Adapted from UNCTAD 2012. ownership (WDI, national statistics) Note: GVC = global value chain; UNCTAD = United Nations Conference on Trade • Adoption of international standards and Development. • International accreditation of domestic accreditation/certifi- cation agencies obstacles to trade or disguised protectionism. Several Sources: Cattaneo and others 2013, based on OECD 2012. Note: More specific policy objectives and measures of labor and social standards questions have been raised, for example, over eco- are presented in table 10.1 in chapter 10. ISO = International Organization for labeling and border adjustment taxes (so-called car- Standardization; WDI = World Development Indicators. bon taxes).78 Where local standards and certification and possibilities (see the subsection “Which Type of accreditation meet international standards and best Economic Upgrading?” in this chapter). Task bun- practice, the costs of value chain management are dling is necessary for consolidating GVCs, in which sharply reduced, which increases a country’s attrac- lead firms reduce the number of intermediates and tiveness to FDI. GVCs therefore make a strong case expect their suppliers to provide a more comprehen- for regulatory convergence, harmonization, mutual sive package with larger services content. Task bun- recognition, and diffusion of international standards. dling might also be necessary for potential offshore Imposing respect for standards is very difficult and locations to attract the production of some tasks that costly for lead firms on their own, although many do cannot be performed independently.80 For example, so (figure 9.3): some transparency mechanisms, such some tasks that can be easily offshored may be bun- as mapping pollution at the micro level in China, dled with tasks that cannot, making it possible to off- help to enforce green supply chains by providing an shore the first set of tasks only to countries that can independent monitoring mechanism to lead firms’ also perform the second set. subcontracting production in China, as well as to civil society (IPE.org.cn). Figure 9.4. Tasks Performed by Apparel Industries in Considering the risks associated with the preva- Torreon, Mexico lence of private standards in GVCs—particularly Retail for smallholders and producers in LMICs, as well as Marketing consumers—the case is strong for multi-stakeholder dialogue and cooperation in defining and enforcing Distribution standards (table 9.2).79 Upgrading Laundry and finishing Bundling Tasks Assembly The trend toward GVC consolidation suggests that a country cannot offer a single task but must Cut offer a bundle of tasks. Economic upgrading tra- Trims and labels jectories often reflect performing new tasks that build on existing ones (figure 9.4), which this book Textiles defines as functional upgrading. Economic upgrad- ing does not always mean to perform or integrate 1993 1996 2000 downstream activities, but offers a wider range of Source: Adapted from Bair and Gereffi 2001. 192 Making Global Value Chains Work for Development Creating a World-Class Workforce81 • A new and evolving set of workforce skills is needed to participate in GVCs. Skill development is a key element of competitive- • Required skills and workforce development ness, participation in GVCs, and economic and social needs vary substantially by stage within industry- upgrading within GVCs. A positive and statistically specific upgrading trajectories. significant correlation exists between human capi- • Workers need soft skills (in addition to hard skills, tal and services exports, for instance.82 Economic which are more easily quantifiable and directly upgrading requires new skills and knowledge, either linked to the job) in today’s world of work. by increasing the skill content of a country’s activities • In LMICs, managerial skills for GVCs are in short (and thus workforce) or by developing competen- supply. cies in niche market segments.83 In other words, eco- • Upgrading in GVCs requires more and better pro- nomic upgrading and social upgrading are mutually fessionals and technicians in bottleneck positions. dependent. Skill shortages can impede a country’s upgrading. On stakeholders and institutions: In Chile, Costa Rica, Ethiopia, and Rwanda, upgrad- ing strategies in GVCs have been most successful • Local education systems currently do not provide when accompanied by complementary workforce the range of skills required by GVCs. development interventions.84 In Rwanda, economic • Technical training institutions and universities upgrading into the high-quality specialty coffee should coordinate more closely with industry segment required skill development for managing stakeholders. workers, plantations, financial risks, and other areas. • New actors—such as individual firms, industry For workforce development to be successful, it must associations, nongovernmental organizations, be part of a coherent overall upgrading strategy that and special government programs—can provide involves key stakeholders. In addition, workforce many of the skills required by GVCs. development must be customized to the specific job • Private sector intermediaries can facilitate requirements.85 upgrading and skills development. GVCs contribute to skill development through • Public-private partnerships have emerged as an lead-firm transfers. Lead firms indeed have strong efficient and effective method for skill development. incentives to train their workforce to comply with their standards. Beyond private initiatives, a strong On global standards: case exists for public investment in skill development to meet the needs of international trade and partici- • Global standards define the upgrading require- pation in GVCs. Table 9.1 lists performance indica- ments for the local workforce. tors related to skill development. • Multi-stakeholder partnerships in LMICs coalesce A look at the link between economic upgrading in response to global standards. and skill development in four GVCs (apparel, tour- • National certification of skills can be a powerful ism, offshore services, and fruits and vegetables) in tool for GVC labor markets in LMICs. some 20 LMICs reached the following conclusions.86 On workforce skills: The successful upgrading in GVCs through devel- • Appropriate worker skills are essential to eco- oping a more skilled workforce is illustrated by the nomic upgrading. apparel industry in Turkey (box 9.3). • The focus of skill development must reflect local needs and those of the global economy. Expanding and Strengthening GVC Participation 193 Box 9.3. Own Design and Branding in Turkey Turkish firms moved into the design segment of the value chain The Turkish government supported upgrading into own brand- as part of a broader strategy to establish the country as a fashion ing, the next stage after own design, by granting incentives for firms center. Industry associations and government agencies collabo- willing to upgrade into branding. The incentives included reimburse- rated to promote Istanbul, targeting it to become a top-five global ments up to 60 percent of the cost for a maximum of three years of fashion center by 2023. personnel expenses, machinery, equipment, software, consultancy, Tight relationships of local manufacturers with large global and research and development material. Leading local firms with retailers, such as the United Kingdom–based Marks & Spencer, their own brands and retail outlets abroad include Sarar, Mithat, facilitated upgrading into design services. In 2007, Denizli was and Bilsar. Erak clothing, which was originally a full-package sup- designing 10 percent of Marks & Spencer’s garments made in plier with international brands such as Guess, Esprit, and Calvin Turkey. Upgrading into own-design manufacturing required a Klein, is now selling its own brand, Mavi Jeans, in 4,600 specialty specialized workforce, which was built with government support. stores in 28 countries. Developing own branding has required addi- Organizations such as the Istanbul Textile and Apparel Exporter tional efforts to foster workforce development, from bodies such Association (IKTIB) worked with the private sector and govern- as IKTIB and KOSGEB (Small and Medium Enterprises Development ment agencies to establish fashion design vocational training Organization), a quasi-governmental organization affiliated with the schools. Istanbul Fashion Academy, established by the European Ministry of Industry and Trade. Union and IKTIB, trains students. Source: Fernandez-Stark, Frederick, and Gereffi 2011. Notes 28. Hale and Long (2011). 29. Gorodnichenko, Svejnar, and Terrell (2007); Javorcik 1. The discussion on mechanisms triggered by GVC (2004a). participation partially evolves from the taxonomy intro- 30. Havranek and Irsova (2011). duced by Farole, Staritz, and Winkler (2014). 31. Javorcik (2004b). 2. In the short run, average productivity may decrease 32. Alfaro and others (2010). and local firms may lose market shares as a result of inten- 33. Harrison, Love, and McMillan (2004). sified competition. 34. Aggarwal, Milner, and Riaño (2011). 3. WEF (2013). 35. Crespo and Fontoura (2007). 4. Humphrey (2004); Humphrey and Schmitz (2002). 36. Meyer and Sinani (2009). 5. Humphrey (2004); Humphrey and Schmitz (2002). 37. Havranek and Irsova (2011). 6. Humphrey (2004); Humphrey and Schmitz (2002). 38. Havranek and Irsova (2011). 7. Humphrey and Schmitz (2002). 39. Du, Harrison, and Jefferson (2011). 8. Milberg and Winkler (2011, 349). 40. Harding and Javorcik (2012). 9. Humphrey (2004); Humphrey and Schmitz (2002). 41. Du, Harrison, and Jefferson (2011). 10. Havranek and Irsova (2011). 42. Abraham, Konings, and Slootmaekers (2010). 11. This section draws on Farole, Staritz, and Winkler 43. Gorodnichenko, Svejnar, and Terrell (2007). (2014). 44. Meyer and Sinani (2009). 12. Takii (2005). 45. Farole and Winkler (2014a). 13. Crespo and Fontoura (2007). 46. This section draws on Farole, Staritz, and Winkler 14. Abraham, Konings, and Slootmaekers (2010). (2014). 15. Javorcik and Spatareanu (2008). 47. Kokko, Tansini, and Zejan (1996); Grünfeld (2006). 16. Paus and Gallagher (2008). The technology gap usually is measured as a domestic 17. Javorcik and Spatareanu (2011). firm’s productivity level relative to a benchmark produc- 18. Lin, Liu, and Zhang (2009). tivity level within the same sector, often of the lead firm’s 19. Paus and Gallagher (2008). or foreign firms. 20. Staritz and Morris (2013). 48. Smeets (2008); Jordaan (2011). 21. Zhang and others (2010). 49. Winkler (2014); Blalock and Gertler (2009). 22. See, for instance, Javorcik and Spatareanu (2011). 50. Girma and Görg (2007). 23. Crespo and Fontoura (2007). 51. Barrios and others (2004). 24. Wang and others (2012). 52. Barrios and others (2004); Karpaty and Lundberg 25. Gorodnichenko, Svejnar, and Terrell (2007). (2004); Keller and Yeaple (2009). 26. This section draws on Farole, Staritz, and Winkler 53. Kinoshita (2001); Kanthuria (2000, 2001, 2002); (2014). Damijan and others (2003); Blalock and Gertler (2009); 27. Javorcik and Spatareanu (2005). Farole and Winkler (2014a). 194 Making Global Value Chains Work for Development 54. Blalock and Gertler (2009). Aggarwal, Natasha, Chris Milner, and Alejandro Riaño. 55. Girma and Wakelin (2007). 2011. “Credit Constraints and FDI Spillovers in 56. Farole and Winkler (2014a). China.” China and the World Economy Research Paper 57. Sinani and Meyer (2004); Winkler (2014). Series No. 2011/21, University of Nottingham, United 58. See Jordaan (2011) for Mexico and Farole and Kingdom. Winkler (2014) for the 78 LMICs. Alfaro, Laura, Areendam Chanda, Sebnem Kalemli- 59. Crespo and Fontoura (2007). Ozcan, and Selin Sayek. 2010. “Does Foreign Direct 60. Barrios, Bertinelli, and Strobl (2006); Farole and Investment Promote Growth? Exploring the Role of Winkler (2014a). Financial Markets on Linkages.” Journal of Development 61. Girma and Wakelin (2007); Winkler (2014). Economics 91 (2): 242–56. 62. Crespo and Fontoura (2007). Bair, Jennifer, and Gary Gereffi. 2001. “Local Clusters in 63. Blomström and Sjöholm (1999) for Indonesia, Global Value Chains: The Causes and Consequences Ponomareva (2000) for Russia, Sinani and Meyer (2004) of Export Dynamism in Torreon’s Blue Jeans Industry.” for Estonia, and Abraham, Konings, and Slootmaekers World Development 29 (11): 1885–903. (2010) and Du, Harrison, and Jefferson (2011) for China Barrios, Salvador, Luisito Bertinelli, and Eric Strobl. 2006. confirm that the potential for positive productivity spill- “Coagglomeration and Spillovers.” Regional Science and overs is less pronounced for exporters compared with non- Urban Economics 36 (4): 467–81. exporters or firms exporting little. By contrast, Jordaan Barrios, Salvador, Sophia Dimelis, Helen Louri, and Eric (2011) for Mexico, Barrios and Strobl (2002) for Spain, Strobl. 2004. “Efficiency Spillovers from Foreign Direct Schoors and van der Tol (2002) for Hungary, Lin, Liu, and Investment in the EU Periphery: A Comparative Zhang (2009) for China, and Farole and Winkler (2014a) Study of Greece, Ireland, and Spain.” Review of World for a large sample of LMICs find positive spillovers from Economics (Weltwirtschaftliches Archiv) 140 (4): exporting or operating in more open sectors. 688–705. 64. Temenggung (2007); Suyanto and Salim (2010). Barrios, S., and E. Strobl. 2002. “Foreign Direct Investment 65. Buckley, Wang, and Clegg (2007); Keller and Yeaple and Productivity Spillovers: Evidence from the Spanish (2009). Experience.” Review of World Economics 138 (3): 66. Sinani and Meyer (2004). 459–81. 67. Du, Harrison, and Jefferson (2011). Blalock, G., and P. Gertler. 2009. “How Firm Capabilities 68. Barrios and Strobl (2002); Farole and Winkler Affect Who Benefits from Foreign Technology.” Journal (2014). of Development Economics 90 (2): 192–99. 69. This section draws on Farole and Winkler (2014b). Blomström, M., and F. Sjöholm. 1999. “Technology 70. Morris, Kaplinsky, and Kaplan (2011). Transfer and Spillovers: Does Local Participation with 71. This section draws on Farole and Winkler (2014b). Multinationals Matter?” European Economic Review 43 72. The following three sections draw on Cattaneo and (4–6): 915–23. others (2013). Brenton, Paul, Gareth Edwards-Jones, and Michael Friis 73. World Bank (2010). Jensen. 2009. “Carbon Labeling and Low-income 74. Govindarajan and Trimble (2012). Country Exports: A Review of the Development Issues.” 75. Kaplinsky, Terheggen, and Tijaja (2010); Gereffi, Development Policy Review 27 (3): 243–67. Fernandez-Stark, and Psilos (2011, 243). Buckley, Peter J., Chengqi Wang, and Jeremy Clegg. 2007. 76. OECD-WTO (2013). “The Impact of Foreign Ownership, Local Ownership 77. Lee, Gereffi, and Beauvais (2012). and Industry Characteristics on Spillover Benefits from 78. Brenton, Edwards-Jones, and Jensen (2009). Foreign Direct Investment in China.” International 79. Lee, Gereffi, and Beauvais (2012); Cadot, Malouche, Business Review 16 (2): 142–58. and Saez (2012). Cadot, Olivier, Mariem Malouche, and Sebastian Saez. 80. Gereffi and Frederick (2010); Lanz, Miroudot, and 2012. Streamlining Non-Tariff Measures: A Toolkit for Nordås (2013). Policy Makers. Washington, DC: World Bank. 81. This section largely draws on Cattaneo and others Cattaneo, Olivier, Gary Gereffi, Sébastien Miroudot, and (2013). Daria Taglioni. 2013. “Joining, Upgrading and Being 82. Saez and Goswami (2010). Competitive in Global Value Chains: A Strategic 83. Humphrey and Schmitz (2002). Framework.” Policy Research Working Paper 6406, 84. World Bank (2014a). World Bank, Washington, DC. 85. World Bank (2014b). Crespo, N., and M. P. Fontoura. 2007. “Determinant 86. Gereffi, Fernandez-Stark, and Psilos (2011). Factors of FDI Spillovers—What Do We Really Know?” World Development 35 (3): 410–25. Damijan, J., M. Knell, B. Majcen, and M. Rojec. 2003. References “Technology Transfer through FDI in Top-10 Abraham, Filip, Jozef Konings, and Veerle Slootmaekers. Transition Countries: How Important Are Direct 2010. “FDI Spillovers in the Chinese Manufacturing Effects, Horizontal and Vertical Spillovers?” Working Sector Evidence of Firm Heterogeneity.” Economies of Paper 549, William Davidson Institute, University of Transition 18: 143–82. Michigan, Ann Arbor, MI. Expanding and Strengthening GVC Participation 195 Du, L., A. Harrison, and G. Jefferson. 2011. “Do Institutions Grünfeld, Leo A. 2006. “Multinational Production, Matter For FDI Spillovers? The Implications of China’s Absorptive Capacity, and Endogenous R&D Spillovers.” ‘Special Characteristics.’” NBER Working Paper 16767, Review of International Economics 14 (5): 922–40. National Bureau of Economic Research, Cambridge, Hale, G., and C. Long. 2011. “Did Foreign Direct MA. Investment Put an Upward Pressure on Wages in Farole, Thomas, C. Staritz, and D. Winkler. 2014. China?” IMF Economic Review 59: 404–30. “Conceptual Framework.” In Making Foreign Direct Harding, Torfinn, and Beata S. Javorcik. 2012. “Foreign Investment Work for Sub-Saharan Africa: Local Spillovers Direct Investment and Export Upgrading.” Review of and Competitiveness in Global Value Chains, edited by Economics and Statistics 94 (4): 964–80. T. Farole and D. Winkler, 23–55. Washington, DC: Harrison, Ann E., Inessa Love, and Margaret S. McMillan. World Bank. 2004. “Global Capital Flows and Financing Constraints.” Farole, T., and D. Winkler. 2014a. “The Role of Mediating Journal of Development Economics 75 (1): 269–301. Factors for FDI Spillovers in Developing Countries: Havranek, Tomas, and Zuzana Irsova. 2011. “Estimating Evidence from a Global Dataset.” In Making Foreign Vertical Spillovers from FDI: Why Results Vary and Direct Investment Work for Sub-Saharan Africa: Local What the True Effect Is.” Journal of International Spillovers and Competitiveness in Global Value Chains, Economics 85 (2): 234–44. edited by Thomas Farole and Deborah Winkler, 59–86. Humphrey, John. 2004. “Upgrading in Global Value Washington, DC: World Bank. Chains.” Working Paper 28, Policy Integration ———. 2014b. “Policy Implications.” In Making Foreign Department, World Commission on the Social Direct Investment Work for Sub-Saharan Africa: Local Dimension of Globalization, International Labour Spillovers and Competitiveness in Global Value Chains, Organization, Geneva. edited by Thomas Farole and Deborah Winkler, Humphrey, J., and H. Schmitz. 2002. “How Does Insertion 263–79. Washington, DC: World Bank. in Global Value Chains Affect Upgrading in Industrial Clusters?” Regional Studies 36 (9): 1017–27. Fernandez-Stark, Karina, Stacey Frederick, and Gary Javorcik, Beata Smarzynska. 2004a. “The Composition Gereffi. 2011. The Apparel Global Value Chain: Economic of Foreign Direct Investment and Protection of Upgrading and Workforce Development. Durham, NC: Intellectual Property Rights: Evidence from Transition Duke University. http://www.cggc.duke.edu/pdfs/2011 Economies.” European Economic Review 48 (1): 39–62. -11-11_CGGC_Apparel-Global-Value-Chain.pdf. ———. 2004b. “Does Foreign Direct Investment Increase Gereffi, Gary, Karina Fernandez-Stark, and Phil Psilos. 2011. the Productivity of Domestic Firms? In Search of “Skills for Upgrading: Workforce Development and Spillovers through Backward Linkages.” American Global Value Chains in Developing Countries.” Center Economic Review 94 (3): 605–27. on Globalization, Governance & Competitiveness, Javorcik, Beata Smarzynska, and Mariana Spatareanu. Duke University; and RTI International. http://www 2005. “Do Foreign Investors Care about Labor Market .cggc.duke.edu/pdfs/Skills-for-Upgrading-Workforce Regulations?” Review of World Economics 141 (3): -Development-and-GVC-in-Developing-Countries 375–403. _FullBook.pdf. ———. 2008. “To Share or Not to Share: Does Local Gereffi, Gary, and Stacey Frederick. 2010. “The Global Participation Matter for Spillovers from Foreign Direct Apparel Value Chain, Trade, and the Crisis: Challenges Investment?” Journal of Development Economics 85: and Opportunities for Developing Countries.” In Global 194–217. Value Chains in a Postcrisis World: A Development ———. 2011. “Does It Matter Where You Come From? Perspective, edited by Olivier Cattaneo, Gary Gereffi, Vertical Spillovers from Foreign Direct Investment and Cornelia Staritz, 157–208. Washington, DC: World and the Origin of Investors.” Journal of Development Bank. Economics 96 (1): 126–38. Girma, Sourafel, and Holger Görg. 2007. “The Role of the Jordaan, Jacob A. 2011. “Local Sourcing and Technology Efficiency Gap for Spillovers from FDI: Evidence from Spillovers to Mexican Suppliers: How Important are the UK Electronics and Engineering Sectors.” Open FDI and Supplier Characteristics?” Growth and Change Economies Review 18 (2): 215–32. 42 (3): 287–319. Girma, Sourafel, and Katharine Wakelin. 2007. “Local Kanthuria, Vinish. 2000. “Productivity Spillovers from Productivity Spillovers from Foreign Direct Investment Technology Transfer to Indian Manufacturing Firms.” in the U.K. Electronics Industry.” Regional Science and Journal of International Development 12 (3): 343–69. Urban Economics 37 (3): 399–412. ———. 2001. “Foreign Firms, Technology Transfer and Gorodnichenko, Yuriy, Jan Svejnar, and Katherine Terrell. Knowledge Spillovers to Indian Manufacturing Firms: 2007. “When Does FDI Have Positive Spillovers? A Stochastic Frontier Analysis.” Applied Economics 33 Evidence from 17 Emerging Market Economies.” (5): 625–42. Discussion Paper 3079, Institute for the Study of Labor, ———. 2002. “Liberalisation, FDI, and Productivity Bonn, Germany. Spillovers—An Analysis of Indian Manufacturing Govindarajan, Vijay, and Chris Trimble. 2012. Reverse Firms.” Oxford Economic Papers 54 (4): 688–718. Innovation: Create Far from Home, Win Everywhere. Kaplinsky, Raphael, Anne Terheggen, and Julia Tijaja. Cambridge, MA: Harvard Business Review Press. 2010. “What Happens When the Market Shifts to 196 Making Global Value Chains Work for Development China? The Gabon Timber and Thai Cassava Value 2013. “Aid for Trade at a Glance 2013: Linking to Value Chains.” In Global Value Chains in a Postcrisis World: Chains.” OECD, Paris. A Development Perspective, edited by Olivier Cattaneo, Paus, E. A., and K. P. Gallagher. 2008. “Missing Links: Gary Gereffi, and Cornelia Staritz, 303–34. Washington, Foreign Investment and Industrial Development DC: World Bank. in Costa Rica and Mexico.” Studies of Comparative Karpaty, Patrik, and Lars Lundberg. 2004. “Foreign Direct International Development 43 (1): 53–80. Investment and Productivity Spillovers in Swedish Ponomareva, N. 2000. “Are There Positive or Negative Manufacturing.” Working Paper 194, Trade Union Spillovers from Foreign-Owned to Commodities and Institute for Economic Research, Stockholm. Linkages: Meeting the Policy Challenge.” Discussion Keller, Wolfgang, and Stephen R. Yeaple. 2009. Paper 14, Making the Most of Commodities “Multinational Enterprises, International Trade, and Programme, The Open University, United Kingdom Productivity Growth: Firm-Level Evidence from the Domestic Firms.” Working Paper BSP/00/042, New United States.” Review of Economics and Statistics 91 (4): Economics School, Moscow. 821–31. Potter, Jonathan. 2001. Embedding Foreign Direct Kinoshita, Y. 2001. “R&D and Technology Spillovers Investment . Paris: Organisation for Economic through FDI: Innovation and Absorptive Capacity.” Co-operation and Development. http://www.icpr3.org CEPR Discussion Paper 2775, Centre for Economic /en/foreign-direct-investment/czech-republic/. Policy Research, London. Saez, Sebastian, and Arti Grover Goswami. 2010. Kokko, Ari, Rúben Tansini, and Mario Carlos Zejan. 1996. “Uncovering Developing Countries’ Performance in “Local Technological Capability and Productivity Trade.” Economic Premise 39, World Bank, Washington, Spillovers from FDI in the Uruguayan Manufacturing DC. Sector.” Journal of Development Studies 32 (4): 602–11. Schoors, K., and B. van der Tol. 2002. “Foreign Direct Lall, Sanjaya. 2000. “The Technological Structure and Investment Spillovers within and between Sectors: Performance of Developing Country Manufactured Evidence from Hungarian Data.” Working Paper Exports, 1985–1998.” Working Paper 44, Queen 2002/157, University of Gent, Gent, Belgium. Elizabeth House, University of Oxford, United Sinani, Evis, and Klaus E. Meyer. 2004. “Spillovers of Kingdom. Technology Transfer from FDI: The Case of Estonia.” Lanz, Rainer, Sebastien Miroudot, and Hildegunn Nordås. Journal of Comparative Economics 32 (3): 445–66. 2013. “Offshoring of Tasks: Taylorism versus Toyotism.” Smeets, Roger. 2008. “Collecting the Pieces of the FDI World Economy 36 (2): 194–212. Knowledge Spillovers Puzzle.” The World Bank Research Lee, Joonkoo, Gary Gereffi, and Janet Beauvais. 2012. Observer 23 (2): 107–38. “Global Value Chains and Agrifood Standards: Staritz, Cornelia, and Mike Morris. 2013. “Local Challenges and Possibilities for Smallholders in Embeddedness and Economic and Social Upgrading in Developing Countries.” Proceedings of the National Madagascar’s Export Apparel Industry.” ÖFSE Working Academy of Sciences of the United States of America 109 Paper 38, Austrian Foundation for Development (31): 12326–31. Research, Vienna. Lin, Ping, Zhuomin Liu, and Yifan Zhang. 2009. “Do Sturgeon, Timothy, and E. Zylberberg. 2012. “Global Value Chinese Domestic Firms Benefit from FDI Inflow? Chains, Special Economic Zones and Domestic Links: Evidence of Horizontal and Vertical Spillovers.” China The Extension of Automotive Global Value Chains into Economic Review 20 (4): 677–91. Nicaragua and Macedonia.” Presentation to the World Meyer, Klaus E., and Evis Sinani. 2009. “When and Where Bank’s International Trade Department, Poverty Re- Does Foreign Direct Investment Generate Positive duction and Economic Management (PREM) Network. Spillovers? A Meta-Analysis.” Journal of International World Bank, Washington, DC. Business Studies 40 (7): 1075–94. Suyanto, Suyanto, and Ruhul A. Salim. 2010. “Sources Milberg, William, and Deborah Winkler. 2011. “Economic of Productivity Gains from FDI in Indonesia: Is It and Social Upgrading in Global Production Networks: Efficiency Improvement or Technological Progress?” Problems of Theory and Measurement.” International Developing Economies 48 (4): 450–72. Labour Review 150 (3–4): 341–65. Takii, Sadayuki. 2005. “Productivity Spillovers and Morris, M., R. Kaplinsky, and D. Kaplan. 2011. Characteristics of Foreign Multinational Plants in “Commodities and Linkages: Meeting the Policy Indonesian Manufacturing 1990–1995.” Journal of Challenge.” Discussion Paper 14, Making the Most Development Economics 76: 521–42. of Commodities Programme, The Open University, Temenggung, Della. 2007. “Productivity Spillovers from United Kingdom. Foreign Direct Investment: Indonesian Manufacturing OECD (Organisation for Economic Co-operation and Industry’s Experience 1975–2000.” Dynamics, Development). 2012. Managing Aid to Achieve Trade Economic Growth, and International Trade Conference and Development Results: An Analysis of Trade-Related Paper c012_048, Kiel Institute for the World Economy, Targets . COM/DCD/TAD(2012)12/FINAL. Paris: Kiel, Germany. OECD. UNCTAD (United Nations Conference on Trade and OECD-WTO (Organisation for Economic Co-operation Development). 2012. Corporate Social Responsibility and Development and World Trade Organization). in Global Value Chains: Evaluation and Monitoring Expanding and Strengthening GVC Participation 197 Challenges for Small and Medium Sized Suppliers in World Bank. 2010. Innovation Policy: A Guide for Developing Developing Countries. Geneva: UNCTAD. Countries. Washington, DC: World Bank. Wang, Chengqi, Ziliang Deng, Mario I. Kafouros, and Yan ———. 2014a. Republic of Burundi: Skills Development for Chen. 2012. “Reconceptualizing the Spillover Effects Growth—Building Skills for Coffee and Other Priority of Foreign Direct Investment: A Process-Dependent Sectors. Washington, DC: World Bank. Approach.” International Business Review 21 (3): ———. 2014b. “Turkey: Trading up to High Income: 452–64. Country Economic Memorandum.” Report 82307, WEF (World Economic Forum). 2013. “Enabling Trade: World Bank, Washington, DC. Valuing Growth Opportunities” WEF, Geneva. Zhang, Yan, Haiyang Li, Yu Li, and Li-An Zhou. 2010. “FDI Winkler, D. 2014. “Determining the Nature and Extent of Spillovers in an Emerging Market: The Role of Foreign Spillovers: Empirical Assessment.” In Making Foreign Firms’ Country Origin Diversity and Domestic Firms’ Direct Investment Work for Sub-Saharan Africa: Local Absorptive Capacity.” Strategic Management Journal 31 Spillovers and Competitiveness in Global Value Chains, (9): 969–89. edited by Thomas Farole and Deborah Winkler, 87–114. Washington, DC: World Bank. Focus area Objectives Strategic questions Policy options Which relationship between economic and social upgrading? Creating a world-class workforce Promoting social Which type of social upgrading? – Developing skills upgrading and Is downgrading a possibility? – Promoting social upgrading Turning GVC cohesion – Engineering equitable distribution of participation into Which links between social upgrading and cohesion? opportunities and outcomes sustainable development Promoting environmental What benefits from environmental Implementing climate-smart policies sustainability regulation? and infrastructure Chapter 10 TURNING GVC PARTICIPATION INTO SUSTAINABLE DEVELOPMENT Introduction shows that economic upgrading may lead to social downgrading. That is, economic upgrading may This chapter tackles the challenge of how to turn lead to lower-value economic activities and weaken global value chain (GVC) participation into sus- workers’ employment, wages, rights, and protection, tainable development. Three areas of sustainable strongly suggesting a role for policy to counter this development are important: macroeconomic sus- possibility. tainability, social sustainability, and environmental sustainability. Not only are they important develop- ment objectives per se, they also ensure the sustain- Which Relationship between Economic and ability of a GVC-centric approach to development. Social Upgrading? This chapter focuses on social and environmental An often implicit assumption is that economic sustainability and leaves the discussion on the mac- upgrading in GVCs will automatically translate roeconomic implications of GVCs for further work. into social upgrading through greater employment Labor market–enhancing outcomes for workers at opportunities and higher wages. However, the link home and more equitable distribution of opportuni- between those elements is unclear from a theoreti- ties and outcomes create social support for a reform cal standpoint. If productivity growth is a proxy for agenda aimed at strengthening a country’s GVC par- economic upgrading and wage growth is a reason- ticipation. Climate-smart policy prescriptions and able representation of social upgrading, economic infrastructure can mitigate the challenges for firms theory can explain the relationship between the from climatic disruptions, as the firms seek to ensure two. Neoclassical theory implies that, other things the long-term predictability, reliability, and time- being equal, social upgrading will result from sensitive delivery of goods necessary to participate in economic upgrading. From an institutional perspec- GVCs. Because climatic disruption can impair firms’ tive, however, social upgrading is de-linked from ability to access inputs and deliver final products, technological change and associated with social insti- countries’ preparedness is an increasingly critical fac- tutions, including union density, bargaining rights, tor in firms’ location decisions. minimum wages, and active labor market policies.1 Empirical research also shows that economic Promoting Social Upgrading and Cohesion: upgrading can translate into social upgrading, but Strategic Questions not necessarily. Therefore, it is important to know the circumstances for economic upgrading to lead to its The issues of social upgrading are not new and social equivalent. Conversely, it is necessary to under- have been discussed in the literature under the role stand how to stanch economic and social downgrad- of multinational corporations in development. ing. If economic upgrading does not automatically However, linking economic and social upgrading lead to social upgrading, policy has a clear role.2 199 200 Making Global Value Chains Work for Development Figure 10.1. Social Cohesion as an End of and a Means for Development GVC Social Social participation upgrading cohesion What does economic upgrading through GVCs good and the working conditions. Although lead mean for living standards, including employment, firms are largely able to require the implementa- wages, working conditions, economic rights, gender tion of similar codes of conduct from their first- equality, and economic security? Improvements in tier suppliers or contract manufacturers through “the terms, conditions, and remuneration of employ- monitoring or audits, monitoring and improving ment and respect for workers’ rights, as embodied working conditions at lower-tier suppliers becomes in the concept of decent work” can be referred to as increasingly difficult. The lead firm’s ability to influ- social upgrading.3 But although substantial research ence suppliers also depends on the power rela- has been done on economic upgrading in GVCs, lit- tions in a GVC. Increased price pressures from the tle systematic research exists on what such economic lead firm create negative incentives for first- and upgrading actually means for employment and living lower-tier suppliers to cut labor and other costs by standards, despite growing interest in understand- violating international labor standards (failure to ing the social spillovers to the domestic economy pay minimum wages, requiring illegal overtime, or of countries already participating in—or thinking using forced and child labor) and other HSE stan- about joining—GVCs. dards (failure to install ventilation systems or fire Evidence and intuition suggest that the impact of safety features, as the 2013 Rana Plaza disaster in GVC participation on living standards depends on Bangladesh demonstrated).4 many factors. One factor is where a country, indus- Social upgrading is linked to a country’s social try, or firm is positioned in the value chain. The effect cohesion, which can be understood as working for countries performing assembly tasks is likely to be toward the well-being of all the members of a society different from that for countries specializing in pre- by (1) creating a sense of belonging and active partic- production stages. The gains may also differ by the ipation, (2) promoting trust, (3) offering the oppor- type of value chain, because some industries are more tunity of upward social mobility, and (4) fighting labor intensive than others (as are some product lines inequality and exclusion. Living standards—notably within the same industry). Different GVCs also may the result of jobs—are major elements linking social involve different combinations of low-skilled, labor- upgrading and cohesion. Although social cohesion intensive, and higher-skilled technology-intensive can be an end (or goal) of development outcomes, it workers. And the spillovers generated by trade flows is also a means for development, especially as greater in GVCs in a specific sector may differ across coun- social cohesion and political stability make countries tries, depending on how integrated the sector is with more attractive for investment (figure 10.1).5 the rest of the economy of each country. Multinationals and large global buyers are under Which Type of Social Upgrading? increasing pressure to comply with international labor and health, safety, and environmental (HSE) The literature divides social upgrading into two standards, which apply particularly to electronics, mutually complementary parts. Measurable stan- apparel, and food GVCs, in which final consumers dards refer to aspects of worker well-being that are perceive a more direct link between the consumer more easily observed and quantified. The most basic Turning GVC Participation into Sustainable Development 201 Figure 10.2. Social “Grading” of Jobs Knowledge intensive High skilled, technology intensive Medium skilled, mixed technologies Low skilled, labor intensive Small scale, household based Source: Adapted from Barrientos, Gereffi, and Rossi 2011. expressions are employment and wages, but oth- Which Measures of Social Upgrading? ers include physical well-being and working condi- tions, such as health and safety, working hours, and The most basic indicator of social upgrading is employment security. Enabling rights are less easily employment growth. Employment growth could also quantified; they include empowerment, nondis- be correlated with various measures of GVC inte- crimination, rights to bargaining, and freedom of gration, but chapter 7 presents instead more direct association.6 measurements of the link by drawing on various Different types of work can be given a different indicators already developed in the literature. The “grade” subjectively, with knowledge-intensive activ- expansion of global production in labor-intensive ities valued at the top and household-based, small- industries has been an important source of employ- scale activities at the bottom (figure 10.2).7 ment generation and other positive impacts through In this framework, three possible “trajectories” of strengthened formal job opportunities. Similarly, improved measurable standards are possible:8 migrant workers and women who previously had dif- ficulty accessing that type of wage work have filled • Small-scale worker upgrading. Workers in home- many of those jobs.9 based production can experience improvements GVC-enhanced employment is a necessary but in their working conditions. Establishing pro- not sufficient condition for social upgrading, because ducer organizations and providing more secure employment gains may be undermined in other areas. contracts, better payment, and upgraded personal Where employment generates better rights and protec- health and safety equipment can support this goal. tion for workers, it can enhance social upgrading. Such • Labor-intensive upgrading. Less-skilled workers employment, however, often is insecure and unpro- can move to other types of labor-intensive work tected, which presents multiple challenges in ensuring characterized by better working conditions. An decent work and wages for more vulnerable workers. example is the move of female workers in Ban- The downward pricing pressure in many GVCs has gladesh or Sri Lanka from subsistence farming to simultaneously led to negative social impacts. wage employment in apparel firms (if the latter With the increasing complexity of trade in GVCs, have buyers’ codes of labor practice). the relationship between trade and employment • Higher-skill upgrading. Workers can move to more becomes more complicated. Rather than exports skilled and better-paid jobs (for example, in infor- generating only domestic employment (as would be mation technology) if they have been trained at the case if countries were selling only intermediate their previous workplace and the firm had higher or final goods abroad), they may generate employ- labor standards. ment in other countries from importing (or buying) 202 Making Global Value Chains Work for Development as between lowering wages and raising labor pro- intermediate goods. The discussion in chapter 7 is ductivity—a low road and a high road. Although the framed on this basis. high road does not guarantee that wage growth (part Chapter 7 presents five indicators that link of social upgrading) will follow, the low road of low- employment and GVC participation in countries ering wages has limits because of considerations of and industries: (1) the labor content of exports, political stability and human subsistence. (2) the labor component of domestic value added in Pressures for upgrading and downgrading com- exports, (3) jobs sustained by foreign final demand, pete within GVCs as suppliers balance higher quality (4) jobs generated by foreign trade in GVCs, and with lower costs. Economic and social upgrading can (5) jobs in GVC manufacturing. Employment can be positively correlated with improved production be a deceptive measure of social upgrading, because when it increases workers’ productivity. For example, jobs created by GVCs can vary in quality in areas pay (an indicator of social upgrading) and produc- such as pay, work hours, conditions, and so on. tivity growth (an indicator of economic upgrading) Much broader than employment, skills, or wages, show an extremely high correlation in a 45-country the concept of social upgrading captures more gen- sample for the apparel and footwear sectors in 1995 erally the gains in living standards and working con- to 1999.14 However, pressure to reduce costs might ditions over time. Other measures include growth lead employers to combine economic upgrading in employment; growth in wages; growth in labor with social downgrading, although that challenge is share; increased formal employment; decline in youth not limited to GVCs. In many labor-intensive indus- unemployment; increased gender equality of employ- tries, the pressure to reduce costs puts significant ment and wages; poverty reduction; higher share of downward pressure on labor costs, including wages wage employment in nonagricultural employment; and working conditions. improved labor standards; improved job safety; abo- In theory, four combinations of outcomes are lition of child labor, forced labor, and employment possible (figure 10.3). Economic upgrading may be discrimination; regulation of monitoring; improved combined with social upgrading or downgrading. If political rights; improved human development indi- labor productivity growth is driven by employment cators; improved standards in plant monitoring; and declines rather than increased value added, eco- a higher number of workers per job.10 These indica- nomic upgrading in fact leads to social downgrad- tors are usually measured at different levels of analy- ing. Similarly, a decline in relative unit labor costs sis, such as country, sector, GVC, and firm, and are can be driven by wage declines rather than produc- compiled from sector-based case studies. tivity increases. Social upgrading may also occur in Measures of social upgrading are likely particular the absence of economic upgrading, and a country to trade within GVCs. For example, the employment may experience simultaneous economic and social rate of women has been rising in export-oriented downgrading. manufacturing industries, services, and agriculture. Bernhardt and Milberg (2011) find that the trans- However, the relative dynamism of female employ- lation is quite varied across countries and GVCs. ment growth tends to decrease as countries upgrade Their study proposes a simple method for combining economically.11 In addition, if exogenous changes in economic and social upgrading. To get an indicator external demand are perpetuated along value chains, for economic upgrading, a weight of 50 percent is the stability of employment in GVCs may also be assigned to the percentage change in export market lower than non-GVC employment.12 share and the percentage change in export unit value. The indicator for social upgrading is obtained analo- gously, assigning a weight of 50 percent to the per- Is Downgrading a Possibility?13 centage change in employment and in real wages. If economic upgrading is a possibility, is downgrad- The development of the economic and social ing also a possibility? If international competitive- realms between the 1990s and the 2000s for several ness depends in part on production costs, there are low- and middle-income countries (LMICs) in the two routes to improve competitiveness: lowering apparel sector shows many cases of overall upgrad- the payment to factors of production (in particular, ing (figure 10.4). Five of the eight countries with labor and capital) and raising productivity. Without data appear in the first quadrant of clear overall considering capital costs, the issue can be simplified upgraders. Among them, Cambodia has been the Turning GVC Participation into Sustainable Development 203 Figure 10.3. Upgrading and Downgrading Economic upgrading High-road Low-road growth growth Social upgrading Social downgrading High-road Low-road decline decline Economic downgrading Source: Adapted from Milberg and Winkler 2011, 345. Figure 10.4. Economic and Social Upgrading and Downgrading in Apparel, 1990s to 2000s 500 Mexico Cambodia (2,797 / 2,955) 400 Social upgrading, 1990s–2000s (% change) 300 200 100 Lesotho Vietnam China (654 / 18) 0 Mauritius India –50 South Africa –50 0 100 200 300 400 500 Economic upgrading, 1990s–2000s (% change) Source: Adapted from Bernhardt and Milberg 2013. Note: Economic upgrading = average of the percentage change in export market share and the percentage change in export unit value. Social upgrading = average of the percentage change in employment and the percentage change in real wages. 204 Making Global Value Chains Work for Development prime performer, with formidable upgrading in labor standards—such as higher minimum wages economic and social terms. Other outstanding per- and more gender equality—can help fight inequality formers include Vietnam (on the economic front) and enhance upward social mobility, thereby foster- and Mexico (on the social front). The progress of the ing social cohesion. remaining two upgraders has been less pronounced, particularly China’s. Lesotho exhibits social upgrad- Education and Skill Building ing without economic upgrading. Mauritius is the This link enables equal opportunities and upward single case of full-fledged overall downgrading. The social mobility. It can result from the lead firm’s ini- remaining two countries, India and South Africa, are tiative to train its own or its suppliers’ workforce, but intermediate. Both have experienced upgrading in also from providing learning on the job. Such train- the economic sphere but downgrading in the social ing allows workers in GVCs to build their knowledge sphere. Overall, there seems to be a positive rela- and perform tasks that require more skills and pay tionship between economic upgrading and social higher wages. upgrading. Skill building also can raise aspirations for work- ers. For example, some workers in Ghana who had previously worked for a multinational company in Which Links between Social Upgrading and the agribusiness sector exhibited entrepreneurism Cohesion? and started their own business.16 When workers’ Social upgrading can enhance social cohesion in a education and skill upgrading lead to better living country. Social upgrading here includes better liv- standards, that link can also create higher education ing standards in the form of more employment, ambitions for their offspring. higher wages, better working conditions and educa- In addition, training initiatives at the firm can tion (including skills development), more economic enhance a sense of active engagement and trust in the rights, more gender equality, and more economic company, especially if that training covers a broader security (including health insurance and pensions). set of skills. Supplier assistance, including training, The following subsections focus on three types of is associated with formal contracting because of the links: jobs and working conditions, education and risk that informal suppliers may side-sell products to skill building, and health insurance and pensions. other clients. A large share of contracts, especially in agriculture, is informal, which limits skill building Jobs and Working Conditions through training.17 Because training measures target Jobs are perhaps the most important link between only parts of the population, they should not be con- social upgrading and cohesion, because unemploy- sidered a substitute for addressing the deeper chal- ment—especially among the youth—can be related lenges of a country’s education system. to social unrest, such as during the Arab Spring. Jobs can help alleviate social tensions because they Health Insurance and Pensions create trust in other people and institutions, as well LMICs have low health insurance and pension cov- as contribute to more civic engagement and thus erage rates (less than one-quarter, on average), espe- social cohesion. In addition, jobs can shape social cially in Africa and Asia. Coverage is particularly low interactions by providing social identity to workers, for low-income workers, often less than 10 percent. connecting people of different socioeconomic and Social upgrading in GVCs can lead to more economic ethnic backgrounds, raising awareness of different security for workers in the form of health insurance views, and influencing people’s aspirations. By con- and pensions. Access to health insurance and pen- trast, social networks can have a negative impact on sion programs usually is linked to jobs because those social cohesion by excluding people who are not part programs are largely financed through payroll taxes of the network.15 (from employers, employees, or both). Working conditions in GVCs also contribute to On the downside, financing social insurance more social cohesion. Better working conditions or programs through payroll taxes excludes informal corporate social responsibility standards, including workers, which in turn discourages employers from economic rights (such as freedom of association) and creating more formal jobs (if the taxes are fully or more security at the workplace (such as increased partially paid by the employer) and discourages HSE standards), promote trust and inclusion. Higher employees from working in the formal sector (if the Turning GVC Participation into Sustainable Development 205 taxes are fully or partially paid by the employee). percent by 2050, but that number doubles under According to recent surveys, workers in LMICs more extreme projections.21 highly value access to health insurance and pensions, The global trade landscape is trending toward and would be willing to contribute a significant share more climate-friendly international standards and of their income to social insurance.18 mandatory sustainability reporting regimes. Some Unequal coverage can also discourage workers of the issues affected include wildlife trafficking, ille- who enjoy social insurance from moving to other gal logging, sustainable management of ocean and firms that do not offer social insurance, which limits coastal resources, energy efficiency, infrastructure for positive knowledge spillovers through labor mobil- electric vehicles, responsible mining practices, chem- ity. A recent World Bank survey in Sub-Saharan ical health and safety cooperation, trade in environ- African countries, for example, confirmed that work- mental goods, and aviation emissions. ing for multinationals in the mining sector seems For countries to comply with such standards to be attractive to local workers, so they tend to stay long term, strategic policy responses are necessary. there rather than move to other firms or start their That will require the mainstreaming of a triple own businesses. That finding has a double negative bottom-line approach to planning that accounts for impact: such firms attract and keep the best workers, financial, social, and environmental policy implica- which leads to skills shortages elsewhere in the local tions. The world’s most successful firms are already labor market, and the reluctance of those workers embracing a culture of “disruptive thinking” when to move on inhibits labor turnover and knowledge envisioning how best to plan for the future. More spillovers.19 participation in GVCs can have a “pro-competition Equalizing opportunities in access to health insur- effect,” leading to increased competition for lim- ance and pensions in a country therefore enhances ited or vulnerable resources. Increasing the scale of social cohesion by integrating the disadvantaged and production can further amplify that effect, requir- helping people build an encompassing social con- ing carefully planned investments in infrastructure. tract. It also helps to reduce inequalities and fosters With an effective strategic vision, countries can (generational or intergenerational) social upward strengthen the ability of their firms to sustain GVC mobility, which contributes to a sense of well-being.20 participation over time. Promoting Environmental Sustainability: Policy Options What Benefits from Environmental Policy has a role in promoting social upgrading and Regulation? cohesion, and environmental sustainability through Firms today are more vulnerable than ever to shifts in GVCs. This section presents complementary precon- the economy and exogenous disruptions. The chang- ditions and policies for government to maximize the ing climate and the resulting changing policy land- sustainable development impact of GVC activities.22 scape are creating new challenges for firms as they seek to ensure the long-term predictability, reliabil- Creating a World-Class Workforce ity, and time-sensitive delivery of goods necessary to participate in GVCs. Climatic disruption can impair Developing Skills firms’ ability to access inputs and deliver final prod- Skill development is a key element not only of com- ucts, making countries’ preparedness an increasingly petitiveness and economic upgrading, but also of critical factor in firms’ location decisions. social upgrading. In other words, economic and Climate change is a multi-sector and uncertain social upgrading are linked and dependent on each phenomenon. Those attributes make evaluating other. Skill shortages can impede upward social economic impacts and designing robust and appro- mobility, and low social mobility can impede eco- priately prioritized adaptation strategies difficult for nomic upgrading. In Chile, Costa Rica, Ethiopia, countries. For example, estimates for Vietnam—one and Rwanda, upgrading strategies in GVCs have of the world’s five most vulnerable countries to cli- been most successful when accompanied by comple- mate change—suggest that climate change is likely mentary workforce development interventions. For to reduce the country’s national income by 1 to 2 workforce development to succeed, it must be part of 206 Making Global Value Chains Work for Development Box 10.1. Succeeding in New Knowledge-Intensive Niche Sectors Nordic Europe has produced many global niche players. Its gov- Nokia fostered multiple startups that produced goods and services ernments recognize the need to encourage more entrepreneurs as diverse as online gaming, automatic recycling systems, do-it- if they want to provide their people with highly paid jobs. They yourself family dining services, and devices that improve people’s therefore encourage universities to commercialize their ideas, moods by firing bright light into the ear canal. The company cre- generate startups, and invest in promoting entrepreneurship— ated an agency that focused on fostering entrepreneurship, Tekes, rather than rely on large local companies to generate business and endowed it with a large staff and budget. A venture capital ecosystems on their own. fund, Finnvera, found early-stage companies and helped them get Three main factors explain the ability of firms in those coun- established. Finally, a large network of business accelerators was tries to develop successful ventures in knowledge-intensive niche financed either with fully public money or through public-private sectors.a First is a commitment to relentless innovation and its partnerships. application to even the most basic industry. Innovation explains Innovation in Finland and other Nordic countries goes well the continuing success of the Danish toy company Lego and beyond the generation of high-tech. Bridging the gap between the ability of a small country such as Denmark to be the world’s engineering and design, innovation in marketing and financing is eighth largest exporter of food products in the world. Second, and equally important. The success of Rovio Entertainment’s Angry related to the first, those countries make a continuing effort to Birds, for example, comes largely from combining skilled mastery upgrade processes through capital-intensive inputs, adding value. of technology with red-hot business acumen. Indeed, innovative Third, flat governance structures and a culture that promotes trust business models explain much of the success of recent Nordic and cooperation allow for consensus-based decisions and long- startups. term planning, thereby creating a business-friendly environment. Particularly instructive is the way in which Finland responded to the decline of Nokia, on which it had become overly dependent. a. The Economist 2013. a coherent overall upgrading strategy.23 An example real wages improve, some workers may lose their jobs is the case study from Nordic Europe in box 10.1. For or see their wages decline when they switch jobs. To a more detailed discussion on the importance of skill facilitate that adjustment, governments can, first, development, see section “Creating a World-Class reduce frictions that increase the costs to workers of Workforce” in chapter 9. moving between jobs and, second, put in place social assistance programs designed to accelerate the tran- Promoting Social Upgrading sition.24 Introducing minimum wages also can pro- Social upgrading can be supported through labor reg- mote social upgrading and cohesion (box 10.2). ulation and monitoring. Host countries must ensure Some countries, such as Brazil, improved the that GVC partners observe the local and national living standards of workers and fought income dis- labor regimes, which should meet core international parities by raising minimum wages in the 2000s. labor standards (for example, the Organisation And although those increases target only the formal for Economic Co-operation and Development sector, the outcomes can spill over to the informal Guidelines for Multinational Enterprises, the Core sector through labor turnover. Misuse of minimum Labour Standards of the International Labour wages can also lead to negative employment effects, Organization (ILO), and the United Nations however, especially if wages are raised in economic Guiding Principles on Business and Human Rights). downturns (such as in Colombia in the late 1990s) However, adopting such standards does not ensure or too quickly (for example, in Indonesia in the early implementation—let alone enforcement—and gov- 1990s). Moreover, the impact on workers is unequal ernments should also ensure comprehensive and sys- and depends on enforcement and compliance, as tematic monitoring with assistance from watchdog well as the labor market segmentation between for- organizations. mal and informal workers. Minimum wages should Well-functioning labor markets are also important, therefore not be seen as a substitute for an effective because the process of integrating into GVCs neces- social policy to mitigate inequality in outcomes.25 sarily entails a reallocation of resources, including Many other factors beyond labor markets and labor, among firms or economic sectors or between social policies contribute to social upgrading and can both. Even as employment opportunities and average be addressed by three sets of initiatives.26 Turning GVC Participation into Sustainable Development 207 Box 10.2. Bangladesh’s Minimum Wage in the Apparel Industry Figure B10.2.1. Minimum Wage per Month for Selected In 2010, following months of violent protests over labor and Countries safety standards, the Government of Bangladesh raised the China monthly minimum wage in the apparel industry from Tk 1,662.50 to Tk 3,000 (about US$38 today). The increase of roughly 80 per- Cambodia cent—the first in the industry since 2006—includes an allowance Indonesia for housing (Tk 800) and medical expenses (Tk 200). Following the collapse of Rana Plaza in April 2013, the Gov- Vietnam ernment of Bangladesh faced even stronger pressure to increase safety and labor standards. As of 2013, Bangladesh had the India world’s lowest minimum wage, one-half the level of Cambodia Bangladesh (US$75) and US$100 less than China (see figure B10.2.1). The government decided to lift the minimum wage to Tk 5,300 (about 0 35 70 105 140 US$68)—a 77 percent raise. U.S. dollars Sources: Bajaj 2010; Mahmood 2013; Yardley 2013. Source: Data from State Department, taken from The Wall Street Journal. Nonstate Initiatives and other contract conditions, including insurance Social upgrading can be promoted through private and pension requirements. governance in the form of (1) corporate policies that exceed minimum standards, (2) negotiated arrange- International Initiatives ments between the corporate sector and labor rep- These initiatives have been fostered at various levels: resentatives, and (3) civil society and consumer campaigns. Social upgrading can be promoted through • Multilaterally, the Policy and Performance Stan- voluntary or semi-voluntary agreements by firms to dards of the International Finance Corporation pay living wages and provide other benefits, as well (IFC) have included reference to the ILO’s core as social institutions that provide services to unem- standards and other labor standards. ployed workers and the working poor. Such initiatives • Coordinated or collaborative multi-stakeholder include standards adopted by industry groups, activi- approaches include the ILO/IFC Better Work Pro- ties of business associations and chambers of com- gramme, Ethical Trade Initiative, Social Account- merce, framework agreements that establish norms of ability International, and United Nations Special trust and conduct, efforts by development associations Representative Ruggie’s Guiding Principles on to attract certain forms of foreign investment or coop- Business and Human Rights. erate with greenfield startups, direct changes in the • As part of its regional trade agreements, the Euro- production process or the structure of buyer-driven pean Union grants bilateral trade concessions to value chains and production networks, and corporate countries that implement the ILO’s core labor social responsibility initiatives by leading brands. standards and other basic rights. • Although regional free trade agreements, such as Government Initiatives the North American Free Trade Agreement and Governments in LMICs can address social upgrad- the Central American Free Trade Agreement, ing by strengthening public institutions for labor include side agreements regarding labor, their regulation (such as labor inspectorates or health and coverage is more limited and they do not explic- safety inspectorates); developing governance capaci- itly refer to ILO standards.27 ties, including social safety nets and other income transfer mechanisms; enforcing labor laws, includ- Policy Objectives and Performance Indicators ing working time and child labor laws (as in Brazil, Policies to support social upgrading should be indi- Chile, Costa Rica, and the Dominican Republic); vidually tailored to the country’s specific situation increasing minimum wages (as in Bangladesh’s and consistent with its overall development strat- apparel industry in 2010); and regulating overtime egy. To comply with those frameworks, local firms 208 Making Global Value Chains Work for Development Table 10.1. Promoting Social Upgrading: Policy Objectives and Performance Indicators Policy objectives • Adopt core international labor standards, and ensure implementation and enforcement, as well as comprehensive and systematic monitoring. • Reduce frictions that increase the costs to workers of moving between jobs, and put in place social assistance programs to accelerate the transition. • Introduce and raise minimum wages to improve living standards, and ensure enforcement and compliance. • Strengthen public institutions for labor regulation, and develop governance capacities. Performance indicators Labor standards: • ILO NORMLEX Labor market frictions and social assistance: • Skills mismatch (ILO KILM) • Employment protection legislation (ILO EPLex; OECD EPL; IFC Doing Business Indicators—Employing Workers) • OECD public expenditure on labor market programs—Public employment services and administration; training; employment incentives; sheltered and supported employment and rehabilitation; direct job creation; startup incentives; and so forth Minimum wages/working poor: • ILO Working Conditions Laws Database • OECD Labour Force Statistics (LFS) • Working poor statistics (ILOSTAT) Implementation and institutional/governance capacity: • Labor inspection indicators (ILO ILOSTAT) • World Bank CPIA—Quality of public administration rating • World Bank Actionable Governance Indicators (AGI) • Sustainable Governance Indicators (SGI) Note: CPIA = Country Policy and Institutional Assessment; EPL = employment protection legislation; EPLex = ILO employment protection database; IFC = International Finance Corporation; ILO = International Labour Office; ILOSTAT = ILO database of labor statistics; KILM = Key Indicators of the Labour Market; NORMLEX = Information System on International Labour Standards; OECD = Organisation for Economic Co-operation and Development. generally require a well-functioning labor market opportunities and outcomes can be complemen- and a strong social governance framework with tary.29 Income-based scholarships, for example, are regulation and capacity building. Table 10.1 lists the cash transfers (promoting equality of outcomes) that policy objectives discussed in this section and the are conditional on education for students (promot- available performance indicators. ing equality of opportunity in the future). Three policy options are recommended: facilitate Engineering Equitable Distribution of access to information, remove discriminatory social Opportunities and Outcomes institutions and establishing rights, and reform For social upgrading to translate into social cohe- social insurance. sion through better living standards, a country must ensure equality of opportunity and outcomes. A soci- Facilitating Access to Information ety can support the relative poor financially (through Equality of opportunity requires including groups of income support or progressive taxes) and through the society that face obstacles to seizing opportuni- provision of services.28 Of particular relevance to ties because they lack information about opportuni- GVCs is the minimum wage. ties or their roles, rights, and entitlements. Equality Promoting equality of opportunities targets of access to jobs is the most important opportunity excluded groups of the society—such as women, for GVCs. Providing access to widely advertised informal workers, rural inhabitants, and minori- information about job vacancies and practical advice ties—by reducing inequalities and discrimination. about how to get those jobs is a precondition of Relevant policies for GVCs include granting equal equality of access. A common program is job search access to jobs, education, health insurance, and pen- assistance, which makes job matching more effective sions. In practice, policies that engineer equality of by providing information about job vacancies and Turning GVC Participation into Sustainable Development 209 job seekers. Assistance can also include job place- to lower-paid, informal jobs. Discriminatory barriers ment and counseling. include formal social institutions and informal social Workers must be informed about their rights institutions, such as norms, values, and traditions. and entitlements. Farmers, the self-employed, and Those informal barriers are reflected in gender- informal workers often are unaware of their rights related stereotyping that discourages women (and in relation to landowners, traders, or employers. men) from choosing untraditional professions.35 Cooperatives, associations of informal workers, Establishing the rights of freedom of association and trade unions can be effective channels of infor- (say, in organizations or trade unions) and collective mation and expression.30 The need to have a voice bargaining enhances social cohesion, thanks to the extends to formal workers and requires that freedom possibility for social dialogue that can address tensions of association and collective bargaining rights be before they lead to conflict. In an attempt to maintain implemented. Policy makers also have to raise aware- social cohesion during the labor market transition, ness of social assistance and other social entitlement China has had collective bargaining mechanisms since programs, especially pensions and health insurance. the mid-1990s, leading to the Labor Contract Law Skill development includes clearly communi- of 2008, which regulates the governance of collective cating to workers about their specific role in the contracts. The establishment of coordination bodies value chain. Female workers in Chinese factories at the province, city, and prefecture levels accompa- often were unable to explain exactly what they were nied that law.36 doing.31 Understanding one’s role and contribu- Although trade unions provide voice to employed tion to the overall good promotes a sense of social workers, they do not cover self-employed or informal identity and belonging, which in turn contributes to workers, who still make up a large share of the work- social cohesion. In addition, workers and firms need force in LMICs. The demand for alternative insti- to be given access to information about accredited tutions of collective representation resulted in the training programs. Training may be provided by pri- emergence of associations of self-employed workers, vate firms, donor programs (such as the U.S. Agency who united to demand better working conditions, for International Development), the public sector, including the protection of rights. Anecdotal evi- and, in some cases, private trainers. For example, in dence shows that in some cases, those efforts include Burundi and Rwanda, private trainers in the infor- filing claims at court, as with street vendors in Lima, mal sector provide fee-based training.32 Peru, and Durban, South Africa.37 Managing information is particularly impor- tant for social insurance, because many LMICs lack Reforming Social Insurance instruments for identifying people. Technological One right is granting universal access to social insur- advances such as biometric technology can help ance. Reforming a country’s social insurance systems overcome such challenges and reduce costs, leakages, can facilitate wider coverage of health insurance and and corruption. Information management systems pensions. To enable knowledge spillovers through must also track people’s medical or work history to the labor mobility effect, it is important to ensure align benefits with contributions.33 portable health and pension benefits across jobs. In Indonesia, some provinces extend noncontribu- Removing Discriminatory Social Institutions tory social health protection to uninsured groups. and Establishing Rights Because funds are pooled at the province level (or Facilitating access to jobs for excluded or disadvan- even at the district level, as in South Sumatra), the taged groups of society, especially women and minor- portability of health benefits is limited.38 In addition, ities, helps economies tap a large productive potential minimum social insurance—notably pensions— and tightens social cohesion. Antidiscrimination can help alleviate economic insecurity. A simula- laws and mandatory or voluntary affirmative action tion model of 18 Latin American countries based on programs are a prerequisite for greater equality of household survey data revealed that universal mini- opportunities.34 mum pensions would substantially reduce poverty Guaranteeing women their property and inheri- among the elderly in most of the countries.39 tance rights enhances their security and equal- The challenge is considerable when social insur- ity, and can enable them to take advantage of ance systems differentiate between formal and formal job opportunities instead of being confined informal jobs, especially if the financing for formal 210 Making Global Value Chains Work for Development Table 10.2. Engineering Equitable Distribution of Opportunities and Outcomes: Policy Objectives and Performance Indicators Policy objectives • Facilitate access to information about opportunities, roles, rights, and entitlements. • Remove discriminatory social institutions by putting in place antidiscrimination laws and mandatory or voluntary affirmative action programs; establish women’s rights (for example, property and inheritance rights) and the rights for freedom of association and collective bargaining. • Reform social insurance systems and combine them with more traditional social assistance programs. Performance indicators Access to information: • IFC Women, business and the law indicators—Accessing institutions • OECD Public expenditure on labor market programs—Placement and related services Antidiscrimination laws and rights: • ILO NATLEX • IFC Women, business and the law indicators—Using property, getting a job, building credit, going to court, and so forth • FAO Gender and land rights database—Property and use rights; inheritance rights; and so forth • World Bank CPIA—Property rights and rule-based governance ratings; gender equality rating • ILO NORMLEX—Freedom of association cases • Trade union density and collective bargaining coverage (ILOSTAT) Social insurance and assistance: • ILO NATLEX • Social security indicators (ILOSTAT)—Social protection coverage; public social protection expenditure; and so on • OECD Social Expenditure Database—Labor market programs; health; old age; and so forth • World Bank CPIA—Policies for social inclusion/equity; social protection rating • WDI—Benefits held by first 20 percent of population and program participation (all social insurance; all social protection; all social safety nets; unemployment benefits; and ALMP) Note: ALMP = active labor market policies; CPIA = Country Policy and Institutional Assessment; FAO = Food and Agriculture Organization; IFC = International Finance Corporation; ILO = International Labour Office; ILOSTAT = ILO database of labor statistics; NATLEX = Database of National Labour, Social Security and Related Human Rights Legislation; NORMLEX = Information System on International Labour Standards; OECD = Organisation for Economic Co-operation and Development; WDI = World Development Indicators. workers is based on contributions and that for workers not only increase overall coverage rates, but informal workers is based on taxes. Tax-based social also facilitate knowledge spillovers through labor assistance programs for informal workers de facto turnover in a country.42 “subsidize” informal work by taxing formal workers Such policies can be combined with more tradi- twice. The portability of social benefits across firms tional social assistance that targets other uninsured therefore requires more innovative instruments sectors of the population (such as the unemployed that target informal workers—who often have the and the elderly). Progress has been substantial in means to contribute to social insurance systems— offering universal entitlement in health, often by as well as a country’s capacity to manage worker creating a parallel system to cover the uninsured. transitions.40 Thailand’s health insurance coverage, for example, One non-tax-based possibility to include infor- reached 98 percent in 2007, although universal cov- mal workers is to offer “unbundled individualized erage was introduced only in 2001. Before the health instruments,” such as individual retirement savings reform, only employees in the public sector or in accounts, which would allow informal workers or firms with more than 20 employees were covered. workers who switch between formal and informal Social pensions also help narrow the coverage gap, jobs to contribute. Subsidized contributions by the although transfers tend to be small (such as US$2.30 state could complement the program. Fairly high per month in Bangladesh). Nevertheless, social pen- contribution rates by informal workers in Mexico sions have coverage rates of about 90 percent in have related pension reforms along those lines. Kyrgyz Republic and Lesotho.43 Table 10.2 shows the Similar approaches are plausible in health insur- policy objectives discussed here and possible perfor- ance.41 Social insurance reforms that target informal mance indicators. Turning GVC Participation into Sustainable Development 211 Implementing Climate-Smart Policies and 12. OECD, WTO, and UNCTAD (2013). Infrastructure 13. This section draws on Milberg and Winkler (2013). 14. Flanagan (2005). Climate-smart policy prescriptions can strengthen 15. World Bank (2013b). global competitiveness. Recent research44, 45, 46, and 47 16. Kaiser Associates Partners (2014a). 17. Kaiser Associates Partners (2014a). has shown that the benefits of environmental regu- 18. World Bank (2013b). lation often vastly outweigh the costs. Proper regu- 19. Kaiser Associates Partners (2014b). lation can induce innovation in green technologies 20. OECD (2011); World Bank (2013b). and produce economy-wide benefits. 21. Channing, Tarp, and Thurlow. 2015. Policy responses to ensure the sustainability of 22. See figure O.1 in the book’s Overview for the strate- gic questions and policy options for this focus area. GVCs and sustained economic growth amid climate 23. World Bank (2014). change include the following examples: 24. Hollweg and others (2014). 25. OECD (2011). • Government should continue to invest in informa- 26. Barrientos, Gereffi, and Rossi (2011). tion systems to monitor climate change impacts 27. Barrientos, Gereffi, and Rossi (2011). in agribusiness; cooperate with the global com- 28. Shared prosperity refers to expanding a country’s income and sharing it such that people in the bottom 40 munity on the development of heat-resistant crop percent of the income distribution increase their welfare as varieties; work to improve water use efficiency; quickly as possible (World Bank 2013a). and ensure that standards for infrastructure, such 29. OECD (2011). as roads, are designed to endure more extremes 30. World Bank (2013b). (warmer/colder) and a more variable climate. 31. Chang (2012). 32. World Bank (2014). • Governments in countries with extensive coastal 33. World Bank (2013b). lowlands should consider the gradual channel- 34. World Bank (2013b). ing of economic activity to safer, higher elevation 35. OECD (2011). zones. The location and vulnerability of the capi- 36. OECD (2011); World Bank (2013b). tal stock in many LMICs in the coming decades is 37. World Bank (2013b). still a matter of choice. Sea-level rise and the con- 38. ILO (2014). 39. Dethier, Pestieau, and Ali (2010). tinual threat of cyclones make this consideration 40. OECD (2011). critical. 41. OECD (2011). • The development of disaster risk mechanisms will 42. For a collection of studies describing strategies to be key to sustainable economic growth. Disaster promote universal health coverage and its effects in 22 risk financing markets are a crucial new develop- countries, see the “Universal Health Coverage Study Series” (World Bank 2013a). ment that merits closer attention for most coun- 43. OECD (2011). tries. Strengthening financial resilience should 44. Aguado, Alvarez, and Domingo (2013). include enhancing technical and institutional 45. Ambec and others (2013). capacities related to crisis management, coordi- 46. Di Marchi, Di Maria, and Micelli (2013). nating various governmental authorities across all 47. Eccles, Ioannou, and Serafeim (2014). levels, and supporting the continuity planning of business. References Ambec, S., M.A. Cohen, S. Elgie, and P. Lanoie. 2013. Notes “The Porter Hypothesis at 20: Can Environmental Regulation Enhance Innovation and Competitiveness?” 1. Milberg and Winkler (2013). Review of Environmental Economics and Policy. 2. Milberg and Winkler (2013). Oxford Journals: First published online January 4, 2013. 3. From Barrientos and others (2011, 301). Aguado, S., R. Alvarez, and R. Domingo. 2013. “Model 4. UNCTAD (2013). of Efficient and Sustainable Improvements in a Lean 5. OECD (2011). Production system through Processes of Environmental 6. Barrientos, Gereffi, and Rossi (2010); Rossi (2013). Innovation.” Journal of Cleaner Production 47 (May): 7. Barrientos, Gereffi, and Rossi (2011). 141–148. 8. Barrientos, Gereffi, and Rossi (2011). Arndt, Channing, Finn Tarp, and James Thurlow. 2015. 9. Barrientos, Gereffi, and Rossi (2010). “The Economic Costs of Climate Change: A Multi- 10. Milberg and Winkler (2011). Sector Impact Assessment for Vietnam.” Sustainability 11. UNCTAD (2013). 7: 4131–45. 212 Making Global Value Chains Work for Development Bajaj, Vikas. 2010. “Bangladesh Garment Workers Awarded Social Protection Floors and Comprehensive Social Higher Pay.” The New York Times, July 28. http://www Security Systems . Geneva: ILO. http://www.social .nytimes.com/2010/07/29/business/global/29garment -protection.org/gimi/gess/ShowWiki.action?wiki .html?_r=0. .wikiId=869. Barrientos, Stephanie, Gary Gereffi, and Arianna Rossi. Kaiser Associates Partners. 2014a. “Sector Case Study: 2010. “Economic and Social Upgrading in Global Agribusiness.” In Making Foreign Direct Investment Production Networks: Developing a Framework for Work for Sub-Saharan Africa: Local Spillovers and Analysis.” Capturing the Gains Working Paper 2010/03, Competitiveness in Global Value Chains, edited by University of Manchester, United Kingdom. Thomas Farole and Deborah Winkler, 163–207. ———. 2011. “Economic and Social Upgrading in Global Washington, DC: World Bank. Production Networks: A New Paradigm for a Changing ———. 2014b. “Sector Case Study: Mining.” In Making World.” International Labour Review 150 (3–4): 319–40. Foreign Direct Investment Work for Sub-Saharan Africa: Barrientos, Stephanie, Frederick Mayer, John Pickles, Local Spillovers and Competitiveness in Global Value and Anne Posthuma. 2011. “Decent Work in Global Chains, edited by Thomas Farole and Deborah Winkler, Production Networks: Framing the Policy Debate.” 117–62. Washington, DC: World Bank. International Labour Review 150 (3–4): 299–317. Mahmood, Syed Zain Al-. 2013. “Bangladesh to Raise Pay Bernhardt, Thomas, and William Milberg. 2011. for Garment Workers.” The Wall Street Journal, May 12. “Economic and Social Upgrading in Global Value http://online.wsj.com/news/articles/SB1000142412788 Chains: Analysis of Horticulture, Apparel, Tourism 7324715704578479231065424630. and Mobile Telephones, Capturing the Gains.” Working Milberg, William, and Deborah Winkler. 2011. “Economic Paper 2011/06, University of Manchester, United and Social Upgrading in Global Production Networks: Kingdom. Problems of Theory and Measurement.” International ———. 2013. “Does Industrial Upgrading Generate Labour Review 150 (3–4): 341–65. Employment and Wage Gains?” In The Oxford ———. 2013. Outsourcing Economics: Global Value Chains Handbook of Offshoring and Global Employment, edited in Capitalist Development. New York: Cambridge by A. Bardhan, D. Jaffee, and C. Kroll, 490–533. New University Press. York: Oxford University Press. OECD (Organisation for Economic Co-operation Chang, Leslie T. 2012. “The Voices of China’s Workers.” and Development). 2011. Perspectives on Global TED (Technology, Entertainment and Design) Talk Development 2012: Social Cohesion in a Shifting World. (online video). http://www.ted.com/talks/leslie Paris: OECD. _t_chang_the_voices_of_china_s_workers. OECD, WTO, and UNCTAD (Organisation for Economic Dethier, J. J., P. Pestieau, and R. Ali. 2010. “Universal Co-operation and Development, World Trade Minimum Old Age Pensions: Impact on Poverty Organization, and United Nations Conference on Trade and Fiscal Cost in 18 Latin American Countries.” and Development). 2013. Implications of Global Value Policy Research Working Paper 5292, World Bank, Chains for Trade, Investment, Development, and Jobs. Washington, DC. Prepared for the G-20 Leaders Summit, Saint Petersburg Di Marchi, V., E. Di Maria, and S. Micelli. 2013. (Russian Federation). http://www.oecd.org/trade/G20 “Environmental Strategies, Upgrading and Competitive -Global-Value-Chains-2013.pdf. Advantage in Global Value Chains.” Business Strategy Rossi, Arianna. 2013. “Does Economic Upgrading Lead and the Environment 22: 62–72. Published online May to Social Upgrading in Global Production Networks? 14, 2012 in Wiley Online Library. Evidence from Morocco.” World Development 46: Eccles, R.G., I. Ioannou, and G. Serafeim. 2014. “The 223–33. Impact of Corporate Sustainability on Organizational UNCTAD (United Nations Conference on Trade and Processes and Performance.” Management Science 60 Development). 2013. “World Investment Report (11): 2835–57. 2013—Global Value Chains: Investment and Trade for The Economist . 2013. “Special Report—The Nordic Development.” UNCTAD, Geneva. Countries: Northern Lights.” The Economist, Febru- World Bank. 2013a. The World Bank Group Goals: End ary 2, 2013 http://www.economist.com/sites/default Extreme Poverty and Promote Shared Prosperity. /files/20130202_nordic_countries.pdf. Washington, DC: World Bank. Flanagan, Robert J. 2005. Globalization and Labor ———. 2013b. World Development Report: Jobs . Conditions: Working Conditions and Worker Rights in Washington, DC: World Bank. a Global Economy. New York: Oxford University Press. ———. 2014. Republic of Burundi: Skills Development for Hollweg, Claire H., Daniel Lederman, Diego Rojas, and Growth—Building Skills for Coffee and Other Priority Elizabeth Ruppert Bulmer. 2014. Sticky Feet: How Labor Sectors. Washington, DC: World Bank. Market Frictions Shape the Impact of International Trade Yardley, Jim. 2013. “Bangladesh Takes Step to Increase on Jobs and Wages, vol. 1. Directions in Development: Lowest Pay.” The New York Times, November 4. http:// Trade, Report 88890, World Bank, Washington, DC. www.nytimes.com/2013/11/05/world/asia/bangladesh ILO (International Labour Organization). 2014. -takes-step-toward-raising-38-a-month-minimum “Indonesia—Health.” In Social Protection: Building -wage.html. PART IV COUNTRY ENGAGEMENT Chapter 11 DESIGNING A COUNTRY ENGAGEMENT STRATEGY BASED ON SOUND ANALYTICS What Is the Goal of This Guide? • Interventions need to build on analytical founda- tions and follow well-targeted and action-bound Use of This Book to Help Design a Country action plans, but they do not need to follow a Engagement Strategy to Achieve GVC-Led standard sequence or timeline abstracting from Development country-specific and context-specific conditions. The coordination, information sharing, and lever- To complement parts I to III of this book, this chapter aging synergies between different interventions are offers guidelines on engaging with country stakehold- important. Coordination demands are high within ers to implement a national strategy to achieve eco- government agencies, GVC stakeholders, and nomic and social development through global value donor partners. chain (GVC) participation. Policy and its implemen- • A participative approach, with alignment on tation in a wide range of influencing areas affect the and ownership of the agenda by all stakehold- odds of success in GVCs. Those areas are as different ers is critical. Effective stakeholder engagement as trade and trade policy, domestic services regula- mechanisms are a central anchor for continued, tions, investment regulations and incentives, compli- long-lasting results (but often the least funded). ance with process and product standards, innovation, Successful sector-specific, public-private coop- industry, entrepreneurship, labor markets, education, eration and dialogue are required to inform and infrastructure and connectivity, as discussed in national competitiveness strategies, investment detail in part III of this book. Thus, creating synergies climate reforms, and investment attraction with on the ground requires multiple interventions and opportunities and challenges at the micro level. long-lasting engagement with a variety of stakehold- Leveraging and reinforcing existing coopera- ers within and outside the country. A few important tion through systematic consultations and for- recommendations and lessons learned for interven- mal mechanisms of bottom-up policy making is tions at the country level need to be kept in mind: fundamental. • Network effects and positive spillovers from • The creation of synergies on the ground requires GVC participation across sectors, based on inte- multiple interventions (advisory, analytics, financ- grated solution packages, are achievable over ing, advocacy) and long-lasting engagement. time. Dynamic learning, replication, and scale- • Policy advice supporting GVC-based growth mod- up can be fostered through global/cross-country els requires sound analytics, evidence, and data. It platforms. also requires 360-degree assessment of the com- • A shared vision and a common understanding of petitiveness of a country’s economy, in its entirety, the project goals and objectives between imple- and drilling down to specific sectors, GVCs, tasks, menting teams, local and international stake- and activities, to identify, prepare, and inform all holders, and other development partners are interventions. important for success. 215 216 Making Global Value Chains Work for Development This chapter brings attention to the synergies intervention should be completed with method- between these different areas and helps support ologies that drill down within sectors, GVCs, and countries’ efforts to identify the necessary reforms specific tasks and activities (see box 11.1 for a dis- to trigger a virtuous cycle of “reform-GVC-entry cussion of complementary work that drills down and upgrading-development.” The cycle would within GVCs). encourage the private sector to keep investing Turning back to the pre-engagement, economy- retained earnings in the continued improvement of wide analysis, the successful implementation of full existing and new activities and tasks of comparative GVC diagnostics begins with effective planning and advantage in countries’ agriculture, manufacturing, management, and an understanding of how this and services sectors. The strategic framework for feeds into the overall country engagement strategy of GVC participation developed in this book—map- GVC participation. ping focus areas for policy with relevant objectives, A three-step process can be envisaged for the strategic questions, and policy options—can guide overall country engagement strategy. policy makers in identifying policy options and pri- orities for fostering GVC-led development (see fig- Component 1. Pre-Project Assessment: From Macro ure O.1 in the Overview). The framework is to be to Micro (2 to 3 months) a first step toward a full GVC participation assess- ment and strategy. This pre-engagement analytical The objective of component 1 of a country’s GVC work and identification of priority policy areas for engagement strategy is to provide a comprehensive, Box 11.1 World Bank Group Approach to Diagnostic Work and Formulation of Action Plans to Strengthen a Country’s Position within Specific GVCs The World Bank Group uses a range of instruments, including The need to cover horizontal policies and vertical interventions advisory services and capacity building, lending, investment in a coherent manner means that the World Bank Group approach support, and guarantees, to help countries, their industrial sec- complements economy-wide assessment, as presented in this tors, and firms in the efforts to enter GVCs, upgrade and densify book, with methodologies that drill down to tasks and activities participation in GVCs, and sustain the engagement over time within individual GVCs. The vertical analysis focuses on identifying at the macroeconomic, social, and environmental levels. This strategic segments and business models that deliver high-value- is achieved by supporting countries’ efforts to improve macro- added dividends and development prospects in selected industries. economic and horizontal policies as well as their vertical inter- These strategic segments of focus are determined through diag- ventions targeting specific sectors, GVCs, products, and firms. nostic work and consultations with the private sector, government Targeted challenges and market failures are grouped into three agencies and ministries, global buyers, lead firms, and advanced broad areas: consumers. Where feasible, the approach favors a participative process, so that after its completion, local stakeholders are trained • Internal to the firm (firm capabilities) and empowered with the necessary know-how to drive the process  Managerial capabilities and workforce skills of supporting the competitiveness of the country.  Technology adoption The success of value chain competitiveness reinforcement  Innovation capabilities strategies requires continued and lasting effort, as opposed • Domestic environment to one-off initiatives. For this reason, the capacity-building and  Business climate and institutions training component and direct involvement in the project of local  Financial and labor markets stakeholders need critically to be built within countries’ public  Quality and conditions of output and input factors sector and/or relevant partner institutions. ▪ Education and skills Finally, the World Bank Group approach emphasizes results ▪ Public policies for innovation and impact measures. Although this is often not a request of gov- ▪ Product and process standards ernments, World Bank Group engagements tend to include frame- ▪ Labor and social conditions works to measure results, reforms, and development impacts. • International dimension This is an important way to assess reforms and correct action in a  Infrastructure and policies for connectivity (physical and timely manner if needed. Monitoring and evaluation frameworks information and communications technology) developed over the years by the World Bank Group systems are  International investment used for establishing monitoring and evaluation protocols and  Trade costs and openness follow-up of results during and after the completion of project and embedded technical assistance. Designing a Country Engagement Strategy Based on Sound Analytics 217 fact-based, and independent preliminary view of the in GVCs that offer the promise of the highest value- country’s trade competitiveness (particularly mea- added growth (see also chapter 4 for further illus- sured in value added), performance in GVC integra- trations), as well as further investigation and/or tion, economic upgrading, and the role of country validation of possible binding constraints and solu- characteristics, including the business climate, invest- tions, building on those identified in component 1 ment climate, and drivers of competitiveness across of the engagement strategy and drilling down within economic, regulatory, operational, and infrastruc- GVCs; and (2) creating a detailed road map for start- tural dimensions. This preliminary view is devel- ing to implement reforms. For example, a possible oped through a desk-based analysis followed by a strategy could be to identify a list of four to six major field-based qualitative assessment and discussion of initiatives to maximize shared value added in incor- the identified challenges, opportunities, and policy porating global best practices and placing a priority options with local public and private sector stake- on “quick wins.” Various governance models can be holders. Planning of the pre-project phase should used for designing the appropriate institutional set- focus on the economy as a whole, but also zoom into ting—for instance, by establishing a working group key industries, strategic segments therein, and indi- to work closely with the president’s or prime min- vidual value chains (as narrowly defined as the avail- ister’s office, or by devising a plan for strengthen- ability of quantitative and qualitative information ing the coordinating mandate of one key ministry. allows). A limited number of key industries (three or Participants can be selected from relevant public four) and/or value chains (eight or nine)—existing institutions, including ministries of economy; min- ones that exemplify critical and/or broader opportu- istries in charge of entrepreneurship and domestic nities and challenges, or new ones that are considered economic development; national and subnational important by the local stakeholders, as well as sub- agencies for the promotion of trade, investment, and national specificities—may also be identified at this competitiveness; chambers of commerce; associa- stage for deeper analysis and discussion of challenges tions of employers; regional development agencies; and opportunities. Component 1 provides a first-pass etc. The established governance body will participate analysis of sector- and GVC-specific issues, which can in the work of component 2 and may oversee the be the object of more focused and deeper assessments work of component 3. in component 2 of the engagement strategy. Assessments in component 1 must be based on the Component 3. Execution Phase of Interventions widest range of available and applicable methodolo- (6 to 18 Months) gies. This process allows for customizing the analysis to country-specific needs and overcoming the limits Component 3, which needs to start after the comple- inherent in specific methodologies. As discussed in tion of component 1, but can start as early as six parts II and III of this book, none of the available months after the beginning of component 2 and methodologies allows a full and balanced assessment delivery of early results, covers the execution phase of a country’s participation in GVCs. Each tool illus- of interventions. It includes revising regulations, re- trated in this book was developed for application to engineering processes, and investing in infrastruc- comprehensive analyses. Together, the tools form a ture to achieve measurable improvements across all suite of analytical frameworks and instruments for key dimensions and areas of binding constraint iden- linking performance (outcomes and potential) to tified at the macro and micro levels. diagnostics of countries’ and regions’ competitive- The material in this book focuses on providing ness in goods and services GVCs. Table 11.1 provides tools to support the assessment in component 1. a summary of the methodologies available to carry Component 1 should be treated as a project within out the assessment and their content. the overall GVC engagement strategy, and therefore should be managed accordingly. Preparatory steps must be considered, establishing objectives and roles Component 2. Drilling within GVCs and Capacity and determining the main actors and scope. Building (12 to 24 Months) Component 2, which can start one month after the Who Is This Guide For? start of component 1, includes (1) establishing the model of country engagement and the appropriate The GVC participation assessment in component 1 is institutional setting for identification of strategies to be led by a small core team, most likely comprising 218 Making Global Value Chains Work for Development Table 11.1. Desk-Based Analysis Component Content 1. Macroeconomic trends Value added by broad sector, employment by broad sector, labor productivity by sector, FDI, exports and imports (% of GDP), exports and imports by broad economic category, and other informed classifications 2. Export market share growth, push, and pull Export market share growth; decomposition in push and pull factors using factors shift-share methodologies 3a. World Bank MC-GVC Dashboard (short or long Trade in main GVCs, exports of GVC products relevant to country, top five version) exports (see chapter 3) 3b. World Bank MC-GVC Dashboard (long Extension of 3a, including country dimensions and follow-up analysis of version) interesting patterns (such as product-specific analysis) (see chapter 3) 4a. Network analysis (short version) Worldwide trade network, country trade network for sector of interest (main buyers), country trade network for sector of interest (main suppliers) (see chapter 6) 4b. Network analysis (long version) Extension to more sectors (four or five, maximum) (see chapter 6) 5a. Trade in value-added indicators (EORA) Foreign value added in gross exports, domestic value added in third countries’ exports, GVC participation index (see chapters 4–6) 5b. Trade in value-added indicators (OECD-TiVA Domestic value added in gross exports (total growth and by sector), or WIOD), other GVC indicators, and econometric decomposition, foreign value added in gross exports, domestic value added decomposition of gross exports, for countries in third countries’ exports, sourcing and selling patterns, value added by covered by these more sophisticated databases destination, import and export upstreamness and gap, contribution of direct and indirect domestic value added and foreign value added to gross export growth (see chapters 4–7) 6a. Econometric assessment: structural Impact of structural integration in GVCs (network measure) on domestic value integration into GVCs and economic upgrading added embodied in exports and gross exports (see chapter 7) 6b. Econometric assessment: probability of entry Probabilistic model of entry in GVCs (see chapter 7) in GVCs 6c. Econometric assessment: economic upgrading Impact of GVC integration (foreign value added in gross exports, domestic value added in third countries’ exports) on value added and the role of national policies (see chapter 7) 7a. Econometric assessment: economic upgrading Impact of GVC integration (imported input share, export share, etc.) on labor using Enterprise Surveys productivity and the role of absorptive capacity (see chapter 7) 7b. Econometric assessment: economic upgrading Impact of GVC integration (imported input share, export share, etc.) on labor using national firm-level data productivity and the role of absorptive capacity (depends on whether data need to be cleaned and the extent of analysis, etc.) 8a. Role of services in GVCs (short version) Zoom into the services dimension of analysis in sections 1 to 7 (see chapter 6) 8b. Role of services in GVCs (long version) 9. Country- or product-specific case study Value chain mapping and country positioning, historical/current trends, stakeholder/actor analysis, challenges and opportunities, future implications, policy implications 10a. Policy section (short version) Policy suggestions based on short GVC analysis (and additional research) (see also part III) 10b. Policy section (long version) Application of strategic policy framework, policy suggestions based on long GVC analysis, screening of policy performance indicators, additional research (potentially drawing on information from mission trip) Note: EORA = Environmental Accounting Framework Using Externality Data and Input–Output Tools for Policy Analysis; FDI = foreign direct investment; GDP = gross domestic product; GVC = global value chain; MC-GVCs = Measuring Competitiveness in GVCs Dashboard; OECD = Organisation for Economic Co-operation and Development; TiVA = Trade in Value Added; WIOD = World Input-Output Database. three to four people. At least one team member experience in GVC analysis and, most important, in- should have sound past experience in policy and depth country knowledge and experience. strategic issues related to trade in general and GVCs If the study participants intend to do in-depth in particular, as well as technical skills in analyzing technical analysis as well, technical experts must trade and production data at the macro and firm be involved. If the team considers certain method- levels. Ideally, the task team leader should have some ologies essential from the outset, then bringing in Designing a Country Engagement Strategy Based on Sound Analytics 219 specialized technical expertise to lead those compo- 1. Assessment and research of the country’s partici- nents may be useful. pation in GVCs, through desk research. The top- The success of component 1 of the GVC participa- ics to be covered and methodologies are listed in tion strategy will depend on combining desk-based table 11.1 and discussed in part II of this book. assessments with inputs and qualitative information An example, applied to Bulgaria, of pre-mission from a wide variety of stakeholders in the country, desk research is provided in chapter 2. The follow- including government officials and the private sector, ing are some of the main tasks to consider for the by conducting individual consultations and focus assessment: group interviews and directly involving counterparts a. Preliminary assessment of the country’s in government in the design of the analysis. In some growth in value added over time, sources of countries, with sufficiently sophisticated human value added, location of final demand, and resources in the public sector, it is also possible to actors that drive the country’s participation in establish a partnership with the local government for GVCs (see chapters 4 to 6). conducting the analysis jointly. In some cases, assem- b. Identification and first-cut analysis of key bling a steering group of key stakeholders (govern- sectors, GVCs, and firms that demand closer ment, business, and labor) to provide inputs and investigation (see chapter 3). feedback at some well-identified stages of the assess- c. Identification of peer countries—for bench- ment may also be useful. marking purposes (see chapter 4). Core staff and/or consultants should include the d. Preliminary identification of challenges and following: needs at the micro (firm) and macro (coun- try) levels to support entry and strengthening 1. One task-team leader, preferably based in the of GVC participation and long-term sustain- country ability of the country’s GVC strategy (see 2. One project coordinator and/or senior analyst chapter 7). 3. One or two junior analysts. e. Preliminary identification of policy areas for intervention and collection of evidence from international best practices (see the strate- Steps in Component 1 gic policy framework in figure O.1 in the A full GVC participation study for component 1 of Overview, as well as chapter 7 and part III). a country engagement strategy is conducted in four key steps: 2. Preparation for in-country fieldwork. In-country fieldwork is an exercise that has three key objec- 1. Prepare a preliminary GVC participation assess- tives. It helps in vetting preliminary findings from ment based on available data. the desk analysis. It allows interacting directly with 2. Conduct initial desk research and prepare a pre- key stakeholders, with a view to understand their liminary GVC participation strategy for field- objectives, concerns, and operating environment, work. as well as for further reference. Finally, it repre- 3. Perform in-country field research and document sents a means to identify and analyze qualitative findings. information that complements the quantitative, 4. Refine the policy recommendations. desk-based research. Table 11.2 provides a full list of key private and public sector stakeholders that The study is to be completed in two to three should be consulted, including exporters, local months, including four to six weeks of fieldwork suppliers, lead firms and global buyers, represen- (although a small country could complete the field- tatives of advanced consumers, intermediaries, work in as few as two weeks). Some steps can overlap key equipment and service providers, distribu- chronologically. tors, retailers, standard-setting bodies, key agencies and ministries, leading consulting companies, and Main Activities before Fieldwork international law firms. Meanwhile, the following are examples of information (and relevant target The following are the main activities that should be stakeholders) that desk research is unlikely to doc- done before beginning the fieldwork: ument and that require field interviews: 220 Making Global Value Chains Work for Development Table 11.2. Stakeholders to Target during Fieldwork Organization Persons to meet Government agencies and industry associations  Representatives of key agencies in charge of regulations that affect the    key GVCs identified: – Line ministries – Economic development agencies – Administrations in charge of industrial and business development – Administrations in charge of trade and trade policy – Finance ministry (or agency in charge of tax and incentive policies) – FDI and investment promotion agencies – Chambers of commerce – Standard-setting bodies – Others, when relevant: agencies in charge of domestic services regulations, science, technology, innovation, entrepreneurship, labor markets, education, and infrastructure and connectivity Firms involved in GVCs in selected industries—for  Senior management of a representative sample of firms:    example: – Domestically owned exporters – Automotive – Domestically owned suppliers located in the country – Electrical and electronics – Domestically owned final producers or assemblers that rely – ICT predominantly on imported inputs – Food and agriculture – Local subsidiaries of MNCs in the country that rely predominantly on – Light manufacturing (textiles, apparel, leather imported intermediates and footwear) – Local subsidiaries of MNCs in the country that rely predominantly on – Chemicals and pharmaceuticals domestically produced inputs – Business processing and back office – A few marginal local firms with high growth potential (for example, innovative SMEs or firms with key capabilities for development and upgrading) – A few examples of firms exemplifying failed GVC participation strategies – International HQ of global lead firms; global buyers; intermediaries; global suppliers; representatives of advanced users that can drive outcomes through their decisions on demand, investment, technology, business models, operational processes, standard setting, and so forth – Very few, carefully selected factory visits may be of interest, but the key priority is to speak with top managers rather than observing production processes; some types of factories (such as assembly plants) can be avoided, as they do not provide major insights about a country’s strategic positioning in GVCs or scope for upgrading Key equipment and services providers (including  Senior management    design, R&D, transport and logistics, BPO, and software providers) Venture capital and private equity or other finance  Senior management    providers (especially those that offer finance to SMEs) Distribution sector: wholesalers and large retail  Senior management    chains Management firms of SEZs and competitive  Senior management    spaces Business incubators and accelerators  Manager or business development person    Technological and industrial parks  Manager or business development person    Management consulting firms, consulting firms  Senior management    specialized in IT and software, and international law firms Note: BPO = business process outsourcing; FDI = foreign direct investment; GVC = global value chain; HQ = headquarters; ICT = information and communications technology; IT = information technology; MNCs = multinational corporations; R&D = research and development; SEZs = special economic zones; SMEs = small and medium enterprises. Designing a Country Engagement Strategy Based on Sound Analytics 221 a. Qualitative characterization of the country’s placing large orders, making direct invest- participation in GVC manufacturing and ments, and introducing requirements services functions, challenges and opportu- that demand technologies, processes, and nities in specific business models, power rela- business models that are more advanced tions within GVCs (including governance compared with what goes on locally. structures), business climate, and recent gov- ii. A few innovative local private firms in the ernment initiatives likely to influence develop- activity of interest or in activities that can ments in GVC participation. easily upgrade to the activity of interest. i. Key policy makers and influential deci- c. Involvement of domestic firms in key stra- sion makers from the private sector. The tegic value chain segments and the nature of priority is to speak with the top managers production, including inputs, intermediate in the country, who are able to provide a goods, final goods, and services (for example, clear understanding of the strategic posi- design, logistics, finance, and business process tioning of the country in GVCs. Carefully outsourcing), again being sure to include rep- selected factory visits may be of inter- resentatives from the key categories: est, but some types of factories (such as i. Most important global lead firms or assembly plants) can be avoided, as they OEMs do not provide major insights about the ii. Contract manufacturers and service country’s strategic positioning in GVCs providers, since they work for a variety or on scope for upgrading. of customers and can provide a broader ii. Large consulting firms or international view of the country’s participation in law firms in the country. These can pro- GVCs than global lead firms or OEMs vide a good assessment of the key factors iii. Services and logistics providers, as these for attracting GVC lead firms to the coun- sectors represent the areas where the try and areas for improvement. most sophisticated and higher-value- iii. Decision makers and investors outside the added segments of a value chain usually lie. country. The crucial question of “where does the country fit in the GVC?” is often Fact-checking prior to the fieldwork on the firms best answered by those outside the coun- and key public sector stakeholders targeted for the try—in the regional or global headquar- interviews can increase the odds of a successful inter- ters of the key global firms, private equity view process. Activities carried out locally are not the firms, and large consulting firms and law only important ones. The sector-, firm-, and cluster- firms. iv. Representatives of advanced users. There level analysis illustrated in part II of this book may may be some key in-country consumers help to reveal the upstream and downstream activi- demanding above-average goods and ser- ties in specific value chains. With this information at vices. Interviews with representatives of hand, it is suggested to undertake the following steps advanced users outside the country also before conducting fieldwork: help in understanding demand trends and likely patterns of future GVC devel- a. Collect key facts about key firms present in the opment. Advanced users could be in the country and immediately upstream and down- public or private sector. stream activities, which are relevant to understand b. The most important global and local players upgrading patterns and potential. For large firms, that form a GVC in which the country has a the key facts available from public sources, includ- presence, and which of these are present in the ing the Internet and sector registries, include country, as well as some of the marginal actors – Firm age and ownership, key milestones, merg- that have high potential for growth (for exam- ers and acquisitions, problems encountered, ple, innovative small and medium enterprises): main products, main downstream markets and i. Global lead firms (or original equipment customers, and financial history manufacturers [OEMs]), global buyers, – Role played by the firm in GVCs, if any and global suppliers are a good starting – Upstream supply base: domestic, regional, and point. They can drive development by global 222 Making Global Value Chains Work for Development b. If possible, prepare synthesis tables of the stan- firms (see table 11.2 for a tentative list). The most dard firm-level information illustrated in part II important topics to raise for discussion during of this book (box 6.2 in chapter 6), which is avail- firm interviews are included in the list of relevant able from the World Bank Enterprise Surveys, or questions for each of these topics, as proposed in other relevant firm-level information, including annex 11A. data on outsourcing and offshoring by business 3. Preparation of an aide-memoire and wrap-up function, if available. The tables will be useful in meeting at the end of the mission to share with assessing whether the interviewed firms are out- key government counterparts, laying out the key performers, low performers, or average perform- findings, gaps in knowledge, lessons learned from ers against relevant benchmarks. stakeholders regarding a strategic vision for the country’s GVC-led development, and priority policy interventions identified. Main Activities during Fieldwork The aim of the fieldwork is to refine the identifi- Activities after Fieldwork cation of the key challenges to GVC participation and the operating context, using qualitative infor- The following are the main activities to conduct after mation that statistics are not able to capture. The the fieldwork: analysis will help in the design of the necessary policy and regulatory interventions that will target 1. Finalize the report. the challenges in entering GVCs and upgrading and 2. Plan and prepare an action plan for implementing densifying participation, while also identifying mac- the reform (for component 2 of the process). roeconomic, social, and environmental conditions 3. Prepare an execution plan (for component 3 of to ensure the sustainability of a GVC-based model the process). of development. Activities during the fieldwork include After validating with interviews, pre-fieldwork desk analysis, and sector selection, the last step 1. Sharing desk analysis with local policy makers. consists of providing a detailed understanding of Present the findings from desk analysis and dis- global, regional, and local features of the GVCs of cuss with policy makers (including government interest, main products, processes, key technology officials) the strategic objectives of the country’s trends, actors, regulations, typical transactions, and GVC participation and pre-fieldwork hypotheses all other factors that shape the country’s ability to emerging from the desk analysis. join the GVCs, expand and strengthen its GVC par- 2. Interviews with key stakeholders. A team will ticipation, and turn GVC participation into sustain- conduct field interviews with representative key able development. Annex 11B provides a checklist of players in GVCs, including global lead firms and topics that should be covered in detail by a combi- suppliers, leading domestic suppliers, contract nation of deskwork and fieldwork, offering a quan- manufacturers and services providers, policy titative foundation for starting an effective country makers, professional associations, and services engagement strategy. Designing a Country Engagement Strategy Based on Sound Analytics 223 Annex 11A Interview Guide for Fieldwork g. Top five origin countries of intermediate inputs used by the firm Firm Interviews h. What does the firm do with imported 1. Firm profile intermediate inputs? Transform them for a. Description of the firm profile and its role manufacturing of products that are also in global value chains (GVCs) semi-finished products, parts, and compo- b. Perspective on structure of the sector, GVC, nents? Transform them into manufacturing segment within the GVC, and where the of final products? Use them for assembly of firm fits in semi-finished products, parts, and compo- c. Review of general economic trends experi- nents? Use them for assembly of final prod- enced by the firm, with a focus on economic ucts? Package them as finished products, upgrading (growth of domestic value added parts, and components? in the past) i. If the firm exports or sells semi-finished d. Perspective on competitors (countries and products to local exporters, are those semi- firms): comparative/competitive advantage/ finished products incorporated by the buy- disadvantage within the sector, GVC, and ers into the production or assembly of final segment within the GVC products or into the production or assem- e. Description of required and prospective bly of further intermediates? human capital and technology to carry out j. For firms that sell to local consumers and current tasks and upgrade. export directly and/or sell to local export- ers, are products/services similar regard- 2. Domestic sales and trade less of the customers, or are those sold to a. Focus on the most recent two years, indi- local consumers different in variety/quality/ cation of export mix (top five products), nature? destination markets (top five destinations), and proportion of exports versus domestic 3. Contractual relationships and issues of con- sales and share of domestic sales going to tract enforcement other local exporters a. Is the firm engaged in long-term relation- b. Nature of exports: final products; semi- ships with the main buyers? Through what finished products, parts, and components; type of contractual arrangement (arm’s packaging materials; raw materials; services; length trade, non-equity relationships, or other equity stake)? c. Hypothetical question for top export prod- b. Local versus global procurement and sup- ucts of the firm: if your firm would increase ply challenges, strategies, strengths prices by 10 percent today, what would be c. Processes, authorities, regulatory environ- the percentage variation in revenues in ment for contract enforcement the domestic market as well as the foreign d. Are the exports/sales made to fit exclusively market, provided that competitors did not to suit unique specifications required by adjust their pricing and all other things buyers? being equal? e. Do buyers provide precise information on d. Expenditure, relative to total, in raw materi- product design and/or quality standards? als and intermediate inputs f. When the contract includes specific quality e. Reliance on imported raw materials and standards, are there penalties if the quality intermediate inputs for the production of standards are not met? goods to be exported versus the production g. Are the prices in the contract negotiated or of goods to be sold domestically imposed by the buyers? f. Nature of imports: final products; semi- h. What is the firm’s degree of participation in finished products, parts, and components; the design of the products it manufactures? packaging materials; raw materials; services; i. What type of assistance is provided by and other which type of collaboration is established 224 Making Global Value Chains Work for Development with the buyer? (The possibilities include etc.), or customs and border agency proce- assisting the firm to achieve a particular dures on imports. design of quality standards; assisting the o. For firms that do not engage extensively in firm with manufacturing technologies; international trade (export or import side), lending, renting, or leasing machinery and what is the main reason? Possible options equipment to the firm; providing the firm include lack of information on foreign buy- with raw materials or intermediate inputs; ers or foreign suppliers, prices offered on engaging the firm in process or product international markets are too low, insuf- research and development activities; orga- ficient access to technologies, insufficient nizing exchanges of personnel with the quality of the firm’s own products, lack firm as a way to disseminate and diffuse of scale to meet a buyer’s contract, lack of new technologies to the firm’s production capacity to meet shipping schedules, lack of facilities; assisting financially with pay- finance, or other. ments in advance or other types of financial assistance; assisting with firm management 4. Control over decisions/power structure within practices, such as financial planning, inven- the GVC tory management, or personnel practices; a. Is the headquarters at the same location as assisting the firm to achieve international the production units? quality certifications and meet regulatory b. Is the headquarters located in the country standards; assisting the firm to meet inter- or abroad? If abroad, what is the time to national labor standards; and assisting the travel to the headquarters door-to-door? firm in meeting international environmen- c. Where are the following decisions made tal standards. (within the firm, at the headquarters, in the j. Who is responsible for the logistics (certifi- production establishment, or by the global cations, customs, and transport)—the firm buyer/supplier)? interviewed or its buyers? i. Hiring full-time employees k. Does the firm have long-lasting and sub- ii. Providing an employee a pay increase stantive collaborative relationships with of 10 percent foreign firms (other than those with its iii. Introducing new products buyers, parent firm, or affiliates of the same iv. Advertising products group)? v. Sourcing decisions for machinery, l. If the firm exports, how was the first for- key parts and components, standard eign order obtained? Possible options are parts and components, packaging through active search for a foreign buyer, a materials, and raw materials contact at a trade fair, an unsolicited order, d. Thinking of all the intermediates that are a foreign supplier of intermediate inputs, a used to produce a final good, what share of government export assistance program, or the purchase value can the firm decide from other. what firm to buy, that is, which suppliers, m. If the firm exports, what obstacles are faced with the understanding that the remaining mostly? Possible options include the reli- fraction is decided by headquarters or the ability of foreign customers; language and global buyer/supplier? cultural barriers; access to credit/financing; e. Does the firm have exclusivity clauses in the exchange rate risk; political risk; intellectual contracts with its buyers/suppliers? property risk; contracting problems; trans- f. Does the firm use data to support decision- portation costs; customs taxes, fees, and making? What type of data? procedures; rules of origin; different stan- dards and regulations; or other. 5. Determinants of upgrading n. If the firm imports intermediate inputs, a. What features of the firm have been impor- what obstacles are faced mostly? Possible tant for connecting with the global lead options include custom tariffs, nontariff firm/global buyer? measures (quotas, technical regulations, i. Product design Designing a Country Engagement Strategy Based on Sound Analytics 225 ii.Brand recognition iv. Technology, including low return iii. Low price on investment; difficulty in inte- iv. Product quality grating new advanced technologies v. Advanced production processes and with existing systems, standards, technology and processes; low information and vi. Advanced material handling communications technology (ICT) vii. Supply chain and logistics capabilities; insufficient scale of technologies production; lack of technical sup- viii. Advanced use of information and port from consultants, customers, or communications technologies suppliers ix. Management practices, including v. Captive relation in GVC, exclusivity explicit and systematic performance contracts tracking, reviews, and/or bonuses at vi. Information, including limited the managerial level information on export markets and x. Agile and flexible organization difficulty in identifying business xi. Total quality management, lean opportunities management vii. Procedural, including excessive and xii. Access to skilled workers complicated paperwork, difficulty xiii. Access to cheap intermediate inputs communicating with foreign cus- tomers/partners, slow collection of xiv. Access to high-quality intermediate payments, and difficulty in enforcing inputs and/or raw materials contracts xv. Product quality certifications viii. Government barriers, including lack xvi. Environmental certifications of support from government and xvii. Managerial certifications unfavorable domestic regulations xviii. Labor standards compliance ix. Business environment, including b. Did you take measures to improve your per- unfamiliar business practices of lead formance against any of the above param- firms/global buyers and inadequate eters in the past two years? infrastructure and regulatory envi- c. What are the obstacles at your firm to ronment for e-commerce obtain the capabilities that would allow you x. Tariffs and nontariff barriers, inad- to upgrade? equate property rights protection i. Human resource issues, including xi. Customer and foreign competitor lack of employee training; employee barriers, including too fierce com- resistance to change; organizational petition or different habits/attitudes/ rigidity of the enterprise; difficulty tastes in recruiting blue collar plant techni- cians, white collar workers/managers 6. Productivity and worker skills in finance, managers in marketing, a. What share of the production costs go into managers in research and develop- electricity; fuel, energy, and gasoline; wages ment (R&D), product design, pro- and workers’ compensation; other? duction, or information technology; b. What is the approximate value of the firm’s difficulty in recruiting a chief execu- investment in machinery, vehicles, and tive officer equipment? What is the value of the invest- ii. Product/price issues, including low- ment in land and buildings? quality products, high prices/costs, c. Hypothetically, if the firm had to purchase inadequate product design, packag- the assets it uses now, in their current con- ing, labeling dition and regardless of whether the estab- iii. Finance, including difficulty in lishment uses them or not, how much accessing private funds or public would they cost, independently of whether funds, or obtaining guarantees they are owned, rented, or leased? 226 Making Global Value Chains Work for Development d. What is the composition of workers in the j. Do you monitor key performance indica- firm? What share of production workers, tors? What types? (Possible options: met- professional/technical workers, managers, rics on production, cost, waste, quality, and other types of workers does the firm inventory, energy, absenteeism, deliveries have? What share of white-collar, blue-col- on time, or other). How many? (Possible lar nonqualified, and blue-collar qualified options: 1-2; 3-9; or 10 or more, on key per- employees? How many employees are at the formance indicators.) How often are they supervisory level? reviewed? (Possible options: yearly, quar- e. For white-collar employees: terly, monthly, weekly, daily, or never). i. How much of their working day goes k. When production targets are met, what into routine tasks? percent of workers receives performance ii. How much into creative thinking bonuses? (understood as developing, designing, l. Does the firm use data to support decision or creating new applications, ideas, making? What type of data? relationships, systems, or products)? iii. How much into making decisions and 7. Innovation capacity solving problems (understood as ana- a. What are the past and future investments in lyzing information and evaluating R&D? results to choose the best solutions and b. What are the areas of innovation/economic solving problems)? upgrading in which the firm is engaged/ iv. How much into unanticipated plans to engage? (Possible options: busi- situations? ness processes, functions, products, or new v. How much into communicating inside sectors.) the organization versus outside the c. What is the skill upgrading potential within organization versus working with the firm? computers? d. What is the current and projected use of f. For supervisors and top management, ICT (and results)? how much of their working day goes into e. What are the requirements of new or exter- creative thinking, making decisions and nal (within and outside the country) exper- solving problems, and unanticipated situ- tise to support innovation? ations? How much into communicating f. What are the firm’s strategies to import inside the organization versus outside the technology and know-how and learn new organization? How much into working with ways of doing business? computers? g. What are the strategic partnerships for g. For blue-collar workers, white-collar work- innovation? ers, supervisors, and managers, how much h. What are the links between innovation and on-the-job training is required after hiring workers/wages? What institutional sup- (none, short demonstration, 1 month, 3 port is required to accelerate the gains from months, 6 months, 1 year, 2 years, 7 years, these links? 10 years, or more)? i. What are the risks in innovation? h. Do you face a shortage of skills? For which j. What is the firm doing/planning to do to tasks in particular? facilitate the transfer of knowledge or other i. To what extent is it possible for workers at tangible or intangible assets from the lead your firm to do the work on their job with- firm/global buyer? Options include no role; out being physically present (all the work, joint research projects; joint training pro- most of the work, some of the work, or none grams for blue collar workers/white col- at all)? What tasks, if any, can be carried out lar workers/managers; technical assistance from a remote location (via computer or from the buyer/lead firm to the suppliers; telephone)? What tasks require face-to-face assistance in product design and develop- physical presence with the other co-workers, ment for suppliers; equipment and special- machinery, or locally installed software? ized machinery; assistance in the design Designing a Country Engagement Strategy Based on Sound Analytics 227 of equipment and specialized machinery; integration in GVCs, and growth in the sec- finance for investment in technology and/or tor (barrier, enabler, or accelerator) innovation from global buyers/lead firms to b. Sufficiency of services in performance, suppliers. access/availability, and resilience, and sug- gested improvements c. Local, regional, and global connectivity of Firm and Public Sector Entity Interviews the domestic country’s infrastructure and 8. Role of exports, imports, foreign direct invest- services ment (FDI), and other non-equity types of real d. Inter-sector use of the domestic coun- (non-financial) investment in the country try’s infrastructure and services (and a. Overall and sector trends (temporal, spa- implications) tial) as they are perceived by the interviewed stakeholders 10. Skills and innovation capacity in the country b. Drivers of trade, FDI, or other non-equity a. Are shortages of skilled workers perceived? real investment, today and projected (polit- b. Does the education system produce skills ical, economic, or social) matching the demand for labor? c. Impact of regional and global competition c. How is vocational training organized? on local firms’ productivity/performance d. What are the past and future investments in (within sector and substitute sectors) R&D in the country/region? d. Recommendations for improvement e. What are the areas of innovation/economic (within firms, sector, and government) upgrading in which the firm is engaged/ e. Strategy to maximize benefits from open plans to engage? (Possible options: business trade environment, FDI, or other non- processes, functions, products, and new equity forms of investment sectors.) f. What are the national strategies to support f. What is the skill upgrading potential within firms’ ability to import technology and the firm? know-how and learn new ways of doing g. What is the current and projected use of business? ICT (and results)? g. What are the strategic partnerships for h. What are the requirements of new or exter- innovation in which the country engages/ nal (within and outside of country) exper- should engage? tise to support innovation? h. What are the links between innovation and i. What are the public strategies to import workers/wages? What institutional sup- technology and know-how and firms in port is required to accelerate the gains from their efforts to learn new ways of doing these links? business? j. What are the strategic partnerships for 9. Quality and competence of the domestic coun- innovation? try’s infrastructure and services k. What are the links between innovation and a. Role of the domestic country’s infrastruc- workers/wages? What institutional sup- ture and services (not just logistics/trans- port is required to accelerate the gains from portation, but also business and technical these links? services) supporting firm entry and deeper l. What are the risks in innovation? 228 Making Global Value Chains Work for Development Annex 11B Checklist of Topics from e. Identify the potential for employment Combined Desk Research and Fieldwork – Which sectors, GVCs, and segments within GVCs have high levels of The aim of the field interviews is to confirm and employment or potential employment? deepen knowledge on the points below (which will Are these low-skilled or high-skilled? Is have already been partly addressed via the desk anal- the average wage above or below that in ysis). These include the following. other countries, in comparable tasks? f. Identify sectors, GVCs, and segments within 1. General overview GVCs where the target country has a plau- a. What is [INSERT COUNTRY NAME]’s sible chance of success position in global value chains (GVCs)? – Which sectors, GVCs, and segments b. What can [INSERT COUNTRY NAME] do within GVCs have a current, nascent, or now to ensure that it increases its share of potential presence in the target country, domestic value added in these GVCs? including foreign direct investment (FDI), c. How can [INSERT COUNTRY NAME] local firms, exports, and employment? leverage its participation in GVCs to increase productivity and innovation? 3. Analysis of global trends d. What are best-practice examples, but also a. Product mapping negative lessons that [INSERT COUNTRY – What are the main final and intermediate NAME] can draw from? goods that comprise the sectors, GVCs, e. Chain upgrading: what is the transferability and segments within the GVCs? of skills across sectors? What key skills (such – What are the main activities in the as skills related to information and com- sectors, GVCs, and segments within the munications technology) are transferable GVCs? to other sectors, including soft skills related – What comprises the chain of value- to functional upgrading (logistics coordina- adding activities in the main product tion, services, design, branding, and prod- areas, and in the case study products in uct development)? particular? b. Key technology trends 2. Drilling down into specific sectors, GVCs, and – What is the impact of technology on market segments within GVCs the sectors, GVCs, and segments within a. Value chain mapping the GVCs, in effects on produced goods – What are the main sectors, product sets, and services, processes, logistics, and clusters, and markets that comprise automation of business functions? the industry in which the country – How is technology flowing to domestic participates? actors in the sectors, GVCs, and segments b. Identify growth sectors, GVCs, and seg- within the GVCs? If it does not, what are ments within GVCs the key bottlenecks? – Which sectors, GVCs, and segments – How is technology impacting the within GVCs have high growth rates and organization of work within sectors, high growth potential? GVCs, and segments within the GVCs? c. Identify the potential for technological Are these impacts aligned to global learning trends, or do they reflect local challenges – What sectors, GVCs, and segments and specificities? within GVCs have high potential for c. Identification of GVC actors technological learning? – Which firms and organizations “drive” d. Identify potential for specialization within the sector as big buyers, market share GVCs leaders, foreign investors, owners of – What sectors, GVCs, and segments within key technologies and other intellectual GVCs have the potential for countries property, and initiate or “lead” the and regions to play specialized roles? process of global engagement (sourcing, Designing a Country Engagement Strategy Based on Sound Analytics 229 FDI, contracting out, licensing, – Role of information technology franchising, etc.)? – New markets – Have specific processes in the sectors, – New product areas GVCs, and segments within the GVCs – New standards, including labor and been outsourced and offshored (for environmental standards example, final assembly or call centers)? – Changes in sector structure (vertical/ – Has a set of large supplier firms grown horizontal integration/disintegration) on the basis of large-scale outsourcing – Market share changes (consolidation or and/or offshoring? How have business fragmentation) models in the sector changed over the – Regulations past 20 years? – Resource constraints – Are there key service providers (such as logistics), equipment vendors, or 4. Identification of key global actors (retailers, component suppliers that strongly lead firms and global buyers, intermediaries, influence the development of the sector key equipment and service providers, distri- as “platform leaders”? butors, central exchanges, standard-setting – Are there “intermediaries” in the GVC, bodies, etc.) such as commodity exchanges, powerful a. Discover how the target country, sector, and trading companies, or large processors cases are perceived by key sector actors that influence international prices and b. Discover how the target country, sector, and standards? cases fit (or do not fit) into the past, present, – Are there institutional actors that have and future sourcing strategies of powerful buy- influenced the structure, location, and ers and lead firms in the sector growth of the global sector, including governments, multilateral agencies, non- 5. Evaluation of the domestic sector and its links governmental organizations, certifica- to GVCs tion and compliance bodies, and sector a. Which of the actors identified above are standard-setting bodies? present in the country? d. Transaction mapping b. Does the domestic sector play a specific role – How is information exchanged across within GVCs? key activities (arms-length, codified, or c. Quantify the domestic sector in terms of: tacit)? – Number and size of firms – Do certain activities need to be – GVC roles played by local firms and co-located? foreign affiliates – Can other activities be accomplished at a – Employment and wages distance? – Trade (imports and exports) – What role does technology play in the – Productivity, value added, etc. coordination of value chain activities? d. Identify and evaluate the effects of gov- – How has the global sector changed over ernment programs and regulations on the the past 20 years? sector. – Outsourcing and offshoring APPENDIXES APPENDIX A DIMENSIONS OF GVC PARTICIPATION: A TENTATIVE CHECKLIST Global value chains (GVCs) are a multidimensional performance as a seller in GVCs may be associated phenomenon that involves flows of goods and ser- with sellers that are also owners of GVC assets, tech- vices (discussed at length in part II of this book). nology, and know-how. High wages are likely to indi- Flows of factors of production (workers, ideas, and cate that the buyer is also a producer, close to final investment) are also important, as shown in chap- demand, and able to generate high value added. ter 1. Table A.1 reports a wide set of examples and Meanwhile, low wages are likely to be predominant in measures that can be analyzed to complement the buyers that are mainly assemblers or involved in activ- task-based and value-added data assessments dis- ities with little transformation. Indicators on flows cussed in part II of the book. of international patents, foreign technology licenses, A few examples demonstrate how to use the table. royalties, and fee services are also important indica- In the “workers” column, the second block of rows tors of countries’ specialization in higher or lower (specialization/value addition) in the table indi- value-added activities, likely position in the GVC net- cates that high wages in a country that shows strong work, and other aspects of GVC participation. 233 234 Table A.1. A Multidimensional Checklist of a Country’s Participation in GVCs Tasks Factors of production GVC component Goods Services Workers Ideas and know-how Investment Role of domestic firms and FDI hosted in the domestic territory in GVCs Buyer Any buyer High share of imported High share of wholesale Depends on the type of buyer. Buyer of licenses and Dominance of inward FDI intermediates from countries service imports, and transport Low-skilled, manufacturing other NEM agreements for and NEMs. FDI/NEMs in that have a comparative and logistics service jobs and shortage of nonstrategic assets. Direct sectors of specialization advantage in the same production and/or imports. management skills for buyers capital control (equity stakes) and with partners relevant intermediates or immediate specializing in low-value- for strategic assets. to GVC trade. FDI versus upstream or downstream added activities. High-skilled Technical cooperation and NEMs measures may sectors. High foreign and service jobs for buyers arm’s length trade may indicate strategic importance Making Global Value Chains Work for Development value-added content in own specializing in high-value- signal either looser forms of assets, degree of exports. added tasks. of collaboration or less collaboration, or hierarchical hierarchical power structure power structure in GVCs (see in GVCs. ideas column). Buyer of high-value-added High share of imported Imports IT and BPO services High wages (relative to Domestic companies tend to Investment in knowledge tasks (for example, final intermediates or unbranded from countries with supplier, and often absolutely) maintain high-value-added creation. Outbound producer, brand owner, or finished products from lower costs and/or other because of higher value-added activities (such as R&D or FDI/NEMs (often cost high-end retailer) countries with lower costs comparative advantages. activities. design) in-house. Potential or resource based) in and/or other comparative Exports final services to purchasing of international countries with costs or other advantages. Exports final consumers abroad. patents or licensing for parts comparative advantages for product to consumers abroad. of the product or technology. the production of intermediate Specializes in postproduction Sizable revenues through inputs. FDI versus NEMs services. trademarks, royalties, license measures may indicate fees, and so forth. strategic importance of assets. Seller Produces and exports a high Exports IT and BPO services Wages (relative to final Purchasing international Larger share of outward share of intermediates to that often have a lower value producer, and often patents or licensing (for investment than in buyer countries that buy imported added. absolutely) are a priori example, of technology), also countries. Investment in scale intermediates. indeterminate: low/high from within the multinational production. Inbound FDI/NEMs depending on position in GVC. group (lead firm or other (often cost or resource based) Dependence on expatriates entities) if supplier is a foreign mainly in sectors of high GVC (managers, technicians) if affiliate. participation. supplier is foreign affiliate of final producer firm. (Table continues next page) Table A.1. (continued) Tasks Factors of production GVC component Goods Services Workers Ideas and know-how Investment Specialization/value addition Mass or primary production/ Low value added from Importance of logistics High dependence on low-skill Licensing (for example, of Investment in scale assembly activities assembly activities and mass performance. High imports workers (including immigrants technology); low patent production. Inbound FDI/NEMs or primary production destined of logistics services. Possible from other low-wage applications. Low rates of (often cost or resource based) to export markets. Buyer-side exports of wholesale services; countries). private R&D activity. from countries that specialize integration higher than transportation high. in preproduction. seller-side integration. Specialization in high- Strong structural integration High value added High wages for researchers, High R&D intensity and Investment in knowledge value-added support and in GVCs, particularly on the from production and designers, and specialized innovation. High returns creation. Outbound market- knowledge activities (R&D, seller side. exports of R&D, design, personnel (domestic and from from royalties and fees. oriented FDI/NEMs in technological development, commercialization services, abroad). High services content in countries that specialize in or specialized services) and so forth. High exports of manufacturing. High domestic postproduction. royalties. content of value added in exports. Specialization in medium- High value added from High wages for specialized Investment in distribution value-added support production and exports of workers (domestic and from channels. Inward market- activities and sales branding, logistics, after-sales abroad), particularly mid-skill oriented FDI/NEMs from services, and so on. workers. countries that specialize in preproduction. Outward market-based FDI/NEMs to support sales abroad. Type of GVC node Headquarter node High share of imported value High share of imported High share of skilled workers. High private R&D intensity; Investment in knowledge added used in production of services value added used in high rates of patents and creation. Outbound FDI/NEMs output. production of output. patent applications by private (often cost or resource based) sector. in factory nodes. Factory node High share of intermediate High share of intermediate Low share of skilled workers. Lower private R&D intensity Investment in scale goods exported to headquarter services exported to than headquarters. Lower than production. Inbound FDI/NEMs node. headquarter node. headquarter rates of private (often cost or resource based) patent applications. from headquarter node. (Table continues next page) Appendix A: Dimensions of GVC Participation: A Tentative Checklist 235 236 Table A.1. (continued) Tasks Factors of production GVC component Goods Services Workers Ideas and know-how Investment Position in the GVC network Incoming spoke High share of intermediate High share of two-way trade Lower wages (relative to Purchasing international Investment in scale goods exported to hubs or in business, transport, and hub, and often absolutely). patents or licensing (for production. Inbound FDI/NEMs central nodes, as identified by distribution services with hubs Dependence on expatriates example, of technology), also (often cost or resource based), Making Global Value Chains Work for Development network metrics. or central nodes, as identified (managers, technicians) if from hub or central node. High mainly in sectors of high GVC by network metrics. incoming spoke is foreign imports of royalties and fees. participation. affiliate of hub or central node. Hub High share of imported value High share of services High wages (relative to Maintains high-value-added Investment in knowledge added used in production value added in exports of spokes, and often absolutely) activities (such as R&D or creation. Outbound FDI/NEMs of output for consumption manufacturing. Good logistics because of higher value-added design) in-house. Potential (often cost or resource based) abroad. performance. activities. Expatriates purchasing of international from hubs to incoming spokes (managers, technicians) if hub patents or licensing. Licensing in upstream sectors. has outward FDI. of technology to incoming spoke. Outgoing spoke High share of final goods High share of wholesale Distribution agreements and Market-based inbound FDI/ imports from hubs or services production or trade. licenses. High imports of NEMs from hub or central central nodes for domestic May act as a regional hub for royalties and fees. node to support sales. consumption or distribution distribution services. to more peripheral locations. May act as a regional hub. Note: BPO = business process outsourcing; FDI = foreign direct investment; GVC = global value chain; IT = information technology; NEMs = non-equity modes of investment; R&D = research and development. APPENDIX B BROAD ECONOMIC CATEGORIES CLASSIFICATION The Broad Economic Categories (BEC) classification, be done. In sum, all the items indicated in bold are as defined by the United Nations (UN), comprises classified as intermediate (see table B.1). Concordance 19 basic categories that are assigned to the final use tables to match the BEC categories to trade data are of the good—capital, consumption, and intermedi- widely available. One common source in the UN ate. Three categories—motor spirit, passenger motor website, which reports concordance tables between cars, and goods not elsewhere specified—are not BEC and the Harmonized System classifications and assigned to any of those 19 categories. The authors Standard International Trade Classification classifica- suggest classifying motor spirit as intermediate goods tions, is http://unstats.un.org/unsd/cr/registry/regot and passenger motorcars as consumption goods. The .asp?Lg=1. assignment of goods not specified elsewhere cannot Table B.1. BEC Classification Broad economic category Final use 1 Food and beverages 11 Primary 111 Mainly for industry Intermediate goods 112 Mainly for household consumption Consumption goods 12 Processed 121 Mainly for industry Intermediate goods 122 Mainly for household consumption Consumption goods 2 Industrial supplies not elsewhere specified 21 Primary Intermediate goods 22 Processed Intermediate goods 3 Fuels and lubricants 31 Primary Intermediate goods 32 Processed 321 Motor spirit Intermediate and consumption goods 322 Other Intermediate goods 4 Capital goods (excluding transport equipment) 41 Capital goods Capital goods 42 Parts and accessories Intermediate goods 5 Transport equipment 51 Passenger motor cars Intermediate and consumption goods 52 Other 521 Industrial Capital goods 522 Nonindustrial Consumption goods 53 Parts and accessories Intermediate goods 6 Consumer goods not elsewhere specified 61 Durable Consumption goods 62 Semi-durable Consumption goods 63 Nondurable Consumption goods 7 Goods not elsewhere specified Intermediate, consumption, and capital goods Source: Based on UN 2002. Note: BEC = Broad Economic Category. 237 APPENDIX C CUSTOMIZED VERSUS GENERIC INTERMEDIATES Sturgeon and Memedovic (2011) differentiate • Generic intermediates are products that are likely between customized and generic intermediates in to be used in a wide range of final products, as three sectors: apparel and footwear, electronics, and well as in products made in large, standardized passenger vehicles. batches and in continuous-process production methods. • Customized intermediates are products that are likely to be used in specific final products or at Table C.1 exemplifies the range of customized inter- least relatively narrow classes of products. mediates in the apparel and footwear sector (excerpt from original table). Table C.1. Customized Intermediates in the Apparel and Footwear Sector (Excerpt) BEC SITC SITC description 22 65225 Other woven fabrics of cotton, unbleached, weight < 200 g/m2 22 6536 Fabrics, woven, containing 85% or more by weight of artificial staple fibers 22 65112 Yarn of carded wool, containing 85% or more by weight of wool, not put up for retail sale 22 65113 Yarn of combed wool, containing 85% or more by weight of wool, not put up for retail sale 22 65114 Yarn of fine animal hair (carded or combed), not put up for retail sale 22 65115 Yarn of coarse animal hair or of horsehair (including gimped horsehair yarn), whether or not put up for retail sale 22 65117 Yarn of carded wool, containing less than 85% by weight of wool, not put up for retail sale 22 65118 Yarn of combed wool, containing less than 85% by weight of wool, not put up for retail sale 22 65121 Cotton sewing thread, not put up for retail sale 22 65133 Cotton yarn (other than sewing thread), containing 85% or more by weight of cotton, not put up for retail sale 22 65134 Cotton yarn (other than sewing thread), containing less than 85% by weight of cotton, not put up for retail sale 22 65141 Sewing thread of synthetic filaments, whether or not put up for retail sale 22 65142 Sewing thread of artificial filaments, whether or not put up for retail sale 22 65143 Sewing thread of synthetic staple fibers, whether or not put up for retail sale 22 65144 Sewing thread of artificial staple fibers, whether or not put up for retail sale 22 65151 Filament yarn (other than sewing thread), of nylon or other polyamides, not put up for retail sale Source: Sturgeon and Memedovic 2011, 30. Note: BEC = Broad Economic Classification; g = gram; m = meter; SITC = Standard International Trade Classification. 239 APPENDIX D PARTS AND COMPONENTS Athukorala (2010) mapped parts and compo- conducted in Malaysia and Thailand. Data compiled nents for manufacturing sectors in the United at the HS 6-digit level were converted to the Standard Nations (UN) Broad Economic Categories (BEC) International Trade Classification (SITC) for the registry in the product list of the World Trade final analysis, using the UN HS-SITC concordance. Organization Information Technology Agreement For an illustration of 20 (of a list of 525) manufac- with the Harmonized System (HS) of trade classi- turing parts and components at the HS 6-digit level, fication at the 6-digit level. Existing gaps in the list see table D.1. were filled with estimates from firm-level surveys Table D.1. List of Manufacturing Parts and Components (Excerpt) Nomenclature Plates, sheets, etc. nesoi, cellular polyurethanes Plates, sheets, etc. nesoi, cellular plastic nesoi Chemical elements doped, used in electronics, discs, wafers, etc. Articles of leather used in machinery/mechanical appliance Pipe, reinforced/combine w/metal only, w/o fitting Pipe, reinforced/combine w/ textiles, w/o fitting Pipe, reinforced/combine w/ material, w/o fitting Tubes, pipe, etc., vulcanized soft rubber, with fitting Endless transmission belt, trapezoidal, circumference > 60 cm, < 180 cm Endless transmission belt, circumference > 180 cm, < 240 cm Conveyor belts or belting reinforced with metal Conveyor belts reinforced with textile materials Conveyor belts reinforced only with plastics Conveyor belts/belting of vulcanized rubber nesoi Endless synchronous belt, circumference > 60 cm, < 150 cm Endless synchronous belt, circumference > 150 cm, < 198 cm Transmission belt/belting, of vulcanized rubber, nesoi Articles of soft vulcanized rubber nesoi Gasket, washers, & other seals, of vulcanized rubber Textile labels, badges, etc., not embroidered, woven Source: Athukorala 2010, 19. Note: cm = centimeter; nesoi = not elsewhere specified or included. 241 APPENDIX E VALUE CHAIN CATEGORIES The classification by Taymaz, Voylvoda, and Yilmaz table E.1. This system has at least two downsides. (2011) carefully assigns exports (categorized at the First, being a classification that covers only goods 4-digit International Standard Industrial Classifi- exports, it does not identify the services segments cation code) to one of five value chain categories— of value chains, such as research and development, final products, main input/part, standard input/ design, commercialization, distribution, marketing/ track, raw material, and machinery/equipment— branding, logistics, and after-sales services, precisely based on engineering considerations. The classi- the segments that allow for functional upgrading. fication is available for five key Turkish industries Second, it does not account for the domestic dimen- (motor vehicles, TV, food, machinery, and textiles sion of value chains, thus providing only a partial and apparel). The product assignment by value chain overview of the situation. category for the five GVC sectors is reported in Table E.1. Assignment of Products to Five Value Chain Categories in Five Main GVC Sectors Final products Main input/part Standard input/track Raw material Machinery/equipment Motor vehicles 111 Auto 120 Motor 131 Auto parts 140 Flat steel 151 Already designed or engineered parts 112 Camion 132 Other components 152 Other 113 Autobus 114 Tractors TV 211 Radio 221 CRT 231 Electronic components 240 Plastic 250 Injection machines 212 TV Food 411 Meat/fish products 422 Sugar 441 Meat/fish 450 Food machines 412 Confectionery 423 Cocoa 442 Sugar beet 413 Chocolate 424 Milk powder, flour, 443 Cocoa powder etc. 414 Flour products 425 Frozen products 444 Milk, wheat, etc. 415 Canned food 426 Tea, etc. 445 Vegetables, fruit 416 Other food 427 Alcohol, vinegar 446 Coffee, soya 417 Drinks 428 Pulp, waste 447 Mineral water 418 Waste products (Table continues next page) 243 244 Making Global Value Chains Work for Development Table E.1. (continued) Machinery 511 Consumer machinery 521 Motors, turbines 531 Metal plates, pipes, etc. 541 Iron, copper, etc. 151 Already designed or engineered parts 512 Industrial machinery 152 Other Textiles and apparel Cotton 311.1 Apparel 321 Fabric 331 Yarn 341 Cotton 350 Textile machinery 311.2 Pajama, t-shirts 311.3 Sheets, etc. Wool 312.1 Apparel 322 Fabric 332 Yarn 342 Wool 350 Textile machinery 312.2 Pajama, t-shirts 312.3 Sheets, etc. 312.4 Carpets Synthetic 313.1 Apparel 323 Fabric 333 Yarn 343 Polyester 350 Textile machinery 313.2 Pajama, t-shirts 313.3 Sheets, etc. 313.4 Carpets Other 314.1 Apparel 324 Fabric 334 Yarn 344 Other 350 Textile machinery 314.2 Pajama, t-shirts 314.3 Sheets, etc. 314.4 Carpets Note: CRT = cathode ray tube. APPENDIX F SECTOR AND PRODUCT CLUSTERS Sectors are a statistics artifact. The same task can be presents an attempt, by the ARD Vest Development deployed in different sectors (for example, threading Agency in Romania, to identify clusters of activities and sewing expertise can be deployed in the textiles needed in agrifood, construction, energy, health, and apparel sector and in the auto sector), and suc- information and communications technology, tex- cess in a given sector may depend on comparative tiles and apparel, and tourism. To some degree, the advantages in tasks that do not belong to that sector detailed activities included in each cluster may be according to official statistics. Services inputs, such as influenced by locational factors, so that application transport and finance, are important for the compet- to other countries may require small modifications itiveness of many manufacturing products. Table F.1 to the compositions of the clusters. Table F.1. GVC Clusters AGRIFOOD NACE rev. 2 code NACE Rev. 2 description 111 Growing of cereals (except rice), leguminous crops, and oil seeds [01.11] 112 Growing of rice [01.12] 113 Growing of vegetables and melons, roots and tubers [01.13] 114 Growing of sugar cane [01.14] 115 Growing of tobacco [01.15] 116 Growing of fiber crops [01.16] 119 Growing of other non-perennial crops [01.19] 121 Growing of grapes [01.21] 122 Growing of tropical and subtropical fruits [01.22] 123 Growing of citrus fruits [01.23] 124 Growing of pome fruits and stone fruits [01.24] 125 Growing of other tree and bush fruits and nuts [01.25] 126 Growing of oleaginous fruits [01.26] 127 Growing of beverage crops [01.27] 128 Growing of spices, aromatic, drug and pharmaceutical crops [01.28] 129 Growing of other perennial crops [01.29] 130 Plant propagation [01.30] 141 Raising of dairy cattle [01.41] 142 Raising of other cattle and buffaloes [01.42] 143 Raising of horses and other equines [01.43] 144 Raising of camels and camelids [01.44] 145 Raising of sheep and goats [01.45] (Table continues next page) 245 246 Making Global Value Chains Work for Development Table F.1. (continued) AGRIFOOD NACE rev. 2 code NACE Rev. 2 description 146 Raising of swine/pigs [01.46] 147 Raising of poultry [01.47] 149 Raising of other animals [01.49] 150 Mixed farming [01.50] 161 Support activities for crop production [01.61] 162 Support activities for animal production [01.62] 163 Post-harvest crop activities [01.63] 164 Seed processing for propagation [01.64] 311 Marine fishing [03.11] 312 Freshwater fishing [03.12] 321 Marine aquaculture [03.21] 322 Freshwater aquaculture [03.22] 1011 Processing and preserving of meat [10.11] 1012 Processing and preserving of poultry meat [10.12] 1013 Production of meat and poultry meat products [10.13] 1020 Processing and preserving of fish, crustaceans, and mollusks [10.20] 1031 Processing and preserving of potatoes [10.31] 1032 Manufacture of fruit and vegetable juice [10.32] 1039 Other processing and preserving of fruit and vegetables [10.39] 1041 Manufacture of oils and fats [10.41] 1042 Manufacture of margarine and similar edible fats [10.42] 1051 Operation of dairies and cheese making [10.51] 1052 Manufacture of ice cream [10.52] 1061 Manufacture of grain mill products [10.61] 1062 Manufacture of starches and starch products [10.62] 1071 Manufacture of bread; manufacture of fresh pastry goods and cakes [10.71] 1072 Manufacture of rusks and biscuits; manufacture of preserved pastry goods and cakes [10.72] 1073 Manufacture of macaroni, noodles, couscous, and similar farinaceous products [10.73] 1081 Manufacture of sugar [10.81] 1082 Manufacture of cocoa, chocolate, and sugar confectionery [10.82] 1083 Processing of tea and coffee [10.83] 1084 Manufacture of condiments and seasonings [10.84] 1085 Manufacture of prepared meals and dishes [10.85] 1086 Manufacture of homogenized food preparations and dietetic food [10.86] 1089 Manufacture of other food products nec [10.89] 1091 Manufacture of prepared feeds for farm animals [10.91] 1092 Manufacture of prepared pet foods [10.92] 1101 Distilling, rectifying, and blending of spirits [11.01] 1102 Manufacture of wine from grape [11.02] 1103 Manufacture of cider and other fruit wines [11.03] 1104 Manufacture of other non-distilled fermented beverages [11.04] 1105 Manufacture of beer [11.05] 1106 Manufacture of malt [11.06] 1107 Manufacture of soft drinks; production of mineral waters and other bottled waters [11.07] CONSTRUCTION NACE 2 NACE 2 description 4110 Development of building projects 4120 Construction of residential and non-residential buildings 4200 Civil engineering (Table continues next page) Appendix F: Sector and Product Clusters 247 Table F.1. (continued) CONSTRUCTION NACE 2 NACE 2 description 4211 Construction of roads and motorways 4212 Construction of railways and underground railways 4213 Construction of bridges and tunnels 4221 Construction of utility projects for fluids 4222 Construction of utility projects for electricity and telecommunications 4291 Construction of water projects 4299 Construction of other civil engineering projects nec 4311 Demolition 4312 Site preparation 4313 Test drilling and boring 4321 Electrical installation 4322 Plumbing, heat, and air conditioning installation 4329 Other construction installation 4331 Plastering 4332 Joinery installation 4333 Floor and wall covering 4334 Painting and glazing 4339 Other building completion and finishing 4391 Roofing activities 4399 Other specialized construction activities nec ENERGY NACE 2 NACE 2 description 3511 Production of electricity 3512 Transmission of electricity 3513 Distribution of electricity 3514 Trade of electricity 3521 Manufacture of gas 3522 Distribution of gaseous fuels through mains 3523 Trade of gas through mains 3530 Steam and air conditioning supply HEALTH NACE 2 NACE 2 description 8610 Hospital activities 8621 General medical practice activities 8622 Specialist medical practice activities 8623 Dental practice activities 8690 Other human health activities 8710 Residential nursing care activities 8730 Residential care activities for the elderly and disabled 8790 Other residential care activities 8810 Social work activities without accommodation for the elderly and disabled 8891 Child day-care activities 8899 Other social work activities without accommodation nec INFORMATION AND COMMUNICATIONS TECHNOLOGY (ICT) NACE 2 NACE Rev. 2 description 2611 Manufacture of electronic components [26.11] 2612 Manufacture of loaded electronic boards [26.12] 2620 Manufacture of computers and peripheral equipment [26.20] (Table continues next page) 248 Making Global Value Chains Work for Development Table F.1. (continued) INFORMATION AND COMMUNICATIONS TECHNOLOGY (ICT) NACE 2 NACE Rev. 2 description 2630 Manufacture of communication equipment [26.30] 2640 Manufacture of consumer electronics [26.40] 2680 Manufacture of magnetic and optical media [26.80] 4742 Retail sale of telecommunications equipment in specialized stores [47.42] 4743 Retail sale of audio and video equipment in specialized stores [47.43] 5821 Publishing of computer games [58.21] 5829 Other software publishing [58.29] 6110 Wired telecommunications activities [61.10] 6120 Wireless telecommunications activities [61.20] 6201 Computer programming activities [62.01] 6202 Computer consultancy activities [62.02] 6203 Computer facilities management activities [62.03] 6209 Other information technology and computer service activities [62.09] 6311 Data processing, hosting, and related activities [63.11] 6312 Web portals [63.12] 6391 News agency activities [63.91] 6399 Other information service activities nec [63.99] 9511 Repair of computers and peripheral equipment [95.11] 9512 Repair of communication equipment [95.12] TEXTILES AND APPAREL NACE 2 NACE Rev. 2 description 1310 Preparation and spinning of textile fibers [13.10] 1320 Weaving of textiles [13.20] 1330 Finishing of textiles [13.30] 1391 Manufacture of knitted and crocheted fabrics [13.91] 1393 Manufacture of carpets and rugs [13.93] 1394 Manufacture of cordage, rope, twine, and netting [13.94] 1395 Manufacture of non-wovens and articles made from non-wovens, except apparel [13.95] 1396 Manufacture of other technical and industrial textiles [13.96] 1399 Manufacture of other textiles nec [13.99] 1411 Manufacture of leather clothes [14.11] 1412 Manufacture of workwear [14.12] 1413 Manufacture of other outerwear [14.13] 1414 Manufacture of underwear [14.14] 1419 Manufacture of other wearing apparel and accessories [14.19] 1420 Manufacture of articles of fur [14.20] 1431 Manufacture of knitted and crocheted hosiery [14.31] 1439 Manufacture of other knitted and crocheted apparel [14.39] 1511 Tanning and dressing of leather; dressing and dyeing of fur [15.11] 1512 Manufacture of luggage, handbags and the like, saddler, and harness [15.12] 1520 Manufacture of footwear [15.20] TOURISM NACE 2 NACE 2 description 5510 Hotels and similar accommodation 5520 Holiday and other short-stay accommodation 5530 Camping grounds, recreational vehicle parks and trailer parks 5590 Other accommodation 5610 Restaurants and mobile food service activities (Table continues next page) Appendix F: Sector and Product Clusters 249 Table F.1. (continued) TOURISM NACE 2 NACE 2 description 5621 Event catering activities 5629 Other food service activities 5630 Beverage serving activities 7911 Travel agency activities 7912 Tour operator activities 7990 Other reservation service and related activities 9321 Activities of amusement parks and theme parks 9329 Other amusement and recreation activities Source: Strategic cluster definition by ARD Vest, Romania. Note: GVC = global value chain; NACE = General Industrial Classification of Economic Activities within the European Communities; nec = not elsewhere classified. APPENDIX G MAIN TYPES OF DATA USED TO MEASURE GVC PARTICIPATION This appendix describes the five main types of data more carefully about this topic, several organizations used to measure global value chain (GVC) partici- recently released new international input-output pation, as well as their capabilities and limitations. (I-O) tables. Box G.1 gives an overview of the avail- These data help in assessing a country’s participation able international I-O data. in GVCs along various dimensions. International I-O data allow the assessment of how primary inputs (workers and capital) in a country are used and how much income they gener- 1. Production Data and Gross Trade Data ate through GVC participation. The key concept is Production data and gross trade data measure the “value-added” versus “gross” exports. Gross exports amount of goods and services produced and traded, are the traditional measure that has been used for respectively. The data do not indicate the domestic or decades—the value of goods when they leave the foreign source of the inputs or the value addition gen- country. Value-added exports strip out the value that erated in a country. Production data from international was added in some other country. Therefore, they sources provide a limited breakdown in industries. enable the identification of the ultimate source or Firm-level sources (for example, customs-level trade destination of the value added that is generated by data) tend to provide a finer disaggregation, as do sec- eliminating double counting in gross export data.1 tor trade data. These data, categorized using informed Gross and value-added exports are important indi- classifications (Broad Economic Classification, parts cators. For instance, most trade policy is applied to and components, or technical classifications) or clas- gross trade, but the number of jobs linked to exports sified into product and sector clusters allow inves- depends on value-added exports. tigation of specific aspects of GVC participation. A Although value-added trade data provide essen- discussion of the main classifications that can be used tial information about GVCs, the data are subject to is provided in appendixes B through F. restrictive assumptions that may significantly bias the quantification of technology coefficients for dif- ferent types of production.2 Being based on non- 2. Value-Added Data harmonized I-O tables from national sources, any In the GVC context, a country’s exports cannot be errors in an I-O national table will produce errors in competitive if the country does not import parts all the value-added trade flows. and components from the most competitive sup- Several organizations have started to produce pliers. Put simply, imports are functionally linked international I-O tables and measures derived from to exports. In this new GVC world, exports are not them, such as the World Input-Output Database, all equal from a national perspective. A million dol- the Organisation for Economic Co-operation and lars of exports may create few jobs or many, and it Development–World Trade Organization Trade in may have few links with the rest of the economy Value Added database, the United Nations Confer- or many. To help governments and scholars think ence on Trade and Development (UNCTAD)–EORA 251 252 Making Global Value Chains Work for Development Box G.1. Major Input-Output Databases Table BG.1.1. International Input-Output Databases Countries/ Sectors/ Database Data source regions products Years WIOD http://www.wiod.org/ 40 35 1995–2011 OECD-WTO TiVA database http://www.oecd.org/ 61 34 1995, 2000, 2005, http://www.wto.org/ 2008–2011 EORA MRIO database http://www.worldmrio.com/ 187 26 1990–2012 Asian Development Bank MRIO Available through WIOD database 18 35 2000, 2005–2008, 2011 database IDE JETRO, Asian International http://www.ide.go.jp 10 78 1985, 1990, 1995, 2000, Input-Output Tables 2005 EXIOPOL Multi-regional Database http://www.exiobase.eu/ 43 200 2007 World Bank Export of Value Added http://data.worldbank.org/data-catalog/export 120 27 1997, 2001, 2004, 2007, database -value-added and http://wits.worldbank.org 2011 /datadownload.aspx World Bank LACEX http://wits.worldbank.org/datadownload.aspx 120 24 or 57 1995, 1997, 2001, 2004, Database 2007, 2011 Daudin, Rifflart, and Schweisguth From authors 66 or 113 55 1997, 2001, 2004 (2011) Johnson and Noguera (2012) From authors 94 57 2004 Koopman, Wang, and Wei (2014) From authors 26 41 2004 Note: EXIOPOL = Environmental Accounting Framework Using Externality Data and Input–Output Tools for Policy Analysis; IDE = Institute of Developing Economies; JETRO = Japan External Trade Organization; LACEX = Labor Content of Exports; MRIO: multi-regional input-output tables; OECD = Organisation for Economic Co-operation and Development; TiVA = Trade in Value Added; WIOD = World Input-Output Database; WTO = World Trade Organization. WIOD providing data based on a bilateral trade database and end- The World Input-Output Database (WIOD), compiled by a con- use categories as well as several derived statistical measures. sortium of 11 institutions, was funded by the European Commis- The 2015 edition includes 61 economies,b plus a “the rest of sion from 2009 to 2012. Based on supply-use tables from official the world” region, with a breakdown into 37 industries. The I-O national statistics, WIOD identifies the I-O links between 27 Euro- tables cover the years 1995, 2000, 2005, and 2008–11, but the pean Union members and 13 other major economies (plus “the goal is to produce a complete time series and extend the country rest of the world”). The database covers 35 industries and all coverage.c years from 1995 to 2011.a WIOD was the first freely available I-O EORA MRIO Database database and one of the highest quality sources; it is available to The EORA global multi-region I-O (MRIO) tables, produced by download at www.wiod.org. the University of Sydney and funded by the Australian Research ADB-MRIO and IDE JETRO Database Council, bring together a variety of primary data sources that are The Asian Development Bank multi-regional input-output tables combined in a single data set that uses interpolation and esti- (ADB-MRIO) complements the WIOD, since it has included addi- mation techniques to provide a contiguous, continuous data set tional Asian countries to WIOD to facilitate analysis related to the for 1990–2012 for 187 countries. Because the project focuses Asia and Pacific Region. Five Asian countries (Bangladesh, Malay- on environmental issues, it displays important deviations from sia, the Philippines, Thailand, and Vietnam) have been added for observed trade flows and gross domestic product, and the tables the years 2000, 2005–08, and 2011. are balanced to match, principally, data from large economies. The Institute of Developing Economies, Japan External Trade The EORA MRIO was used by UNCTAD to produce trade in Organization (IDE JETRO) database is the ancestor of all inter- value-added indicators. Because standard I-O calculations require national input-output tables. Since 1975, IDE-JETRO has been a balanced table, the EORA database contains only 26 harmonized producing international input-output tables with a focus on Asia- sectors. Although it has the most extensive country coverage of Pacific countries. the databases discussed, I-O tables for many countries in EORA OECD-WTO TiVA Database are not available and have been estimated from the United Nations The Organisation for Economic Co-operation and Development System of National Accounts and other, more aggregated data.d (OECD) Inter-Country I-O tables represent a similar effort to WIOD, (Box continues next page) Appendix G: Main Types of Data Used to Measure GVC Participation 253 Box G.1. (continued) World Bank Export of Value Added Database and GTAP exports. The labor value added is further decomposed between Input-Output Tables skilled and unskilled labor. The database covers the years 1995, The World Bank Export of Value Added database is based on the 1997, 2001, 2004, 2007, and 2011 for 24 or 57 sectors. input-output data from the Global Trade Analysis Project (GTAP), Other Databases which uses alternative sources and estimates missing data. The Other GTAP-based global value chain studies include Daudin and GTAP I-O data include social accounting matrices that cover addi- others (2006), Johnson and Noguera (2012), and Koopman and tional countries where no official I-O tables are available. The 2001, others (2011, 2014). In general, those studies have not included 2004, 2007, and 2011 GTAP data cover 57 sectors. The 2007 and the full dimensionality of the GTAP data set. 2011 GTAP data cover 129 countries or regions, whereas the 2001 Finally, the Environmental Accounting Framework Using Exter- and 2004 data cover fewer countries. Developed by Francois, Man- nality Data and Input–Output Tools for Policy Analysis (EXIOPOL) chin, and Tomberger (2013), the World Bank Export of Value Added provides international input-output tables (EXIOBASE) that can be data set allows for exploiting social accounting matrices spanning used for analysis of the environmental impacts associated with the intermittent years from 1992 to 2011, to construct country-specific final consumption of product groups. EXIOBASE covers 43 coun- measures of the direct and indirect contribution of goods and ser- tries, five world regions, and 200 products in 163 industries. It also vices to the value added contained in a given country’s domestic provides information on 15 land use types, employment per three production and exports. The Export of Value Added database cov- skill levels, 48 types of raw materials, and 172 types of water uses. ers 27 sectors, 120 countries, and the years 1997, 2001, 2004, 2007, and 2011. (See appendix H for further discussion.)e World Bank LACEX Database a. For more detailed information, see Timmer (2012). The World Bank Labor Content of Export (LACEX) Database was b. For better estimation of the value added in Chinese production and trade, the database includes additional I-O tables for three types of economic activity: developed based on input-output data from GTAP (Calì and others processing trade, ordinary trade, and domestic-only enterprises. forthcoming). The database is similar to the World Bank Export of c. For more information on the international I-O tables and TiVA measures, see Value Added database, except it looks at the value added remu- OECD and WTO (2012). nerated to labor (instead of total value added). As such, it allows d. Lenzen and others (2013) describe the construction of the underlying EORA database, and UNCTAD (2013a, 2013b) describes the related UNCTAD-EORA the user to compare the direct and indirect contributions of labor database. in different sectors (goods and services) of an economy with the e. Tsigas, Wang, and Gehlhar (2012) provide a detailed example of how to value added contained in the economy’s domestic production and construct an international I-O table from GTAP. database, and the World Bank.3 Table BG.1.1 summa- 4. Re-import and Re-export Data rizes and describes in more detail the major interna- tional I-O databases. In GVC analysis, re-import data concern countries’ exports of intermediate goods that are then re- imported after some processing in the partner coun- 3. Imports of Intermediates try. Re-export data track the inverse flow; they are Data on imports of intermediates indicate the reliance a mirror concept of re-import data. These data are of domestic production on imported intermediates. available at the same (fine) level of disaggrega- Combined with export data through the use of inter- tion as are gross exports data. The concepts of national I-O data, intermediate imports data allow re-importing and re-exporting in the compilation of the measurement of the import content of exports. official trade data are different. Re-exports in official The data are subject to the same caveats highlighted trade statistics are goods that have not been trans- in the section on value-added data. A different type formed since they were imported, whereas re-imports of data exists for some countries: data from special are goods that are returned without being sold.4 customs regimes for processing trade. Those data are collected in cases in which a country suspends tariffs 5. Firm-Level Data on imported intermediates if all the intermediates are used to make goods that are subsequently exported. Firm-level data capture the main actors in a value Finally, changes in I-O patterns in I-O tables can reveal chain: final producers and suppliers. Using firm- if the country is entering a new production or tech- level survey data about direct links in GVCs allows nology. However, the level of sector disaggregation is the detection of additional aspects of GVC participa- limited. tion—beyond what the aggregate data reveal. Those 254 Making Global Value Chains Work for Development aspects include the potential for gains (and also risks) domestic producers’ reliance on imported interme- from increased GVC integration. The Enterprise diates, and domestic suppliers’ share of exports. In Surveys, which are available from the World Bank— combination with data on firm-level output, value and have wide country coverage—allow for cross- added, capital stock, productivity, and employment, country comparisons. Stratified at sector, firm which are available from the World Bank Enterprise size, and sub-national geographic levels, the sur- Analysis Unit or from national sources, the surveys veys provide information on multinational sourc- allow for relating GVC participation with firm-level ing strategies (domestic versus foreign suppliers), characteristics Appendix H THE WORLD BANK EXPORT OF VALUE ADDED DATABASE The World Bank Export of Value Added database pro- Data Source vides information on the domestic value–added con- tent of domestic output and exports for 120 countries The underlying data for the World Bank Export of across 27 sectors of the economy, including 10 com- Value Added database come from the Global Trade mercial services sectors, 14 manufacturing sectors, and Analysis Project data set (GTAP, https://www.gtap 3 primary sectors, spanning intermittent years between .agecon.purdue.edu). The GTAP database is a global 1997 and 2011. Trade data usually are measured at database that measures and describes bilateral trade transaction value, which is the price actually paid or patterns, production, consumption, and interme- payable for goods and services. Thus, the transaction diate input use of goods and services. Because the value of goods and services is a gross value (or value data are bilateral and developed for various years, added plus intermediate inputs), which may overesti- they can be used to obtain data for cross-border mate or underestimate the real contribution of goods links in recent years as well as the way those links and services to trade. To overcome this shortcoming, have changed over time. The data set thus repre- a calculation of trade measured in value added, using sents an advanced input-output panel—also known data based on input-output tables, has been devel- as social accounting—of incomes and expenditures oped. The measure includes the direct value-added linked to trade from 1992 to 2011, covering not only contribution of sectors to total output and exports as key Organisation for Economic Co-operation and well as their indirect contribution through links. Development (OECD) economies, but also a wide This includes both forward links—the value-added range of developing countries divided by sectors. contribution of a particular sector as an input to other The basic structure of GTAP is explained by sectors’ output or exports—and backward links—the McDougall (2001) and McDougall and Hagemejer value-added contribution of all other sectors to a (2005). The database provides explanations of particular sector’s output or exports. Specifically, the how it was built, its underlying data structure, and forward links in the database show the importance of its data sources with each new release. The GTAP services as inputs to other sectors’ output and exports. database is produced by a consortium of institu- Services are an input for many other economic activi- tions, including the World Bank, U.S. International ties, which implies the importance of their efficient Trade Commission, World Trade Organization, delivery to increase competitiveness, although some OECD, United Nations Conference on Trade and services are also for final consumption. In sum, this Development, United Nations Food and Agriculture type of data set provides information on what the Organization, and universities and independent direct and indirect value-added contribution of goods research institutes. The project is based at Purdue and services to output and exports looks like over University. A public good, GTAP is an open-source time. As such, the database follows up on the pioneer- input that is often used for policy modeling on issues ing work of Christen, Francois, and Hoekman (2012) such as climate change, regional trade agreements, and Francois, Manchin, and Tomberger (2013). and even food security. 255 256 Making Global Value Chains Work for Development Although the GTAP database has been updated the latter. To account for this, the database also mea- continuously with each new release, the authors of sures the indirect contribution of services to value the database have maintained a consistent set of added through forward and backward links. regions, countries, and sectors to ensure strong com- Forward links represent the contribution of a spe- patibility of the data over time. cific sector as an input to other sectors’ value added. More specifically, they are the supply response of a particular sector to all other (downstream) sectors’ Why This Database? demand for more inputs. Typically, when all other The World Bank Export of Value Added database sectors are expanding, they demand more financial provides a comprehensive picture of economic rela- services. The forward link will indicate how much tionships between sectors within countries that are those sectors are using inputs from the financial sec- accounted for in national income or gross domestic tor—in other words, how much the financial sector product. The underlying theory stems from econom- output will expand. Looking again at Bangladesh, ics and is based on a “general equilibrium principle,” trade and transport services represent 28.9 percent of meaning that every income of an economy has a the value added of the domestic economy, and 24.2 counterpart in expenditure, so that all receipts and percent of the value added of exports, after consider- outlays correspond with each other. ing this sector’s contribution as inputs in other sec- The strength of this database, therefore, is in its tors’ value added. comprehensiveness in setting out the interrelation- Likewise, backward links represent the contribu- ships within an economy between links that record tion of all other sectors to a specific sector’s value intermediate input use and final demand. Importantly, added (upstream)—in other words, they represent consumption and trade patterns are also recorded in how much the demand of a particular sector will pull this database, so the value added produced in one the supply of all other sectors. For example, the con- sector can be linked in intermediate or final demand struction sector demands from other sectors cement, to the domestic consumption and external trade glass, bricks, and steel, which in turn contribute demand patterns of other sectors. Those links provide to the value added of the construction sector. In a fuller analysis than national input-output tables. Bangladesh, trade and transport services demanded The undertaking of this database eventually only 11.5 percent of the value added of the domes- enables analysis of the complex structure of the tic economy, and 4.9 percent of the value added of value-added content of final output and demand, as exports across all sectors. well as exports. The structure of value-added can be Links represent the interdependence of sectors further broken down into direct and indirect value of the economy. Industries with strong backward added. Direct value added captures the true sector- and forward links play an important role in a coun- specific value added that is generated within an try’s development strategy. A sector with strong economy and nets out domestic and foreign inputs, backward links means that an increase in the final whereas indirect value added adds to the direct mea- demand of those industries’ output will have a large sure the portion of value added of the inputs pro- effect on industries that supply inputs in the produc- duced domestically. Both types of value added can tion of those industries’ output. And a sector with then be expressed as part of the domestic output per strong forward links means that an increase in the sector, as well as a sector’s share of exports. final demand of other industries’ output will have a An economy’s value added is measured as the large effect on the industry. Naturally, strong links “net” output after adding all sectoral outputs and to export value added suggest an important role in a subtracting all intermediate inputs. The direct country’s export strategy. contribution of each sector to value added, of the From a policy point of view, when looking at ser- domestic economy as a whole or only of exports, is vices, what is of most interest is their services links. the value added sold directly to final consumers. For But the backward links of dominant manufacturing example, in Bangladesh, trade and transport services sectors to upstream services sectors can also reveal represent 10 percent of the direct value added of the the effect of manufacturing demand on the supply domestic economy and 1 percent of the direct value of services. added of exports. However, the contribution of that sector to the overall value added or export value Methodology of the Database added of the economy is much higher, because those services enter as inputs into the production function To obtain all the links of value added requires first of downstream sectors, continuing to add value to calculating how much input is contained in one unit Appendix H: The World Bank Export of Value Added Database 257 of final output. That is the intermediate multiplier database also contains figures for total final demand matrix, also commonly known as the Leontief matrix and exports taken from GTAP, we can further split (M). The M-matrix holds information on direct the value-added matrix into value-added flows that and indirect input use in a sector. Second, informa- result from domestic activities, which is the primary tion about the shares of value added in total output, component of gross national product and value- which is then consumed domestically or otherwise added flows that are recovered from exports. More exported, is organized in what we call the B-matrix. technically, we can call the former type of value added The intermediate input use shares of the M-matrix the G-matrix, whereas the latter is the H-matrix. For are multiplied by the final output and exports figures both matrices, the figures can be further divided into (B-matrix) to obtain the corresponding value-added direct value added and indirect value added. shares. Writing this more formally, a domestic economy’s Example gross output can be represented in terms of final demand plus intermediate input demand. Let FD To apply the methodology, we can provide an exam- denote final demand, INT denote intermediate input ple by examining Turkey for 2007. Table H.1 pro- demand, and GO denote gross output: vides five measures based on the Export of Value Added database, for the domestic economy as well as GO = FD + INT (H.1) for exports. INT = a GO (H.2) The first column shows how much value added (as where a is a coefficient that indicates the shares of share of a sector’s gross output) is produced in each each intermediate input used in a unit of gross output, sector. The second column shows the share of a sec- which can be thought of as being used across different tor’s exports in gross terms as part of the economy’s sectors and hence can be expressed in the form of a total exports. Similarly, the third column shows the matrix. Equation (H.1) defines final output by inter- share of exports in value-added terms as part of the mediate input use requirements, which is in effect economy’s total value-added exports. We can see that nothing other than the M-matrix. Because we know transport services accounts for an important sector, which sector’s value added goes into another sector, because it generates 19 percent of export value added we can thus also identify the direct and indirect inputs in Turkey’s economy, although it occupies only 15 between sectors expressed in a certain currency. That percent of gross exports. If we look at clothing, for implies that we can analyze real production activities instance, those figures are reversed. and measure it by the value of output. With this database, as previously explained, we Each industry uses a different amount of inter- can estimate the value added embodied in exports for mediate inputs; thus, each link between industries Turkey’s commercial services sector. Because we can will vary. In addition, final demand (FD) as such is account for all links across sectors, we can estimate measured in gross terms, whereas the goal of this in two directions—forward and backward. As for database is to obtain a cleaner “net” measure of domestic production, forward links tell us how much value added. Focusing on value added, we have to value added each sector contains, which is exported disentangle the flow of gross activities and the value- directly and indirectly in the production of other added activities from intermediate to final use. Some goods and services. Hence, forward links represent share of the gross output measure also involves the the contribution of a specific sector as an input to value added that is created within each sector. Once other sectors’ value added. Backward links represent we have a clean measure of value-added shares of the contribution of all other sectors to a specific sec- final demand, we can call that measure β, which then tor’s value added (upstream). The shares of forward must be integrated with the M-matrix to obtain flows links and direct exports are shown in the fourth col- of value added broken down across sector activities. umn in table H.1, while the shares of backward links That leads to a new value-added matrix, VA: and direct exports are shown in the fifth column. Turkey’s business services, for instance, account VA = βM (H.3) for 8 percent of total value added exported directly In effect, β is a coefficient that indicates the shares and 14 percent when including forward links; that of each intermediate input used in a unit of value number is the value of services used as inputs to added, which can be thought of as being used across other sectors. Backward links represent only 6 per- different sectors. Because the Export of Value Added cent of total exports (box H.1). 258 Making Global Value Chains Work for Development Table H.1 Turkey’s Exports, Gross and Value-Added Measures, by Sector, 2007 (%) Share of domestic Value added: Value-added: value-added direct exports direct exports in gross output Gross value: Value added: and forward and backward Sector by sector direct exports direct exports links links (1) (2) (3) (4) 1 Primary agricultural goods 65 3 5 5 4 2 Other primary goods 76 2 3 3 2 3 Energy 27 2 1 4 1 4 Processed foods 43 3 3 3 4 5 Beverages and tobacco 54 0 0 0 0 6 Textiles 38 10 9 8 10 7 Clothing 32 7 5 3 7 8 Leather 26 0 0 0 0 9 Lumber 27 1 1 1 1 10 Paper 43 1 1 1 1 11 Chemicals 29 6 4 4 5 12 Non-metal minerals 51 3 3 2 3 13 Metals 22 9 5 5 6 14 Fabricated metals 35 3 3 4 3 15 Transport equipment 39 14 13 8 14 16 Machinery 40 12 11 8 12 17 Other manufacturing 35 2 1 1 1 18 Water 84 0 0 0 0 19 Construction 47 1 1 0 1 20 Distribution 74 1 3 10 2 21 Transport 56 15 19 16 16 22 Communication 86 0 1 2 1 23 Finance 66 1 1 5 1 24 Insurance 66 1 1 1 1 25 Other business services and ICT 61 1 1 3 1 26 Other consumer services 58 1 2 1 1 27 Other services 89 1 3 2 2 Total 100 100 100 100 Total business services (22–27) 5 8 14 6 Source: Based on the World Bank Export of Value Added database. Note: ICT = information and communications technology. Box H.1 Value Added in Exports We measure the value added contained in exports as follows. Total (direct and indirect) value added in exports, based on back- First, we calculate direct cost shares linked to demand for inter- ward links: mediate inputs: B z = α z + ∑ .01× θi ,zv i x z , e i ≠z θ z ,i = z ,i × 100. ∑e j ,i j where ei,j represents expenditure in sector j on inputs indexed by i, including value added, or primary inputs (capital, labor, and Direct value added in exports: land), and intermediate inputs; vj represents expenditure on pri- αz = vzxz . mary inputs as a share of total costs of production in sector j ; and xj represents the gross value of exports from sector j. Total (direct and indirect) value added in exports, based on forward links: Fz = α z + ∑ .01× θ z ,iv z x i , i ≠z Appendix I SURVEY YEAR AND NUMBER OF DOMESTIC AND FOREIGN MANUFACTURING FIRMS, BY COUNTRY Farole and Winkler (2014) have developed an econo- from foreign direct investment (chapter 7). The metric analysis to assess how foreign investor char- econometric estimation uses a cross-section of more acteristics, domestic firms’ absorptive capacity, and than 25,000 domestic manufacturing firms in 76 a country’s institutional variables influence intra- low- and middle-income economies from the World industry productivity spillovers to domestic firms Bank’s Enterprise Surveys, as listed in table I.1. Table I.1 Survey Year and Number of Domestic and Foreign Manufacturing Firms, by Country Domestic Foreign Country Survey year manufacturing firms manufacturing firms Albania 2007 87 23 Algeria 2007 374 10 Argentina 2010 681 111 Armenia 2009 102 12 Azerbaijan 2009 95 23 Belarus 2008 86 13 Bolivia 2010 118 22 Bosnia and Herzegovina 2009 113 10 Brazil 2009 851 59 Burkina Faso 2009 81 16 Burundi 2006 110 28 Cameroon 2009 87 29 Chile 2010 673 102 Colombia 2010 641 65 Costa Rica 2010 272 53 Côte d’Ivoire 2009 144 49 Croatia 2007 364 47 Ecuador 2010 97 22 Egypt, Arab Rep. 2008 1,103 36 El Salvador 2010 104 20 Ethiopia 2006 337 22 Georgia 2008 109 14 Ghana 2007 271 19 Guatemala 2010 315 40 Guinea 2006 122 15 Guinea-Bissau 2006 72 9 (Table continues next page) 259 260 Making GVCs Work for Development Table I.1 (continued) Domestic Foreign Country Survey year manufacturing firms manufacturing firms Honduras 2010 130 18 India 2006 2,134 37 Indonesia 2009 1,067 89 Jamaica 2010 103 16 Jordan 2006 305 48 Kazakhstan 2009 169 14 Kenya 2007 331 65 Kyrgyz Republic 2009 80 16 Latvia 2009 65 24 Lithuania 2009 85 19 Macedonia, FYR 2009 98 23 Madagascar 2009 125 79 Malaysia 2007 773 321 Mauritania 2006 112 16 Mauritius 2009 157 18 Mexico 2010 1,046 110 Moldova 2009 86 22 Mongolia 2009 115 15 Morocco 2007 354 103 Mozambique 2007 284 56 Namibia 2006 114 37 Nepal 2009 124 4 Nicaragua 2010 115 10 Nigeria 2007 938 10 Pakistan 2007 763 20 Panama 2010 101 14 Paraguay 2010 108 17 Peru 2010 673 87 Philippines 2009 692 262 Poland 2009 137 14 Romania 2009 143 36 Russian Federation 2009 660 39 Rwanda 2006 54 14 Senegal 2007 243 16 Serbia 2009 117 19 South Africa 2007 579 101 St. Vincent and the Grenadines 2010 124 24 Swaziland 2006 68 38 Tajikistan 2008 98 15 Tanzania 2006 247 39 Thailand 2006 800 230 Turkey 2008 866 29 Uganda 2006 276 58 Ukraine 2008 523 46 Uruguay 2010 327 32 Uzbekistan 2008 87 34 Venezuela, RB 2010 71 14 Vietnam 2009 649 130 Yemen, Rep. 2010 238 5 Zambia 2007 236 68 Total 25,199 3,440 Source: Farole and Winkler 2014, 80. Appendix I: Survey Year and Number of Domestic and Foreign Manufacturing Firms, by Country 261 Because the effect of GVC participation is the the measures of structural integration are based on topic of interest in chapter 2, we merged the data set the OECD-WTO TiVA database, fewer observations with two sector measures of structural integration in are available; there are a total of more than 14,000 GVCs: BONwin (buyer’s perspective) and BONwout manufacturing firms in 22 low- and middle-income (seller’s perspective), as described in chapter 6 and countries, as listed in Table I.2. computed by Santoni and Taglioni (2015). Because Table I.2 Survey Year and Number of Domestic and Foreign Manufacturing Firms, by Country, Select Countries Domestic Foreign Country Survey year manufacturing firms manufacturing firms Total Argentina 2010 681 111 792 Brazil 2009 851 59 910 Bulgaria 2007 788 114 902 Chile 2010 673 102 775 Czech Republic 2009 79 22 101 Estonia 2009 65 25 90 Hungary 2009 87 29 116 India 2006 2,134 37 2,171 Indonesia 2009 1,067 89 1,156 Latvia 2009 60 23 83 Lithuania 2009 77 19 96 Malaysia 2007 773 321 1,094 Mexico 2010 1,046 110 1,156 Philippines 2009 692 262 954 Poland 2009 137 14 151 Romania 2009 143 36 179 Russia 2009 660 39 699 Slovak Republic 2009 74 15 89 South Africa 2007 579 101 680 Thailand 2006 586 187 773 Turkey 2008 866 29 895 Vietnam 2009 649 130 779 Total 12,767 1,874 14,641 Notes 1. For example, in gross trade data, if a disk drive from the rest of the world). Similarly, Feenstra and Jensen (2012), Thailand is exported to China, where it is used to make a using confidential U.S. firm data, report that estimates of laptop, which is exported to the United States, the value of intermediate imports based on the proportionality assump- the disk drive is included in Thailand’s and China’s export tion introduce nontrivial errors in some cases. data. Value-added trade data assign the wages and profits 3. The Asian I-O database by IDE-JETRO covers 76 in Thai activities and Chinese activities separately, so it can industries in intermittent years between 1975 and 2005, be seen how each stage of the production process contrib- but it focuses on Asian economies (China; Indonesia; utes to total incomes. Japan; Malaysia; the Philippines; the Republic of Korea; 2. For example, Baldwin and Lopez-Gonzalez (2013) Singapore; Thailand; and Taiwan, China) and the United find that the proportionality in the use of intermediates and States only. by different industries in one of the most important sources 4. See, for example, UN (United Nations). 2004. of these data (the WIOD database) leads to substantially International Merchandise Trade Statistics: Compilers different measures of import content of exports, depend- Manual. New York: UN, 28. http://unstats.un.org/unsd ing on the aggregate chosen (with or without the data for /publication/SeriesF/seriesf_87e.pdf, 28. 262 Making Global Value Chains Work for Development References Athukorala, Prema-chandra. 2010. “Production Net- Lenzen, Manfred, Daniel Moran, Keiichiro Kanemoto, and works and Trade Patterns in East Asia: Regionalization Arne Geschke. 2013. “Building Eora: A Global Multi- or Globalization?” Asian Development Bank (ADB) regional Input-Output Database at High Country and Working Paper Series on Regional Economic Integra- Sector Resolution.” Economic Systems Research 25 (1): tion 56, ADB, Manila. 20–49. Baldwin, Richard, and Javier Lopez-Gonzalez. 2013. McDougall, Robert. 2001. “The GTAP Database, Version “Supply-Chain Trade: A Portrait of Global Patterns 5.” Center for Global Trade Analysis, Purdue University, and Several Testable Hypotheses.” NBER Working West Lafayette, IN. Paper 18957, National Bureau of Economic Research, McDougall, Robert, and Jan Hagemejer. 2005. “Services Cambridge, MA. Trade Data.” In Global Trade, Assistance, and Production: Calì, M., J. Francois, C. Hollweg, M. Manchin, D. A. The GTAP 6 Data Base, edited by Betina V. Dimaranan. Oberdabernig, H. Rojas-Romagosa, S. Rubinova, and Center for Global Trade Analysis, Purdue University, P. Tomberger. Forthcoming. Jobs and Wage Content of West Lafayette, IN. Trade. Washington, DC: World Bank. OECD (Organisation for Economic Co-operation and Christen, Elisabeth, Joseph Francois, and Bernard Development) and WTO (World Trade Organization). Hoekman. 2012. “Computable General Equilibrium 2012. “Trade in Value-Added: Concept, Methodologies, Modeling of Market Access in Services.” In Handbook and Challenges.” Joint OCED-WTO note. http://www of Computable General Equilibrium Modeling, edited .oecd.org/sti/ind/49894138.pdf. by Peter B. Dixon and Dale W. Jorgenson, 1601–43. Sturgeon, Timothy J., and Olga Memedovic. 2011. Amsterdam: Elsevier. “Mapping Global Value Chains: Intermediate Goods Daudin, Guillaume, Christine Rifflart, and Danielle Trade and Structural Change in the World Economy.” Schweisguth. 2011. “Who Produces for Whom in the Working Paper 10/2010, United Nations Industrial World Economy?” Canadian Journal of Economics/ Development Organization, Vienna. Revue canadienne d’économique 44 (4): 1403–37. Taymaz, Erol, Ebru Voylvoda, and Kamil Yilmaz. 2011. Daudin, Guillaume, Christine Rifflart, Danielle “Uluslararasi Uretim Zincirlerinde Donusum ve Schweisguth, and Paola Monperrus-Veroni. 2006. “Le Turkiye’nin Konumu.” TUSIAD-Koc Universitesi Commerce Extérieur en Valeur Ajoutée.” Revue de Ekonomik Arastirma Forumu Calisma Raporlari l’OFCE: Observations et Diagnostics Économiques 98: Serisi. Turkish Industry and Business Association, 129–65. Istanbul. Farole, Thomas, and Deborah Winkler. 2014. “The Role Timmer, Marcel P., ed. 2012. “The World Input-Output of Mediating Factors for FDI Spillovers in Developing Database (WIOD): Contents, Sources and Methods.” Countries: Evidence from a Global Dataset.” In Making World Input-Output Database Working Paper 10. Foreign Direct Investment Work for Sub-Saharan Africa: http://www.wiod.org/publications/source_docs Local Spillovers and Competitiveness in Global Value /WIOD_sources.pdf. Chains, edited by Thomas Farole and Deborah Winkler, Tsigas, Marinos, Zhi Wang, and Mark Gehlhar. 2012. 59–86. Washington, DC: World Bank. “How a Global Inter-Country Input-Output Table with Feenstra, Robert C., and J. Jensen. 2012. “Evaluating Processing Trade Account Can Be Constructed from Estimates of Materials Offshoring from U.S. GTAP Database.” Paper presented at the 15th Annual Manufacturing.” NBER Working Paper 17916, National Conference on Global Economic Analysis, Geneva, June Bureau of Economic Research, Cambridge, MA. 27–29. https://www.gtap.agecon.purdue.edu/resources Francois, J., M. Manchin, and P. Tomberger. 2013. /download/5998.pdf. “Services Links and the Value Added Content of Trade.” UN (United Nations). 2002. “Classification by Broad Policy Research Working Paper 6432, World Bank, Economic Categories.” United Nations, Department Washington, DC. of Economic and Social Affairs, Statistics Division, Johnson, Robert C., and Guillermo Noguera. 2012. Statistical Papers, Series M No. 53, Rev. 4. United “Accounting for Intermediates: Production Sharing Nations, New York. and Trade in Value Added.” Journal of International UNCTAD (United Nations Conference on Trade and Economics 86 (2): 224–36. Development). 2013a. Global Value Chains and Koopman, Robert, William Powers, Zhi Wang, and Shang- Development: Investment and Value Added Trade in Jin Wei. 2011. “Giving Credit Where Credit Is Due: the Global Economy: A Preliminary Analysis. New York Tracing Value Added in Global Production Chains.” and Geneva: United Nations. http://unctad.org/en NBER Working Paper 16426, National Bureau of /PublicationsLibrary/diae2013d1_en.pdf. Economic Research, Cambridge, MA. ———. 2013b. “World Investment Report 2013—Global Koopman, Robert, Zhi Wang, and Shang-Jin Wei. 2014. Value Chains: Investment and Trade for Development.” “Tracing Value-Added and Double Counting in UNCTAD, Geneva. Gross Exports.” American Economic Review 104 (2): 459–94. ECO-AUDIT Environmental Benefits Statement The World Bank Group is committed to reducing its environmental footprint. In support of this commitment, the Publishing and Knowledge Division leverages electronic publishing options and print-on-demand technology, which is located in regional hubs worldwide. Together, these initiatives enable print runs to be lowered and shipping distances decreased, resulting in reduced paper consump- tion, chemical use, greenhouse gas emissions, and waste. The Publishing and Knowledge Division follows the recommended standards for paper use set by the Green Press Initiative. The majority of our books are printed on Forest Stewardship Council (FSC)–certified paper, with nearly all con- taining 50–100 percent recycled content. The recycled fiber in our book paper is either unbleached or bleached using totally chlorine-free (TCF), processed chlorine-free (PCF), or enhanced elemental chlorine-free (EECF) processes. More information about the Bank’s environmental philosophy can be found at http://www.worldbank.org/corporateresponsibility.