G L O B A L VA L U E C H A I N D E V E L O P M E N T R E P O R T 2 0 19 TECHNOLOGICAL INNOVATION, SUPPLY CHAIN TR ADE, AND WORKERS IN A GLOBALIZED WORLD © 2019 World Trade Organization World Trade Organization Centre William Rappard Rue de Lausanne 154 1211 Geneva 2 Switzerland Telephone: +41 (0)22 739 51 11 Internet: www.wto.org This work is a product of the World Trade Organization, the Institute of Developing Economies (IDE-JETRO), the Organisation for Economic Co-operation and Development, the Research Center of Global Value Chains headquartered at the University of International Business and Economics (RCGVC-UIBE), the World Bank Group, and the China Development Research Foundation. It is based on joint research efforts to better understand the ongoing development and evolution of global value chains and their implications for economic development. The findings, interpretations, and conclusions expressed in this work are those of the authors and do not necessarily reflect the views of the co-publishing partners, their Boards of Executive Directors, or the governments they represent. The co-publishing partners do 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 judgment on the part of the co-publishing partners concerning the legal status of any territory or the endorsement or acceptance of such boundaries. Rights and Permissions The material in this work is subject to copyright. Because the co-publishing partners encourage dissemination of their knowledge, this work may be reproduced, in whole or in part, for non-commercial purposes as long as full attribution to this work is given. Any queries on rights and licenses, including subsidiary rights, should be addressed to World Bank Publications, The World Bank Group, 1818 H Street NW, Washington, DC 20433, USA; fax: 202–522–2625; e-mail: pubrights@worldbank.org. WTO print ISBN 978-92-870-4771-7 WTO web ISBN 978-92-870-4968-1 This publication uses US spelling. All mentions of dollars refer to US dollars, unless otherwise indicated. The term “billion” refers to a thou- sand million. The Research Center of Global Value Chains acknowledges the financial support from the Bill & Melinda Gates Foundation. iii  Contents Foreword by Michael Spence. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . v Co-publishing partners. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . vi Contributors. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . vii Acknowledgements. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . viii Abbreviations and acronyms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ix Key messages . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . x Executive summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 David Dollar Chapter 1 Recent patterns of global production and GVC participation. . . . . . . . . . . . . . . . . . 9 Xin Li (Beijing Normal University), Bo Meng (IDE-JETRO), and Zhi Wang (RCGVC-UIBE) Chapter 2 Trade, value chains and labor markets in advanced economies . . . . . . . . . . . . . . . . 45 Marc Bacchetta (WTO) and Victor Stolzenburg (WTO) Chapter 3 Global value chains and employment in developing economies. . . . . . . . . . . . . . . . 63 Claire H. Hollweg (World Bank Group) Chapter 4 Technological progress, diffusion, and opportunities for developing countries: Lessons from China. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 83 Satoshi Inomata (IDE-JETRO) and Daria Taglioni (World Bank Group) Chapter 5 Understanding Supply Chain 4.0 and its potential impact on global value chains . . . . . . 103 Michael J. Ferrantino (World Bank Group) and Emine Elcin Koten (World Bank Group) iv • Technological innovation, supply chain trade, and workers in a globalized world Chapter 6 The digital economy, GVCs and SMEs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 121 Emmanuelle Ganne (WTO) and Kathryn Lundquist (WTO) Chapter 7 Should high domestic value added in exports be an objective of policy?. . . . . . . . . . . 141 David Dollar (Brookings Institution), Bilal Khan (RCGVC-UIBE), and Jiansuo Pei (SITE-UIBE) Chapter 8 Improving the accounting frameworks for analyses of global value chains . . . . . . . . . . 155 Nadim Ahmad (OECD) Appendix 1 Chapter Authors’ Conference: Final programme . . . . . . . . . . . . . . . . . . . . . . . 179 Appendix 2 Technological Innovation, Supply Chain Trade, and Workers in a Globalized World: Global Value Chain Development Report 2019 Background Paper Conference. . . . . . . . . . . . 181 v  Foreword T here are different ways to analyze the global economy. The mobile-internet- and platform-centered open ecosystems, One is to view it through the lens of growth and struc- along with mobile payment systems and enabled financial services, tural change in individual economies, developed and have the potential to support inclusive growth patterns and expand developing. A second is to use the lens of global value the channels, opportunities, and accessible markets for SMEs. Data chains (GVCs), the complex network structure of flows of goods, from China’s domestic economy experience supports these trends. services, capital and technology across national borders. Both are Exploiting the international potential of these platforms to expand useful and they are complementary to one another. trade and access for SMEs requires investment and infrastructure in The 2019 edition of the GVC Development Report is enor- developing countries, but also new trade regimes that increase the mously valuable, in part because it captures the underlying tech- openness of the ecosystems. In other words, the potential to sup- nological and economic forces that are transforming the patterns port growth and employment in SMEs via access to global markets of global interconnectedness. on digital platforms is as yet largely unexploited. The report notes that there are two megatrends in process. The report supports and adds to a broad range of studies that One is the growth of developing countries, the expansion of the suggest that the combination of trade and various aspects of digital middle classes in them, and the shift in the share of global purchas- transformation has contributed to job and income polarization, and ing power toward the developing economies. By itself this would that vigorous policies (by government and business) are required to produce major shifts in the characteristics of GVCs. Regional trade restore more inclusiveness to the observed growth patterns. This is rises as a share, especially in Asia. More production now goes to especially true in developed economies. Key policies are those that rapidly growing domestic markets in developing countries instead support the workforce in transitions as a growing range of tasks are of being exported outside the region. Trade is shifting from a automated and jobs shift toward a mix of tasks that are complemen- stark version of comparative advantage based on differential labor tary to the machines. costs and labor arbitrage, toward something that more closely In developing countries, especially those in the middle-income resembles the intra-industry model of trade among developed category, while the pressures on the structure of jobs and employ- economies based on product and technological differentiation. Of ment are similar to developed economies, the net impact of dig- course, that process is far from complete, and there remain ear- ital technology appears to have been positive for growth and for ly-stage, and relatively low-income developing countries for which employment. the growth models will continue to depend on accessing global There is an important caution in the report. The long-run goal demand via labor-intensive, process-oriented manufacturing. of development is of course to increase productivity, employment The second megatrend is the digitization of the underpinnings and incomes. But in the context of GVCs, attempts to artificially of entire economies and, by implication, GVCs and the global economy. This too is a process that is underway and one that has increase the domestic value-added content of exports, ahead much further to go. It is difficult if not impossible to accurately pre- of the technological deepening of the economy, are likely to be dict the endpoint, if there is one. But there are important insights counterproductive. that the second GVC report highlights. At a more macro level, while trade continues to grow, especially One clear message is that as economies move to being built in services (where there remain challenging measurement prob- in part on digital foundations, trade, GVCs and digital technology lems) the declines in trade relative to global GDP and the rising cannot be separated and dealt with as independent trends and share of intraregional trade are understood to be largely the natural forces. consequences of economic development and the early stages of For early-stage developing countries, automation will at some the digital transformation of economies, and not mainly the result point displace the labor-intensive technologies that underpinned of trade frictions and resistance to globalization engendered by the earlier Asian growth stories. That shift will occur differentially the adverse distributional features of growth patterns. by sector, with textiles and more generally the sewing trades The second GVC report is carefully researched and deep in being the least vulnerable in the short run. The message is two- insights. It does an admirable job of capturing the complexity of fold: don’t give up on the traditional growth model but move rap- a global economy in rapid transition, and especially of bringing idly to expand internet capability and the digital underpinnings into focus the major forces and trends and their impacts. and infrastructure of the economy. Michael Spence Nobel Laureate in Economics vi • Technological innovation, supply chain trade, and workers in a globalized world Co-publishing partners T his work has been co-published by the World Trade economies whose mission is to promote policies that will improve Organization, the Institute of Developing Economies the economic and social well-being of people around the world: (IDE-JETRO), the Organisation for Economic Co-oper- “Better Policies for Better Lives”. The OECD does this by provid- ation and Development, the Research Center of Global ing a forum for governments to share experiences and by seek- Value Chains headquartered at the University of International ing solutions to common problems. Business and Economics (RCGVC-UIBE), the World Bank Group, The Research Center of Global Value Chains (RCGVC) is a and the China Development Research Foundation. global academic think tank headquartered at the University of The World Trade Organization (WTO) is an international International Business and Economics, focusing on basic and organization that deals with the global rules of trade between interdisciplinary research activities on the development of global nations. The WTO administers agreements, negotiated and value chains (GVCs) and its implication on global economies. signed by its members, which provide the legal ground rules for The World Bank is an international development institution international commerce. Their purpose is to help trade flow as established by Articles of Agreement adopted by its member freely as possible for the economic development and the welfare countries. The World Bank’s overarching mission is to reduce of its members’ citizens. The WTO is serviced by a secretariat poverty, improve living conditions, and promote sustainable and which provides expert, impartial and independent support to comprehensive development in its developing member coun- member governments, including research, analysis and statistical tries. It has established two ambitious goals to anchor its mis- information related to the role and developments of trade in the sion: end extreme poverty within a generation and boost shared global economy. prosperity. The World Bank will achieve these goals by providing IDE-JETRO is a government-affiliated research institute that loans, concessional financing, technical assistance, and knowl- conducts basic and comprehensive research on economics, edge sharing services to its developing member countries and politics, and social issues in developing countries. Through its through partnerships with other organizations. research, IDE-JETRO contributes knowledge of developing The China Development Research Foundation (CDRF) is a economies and better understanding of the regions to the gov- public foundation initiated by the Development Research Center ernment and public. of the State Council (DRC). Its mission is to advance good gover- The Organisation for Economic Co-operation and Devel- nance and public policy to promote economic development and opment (OECD) is an international and inter-governmental social progress. organization comprising the world’s main industrialized market vii  Contributors Co-editors Claire H. Hollweg Senior Economist, World Bank Group David Dollar Senior Fellow, China Center, Brookings Institution Satoshi Inomata Chief Senior Researcher, Institute of Developing Economies – Emmanuelle Ganne Japan External Trade Organization Senior Analyst, World Trade Organization Bilal M. Khan Victor Stolzenburg Assistant Professor, Research Center for Global Value Chains, Research Economist, World Trade Organization University of International Business and Economics Zhi Wang Emine Elcin Koten Professor and Director, Research Center for Global Value Chains, Consultant, World Bank Group University of International Business and Economics (RCGVC- UIBE); Research Faculty and Senior Policy Fellow, Schar School Xin Li of Policy and Government, George Mason University Professor, School of Statistics, Beijing Normal University Other contributors Kathryn Lundquist Statistician, World Trade Organization Nadim Ahmad Head of Trade and Competitiveness Statistics Division, Bo Meng Organisation for Economic Co-operation and Development Senior Overseas Research Fellow (New York), Institute of Developing Economies - Japan External Trade Organization Marc Bacchetta Counsellor, World Trade Organization Jiansuo Pei Associate Professor, School of International Trade and Michael J. Ferrantino Economics, University of International Business and Economics Lead Economist for Trade Policy, World Bank Group Daria Taglioni Senior Economist, World Bank Group viii • Technological innovation, supply chain trade, and workers in a globalized world Acknowledgments T he Global Value Chains Development Report is a joint (Penn State University), Gary Hufbauer (Peterson Institute for publication of the World Trade Organization (WTO), International Economics), Alonso de Gortari (Princeton and Dart- the Institute of Developing Economies (IDE–JETRO), mouth), Kalina Manova (University College London), Maurice D the Organisation for Economic Co-operation and Kugler (George Mason University), Marcel Timmer (the University Development (OECD), the Research Center of Global Value of Groningen), and Felix Tintelnot (University of Chicago). The Chains (RCGVC-UIBE), the World Bank Group, and the China editors are grateful to Michael Spence for his keynote speech Development Research Foundation, based on joint research at the background paper conference in Beijing and his invalu- efforts to better understand the ongoing development and evo- able expertise and advice on the overall narrative of the report. lution of global value chains and their implications for economic The editors also thank Robert Koopman, chief economist of the development. World Trade Organization and Caroline Freund, director of the This second report draws contributions from 23 background World Bank Group’s Trade, Regional Integration and Investment papers; 16 of them were presented and discussed at the confer- Climate, for their guidance and support during the joint research ence “Technological Innovation, Supply Chain Trade, and Work- process as well as research and data contributions from the Asian ers in a Globalized World” in Beijing during March 22–23, 2018, Development Bank. organized by the RCGVC and the China Development Research The editors are grateful to William Shaw, who copy-edited the Foundation. Drafts of the eight chapters of the report were pre- report, and to Anthony Martin, Head of WTO Publications, and sented and discussed at the second Authors’ Conference in Helen Swain, WTO Publications Editor, who were responsible for Geneva on October 8, 2018, organized by the WTO. The editors the production of the report. thank the authors of background papers and individual chap- The RCGVC would like to acknowledge the financial support ters and the discussants and participants in the two conferences from the Bill & Melinda Gates Foundation and National Science for insightful comments and suggestions that helped draft and Foundation of China (grant No. G0304-71733002). improve the chapters (see appendices 1 and 2 for the programs). The report’s co-editors are David Dollar, Emmanuelle Ganne, Special thanks go to our external reviewers: Jonathan Eaton Victor Stolzenburg and Zhi Wang. ix  Abbreviations and acronyms ADB Asian Development Bank IDE–JETRO Institute of Developing Economies AI artificial intelligence IMF International Monetary Fund ASEAN Association of Southeast Asian Nations IoT Internet of Things B2B business-to-business IPR intellectual property rights B2C business-to-consumer ITC International Trade Centre CDRF China Development Research Foundation LACEX World Bank’s Labor Content of Exports CEO Chief Executive Officer LDC least-developed country CIF ost, insurance and freight LMIC low/medium-income country CMI customer-managed inventory MFN most-favored nation DRC Development Research Center of the State MNE multinational enterprise Council OECD Organisation for Economic Co-operation DVA domestic value-added and Development DVAR domestic value-added ratio R&D research and development ECU electronic control unit RCGVC Research Center of Global Value Chains EDI electronic data interchange RFID radio frequency identification ESUT extended supply-use table SME small and medium-sized enterprise FATS foreign affiliates statistics SNA System of National Accounts FDI foreign direct investment STEM science, technology, engineering, and F.O.B. free-on-board mathematics FTZ foreign trade zone SUT supply and use table GDP gross domestic product TFP total factor productivity GM global manufacturing TiVA trade in value-added GSM global system for mobile communication UIBE University of International Business and Economics GTH Global Trade Helpdesk UNCTAD United Nations Conference on Trade GVC global value chain and Development HIC high-income country VMI vendor-managed inventory ICIO inter-country input-output WB World Bank ICT information and communication technology WIOD World Input-Output Tables IDB Inter-American Development Bank WTO World Trade Organization x • Technological innovation, supply chain trade, and workers in a globalized world Key messages • The growth of global value chains has slowed since the Global the re-shoring of manufacturing production, thus reducing Financial Crisis of 2008-09 but not stopped. In fact, complex opportunities for developing countries to participate in GVCs, global value chains (GVCs) grew faster than GDP in 2017. or they may strengthen GVCs by reducing coordination and • Factoring in GVCs when studying the impact of trade on matching costs between buyers and suppliers. labor markets reveals that trade has not been a significant • Despite the aggregate gains they create, trade, automation contributor to declines in manufacturing jobs in advanced and digital technologies can cause disruption and widen exist- economies, and that job gains in services have offset job ing disparities across regions and individuals. This calls for losses in manufacturing. broad and comprehensive adjustment policies. • The emergence of GVCs has offered developing countries • While small and medium-sized enterprises (SMEs) are opportunities to integrate into the global economy by deliver- under-represented in GVCs, the digital economy provides new ing jobs and higher income. opportunities for SMEs to play a more active role. • The impacts of technological change and increased produc- • Open and transparent policies tend to promote GVC-led tivity on employment linked to GVCs have been offset by growth more than import-reducing policies targeted at raising growing consumer demand, and in the short term, automa- the share of domestic value-added in exports. tion will not dramatically reduce the attractiveness of low- • Using value-added trade rather than gross trade statistics wage destinations, especially for labor-intensive tasks that is crucial to understanding GVCs and their impact on jobs. require human dexterity. Efforts to continue to improve the quality of these estimates • The impact of new digital technologies on GVCs is uncertain: are strongly encouraged. they may reduce the length of supply chains by encouraging Executive summary DAVID DOLLAR M ore than two-thirds of world trade occurs through other components of GDP. During the global financial crisis there global value chains (GVCs), in which production was naturally some retrenchment of GVCs, followed by quick crosses at least one border, and typically many recovery (2010-2011) but since then, with the exception of 2017, borders, before final assembly. The phenomenal growth has, in the main, slowed. In 2017 expansion of complex growth in GVC-related trade has translated into significant eco - GVCs was faster than GDP growth, but it is too early to say if this nomic growth in many countries across the globe over the last is a new trend or just a one-year blip. two decades, fueled by reductions in transportation and com- Concerning which sectors are particularly amenable to munication costs and declining trade barriers. But, at the same GVCs, over a long period we found that, the higher the technol- time, it has contributed to distributional effects that mean that ogy (knowledge) intensity of a sector, the more significant the the benefits of trade have not always accrued to all, which has, at increase of complex GVC activities. Thus, GVC linkages are espe - least in part, been a driver in the backlash against globalization cially important for high-tech sectors and it is in these areas that and the rise of protectionism and threats to global and regional we see highly complex value chains involving many countries. trade agreements. In addition, new technological developments We also distinguish between intra-regional GVC activities and such as robotics, big data, and the Internet of Things (IoT) are inter-regional ones. Activities within North American economies beginning to reshape and further transform GVCs. This second would be an example of the former, whereas China’s growing GVC development report takes stock of the recent evolution of contribution to value chains centered on the U.S. or Germany GVC trade in light of these developments. would be examples of the latter. Between 2000 and 2017, the weight of intra-regional GVC activities in “Factory Asia” came to exceed that of “Factory North America”. In contrast, the share of Update on trends in GVCs intra-regional GVC activities declined relatively in both Europe and North America and their share of inter-regional production The growth of global value chains has slowed since the global sharing activities increased, especially their GVC linkages with financial crisis. A country’s GDP (value added) can be decom- “Factory Asia”, reflecting in large part increased inter-connect- posed into purely domestic, traditional trade, in which a product edness with China. China is increasingly playing an important is made in one country and consumed in another, simple value role as both a supply and demand hub in traditional trade and chain trade, in which a good made in one country crosses one simple GVC networks, although the U.S. and Germany are still border and is used in production in the partner country before the most important hubs in complex GVC networks. consumption there, and complex value chain trade, in which GVC analysis also provides some insight into bilateral trade production crosses multiple borders. From 2000-2007, GVCs, balances and how they should be interpreted. In a world in which especially complex ones, were expanding at a faster rate than most trade consists of parts and components, bilateral trade 1 2 • Technological innovation, supply chain trade, and workers in a globalized world balances are significantly affected by the supply and demand of nationwide and sectoral level to regional and individual outcomes third countries; and, net imports are no longer a proper measure reveals substantial heterogeneity in how aggregate effects map of the impact of an international trade shock on the domestic out. For instance, when local labor markets within countries are economy in the age of GVCs, compared to the time when final not sufficiently diversified, trade can widen regional disparities. goods trade dominated. China happens to be at the end of Regions specialized in import-competing and upstream industries many Asian value chains, taking sophisticated components from can fall behind, while areas with industries that export or benefit Japan, the Republic of Korea, and Chinese Taipei and assembling from cost savings due to cheaper imported inputs pull away. these into final products. Two-thirds of all intermediate imports Similarly, trade may work in the same direction as other drivers of information and communication technology (ICT) products, in contributing to labor market polarization. In particular, automa- coming from other countries in Factory Asia, but also with signif- tion has impacted jobs in the middle of the skills distribution, with icant contributions from Europe and North America, are used as remaining jobs concentrated at the high and low ends. Between inputs into Chinese exports. Indeed, the Chinese domestic value 1999 and 2007, the years when China was reducing barriers and content of their exports of ICT products accounts for only around entering the WTO, nearly all advanced economies had increases half of the total export value. As such, trade balances look very in employment shares for high- and low-skilled jobs, and declines different in value-added terms. For example, the U.S. trade defi- for middle-skill work (see Figure 1). cit in ICT products with China is roughly cut in half if the calcula- While trade and automation are making a country as a whole tion is made in value added terms. richer, there is a need for adjustment policies to ensure a more even distribution of these gains. This is especially the case as value chains magnify trade-induced changes in skill requirements Labor market effects of GVCs in developed and thereby raise the demand for worker flexibility and the need countries for training support. With regard to the optimal design of such policies, value chains make targeted or specific labor market One of the main controversies of globalization is its effect on interventions increasingly difficult. As input-output linkages labor markets in both developed and developing economies. cause trade shocks to spread more widely within economies, Across advanced economies, the real median wage has grown import competition is less and less limited in terms of industries, slowly over the past two decades and manufacturing employ- regions, or skill levels. As a result, it can become more difficult to ment has been on the decline, while incomes of highly skilled identify the exact reason for individual displacement. Therefore, workers and owners of capital have soared. There are of course adjustment policies should not differentiate between the various many factors at work here, and not all are related to globaliza- reasons for worker displacement, such as automation or trade, tion, especially countries’ own domestic tax and transfer policies, and be less dependent on affected workers fulfilling certain but one additional factor has been big developing countries, especially China and Eastern European economies, opening up and joining the global economy. FIGURE 1 Percentage point changes in employment shares A number of studies have concluded that, in particular, the by skill level between 1995-2015 impact of Chinese import competition on the U.S. labor market, Percent especially after China joined the WTO, was a significant factor behind U.S. manufacturing employment dropping sharply after 10 2000. But these analyses have typically only provided a partial view of the overall impact on employment, by and large ignoring 5 the reality of value chains. A full view requires that we account for the fact that the development of value chains results in churning 0 across economies, as firms and countries specialize and create certain types of jobs while eliminating others. General equilib- -5 rium analyses of the so-called “China shock” that take account of GVCs find that, for the U.S., trade was not a main contribu- -10 tor to the loss of manufacturing jobs and has only minor aggre- gate employment effects. One important reason for this more -15 nuanced effect is that while some industries contracted because Southern Northern Western Total North Japan Central of increased competition, others expanded thanks to the cost Europe Europe Europe America Europe savings that GVC linkages provided, counterbalancing jobs lost in contracting industries. This is consistent with economic theory, Low skill Middle skill High skill which suggests that trade should not have a large net effect on employment. Source: OECD (2017). See chapter 2 for details. That being said, the effects vary considerably across regions and individuals with different skill levels. Moving from the Executive summary • 3 FIGURE 2 Many jobs in Viet Nam are tied to exports, directly and indirectly 25000 Manufacturing Agriculture, forestry and fishery 20000 Trade 15000 Hotels and restaurants 10000 Mining Transport, storage and communications 5000 Education and training; health and social work; recreational, cultural and sporting activities 0 Financial intermediation 1989 1996 2000 2007 2012 Real estate, renting and business activities; Year scientific activities and technology Other services activities Electricity, gas and water supply Total JOBX Construction Direct JOBX Public administration, activities of party and membership organizations and defence Indirect JOBX 0 2000 4000 6000 8000 10000 12000 Direct JOBX Indirect JOBX Source: Hollweg (2017). See chapter 3 for details. conditions. In addition, mobility and place-based policies could indirect exporting firms that supply inputs to the direct exporters usefully complement general labor market policies to address (see Figure 2). regional divergence. The relationship between GVC integration and level of employment though is not necessarily positive in all contexts. Imports of goods and services (backward GVC participation) Labor market effects of GVCs in developing matter as much as exports of intermediates (forward GVC partici- countries pation) to be successful in GVCs, where opening up to imports is often a pre-condition to successfully export. However, there may The emergence of global value chains has offered developing be import-competing effects in labor markets. countries new opportunities to integrate into the global econ- Evidence as well as intuition suggests that GVC participation omy. This has fundamental impacts on where jobs go, who gets will have other distributional implications. Greater participation them, and what type of jobs they are. Significant parts of the of developing countries in global trade is expected to integrate developing world are deeply involved in GVCs. Their input has not only markets for products, services, finance and technology, been initially concentrated in labor-intensive activities, which may but also, directly and indirectly, markets for labor. The hallmark have had important impacts on poverty in developing countries. of globalization is big developing countries opening up and join- For example, the boom in exports to the United States following ing global trade. In general, such economies are abundant in the US–Vietnam Bilateral Trade Agreement of 2001 was partic- unskilled labor and scarce in skilled labor and capital relative to ularly beneficial to wages of unskilled workers, reduced the skill global averages. The factor-endowment theory of trade predicts premium, and was a key driver of poverty reduction in Viet Nam that trade will reduce returns to unskilled labor in advanced econ- because it was concentrated in unskilled, labor-intensive GVC omies while raising returns to capital and skilled labor. This trend sectors, most notably textiles. has generally been observed. But the opposite trend should There is a positive association between output growth and occur in developing countries that open up: wages of unskilled employment growth within GVC sectors, which increased over- workers, clearly the most abundant factor in many developing all welfare as workers moved out of agriculture or the informal countries, should rise faster than other factor rewards. This has sector toward better paying, higher value-added jobs. Women not happened in most developing countries; rather, employment who previously had difficulty accessing this type of wage work creation and wage gains have been biased towards more skilled have filled many of these jobs. Employment and wage impacts workers. GVC expansion in developing countries is associated can happen both directly within exporting firms as well as indi- with higher relative demand for skilled workers. Characteristics rectly through these firms’ demand for goods and services from of GVCs themselves, by supporting more complex industrial the domestic economy. The extent to which GVCs interact with organization, as well as services inputs that are complementary domestic labor thus depends on the linkages of exporting firms to value chains, can be skill-biased. to domestic, input-supplying firms. The firms that export directly Automation may be threatening GVC jobs in developing account for only a small part of GVC jobs. In Viet Nam, most countries in the long term, where the routine tasks more suscep- of the job creation results from backward linkages – that is, in tible to automation are increasingly performed. Technological 4 • Technological innovation, supply chain trade, and workers in a globalized world advancements that largely get diffused through global value firms to be full partners in global technology ecosystems and to chains are affecting how GVCs support jobs in developing coun- pursue open source innovation solutions. tries. Evidence suggests that changes in efficiency in GVCs has The question that now remains is whether firms from other negative impacts on employment linked to countries’ participa- countries, especially in less/least-developed regions, can repli- tion in the global production of products, all else equal. Tech- cate the positive experience of leveraging platforms by Chinese nological innovation has also lowered the demand for low-skilled firms as demonstrated in this chapter. And does automation of workers relatively more than compared to high-skilled workers. production even prevent initial entry based on low wages? Nevertheless, the adverse effects of changing production tech- Robotics, 3D printing, the IoT, Big Data, and cloud comput- nologies and efficiencies on employment have been offset by ing, among others, are transforming entire industries. The evi- increased consumer demand, whereby the domestic consump- dence suggests that automation reduces some of the incen- tion expenditures in large emerging economies such as China and tives for GVCs to relocate to lower-wage countries. However, it India will generate new demand for labor for the global economy is also seen that automation does not necessarily dampen the These distributional consequences of trade and other forces attractiveness of low-wage destinations, especially for labor-in- are a principal concern to policymakers. Policies also play an tensive tasks that require human dexterity. In the apparel indus- important role in mediating the relationship between GVCs and try, for example, soft materials like fabrics are difficult to handle employment in developing countries. These include policies that through automation compared to solid materials such as metal support (i) participation of developing countries in GVCs, (ii) fos- or wooden objects, and sewing/stitching can still be out of the tering positive spillovers from GVC participation, (iii) upgrading reach of “robots’ hands”. to higher value-added tasks within GVCs, and (iv) mediating neg- In this sense, automation is likely to have only a limited impact ative effects from winners, such as inequality. on developing countries’ opportunities to participate in value chains through the offshoring of production by high-income countries, at least in the short term. Foreign direct investment Technological progress, diffusion, and flows (greenfield investment) from high-income countries to opportunities for developing countries low- and middle-income countries has declined since 2010. Nev- ertheless, there are important differences across industries and The nature of technology used in products plays a major role in between production and assembly tasks within industries. The determining the governance structure of value chains and the pattern across countries also suggests that some FDI may have benefits of participation for developing countries. Standardization migrated from China to low-income and middle-income coun- through breaking production into modules with a high degree of tries in Asia and Africa and from higher- to lower-income coun- functional autonomy (limited mutual interference between mod- tries in the Europe and Central Asia region. ules) can dramatically reduce the amount of R&D, learning by While automation does not pose immediate risks to shut the doing, and the number of complementary skills needed to pro- door to labor-intensive exports from developing countries, gov- duce a good. This greatly increases opportunities for developing ernments need to develop a comprehensive digital strategy. Our country firms to participate in formerly capital-intensive industries economies are increasingly sitting on a digital foundation, one through reducing entry costs into global value chains. that is generating high-speed growth and disruptive change. However, widespread access to standardized products with The employment and investment of tomorrow will be data inten- little ability to modify technical features can lead to an exces- sive, and value in a knowledge economy is increasingly created sive supply of homogeneous products in a local market, resulting by innovative ideas and data. in intense price competition and limited technology transfer. By Not only is embracing digital technologies good for the contrast, technology that facilitates scope for product modifica- economy, but it is also good for society. The digitally powered, tion and greater interaction with lead manufacturers can help knowledge-intensive GVCs that are emerging and are likely to boost technology transfer and product upgrading by develop- dominate in the coming years have a strong potential for inclu- ing country firms. Chapter 4 illustrates this interaction between sion. As Nobel Prize winner Mike Spence points out, they have changes in technology and opportunities for developing coun- low marginal costs of production and are non-rival. Moreover, tries through developments in the automotive and cell phone they can expand markets for small businesses beyond traditional industries in China. geographies. They can also expand financial inclusion, as data The chapter argues that policies for helping domestically on e-commerce can be used as collateral, and smartphones link owned firms to become technologically standalone – what some up poorer countries to these opportunities. might refer to as “techno-nationalism” – do not necessarily deliver the expected results in terms of upgrading. The world’s most powerful technology companies, both from emerging and GVCs and digital technology advanced countries, work with global suppliers and even with competitors in “open innovation” environments. Hence, the “Supply Chain 4.0” is the re-organization of supply chains – advice to policymakers seeking to upgrade toward the global design and planning, production, distribution, consumption, technology frontier is to prioritize measures that encourage and reverse logistics – using technologies that are known as Executive summary • 5 “Industry 4.0”. These technologies emerged in the 21st century New technologies gather prodigious amounts of data. Big and are largely implemented by firms that are at the frontier of data analytics is about using data to drive useful business intel- supply chain management in high-income countries. The most ligence, answering the questions, “What just happened?”, “Why frequently mentioned supply management techniques are the did it happen?”, and “What are we going to do next?” The ability IoT, big data analytics, 3D printing, advanced (autonomous) to collect and analyze data gathered in the whole supply chain robotics, smart sensors, augmented reality, artificial intelligence, makes it possible to “run scenarios within the platform”, where and cloud computing. Through these advanced techniques, a the platform is conceived of as an overarching software solution continuous flow of information between the retailer and supplier within the supply chain control center. Besides saving time and keeps the shelves stocked and there is no longer a “back room” labor, and reducing errors, EDI enables a large amount of data in stores where inventory is kept. capture about customer behavior which can be the basis for “Supply Chain 4.0” is about transforming the model of supply supply chain analytics using either “big data” or “small data” chain management from a linear model in which instructions techniques. flow from supplier to producer to distributor to consumer, and The use of modern technology and human labor in warehouses back, to a more integrated model in which information flows in are often complements, rather than substitutes, especially in con- multiple directions. While lead firms are increasingly analyzing ditions where e-commerce is substantially increasing demand this information through “supply chain control towers,” the end for certain goods and services. E-commerce is a mechanism for effect of this development is making the goods economy more translating unpaid household shopping time into paid market responsive to consumer demand. According to a recent PwC time. Instead of consumers spending time shopping, workers in (2016a) study on the rise of Industry 4.0, a third of the more than warehouses and on delivery trucks are picking goods off ware- 2,000 respondents say their companies have started to digitize house shelves and bringing them to the consumer’s front door. their supply chains, and fully 72% expect to have done so five Most of the jobs being created involve moving goods around years from now. either in warehouses or delivery vehicles and have many of the In “Supply Chain 4.0”, the internet makes the warehouse vis- characteristics of factory work. A study using U.S. data gathered ible to the customer and within the warehouse, some technol- in the Occupational Employment Statistics of the Bureau of Labor ogies such as autonomous logistics and robotic transport can Statistics, shows that employment in the most dynamic parts of be employed to substantially improve pick-and-pack perfor- the supply chain has grown at a rate substantially exceeding that mance. Business-to-business (B2B) e-commerce consists of links of the overall economy since 2011. These sectors include ware- in supply chains – whether transactions between parts suppliers housing and storage, couriers and messengers (i.e. express deliv- and assemblers, between distribution centers and retailers, or ery), and non-store retailers (i.e. e-commerce companies). online purchases of services which in many cases support the Digital technologies and the internet are becoming the foun- supply chain. B2B commerce can be implemented either through dation of entire economies. There are huge benefits in terms websites, much like business-to-consumer (B2C) e-commerce, or of inclusive patterns of growth, innovation and entrepreneurial through electronic data interchange (EDI) which is a mature tech- opportunities, but the downside risks are much larger than was nology through which the computer systems of the buyer and initially understood. Trade and investment will be vulnerable in seller are directly connected using a common record format. the near complete absence of international agreements on the To rapidly assess and respond to changes in customer uses and prohibited abuses of the internet and data. demand, tracking and tracing throughout the supply chain is enabled through sensing technologies underlying the IoT, including radio frequency identification (RFID), Bluetooth, and The digital economy and SMEs global system for mobile communication (GSM). Applications of IoT are increasingly used to facilitate the management strategies Small and medium-sized enterprises in general have low direct of “customer-managed inventory” (CMI) or “vendor-managed participation in international trade, compared to large enter- inventory” (VMI) in which information is initially provided by a prises. This result makes economic sense as long as there are customer and then transmitted up the supply chain to the ware- fixed costs in exporting, such as learning about foreign markets house. Technologies such as RFID tags then transmit information or rules and minimum scales for shipping. In theory, the spread to the distribution center so that orders can be fulfilled. An EDI of GVCs should reduce these effects and make it easier for SMEs system causes an order created electronically by the customer to to participate in trade as the break-up of the production process be instantly duplicated without error in the vendor’s computer makes it feasible for a specialized firm to find niche markets. Yet, system, and the invoice to be similarly electronically duplicated SMEs are underrepresented in GVCs. in the customer’s computer system. Some of these processes This may be changing, however, as access to information and are being implemented through blockchain, a distributed ledger communication technology (ICT) continues to grow. For example, technology that allows multiple parties to maintain copies of the there is evidence that the internet reduces search costs, facilitat- same information in various locations, either in an open manner ing more exchange and increasing firm productivity. Cross-bor- or requiring individual entities’ permission to access the network. der e-commerce platforms are also providing new opportunities Its special feature is that historical entries cannot be altered. for SMEs and even micro firms. 6 • Technological innovation, supply chain trade, and workers in a globalized world Using firm-level data from the World Bank’s Enterprise Sur- countries to specialize in a particular activity and join a global veys, new research finds that whether a firm has a website on production network. As a developing country moves from export the internet facilitates the participation of manufacturing SMEs of primary products to export of manufactures and services via in GVCs and trade. In particular, such SMEs are more likely to use GVCs, the ratio of domestic value added to gross export value foreign inputs for production and export their output. Further, tends to fall. Developing countries often start out at the end of ICT connectivity is found to be more important for small firms value chains, with labor-intensive assembly of parts produced than for large ones when considering whether or not a firm par- elsewhere. For some individual products, the ratio of domestic ticipates in trade. value added to gross export value can be very small, maybe only Evidence underlines the importance of ICT access for SMEs to a few percentage points. The gross exports from the country can join GVCs in the digital economy, however, access to new tech- be very large, but this is an artifact of the position in the value nology varies not only between firm size, but also regionally by chain. The country’s value added contribution to the export is level of development. Infrastructure constraints faced by devel- much smaller. Many developing countries worry about this phe - oping countries in e-commerce range from the most basic, such nomenon and aspire to increase their value added contribution as access to a steady supply of electricity, to the more complex, to exports. There are a number of reasons why this objective such as not having access to electronic payment systems or a should be approached cautiously. It may seem like simple math lack of high-speed internet cables. This is a particular problem, that a higher domestic value added share means more total not only because information communication technology (ICT) is value added exported and hence more GDP. But that simple necessary for e-commerce, but also because ICT is now consid- idea ignores the reality that imported goods and services are ered a pre-requisite for joining most GVCs. No matter the inter- a key support to a country’s competitiveness. If a country artifi- net’s functionality, regardless of lacking features such as broad- cially replaces key inputs with inferior domestic versions, the end band connection and e-commerce platforms, e-commerce can result is likely to be fewer gross exports and less, not more, total only develop if the internet is present. This is in line with empir- value added exports. ical studies showing that access to the internet improves export History provides a number of interesting lessons about this performance in developing countries across manufacturing and issue. First, in almost all countries, developed and developing services sectors through reduced search costs and decreased alike, the share of domestic value added in exports has tended to distance barriers. Furthermore, the internet has also been shown trend downwards over time. This reflects the expansion of global to increase firm productivity, especially of smaller and less inno - value chains. Even the countries best known for final products in vative firms. key sectors such as autos, machinery, and electronics rely heav- However, SMEs face a number of additional challenges inte - ily on imported inputs, both manufactures and services. Many of grating into GVCs with the digital economy. On top of lagging the iconic products in the world, such as BMW cars and iPhones, behind large firms in terms of overall digital technology use and have large amounts of imported inputs that go into final assem- capability, small businesses may also find it difficult to access bly. Developing countries have learned part of this lesson and e-commerce platforms and payment systems. National policy are generally quite open to imports of parts and components. may also be inadvertently preventing successful internationaliza- However, imported services are also a key input into manufac- tion of SMEs via GVCs. Complex customs procedures, regulatory tures, and developing countries tend to be more protectionist uncertainty and barriers to services trade all adversely affect vis-à-vis services. Both trade in services and investment in ser- SMEs and pose challenges to SME participation in GVCs, despite vices (often needed in order to trade the services) tend to be the opportunities provided by e-commerce. more restricted in developing countries, than policies towards These findings underscore the continuing need to improve manufactures. Developing countries that have more imported the ICT environment/infrastructure and to expand services such service content in their exports tend to be more persistent and as e-payment and e-commerce, all of which benefit SMEs dispro - successful exporters of manufactures. portionally, but they also highlight the lack of information regard- A second point about the ratio of domestic content to gross ing SMEs. In theory the digital economy holds potential for SME value of exports is that the early East Asian industrializers show participation in GVCs, but for effective policies to be developed, a highly non-linear trend in this variable. In the case of Japan, better data will need to be collected. this ratio fell in the early post-war period as the country opened up and began to use imported inputs. In the 1980s, however, the ratio increased as Japan became a capable producer of a Should high domestic value added in exports wide range of manufactured intermediates and parts. Since 1990 be an objective of policy? there has again been a sharp trend downwards in domestic con- tent as complex value chains developed throughout Asia. Japan Global value chains make it easier for developing countries to is an industrial powerhouse with many successful brands, and it move away from export reliance on unprocessed primary prod - is revealing that the domestic content ratio in the most recent ucts to become exporters of manufactures and services. Before year is the lowest ever recorded. Being an industrial powerhouse the development of GVCs, a country had to master the pro- does not mean that all activities take place within the border. duction of a whole product in order to export it. GVCs allow Japanese firms use imported goods and services in a highly Executive summary • 7 efficient manner. The Republic of Korea’s and Chinese Taipei’s trade faced by smaller firms compared to larger firms). The same experiences are very similar to Japan’s, but with a lag. generalizations hold true for foreign-owned enterprises, or enter- China’s recent experience is an important counter-example. prises with affiliates abroad, compared to purely domestic firms; At the beginning of economic reform there was a sharp drop for example, the foreign content of exports by foreign-owned in the domestic value added ratio as the country moved from firms in the transport sector in the United States is twice that of exporting primary products to assembling apparel and elec- domestically owned firms. But TiVA estimates, relying as they do tronics using parts produced in other countries. However, over on national supply-use and input-output tables, cannot reflect the past decade the ratio has been rising, catching the attention these heterogeneities; thus, key measures, such as the import of other developing countries. Our research indicates that this content of exports, are downward biased. trend is primarily the result of technological advance in China, Additional complexities can create significant interpretative not the result of restrictive trade policy. What is happening to challenges for users of TiVA type statistics. Because inter-coun- China now is analogous to what happened to Japan in the 1980s try input-output tables value transactions at basic, and not and the Republic of Korea in the 1990s, as their technological market, prices, many of the related TiVA analyses reveal only capability advanced. If China’s experience continues to be simi- some of the story. What is often not fully understood in the use lar to the earlier industrializers, then the ratio can be expected to of tables valued in basic prices is that they exclude the value that peak and later decline as labor-intensive activities are off-shored is added at the end of the chain by distribution sectors (in par- to lower wage locations and more imported components and ticular retail and wholesale, which often include value associated parts are used in production to keep Chinese firms competitive with marketing activities and brands). At the heart of the debate, in international markets. China’s development is likely to be influ- and indeed confusion, is that input-output tables in basic prices enced by its “Made in China 2025” industrial policy. This policy are in essence a mechanism to provide a view of production, and aims to make China a technology leader in ten advanced manu- because they remove significant distribution margins at the end facturing sectors. China has set indicative targets for domestic of the chain, they are less well equipped to provide a perspec- value added in these sectors. In semiconductors, for example, tive from the consumption point of view. This has a direct impact China currently imports 90% of usage, but plans to produce 70% on smile-curve type analyses that describe where sectors are in of usage by 2025, which would be an extraordinary shift. What value chains and how far they are from final demand. Moreover, remains unclear is what policy tools China will use. If it restricts although the basic price concept may provide a correct view of, imports or direct investment in these sectors, it will make its firms for example, the domestic value-added or services content of a less competitive, not to mention inflaming global trade tensions. country’s total exports, it provides an arguably distorted view of the same measure of a given good seen from a consumption, or free-on-board (FOB) perspective. This is because basic prices Issues in GVC measurement exclude often significant distribution margins related to trans- portation from the factory gate to the customs frontier, which The proliferation and development of global input-output tables may also reflect significant contributions from activities related to in recent years has significantly transformed our ability to inter- brand, R&D, design, and marketing. For example, the US domes- pret global production. But, important though such initiatives tic value-added content of its exports of textiles and clothing in have been, they are typically silent on the role of multinationals FOB prices was around 20% in 2016 compared to 3% using the in this new landscape. In addition, the policy debate in recent pure basic price approach. years has increasingly focused on ‘inclusive globalization’, refer- The basic price approach also limits the scope to reveal addi- ring to the growing realization that the benefits of globalization tional dependencies related to globalization, for example jobs may not have accrued to all members of society equally, even if sustained in retailers through sales of imports. A complementary only as a process of transition. accounting framework is developed in “market” prices to illus- With traditional macroeconomic statistics, it is not immediately trate the insights that can be gained through such an approach. clear, for example, which categories of workers in which countries In the United States, for example, the sale of imports supported benefit from globalization (and how) and which may have been, 9.0 million jobs. even if only temporarily, left behind. Moreover, trade in val- ue-added (TiVA) estimates, derived through the construction of a global input-output table, implicitly assume that all firms within a given sector have the same production function (input-output technical coefficients), import intensity and export intensity. This of course has never been true. We know for example that larger firms will typically have different production functions compared to smaller firms, because of economies of scale, and also higher labor productivity. And these firms will also typically be more export and, indeed, import orientated than their smaller counterparts (reflecting in part the disproportionate costs of CHAPTER 1 Recent patterns of global production and GVC participation Xin Li (Beijing Normal University), Bo Meng (IDE-JETRO), and Zhi Wang (RCGVC-UIBE) ABSTRACT Taking advantage of a new accounting method to decom- chains still remain largely regional; China is increasingly pose GDP production into pure domestic production, tra- playing an important role as both a supply and demand hub ditional trade, simple and complex GVC activities, this in traditional trade and simple GVC networks, although the chapter examines recent trends in global value chain (GVC) US and Germany are still the most important hubs in com- activities across the world. Our main findings show that the plex GVC networks; bilateral trade balances are significantly pace of GVC activities picked up in 2017 after a period of affected by the supply and demand of third countries; and slow down since 2012; intra-North American and intra-Euro- net imports are no longer a proper measure of the impact pean GVC activities declined relative to inter-regional trans- of international trade on the domestic economy in the age actions due to higher penetration via Factory Asia but value of GVCs. • The growth of global value chains has slowed since the 2008-09 Global Financial Crisis but has not stopped. From 2000 to 2007, global value chains (GVCs), especially complex ones, expanded at a faster rate than GDP. During the global financial crisis there was naturally some retrenchment of GVCs, followed by quick recovery (2010-2011), but since then growth has mostly slowed. However, most recent data for 2017 show that complex GVCs grew faster than GDP. • Value chains remain largely regional but they are not static. Between 2000 and 2017, intra-regional GVC trade increased in “Factory Asia” reflecting, in part, upgrading by China and other Asian economies. In contrast, intra-regional GVC trade in “Factory Europe” and “Factory North America” decreased slightly relative to inter-regional GVC trade reflecting stronger linkages with “Factory Asia”. • China has emerged as an important hub in traditional trade and simple GVC networks, but the United States and Germany remain the most important hubs in complex GVC networks. 9 10 • Technological innovation, supply chain trade, and workers in a globalized world G lobal value chains, where firms specialize in a particu- based on whether it crosses national borders or not. Value-added lar set of activities in one country to produce parts and creation is only classified as a GVC activity when the embodied components for other countries, have spread the pro- factor content in a product crosses a national border for produc- duction process across countries; their share of world tion purposes (Box 1.1). production and trade has expanded greatly over the past three The chapter is organized as follows. Section 1 describes decades. In the years immediately after the global financial crisis, the changing pattern of global production activities and GVC however, the expansion of GVCs significantly slowed, according participation across countries and industries based on global to GVC production measures reported in the 2017 GVC develop- inter-country input-output (ICIO) tables constructed by Asian ment report. At the same time, the world has seen the emergence Development Bank, which covers 62 economies and 35 indus- of populist, protectionist movements in many advanced countries. tries up to 2017. Section 2 demonstrates the changing distribu- The looming trade tension between the United States and its tion of value-added production activities along typical global major trading partners, especially China, the second largest econ- value chains, as more developing countries have been integrated omy in the world, will have significant consequences for growth into the global production network. Section 3 uses network opportunities in developing countries, but also, in a world of high analysis to demonstrate the topology of the global production levels of interdependence, developed economies. network structure of traditional trade, simple and complex GVC The first chapter of this report updates trends in GVC pro - activities, and their evolution between 2000 and 2017. Section 4 duction and trade activities in both developed and developing analyzes the multilateral nature of bilateral trade and focuses on economies by technology (knowledge) intensity and income three sensitive bilateral trade relations (US-China, US-Germany, level, according to the production decomposition method pro- US-Japan) to demonstrate the roles third countries have played posed by Wang et al (2017). This approach classifies the embod - in determining bilateral trade balances in the age of global value ied factor content in a product into GVC and non-GVC activities chains. Section 5 concludes. BOX 1.1 A production decomposition to identify and measure GVC activities In Wang et. al. (2017), production activities are divided into there is no cross-country production sharing; the difference 4 broad types depending on whether they involve produc- between the two is whether they satisfy either domestic or tion sharing between two or more countries. The first type foreign final demand. The last two types are cross-country is value added produced at home and absorbed by domes- production sharing activities; the differences between the tic final demand without involving international trade. No two are whether they satisfy partner country or other coun- factor content crosses national borders in the entire produc- tries’ final demand, and the number of times factor content tion and consumption process. The second type is domestic crosses national borders. Domestic and import input-output value added embodied in final product exports, that is, tra - coefficient matrixes in ICIO tables are used to distinguish ditional trade: products are made completely by domestic domestic and foreign factor content in various production factors and factor content crosses a national border once activities. The classification and relation among the four for consumption only. The third type is domestic value types of production are depicted in Figure 1.1. added embodied in a country-sector’s intermediate trade According to this decomposition method, GVC activities that is used by the partner country to produce its domestic as a share of total production activities can be used to mea- products consumed locally, or is foreign value added that sure the intensity of each country-sector’s participation in is imported directly from partner countries and used for cross-country production sharing activities. Essentially, this domestically consumed products. Factor content is used in approach measures the percentage of production in a par- production outside the home country and crosses a national ticular country-sector that has been engaged in global pro- border once for production. Therefore, it is referred to duction networks. The forward GVC participation indicator as “simple GVC activities”. The last type is value added is based on a decomposition of GDP production; it shows embodied in intermediate exports/imports that is used the percentage of production factors employed in a coun- by a partner country to produce exports (intermediate or try-sector that have been involved in cross-country produc- final) for other countries. In this case, factor content crosses tion sharing activities. The backward participation indicator a national border at least twice, so is referred to as “com- is computed based on a decomposition of final goods pro - plex GVC activities.” Production activities in the first two duction; it shows the percentage of final products produced types are entirely conducted within national borders, and by a country-sector coming from GVC activities. Recent patterns of global production and GVC participation • 11 FIGURE 1.1 Decomposition of production activities1 Production of Value-added or GVCs Final products Production sharing between two or more countries Pure Domestic Traditional Trade GVCs No border Cross border Cross border crossing for consumption for production Hair cut Portugal wine in Intermediate trade exchange for England cloth Simple GVCs Complex GVCs Cross border once for production Cross border Chinese steel in US building at least twice iPhone/Auto 1. The changing pattern of global production 10% increase in complex GVC activities led the growth. However, activities and GVC participation2 rising trade tensions between the United States and its major trading partners, especially China, has introduced tremendous GVC activities as a share of global GDP fell from 2011 to 2016, uncertainty in the global economy recovery process. Determin- as the share of purely domestic production activities rose (see ing whether the recovery of cross-country production sharing Figure 1.2, which is an update of Figure 2.3 in the 2017 GVC activities in 2017 has started a new trend requires more years of Development Report based on the newly released ICIO tables data and further analysis. by the Asian Development Bank). This continues the downward A first step is to measure the impact of the recent, sharp trend in GVC activities shown in the 2017 GVC report based changes in commodity prices on nominal growth rates of pro- on data through 2014. However, the growth of global trade duction activities shown above. The global prices of crude oil surpassed the growth of global GDP for the first time in nearly and other bulk commodities have gone through a “super circle” six years in 2017, and there were signs of a recovery of GVC since 2000. For example, the per barrel crude oil price (dated activities. Brent) fluctuated dramatically during 2000-2018, rising from less The nominal growth rate of all types of production activi- than 30 US dollars in 2000 to over 110 dollars in 2011, falling to ties (the four activities are defined in Box 1.1) fell sharply during less than 50 dollars by 2016, and then rebounding to about 70 2012-2016, with a much sharp slowdown in cross-country, pro - dollars since early in 2018. Because crude oil and other bulk com- duction-sharing GVC activities. The decline was the steepest for modities are important intermediate inputs in global production, complex GVC activities, followed by simple GVC activities, tra- these price fluctuations may affect the relative nominal growth ditional trade and domestic production activities; the average patterns of different types of value-added creation activities annual changes for these four types of activities during 2012- measured in current US dollars shown in Figure 1.3. 2016 were -1.65%, -1.00%, -0.28% and 1.49% respectively (indi- It appears, however, that the more rapid decline in the nom- vidual year data are reported in Figure 1.3, which is an update inal value of GVCs than other activities as a share of GDP from of Figure 2.5 in the 2017 GVC report). Thus, the limited increase 2011-2016 was not due simply to price changes. Figure 1.4 shows in global GDP from 2012-2016 was almost entirely accounted the growth rate of the volume of world merchandise trade, world by the growth of pure domestic production; international trade real GDP and their ratio during 1995-2017. For each year when contributed very little during this slow recovery period. In 2017, global real trade growth was faster than global real GDP growth, the growth rate of global trade exceeded that of global GDP, a complex GVC activities had the highest nominal rate of growth 12 • Technological innovation, supply chain trade, and workers in a globalized world FIGURE 1.2 Trends in production activities as a share of global GDP, by type of value-added creation activity, 1995-2017 Percent 86 Asian Financial Crisis 84 Financial Crisis 82 Dotcom Bust 80 78 Domestic 76 10 Simple GVC 8 6 Traditional GVC 4 Complex GVC 2 0 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 Source: 1995-2009 are based on the University of International Business and Economics (UIBE) GVC indexes derived from the 2016 World Input-Output Table, and 2010-2017 are based on the UIBE GVC indexes derived from the Asian Development Bank (ADB) 2018 ICIO tables. FIGURE 1.3 Nominal growth rates of different value added creation activities, global level, 2000-2017  30% 15% 20% 10% 10% 5% 0% 0% 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 -10% -5% -20% -10% -30% -15% Domestic Traditional Trade Simple GVC Complex GVC The nominal growth of GDP (right) Source: 2000-2010 are based the UIBE GVC indexes derived from the 2016 World Input-output table, and 2011-2017 are based on the UIBE GVC indexes derived from the ADB 2018 ICIO tables. Recent patterns of global production and GVC participation • 13 FIGURE 1.4 The growth rate of merchandise trade volume and real global GDP, 1995-2017, %  20.0 4.0 3.2 15.0 3.0 2.5 2.5 2.5 2.3 1.9 1.9 1.8 1.7 1.7 10.0 1.6 2.0 1.5 1.4 1.5 1.5 1.1 0.9 0.9 0.9 0.8 0.7 5.0 1.0 0.0 0.0 -0.2 -5.0 -1.0 -10.0 -2.0 -15.0 -3.0 95 96 97 98 99 00 01 02 03 04 05 06 07 08 09 10 11 12 13 14 15 16 17 19 19 19 19 19 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 World merchandise World GDP growth (left) Ratio of trade growth volume trade growth (left) to GDP growth (right) Source: Global GDP is from World Development Indicators, WB and World Economy Outlook, IMF; Merchandise volume trade is from UNCTAD. The ratio of trade decline to GDP decline in 2009 is 7.4, out of scale shown in the graph. among the four type activities shown in Figure 1.3. And when Across sectors3, mining (represented by the purple dots) is in world trade grew slower than world GDP, complex GVC activi- the upper left corner, indicating a high degree of forward GVC ties grew more slowly than other activities. This can be under- participation but a low degree of backward GVC participation. stood intuitively, because complex GVC activities are the only Most service sectors, especially for sectors in the other services one of these four components of value added production where group (utility, education, health care and personal services, rep- factor content embedded in products cross a national border at resented by the blue dots) tend to be in the lower left corner, least twice. When complex GVC activities grow slower than pure meaning that they have low participation in GVC activities by domestic production activities, as happened during 2012-2016, both measures. In comparison, high research and development world trade grows slower than GDP. (R&D) intensity manufacturing sectors (red dots) tend to be in the To evaluate the impact of the shift in production patterns upper right quarter of the graph, reflecting their active partic- after the global financial crisis to GVC participation across coun- ipation in GVCs as both producers and buyers of intermediate tries and industries, we plot the forward and backward GVC par- products. ticipation indicators jointly in a scatterplot based on ADB ICIO Ten years after the global financial crisis, global GVC partic- tables (Figure 1.5). The two red dotted lines indicate the world’s ipation has not returned to pre-crisis levels: the global average average forward and backward participation rates and divide GVC participation rate (as a share of GDP) was 0.1289 in 2017, the figure into four quadrants. Most countries fall along the compared to 0.1343 in 2007. GVC activities recovered faster in 45-degree line, indicating that countries that have a high degree high-income countries than in middle-income countries. The of forward participation also tend to have a high degree of back- recovery of specific GVC activities (backward versus forward par- ward participation. Major resource exporters, such as Mongolia, ticipation) also differs across income groups. Forward GVC par- Russia and Norway, fall above the 45-degree line (Figure 1.5, ticipation increased more rapidly than backward participation in upper left). Since natural resources are the most upstream sec- the high-income countries, especially in the high-income Eastern tors, these economies tend to have much higher degree of for- European countries (the forward participation rate of the Czech ward GVC participation than backward GVC participation. Republic rose from 0.2355 in 2007 to 0.2812 in 2017, of Estonia 14 • Technological innovation, supply chain trade, and workers in a globalized world FIGURE 1.5 GVC participation indicators, country levels and sector levels  0.7 0.7 Forward -Linkage 2017 Forward -Linkage 0.1343 2007 0.1289 0.6 0.6 AGR AGR 0.5 0.5 FBS FBS HTI HTI 0.4 0.4 LTI LTI MIN MIN 0.3 0.3 MTI MTI OSE OSE 0.2 TTC 0.2 TTC 0.1 0.1343 0.1 0.1289 Backward-Linkage Backward -Linkage 0.0 0.0 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.7 0.7 0.1343 2007 0.1289 2017 Forward-Linkage Forward-Linkage 0.6 0.6 BRN 0.5 0.5 SIN LUX MAL LUX BRN 0.4 MON 0.4 High income KAZ High income MON Lower middle NOR Lower middle income 0.3 income 0.3 RUS LAO Upper middle Upper middle RUS MYS income 0.2 VNM income 0.2 NOR VNM 0.1343 0.1289 0.1 0.1 USA Backward-Linkage USA Backward -Linkage 0.0 0.0 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 Note: AGR is an abbreviation of Agriculture, MIN is Mining; HTI, MTI and LTI are High, middle and low R&D intensive industries respectively; TTC is Trade and Transportation; FBS is Financial and Business services; OSE is other services. Source: the UIBE GVC indexes derived from the ADB 2018 ICIO tables. from 0.2536 to 0.3151, of Hungary from 0.2298 to 0.2777, and of return to pre-crisis levels. For instance, India’s forward and back- Latvia from 0.1818 to 0.2712). A higher growth rate of forward ward participation rate dropped from 0.1006 and 0.1382 in 2007 participation in manufacturing and service sectors often implies to 0.0655 and 0.0991in 2017, respectively. China, Indonesia and faster upgrade of GVC production activities4 as well as the deep- Philippines also were subjected to similar declines. ening of intra-product specialization brought about by the recov- Comparing the development of different GVC activities in dif- ery of cross-country production sharing activities. At the same ferent income groups in longer period, significant growth of GVC time, some middle-income economies such as Mexico, Romania participation only occurred in high-income countries. In partic- and Viet Nam moved up faster in backward participation, which ular, their forward GVC participation rate increased from 9.5 in mirrors what happened in developed countries. Finally, some 2000 to 12.7 in 2017, while simple and complex activities con- Asian developing economies that experienced a decline in both tributed approximately equal shares (Table 1.1). The GVC partic- forward and backward GVC participation have not yet seen a ipation rate actually declined in upper middle income countries. Recent patterns of global production and GVC participation • 15 TABLE 1.1A Forward GVC participation indexes by country groups (Percent of GDP) GVC participation Simple GVC Complex GVC Income level 2000 2007 2017 2000 2007 2017 2000 2007 2017 High 9.5 11.8 12.4 5.6 6.8 7.1 3.8 5.0 5.3 Upper middle 11.4 14.1 10.5 7.2 8.4 6.4 4.2 5.6 4.2 Lower middle 10.8 12.4 9.1 6.9 7.6 5.7 3.9 4.8 3.4 TABLE 1.1B Backward GVC participation indexes by country groups (percent of final goods production) GVC participation Simple GVC Complex GVC Income level 2000 2007 2017 2000 2007 2017 2000 2007 2017 High 9.3 11.7 11.8 5.8 6.8 6.5 3.5 4.9 5.3 Upper middle 12.5 14.1 10.5 7.3 7.7 6.3 5.2 6.4 4.2 Lower middle 11.7 14.2 11.8 7.9 9.3 7.6 3.8 4.8 4.2 Source: the UIBE GVC indexes derived from the ADB 2018 ICIO tables. This is because participation in cross-border production sharing domestic value chains. The proper combination of cross-border is only one kind of division of labor that can contribute to indus- and domestic value chains, or domestic and foreign factor con- trialization. The substitution of imported intermediate inputs by tent in a particular product, should be determined by market domestically-produced intermediate inputs in advanced devel- forces (this issue is examined in detail in Chapter 7). oping economies, such as the industrial upgrading in China, The 2008/2009 global financial crisis had a dramatic, nega- may also reduce the intensity of GVC participation due to the tive impact on GVC participation for all countries in the world deepening of domestic division of labor and the lengthening of (Figure  1.6). The GVC participation rate increased by 4.3% per FIGURE 1.6 The changing intensity of GVC participation by income groups, 1995-2017  Forward participation Backward participation 16.0 16.0 15.0 15.0 14.0 14.0 13.0 13.0 12.0 12.0 11.0 11.0 10.0 10.0 9.0 9.0 8.0 8.0 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 World average High income World average High income Lower middle income Upper middle income Lower middle income Upper middle income Note: 1995-2011 are from WIOD 2014ed, 2012-2017 are from ADB ICIO database. The global average GVC participation ratio may above all three country groups in some years, due to the incomplete country coverage in both ADB and WIOD database. ADB ICIO table only covers 62 countries and the WIOD ICIO table only covers 43 countries, rest countries in the world are classified as rest of the world in both databases. Therefore, when the GVC participation in those not individually identified countries increase, the global average will be higher than the three country groups reported here, and this is confirmed by the analysis of Figure 1.6-1.8 below. Source: The UIBE GVC indexes derived from the WIOD and ADB 2018 ICIO tables. In particular, the data from 1995 to 2011 derived from the WIOD, and the data from 2012-2017 derived from ADB. 16 • Technological innovation, supply chain trade, and workers in a globalized world year during the pre-crisis GVC expansion period (2000-2008). activities rate in most industry groups, indicating complex GVC This rate declined by 14.9% during the crisis in 2009, but recov- activities are more sensitive to external economic shocks. ered by 9.0% during 2010-2011. However, the world average GVC Analysis over a longer period shows that GVC activities of all participation rate declined by 1.6% per year with the sharp slow- sectors increased from 2000 to 2017. The higher the technol- down of global trade from 2012 on, mainly driven by middle-in- ogy (knowledge) intensity of the sector, the larger the increase come countries (the complex GVC participation rate of high-in- in complex GVC activities. For instance, the forward GVC par- come countries was higher in 2017 than in 2007). In particular, the ticipation rate of the high, middle and low technology-intensive GVC participation rate of the lower middle-income and upper manufacturing sectors increased by 4.2, 3.8 and 3.2 percentage middle-income groups in 2017 was still approximately 2.6 and 3.7 points during 2000 to 2017. Complex GVC activities contributed percentage points lower than their participation rate in 2007. 58.1% of these increases, on average, with a particularly high According to the table 1.2a and table 1.2b, the participation contribution (76.4%) to the 4.2 percentage point increase of rates of most industry groups are still lower than their pre-cri- the GVC participation rate in the high-tech sector. The forward/ sis levels, especially for all the goods producing industries. backward GVC participation rates in the business and financial The tables also indicate that the complex GVC activities rate services sector, which also is relatively knowledge intensive, increased more (or declined more) than did the simple GVC also increased from 10.7/5.8 to 15.2/9.4, respectively (Table 1.2). TABLE 1.2A Forward GVC participation indexes by industry groups (percent of value added) GVC participation Simple GVC Complex GVC Sector level 2000 2007 2017 2000 2007 2017 2000 2007 2017 High Tech 25.3 30.7 28.8 13.8 16.1 15.6 11.5 14.6 13.2 Middle Tech 22.5 21.6 23.7 14.5 16.4 14.7 8.0 9.7 9.1 Low tech 12.4 15.8 15.3 7.9 9.9 9.5 4.5 5.9 5.8 Business & financial 10.7 14.9 15.2 6.6 9.1 9.0 4.0 5.8 6.2 Trade and transportation 10.2 13.4 13.4 6.2 7.9 8.0 4.0 5.5 5.4 Other services 2.3 3.5 3.3 1.4 2.1 2.0 0.9 1.4 1.3 Agriculture 8.3 11.4 10.6 5.8 7.8 7.2 2.4 3.6 3.5 Mining 39.9 54.3 48.3 25.6 34.5 29.6 14.3 19.8 18.8 Source: the UIBE GVC indexes derived from the ADB 2018 ICIO tables. TABLE 1.2B Backward GVC participation indexes by industry groups (percent of final goods production) GVC participation Simple GVC Complex GVC Sector level 2000 2007 2017 2000 2007 2017 2000 2007 2017 High Tech 22.3 28.8 26.8 8.4 9.8 9.6 13.9 19.0 17.3 Middle Tech 19.1 26.9 25.9 10.0 14.4 13.2 9.1 12.5 12.7 Low tech 16.6 21.8 20.5 9.9 11.7 10.5 6.7 10.1 10.0 Business & financial 5.8 8.7 9.4 4.2 5.7 5.9 1.7 2.9 3.6 Trade and transportation 7.1 10.3 10.4 4.9 6.8 6.7 2.2 3.4 3.7 Other services 6.9 10.2 10.0 5.3 7.6 7.3 1.6 2.5 2.6 Agriculture 8.4 11.3 9.6 5.7 7.5 6.2 2.7 3.8 3.4 Mining 10.2 12.1 11.4 6.5 6.1 7.6 3.7 5.9 3.8 Source: the UIBE GVC indexes derived from the ADB 2018 ICIO tables. Recent patterns of global production and GVC participation • 17 Higher GVC intensity in the high-tech, knowledge intensive sec- generating new sources of comparative advantage for interna- tors in part reflects the role of GVCs in the dissemination of tech- tional exchange. The organization of production based on tasks nology from the lead firms to their suppliers (Rodrik, D., 2018). by multinational enterprises, in which parts and components of The high intensity of complex GVC activities in high-tech special products (such as computers, automobiles and airplanes) sectors indicates R&D and other technology inputs have pro - cross national borders several times (complex GVC activities) moted intra-product specialization and the extension of global is the fundamental force that drove global trade growth faster production networks. Slicing the production process into differ- than global GDP growth before the global financial crisis. It ent tasks has greatly extended the depth and scope of interna- also provided new opportunities for developing countries to be tional exchange and division of labor, from between products to integrated into global economy by specializing in some simple between stages of the production of individual products, thus tasks in which they have a comparative advantage, thus enabling FIGURE 1.7 GVC participation indicators by countries and sectors, 2007 and 2017, manufactures  1.0 1.0 0.2688 2007_MTI 2017_MTI Forward-Linkage 0.2586 Forward-Linkage 0.9 0.9 0.8 G HKG 0.8 HKG LTU LUX 0.7 0.7 0.6 MYS 0.6 THA High income High income DEU 0.5 Lower middle 0.5 Lower middle DEU income income 0.4 TWN 0.4 ITA Upper middle VNM Upper middle ITA income MEX income 0.3 0.3 0.2 MEX 0.2 0.2605 USA 0.2372 CYP 0.1 VNM 0.1 CAM IND IND Backward-Linkage Backward-Linkage 0.0 0.0 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 1.0 1.0 0.2179 2007_LTI 0.2051 2017_LTI Forward-Linkage Forward -Linkage 0.9 0.9 High income Lower middle 0.8 0.8 income BRN Upper middle 0.7 BRN 0.7 income 0.6 0.6 SGP FIN 0.5 MYS 0.5 High income TUR SGP 0.4 Lower middle 0.4 income 0.3 Upper middle 0.3 THA IDN income RUS 0.2 0.1581 0.2 IDN LAO 0.1527 VNM VNM 0.1 LAO C AM 0.1 JPN CYP MEX Backward -Linkage MEX Backward-Linkage 0.0 0.0 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 Note: the country abbreviation is according to the ISO 3166-1 alpha-3, and a complete list of the current officially assigned ISO 3166-1 alpha-3 codes is available on the United Nations International Trade Statistics: https://unstats.un.org/unsd/tradekb/knowledgebase/country-code. Source: the UIBE GVC indexes derived from the ADB 2018 ICIO tables. 18 • Technological innovation, supply chain trade, and workers in a globalized world FIGURE 1.8 GVC participation indicators by countries and sectors, 2007 and 2017, services  1.0 1.0 0.0866 Forward-Linkage 2007_FBS 2017_FBS Forward-Linkage 0.0944 0.9 0.9 0.8 0.8 0.7 0.7 LUX LUX 0.6 0.6 IRL SGP 0.5 IRL High income 0.5 NLD MLT 0.4 MYS Lower middle 0.4 CYP income High income Upper middle THA income 0.3 Upper middle 0.3 Lower middle NLD income KAZ income 0.2 0.2 MYS FRA 0.1493 FRA 0.1516 0.1 0.1 KGZ Backward-Linkage Backward-Linkage 0.0 0.0 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 1.0 1.0 2017_OSE Forward-Linkage 2007_OSE Forward-Linkage 0.1026 0.0996 0.9 0.9 0.8 0.8 0.7 0.7 0.6 0.6 High income 0.5 0.5 High income Lower middle 0.4 income 0.4 Lower middle Upper middle income BTN 0.3 income 0.3 Upper middle income 0.2 0.2 NLD KGZ Backward-Linkage 0.1 0.0349 0.1 VNM MLT 0.0331 MLT USA Backward-Linkage 0.0 0.0 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 1.0 1.0 Forward-Linkage Forward-Linkage 0.9 0.1026 2007_TTC 0.9 0.1039 2017_TTC 0.8 0.8 0.7 0.7 0.6 SGP 0.6 SGP LUX MDV 0.5 High income 0.5 High income KAZ MYS LUX Lower middle LTU Lower middle 0.4 MON income 0.4 income VNM RUS TWN Upper middle 0.3 Upper middle 0.3 RUS BEL BEL income income 0.2 VNM 0.2 0.1340 0.1337 CHN 0.1 0.1 Backward-Linkage Backward-Linkage 0.0 0.0 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 Source: the UIBE GVC indexes derived from the ADB 2018 ICIO table. Recent patterns of global production and GVC participation • 19 FIGURE 1.9 GVC participation indicators by countries and sectors, 2007 and 2017, agriculture and mining 1.0 1.0 0.1131 Forward-Linkage Forward -Linkage 0.0961 0.9 2007_AGR 0.9 2017_AGR 0.8 0.8 0.7 0.7 LUX 0.6 0.6 High income 0.5 0.5 MYS Lower middle High income income 0.4 0.4 Lower middle Upper middle BEL income income MYS 0.3 0.3 THA Upper middle VNM LAO income LUX 0.2 0.2 THA 0.1137 0.1064 0.1 0.1 LAO Backward -Linkage Backward-Linkage 0.0 0.0 ID 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 1.0 1.0 NLD 0.1138 Forward-Linkage 0.1207 2017_Min Forward -Linkage 0.9 NLD 2007_Min 0.9 AUS MON VNM 0.8 NOR 0.8 DEU KAZ RUS MON 0.7 0.7 NOR DNK 0.6 0.6 IRL 0.5 0.5429 0.5 GBR 0.4835 GBR MLT 0.4 0.4 High income High income 0.3 KGZ Lower middle 0.3 income Lower middle income 0.2 Upper middle 0.2 JPN Upper middle income CHN income 0.1 0.1 Backward -Linkage Backward-Linkage 0.0 0.0 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 Source: the UIBE GVC indexes derived from the ADB 2018 ICIO table. developing countries to achieve rapid industrialization through heterogeneous across industries. Communication, financial and joining GVCs. business services, as well as trade and transportation services, Generally speaking, industry groups in manufactures have have much higher GVC participation rates than other domestic higher average GVC participation intensity than industry groups services such as education, health care and personal services, in mining and services (see the scatter plots of backward and because the former are critical inputs in the modern production forward participation rates across countries—Figures 1.7, 1.8 process. and 1.9). In the mining sector, which is the main source of raw GVC participation rates also differ significantly by geo - materials input in the early stages of production, the forward graphic region5. Figures 1.10-1.12 report both forward and back- participation ratio is generally higher than backward participa- ward GVC participation intensities and their inter- and intra-re- tion for most countries, while in other services (utilities, educa- gional shares for manufacture industries in the three major tion, health care and domestic services), which are closer to the supply chain blocks (North America, Europe and Asia). In each final consumer and placed at the final stage of the production figure, the very last pair of columns are the GVC participation chain, the backward participation is higher than forward par- rates in levels and the previous columns are the decomposition ticipation for most countries. In manufactures, higher R&D and across regions. For example, in Figure 1.10, which pertains to knowledge intensities are associated with a higher GVC partic- Asia, the bar for Asia shows the share of intra-regional activities ipation rate (see above). In services, GVC participation is also in Asia’s total GVC participation, while the other bars show the 20 • Technological innovation, supply chain trade, and workers in a globalized world participation of other regions in Asian GVCs, either as suppliers In “Factory Asia”, the increase of cross-country production (backward linkages for Asia) or purchasers (forward linkages for sharing activities in the last decade was led by intra-regional Asia). The light- and dark-colored portions of the bar show the complex GVC activities. This share increased from 38.5%/39.6% shares of different groups inside the region (the light-colored of Asia’s total forward/backward complex GVC activities in portions represent East Asia and Western Europe, and the dark 2000 to 43.9%/46.2% in 2017. Another notable development color portions represent the Rest of Asia and Eastern Europe). was the market-driven enlargement of “Factory Asia”, as more Generally speaking, the higher the degree of economic inte- Asian lower middle-income countries were integrated into gration in a regional production network, the higher the intra-re- Asian production network during this period. In the “Rest of gional GVC activities. In 2000, “Factory Europe” had the highest Asia”, the shares of forward and backward GVC activities rose degree of economic integration, so its share of intra-regional from 10.2% to 11.8% and from 16.6% to 19.4%, respectively. GVC activities is the highest among the 3 regional production However, the importance of North America and Europe as networks; North America ranks second and Asia third. However, both destinations of Asia’s GVC exports (Figure 1.10, forward ten years after the financial crisis, along with the rising scale of GVC activities) and sources of Asia’s GVC imports (Figure 1.10, the regional economy, the share of intra-regional GVC activities backward GVC activities) has declined. in “Factory Asia” exceeded that of “Factory North America”, In Europe, the decline in complex GVC activities represent- especially in complex GVC participation. In contrast, the share ing the breadth of regional production linkages is much more of intra-regional GVC activities has declined in both “Factory than that of simple GVC activities. In particular, the share of Europe” and “Factory North America” and their share of inter-re- intra-regional complex forward GVC participation decreased gional production sharing activities has increased, especially by 6.7 percentage points in the last decade, from 47.6% to their GVC linkage with “Factory Asia”. 40.9%, and intra-regional backward complex cross-border FIGURE 1.10 Forward and backward (simple/complex) GVC participation, share of intra-and inter-regional GVC activities in manufacturing, (%), 2000 and 2017, Asia Simple forward GVC activities in manufacturing Complex forward GVC activities in manufacturing 50.0 50.0 45.5 43.9 2000 2017 45.0 42.0 45.0 38.5 40.0 The rest of Asia 40.0 The rest of Asia 35.0 35.0 30.0 27.2 27.4 30.0 Eastern EU 25.0 East Asia 25.0 22.8 22.5 20.5 20.0 Eastern EU 18.1 17.5 20.0 East Asia 18.2 18.0 15.6 15.0 13.3 15.0 9.0 10.4 11.2 10.0 10.0 7.7 7.8 Western EU 5.0 Western EU 5.0 0.0 0.0 Asia EU NAFTA ROW Simple Asia EU NAFTA ROW Complex GVCPt_F GVCPt_F Simple backward GVC activities in manufacturing Complex backward GVC activities in manufacturing 50.0 48.5 50.0 46.2 45.0 41.1 2000 2017 39.6 40.0 The rest of Asia 40.0 The rest of Asia 35.0 29.6 Eastern EU 30.0 30.0 24.8 23.7 25.0 East Asia 21.9 21.8 21.8 20.0 Eastern EU 20.0 East Asia 16.8 17.8 15.0 15.0 12.2 12.8 9.7 9.3 10.0 6.9 7.8 7.4 10.0 5.0 Western EU Western EU 0.0 0.0 Asia EU NAFTA ROW Simple Asia EU NAFTA ROW Complex GVCPt_B GVCPt_B Note: the last set of bars represent the overall GVC participation ratios for Asia in 2000 and 2017. The country groups refer to footnote 5. Source: the UIBE GVC indexes derived from ADB 2018 ICIO tables. Recent patterns of global production and GVC participation • 21 FIGURE 1.11 Forward and backward (simple/complex) GVC participation, share of intra-and inter-regional GVC activities in manufacturing, (%), 2000 and 2017, Europe Simple forward GVC activities in manufacturing Complex forward GVC activities in manufacturing 62.4 62.5 65.0 Eastern EU 65.0 59.6 2000 2017 60.0 60.0 Eastern EU 55.0 50.0 55.0 50.0 50.0 45.0 45.0 40.0 40.0 35.0 35.0 The rest The rest 30.0 30.0 Western EU of Asia 25.0 of Asia 20.8 19.7 25.0 Western EU 18.9 20.0 16.4 15.3 16.6 20.0 12.9 17.3 15.0 10.2 12.9 12.1 15.0 12.8 11.8 13.3 12.6 9.7 10.0 10.0 5.0 East Asia 5.0 East Asia 0.0 0.0 Asia EU NAFTA ROW Simple Asia EU NAFTA ROW Complex GVCPt_F GVCPt_F Simple backward GVC activities in manufacturing Complex backward GVC activities in manufacturing 63.6 64.6 65.0 60.0 65.0 Eastern EU 60.0 Eastern EU 60.0 2000 2017 55.0 55.0 50.0 46.3 50.0 45.0 45.0 40.0 40.0 35.0 35.0 The rest The rest 30.0 of Asia 25.0 30.0 23.1 25.0 25.0 of Asia Western EU 20.0 16.2 15.3 15.0 20.0 16.6 Western EU 15.1 12.5 11.2 11.8 12.3 13.3 11.5 15.0 9.7 15.0 10.7 10.0 10.0 7.4 5.0 5.0 East Asia East Asia 0.0 0.0 Asia EU NAFTA ROW Simple Asia EU NAFTA ROW Complex GVCPt_B GVCPt_B Note: the last set of bars represent the overall GVC participation ratios for Europe. Source: the UIBE GVC indexes derived from the ADB 2018 ICIO tables. production sharing activities fell by more than 8 percent- In North America, the share of intra-regional complex GVC age points, from 41.1% to 33.0%. This was mainly due to the activities in forward/backward linkages fell by 6.7% and 8.1% relative decline of intra-regional GVC linkages in Western from 2000 to 2017, respectively, although the share of intra-re- Europe, since this share in Eastern Europe increased during gional simple GVC activities changed slightly. The concomitant this period. The shares of inter-regional production sharing rise in the share of inter-regional complex activities reflects the activities between Europe and Asia and Rest of the World also more globalized supply chains in North America today compared increased; the manufacturing links between Europe and Asia to 17 years ago (recall that complex GVC activities involves prod - are more reflected in the complex GVC activities, and the man - ucts that cross national borders at least twice, which has been ufacturing links with Rest of World are more reflected in the the most important driving force behind globalization). More - simple GVC activities. For instance, the share of Asia as the over, the development is not only reflected in the manufacturing destination of Europe’s complex GVC exports and the share sectors, but also in services sectors. For instance, in telecommu- of Asia as the source of Europe’s complex GVC imports both nication, financial and business services, North America’s share increased by over 4 percentage points, from 12.9% to 17.3% of both GVC exports to and GVC imports from Asia and Europe and 12.3% to 16.6%, respectively. East Asia contributed 79.9% exceeded its share of intra-regional GVC activities in 2017, par- and 81.4% of these changes, respectively. The share of Rest of ticularly for complex GVC activities.6 This reflects the intensive the World as the destination of Europe’s simple GVC exports outsourcing of services from the United States to Asian countries and as the source of Europe’s simple GVC imports increased (such as India and Philippines), and the tightly linked financial and from 12.1% to 20.8% and 15.0% to 25.0%, respectively during business service supply chain activities between North America this period. and Europe. 22 • Technological innovation, supply chain trade, and workers in a globalized world FIGURE 1.12 Forward and backward (simple/complex) GVC participation, share of intra-and inter regional GVC activities in manufacturing, (%) 2000 and 2017, North America Simple forward GVC activities in manufacturing Complex forward GVC activities in manufacturing 60.0 60.0 51.5 49.0 50.0 50.0 47.6 2000 2017 40.9 40.0 40.0 The rest The rest 30.0 of Asia 30.0 of Asia Eastern EU Eastern EU 22.7 22.2 23.6 18.6 19.8 17.0 19.8 19.4 20.0 20.0 12.9 11.4 8.7 10.2 10.8 12.7 10.0 East Asia 10.0 East Asia Western EU 6.5 7.0 Western EU 0.0 0.0 Asia EU NAFTA ROW Simple Asia EU NAFTA ROW Complex GVCPt_F GVCPt_F Simple backward GVC activities in manufacturing Complex backward GVC activities in manufacturing 60.0 60.0 50.0 48.2 49.0 2000 2017 50.0 41.1 40.0 40.0 Eastern EU The rest 33.0 The rest 30.7 30.0 of Asia of Asia 30.0 25.5 Eastern EU 21.2 23.0 19.4 20.9 19.8 20.0 16.1 16.3 20.0 13.5 13.0 10.0 8.9 6.3 7.5 10.0 8.2 9.6 East Asia East Asia Western EU Western EU 0.0 0.0 Asia EU NAFTA ROW Simple Asia EU NAFTA ROW Complex GVCPt_B GVCPt_B Note: the last set of bars represent the overall GVC participation ratios for North America. Source: the UIBE GVC indexes derived from the ADB 2018 ICIO table. 2. The changing distribution of value-added through a recently-developed input-output based methodology along typical GVCs7 (see Ye, Meng et.al., 2015; Meng, Ye et.al., 2017), the relationship between value-added distribution and GVC participation can be This section uses “smile curve” analysis to discuss how the distri- empirically identified and drawn for various GVCs.8 bution of value added across countries and industries via GVCs In Figures 1.13 and 1.14, we take the final goods exports of changes when more and more developing countries are partici- Mexico’s ICT industry and Japan’s auto industry as examples. pating in global production networks. The y-axis of these figures shows compensation per employee The concept of the smile curve was first proposed around (a proxy for technology level or a first-order approximation of 1992 by Stan Shih, the founder of Acer, a technology company labor productivity)9 in constant 2000 U.S. dollars, and the x-axis headquartered in Chinese Taipei (Shih 1996). In the personal denotes distance showing how far a specific participating coun- computer industry, Shih observed that both ends of the value try and industry pair in the particular GVC of interest is away from chain bring higher value added to the product than the middle global consumers.10 The data used is from the WIOD (2016 ver- part. In business management theory, the smile curve is a graph- sion), which covers 43 economies and 56 industries over 15 years ical depiction of how value added varies across the different (2000-2014), with the total number of GVC participants (43 × 56 stages of bringing a product to the market in a manufacturing = 2,408) represented as circles in these figures. The size of the industry. The logic of the smile curve has been widely used in circle represents the absolute value added created by joining the case studies of individual firms, but rarely identified, measured, corresponding GVC (the minimum threshold for inclusion in the and evaluated at the country level by using real data with explicit figure is 0.1% of the total value-added gain measured in million consideration of GVCs. As we show in the 2017 GVC Develop- U.S. dollars). The smooth line is fitted by local polynomial regres- ment Report, by borrowing the image of the smile curve and con- sion–smoothing weighted by its value-added volume, and the sistently measuring both the value-added gains from GVC par- shadowed area represents the confidence interval around the ticipation and the distance between producers and consumers smooth line. Using the estimated smile curve can enhance our Recent patterns of global production and GVC participation • 23 FIGURE 1.13 Mexico’s ICT final goods exports related value chain, 2000 and 2014 Note: y-axis represents the compensation per employee in constant thousand U.S. dollars (base year: 2000); the x-axis represents the length of the correspond - ing production chain in average stages of production. 24 • Technological innovation, supply chain trade, and workers in a globalized world FIGURE 1.14 Japan’s auto final goods exports related value chain, 2000 and 2014 Note: -axis represents the compensation per employee in constant thousand U.S. dollars (base year: 2000); the x-axis represents the length of the corresponding production chain in average stages of production. Recent patterns of global production and GVC participation • 25 understanding of the participants (countries and industries) of a of Japan’s auto industry from traditional mass producer to mass specific GVC as well as their positions and economic gains from customizer, based on digital technology and artificial intelli- the chain. gence, similar to what happened in German’s auto industry (as The plotted GVC for Mexico’s ICT (MEX17) final goods exports reported in the 2017 GVC Development Report). The mass cus- to the world market in 2000 clearly appears as a smile curve (see tomized manufacturing stage accounted for a relatively large Figure 1.13). The main participants in the pre-fabrication stages portion of the total value gain, while the traditional high-end (upstream) of this value chain comprise many US industries, such design and sales functions accounted for only a small portion as ICT (USA17), wholesale trade (USA29), legal accounting, head of total value-added creation, mostly by producers from foreign offices, management consultancy activities (USA45), electrical countries. This is contrary to the typical intuition from the smile equipment (USA18), fabricated metal products (USA16), machin- curve, in which traditional manufacturing stands only at the low ery and equipment n.e.c. (USA19), and chemicals (USA11); some end of the value chain, such as Mexico’s ICT final goods exports Mexican domestic industries, such as chemicals (MEX11), machin- in 2000. But it could also reflect the ongoing structural change in ery and equipment n.e.c. (MEX19), electrical equipment (MEX18); GVCs, such as the emergence of the customer to manufacturing and several Japanese industries such as ICT (JPN17), basic metals (C2M) business model in several industries. The most important (JPN15), and fabricated metal products (JPN16). The main partic- changes between 2000 and 2014 were the increasing number ipants in the post-fabrication stages (downstream) comprise US and variation of foreign participants and the increasing length industries such as wholesale trade (USA29), retail trade (USA30), of the curve. In 2000, the United States and Germany dominated warehousing (USA34) and so on. Most participating industries foreign participants upstream and downstream, while in 2014, upstream and downstream in Mexico’s ICT exports-related value more industries from foreign countries were involved, especially chain are from the US and Japan, countries with high levels of industries from China. This clearly reflects the increasing diversity labor compensation, while most participating industries in the and complexity of international fragmentation in Japan’s auto middle of the value chain are from the Mexico’s domestic indus- production. In addition, given the increase in labor compensa- tries with low levels of labor compensation. Therefore, the whole tion and absolute volume of value-added gain in Japan’s auto chain naturally appears as a smile curve. industry, along with the relatively low level of labor compensa- However, the shape of the curve changed significantly in 2014, tion of upstream and downstream participants from China, the from a smile curve to a kind of “W” curve. At least three fac- slope of the entire curve became much steeper. This implies that tors contributed to the remarkable changes in the shape of this Japan’s auto sector has enhanced its comparative advantage by smile curve. One was the rapidly increasing presence of Chinese outsourcing more upstream and downstream tasks that were for- industries in Mexico’s value chain upstream. As seen in 2014, merly done by Japanese employees to China through GVCs. many Chinese industries with low compensation per employee, such as ICT (CHN17), wholesale trade (CHN29), mining (CHN4), electrical equipment (CHN18), machinery and equipment n.e.c. 3. The topology and structure change of GVC (CHN19), and basic metals (CHN15), replaced other countries’ production and international trade11 positions in the Mexican value chain. Those Chinese industries became some of the main players, with a large value-added gain Network analyses have been used widely to visually simplify the in the pre-fabrication stage of this value chain. This reflects the image of GVC activities given their increasing complexity (see fact that producing ICT exports in Mexico used more Chinese Ferrarini, 2013; Ferrantino and Taglioni, 2014; Zhou, 2016; Xiao et intermediate inputs directly and indirectly. The second factor al., 2017). Unlike the literature in international trade-related net- was the rapid technological upgrades that occurred in the US work analyses, we separate bilateral trade flows across countries ICT industry (USA17), indicated by the simultaneous increase in into three types of networks (traditional trade networks, simple compensation per employee and maintenance of a large volume GVC networks and complex GVC networks) based on the pro - of value-added gain. This implies that Mexico’s ICT produc- duction activity decomposition method proposed by Wang et tion was highly dependent on high-tech US intermediates. The al. (2017) (see Box 1.2).12 The network analysis in this section third factor was the increasing volume of value-added gain by provides a new view about how trade and production sharing Mexico’s service industries (legal accounting, head offices, man- activities are concentrated among bilateral trade partners, as agement consultancy activities (MEX45); other professional, well as the changing interdependency among trading partners scientific, technical, and veterinary activities (MEX49)) in the in different networks. pre-fabrication stage. All these developments may have also One conclusion of the network analysis, which covers 62 coun- contributed to the overall expansion of Mexico’s ICT value chain, tries and 35 sectors from 2000 to 2017, is that the topology struc- as the entire length (x-axis) of this chain increased from 6.8 to 8.3 ture of networks (at the aggregate and individual sector levels) between 2000 and 2014. changes only gradually. Even the financial crisis of 2008 did not Japan’s final auto (JPN20) products exports-related value result in a significant change in the network topology in 2009. chain also experienced a dramatic change from a smile curve to This implies that the structure of global production networks an inverted smile curve-a frown from 2000 to 2014 (Figure 1.14). expressed by the topology of country to country relationships To some extent, this may have reflected the successful transition is resilient, even when economic shocks of a large magnitude hit 26 • Technological innovation, supply chain trade, and workers in a globalized world BOX 1.2 How network graphs are drawn in our GVC analysis We draw two types of networks from ADB ICIO data to arrow of the linkage shows the direction of the value-added identify the hubs of various networks from importer and flow. Two tests are used to determine whether a linkage exporter perspectives. One takes a specific country as a line appears between trading patterns (taking the case of supply hub if the majority of the imports by most countries supply hub related networks as an example): 1) if country A in the network is from that country. Another takes a spe- takes the largest share in country B’s value-added imports, cific country as a demand hub if the majority of the exports a linkage will be shown from A to B; or 2) if country A’s share from most countries in the network goes to that country. In in country B’s value-added imports is larger than 25%, a our network figures, the size of the bubble represents the linkage will be shown from A to B. The first standard is the share of a country’s value-added exports or imports in total so-called “Top1” threshold widely used in network analyses value-added exports or imports. The share of value-added to identify the most important linkages. The second stan- flow between each trading partner in total value-added dard is used to adjust the density of the network, in order to flow is represented by the thickness of the linkage. The avoid losing other important linkages.14 the global economy. Therefore, this analysis considers only the (Russia, Brazil, India). When comparing the magnitude of the long-term change from 2000 to 2017.13 We consider both the value-added flows across countries over time, it is easy to see networks for the aggregate economy (all goods and all services), that the linkages between China and other main regional hubs as well as selected typical GVC sectors (textile, ICT, and ser- as well as its surrounding countries became much thicker. vices) as examples. The middle-left part of Figure 1.15 shows the simple GVC trade networks for all goods and services in 2000. Compared 3.1 Supply hubs of value-added trade to the traditional trade networks, the US was a global supply Supply hubs of value-added trade at the aggregate level hub with important outflow linkages to the other two regional As shown in the upper-left part of Figure 1.15, the three large hubs, Germany and Japan. Some remarkable differences can regional supply hubs in the traditional trade networks in 2000 be observed within each region. For example, compared to the were the US, Germany and Japan. Obviously, these three hubs traditional trade networks, more extra-regional countries had have very important linkages with their neighbor countries. the US as their main supplier of value added through simple The US has strong linkages to its two North American part- GVC trade. This also reflects the fact that US intermediate prod - ners, Canada and Mexico, the two large Asian countries, Japan ucts were greatly used as inputs for many countries to produce and the Republic of Korea, and Brazil, India and Australia. domestically-used final products. The UK, which was a sub-hub Japan can also be considered as a regional supply hub in the in Europe in the traditional trade networks, becomes a sub-hub Asia-Pacific region, since the US, China, the Republic of Korea, with important linkage with the US in the simple GVC trade Chinese Taipei and many Asian countries have Japan as their networks. most important value-added supplier through final product A remarkable structural change in the simple GVC trade net- trade. Germany is the largest supply hub in the European area, works occurred between 2000 and 2017 (the middle-right part because the majority of value-added imports in final products of Figure 1.15). In 2017 there was no longer any important link- by almost all European countries is from Germany. When zoom- age between any two hubs, as simple GVC activities became ing in the figure, we can also find some small regional hubs in more concentrated within Europe, North America and Asia. The the European area, such as the UK, France, Italy, Spain, Bel - US and Germany connected to each other indirectly through gium, and Russia, and in the Asia-Pacific region, such as China, the Netherlands. The number of countries with strong linkages the Republic of Korea, India, Thailand, and Singapore, who have to the US decreased dramatically, as most of the surrounding more than two linkages with other countries. linkages moved to China. Germany maintained its position as Comparing the situation of 2000 to that for 2017 (the upper- a regional supply hub in Europe with strong linkages to more right part of Figure 1.15), it seems there was no significant countries. China replaced Japan and part of the US position change in the network topology in Europe and North America, and became the second largest supply hub in terms of both the but dramatic changes occurred for Asia: China took over Japan’s magnitude of its value-added exports and the number of strong position and became a global supply hub of value-added export linkages to other countries. through final products trade. China not only had important link- Looking at the evolution of the complex GVC trade networks ages with other hubs (the US and Germany), but also with its from 2000 to 2017 (see the bottom panel of Figure 1.15), trade Asian neighbors (Japan, the Republic of Korea, Chinese Taipei, became more concentrated among regional trading partners, and almost all Asian countries) and other emerging countries and there was no important direct linkage among regional hubs. Recent patterns of global production and GVC participation • 27 The US connected with Germany indirectly through two coun- Obviously, there were many regional supply hubs in the traditional tries, Luxembourg and the UK. In addition, the volume of Chi- trade networks in 2000. There were three main regional supply na-made intermediates used as inputs for its downstream coun- hubs in Europe, Germany, Italy and the UK, who exported textile tries to further produce exporting products increased rapidly sector value-added to their trading partners through final goods over the period as seen from the bubble size change for China. trade. Germany and the UK connected indirectly through Turkey. India was also a sub-supply hub with inflow linkage from the UK Supply hubs of value-added trade in various networks for and outflow linkages to Nepal and Bangladesh. The presence of selected sectors Italy, as the most traditional country with strong fashion sectors, The topologies and changes in structure over time in individ- can be clearly identified in these networks. This is very different ual sectors may differ considerably from the aggregate patterns from the networks at the aggregated level shown in Figure 1.15, in shown above. Figure 1.16 shows the textile sector related networks. which Italy’s presence in the textile sector is largely masked. FIGURE 1.15 Supply hubs of trade in value-added in various networks at the aggregate level MLT PRT Traditional trade networks (all goods and services) NOR GRC LUX NLD ESP PRT MLT SWE ITA CYP BEL FRA LTU ITA FRA CHE IRL ESP NLD GBR PHI PAK IRL GBR LVA HKG AUT CYP LAO BRA MEX CHE MON KAZ DNK KGZ FIJ BEL POL CAN NOR RUS BRA CAN SWE FIN DEU USA CHN USA EST CHN FIN DEU IND AUT MAL CZE BTN JPN MEX TUR PAK HKG AUS JPN DNR KOR NPL VIE HUN AUS KOR SIN RUS CZE KAZ ROM EST SRI TAP FIJ HUN IDN IND BRN MDV SYN BGR LTU SRI SIN THA MAL SVN TAPTUR HRV VIE IDN ROM THA BTN NPL LVA HRV KGZ MON SVK SVK PHI BRN LUX BGR GRC CAM CAM BAN MDV 2000 BAN LAO 2017 Simple GVC trade networks (all goods and services) LUX PRT ESP MLT IRL CHE BEL PRT CHE BRA AUT ITA LUX BEL NOR AUT ITA GRC MLT BRA ESP IRL GBR NLD CYP NOR SWE FRA SWE FRA GBR CAN USA CAN FIN DNK NLD DNK MEX CZE DEU SIN USA SR ROM DEU MEX POL CHN HKG FIN CZE CHN KGZ JPN MDV KAZ JPN PHI EST MON IND CAM HUN CAM AUS HUN TAP KGZ SVK LAO BAN LAO KOR ROM SVK SRI RUS BGR SIN NPL PAK BAN AUS THA HRV VIE KOR FIJ BGR RUS MON SVN TUR EST BTN GRC IDN LVA THA TAP MDV KAZ FIJ MAL IDN PAK SVN HRV TUR IND LTU VIE HKG CYP LTU PHI MAL BRN LVA BTN BTN NPL 2000 2017 Complex GVC trade networks (all goods and services) NOR BEL PRT DNK BEL BGR ESP SWE PRT CHE NLD MLT AUT ITA LUX GRC BRA POL ESP IRL ITA NLD FRA EST CAN AUT LUX CYP GBR IRL FIN FRA PAK MON SRI GRC GBR MEX CHE PHI POL VIE LAO SWE BRA NOR CAM CZE HKG DEU USA BRN CZE KAZ DEU CHN USA CHN HKG DNK KAZ LTU CAN HUN HRV JPN ROM LVA JPN MON PHI TAP ROM BAN KOR IND KGZ RUS LAO CYP FIN BGR SIN BTN RUS KOR BAN MEX SVN TUR VIE AUS HUN KGZ THA SIN SVK LVA TAP NPL LTU PAK SVK MAL AUS MLT THA FIJ SVN HRV EST TUR IDN MAL CAM IND IDN FIJ BRN NPL MDV BTN 2000 2017 Note: the size of the circles represents the magnitude of value-added exports. The volume of value-added flow between each pair of trading partners is repre - sented by the thickness of the line linking the two. Source: Meng et al. (2018) based on the UIBE GVC indexes derived from the ADB 2018 ICIO table. 28 • Technological innovation, supply chain trade, and workers in a globalized world FIGURE 1.16 Supply hubs of trade in value-added in various networks for the textile sector Traditional trade networks (textile sector) NLD CZE DNK SVE BEL FIN PRT POL SVN PRT ESP LVA FIN IRL GBR NOR ROM ESP BEL EST KGZ BGR FRA FRA SWE KAZ CYP KAZ LTU GRC FIJ GRC BRA ITA NOR DNK SVN AUS RUS SRI HUN HRV ITA HKG BRA CHE MEX HRV VIE AUT POL NLD DEU CHN USA KGZ SVK HKG USA MEX CAN AUT DEU LUX MLT ROM CHN FIJ TUR BGR HUN TUR PAK KOR CAN MON SIN LTU JPN VIE CZE SIN MON NPL EST GBR MAL CHE PHI THA LAO CAM IRL TAP IDN SVK LVA LUX IND BAN BRN THA RUS MLT LAO IND NPL MAL KOR CYP CAM BTN IDN MDV AUS MDV BTN PHI TAP JPN PAK BRN BAN 2000 2017 Simple GVC trade networks (textile sector) FIN DNK LTU SVE EST LVA EST IRL NOR KAZ DNK PRT FRA NLD SWE HKG NOR MLT ROM BEL BEL FIJ HRV ESP GBR TUR IRL KGZ FRA PRT SVN ITA GBR SIN ESP BAN MAL MON HKG AUT BRA GRC ITA MLT CZE NPL MEX AUT KGZ POL ROM BRN MEX DEU USA LUX BOR CHE CHN NLD BGR SVK HUN CAN LUX IDN CHE DEU CHN BRA USA RUS TUR POL AUS IDN CAN GRC KOR BRN VIE SIN CYP CZE FIJ KOR LAO RUS JAP IND MON CAM HUN JAP TAP VIE KAZ PHI NPL TAP FIN LVA BTN CYP SRI SVN LTU HRV SVK MAL IND LAO MDV AUS THA PHI THA PAK SRI CAM PAK MDV BTN BAN 2000 2017 EST Complex GVC trade networks (textile sector) DNK IRL IRL NOR GRC LUX LTU BEL NLD BEL NOR ROM BGR SWE SWE ESP MLT DNK PRT GBR PRT GBR FRA SVN TUR FIN ESP TUR FRA LVA GRC HRV AUT ITA EST BGR ITA NPL CHE CYP NLD MAL KAZ ROM MON HKG RUS MEX POL POL KGZ HKG MLT LUX CZE KGZ DEU FIJ CHN USA DEU CZE BRN CHN MEX BRA CHE BRA SIN USA CAN IDN AUT IDN LAO BAN KOR HUN KOR HUN SVK MON CAM FIN LVA AUS SIN CAN PHI JAP PAK HRV JAP VIE SVN LTU SVK FIJ IND TAP KAZ MAL NPL THA CYP VIE RUS TAP BTN SRI PHI AUS IND BRN SRI PAK THA LAO CAM MDV MDV BTN BAN 2000 2017 Note: the size of the circles represents the magnitude of value-added exports. The volume of value-added flow between each pair of trading partners is repre - sented by the thickness of the line linking the two. Source: Meng et al. (2018) based on the UIBE GVC indexes derived from the ADB 2018 ICIO table. The structure of textile networks changed dramatically from China had a significant comparative advantage in exporting appar- 2000 to 2017. China became the largest and the unique global els, given its large labor force with lower wages, while FDI inflows supply hub; in the figure China has pushed away all the other from developed countries helped make China the largest exporter regional hubs and surrounding countries to the periphery of the of textile and apparel products in the world. By 2017, China’s tex- traditional trade networks. This phenomenon is consistent with tile sector played a dominant role in traditional trade networks as the fact that textile final goods made in China can be found every- well as the simple and complex GVC trade networks. This implies where in the world. Mixed reasons may explain this phenomenon. that China is gradually upgrading its textile sector, and thus can China already had substantial textile production capacity in its export more intermediates to other countries through GVC trade. early stage of development. Thus it easily joined GVCs by export- Although China has grown to become a new rival in GVC trade ing more final textile products when tariff and non-tariff barriers through upgrading of intermediate exports of textile, Italy can still decreased in other countries after its WTO accession. Moreover, maintain its position as a regional hub especially in the complex Recent patterns of global production and GVC participation • 29 FIGURE 1.17 Supply hubs of trade in value-added in various networks for the ICT sector EST NPL MLT Traditional trade networks (ICT sector) LVA NOR NLD FIN FIN ITA SWE DNK AUT PHI GRC PAK CYP NOR MLT HKG BEL CHE FRA IRL ITA FRA LUX CAM ESP LVA AUT CHE EST LAO DNK GBR HKG BRA CAN SWE NLD FIJ POL HUN GBR BRA CAN HUN POL IRL DEU CHN MEX PRT CZE CZE MEX USA ESP PRT DEU CHN ROM USA ROM KAZ IND MON SVN BEL SVK FIJ JPN MAL BAN KOR KOR AUS BTN SVN SIN LTU BAN SRI BRN LUX BGR LAO THA HRV TAP HRV JPN SVK MDV LTU RUS GRC BGR CYP TUR SRI IDN NPL VIE SIN PAK KGZ AUS TAP KAZ KGZ THA VIE IND PHI CAM MDV TUR IDN MAL BTN BRN MON 2000 2017 EST Simple GVC trade networks (ICT sector) LVA SWE NLD FIN CHE AUT NOR MLT FIJ SRI ITA CYP AUT GBR MLT PAK BAN ESP IRL KGZ SWE MON LAO BEL FRA ESP ITA GBR ROM KAZ BRA CHE KGZ IRL DNK POL CAN HKG CAM POL MEX CZE MEX NOR DNK NLD DEU USA FRA HUN CAN FIN CHN DEU CHN USA BRN MON THA HUN CZE RUS TUR RUS AUS BEL THA JPN CYP KOR BRA ROM NPL HKG PHI KOR SVN AUS PRT TUR LUX PRT LUX NPL SVN BGR SRI BRN TAP MAL MDV SVK HRV GRC TAP BGR GRC IND IND JPN VIE SVK HRV BTN IDN LAO SIN LVA EST LTU PHI FIJ BTN IDN MAL BAN PAK MDV VIE CAM 2000 2017 NLD Complex GVC trade networks (ICT sector) AUT SWE CHE EST BEL ITA MLT IRL NOR LVA CAN CZE GBR AUT MLT PAK LAO FIJ BRA MON SRI LUX SWE DNK FIN FRA ITA MDV KGZ CYP ESP CHE ESP GBR MEX CAM MEX POL HKG IRL POL BRA CAN ROM NOR NLD DEU USA DNK FRA CZE KGZ CHN HUN HUN DEU CHN USA BRN PRT BAN IND BEL HKG JPN PHI ROM NPL FIN KAZ LUX MAL BTN IDN TAP SVN MON RUS PRT RUS SIN KAZ PAK AUS SVK SRI BGR KOR HRV TUR KOR SVK PHI LTU GRC BGR VIE SVN HRV CYP BAN TAP LVA JPN MAL THA GRC AUS TUR NPL IND SIN EST LTU IDN BRN VIE BTN LAO THA CAM 2000 MDV FIJ 2017 Note: the size of the circles represents the magnitude of value-added exports. The volume of value-added flow between each pair of trading partners is repre - sented by the thickness of the line linking the two. Source: Meng et al. (2018) based on the UIBE GVC indexes derived from the ADB 2018 ICIO table. GVC trade networks. This indirectly reflects the strength of Italy’s keeping just important linkages with a limited number of coun- technology in producing complex textile products compared to tries. Japan’s presence decreased dramatically, as it moved from other European countries whose presences have declined in com- a global supply hub in the traditional trade networks and regional plex GVC trade networks over time. supply hub in the simple GVC networks in 2000 to the periph- The network topology for ICT experienced dramatic changes ery of the Asia-Pacific region in 2017. These changes reflect the from 2000 to 2017 (Figure 1.17 shows the ICT sector’s val- so-called industrial hollowing15 out in the US and Japan’s ICT ue-added exports related networks). In 2017, China took over sectors (especially for final goods production), accompanied by Japan’s position, becoming a global supply hub in both tradi- large scale FDI from these countries to China. The latter made an tional trade and simple GVC networks. Inside Asia in 2017, Japan, important contribution to China’s ICT development, since even the Republic of Korea and Chinese Taipei played very important in recent years more than half of China’s ICT exports were pro - roles as sub-hubs. The US became a largely regional supply hub, duced by foreign-owned enterprises. 30 • Technological innovation, supply chain trade, and workers in a globalized world FIGURE 1.18 Supply hubs of trade in value-added in various networks for the services sector Traditional trade networks (services sector) LUX PRT LVA PRT EST NLD MLT LTU NOR DNK BEL GRC ESP MLT BEL ITA CYP FIN POL ESP FRA CYP ITA FRA IRL CHE IRL CHE NLD GBR AUT SWE KAZ BRA NOR MEX LUX BRN GBR CAN SWE DNK PHI ROM MEX DEU USA AUT BAN DEU SIN HUN USA EST CHN CHN JPN TAP FIN CZE BRA CAN KOR POL KOR MDV FIJ THA SRI SIN JPN IND HKG SVK PAK IND HUN TUR MDV CZE RUS LTU BGR LAO SRI PAK AUS NPL MAL ROM SVN THA NPL SVK BGR KAZ VIE TAP BTN HRV GRC TUR VIE BTN SVN HRV LVA MAL FIJ IDN MON KGZ IDN THA BRN CAM MON AUS CAM RUS HKG LAO BAN KGZ 2000 2017 Simple GVC trade networks (services sector) PRT BEL NOR ESP MLT CYP PRT IRL EST CHE LUX LVA POL ITA FRA EST CHE NLD GBR AUT ESP CYP NOR IRL GRC ITA GRC BRA LUX SWE FIN AUT CZE FRA MLT SWE NLD GBR BRA CZE BEL DNK CAN HUN MEX ROM MEX DEU DEU USA DNK CHN USA CHN CAN BRN ROM KAZ POL FIN JPN PHI PHI SVN TUR IND SVK KGZ JPN VIE KOR BGR SVK CAM TAP MAL RUS TAP AUS MAL BTN HUN LAO KOR SIN IDN BGR IND THA LAO KGZ HRV PAK SIN HKG HRV RUS MON THA BAN MDV MON SVN TUR AUS LTU LVA CAM BRN FIJ PAK MDV SRI NPL KAZ VIE IDN FIJ BAN LTU HKG SRI 2000 2017 BTN NPL Complex GVC trade networks (services sector) BGR PRT DNK NOR PRT EST ESP SWE IRL EST CHE BEL LUX ESP ITA LUX CYP NOR FRA GBR IRL AUT FIN SWE ITA GRC NLD BEL GRC BRN FRA POL CYP NLD POL BRA GBR BRA MLT MEX CHE HUN DEU MEX USA DEU KAZ LVA USA AUT CHN CAN LTU CHN CZE PHI CZE TUR MON JPN BRN CAN KAZ THA SVN BAN ROM RUS MLT MDV KGZ JPN BAN SVK MAL MAL ROM FIJ IND BTN DNK TUR VIE SIN KGZ HKG FIN LAO TAP HRV SIN NPL AUS RUS IDN KOR KOR MDV LTU THA TAP SRI LAO HUN SVK SVN HRV CAM IDN LVA AUS CAM SRI FIJ BGR MON PAK PAK VIE HKG PHI BTN NPL 2000 2017 Note: the size of the circles represents the magnitude of value-added exports. The volume of value-added flow between each pair of trading partners is repre - sented by the thickness of the line linking the two. Source: Meng et al. (2018) based on the UIBE GVC indexes derived from the ADB 2018 ICIO table. Nevertheless, the US and Japan remained important hubs The US was the largest supply hub for services in 2000 in the in complex GVC networks in 2017, in terms of both the volume traditional trade networks (Figure 1.18 shows the services sec- of value added traded and the number of countries with strong tor’s value-added exports related networks). The US had sig- linkages. The US and Japan were still the main suppliers of com- nificant outflow linkages to Canada and Japan, and indirectly plex intermediate goods used by downstream countries through connected with the other supply hub, Germany, through third complex GVC activities. At the same time, China’s ICT sector countries (Ireland and the UK) in 2000. In 2017, however, the US exported more value added through both simple and complex had few direct outflow linkages going to Asia. In 2017, Germany GVC trades. This provides some evidence of the ongoing indus- maintained its presence as a regional supply hub with import- trial upgrading in China’s ICT industries, since more intermediate ant linkages to other sub-regional hubs (France and Italy), lost its products have been made in China. linkage with the sub-regional hub Russia, and added a linkage Recent patterns of global production and GVC participation • 31 with the new sub-regional hub, Poland, in Europe. China took partners, such as the UK (which joined the European networks as over Japan’s position in Asia and became a large supply hub with a sub-supply hub), as well as Japan, the Republic of Korea and an important presence in exporting service sector value added Hong Kong, China (which have joined the Asia networks as sub- to the US and other Asian economies in the traditional trade net- hubs surrounding China). There was no longer any direct link- works. While China did not export a large amount of services to age between the US and Germany in 2017, but they indirectly the global market directly, China was the largest manufacturing linked to each other through the Netherlands. China took over final goods exporter and the value added of China’s domestic Japan’s role, becoming a regional supply hub with an important services were embodied in these exports. inflow linkage from the US and outflow linkages to other Asian In the simple GVC trade networks, the US maintained its role economies. This implies that China’s services sector directly as the largest supply hub in 2017, but lost some important trading and indirectly exported value added to other Asian economies FIGURE 1.19 Demand hubs of trade in value-added in various networks at the aggregate level Traditional trade networks (all goods and services) NOR SWE FIN LUX BRA ROM BEL MON CYP KAZ AUT NLD ROM EST CYP CAN HRV MEX POL CHE LTU RUS POL AUT EST ESP FIN ESP MEX BRA NOR CHE ITA SRI SWE CAM CZE NLD LVA MDV PRT PRT FRA IRL PAK DNK CAN FRA BEL CAM GRC GBR DNK BAN SVK DEU USA PHI HUN HUN DEU VIE BTN USA CZE RUS CHN BRN LUX IRL SVK KGZ NPL GBR PAK FIJ BGR LTU SIN CHN PHI JPN FIJ AUS JPN SVN TUR KAZ SVN TUR MLT IND VIE AUS IDN IDN MLT BGR HRV LVA MON MDV HKG THA BAN SRI KOR TAP LAO TAP HKG MAL BTN MAL KOR THA KGZ NPL GRC LAO BRN SIN 2000 2017 Simple GVC trade networks (all goods and services) DNK SWE ROM LVA KGZ BGR MLT EST EST PRT HRV IRL SWE MDV MLT NLD RUS IRL HRV GBR AUT BEL ITA CHE MEX CHE ESP ITA CYP POL BEL KAZ FIN FIN MDV PRT NLD FRA BRA AUT DNK LVA FRA CAN POL ESP GBR MEX ROM GRC NOR USA SRI CZE DEU CZE DEU USA CAN PHI LUX TUR SRI CHN BGR HUN BRA JPN BRN IND HUN MON CHN BRN BAN SVK LUX CAM GRC CAM VIE JPN SVN KOR PAK MON BTN SVK KGZ SVN LTU MAL KOR IND PHI RUS NPL BAN THA SIN FIJ PAK AUS TAP TUR HKG THA KAZ IDN VIE LAO IDN CYP FIJ BTN NPL HKG LTU SIN MAL TAP LAO 2000 2017 Complex GVC trade networks (all goods and services) PRT SRI LTU EST KAZ SRI MDV BAN CYP MDV CAM PHI BGR NOR FIN BTN LAO GBR ROM DNK SIN ESP PRT POL ITA BEL HRV IRL IRL LUX NLD FRA LUX BEL GBR MAL ITA KGZ NLD SWE THA CHE FRA ESP CZE AUT CHE EST SWE HUN BRA LVA HUN MEX CHN POL CYP DNK DEU CHN USA DEU USA MEX CAN BRA CZE MLT KOR MON TAP AUT IDN CAN PAK JPN SVN HKG SVK MON JPN TUR THA VIE SIN PHI RUS BRN KOR GRC FIN AUS TAP IND MAL BRN SVK PAK VIE SVN IND KAZ NOR BAN HKG MLT FIJ RUS LVA KGZ TUR NPL ROM BGR LTU GRC FIJ LAO CAM NPL AUS BTN IDN HRV 2000 2017 Note: the size of the circles represents the magnitude of value-added imports. The volume of value-added flow between each pair of trading partners is repre - sented by the thickness of the line linking the two. Source: Meng et al. (2018) based on the UIBE GVC indexes derived from the ADB 2018 ICIO table. 32 • Technological innovation, supply chain trade, and workers in a globalized world FIGURE 1.20 Demand hubs of trade in value-added in various networks for the textile sector Traditional trade networks (textile sector) LAO NPL BRN LAO NPL MDV MON CAM HUN SVK AUT POL AUT CAN SRI GRC KAZ CZE BRN BRA SVN POL MEX CAM CHE ROM ITA BEL SRI NLD CHN MEX IRL BRA PAK HUN NLD PHI LUX CHN HRV BTN MLT DEU CZE CAN USA VIE DEU CHN MDV LUX SIN GBR USA BAN JPN THA SVK BAN MON TUR JPN FRA SIN SVN MLT IDN HKG KOR AUS IDN ITA RUS TUR IND BGR GBR VIE ROM MAL LTU LVA RUS IND FIN PRT TAP PHI CYP THA HRV CHE ESP FIJ BTN FRA SWE KAZ KOR MAL BGR GRC IRL KGZ AUS FIJ TAP DNK EST FIN SWE PRT ESP DNK LTU LVA EST CYP KGZ NOR BEL NOR 2000 2017 BEL Simple GVC trade networks (textile sector) ROM ESP BGR PRT LVA LAO FRA ESP SVN ITA NOR SWE DNK VIE BRA MEX HRV PRT MDV CAN ROM BGR NPL FRA BRA MLT ITA EST IRL POL NLD MON CHE HUN IRL GBR SRI GBR MEX BTN CZE MDV BAN BEL LAO NLD HKG DEU AUT DEU CHN USA CHN USA PAK SIN PAK VIE POL HUN CHE BRN CAN SVK MAL SVK HRV LUX PHI MAL AUT CZE MON AUS JPN FIN JPN THA SIN IND CAM SRI SWE PHI KGZ AUS NOR DNK TAP SVN RUS TAP IND KOR IND KAZ RUS KOR IDN GRC KAZ LUX IDN TUR BRN CYP NPL FIJ CAM TUR BTN FIJ LTU BAN FIN KGZ LVA LTU CYP EST THA GRC HKG 2000 2017 Complex GVC trade networks (textile sector) CYP KAZ SVN MDV MON MLT ROM BGR RUS BTN BGR LVA GBR EST IRL HKG CAM MAL BEL RUS SRI MLT BRN IRL ITA MDV ESP BAN NPL KAZ ESP ITA SVK CZE POL GBR BEL NLD SIN KGZ FRA THA FRA CHE PHI PRT SRI HRV AUT POL ROM USA MEX BRA USA MEX BTN CHN DEU HUN DEU GRC NPL PAK PRT CHN AUS CAN DNK IDN IND KOR NOR SWE DNK KOR MON CAN BRA LTU FIN FIJ IDN THA TUR VIE TAP AUS SWE TAP HUN LUX TUR HKG JPN VIE KGZ NOR MAL FIN JPN CYP AUT SVK LAO PAK IND FIJ EST NLD BAN SIN CZE LUX LTU GRC LVA SVN HRV CAM PHI BRN LAO 2000 2017 Note: the size of the circles represents the magnitude of value-added imports. The volume of value-added flow between each pair of trading partners is repre - sented by the thickness of the line linking the two. Source: Meng et al. (2018) based on the UIBE GVC indexes derived from the ADB 2018 ICIO table. used to produce final goods. However, China still largely relied 3.2 Demand hubs of value-added trade in various networks on US-made intermediate services when producing domestically- Demand hubs of value-added trade in various networks at the used final goods. aggregate level A very similar pattern can also be found in the complex GVC The US was the unique global import demand hub in 2000, with trade networks. One difference is that Germany’s services sector connections to several Asia Pacific economies and some Euro - had a much larger presence in exporting value added through pean counties, and stronger linkages with the regional demand multiple cross-border transactions of intermediate goods in hubs of Germany, the UK and Japan (upper part of Figure 1.19). GVCs. This is probably due to the following fact: Germany has a The structure didn’t change greatly in 2017, except for the dra- high comparative advantage in exporting high-tech and complex matic rise of China as a new regional demand hub in Asia with intermediate goods, which embody value added from the domes- the strongest outflow linkage to the US. A similar pattern can be tic services sector, since producing these high-tech intermediate seen in the change in the simple GVC trade networks (the middle exports requires inputs from the domestic services sectors, such part of Figure 1.19) from 2000 to 2017, except that China became as business supporting services and financial intermediaries. a regional demand hub with more inflow linkages from Asian Recent patterns of global production and GVC participation • 33 economies, as well as from some emerging countries outside Asia still matters in globally fragmented production, especially in com- (Russia and Brazil). However, there was no global demand hub plex GVCs. This is because regional trade agreements recently have in the complex GVC trade networks (the bottom part of Figure made greater progress than WTO negotiations in reducing the trans- 1.19) in either 2000 or 2017, as GVC imports of Germany, the US action costs, including tariffs and non-tariff barriers, involved in each and China were concentrated with their regional trading partners. border crossing. At the same time, regional trade agreements also Germany’s presence increased by 2017 to larger than that of the follow rules-of-origin which likely promote complex GVC activities. US, and China expanded rapidly. The US only maintained import- ant linkages with its two regional partners, Canada and Mexico. Demand hubs of value-added trade in various networks for All the above observations imply that the more complex the selected sectors network, the more concentrated the cross-border transactions of Greater variation in the structural change in networks can be intermediate goods in GVCs. In other words, geographic distance found at the sector level. In the textile sector, the volume of FIGURE 1.21 Demand hubs of trade in value-added in various networks for the ICT sector Traditional trade networks (ICT sector) KGZ CYP AUS NPL NOR RUS ROM LVA MDV BEL POL ESP CHE ITA NLD LUX SWE BRN RUS FIN KAZ KGZ EST MON LUX ITA IRL KAZ EST FRA PAK BRN BRA ESP CAN BEL FRA SWE CAN CHE CYP SRI FIN AUT MLT MEX GBR MEX NLD THA NOR IRL GBR DNK AUT HKG IDN PAK BRA CZE DEU CHN DNK USA USA DEU CHN SIN THA POL HUN TUR PHI JPN PRT CZE IND MAL VIE TUR MON HUN LAO MAL ROM SIN PHI IND HKG BAN PRT MLT KOR BGR LTU KOR JPN CAM TAP IDN CAM LTU BTN MDV TAP BTN NPL GRC AUS GRC FIJ SVK HRV BGR SVN FIJ SVK SVN HRV LAO SRI LVA VIE BAN 2000 2017 IRL GRC FIN Simple GVC trade networks (ICT sector) CYP EST HRV HRV NLD ITA EST FIJ MDV CYP IRL SWE ESP GBR BRA ITA AUT MLT POL PRT NOR DNK ROM GRC CHE CHE BEL NOR CAN ESP NLD FRA KAZ KGZ FRA FIN LVA POL BRN AUT LUX BEL GBR SWE CZE TUR DNK HUN MEX HUN BRA DEU CHN DEU PRT USA PHI SVK CHN USA IDN BTN TUR MON VIE NPL BGR MAL MEX IND MDV MON LUX BRN CZE VIE THA HKG THA SVN IND CAN LTU CAM JPN KOR BAN RUS HKG ROM RUS MLT TAP MAL SRI JPN LAO SRI SVK SVN BGR LVA KAZ LTU TAP KOR IDN PAK SIN BTN AUS LAO NPL AUS FIJ KGZ PAK BAN CAM PHI 2000 2017 Complex GVC trade networks (ICT sector) CYP GRC SRI NOR DNK IRL EST IRL HUN MDV EST ESP FIJ ITA CZE GBR NOR SRI FIN BEL BEL KAZ GBR ITA POL KGZ MLT PAK FRA SWE RUS AUT SWE ESP SVK LAO NLD DNK LVA SIN HRV FRA LUX AUT THA KAZ KGZ HUN BRA FIJ NLD BRN CHE IDN BGR ROM MAL MDV DEU LVA USA DEU PHI BTN CHN MEX FIN CHN USA MEX TUR CHE MON AUS BAN KOR LUX LTU KOR IND CAM CZE CAN PRT VIE VIE ROM TAP BRA NPL POL JPN TAP BGR JPN CAN PRT SVK SVN MON TUR SIN HKG SVN GRC MAL BRN RUS MLT THA BAN LTU HRV CYP CAM IND PHI AUS IDN LAO BTN NPL PAK 2000 2017 Note: the size of the circles represents the magnitude of value-added imports. The volume of value-added flow between each pair of trading partners is repre - sented by the thickness of the line linking the two. Source: Meng et al. (2018) based on the UIBE GVC indexes derived from the ADB 2018 ICIO table. 34 • Technological innovation, supply chain trade, and workers in a globalized world FIGURE 1.22 Demand hubs of trade in value-added in various networks for the services sector Traditional trade networks (services sector) ROM EST HRV ESP POL CHE PRT FIN FIN KAZ AUT ITA ESP BEL RUS NLD NOR NLD FRA GRC ITA BEL BTN SWE NOR DNK AUT MLT FRA EST LUX KGZ CYP HUN POL MDV CAN IRL SWE MEX NPL BRA ROM DNK GBR SRI CZE BRA PAK IND PRT CHE CAN DEU SVK USA DEU USA IRL IDN CHN MEX LUX CZE MON BGR GBR IND TUR SRI BTN BRN LTU MLT JPN HUN CAM CHN HKG JPN TUR FIJ VIE SVN LVA PHI BAN LVA SVN RUS CAM MDV AUS THA NPL HRV SVK PAK BAN MON LTU BGR LAO THA SIN CYP PHI GRC KAZ MAL KOR LAO TAP SIN AUS KOR BRN FIJ TAP HKG MAL IDN KGZ 2000 2017 Simple GVC trade networks (services sector) MLT LVA MDV HRV SVK MLT DNK SWE CHE IRL EST SWE SRI NOR NLD MDV FIN BEL AUT BEL ITA NOR GBR LVA CYP POL PRT FIN GRC FRA CAM ESP LUX DNK AUT HRV MEX CHE NLD CAN GBR ROM ITA FRA ESP CZE CAN RUS PRT DEU IRL MEX LUX USA DEU CAM POL USA HUN CHN MON BGR BRA SRI KAZ HUN CHN BRA JPN BRN AUS SVN LAO BAN BAN IND CZE ROM JPN LTU MON KGZ KOR PHI SVN PAK LTU PAK SIN SVK LAO KOR PHI RUS AUS VIE FIJ TUR TAP SIN NPL BTN EST IDN BRN MAL THA FIJ TAP MAL IDN HKG HKG NPL KGZ KAZ BGR 2000 BTN CYP Complex GVC trade networks (services sector) 2017 LTU LVA SRI BGR CAM EST ESP BAN BGR CYP BTN IDN MLT SRI GBR MDV DNK PHI MDV PRT BEL FRA BEL POL ITA NLD NLD IRL SIN ITA SWE LUX IRL GBR THA CYP CHE MEX LUX ESP FRA MAL CZE LVA CHE NOR FIN AUT USA CAN DEU CHN EST HUN DEU CHN SWE AUT LAO CAN USA CAM BRA THA POL MEX NOR HUN KOR RUS MON HKG RUS BRN CZE BRA FIN KAZ AUS HKG TAP MON TAP JPN IND SVN BTN ROM KGZ KAZ SVK SVK KOR ROM LTU TUR BAN SIN MAL TUR PAK NPL GRC BRN JPN FIJ VIE PAK IDN SVN HRV KGZ AUS HRV IND FIJ LAO GRC PHI NPL 2000 2017 Note: the size of the circles represents the magnitude of value-added imports. The volume of value-added flow between each pair of trading partners is repre - sented by the thickness of the line linking the two. Source: Meng et al. (2018) based on the UIBE GVC indexes derived from the ADB 2018 ICIO table. China’s trade increased sharply from 2000 to 2017, but its only the complex GVC networks, the connection in Europe, Asia and important outflow linkage was to the US (Figure 1.20). Germany’s North America became more concentrated with their regional presence as a regional demand hub fell from 2000 to 2017, while partners. The importance of France, Turkey and Viet Nam as Russia became an important regional demand hub in Europe with sub-regional demand hubs increased substantially by 2017. Com- inflow linkages from some Eastern European and Central Asian pared to the position in simple GVC trade networks, Russia’s pres- countries. In the simple GVC networks, China’s importance as a ence was very low in the complex GVC trade networks. regional demand hub increased, with an important outflow link- In the ICT sector, China became the largest demand hub for the age to the US and inflow linkages from most Asian economies. On traditional trade networks. In 2017, China had the largest magni- the other hand, Italy changed from the largest regional demand tude of imports (indicated by the size of the circle) and important hub in Europe to an isolated country, as Italy’s participation pat- inflow linkages from Germany, Japan, the Republic of Korea, Chi- tern in simple GVCs changed from an intermediate goods-ori- nese Taipei, and outflow linkages to the US (Figure 1.21). A very ented importer to an intermediate goods-oriented exporter. In similar pattern for China can also be found in the simple GVC trade Recent patterns of global production and GVC participation • 35 networks. By 2017 the US had lost many inflow linkages from Asia, a greater role in meeting this demand. Because of this and due but still maintained many inflow linkages from other economies in to China’s rapid increase in purchasing power, China has become the simple GVC trade networks. In the complex GVC trade net- one of the most important demanders of value-added through works, Europe, Asia and North America had become more sepa- final goods trade for several other countries. While China’s per rated, as there was no longer any direct or indirect linkage among capita GDP is still lower than most developed countries (US$8,827 the regional hubs Germany, China and the US. Europe changed for China versus US$59,532 for the US in 2017 according to data from multi-hubs to a single hub type network, while Asia changed from the World Bank Group), given China’s potential for positive from a single hub to a multi-hub type network. economic growth, the ongoing process of further opening-up, and The most important structural change in the services sector was its large population size, it is not difficult to imagine that China will the rise of China, which in 2017 became a regional demand hub become an important demand hub even in traditional trade net- in all three networks (Figure 1.22). The US was still the only global works as a large buyer of final goods in the near future. No doubt, demand hub in services for both traditional and simple GVC trade this will also significantly change the world map of economic inter- networks. The complex GVC trade networks are largely separated, dependence, as well as the distribution pattern of countries’ influ- since there was no direct linkage among regional hubs in both ential power in many senses. 2000 and 2017. Germany’s presence in the complex GVC trade networks had increased by 2017, reflecting the significant depen- dence of most European countries’ services sectors on German 4. The multilateral nature of bilateral trade demand for intermediate imports. balances in the age of GVCs16 From the perspective of global production networks, we can see that the rise of China has dramatically changed the whole topology Discussions of the US trade deficit in the press often focus on the of GVCs from both the demand and supply sides at both the aggre- aggregate deficit. The US has run huge trade deficits in manufac- gated and individual sector levels. This clearly reflects the fact that turing products, but has enjoyed a trade surplus in agricultural China is no longer just a “factory” exporting huge amounts of final products and services (Figure 1.23). The US trade deficit in man- goods to the world; China has emerged as a new “superpower” ufacturing products increased sharply in the late 1990s, acceler- through rapid industrial upgrading, which is reflected in the large ated after China joined the WTO in 2001, and further widened a scale of its exports and imports of intermediate goods and services few years after the global financial crisis. via both simple and complex GVC trade networks. In other words, The dramatic increase in the U.S. manufacturing trade defi- more countries, especially in Asia, have become highly dependent cit with China since China’s WTO accession is largely a result of on China’s supply of value-added and its demand for value-added the movement of production facilities from other industrialized directly and indirectly via GVCs. Another interesting finding that countries (mainly Japan and the Asian NICs) to China (Table 1.3 is not so remarkable, but can be clearly observed in our results, is reports the share of U.S. major trading partners’ contribution to that most of China’s final demand in the past was previously satis- the U.S. trade deficit in manufactured products between 1990 fied by its own domestic suppliers, whereas nowadays imports play and 2017). For example, in 1990, Japan and the four Asian Tigers FIGURE 1.23 United States worldwide trade balance in broad economic sectors % GDP 2 Agriculture, hunting, forestry and fishing Mining Manufacturing 1 Service 0 -1 -2 -3 -4 -5 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 Source: Bureau of Economic Analysis, US Department of Commerce, available online at https://www.bea.gov/ 36 • Technological innovation, supply chain trade, and workers in a globalized world TABLE 1.3 Share of U.S. trade deficit in manufacturing products with partners (percent) Four Asian Rest of CAN JPN DEU MEX ASEAN9 CHN ROW G7 NICs OECD 1990 7.5 49.4 25.7 10.2 -1.9 6.1 10.3 -6.7 -0.5 70.4 1995 6.7 45.3 10.8 10.0 5.7 13.1 24.2 -6.0 -9.8 66.9 2000 6.2 24.9 9.1 8.3 3.6 11.5 25.5 7.3 3.7 45.1 2005 4.1 15.1 3.7 8.5 4.2 9.4 37.0 12.9 5.1 34.3 2008 -2.2 16.5 1.6 8.6 5.8 11.1 57.6 9.2 -8.2 31.0 2009 -7.0 14.6 1.7 8.0 6.8 13.0 70.5 7.9 -15.5 21.7 2010 -5.9 15.1 -0.2 7.9 8.1 11.0 67.1 9.7 -12.8 22.9 2011 -7.5 15.0 -0.2 10.4 6.1 11.1 67.6 9.0 -11.5 23.4 2012 -8.6 16.9 0.1 12.2 5.6 11.7 70.6 7.3 -16.0 27.0 2013 -9.5 16.4 -1.4 13.8 5.2 12.8 72.6 8.5 -18.4 28.4 2014 -8.4 13.5 0.4 13.2 5.2 13.6 67.7 11.0 -16.2 25.5 2015 -5.1 11.2 1.7 11.1 7.3 13.6 59.5 11.3 -10.5 23.5 2016 -4.5 11.0 1.9 9.2 8.2 14.3 55.4 11.9 -7.4 21.4 2017 -4.7 10.5 1.2 8.5 8.1 14.7 56.1 12.8 -7.2 20.1 Data Source: OECD Bilateral Trade in Goods by Industry and End-use (BTDI*E), ISIC, Rev.4, available online: https://stats.oecd.org/Index.aspx?DataSetCode=BT- DIXE_I4. ASEAN 9 include MYS, PHL, THA, IDN, VNM, BRN, KHM, MMR and LAO. SGP is included in Four Asian NICs. BOX 1.3 Identifying and measuring the third country effect in bilateral trade An integrated mathematical framework to trace value added in exports that also crosses national borders at least twice; and identify double counted items in gross trade flows is pro - PDC counts in no country’s GDP, as it is the factor content vided in Koopman, Wang and Wei (KWW, 2014). A country’s that has already been counted by at least one of the three gross exports can be decomposed into the sum of four con- components above and crosses national borders at least ceptually different components: (a) domestic value added three times but is recorded in gross trade statistics by each that is ultimately absorbed abroad, or value-added exports country’s custom authority. (VAX) as named by Johnson and Noguera (2012); (b) domes- By identifying which parts of the gross trade transac- tic value added that is exported (as intermediate exports) tions are double counted relative to GDP statistics, the and then returned home (RDV); (c) foreign value added used KWW method provides a way to correctly interpret gross in the production of exports (FVA); and (d) multiple counted trade data in value added terms (relative to GDP) and links value added due to back and forth cross-border intermedi- gross trade and GDP statistics (the two most important ate trade (PDC). KWW further shows that these components and popular used economic statistics today) based on the of gross exports all have specific types of relationships with System of National Accounts standard (SNA). Wang, Wei, GDP statistics: VAX is the home country’s GDP used to sat- and Zhu (2014) extend the KWW accounting framework isfy foreign demand, in which the factor content embodied to trade at the bilateral, sector, and bilateral sector levels in gross exports crosses national borders at least once; RDV and provide a consistent accounting framework that resem- is not part of home country’s value added exports, but is bles in spirit that of KWW (2014) across different levels of part of home country’s GDP that is eventually absorbed at aggregation. By splitting these four broad components into home as the country’s final demand, through which domes- more detailed items, the roles of third countries in bilateral tic factor content crosses national borders at least twice; trade can be clearly identified and measured, as indicated FVA is a part of other countries’ GDP, or the factor content by Table 1.4. Recent patterns of global production and GVC participation • 37 BOX 1.3 (continued) Identifying and measuring the third country effect in bilateral trade TABLE 1.4 Decomposition of bilateral gross trade to identify and measure the roles of third counties in bilateral trade (percent) Core KWW Relation to GDP Number of border Detailed Decomposition Economic interpretation decomposition statistics crossings VAX_G DVA_DIR Domestic VA in Home GDP satisfies Value added exports production of exports final demand in partner At least once that is finally absorbed by country trading partner DVA_IND Domestic VA in Home GDP satisfies final production of exports demand in third countries that is finally absorbed by third countries RDV_G RDV_G Domestic VA first Home GDP satisfies own Returned DVA exported but finally domestic final demand returned home and through international consumed there trade At least twice FVA MVA Trading partner’s VA Partner’s GDP satisfies Foreign value added used in production of final demand in partner exports that return to and country is absorbed by partner OVA Third countries’ VA used Third countries’ GDP in production of exports satisfies final demand in that is finally absorbed by partner country partner PDC ODC Pure double counting in No country’s GDP Pure double counting gross exports sourced from third countries DDC Pure double counting in No country’s GDP gross exports sourced At least three times from home MDC Pure double counting in No country’s GDP gross exports sourced from partner The decomposition of bilateral trade at a detailed level This ratio is driven by final demand in the partner coun - shows that the role of third countries in bilateral trade can try and the production sharing arrangement between be measured by 3 of the 8 detailed components (in blue the home and third countries. Finally, the ratio of ODC font): DVA_IND, OVA and ODC. The ratio of DVA_IND to to gross trade is used to measure the complexity of the gross trade is used to measure the importance of a part- third-country effect. This ratio is determined by the pro- ner country as a transfer platform for the home country’s duction arrangement among home, partner and third DVA absorbed in third countries. This ratio is determined countries. ODC refers only to intermediate inputs that by the production sharing arrangement between the cross a national border at least three times (a firm uses home and partner country, as well as by final demand in intermediate inputs from a country to produce intermedi- third countries. Similarly, the ratio of OVA to gross trade ate inputs in another country for production of exports to is used to measure the importance of third countries’ a third country, involving production sharing activities of factor content for the home country’s export production. at least among 3 countries). 38 • Technological innovation, supply chain trade, and workers in a globalized world were the source of about 75% of the U.S. worldwide trade deficit global production networks over the last two decades and have in manufactured products, but by 2017 their share had declined increased their share of the US global trade deficit in manufac- to less than 12%. Over the same period, China’s share of the U.S. tured goods (Table 1.3). This suggests that the development of trade deficit in manufacturing products increased dramatically various global production chains is one of the fundamental driv- from 10% to about 73% in 2013, and has declined since then. In ing forces of the growing U.S. bilateral trade deficit with China in other words, while China was becoming an increasingly import- manufactured products during the past two decades. ant source of manufactured goods, the relative importance of To examine the role GVCs have played in the geographical the rest of the industrialized world as a whole was declining (see shifting of the US trade deficit in manufacturing products, this the last column of Table 1.3), because many firms in these econ- section analyzes the value-added structure of the three trade omies were shifting their manufacturing and assembly facilities routes where the US has the largest deficit, namely US trade to China via their FDI to China. Trade statistics by ownership with China, Japan and Germany, using the gross trade account- from China Customs confirm that China’s trade surplus in man- ing method proposed by Koopman et. al (2014, see Box 1.3 for ufacturing products with the US was mainly generated by wholly details). foreign-owned enterprises (FIE) and joint venture companies We first look at the value-added structure for US net imports (JOV), although Chinese-owned private firms (PRI) have played of computer, electronic and optical equipment (OECD-ICIO C30, an increasing role in recent years17. 32 and 33) from China as an example. The decomposition results Along with China, other emerging economies, such as Mexico are reported in Table 1.5. Column (1) reports gross exports in and the ASEAN countries, have been increasingly integrated into millions of dollars (current prices). Column (2) reports value TABLE 1.5 US-China trade of computer, electronic and optical equipment (million USD) Year TEXP VAX_G DVA_DIR DVA_IND RDV_G DDC MC OVA ODC (1)=2+3+4 (2)=2a+2b (2a) (2b) (3) (4) (5) (6) (7) +5+6+7 China exports to the United States Value 17,553 4,356 3,652 704 21 64 1,785 9,385 1,942 2000 Share 100 24.8 20.8 4 0.1 0.4 10.2 53.5 11.1 Value 94,153 33,869 29,826 4,043 195 1,003 6,229 47,502 5,356 2007 Share 100 36 31.7 4.3 0.2 1.1 6.6 50.5 5.7 Value 166,296 76,573 67,422 9,151 675 2,537 9,301 69,035 8,176 2014 Share 100 46 40.5 5.5 0.4 1.5 5.6 41.5 4.9 US exports to China Value 5,362 3,441 2,504 936 572 139 46 725 440 2000 Share 100 64.2 46.7 17.5 10.7 2.6 0.9 13.5 8.2 Value 13,930 9,182 4,891 4,291 2,016 237 427 886 1,182 2007 Share 100 65.9 35.1 30.8 14.5 1.7 3.1 6.4 8.5 Value 25,054 18,544 11,099 7,445 3,346 317 754 1,033 1,061 2014 Share 100 74 44.3 29.7 13.4 1.3 3.0 4.1 4.2 US net imports from China Value 12,191 915 1,148 -232 -551 -75 1,739 8,661 1,502 2000 Share 100 7.5 9.4 -1.9 -4.5 -0.6 14.3 71.0 12.3 Value 80,223 24,687 24,935 -248 -1,821 765 5,802 46,616 4174 2007 Share 100 30.8 31.1 -0.3 -2.3 1.0 7.2 58.1 5.2 Value 141,242 58,029 56,323 1,706 -2,671 2,220 8,547 68,002 7,114 2014 Share 100 41.1 39.9 1.2 -1.9 1.6 6.1 48.1 5.0 Source: The UIBE GVC indexes derived from the 2017 OECD ICIO table. Recent patterns of global production and GVC participation • 39 added exports (VAX_G) associated with these gross trade flows. the global production network. By contrast, Chinese firms began In the next five columns, major components of gross exports to join the global value chains since deregulation of foreign are reported: domestic value added that is ultimately absorbed investment in 1992, undertaking processing and assembly tasks, by partner country ((2a) DVA_DIR); domestic value added that so that the ratio of domestic value added to gross exports was is ultimately absorbed by third countries ((2b) DVA_IND), which very low; a great deal of value came from foreign upstream sup- depends upon final demand in the third country; domestic value pliers of raw materials, parts and components. In 2000, 98.7% of added in exports that is ultimately returned and consumed at China’s exports of computer, electronic and optical equipment to home (column (3) RDV_G), which is part of home country’s GDP the US were processing exports. After China entered the WTO, and final demand; loop effects between bilateral trading part- Chinese firms started to move up the global value chains. More ners (Column (4) and (5) DDC and MC), third countries’ value Chinese firms upgraded to general trade, and the proportion of added in gross exports (column (6), OVA) and pure double processing trade fell (from 87.3% in 2007 to 77.4% in 2014). counting sourced from third countries (column (7), ODC). Such a value-added structure of US net imports from China The decomposition results not only reveal the mislead- is not uncommon. The important role played by third countries ing nature of the balance of trade computed from gross trade also can be observed in US net imports from Germany, Japan statistics, but also the sources of such statistical illusion. Value and many other trading partners. Figure 1.24 shows the val- added in exports (VAX_G) accounted for only 25% of China’s ue-added structure of US total net imports from Germany. A exports of computer, electronic and optical equipment to the much larger portion of US intermediate goods exports to Ger- US before China’s WTO accession. This share increased after- many were re-exported to third countries compared to the share wards, but remained lower than 50% in 2014. The value added of US imports from Germany (DVA_IND, which depends on final in exports from third countries consistently accounted for more demand in third countries) that was re-exported. Thus, in this than 50% of China’s exports of these goods throughout the re-exported portion, the US actually ran a large surplus with sample period. The composition of US exports to China was Germany in terms of value added, especially in services sectors. the opposite, as the share of VAX_G dominated throughout the Compared to US net imports from China, US net imports from sample period (between 65-75%). The value added content from Germany contain a much higher share of Germany’s factor con- third countries (OVA+ODC) accounted for less than 20% of US tent (around 80%), but third countries’ suppliers also accounted gross exports of these goods, and declined to only about 8% in for around 40% (third countries’ final demand accounted for a 2014. MC+OVA+ODC accounts for the largest portion of China’s negative 20%, implying that Germany’s imports from the US exports, as China used upstream inputs from the US and third depended more on third countries’ final demand for Germany’s countries to produce its exports; DVA_IND+RDV+DDC is the products that use US intermediate inputs). All of this demon- largest portion of US exports, which are US products imported strates the complex composition and offsetting factors involved by China used as inputs to produce China’s exports for US and in gross net trade flows. third country markets. Therefore, the main source of the trade To further demonstrate the differing roles of third countries imbalance in China-US bilateral trade in computer, electronic across bilateral trade routes, Figure 1.25 compares the changing and optical equipment was the third countries’ value added in value-added structure of: US net imports of computer, electronic gross trade flows. Third countries accounted for 80.3% of the and optical equipment from China and US net imports of trans- total trade imbalance in 2000, falling to 53.1% in 2014. port and storage services from Germany; and US net imports of Bilateral trade balances (net imports) are often used by trade motor vehicles from Germany and Japan. and labor economists as a measure of import penetration and US net imports of ICT products from China increased rapidly the impact of external trade on domestic economic activity. after China joined the WTO, jumping from less than 10% of US When traditional (final goods) trade dominated international sector value-added (11 billion USD) in 2001 (right scale of Figure trade flows, the net imports captured the imported factor con- 1.25, top left) to over 60% (141 billion USD) in 2014. Factor content tent from the surplus economy to the deficit economy. However, from third countries played the most important role in this dramatic when global trade is dominated by global value chains, gross growth (well above 50%). This reflected other countries using China trade balance is no longer a reliable measure of import penetra- as an assembling hub to re-export their domestic value added to tion. As shown in the bottom panel of Table 1.5, US net imports satisfy US final demand. Similarly, demand for German goods by of computer, electronic and optical equipment only contain a third countries, mostly nearby European economies, were the driv- very small portion of Chinese factor content. In 2000, Chinese ing force behind the rise in US net exports of transport and storage value added (factor content) only constituted 7.5% of US total net services to Germany (Figure 1.25, bottom left). imports from China. This share increased rapidly after China join Third countries’ production significantly affected US deficits the WTO, reaching 30.8% in 2007 and 41.1% in 2014. in motor vehicles with Germany and Japan from 1995 to 2014. A Differences in the value-added structure of exports between substantial portion of US net imports from Germany (more than China and the US reflects the different role that the two coun- one fourth of US net imports in 2014) reflected factor content tries’ firms played in this sector. With high design and system from third countries, mostly Eastern EU countries and China, integration capacities, US multinationals were the lead firms of while final demand in third countries accounted for only about global value chains and occupied a top and central position in 5% of US net imports over this period (Figure 1.25, bottom right). 40 • Technological innovation, supply chain trade, and workers in a globalized world FIGURE 1.24 Value-added structure of US net imports from Germany Share % Billions USD 100 50 80 40 60 30 40 20 20 10 0 0 95 96 97 98 99 00 01 02 03 04 05 06 07 08 09 10 11 12 13 14 19 19 19 19 19 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 -20 -10 Net imports (right) DVA_DIR DVA_IND -40 -20 RDV_G DDC+MC OVA ODC Note: Refer to Box 1.3 for symbol definitions. Source: The UIBE GVC indexes derived from the 2017 OECD ICIO tables. The importance of third countries’ factor content supply and Several findings emerge from this chapter: final demand in US net imports of motor vehicles from Japan First, the globalization of production slowed after 2011, indi- increased towards the end of this period, but remained at a cated by the increase of purely domestic production and the lower level than in Germany. decline of GVC activities as a share of total production activities. This analysis illustrates that in the age of global value chains, As the growth of global trade surpassed the growth of global when embodied factor content and sources of final demand of GDP for the first time in nearly six years, there were some signs gross trade flows vary significantly across trade routes by coun- of recovery of GVC activities in 2017, especially for complex tries and products, net bilateral imports are no longer a reliable GVCs activities. However, 10 years after the global financial crisis, measure of the impact of trade with a partner country on domes- global GVC participation has not returned to the pre-crisis level. tic prices and wages. This also implies that any change in bilateral Considering a longer period, the higher technology (knowledge) trade policy can have a significant impact on third countries that intensity of a sector, the more significant the increase of complex should not be overlooked in dealing with bilateral trade issues. GVC activities. Second, while the share of intra-regional GVC activities in total GVC activities increased in Asia from 2000 to 2017, the 5. Conclusions share of intra-regional GVC activities declined in both Europe and North America and their share of inter-regional production The rise of GVCs has significantly changed the nature and struc- sharing activities increased, especially their GVC linkages with ture of the world economy. The increasing complexity of GVCs “Factory Asia”. GVC trade become more global in 2017 compare also brings great challenges to policy making in both developed to 2000. and developing countries. This chapter has presented trends Third, from the view of global production network topology, in GVC production and trade up to 2017 from various perspec- China played an increasingly important role as both a supply and tives, based on a recently developed production decomposition demand hub in traditional trade and simple GVC activities, while method that classifies factor contents embodied in a product the US and Germany remained the most important hubs in com- into GVC and non-GVC activities depending on whether they plex GVC networks. China has emerged as a new hub through cross national borders. rapid industrial upgrading, represented by its more high-tech Recent patterns of global production and GVC participation • 41 FIGURE 1.25 The roles of third countries can be very different in different bilateral trading routes Share % Billions USD 90 (a) US net imports from CHN, ICT 160 14 (a) US net imports from JPN, motor vehicles 50 80 140 45 12 70 40 120 10 35 60 100 8 30 50 80 25 40 6 60 20 30 4 15 40 20 10 10 20 2 5 0 0 0 0 -10 -20 19 5 19 6 19 7 19 8 20 9 20 0 20 1 20 2 20 3 20 4 20 5 20 6 20 7 20 8 20 9 20 0 20 1 20 2 20 3 14 9 9 9 9 9 0 0 0 0 0 0 0 0 0 0 1 1 1 1 19 5 19 6 19 7 19 8 20 9 20 0 20 1 20 2 20 3 20 4 20 5 20 6 20 7 20 8 20 9 20 0 20 1 20 2 20 3 14 19 9 9 9 9 9 0 0 0 0 0 0 0 0 0 0 1 1 1 1 19 (a) US net imports from DEU, transport and storage (a) US net imports from DEU, motor vehicles 35 0 30 30 30 -1 25 25 25 -2 20 20 20 -3 15 -4 15 15 10 -5 10 10 5 -6 5 5 0 -7 -5 -8 0 0 19 5 96 19 7 19 8 20 9 20 0 20 1 20 2 20 3 20 4 20 5 20 6 20 7 20 8 20 9 20 0 20 1 20 2 20 3 14 -10 -9 19 5 19 6 19 7 19 8 20 9 20 0 01 20 2 03 20 4 20 5 20 6 20 7 20 8 09 20 0 11 20 2 20 3 14 9 9 9 9 0 0 0 0 0 0 0 0 0 0 1 1 1 1 9 9 9 9 9 0 0 0 0 0 0 0 1 1 1 19 19 19 20 20 20 20 Net Imports(Right) DVA_IND OVA ODC Note: The blue bars represent net gross imports, measured in percent of US sector value-added according to the scale in vertical axes on the right, a positive number indicates US trade deficit, a negative number indicates US surplus; The lines represent third countries’ roles, measured in percentage point according to scale in vertical axes on the left. The y-axes indicates calender years. Refer to Box 1.3 for symbol definitions. Source: The UIBE GVC indexes derived from the 2017 OECD ICIO tables. intermediate exports and imports. Bilateral trade, especially More policy analyses on the impact of technology changes and complex GVC trade, became more concentrated among major GVC trade on labor markets in developed and developing coun- regional trading partners, indicating distance matters even for tries will be discussed in detail in other chapters of this report. value-added trade and GVCs. Current residence-based national account rules treat all firms Fourth, in the age of GVCs, bilateral trade balances are no within national borders as domestic firms, so the value-added longer a reliable measure of the impact of partner countries on creation of foreign affiliates is treated as part of purely domestic domestic economic activities. For example, production and final production activities if they do not engage in cross border trade. demand from third countries have had a significant impact on However, some of their production may also be a type of GVC US net imports from China, Germany and Japan. And factor con- activity, especially in services because the supply of services tent from third countries accounted for more than half of the bur- through commercial presence abroad is an important way of geoning deficit in US net imports of ICT products from China, conducting international transactions in services (mode 3 – com- which increased 12.8 times in the 15 years up to 2014 to reach mercial presence). The distinction between foreign and domestic 141 billion USD. owned firms is particularly relevant. However, no ICIO table cur- One important policy implication is that changes in trade rently available is able to separate production activities between policy can have broad and unanticipated effects. Unilateral domestic firms and foreign affiliates to allow us to develop GVC imposition of trade protection on exports from a partner coun- measurement for such activities. Initiatives in this direction are try can have a significant impact on third countries when trade is being taken in the international statistical community. Chapter carried out through GVCs, particularly complex GVCs. Indeed, as 8 of this report will discuss this and related GVC measurement many products today are already “made in the world”, increasing issues in more details. import protection can even harm exports from the home country. 42 • Technological innovation, supply chain trade, and workers in a globalized world Notes 6. Its GVC exports share to Europe and Asia was 40.4% and 20.4% respectively, higher than its share of intra-regional complex GVC 1. “Pure domestic” means domestic value-added in domestically pro- activities at 18.1%; Its complex GVC imports share from Europe and duced final products that satisfy domestic final demand without Asia was 31.2% and 27.8% respectively, also higher than its share of involving cross border trade and production sharing activities, it can intra-regional complex GVC activities at only 20.7%, also be phrased as “not traded internationally”; “Traditional trade” 7. This section was written by Bo Meng and Ming Ye. is final goods and services produced for exports with only domestic 8. Some care is needed in interpreting smile curves produced using factor content, it can also be phrased as “Trade in final products” or input-output tables in basic prices, see also Chapter 8. “Ricardian Trade”; “GVCs” are basically “trade in intermediate prod- 9. The data for compensation per employee is from the WIOD Socio ucts”. The distinction between simple and complex GVC activities Economic Accounts 2016 version (compensation of employees / in our estimates are determined by the number of national border number of employees). crossing, not the differences in technology or the complexity of actual 10. The distance is measured by a value-added weighted average of pro- production process (although there is a correlation between them), duction stages. For detailed methodology, one can refer to Ye, Meng so they can be phrased as “value-added activities cross one or more et.al. (2015). than one national borders”. Some care is needed in interpretation, 11. This section was written by Bo Meng, Hao Xiao and Jiabai Ye. for example a large economy is likely to see lower levels of estimated 12. Data are from the ADB ICIO database (the 2018 version). complex GVCs than would be the case if the same economy was split 13. It should be noted, these types of plots are better for capturing long- into a series of smaller economies. run changes on the extensive margin rather than short-run changes 2. This section was written by Xin Li and Zhi Wang. that occur on the intensive margin. 3. We aggregate the 65 WIOD industries into 8 industry groups: (1) 14. It should be noted that country size may result in some bias in our AGR: Agriculture, Hunting, Forestry and Fishing (ISIC rev.3 “01T05”); analysis. For example, countries exporting to the US are more likely to (2)Min: Mining and Quarrying (ISIC rev.3 “10T14”); (3) HTI: High R&D see their exports classified as ‘simple’ than ‘complex’ GVC activities, intensive industries (ISIC rev.3 “24, 29T34, 352,353, 359”);(4) MTI: compared to exports within a ‘fragmented’ region of smaller countries Medium R&D intensive industries (ISIC rev.3 “25T28, 351, 37”); (5) LTI: (e.g. EU). low R&D intensive industries (ISIC rev.3 “15T23, 36”), (6) TTS: Trade 15. A large number of studies have argued that due to rising manufac- and Transportation (ISIC rev.3 “50T52”, 55, “60T63”);(B)FBS Post and turing costs in developed nations, many companies are looking to Telecommunications, Financial and Business services (ISIC rev.3 “64, less-developed nations to set up manufacturing facilities in hopes of 65T67, 71T74”); (8) OSE: Real Estate Activities, Utility, Construction, reducing costs. These developed countries are being “hollowed out”, and: other services (ISIC rev.3 “70, 75, 80, 85, 90T93, 95, 40,41, 45”). which poses a threat to many factory workers because they could 4. The relative value of the forward and backward participation indi- lose their job to someone in another country. The level of industrial ces indicates a country-sector’s position in the global production hollowing out can be measured by net FDI outflows, unemployment network. A higher degree of forward participation than backward rates, the share of manufacturing industries in GDP, and other means. participation implies that the country is more actively engaged 16. This section was written by Fei Wang, Zhi Wang and Kunfu Zhu. in upstream production activities in GVCs. Some care is needed in 17. Based on trade statistics collected by the General Administration interpretation however, see Ahmad, N., et al. (2017), “Indicators on of Customs of the People’s Republic of China (GACC), China had a global value chains: A guide for empirical work”, OECD Statistics 304.8 billion USD trade surplus in manufacturing products with the Working Papers, No. 2017/08, OECD Publishing, Paris, https://doi. United States in 2017. The share of FIE and JOV was 55%, the share of org/10.1787/8502992f-en. PRI was 41%, while SOE and other firms represented only about 4%. 5. As a result, industry became an inappropriate analytical unit for the study of international trade. See the discussion on firm heterogeneity for the empirical challenges to tackle this problem in Chapter 8. Recent patterns of global production and GVC participation • 43 References Mudambi, R. (2008). Location, Control and Innovation in Knowledge-Inten- sive Industries. Journal of Economic Geography, 8(5), 699-725. Baldwin, Richard E. and Javier López- González (2013). Supply-Chain trade: Rodrik, D. 2018. New Technologies, Global Value Chains, and the Develop- A portrait of global patterns and several testable hypotheses, NBER ing Economics. Pathways for Prosperity Commission Background Paper Working paper 18957, National Bureau of Economic Research. Series: No.1 Oxford, United Kingdom. Ferrantino M J, Taglioni D. (2014). Global value chains in the current trade Shin, N., Kraemer, K.L., Dedrick, J. (2012). Value Capture in the Global Elec- slowdown, World Bank Economic Premise, 137. tronics Industry: Empirical Evidence for the “Smiling Curve” Concept. Ferrarini B. (2013). Vertical trade maps, Asian Economic Journal, 27(2): Industry and Innovation, 19(2), 89-107. 105–123. Shih, S. (1996). Me-Too is Not My Style: Challenge Difficulties, Break through Johnson, R. and Noguera, G. (2012). Accounting for Intermediates: Pro- Bottlenecks, Create Values (Chinese Taipei: The Acer Foundation). duction Sharing and Trade in Value-added, Journal of International Wang, Zhi, Shang-Jin Wei, and Kunfu Zhu, 2014, “Quantifying International Economics, 86(2):224-236. Production Sharing at the Bilateral and Sector Levels,” NBER Working Koopman, R., Wang, Z., Wei, S-J. (2014). Tracing Value-Added and Double Paper 19677. Counting in Gross Exports. The American Economic Review, 104(2): Wang, Zhi, Shang-Jin Wei, Xinding Yu, and Kunfu Zhu, 2017a, “Measures of 459-494. Participation in GlobalValue Chains and Global Business Cycles”, NBER Los B., Timmer M.P., de Vries G.J., (2015). How Global are Global Value Working Paper No. 23222. Chains? A New Approach to Measure International Fragmentation, Xiao, H., Sun, T., Meng, B., Cheng, L. (2017). Complex network analysis Journal of Regional Science, 55(1): 66-92. for characterizing global value chains in equipment manufacturing, Meng, B., M. Ye, Wei, S-J. 2017. Value-added Gains and Job Opportunities PlosOne, 12 (1): e0169549. in Global Value Chains, IDE Discussion Paper, 668, IDE-JETRO, Chiba Ye, M., B. Meng, and S. Wei. 2015. “Measuring Smile Curves in Global City, Japan. Value Chains.” IDE Discussion Paper 530, IDE-JETRO, Chiba City, Meng, B., Xiao, H., Ye, J., Li, S. (2018). Are global value chains truly global? Japan. A new perspective based on the measure of trade in value-added, IDE Zhou, M., Wu, G., Xu, H. (2016). Structure and formation of top networks in Discussion Paper, 736, IDE-JETRO, Chiba City, Japan. international trade, 2001–2010, Social Networks, 44: 9–21. CHAPTER 2 Trade, value chains and labor markets in advanced economies* By Marc Bacchetta (WTO) and Victor Stolzenburg (WTO) ABSTRACT Trade is a major source of employment. Nevertheless, trade vary considerably across regions and individuals with dif- has recently been caught in the crossfire in discussions ferent skill levels. This implies that policy has a central role around the decline of manufacturing employment and the to play in making sure that the gains from trade are shared polarization of labor markets in advanced economies. In this evenly. Our findings highlight that a value chain perspec- chapter we examine what the academic literature has to say tive is important for assessing the impact of trade on labor on the relationship between trade and labor markets, with markets. The emergence of value chains has strengthened a specific focus on studies with a value chain perspective. linkages between sectors, magnified trade’s impact on skill We find that trade has only modest effects on aggregate demand and requires novel trade statistics. Ignoring this employment and is unlikely to have been a major contrib- leads to a biased view of trade and overestimates its role in utor to the decline of manufacturing. However, the effects the decline of manufacturing employment. • Factoring in GVCs when studying the impact of trade on labor markets reveals that trade has not been a significant contributor to declines in manufacturing jobs in advanced economies, and that job gains in services have offset job losses in manufacturing. • However, the effects of trade can vary considerably across regions and individuals with different skill levels, compounding regional disparities and labor market polarization driven by other factors such as automation. • Adjustment policies should not differentiate between the various reasons for worker displacement, such as automation or trade, and should be less dependent on affected workers fulfilling certain conditions. * We would like to thank Klaus Michael Keller for excellent research assistance and inputs to this chapter. We further would like to thank Michael Ferrantino, Bill Shaw and participants at the second conference for the GVC Development Report 2019 for helpful comments and suggestions. 45 46 • Technological innovation, supply chain trade, and workers in a globalized world 1. Introduction T A prominent role in this regard has been given to China’s WTO accession in 2001 (Autor et al., 2013; Pierce and Schott, 2016), rade is a major source of employment in advanced the conclusion of NAFTA in 1994 (Hakobyan and McLaren, 2016), economies. Estimates suggest that exports supported and the EU enlargement in 2004 (Braakmann and Vogel, 2011). 12 million jobs in 2014 in the United States alone.1 The Since the results of these studies are contrary to the common importance of exports for employment becomes even view among economists that trade has only minor employment more apparent when expressed in terms of shares. Figure 2.1 impacts,3 several literature surveys have recently re-examined the shows that exports may account for almost 50% of jobs in Ireland role of international trade for labor market outcomes to reach a and around 30% of jobs in Germany. Similarly, imports can con- conclusive and comprehensive assessment of trade’s impact on tribute to employment. By reducing the costs of production, they employment and wages.4 This chapter summarizes the findings can lead to higher demand which can translate into more jobs.2 of these surveys but, crucially, also sheds light on a topic that has In addition, jobs supported by imports or exports pay a signif- not received sufficient attention in recent articles: the impact of icant premium when compared to jobs supported by domestic the expansion of value chains on the relationship between trade demand. Martins and Opromolla (2011) find that average wages and labor markets. are up to 30% higher in exporting and importing plants com- As discussed in the first chapter of this report and its 2017 pared to non-trading plants. predecessor, the rise of domestic and international production Despite this positive role, trade has recently been caught in the fragmentation has proceeded rapidly in recent decades. Firms crossfire in the discussion around the decline of manufacturing have unbundled their factories and outsourced production employment in advanced economies. In fact, foreign competition stages across the globe. This has major implications for the inter- has been blamed for job and income losses for many decades. actions between trade and labor markets, and it is important to In particular the rise of new economic powerhouses has tradi- examine them separately from the wider trade and labor market tionally caused popular backlashes against liberal trade regimes discussion to fully understand their effects. Three main conse- in incumbent countries leading to policies that increase barriers quences of value chains can be highlighted. to imports. Examples reach from the British Merchandise Marks Firstly, the impact of import competition on labor markets is Act of 1887 targeted at German imports to Japan’s 1981 volun- not limited to import-competing industries anymore. Instead, tary export restraints. Today, import competition from emerging downstream customers and upstream suppliers are affected markets and formerly planned Eastern European economies has much more than in a non-fragmented economy. Therefore, been mentioned among the main factors behind the labor market trade shocks propagate more widely through the economy than adjustments that have taken place over the last decades. A series in the past. For instance, when an industry contracts due to for- of recent empirical studies find that trade liberalization epi- eign competition this will hurt its upstream supplier industries sodes have had a detrimental impact on labor market outcomes. and the suppliers’ suppliers since they will face lower demand. FIGURE 2.1 Domestic employment supported by exports  50% 40% 30% Indirect 20% Direct 10% 0% AUS CAN CHE DEU DNK ESP EU FIN FRA GBR IRL ITA JPN NLD NOR SWE USA Notes: Authors’ calculations based on 2014 data from the World Input Output Database (WIOD). For the methodology please refer to Calì et al. (2016). Direct refers to employment supported in exporting establishments using the share of exports in total output of these establishments. Indirect refers to employment supported in establishments that supply exporters and depend on foreign demand via these supply linkages. In many countries more than half of the jobs sup - ported by exports are due to indirect value chain linkages. Trade, value chains and labor markets in advanced economies • 47 Downstream customers, on the other hand, could benefit from The chapter shows that trade is likely to raise aggregate cheaper inputs that foreign competition implies. As a result, it is employment and real wages and that taking a value chain per- necessary to take into account input-output linkages between spective is important. In particular, cost savings due to cheap industries as well as the position of industries in the value imports and export opportunities create employment in many chain when analyzing the effect of trade on labor markets. This sectors of the economy that do not trade directly but benefit becomes strikingly clear when looking at Figure 2.1. It shows that from trade through input-output linkages. Even when the focus is in several advanced economies more than half of the jobs sup- on the manufacturing sector, the evidence paints a more benign ported by exports are not within exporting establishments but picture of trade which contrasts with popular perception. Once within supplier establishments. the rise of value chains is properly accounted for, it suggests Secondly, the expansion of value chains entails not only sec- that trade has contributed at best a relatively small share to the tors that compete across countries but also tasks and stages. This decline of manufacturing employment in advanced economies. means that competition happens at a much finer level with severe Trade has, however, contributed to regional and individual consequences for skill demand within countries. The production disparities. Since industries tend to cluster regionally, studies of many goods takes place in various stages that require different show that the impact of trade is very heterogeneous across geo- levels of skills. Value chains allow unbundling these stages spa- graphic areas. While trade benefits labor markets in regions with tially so that countries well-endowed in skilled (unskilled) labor exporting industries and industries that rely on imported inputs, can specialize in skilled (unskilled)-intensive stages. This shifts it might hurt regions that compete directly with foreign pro- aggregate skill demand much more than traditional trade, which ducers. Therefore, it leads to a spatial divergence in economic required all stages with their different intensities to be performed activity. Similarly, trade is shown to increase the demand for skills domestically, and might lead to an increased polarization of the and, thus, has uneven effects across individuals, a trend that labor market. has been aggravated by the rise of value chains. This is where Thirdly, value chains imply that traditional gross trade sta- policy intervention has the potential to play an important role. tistics are insufficient to properly assess the impact of trade on The right interventions can spread the gains from trade more labor markets because they mis-measure the scale and scope of evenly and guarantee that regions and individuals are not hurt import competition. For example, when firms offshore assembly by globalization. stages but keep upstream stages domestic, gross import statis- The chapter proceeds as follows. Section 2 discusses the role tics heavily overstate import competition because they suggest of trade in the decline of manufacturing employment. Section 3 that the complete value chain was offshored. Even if all stages are assesses how trade affects aggregate nationwide labor market offshored, gross import statistics tend to falsely assign competi- outcomes. Section 4 examines the impact of trade on regional tion to downstream industries because they suggest that the full inequality. Section 5 analyses the impact on skill demand and the value of an imported good has been created by the downstream polarization of the labor market. Section 6 proposes potential exporting industry when in fact much of the value is supplied by policy responses. Section 7 concludes. foreign upstream industries. This causes competition to be over- stated downstream but understated upstream. Due to these three changes, a comprehensive overview of the 2. Trade, GVCs, and the decline of impact of trade on labor markets requires a value chain perspec- manufacturing employment tive. Moreover, a value chain perspective is not only relevant to correctly assess whether trade boosts employment and wages One of the most contested issues in the trade and labor market or not, but it also changes subsequent policy recommendations. debate is whether and by how much imports have contributed For example, when trade shocks spread more widely within econ- to the decline of manufacturing employment in advanced econ- omies and when competition moves to ever finer degrees, it omies vis-à-vis alternative factors such as technology-driven becomes increasingly difficult to target individuals hurt by trade. productivity improvements or changes in preferences towards This chapter discusses the three changes and their implica- services. Across all high-income economies the share of man- tions for the relationship between trade and labor markets and ufacturing employment in total employment has been steadily for adjustment policies along four major debates surrounding declining for decades which has attracted considerable atten- labor markets in advanced economies. The four debates concern tion, potentially due to the fact the manufacturing jobs pay a the role of trade in: premium even after controlling for a variety of worker character- • the decline of manufacturing employment, istics (Langdon and Lehrman, 2012). In the public debate, trade • nation-wide employment trends, has been and continues to be listed as a prime culprit behind job • the rise in regional inequality, and losses in the manufacturing sector. • the increase in labor market polarization. Economic studies from the 1990s and early 2000s show in this In each case, the discussion starts with a summary of the regard that after a trade shock, employment in import-compet- results of studies that do not take value chains into account and ing industries suffers relative to employment in export-oriented then highlights the additional insights that value chain studies can sectors (e.g. Revenga, 1992). More recent studies focusing on the add. effects of rising Chinese import competition on US labor markets 48 • Technological innovation, supply chain trade, and workers in a globalized world find similar results and have sparked a heated debate around remarkable is that half of the decline of manufacturing employ- the role of trade in explaining the loss of manufacturing jobs. US ment in Denmark between 1994 and 2007 is due to firms switch- manufacturing employment was stable around 18 million workers ing their sectoral affiliation from manufacturing to services which between 1965 and 2000 before falling by 18 percent between implies that no actual job loss has taken place in these instances 2001 and 2007. Estimates of the share of this loss of jobs due (Bernard et al., 2017). to trade based on “back-of-the-envelope” calculations in popu- Importantly, trade implies not just import competition but lar media outlets, blogs and policy briefs range between 1 and also export opportunities and cheaper inputs. According to 20 per cent (De Long, 2017; Krugman, 2016b; Hicks and Devaraj, recent work, the global export expansion of US manufactur- 2015) with one author going as far as to claim that the growing ing products, which was supported by cheap Chinese inputs manufacturing trade deficit of the United States can explain into US production, almost completely offset job losses due to almost all of the manufacturing jobs lost in the period between import competition from China (Feenstra et al., 2017). In Ger- 2000 and 2007 (Scott, 2015). many, new export opportunities in Central Eastern Europe More rigorous economic analyses support the claim that have even slowed down the decline of manufacturing employ- trade has played a limited role, explaining at the very most one ment despite rising import competition (Dauth et al., 2018). In quarter of the recent decline. Seminal work in this area by Autor addition, Chinese imports and offshoring have reduced prices et al. (2013), who examine the increase in Chinese import compe - in the US considerably (Amiti et al., 2017; Handley and Limao, tition by comparing more and less exposed local labor markets 2018). This is likely to lead to significant cost savings for firms in the United States, finds that it can explain around 25 per cent and higher consumer spending. Evidence shows that these cost of the manufacturing decline. Studies based on this work show reductions enabled import-competing US manufacturing firms that similar but less pronounced trends can be observed in sev- to shift resources to industries in which they enjoy a comparative eral European countries (Donoso et al., 2015; Balsvik et al., 2015; advantage relative to China. This has led, in turn, to an increase Malgouyres, 2017). There is also evidence of large productivity in overall manufacturing employment and wages among these gains in advanced economies from trading with China, however firms (Magyari, 2017). Cost savings have also enabled offshor- with substantial job losses in exposed industries (Ahn and Duval, ing firms to expand onshore employment, leading to overall 2017). Another study shows that detrimental effects on employ- employment gains in offshoring industries (Kovak et al., 2017). ment in manufacturing in the United States arose because of the According to recent evidence, higher consumer spending in elimination of tariff uncertainty rather than tariff reduction after addition with other indirect effects, such as the between-region China’s WTO entry (Pierce and Schott, 2016). elasticity of labor supply, can even fully cancel out manufac- One explanation could be that advanced economies react to turing employment losses due to Chinese import competition increased import-competition in manufactured goods by spe- (Adao et al., 2019). cializing in the tradable services sector, such as business ser- Another important issue brings us finally to the importance vices, R&D, design or financial services, in which they typically of a value chain perspective in the analysis. Value chains have have a comparative advantage (Spence and Hlatshwayo, 2012). spread considerably over the last decades both domestically and The trade-induced shift from a manufacturing- to services-based internationally. The average share of in-house production in total economy is however not found to happen smoothly at the micro- output decreased by 8% from 1995 to 2011, indicating an on-go- level. Indeed, considering the movements of workers between ing fragmentation process.5 This spatial unbundling of produc- sectors in Germany, there is little evidence that the increasing tion has profound implications for the impact of trade on labor employment in service industries comes from incumbent man- market outcomes. Ignoring these implications can lead to severe ufacturing workers who directly switch jobs without undergoing mismeasurement of the impact of trade shocks on employment an unemployment spell. Instead, the rise of services is found to and wages. be driven by young labor market entrants who exhibit different The foremost reason why it is important to take value chains sectoral entry behaviours than previous generations, and by into account when assessing the impact of trade on labor market returnees out of non-employment who take up jobs in different outcomes is that the spread of value chains within and across industries than their previous one (Dauth et al., 2018). countries has strengthened inter- and intra-industry linkages However, many factors other than trade have been boosting considerably over the last decades. These linkages imply that the non-tradeable sector in advanced economies over time. For trade shocks propagate through the economy much more than instance, changing demand patterns caused by demographic in a vertically integrated world. Evidence from Belgium shows change and increased incomes favour services over manufac- for instance that while only 7.3% of Belgian firms export, 42.4% turing. Moreover, Bernard and Fort (2017) suggest that part of supply exporters directly or indirectly and are thus dependent the manufacturing decline is due to a statistical misconception, on foreign demand (Dhyne and Rubinova, 2016). Even more strik- because some firms are counted as wholesalers despite their ingly, 97% of all Belgian firms are dependent on foreign inputs involvement in the production of goods, with Apple Inc. serv- even though only 15% import directly (Tintelnot et al., 2017). As ing as a prime example. By re-classifying these firms as manu- a result, looking only at exporting and import-competing firms facturers, up to two million jobs in the US can be shown to have or industries when examining the trade and labor market rela- switched from services to manufacturing in 2007. Even more tionship means missing a major part of the picture. Trade, value chains and labor markets in advanced economies • 49 In this context, researchers have revisited the seminal work and upstream industries face relative employment and wage by Autor et al. (2013) that examines the effect of Chinese import losses. These combined losses are however balanced out by the competition on labor market outcomes in US commuting zones. downstream gains, leading to a small negative impact of imports In contrast to the original study, the new work considers as on manufacturing. exposed not only industries that produce the products that the Other studies have performed similar exercises with different US imports from China but also these industries’ upstream sup- methodologies that take input-output linkages into account but plier and downstream customer industries (Wang et al., 2018). have failed to observe positive effects on downstream indus- It thus takes a value chain perspective. The hypothesis is that tries of the same magnitude (Acemoglu et al., 2016; Caliendo supplier industries are hurt by import competition when their et al., 2018). They nevertheless suggest that the contribution customers contract because the demand for their products of Chinese import competition to the decline of US manufac- decreases. Customer industries, on the other hand, might ben- turing from 2000 to 2007 is about one third smaller than the efit from import competition affecting their suppliers since they corresponding value by Autor et al. (2013) who do not have a can reduce their costs by switching from domestic to cheaper value chain perspective. The difference regarding the magnitude foreign suppliers, which can raise demand for their products of downstream effects is likely due to an improper measure of and subsequently employment. This would be in line with recent downstream and upstream exposure in the latter studies.6,7 findings showing that firms that source more inputs from abroad The difference in results across these studies raises a second expand production and increase domestic sourcing as well important point. In the age of GVCs, gross trade statistics can (Antràs et al., 2017). be misleading because they ignore complex cross-border pro - Extending the definition of import exposure in this value duction linkages which are better accounted for by trade in chain consistent manner attenuates the findings by Autor et al. value added statistics as highlighted by chapter 1 of this report. (2013) for manufacturing employment. Wang et al. (2018) find In particular, recent research highlights that by relying on gross that commuting zones more exposed to Chinese imports fare trade data many studies on Chinese import competition ignore only slightly worse in terms of manufacturing employment and the high amount of US value added in Chinese exports to the real wage growth than less exposed regions. This is mainly due US, the high services and primary sector content in manufac- to employment creation in downstream industries that expand, turing exports, as well as double counting due to back-and- potentially due to cheaper inputs. Directly exposed industries forth trade.8 Figure 2.2 illustrates this point by showing that the FIGURE 2.2 Value added sources of Chinese manufacturing exports (% share of exports) Manufacturing, nec; recycling Transport equipment Own industry Electrical and optical equipment Machinery, nec Other manufacturing Basic metals and fabricated metal industries Other non-metallic mineral Rubber and plastics Primary industries Chemicals and chemical products Service industries Coke, refined petroleum nuclear fuel Pulp, paper, printing and publishing Wood and products of wood and cork Leather, leather and footwear Textiles and textile products Food, beverages and tobacco 0 25 50 75 100 Notes: Author’s calculations based on ADB-WIOD data for 2015. The chart highlights that the majority of value added in an industry’s manufacturing exports is sourced from other industries, in many cases outside of manufacturing. 50 • Technological innovation, supply chain trade, and workers in a globalized world majority of value added in manufacturing exports is not added Of course, theory can be based on false assumptions and in the exporting industry but in upstream industries, including descriptive statistics can be misleading. Therefore, it is important services. These three factors limit import competition exposure to examine the empirical literature to see if it confirms the pre - of manufacturing industries considerably. When using the more dictions. Different methodologies have been applied to assess appropriate value-added trade statistics, the effect of Chinese the relationship between trade and aggregate employment, and import competition is reduced by about one third (Jakubik and all are broadly in line with theory and simple correlations. For Stolzenburg, 2018). instance, cross-country econometric studies that estimate the In sum, once exports, input-output linkages, and value-added effect of changes in trade policy or trade openness on changes in trade statistics are accounted for, trade seems to have contrib- employment find that trade shocks reduce unemployment mod - uted to the recent decline in manufacturing employment in estly. A 1 per cent decrease in tariffs is estimated to lower unem- advanced economies only to a very small degree, if at all. In fact, ployment by about 0.35 per cent, while a 10 percentage point trade even slowed the decline down in some countries, such as increase in trade openness is found to reduce aggregate unem- Germany. Separate from this, and potentially more important, ployment by about three-quarters of a percentage point (Dutt et is the question of how trade has affected the overall number of al., 2009; Felbermayr et al., 2011). Using novel value-added trade jobs across all sectors. After all, manufacturing is only responsi- statistics, one study finds that the 2004 EU Enlargement led to ble for a minor share of overall employment in most advanced employment gains of up to 0.11% in EU15 countries (Kaplan et economies. This question is discussed in the next section. al., 2018). Similarly, an input-output analysis of trade-induced labor demand changes finds that trade added close to one million 3. Trade, GVCs, and nation-wide labor market jobs to the US economy from 1995 to 2011, mostly due to an outcomes increase in services exports (Feenstra and Sasahara, 2017). An important feature of the study for the purposes of this chapter As said above, other sectors than manufacturing make up for is that it highlights how important input-output linkages are for most of employment in advanced economies, in particular the the outcome. For instance, it shows that exports generated 4.1 services sector. Since more and more services become tradable million additional jobs in services sectors, of which about 1.4 or linked to foreign competition and demand through input-out- million were not due to services exports but rather to services put linkages, aggregate labor market effects of trade have embedded in manufacturing exports through cross-sector value become increasingly distinct from its effect on manufacturing. chain linkages. This further highlights the importance of a value chain perspec- Other studies have estimated the effect of trade on aggre- tive for assessments of the trade and labor market relationship. gate employment by simulating the effects of trade flow or Theories of international trade typically suggest that trade policy changes in structural models of trade. Applications of this should not have a major effect on the aggregate level of employ- approach lead to remarkably similar conclusions. For example, ment. They tend to predict that trade has second order effects an analysis of the rise of Chinese import competition suggests by shifting resources across firms and sectors, which can affect that the US has experienced aggregate employment gains due aggregate employment if labor market frictions are sector or to the expansion of services industries that benefit from cheap firm-specific (Helpman and Itskhoki, 2010; Davis and Harrigan, imported inputs (Adao et al., 2019; Caliendo et al., 2018). A 2011; Carrère et al., 2015). They stress however that the dominant related study on NAFTA emphasizes the relevance of taking determinant of unemployment is country-, sector- and firm-spe - input-output linkages into account in this approach by showing cific labor market institutions. that in their absence some effects are underestimated by 50% Descriptive statistics are broadly in line with what economic (Caliendo and Parro, 2015). theory predicts. In contrast to a relatively widespread perception One study with a similar approach finds that a return to in developed countries, trends in aggregate labor market out- most-favored-nation (MFN) tariffs between Britain and the comes, such as labor force participation, employment-to-pop- EU-27 would lead to significant job losses in both Britain and ulation ratios, the unemployment rate or real wages have not the EU27. The same model predicts that a potential EU-US trade shown dramatic changes since the early 1990s, other than agreement, which would eliminate all import tariffs and reduce those related to the Great Recession (World Trade Organization non-tariff barriers, would create about 0.35 million jobs in the US (WTO), 2017). There is no increasing trend in unemployment or and over 1 million jobs in the EU. It then highlights that between decreasing trend in labor force participation that is common to 60% and 72% of the employment effects of these policy shocks developed countries and could potentially be related to glo- would be due to indirect effects caused by value chain linkages balization or more specifically to the expansion of North-South rather than due to direct effects (Vandenbussche et al., 2017; trade. What can be observed however, are differences in the Vandenbussche et al., 2018).9 An advantage of these two studies levels of employment indicators across countries, which suggests is that they rely on value added trade data and, therefore, avoid that country-specific factors play an important role in explaining the pitfalls of gross trade data pointed out by Jakubik and Stol- labor market outcomes. zenburg (2018). Trade, value chains and labor markets in advanced economies • 51 National effects could also be inferred from intra-coun- However, as was the case with manufacturing employment, try studies. A study exploiting differences in exposure of US once the effects of export expansion, cheaper inputs, and value commuting zones to Chinese import competition finds that chain linkages are added to the equation, the picture changes. labor markets in more exposed regions perform better than Regions home to offshoring firms have benefitted from local less exposed labor markets as negative effects of trade on employment gains due to employment expansion in offshoring manufacturing employment are more than offset by trade-in- firms and positive spill-over effects (Kovak et al., 2017). Moreover, duced gains in services employment (Wang et al., 2018). If the effect of import competition shocks can be attenuated at the less exposed labor markets incur only minimal employment regional level if job losses in some sectors or firms are compen- and wage changes due to trade, then the effects of trade on sated by job creation in other sectors or firms in the same com- labor market outcomes at the national level should be positive. muting zone. Trade opening often means both opening of the Consequently, this approach further confirms the finding that domestic market and improved access to export markets at the trade leads to moderate but positive changes in labor market same time, and firms that gain access to foreign markets raise outcomes. Interestingly, the study also highlights, similar to their exports and generate new jobs. One study concludes that Feenstra and Sasahara (2017) and Caliendo et al. (2018), that since many import-competing regions in the US also export or local job gains in services industries that benefit from cheap benefit from cheaper inputs, the effects basically balance out Chinese manufacturing inputs are of paramount importance for and exposed and unexposed local labor markets follow a similar this finding. trend (Feenstra et al., 2017). Industrial diversification is therefore Given the relative unanimity in the findings of studies that a key aspect for a fast and smooth regional adjustment to trade. apply a large variety of methods, it is safe to conclude that trade Evidence from Germany shows that when regions are too con- has a small but positive effect on aggregate labor market out- centrated, trade can widen regional disparities despite its posi- comes in advanced economies. Due to the changing industrial tive aggregate effect (Yi et al., 2017). structure of advanced economies, these gains mainly accrue As with exports, analyses of input-output linkages can in the services sector where many high-income countries have uncover positive effects of import competition by illustrating a comparative advantage. Once again, assessing the question how industries might benefit from cheaper inputs. The advan- from a value chain perspective has a major impact on the results tage of analysis at the regional level compared to the with- and is central to the conclusion. in-manufacturing level is that one can capture a wider set of industries that benefit from cheaper inputs. After all, many ser- vices industries rely on manufacturing inputs as well. One study 4. Trade, GVCs, and regional divergence finds in this regard that Chinese imports raised US employment in the construction industry alone by 50,000 workers (Caliendo The finding that trade leads to small positive labor market out- et al., 2018). More generally, the inclusion of the services sector comes at the aggregate level could conceal substantial het- in the analysis made possible by comparing regions is important erogeneity in effects at the regional level. Since trade shifts because manufacturing represents only a small share of employ- resources across sectors and firms and since the distribution of ment. According to data from the Federal Reserve Bank of St. sectors and firms across regions is not uniform, trade shocks Louis, manufacturing accounted for only 20% of employment in should affect regions differently depending on their industrial Germany and 10% in the US in 2012. In contrast, industries that structure. Indeed, the same research that was discussed for the use manufacturing imports as inputs cover almost the entirety of debate on the decline in manufacturing shows that the effects of employment. This observation is highlighted in Figure 2.3 which trade on labor markets vary considerably by region. shows by how much industries are exposed to manufacturing Initial work found that in the absence of accounting for import competition, differentiated into three different forms of input-output linkages, rising imports lead to higher unemploy- exposure. The figure shows that while direct import competi- ment, lower labor force participation, and reduced wages in US tion is limited mainly to manufacturing, downstream exposure local labor markets that are more exposed to Chinese imports is prevalent in all sectors. This can explain why downstream relative to less exposed labor markets. This applies to directly employment creation can more than offset employment losses in exposed manufacturing workers, and in terms of wage losses also upstream and directly exposed industries within the same local to workers in non-tradable services industries whose output suf- labor market, as shown by Wang et al. (2018).10 fers from lower regional demand (Autor et al., 2013). Moreover, Finally, using value added data can show a very different workers in exposed local labor markets appear to be reallocated picture of the geography of trade shocks than that indicated by to non-exposed industries and therefore experience greater analysis based on gross trade flows as Figure 2.4 shows. Loca- job churning and reduced lifetime income as a consequence of tions specialized in downstream industries, in particular electrical increased imports from China (Autor et al., 2016; Asquith et al., machinery and electronic equipment, are much less exposed to 2017). These results are also corroborated by studies that look at import competition than what gross imports would suggest. On other advanced economies, including France, Germany, Norway the other hand, the opposite holds for certain locations special- and Spain (Autor et al., 2016; Malgouyres, 2017; Dauth et al., ized in upstream manufacturing, including steel. Two extreme 2014). cases in this regard are San Jose, California, home to Silicon 52 • Technological innovation, supply chain trade, and workers in a globalized world FIGURE 2.3 The different forms of import exposure Direct Competition Upstream Channel Downstream Channel Agriculture Mining Food Textiles Wood Products Paper Products & Printing Refined Petroleum Chemical Products Rubber and Plastics Non-Metallic Products Basic Metals Fabricated Metal Machinery & Equipment Office Machinery Electrical Machinery Motor Vehicles Other Transport Equipment Recycling Electricity, Gas & Water Supply Construction Wholesale & Retail Trade Hotels & Restaurants Transport & Storage Post & Telecommunications Financial Intermediation Real Estate Renting Computer Activities R&D Public Admin Education Health & Social Work Social Services Private Households -0,5 0 0,5 1 1,5 2 0 0,5 1 0 0,5 1 Primary Manufacturing Services Notes: Taken from Wang et al. (2018). Horizontal units indicate annualized percentage changes in exposure to imports from China from 2000 to 2007. The graph shows that downstream import exposure, i.e. exposure stemming from the use of imported inputs, is much larger in scale and scope than upstream and direct exposure. Valley and many of the US’ main electronic equipment manufac- use foreign steel as an input because it implies lower demand for turers, and North-West Indiana, home to the largest steel mill domestic steel. in the US and large aluminium producers. In these commuting To sum up, the effects of trade can differ markedly by region. zones, import competition in value added terms is more than a Areas that benefit from export expansion or cheaper inputs standard deviation different from gross import exposure (Jakubik experience wage and employment growth while areas that com- and Stolzenburg, 2018). US regions specialized in consumer elec- pete with imports or have no access to foreign markets might fall tronic production are clearly less exposed to import competition behind. This creates considerable inequalities between regions, when it is measured in value added terms than what gross trade especially when regions are not sufficiently diversified. Although statistics would suggest, since high-tech imports contain both the general equilibrium effects of trade for aggregate employ- a high amount of US upstream content and inputs from other ment are found to be positive, this highlights the need for policy industries. On the other hand, gross trade statistics miss that intervention to facilitate adjustment in areas most affected by upstream steel producers suffer when goods are imported that import competition. Trade, value chains and labor markets in advanced economies • 53 FIGURE 2.4 Difference in import exposure between gross trade and value added trade 1.85 − 4.76 0.92 − 1.85 0.00 − 0.92 −0.92− 0.00 −1.21− −0.92 Notes: Taken from Jakubik and Stolzenburg (2018). The colour scale differentiates between below and above one standard deviation of gross trade based ex- posure differences. The units are in USD 1000 per worker. The graph shows that a value added import based exposure measure differs in some regions sharply from a gross trade one. 5. Trade, GVCs, and labor market polarization channels through which trade can lead to an increasing demand for skills, not only in developed countries. For instance, an This section finally turns to the question whether trade has con- increase in the relative demand for high-skilled workers can come tributed to a polarization of labor markets in advanced economies. from a trade-induced change in the firm composition. When Polarization refers to a rise in low- and high-skilled employment trade liberalization opens new trading opportunities, the most at the expense of medium-skilled jobs and has been observed productive firms try to seize them and expand their production. across a wide range of advanced economies, as can be seen in At the same time, international trade stiffens competition in the Figure 2.5. Trade may play a role here because it can significantly domestic market, leading the least efficient firms to reduce their affect the composition of employment by task and by occupation, sales or close down. High-productivity expanding firms tend to in addition to trade’s impact on the structure of employment by be more skill-intensive than low-productivity downsizing firms, sector or by region. The new jobs that trade creates in more pro- and therefore this change in firm composition may translate into ductive sectors and firms are not necessarily the same as those an increase in the relative demand for high-skilled workers irre- that disappear in import-competing industries or firms. Rather, spective of the industry specialization (Helpman et al., 2010). In together with technology, trade tends to increase the demand for addition, trade may increase the rewards for skill-biased technical high-skilled workers compared to mid- and low-skilled workers, change, which further raises skill demand (Bustos, 2011). and to decrease the demand for mid-skilled workers performing The rise of value chains is likely to have exacerbated this phe- routine tasks compared to both high- and low-skilled categories. nomenon since it allows for the offshoring of not just complete In order to account for this, it is necessary to examine the effect of production processes but production stages (Baldwin, 2016). trade on the demand for specific tasks, distinguishing in particu- Theory suggests that as offshoring costs fall, firms in developed lar between routine and non-routine tasks. economies can relocate more production stages to developing Both traditional and more recent trade theories predict that economies which will allow the former to technologically upgrade trade should raise the demand for high-skilled relative to low- and specialize in a narrower set of stages that are relatively skill-in- skilled workers. The traditional factor-endowment theory of tensive corresponding to their comparative advantage (Feenstra comparative advantage predicts that trade will increase the rel- and Hanson, 1995; Grossman and Rossi-Hansberg, 2008; Lee ative demand for skills in an advanced economy that is relatively and Yi, 2018). Moreover, foreign demand for high quality goods skill-abundant. More recent theories point out several additional causes exporters to demand technological upgrading also from 54 • Technological innovation, supply chain trade, and workers in a globalized world FIGURE 2.5 Percentage point changes in employment shares by skill level between 1995 and 2015 10 Low skill Middle skill High skill 5 0 -5 -10 -15 Southern Northern Western Total North Japan Central Europe Europe Europe America Europe Notes: Taken from OECD (2017). In many advanced economies the share of low- and high-skilled jobs expanded at the expense of middle-skilled jobs. their suppliers which in turn have access to cheaper but more skill goods, and which were therefore likely to engage in offshoring demanding inputs from abroad. This causes the upskilling effect of the assembly stage, changed their labor force composition to propagate through the supply chain. As a result, skill intensity towards non-production activities such as marketing or distribu- has been shown to increase in close to one third of firms that nei- tion (Biscourp and Kramarz, 2007). Evidence from the same study ther import nor export (Fieler et al., 2018). also shows that all types of offshoring, whether foreign sourc- Empirical research supports the view that international trade ing of final goods or intermediate inputs, are associated with an increases the relative employment of skilled workers in devel- increase in the share of skilled workers such as engineers or tech- oped countries. Detailed information on the skill structure within nicians among the remaining production workers. Interestingly, French manufacturing firms shows that firms employ relatively the employment changes in this study were due to offshoring to more skilled workers in marketing and development when they other OECD countries, suggesting that skills upgrading within sell their products outside of France (Maurin et al., 2002). Other firms from high-income countries is not necessarily linked to off- studies show that import competition leads to skill upgrading shoring to low-wage countries. Rather, it appears to be associ- through its impact on product and process innovation. Using firm- ated with increases in sourcing from foreign markets in general. level data for twelve European countries over the period 1996 to Evidence from the United States further shows that imports of 2007, Bloom et al. (2016) estimate that increased trade with China intermediate inputs from China mildly increased the relative accounted for about 15 per cent of the technology upgrading in employment of non-production workers compared to production Europe between 2000 and 2007. They also show that technology workers (Wright, 2014). This upskilling effect has been magnified upgrading has had a significant impact on the relative employment as domestic value chains have developed in China allowing for an of skilled workers. Supporting this evidence, an analysis of Belgian even finer degree of specialization (Dollar et al., 2018). firms in the same period, 1996 to 2007, shows that import compe- Newly available data on occupational characteristics allows tition from China led to skill upgrading in low-tech industries (Mion researchers to better characterize recent changes in the nature and Zhu, 2013). The findings suggest that the response to imports of work and the tasks required in each occupation beyond the from China accounted for 27 per cent of the increase in the share high- vs low-skilled dichotomy. The types of tasks performed by of non-production workers, and for almost half of the increase in a worker also determine whether a job is suitable to be offshored the share of highly-educated workers in the low-tech industries. and whether it is susceptible to import competition from low- Firm-level evidence from France shows that offshoring is wage countries. Occupations that require repetitive, easily cod- associated with a lower relative demand for production work- ifiable tasks are easy to relocate or automate. Non-routine and ers, especially for the less-skilled ones. Between 1986 and 1992, manual occupations that require abstract thinking, face-to-face French manufacturing firms that increased their imports of final communication, or physical presence are much less tradeable Trade, value chains and labor markets in advanced economies • 55 and automatable. Since routine tasks tend to be medium-skilled, premium, i.e. the ratio of wages commanded by high-skilled and manual tasks low-skilled and abstract tasks high-skilled, labor low-skilled workers. This higher skill premium acts as a signal for market polarization can arise with trade liberalization and techno- workers to increase their skill levels and acquire the appropriate logical progress. type of skills. When skill supply responds to market changes, Firm- and worker-level evidence shows that offshoring and employment of high-skilled workers increases and the skill pre- import competition have a small positive impact on the demand mium tends to decrease. Increases in the skill premium can there- for non-routine occupations and thus on job polarization. Off- fore be an important mechanism in upskilling the labor force and shoring by German multinational enterprises for example is asso - consequently in advancing economic development. Due to labor ciated with an increase in non-routine and interactive tasks per- market rigidities, the response of skill supply to an increased skill formed in the onshore plants, and a higher share of high-skilled premium can take several years, leading to a sustained wage workers, accounting for about 10 to 15 per cent of these changes polarization. The ease of adjustment chiefly depends on workers’ (Becker et al., 2013). Another recent study shows that in Denmark characteristics. While high-skilled workers can adjust to changes import competition from low-wage countries has led to a decline in skill demand more promptly than low-skilled workers, upskilling in routine, mid-skilled manufacturing occupations, and has there- or re-skilling of low-skilled workers takes more time. fore contributed to an overall shift in employment towards both Evidence from the U.S. labor market suggests that low-wage high- and low-skilled occupations (Keller and Utar, 2016). Evi- workers churn primarily among manufacturing sectors, where dence from the United States and Western Europe suggests that they are repeatedly exposed to subsequent trade shocks, while increased import competition measured at the occupation-level high-wage workers are better able to move across employers with (rather than at the industry-level), and offshoring to low-income minimal earnings losses and are more likely to move out of man- countries have brought about a relative decline in real wages, ufacturing conditional on separation (Autor et al., 2014; Krishna especially for low- and mid-skilled occupations intensive in rou- and Senses, 2014). Even when they move outside manufacturing, tine tasks (Ebenstein et al., 2014). Import competition and off- many workers faced with import competition have been shown shoring are also found to reduce employment probabilities and to incur income losses as they land in low-skilled services jobs wages for workers in exposed firms relative to those in export- (Ebenstein et al., 2014). Danish evidence shows that workers in ing firms, thus leading to wage polarization between skill groups occupations that require cognitive skills either stay in mid-wage and firms (Hakkala and Huttunen, 2016; Utar, 2016; Hummels et jobs or move upwards, and therefore are unaffected or benefit al., 2014). Finally, services offshoring also increases the relative from import competition (Keller and Utar, 2016). It also shows demand for high-skilled workers in non-routine occupations but that vocational training with a manufacturing focus makes mid- the effect is economically small (Crinò, 2010; Crinò, 2012). wage workers less vulnerable to wage declines if they stay in their However, studies that take a wider range of potential drivers of job but it does not shield them from being obliged to move into polarization into account find that technology is significantly more low-wage jobs. Post-secondary education and vocational training important in driving polarization than import competition or off- with an information technology focus, on the other hand, pre- shoring in value chains (Goos et al., 2014; Autor et al., 2015; Zhu, vents workers from having to move to low-wage jobs and strongly 2017). Two recent studies stand out for accounting explicitly for increases their chances of moving to high-wage jobs if they face the rise of GVCs. The first builds a task-based model of produc- import competition from a low-wage country. tion in global value chains and decomposes observed changes in To conclude, trade has contributed to an increase in the occupational labor demand into an automation and an offshoring demand for skills and labor market polarization in advanced econ- component. It finds that while both factors have contributed to omies. While it is by far not the most important factor behind polarization in advanced economies, the effect of automation is these trends, its role is relevant. To make sure that the gains from dominant (Reijnders and de Vries, 2018). The second study goes trade are shared more widely across individuals with different a step further and decomposes changes in US labor demand into skills, policy interventions are thus necessary. This is the focus of that due to participation in GVCs, competition from imports of the next section. Chinese final goods, and automation. The results suggest that import competition from China increased the share of low-skilled employment, while participation in GVCs increased the share of 6. Facilitating labor market adjustment to high-skilled employment. Trade as a combination of the two has trade with GVCs thus contributed to polarization. The results for trade are however dwarfed by the estimates for the role of technology (Beverelli et Economic openness, increased trade and investment, further inte- al., 2018). gration in GVCs, and the diffusion of technology create greater Independent of the exact driver, an increase in the demand wealth and opportunities, but they also induce job displacement for high- relative to low- or medium-skilled workers can translate and political discontent. By slowing the adjustment process, into an increase in the share of skilled workers, an increase in the labor and capital market frictions generate an efficiency loss at skill premium or a combination of both. In the short term, the the aggregate level which corresponds to the income and welfare supply of workers with a given skillset tends to be fixed and an that is foregone as the economy performs below its potential. increased demand for skills translates into increases in the skill Evidence suggests that following trade opening, unemployment 56 • Technological innovation, supply chain trade, and workers in a globalized world tends to increase before it decreases. A time-frame of 7 to 10 foreign firms have caused trade shocks to spread more widely in years appears to be necessary for economies to return to their an economy, leading to indirect employment effects up and down new steady state (Arias et al., 2013; Artuç et al., 2010). Three fac- the value chains. Therefore, it is increasingly difficult for adversely tors have been found to determine how easily countries adjust to affected workers in up- or downstream firms to qualify for spe- trade, namely trade balances, the pattern of trade opening and cific adjustment assistance. This means that not all of the adversely the degree of regional diversification (Krugman, 2016a; Hakobyan affected workers may be granted adjustment support, lowering and McLaren, 2016; Yi et al., 2017). the policy’s effectiveness in facilitating adjustment. Although spe- Adjustment processes also raise issues of equity, affecting the cific adjustment policies (e.g., the US Trade Adjustment Assistance political support for an open economy. Even if on average the program or the European Global Adjustment Fund) do cover effects of trade are positive, workers with the wrong skills in neg- workers from adversely-affected, first-tier upstream or down- atively affected regions and/or sectors can suffer important and stream producers, they fail to account for linkages between firms persistent losses. Evidence on the effects of NAFTA on the US further up or down the value chain as well as linkages across bor- labor market suggests, for instance, that despite average nomi- ders. General adjustment policies have the advantage that they nal wages and overall employment remaining largely unaffected, can also support workers in those firms that are indirectly affected certain workers who lived in more exposed areas or worked in but do not qualify for specific adjustment assistance due to size more exposed sectors incurred earnings losses relative to less thresholds or the difficulty to establish a clear chain of causality. exposed peers (Hakobyan and McLaren, 2016). The combined role More generally, non-specific adjustment policies also support of location and industry exposure implied that a blue-collar foot- workers adversely affected by technological change and other wear worker without a high school degree in a town specializing in shocks which induce adjustment processes that are difficult to dis- footwear production was hit across several dimensions. The study entangle from, similar to and easy to confuse with those induced reports that in the most vulnerable regions and industries, high- by trade. Figure 2.6 highlights that the scale and scope of these school dropouts experienced a decrease in wage growth over the policies differs widely across advanced economies. decade of respectively 4 and 17 percentage points compared to Training assistance and education programmes have an similar workers that were less exposed. Evidence for Germany fur- increasingly important role to play in facilitating adjustment to ther shows that the expansion of export-oriented sectors did not trade in global value chains. An important implication of value benefit workers displaced by import competition. Instead, gains chain trade for labor markets is that it has transformed interna- in these industries accrued primarily to workers from the same tional competition, which now impacts economies at a much finer sector, new labor market entrants, or previously unemployed resolution (Baldwin, 2016). Traditionally, countries specialized in workers (Dauth et al., 2016). This is confirmed by Danish data industries in which they were most competitive. With the rise of (Keller and Utar, 2016), which suggests that many displaced mid- global value chains, however, comparative advantage has shifted wage manufacturing workers moved into low-wage services jobs. towards the level of production stages and specific tasks within As a result of these efficiency, equity, and political economy value chains. This has important implications for workers that lose issues, there is a strong rationale for governments to take a closer their jobs in the adjustment process. While before the age of value look at the concerns associated with adjustment processes and chains it was easier to transition from sunset to sunrise sectors to take the necessary action to address them. This section dis- offering their initial human capital, these workers may now face cusses how governments can facilitate adjustment to trade liber- difficulties marketing their initial skill set which might have become alization with a specific focus on the challenges that arise due to obsolete. As their old task might have disappeared altogether, value chains. workers either upgrade their skill sets to perform new different Policies that governments can implement to lower the cost of tasks with equal or better pay or transition without training into adjustment to a changing trade environment fall into three main low-wage jobs (Keller and Utar, 2016). Therefore, training assis- categories: general adjustment policies, which typically consist tance programmes have become increasingly important in adjust- of some combination of active labor-market policies (such as job ment policies compared to employment subsidies or job search search assistance and training) and passive labor-market policies assistance, as they help displaced workers to better respond to (including income support and social insurance programs); spe- the changing demand for skills. Effective training assistance and cific adjustment programmes; and other policies (including educa- education policies promote skills that are relevant for multiple tion, infrastructure, credit market, trade, mobility and place-based industries, increasing workers’ flexibility and resilience in an unpre- policies) that do not directly intervene in labor markets. Available dictable job market (Baldwin, 2016). evidence on the effectiveness of these policies suggests that Taking a value chain perspective and more broadly accounting there is no one-size-fits-all recipe to reduce trade-related adjust- for input-output linkages and exports when assessing the effects ment costs (World Trade Organization (WTO), 2017). of trade on labor markets at the regional level does not alter General adjustment policies – which aim at addressing adjust- the conclusion that these effects are likely to differ considerably ment problems independently of their cause - appear to be more between regions. What it does is to help identify these effects adequate than specific trade adjustment policies for facilitat- with more accuracy, thereby helping with the design of appro- ing workers’ adjustment to trade in the presence of global value priate adjustment policies. Most adjustment policies have a role chains. Increasing input-output linkages between domestic and to play in addressing regional adjustment difficulties but mobility Trade, value chains and labor markets in advanced economies • 57 FIGURE 2.6 Public Spending on Labor Market Policies in OECD countries (% of GDP) Active Labor Market Policies Passive Labor Market Policies Denmark Sweden Finland France Hungary Netherlands Austria Belgium Luxemburg Germany Spain Switzerland Ireland Portugal Italy Norway Poland Employment Incentives Czech Republic Rep. of Korea Start−Up Incentives New Zealand Direct Job Creation United Kingdom Australia Public Employment Services Slovenia and Administration Canada Estonia Sheltered and Supported Employment Slovakia Out−of−work income maintenance and Rehabilitation Chile Early Retirement Israel Training Japan United States 2.00 1.75 1.50 1.25 1.00 0.75 0.50 0.25 0.00 0.00 0.25 0.50 0.75 1.00 1.25 1.50 1.75 2.00 Notes: Data from the OECD Labour Market Programmes database for 2015. UK data for 2011. The graph shows wide variation in scale and scope of labor market policies across OECD countries. and place-based policies are the instruments of choice to address in human capital with policies linked to education, training and regional disparities in adjustment costs. When a region is nega- retraining, and in infrastructure and social capital. However, as tively affected by import competition, some of the workers who production factors, in particular financial and knowledge capital, lose their jobs and cannot find a new one may be willing to move have become much more mobile, competitiveness policies need to a region where they can be re-employed. However, because of to be targeted at those factors that are naturally more “sticky” mobility frictions they may not be able to do so. Mobility policies such as certain types of human, social and physical capital as consist in various measures aimed at lowering or eliminating such well as infrastructure in order to retain the investment’s benefits. frictions. Place-based policies can usefully complement mobil- Moretti (2012) finds that highly skilled labor presents an attractive ity policies by helping those who are negatively affected but are combination of low mobility with high spill-overs whereas finan- not willing to move. They can dampen negative effects of trade cial capital gained for instance through tax breaks will likely flow openness on local labor markets. Finally, new technologies can be to the place with highest return. As production structures are utilized to bring regions that have fallen behind closer to hubs of increasingly fragmented and productive factors increasingly more innovation by reducing face-to-face constraints that are inherent to mobile, sectors have become the wrong operational unit with services delivery and many manufacturing processes. Technology which to frame competitiveness and industrial policies. The focus also can enable training and education programs to reach a much now may rather be on cities as centres of excellence in a particular wider and diverse audience at little cost. This can help to counter- stage of production or developing to become a first-class centre act the forces that promote regional disparities. Importantly, suc- (Baldwin and Evenett, 2012). In this perspective, cities should cessful coordination of the various policies requires cooperation be seen as production hubs that provide a fast-adjusting range between the different levels of government (Alden, 2017). of diverse world-class services, including in particular inputs into Trade in global value chains significantly affects the way gov- manufactured exports, and a corresponding range of good jobs ernments can promote their economies’ competitiveness. Tradi- with a reasonably high level of resilience to international compe- tionally, competitiveness policies aimed at fostering industries tition (Baldwin, 2016). With their localized social capital, cities can with the biggest spill-overs or at correcting market failures. They serve as the breeding ground of innovation and can be seen as the promoted investment in knowledge capital with government competitive frontier of developed economies in the 21st century. sponsored research, private-sector R&D subsidies and tax breaks, With more diversified labor markets they also allow for a greater 58 • Technological innovation, supply chain trade, and workers in a globalized world resilience of workers to adjust to changing economic conditions suppliers, cheaper imports lower costs in downstream firms which and can thereby improve the adjustment process (Yi et al., 2017). allows them to expand. In addition, export expansion has bene- Last but not least, with global value chains trade policy has fitted several manufacturing industries such that the combined become an even more problematic trade adjustment instrument effect of trade on manufacturing employment is likely to be minor. than before, as its consequences for employment have become However, moving from the nation-wide and sectoral level to increasingly difficult to assess. While trade restrictions, for exam- regional and individual outcomes reveals substantial heterogene- ple in the form of safeguards, can help domestic firms to adjust ity in how these aggregate effects map out. For instance, when by temporarily limiting import competition and increasing their local labor markets within countries are not sufficiently diversi- share in the domestic market, they also penalize export-oriented fied, trade can widen regional disparities. Regions specialized industries and industries that benefit from cheaper inputs. Tariffs in import-competing and upstream industries can fall behind, on imports of intermediates, for instance, can increase the sourc- while areas with industries that export or benefit from cost sav- ing cost of domestic exports and thereby worsen their competi- ings pull away. Similarly, trade can lead to labor market polariza- tiveness. Moreover, increased import tariffs can also have a nega- tion by favouring high-skilled employment over medium-skilled tive effect on domestic exporters if these are upstream suppliers employment. While other factors like technological progress of foreign firms adversely affected by the raised tariffs (Barbe and have contributed more significantly to these phenomena, policy Riker, 2017). Vandenbussche et al. (2017) also point at the impor- can ensure a more even distribution of the gains from trade by tance of such cross-border linkages for domestic employment. addressing these inequalities. They estimate that a return to MFN tariffs between the UK and The chapter finds that value chains make targeted or spe- the EU would cause job losses since many British firms are import- cific interventions increasingly difficult. As input-output linkages ant suppliers and buyers of continental European firms. Along the cause trade shocks to spread more widely within economies, same lines, recent research suggests that spill-over effects due import competition is less and less limited in terms of industries, to supply chain linkages between manufacturing and non-manu- regions, or skill levels. As a consequence, it becomes important facturing industries have become more important over time and for policies that ease adjustment to trade to be more general and should be taken into account when shaping trade policy (Kühn less dependent on affected workers fulfilling certain conditions. and Viegelahn, 2018). This is especially the case as value chains magnify trade-induced changes in skill requirements and thereby raise the demand for worker flexibility and the need for training support. 7. Conclusions In sum, this review shows that trade benefits on average not only consumers but also workers. This finding goes against common This chapter reviews research on the relationship between trade views in public discussions and highlights the need for better com- and labor markets in order to assess how trade has affected man- munication on the benefits of trade. It also shows that there is an ufacturing and aggregate employment, as well as regional and important part to play for policy as these benefits tend to cluster skill-based inequality in advanced economies. It adds to the regionally and among individuals with the right characteristics. existing literature by taking a value chain perspective which prop - As always, a number of caveats apply. Beyond those that per- erly accounts for the rise of global value chains in the last three tain to the generalization of country-specific results, two caveats decades. Based on the review, it discusses how policy can facili- are worth emphasizing here. First, this paper does not discuss tate adjustment to international trade when production is increas- the effects of trade in the presence of value chains on other out- ingly fragmented across and within borders. comes such as, for example, labor force participation, employ- It highlights that value chain perspectives in labor market stud- ment volatility, the geographical mobility of workers, the labor ies of international trade are crucial due to three factors. Firstly, share of income, or indirect effects on political, sociological, or value chains imply that trade shocks are felt much more broadly health outcomes, etc. This is mostly because of a lack of evidence, in economies since firms and industries are connected through but also because there seems to be considerable heterogeneity input-output linkages. Secondly, value chains have magnified even among developed countries with regard to the evolution trade’s impact on skill demand by allowing for specialization of these variables. Note, however, that it is not unreasonable to not only across but also within sectors according to compara- assume that evidence concerning these other effects would be tive advantage. Thirdly, value chains make it necessary to com- in line with the evidence on sectoral and regional employment. plement traditional gross trade statistics with novel value added Second, some indirect effects of trade, notably on technology trade statistics in order to correctly measure the volume and geo- and productivity, are not taken into account. This means that a graphical incidence of trade shocks. clean separation of the effects of trade from those of technology, Taking these factors into account shows that trade leads to a notoriously difficult objective to achieve given the strong inter- employment and wage gains at the national level, although in actions between trade and technology and the isomorphic nature the case of employment these are small. At the sectoral level, it of their effects, remains somewhat elusive. This should certainly stresses that trade is unlikely to be a major driver of employment not be seen as an argument against trade adjustment policies but losses in manufacturing due to offsetting factors. While import rather as another reason why general adjustment policies should competition can hurt employment in exposed industries and their be prefered to specific trade adjustment programs. Trade, value chains and labor markets in advanced economies • 59 Notes References 1. These numbers are based on an accounting exercise. They are not Acemoglu, D., Autor, D. H., Dorn, D., Hanson, G. H. and Price, B. (2016), meant to suggest that an equivalent number of jobs would disappear ‘Import Competition and the Great US Employment Sag of the 2000s’, in autarky. Note that the estimates are likely upward biased as export- Journal of Labor Economics 34(S1): S141-S198. ing firms tend to have a higher import content than non-exporting Adao, R., Arkolakis, C., and Esposito, F. (2019), ‘Spatial Linkages, Global firms and higher productivity and current input-output tables are not Shocks, and Local Labor Markets: Theory and Evidence’, NBER Working able to differentiate between the two. Paper No. 25544. 2. Sales and distribution of imports in particular to households are also Ahn, J. and Duval, R. (2017), ‘Trading wtih China: Productivity Gains, Job important contributors to jobs but not covered in this section. See Losses’, IMF Working Paper No. 122, International Monetary Fund, Chapter 8 for details. Washington D.C. 3. See for instance, Krugman’s (1994, p. 25) famous quote: “It should Alden, E. (2017), Failure to Adjust: How Americans Got Left Behind in the be possible to emphasize to students that the level of employment Global Economy, Rowman & Littlefield. is a macroeconomic issue […] with microeconomic policies like tariffs Amiti, M., Dai, M., Feenstra, R. C. and Romalis, J. C. (2017), ‘How Did China’s having little net effect.” WTO Entry Benefit U.S. Consumers?’, NBER Working Paper No. 23487. 4. See, e.g., Helpman (2016), Muendler (2017), and WTO (2017). Antràs, P., Fort, T. C. and Tintelnot, F. (2017), ‘The margins of global sourcing: 5. Authors’ calculation based on OECD-WTO TiVA data. Theory and evidence from US firms’, American Economic Review 107(9): 6. Positive downstream effects of Chinese imports have also been shown 2514-2564. for Japanese firm sales (Fabinger et al., 2017). Arias, J., Artuc, E., Lederman, D. and Rojas, D. (2013), ‘Trade, informal 7. Wang et al. (2018) argue that to properly capture the impact on down- employment and labor adjustment costs’, World Bank Policy Research stream producers it is for instance central to differentiate between Working Paper No. 6614. intermediate and final goods imports, since only the former has the Artuç, E., Chaudhuri, S. and McLaren, J. (2010), ‘Trade shocks and labor potential to reduce input prices. Failing to do so will lead to mea- adjustment: A structural empirical approach’, The American Economic surement error biasing estimates towards zero. Similarly, cross-border Review 100(3): 1008-1045. input trade might differ from intra-country input trade. The assump- Asquith, B. J., Goswami, S., Neumark, D. and Rodriguez-Lopez, A. (2017), tion that US industries source nationally in the same way as from ‘U.S. Job Flows and the China Shock’, NBER Working Paper No. 24080. abroad ignores that countries specialize within value chains and thus Autor, D. H., Dorn, D. and Hanson, G. H. (2013), ‘The China Syndrome: provide different types of inputs. This equally causes measurement Local Labor Market Effects of Import Competition in the United States’, error. American Economic Review 103(6): 2121-2168. 8. In the age of value chains gross trade data suffers from double count- Autor, D. H., Dorn, D. and Hanson, G. H. (2014), ‘Trade adjustment: Worker ing when intermediates cross the same border twice. If for example, level evidence’, The Quarterly Journal of Economics 129(4): 1553-1623. China produces phone cases and ships them to the US where high- Autor, D. H., Dorn, D. and Hanson, G. H. (2015), ‘Untangling Trade and tech components are inserted before the phone travels back to China Technology: Evidence from Local Labour Markets’, Economic Journal for final assembly, then the phone case would be counted twice by 125(584): 621-646. gross trade data. Autor, D. H., Dorn, D. and Hanson, G. H. (2016), ‘The China shock: Learn- 9. Carrere et al. (2015) find a small increase in the US unemployment ing from labor market adjustment to large changes in trade’, Annual rate due to a potential EU-US trade agreement but they don’t account Review of Economics 8: 205-240. for input-output linkages. Moreover, they show that replacing NAFTA Baldwin, R. and Evenett, S. (2012), ‘Value creation and trade in 21st century with 20% import tariffs would increase unemployment by 6% in the manufacturing: What policies for UK manufacturing?’, in Greenaway, D. US and 21% in Mexico, an effect that might become even larger with (ed), The UK in a Global World, [4] London: Center for Economic Policy input-output linkages that magnify sectoral effects. Research: 71-128. 10. As was the case for manufacturing employment, Acemoglu et al. Baldwin, R. (2016), The Great Convergence: Information Technology and (2016) and Caliendo et al. (2018) find a more muted balancing effect the New Globalization, Cambridge (MA): Belknap Press. of input-output linkages on regional disparity due to the issues out- Balsvik, R., Jensen, S. and Salvanes, K. G. (2015), ‘Made in China, sold lined in the section on manufacturing employment. in Norway: Local labor market effects of an import shock’, Journal of Public Economics 127: 137-144. Barbe, A. and Riker, D. (2017), ‘The Effects of Tariffs on Employment in Global Value Chains’, USITC Economics Working Paper 2017-07-A. Becker, S., Ekholm, K. and Muendler, M. A. (2013), ‘Offshoring and the onshore composition of tasks and skills’, Journal of International Economics 90(1): 91-106. Bernard, A. B. and Fort, T. C. (2017), ‘Factoryless Goods Producers in the USA’, in Fontagné, L. and Harrison, A. (eds) , The factory-free economy: Outsourcing, Servitization, and the Future of Industry, [5] Oxford: Oxford Univeristy Press: 136-168. 60 • Technological innovation, supply chain trade, and workers in a globalized world Bernard, A. B., Smeets, V. and Warzynski, F. (2017), ‘Rethinking De-Industri- Ebenstein, A., Harrison, A., McMillan, M. and Phillips, S. (2014), ‘Estimat- alization’, Economic Policy 32(89), 5-38. ing the impact of trade and offshoring on American workers using the Beverelli, C., Rubinova, S., Stolzenburg, V. and Woessner, N. (2018), ‘Did current population surveys’, Review of Economics and Statistics 96(4): Global Value Chains contribute to rising labour market polarization?’, 581-595. WTO, unpublished manuscript. Fabinger, M., Shibuya, Y. and Taniguchi, M. (2017), ‘International Influences Biscourp, P. and Kramarz, F. (2007), ‘Employment, Skill Structure, and Inter- on Japanese Supply Chains’, RIETI Discussion Paper Series 17-E-022. national Trade: Firm-Level Evidence for France’, Journal of International Feenstra, R. C. and Hanson, G. H. (1995), ‘Foreign Investment, Outsourcing Economics 72(1): 22-51. and Relative Wages’, NBER Working Paper No. 5121. Bloom, N., Draka, M. and Van Reenen, J. (2016), ‘Trade induced technical Feenstra, R. C., Ma, H. and Xu, Y. (2017), ‘US Exports and Employment’, change? The impact of Chinese imports on innovation, IT and produc- NBER Working Paper No. 24056. tivity.’, The Review of Economic Studies 83(1): 87-117. Feenstra, R. C. and Sasahara, A. (2017), ‘The ‘China Shock’, Exports and Braakmann, N. and Vogel, A. (2011), ‘How does economic integration influ- U.S. Employment: A Global Input-Output Analysis’, NBER Working ence employment and wages in border regions? The case of the EU Paper No. 24022. enlargement 2004 and Germany’s eastern border’, Review of World Felbermayr, G., Prat, J. and Schmerer, H.-J. (2011), ‘Trade and unem- Economics 147(2): 303-323. ployment: What do the data say?’, European Economic Review 55(6): Bustos, P. (2011), ‘Trade Liberalization, Exports, and Technology Upgrad- 741-758. ing: Evidence on the Impact of MERCOSUR on Argentinian Firms’, Fieler, A. C., Eslava, M. and Xu, D. Y. (2018), ‘Trade, Quality Upgrading, and American Economic Review 101(1): 304-340. Input Linkages: Theory and Evidence from Colombia’, American Eco- Calì, M., Francois, J., Hollweg, C. H., Manchin, M., Oberdabernig, D. A., nomic Review 108(1): 109-146. Rojas-Romagos, H., Rubinova, S., and Tomberger, P. (2016), ‘The Labor Goos, M., Manning, A. and Salomons, A. (2014), ‘Explaining job polariza- Content of Exports Database’, World Bank Policy Research Working tion: routine-biased technological change and offshoring’, American Paper 7615. Economic Review 104(8): 2509-2526. Caliendo, L., Dvorkin, M. and Parro, F. (2018), ‘Trade and Labor Market Grossman, G. M. and Rossi-Hansberg, E. (2008), ‘Trading tasks: A simple Dynamics: General Equilibrium Analysis of the China Trade Shock’, Yale theory of offshoring’, The American Economic Review 98(5): 1978-1997. University, unpublished manuscript. Hakkala, K. N. and Huttunen, K. (2016), ‘Worker-level consequences of Caliendo, L. and Parro, F. (2015), ‘Estimates of the Trade and Welfare import shocks’, VATT Institute for Economic Research Working Paper Effects of NAFTA’, The Review of Economic Studies 82(1): 1-44. No. 74. Crinò, R. (2010), ‘Service Offshoring and White-Collar Employment’, Review Hakobyan, S. and McLaren, J. (2016), ‘Looking for local labor market effects of Economic Studies 77(2): 595-632. of NAFTA’, Review of Economics and Statistics 98(4): 728-741. Crinò, R. (2012), ‘Service Offshoring and the Skill Composition of Labor Handley, K. and Limao, N. (2018), ‘Policy Uncertainty, Trade and Welfare: Demand’, Oxford Bulletin of Economics and Statistics 74(1): 20-57. Theory and Evidence for China and the U.S.’, American Economic Dauth, W., Findeisen, S. and Suedekum, J. (2014), ‘The rise of the East and Review 107(9): 2731-2783. the Far East: German labor markets and trade integration’, Journal of Helpman, E. (2016), ‘Globalization and Wage Inequality’, NBER Working the European Economic Association 12(6): 1643-1675. Paper No. 22944. Dauth, W., Findeisen, S. and Suedekum, J. (2018), ‘Trade and Manufactur- Helpman, E., Itskhoki, O. and Redding, S. (2010), ‘Inequality and unem- ing Jobs in Germany’, American Economic Review 107(5): 337-342. ployment in a global economy’, Econometrica 78(4): 1239-1283. Dauth, W. and Suedekum, J. (2016), ‘Globalization and local profiles of eco- Hicks, M. and Devaraj, S. (2015), ‘The Myth and the Reality of Manufactur- nomic growth and industrial change’, Journal of Economic Geography ing in America’, Ball State University, unpublished manuscript. 16(5): 1007-1034. Hummels, D. L., Jørgensen, R., Munch, J. R. and Xiang, C. (2014), ‘The De Long, J. B. (2017), ‘NAFTA and other trade deals have not gutted Amer- Wage Effects of Offshoring: Evidence from Danish Matched Work- ican manufacturing — period’, https://www.vox.com/the-big-idea/ er-Firm Data’, American Economic Review 104(6): 1597-1629. 2017/1/24/14363148/trade-deals-nafta-wto-china-job-loss-trump. Jakubik, A. and Stolzenburg, V. (2018), ‘The “China Shock” Revis- Dhyne, E. and Rubinova, S. (2016), ‘The supplier network of exporters: Con- ited: Insights From Value Added Trade Flows’, WTO Working Paper necting the dots’, National Bank of Belgium Working Paper Research ERSD-2018-10. No. 296. Kaplan, L. C., Kohl, T. and Martinez-Zarzoso, I. (2018), ‘Supply-chain trade Dollar, D., Kidder, M. and Khan, B. M. (2018), ‘The effect of production frag- and labor market outcomes: The case of the 2004 European Union mentation on skills reallocation: Is it felt equally across levels of devel- enlargement’, Review of International Economics 26(2): 481-506. opment?’, UIBE, unpublished manuscript. 9-25-2018. Donoso, V., Martin, V. and Minondo, A. (2015), ‘Do differences in the expo- Keller, W. and Utar, H. (2016), ‘International trade and job polarization: evi- sure to Chinese imports lead to differences in local labour market out- dence at the worker level’, CESIFO Working Paper No. 5978. comes? An analysis for Spanish provinces’, Regional Studies 49(10): Kovak, B. K., Oldenski, L. and Sly, N. (2017), ‘The Labor Market Effects 1746-1764. of Offshoring by U.S. Multinational Firms: Evidence from Changes in Dutt, P., Mitra, D. and Ranjan, P. (2009), ‘International trade and unemploy- Global Tax Policies’, NBER Working Paper Series No. 23947. ment: Theory and cross-national evidence’, Journal of International Krishna, P. and Senses, M. Z. (2014), ‘International Trade and Labor Income Economics 78(1): 32-44. Risk in the U.S.’, Review of Economic Studies 81(1): 186-218. Trade, value chains and labor markets in advanced economies • 61 Krugman, P. (1994), ‘What Do Undergrads Need to Know About Trade?’, Wang, Z., Wei, S.-J., Yu, X. and Zhu, K. (2018), ‘Re-examining the Effects of American Economic Review 83(2): 23-26. Trading with China on Local Labor Markets: A Supply Chain Perspec- Krugman, P. (2016a), ‘Analytics of Trade Deficits and Manufacturing Employ- tive’, NBER Working Paper No. 24886. ment (Very Wonkish)’, https://krugman.blogs.nytimes.com/2016/12/26/ World Trade Organization (WTO) (2017) World Trade Report 2017: Trade, analytics-of-trade-deficits-and-manufacturing-employment-very-wonkish/. Technology and Jobs, Geneva: WTO. Krugman, P. (2016b), ‘Trade and Manufacturing Employment: No Real Dis- Wright, G. (2014), ‘Revisiting the employment impact of offshoring’, Euro- agreement’, City University of New York, unpublished manuscript. pean Economic Review 66: 63-83. Kühn, S. and Viegelahn, C. (2018), ‘Foreign trade barriers and jobs in global Yi, M., Müller, S. and Stegmaier, J. (2017), ‘Industry Mix, Local Labor Mar- supply chains’, International Labour Review. kets, and the Incidence of Trade Shocks’, US Census Bureau, unpub- Langdon, D. and Lehrman, R. (2012), ‘The Benefits of Manufacturing Jobs’, lished manuscript. U.S. Department of Commerce ESA Issue Brief #01-12. Zhu, W. (2017) Heterogeneous Technological Changes and Their Impacts Lee, E. and Yi, K. M. (2018), ‘Global Value chains and Inequality with Endog- on Income Inequality and Productivity Growth. Queen’s University, enous Labor Supply’, Journal of International Economics (forthcoming). Kingston, Ontario. Magyari, I. (2017), ‘Firm Reorganization, Chinese Imports, and U.S. Manu- facturing Employment’, Columbia University, unpublished manuscript. Malgouyres, C. (2017), ‘The Impact of Chinese Import Competition on the Local Structure of Employment and Wages: Evidence from France’, Journal of Regional Science 57(3): 411-441. Maurin, E., Thoenig, M. and Thesmar, D. (2002), ‘Globalization and the demand for skill: an export based channel’, CEPR Discussion Paper No. 3406. Mion, G. and Zhu, L. (2013), ‘Import competition from and offshoring to China: A curse or blessing for firms?’, Journal of International Economics 89(1): 202-215. Moretti, E. (2012), The New Geography of Jobs, Houghton Mifflin Harcourt. Muendler, M. A. (2017), ‘Trade, Technology, and Prosperity: An Account of Evidence from a Labor-market Perspective’, WTO Working Paper ERSD-2017-15. OECD (2017), ‘OECD Employment Outlook2017’, OECD Publishing, Paris. Pierce, J. R. and Schott, P. K. (2016), ‘The surprisingly swift decline of US manufacturing employment’, The American Economic Review 106(7): 1632-1662. Reijnders, L. S. M. and de Vries, G. J. (2018), ‘Technology, offshoring and the rise of non-routine jobs’, Journal of Development Economics 135: 412-432. Revenga, A. (1992), ‘Exporting Jobs? The Impact of Import Competition on Employment and Wages in U.S. Manufacturing’, Quarterly Journal of Economics 107(1): 255-284. Scott, R. E. (2015), ‘Manufacturing Job Loss’, Economic Policy Institute Issue Brief 402. Spence, M. and Hlatshwayo, S. (2012), ‘The Evolving Structure of the Amer- ican Economy and the Employment Challenge’, Comparative Economic Studies 54(4): 703-738. Tintelnot, F., Kikkawa, A. K., Magne, M. and Dhyne, E. (2017), ‘Trade and Domestic Production Networks’, Society for Economic Dynamics, 2017 Meeting Papers 381. Utar, H. (2016), ‘Workers beneath the Floodgates: Impact of Low-Wage Import Competition and Workers’ Adjustment’, CESifo Working Paper Series No. 6224. Vandenbussche, H., Connell, W. and Simons, W. (2017), ‘Global Value Chains, Trade Shocks and Jobs: An Application to Brexit’, KU Leuven Discussion Paper Series DPS17.13. Vandenbussche, H., Connell, W. and Simons, W. (2018), ‘The cost of non- TTIP: A Global Value Chain Approach’, CEPR Discussion Papers No. 12705. CHAPTER 3 Global value chains and employment in developing economies Claire H. Hollweg (World Bank Group) ABSTRACT The emergence of global value chains – whereby goods that linkages of exporting firms to domestic, input-supplying used to be produced within one country are now fragmented firms. Employment creation and wage gains have been and distributed across global networks of production – has biased towards more skilled workers in developing countries, offered developing countries new opportunities to integrate which contrasts with the predictions of trade theory. The into the global economy. This has also had fundamental skill-biased nature of GVC trade is associated with increased impacts for workers in developing countries. The chapter complexity of global supply chains as well as increased use shows that higher earnings and employment within sectors of skill-intensive inputs, notably services. New emerging and firms is associated with GVC integration, which also sup- trends, including automation and digitization, may further ports other spillovers that operate through labor markets. determine how employment in developing countries will be But it has also had distributional implications of where jobs affected by GVC trade in the future. The findings point to go and the types of jobs they are. Jobs growth has occurred education as well as trade and labor policies as important directly in the export sector, as well as indirectly through factors for strengthening the GVC-labor relationship. • The emergence of GVCs has offered developing countries opportunities to integrate into the global economy, which has had a significant impact on jobs and income in GVC sectors and firms. Integration can have additional benefits for the wider economy as most jobs are generated through upstream domestic supply chains. • Across the developing world, demand for skilled labor is rising. GVCs reinforce this trend by supporting more complex industrial organization and by relying on complementary skill-intensive services inputs. • The impacts of technological change and increased productivity on employment linked to GVCs have been offset by growing consumer demand. 63 64 • Technological innovation, supply chain trade, and workers in a globalized world 1. Introduction T and technology dissemination. Whether and how these spillovers occur through GVC trade is important for policy. he emergence of global value chains – whereby goods Third, is GVC trade associated with increased demand for that used to be produced within one country are now skilled labor in developing countries? And if so, through what fragmented and distributed across global networks channels? The factor-endowment theory of trade predicts that of production – has offered developing countries new trade will reduce returns to unskilled labor in advanced econo- opportunities to integrate into the global economy. Countries no mies while raising returns to capital and skilled labor. However, longer need to develop entire industries to export; firms instead the developing world shows rising relative demand for skilled can access global markets by specializing in specific products or labor, similar to advanced economies. Whether these trends are tasks within a value chain. Today, significant parts of the devel- associated with GVC participation is important to understanding oping world are deeply involved in GVCs, with developing coun- the implications of GVCs for workers. tries’ share in global GVC trade estimated at about 33 percent in Fourth, how are technological advancements within GVC 2011 (Kummritz and Quast 2017). trade affecting jobs and skills in developing countries? New Though powered by new technologies, the economic incen- technologies are transforming the production process and alter- tives for GVC trade were largely driven by access to lower-cost ing our world of work (Hallward-Driemeier and Nayyar 2017), and labor. Offshoring happened initially in assembly activities in many policy makers are deeply concerned about the impact of light manufacturing, but GVCs have since expanded to agricul- automation on assembly jobs in important GVC sectors. tural and services sectors, as well as higher technology and more The chapter shows that GVC integration has supported jobs knowledge intensive manufacturing industries. This unbundling and earnings, as well as other spillovers that operate through of production is expected to have implications for labor markets labor markets. Job and wage gains have been achieved not only – where jobs go, who gets them, and what type of jobs they are within the exporting sector, but indirectly through linkages of (Farole 2015). exporting firms to domestic, input-supplying firms. However, There are reasons to expect that the nationwide employment GVC expansion in developing countries is also associated with effects of GVC integration are different in developing than in higher relative demand for skilled workers. The chapter illus- developed countries. For example, workers in developing coun- trates that characteristics of GVCs themselves, by supporting tries often participate in different segments and tasks within more complex industrial organization, as well as services inputs GVCs than do workers in developed countries. Similarly, the that are complementary to value chains, can be skill-biased. introduction of new technologies may impact the GVC partici- The chapter also shows that technological advancements that pation of developed and developing countries differently. GVCs largely get diffused through global value chains are affecting are a channel for the transmission of new technologies from how GVCs support jobs in developing countries. Evidence sug- developed to developing countries, which could also result in gests that changes in efficiency in GVCs has negative impacts additional spillover effects of participation in terms of learning. on employment linked to countries’ participation in the global Trade is also shown to increase the demand for skills, but the production of products, all else equal. Technological innovation implications for workers may be different in developing countries has also lowered the demand for low-skilled workers relatively that are abundant in unskilled labor and tend to participate in more than compared to high-skilled workers. Nevertheless, the lower-skilled segments of value chains. adverse effects of changing production technologies and effi- This chapter focuses on the nationwide implications of GVC ciencies on employment have been offset by increased con- integration for workers in developing countries, in terms of jobs sumer demand, whereby the domestic consumption expendi- and wages, sector of employment, and skills. It also explores tures in large emerging economies such as China and India will technological change from the perspective of GVCs and its impli- generate new demand for labor for the global economy. cation for jobs and skills going forward. The chapter considers The chapter proceeds as follows. Section 2 surveys the empir- that GVC trade may not have the same effects on developing ical literature linking GVC participation to nationwide job and countries as non-GVC trade does, and takes a GVC focus when wage trends in developing countries. Section 3 discusses other looking at these impacts. In doing so, it addresses four poli- spillovers resulting from features specific to GVCs that operate cy-relevant questions. through countries’ labor markets. The potential links between First, how do GVCs impact jobs and earnings in developing GVC participation and the relative demand for skilled labor in countries? Supporting better jobs and higher wages are a pri- developing countries are examined in Section 4, including spe - mary policy objective, which necessitates an understanding cific features of GVCs that could be behind this trend. Section of the relationship between GVC participation and these labor 5 analyzes the potential implications of new technologies for market outcomes. GVCs from the perspective of labor markets. Section 6 identi- Second, does GVC trade have other development impacts fies policy considerations for developing countries to achieve on workers? There is a large literature showing that export- better labor-market outcomes from GVC participation. Section ing and importing raise productivity. Many of these impacts 7 concludes. happen through labor markets, for example through learning Global value chains and employment in developing economies • 65 2. GVCs and nationwide jobs and earnings in Using the structural decomposition methodology, the change developing economies in employment between two different points of time can be explained by several factors, such as the change in domestic final A first step in analyzing the employment impact of GVCs on demand, domestic production technology, exports and imports. developing countries should consider not only the direct impact In addition, this tool can be used to measure how much of the on jobs and wages, but also the nationwide implications for change in employment is caused by each of these factors. Unlike workers across sectors of the economy. previous work, this chapter follows the approach used in Wang et How do GVCs impact jobs and earnings in developing coun- al. (2017) to explicitly separate the impact of international trade tries? While there is a decades-long catalogue of literature on into that related to traditional trade, simple GVCs and complex the labor market impact of trade, the literature on the relation- GVCs. This not only helps identify how changes in GVC participa- ship between GVCs and labor market outcomes is more nascent, tion affect employment, but also helps measure how other fac- though growing. Empirical studies are often limited to individual tors impact employment via various GVC routes. country studies and have focused mainly on high-income coun- In the model, a change in a country’s employment is decom- tries, although some contributions have focused on developing posed into the change in labor productivity, the change in GVC countries. production networks (further explained by the change in pure Meng, Xiao, and Ye (2018) use a GVC-based structure decom- domestic value chains, simple GVCs, and complex GVCs), and position analysis to identify determinant factors associated with the change in final demand (further explained by the change in the change in employment at the country level. Their decom- the level of final demand, the change in household / government position is based on the World Input-Output Database (WIOD) / investment preferences, and the change in the structure of at constant prices for the period of 2002-2007, 2007-2009, and domestic expenditure). The results are presented in Figure 3.1 2009-2014. The objective is to better understand how GVCs for the most recent period 2009 to 2014 for selected countries affect employment at the country level. including China, India, the United States, Germany and Japan. FIGURE 3.1 Decomposition of the change of a country’s labor between 2009 and 2014  100% 60% Scale of foreign final demand Scale of domestic final demand 80% 48% Structure of foreign final demand Structure of dommestic final demand 60% 36% Preference of foreign investment Preference of foreign government consumption 40% 24% Preference of foreign household consumption 20% 12% Preference of domestic investment Preference of domestic government consumption 0% 0% CHN IND USA DEU JPN Preference of domestic household consumption -20% -12% Complex GVCs Simple GVCs -40% -24% Domestic value chains Labor intensity -60% -36% Total change (right axis) Source: Meng, Xiao, and Ye (2018). 66 • Technological innovation, supply chain trade, and workers in a globalized world For each country, the figure plots the percentage change in jobs influenced employment growth in India, more so in the non-man- due to each of the separate effects, holding all else equal. Sum- ufacturing industries. However, higher forward linkages did not ming across all effects gives the total percent change in jobs at have any significant impact on employment. the country level. Evidence as well as intuition suggests that GVC participation Increased participation in GVCs – especially in complex GVCs will have other distributional implications with respect to where whereby cross-border transactions happen more than twice – is jobs go, the types of jobs they are, and who gets them. For associated with employment increases in all five countries. How- example, GVC participation has also had important implications ever, the positive impacts from increased participation in complex for gender outcomes (see Box 3.1). Shepherd and Stone (2012) GVCs from 2009 to 2014 are getting smaller compared to the ear- find that firms with international linkages hire a larger share of lier periods analyzed in Meng, Xiao and Ye (2018). The impact of female workers, providing evidence that international linkages the change in simple GVCs is no longer positive for all countries, as provide greater opportunities for women to enter the formal it had previously been. Thus, the role played by GVCs in increasing labor market. Women who previously had difficulty accessing employment has fallen since 2002-2007 and 2007-2009. The most this type of wage work have filled many of these jobs (Barrientos, important factors for employment growth are increases in domes- Gereffi, and Rossi 2010). tic and foreign final demand, while labor productivity (output per Employment and wage growth can happen both directly worker) is the most important factor associated with reduced within exporting firms as well as indirectly through these firms’ employment. demand for goods and services from the domestic economy, Other literature confirms the positive association between GVC suggesting other distributional consequences. The extent to participation and employment growth. The 2016 World Bank book which GVCs interact with domestic labor thus depends on the Stitches to Riches (Lopez-Acevedo and Robertson, 2016) shows, linkages of exporting firms to domestic, input-supplying firms. based on data on the apparel sector in South Asia between 2000 Viet Nam is an example of a country that has benefitted greatly and 2010, that when a country experiences a 1 percent increase from trade opportunities, where exports support jobs both in apparel output (which is used as a proxy for apparel exports), directly and indirectly (see Box 3.2). there is a 0.3-0.4 percent increase in employment. This increased The type of jobs also depends on the type of activities under- overall welfare as workers moved out of agriculture or the infor- taken by firms within value chains, which also matters for GVCs mal sector toward these better paying, higher value-added jobs. development impact. As noted by Shephard and Stone (2012), Shingal (2015) summarizes case studies on Vietnamese and Ban- the labor market impacts of assembly operations, which are rel- gladeshi garments, Vietnamese and South African textiles, and atively low wage and low skill, are different from those of more Kenyan and South African horticulture. Overall, these case studies high technology production processes, which tend to be associ- show that GVC-participation is welfare improving, in the sense that ated with stronger relative demand for skilled labor and higher it provides opportunities for employment and income gains. relative wages. For example, the boom in exports to the United Shepherd and Stone (2012) empirically test the relationship States following the US–Viet Nam Bilateral Trade Agreement between labor outcomes and GVC participation using a cross-sec- of 2001 was particularly beneficial to wages of unskilled work- tion of firm-level data for OECD and emerging economies of Brazil, ers, reduced the skill premium, and was a key driver of poverty India, Indonesia, China and South Africa and fixed effect regres- reduction in Viet Nam because it was concentrated in unskilled, sions. They show that firms with the strongest international link- labor-intensive GVC sectors, most notably textiles (Fukase 2013; ages, i.e. that import, export, and are foreign-owned, which serves McCaig 2011). as a proxy for GVC participation, show the highest employment levels. The positive effects of internationalization on labor demand is stronger for emerging markets than for OECD countries. Shep- 3. GVCs and nationwide spillovers through herd and Stone (2012) also find that firms with the strongest inter- labor markets national linkages pay higher wages. The relationship between GVC integration and level of employ- Are there other development impacts for workers of GVC trade? ment is not necessarily positive in all contexts. Imports of goods The opportunity for GVCs to impact labor markets goes beyond and services (backward GVC participation) matter as much as their direct and indirect impact on jobs and wages. There exists exports of intermediates (forward GVC participation) to be success- a large literature on spillover effects of learning by exporting as ful in GVCs, where opening up to imports is often a pre-condition well as learning by importing. To the extent that GVC participa- to successfully export. However, there may be import-competing tion supports domestic firms in developing countries to import effects in labor markets. In the case of India, Banga (2016) examines and export, this can be a key channel by which GVC participation the industry-level impact of participation in GVCs on employment supports spillovers. during 1995-2011. Using fixed effects and GMM estimation tech- First, this can happen through access to information and open niques, the author analyzes how increasing foreign value added in markets (Shepherd and Stone 2012), or by importing inputs that output, foreign value added in exports, and domestic value added contain knowledge and technology. The governance structure in exports of intermediate goods can affect employment growth. of GVC relationships between lead firms and suppliers suggest The results reveal that higher backward linkages have negatively additional mechanisms for knowledge spillovers. Buyers and Global value chains and employment in developing economies • 67 BOX 3.1 GVCs and gender outcomes Evidence shows that women take on a larger share of jobs in sectors. Women workers are more likely to be located in lower labor-intensive value chains than do men. In sectors most value-added components of value chains, are paid less than men, intensively traded in GVCs – such as apparel, footwear, and and are more likely to face problematic working conditions. The electronics – lower-skilled, young, female workers account for benefits of women’s participation in GVCs largely reflect greater the largest share of employment. In the Kenyan and Ugandan numbers of jobs rather than opportunities to work in higher-pay- floriculture GVCs, women represent 65-75 percent of the labor ing jobs within a sector (Bamber and Staritz 2016). force, working mostly in packhouses that offer higher incomes The disadvantages women often face in endowments, than on-farm labor does. In the apparel sector globally, women including assets, education, experience, and social capital, make are a majority of the workforce. For example, in Turkey, two it difficult for them to access the better jobs resulting from par- million of the three million workers in this sector in 2008 were ticipation in GVCs. This was found to be the case in call centers women. In Bangladesh, approximately 80 percent of the gar- in Egypt, where limited access to education, training, promotion ment workers in the same year were women (Kumar 2017). and networks made it difficult for women to take advantage of Yet global value chain participation does not necessarily lead the rising demand for higher technical skills generated by prod- to gender equality. Gender inequalities in GVC participation can uct upgrading (Ahmed 2013). These gender-intensified con- manifest in a number of ways and for a number of reasons. For straints can restrict a country’s ability to remain competitive and example, there is gender segregation both across and within upgrade to higher-value segments of the chain. BOX 3.2 Labor content of Viet Nam’s exports1 Exports have become increasingly important for jobs in Viet exports – taking into account both the direct and indirect Nam. Following the methodology of Calì et al. (2016), Holl- jobs – increased from 5.3 million to 20.5 million, or from less weg (2017) computes the jobs content of Viet Nam’s exports than 1 in 5 jobs to more than 2 in 5 (19% employed directly using a panel of Viet Nam’s input-output tables between by exports and another 21% of the workforce indirectly). 1989 and 2012, matched with sectoral employment data The manufacturing sector dominates the number of jobs from Viet Nam’s statistical yearbook. The calculations yield a in exports. In 2012, manufacturing exports supported 11.8 set of linkages, both direct and indirect, across sectors of the million jobs, or 58 percent of total export jobs, more than a number of jobs employed in export production. From 1989 12-fold increase since 1989. However, most of these export to 2012, the total number of jobs supported by Viet Nam’s jobs are indirect—about 71 percent in 2012. FIGURE 3.2 Direct, indirect (backward) and total jobs content of exports, total (left) and across sectors (right) Jobs (thousands) Manufacturing 25000 Agriculture, forestry and fishery 20000 Trade Hotels and restaurants 15000 Mining Transport, storage and communications 10000 Education and training; health and social work; recreational, cultural and sporting activities Financial intermediation 5000 Real estate, renting and business activities; scientific activities and technology 0 Other services activities 1989 1996 2000 2007 2012 Electricity, gas and water supply Construction Year Public administration, activities of party and membership organizations and defence 0 2000 4000 6000 8000 10000 12000 Total JOBX Direct JOBX Indirect JOBX Direct JOBX Indirect JOBX Source: Hollweg (2017). Calculations using Viet Nam’s IO tables and employment data from Viet Nam’s Statistical Yearbook. 68 • Technological innovation, supply chain trade, and workers in a globalized world BOX 3.3 China’s import penetration in developed vs. developing countries What has been the impact of increasing imports of Chinese enous), an orthogonal set of weights are constructed using intermediate goods, including on labor markets, in other a gravity-model specification to predict cross-country inter- countries? Increased availability of Chinese goods and mediate flows based on bilateral exogenous determinants. services is argued to have put pressure on employment in The authors show that China’s import penetration shock import-competing sectors in other countries. Autor, Dorn, matters for output and value added in trading partners, but and Hanson (2013) and Acemoglu et al. (2016) have shown with differing effects across income groups with winners a negative employment effect from Chinese import pene- and losers (Figure 3.3). For high-income countries, higher tration. Bernard, Jensen, and Schott (2006), Mion and Zhu import penetration of Chinese goods is associated with (2013) and Bloom, Draca, and Van Reenen (2016) also show declines in gross output and value added. However, upper import competition effects on firm employment, survival, middle-income countries appear to have benefitted through technology and innovation. deepening trade integration with China, where higher In the context of shared international production, how- import penetration is associated with higher growth in ever, cheaper intermediate goods from China may also offer gross output as well as value added. For low-income coun- some countries competitive opportunities. Boffa, Santoni, tries, the results are inconclusive, as there is no relationship and Taglioni (2018) quantify the impact of China’s increased once additional controls are added. Overall, China’s market import penetration of intermediate goods in terms of output penetration has been an opportunity, rather than a threat, and value added of partner countries. The authors build a for some developing economies. trade-weighted measure of China’s import penetration, The degree of complementarity or substitution of and empirically relate it with the growth rate of output and domestic production with China’s imports is one factor value added in partner countries. Since countries’ imports potentially driving these varying results. If production struc- from China depend on their own domestic production struc- tures are substitutes, Chinese import penetration may dis- ture, this relationship may be endogenous. To overcome this place local producers. On the other hand, China’s trading endogeneity, Boffa, Santoni, and Taglioni (2018) instrument partners may benefit in terms of value added and output if China’s import penetration with the weighted average Chi- their production structures are complementary to China’s. nese intermediate import penetration in all trade partners. China requires inputs for its own production, which may Rather than using trade weights (which may also be endog- stimulate foreign supply due to interregional linkages. FIGURE 3.3 Estimated effect of China import penetration on gross output and value added  High income Middle income Low income -0,2 -0,1 0 0,1 0,2 0,3 0,4 0,5 0,6 0,7 Value added Output Note: Results of two-stage least squares estimates reported. Only significant estimates reported. Source: Boffa, Santoni and Taglioni (2018). Global value chains and employment in developing economies • 69 suppliers exchange not only goods and services, but also know- opposite trend should occur in developing countries that open how and technology. Based on qualitative data, Gyeke-Dako et up: wages of unskilled workers, clearly the most abundant factor al. (2017) find that firms that are inserted into GVCs in Ghana, in many developing countries, should rise faster than other factor whereby governance structures are characterized by lead firms, rewards. This has not happened in most developing countries; are more likely to have employment strategies that improve the rather, employment creation and wage gains have been biased quality of employment compared with firms that do not have towards more skilled workers. links with lead firms. Learnings effects and feedback loops in Is GVC trade associated with increased demand for skilled tacit knowledge also occurs from using more sophisticated tech- labor in developing countries? And if so, through what channels? nology (MacGarvie, 2006). Similarly, there exist self-reinforcing A recent paper by Farole, Hollweg, and Winkler (2018) focuses complementarities between importing and innovation capabili- on two specific patterns of GVC integration – backward (or ‘buy- ties (Boeler, Moxnes, and Ulltveit-Moe, 2015). ing-side’) and forward (or ‘selling-side’) – to empirically relate Second, employer-sponsored training within GVCs can also changes in GVC integration to changes in the relative demand be an effective mechanism for skills development. In Cambodia, for skilled labor. exporters and foreign firms have a higher incidence of provid - Forward integration is defined by the incorporation of a ing training to workers than non-exporters or domestic firms. A firm’s exports in the production of exports by a third country, in 2012 Employer Skills Needs Survey undertaken by the ILO and other words, supplying intermediate inputs for other countries’ National Employment Agency (NEA) have information on train- exports. For example, the Czech Republic may produce exhaust ing by ownership (foreign, Cambodian), and main market (inter- systems that are incorporated into an automobile produced in national, national, local). Nearly three quarters of foreign firms Germany, or Malaysia may produce microchips that are included and export firms provide training to workers, compared to 57 in US-developed iPhone manufactured in China. This is often percent of domestic firms and 61 percent of firms that service the measured at the country-sector level in terms of overall levels national market. (domestic value added embodied in third-country exports) and Third, better working conditions may also result from GVC in terms of intensity (share of domestic value added embodied in participation, as governments seek to comply with buyers’ stan- third-country exports). dards on health, safety and treatment of workers. Where GVC Backward integration is defined by the use of foreign inputs employment generates better rights and protection for workers, in production that is exported; in other words, buying foreign it can enhance social upgrading. But often this employment is inputs in order to export. For example, Bangladesh may import insecure and unprotected, and there are significant challenges textile fabric produced in Pakistan that is used to make clothing ensuring decent work and pay for more vulnerable workers. The exported by Bangladesh. This is often measured at the coun- downward pricing pressure found in many GVCs has simultane- try-sector level in terms of overall levels (foreign value added ously led to negative social impacts. However, these outcomes in exports) and in terms of intensity (foreign share of total value do not necessarily occur automatically, and policies can support added in exports). better working conditions (discussed below). Farole, Hollweg, and Winkler (2018) use data from the World Fourth, growth and productivity spillovers can also material- Bank’s Labor Content of Exports (LACEX) database of 57 sec- ize for developing countries that participate in GVCs. For exam- tors for the years 2001, 2004, 2007, 2011, and 2014 and about ple, access to cheaper or more diversified varieties and comple - 120 countries. The database uses input-output data from the mentarities between imported inputs and domestic products Global Trade Analysis Project to measure the direct and indirect leads to gains in scope and productivity, which is found to matter wages paid to produce exports by worker type. The labor market more than direct benefits from lower prices or higher quality of outcome is the relative demand for skilled labor (measured as foreign inputs (Goldberg et al., 2010; Halpern, Koren, and Szeidl, wages paid to produce exports to skilled versus unskilled labor). 2015). Boffa, Santoni and Taglioni (2018) show that the increased It includes both the direct wages paid to workers in the export- supply of intermediate products by China has had output and ing sector, as well as the indirect wages paid to workers supply- value added gains for middle-income countries (see Box 3.3). ing domestic inputs to exports. The authors regress the labor market outcome on the log of the measure of GVC participation controlling for log of output at the country, sector and year level 4. GVCs and the relative demand for skilled as well as country-sector, sector-year, and country-year fixed labor effects. The authors then interact the trade measure with a series of dummy variables that take the value of 1 to reflect country The hallmark of globalization is big developing countries open- income level (high income, upper middle income, lower middle ing up and joining global trade. In general, such economies are income, low income) and 0 if not, to detect the joint effect for abundant in unskilled labor and scarce in skilled labor and capi- that dummy of interest. tal relative to global averages. The factor-endowment theory of The estimated coefficient on the measure of GVC partici- trade predicts that trade will reduce returns to unskilled labor pation is presented in Figure 3.4. The authors find that greater in advanced economies while raising returns to capital and returns to skilled labor is correlated with GVC expansion on the skilled labor. This trend has generally been observed. But the buying side. This holds across all income categories, in particular 70 • Technological innovation, supply chain trade, and workers in a globalized world FIGURE 3.4 Relationship between relative demand for skilled labor and GVC participation (a) Buyer (b) Seller Foreign value added share Domestic value added in 3rd country export share hic 0,025 0,0003 hic umic 0,012 -0,0007 umic lmic 0,009 lmic lic 0,02 -0,0001 lic Note: Coefficient estimate reported when statistically significant. hic = high income country; umic = upper middle income country; lmic = lower middle income country; lic = low income country. Source: Farole, Hollweg, and Winkler (2018). high- and low-income countries, resulting in a U-shaped effect. and therefore this change in firm composition may translate into The strong correlation in low-income countries runs counter to an increase in the relative demand for high-skill workers irre- the discourse that GVC integration leads to specialization based spective of the industry specialization (Helpman, Itskhoki, and on comparative advantage, which would presumably result in an Redding 2010). In addition, trade may increase the rewards for increased demand for unskilled (low wage) labor in these coun- skill-biased technical change which further raises skill demand tries. The positive skill-biased affect associated with GVC partic- (Bustos 2011). Some studies attribute the labor demand bias ipation as a buyer happens primarily in input-supplying sectors, against less-educated workers to both GVC participation and with lower relative demand for skills directly in the exporting technological progress (discussed below). sectors. The authors find no correlation in the overall sample This chapter explores three additional channels specific to between GVC participation as a seller and the relative demand GVC trade: (1) the importance of services for GVC trade, (2) the for skilled labor. However, high income countries show a positive increased complexity of industrial organization in GVC trade, relationship. and (3) the skills composition of different activities performed Firm-level analysis also confirms a positive and significant rela- within GVCs. tionship between GVCs and skilled labor. Shepherd and Stone (2012) find a positive and significant relationship between the GVCs and services inputs number of skilled workers and firms with international linkages The emergence of GVCs has been accompanied by important (that import, export, and are foreign owned). Applying propen- changes in the services sector, where services have become crit- sity score matching techniques to firm-level data in a sample of ical for countries’ trade, including participation and upgrading 27 transition economies, Crinò (2012) find that importing inputs in global value chains. Services play a dual role—as inputs into increases the relative demand for skilled labor. Specifically, it manufacturing and agriculture value chains and as value chains of explains more than one quarter of the unconditional difference their own. Much of the value of manufactured goods comes from between importers and non-importers in the employment share inputs of services industries; some studies estimate that services of high skill workers. account for 40 percent of the value added of world trade (Lanz Recent theories point out several channels through which and Maurer 2015). trade can lead to an increasing demand for skills. For instance, an Differences in skills intensity across sectors mean distribu- increase in the relative demand for high-skill workers can come tional implications of GVC participation on labor markets when from a trade-induced change in the firm composition. When GVC participation changes the relative demand for inputs trade liberalization opens new trading opportunities, the most across more- or less-skill intensive sectors. For example, ser- productive firms try to seize them and expand their production. vices jobs are needed to manage the complexity of the supply At the same time, international trade stiffens competition in the chain and preserve production throughout the chain. Exam- domestic market, leading the least efficient firms to reduce their ples include management, financial services, telecommunica- sales or close down. High-productivity expanding firms tend to tions, and other services such as auditors and lawyers (Tagli- be more skill-intensive than low-productivity downsizing firms, oni and Winkler, 2016), which tend to be more skill intensive. Global value chains and employment in developing economies • 71 Calì and Hollweg (2017) show for South Africa that enhanced from the source country as a share of total foreign intermediates GVC participation in GVC-intensive sectors had distributional (i.e. the import share) in Mexico, China, Germany and the United implications for skilled versus unskilled labor. Using the World States. There is a clear positive association between the domes- Bank Labor Content of Exports database, the authors measure tic value chain length of the supplier and the import share from the direct and indirect labor content embodied within South Afri- that supplier. Moreover, the strength of this correlation varies ca’s exports between 2001 and 2011. They find that enhanced across levels of economic development; the correlation is stron- GVC participation in automotives and wearing apparel was asso- ger when the seller is from a high-income country. In sum, buyers ciated with a decline in the relative demand for skilled labor import more from partners who have longer domestic value directly employed in these GVC sectors, and an increase in the chains for all levels of development, but longer supply chains are relative demand for skilled labor indirectly employed in sectors more strongly correlated for high-income sellers. that produce inputs for GVC sectors, in particular services sectors. What could drive this association? Buyers may prefer to estab- Other literature confirms that GVC integration entails the use lish relationships with suppliers that have longer value chains, of upstream inputs that are not only more labor intensive than as a way to reduce the transaction costs in international trade. non-GVC exports, but also more skills intensive, as non-GVC External transaction costs are high in international trade, and exports rely on relatively more commodified upstream inputs they are incurred at both the product level as well as the trading (Farole and Pathikonda, 2017; Taglioni and Winkler, 2016). A case partner level. At the trading partner level, these transaction costs study of the impact of the Japanese Multinational Company could include asymmetric information problems, language barri- on skilled labor in Malaysia shows that the integration of the ers, cultural differences, unfamiliar foreign contractual enforce- subsidiary’s production network into its GVC spurred increas- ment institutions, among others. Longer value chains within the ing needs for skill development, particularly in management supplier therefore reduces transaction costs for the importer. and engineering services (Iberahim 2013). Fernandez-Stark, Second, Kidder and Dollar (2018) empirically test the relation- Bamber, and Gereffi (2010) show that the Chilean offshore ser- ship between domestic value chain length and the skill composi- vices sector typically employs more skilled workers than other tion of the labor force, to determine to what extent lengthening sectors: employees are typically younger, more likely to be of the value chain might be capable of explaining patterns in the male, and hold some level of tertiary education, most often distribution of skills. They regress the share of skilled labor in a from a technical education institution rather than a university. country, industry and year on the upstream global value chain length of the industry, as well as other control variables for 2000- GVCs and complex industrial organization 2008, and find a positive correlation between higher skills and A recent study by Kidder and Dollar (2018) shows that GVC inte - longer global value chains (skills are concentrated in sectors with gration can be biased towards more skilled workers in devel- longer value chain length). To identify the direction of causality oping countries. This happens because (i) GVCs are associated of this relationship, the authors then use an instrumental vari- with more complex industrial organization, and (ii) more complex ables approach by instrumenting value chain length with China’s industrial organization is associated with more skilled labor in trade liberalization, treated as an exogenous “China-shock”. countries that export in GVCs. China is used as an exogenous shock to global value chain First, Kidder and Dollar (2018) construct an average mea- length of Chinese trading partners who are downstream of Chi- sure of the value chain length, proposed by Wang et al. (2017), nese production. Imports into China are influenced by Chinese as a measure of complex industrial organization. That is, indus- import tariffs. The impact of reducing import tariffs on partner tries with longer value chains are considered to be more com- countries’ value chain length will vary with the inherent tradabil- plex. The measure is the weighted average number of produc- ity of sectors, as well as the distance to China. Kee and Tang tion processes within the chain, starting from the product – for (2015) show that the reduction in costs due to lower import tar- example, machinery – to the product’s raw inputs – for example, iffs led Chinese manufactures to substitute out of foreign goods metal. The authors use global input-output data for 2008 from and into domestic varieties. This in turn increased the value chain the World Input-Output Database. They also consider the length length of Chinese goods, which increases the global value chain of the domestic portion of the value chain. Value chain length length of trade partners who import intermediates from China. serves as an indication of how complex the production process Thus, the import tariff reduction had a direct impact on both the has become. Chinese domestic and Chinese global value chains. By interact- The authors find a positive association between firms’ sourc- ing the Chinese tariffs with measures of tradability and distance, ing decisions of foreign intermediate inputs and the foreign mar- Kidder and Dollar (2018) come up with an instrument for value ket’s value chain length. That is, trading partners are more likely chain length that varies across sectors and trading partners. The to import intermediate products from country-industries that authors show that the instrument is a good predictor of trading have longer value chains. As an example, Figure 3.5 shows the partners’ imports of intermediate products from China. destinations in which the manufacturing sectors in Mexico, China, Kidder and Dollar (2018) find that value chain length itself Germany and the United States are sourcing their machinery. On affects the skills composition of the work force. In developed the x-axis is the domestic value chain length of the source coun- economies, there are strong positive effects on high-skilled try. On the y-axis is the log of the share of foreign intermediates labor, as well as moderate positive effects on medium-skilled 72 • Technological innovation, supply chain trade, and workers in a globalized world FIGURE 3.5 Source destination and value chain length  United States Manufacturing Industry 16 China Manufacturing Industry 16 4 RUS Source Country Intermediate Good Penetration Source Country Intermediate Good Penetration 4 TUR SVK ROU TUR AUS POL BRA EST SVN JPN JPN CAN ESP KOR 2 AUT DEU RUS LTU MEX BGR TWN AUT DEU TWN AUS MEX GBRGRC KOR 2 SWE GRCFIN SWE FRA POL EST ROU IDN HUN FRA ITA IND LTU ITA CZE NLD CAN CZE GBR DNK FIN ESP USA DNK NLD PRT 0 IDN HUN BRA 0 IND IRL BGR SVK IRL PRT −2 SVN −2 MLT MLT −4 −4 1.4 1.6 1.8 2 2.2 2.4 1.4 1.6 1.8 2 2.2 2.4 Source Country Value Chain Length Source Country Value Chain Length Germany Manufacturing Industry 16 Mexico Manufacturing Industry 16 3 4 Source Country Intermediate Good Penetration Source Country Intermediate Good Penetration LTU ROU RUS SVN CZE TUR ROU SVN BRA JPN CAN ESP CZE KOR 2 2 DNK POL SVK DEU ITA AUT TWN SWE FIN TWNUSA EST BGR ITA FRA JPN PRT SWE HUN FRA ESP AUT POL IND TUR DNK GBR NLD MLT NLD AUS GRC BGR 1 0 PRT IDN GBR FIN KOR CAN RUS AUS HUN MEX EST BRA −2 IND 0 USA GRC IRL LTU IRL SVK MLT IDN −1 −4 1.4 1.6 1.8 2 2.2 2.4 1.4 1.6 1.8 2 2.2 2.4 Source Country Value Chain Length Source Country Value Chain Length Note: Machinery industry 13 refers to WIOD industry code 13 ‘machinery nec’ (which corresponds to NACE industry 29). Manufacturing industry 16 refers to WIOD industry code 16 ‘manufacturing nec’ (which corresponds to NACE industry 34135). The graph plots the share of machinery industry 13 from each source country used by the manufacturing industry 16 in the United States, China, Germany and Mexico. Source: Kidder and Dollar (2018). labor. In developing economies, the results are a modest pos- China offers an interesting case study. Unlike the experience in itive effect on both high and medium skills. The low-skill labor other developing countries, China’s domestic value-added content share is negatively affected by value chain length in both devel- in exports increased during the 2000s. Chen et al. (2018) analyze oped and developing countries. China’s domestic value added from activities in exports between The results are consistent with the idea that expansion of GVCs 2002 and 2012, to understand whether this increase reflects a modified the usual effects of trade on the demand for factors. movement up the value chain towards higher value-added, Services are the likely mechanism at play, as discussed above. skill-intensive activities. First, skilled labor is needed to manage the value chain, in sectors The authors analyze China’s domestic value added from the like logistics and transport. Second, skilled labor is also needed perspective of activities in exports between 2002 and 2012. in services inputs that are complementary to value chains, such as The authors used occupational data as well as inter-provincial finance, telecommunications, and business services. input-output tables to distinguish between four possible busi- ness activities: R&D, fabrication, marketing, and other support GVCs and upgrading services. The contribution of an activity is the wage income The changing nature of tasks and activities performed within value of workers that perform it, based on their occupation. Data is chains can also impact the relationship between GVC participation available for 42 industries in 31 provinces of China. A functional and the relative demand for skilled labor. For example, higher-value index of specialization (FS index) is constructed, which measures added activities such as research and development, design, brand- the relative specialization of a province for each of the possible ing, sourcing, and customer support that make up important com- business activities. A province is considered to have a relative ponents of GVCs are also relatively more skill intensive. specialization in the business activity if the FS index is above 1. Global value chains and employment in developing economies • 73 FIGURE 3.6 Specialization by Chinese provinces and Figure 3.6 plots the FS index against provincial GDP per capita functions, 2012 in 2012. FS index R&D activities Findings suggest that the increase in China’s domestic value added in exports arose from an expansion of fabrication activ- 2.50 ities in provinces such as Guangdong, Jiangsu, and Zhejiang. Beijing Moreover, there is a negative relationship between specializa- 2.00 tion in fabrication activities and the GDP per capita of Chinese Jiangxi Tianjin provinces (bottom panel of Figure 3.6). However, there is a clear sub-national variation in domestic value added from activities in 1.50 exports. Richer provinces increasingly specialized in higher value Jiangsu Zhejiang Shanghai added activities including R&D and sales and marketing, partic- Tibet Inner Mongolia 1.00 Guizhou Anhui Fujian ularly Beijing, Tianjin, and Shanghai (top and middle panels of Sichuan Gansu Shaanxi Chongqing Shandong Figure 3.6). The changing nature of global value chain activities Shanxi Ningxia Hebei Guangdong Henan Heilongjiang Qinghai Jilin Liaoning that the economy performs therefore has skill implications. 0.50 Guangxi Xinjiang Yunnan Hainan 0 20,000 40,000 60,000 80,000 5. GVCs, technological change, and labor Regional GDP per capita (in yuan) markets FS index Sales and marketing Beijing 2.00 Technology is advancing rapidly, and innovations are increasingly disrupting production patterns around the world. Recent World Bank research shows that the increasing adoption of industrial 1.50 Shanghai automation, data exchange, advanced robotics, smart factories, Inner Mongolia the Internet of Things, and 3D printing – referred to as “Indus- Jiangxi Tianjin try 4.0”, or the fourth industrial revolution – are transforming Chongqing Zhejiang Tibet the manufacturing process and altering our world of work (Hall- 1.00 Fujian ward-Driemeier and Nayyar 2017). Liaoning Jiangsu Guizhou Anhui Guangdong GansuSichuanNingxia Xinjiang Shandong The introduction of these new technologies in production in Yunnan Shanxi Shaanxi Henan Hebei Qinghai Jilin developing countries often takes place through GVCs, where 0.50 Guangxi Heilongjiang Hainan lead firms disseminate technology to their suppliers (Rodrik, 0 20,000 40,000 60,000 80,000 2018). As discussed throughout this report, these technolog- Regional GDP per capita (in yuan) ical advancements are also shaping global value chains, and ultimately the domestic implications of GVC participation. Pol- FS index Fabrication icymakers in many developing countries engaged in GVCs are 1.40 concerned about the impact of Industry 4.0, such as automation Hainan GuangxiHeilongjiang and digitization, on manufacturing assembly jobs and skills, and Yunnan Qinghai Jilin 1.20 Xinjiang Shaanxi Shanxi Hebei ultimately the welfare of their citizens. Henan Shandong Ningxia GansuSichuan Liaoning Innovation will always be disruptive, and for the most part, Guangdong Guizhou Anhui 1.00 Fujian Industry 4.0 can bring new opportunities for developing coun- Chongqing Jiangsu Tibet tries to engage in and achieve the benefits of GVC participation. Zhejiang Inner Mongolia Digital technologies are reducing entry costs into manufactur- 0.80 Shanghai Tianjin ing by reducing the impact of distance. 3-D printing may lower Jiangxi transport costs, lessen the importance of achieving economies 0.60 of scale for manufacturers, and make it easier to manufacture high-quality products. New technologies in the production pro- 0.40 Beijing cess can boost productivity, drive down costs, and support the 0 20,000 40,000 60,000 80,000 speed of technological diffusion and catch-up. E-commerce Regional GDP per capita (in yuan) platforms allow small-scale producers to sell goods directly to consumers, both domestically and for export. The information Note: The functional specialization (FS) index measures the income share revolution has provided new opportunities for developing coun- of the function in exports in a province relative to the income share of the tries to go beyond traditional services exports, such as tourism function in exports in all Chinese provinces. If the FS index is above one, and transport, to export modern services. Many ICT-enabled the province is said to be specialized in that function. professional services – which can be developed without a man- Source: Chen et al. (2018). ufacturing core – can be exported electronically and also be a source of innovation or technology diffusion (Nayyar 2017). 74 • Technological innovation, supply chain trade, and workers in a globalized world Industry 4.0 also poses risks for developing countries, many of inputs. Increased efficiency in a GVC will, ceteris paribus, lead which would be felt within labor markets. Looking ahead, to the to lower demand for jobs in a particular skill type or activity. For degree that new technologies associated with Industry 4.0 may example, if machines replace workers in some of the produc- be labor-saving, the concern is that robotics will replace low- tion tasks in the supply chain, then this will lower the number of skilled assembly jobs in developing countries. FoxConn recently GVC jobs, all else equal (technology within GVCs). On the other replaced 60,000 Chinese factory workers with industrial robots hand, if garment manufactures decide to outsource some jobs (Wakefield 2016). In addition to being a risk of job loss for some to another economy, then the number of jobs is unchanged, but countries, Industry 4.0 may also be one of missed opportunities fewer workers are employed in the outsourcing economy, and for other countries. Robotics, 3-D printing, and other advances more in the receiving economy (task relocation). Efficiency is also raise the possibility of “re-shoring” of routine activities from allowed to vary across economies. If productivity in an economy labor-abundant developing economies back to developed econ- catches up to the productivity leader then fewer jobs would also omies. For instance, there is a concern that the expected migra- be needed to produce the same amount of output (country-level tion of labor-intensive activities from China to poorer economies efficiency). with lower labor costs, such as those in Sub-Saharan Africa, The results of the structural decomposition are presented in might not happen; a re-shoring of manufacturing activity from the bottom panel of Figure 3.7. For each economy, the figure developing back to developed countries may take place instead. plots the percentage change in GVC jobs due to each of the sep- How are technological advancements within GVC trade arate effects, holding all else equal. Summary across all effects affecting jobs and skills in developing countries? A recent study give the total percent change in GVC jobs at the country level. by Bertulfo, Gentile, and de Vries (2019) provides analytical evi- The study finds that technology within GVCs, or changes in dence on this question, focusing specifically on the impact of the efficiency within a specific GVC, is associated with a decrease in acceleration of technological progress within GVCs on jobs. The the levels of employment across all sectors in developing Asia. authors apply a structural decomposition method to examine For example, increases in efficiency would have reduced GVC the drivers of the changes in GVC jobs in 12 developing Asian jobs by about 50 percent in developing Asia, holding all else economies between 2005 and 2015 using regional input output equal. The estimated effect is smaller in services than in agricul- tables and labor force survey data.2 GVC jobs are defined as the ture and manufacturing. Efficiency gains within economies has jobs in a country linked to participation in global production of also negatively affected employment levels. For example, GVC a particular set of products (Timmer et al. 2014; Meng, Peters, jobs would have been lower by about 20 percent holding all else and Wang 2015). GVC jobs are further disaggregated by skill equal in developing Asia. type (low, medium, high skilled) and business activity (R&D, pro - Nevertheless, the adverse effects of changing production duction, logistics / sales / marketing, administration / back office, technologies and efficiencies on employment have been offset and headquarter workers). by increased consumer demand. Demand for goods and ser- The decomposition of the change of employment is presented vices from a new Asian middle class in particular has increased in the top panel of Figure 3.7. The total change in employment employment levels. In developing Asia, for example, the increase in each economy can be separated into changes due to: (i) shifts in employment associated with own-country income is 80 per- within GVCs, resulting from changes within the production struc- cent, versus 8 per cent due to increased income from the rest of ture of the GVC of a specific final product; (ii) shifts between the world. The findings also suggest that the domestic consump - GVCs, resulting from changes in consumer demand for differ- tion expenditures in large emerging economies such as China ent products; or (iii) shifts due to changes in global demand for and India will generate new demand for labor for the global goods and services, which is separated between domestic or economy. The impact on employment of task relocation between foreign demand (ADB 2018). For example, if consumers increase economies that participate in GVCs is smaller and mixed. their demand for services and lower their demand for manufac- The interaction of GVC expansion and technological change tured goods, then employment would rise in services sectors has distributional consequences. The results of Bertulfo, Gen- and decline in manufacturing (shifts between GVCs). And if tile, and de Vries (2019) suggest that technological change in income increases in either the domestic or foreign economies, GVCs has been skill biased in developing Asian economies. then employment will rise to meet the higher demand for goods The authors separate employment into routine and non-routine (shifts due to changes in demand driven by income growth). occupations, to understand how technology is impacting the Shifts within GVCs are further separated into: (i) technology skill profile of GVC-related jobs. Routine tasks include occupa- within GVCs, or changes in employment associated with changes tions such as craft and related trade workers, plant and machine in efficiency within a specific GVC; (ii) task relocation, or changes operators, or clerical support workers. Nonroutine tasks include in employment as the location changes for one or more produc- services and sales workers, managers, professionals, or techni- tion tasks; and (iii) country-level efficiency, or changes in employ- cians. Job losses due to the implementation of technology along ment from efficiency changes in an economy that participates GVCs have been associated with a decline in both routine and in GVCs (ADB 2018). Technological progress is measured as a nonroutine employment levels. However, the share of nonroutine change in the efficiency units of labor, which is determined by (cognitive) employment has increased, meaning these types of the technical production requirements in terms of intermediate occupations are becoming relatively more important in GVCs. Global value chains and employment in developing economies • 75 Using a similar methodology and country coverage but for Recent advances in automation have also sparked concerns the periods 2000-2011, de Vries et al. (2016) look at changes in over the impact on jobs within global value chains, particularly GVC jobs by business activities due to technological change. The assembly. For example, a recent study by Chang et al. (2016) results suggest that technological change within GVCs had a big estimate that as many as 88% of Cambodian, 86% of Vietnam- effect on the relative demand for higher value added and more ese and 64% of Indonesian wage workers could face possible skill-intensive activities, such as R&D and headquarter jobs. For replacement by automation. Nevertheless, tradable goods such example, improvements in GVC technology lowered demand as textiles, garments, and footwear continue to be labor inten- for production workers by about 55 million workers in China, sive and do not feature much automation yet (Box 3.4). but hardly affected demand for R&D jobs. This is suggestive evidence of technological change being to the benefit of more knowledge-intensive activities, as well as changing the sectoral 6. Policy considerations for strengthening the profile. While automation will likely reduce the number of tradi- labor-GVC relationship tional manufacturing jobs, new service jobs will also be created within the chain. The skills required for workers in these positions The evidence above illustrates how integration and upgrading is less industry-specific, and more related to learning the soft- in GVCs has implications for labor markets in developing coun- ware programs and basic computer skills (Frederick 2018). tries. Policy can support the relationship between labor and GVC FIGURE 3.7 Decomposing changes in labor demand  Technology Own country Change in Task relocation Within GVC Income Employment Country-level efficiency Rest of the world Between GVC First level Second level Bangladesh India Indonesia Malaysia Philippines China Republic of Korea Chinese Taipei Thailand Viet Nam Developing Asia % -80 -40 0 40 80 -60 -40 -20 0 20 40 0 20 40 60 80 100 Technology within a GVC Task relocation Income from own country Country-level efficiency Income from the rest of the world Note: Because manufacturing excludes the industry subsectors electricity, gas and water supply and construction, “all sectors” is larger than the sum of agricul - ture, manufacturing and services. Developing Asia in the decomposition analysis includes Bangladesh, China, India, Indonesia, the Republic of Korea, Malaysia, Mongolia, the Philippines, Sri Lanka, Chinese Taipei, Thailand, and Viet Nam. Source: Bertulfo, Gentile, and de Vries (2019) and ADB (2018). 76 • Technological innovation, supply chain trade, and workers in a globalized world BOX 3.4 Automation in the apparel value chain Automation in the apparel industry is complex. Automat- considered. Flextronics is an example of a company in elec- ing a production process typically occurs because: (1) it is tronics that is now rapidly pursuing how to use their operat- expensive to hire people to do the job; (2) the product has ing model in industries such as footwear. the potential to be contaminated if handled; or (3) the task There are, however, reasons to believe that the spread of is repetitive with minimal changes. Apparel, particularly the automation in the apparel industry will not be widespread sewing segment, does not meet these requirements. There in the immediate future. The Sewbot is designed primarily has historically been a pool of low cost labor from a global for t-shirts with possible expansion to jeans. T-shirts are perspective, contamination is not an issue, and whereas the relatively simple to produce, with production consisting of task is repetitive, it changes often. For these reasons, there 13 separate tasks, from quality inspection to heat transfer has been minimal demand to automate most parts of the to collar and label attachment to steaming and hemming. apparel supply chain. Jeans, on the other hand, involve some 30-40 operations, Recent advances in automation have led to assessments while a dress shirt with pockets has 78 steps, or six times that the apparel sector could be vulnerable to disruptive the number of operations to make a t-shirt (Barrie, 2017). change. One report concludes that the broader textiles, Moreover, Sewbot’s executives believe the technology is ill- clothing and footwear industry faced higher automation suited for more complex designs and concede that low-cost risks than workers in automotive and auto parts; electron- locations such as Bangladesh are likely to still have signifi- ics and electrical parts; business process outsourcing; and cant competitive advantages (Bain, 2018; Peters, 2017). retail value chains (Chang et al. 2016). ASEAN nations could Another limitation is that these machines are expensive be in a precarious situation, according to this analysis. and are often only purchased by multinational enterprises The first robotic, automated production line for apparel (MNEs) that serve global buyers from various destinations could be operational by the end of 2018 with larger-scale around the world. They do this to ensure products are implementation further ahead (Stacey and Nicolaou, 2017). the same and because their production volumes are high Softwear Automation produces a clothes-making robot enough to warrant the capital expenditures. called “Sewbot.” The system was being installed in a facil- A final critical element that makes robotic automation ity in the United States with the expectation of producing more difficult in apparel is the fact that fabrics are flexible 1.2 million t-shirts per year at a price that is competitive (drape, soft). This makes them difficult to handle through with manufacturing and shipping the same material in low- automation because tension needs to be applied in various wage locations (Peters, 2017). Large-volume buyers such as degrees depending on the desired aesthetics of the prod- Walmart have expressed interest in the technology, part- uct. The look and feel are key drivers of apparel purchasing, nering with Softwear on trial projects (Stacey and Nicolaou, so if these important elements are compromised by auto- 2017). In the case of the Sewbot for t-shirts, its developers mation, they are far less likely to be used. Similarly, given claim that one operator overseeing a t-shirt line can carry the high cost of investment, apparel manufacturers will be out the tasks of 10 operators in approximately half the hesitant to purchase machinery until it is proven to be a time (Barrie, 2017). There have been similar advancements reliable replacement for human workers. There are nascent in related industries, such as footwear, in which companies technologies that might enable firms to address some of such as Adidas have used computerized knitting technolo - the challenges associated with handling soft materials. Yet gies for shoe mesh at facilities in Europe and North America even with these developments, Crystal Group, the largest (Emont, 2018). In the longer-term, firms from other indus- clothing manufacturer in the world, has expanded produc- tries that have been more modular and automated for tion in Bangladesh and Viet Nam, with its CEO specifically decades may also move into the apparel industry and bring stating that robots could not compete with humans (Bain, production model knowledge that creates new oppor- 2018). tunities for automation that have not yet been seriously Source: Frederick (2018). participation in different ways. This chapter considers policies First, to the extent the GVC participation supports better that (i) support participation of developing countries in GVCs, jobs and higher wages, policies that support developing coun- (ii) foster positive spillovers from GVC participation, (iii) induce tries to participate in GVCs can support these outcomes. Tagli- upgrading to higher value added tasks within GVCs, and (iv) oni and Winkler (2016) provide a thorough review of policies mediate negative effects, such as increased inequality. that can foster GVC participation of developing countries. Key Global value chains and employment in developing economies • 77 areas include trade policy, trade infrastructure, and labor market international trade, which prepares them to better absorb FDI flexibility. spillovers. Overall, studies confirm that FDI spillovers are larger A country’s trade policy shapes the amount and type of for- in countries that are more open towards trade (Meyer and Sinani eign investment and, thus, influences the potential of GVC inte- 2009; Du, Harrison, and Jefferson, 2011; Havranek and Irsova 2011). gration to influence labor markets. Open trade regimes may be Labor market regulations may also influence the effect of GVC more likely to attract foreign investors than inward-oriented integration on domestic firms through various channels. Labor regimes, since they are less constrained by the size and efficiency market regulations, and, in particular, wage constraints, can affect of the local market (Crespo and Fontoura 2007). Foreign investors skills in a firm, and hence their absorptive capacity (Hale and Long might also be more export-oriented in an open setting, increas- 2011). Overly rigid labor markets can reduce the likelihood of labor ing chances for local suppliers to become exporters, too. More- turnover and GVC spillovers. Conversely, overly flexible labor mar- over, foreign investors in an open trade setting are globally more kets may generate frequent labor turnover, which reduces the time integrated and therefore adopt the newest technologies (Meyer for domestic workers to acquire skills and knowledge from foreign and Sinani 2009). Others, however, argue that foreign investors in firms. Kummritz, Taglioni, and Winkler (2017) find that labor free- an outward-oriented trade setting might focus more strongly on dom tends to increase the value-added gains from GVC integra- international distribution and marketing, while foreign firms in an tion as a buyer and seller. Thus, the policy environment can medi- inward-oriented policy regime might bring newer technologies to ate better labor market outcomes from GVC participation. the host countries (Crespo and Fontoura 2007). Third, policies can influence the activities that countries under- Connective trade infrastructure, firm capabilities, and devel- take in GVCs. If the nature of GVC participation matters for the oping standards also influence GVC participation. Improving types of jobs it supports, then policies can support better types trade facilitation and developing more competitive trade logistics of GVC participation. As noted by Shepherd and Stone (2012), sectors to compete effectively in an environment that requires policies that are designed to help firms — in a non-distortionary seamless importing and exporting are key recommendations for way — to move through GVCs to positions of higher value added strengthening GVC participation (Taglioni and Winkler 2016). are likely to help promote the beneficial labor market effects of Higher labor market flexibility – in absolute terms as well as rel- GVC participation. Education and training, as well as infrastructure ative to the level in the foreign investor’s home country – is also development, and backbone services sectors can also help firms shown to have a positive effect on the chances of securing initial to successfully internationalize in higher value added activities. foreign investment (Javorcik and Spatareanu 2005). Human capital, for example, may influence the quality and avail- Second, how GVC participation and labor markets inter- ability of workers. act can also be influenced by policy. Different structural con- However, better conditions for workers within GVCs does not ditions, endowments, and policies may play a role in mediating necessarily follow from greater GVC participation (Milberg and the trade-labor relationship and fostering the positive spillovers Winkler 2011). Social upgrading can be fostered by labor reg- that can occur. The literature confirms that education and skills ulations, such as those for occupational safety, health and envi- influence the share of human capital in firms and are particularly ronment standards in GVC sites. For example, Hollweg and Kanz important for expanding trade integration and spillovers from (2018) use firm-level data from the ILO-IFC Better Work Viet Nam FDI in developing countries (Farole and Winkler 2014). Meyer and program to assess the relationship between transparency on Sinani (2009) show evidence that the share of workers with tertiary working conditions and firm compliance in the apparel sector in education significantly affects FDI spillovers. This relationship Viet Nam between 2010 and 2018. The authors find that while takes a U-shaped form, that is, only below or above certain thresh- continued participation in the Better Work Viet Nam program old levels of human capital does the extent of spillovers increase has the strongest effect on changes in firm compliance with labor (Meyer and Sinani 2009). Tytell and Yudaeva (2007) find for Roma- standards over time, the public disclosure of firms’ names that fail nia that productivity spillovers from foreign direct investment (FDI) to comply with critical labor issues is also associated with increase in manufacturing are significantly lower in regions with a low share compliance. The effects are stronger in some compliance points of education. Farole and Winkler (2014) confirm for a sample of including occupational health and safety, work time, and child 78 low- and middle-income countries that a country’s government labor. spending on education as a percentage of GDP has a strongly Fourth, if GVCs tend to be associated with greater inequality positive productivity effect. Kummritz, Taglioni, and Winkler (2017) by increasing the relative demand for skilled labor in develop- measure the positive impact of skills building on the value-added ing countries, then policy has an important role to pay in ensur- gains from GVC integration as a seller in a sample of 61 countries. ing that the gains from trade are shared evenly (see Chapter 2). A higher expected number of years of schooling (Barro and Lee Complementary policies are likely to play a vital role. As noted 2013), the share of workers with a secondary degree, or higher, by Shepherd and Stone (2012), GVCs could have stronger effects in the total workforce (WDI 2018), as well as better educational on inequality in the absence of education and training policies quality (WEF 2018) all show positive interaction terms with GVC designed to promote workforce and human capital development. integration. Well-functioning labor markets are also important, because inte- Trade policy also affects domestic firms. Local firms in an open grating into GVCs generates faster growth and transformation, trade regime are more exposed to competitive pressures through and require economy-wide adjustment (see Chapter 2). 78 • Technological innovation, supply chain trade, and workers in a globalized world 7. Conclusions other development impacts that operate through labor mar- kets. Job and wage gains have been achieved not only within the Today, significant parts of the developing world are deeply exporting sector, but indirectly through linkages of exporting involved in GVCs. The unbundling of production is expected to firms to domestic, input-supplying firms. Employment and wage have significant implications for labor markets – where jobs go, gains have been biased towards more skilled workers, which who gets them, and what type of jobs they are (Farole 2015). The contrasts with the predictions of trade theory. The skill-biased labor market impacts of GVC participation, as well as the impact nature of GVC trade is also associated with increased complexity that future megatrends will have on labor markets, are a principal of global supply chains as well as increased use of skill-intensive concern to policymakers in developing countries. inputs, notably services. New emerging trends, including auto- The chapter focused on nationwide implications of GVC inte- mation and digitization, may further determine how developing gration for workers in developing countries, from the perspec- countries will be affected by GVC trade in the future. tive of nationwide jobs and wages, nationwide spillovers, and the Policies also play an important role in mediating the relation- relative demand for skilled versus unskilled labor. It also analyzed ship between GVCs and employment in developing countries. the potential implications of new technologies for GVCs from the These include policies that support (i) participation of develop- perspective of labor markets. ing countries in GVCs, (ii) fostering positive spillovers from GVC The chapter showed that, while GVC participation has been participation, (iii) upgrading to higher value-added tasks within important for jobs and wages, it also has had distributional con- GVCs, and (iv) mediating negative effects from winners, such as sequences for where jobs go and the types of jobs available. skilled versus unskilled labor. GVC integration has supported jobs and earnings, as well as Global value chains and employment in developing economies • 79 Notes References 1. Note that the estimates are likely upward biased as exporting firms in ADB (2018). Asian Development Outlook 2018: How Technology Affects particular are likely to have higher import content than non-exporting Jobs. Manila: Asian Development Bank. firms and higher productivity and current input-output based tables Acemoglu, D., D. Autor, D. Dorn, and G. H. Hanson (2016). “Import are not able to differentiate between the two. Competition and the Great US Employment Sag of the 2000s.” Jour- 2. Bangladesh, China, India, Indonesia, the Republic of Korea, Malaysia, nal of Labor Economics 34(1): 141-98. Mongolia, the Philippines, Sri Lanka, Chinese Taipei, Thailand, and Ahmed (2013). “Global Value Chains, Economic Upgrading, and Gender Viet Nam. in the Call Center Industry.” In C. Staritz and J. G. Reis (Eds.), Global Value Chains, Economic Upgrading and Gender: Case studies of the horticulture, tourism and call center industries (2013). Washington, DC: World Bank. Autor, D. H., D. Dorn, and G. H. Hanson (2013). “The China Syndrome: Local labor market effects of import competition in the United States.” American Economic Review 103(6): 2121-68. Bain, M. (2018). “The World’s Largest Clothing Maker Isn’t Betting on Auto- mation Replacing Cheap Human Labor.” Quartz. Available at: https:// qz.com/1169397/crystal-group-is-investing-in-low-wage-labor-not- robots-after-its-ipo/ Bamber, P., and C. Staritz (2016). The Gender Dimensions of Global Value Chains. Geneva: International Center for Trade and Sustainable Development. Banga, K. (2016). “Impact of Global Value Chains on Employment in India.” Journal of Economic Integration 31(3): 631-73. Barrie, L. (2017). “Sewbots Pave the Way for the Apparel Factory of the Future.” Just-Style. Available at: https://www.just-style.com/anal- ysis/sewbots-pave-the-way-for-the-apparel-factory-of-the-future_ id130877.aspx. Barrientos, S., G. Gereffi, and A. Rossi (2010). “Economic and Social Upgrading in Global Production Networks: Developing a framework for analysis.” Capturing the Gains Working Paper 2010/-3, University of Manchester, United Kingdom. Bernard, A. B., J. B. Jensen, and P. K. Schott (2006). “Survival of the Best Fit: Exposure to low-wage countries and the (uneven) growth of U.S. manufacturing plants.” Journal of International Economics 68(1): 219-37. Bertulfo, D., E. Gentile, and G. de Vries (2019). “The Employment Effects of Technological Innovation and Participation in Global Value Chains: Evidence from Asia.” Economics Working Paper Series No. 572, Asian Development Bank, Manila. Bloom, N., M. Draca, and J. Van Reenen (2016). “Trade Induced Tech- nical Change? The impact of Chinese imports on innovation, IT and productivity.” The Review of Economic Studies 83(1): 87-117. Boeler, E. A., A. Moxnes, and K. H. Ulltveit-Moe (2015). “R&D, Interna- tional Sourcing, and the Joint Impact on Firm Performance.” American Economic Review 105(12): 3704-39. Boffa, M., G. Santoni, and D. Taglioni (2018). “From China with Love.” Mimeo, World Bank, Washington, DC. Bustos, P. (2011). “Trade Liberalization, Exports and Technology Upgrad- ing: Evidence on the impact of MERCOSUR on Argentinian firms.” American Economic Review 101(1): 304-40. Calì, M., J. Francois, C. H. Hollweg, M. Manchin, D. A. Oberdabernig, H. Rojas-Romagosa, S. Rubinova, and P. Tomberger (2016). “The Labor Content of Exports Database.” Policy Research Working Paper No. 7615, World Bank Group, Washington, DC. 80 • Technological innovation, supply chain trade, and workers in a globalized world Calì, M., and C. H. Hollweg (2016). “How Much Labor Do South African Hale, G., and C. Long (2011). “Did Foreign Direct Investment Put an Exports Contain?” Policy Research Working Paper No. 8037, World Upward Pressure on Wages in China?” IMF Economic Review 59: Bank Group, Washington, DC. 404-30. Chang, J.-H., G. Rynhart, and P. Huynh (2016). ASEAN in Transformation: Hallward-Driemeier, M., and G. Nayyar (2017). Trouble in the Making? The Textiles, clothing and footwear: Refashioning the future. Geneva: Inter- future of manufacturing-led development. Washington, DC: World national Labor Organization (ILO), Bureau for Employers’ Activities. Bank. Chen, Q., Y. Gao, J. Pei, G. de Vries, and F. Wang (2018). “Formerly Assem- Halpern, L., M. Koren, and A. Szeidl (2015). “Imported Inputs and Produc- bled, But Now Designed in China? Assessing the domestic value-added tivity.” American Economic Review 105(12): 3660-703. of activities in gross exports.” GGDC Research Memorandum 180, Uni- Havranek, T., and Z. Irsova (2011). “Estimating Vertical Spillovers from FDI: versity of Groningen, Groningen, Netherlands. Why results vary and what the true effect is.” Journal of International Crespo, N., and M. Fontoura (2007). “Determinant Factors of FDI Spill- Economics 85: 234-44. overs: What do we really know?” World Development 35(3): 410-25. Helpman, E., O. Itskhoki, and S. Redding (2010). “Inequality and Unem- Crinò, R. (2012). “Service Offshoring and the Skill Composition of Labour ployment in a Global Economy.” Econometrica 78(4): 1239-83. Demand.” Oxford Bulletin of Economics and Statistics 74(1): 20-57. Hollweg, C. H. (2017). “How Do Exports Support Labor in Vietnam?” de Vries, G., Q. Chen, R. Hasan, and Z. Li (2016). “Skills and Activity Mimeo, World Bank, Washington, DC. Upgrading in Global Value Chains: Trends and drivers for Asia.” Eco- Hollweg, C. H., and M. Kanz (2018). “Firm Compliance and Public Disclo- nomics Working Paper Series No. 496, Asian Development Bank, sure in Vietnam.” Mimeo, World Bank, Washington, DC. Iberahim, H. (2013). “Vertical Integration into Global Value Chains and Its Manila. Effects on Skill Development.” IEEE Business Engineering and Industrial Du, L., A. Harrison, and G. Jefferson (2011). “FDI Spillovers and Industrial Applications Colloquium: 28-33. Policy: The role of tariffs and tax holidays.” NBER Working Paper No. Javorcik, B., and M. Spatareanu (2005). “Do Foreign Investors Care about 16767, National Bureau of Economic Research, Cambridge, MA. Labor Market Regulations?” Review of World Economics 141(3): Emont, J. (2018). “The Robots are Coming for Garment Workers: That’s 375-403. good for the US, bad for poor countries.” Washington Street Journal. Kee, H. L., and H. Tang (2015). “Domestic Value Added in Exports: Theory Available at: www.wsj.com/articles/the-robots-are-coming-for-garment- and firm evidence from China.” Policy Research Working Paper No. workers-thats-good-for-the-u-s-bad-for-poor-countries-1518797631. 7491, World Bank Group, Washington, DC. Farole, T. (2015). Factory South Africa? SACU in Global Value Chains. World Kidder, M., and D. Dollar (2018). “The Effect of Production Fragmentation Bank: Washington, DC. on Skills Reallocation: Is it felt equally across levels of development?” Farole, T., C. H. Hollweg, and D. Winkler (2018). “Trade in Global Value Mimeo, The Research Institute of Global Value Chains, Beijing. Chains: An Assessment of Labor Market Implications.” Mimeo, World Kumar, R. (2017). “Global Value Chains: A way to create more, better and Bank, Washington, DC. inclusive jobs.” Jobs Development Blog, World Bank, Washington, DC. Farole, T., and V. Pathikonda (2017). “The Capabilities Driving Participation Kummritz, V., and B. Quast (2017). “Global value chains in develop- in Global Value Chains.” Journal of International Commerce, Econom- ing economies.” Vox EU. Available at: https://voxeu.org/article/ ics and Policy 8(1): 1750006. globa-value-chains-developing-economies. Farole, T., and D. Winkler (2014). “The Role of Mediating Factors for FDI Kummritz, V., D. Taglioni, and D. Winkler (2017). “Economic Upgrading Spillovers in Developing Countries: Evidence from a global dataset.” In through Global Value Chain Participation Which Policies Increase the T. Farole and D. Winkler (Eds.), Making Foreign Direct Investment Work Value Added Gains?” Policy Research Working Paper No. 8007, World for Sub-Saharan Africa: Local Spillovers and Competitiveness in Global Bank, Washington, DC. Value Chains. Washington, DC: World Bank. Lanz, R., and A. Maurer (2015). “Services and Global Value Chains: Some Fernandez-Stark, K., P. Bamber, and G. Gereffi (2010). “Workforce Devel- evidence on servicification of manufacturing and services networks.” opment in Chile’s Offshore Services Value Chain.” Center on Globaliza- Working Paper ERSD-2015003, World Trade Organization, Geneva. tion, Governance and Competitiveness, Duke University, Durham, NC. Lopez-Acevedo, G., and R. Robertson (2016). Stitches to Riches? Apparel Frederick, S. (2018). “Apparel Skills Mapping and Functional Upgrading in Employment, Trade, and Economic Development in South Asia. Wash- Cambodia: Jobs diagnostic.” Mimeo, World Bank, Washington, DC. ington, DC: World Bank. Fukase, E. (2013). “Export Liberalization, Job Creation and the Skill Pre- MacGarvie, M. (2006). “Do Firms Learn from International Trade?” Review mium: Evidence from the US-Vietnam Bilateral Trade Agreement.” of Economics and Statistics 88(1): 46-60. Policy Research Working Paper No. 6419, World Bank, Washington, DC. McCaig, B. (2011). “Exporting Out of Poverty: Provincial poverty in Viet- Goldberg, P. K., A. K. Khandelwal, N. Pavcnik, and P. Topalova (2010). nam and US market access.” Journal of International Economics 85(1): “Imported Intermediate Inputs and Domestic Product Growth: Evidence 102-13. from India.” The Quarterly Journal of Economics 125(4): 1727-67. Meng, B., G. Peters, and Z. Wang (2015). “Tracing CO2 Emissions in Global Gyeke-Dako, A., A. D. Oduro, F. E. Turkson, P. T. Baffour, and E. N. Abbey Value Chains.” Discussion Paper No. 486, Institute of Developing Econ- (2017). “Ghana’s Participation in Global Value Chains: The employment omies, Japan External Trade Organization, Tokyo. effects.” R4D Working Paper No. 2017/05, World Trade Institute, Uni- Meng, B., H. Xiao, and J. Ye (2018). “A Global Value Chain Based Structure versity of Bern, Switzerland. Decomposition Analysis on the Change of Employment.” Mimeo. Global value chains and employment in developing economies • 81 Meyer, K., and E. Sinani (2009). “When and Where Does Foreign Direct Stacey, K., and A. Nicolaou (2017). “Stitched Up by Robots: The threat Investment Generate Positive Spillovers? A meta-analysis.” Journal of to emerging economies.” Financial Times. Available at: www.ft.com/ International Business Studies 40(7): 1075-94. content/9f146ab6-621c-11e7-91a7-502f7ee26895. Milberg, W., and D. Winkler (2011). “Economic and Social Upgrading in Taglioni, D., and D. Winkler (2016). Making Global Value Chains Work for Global Production Networks: Problems of theory and measurement.” Development. Washington, DC: World Bank. International Labour Review 150(3-4): 341-65. Timmer, M. P., A. A. Erumban, B. Los, R. Stehrer, and G. J. de Vries (2014). Mion, G., and L. Zhu (2013). “Import Competition from and Offshoring to “Slicing Up Global Value Chains.” Journal of Economic Perspectives China: A curse or blessing for firms?” Journal of International Econom- 28(2): 99-118. ics 89(1): 202-15. Tytell, I., and K. Yudaeva (2007). “The Role of FDI in Eastern Europe and Nayyar, G. (2017). “The Future of Manufacturing-Led Development in New Independent States: New channels for the spillover effect.” In K. Rwanda.” Mimeo, World Bank, Washington, DC. Liebscher, J. Christl, P. Mooslechner and D. Ritzberger-Grünwald (Eds.), Peters, A. (2017). “This T-shirt Sewing Robot Could Radically Shift the Apparel Foreign Direct Investment in Europe: A changing landscape, Edward Industry.” Fast Company. Available at: www.fastcompany.com/40454692/ Elgar, Cheltenham, UK and Northampton, MA, USA. this-t-shirt-sewing-robot-could-radically-shift-the-apparel-industry. Wakefield, J. (2016). “Foxconn replaces ‘60,000 factory workers with Rodrik, D. (2018). “New Technologies, Global Value Chains, and Develop- robots.” BBC News. Available at: https://www.bbc.com/news/ ing Economies.” NBER Working Paper No. 25164, National Bureau of technology-36376966. Economic Research, Cambridge, MA. Wang, W., S.-J. Wei, X. Yu, and K. Zhu (2017). “Characterizing Global Value Shepherd, B., and S. Stone (2012). “Global Production Networks and Chains: Production length and upstreamness.” NBER Working Paper Employment: A developing country perspective.” Working Paper TAD/ No. 23261, National Bureau of Economic Research, Cambridge, MA. TC/WP(2012)29, Trade and Agriculture Directorate, Trade Committee , Organisation for Economic Co-operation and Development, Paris. Shingal, A. (2015). “Labour Market Effects of Integration into GVCs: Review of literature.” R4D Working Paper 2015/10, World Trade Institute, Uni- versity of Bern, Switzerland. CHAPTER 4 Technological progress, diffusion, and opportunities for developing countries: lessons from China* Satoshi Inomata (IDE-JETRO) and Daria Taglioni (World Bank Group) ABSTRACT The nature of technology used in products plays a major products in a local market, resulting in intense price com- role in determining the governance structure of value chains petition and limited technology transfer. By contrast, tech- and the benefits of participation for developing countries. nology that facilitates scope for product modification and Standardization through breaking production into modules greater interaction with technology owners can help boost with a high degree of functional autonomy (limited mutual technology transfer and product upgrading by develop- interference between modules) can dramatically reduce ing country firms. The chapter illustrates this interaction the amount of research and development (R&D), learning between changes in technology and opportunities for by doing, and the number of complementary skills needed developing countries through developments in the automo- to produce a good. This greatly increases opportunities for tive and mobile phone handset industries, with a particular developing country firms to participate in formerly capi- reference to China’s growth experience. It also finds that tal-intensive industries through reducing entry costs into automation is likely to have only a limited impact on devel- global value chains. However, widespread access to stan- oping countries’ opportunities to participate in value chains dardized products with little ability to modify technical through the offshoring of production by high-income coun- features can lead to an excessive supply of homogeneous tries, at least in the short term. • Policies for helping domestically owned firms become technologically standalone – what some might refer to as “techno-nationalism” – do not necessarily help countries move into higher value- added production within GVCs. Instead, policymakers should encourage firms to be full partners in global technology ecosystems and to pursue open source innovation solutions. • Automation might become a threat to developing country employment in the long term if consumption does not increase fast enough to generate sufficient additional labor demand to offset the labor-saving impact of technological change. In the short term, however, automation will not dramatically reduce the attractiveness of low-wage destinations, especially for labor-intensive tasks that require human dexterity, such as in the apparel industry. • While automation does not pose immediate risks, governments need to develop a comprehensive digital strategy to maximize the gains from GVCs. * This chapter draws from background studies and ongoing research collaboration with the following researchers: Chiara Criscuolo, Yoshihiro Hashiguchi, Keiko Ito, Jonathan Timmis, Ke Ding, Shiro Hioki, Mai Fujita, Tim Sturgeon, Eric Thun, Yuqing Xing, Satoshi Nakano, Kazuhiko Nishimura and Jiyoung Kim. 83 84 • Technological innovation, supply chain trade, and workers in a globalized world 1. Introduction A mildly associated with lower foreign direct investment (FDI) flows from richer to poorer countries. Yet, automation is not going to n increasing number of developing countries is rec- dramatically reduce the attractiveness of low-wage destinations ognizing that participation in global value chains in the near term, especially for labor-intensive tasks that require (GVCs) is an important prerequisite for economic human dexterity. In the apparel industry, for example, soft mate - development. At the same time, however, they fear rials like fabrics are difficult to handle through automation com - that the prospects for value chain upgrading is limited, because pared to solid materials such as metal or wooden objects, and once they join a value chain their production activities become sewing/stitching can still be out of the reach of robots’ hands “locked in” to the lower value-added segments of global pro- (see the evidence in Section 4.) And even in highly automatized duction systems. Added to these concerns, they fear that new industries such as electronics, human fingers are still needed for labor-saving technologies, such as robotization and automation the assembly of devices made of thousands of tiny components. in manufacturing, could erode the previous attractiveness of a This is, for example, the case for smartphones, as discussed in cheap labor force as a source of comparative advantage. The interviews with manufacturers. A bigger challenge for GVC new- analysis presented in this chapter shows that joining and upgrad- comers is rather to be competitive vis-à-vis existing production ing in GVCs is still possible, provided that firms’ strategies and clusters and countries with high density of supply chains. These policy interventions adapt themselves to the new and evolving induce to lower costs for various support functions and services, technology environment. beyond automation, and the density of supply chains supports Standardization, modularity and digitalization have made responsiveness. even complex technologies progressively more “diffusable” over The rest of the chapter is organized as follows. In Section 2, the years, and this represents a new opportunity for firms from we discuss the role of the GVC power relations in determining developing countries to join and move up the value chain. Stan- the way that technological progress creates opportunities for new dardization and modularity tend to increase as a technologies entrants. In Section 3, we illustrate how two specific technologi- and products mature, managers try to reduce uncertainty and cal and business innovations, i.e. production modularization and lower costs, and best practices get codified in the supply base. platforms (a digital evolution of modularization), have played a Today digital technologies enable standardization and modular- fundamental role in opening up opportunities for new entrants ity in increasingly complex features, products, and transactions. in technology-intensive industries such as automotive and mobile The greater spillover resulting from standardization and modu- phones. Section 4 discusses what strategies have allowed new larity allows faster diffusion of technology. The digitalization of manufacturers to leverage the opportunities from production many complex industrial productions enables even more firms modularization and platforms to upgrade because entry per se and countries to leapfrog to more advanced technology. We does not translate into immediate technological progress for use case study evidence from the automotive and mobile phone these entrants. Then, Section 5 discusses one area of great public industries to support this thesis. We show that more standard- concern recently: robots and automation. Section 6 presents ization, or less complexity, of both products and production some policy implications from the discussion. processes in value chains that are typically technology-intensive, such as the automotive and mobile phone ones lowered the entry costs into complex, technology intensive, products. Stan- 2. Technological progress and value chain dardization has dramatically reduced the amount of R&D, learn- dynamics ing by doing, and the number of complementary skills needed to produce a car or a phone handset. The extent to which technological progress will disrupt the Modularization and standardization lower the entry costs to present configuration of supply chains and open them to new product upgrading, but this does not translate automatically players depends, in part, on the form of GVC power relations. into technological advancement for the manufacturers. To move Inomata (2017) employs the analytical framework developed up to high value-added segments of technologically advanced by Gereffi, Humphrey, and Sturgeon (2005) to show how power value chains requires learning additional and complementary relations between buyers and suppliers, as determined by the skills, even though they may be unrelated to some parts of man- nature of transactions and the capabilities in the supply base, ufacturing activities (e.g. marketing, sales, etc.). Our discussion affect opportunities for new participants in GVCs. Gereffi et al. on the success and upgrading of the Chinese smartphone indus- (ibid.) define five forms of GVC power relations: market-type, try offers an example of the strategies that have allowed some modular-type, relational-type, captive-type, and hierarchy-type firms to leverage technological progress to upgrade, get closer (see Annex), and among them the modular-type GVCs are par- to the global technology frontier, and become global brands. ticularly interesting for our discussion. A “module” generally In this chapter we also conclude that the need to graduate refers to a composite of subcomponents grouped by the type of from labor-intensive production is not urgent. We show that function assumed in the final product. Each module has a high automation reduces some of the incentives for GVCs to relocate degree of functional autonomy (namely, the mutual interference to lower wage countries: the rising stock of industrial robots in between modules is small), while the standardized architec- high-income countries over the period 2003-2015 appears to be ture of a module’s interface makes it easy to combine multiple Technological progress, diffusion, and opportunities for developing countries: lessons from China • 85 modules. Modularization can be employed in manufacturing However, developments in design schemes have spurred of complex products, where production processes are simpli- changes that have increased modularity in the auto industry. fied and partitioned. In modular types of production, knowl- Large scale modularization in the automotive sector is already edge-intensive segments (such as the harmonization of core two decades old (Takeishi and Fujimoto, 2001). 2 Here however, components) are limited to only a few stages of the production we focus on a few recent examples. In 2013, Nissan introduced process. a design scheme called the “common module family” into the Accordingly, modularization reduces technological barri- production lines of several key models. The scheme’s objective ers to entry. It lowers the amount of R&D and learning-by-do - was to reconfigure the production system so as to reduce costs ing necessary to integrate into skill- and capital-intensive value yet also maintain the variety of product line-ups. This was pur- chains (Sturgeon and Thun, 2019, and Xing, 2018). Chesebrough sued through the modularization of products, which increased and Kusunoki (2001) further note that modularization also tends the proportion of standardized common components that to reduce product uniqueness – a feature associated with high can be shared among different models, while also reducing value added – which they refer to as “the modularity trap”. Firms costs through bulk purchases of common inputs. Even before adopting the same modules basically produce very similar prod- the introduction of Nissan’s scheme, Volkswagen devised the ucts. This undermines firms’ profitability, mainly due to the high “modular transverse matrix platform” to develop a wide range levels of competition. Section 2 will discuss this in more detail. of different products, including its standard models, such Therefore, adoption of advanced modules alone does not as the Golf, as well as luxury cars, such as Audi. Toyota later generate technological progress in manufacturers. Modulariza - adopted the “Toyota new global architecture” for Prius in 2015, tion helps to move into more complex value chains. But, in order while Hyundai Motors, aided by its fully-automated assembly to capture more value and increase profit margins, firms also system, engaged in the large-scale outsourcing of its main car need to learn to manage more complex processes (i.e. a pro - components, including the cockpit and chassis (Nikkei Busi- cess where a higher number of complementary skills is needed), ness, 2013). and to master more complex tasks (i.e. tasks with some features The implementation of modularization schemes has opened that makes them unique). The smile curve shows there compe- up new opportunities for firms from developing countries. As tences need to be developed to escape the modularity trap: the discussed earlier, modularization simplifies the production right-hand side (downstream) edge of the curve, where local of a complex product by reducing knowledge-intensive seg - firms capture value through branding and product ownership. ments of production (such as the harmonization of car compo- This requires developing expertise in business functions such nents), with the effect of substantially lowering technological as design and marketing. These are capabilities very different barriers to market entry. For example, Shenyang Aerospace from production skills, as are the features of the ecosystem and Mitsubishi Motors China ran a joint business with a US auto- institutions that support them. These topics will be the subject parts supplier, Delphi, to sell engines, transmissions, and other of Section 3. core system components to local car manufacturers in China (Oshika et al., 2009). Engines and transmission systems were generally produced in-house. However, digital technology now 3. Opportunities from modularization and makes it possible to pre-adjust the components to the specifi - platforms: examples from the car and mobile cations of a customer’s individual car models with the help of phone industries electronic control units (ECUs).3 Local manufacturers were thus able to enter the low-end Chinese car market without the need 3.1 Automotive industry: the modularization of cars to develop sophisticated in-house technology. Firms such as Manufacturers in the automotive industry tend to show hierar- Chery, BYD, and Geely were able to produce inexpensive, small chical power relations. A car is an extremely complex system cars that meet the needs of first-time car buyers. Between 1995 containing over 15,000 different components, including key and 2010, domestic firms increased their share of the Chinese components that are often design-specific and difficult to sub - car market by 31.9 percent (Brandt and Thun, 2016). Obviously, stitute. During the assembly stage, the parts must be carefully the lower barriers to entry generated by opening up access aligned with one another in harmony, and the risk of interfer- to platform technology were only partly responsible for this ence between parts is not uncommon. For example, the “com - spectacular rise in market share. Market structure and compe - puterization” of modern cars has increased the risk that the tition, ownership, and the mode of foreign entry are also cru- air-conditioning system will interfere with electronic-intensive cially important in determining the scope for innovation and modules, which need to be located nearby within a narrow upgrading. space between the engine and the instrument panel. Because The gradual transformation of the automotive industry’s of the high degree of manufacturing complexity, the automo - value chains from the hierarchical-type to the modular-type tive industry is highly prone to vertical integration and therefore was associated with an increased ability to codify transactions.4 to adopting a GVC power relation of the hierarchical type. This Codification has enabled firms to unbundle tasks (design, fab - ensures a holistic and systematic coordination of every aspect of rication, assembly, and marketing), and for competition to take production from start to finish.1 place in specific segments of production, rather than at the 86 • Technological innovation, supply chain trade, and workers in a globalized world level of the whole industry, as traditionally envisaged in classi- as Shanzhai, to capture significant market shares despite having cal theories. As a result, the automotive industry changed from low expertise in core aspects of mobile phones technology. a vertically-integrated production system to one where value These brands grew from a share of less than 5 percent of the chains operate in a more open environment, thus increas- domestic mobile-phone market in 1999 to more than 50 percent ing opportunities for emerging companies in developing by 2003 (see Figure 4.1). Their business model consisted in cater- countries.5 ing the domestic markets with low-priced handsets, which they were able to produce at low cost by leveraging the platforms’ 3.2 Electronic equipment industry: from modularization digital technology. to the emergence of platforms and platform leaders The electronic equipment industry covers a wide range of The modularization of the final products’ architecture is what products, from personal computers (PCs) to mobile commu- makes platforms effective in allowing newcomers to the GVCs. A nication devices. Typically, the industry’s supply chains are platform is a complete module on its own that does not require characterized by long supply lines that connect global buyers surrounding components to have any product-specific attributes, with electronic hardware manufacturers and assemblers. except those regarding connection. Accordingly, any parts sup - Global buyers are manufacturers of final consumer products, pliers that have adopted the platform’s interface can enter the such as Apple, Hewlett-Packard, Toshiba, NEC, Samsung, and market. Correspondingly, this tends to invite a massive entry of LG, that organize and preside over their own global produc- producers into the industry. In the case of China’s mobile-phone tion networks and tend to be located in traditional knowledge industry, the marketing strategy of MediaTek, the chip vendor clusters. Suppliers tend to be dispersed nationally, region- from Chinese Taipei mentioned earlier that came into the inte- ally, and globally. 6 The long supply lines are the result of grated circuit chip market in China. Shiu and Imai (2009) argue the delinking of innovation, design, heavy engineering, and that the company boosted their influence in the industry by standard-setting from production and assembly (a GVC pat- devising a unique marketing strategy. Alongside the production tern common in technologically-intensive industries).7 This, together with the standardization of many information-com- and sales of chipsets, they also offered an assembly blueprint for munication technology (ICT) processes, including important mobile phone terminals as a package bundle. The blueprint pro- ones, 8 led to a modular type of power relations for electronic vided a thorough how-to guideline for producing mobile phones equipment GVCs. that embody its chipsets, such as the layout of parts configu- At the turn of the new millennium, platforms and platform ration and electrical wiring, and even included a list of recom- leaders emerged as dominant new players in the electronics mended parts suppliers.10 equipment GVCs.9 A “platform” is defined as “a set of common The turnkey solution of MediaTek’s platform, however, components, modules, or parts from which a stream of deriva- turned out to be a double-edged sword.11 While it enabled tive products can be efficiently created and launched” by “con- local manufacturers with limited knowledge and experience straining the linkages among the other components” (Baldwin to enter the mobile handset market, it also became difficult and Woodard, 2009). Platforms are built on core technology for them to differentiate their final products, and little tech- modules which define the fundamental technical parameters of nology and know-how was transferred to the manufactures of the products manufactured through the platform. A large-scale the low-cost handsets. There are two reasons for this. First, as integrated circuit, which often determines the performance level of the final product in which it is embedded, provides a good part of its marketing strategy, MediaTek decided to disclose example of a core technology module. A platform leader is a firm only about 20 percent of its software source code, leaving that controls core technology modules, and therefore governs the remaining 80% “black-boxed”. This meant that users of the final product’s functions and performance. Such companies their platform ecosystem were bound to produce products are still predominantly from rich countries. Yet, over the years, whose designs were highly subordinate to the platform’s platform leaders from developing countries have also emerged. interface specification. The second factor was that the plat- Particularly notable is the emergence of MediaTek as one of five form invited massive entry of producers into the market, as dominant global players in mobile phone processors applica- discussed earlier. This resulted in excessive supply of homo - tions, and the associated dominance in China’s mobile-phone geneous goods for those manufacturers using the platform, market. We will discuss their role in what follows. as well as market fragmentation, severe price competition The advent of platforms has significantly destabilized the tra- and low profit margins. Under these conditions, producers ditional set-up of electronic equipment GVCs. The mobile phone had very limited room for expenses in R&D or innovation that industry in China illustrates well the potential for disruption by platforms. By integrating most of the mobile phone’s function- could have encouraged upgrading.12 The GVC power rela- alities, platform solutions (sometimes referred to as “reference tions of the industry was also affected. As China experienced designs”) have lowered the cost and time required by manufac- an excess supply of undifferentiated mobile-phone terminals, turers to design low-end mobile phones. This has allowed Chi- the industry’s value chains went from the modular type to the nese brands, especially producers of imitative products, known market type.13 Technological progress, diffusion, and opportunities for developing countries: lessons from China • 87 FIGURE 4.1 The mobile-phone market in China prior to the smartphone  100% 90% 2007 China market share 80% Foreign brands Top 5 local brands 70% Nokia 28.9 Lenovo 6.5 60% Motorola 18.8 Nigbo-Bird 4.3 50% 40% Samsung 11.1 Amoi 3.2 30% Sony-Ericsson 5.5 ZTE 2.3 20% LG 3.1 Konka 1.8 10% Top 5 67.4 Top 5 18.1 0% foreign Share local Share 1999 2000 2001 2002 2003 2004 2005 2006 2007 Q3 Foreign Brands Local Brands (Authorized) Estimated Share of Local Brands Including Unauthorized Market Note: Because of the lack of coverage of unauthorized makers’ shipment, the share of local brands is biased downward. The broken line represents a rough estimate of the share of local brands including unauthorized makers. Source: Imai and Shiu (2011), compiled from estimates by MII, CCID, and Pday Research Center. 4. Upgrading options smartphone market. Furthermore, the domestic market posi- tions of Chinese firms have also changed significantly (Table 4.1 Surviving the price wars: options for firms in 4.2). From 2010 onward, Chinese products gained market share developing countries in products with mid-range prices, while still keeping their abso- What are the options for firms in developing countries to avoid lute advantages in the low price market. Some Chinese firms excessive price competition at the low end and upgrade their even began to enter the high-end segment of the smartphone value chains? The previous section illustrates two examples market. showing that the introduction of new technologies can disrupt These trends were triggered, in part, by changes in con- the existing form of supply chains and stimulate market entry sumer preferences regarding technology features. MediaTek of emerging firms in GVCs, but can also lead to an excessive maintained its advantage during the 3G era, yet it was not able supply of undifferentiated products. This significantly reduces to keep its dominant position when 4G was introduced.15 Qual- the profitability of the industry, leading to a high level of market comm, as the world’s largest owner of 3G and 4G technology fragmentation, falling prices, and little scope for innovation and patents, increased its share of China’s smartphone-baseband upgrading. Given this background, what are the options for IC market. Its shipment share in China’s 4G market accounted firms in developing countries? Can excess capacity and falling for more than 50% in 2015. Four Chinese companies in the top prices be avoided? What are good approaches to upgrading the ten list in Table 4.1. primarily adopted Qualcomm’s platforms: position of emerging market firms in high technology areas? Xiaomi (70% of all models, as of 2015), OPPO (70%), VIVO (60%), Some local manufacturers have upgraded their own value and ZTE (50%). The high demand for Qualcomm’s 4G technol- chains through a commitment to active learning, enabled by ogy was primarily driven by the dramatic increase in consumer open platforms and a shift in consumer demand. Ding and Hioki demand for products of greater quality, functionality, and better (2017) illustrate how technological transfer and value-chain data transmission. The increase in internet users interested in upgrading happened in the Chinese mobile phone industry accessing communication platforms (WeChat, Taobao, and Didi) over the course of the last 15 years. As described earlier, in the along with the upgrading of preferences that is consistent with early 2000s, the mobile-phone industry in China left little room a wealthier society, led to a surge in the demand for mid-range for upgrading, dominated as it was by the “shanzhai sector”. and high-end products. In particular, consumers demanded In recent years, however, Chinese companies in the industry handsets with 4G technology, for their ability to provide faster have achieved remarkable growth (Table 4.1) and some of them and more stable transmission. Qualcomm’s strategy was to focus have rapidly achieved international brand status in the global on these middle-range and high-end segments of the market by 88 • Technological innovation, supply chain trade, and workers in a globalized world TABLE 4.1 Shipments of major smartphone makers in the global market, million units Vendors 2011 2012 2013 2014 2015 2016 2017 1 Samsung 95 198 299 308 320 310 316 2 Apple 93 136 153 193 232 216 216 3 Huawei 17 31 52 75 108 139 153 4 OPPO N/A 5 18 31 45 95 118 5 VIVO N/A 3 12 30 44 82 95 6 Xiaomi N/A 7 19 65 73 58 92 7 LG 19 26 48 59 60 N/A 56 8 ZTE 17 31 42 45 51 57 46 9 Lenovo 4 23 45 N/A 45 50 39 10 Gionee N/A N/A N/A N/A N/A N/A 24 Note: N/A means the relevant information is not available. Source: Ding and Hioki (2018), compiled from data by IHS iSuppli, a market research firm. serving a few emerging local firms with production capabilities new model, and keep continuous communication with their cus- that can accommodate the technological profiles of Qualcomm’s tomers to learn about end-consumers’ demand and preferences. platform. This is in a sharp contrast with MediaTek’s strategy of Reducing product modularity by opening the platform source providing turnkey solutions to numerous undifferentiated manu- codes to its users, allowed Qualcomm to offer them the possibil- facturers with minimum production capabilities. ity to undertake significant product differentiation on their own. Qualcomm also adopted an open platform approach and It is reported that Qualcomm has opened approximately 80% became highly proactive in developing new products and resolv- of its hardware driver source code, compared to only 20% by ing problems jointly with its customers. This is because deep- MediaTek, as pointed out earlier. Under certain circumstances, ening technological complexities now entails much closer col- the company even allows its customers to adjust the platform’s laboration between platform vendors and mobile phone makers design parameters (such as radio frequency specifications). Qual- as well as the suppliers of other relevant components (ampli- comm offers regular support to its platform users and assists fier and antenna, etc.). Furthermore, the life cycle of a mobile them in conducting co-marketing, often jointly holding product phone became much shorter (from 2 years in the 2/3G era to 6 release conferences or introducing them to overseas carriers. In months in the 4G era) while the expected time span for investing this way, the company constantly exchanges technological and in research and development of IC chipsets became consider- marketing information with its customers. Such interactions are ably longer. Platform vendors must therefore predict the future highly relevant for developing the competitive advantages of market trend two or three years in advance of the release of a local manufacturers.16 TABLE 4.2 Market share of local smartphone brands in China 2014 Q4 2015 Q3 Share of local Share of local Share of local Share of local Share of total brands in each Share of total brands in each top 3 top 3 segment segment High-end 16% Information 4.2% 13.5% Information 9.4% (>500$) unavailable unavailable Mid-range 20.4% 76.5% 44.6% 24.8% 81.9% 58.8% (250-500$) Low-end 63.6% 100% 45.4% 61.7% 100% 48% (<250$) Source: Ding and Hioki (2018), compiled from data by GFK market research. Technological progress, diffusion, and opportunities for developing countries: lessons from China • 89 Qualcomm also provided its customers with various oppor- Google’s search engine would create more revenue than would tunities to cooperate with global suppliers, which further helped fees from Android licenses. to accelerate the upgrading of the Chinese mobile phone indus- The impact of Android on the mobile telecom value chain was try. For example, Xiaomi collaborated with Biel Crystal (for cover profound, as it caused the composition of the handset industry glass), OPPO with Texas Instruments (for power chips), VIVO to shift dramatically. By 2016, all operating systems that pre- with Sony (for a front dual camera) and ArcSoft (for the camera dated the Apple iOS were reduced to single-digit market shares software). (see Figure 4.2, left-hand panel). A parallel shift also occurred in In summary, over time technological innovation and the strat- the market of phone manufacturers. The profits for handset sales egies of major firms have driven dramatic changes in the partic- were almost entirely taken by two firms: Apple, with 75 percent, ipation of domestic firms in the Chinese mobile phone indus- and Samsung, with 25 percent (Reisinger, 2016). Incumbents try. MediaTek from Chinese Taipei enabled local firms to enter (producers of pre-smartphone-era feature phones) collapsed the market by providing a highly standardized platform that from a 60 percent market share to less than 10 percent. In fact, of gave a turnkey solution for those without sufficient knowledge all the incumbent firms, only Samsung was able to make the tran- and experience to manufacture high-tech mobile phone termi- sition to Android (and to the smartphone market) successfully nals. However, the lower technological barrier to entry caused (see Figure 4.2, right-hand panel). At the same time, a plethora of an excessive supply of undifferentiated products in the low-end new firms, mostly Chinese, emerged. market, leading to intense price competition. Referring back to The same pattern seen in previous waves of technological the GVC typology, MediaTek’s platform transformed the indus- progress was observed for smartphones. With the availability of try’s value chains from the modular type to the market type. highly-integrated chip sets linked to an open-source operating Subsequently, Qualcomm’s higher level of commitment and col- system, Google’s Android lowered the barriers to entry for new laboration with customers (through opening up most of the plat- firms with lower capabilities, and also reduced product distinc- form’s software source code, technical assistance, and joint prod- tiveness and the value-added from manufacturing handsets. The uct development or product promotion) enabled local firms that two leading brands, Samsung and Apple, which covered 35 per- had accumulated the minimum expertise to accommodate Qual- cent of the market in 2016, had relatively stable market shares. comm’s technological profiles to upgrade their final products. The remainder of the market was very fragmented and unsta- Thus, Qualcomm’s platform further changed power relations of ble. In particular, firms outside of the top five, which account for this value chain into the relational type, in line with the increas- nearly half of the world market, were subject to high volatility and ing complexity of final product characteristics.17 These develop- short spells (Table 4.3). This is typical of the so-called “modular- ments required continuous efforts by local emerging companies ity trap” (Chesbrough and Kusunoki, 2001). Only handset makers to learn through active interactions with more advanced firms. with significant software development capabilities were able to differentiate themselves and achieve stable market share. This 4.2 Implications of market shifts from feature phones to evolution was unlike the pre-smartphone era, described in Stur- smartphones geon and Linden, 2011. At the time, the top five firms, includ - So far, we have focused on the impact of disruptive technology ing industry pioneers Nokia and Motorola, dominated for many embedded in key hardware components (such as IC chipsets). years with relatively stable market shares. However, disruption of value chains also can be driven by soft- Global manufacturing of mobile handsets moved mostly to ware evolution. Sturgeon and Thun (2019) show how the smart- China, driven by both supply and demand factors. In 2016, China phone market provides an opportunity to assess how companies accounted for more than three-quarters of global production can upgrade in manufacturing GVCs following disruptive techno- (HIS Markit Data). China remains the main assembly location logical change. The introduction of smartphones in 2007 opened for all top firms, with the exception of the two brands from the up opportunities for upgrading by Chinese firms. With its iPhone Republic of Korea, namely Samsung and LG.18 A key attractive- handset, launched in 2007, Apple established a platform with a ness of China as assembly location is the fact that it accounts for partly open architecture, the Apple iOS. Third-party developers about one-third of total global demand, representing the largest can access the platform, and design tools and sell applications mobile phone market worldwide. Moreover, a number of Chi- (apps) on Apple’s online store, but governance of the resulting nese brands, including Huawei, Xiaomi and Oppo, have emerged ecosystem is closed (Parker et al., 2016). Partly in response to the as increasingly popular, first among Chinese consumers, and iPhone, Google launched the Android OS for mobile handsets increasingly in foreign markets (HIS Markit Data; Xing, 2018). one year later. In contrast to iOS, Android has an open technol- Chinese smartphone producers are upgrading through build- ogy architecture and largely open governance. The Android OS ing their own brands and being strategic on what components to was licensed for free and its “source code” published through build. This is different from the traditional view that firms upgrade the Android Open Source Project for all to use or modify as along a predetermined sequence of manufacturing tasks. They needed. As the leading internet search company, with revenues are no longer participating only as suppliers of global brands coming mainly from online ad placement fees, Google wanted or producers of low-cost undifferentiated devices. Rather, they more people to access the internet (and thus the Google search succeeded in unseating the market leaders, Samsung and Apple, engine). The expectation was that continued growth in the use of from the Chinese domestic market by focusing on customer 90 • Technological innovation, supply chain trade, and workers in a globalized world FIGURE 4.2 Shift from feature phones to smartphones  100% 100% 90% 90% 80% 80% 70% 70% 60% 60% 50% 50% 40% 40% 30% 30% iPhone introduced 20% 20% 10% 10% Android introduced 0% 0% 2008 2009 2010 2011 2012 2013 2014 2015 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 Android iOS Windows Feature phone Others** Samsung (Google) Mobile incumbents*** (Rep. of Korea) Blackberry Symbian Others LG Apple Other major (RIM) (Nokia**) (Rep. of Korea) (USA) Chinese brands* Lenovo Huawei Xiaomi (China) (China) (China) Notes: Based on unit sales to end users. *Others include Linex (open source) Notes: Based on unit sales to end users. Data are indicative only since not WebOs (Hewlett Packard and others), and Bada (Samsung). ** Symbian, orig- all firms were identified in all years. From 2003-2008 only the top five firms inally developed by a UK-based software company and compatible only with were identified. In subsequent years the top ten were identified. *Unspecified (UK-based) ARM processors, pushed the hardest by Nokia but also used in Chinese brands identified in the data in various years include ZTE, TCL, and keyboard-based “smartphones” made by Motorola and Sony Ericsson. Yulong. **Brands identified in the data in various years include HTC (Chinese Source : Gartner Newsroom Press Releases (Various Years). Taipei), BenQ (Chinese Taipei), and RIM (Canada) are included the remainder “Others” category. ***Feature phone incumbents include Nokia (Finland), Motorola (USA), Sony-Ericsson (Japan-Sweden), and Siemens (Germany). Source : Gartner Newsroom Press Releases (Various Years). Source: Sturgeon and Thun (2019), and HIS Markit Data. orientation, and by growing their design and marketing capabili- truly global R&D footprint, where countries globally attract tasks ties (Brandt and Thun, 2010, 2011, 2016; Thun, 2018; Xing, 2018). in which there is local expertise. Moreover, all major handset By building their brands, these firms moved from their original producers mostly source their inputs from the same technology focus on cost-conscious customers, and increasingly toward mid- suppliers. Key technology suppliers include mostly firms from range consumers demanding value for money (Brandt and Thun, developed countries such as Google, Samsung, Qualcomm, 2010 and 2016, refer to this progression as the “fight for the Broadcom, and leading semiconductor companies ARM and middle”). In so doing, they managed to upgrade their position in NXP. the mobile phone value chain, serving the Chinese market first, As shown above, the smartphone market makes the case and then becoming increasingly successful in other markets (see that, following disruptive technological change, one key reason Figure 4.3). As a result, by 2017, Chinese brands had captured 87 for Chinese firms’ upgrading was the strong connectivity to percent of the domestic market. global technology ecosystems. Growing own design and mar- Successful Chinese firms also rely on knowledge-intensive keting capabilities allowed Chinese firms to respond rapidly to intermediates and globally available technology. None of the changes in market demand and consumer taste. Their reliance top Chinese brands (Huawei, Oppo, Vivo and Xiaomi) has core on GVCs allowed them to develop products that are interoper- technological capacity in-house. These firms rely on GVCs for able and compatible with global markets. Incidentally, the local technology and develop products that depend on interopera- presence of foreign firms enhanced the mutually beneficial rela - bility and compatibility with global markets. Successful Chinese tionship between foreign core technology providers and local brands have not indigenized production in China. They have a manufacturers. Domestically-owned firms had better and faster Technological progress, diffusion, and opportunities for developing countries: lessons from China • 91 TABLE 4.3 Top five mobile handset brand market share in five-year intervals (millions of units) Feature phone era (through 2007) 2003 2007 Company Home country Sales Market share Company Home country Sales Market share Nokia Finland 180,672 35 Nokia Finland 435,453 38 Motorola United States 75,177 15 Motorola United States 164,307 14 Samsung Republic of Korea 54,475 11 Samsung Republic of Korea 154,541 13 Siemens Germany 43,754 8 Sony Ericsson Japan/Germany 101,358 9 LG Republic of Korea 26,214 5 LG Republic of Korea 78,576 7 Others 139,696 27 Others 218,604 19 TOTAL 519,989 100 TOTAL 1,152,840 100 Smart phone era (after 2007) 2011 2016 Company Home country Sales Market share Company Home country Sales Market share Nokia Finland 422,478 24 Samsung Republic of Korea 306,447 21 Samsung Republic of Korea 313,904 18 Apple United States 216,064 14 Apple United States 89,263 5 Huawei China 132,825 9 LG Republic of Korea 86,371 5 Oppo China 85,300 6 ZTE China 56,882 3 Vivo China 72,409 5 Others 805,666 45 Others 682,314 46 TOTAL 1,774,564 100 TOTAL 1,495,358 100 Source: Sturgeon and Thun (unpublished), using HIS Markit Data. access to technology inputs that boosted the competitiveness The importance of connectivity to key players is also demon- of their products, and owners of core technology benefited strated through network analysis. Criscuolo et al. (2017) apply from expanding their sales in a large and growing Chinese network theory to an examination of foreign peer effects on market. firm-level total factor productivity (TFP). Based on Chinese FIGURE 4.3 Emergence of Chinese smartphone brands, in the domestic and foreign markets, percent Top manufacturers’ market shares in the Chinese market, 2017 Chinese brands’ market share in foreign markets, 2017 25 100 20 80 15 60 10 40 5 20 0 0 Huawei OPPO Vivo Xiaomi Apple Samsung China India Bangladash Russia US World Source: Xing (2018). 92 • Technological innovation, supply chain trade, and workers in a globalized world FIGURE 4.4 TFP elasticities of Chinese firms with respect to centrality index and average productivity of their buyers/sellers (a) Comparison among firms with different levels of initial productivity (b) Comparison among firms with different sizes (Percentage)(Percentage) 2.5 0.8 2.0 0.6 1.5 0.4 1.0 0.2 0.5 0.0 0.0 -0.2 -0.5 -1.0 -0.4 -1.5 -0.6 -2.0 -0.8 1st quantile 2nd quantile 3rd quantile 4th quantile 1st quantile 2nd quantile 3rd quantile 4th quantile Less productive More productive Smaller Larger Network centrality Productivity of buyers/sellers Network centrality Productivity of buyers/sellers Source: Criscuolo et al. (2018). BOX 4.1 “Value chain migration” — Can it be another scenario for surviving price wars? In the face of increasing competition and attrition in home attributed to their local partners; namely, the knowledge of markets, some small-scale Chinese firms established new the local demand profiles and the capacity to handle indi- value chains with other developing countries by tapping vidual dealers scattered around the country. The latter into the uncultivated low-end markets at destination. Fujita property was especially important because low-priced (2017) presents the case of motorcycle industry in Viet Nam. motorcycles mainly catered to consumers in rural provinces. Prior to the entry of Chinese firms, the motorcycle indus- Business statistics reveal that these teams of Chinese parts try in Viet Nam had been dominated by a handful of Japa- suppliers and Vietnamese assemblers collectively outper- nese- and Chinese Taipei-invested manufacturers producing formed Lifan, a big Chinese-invested motorcycle manufac- sophisticated yet expensive models that were far beyond turer which entered the Vietnamese market with its own the reach of the majority of the population. In this context, brand name. Chinese firms, faced with saturated consumer demand in While this “value chain migration” strategy provided the home market, saw Viet Nam, a low-income country with a quick route for small-scale Chinese firms to escape from only expensive models available, to be a promising outlet intense competition at home, there is a problem of sus- for their low-priced Chinese products. tainability in the targeted low-end market at destination. The penetration of Chinese firms into the motorcycle Indeed, with rapidly rising incomes in Viet Nam, the market industry in Viet Nam started with the massive export of fin - for low-priced motorcycles in the country nearly disap- ished vehicles. However, in 2002, the Vietnamese govern- peared by the early 2010s, only a decade after its emer- ment enforced a measure against the imports of assembled gence. The teams of Chinese suppliers and local assemblers vehicles and implemented high local content rules. As a failed to keep up product development in order to meet result, firms’ market entry mode in Viet Nam shifted from changes in consumer demand, primarily due to the lack of vehicle exports to component exports, and then to FDI, the technology required to upgrade their products. In the giving rise to a new form of China’s GVCs serving the low- end, the entire market is dominated by five foreign-invested end market in Viet Nam. manufacturers from Japan, Chinese Taipei and Italy, collec- Particularly notable was the performance of Chinese-in- tively accounting for a 98% share, including Honda’s 63% vested parts suppliers who teamed up with Vietnamese (Nguyen Thi Thu Ha and Ho 2013). assemblers. They capitalized on the competitive advantages Source: Fujita (2017) Technological progress, diffusion, and opportunities for developing countries: lessons from China • 93 and Japanese firm-level microdata as well as multi-country China too is rapidly automating production through robot- input-output tables, the study investigates the relationship ization to address declining wage competitiveness. Standard between firm performance and the position of firm operation Chartered Global Research (2016) found that 48 percent of within the GVCs. The GVC position of firm operation is deter- 290 manufacturers surveyed in the Pearl River Delta would mined by two factors: network centrality, which represents consider automation or streamlining processes as a response particular firm’s interconnectedness with other players in the to labor shortages; less than a third would consider moving network, and (weighted) average productivity of its buyers/sell- capacity either inland or out of China. Some high-profile firms ers, which indicates the relative importance of the firm’s peers. are already substituting a substantial number of workers with TFP elasticities with respect to these two factors define over- industrial robots. For example, Foxconn, producing Apple all peer effects on the firm in question. Their estimation results and Samsung products in China’s Jiangsu province, recently for China, presented in Figure 4.4, reveal that firms that are ini - replaced 60,000 factory workers with industrial robots (South tially (i.e. at the beginning point of observation) less productive China Morning Post 2016). If China moves into more sophis - or smaller are likely to improve their productivity faster than ticated exports while automating and retaining market share others when they are connected to the key players in the pro- of the less sophisticated exports, then the expected en masse duction networks.19 This implies that, for small emerging com- migration of manufacturing jobs may not occur. panies in developing countries, “to whom to be connected” in More systematic evidence on robots and reduced offshor- the international production networks is highly relevant, at least ing, as manifested in FDI flows from high-income countries in the long run, when we consider the impact of technological to low- and middle-income countries, has emerged recently. progress on economic development (see Box 4.1). 20 Based on firm-level data for 3,313 manufacturing companies The question that remains unanswered is whether firms from across seven European countries, Kinkel, Jager and Zanker other countries can replicate the positive experience of these (2015) find that firms using industrial robots in their manufac - Chinese firms. Are firms from smaller countries precluded this turing processes are less likely to offshore production activities opportunity? And does automation of production even pre- outside Europe. Hallward-Driemeier and Nayyar (2018) find a vent initial entry based on low wages? The next section will non-linear relationship between the intensity of robot use22 in discuss the impact of automation on offshore potential of low high-income countries (HICs) and FDI from HICs to low/medi- cost-locations. um-income countries (LMICs) between 2003 and 2015. For some time, the increasing intensity of industrial robots moved together with flows of FDI from HICs to LMICs. This is consis - 5. Is automation reducing the offshoring tent with the literature which argues that many of the tasks potential of low-cost locations? that are suitable for automation are also suitable for offshoring (Autor, Dorn and Hanson 2015). For instance, routine tasks that Historically, new technologies and changing trade patterns follow explicit codifiable procedures are well suited to auto - have tended to widen the circle of countries benefiting from mation because they can be computerized, and well suited to expanding production. As countries’ costs rise, production offshoring because they can be performed at a distance with- tends to move into more capital-intensive goods, with the more out substantial loss of quality (Autor, Levy and Murnane, 2003). labor-intensive tasks moving to lower-cost locations offshore. The non-linearity – whereby beyond a threshold level of robot This “flying geese” model21 of industrialization and trade has intensity there is a negative association between robot use in been observed for several decades, as the more labor-inten- HICs and FDI flows from HICs to LMICs – reflects the fact that sive tasks have shifted from developed economies to the newly the scale of use may be a significant factor in making robots industrialized economies of East Asia and China. The question economically attractive. now is whether automation in established manufacturing cen- The relationship between robots and offshoring, however, ters may reverse this process by reducing offshoring. varies across sectors. Hallward-Driemeier and Nayyar (2018) There is an increasing amount of anecdotal evidence on show that the use of robots in high-income countries has how increased automation has already enabled some leading increased steadily over the past two decades, with the steep- firms to reshore labor-intensive manufacturing activities back to est increases in motor vehicles and other transport equipment, high-income economies. Foxconn, the world’s largest contract and electrical machinery and electronics 23 (see Figure 4.5). electronics manufacturer best known for manufacturing Apple’s As automation increases, penetration rates are starting to iPhone, has recently announced it will spend $40 million at a increase even in other manufacturing and services industries, new factory in Pennsylvania, using advanced robots and creat- such as logistics and food production. However, the textiles ing 500 jobs (Lewis 2014). Adidas, the German sporting goods and apparel sector still remains amongst the least automated, company, has established “Speedfactories” in Ansbach, Ger- especially apparel. A lower rate of robot intensity in this many, and Atlanta, which will use computerized knitting, robotic sector is associated with rates of new FDI from high income cutting, and 3-D printing almost exclusively to produce athletic to low- and middle-income countries that are greater than footwear (Assembly 2012; Bloomberg 2012; Economist 2017a, those of highly automated industries such as automotive and 2017b; Financial Times 2016). electronics (see Figure 4.6). Data on FDI (not reported here) 94 • Technological innovation, supply chain trade, and workers in a globalized world FIGURE 4.5 Operational stock of robots in high-income countries, 1993-2015  1,200,000 All other manufacturing branches Motor vehicles and 1,000,000 other transport equipment Industrial machinery 800,000 Electrical/electronics Metals and metal products 600,000 Glass, ceramics, stone, mineral products Rubber and Plastics 400,000 Other chemical products n.e.c. Pharmaceuticals, cosmetics 200,000 Paper Wood and Furniture 0 Textiles 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 Food and Beverages Source: Hallward-Driemeier and Nayyar (2018), using International Federation of Robotics Database. also suggests that some FDI may have migrated from China Here, the development of mutually beneficial relationships to LMICs in Asia and Africa, and from higher- to lower-income between foreign core technology providers and local manufac- countries in the Europe and Central Asia region (Hallward-Drie- turers is the key. Local firms have better (and faster) access to meier and Nayyar 2017). technology inputs that boost the competitiveness of their prod- ucts, and the owners of core technology benefit from expand - ing their sales in large and growing markets. The ability of gov- 6. Policy implications ernments in developing countries to nurture such relationships depends on their ability to reform the domestic investment envi- This chapter draws several lessons on how to achieve upgrad- ronment in a manner to stimulate and rationalize technological ing to move closer to the global technological frontier, largely transfer/sharing by advanced firms within a sequence of local based on the experience of China’s automotive and electrical supply chains. equipment industries.24 The successful firms depend on access One important aspect of the reform is building capabilities to constantly evolving global technology and knowledge-inten- of local manufacturers. Manufacturing can no longer thrive with sive intermediates. A number of Chinese smartphone manufac- unskilled workers alone, and many tradable services are skill turers, for example, have succeeded in entering and upgrading intensive. Recourse to industrial polices to stimulate GVCs, how- in GVCs by leveraging global technology ecosystems and by ever, can have unintended consequences. Some incentives may responding rapidly to changes in market demands and con- take the form of implicit or explicit subsidies, and lead to trade sumer tastes. tensions. Weaker bargaining power of governments, compared Technological progress triggered these changes. Modu- to large lead firms in GVCs, also means that there is the risk that larization of product architecture offered a new entry point to incentives result in sizeable transfers of rents to the firms, reduc- GVCs for small-scale firms in developing countries. The import- ing the social dividend of being in GVCs. ant message of our study, however, is that entry into GVCs alone Another important dimension of domestic reform is the does not translate automatically into technological upgrad- development of legal/institutional bases. Creating an attractive ing. To move up to high value-added tasks in technologically investment environment is a multi-faceted task. Policy-planners advanced value chains requires additional and complementary have to consider various domestic factors that might affect firms’ efforts by local actors. investment decisions: physical infrastructure, trade policies, Technological progress, diffusion, and opportunities for developing countries: lessons from China • 95 FIGURE 4.6 Robot stock in electronics and automotive relative to apparel in high-income countries (ratio) vs FDI flows from high-income to middle- and lower-income countries in electronics and automotive relative to apparel (ratio), 2003-15 (a) Number of FDI Projects Electronics and automotive products Textiles, apparel and leather products 30 000 3 500 25 000 3 000 2 500 20 000 2 000 15 000 1 500 10 000 1 000 5 000 500 0 0 2003 2005 2007 2009 2011 2013 2015 03 04 05 06 07 08 09 10 11 12 13 14 15 20 20 20 20 20 20 20 20 20 20 20 20 20 (b) Stock of Industrial Robots Electronics and automotive products Textiles, apparel and leather products 700 000 1 400 600 000 1 200 500 000 1 000 400 000 800 300 000 600 200 000 400 100 000 200 0 0 03 04 05 06 07 08 09 10 11 12 13 14 15 03 04 05 06 07 08 09 10 11 12 13 14 15 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 Source: Hallward-Driemeier and Nayyar (2018), using International Federation of Robotics Database. competition policies, wage levels, workers’ educational attain- products. And most importantly, the study also shows that “con- ment, and so on. Among them, increasing attention is being tract-intensive” products are often skill-intensive as well, and paid to the role of the legal system in facilitating capital inflows, hence likely to be of high value-added. especially into developing economies. Even though the issue is Closely related to this issue is the evidence that patent not touched in the preceding argument, it is worth shedding a laws in offshore destinations influence global firms’ innovation light on this important aspect of globalization. decisions. Bilir and Sakamoto (2018), using detailed data on Nunn (2007), for example, introduces the concept “con - US patent grants/citations and US multinational firms’ affiliate tract-intensive” products, which rely on production processes R&D investment, show that the presence of imitation risk from with complicated interactions between clients and suppliers at potential rivals at offshore destinations can drive leading mul- various stages of a production sequence. Such a product attri- tinational firms to innovate selectively. They do so by shifting bute is especially salient in industries with a high degree of development resources toward relatively short-lived products market differentiation, for example airplanes or special indus - that are difficult to imitate before they become obsolete. Here, trial machinery. Accordingly, the countries with well-established by reducing imitation risk, patent reforms at offshore destina- legal systems and high-quality institutions are considered to tion facilitates innovation by multinational firms, but at the same have comparative advantages in producing this type of prod- time also increase the average economic lifespan of the prod- uct, just in the same way that countries with an abundant cheap ucts they seek to develop. This implies that a policy reform of labor force are more competitive in producing labor-intensive intellectual property rights in less developed countries affects 96 • Technological innovation, supply chain trade, and workers in a globalized world not only the level of innovation but also the type of innovation Finally, while automation does not pose immediate risks to associated with offshoring activities, encouraging the develop- shut the door to labor intensive exports from developing coun - ment of technology with more sustainable economic values. tries, governments need to develop a comprehensive digital In conclusion, putting a foot in the door (of new industries) is strategy. Our economies are increasingly sitting on a digital not enough to thrive in GVCs. The examples of China and earlier foundation, one that is generating high-speed growth and dis- developers show that developing through GVCs is a decades ruptive change. The employment and investment of tomorrow long journey that requires a reform effort sustained over time. will be data intensive. Value in a knowledge economy is created While international connectivity is a key for entering GVCs, by innovative ideas and data. As economies and firms from dif- domestic governance matters for upgrading therein. Local gov- ferent countries grow similar in size, international trade will inten- ernments need to offer well developed domestic legal systems, sify. But trade may tilt away from physical goods and towards and guarantee the rule of law and high quality institutions. Good data. Importantly, the digitally-powered, knowledge-intensive governance is crucial for attracting high value-added segments GVCs that are emerging and are likely to dominate the future of global supply chains, where technological transfer/sharing have a strong potential for inclusion. Moreover, they can contrib - between global firms and local suppliers is considered more ute to expand markets for small businesses beyond traditional solid and sustainable. geographies. They can also expand financial inclusion, as data Industrial policy can have unintended consequences, and on e-commerce can be used as collateral, and smartphones link therefore should be carefully crafted. Policies for helping up the bottom half of world incomes to these opportunities. domestically-owned firms to become technologically standalone – what some might refer to as “techno-nationalism” – do not necessarily deliver the expected results. The world’s most pow- erful technology companies, both from emerging and advanced countries, work with global suppliers and even with competitors in “open innovation” environments. Hence, the advice to policy- makers seeking to upgrade toward the global technology fron- tier is to prioritize measures that encourage firms to be full part- ners in global technology ecosystems, rather than champions of domestic technology, or of so-called techno-nationalism. Technological progress, diffusion, and opportunities for developing countries: lessons from China • 97 ANNEX 4.1 Typology of global value chains25 Gereffi, Humphrey, and Sturgeon (2005) set out a typology of Relational-type GVC five global value chains (GVCs) on the basis of the structure of When the manufacturing process involves specialized equip- power relations between the contracting parties. ment (for example, the mold for a product of a particular shape), transactions become asset-specific, and the contracting par- Market-type GVC ties become mutually dependent. The equipment for a specific Producing a commodity of a generic nature does not require any purpose has limited scope for alternative uses, so its productiv- specific investment in production facilities for a particular trans- ity will drop considerably when it is applied in other contexts. action, so both customers and suppliers have countless choices Accordingly, the service suppliers (the holders of the specialized for alternative partners. They are connected mainly through open equipment) are not motivated to look for other potential clients. spot-market transactions in a shoulder-to-shoulder relationship. But it is also difficult, or at least costly, for the client to expect Also, the procurement of a generic commodity will not neces- the same level of performance from other third suppliers with- sitate an exchange of detailed product specification between out these specialized facilities. As a result, both parties have little contractors because the key information is mostly reduced to incentive to search for alternative business relations. Further, the preset price of the product that can be found in a book of reinvestment in the specialized equipment for raising productiv- catalogs. The transaction cost for changing business partners is ity deepens the asset-specificity of the transaction, thus trapping almost negligible, leaving the value chains in a constant state of the parties in even more mutually dependent relationships. flux because of their high price elasticity. Captive-type GVC Modular-type GVC This type of transaction assumes an overwhelming disparity in In business management or industrial engineering the word power exercise among the parties, as seen in the business rela- “module” generally refers to a composite of subcomponents tions between a lead firm of global brands and its subcontracting grouped by the types of functions that are assumed in making up local small companies. Service suppliers are expected to follow the final product. The possibility of different combinations of dif- the client’s instructions word for word and are subject to strict ferentiated modules enables producers to design multiple vari- surveillance on product quality and delivery times. Unlike sup- ants of a product. By the same token, if a complex transaction pliers in the market-type GVC, the captive service suppliers have can be accommodated in the supply base by adjusting the com- neither sufficient productive capacity to enjoy the scale of mass bination of multipurpose equipment, the supplier will not have to production, nor the specialized production facilities needed to incur transaction-specific investment (no hold-up problem) and claim its uniqueness, as attributed to the suppliers in the rela- is thus able to spread the equipment’s use across a wide range tional-type GVC. The availability of only mediocre production of potential clients. Even though the information to be delivered capability greatly narrows their opportunities to look for alterna- between the contractors may be considerable (say, for producing tive business relations, imposing a captive position toward their a complex product), the relative easiness to codify transactions, clients. as presumed in this type of GVC power relations, compresses the volume of interventions, and the supplier is able to take overall Hierarchy-type GVC control of the production process. This implies that the transac- This type of GVC generally refers to the relations within a verti- tion cost for changing business partners remains relatively low. cally integrated firm, as with multinational corporations. 98 • Technological innovation, supply chain trade, and workers in a globalized world Notes and even strengthen, their roles in GVCs, thanks to the fact that first- tier suppliers have co-located with lead firms. 1. Sturgeon and Thun (2019) note that there are three reasons for the rel- 8. From the beginning, the industry has had close links with the devel- atively short supply lines in the automotive industry. First, motor vehi- opment of military technology; hence, the standardization of its major cles comprise several heavy, bulky and sometimes easily damaged product lines was advanced under strong military influence. Product components (engines, large metal parts, seats and painted items) standardization was further facilitated by the introduction of com- that increase shipping costs. Second, the adoption of low-inventory, puter-aided design systems, which allowed information on product just-in-time assembly techniques and high product variety (vehicles designs and specifications to be digitized and stored for repeated use can have dozens or hundreds of options) increase the motivation to in the industry. In addition, the Information Technology Agreement, locate module and sub-subsystem assembly close to or even adjacent a high-level plurilateral free-trade agreement, was adopted by many to final assembly. Third, many countries, including the United States, countries including emerging economies, and thus became another China, Brazil, India, South Africa, and many others, have long-stand- important driver of standardization and modularization of the indus- ing policies, both explicit and implied, that have encouraged FDI and try’s value chains. high local content levels — and more recently, R&D and engineering 9. Platforms can exist at all levels of a value chain and in all industries, investments—in return for market access. Because of their relatively and are ideal to help latecomers to join capital- and skill-intensive recent importance in the industry, this has meant a wave of FDI by value chains. Platforms provide a wide range of functionalities and suppliers to provide local content. flexibility. As such, platforms have played a key role in disrupting var- 2. “It is two German automakers, Volkswagen and Mercedes-Benz (pres- ious industries, from consumer electronics such as LCD TVs to special industrial machinery such as numerically-controlled machine tools. ently DaimlerChrysler), that geared up the auto industry’s modulariza- In the PC industry, the most prominent example is “Win-tel”, which tion in the mid-1990s. Their new assembly plants, which started pro- is a coinage from Microsoft’s operating system Windows and chip duction in 1996 and 1997, introduced modularization on a large scale, designer/vender Intel. specifically at Volkswagen’s plants in Resende (Brazil), Boleslav (Czech 10. According to Shiu and Imai (ibid.), sales promotion through blueprint Republic), and Mosel (former East Germany), and Mercedes-Benz’s bundling is known to have originated in the business model of US/ plants in Vance (U.S.) and Hambach (France)”. Source: Takeishi and European chip vendors who sought marketing opportunities in China. Fujimoto, 2001. However, the production guidelines in the blueprints of the US and 3. Shenyang Aerospace Mitsubishi Motors designed engines and trans- European vendors covered only basic aspects of terminal assembly. missions, and then Delphi took charge of ECU adjustment to custom- Lacking detailed explanations, these blueprints were not sufficiently ize these system components according to the individual designs of user-friendly for Chinese manufacturers with limited experience in customers’ vehicles (Oshika et al., 2009). The company codified the the production of high-tech equipment such as mobile phones. In harmonization expertise and encapsulated it in a chip as a set of dig- contrast, MediaTek from Chinese Taipei provided full guidelines for ital information, whereby potential conflicts among parts arising from every aspect of assembly tasks, even covering multimedia functions variations in car bodies can be mediated through a mere parameter for music/video playback, and offered a package with a considerably adjustment of ECUs. cheaper license fee than those of US/European rivals. As a result, 4. This is consistent with the 3Cs model of Gereffi et al. (ibid.). MediaTek contributed to Chinese manufacturers’ ability to produce 5. The modularization of car architecture has also invited new entrants at low cost while still providing highly appealing products for local from other industries. Panasonic’s subsidiary Automotive & Industrial consumers. Systems develops system component packages in three areas: cock- 11. In a general equilibrium perspective, the price competition benefits pit systems (displays, gauges, and car navigation devices); drive-as- downstream users, especially final consumers. Here, we consider sist systems (sensors, cameras, and LEDs); and power management costs and benefits only from the viewpoint of mobile phone produc- systems (compressors and charge controls). Panasonic’s technological ers in relation to their development potentials. know-how from manufacturing electrical equipment is fully applied to 12. The case of MediaTek illustrates how a platform leader can use its and embodied in the car production schemes (Nikkei Business, 2013). leadership position to impose a closed system of governance on the 6. Suppliers are located in places as diverse as the United States, resulting ecosystem (as opposed to allowing it to be open source). Mexico, Brazil, Viet Nam, Malaysia, India, China, and various locations When this happens, the platform leader can impose structural con- in Europe. On the ICT services side, countries such as India, Philip- straints on the design and specification of other auxiliary components, pines, and Ukraine provide routine software coding and the provision with the effect that suppliers and other firms in the platform ecosys- of remote ICT-enabled services. tem may be forced to produce products whose designs are highly 7. There are two main reasons for the delinking of production from subordinate to the platform’s interface specification. The platform design and innovation activities. First, the deep technical, manage- leader can also completely “black-box” the interior of the platform ment, and financial expertise needed to develop and launch new module itself, which gives it potentially an overwhelming power to products and alter the technological trajectory and evolution of influence the way supply chains are organized in the industry. knowledge-intensive industries takes a long time to develop and 13. To capture consumers’ attention, local manufacturers rushed to intro- therefore tends to be place-specific. Second, because of fragmenta- duce multiple models with very similar functionalities. As a result, the tion in GVCs, traditional design clusters have been able to maintain, market was flooded with undifferentiated products and the industry’s Technological progress, diffusion, and opportunities for developing countries: lessons from China • 99 profitability declined significantly. The emergence of MediaTek pro- 19. Note that this finding was also captured by the earlier work of vided local manufacturers with the opportunity to produce high-tech Santoni and Taglioni (2016). Also, the demonstrated empiri- mobile-phone terminals, but also induced the side-effect of rapid cal result is consistent with Criscuolo and Timmis’s (2017) based commoditization of the industry. Commoditization of mobile phones on large-scale multi-country firm-level data collected from the into undifferentiated products significantly lowered the complexity of ORBIS. Although the ORBIS data include Japanese and Chinese transactions between parties. firms, the coverage of these Asian firms is not large in the analysis 14. As discussed in the text, the value chain of the shanzai sector is typi- by Criscuolo and Timmis (2017) mainly because the value added cally arm’s length, prone to feature numerous undifferentiated prod- information is not available for many Asian firms. Therefore, the ucts, i.e. characterized by a highly disintegrated market, with dozens firm-level data used by Criscuolo and Timmis (2017) cover more of independent firms specializing in the same narrow and low-value European firms than Asian firms such as Japanese and Chinese firms. added segments of production and competing with each other In Nakano, Nishimura, and Kim (2018), a parallel approach is pro- harshly on prices. jected to address the issue of technological diffusion by employing 15. The “G”s of 3G and 4G stand for a generation of mobile phone tech- a general equilibrium framework, again using input-output accounts. nology, and hence the terminals with 4G generally assume higher per- The technological diffusion is considered to transform the input sub- formance than those with 3G in terms of data transmission speed and stitution structure as prompted by the change in relative prices of reliability. For the previous generations, 1G was analogue technology, products. Such “structural propagation” was quantified by using a which turned into digital technology from 2G. Today, we are now system of various cost functions whose parameters were estimated via talking about 5G, which is considered to have a significant impact on two timely-distant input-output accounts and deflators. the way of our life. Compared to the earlier technologies, 5G realizes 20. Fujita (2017) presents a case of the motorcycle industry in Viet Nam greater speed, lower latency, and simultaneous connection to larger in which Chinese suppliers chose to engage in the relations with local number of devices. Such features brought a wider prospect for the Vietnamese manufacturers without strong technological bases. See high level of applications in the areas of Internet of Things, remote Box 4.1 for the motivation of the strategy and its consequences. services, self-driving systems, virtual/augmented realities, and so on. 21. The well-known theories of Vernon’s “product life-cycle” and Aka- 16. Also, Qualcomm’s unique patent licensing model, based on a rev- matsu’s “flying geese” depict a process in which technologies orig- enue-sharing scheme, provided its own incentive to care about the inating in advanced economies become obsolete and are passed performance of its customers, thus making further motivation to on to less-developed countries, thereby promoting their economic closely collaborate with them. development. 17. Ding and Hioki (2017) do not consider this form of new value chains 22. Measured as the stock of industrial robots per 1000 persons as relational since it does not involve asset-specific transactions. How- employed. ever, one might consider that the human/organizational relationships 23. This sector-level data obfuscates the fact that certain tasks even in developed through collaboration are specific (intangible) assets, as these highly automated sectors will continue to be labor-intensive. frequently observed in Japanese firms’ practices in keiretsu networks. 24. Even though these two industries are highly prominent in GVCs, it 18. These two firms together assemble about 22 percent of their hand- is also interesting and worthwhile to consider how representative sets in the Republic of Korea, and rely on Viet Nam and Indonesia as they are in terms of their development experience vis-à-vis other secondary sources to China. Only HTC, from Chinese Taipei, and a industries, especially those relevant for developing countries such as relatively minor player, produces entirely at home. India and Brazil are apparels or agro-business. This will be the topic of future research. significant assembly locations for many brands, in part to meet strong 25. This annex is reprinted from GVC Development Report 2017, chapter 1. local content requirements in these large markets. 100 • Technological innovation, supply chain trade, and workers in a globalized world References Economist. 2017b. “3D Printers Start to Build Factories of the Future.” June 29. Assembly. 2012. “Automation Profiles: Robots Help Philips Shave Assem- Financial Times. 2016. “Robot Revolution Helps Adidas Bring Shoemaking bly Costs.” Assembly Magazine, June 1. Back to Germany.” June 8. Autor David H., David Dorn & Gordon H. Hanson, 2016. “The China Shock: Fujita, M. 2018. “How suppliers penetrate overseas market: Internationaliza- Learning from Labor-Market Adjustment to Large Changes in Trade,” tion of Chinese firms from the value chain perspective.” IDE Discussion Annual Review of Economics, Annual Reviews, vol. 8(1), pages 205-240, Paper Series No. 727. Institute of Developing Economies, JETRO. October. Gereffi, G., J. Humphrey, and T. Sturgeon. 2005. “The Governance of Global Autor, David H., Frank Levy, and Richard J. Murnane. 2003. “The Skill Value Chains.” Review of International Political Economy 12 (1): 78–104. Content of Recent Technological Change: An Empirical Exploration.” Hallward-Driemeier, M., and Gaurav Nayyar. 2018. Have Robots Grounded Quarterly Journal of Economics 118(4): 1279–1333. the Flying Geese: Evidence from Greenfield FDI in Manufacturing. World Baldwin, C.Y. and Woodard, C.J.: The architecture of platforms: a unified Bank, Washington D.C. view. In: Gawer, A. (ed.) Platforms, Markets and Innovation, pp. 19--44. Hallward-Driemeier, M., and Gaurav Nayyar. 2017. Trouble in the Making? Edward Elgar Publishing (2009) The Future of Manufacturing-Led Development. World Bank, Washing- Bilir, L.K, and Y. Sakamoto. 2018. “Patent Laws and Innovation Selection in ton D.C. the Global Firm.” a paper presented at the conference of The Japan Kinkel, Steffen., A. Jager, and Christoph Zanker. 2015. “The effects of robot Society of International Economics, held 13-14 October, Nishinomiya, use in European manufacturing companies on production off-shoring Japan. outside the EU”. June. 22nd International Annual EurOMA Conference, Brandt, L., J. Van Biesebroeck and Y. Zhang (2008), ‘Creative accounting or At Neuchâtel, Switzerland. creative destruction?” Imai K., Shiu J.M. (2011) Value Chain Creation and Reorganization: The Brandt, L. & E. Thun. 2010. “The Fight for the Middle: Upgrading, Compe- Growth Path of China’s Mobile Phone Handset Industry. In: Kawakami tition, and Industrial Development in China,” World Development, vol. M., Sturgeon T.J. (eds) The Dynamics of Local Learning in Global Value 38, no. 11, November. Chains. IDE-JETRO Series. Palgrave Macmillan, London. https://link. Brandt, L. & E. Thun. 2011. “Going Mobile in China: Shifting Value Chains springer.com/chapter/10.1057/9780230281783_3 and Upgrading in the Mobile Telecom Sector,” International Journal of Inomata, S. 2017. “Impact of new technologies on the organization of Technological Learning, Innovation and Development, Vol. 4, No.1/2/3. global value chains (1): dynamic transformations within global value Brandt, L. & E. Thun, 2016. “Constructing a Ladder for Growth: Policy, Mar- chain typology.” IDE Discussion Paper Series No.682. Institute of Devel- kets, and Industrial Upgrading in China,” World Development, Elsevier, oping Economies, JETRO. vol. 80(C), pages 78-95. Lewis, Colin. 2014. “Robots Are Starting to Make Offshoring Less Attrac- Bloomberg. 2012. “China No Match for Dutch Plants as Philips Shavers Come tive.” Harvard Business Review. May 12. Home.” Bloomberg Technology, January 19. http:// www.bloomberg.com/ Nakano, S., K. Nishimura, and J. Kim. 2018. “Multi-Sectoral Value Chain in news/articles/2012-01-19/china-no-match- for-dutch-plants-as-philips- a Bilateral General Equilibrium.” IDE Discussion Paper Series No. 691. shavers-come-home-1-. Institute of Developing Economies, JETRO. Chesbrough, H. and K. Kusunoki, (2001) ‘The modularity trap: innovation, Nguyen Thi Thu Ha, and Chenger Ted Ho. 2013. “Localization Strategy of technology phase shifts and the resulting limits of virtual organiza- Motorcycle Industry in Vietnam.” Presentation at the International Con- tions’, in I. Nonaka and D. Teece (Eds.) Managing Industrial Knowledge, ference on Social Science and Business, Chinese Taipei, January 2013. London: Sage. Nikkei Business. 2013. Nihonkeizai-shinbun shuppan-sha, 22 July 2013. Criscuolo, C., Y. Hashiguchi, K. Ito, and J. Timmis. 2018. “Changing Struc- Nunn, N. 2007. “Relationship-Specificity, Incomplete Contracts, and the ture of Global Value Chains and Domestic Firms’ Productivity: Evidence Pattern of Trade.” Quarterly Journal of Economics 122 (2): 569–600. from Japanese and Chinese Firm-Level Data.” A background paper for Parker, G.G., M.W. Van Alstyne and S.P. Choudary (2016), ‘Platform Revolu- Global Value Chain Development Report 2017: Measuring and Analyz- tion: How Networked Markets Are Transforming the Economy and How ing the Impact of GVCs on Economic Development. The World Bank to Make Them Work for You’, WW Norton & Company. Group. Oshika, T., R. Inoue, J. Oh, S. Orihashi. 2009. “Automotive industry: com- Criscuolo, C. and J. Timmis (2017a) The Changing Structure of GVCs: Are parison of Asian industries based on architecture analysis.” in Shintaku. Central Hubs Key for Productivity? A Background Paper for the 2017 J., and T. Amano eds., Global Strategy of Manufacturing Management: Conference of the Global Forum on Productivity, June 26-27, Budapest. Industrial Geography in Asia. Yuhikaku, Japan: 163-84. Criscuolo, C., and J. Timmis (2017b) The Relationship between Global Santoni, G. and D. Taglioni (2015), “Networks and structural integration in Value Chains and Productivity, International Productivity Monitor, 32: global value chains” in Amador, J. and F. di Mauro (eds.), The Age of 61-83. Global Value Chains: Maps and Policy Issues, pp. 68-84, Centre for Eco- Ding, K., and S. Hioki. 2018. “The role of a technological platform in facil- nomic Policy Research, voxEU eBook. itating innovation in the global value chain : a case study of China’s Shiu, K., and K. Imai. 2009. “Mobile phone industry: the transformation of mobile phone industry.” IDE Discussion Paper Series No. 692. Institute architecture and competition structure in Chinese markets.” in Shintaku. of Developing Economies, JETRO. J., and T. Amano eds., Global Strategy of Manufacturing Management: Economist. 2017a. “3D Printers Will Change Manufacturing.” June 29. Industrial Geography in Asia. Yuhikaku, Japan: 111-35. Technological progress, diffusion, and opportunities for developing countries: lessons from China • 101 South China Morning Post. 2015. “Rise of the Robots: 60,000 Workers Thun, Eric and Sturgeon, Tim (2019) When Global Technology Meets Local Culled from Just One Factory as China’s Struggling Electronics Hub Standards: Reassessing the China’s Mobile Telecom Policy in the Age of Turns to Artificial Intelligence.” September 21. Platform Innovation. In: Rawski, Thomas and Brandt, Loren, (eds.) Policy, Standard Chartered Global Research. 2016. “Global Supply Chains: New Regulation and Innovation in China’s Electricity and Telecom Industries. Directions.” Special Report, Standard Chartered Global Research, Cambridge University Press. ISBN 978-1108703697 (Accepted) London. Takeishi, Akira and Takahiro Fujimoto. (2001) “Modularization in the Auto Sturgeon T.J., Linden G. (2011) Learning and Earning in Global Value Chains: Industry: Interlinked Multiple Hierarchies of Product, Production, and Lessons in Supplier Competence Building in East Asia. In: Kawakami Supplier Systems, first draft” M., Sturgeon T.J. (eds) The Dynamics of Local Learning in Global Value Xing, Y., and Y. He, 2018. “Domestic Value Added of Chinese Brand Mobile Chains. IDE-JETRO Series. Palgrave Macmillan, London Phones,” GRIPS Discussion Papers 18-09, National Graduate Institute for Policy Studies. CHAPTER 5 Understanding Supply Chain 4.0 and its potential impact on global value chains Michael J. Ferrantino (World Bank Group) and Emine Elcin Koten (World Bank Group)* ABSTRACT The reorganization of supply chains using advanced technol- more responsive to consumer demand, boosting employment ogies, such as the Internet of Things (IoT), big data analytics, (employment in supply chain sectors where such technologies and autonomous robotics, is transforming the model of supply are most likely to be applied has grown much more rapidly chain management from a linear one, in which instructions than in other supply chain sectors and in the economy as a flow from supplier to producer to distributor to consumer, whole) and saving consumers’ time. The impact of these tech- and back, to a more integrated model in which information nologies on the length of supply chains is uncertain: they may flows in an omnidirectional manner to the supply chain. While reduce the length of supply chains by encouraging the reshor- e-commerce is uniquely suited to many of these techniques, ing of manufacturing production to high-income economies, they also hold the promise of improving efficiency in brick- thus reducing opportunities for developing countries to par- and-mortar stores. These technologies are generating enor- ticipate in GVCs, or they may strengthen GVCs by reducing mous benefits through reducing costs, making production coordination and matching costs. • Digital technologies are transforming supply chain management from a linear model in which instructions flow from supplier to producer to distributor to consumer, and back, to a more integrated model in which information flows in multiple directions (sometimes referred to as Supply Chain 4.0). • Digital technologies offer huge benefits in terms of inclusive patterns of growth, innovation and entrepreneurial opportunities • The impact of new digital technologies on GVCs is uncertain: they may reduce the length of supply chains by encouraging the reshoring of manufacturing production, thus reducing opportunities for developing countries to participate in GVCs, or they may strengthen GVCs by reducing coordination and matching costs. * We are grateful for helpful comments by Gary Hufbauer, Satoshi Inomata, Kalina Manova, William Shaw, Emmanuelle Ganne, and Lauren Deason. All errors and omissions remain the responsibility of the authors. 103 104 • Technological innovation, supply chain trade, and workers in a globalized world 1. Introduction “S chapter is already being applied in actual supply chains, or is at least being piloted. While the literature includes many ideas upply Chain 4.0” is the re-organization of supply for emergent technologies that might be available by 2030 chains – design and planning, production, dis- (for example, vast fleets of self-driving delivery vehicles, or the tribution, consumption, and reverse logistics – “smart mirror” in the local clothing store that supposedly will using technologies that are known as “Industry allow you to virtually try on clothes just by scanning their bar 4.0”. These technologies, which emerged in the 21st century, codes), this argument does not depend on the deployment are largely implemented by firms that are at the frontier of of technologies that do not really exist yet. The diffusion of supply chain management in high-income countries. Though, already existing Supply Chain 4.0 technologies will already have as we will argue, this classification is somewhat artificial, it does a substantial impact. in fact capture certain prevailing ideas about what firms need When we say that Supply Chain 4.0 is here, we mean that it to do, and are doing, in order to maintain competitive supply is here at the frontier of supply applications and being more chains. widely adopted, not that it is universal. Even in high-income countries, the principles of Supply Chain 4.02 are unequally 1.1 Supply Chain 4.0 is here already applied. Advanced supply management techniques are more While much of the literature we will review is forward-looking, likely to be observed in sectors such as electronics where earlier and indeed has emerged only in the last two or three years, waves of management techniques took hold first, or in big-box almost all of the technologies we discuss are being imple- retailers such as Walmart. As recently as February 2018, supply mented today, at least by firms at the frontier of supply chain chain problems caused two-thirds of the 900 Kentucky Fried management, which by and large are in high-income countries.1 Chicken restaurants in the United Kingdom to close because With only one or two exceptions, everything described in this they had run out of chicken.3 FIGURE 5.1 US employment by sector, supply chain sectors, manufacturing, transportation, post office and other, percent change (2011-2016)  45 40 35 30 Supply-chain sectors Aggregates 25 20 15 10 5 0 -5 . rs e g es rs n e s s s rs y tc er er ic or tio ag in om ile ile ile ,e or rv g al ur ct or ta ta ta St ta en Se on ts es ct se Re or St re Re ke s ise a l ec ho al es sp n uf d e ar e ai st e nd or an an M w an S or M or ch Po lU st ha nt st M Tr d St g c ly on an ta ty ha in ni c al s pp er al To s ro ou N c er ou rs M ci er t su ie ec ne en e eh M al ur Sp al El la G er ar Co t el To le W en isc a G es M l ho W S Understanding Supply Chain 4.0 and its potential impact on global value chains • 105 1.2 It transforms business models, making supply more FIGURE 5.2 Shopping and e-commerce occupations, customer-driven gender division (2017) While Supply Chain 4.0 involves the deployment of such contem- porary tools as the Internet of Things (IoT), big data analytics, Percentages autonomous robotics, and the like, it is not really about any of these things. It is about transforming the model of supply chain 100 management from a linear model in which instructions flow from supplier to producer to distributor to consumer, and back, to a 80 more integrated model in which information flows in an omnidi- rectional manner to the supply chain. While lead firms are increas- 60 ingly analyzing this information through “supply chain control towers,” the end effect of this development could be making the 40 goods economy more responsive to consumer demand. 20 1.3 E-Commerce is ideally, but not uniquely, suited for Supply Chain 4.0 0 The ability to capture data in e-commerce empowers many of Time spent purchasing Transportation the data-driven methods we will discuss. In particular, older goods and services and material technologies (electronic data interchange) were already gather- moving occupations, ing large amounts of information in business-to-business (B2B) Men Women employment e-commerce, which can be used to improve supply chain perfor- mance. At the same time, most of the developments discussed here can be used to improve the performance of traditional Source: BLS American Time Use Survey, BLS Current Population Survey, and brick-and-mortar stores, where the large majority of retailing still authors’ calculations. takes place, as well as in an e-commerce setting. 1.4 It generates jobs, which substitute for household below, these workers, concentrated primarily in warehouses and labor and promote human well being express delivery companies, are paid to do the picking, packing, In an exercise using U.S. data gathered in the Occupational and driving that would otherwise be done by household shop- Employment Statistics of the Bureau of Labor Statistics, this study pers in the absence of e-commerce. shows that employment in the most dynamic parts of the supply chain has grown at a rate substantially exceeding that of the 1.5 It can transform the operation of global value chains overall economy since 2011. These sectors include warehousing Whether conceived of as an advanced management practice, or and storage (used by all retailers, Walmart as well as Amazon), simply as a cluster of technologies to be deployed by advanced couriers and messengers (the sector including UPS and Federal management practices, Supply Chain 4.0 provides substantial Express, commonly known as “express carriers”), and non-store opportunities for firms to enhance productivity, profitability, retailers (particularly electronic shopping and mail-order houses, product quality, and performance in international trade. Because the sector inhabited by Amazon and eBay) (see Figure 5.1). Most Supply Chain 4.0 diffuses at an unequal rate, it can also influence of the jobs being created involve moving goods around either in the size distribution of firms within industries as well as income warehouses or delivery vehicles and have many of the character- distribution across countries. The enhanced ability to track both istics of factory work. Though robots are used in many of these physical and financial information also has implications for activ- applications, they appear, at present, to be complementary with ities of government which depend on highly disaggregated firm human labor. data, such as tax enforcement and monitoring of rules of origin in Most importantly, e-commerce, powered by Supply Chain 4.0, international trade. involves a great substitution of market labor for household shop- ping time. Traditional shopping is a time-consuming and, for many, tedious activity. Because household time is an intrinsically 2. The impact of Supply Chain 4.0 on firms scarce resource, Supply Chain 4.0 is already having profound impacts on human well-being. However, time saved as a result 2.1 Technologies and management strategies of e-commerce also has increased employment in the transpor- One way to approach Supply Chain 4.0 is to treat it as simply tation and material moving occupations. As shown in Figure 5.2, the application of Industry 4.04 to the supply chain.5 And a men account for 42 percent of the time spent shopping, while common way to approach Industry 4.0 is to treat it as simply a women account for 58 percent, whereas men account 82 per- bundle of technologies that have emerged, or are emerging, in cent of employees in transportation and warehousing jobs, while the 21st century (see Figure 5.3). Then the task might be simply women account for 18 percent. As discussed further in section vi to map the technologies in Industry 4.0 to each of the steps of 106 • Technological innovation, supply chain trade, and workers in a globalized world the supply chain – design and planning, production, distribution, long time to determine what the most successful technologies and consumption. will be in any given area. For example, during 1880-1920 it was While each of the “industrial revolutions” is generally char- not at all obvious how three available forms of energy, steam, acterized by a cluster of typical technologies, the list of these electricity and gasoline, were to be applied to two areas of technologies varies from one author to another. Cirera et al. activity, factories and motor vehicles. Eventually a consensus (2017) identify 17 technologies that are said to characterize emerged that factories ought to be run by electricity and motor Industry 4.0 (see Figure 5.4), which are referenced two or more vehicles by gasoline, but not before every other combination of times in a corpus of underlying sources, of which the most fre- power and activity had been experimented with extensively, and quently mentioned are the IoT; big data analytics; 3D printing; with some success (Freeman and Soete 1997 75-80, 139-140). advanced (autonomous) robotics; sensor-using smart factories6; Fortunately, there is a more fruitful way to approach the prob- augmented reality7; artificial intelligence8; and cloud computing9. lem, because the broad functional outline of how Industry 4.0 Pfohl et al. (2015) identify over 50 technologies associated with affects supply chains is already apparent. Industry 4.0, mind-mapped to such underlying attributes as “dig- Supply Chain 4.0 fundamentally changes the way informa- italization” (which applies to everything), “mobility”, “modular- tion flows through the supply chain. Traditional supply chains ization,” “network collaboration,” “autonomization”, “transpar- link suppliers to customers in a linear manner, with each firm ency,” and “socialization”. sourcing inputs from suppliers and in turn delivering its products It is tempting, as noted above, to attempt to understand to customers (Figure 5.5). The planning process of each firm is Supply Chain 4.0 as the application of Industry 4.0 to supply designed to ensure that deliveries are coordinated with the cus- chains, and then to map each of the stages of the supply chain tomers’ sourcing activities, and that sourcing activities are coor- (planning and design, production, distribution, consumption, dinated with the suppliers’ delivery activities, and that returns of reverse logistics) to one or more of the iconic technologies said unwanted or unneeded products are accounted for (PWC 2016b). to be typical of Industry 4.0: the IoT, cloud computing, artificial The processes by which this is done have been codified in the intelligence, etc. The difficulty immediately arises that the appli- Supply-Chain Operations Reference (SCOR) model, originally cation of technologies to sets of problems is fluid, and it takes a developed in 1996 by the management consulting firm PRTM FIGURE 5.3 The currently fashionable model of Industry 4.0 is over-simplified, but it reflects current thinking about what’s happening now (2018)  Cyber Physical Mechanization, Mass production, Automation, Systems, steam power, assembly line, computers and internet of things, weaving loom electrical energy electronics networks Source: https://www.hammelscale.com/industry-4-0/ Understanding Supply Chain 4.0 and its potential impact on global value chains • 107 FIGURE 5.4 Industry 4.0 technologies, by relative emphasis in recent studies  Internet of Things Big data analytics 3D printing Advanced (autonomous) robotics Smart sensors Augmented reality Cloud computing Energy storage AI or machine learning Nanotechnology Synthetic biology Simulation Human-machines interface Mobile devices Cyber security Quantam computing Horizontal and vertical integration 0 1 2 3 4 5 6 7 8 9 10 Number of references Source: Graphic from Hallward-Driemeier and Nayyar (2017). (now part of PriceWaterhouseCooper) and AMR Research (now part of Gartner) (Lambert 2008, p. 305), and are now part of a FIGURE 5.5 Traditional supply chain model de facto standard strategic, management, and process improve- ment methodology for supply chain management. The ideas behind SCOR, and their implementation, have been important Plan Plan Plan for the development of global value chains and for supply coor- dination among networks of firms. As successful as this method of supply chain management has Supplier Production Distribution Customer/ been, it has limitations. Flows of information tend to primarily Consumer link each firm to its immediate suppliers and customers, not to firms further down the chain. In supply chains with multiple links, this leads to delays in the processing of information. In particu- Order and Order and Order and confirm confirm confirm lar, changes in the system flowing from changes in final demand, which are often unpredictable, become distorted as they pass upstream, analogous to the old child’s game in which a message Source: PWC 2016b. whispered from one player to another becomes more and more different from its original content. Even with a lead firm acting as “impresario” of a network of firms, one actor is unlikely to have disk drives in Thailand, which are in turn shipped to assemblers full information about everything that is going on in the supply of laptop computers in China, cannot see changes in consumer chain. Managers at Walmart, planning for the fall apparel season, demand visible to Best Buy, a retailer in the United States (Hirat- are in some sense leaders of their global supply chains (USITC suka 2005). The term Supply Chain 4.0 can be usefully applied to 2011, 3-33 ff). But they are unlikely to actually know what is hap- an integrated supply chain ecosystem, in which information flows pening in button and zipper factories in Bangladesh which are in all directions, analytics enable adjustment throughout the part of their supply chain. That information is held by middlemen. supply chain, and response takes place in real time (PWC 2016b) Firms in Singapore, which ship small screws to manufacturers of (see Figure 5.6).10 To rapidly assess and respond to changes in 108 • Technological innovation, supply chain trade, and workers in a globalized world FIGURE 5.6 Integrated supply chain ecosystem  Track and trace Customer/ Distribution Consumer Pla n an Pl Q ua y an lit lit d y ua us st Q at at us st d Or an d co er a nd nf nd irm e ra m Supply chain rd ir O onf control tower c lan Pla P n y Q lit ua ua t us an lit Q ta d d s sta y an tu s O r nd co der e ra m nf an Supplier rd fir irm d Track and trace O con Production Source: PWC 2016b. customer demand, tracking and tracing throughout the supply of nearly 40 times, processing costs have declined almost 60 chain is enabled through sensing technologies underlying the times, and many of the sensors used in IoT technology cost Internet of Things (IoT), including radio frequency identification no more than 60 cents (CGI 2016). These data are only useful (RFID), Bluetooth, and GSM (global system for mobile commu - if they can be reduced to information useful for making deci- nication), which links maritime transport to satellites. In par- sions in real time that create business value. Big data analytics ticular, changes in customer demand can be rapidly assessed thus are about using data to drive useful business intelligence, and responded to. This technology has had a wide uptake. answering the questions, “What just happened?”, “Why did it According to a recent PwC study on the rise of Industry 4.0, a happen?”, and “What are we going to do next?”. Specific appli - third of the more than 2,000 respondents say their companies cations of big data analytics include early warning algorithms have started to digitize their supply chains, and fully 72 percent (are we about to run out of something or hit a bottleneck? Did expect to have done so five years from now (PWC 2016a, p. 11). prices we care about just rise?), predictive algorithms (what is demand likely to look like next spring, or five years from now?), 2.2 Big data and supply chain analytics – running stock-keeping unit (SKU) rationalization (the decision about scenarios from a supply chain control tower the optimal set of products, or SKUs, to offer to consumers at New technologies gather prodigious amounts of data. In the any given time), channel assessment (the decision about the last decade, the cost of bandwidth has decreased by a factor optimal way to get product to end market, e.g. e-commerce/ Understanding Supply Chain 4.0 and its potential impact on global value chains • 109 distributors/company-owned outlets/large and small retailers/ reduces maintenance costs. (CGI 2017) For example, Microsoft mail-order/etc.), and dashboards (user-friendly quick visualiza- and CGI developed a smart-sensor based solution for a com- tion in “supply chain control centers”). The ability to collect and pany that maintains more than 1.2 million elevators worldwide. analyze data gathered in the whole supply chain makes it possi- Information from the sensors is made available to service tech- ble to “run scenarios within the platform” (PWC 2016b), where nicians and their supervisors through cloud-based dashboards. the platform is conceived of as an overarching software solution Manyika et al. (2015) estimate that predictive maintenance within the supply chain control center. using IoT can reduce maintenance costs of factory equipment The desire to collect and distribute data rapidly across a by 10-40 percent and reduce equipment downtime by up to 50 supply chain explains much of the recent enthusiasm for block- percent. Similarly, the use of predictive analytics and IoT can chain technologies in the context of supply chains (Petersen have a big impact on energy maintenance, both by using energy et al. 2017). Blockchain is a distributed ledger technology that consumption data to detect potential equipment failures and by allows multiple parties to maintain copies of the same informa- continuously modifying equipment settings and process param- tion in different locations, either in an open manner or requiring eters in real time (CGI 2017). individual entities’ permission to access the network. Blockchain protocols encode information such as numbers or programs, 2.4 Smart logistics and the warehouse of the future time-stamp them, and enter them as a block into a continu- Smart logistics encompasses not only scheduling of transport, ous chain of previous blocks linked to the same transaction but also activities within the warehouse. It is within the ware- (Niforos, Ramachandran and Reherman, 2017). Such attributes house that many of the most profound changes are already make blockchain attractive for supply chain management, as taking place. As noted above, one of the big changes is that well as for other uses such as fintech, cryptocurrencies, smart the warehouse and the customer become more visible to each contracts, and security. Blockchain technology also has poten- other, so that customer final purchases trigger not only prod - tial application in port logistics, improving tracking and trac- uct moves from the warehouse but also product moves from the ing of containers and coordination among the diverse actors manufacturer to the warehouse. in ports such as carriers, ship agents, terminal operators, insur- In e-commerce, the Internet makes the warehouse visible ers, customs agents, financial institutions and inland transport to the customer. A familiar example of this is the notice one (Weernink et al., 2017). While there is a great deal of hype about encounters at Amazon.com, “Only three left! Hurry!”, which blockchain and supply chains at the present moment, pilot proj- can be used to influence both consumer behavior and trigger ects involving establishing origin of Australian oats, preventing re-stocking. At Taobao.com, the giant Chinese e-commerce counterfeiting of Italian wine, combating fraud in diamond mar- platform, customers are presented with both inventory and kets, and tracing the provenance of geological samples have sales data for products. Alibaba is another platform that func- demonstrated proof-of-concept (Petersen et al. 2017). It should tions as the architect of an increasing complex eco-system, that be noted that many of these coordination functions can be per- includes designers/entrepreneurs, marketers, payments, financ- formed by combinations of technologies that do not involve ing (credit) logistics suppliers, integration of on- and offline blockchain. retail, supply chains and manufacturing, all of which are com- plementary players in the eco-system interacting on the net- 2.3 Smart factories/fractal factories/M2M work, in rapid-response, data-driven, algorithm-guided mode communications/driverless programmable vehicles (Spence, 2018). Improved data gathering within the IoT, combined with analyt- The predictive maintenance techniques discussed above can ics, enables process optimization within the factory as well, in reach into the warehouse as well, which can similarly optimize order to enable timely business decisions. The application of delivery of spare parts to factories. Indeed, with a flexible 3D Supply Chain 4.0 within manufacturing facilities is sometimes printer, spare parts can be produced in the warehouse, trig- referred to as the “smart factory” (Pfohl et al. 2015). Embedded gered by demand. Some analysts project that 3D printers, which data collection units, using both automatic identification and can be placed in any environment including delivery trucks, may data collection and radio-frequency identification (RFID) tech - make warehouses obsolete. nologies, can be embedded in most pieces of factory equip- A traditional warehouse involves a good deal of “pick and ment. The information can be passed from machine to machine pack” activity. Employees search around in the warehouse for (M2M) and handed to a supply chain control tower for deci- products that have been ordered, take them off the shelves, and sion making. Autonomous robotics simply refers to the control pack them. If the warehouse serves several firms, the packing and reprogramming of robotics using bilateral and multilateral may involve selecting packing materials marked with the logo of machine communication. Intra-logistics within factories includes a particular firm. Clearly knowing where the products are located the use of driverless vehicles to move materials based on exter- in a large warehouse, and moving through the warehouse in a nally-provided information. time-minimizing manner, can speed up delivery time substan- One of the most important features of the Smart Factory is tially and reduce errors. Within the warehouse, autonomous the ability to do predictive maintenance. The use of sensors to logistics and robotic transport can be employed to substantially identify maintenance needs in advance of potential breakdowns improve pick-and-pack performance. Other technologies can be 110 • Technological innovation, supply chain trade, and workers in a globalized world used as well. Here’s one example of the use of augmented reality 2.5 E-Commerce is ideally, but not uniquely, suited for in a warehouse: Supply Chain 4.0 As we have seen, many of the tools of Supply Chain 4.0 can “DHL recently conducted tests on an augmented reality system be applied to traditional store-based retailing. The expan - at a warehouse in the Netherlands owned by Ricoh, the Japanese sion of e-commerce, however, allows additional ways in which imaging and electronics company. Equipped with smart glasses new technologies can be implemented. One obvious feature containing software from Ubimax, employees navigated through of B2C-commerce is that the process of purchasing involves the warehouse along optimized routes via the glasses’ graphics electronic data entry on the part of the consumer. This enables display, enabling them to find the right quantity of the right item information to be captured, preferences to be assessed, and much more efficiently, and with reduced training time. Over the strategies to target the consumer to be implemented, such as three weeks of the test, 10 order pickers succeeded in fulfilling the ubiquitous pop-ups which now follow one around the Inter- 9,000 separate orders by picking more than 20,000 items. The net after having viewed a product in a given category. resulting productivity improvements and reduction in errors Although most of the popular discussion of e-commerce increased the overall picking efficiency by 25 percent” (PWC is on B2C, nearly 90 percent of e-commerce is in fact busi- 2016b, p. 22).” ness-to-business (B2B) (UNCTAD 2017, from which Table 5.1). This means by definition that it consists of links in supply chains This example highlights a feature of many Supply Chain 4.0 – whether transactions between parts suppliers and assem - technologies which will be important for understanding their blers, between distribution centers and retailers, or online employment effect. The use of new technology and human labor purchases of services which in many cases support the supply are often complements, rather than substitutes, especially in chain. B2B commerce can be implemented either through conditions where e-commerce is substantially increasing demand websites, much like B2C e-commerce, or through electronic for certain goods. Rugaber (2018) reports that the online retailer data interchange (EDI). EDI is a mature technology12 through Boxed in Edison, New Jersey opened up an automated ware - which the computer systems of the buyer and seller are directly house in Union, New Jersey. Demand for goods was such that connected using a common record format.13 As an example the firm ended up employing more humans, adding a third shift, of the pervasiveness of EDI, the United Kingdom’s Office of as well as more robots. The new jobs are less physically demand- National Statistics finds that a majority of all e-commerce in ing as well. Rather than taking thousands of steps a day loading the U.K. consisted of B2B e-commerce conducted through EDI, items onto carts, employees can stand at stations as conveyor as opposed to over websites that resemble B2C e-commerce belts bring goods to them.11 (Table 5.2). TABLE 5.1 Top 10 economies by total, B2B and B2C e-commerce, 2015, unless otherwise indicated Total B2B B2C Share in total Economy $ billion Share in GDP (%) $ billion $ billion e-commerce (%) 1 United States 7,055 39 6,443 91 612 2 Japan 2,495 60 2,382 96 114 3 China 1,991 18 1,374 69 617 4 Republic of Korea 1,161 84 1,113 96 48 5 Germany (2014) 1,037 27 944 91 93 6 United Kingdom 845 30 645 76 200 7 France (2014) 661 23 588 89 73 8 Canada (2014) 470 26 422 90 48 9 Spain 242 20 217 90 25 10 Australia 216 16 188 87 28 Total for top 10 16,174 34 14,317 89 1,857 World 25,293 - 22,389 - 2,904 Source: UNCTAD Information Technology Report 2017. Understanding Supply Chain 4.0 and its potential impact on global value chains • 111 TABLE 5.2 The United Kingdom reported that about 50 3. The impact of Supply Chain 4.0 on percent of e-commerce in 2017 was electronic data consumers – customer fulfilment increasingly interchange B2B resembles magic Value in 2015 Grand Total In a traditional consumer supply chain, the final step is an Mode Sector (billion UK £) (%) in-store retail establishment. Consumers frequently experience All modes Total 560 100 the frustration of goods being out of stock, either goods that are usually on the shelves but are not there on the day the con- of which B2B 400 71.4 sumer is in the store, or goods that the consumer would like of which B2C 160 28.6 to buy and knows that they exist, but that the store does not carry. In such cases, the remedies are familiar. Do you have Electronic Data Interchange (EDI) any more in the back room? May I speak to a manager? For a Total (B2B)* 281 50.2 particularly vigorous consumer inquiry, the manager might be Website Total 279 49.8 prevailed upon to call another store in the chain, or a regional of which warehouse or distribution center. By this time, the consumer B2B 119 21.3 may well have given up and not made the purchase at all, or of which B2C 160 28.6 gone to a competitor. Applications of IoT are increasingly used to facilitate the Source: UK Office for National Statistics. management strategies of “customer-managed inventory” * EDI can be explained as an automated transaction between businesses and (CMI) or “vendor-managed inventory” (VMI). These strategies therefore EDI sales are classed as business-to-business sales. represent a revolution in supply chain management of compa- rable importance to the “just-in-time” revolution in manufac- turing pioneered at Toyota and other companies in the 1960s. In such models, information is initially provided by a customer, for example by scanning a bar code associated with a purchase, Transactions between businesses which take place without and then transmitted up the supply chain to the warehouse/dis- EDI involve multiple processes of transmission and re-copying tribution center.15 Technologies such as RFID tags then transmit of data. A customer creates an order manually, perhaps using information to the distribution center so that orders can be ful- a computer. The order is transmitted by telephone or fax. It is filled. The information involved is mediated by EDI (see above manually keyed into the vendor’s computer system. When the under e-commerce). Since demand still cannot be fully forecast, order is fulfilled an invoice is created manually (with or without models of inventory management such as scan-based trading the aid of a computer). The invoice is sent back to the cus- or consignment distribute the risk between suppliers and retail- tomer, who enters the data on the invoice manually. ers by enabling retailers to take physical possession of inven- Each of these steps in the process is time-consuming. More - tory while suppliers retain ownership, so that the sale between over, each step is a place at which error can be introduced into the supplier and retailer does not actually take place until the the system, leading not only to slow order fulfilment but to final consumer checks out at the register. More complex ver- lack of fulfilment or mis-fulfilment. An EDI system causes an sions of this transaction are possible. order created electronically by the customer to be instantly By mediating a series of linkages between retailers, ware- duplicated without error in the vendor’s computer system, and houses, manufacturers, and suppliers of inputs to manufac- the invoice to be similarly electronically duplicated in the cus- turing, EDI-driven CMI minimizes forecasting errors along the tomer’s computer system. supply chain. As a hypothetical example, a consumer checking Besides saving time and labor, and reducing errors, EDI out of an AT&T store in California with a newly purchased Sam - enables a large amount of data capture about customer sung smartphone may, by the single act of purchase, trigger a behavior. Thus, data captured in EDI can be the basis for chain of information going all the way back to a company that supply chain analytics using either big data or “small data” supplies Samsung with touch screens relatively quickly, with techniques. One study of manufacturers in the Czech Repub- tight linkages between the “supply chain control towers” of lic finds that firms using EDI were also more likely to adopt Samsung and AT&T. advanced techniques of inventory management, such as con- Future developments in in-store retailing, enabled by IoT signment stocks, buffer stocks, and safety stocks14 (Vrbová et technology, will enhance both the customer experience and al 2016). The same study reports that sectors with above-av - the ability of stores to pursue advanced management strate- erage use of EDI include auto parts, electronics, engineering gies (Gregory, 2015). Using their cell phones, customers may be industries, plastics, retailing and textiles. These are all sectors able to scan barcodes on items to obtain product information associated around the world with well-organized value chains, or identify other colors or sizes available on the retailer’s web- showing the use of EDI-driven data capture and analysis in site. VIP customers may be offered virtual coupons on enter- value chains. ing the store. Smart mirrors may allow customers to “try on” 112 • Technological innovation, supply chain trade, and workers in a globalized world different clothing virtually. This experience, which immerses FIGURE 5.7 Hours spent per day shopping in the U.S. the customer in a retail environment with the aid of both mobile (2002-2018) and in-store devices, may be known as the Internet of Me. From the management standpoint, smart price tags can be changed in 1.2 real time based on demand or other needs, and “smart shelves” in store could detect low inventory, thus providing further sup- 1.0 port for CMI and VMI strategies. Of course, many of these same principles apply in markets for intermediate goods – B2B mar- 0.8 kets. In these markets, the ability to use analytics and advanced supply chain management to improve performance is in many 0.6 ways more advanced than in business-to-consumer (B2C) mar- kets, especially in sectors such as electronics, apparel, and motor 0.4 vehicles where sophisticated supply chain methods have been in existence for an extended period of time. This is also discussed 0.2 in the section on e-commerce. 0 2002 2004 2006 2008 2010 2012 2014 2016 2018 4. The impact of Supply Chain 4.0 on workers Total Men Women 4.1 Physical labor in warehousing and driving substitutes for household time Source: BLS American Time Use Survey. In an important recent contribution, Mandel (2017) demon - strates that U.S. sectors involved in supply chain activities associated with e-commerce have generated a significant amount of employment over the last decade – over twice as particularly in an affluent society, the implications of e-com - much as the reduction in employment in store-based retailing merce for social welfare are potentially profound. This includes occurring at the same time. Moreover, the jobs involved are implications for the gender distribution of labor. A reasonable reasonably well-paying, and to some extent look like the old hypothesis is that a further examination of the American Time factory jobs in manufacturing which became less numerous Use Survey would reveal that the hours spent in shopping during the period 1979-2010. activities are disproportionately female, while the employ- Specifically, Mandel finds that from December 2007 to June ment in supply chain activities are likely to be relatively more 2017, e-commerce jobs in fulfilment centers and e-commerce those of male mail workers. We leave this hypothesis for future companies rose by 400,000, substantially exceeding the examination. 140,000 decline in brick-and-mortar retail jobs. On a country by country basis, fulfilment center jobs pay 31 percent more 4.2 Overall trends in supply chain employment than brick-and-mortar retail jobs in the same area. Data from the American Time Use Survey (BLS) imply that Data in 2016, Americans age 15 and over spent 1.2 billion hours We analyze a group of sectors particularly involved in the dis- per week driving to the mall, finding a parking place, wander- tribution of goods, including wholesaling (both traditional and ing around the aisles, checking out, and driving home. The electronic), retailing (both store-based and non-store based), number of hours spent by each such individual shopping per couriers and messengers, and warehousing and storage week declined from 4.9 in 2005 to 4.4 in 2012, recovering (Table 5.3). We call the aggregate of these data the “supply slightly to 4.5 in 2016. Due to online shopping, in the years chain sectors.’ We then use data from the Occupational Employ- between 2006 and 2012 each individual over age 15 spent 6 ment Statistics (OES) of the U.S. Bureau of Labor Statistics to minutes fewer a day in the purchase of goods and services, track sector-level employment as well as employment in individ- which adds up to 11.8 billion leisure hours a year to spend on ual occupations in each sector. For contrast, we compare the something else (see Figure 5.7). At the same time, the brick- results with trends in manufacturing and in the U.S. economy as and-mortar share of retail sales declined from 98 percent to a whole. 92 percent. We focus on the period from 2011-2016. Even though it is a Thus, e-commerce is a mechanism for translating unpaid very recent period, it corresponds roughly to the period during household shopping time (which has valuable alternate uses) which the discussion of “Industry 4.0” (and thus, eventually, to paid market time. Instead of consumers spending time “Supply Chain 4.0”) crept into the public awareness. This is a shopping, workers in warehouses and on delivery trucks are shorter period of time than covered in Mandel (2017). Moreover, picking goods off warehouse shelves and bringing them to we have a greater focus on the occupational composition of the consumer’s front door. Since time is a scarce resource, employment. Understanding Supply Chain 4.0 and its potential impact on global value chains • 113 TABLE 5.3 Sectors of employment defined as U.S. “supply for substantially more activity than general merchandise stores. chain sectors” Among the rapidly growing supply chain sectors, the largest job gains have been in warehousing and storage. Total supply chain sectors: 423 Merchant Wholesalers (durable & non-durable goods) Types of employment increasing in supply chain sectors 425 Wholesale Electronic Markets and Agents and Brokers The dominant category of employment that has expanded in 441-8 & 451 Specialty Store Retailers the current supply chain boom is “transportation and mate- rial moving operations.” Over 2011-2016, these occupations of which motor vehicles and parts dealers; furniture and home furnishings; electronics and appliance; accounted for an increase in employment of over 350,000 in warehouses and courier services (Figures 5.10 and 5.11). These building materials and garden equipment and types of jobs involve a combination of physical and mental activ- suppliers; food and beverage; ity comparable to that of Industry 2.0, but less strenuous because health and personal care; gasoline stations; clothing of the effects of mechanization. and accessories; Among e-commerce firms proper (electronic shopping and sporting goods, hobby, book and music mail-order houses), the greatest absolute growth in employment 452 General Merchandise Stores has been in office and administrative support occupations, with 453 Miscellaneous Store Retailers (e.g. dollar stores) the second largest absolute growth (and largest percentage change growth) being in business and financial operations occu- 454 Non-store Retailers pations) (see Figure 5.12). Among specialty stores, employment in of which Electronic Shopping and Mail Order Houses many of the back-office occupations has declined, and the gains 493 Warehousing and Storage have come in customer-facing occupations – sales and related occupations, and health care practitioners and technical occu- 481-4 General Transportation pations (see Figure 5.13). The gain in health care workers can be of which Air, rail, water, and truck transportation attributed to a single category of specialty stores, pharmacies. 491 Postal Service Companies such as Walgreens and CVS are increasingly offer- ing vaccinations and other basic health care services hands-on in 492 Couriers and Messengers their retail establishments, which carry many of the same items For comparison: available in food stores and general merchandise retailers. 31-33 Total manufacturing Total supply chain sectors 5. The impact of Supply Chain 4.0 on GVCs Total U.S. economy Note: Sectors of employment defined using the Occupational Employment Supply Chain 4.0 can be seen either as an advanced manage- Statistics (OES) data from BLS from 2011-2016. ment practice, or as a cluster of technologies more likely to be adopted as the result of advanced management practices. As shown by recent survey-based research, improvement of man- Supply chain sectors associated with e-commerce expe - agement practices – such as may be associated with adoption rienced rapid employment growth from 2011 to 2016. While of Supply Chain 4.0 – is likely to enhance productivity and prof- employment growth in U.S. supply chain sectors as a whole (8.7 itability, lead to higher-quality outputs produced using high- percent) was below that of overall employment (9.2 percent), er-quality inputs (Bloom, Manova, Sun, Van Reenen and Yu 2018). employment growth was much higher in the subcomponents of Supply Chain 4.0 is designed to enhance key management com- warehousing and storage (28.9 percent), non-store retailers (20.3 petencies, such as effective target setting, collecting and ana- percent), and couriers and messengers (16.0 percent). Within the lyzing data to monitor progress towards these targets, inventory subcategory of non-store retailers, employment in the category management, coordination of targets/progress across produc- of electronic shopping and mail-order houses, which approaches tion stages, and worker supervision and incentives. most closely the usual conception of e-commerce, grew even Supply Chain 4.0 technologies may enable firms to reduce the more rapidly at 41 percent.16 The time profile of employment number of stages in supply chains by reshoring routine labor-in- increase shows that while jobs in the “couriers and messengers” tensive activities in developing countries back to the developed sector grew steadily throughout the period, those at non-store countries. These technologies make undertaking some produc- retailers experienced an acceleration after 2013, while in ware - tion stages in high-wage countries more profitable by reducing housing and storage the acceleration kicked in after 2014 (Figure the amount of labor required, thus weakening the incentive for 5.8). In terms of absolute job gains in the supply chain sectors, firms to locate in low-wage countries and reducing the impor- these were mainly in specialty stores – that is, stores that spe - tance of low labor costs in determining comparative advantage, cialize in one type of merchandise such as food, apparel, elec- providing instead an advantage to integrating multiple stages tronics, cars, or sporting goods (Figure 5.9). Such stores account of production at a single automated location (Dachs et al. 2017). 114 • Technological innovation, supply chain trade, and workers in a globalized world FIGURE 5.8 Employment growth in U.S. supply chain sectors and overall economy, index, 2011 = 100 150 Warehousing and Storage 145 Nonstore Retailers 140 Couriers and Messengers Total US economy 135 Total supply chain sectors 130 Specialty store retailers 125 Wholesale Electronic Markets and Agents and Brokers 120 General Merchandise Stores 115 Manufacturing 110 Miscellaneous Store Retailers Merchant wholesalers 105 General Transportation 100 2011 2012 2013 2014 2015 2016 FIGURE 5.9 Absolute changes in U.S. employment, supply chain sectors, and transportation (2011-2016) 1,000,000 900,000 800,000 700,000 600,000 500,000 400,000 300,000 200,000 100,000 0 -100,000 ok s n ge rs s rs s es rs rs Br et er tio er le ile ile ile or a d rk ng sa or rta ta ta ta St an a le ts ic M se St re Re Re po se ho es d e ns di re e an en n or tw M or o an ra to st Ag tr d g st an lT sS ch d lec an in ty on ch a s er al ou er an E ou N rs er i lM ec en le ie ne eh M sa ur Sp a G la ar er Co le el W en ho isc G W M Understanding Supply Chain 4.0 and its potential impact on global value chains • 115 FIGURE 5.10 Warehousing and storage – changes in employment in selected occupations (2011-2016) 200,000 35% 180,000 35% 160,000 30% 140,000 25% 120,000 100,000 20% 80,000 15% 60,000 10% 40,000 5% 20,000 0 0 Transportation Office and Production Management Installation Sales and Business Computer and material administrative occupations occupations maintenance, related and financial and moving support and repair occupations operations mathematical occupations occupations occupations occupations occupations Gross change Percent change FIGURE 5.11 Couriers and messengers – changes in employment in selected occupations (2011-2016) 200,000 60% 150,000 40% 100,000 20% 50,000 0% 0 -20% -50,000 -40% -100,000 -60% Transportation Office and Management Installation Computer and material administrative occupations maintenance, and moving support and repair mathematical occupations occupations occupations occupations Gross change Percent change 116 • Technological innovation, supply chain trade, and workers in a globalized world FIGURE 5.12 Electronic shopping and mail order houses – changes in employment in selected occupations (2011-2016) 40,000 120% 35,000 100% 30,000 80% 25,000 20,000 60% 15,000 40% 10,000 20% 5,000 0 0 Office and Sales and Transportation Management Computer Business Production Healthcare administrative related and material occupations and and financial occupations practitioners support occupations moving mathematical operations and technical occupations occupations occupations occupations occupations Gross change Percent change FIGURE 5.13 Specialty stores – changes in employment in selected occupations (2011-2016) 20,000 300% 15,000 250% 10,000 200% 5,000 150% 0 100% -5,000 50% -10,000 0% -15,000 -50% at ted io e tio e, n n tio ial tio nt, tio al tio al at ers at tiv io tio pa tic pa nc pa nc pa eri pa e at ns ns ns ns ns ns ns ns up la up n up ra pa cu nm cu ma cu na cu na cc itio cu at up cc d re io io cc ist cu oc te oc m oc d fi oc tai oc he cc t o in l o ct ir ain oc n d ia ter m at ca ra to ns n sa in an a d m pa , m p n ed n en pp a tio le s m ,e ch are ov n o nd s su and d ion Sa m tio em er ine uc g d gn or ra c ta an lat lth od io ag op us an si ni ce te or re at ts de al B an ea Pr an pu ffi sp st te H M O or s, In m an sp rt d Co Tr A Gross change Percent change Understanding Supply Chain 4.0 and its potential impact on global value chains • 117 It has been argued that 3-D printing works in this way. Accord- concentration in sectors where it is important, and to increased ing to one estimate it is expected that 3-D printing will disrupt income inequality across countries. Countries with higher internet between 4.6 percent and 14.9 percent of global trade flows (Arvis penetration, firms and countries with greater digital entrepreneurial et al. 2017). By shortening GVCs, 3-D printing may eliminate the skills, and firms which have mastered previous generations of supply productivity benefits associated with international trade in manu- management practices (such as the SCOR model of the 1990s) are factured goods by reducing the need for unskilled labor-intensive likely to have advantages in adopting Supply Chain 4.0 methods. tasks. On the other hand, 3-D printing has actually been associ- Conversely, attempts by developing countries to promote ated with increased trade in at least one sector – hearing aids – entry into new manufacturing sectors, particularly using strategies where the technology has been rapidly adopted (Freund, Mulab - promoting domestic firms with subsidies, incentives, and special dic and Ruta 2018). zones, might not take into account whether key players in the supply The new digital technologies are driving a revolution in the way chain are using the most advanced technologies, and thus be at a firms are shaping the organization of their production processes. competitive disadvantage relative to strategies which successfully For example, in 2016, Adidas opened a fully-automated shoe fac- attract FDI from firms which have mastered Supply Chain 4.0. tory using 3-D technology and robotics in Germany. The goal was Differences in the rate of diffusion and adoption of Supply to individualize its products and react more promptly to consumer Chain 4.0 may not necessarily have negative impacts for poverty needs by bringing manufacturing closer to its clients and speed- alleviation or income growth of people with lower incomes in ing up delivery. The number of workers required in this factory is a developing countries. It depends on how the gains from the new fraction of the number of people working in emerging economies management practices are distributed along the supply chain. For in the production of the same sportswear (Backer and Flaig 2017). example, in some cases the application of advanced supply man- Thus, this form of innovation may slow the growth of GVCs and agement practices to an agriculturally-based supply chain origi- increase the importance of skills development. nating in developing countries could enable additional steps of The reshoring of production by high-income countries could food processing in those countries, while in other cases they could reduce demand for the products of manufacturing exporters and lead to increased export of raw materials. In the case of increased stifle the potential entry of newcomers into manufacturing GVCs export, whether farmers capture any of the gains may depend (Hallward-Driemeier and Nayyar 2017). The higher and more spe- on whether improved (likely foreign) management of the overall cific investments in advanced production technology are, the supply chain induces farmers to produce higher-quality produce greater the possibility to integrate manufacturing operations at at higher prices, or to have higher rejection rates. The effects of one focal plant, favoring reshoring (Dachs et al. 2017). A report Supply Chain 4.0 on poverty and shared prosperity are thus likely by Citigroup and the University of Oxford’s Oxford Martin School to be contingent on a variety of local circumstances. finds that 70 percent of Citi institutional clients surveyed believe Another potential impact of Supply Chain 4.0 relates to the that automation will encourage companies to move their man- interactions between firms and governments. Improved supply ufacturing closer to home, with North America having the most chain management can lead to increased traceability of goods and to gain from automation, followed by Europe and Japan. By con- financial information. This could make it easier for firms engaged trast, the authors estimated that China, Association of Southeast in international trade to satisfy rules of origin by providing a com- Asian Nations (ASEAN) member countries, and Latin America prehensive audit trail, and it could make it easier for governments have the most to lose from automation (Citigroup 2016). Hence, to monitor some types of tax evasion. the increased use of labor-saving technologies will change the patterns of comparative advantage of manufacturing in the global market. 6. Conclusion On the other hand, developments in the technologies such as IoT, big data and cloud computing can strengthen the current It is dangerous to take a snapshot of recent history, whether of structure of GVCs by reducing the costs of tracking and monitor- technologies, institutions, or economic trends, and project it ing the components of production, thus lowering coordination very far in the future. Current developments in supply chains and matching costs. A survey of 152 decision-makers in auto- appear to be employment-generating, but this could reverse motive, aerospace, electronics, and industrial equipment manu- if developments in robotics advance in certain directions. The facturing companies in Germany, France, and the U.S. finds that technology could evolve in entirely unpredictable ways. Or, more the biggest benefit of cloud computing is to reduce the cost of pessimistically, its diffusion could stall, limiting the application of optimizing infrastructure (48.3 percent of respondents), followed Supply Chain 4.0 to already high-income countries and becom- closely by efficient collaboration across geographies (47.7 per- ing another contributor to global income divergence, which may cent) and the ability to respond quickly to business demands (38.4 already be the case with Industry 4.0. Concerns about consumer percent) (the Microsoft Discrete Manufacturing Cloud Computing privacy could easily cause governments to act to forestall some Survey, Microsoft Corporation 2011). of the developments discussed here. For the present, though, The degree of adoption and diffusion of Supply Chain 4.0 jobs are being created in supply chains, and advances in supply processes is likely to vary across both firms and countries. As a chains are creating benefits for consumers. This can be taken as result, in the medium run it could give rise both to more industrial at least a small cause of optimism. 118 • Technological innovation, supply chain trade, and workers in a globalized world Notes from dispersed locations to be assembled at a common location (cf. Hiratsuka 2005). The concept of a supply chain control tower applies 1. Or, likely as not, in China, though this chapter does not attempt to equally well to “snake” and “spider” type supply chains. In a more document that specifically. elaborate chain, in which some lead firm orders major assemblies 2. Six principles of Industry 4.0 are: 1) Interoperability: the ability for from Tier I suppliers, which in turn order sub-assemblies from Tier II plant equipment (i.e., workpiece carriers, assembly stations and prod- suppliers, and so on, the flow of material may resemble a combination ucts), humans or smart factories to connect and communicate with of “spiders” and “snakes”. In such a complex supply chain, it might each other via the IoT and the Internet of Services; 2) Virtualization: a make sense for each Tier I supplier to have its own supply chain con- virtual copy of the smart factory created by linking sensor data (from trol tower, with information being further aggregated at the level of monitoring physical processes) with virtual plant models and simula- the lead firm. tion analytics; 3) Decentralization: the ability of cyber-physical systems 11. Not every development in robotics is complementary to human labor. within smart factories to make decisions on their own; 4) Real-time The development of prototype robots that can pick goods from capability: the capability to collect and analyze data and provide the shelves could lead to robots that would easily replace some workers. derived insights immediately; 5) Service orientation: offering of ser- However, the dexterous movements of the human hand and arm have vices (of cyber-physical systems, humans or smart factories) via the proved difficult to replicate mechanically. This replicates the expe- Internet of Services; and 6) Modularity: flexible adaptation of smart rience of the first Industrial Revolution, in which there was approx- factories to changing requirements by replacing or expanding indi- imately an 80-year gap between the development of mechanical vidual models. spinning and the invention of the sewing machine (which still needed 3. Available at: https://www.usatoday.com/story/money/2018/02/19/ dexterous human labor). Gordon (2016) reports that in advanced kfc-chicken-shortage-u-k-restaurants-close-amid-delivery-mis- robotics competitions, robots still have difficulty opening doorknobs. haps/350698002/,February 19, 2018. 12. International organizations began developing record formats for EDI 4. The term “Industry 4.0” is of German origin. It arises from the German as early as the 1960s (UN/CEFACT et al 2017). By the 1980s the use of Government’s High Tech 2020 strategy, an initiative launched in EDI for firm-to-firm transactions, both nationally and across borders, 2011 and conducted through the Ministry of Education and Research was widespread. (BMBF) and the Ministry for Economic Affairs and Energy (BMWI) 13. The connection for EDI can either be a direct physical (hardwired) (European Commission 2017). As popularized, Industry 4.0 refers to connection, or implemented over the Internet, or, more recently, take the most recent in a sequence of “industrial revolutions” in historical the form of a cloud-based solution. time (e.g. Hallward-Driemeyer and Nayyar 2018, 40-41). 14. According to Vrbová et al. (2016) in consignment stock the vendor, 5. The definition by analogy to Industry 4.0 corresponds to the most instead of the buyer, is in charge of managing the buyer’s inventory common usage of “Supply Chain 4.0”, e.g. Alicke et al. 2016, Asthana and triggering replenishment orders; in buffer stock the placement 2018. To our knowledge, nobody has attempted to provide histori- takes place at a particular critical stage of supply chain; and in safety cally-based definitions of “Supply Chain 1.0,” “Supply Chain 2.0” or stock it is stored in the final stage of the supply chain. “Supply Chain 3.0.” 15. This paper will use the older term “warehouse” and the more modern 6. A “smart factory” is a highly digitized and connected production facil- term “distribution center” interchangeably, as synonyms. Increas- ity of the type associated with Industry 4.0. The idea of a “smart fac- ing use of “distribution center” in place of “warehouse” is associ- tory” is still in its infancy and does not refer to a tightly standardized ated with the spread of more advanced techniques of supply chain specification of operations. management. 7. “Augmented reality” refers to a technology that superimposes a 16. Besides e-commerce, “non-store retailers” includes such firms as computer-generated image on a user’s view of the real world, thus direct sales (i.e. door-to-door or house parties) and vending machines. providing a composite view. It includes as a subcategory “virtual real- ity,” displays of information of a “3D” or “real” character mediated by such hardware as special headsets or gloves. 8. “Artificial intelligence” (AI) refers to the theory and development of computer systems able to perform tasks that normally require human intelligence, such as visual perception, speech recognition, deci- sion-making, and translation between languages. It is closely related to the concept of “machine learning,” i.e. computer systems that improve their performance with accumulated experience. 9. “Cloud computing” denotes the practice of using a network of remote servers hosted on the Internet to store, manage, and process data, rather than a local server or a personal computer. 10. Figures 5.5 and 5.6 portray a linear supply chain where goods are moved from one location to another sequentially – a structure often referred to as a “snake” supply chain. “Snakes” are contrasted with “spider” supply chains, in which parts and components are brought Understanding Supply Chain 4.0 and its potential impact on global value chains • 119 References Lambert, Douglas M. (2008) Supply Chain Management: Processes, Part- nerships, Performance. 3rd Edition. Sarasota, Florida: Supply Chain Alicke, Knut, Jürgen Rachor, and Andreas Seyfert (2016), Supply Chain 4.0 Management Institute. – The Next-Generation Digital Supply Chain. McKinsey & Co. Supply Mandel, Michael (2017). How Ecommerce Creates Jobs and Reduces Chain Management Practice (June). Income Inequality. Washington, DC: Progressive Policy Institute, Arvis, Jean-Francois, Paul E. Kent, Ben Shepherd, and Rajiv Nair (2017), September. Additive Manufacturing and the Diffusion of 3D Printing: Impact on Manyika, James, Michael Chui, Peter Bisson, Jonathan Woetzel, Richard International Trade. Unpublished manuscript, World Bank, Washington, Dobbs, JacquesBughin, and Dan Aharon (2015), The Internet of Things: DC. Mapping the Value Beyond the Hype. McKinsey Global Institute 3. Asthana, Rahul (2018), Making Sense of Supply Chain 4.0. https://www. Microsoft Corporation (2011), Digital Infrastructure, Cloud Computing industryweek.com/supply-chain/making-sense-supply-chain-40 Transforming Fragmented Manufacturing Industry Value Chain Accord- (November 2). ing to Microsoft Study. Hannover, Germany. https:// news.microsoft. Bloom, Nicholas, Kalina Manova, John Van Reenen, Stephen Teng Sun, com/2011/04/03/digital-infrastructure-cloud-computing-transform- and Zhihong Yu (2018), Managing Trade: Evidence from China and ing-fragmented-manufacturing-industry-value-chain-according-to-mi- the United States. National Bureau of Economics Working Paper No. crosoft-study/. 24718, June. Niforos, Marina, Vijaya Ramachandran and Thomas Rehermann (2017), Ciera, Xavier, and William F. Maloney (2018), The Innovation Paradox: Blockchain: Opportunities for Private Enterprise in Emerging Markets. Developing Country Capabilities and the Unrealized Promise of Techno- Washington, DC: International Finance Corporation, World Bank Group. logical Catch-Up. Washington, DC: World Bank. Petersen, Moritz, Niels Hackius, and Birgit von See (2017), Mapping the Citigroup (2016), Technology at Work v2.0: The Future Is Not What It Used Sea of Opportunities: Blockchain in Supply Chain and Logistics. Work- to Be. Citi GPS: Global Perspectives & Solutions, Joint Report with ing Paper. Oxford Martin School. Pfohl, Hans-Christian, Burak Yahsi and Tamer Kurnaz (2017), The Impact of CGI (2017), Industry 4.0: Making Your Business More Competitive. CGI Industry 4.0 on the Supply Chain. Proceedings of the Hamburg Interna- Group Inc. tional Conference of Logistics (HICL) – 20. Dachs, Bernhard, Steffen Kinkel, and Angela Jäger (2017), Bringing it All PWC (2016a), Industry 4.0: Building the Digital Enterprise. PriceWater- Back Home? Backshoring of Manufacturing Activities and the Adoption houseCoopers LLP. of Industry 4.0 Technologies. PWC (2016b), Industry 4.0; How Digitization Makes the Supply Chain More De Backer, Koen, and Dorothee Flaig (2017). The Future of Global Value Efficient, Agile, and Customer-Focused. PriceWaterhouseCooper LLP. Chains. Rugaber, Christopher S. (2017), Robots and Automation Likely to Create Freund, Caroline, Alen Mulabdic (2018), Is 3-D Printing A Threat to Global More Jobs in E-Commerce. Associated Press, reprinted in Inc. Brand Trade? The Trade Affects You Didn’t Hear About. World Bank. Prelimi- View, October 30. nary: processed. Spence, Mike (2018), Jobs, Digital Technology and Structural change, GVC Freeman, Christopher and Luc Soete (1997), The Economics of Industrial Conference. Beijing, March 2018. Organization. 3rd Edition, London and New York: Routledge. UN/CEFACT – ITDP (United Nations Centre for Trade Facilitation and Elec- Garbee, Elizabeth (2016), This Is Not the Fourth Industrial Revolution. Slate. tronic Business) and World Economic Forum (2017), Paperless Trading: com, January 29, http://www.slate.com/articles/technology/future_ How Does It impact The Trade System?. UN/CEFACT – ITPD/Paperless tense/2016/01/the_world_economic_forum_is_wrong_this_isn_t_the_ Trade – Po051, Final Draft. fourth_industrial_revolution.html USITC (2011), The Economic Effects of Significant U.S. Import Restraints. Gereffi, Gary, John Humphrey, and Timothy Sturgeon (2005), The Gover- Seventh Update 2011. Special Topic: Global Supply Chain. Publication nance of Global Value Chains. Review of International Political Econ- 4253. Washington, DC: U.S. International Trade Commission. omy. 12(1), pp.78-104. Vrbová, Petra, Jiří Alina and Václav Cempirek (2016), Usage of EDI (Elec- Gregory, Jonathan (2015), The Internet of Things: Revolutionizing the Retail tronic Data Interchange) in the Czech Republic. Paper for the 10th Inter- Industry. AccentureStrategy. national Days of Statistics and Economics, Prague, September 8-10. Gordon, Robert J. (2017), The Rise and Fall of American Growth: The U.S. Weernink, Marissa Oude, Willem van den Engh, Mattia Fransisconi and Standard of Living Since the Civil War (The Princeton Economic History Frida Thorborg (2017), The Blockchain Potential for Port Logistics. of the Western World). Princeton: Princeton University Press. Hackius, Niels, and Petersen, Moritz (2017), Blockchain in Logistics and Supply Chain: Trick or Treat?. In Proceedings of the Hamburg Interna- tional Conference of Logistics (HICL), pp. 3-18, epubli. Hallward-Driemeier, Mary, and Gaurav Nayyar (2017). Trouble in the Making? The Future of Manufacturing-Led Development. Washington, DC: World Bank. Hiratsuka, Daisuke (2013), Vertical Intra-Regional Production Networks in East Asia: A Case Study of Hard Disk Drive Industry. IDE Working Paper. CHAPTER 6 The digital economy, GVCs and SMEs By Emmanuelle Ganne (WTO) and Kathryn Lundquist (WTO) ABSTRACT Although small and medium-sized enterprises (SMEs) repre - that this positive digital effect is greater for SMEs than it is sent the vast majority of firms worldwide, their participation for large firms. The chapter reviews the various opportuni- in international trade remains limited relative to their share ties that the digital economy opens for SMEs, especially in of overall economic activity and employment as compared to terms of cost reductions and the emergence of new busi- large firms. The rise of the digital economy could, however, ness models, but also discusses policy measures that could open a range of new opportunities for small firms to play a be taken to promote SME participation in GVCs. Indeed, sig - more active role in global value chains (GVCs). This chapter nificant challenges remain for SMEs to enter GVCs, some of reviews evidence of SME participation in international trade which are exacerbated by the new digital economy. A holis- and production networks and looks at how the digitalization tic approach that combines investment in ICT infrastructure of our economies is already affecting, or could affect future, and human capital with trade policy measures and measures SME contributions to GVCs. New research by Lanz et al. to improve the business environment, access to finance and (2018) finds evidence that digitally-connected SMEs in devel- logistics, and promote innovation and R&D is necessary. oping countries tend to import a higher share of their inputs Improving the availability of data would also help to better than non-digitally-connected firms. Additionally, it is shown understand and integrate SMEs in GVCs. • Although small and medium-sized enterprises (SMEs) represent the vast majority of firms worldwide, their participation in international trade remains limited relative to their share of overall economic activity and employment as compared to large firms. • The rise of the digital economy could, however, open a range of new opportunities for small firms to play a more active role in global value chains (GVCs). • New research finds that when a manufacturing SME has a website, this facilitates its participation in GVCs and trade. In particular, such SMEs are more likely to use foreign inputs for production and export their output. Further, information and communication technology (ICT) connectivity is found to be more important for small firms than for large ones in whether or not a firm participates in trade. • However, SMEs continue to face important challenges when integrating into GVCs. A holistic approach that combines investment in ICT infrastructure and human capital with trade policy measures and with measures to improve the business environment and access to finance and logistics, and promote innovation and R&D, is necessary. • Improving the availability of data would also help to better understand and integrate SMEs in GVCs. 121 122 • Technological innovation, supply chain trade, and workers in a globalized world 1. Introduction G The internet and digital technologies that leverage the inter- net to collect, store and process data, such as artificial intelli- lobal value chains (GVCs) are often considered the gence (AI), the Internet of Things (IOT) and blockchain, open new lead story of trade in the modern world, with an opportunities for SMEs, not only for market entry, but also for estimated 80 per cent of global trade taking place participation in GVCs and international trade (WTO, 2018). This is through them (UNCTAD, 2013). At the same time, particularly true in the services sector where SMEs are most likely a growing understanding of the importance of small and medi- to engage in trade (ABAC, 2018). um-sized enterprises (SMEs) to the global economy, and their Given the pervasiveness of SMEs throughout the global econ- roles within the digital economy, has been emerging. However, omy, the substantial role of GVCs for international trade and the SMEs have been shown to participate less in international trade, changes ICT is bringing through the new digital economy, further including GVC trade, than large businesses despite being the consideration ought to be given to how digital technology could largest firm segment by numbers in the world. Given the sub - be altering SME GVC participation. Firms of all sizes inherently stantial changes that the internet and digital technologies that seek to maximize profits, be it through the use of digital tech- leverage the internet to store and process data (sometimes nology or sales via international exports. Given the potential referred to as Industry 4.0) have made or are making to the for digital technology to reduce fixed trade costs, it is import- global economy, the following questions arise: how has the digi- ant to understand how technological change affects SME deci- tal economy changed the landscape for SMEs? What are the new sion-making with regards to both imports and exports. This chap - opportunities and challenges they face when it comes to partici- ter explores SME participation, and lack thereof, in international pating in GVCs in the digital era? And what policy changes could trade, including GVCs; discusses how digital technologies can be made to support these firms? help SMEs integrate into GVCs; considers the various constraints SMEs are estimated to account for between 80-99 per cent that restrict SMEs’ ability to embrace new technologies and par- of firms in any given country as well as between 60-70 per cent ticipate in global production networks; investigates how the digi- tal economy has re-shaped international trade for SMEs as well as of global employment (WTO, 2016; IFC, 2013). They also have a its potential effects on SMEs in GVCs; and lastly looks into ways higher rate of sales growth than large firms (Cusolito et al., 2016). the policy environment could be changed to better support SME This implies a substantial share of any nation’s economy is sup- access to GVCs in the context of the digital economy. ported by SMEs. However, a more precise estimate of SMEs’ con- tribution to GDP is hampered by the lack of a standard definition for what, exactly, constitutes an SME. Definitions for small firms 2. SME participation in international trade and range from those solely based on number of employees and rev- GVCs enue generated (the European Union defines SMEs as firms with up to 250 employees and turnover of no more than 50 million In theory, global value chains open new prospects for SMEs to euros), to one dependent on the industry of operation (in China, participate in international trade. The international fragmenta- SMEs can include firms of up to 3,000 employees and total rev- tion of production increases the opportunities for SMEs to spe - enues up to 300 million yen, depending on the industry).1 These cialize in niche markets and narrow activities at various stages of differences in definition make certain comparisons more chal- the production chain. Nevertheless, in spite of the key economic lenging and must be considered when drawing conclusions. role played by SMEs in terms of economic output, participation Regardless of the nebulous way SMEs are defined, they are of SMEs in global value chains remains low compared to that of not well represented in international trade and GVCs (WTO, large firms. 2016). This is in spite of the fact that the international fragmen- tation of production would seem to have increased the oppor- 2.1 SMEs, international trade and GVCs: direct vs. indirect tunities for SMEs, given that production is broken into smaller, participation more specialized pieces. Yet SMEs face a number of size-related SMEs can join global value chains by exporting intermediate constraints, from limitations related to quantity of production, goods or services directly (direct forward participation) or by sup- to in-house administrative resources, that prevent many of them plying inputs to a local firm or multinational company – indirect from achieving the full potential of GVC participation (Cusolito et exports (indirect forward participation). These forms of integra- al., 2016). tion into GVCs are not necessarily exclusive. Some SMEs export Given the positive effects GVCs have been shown to bring, both directly and indirectly, highlighting the potential comple- it is worth considering how to include more small firms in global mentarity of these foreign market entry modes (Nguyen et al., production networks. For example, participation in GVCs is asso - 2012). SMEs can also participate in GVCs by importing products ciated with increased productivity, the export of more sophisti- as inputs into their own production processes (direct backward cated (and frequently higher value) products, and a more diver- participation) or sourcing products from local firms that use sified national export basket. Additionally, GVCs have been imported inputs (see Figure 6.1).2 Forward linkages represent the demonstrated to be a pathway for economic development for seller’s perspective, or supply side, while backward linkages rep- countries (Kowalski et al., 2015). resent the buyer’s perspective, or sourcing side, of GVCs. The digital economy, GVCs and SMEs • 123 FIGURE 6.1 How SMEs can benefit from GVCs of SMEs, etc. In addition, while GVC trade is usually understood as trade in intermediates, available data sets do not necessarily distinguish between direct exports of final products and direct exports of intermediates. An analysis of data on gross direct Direct or indirect Direct or indirect backward participation forward participation exports can, however, provide some indication of SME forward participation in global value chains as such exports, which cover = direct imports of inputs = direct exports of intermediate both final and intermediate products, necessarily represent an or indirect imports through products or indirect exports upper bound. a domestic firm through a domestic firm Keeping these limitations in mind, firm-level evidence reveals that despite SMEs making up the vast majority of firms in both Wider access to: Possibility to focus on developed and developing countries, SME direct and indirect • more sophisticated and specific segments of the competitively priced production chain (no need participation in GVCs remains limited relative to their share of imports to master the entire overall activity and employment compared to large firms. • new technologies production process). • inputs that may not be Direct participation in GVCs: a “big firm story”? accessible domestically While in most OECD economies SMEs account for 99 per Tech transfers from lead firms cent of all firms, around two-thirds of total employment and over half of business sector value-added, their contribution to over- all exports is much lower than their economic weight in terms of Source: Adapted from López González (2017). value creation and employment, with only a handful of excep - tions (OECD, 2018c) (see Figure 6.2). In countries such as France, Germany, Slovakia and Sweden, SMEs account for only 30 to 40 The extent to which SMEs participate in GVCs is, however, dif- per cent of gross exports, well below their contribution to value ficult to assess thoroughly. The availability of international trade creation and employment. Not only is SME participation in direct data by enterprise size remains limited, making analysis rather exports low compared to their economic weight, only a fraction difficult and often partial. Most studies rely on a mix of enter- of SMEs export at all. This is a distinct difference between large prise surveys, case studies, and administrative data, with all the and small firms, given that the majority of large businesses are compromises that such approaches entail in terms of incomplete also international exporters. Evidence from OECD countries country coverage, different time series, inconsistent definitions shows that only 5 to 40 per cent of SMEs export, while more FIGURE 6.2 SME export activity, value added and employment shares (2015 or latest available year, as a percentage) (Percentage) % of exports % of value added % of business employment 100 90 80 70 60 50 40 30 20 10 0 A U T RV X D P LD N T L EU R BR UN R A K N N K RC R R R L A E KOX E H L U L G T E BE ES ES PR LU IR TU IT IS PO LT IS N O SV BG CH LV AU CZ SW E FR SV FI RO H N G D M Source: Figure 1.20 of OECD (2018c). 124 • Technological innovation, supply chain trade, and workers in a globalized world FIGURE 6.3 Industrial firms engaged in exports (2015 or latest available year, as a percentage of total firm size by size class)  100 90 80 70 60 50 40 30 20 10 0 L A L L A A A LD R BR N RC UN R E N EU P U T U K N E RV X K T T IR BE PO LV BG IT FR US O SW ES AU N CZ RO PR LT SV LU SV ES FI CA G N G D H N H D SMEs Large enterprises Source: Figure 1.21 of OECD (2018c). than 80 per cent of large firms do (see Figure 6.3). Other studies ranging from an estimated 22 per cent in the Middle East to 70 confirm these numbers. Mayer and Ottaviano (2007) showed, for per cent in some African countries (ITC, 2015a), SMEs’ exports example, that 1, 5, and 10 per cent of companies account for no are significantly less. For example, SME exports account for 28 less than 40, 70, and 80 per cent, respectively, of Europe’s aggre- per cent of overall exports in developing Europe, 16 per cent in gate exports. These numbers would tend to support the view that the Middle East, 8.7 per cent in developing Asia, and only 3 per direct insertion into GVCs via exports is “a big firm story” (Cusolito cent in Africa. As in the case of developed economies, big firms et al., 2016). account for the bulk of exports. Cebeci et al. (2012) find that the However, these numbers hide considerable heterogene- top 5 per cent of firms account, on average, for 80 per cent of ity across firm size classes. The smaller the company, the less exports in low-income countries. export-oriented it is (see Figure 6.4). Only a marginal number of However, while GVC direct participation would seem to micro companies export, while the participation of medium-sized be above all a big firm story when considering gross exports, companies in exports and imports approaches that of large busi- studies that examine exports of intermediates seem to show a nesses. Participation in exports remains, to a large extent, a big more nuanced picture. Evidence from Southeast Asia reveals, firm story in developed economies, except in some niche markets.3 for example, that SME exports of intermediates in Thailand The situation is not much different in developing economies, represent a bigger share of their overall exports than for large with rough estimates of SME contribution to GDP significantly firms – 16 per cent of SME exports are sold to firms abroad for larger than their relative contribution to international trade, and further processing, while only 6 per cent of large firms’ exports estimates of SME contribution to international trade being only a are (López González, 2017). This finding reflects the opportuni- fraction of large firms’ contribution. SMEs in developing countries ties that global value chains open for SMEs to integrate into the are thought to provide about 45 per cent on average of a coun- global economy by specializing in segments of production and try’s GDP (WTO, 2016), but SMEs’ exports represent on average supply of intermediates, rather than having to master the entire just 7.6 per cent of total manufacturing sales, compared to 14.1 per production process of finished products. Opportunities in this cent in the case of large manufacturing firms (WTO, 2016).4 Recent respect might be even bigger in the services sector. In Viet Nam, World Bank micro firm surveys in selected least developed coun- for example, the share of SME exports used by other countries to tries (LDCs) confirm the low level of participation of micro firms (i.e., produce other exports increases from 5 per cent when only man- firms of less than five employees) in international trade. Micro firms ufacturing is considered, to 26 per cent when service firms are engaged in exports represented only 6 per cent of surveyed firms included (López González, 2017). While these numbers cannot in Congo in 2013, around 3 per cent in Ethiopia (2011 data), and be generalized, they provide an interesting new perspective on less than one per cent in Myanmar (2014 data). SME GVC participation in Southeast Asia. SMEs’ contribution to GDP and exports also varies signifi- Another way for SMEs to benefit from GVCs is through cantly across developing regions. Although SME contribution imports of intermediate goods (backward participation), which to GDP is estimated to be relatively high throughout the world, matter for competitiveness (Lopez-Gonzalez, 2016 and 2017). It The digital economy, GVCs and SMEs • 125 FIGURE 6.4 Percentage of industrial firms that are exporting and importing by enterprise size (number of employees), 2013 or latest year  80 70 70 66 62 58 60 50 38 40 40 30 20 20 17 12 9 10 0 Total 0-9 10-49 50-249 250+ Exports Imports Source: WTO (2016). has been shown that firms that use more imported products are 2016). A simpler route for SMEs to engage in GVCs is often to more productive as they can draw on cheaper and more sophis- start by exporting indirectly, through a local firm. ticated inputs as well as benefit from innovation and new tech- nologies embodied in imports (Bas and Strauss-Kahn, 2014 and 2.2 Indirect participation in GVCs 2015). According to WTO estimates, GVC participation by SMEs Smaller firms often participate in global value chains indirectly in the manufacturing sector in developing countries is mainly by supplying intermediates to other local firms – domestic or driven by upstream links (backward participation), with SMEs foreign-owned – that export (indirect forward participation). importing inputs needed in their manufacturing process from The enterprise then behaves like an “indirect exporter” by con - abroad (Lanz et al., 2018; WTO, 2016). This is particularly true in tributing to the production of goods and services exported by countries where companies engage extensively in processing other domestic companies. Likewise, the fixed costs associated trade. Processing trade allows a firm to conduct intermediate with direct importing may lead many SMEs to source inputs stages of production and assembly on behalf of a foreign party. from local enterprises that use imported products (indirect The firm receives the blueprints and imports all, or part of, the backward participation). Evidence on indirect participation of raw and auxiliary materials, parts and components, accessories, SMEs in GVCs is scarce and difficult to collect due to lack of and packaging materials from abroad, and re-exports the fin- data on value-added at the firm level. Only a few studies have ished products after processing or assembly. Engaging in pro- examined SME indirect participation in GVCs, either as suppli - cessing trade requires less technological know-how and working ers (forward participation) or as importers of inputs (backward capital needs – although it may require having certain automated participation). processes in place to ensure quality control and supply reliability, Studies that analyze the role of SMEs as suppliers reveal which may or may not be borne by the foreign party. Evidence that focusing only on direct exports significantly underesti - from China shows that processing trade allows less productive mates the role played by SMEs in GVCs. In an often-quoted and financially constrained firms to participate in GVCs when study, Slaughter (2013) showed, for example, that US multina - they would not have been able to otherwise (Manova and Yu, tional companies typically purchase more than US$3 billion 2016). Such firms tend to be SMEs. in inputs a year from more than 6,000 U.S. SMEs, which rep - Among the factors often put forward to explain why SMEs’ resents almost 25 per cent of the total inputs purchased by direct participation in GVCs is lower compared to that of large those firms. Other estimates from the US International Trade firms is the fact that engaging in international markets can be Commission (USITC) (2010) find that in 2007 the export share costly. Lacking economies of scale, SMEs face higher fixed costs of US SMEs rose from 28 per cent (in gross exports) to more than larger companies and are disproportionately affected by than 40 per cent (in value-added terms) when indirect exports costs associated with the import and export process (WTO, were considered. Calculations using the TiVA database 126 • Technological innovation, supply chain trade, and workers in a globalized world developed by the OECD and the World Trade Organization to direct exports (6.5 per cent vs 2.2 per cent). However, despite show that including the contribution of upstream SME suppli - there being more SMEs that engage in indirect exports, overall ers significantly increases the share of SMEs in total exports SME participation in GVCs is small and they remain largely dom- of domestic value added. In the Slovak Republic, for example, inated by large companies. In the case of Chile, the gap is strik- SMEs account for only 34 per cent of gross domestic exports, ing: more than 46 per cent of large companies engage in direct but for 56 per cent of the total value added in the country’s exports, while only 9 per cent of SMEs export, including both exports when upstream suppliers are considered (OECD, direct and indirect exports. The situation at the product level 2018c) (see Figure 6.5). varies, however. Indeed, the share of indirect exports of SMEs in Indirect exports of SMEs are particularly significant in sec- total sales outpaces that of large firms in some specific sectors, tors where GVCs play an important role and where scale such as certain types of machinery, publishing and printing, and matters, such as in the automobile and transport equipment in paper and paper products manufacturing (WTO, 2016). Ser- manufacturing sector (OECD, 2018b; WTO, 2016), and for inde - vices SMEs were also found to participate more in indirect exports pendent SMEs (i.e., those not owned by a larger domestic firm than direct exports. Overall, however, backward and forward GVC or foreign firm – OECD, 2018c). Evidence shows that SMEs tend participation of SMEs in developing countries remains low (see to channel their indirect exports through large firms rather than Figure 6.6). through other SMEs (Cusolito et al., 2016). The role played by indirect forward participation of SMEs, espe- While evidence based on indirect exports shows a higher cially in developed countries, would tend to suggest that indirect level of integration of SMEs in GVCs in OECD countries, indi - participation serves to a certain extent as a substitute for direct rect exports appear to play a lesser role in developing coun - participation in GVCs. The question then arises as to whether such tries. Using data from World Bank Enterprise Surveys, the WTO indirect participation benefits firms and impacts their performance estimated that indirect exports of manufacturing SMEs from in the same way as direct participation. Assessing the relative developing countries were 2.4 per cent of total sales on aver- impact of direct versus indirect participation on firms’ performance age, or one-third the estimated share of direct exports. Such is an issue that requires further attention from researchers. Like- results, however, hide significant differences across regions, wise, studies that distinguish between direct and indirect partici- within regions, and at the product level. While SME indirect pation usually focus on exports. It would be equally interesting to participation in exports is estimated at more than 9 per cent of examine SMEs’ participation in GVCs through indirect backward total sales in developing Europe, it accounts for 2.4 per cent in participation. Indeed, like in the case of direct exports, the high the Middle East and only 1 per cent in Africa (WTO, 2016). At fixed costs associated with direct imports may lead many SMEs to the country level, a recent study carried out in Chile reveals that source inputs from local companies that use imported intermedi- three times as many SMEs engage in indirect exports compared ates rather than to import directly. FIGURE 6.5 Direct and indirect exporting activity of SMEs in OECD countries, 2014 As a percentage of gross export 70% 60% 50% 40% 30% 20% 10% 0% AUT BEL CZE DEU DNK ESP FIN FRA GBR HUN ITA MEX NLD POL PRT ROU SVK SVN SWE TUR USA SMEs' share in gross exports SMEs' share in value added exports Source: OECD (2018a). The digital economy, GVCs and SMEs • 127 FIGURE 6.6 SMEs in developing economies: backward and forward participation in GVCs 100 90 High forward GVC participation participation High GVC 80 Direct and indirect exports (Forward participation) 70 60 50 40 High backward GVC participation participation Low GVC 30 20 10 0 0 10 20 30 40 50 60 70 80 90 100 Use of foreign inputs in manufacturing (backward participation) Note: Each square represents the average GVC participation of SMEs in a given developing country. Source: WTO estimates based on World Bank Enterprise Surveys (Figure B.19 of WTO (2016)). 2.3 Constraints on SME participation in global value chains Various factors have been found to play a determining role Two key challenges persist in limiting SME GVC integration: the in explaining the size of the informal sector, including the tax challenge of informality5 and the relative resource advantage burden (e.g. Cebula, 1997; Giles and Tedds, 2002); financial that large firms have over SMEs. market development (Straub, 2005); and institutional quality, Global value chains operate in the formal sector, but it is esti- regulatory burden and quality of the legal system (Friedman et mated that around 80 to 90 per cent of SMEs worldwide are infor- al., 2000; Johnson et al. 2000; Botero et al., 2004; Dabla-Norris mal (IFC, 2012). Informality is pervasive in the developing world. et al., 2005). The majority of firms in many developing countries are informal High levels of informality can affect growth and productiv- (Andrade et al. 2015; Bruhn and McKenzie 2014; Cusolito et al. ity of a country and hold back inclusion into GVCs. Informality 2016). In Brazil, for example, nearly two-thirds of businesses, 40 can generate inefficiencies in the production process, as infor- per cent of GDP and 35 per cent of employees are informal (Ulys- mal firms may choose to limit their growth to avoid detection seay, 2015). Similarly, in Sri Lanka only one-fifth of firms operat- (Farazi, 2014) and tend to use less advanced production technol- ing without paid workers are registered and even among firms ogies (Perry et al., 2007). Corruption is also often a side-effect employing paid workers, more than half are unregistered with of informality, and even where it is not, recent work looking at one or more pertinent agencies (de Mel, McKenzie, and Wood - Chinese firms has shown corruption to have a larger negative ruff 2013). Overall, the ILO estimates that the informal economy impact on the productivity of SMEs than on the productivity of comprises more than half of the global labor force (ILO web- large firms (Lu et al., 2018). Informal firms also face greater diffi- site6), with most informal workers in developing countries being culties accessing finance, which can result in sub-optimal levels women.7 Informal firms tend to be much smaller than formal firms of investment in research and development, physical capital, (La Porta and Shleifer, 2014). In fact, the large majority of infor- and training (Farazi, 2014). Informality is a binding constraint to mal firms – up to 90 per cent in Sri Lanka for example – are small, integrating into global value chains, but it is also a constraint for subsistence enterprises with no paid employees. firms operating in the formal sector. A study by the Independent 128 • Technological innovation, supply chain trade, and workers in a globalized world Evaluation Group, a World Bank institution, finds that about 32 per being digitally connected with GVC participation. Evidence backs cent of formal firms with 10–99 employees in a sample of develop- the theory that these digital changes can support SME participa- ing countries report informality as one of the top five constraints tion in GVCs, particularly import-based (backward-linked) GVCs. they face in doing business (cited in Cusolito et al., 2016). This is an important insight given that limited SME participation Other factors commonly cited to explain the difficulties faced by in GVCs continues to restrain participation in international trade. SMEs, in particular those from developing countries, to integrate However, the importance of the divide between firms with access in global value chains range from limited resources and access to to the internet and those without is underscored by this research. finance, to the difficulty some SMEs face in meeting product and Using World Bank Enterprise Survey data, the authors demon- quality standards. Noted constraints that affect SMEs include lack strate that, for firms, having a website (a proxy for being ICT-en- of knowledge about foreign markets as well as missing in-house abled) in a developing country has a larger predicted impact on skills such as marketing; insufficient knowledge of cumbersome both an SME’s share of imported inputs for production and on trade regulations and border procedures; and poor physical and an SME’s share of direct exports, than it does for large firms. An ICT infrastructure that limits distribution and operational support ICT-enabled small firm of 2 employees would have a predicted (ADB, 2015; Cusolito et al., 2016; ITC, 2015b; WTO, 2016). share of imported inputs that is 10 percentage points higher Can the rise of technologies based on the internet and the than a firm of the same size that is not ICT-enabled. Similarly, a remodeling of economic activity that accompanies it open new firm of 12 employees would have a predicted share 8 percentage opportunities for smaller firms to more actively participate in points higher. This is significantly greater than the estimated dif- global production networks? Can the digital economy help small ference for larger firms. In the case of firms of 50 employees, the traders integrate into global value chains? Evidence suggests that predicted effect of being ICT-enabled on the share of imported the potential could be significant. inputs, versus for firms that are not connected, is only 5 percent- age points and for firms of 100 employees it is only 3 percent- age points (see Figure 6.7). For total exports, the effect of being 3. Digital technologies can boost SME trade ICT-enabled is highest for firms with between 15 and 25 employ- and GVC participation ees, with a steep decline as the number of employees grows (see Figure 6.8). In both cases, being ICT-enabled shows a stronger Digital technologies continue to make substantial changes to result for SMEs’ participation in trade than for large firms, mean- the economy with cascading implications for international trade. ing the impact of being digitally-enabled is significantly greater For small firms, the internet has increased access to international for small firms than for large ones. This is in line with evidence markets, with the WTO finding that on average 97 per cent of that small businesses with a website were almost four times more internet-enabled small businesses export (WTO, 2016). Compa- likely to export than those without (Oxford Economics, 2017). nies also acknowledge the importance of new internet-enabled The study also considers country-level digital connectivity and technologies. For example, a study of 600 European SMEs found its effects on participation in trade by firm size. Using the number that more than 70 per cent of those surveyed not only consider of fixed broadband subscriptions in a country to proxy digital con- that they benefit from the ongoing process of digitalization, but nectedness, the paper again demonstrates that for developing also that digitalization makes it easier to integrate foreign cus- countries, increased digital connectivity seems to increase small tomers and suppliers into their own value chains (Abel-Koch, firms’ share of imported inputs used for production more than for 2016). Additionally, a joint OECD and World Bank study (Cusolito large firms. Or, in other words, a small firm’s participation in back- et al., 2016) finds that the use of the internet reduces SME export- ward-linked GVCs will benefit more than a large firm’s if a country ing costs, thereby increasing export participation, and that SMEs has better digital connectivity. Similarly, for total exports, the find- are more likely to be involved in technologically-enabled trade ings suggest that more broadband subscriptions at the country than traditional trade. At the same time, there are also many level leads to a greater positive effect on SME exports than for factors limiting SME participation in GVCs in the context of the large firms. These findings imply that large firms have established digital economy. For example, it has become clear that internet other non-ICT enabled means of communication with overseas access is now often a requirement for joining many GVCs (ADB, suppliers and customers, such as analogue telephones or in-per- 2015) and that the ICT level of operation is one of the key attri- son traveling, that might not be so easily available to SMEs. butes that multinational corporations assess when they want to enter a business relationship with SMEs (APEC, 2014). However, 3.2 How can digital technology support SME trade? few studies have looked directly at the impact the new digital- There are many reasons why access to digital technologies can ly-based economy is having on SME participation in GVCs. increase SMEs’ participation in trade. Internet access can reduce barriers and costs to trade for all firms (but especially for ser- 3.1 The impact of digital connectivity on SME GVC vices SMEs (Cusolito et al., 2016)) as well as increase access to participation foreign markets through online sales and e-commerce. The rise Recent work by Lanz et al. (2018) has looked more closely into the of smartphones has also allowed leapfrogging of some capital- differences between ICT-enabled SMEs and large firms in devel- and/or infrastructure-intensive technology, especially by firms oping countries with regards to trade, as well as the relationship of in developing countries. Additionally, the digital economy itself The digital economy, GVCs and SMEs • 129 FIGURE 6.7 Effect of being ICT-enabled on the predicted FIGURE 6.8 Effect of being ICT-enabled on the predicted share of imported inputs for production share of exports out of total sales 15% 8% Predicted effect on the share of imported inputs Predicted effect on the share of total exports 6% 10% 4% 5% 2% 0 0 2 12 50 100 2 5 15 25 45 100 Number of employees Number of employees Note: Figure marks represent the estimated range for the predicted effect of Note: Figure marks represent the estimated range for the predicted effect of being ICT-enabled on a firm’s imports for inputs. being ICT-enabled on a firm’s total exports. Source: WTO estimates based on WB Enterprise Survey Data. Source: WTO estimates based on WB Enterprise Survey Data. is creating new opportunities by increasing the number of par- to AMTC (2018). Online reviews can also be a powerful tool to ticipants in international trade, as well as creating new business attract potential customers from anywhere in the world (Oxford models that affect the structure of supply chains, including being Economics, 2017) and new adwords, or other targeted advertis- “born global.” For example, there are new services on offer, ing, can help firms with limited resources reach new consumers including programming or logistics, that require only the neces- (AMTC, 2018). In fact, marketing for both SME manufacturing and sary technical skills and being ICT-enabled (ITC, 2015a). Online services firms is projected to have the largest savings in export sales are also making smaller “just-in-time” purchases more related costs in a digital environment (AMTC, 2018). common than large advanced orders, a development that could Additionally, online and mobile banking or finance (which may benefit SMEs (AliResearch, 2017). Only SMEs with resources and even be provided through e-commerce platforms), as well as new managers willing to adopt these new technologies are in the posi- financing tools like online crowdfunding, can supplement tradi- tion to take advantage of these opportunities (ITC, 2015a). All of tional finance for SMEs. For example, services such as Alibaba’s these topics will be explored further in this section. e-Credit Line, or IndiaMart’s Payment Protection insurance, can be important trade finance resources for small companies. Block- Cost-reducing properties of digital technology chain could also open new opportunities for SMEs to access trade Entering international markets is difficult and costly, dispropor- finance by making it easier for small companies to build a credit tionally affecting small firms that face a host of constraints as history as well as by opening up the possibility for small firms and discussed earlier, including higher relative fixed costs than larger producers to make transactions on a peer-to-peer basis without companies, insufficient R&D and skills training, and insufficient the need to secure traditional trade finance or even to go through knowledge of foreign markets and regulations. However, digital banks (Ganne, 2018). technologies can ease a number of these constraints and reduce Besides reducing financial costs, online access to information SME expenditures in a range of areas, from market research to also has significant time saving benefits by reducing the need operational support (see Figure 6.9). New websites and digital for some in-person interactions, such as with banking. This has processing tools can bring services to SMEs that were formerly been shown to save SMEs up to 29 per cent of the time previously unaffordable. The following will look more closely at the ways dig- required (AMTC, 2018). Related to time saving, regional SME net- ital technology supports SMEs. working platforms have also been created to bring information In terms of market research and general marketing, the inter- together in one place and to facilitate networking among SME net provides access to a wide variety of information, including suppliers (see Box 6.1). To expand these benefits, the World SME information related to potential consumers or national regulatory Forum has proposed plans to create eWSF,8 a global equivalent compliance and how to trade across borders. Online marketing to regional networks such as ConnectAmericas. This can result in has also been shown to be important for SMEs, with digital access significant savings related to export activities and benefit SMEs reducing estimated marketing costs by 57 per cent according in the international market. 130 • Technological innovation, supply chain trade, and workers in a globalized world FIGURE 6.9 Ways the digital economy can reduce SME business costs  Export value chain Impact of digital Details Traditional scenario Digital scenario • Identification and quantification • Labour intensive: dedicated • Desktop research Market of foreign business opportunities staff, market research agency, • Digital market research tools research • Obtaining information and a rigorous potential field trip (e.g. online surveys) understanding of the target market • Potential travel to market • Reduced need for travel • Targeting of customers in the foreign • Procurement of local • Digital advertising channels market through advertising advertising space in foreign, (search engine optimisation, Marketing • Dissemination of promotional material market (e.g. newspaper, radio display, social, video) through various advertising channels & TV ads) • Leveraging market platforms • Access to product shipment insurance for • Limited transparency • Product comparison sites Insurance and securing funding for export ventures • Time intensive paper based • Single window view into and • Obtaining information on and procuring approach market to insurance and securing loans • Dedicated brokers • Digital financial products financing • Regulation, rules and laws in the foreign • Time intensive paper based • National single window market the MSME has to comply with Regulatory approach • Costs of complying with foreign regulation • Dedicated consultant compliance such as filing documents and legal costs • Physical delivery of goods to the foreign • Manual management of • Automated and digitalised market supply chains supply chain management Distribution • Product delivery and channels through • Limited information on causes (e.g. Internet of Things) which sales occur of inefficiencies • Day to day operations of the business e.g. • Special IT equipment (e.g. • Cloud computing and Operational processing orders, back offices tasks servers, office software) software support • IT heavy tasks such as database • Communication services • Voice over IP management, accounting, communication • Dedicated travel agents • Online travel services Source: Alphabeta framework (from AMTC, 2018). BOX 6.1 ConnectAmericas, an online network for businesses in the Americas Online networks for businesses are an important tool pro- partner from Curaçao regarding a potential business rela- vided through the internet for SMEs to connect to inter- tionship. Following a Skype conversation, they agreed to national markets. ConnectAmericas, a business network work together, with the new partner in Curaçao using his initiative by the Inter-American Development Bank (IDB) established business to actively promote and advertise Mr. with the support of Google, DHL, Visa and Alibaba, seeks Olivares’ training services. to promote international trade and investment by SMEs in GlamST was created by two telematics engineers, Caro- the Americas through its platform. Two examples illustrate lina Bañales and Augustina Sartori, to enhance the customer its usefulness to small digital businesses working to enter experience, both online and in-store, for retail cosmetics international markets. The first is Rodrigo Olivares and brands through a virtual makeup application they devel- his online engineering training services and the second is oped. ConnectAmericas provided GlamST with a way to GlamST, a virtual makeup application founded by Carolina research and verify potential business clients for the app. Bañales and Augustina Sartori. Further, Ms. Bañales noted that ConnectAmericas provides After registering on ConnectAmericas, Mr. Olivares resources via their platform for accessing start-up capital as quickly received verification of his company by Connect- well as client and product development tools. Americas. Mr. Olivares next indicated his desire to expand his training services beyond his Chilean base. Within a Source: https://connectamericas.com/video/rodrigo-olivares-did-business short amount of time he was contacted online by his now -3-days-thanks-connectamericas The digital economy, GVCs and SMEs • 131 Digital technologies can also help reduce regulatory compli- intermediaries by reducing the costs of international trade, espe- ance costs by making information available online. For example, cially with regards to logistics services. government tax compliance regulations or export requirements At a basic level, digital technology has been crucial in lowering can now often be found on the internet and necessary informa- the cost of operational support needed for business generally, but tion can sometimes be submitted via e-documentation. This is especially for cross-border initiatives. Email, voice over internet pro- important for SMEs, a majority of which were found to outsource tocol (VOIP) systems and online video conferencing now mean that customs-related regulatory compliance in a recent ITC survey firms can be in touch at reasonable cost, especially internationally. (ITC, 2017a). Recognizing the potential for the internet to facilitate Further, the use of machine learning to provide real-time translation SMEs’ access to information for international trade, the ITC with is also bringing down language barriers. the WTO and UNCTAD has developed the Global Trade Helpdesk Altogether, these reduced business and trade costs have the (GTH) as a one-stop shop (see Box 6.2). New technologies like potential to be relatively more beneficial for SMEs, especially SMEs in blockchain can also contribute to greater transparency, making it the services sector, than for large firms with regards to international easier to trace supply chains and prove product origins (Ganne, markets. This is even more true for SMEs in developing countries 2018). It is estimated that manufacturing SMEs can see as much where the relative burden has been noted to be the highest (WTO, as a 40 per cent reduction in compliance costs and a halving of 2018). In fact, it is estimated that digital technologies can lower SME the time required to comply thanks to digital technologies, while export costs by as much as 82 per cent and reduce foreign market services SMEs can see the costs eliminated entirely (AMTC, 2018). operating costs by up to 59 per cent (AMTC, 2018). Digital technolo- Other cost reducing services are available as well, particularly with gies have lowered the cost to internationalize, thereby widening the regard to distribution services. Digital logistics that leverage IoT and scope for SME participation in international trade and GVCs (OECD, artificial intelligence now permit much closer tracking of shipments 2018b; WTO, 2018). It is estimated that the rise of digital technol- and inventory, allowing firms to better assess their production and ogies such as IoT, artificial intelligence, 3D printing and blockchain demand (AMTC, 2018; WTR, 2018. See also chapter 5). Additionally, could lower trade costs by another 10.5 per cent over the next 15 recruitment websites make it easier to list and fill vacancies and price years, with such decline especially benefiting SMEs and firms from comparison sites can significantly reduce firm expenditures on their developing countries, provided appropriate complementary policies required goods and services. Cloud technologies can also reduce a are put in place (WTR, 2018). firm’s expenses on hardware, software, web hosting and the associ- ated administrative costs (AMTC, 2018). These services are used by E-commerce as an enabler or alternative to GVC participation “lean start-ups” to lower their fixed costs, thereby increasing com- Access to online sales platforms has been a very important petitiveness in the fast-changing digital environment (OECD, 2017a). development for SMEs, especially as it relates to GVCs and inter- Related to distribution services, studies have also shown that national supply. Lendle et al. (2014) shows, in a sample of 18 small and financially constrained firms rely heavily on intermedia- countries, that between 88 to 100 per cent of eBay sellers are tion services (Ahn, 2011; Chan, 2015) and often do not have direct merchandise exporters, compared to only ten per cent of small export market access (Felbermayr, 2011). This is especially true firms operating through traditional non-platform methods. Fur- when market access costs are high, leading to trade intermedi- ther, SMEs participating in e-commerce tend to remain export- ation services taking a larger revenue share from exporters than ers longer than those in purely traditional markets (ITC, 2016) would be predicted by standard trade models (Schroder, 2003). and growth of e-commerce yields productivity gains of 6 to 15 Digital technology can work to reduce the distortion posed by per cent for SMEs (ABAC, 2018). Although SMEs with access to BOX 6.2 The Global Trade Helpdesk, international trade information in one location The Global Trade Helpdesk (GTH) is a joint ITC, UNCTAD, non-tariff measures, and notifications of WTO Members; WTO initiative that aims to improve the quality, transparency export/import procedures (e.g. pre-shipment formalities, and accessibility of trade-related information by providing a certification and inspection processes, transport documents); unique entry point to existing trade-related information. The business opportunities (market prices, company directory, GTH specifically targets SMEs who often do not have the upcoming events); and policy outlook (trade statistics, export resources to access fee-based information. potential analysis, trade agreements). The beta version of the GTH was launched at the 11th In the coming years, the GTH will be translated to all six WTO Ministerial Conference in 2017. The GTH integrates official United Nations languages to be accessible to people comprehensive information from various sources on market around the world. requirements, including customs tariffs, taxes, rules of origin, Source: www.helpmetrade.org 132 • Technological innovation, supply chain trade, and workers in a globalized world e-commerce may not immediately participate in GVCs, SMEs generally are also areas that SMEs can take advantage of. SMEs can often enter international trade and supply chains as e-commerce join GVCs as independent service contractors for digital products importers before becoming exporters and suppliers themselves that may be exported indirectly over the web. Besides online and (cited in Lanz et al., 2018). mobile apps, online content creators in general have also sprung In general, small firms that use e-commerce also have more up as ways small businesses, even individuals, are employed. The access to international markets and supply chain demand. For employment share of SMEs in the ICT sector in OECD countries example, SMEs are able to reach 30 different economies on grew from 3.8 per cent to 4.7 per cent between 2010 and 2016, average using online platforms (ARTNeT, 2018). On top of this, and SMEs’ share of value added in this sector increased in nearly all e-commerce has changed supply chains more fundamentally OECD countries, with the most substantial increases in publishing via the noted trend for firms to retain a reduced inventory and activities and telecommunications (OECD, 2018c). instead purchase high frequency, small volume shipments online The opportunities opened by digital technologies are multi- (see chapter 5 and AliResearch, 2017). SMEs might be able to take faceted, and some studies estimate that digitalizing MSMEs is advantage of this change given their potential for greater busi- the largest contributor to kick-starting virtuous cycles, especially ness agility and the evolving need for smaller quantities. for firms engaging in cross-border trade (ABAC, 2018). An estimated 90 per cent of e-commerce transactions are B2B (ITC, 2017a), thereby implying underlying value chain transac- 3.3 Digital challenges for SMEs to enter GVCs tions. Although the majority of e-commerce consists of domestic The digital economy, and ICT generally, are significant enablers transactions, cross-border retail e-commerce is expected to grow for SME participation in GVCs. However, challenges related to at twice the rate of domestic e-commerce, potentially boosting SME participation remain. SMEs lag in terms of digital technology international trade (ARTNeT, 2018). E-commerce and digital plat- adoption for a variety of reasons, from cost of implementation to forms have also been crucial for facilitating international trade by management (OECD, 2017a). Moreover, large firms have imple- SMEs, and e-commerce in general is becoming increasingly inter- mented a wide range of technologies with a diverse set of require- national (ITC, 2017a). Thus, e-commerce can be an alternative to ments (see chapter 5 for a discussion on company use of advanced participation in GVCs by SMEs through multinational companies. supply management techniques). As a result, the ability to interact Depending on the business model employed by a given producer effectively with these technologies has now frequently become a or manufacturer, SMEs may search online for inputs meeting their precondition in big industries for other firms to become suppliers, criteria rather than having a formal agreement with a supplier, thereby becoming a requirement for certain types of GVC par- thereby creating potential opportunities for firms of any size. ticipation and potentially excluding some suppliers (see Box 6.3). Online sales platforms, and e-commerce generally, have also Without the required capital and skills, firms can be left farther and been shown to provide more inclusive environments for SMEs farther behind when it comes to GVC participation. through anonymity given that firms might otherwise be dis- In addition to the difference between small and large firm criminated against based on size or ownership if operating in a adoption of digital technologies, the “digital divide” between traditional market (WTO, 2016; WTO, 2018). For example, wom- developed and developing countries is also a prominent issue for en-owned businesses are frequently better represented in online the new digital economy. Developing countries often have a lower platforms than offline, with the share of women-owned online level of internet access, and the internet that is available may have firms double the share of offline firms (ITC, 2017a). In China, 49.4 a lower bandwidth than that in developed countries. Because of per cent of Alibaba’s active online storeowners are female, and this reduced accessibility, there is also often a deficit of internet Etsy reports that more than 80 per cent of its retailers are women and related technical skills, posing additional barriers for SMEs. (AliResearch, 2017; TechCo, 2015; additionally, see Box 3.1 in LDCs in particular often lack the necessary infrastructure for their Chapter 3). However, despite the possibility for e-commerce to SMEs to access the internet. Additionally, e-commerce platforms open new doors for SMEs, large firms conduct the vast majority may not have expanded to certain developing countries, especially of e-commerce transactions (see Figure 6.10). LDCs, given low demand, lack of online financial infrastructure, and liability concerns (Lanz et al., 2018). Lastly, an issue that affects all New business structures and opportunities online firms, but developing country firms, is visibility. Although Digital technologies can indirectly increase SME trade by ways a firm may have a website, if the firm lacks the skills required to other than reducing costs. The scaling up of small firms, including market the business both online and offline, potential customers the “born global” phenomenon, sometimes referred to as “micro- will not know of the service or product’s availability (AMTC, 2018). multinationals” (Cusolito et al., 2016), is one important way that Even with digital capabilities, firms still face significant barriers SMEs can enter international markets and value chains. Micromul- to participate in the digital economy, such as with access to pay- tinationals achieve scale without mass, which has typically been ment systems and online sales platforms (AMTC, 2018). Further, required to expand abroad in the past (OECD, 2017a). Although e-commerce platform requirements can often be challenging for born global firms can start from any size, given the short time span SMEs to comply with and are sometimes labelled as “gatekeep- for expansion they frequently are SMEs. ers.” These barriers include membership requirements, such as the Separately, fully digital products and their creation services, such use of specific logistics suppliers; the requirement to deliver prod- as electronic games, smartphone applications, or even software ucts to purchasers within tight timeframes; and a return policy that The digital economy, GVCs and SMEs • 133 FIGURE 6.10 Proportion of small and big firms selling online in 2014 and 2015 (Per cent) Singapore (2014) Republic of Korea (2014) Belarus (2015) Indonesia (2014) Mauritius (2013) Ireland (2015) Norway (2015) Sweden (2015) Denmark (2015) Germany (2015) Belgium (2015) Netherlands (2015) Czech Republic (2015) United Kingdom (2015) Slovenia (2015) France (2015) Qatar (2015) Croatia (2015) Lithuania (2015) Portugal (2015) Finland (2015) Spain (2015) Malta (2015) Serbia (2014) Austria (2015) Slovakia (2015) Estonia (2015) Hungary (2015) United Arab Emirates (2013) Turkey (2015) Poland (2015) Italy (2015) Luxembourg (2015) Kazakhstan (2014) Latvia (2015) Bulgaria (2015) Cyprus (2015) Romania (2015) Greece (2015) Thailand (2013) Azerbaijan (2014) Republic of North Macedonia (2015) 0 20 40 60 80 100 Small enterprises (10-49 employees) Large enterprises (+250 employees) Source: (UNCTAD, 2017) based on World Bank data. is often more accommodating than the seller’s default (ARTNeT, As with all change, the digital economy has had, and will con- 2018). Additionally, developing country firms in particular note the tinue to have, significant disruptive effects on traditional mar- high costs associated with many of these platforms, including sales kets. All of these issues have implications for SMEs and their par- commission charges that range from 15 to 40 per cent of the sale ticipation in digitally facilitated trade and changing GVCs. depending on the seller’s location (WTO, 2018). A new technology, blockchain, could help to remedy some of these challenges. The technology is already being used to implement peer-to-peer mar- 4. How to promote SME participation in GVCs? ketplaces that operate without the need for a central actor (such as OpenBazaar). However, such initiatives remain, for the time being, Reaping the benefits of digital trade is not automatic. While the very limited in scope, and it is difficult to tell whether they will offer rise of the internet has opened new opportunities for SMEs to real benefits compared to existing platforms (Ganne, 2018). participate in global value chains, challenges remain that relate Separately, a recent study by the OECD (2017) also notes that both to access and use of digital technologies, and to the the digital economy has led to increased complexity, changes in broader ecosystem in which SMEs evolve. SMEs continue to face required skills and business models, and a “winner-take-all” envi- significant constraints in terms of connectivity and level of dig- ronment whereby the leading player can often dominate a market at ital skills, especially in developing countries, and market barri- the global level. Firms may be locked out of markets by the “instant ers and inefficiencies in the business environment continue to upscale” effect of winner-take-all firms that seize the market. disproportionately affect them. Increasing SME participation in 134 • Technological innovation, supply chain trade, and workers in a globalized world BOX 6.3 Connecting SMEs to the digital supply chain – challenges for the European automotive industry A challenge for SMEs to integrate as suppliers into the Euro - specific EDI software reduced the ICT burden for suppliers, pean automotive industry in the digital economy arose with but there were hidden costs related to data entry errors and the industry’s adoption of automated electronic data inter- employee time requirements on the part of the supplier. change (EDI) systems (see chapter 5 for a more in-depth Although the digital economy has created many oppor- description of EDI). Large automobile manufacturers insisted tunities for SMEs to become suppliers, issues relating to on compliance with their selected EDI standards to avoid integration and specification requirements – not only of the complications and errors. However, these systems often manufactured product but also with the delivery systems required large upfront investment that acted as barriers themselves – will continue to pose challenges for SMEs that to entry for SMEs. The development of WebEDI, a method tend to have relatively smaller resources for compliance. of conducting EDI through an internet browser rather than Source: EC (2012). global value chains requires more than simply the technology. It is regularly pointed out as one of the key impediments to SME par- requires a conducive business environment that allows SMEs to ticipation in e-commerce activities and global value chains more gen- seize the opportunities that the digital economy opens. erally. In a recent ITC survey of 2,200 SMEs in 111 countries, the lack of technical skills was ranked second out of all reported challenges 4.1 Improving SMEs’ access and use of digital technologies for e-commerce participation, behind online visibility. In fact, insuffi- SMEs’ access and use of digital technologies remains constrained cient knowledge of online marketing tools, or technical skills, was one by various factors ranging from the most basic, such as access of the key reasons put forward to explain the lack of online visibility to a steady supply of electricity in many developing countries, to for these firms. Improving online visibility requires more than simply the more complex, such as a lack of high speed internet cables having a webpage or access to an online platform; it requires specific (ITC, 2016; Darsinouei, 2017; Lanz et al., 2018). The development digital skills to master online marketing techniques (ITC, 2017a). of an efficient ICT infrastructure is essential to access global mar- To promote SME participation in global value chains, policy kets (BIAC et al., 2016; OECD, 2017a; OECD, 2017b), and when it makers need to ensure that SMEs and workers have the digital comes to e-commerce, the most important technological require- skills and knowledge to use ICT technologies efficiently in the ment remains basic access to the internet. E-commerce can only different business functions involved in international trade, from develop if the internet is present (Fernandes et al., 2017). It is market research, to product development, sourcing, production, therefore vital that governments provide their business sector, sale, and after-sale services, and actively support the develop- and in particular SMEs, with affordable, high-quality internet ment of ICT (and mobile) infrastructure. infrastructure. Mobile technology is also very important for busi- nesses, in particular in developing countries, and government 4.2 Other policy measures to support SME trade and should support both mobile infrastructure and efforts to create integration into GVCs mobile-friendly, paperless e-government systems (ABAC, 2018). Even when connected online, SMEs face a host of other barri- Key policy aspects include the mobilization of investment in ICT ers that can prevent them from joining GVCs or participating in infrastructure, both public and private, as well as the creation of international trade in the new digital economy. Many of these are a regulatory environment that provides for sound competition in ongoing obstacles, such as informality or access to finance and the telecommunications sector (Lanz et al., 2018). logistics. However, some have become even more relevant in the However, improving access to the internet and mobile technology digital age. For example, de minimis import thresholds are par- is, on its own, not sufficient to support integration of SMEs in inter- ticularly important given the increase in small shipments that has national production networks if they are not aware of the opportuni- come with e-commerce. Closer inspection of these peripheral ties that the digital economy opens and if they lack the digital skills issues can provide an indication of ways to improve SME partici- required to participate in such networks. Awareness among SMEs of pation in GVCs and international trade. how to participate in the digital economy, and how to benefit from the opportunities that digital technologies offer, remains relatively Trade policy limited (OECD, 2018c). In middle-income countries, many SMEs have Trade policy can have important simplifying effects on cross-border internet access but they often have limited understanding or capabil- trade, which can increase the use of GVCs. For example, de minimis ity to leverage the internet as part of their business plan (Cusolito et policies that set thresholds under which shipments are not required al., 2016). Further, the gap in technological adoption by SMEs relative to pay duties can reduce the tariff accumulating impact on trade, or, to large firms remains in part because of other missing components in other words, reduce the effect of adding the tariff cost of every such as insufficient R&D, human resources, and organizational and border crossing to the final product price (ITC, 2017a). This not process innovation (OECD, 2018c). The lack of technical digital skills only makes it less expensive to import intermediate products, but The digital economy, GVCs and SMEs • 135 also can make exports more competitive by reducing the final in international trade have been found to be more innovative mark-up price required for profitability. De minimis thresholds (WTO, 2016). Promoting participation in international trade and vary considerably from one country to another, ranging from none innovation are two sides of the same coin that should be pur- (meaning all imports require a customs duty regardless of value) sued in tandem. Although few SMEs have the resources to invest to 1,000 USD.10 Low de minimis thresholds pose particular barri- in R&D, those that do can contribute significantly to innovation ers for SMEs involved in ecommerce, which may have frequent low (ADB, 2015). Firm R&D spending is closely linked with manager’s volume shipments of sometimes low-value articles that still might education and experience (Gao, 2015; OECD, 2007) and can be be required to pay customs duties (Suominen, 2017). Import tar- supported by investment in areas such as technical skills or pro- iffs in general apply a cost to GVC participation, and areas with tection of IP. Further, as previously mentioned, SME participation low average import tariffs, such as Southeast Asia, are much more can be limited by system incompatibility or lack of R&D (OECD, integrated in manufacturing GVCs compared with areas that have 2007), all of which supports the idea that more R&D by SMEs can high average import tariffs, such as South Asia. Greater use of for- contribute to greater internationalization and GVC participation. eign inputs has been shown to have a positive impact on the level of sophistication and diversification of exports. This suggests that Improving the business environment policies that reduce import tariffs and facilitate border procedures Inconsistencies and uncertainties in regulation are detrimental to are likely to help SME integration into GVCs (Cusolito et al., 2016). businesses, whatever their size, but they affect SMEs more than Indeed, complicated customs procedures have been shown to be large businesses. Indeed, SMEs’ limited resources make it more especially harmful to SMEs (WTO, 2016). difficult for them to follow and deal with regulatory changes. As a Beyond tariff reductions and trade facilitation measures, deep- result, they often incur relatively higher costs to gain market share ening trade integration is positively correlated with value chain (OECD, 2017a). A complex, inconsistent and unstable regulatory activity. In a recent study, the ITC finds that increasing the number environment can hold SMEs back (see Box 6.4). Regulatory costs of trade provisions covered by trade agreements leads to more and administrative burdens can also prevent SMEs from partic- value chain integration between firms of all sizes in the participating ipating in formal sector activities, thereby also preventing them countries, with small firms benefiting the most (ITC, 2017a). Integrat- from expanding their operations internationally (OECD, 2017a). ing investment provisions in a preferential trade agreement rather When it comes to digital trade, particular consideration ought than in a separate bilateral investment treaty was also found to to be given to laws and regulations that relate to the flow of data, increase the level of domestic value added in exports (ITC, 2017b). consumer protection, and the recognition of digital documents Finally, significant barriers remain in the services sector, which and signatures. Although countries may unilaterally enact many affects SMEs disproportionately (WTO, 2016). Further liberalization reforms to improve the trading environment, especially in the area of trade in services, in particular of services that allow companies to of digital trade, other measures related to data privacy rules and connect to global value chains, such as ICT and logistics, could act standards, data movement, and recognition of e-contracts may as important enablers of SME GVC participation. require international cooperation (ARTNeT, 2018; Lanz et al., 2018). Finally, there is no sound business environment without sound Logistics and cost of delivery competition. The rise of the internet has raised new issues in this For physical goods, a key issue for trade participation by SMEs is respect. The “network effect” has enabled some internet com- the cost of delivery. In a survey conducted by the ITC (2017a), SME panies to expand rapidly, often using a subsidized fee model respondents noted their main trade challenges were costly postal whereby they price user access below their own business costs and courier delivery services. Other logistics difficulties for SMEs to gain market share. As a result, smaller firms cannot compete include the costs of shipment warehousing. These issues are par- in, or may be priced out of, the market entirely (ITC, 2017a). ticularly important for developing countries where the share of logistics costs in final prices is estimated at 26 per cent, almost Improving access to finance twice the share for developed countries (ITC, 2017a). While some It is well-established that SMEs are less able to access finance of these issues can be addressed at a regulatory level by further than large firms, be it for trade or other costs. In fact, it is esti- opening services sectors, others require proactive investment mated that the gap in available credit for formal SMEs is around measures. Indeed, a large part of the logistics challenge faced by 1 trillion US dollars, and more than half of formal SMEs in emerg- SMEs is linked to infrastructure. Without developed ports, roads, ing markets do not have adequate access to financial institutions and cargo-handling facilities, shipping costs are more expensive (Salman et al., 2017). For trade finance in particular, the WTO has (Cusolito et al., 2016). For example, it has been estimated that it is found that over half of SME requests are rejected, compared to cheaper to ship goods across the Pacific or Atlantic oceans than it only 7 per cent of large firm requests (WTO, 2018). Much of SMEs’ is to ship within the ASEAN region (ARTNeT, 2018). lack of access to trade finance stems from the cost of SME eval- uation by established lenders using traditional means like credit Promoting innovation and R&D histories. However, new technologies such as Blockchain that Participation in international trade and innovation are closely enhance traceability (Ganne, 2018) or Alibaba’s e-Credit Line that linked. Firms that innovate tend to engage more actively in takes advantage of its large store of transactions history to deter- international trade (Tian et al, 2017) and firms that participate mine credit-worthiness, could help SMEs access trade finance. 136 • Technological innovation, supply chain trade, and workers in a globalized world BOX 6.4 Regulation can hold SMEs back Regulatory standardization not only benefits cross-border business of cloud computing, uncertain quality and reliability, goods trade, but also international trade in digital services. and different operating procedures and infrastructure that Pegaxis, a Singapore-based property management service plat- may require changes within Pegaxis. Additionally, proposals by form that connects property managers with providers of ser- countries like Indonesia to require within-country incorporation vices such as landscaping or building maintenance, has encoun- reduce the geographical benefits of an online business model tered difficulties expanding to new markets in the region. For and impose time and financial costs. Uncertainties about liabil- example, Pegaxis is concerned at potential data server local- ity are also a concern for Pegaxis, such as regarding who would ization requirements. These would impose costly burdens on be considered at fault for defamatory reviews left on their web- the firm, which already has a cloud-based business model using site. This highlights how regulatory consistency, especially in the servers across the globe. Server localization requires switch- digital age, can benefit SMEs seeking to operate internationally. ing to a new provider with potentially less experience in the Source: EC (2012). Further, lack of finance is the primary barrier to SME formal- This could be changing, however, as the rise of the digital econ- ization in developing countries (OECD, 2017a). Without access to omy is reducing information search costs, facilitating exchanges, and finance, SMEs are constrained not only in their ability to export, providing new marketing, finance and networking opportunities. but also to increase their business generally, thereby making New research by Lanz et al. (2018) reveals that in developing coun- GVC access and even formalization substantially more difficult. tries, access to digital technology appears to have a positive effect on SME participation in backward-linked GVCs as well as on total Improving the quality of data exports by SMEs. This is in line with other research that has shown As the previous sections have shown, lack of data and informa- the cost-reducing effects that digital technology can have on busi- tion about SME operations represents an important barrier to ness operations, such as improved access to information or access better understanding and integrating SMEs into GVCs. Without to online services. Additionally, e-commerce provides new ways for good information on SMEs, it is difficult to know where to target firms of all sizes to access global markets, both for buying and selling policies, or whether a particular action has been effective. In intermediate or final products. Lastly, the digital economy has cre- this vein, efforts are being made to develop the Trade in Value ated new business structures that make it possible for small firms to Added (TiVA) database for improved GVC analysis, with initia- scale up in ways previously unattainable, such as the “born global” tives to include firm size breakouts in future editions. However, phenomenon, which can lead to increased SME international trade the number of economies it contains are still limited, and devel- and GVC participation. oping countries, particularly low-income countries, are not well Despite new avenues, such as online platforms that SMEs can now represented. In general, efforts are underway to sensitize coun- use to access international markets and GVCs via the digital economy, tries to break down their statistical information by firm size, as barriers continue to hinder SME access. There are a number of ways recommended by the OECD Expert Group on Extended Sup - policies, and the trading environment, can be changed to better sup- ply-Use tables that was created in 2014 (see chapter 8). Overall, port SMEs in the new digital economy. If internet access is available better information on firm operations within a country, including then an online purchase may be made, but without appropriate ship- the size of the firms, the industries they participate in, and the ping logistics, straightforward customs formalities and processes, a value and volume of trade they conduct (including whether the favorable business and regulatory environment and access to finance trade is direct or indirect) are all crucial pieces of information to a firm will be unable to complete the transaction. understand the basics of SMEs and value chain participation. Overall, reducing barriers to digital trade will require a holis- tic approach. Even though digital technologies can facilitate SMEs’ integration into GVCs, they are only one element of the ecosystem 5. Conclusions required for an SME to reach full trading potential and the develop- ment of coherent national strategies is essential. On a policy level, The international fragmentation of production that has remod- better data is also required in order to understand where the trading eled international trade over the last decades should have made it difficulties are in a given economy. Availability of data by firm size is easier for small companies to participate in global supply chains, critical to allow policy makers to better target their actions and effec- by allowing them to focus on niche markets and narrow segments tively support SMEs’ integration into GVCs. Increased availability and of international production chains. However, evidence suggests quality of data, and further analysis of direct vs indirect backward that participation of SMEs in global value chains remains limited participation and of the impact of direct versus indirect participation relative to their share of overall economic activity and employ- on firms’ performance would help to better understand and integrate ment, especially in developing countries. SMEs into GVCs. The digital economy, GVCs and SMEs • 137 Notes References 1. http://blogs.worldbank.org/psd/a-universal-definition-of-small-enter- Abel-Koch, J. (2016), “SMEs’ value chains are becoming more international prise-a-procrustean-bed-for-smes – Europe remains key”, KfW Research, Research Paper No. 137, 16 2. The terms forward and backward participation are also often referred August 2016. to as “forward linkages” and “backward linkages”. Aliresearch. (2017), “Inclusive Growth and E-commerce: China’s Experi- 3. In Germany, for example, SMEs hold between 70 and 90 per cent of ence,” April 2017. global market shares in some specialized manufacturing segments, Ahn, J., Khandelwal, A., and Wie, S. (2010), “The Role of Intermediaries and SME merchandise exports in textile, apparel and wood manu- in Facilitating Trade,” National Bureau of Economic Research, Working facturing represented more than 60 per cent of total exports across Paper 15706, Cambridge: Massachusetts. twelve OECD countries in 2015 (OECD, 2018b). Andrade, Gustavo Henrique, Miriam Bruhn, and David McKenzie. 2015. 4. This is based on data from World Bank Enterprise Surveys for over “A Helping Hand or the Long Arm of Law? Experimental Evidence on 25,000 SMEs in the manufacturing industry in developing economies. What Governments Can Do to Formalize Firms.” World Bank Economic The World Bank Enterprise Surveys collect data from key manufac- Review 30 (1): 24–54. turing and service sectors in every region of the world. The surveys APEC (2014), “Integrating SMEs into Global Value Chains: Policy Principles are conducted according to the global sampling methodology which and Best Practices”, APEC Policy Support Unit, Issues Paper No. 6, May uses stratified random sampling to minimize measurement error 2014. and to yield data that are comparable across economies. The sam- APEC Business Advisory Council (ABAC) (2018), Realizing the Untapped pling methodology generates a sample representative of the whole Potential of MSMEs in APEC: Practical Recommendations for Enhancing non-agricultural private economy, including services industries, and Cross-Border Trade, University of Southern California, Marshall School generates large enough sample sizes for selected industries to con- of Business: Los Angeles. duct statistically robust analyses with levels of precision at a minimum Asian Development Bank (ADB) (2015), Integrating SMEs into Global Value of 7.5 per cent for 90 per cent confidence intervals. Years covered Chains: Challenges and Policy Actions in Asia, 2015, Manila: ADB. differ from country to country. Asia Pacific MSME Trade Coalition (AMTC) (2018), Micro-Revolution: The 5. A commonly used definition of informality (or informal economy) New Stakeholders of Trade in APAC, AMTC: Singapore. in the literature is the one proposed by Schneider et al. (2010) who Bas, M. and V. Strauss-Kahn (2015), “Input-trade Liberalisation, export define the informal economy as comprising market-based legal pro- prices and quality upgrading”, Journal of International Economics, duction of goods and services deliberately concealed from public 95(2): 250-262. authorities to avoid paying taxes, social security contributions, and Bas, M. and V. Strauss-Kahn (2014), “Does importing more inputs raise meeting legal obligations/requirements and market standards. exports? Firm-level evidence from France”, Review of World Econom- 6. https://www.ilo.org/global/topics/employment-promotion/infor- ics, 150(2): 241-275. mal-economy/lang--en/index.htm BIAC, B20 China, World SME Forum, SME Finance Forum (2016), Financing 7. The share of women in informal employment in developing coun- Growth; SMEs in Global Value Chains. tries according to the latest available data was 4.6 percentage points Botero, J., S. Djankov, R. La Porta, F. Lopez-de-Silanes, and A. Shleifer higher than that of men including agricultural workers and 7.8 per- (2004), “The Regulation of Labor,” Quarterly Journal of Economics, Vol. centage points higher without (ILO, 2018). In some sub-Saharan Afri- 119, No. 4, pp. 1339–82. can countries, the gender gap between the formal and the informal Bruhn, Miriam, and David McKenzie (2014), “Entry Regulations and the For- sector even exceeds 20 per cent (ILO, 2018). malization of Microenterprises in Developing Countries,” World Bank 8. eWSF is intended to be a global platform for SMEs to capture GVC Research Observer 29 (2): 186–201. and B2B opportunities. Although the site is still in development, the Cebeci, T., Fernandes, A. M,. Freund, C. and Pierola, M. D. (2012), goal is to develop modular pieces to come online as each part is cre- “Exporter Dynamics Database”, Policy Research Working Paper 6229, ated. http://www.worldsmeforum.org/wp-content/uploads/2016/12/ World Bank, Washington, DC. EceIdilKasap_CACCI_Nov24.pdf Cebula, R. (1997), “An Empirical Analysis of the Impact of Government Tax 9. Born global firms are generally defined as those that achieve 25 and Auditing Policies on the Size of the Underground Economy: The per cent foreign sales out of their total sales within their first 3 years Case of the United States, 1973–94,” American Journal of Economics (Nordas, 2015). and Sociology, Vol. 56, pp. 73–186. 10. For a list of de minimis levels as of 28 March 2018 please see https:// Chan, J. (2015), “Trade Intermediation, Financial Frictions, and the Gains global-express.org/assets/files/Customs%20Committee/de-minimis/ from Trade,” Stanford Institute for Economic Policy Research, Working GEA%20overview%20on%20de%20minimis_28%20March%202018. Paper 15-009. pdf. Cusolito, A. P., Safadi, R. and Tagioni, D. (2016), Inclusive Global Value Chains – Policy Options for Small and Medium Enterprises and Low-In- come Countries, OECD/World Bank. Dabla-Norris, E., Gradstein, M., Inchauste, G. (2005), “What Causes Firms to Hide Output? The Determinants of Informality”, IMF Working Paper WP/05/160, International Monetary Fund. 138 • Technological innovation, supply chain trade, and workers in a globalized world Darsinouei, A. E. (2017), Understanding E-Commerce Issues in Trade Johnson, S., D. Kaufmann, J. McMillan and C. Woodruff (2000), “Why Do Agreements: A Development Perspective Towards MC11 and Beyond, Firms Hide? Bribes and Unofficial Activity after Communism,” Journal Geneva: CUTS International. of Public Economics, Vol. 76, pp. 495–520. de Mel, S., McKenzie, D., and Woodrufff, C. (2013), “The Demand for, and Kowalski, P., Lopez Gonzalez, J., Ragoussis, A. and Ugarte, C. (2015), “Par- Consequences of, Formalization among Informal Firms in Sri Lanka,” ticipation of Developing Countries in Global Value Chains: Implications American Economic Journal: Applied Economics. for Trade and Trade-Related Policies”, OECD Trade Policy Papers, No. European Commission (EC), (2012), Integrating Automotive, Connecting 179, OECD Publishing, Paris, SMEs to the Digital Supply Chain, Brussels: EU. Lanz, R., Lundquist, K., Mansio, G., Maurer, A., and Teh R. (2018), “E-com- Farazi, S. (2014), “Informal Firms and Financial Inclusion: Status and Deter- merce and developing country-SME participation in global value minants”, Policy Research Working Paper No. 6778, World Bank, Wash- chains”, Background paper. ington, D.C.: World Bank. La Porta, R., and Shleifer, A. (2014), “Informality and Development”, NBER Felbermayr, G. and Jung B. (2011), “Trade Intermediation and the Orga- Working Paper No. 20205, June 2014. nization of Exporters,” Review of International Economics, 19(4), Lendle, A. and Olarreaga, M. (2014), “Can Online Markets Make Trade More pp.634-648. Inclusive?” Discussion Paper No. 349, Washington DC: Inter-American Fernandes, A. M., Mattoo, A., Nguyen, H., & Schiffbauer, M. (2017), “The Development Bank. Internet and Chinese Exports in the Pre-Alibaba Era”, Policy Research López González, J. (2017), “Mapping the participation of ASEAN small- Working Papers. and medium- sized enterprises in global value chains”, OECD Trade Friedman, E., S. Johnson, D. Kaufman, and P. Zoido-Lobatón (2000), Policy Papers, No. 203, OECD Publishing, Paris. “Dodging the Grabbing Hand: The Determinants of Unofficial Activity Lopez-Gonzalez, J. (2016), “Using Foreign Factors to Enhance Domestic in 69 Countries,” Journal of Public Economics, Vol. 76, pp. 459–92. Export Performance: A Focus on Southeast Asia”, OECD Trade Policy Ganne, E. (2018), “Can Blockchain Revolutionize International Trade”, Papers, No. 191, OECD Publishing, Paris. Geneva: WTO. Lu, J. Meng, B., Zhang, H., and Wei, S. (2018), “Corruption, Import Liber- Gao, Y., Hafsi, T., (2015) “R & D spending among Chinese SMEs: the role of alization, and Productivity in China: A Firm-Level Analysis,” Background business owners’ characteristics”, Management Decision, Vol. 53 Issue: paper. 8, pp.1714-1735. Manova, Kalina & Yu, Zhihong (2016), “How firms export: Processing vs. Giles, D., and L. Tedds (2002), “Taxes and the Canadian Underground ordinary trade with financial frictions,” Journal of International Econom- Economy,” Canadian Tax Paper No. 106 (Toronto: Canadian Tax ics, Elsevier, vol. 100(C), pages 120-137. Foundation). Mayer, T., and Ottaviano, G. (2007), The Happy Few: The Internationali- Harvie, C., Narjoko, D., and Oum, S. (2010), “Firm Characteristic Determi- sation of European Firms: New Facts Based on Firm-Level Evidence, nants of SME Participation in Production Networks”, ERIA- Discussion Brussels: Bruegel Blueprint Series. Paper Series, No. 2010–11, October. Nguyen, D. C., Nguyen, N. A., Li, H. A. and Nguyen, T. P. M. (2012), “Inno- Harvie, C., Narjoko, D., and Oum, S. (2015), Economic Integration in East vation and Choice of Exporting Modes under Globalization”, in Hahn, Asia: Production Networks and Small and Medium Enterprises, Rout- C. H. and Narjoko, D.A. (eds), Dynamics of Firm Selection Process in ledge, London and New York. Globalized Economies, ERIA Research Project Report 2011, No. 3. International Finance Corporation (IFC) (2012), Enterprise Finance Gap Organization for Economic Co-operation and Development (OECD) Database, World Bank, Washington DC. (2018a), “Accounting for firm heterogeneity in global value chains: International Finance Corporation (IFC) (2013), IFC Jobs Study: Assessing The role of Small and Medium Sized Enterprises”, STD/CSSP/ Private Sector Contributions to Job Creation and Poverty Reduction, WPTGS(2018)5. World Bank, Washington DC. Organization for Economic Co-operation and Development (OECD) International Labour Organization (ILO) (2018), Women and men in the (2018b), “Fostering greater SME participation in a globally integrated informal economy: A statistical picture, third edition (Geneva). economy”, Discussion Paper, SME Ministerial Conference 22-23 Feb- International Trade Centre (ITC) (2017a), New Pathways to E-Commerce – A ruary 2018. Global MSME Competitiveness Survey, September 2017, Geneva: ITC. Organization for Economic Co-operation and Development (OECD) International Trade Centre (ITC) (2017b), SME Competitiveness Outlook (2018c), “SME and Entrepreneurship Outlook”, Room Document, 2017, Geneva: ITC. OECD Working Party on SMEs and Entrepreneurs (WPSMEE), 10-11 International Trade Center (ITC) (2016), “Bringing SMEs onto the e-Com- October 2018. merce Highway,” Organization for Economic Co-operation and Development (OECD) Geneva: ITC. (2017a), “Enhancing the Contributions of SMEs in a Global and Digi- International Trade Centre (ITC) (2015a), SME Competitiveness Outlook: talized Economy”, Meeting of the OECD Council at Ministerial Level, Connect, Compete and Change for Inclusive Growth, Geneva: ITC. Paris, 7-8 June 2017. International Trade Centre (ITC) (2015b), “How companies experience Organization for Economic Co-operation and Development (OECD) non-tariff measures – Survey-based evidence from developing coun- (2017b), Small, Medium, Strong. Trends in SME Performance and Busi- tries”, ITC technical paper. ness Conditions, OECD Publishing, Paris. The digital economy, GVCs and SMEs • 139 Organization for Economic Co-operation and Development (OECD) (2007), Suominen, K. (2017), “The Silver Bullet for Fueling Small Business Exports “Enhancing the Role of SMEs in Global Value Chains”, Proceedings of in the Ecommerce Era: A plurilateral on De Minimis,” April 2017. OECD global conference. 31 May–1 June 2007, Tokyo. Tech.Co (2015), “The Story of Etsy’s Success is the Story of the Success of Oxford Economics (2017), SME Strategies for Success: A Global Study Women Entrepreneurs,” May 13, 2015. that Reveals the Key Drivers of Success for SMEs in 2017 and Beyond, Tian, W., Yu, M., Zhang, R. (2017), “International trade and firm innovation: Oxford Economics, Oxford. patterns and evidence from the Chinese Employer-Employee Survey Perry, G., Maloney, W. F., Arias, O. S., Fajnzylber, P., Mason, A. D., and data”, China Economic Journal, Volume 10, 2017, Issue 1. Saavedra-Chanduvi, J. (2007), Informality: Exit and Exclusion, World Ulysseay, G. (2015), “Firms, Informality and Development: Theory and evi- Bank, Washington DC. dence from Brazil”, Rede de economia aplicada (REAP) Working Paper Salman Alibhai, Simon Bell, and Gillette Conner, “What’s happening in the 76, March 2015. missing middle? Lessons from financing SMEs,” World Bank Group: United States International Trade Commission (USITC) (2010) Small and Washington DC. Medium-Sized Enterprises: Characteristics and Performance, Publica- Schneider, F., Buehn, A. and Montenegro, C. E. (2010) “Shadow Econo- tion No. 4189, Washington DC: USITC. mies All over the World, New Estimates for 162 Countries from 1999 United Nations Conference on Trade and Development (UNCTAD) (2017), to 2007”, Policy Research Working Paper 5356, World Bank, Wash- Information Economy Report 2017: Digitalization, Trade and Develop- ington DC. ment, Geneva: UNCTAD. Schröder, P., Trabold, H., and Trübswetter, P (2003), “Intermediation in For- United Nations Conference on Trade and Development (UNCTAD) (2013), eign Trade: When do Exporters Rely on Intermediaries?” World Investment Report 2013: Global Value Chains, Investment and Slaughter, M. J. (2013) American Companies and Global Supply Networks: Trade for Development, New York and Geneva: UN. Driving US Economic Growth and Jobs by Connecting with the World, World Trade Organization (WTO) (2016), World Trade Report 2016: Level- Washington DC: Business Roundtable with the United States Council ling the Playing Field for SMEs, Geneva: WTO. for International Business and the United States Council Foundation. World Trade Organization (WTO) (2018), World Trade Report 2018: The Straub, S. (2005), “Informal Sector: The Credit Market Channel,” Journal of Future of World Trade, How Digital Technologies are Transforming Development Economics (forthcoming). Global Commerce, Geneva: WTO. CHAPTER 7 Should high domestic value added in exports be an objective of policy? David Dollar (Brookings Institution), Bilal Khan (RCGVC-UIBE), and Jiansuo Pei (SITE-UIBE) ABSTRACT Global value chains make it easier for developing countries is not good policy. It may seem like simple math that a higher to move away from export reliance on unprocessed primary domestic value added share means more total value added products to become exporters of manufactures and services. exported and hence more GDP. But that simple idea ignores Global value chains (GVCs) allow countries to specialize in a the reality that imported goods and services are a key sup- particular activity and join a global production network. As a port to a country’s competitiveness. The chapter documents developing country moves from export of primary products this via the history of the successful East Asian industrializers to export of manufactures and services via GVCs, the ratio as well as more recent evidence from Association of South- of domestic value added to gross export value tends to fall. east Asian Nations (ASEAN) economies. If a country artifi- Many developing country policy-makers worry about this cially replaces key inputs with inferior domestic versions, the trend and aspire to increase their value added contribution end result is likely to be fewer gross exports and less, not to exports. There are a number of reasons why this objective more, total value added exports. • In almost all countries, developed and developing alike, the share of domestic value added in exports has tended to trend downwards recently. This reflects the expansion of global value chains. • Many developing countries worry about this phenomenon and aspire to increase their value-added contribution to exports. This objective should be approached cautiously. Imported goods and services are a key support to a country’s competitiveness. If a country artificially replaces key inputs with inferior domestic versions, the result is likely to be fewer gross exports and fewer, not more, total value-added exports. • China’s recent experience is often given as an important counter-example, since its domestic value- added ratio has been rising over the past decade, but our research indicates that this trend is primarily the result of technological advances in China. • Consequently, the Chinese ratio can be expected to peak and later decline if China further opens up and follows in the steps of other earlier Asian industrializers, such as Japan and the Republic of Korea. 141 142 • Technological innovation, supply chain trade, and workers in a globalized world 1. Introduction G and will make use of some of them. The decline in the share of domestic value added in exports in many cases is also the result lobal value chains make it easier for developing of structural change in the export basket. Economies that open countries to move away from export reliance on up after an autarkic period often start by exporting primary unprocessed primary products to become exporters products, which tend to be relatively homogeneous worldwide. of manufactures and services. Before the develop- Manufactures produced in closed economies are low-quality and ment of GVCs, a country had to master the production of a whole find few markets. Over time, however, with openness to imported product in order to export it. GVCs allow countries to specialize inputs, a competitive manufacturing sector may emerge. China, in a particular activity and join a global production network. As Mexico, and Viet Nam are all examples of economies that initially a developing country moves from export of primary products exported primary products after opening up, but soon moved to to export of manufactures and services via GVCs, the ratio of manufacturing exports. That structural shift will tend to reduce domestic value added to gross export value tends to fall. Devel- the domestic content share of overall exports because that share oping countries often start out at the end of value chains, with is generally higher for primary products and lower for manu- labor-intensive assembly of parts produced elsewhere. For some factures. So, in general, we observe a declining domestic value individual products the ratio of domestic value added to gross added ratio over time. Further, this indicator does not have direct export value can be very small, maybe only a few percentage welfare implication, so it is not appropriate to formulate policies points. The gross exports from the country can be very large, but around pursuing a higher domestic content ratio in exports. this is an artifact of the position in the value chain. The country’s In this section we examine historical data for three indus- value added contribution to the export is much smaller. Many trializing economies in East Asia to study the development of developing countries worry about this phenomenon and aspire domestic content in exports. We proceed in three steps of analy- to increase their value added contribution to exports. There are sis, namely the aggregate trend for the total economy, the trend a number of reasons why this objective should be approached for manufacturing products, and the trend for the electronics cautiously. It may seem like simple math that a higher domes- industry, a high-tech sector.1 tic value added share means more total value added exported and hence more GDP. But that simple idea ignores the reality 2.1 The case of Japan that imported goods and services are a key support to a coun- For Japan, annual input-output tables date back to 1973. It is evi- try’s competitiveness. If a country artificially replaces key inputs dent that the domestic content in overall exports has declined, with inferior domestic versions, the end result is likely to be fewer decreasing roughly 0.12 points from 1973 through 2014 (Figure gross exports and less, not more, total value added exports. 7.1). Several factors may account for this decline. It can be seen In this chapter we examine this issue. Section 2 looks at the that the domestic value-added ratio (DVAR) in manufactures is historical experience of the successful East Asian industrializers, always below that of total exports, reflecting differences between Japan, the Republic of Korea, and Chinese Taipei. Section 3 then focuses on the more recent experience of ASEAN economies as well as general literature on the issue of domestic value added content of exports. Section 4 looks at policy measures that econ- FIGURE 7.1 Domestic value added in Japanese exports omies can consider in order to move up the technological ladder. In general, artificially trying to boost domestic content is going to be a losing strategy. Countries would do better to focus on 1.0 human capital development, support to R&D, intellectual prop - 0.9 erty rights protection, and opening up remaining closed sectors 0.8 of the economy, especially services. 0.7 0.6 0.5 2. The decline of domestic value added in 0.4 exports in Japan, the Republic of Korea, and 0.3 Chinese Taipei 0.2 0.1 After two decades of dedicated work among international 0 economists in measuring international fragmentation, a consen- 1970 1975 1980 1985 1990 1995 2000 2005 2010 2015 2020 sus has more or less formed, that the trend of domestic value added in exports is declining. Starting from autarky, when the Aggregate Manufacture Electronics economy opens to trade, there are several reasons why the domestic content of exports would begin to decline. Opening up to imports of intermediate goods and services means that a Source: Authors’ calculation based on Japan’s national IO table. country’s producers have access to the most competitive inputs Should high domestic value added in exports be an objective of policy? • 143 manufactures and primary products. The latter have relatively As in Japan’s case, several factors may account for the decline few intermediate inputs, and hence few imported inputs. The in the ratio of domestic value added to export value, such as the structural shift away from primary exports towards manufactured continuing trade liberalization, international production frag- exports would pull down the DVAR in overall exports. In addition, mentation, and structural shift from primary exports to manufac- various waves of trade liberalization gave Japanese producers tures. As with Japan, the DVAR in electronics is particularly low. better access to imported goods and services for production. The Republic of Korea has very successfully developed its own Clearly, this decline trend is more pronounced for manufac- brands in televisions and smartphones, yet the DVAR in elec- tured products, as well as for the high-tech electronics sector, tronics has generally been below 40%. The Republic of Korea’s especially after 1990 when Japan’s serious trade liberalization success results from combining domestic value added with accelerated. The general pattern is similar, while the ratio for imported components and services. manufacturing products and electronics are much lower than the aggregate. In addition, they decrease 0.19 and 0.15, respec- 2.3 The case of Chinese Taipei tively, comparing with 0.12 for the aggregate value. Arguably, For Chinese Taipei, annual input-output tables date back to 1960s, the expanding international production fragmentation is well and the domestic content of exports peaked in 1969 with a ratio of observed in Japan’s case. It is also notable that electronics is the roughly 79 percent (Figure 7.3). Domestic content has fallen sharply most high-technology sector, and here Japan’s domestic value over time, reaching 48 percent in 2011. Hence, the overall decline added in less than 40% of the gross export value in the most was around 30 percentage points (comparing with the world aver- recent year. Thus, success in the high-tech sector goes hand-in- age ratio of value added to exports falling by roughly 10 percent- hand with extensive use of imported inputs and services. age points, as reported in Johnson and Noguera, 2016), which is remarkable. Different from its Asian peers such as Japan and the 2.2 The case of the Republic of Korea Republic of Korea, Chinese Taipei is a typical small open economy. For the Republic of Korea, annual input-output tables date back Given the growth of international production fragmentation, along to 1985. It is observed that the aggregate ratio of the domestic with Chinese Taipei’s steady trade liberalization, it is expected that content in exports declined, with most of the change since 1995 the ratio of domestic content to exports would see a sharp decline. (Figure 7.2). Between 1995 and 2014 the ratio dropped roughly As a strategy for the developing regions to integrate into the 0.15. Next, we observed that the decline was similar for manu- world economy, joining global production is one of the shortcuts. factured products; while the general pattern is similar, the ratio This is particularly true for small open economies. In this way, the for manufacturing products is much lower than the aggregate. domestic industry structure is no longer a prerequisite for produc- DVAR for aggregate exports was about 0.55 in the most recent ing internationally competitive products, as they can specialize in year, compared to less than 0.40 for manufactures. The trend for some particular stage of production, e.g. processing and assembly the electronics sector is similar. activities. FIGURE 7.2 Domestic value added in Korean exports FIGURE 7.3 Domestic value added in exports of Chinese Taipei 0.8 0.9 0.7 0.8 0.6 0.7 0.6 0.5 0.5 0.4 0.4 0.3 0.3 0.2 0.2 0.1 0.1 0 0 1980 1985 1990 1995 2000 2005 2010 2015 2020 1960 1970 1980 1990 2000 2010 2020 Aggregate Manufacture Electronics Aggregate Manufacture Electronics Source: Authors’ calculation based on the Republic of Korea’s national IO table. Source: Authors’ calculation based on Chinese Taipei’s IO table. 144 • Technological innovation, supply chain trade, and workers in a globalized world 3. Developing countries’ experience of joining Intuitively, importing better quality intermediates as well as ser- GVCs vices, increases the competitiveness of the domestic firms in the international market and leads to higher demand for the product This section will analyze the recent experiences of develop - as well as employees in exporting sector. ing countries by comparing the domestic value-added (DVA) in There is no single strategy that works for every economy. exports and its implications for the labor market. In the last few Each country has to realize the economic activity that can be decades, we have noticed that many developing countries have integrated into the GVCs. Figure 7.5 shows backward and for- been joining at the lower end of the value chain and have been ward GVC linkages for Asian economies in 1995 and 2011. Viet able to increase their gross exports, achieve higher GDP per Nam, for example, has increased the backward linkages, that capita growth, and generate employment opportunities despite is, the use of imported goods and services in its production the reduction in DVA in gross exports. However, the policymakers of exports. Viet Nam has primarily participated at the produc- in some of these economies are now targeting to increase the DVA tion and assembly stage of manufacturing sector (light man- in gross exports by using tariff and non-tariff barriers to imported ufacturing, electrical equipment, electronics etc.) in the GVCs. inputs. They believe that the best way to utilize the exporting Viet Nam’s DVA share in gross exports fell from 79.1% to 63.7% activity for development is by increasing the DVA content in gross during this period. During the same period, GDP per capita in exports as it will create more and better job opportunities for Viet Nam increased from $288 in 1995 to more than $1500 in domestic workers, given everything else remains constant. 2011. Viet Nam has been able to shift a significant proportion Figure 7.4 compares the ASEAN exports in 1995 and 2011. of workers from the relatively less productive agricultural sector From 1995 to 2011, DVA share in exports dropped from 71% to to the more productive manufacturing and services sectors. This 67%, though, it can be seen that the absolute value of exports remarkable progress has been achieved by embracing trade and increased many folds. Much of this increase has been attributed investment openness by signing a Bilateral Trade Agreement to increase in the intermediate exports i.e. exports related with USA in 2002 and joining the WTO in 2007. These agree- to the GVCs. This tremendous increase in gross exports also ments encouraged Viet Nam to reduce the import tariffs and led to remarkable growth in jobs. Using empirical evidence, improve infrastructure to attract foreign direct investment (FDI). Lopez-Gonzalez (2016) has shown that the foreign value-added, These policies resulted in importing better quality inputs as well in form of intermediate imports as well as services, plays a sig- as related services and focusing on the stage of production (pri- nificantly positive role in enhancing the employment as well as marily assembling and processing) where the Vietnamese firms/ productivity (value-added per worker) in the ASEAN countries. workers have comparative advantage. In 2017, nearly a third of FIGURE 7.4 Enjoying a smaller share of bigger pie, ASEAN exports Value added content of exports (share) Value added content of exports (Value in billion USD) 120 1,400 100 1,200 28.6% 33.4% 1,000 401 80 800 60 600 40 71.4% 66.6% 400 798.6 20 89.8 200 224.5 0 0 1995 2011 1995 2011 Domestic Foreign Domestic Foreign Source: Lopez-Gonzalez (2016). Should high domestic value added in exports be an objective of policy? • 145 FIGURE 7.5 GVC participation in Factory Asia  50 40 30 20 10 0 m na sia am sia a a re n s a ei nd ne re di pa in al o ip hi ay ne N bo Ko Ch ap ss pi la Ja ,C Ta al do ai et ru ip m ng of M ng Th e Vi ar ill Ca In Si es Ph ic D Ko in bl ei Ch pu g un on Re Br H Backward 1995 Forward 1995 Source: Lopez-Gonzalez (2016). the Viet Nam’s exports as well as imports consisted of electrical In order to highlight the implications of directly targeting and electronics equipment despite having a shortage of skilled the DVA in exports as a national policy for development, we workers in the labor force. This has been possible only by joining can compare Bangladesh and Pakistan’s approach towards the the GVCs in electronics and electrical equipment. exports in the textile and clothing sector. The biggest exports In contrast, Indonesia has not seen much change in backward of both countries have been textiles and clothing. In 1990, Ban- linkages since 1995 as its specialization lies in natural resources gladesh’s exports ($1.09 bn) were a third of Pakistan’s exports and hence, it has strengthened its forward linkages since then. ($3.5 bn) in textiles and clothing. Since then, Pakistan, being a Indonesia joined the global value chains in relatively upstream cotton producer, incentivized the textile producers to use the industries. Indonesia’s GDP per capita has also grown three local inputs and export the finished products. Bangladesh, times during the same period, though, this phenomenal growth mostly importing the raw materials for textile and clothing, has come through forward linkages in GVC participation. We can focused more on the trade reforms and opening up the economy also notice that most of the other countries in ASEAN followed a to foreign investors. Bangladesh integrated its textile and cloth- similar pattern of achieving higher growth in GDP per capita as ing sector in the global value chains, sourcing most of the raw well as exports related to the GVCs. material from abroad and exporting readymade garments to the Figure 7.6 shows that the share of foreign value added in developed world. This helped Bangladesh to slowly convert its gross exports of the domestic economy is not only increasing comparative advantage in clothing into competitive advantage through importing more intermediate products but also through over time by using better quality inputs as well as foreign ser- utilizing foreign services in exports. These foreign service pro - vices by collaborating with the leading garments manufacturers. viders, being much more efficient than the domestic providers, In 2016, Pakistan’s textile and clothing exports ($12.4 bn) were also play a significant role in enhancing the competitiveness of less than half of Bangladesh’s exports ($28.3 bn).2 Using John- the exports. It must be emphasized that different economies in son (2018), we can calculate the DVA in the Textile and Clothing ASEAN entered at different stages of GVCs and specialized in sector for these two economies in 2014.3 Bangladesh’s DVA in different industries (or possibly within the same industry but at the Textile and Clothing sector was 64.5% as compared to 80.3% different stages of production) based on the comparative advan- for Pakistan.4 It can be seen that Bangladesh’s exports have risen tage of the domestic economy. Most of the ASEAN economies much faster as compared to Pakistan despite having lower DVA have integrated well into the regional as well as global value in exports. As the labor costs are rising in China, many garment chains depending on the relative comparative advantage of the producers might look for opportunities abroad to relocate their domestic economy. This integration helped these economies plants. Bangladesh, being well integrated into the GVCs in tex- to achieve much higher GDP growth and create millions of job tile and clothing, the second biggest exporter of garments and opportunities for their workers and helped a significant propor- offering lower wages, will be the first choice of these firms to tion of the population to lift out of poverty. relocate. The textile and garment sector in Bangladesh has also 146 • Technological innovation, supply chain trade, and workers in a globalized world FIGURE 7.6 Service content of exports in Factory Asia  90 80 70 60 50 40 30 20 10 0 a sia na m am n na sia a re es ei nd re pa di al o hi i ip ay n ne Ko Ch N bo la ss ap Ja pi ,C Ta al ai do et ru ip m ng of M Th ng Vi e ar ill Ca In es Si ic Ph D Ko bl in ei pu Ch g un on Re Br H Domestic service content 1995 Foreign service content 1995 Source: Lopez-Gonzalez (2016). helped to achieve twice the female labor force participation rate as well as achieve higher GDP growth rate since joining WTO. as that of India and Pakistan as nearly 85% workers are females Tang et al. (2018) have suggested that the substantial improve- in this sector. Bangladesh’s GDP per capita also surpassed Paki- ment in technology in China, along with falling trade costs, have stan’s GDP per capita in 2017. been the reason for the rising DVA in gross exports recently. Another good example of the contrast between global inte - China has also invested hugely in improving human capital in gration and import substitution comes from the auto sector in the last two decades to complement the advanced technology Malaysia and Thailand, neighboring countries at similar stages adoption by Chinese firms. Though, as we can see from the of development. Malaysia tried to develop an indigenous auto experiences of the Republic of Korea and Japan, this might be a industry and a national champion brand through protectionist short-lived phenomenon. Once the Chinese economy catches up policies, whereas Thailand strove to join GVCs around existing in technology with other economies and achieves the maximum brands by attracting FDI from Japanese and American compa- DVA in exports, rising labor costs and stringent environmental nies. Thailand’s strategy enabled it to integrate into successful standards might push the firms to outsource the low value added value chains and become a significant exporter of auto value segment in production to the other regions in the world to main- added, primarily via parts. Malaysia’s effort did not produce a tain competitiveness by the exporting firms. globally competitive car and eventually had to be abandoned The regional experiences within China provide some addi- (Wad 2009). tional perspective on the relationship between DVA and level of All the examples discussed above suggest that participation development. Across Chinese provinces there is an inverted-U and integration into the GVCs help the economies to improve shape relationship in which DVA tends to rise with per capita their trade competitiveness, achieve higher GDP per capita GDP and then decline beyond a certain threshold (Figure 7.7). In growth and improve female labor force participation despite fall- general, the richest provinces in China have low ratios of domes- ing DVA in gross exports. Global technological advancement as tic value added to exports. Beijing and Shanghai have particu- well as falling trading costs have resulted in the fragmentation of larly low ratios. But the export powerhouse of Guangdong also production across borders. This reduction in trade costs helps has a low domestic value-added ratio, and hence a high ratio the firms to exploit the comparative advantage of each country of imported inputs. This is consistent with the view that export in the specific stage of production and hence, there is a reduc- success requires access to the best inputs in terms of manu- tion in DVA in gross exports. The only country that has been factured parts and supporting services. As more Chinese prov- able to buck the trend of global decline in DVA in gross exports inces develop in the direction of the already successful ones, the despite increasing GDP per capita as well as rising exports has national ratio of domestic value added to gross exports is likely been China. China has been able to increase the DVA in exports to fall. Should high domestic value added in exports be an objective of policy? • 147 FIGURE 7.7 Relationship between per capita GDP and the provincial VA rate of exports Intra-provincial VA rate of export: % 85 Hubei Sichuan Jilin Shandong 75 Fujian Hunan Qinghai Guangxi Shanxi Yunnan 65 Hebei Jiangxi Xinjiang InnerMong Heilongjiang Liaoning Guizhou Shaanxi Henan Zhejiang Chongqing 55 Gansu Jiangsu Guangdong Tianjin Ningxia 45 Anhui Hainan 35 Beijing Shanghai 25 8 8.2 8.4 8.6 8.8 9 9.2 9.4 9.6 9.8 LN(GDP per capita) Source: Li et al. (2018). These results have also been supported by the empirical anal- the import tariffs on intermediates will help the firms to import ysis for other developing countries in the literature. Using the higher quality inputs, it will also help to increase the employ- manufacturing firms’ data from Indonesia, Amiti and Konings ment in the plant, along with improving the quality of the output, (2007) find that a 10% reduction in import (input) tariffs would irrespective of the firm being an exporting firm or producing for lead to 12% productivity gains for the importing firms, at least the domestic market. These results from the existing literature twice as high as gains from reducing the output (final goods) suggest that adding tariffs on the imported inputs will not only tariffs. Interestingly, Goldberg et al. (2010) show that removing adversely affect the quality, employment and number of products (or lowering) the import tariffs on newer inputs also assists the of the firms in exporting sector but also of the domestic firms. domestic firms to introduce newer products. They show that the lower input tariffs accounts for an average of 31 percent of the newer products introduced by domestic firms in India. Kugler 4. Policies for technological upgrading and Verhoogen (2012) study the impact of quality of inputs and outputs on the plant size, in terms of employment, for Colom- Technological upgrading is an important part of the convergence bian manufacturing sector. They found that the more productive process. Developing countries that are integrated into the global plants use higher-quality imported inputs as indicated by com- economy generally have had more rapid total factor productivity paring the price of domestic input in the same category by the growth – our best way to measure technological advance – than same plant. They also showed that there is positive correlation the already developed economies. Developed economies are at between the quality of inputs and the plant size and the price of the frontier and have to invent new technologies, which is costly the output, an indicator of quality of the product. Since reducing and difficult. Developing countries can absorb already existing 148 • Technological innovation, supply chain trade, and workers in a globalized world technologies through direct foreign investment and learning. Firms in competitive economies source goods and services As they progress, it is natural for developing countries to begin internationally if they are superior in quality and/or lower cost. spending resources on inventing new technologies so the more If a country artificially induces firms to source locally, it will advanced developing economies are both absorbing existing reduce their competitiveness and lead to less total value added technologies and innovating new ones. and productivity. From a policy point of view then, developing We have seen earlier in this chapter that, at certain stages countries should encourage technological advance but remain of development, technological advance goes hand-in-hand indifferent to whether inputs are sourced locally or internation- with higher domestic value added content of exports. We see ally. That is a choice best left to the firm. There are policies that the causality here running from technological advance to GVC countries can use to encourage technological innovation, such structure, not from domestic content to technological advance. as support for STEM (science, technology, engineering, and The reason for this conclusion is that, for Japan, the Republic of mathematics) education, subsidies to R&D, intellectual prop - Korea, and Chinese Taipei, as well as for the other technologi - erty rights (IPR) protection, and openness to foreign trade and cally advanced economies such as Germany and the U.S., the investment. clear trend is for DVAR to fall. The most advanced technological The leading countries in the world producing STEM gradu - economies are all extremely open, capitalist economies. Firms ates are now China and India. In 2016 China produced almost are choosing in competitive markets which inputs – goods and 5 million STEM graduates (undergraduate and graduate com - services – to source locally and which to source internationally. bined), and India, nearly 3 million (Figure 7.8). This was far in The result in all of these economies is more international sourc- excess of the 568,000 graduated in the U.S. Russia, Iran, and ing over time. Indonesia also produced significant STEM graduates. This We conclude from these patterns that it is reasonable for increasing pool of technical labor in emerging markets naturally a developing economy to aspire to more rapid technological encourages hi-tech industries to expand there, including the advance, which will contribute to higher living standards both establishment of research centers. Aside from quantity of grad- directly and indirectly (because technological advance raises uates, there is also the important issue of quality. Most of the the return to investment and encourages capital accumulation). top research universities in the world are in the U.S. and West- In certain periods, this may lead to an increase in the DVAR, ern Europe, but Chinese and Indian universities are starting to but in the long run it is likely to lead to declines in DVAR as has climb the ranks. In 2018, 43 of the top 100 research universities been witnessed in all of the advanced economies. It is an easy in the world were in the U.S., followed by Continental Europe, mistake for developing countries to see the causality going the Australia and Canada (Figure 7.9). China came next. The highest other way. If all else were equal, then increasing DVAR would ranked universities from China were Peking University (#27) and mean more total value added and typically higher productiv- Tsinghua University (#30). No Indian university has yet cracked ity. The problem with this thinking is all else will not be equal. the top 100. FIGURE 7.8 Countries with the most STEM graduates (2016)  China India United States Russia Iran Indonesia Japan 0 1,000,000 2,000,000 3,000,000 4,000,000 5,000,000 Source: http://blogs-images.forbes.com/niallmccarthy/files/2017/02/20170202_STEM.jpg Should high domestic value added in exports be an objective of policy? • 149 FIGURE 7.9 Top 100 world research universities, by area (2018)  Singapore Japan China Canada Australia United Kingdom Continental Europe United States 0 5 10 15 20 25 30 35 40 45 50 Source: The Times Higher Education World University Rankings 2018. Emerging markets in general still spend very little on differences between the advanced economies and emerging mar- research and development. India, despite its success in certain kets is in the quality of IPR protection. All of the top innovative hi-tech areas, spends only about half a percent of GDP on R&D. economies score very highly on an index of IPR protection from the In fact, most emerging markets do not even report consistent Intellectual Property Rights Alliance (Figure 7.11). Emerging markets data on R&D because it is such a small part of their economies. such as Argentina, Brazil, China, India, Indonesia, and South Africa China is the notable exception. As recently as 2001 China spent lag well behind. For countries like China that have made progress less than 1% of GDP on R&D, but that figure has been climbing with the inputs of innovation, such as STEM graduates and R&D steadily in recent years and in 2015 China spent more than 2% spending, improving IPR protection should be a key priority in order (Figure 7.10). The advanced economies generally spend between to get the greatest innovation output from the effort. 2 and 3% of GDP on R&D, and China has now joined that club. The most technologically advanced countries have seen their In both China and the U.S., about one-fifth of R&D is financed DVARs decline in recent years as they make proportionally more by the government, with the rest primarily coming from industry. use of imported inputs. These economies also tend to have large This reveals that it is difficult for the government to have much shares of services in their exported value added. This rising direct effect on R&D. Subsidies, usually in the form on tax breaks, service share reflects two factors: first, there is growing service play some role. But, in general, R&D is based on corporate deci- content embodied in manufactured products, such as software sions which are influenced by availability of technical labor and in automobiles and appliances; second, as value chains become other aspects of the policy environment. more fragmented, services such as finance, telecom, and trans- One of the most important aspects of the policy environment for port are increasingly important in managing value chains. Given R&D is intellectual property rights protection. Since the vast major- these trends, it is not surprising that the most technologically ity of R&D funding comes from industry, it is aimed at developing advanced countries tend to be very open to trade and invest- commercial innovations – new technologies for providing goods ment in services. In these sectors trade and investment tend to and services. The logic of IPR protection is to provide a temporary go together because it is hard to trade most services without an monopoly for the innovator. This is necessary to create a financial investment presence. incentive to innovate. If innovations could be instantly copied, then The OECD calculates a direct investment restrictiveness index there would be no incentive for R&D. On the other hand, once inno- for whole economies and for particular sectors. The advanced vations exist it is socially optimal for them to diffuse, and for that economies that make up the OECD are open in virtually all sec- reason IPR protection tends to be temporary and imperfect, allow- tors. Emerging markets, on the other hand, tend to be fairly ing reasonable offshoots to develop quickly. One of the striking open in manufacturing but still somewhat closed in services such 150 • Technological innovation, supply chain trade, and workers in a globalized world FIGURE 7.10 R&D as a share of GDP (2015)  3.5 3 2.5 2 1.5 1 0.5 0 India China Netherlands Germany United States Japan Source: World Bank (2018). as telecom and finance (Figure 7.12). This is particularly true for firm-level evidence that the closed service sectors have low pro - China, which overall is the most closed among major emerging ductivity levels and growth rates. Hence, the protected strategy markets. Countries such as India and Indonesia are not as closed consigns China to poor-quality services that then make it more as China, but far from OECD levels. Argentina, Brazil, and South difficult for other sectors, including manufacturing, to reach Africa all tend to be more open. In the case of China, there is international quality. FIGURE 7.11 IPR protection index (2018)  9 8 7 6 5 4 3 2 1 0 Argentina Brazil China India Indonesia South Africa U.S. Netherlands Japan Source: https://www.internationalpropertyrightsindex.org/ Should high domestic value added in exports be an objective of policy? • 151 FIGURE 7.12 FDI restrictiveness index for manufacturing, telecom, and financial services (2018) (0= open, 1= closed) Mfg OECD South Africa India Telecomms Indonesia China Brazil Argentina Finanserv 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 Source: https://stats.oecd.org/Index.aspx?datasetcode=FDIINDEX# Concerns that China is not as open as other major economies Still the Made in China 2025 program, along with China’s have been compounded by the Made in China 2025 program. ongoing investment restrictions, have created some confusion This plan, from the Ministry of Industry and Information Technol- about the direction of policy. China is likely to get the most out of ogy (MIIT), aims to transform China into a hi-tech powerhouse its impressive investment in STEM students and R&D if it opens and focuses on ten industries: the remaining sectors of the economy and continues to improve • Artificial intelligence and quantum computing IPR protection. If, alternatively, it tries to artificially pump up the • Automated machine tools and robotics domestic content of favored sectors, that is likely to be a recipe • Aerospace for slow technological advance and ongoing trade conflicts. • Maritime equipment Every economy in the world has an opportunity to join GVCs • Modern rail transport equipment irrespective of the type of human and physical capital available • Self-driving and new energy vehicles in the economy. If the domestic economy has relatively higher • Power equipment skilled workers like Singapore or Hong Kong, China, they will join • Agricultural equipment the GVCs at higher value-added segment like designing or high- • New materials end services (like marketing, financial etc.). On the contrary, if the • Biopharma and advanced medicine economy has relatively more unskilled workers, it would join the MIIT’s plans call for rising domestic content for these sectors. GVCs in lower value added segments like assembly and pack- Other Chinese officials emphasize that these are indicative, not aging. Even if the economy joins at the lower value-added seg- mandatory targets. Mandatory domestic content requirements ment, it still helps the economy to generate more and better job would be a WTO violation. The IMF discussed these issues with opportunities for the unskilled workers. Every country needs to Chinese authorities during the 2018 Article IV consultation: “The assess how skilled (or unskilled) the workforce is, which region authorities stressed that their plan to develop strategic sectors it is located in and what comparative advantage it can exploit would be market-based…. The authorities clarified that the gov- to join the GVCs in a specific sector. Once it is integrated, to ernment did not set mandatory targets for domestic content…. enhance the value-addition (or move up the value-chain), follow- They emphasized that domestic and foreign companies would ing the Chinese example, the domestic economy needs to invest be treated equally in China’s effort to update ins industrial in upskilling workers, R&D and technology adoption by firms, as sector, noting that industrial policies needed to be market-ori- well as supporting ICT and physical infrastructure by convert- ented.” (p. 22) ing comparative advantage into competitive advantage. If the 152 • Technological innovation, supply chain trade, and workers in a globalized world economy tries to increase the DVA in exports by artificially sup - in reduced demand for the product as well as workers in the porting the inputs/intermediates by using tariffs and non-tariff exporting sector and will also affect the productivity and quality measures, it will increase the cost of production and make the of the domestic firms as well, adversely affecting the welfare in product less competitive in the international market, resulting the society. Should high domestic value added in exports be an objective of policy? • 153 Notes International Monetary Fund. 2018. Article IV Consultation. 1. It is common practice to present the aggregate trend for domestic Johnson (2018), Measuring Global Value Chains, Annual Review of Eco- content in exports, and then for manufacturing as most trade take nomics, forthcoming. place in this sector. Following Johnson and Noguera (2016), the total Kugler, Maurice and Eric Verhoogen (2012), “Prices, Plant Size, and Product economy is grouped into four categories, Agriculture, Forestry, and Quality,” Review of Economic Studies, vol. 79(1), pages 307-339. Fishing, Non-Manufacturing Industrial Production, Manufacturing, Kummritz, V. (2014), Global Value Chains: Benefiting the Domestic Econ- and Services. omy, IHEID Working Papers, 02-2015, Economics Section, The Gradu- 2. Though, this difference in policy is not the only reason for poor ate Institute of International Studies. exporting performance by Pakistan but one of the primary reasons. Levchenko, A.A. (2007), Institutional Quality and International Trade, During the same period, Pakistan experienced many security issues Review of Economic Studies, Oxford University Press, Vol. 74(3), pp. and crippling power outages. 791-819. 3. We are using 2014 as it is the latest year in GTAP (version 10) data- Li, S., J. He, H. Zhang, and L. Zhu (2018). “Should High Ratio of Value base, recording the data for these two economies Added to Gross Export Value Be an Objective of Policy – Based on 4. The aggregate DVA for Bangladesh (67.6%) is also lower as compared China’s Provincial Level Data,” working paper, Development Research to Pakistan (82.6%) despite having higher total exports as well. Center, Beijing. Lopez-Acevedo, Gladys; Robertson, Raymond. (2016) Stitches to Riches? : Apparel Employment, Trade, and Economic Development in South References Asia. Directions in Development--Poverty;. Washington, DC: World Bank. https://openknowledge.worldbank.org/handle/10986/23961 Amiti, Mary, and Jozef Konings (2007), “Trade Liberalization, Intermediate License: CC BY 3.0 IGO. Inputs, and Productivity: Evidence from Indonesia,” American Eco- Lopez Gonzalez, J. (2016), Using Foreign Factors to Enhance Domestic nomic Review 97 (5): 1611–38. Export Performance: A Focus on Southeast Asia, OECD Trade Policy Bas, M. and V. Strauss-Kahn (2014), Does importing more inputs raise Papers, No. 191, OECD Publishing, Paris. exports? Firm-level evidence from France, Review of World Economics, Melitz, M.J. (2003), The Impact of trade on aggregate industry productivity Vol. 150(2), pp.241-275. and intra-industry reallocations, Econometrica 71, 1695-1725. Bas, M. and V. Strauss-Kahn (2015), Input-trade Liberalisation, export prices OECD. 2014. OECD International Direct Investment Statistics 2014, OECD and quality upgrading, Journal of International Economics, Vol. 95(2), Publishing, Paris, http://dx.doi.org/10.1787/idis-2014-en 250-262. R. Johnson and G. Noguera (2016). A Portrait of Trade in Value Added over Caesar B. Cororaton and David Orden, (2008) Pakistan’s cotton and textile Four Decades. NBER Working Paper No. 22974. economy: Intersectoral linkages and effects on rural and urban poverty Region’s Input-Output Tables for the Republic of Korea, Japan, and Chinese No 158, Research reports, International Food Policy Research Institute Taipei, various years. (IFPRI) Tang, Heiwai; Qi, Steffan; Lai Edwin (2018) Global Sourcing and Domestic Claire Hollweg, Daria Taglioni, Tania Smith (2017) Viet Nam at Crossroads: Value Added in Exports Working paper Engaging in the Next Generation of Global Value Chains Wad, P. 2009. “The Automobile Industry of Southeast Asia: Malaysia and Dollar, D. (1992), Outward-oriented developing economies really do grow Thailand,” Journal of the Asia Pacific Economy, Volume 14, Issue 2. more rapidly: Evidence from 95 LDCs, 1976-85. Economic Development The World Bank. 2018. World Development Indicators 2018. and Cultural Change 40 (3): 523-544. U.S. Chamber of Commerce. 2017. Made in China 2025: Global Ambitions Eaton J. and S. Kortum (2002), Technology, Geography and Trade, Econo- Built on Local Protections. metrica, Vol. 70(5), pp 1741-1779. World Integrated Trade Solution Goldberg, Pinelopi K., Amit K. Khandelwal, Nina Pavcnik, and Petia Topalova (2010), “Imported Intermediate Inputs and Domestic Product Growth: Evidence from India,” Quarterly Journal of Economics 125 (4): 1727–67. CHAPTER 8 Improving the accounting frameworks for analyses of global value chains Nadim Ahmad (OECD) ABSTRACT The use of global input-output tables, and the creation of content of exports for the economy as a whole, a key indica- Trade in Value-Added (TiVA) statistics, has greatly improved tor characterizing global production. Moreover, TiVA analyses our understanding of the fragmentation of global production are based on basic price concepts, which provide an appro- through value chains. However, their application requires a priate view of production through value chains, but are less number of assumptions that, in practice, typically understate well equipped to analyse consumption, particularly as they the degree of interconnectedness. TiVA estimates implicitly exclude significant distribution margins (in particular retail assume identical production functions across firms within an and wholesale activities, often including marketing activities industry, when in reality production functions differ consid- and brands), which add value at the end of the chain. This can erably. Typically, larger (and foreign-owned) firms tend to be distort analyses using “smile curves”, which show the distance more trade oriented than smaller (and domestically-owned) from final demand of different sectors within value chains, firms. As a result, TiVA statistics underestimate the import and in turn understate the scale of jobs supported by trade. • Trade in Value-Added (TiVA) statistics have greatly improved our understanding of GVCs, but they use assumptions that generate typically downward biases in measures of GVC integration, and they give little information regarding the investment strand of GVCs. • Efforts to mainstream key characteristics of different types of firms in the production of tomorrow’s TiVA models, through extended supply-use tables, should be prioritized, to improve not only their relevance, but also their quality. • Efforts to complement TiVA estimates currently based on basic prices with estimates based on market prices should also be initiated, not only to ease interpretability, but also to highlight the significant role played by distributors and to better understand the role played by intellectual property. Market-based approaches, for example, reveal that 9 million jobs are sustained in the United States through sales of imports. 155 156 • Technological innovation, supply chain trade, and workers in a globalized world 1. Introduction T This chapter highlights the importance of developing exten - sions to current TiVA frameworks (Section 2) that are better he proliferation and development of global input-out- able to capture firm heterogeneity, and, in turn, better high - put tables in recent years has significantly transformed light the importance of multinational enterprises (MNEs) within our ability to interpret global production. But import- GVCs. It also explores the development of a complementary ant though such initiatives have been, it is important to accounting framework in “market” prices and tries to illustrate recall that they are analytical tools, requiring implicit and explicit the insights that can be gained through such an approach assumptions on the detailed interactions of consumers and pro- (Section 3). In the United States the sale of imports generated ducers, and indeed, in their current form are silent on many driv- an additional 840 billion USD of US value-added in 2016, sup - ers of globalization, such as the role of multinationals, and on porting 9.0 million jobs. impacts, for example with respect to “inclusive globalization”. Trade in Value-Added (TiVA) estimates, derived through the construction of a global input-output table, implicitly assume 2. Accounting frameworks for global value that all firms within a given sector have the same production chains: extended supply-use tables function (input-output technical coefficients), import intensity and export intensity. 2.1 Overview This of course has never been true. We know for example that The increasing international fragmentation of production that larger firms will typically have different production functions to has occurred in recent decades driven by technological progress, smaller firms, because of economies of scale, and also higher reductions in trade costs, improved access to resources and mar- labor productivity. And these firms will also typically be more kets, trade policy reforms, and indeed cost factors in emerging export- and, indeed, import-orientated than their smaller coun- economies, has challenged our conventional wisdom on how we terparts (reflecting in part the disproportionate costs of trade look at and interpret globalization. For example, traditional mea- faced by smaller firms compared to larger firms). The same gen- sures of trade record gross flows of goods and services each and eralizations hold true for foreign-owned enterprises, or enter- every time they cross borders, leading to what many describe prises with affiliates abroad, compared to purely domestic firms; as a “multiple” counting of trade, which may lead to misguided for example, the foreign content of exports by foreign-owned policy measures in a wide range of policy areas. In response to firms in the other transport equipment sector in the United this, the international statistics community has begun to develop States is twice that of domestically-owned firms. new measures of trade on a value added basis, for example the That is not, of course, to say that the underlying conceptual OECD-WTO TiVA database, WIOD, APEC-TIVA and the Euro- basis for TiVA is incorrect. If, for example, global input-output pean FIGARO initiative. tables were compiled at the firm level, with appropriate break- But important though such initiatives are, they are only able downs to reflect the specific products and the (often differential) to respond to one aspect of the globalization debate. Significant prices paid by consumers (as well as differences in transportation attention, for example, is focused on the role of multinationals costs), then the corresponding results would accurately reflect in this new landscape, and, on this, with the exception of recent the underlying reality they seek to measure. But, for many prac- exploratory initiatives2, current available, and in particular offi- tical reasons1, this is some way off what happens in practice. Fur- cial, statistics that follow the TiVA approach are silent. Of particu- ther, the inability to capture this heterogeneity in current TiVA lar relevance in this context is the ability of multinationals to shift measures is increasingly compounded by additional complexi- intellectual property products, such as software and R&D, from ties, notably the increased scope for multi-nationals (MNEs) to one economic territory to another, raising broader questions on maximize global profits by recording intra-firm transactions in the ability of GDP to accurately describe “meaningful” economic knowledge-based services in a way that is most advantageous to activity, with concomitant impacts on other macro-economic sta- the firm. In practice this means that these types of intrafirm trans- tistics, including TiVA. For example, TiVA measures purport to actions can be recorded explicitly as cross-border trade or (and show how (in which industries) and where (in which territories) so outside of the TiVA system) as primary income flows. value is generated in the production of a good or service. The But this is not the only area where there are challenges simple relocation of an intellectual property product from one with the use of current TiVA statistics. Because inter-country economic territory to another3 can radically alter that view. input-output tables value transactions at basic, and not market, In addition, the policy debate in recent years has increasingly prices, many of the related TiVA analyses reveal only part of the focused on what has become referred to as “inclusive global- story. For example the US domestic value-added content of its ization”, i.e. the growing realization that the benefits of global- exports of textiles and clothing, in free-on-board (F.O.B.) prices, ization may not have accrued to all members of society equally, was around 20% in 2016 using market prices, compared to 3% even if only as a process of transition. The challenges of inclusive using the pure basic prices approach. The basic price approach globalization require that the impacts on people (in other words, also limits the scope to reveal additional dependencies related workers) are also captured in our statistics. This requires infor- to globalization, for example jobs sustained in retailers through mation on skills, occupations, and compensation paid to these sales of imports. categories of workers. Improving the accounting frameworks for analyses of global value chains • 157 2.2 Improved accounting frameworks for GVC analyses This of course has never been true. We know for example that More fundamentally, there is a growing appreciation that the sta- larger firms will typically have different production functions than tistical compilation tools and accounting frameworks designed smaller firms (because of economies of scale) as well as higher and developed over the last 60 years in various manifestations of labor productivity. And these firms will also typically be more the System of National Accounts (SNA), despite their significant export- and, indeed, import-orientated than their smaller coun- advances, may reflect a world that no longer exists. terparts (reflecting in part the disproportionate costs of trade In the early days of the SNA, the rest of the world was faced by smaller firms compared to larger firms). recorded as a separate institutional sector to and from which The same generalizations hold true for foreign-owned enter- goods were sold and bought; and such a view was largely suf- prises, or enterprises with affiliates abroad, compared to purely ficient. But over the years as global production chains and inter- domestic firms; indeed in many countries MNEs account for the connectedness grew, there was a growing realization that addi- lion’s share of overall trade (Figure 8.1). But TiVA estimates, rely- tional information was needed to properly navigate around the ing as they do on national Supply-Use and Input-Output tables, economic landscape, which resulted in the development of new cannot reflect these heterogeneities; meaning that key mea- areas of statistics, such as foreign direct investment measures sures, such as the import content of exports are typically down- and data collections focusing on inward and outward activities ward biased. of foreign affiliates statistics (FATS). More recently new data col- Moreover, the very process of globalization has increased the lections, or rather compilations, have focused on linking trade scale of these heterogeneities, driving coach and horses through and business registers to provide insights on which firms in which the assumption of homogeneity within sectors. As firms within sectors engage in imports and exports (referred to as Trade by sectors increasingly specialize in specific tasks in the production Enterprise Characteristics). process, they also suck in greater imports from the upstream These more recent innovations have significantly improved part of the value chain and have greater export orientation. In our collective understanding of trade, and indeed investment, addition globalization has itself led to an increased prevalence of but they are still to a large extent only a partial solution to the (once rare) categories of firms such as Factoryless Producers and statistical challenges presented by globalization. Processers, where recent changes in the accounting system fur- The development of TiVA type statistics is certainly a step ther weaken the case for assumptions of homogeneity in techni- forward in this area, but these too suffer from the stove-pipe cal coefficients. For example, all other things being equal, a pro - approach. TiVA estimates, derived through the construction of cessing firm in one sector will have significantly less (recorded) a global input-output table, implicitly assume that all firms within imports than a non-processing firm producing the same final a given sector have the same production function (input-out- product. Similarly, a Factoryless Producer will be allocated to put technical coefficients), import intensity and export intensity. the distribution sector (with limited intermediate consumption of FIGURE 8.1 Foreign-owned firms across economies (2011)  90% 80% 70% 60% 50% 40% 30% 20% 10% 0% a ria m y e en ic g y ic ly i nd nd r l nc an ur on bl ga ga bl do Ita st ed bo a pu la a m pu t Au tu un ng Fr nl Es Sw Po er m Re r Re Fi H Po Ki G xe h k d Lu ec a ite ov Cz Un Sl Foreign controlled as % of firms % exports % imports Note: Foreign-owned firms are defined according to FATS/AMNE 50% thresholds. Source: OECD Trade by Enterprise Characteristics. 158 • Technological innovation, supply chain trade, and workers in a globalized world goods) but the same firm that chooses to buy the material goods and out of the economy and, in particular, from which category of used by the processing firms will be allocated to the manufactur- firms (e.g. industrial sector) these arise. ing sector (with significant intermediate consumption of goods). In this sense it is important to note that value added essen- The ability of national (and international) Supply-Use and tially reflects two main components4 – (i) operating surplus Input-Output tables, based on industrial groupings alone, to (including mixed income), or compensation for capital, and (ii) describe how demand and supply relationships are related has compensation for employment. While the latter component therefore become more difficult. Typically, in confronting the largely reflects the direct benefits that accrue and “stick” within problem of heterogeneity, the conventional approach has been the economy through production5 the case is not so clear for the to provide more detail by aggregating firms at lower levels of the former, where foreign affiliates are concerned. industrial classification system, for example 3 or 4 digit group - In perfect markets the operating surplus generated by foreign ings as opposed to two digit groupings, subject to confidenti- affiliates is equivalent to the return on produced “tangible” and ality restrictions being preserved. But this approach may not be “intangible” capital and also non-produced assets used in pro- optimal, neither in terms of reducing heterogeneity within aggre- duction6. While the National Accounts of countries attribute the gations (and in a way that best responds to the policy drivers) nor ownership of this capital to the affiliated enterprise, the ultimate necessarily in terms of processing burdens. beneficiary of the operating surplus is not necessarily the affili- That is not to say that industrial classification systems are ate but its parent. This has raised questions – often in emerging completely obsolete. It would serve little purpose for example economies but also in developed economies – about the actual to devise an optimal system that did not retain some means of benefits of foreign MNEs to the host economy. Indeed, more classifying firms on the basis of their activity, (e.g. manufacturing recently it has begun to raise questions about the meaningful- versus services) if only because these remain the key prisms that ness of GDP itself as a tool for macro-ecomomic policy making. users look through when analyzing production. But it does serve Particularly important in this regard are transactions in intan- to highlight that other approaches to tackling heterogeneity can, gible assets: those recognised as produced in the SNA (such as and should, be considered. research and development, software, etc.), non-produced (such Arguably a more radical approach is needed. Such an approach as brands) and also other knowledge-based capital (such as requires that the role of foreign affiliates in the economic territory, organizational capital, e.g. management competencies). Often, which is significant in many economies, Figure 8.2, and affiliates in international trade in services statistics, payments for the use abroad are captured explicitly (and visibly) in the core accounts of these produced and non-produced assets are recorded as and in the development of GVC-related (i.e. TiVA) indicators. It purchases (intermediate consumption) by one affiliated enter- also requires improved information on the trade relationships prise from another. But often they are not, and instead they of categories of firms (for example exporter and non-exporter). are implicitly recorded under primary income payments (such Equally important is the need to fully articulate income flows in as investment income, or reinvested earnings in the Balance of FIGURE 8.2 Value Added at Factor Cost of Foreign Affiliates – share of national total, 2014 (ISIC B-N, ex K)  60% 50% 40% 30% 20% 10% 0% Sl ria e en Ki tvia ce Be m Fi ny he m Cz Re rg n k Sw y Re lic Po ia Po nia s Es c G y ry er l nd d ec nd ar a ga i ai l bl n do Ita iu ak ou ga b an st an w a ed Sp to e m la re La m ec pu pu lg rla rtu or Au ov ng nl un Fr ov mb en N H G D xe et h d Lu N ite Un Sl Source: OECD AMNE database. Improving the accounting frameworks for analyses of global value chains • 159 Payments). In the former case, the value added of the affiliate 2.3 National examples of extended supply-use tables using the assets is lower, as the value added generated through It is important to stress that the recognition that greater hetero- ownership of the asset appears on the accounts of the affiliate that geneity (disaggregation of firms) within national supply-use and owns it. In the latter case, however, the value added of the affiliate input-output tables is not of course new. It stands to reason that using the asset is higher (as there is no intermediate consumption) more detailed tables will produce better results. Indeed Chap- with the “ultimate” beneficiary (the owning affiliate) recording no ter 14 of the 2008 SNA provides a presentation of Supply-Use value added but instead recieving primary income from the using tables that differentiate production on the basis of market output, affiliate. In both cases, however, the ultimate “income” generated non-market output and production for own-final use. Historically by the asset ends up on the books of the owner (at least in theory, and certainly prior to the explosion in GVCs, capturing hetero - as even the very notion of the ultimate onwer is a complex issue). geneity was typically achieved through more detailed splits of Furthermore, the distinction between the two scenarios above industries. What has changed in recent years is the greater appre- is often clouded by (a) the ability of the statistical information ciation that a focus on the industries of firms is not necessarily system to record the flows and (b) transfer pricing and tax incen- best nor indeed optimal. Indeed, in 2011, even before the OECD- tives of MNEs. Indeed, in some countries where foreign affiliates WTO released their first TiVA database in January 2013, it had generate significant value added and repatriate significant prof- become clear that a new approach to heterogeneity was needed, its back to parent companies the policy focus has switched from in particular one that focused on the role of MNEs.8 GDP to GNI, and indeed in some countries, such as Ireland, to new These earlier discussions, and indeed the first release of TiVA, accounting concepts7. highlighted the importance of looking anew at national statistics This is not however an issue singularly related to knowl- compilation systems, with the OECD moving, in 2014, to create a edge-based assets. Transfer pricing is also prevalent in trans- new expert group of countries that would begin to develop what actions related to goods. Moreover, notwithstanding these have become known as ESUTs; in other words accounting tools for issues, significant income flows generated by an affiliate can be a coherent view of trade, investment, income and production (for repatriated to parents via other means, for example as interest a detailed exposition of the accounting framework of ESUTs see payments. Ahmad 2018). What follows below are national examples9 illustrat- The tool advocated in the SNA for ensuring coherence across ing the potential (and indeed actual for China and Mexico, whose various data sources to assure alignment of GDP estimates cre- extended tables are already integrated into the OECD-WTO TiVA ated by the income, expenditure and production approach is database) impact of improved heterogeneity on TiVA estimates. supply-use tables, the same underlying core statistical input required for TiVA estimates. As shown in this chapter, through (in Results for China principle) simple extensions to conventional supply-use tables, China has worked to develop extended supply-use tables that Extended Supply-Use Tables (ESUT) provide the ideal basis for differentiate between three categories of firms – exporters oper- bringing together these various domains into a single, integrated ating within the Customs Processing regime, other exporters, economic accounting framework that puts the measurement of and non-exporters. Figure 8.3 below reveals significantly dif- the “global” at the heart of the “national”. ferent movements in the trend of the foreign content of China’s FIGURE 8.3 Foreign value-added content of China’s exports  40 With split 30 for processing firms (ESUT) 20 Without split 10 for processing firms 0 1995 2000 2005 2008 2009 2010 2011 Source: OECD-WTO TiVA (May 2013 version). 160 • Technological innovation, supply chain trade, and workers in a globalized world FIGURE 8.4 US value added content of Mexico’s exports % (2011)  40% 35% 30% Domestically 25% owned 20% 15% 10% Global 5% manufacturers 0% m s g s ts s c qu les s s ls t t r od d tic al al en ile en pe e. ic rin uc ta eu oo ic et on c n. as xt Fo m m e Pa tu hi od em l -m tro W M Te r pl ip y ip ve ac t er pr ec qu on Ch Pe d uf in or el an al N le an te ch ot et nd er ca m r a M M po M ra bb tri er ns ec te th Ru tra pu O El m er Co th O Source: Based on Mexico’s Extended SUT. exports over the last two decades when comparing esti - imported intermediate consumption and for 71% of gross mates based on extended SUTs (referred to as ICIO) and pure exports of the Mexican manufacturing sector in 2008. Almost national tables without a breakdown (referred to as national). by definition the import content of Mexico’s global manufac - turing (GM) firms is significantly higher than comparable firms Results for Mexico in the same sector. This can have a significant difference on Mexico (Instituto Nacional de Estadística y Geografía – INEGI) highly policy relevant indicators, for example, on measures have produced a categorization of firms referred to as global of the US content of Mexico’s exports (Figure 8.4), where manufacturers10 that: a) import the majority of their purchases one-quarter of the exports by GM firms in the motor vehicle (imports account for at least 2/3 of their export value); b) sector reflect upstream US contributions, compared to around produce only for exports; and c) are controlled by a foreign half that amount for non-GM firms; this relationship is seen owner. These global firms were responsible for 55% of total across most activities. FIGURE 8.5 Foreign content of US exports, % (2011) (selected industries)  30 25 20 Non-MNE 15 Domestic MNE 10 Foreign MNE 5 0 Computer,electronic Electrical Motor vehicles Other transport Total economy and optical products machinery equipment Source: Based on the US Extended SUT. Improving the accounting frameworks for analyses of global value chains • 161 FIGURE 8.6 Foreign content of Costa Rica’s exports, % (2012)  70% 60% 50% 40% 30% 20% 10% 0% re e s s ts s es s ts es t r ry t t g l m ce en nt pe al en en nc re uc in uc tu ne iti ic er co ra vi pa em ra in m m pa ul rv od iv od hi in er au le M su ip ip ric ct se ap ac m ag pr d pr ss te st qu qu ta in Ag an er M re nd nd an es od al nd d or te le th d et an lm sin & sa sa pp Fo ta ca O oo or m ls Bu al e ile ia tri sp or su W te nc nd ic nc ec xt sp an ho ce em na sa Te na El an Tr ffi Fi l, Ch Fi al Tr ai O et et M ,r le sa le ho W Total RE RD Source: Based on Mexico’s Extended SUT. Results for the United States Results for Canada The United States (Bureau of Economic Analysis) has developed Results from a recent collaboration between the OECD and Extended SUTs with a three-way classification of firms reflecting Statistics Canada reveal that the impact of compiling ESUT esti- ownership structures, that differentiate between foreign-owned mates for the business sector, accounting for either ownership or affiliates operating in the US, domestically-owned MNEs, and trading status, was an increase in the overall foreign value added domestically-owned firms with no affiliates abroad.11 Results for content of Canada’s exports of 4 percentage points. Figure 8.7, the United States also reveal significant differences between the which shows that foreign-owned firms are responsible for a lower foreign content of exports across categories of firms defined by share of exports in value-added terms than in gross terms, high- ownership structure. At the whole economy level the foreign con- lights this higher propensity to import by foreign-owned firms, tent of US exports by foreign-owned firms is almost twice that of and, of course, the importance of capturing improved firm het- domestically-owned non-MNEs. This partly reflects compositional erogeneity in national SUTs. effects, but the foreign content is higher across nearly all activities (Figure 8.5) Results for Nordic countries In a recent collaboration between 5 Nordic countries (Denmark, Results for Costa Rica Finland, Iceland, Norway and Sweden) and the OECD, the OECD A similar picture of strong heterogeneity emerges for Costa Rica, developed extended SUTs with three variants of firm breakdown: whose ESUT differentiates between firms operating from free • By size class: micro, small, medium and large, further broken trade zones (referred to as RE in Figure 8.6) and firms operating down by whether the micro, small and medium firms were outside of foreign trade zones (FTZs) (referred to as RD). The independent or part of a larger enterprise group. results show that RE firms have a higher import content of exports • By trading status: non-traders, two-way traders, importers than RD firms across a range of important export activities. and exporters 162 • Technological innovation, supply chain trade, and workers in a globalized world FIGURE 8.7 Share of gross and value-added exports by ownership status, % (2010), Canada  1.0 0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0 nt ry, n g sin n de s de es ce es tri d r io tie ct il ou tio rin to g us an g ra o hu st te es al ran ic tra tra ct n in ili ec tu eh ta d re xt nd rv ta io tru nd n Ut re u ar r ac an e,fo le il se l i io ss w po se ga d ns ta ns sa ra at uf es e Re i d ans Co an e, ng ur ga in le ltu rm an at in nc d in ho hi lt iv M cu fo Tr us fis ricu an M pr na W In lb Fi er Ag an ta th To O Foreign value-added exports Domestic value-added exports Foreign gross exports Domestic gross exports Source: OECD/Statistics Canada. FIGURE 8.8 Exports in gross and value-added terms, % (2013), by ownership structure  100% 90% 80% 70% 60% 50% 40% 30% 20% 10% 0% Gross exports Value added Gross exports Value added Gross exports Value added Gross exports Value added of exports of exports of exports of exports Denmark Finland Norway Sweden Non-MNEs Non-MNEs, direct exports Non-MNEs, indirect exports via other non-MNEs Non-MNEs, indirect exports via domestic MNEs Non-MNEs, indirect exports via foreign-owned MNEs Domestic MNEs Foreign-owned MNEs Source: Nordic Countries in global value chains, 2017. Improving the accounting frameworks for analyses of global value chains • 163 FIGURE 8.9 Exports in gross and value-added terms, % (2013), by size class  100% 90% 80% 70% 60% 50% 40% 30% 20% 10% 0% Gross exports Value added Gross exports Value added Gross exports Value added Gross exports Value added of exports of exports of exports of exports Denmark Finland Norway Sweden Independent - Micro Independent - Medium Dependent - Small Large Independent - Small Dependent - Micro Dependent - Medium Source: Nordic Countries in global value chains, 2017. • By ownership status: non-MNEs, domestic MNEs and foreign foreign-owned MNEs, in large part reflecting scale. Figure  8.9 MNEs. presents a similar picture showing the higher integration of Highlights from this collaboration are presented below as Fig- smaller firms in GVCs when seen in value-added terms, through ures 8.8-10. Figure 8.8 reveals the significant upstream integration their upstream integration as suppliers to larger exporting firms. of non-MNEs across all countries, compared to integration seen Figure 8.10 presents information on jobs sustained through inte- looking purely at gross trade relationships. Of particular note is gration in GVCs. A significant insight from this presentation is the the fact that in all countries bar Sweden this integration is primarily fact that even within firms that have no direct exports, around one channeled via domestic MNEs; in Sweden the main link is through in six of all jobs in these firms are dependent on foreign markets. FIGURE 8.10 Jobs embodied in exports, % of total (2013), by trading status  70% 60% 50% 40% 30% 20% 10% 0% Exporters Importers Two-way traders Non-Traders Exporters Importers Two-way traders Non-Traders Exporters Importers Two-way traders Non-Traders Exporters Importers Two-way traders Non-Traders Denmark Finland Norway Sweden Direct exports Indirect exports Source: Nordic Countries in global value chains, 2017. 164 • Technological innovation, supply chain trade, and workers in a globalized world 2.4 Concluding comments Extended Supply-Use tables provide an effective tool to The statistical challenges of globalization are profound, and it respond to these developments and growing needs. Increas- has become increasingly clear in recent years that conventional ing globalization of production raises challenging questions for approaches used to understand how economies work can no national statistics. And fundamental and long-standing axioms longer rely solely on national statistics. Increasingly, in order to regarding the nature of production and the way that statistics are understand how economies work and how to target and create necessarily compiled warrant a rethink. Certainly the evidence industrial policies focusing on competitiveness, it is necessary suggests that long-standing assumptions concerning homoge- to see the whole. National statistics build pictures based on neity of firms within industry classifications should be reviewed. interrelationships between producers and consumers and the The evidence also suggests, particularly for those countries with rest of the world. But these relationships, particularly those with FATS and TEC data, that an optimal level of aggregation may the rest of the world, have become increasingly more complex, be achievable without any significant increase in compilation or and, as such, there is an increasing need to consider global reporting burden. But, of course, such reconsiderations need also production within a global accounting framework. This implies take into account constraints such as burdens and confidentiality. a departure from the traditional role of international organi- Supply-Use tables have become the conventional route with zations as compilers of internationally comparable national which coherent estimates of the national accounts, trade and statistics, such as national input-output or supply-use tables. production are now systematically compiled in many countries Instead, it requires that they bring together these national and lend themselves as being the ideal way in which to resolve tables to create a global table. these issues. Extended Supply-Use tables can play a similar role Although TiVA estimates have been able to shed important in responding to questions on globalization. light on our understanding of international trade and its rela- Three final comments, providing a broader perspective, are tion to activity and competitiveness, in particular the impor- worth making in this respect. The first concerns the quality of tance of recognizing the importance of imports to exports, national supply-use tables. In many (most) countries, such tables and, so, the hitherto hidden costs of protectionism as well are derived using a series of assumptions at least in some years, as the benefits of trade liberalization, particularly in services, reflecting in part the often different periodic nature of the large they do not reveal the full picture. With significant shares of number of datasets needed to construct SUTs. Many of these exports being driven by foreign affiliates, TiVA estimates have assumptions are based on some underlying view of stability and also revealed the importance of going beyond just value added homogeneity in production functions. As shown, globalization towards income, in order to capture flows outside of conven - is increasingly undermining the strength of these assumptions. tional international trade statistics, such as the repatriation of Looking again how the homogeneity is likely to manifest itself profits related to the use of non-produced knowledge-based across firms and creating SUTs based around these categori- assets (e.g. brands) and, indeed, the repatriation of profits zations of firms can greatly help to mitigate these effects and related to the use of produced knowledge-based assets (e.g. strengthen these assumptions, which will remain necessary, per- software) that are (often incorrectly) not recorded as receipts haps indefinitely, across most countries. As such, one important from exports of services. benefit of extended SUTs that should not be overlooked is their The emergence of global value chains therefore also raises ability to improve the quality of the core accounts, and indeed arguably profound questions about the way national statistics GDP. In the same way they are also ideally placed to be able to are currently compiled. In the same way that international orga- significantly improve the interpretability of the accounts, in par- nizations increasingly need to think “national” in the way they ticular, when the accounts are affected by phenomena related to present and compile their statistics, where “national” reflects globalization, such as relocations. the single economic territory comprising the “world” or large The second comment concerns the potential momentum parts of it, national statistics institutions need to think global. extended SUTs could provide to the development and improve - In other words, in the construction of national statistics ment of statistical business surveys. The evidence shows that greater emphasis is needed on the role of the rest of the world, significant heterogeneity exists across all categories of firms, both as a source of demand and supplier of demand but also and that the conventional stratification variables used in survey with regards to the role of multinationals. This requires a rethink sampling (typically activity and size) may be sub-optimal. It may of the way that firms are currently aggregated within statistical for example be necessary to include additional, but readily avail- information systems, to move beyond the classic aggregation able, stratification variables, pertaining for example to ownership based almost exclusively on industrial classification systems (e.g. part of a foreign MNE, domestic MNE, an enterprise group, towards more meaningful aggregations that better reflect exporter, non-exporter) in designing tomorrow’s surveys. today’s “global factory”. The third comes back to the issue of the statistical unit. The Such considerations are also essential not only to better current 2008 SNA preference for the establishment should not understand the way that global production is today organized be a barrier to developing extended SUTs. If for example these but also to better understand how investment drives global can only be developed using a different statistical unit, then value chains, and in particular how that very same investment countries are strongly encouraged to consider doing so. There is can lead to difficulties in interpreting trade flows as well as GDP. an increasing recognition that the arguments for the current SNA Improving the accounting frameworks for analyses of global value chains • 165 preference for the establishment have been weakened because service (and which, by definition, are excluded from the F.O.B. of the changing nature of production and indeed because of the price of the exported product, and indeed the cost, insurance changes made in the SNA itself regarding economic ownership. and freight (C.I.F.) price of an imported product). But this con- This is further recognized in the 2008 SNA Research Agenda, vention is by no means a panacea. where explicit references are made for the need to reconsider Larger enterprises within affiliated supply chains for exam- the establishment preference, taking into account the “basic ple are more likely (than independent smaller enterprises say) to source information” and changes in the underlying accounting include the costs of distribution in the basic price they charge principles of “Input-Output” tables, whose emphasis has moved (whether these are produced using in-house services or pur- from a physical perspective to an economic perspective. chased from third parties), and so, in these circumstances, no adjustments will be made to arrive at a basic price estimate, which will be equivalent to the F.O.B. price. So, as can be seen, 3. A new look at trade in value-added sometimes the distribution services are included in basic price and global value chains: a view from the measures and sometimes they are not, depending on how the consumption perspective – what the original producer chose to invoice them. accounting framework doesn’t tell you But this is not the biggest issue here: the removal of the margin generates an alternative perspective of the value of what 3.1 Overview is being traded (and Figure 8.11A reveals that this can be sig- In the SNA the recommended price basis for producers, and so, nificant) both from the exporting country’s perspective and the de facto in input-output tables, consumers, is the concept of importing country’s (exacerbating complications raised by the Basic Price12. In very simple terms this is equivalent to the fac- fact that import prices typically also include international distri- tory gate price, and so excludes any distribution margin not sub - bution margins). sumed in the original invoice price of the producer, and that are For any given export of a good therefore, because the domes- included in the price paid by the final consumer. Also excluded tic content of distribution services is typically high, the share of are any taxes paid or subsidies received on the product sold. domestic content of exports for a given good will be lower when Although superficially benign, the distinction between basic measured on a basic price basis than compared to estimates on a and purchasers prices matters, especially for GVC analysis. F.O.B basis (although, in theory, for exports of total, whole econ- Export prices are measured on a free on board (F.O.B.) basis omy, goods and services, the ratios should align) (see, for example, and include any distribution services related to delivery from the Figure 8.12). Similarly looking at imports of a particular good into factory gate to the port, and organized by the producer, but for an economy, a basic price measure will show a significantly smaller input-output tables in basic prices (when these margins are sep- (often implausibly low) contribution from the distribution and arately invoiced by the producer to the consumer or provided by transportation sector, compared to C.I.F measures. Basic price an intermediary that purchases and then exports the goods) they concepts also complicate and hamper analyses of the multiplica- are removed from the F.O.B. price and are instead re-allocated tive impact of tariffs, as, in a basic price format the rates, which as separate exports of distribution services (typically recorded are usually applicable to a C.I.F. price, will instead be applied to a as output of transportation services and/or output of the retail/ lower basic price; this underestimates the overall impact of tariffs. wholesale sector). Figure 8.1314 reveals the impact that different price bases On average these margins can be significant13, ranging at can have in interpreting the decomposition of value in GVCs around 10 and 15% across countries, and over 30% in Greece, by looking at the domestic services content of textiles exports. with significant differences by specific product, for example In the United Kingdom and Sweden for example the domestic 140% and 216% for textiles and clothing in the United King- services content jumps to around 70% compared to around 20% dom and Sweden respectively and 310% for pharmaceuticals in using the basic price concept. On average, across countries the Greece (Figures 8.11A and B). domestic services content of exports increases by around 15 per- Moreover, with respect to international input-output tables, centage points. a focus on the distribution margin provided in delivering a good Of particular interest in this respect is the contribution made from the factory gate to the customs frontier understates the size by the distribution sector (transport, retail and wholesale) in the of the problem related to the use of the basic price concept, as overall production of a given product, which is noticeably lower global input-output tables will also reallocate (to the distribution using the basic price concept (with well over half of the increase sector/product) the distribution margin related to the transpor- in domestic services value-added content reflecting distribution tation of the good from one frontier to another, and in turn the services in most countries). final distribution margin related to delivery from the frontier to The upshot is that by decoupling the distribution costs the final consumer. involved in transporting a good from the factory gate to the cus- In effect input-output tables at basic prices treat distribution toms frontier from the production costs of the good, the basic services as if they reflected the acquisition of a separate product. price concept creates an arguably downward-biased estimate The rationale is that this creates an equivalence with prices paid of the overall contribution of exports of that good to the local by consumers when they independently organize the distribution economy. Exacerbating this downward bias is the fact that the 166 • Technological innovation, supply chain trade, and workers in a globalized world FIGURE 8.11 Factory gate to exporting country’s customs frontier, recorded distribution margins (% of basic price of recorded exports) A: By product  70% 120 216 179 310 91 60% 50% 40% 30% 20% 10% 0% ve .c. um o re g s s tro d m ls ry te abr ic m ls er spo les sh g ng ts & tals pt pt t g m ic r l n al al pe cc ica in rin a a in e uc t st tu e Fi an t eq as ic tic et hi er ic n. le in rd ba m e re Pa ac cal opt ul tu od em h ot pl in tro M eu ip co Fo ric ,to ac r cl pr & ne equ Ch ac eq pe re Ag uf es er or e s& s lic er d Ba m c el cat & n t ag & ed oo ni al bb O Mo ar ra ile m g et i ry er fin W Ph t tin Ru xt tri m ec ev re Te ec in F n- th th hi ,b Pr & no El O r, od ke M er Fo pu Co th m O Co Average across countries Maximum across countries B: By country 70% 216 120 140 310 60% 50% 40% 30% 20% 10% 0% m . Sl ep. G ce Sw nia da ec le ite Tu n m D Rep La a lia o Li rea H ece Ca m Be tria ia Fi nia s es Es rk Un ing y of ly Ire ry d nd ov al Po d ni e et xic e Cz Chi tv ite do iu Ita a ra an an ga at ed na n rk Sl tug R e ua to Ko s re la lg rla N Me ov st Au St h Fr nl ak un en th Au r he d K p. d Re Un Share of of total exports Minimum across products Maximum across products Note: Data for Australia refers to 2014, Austria 2012, Belgium 2010, Canada 2013, Chile 2015, Czech Republic 2017, Denmark 2014, Estonia 2014, Finland 2014, France 2014, Greece 2010, Hungary 2014, Ireland 2011, Italy 2014, Republic of Korea 2010, Lithuania 2014, Latvia 2010, Mexico 2013, Netherlands 2014, Portugal 2013, Slovak Republic 2012, Slovenia 2015, Sweden 2015, Turkey 2012, United Kingdom 2010, United States 2016. Source: OECD Supply-Use Table database. Improving the accounting frameworks for analyses of global value chains • 167 FIGURE 8.12 International transportation margins basic price of the exported good will include all upstream distri- on US imports bution costs incurred in the production of that good, including cross-border distribution costs on intermediate imports used in production. So, in other words, distribution costs incurred in pro- 70 Thousands of product/country entries ducing a good for export will be reflected in the basic price of that good when they relate to intermediate parts shipped within 60 the country or imported into the country but, typically, not when 50 they relate to transportation of the goods to the customs frontier. In addition, the concept proves problematic for notions of 40 international competitiveness, as the basic price concept de 30 facto gives the impression that countries are engaged in signif- icant direct exports of these distribution activities, as any distri- 20 bution costs related to the transport of a good from the factory gate to the customs frontier will be treated as if they were direct 10 exports of separate distribution services. For example, a country 0 may have restrictions on the provision of these services by for- 0 0,2 0,4 0,6 0,8 1 1,2 eign operators, as well as high relative prices that are absorbed only through the increased international competitiveness of CIF minus FOB margin as a share of CIF price goods-producing sectors purchasing these distribution services. This country is more likely than not to reveal relatively higher Source: OECD International Transport and Insurance Costs of Merchandise measures of revealed comparative advantages (when measured Trade. on the conventional gross basis) in the distribution sector and relatively lower in the goods producing sector, when the com- plete opposite is the more likely scenario. FIGURE 8.13 Domestic services value-added content of textiles exports (basic versus F.O.B. prices)  80% 70% 60% 50% 40% 30% 20% 10% 0% Au ia Be ria Ca m da Re ile D blic Es rk Fr a G e H ce Ire y nd ly a Li tvia M ia he o ov P s Re nd Sl lic Sw ia ite Tu n Un ing y om es nd r e ni re c e et xic Ita l n en a ga iu b at ra h an e rk st ed na la la to ua m Ko d pu La pu C re rla lg e St st ov o un en th Au of d K ite p. h ak d N ec Re Cz Un Sl Basic F.O.B Note: Data for Australia are 2014, Austria 2012, Belgium 2010, Canada 2013, Chile 2015, Czech Republic 2017, Denmark 2014, Estonia 2014, France 2014, Greece 2010, Hungary 2014, Ireland 2011. Italy 2014, Republic of Korea 2010, Latvia 2010, Lithuania 2014, Mexico 2013, Netherlands 2014, Poland 2013, Slovak Republic 2012, Slovenia 2015, Sweden 2015, Turkey 2012, United Kingdom 2010, United States 2016. Source: OECD Supply-Use table database and OECD-WTO TiVA. 168 • Technological innovation, supply chain trade, and workers in a globalized world FIGURE 8.14 Margins on household final consumption, % of basic price A: By product  600% 500% 400% 300% 200% 100% 0% ve .c. Ch um o re xt s,to g ub ceu ls m r & als ls ry te Fab sic als er spo les ish ts g pr ng op ls u f qpt eq ent pe ing m ic fin eco r al e cc in rin a ec ate eta ec ic & eta .e uc t st tu & Pap tic al plas ic tic i r c F le ag Min th rd ba m e re tn e hi ul tu od m m in m tro o ip rt Fo ric ac p e cl ne equ tro d Ag er or a s& lic r d e a m an be an t & ed oo B O Mo ar ac ical ic n ile m ke ting et ry r er r W Ph t tr ev R re Te Co in n- th th hi ,b el Pr & no El O od r, M er Fo th pu O m Co Average across countries Maximum across countries B: By country 600% 500% 400% 300% 200% 100% 0% m . Sl ep. G ce Sw nia da ec le ite Tu n m D Rep La a lia o Li rea H ece Ca m Be ria ia Fi nia s es Es rk Un ing y of aly Ire ry d nd ov al Po nd ni e et xic e Cz hi tv ite do iu a ra an an ga at ed na st rk Sl tug R e ua to Ko It C re la lg rla N Me ov st Au St h Fr nl ak un en th Au r he d K p. d Re Un Share of of total HHFC Minimum across product Maximum across products Note: Data for Australia are 2014, Austria 2012, Belgium 2010, Canada 2013, Chile 2015, Czech Republic 2017, Denmark 2014, Estonia 2014, Finland 2014, France 2014, Greece 2010, Hungary 2014, Italy 2014, Republic of Korea 2010, Lithuania 2014, Latvia 2010, Mexico 2013, Netherlands 2014, Portugal 2013, Slovak Repub - lic 2012, Slovenia 2015, Sweden 2015, Turkey 2012, United Kingdom 2010, United States 2016. Source: OECD Supply-Use table database and OECD-WTO TiVA. Improving the accounting frameworks for analyses of global value chains • 169 But it is equally important to note that this is not only an issue TABLE 8.1 Derivation of Apple’s gross margin on 30GB for decompositions of exports into their sources of value added. video iPod It affects all components of demand. For estimates of intermedi- Retail Price $299 ate consumption (or rather the coefficients of the Leontief matrix) the effects are mitigated by the fact that the distribution costs Distributor Discount (10%) ($30) will always be captured in the costs of production of a good, Retailer Discount ($45) whether embodied in the price of any intermediate used in pro- Sub-total (estimated wholesale price) $224 duction or treated as a separate cost. This reflects the fact that Factory Cost $144 intermediate consumption totals are always measured at market prices even if the components are recorded in basic prices. Estimated Apple gross profit $80 In other words, Leontief coefficients provide a theoretically correct view of the upstream impact of the production of a given Source: Dedrick et al, 2008. good, but only when the application is to determine the full upstream impact of production as opposed to consumption. All current TiVA estimates align with this production view, but many A simple way to illustrate shortcomings in current measures, of the applications are in fact looking at things from a consump- and in particular the basic price concept, is to reconsider how tion perspective. But in basic price Leontief systems, distribution they reflect single case studies, indeed case studies that have margins provided by an intermediary (such as a retailer) or mar- acted as motivators for much of the work, and new statistics on gins that are not part of an all-inclusive price charged directly by GVCs, that exists today. Perhaps the most well-known of these is Dedrick et al.’s sem- the producer, are stripped out of the consumption (market) price. inal 2008 work looking at the decomposition of value creation in Not surprisingly, these charges can make a significant difference an iPod (Table 8.1). to the overall price of a good (see Figure 8.14). As noted in their study, the factory gate price (roughly For products, taking an average across countries’ margins equivalent to what would be recorded in trade statistics) was adds (a low of) 31% to the basic price of printing products and less than half the total retail price, and, indeed, Apple’s con- (a high of) 113% for textile and chemical products (and 560% for tribution (measured as its gross profit), and compensation for basic metal products in Denmark). For countries, looking at total design, marketing and research and development, is com- consumption of goods in basic prices, margins add a further 41% pletely absent from the factory gate price. in Slovakia to 92% in Denmark. In this sense therefore any attempt to assess the full value None of that is to say that basic price approaches are with- chain, including Apple’s contribution, by decomposing only out merit. Far from it, as they provide the conceptually correct the factory gate price, will be severely compromised as the view of the decomposition of costs from a production perspec- high-value activities, R&D and design (which are generally tive. Moreover, as described below, they are also significantly positioned at the beginning of the value chain, Figure 8.15) and easier to calculate from current national accounting systems than marketing and distribution (at the end of the chain) are com- decompositions based on market price concepts. pletely absent from the decomposition. This is what is de facto But it is clear that some care is needed in interpretation. As done in decompositions of value using input-output tables at shown above, for analyses of global value chains, taking a per- basic prices, because, as noted above, the contribution from spective from purchasers’ prices rather than basic prices can distribution services, and very often R&D, marketing and present a significantly different picture of GVCs, for example design are shown as separate expenditure items also in basic concerning the contribution to the domestic economy of exports prices. of a given product. But the purchaser’s prices concept is perhaps An underappreciation of this shortcoming in the basic price also preferable in the derivation of other conventional analyses concept for GVC analyses of lengths and positions of activities in and metrics that rely on input-output based indicators. Perhaps value chains is widespread in the literature. For example, Degain chief in this respect concerns analyses of the now well-known et al. (2017)’s otherwise excellent paper “Recent trends in global Smile Curve, which is looked at in the following section. trade and global value chains” provides a decomposition of value added, showing the contribution made by various indus- 3.2 Looking anew at the Smile Curve tries and countries relative to their distance from the consumer Although, at least in recent years, there has been an improved and by their relative compensation per hour. understanding of the limits of GVC analyses that look at fragmen- Intuitively, all of their charts plotting relationships for various tation of production through the prism of Stan Shih’s Smile Curve, products (see below Figure 8.16, the example for China’s elec- even with these limits it remains an important looking glass. trical and optical equipment) show distribution activities (clas- A greater awareness that conventional statistics concerning sified as industry 20 in the Figure) close to the consumer (with fragmentation of production reflect the basic price rather than relatively high labor costs), where Degain et al. explain: “Post- the market price concept can further help improve our under- fabrication service industries with higher labor compensation per standing and limitations of basic price measures. hour – such as wholesale (20) and inland transportation (23) in 170 • Technological innovation, supply chain trade, and workers in a globalized world FIGURE 8.15 Conceptual framework of the Smile Curve  Value Added Marketing, Advertising Basic and applied and Brand management, R&D, Design, Specialized logistics, Commercialization After-sales services Manufacturing, Standardized services Marketing R&D Knowledge Knowledge Inputs Markets Location 1 Location 2 Location 3 Location 4 Location 5 Source: Mudambi, 2008.17 FIGURE 8.16 Smile Curve for China’s exports of electrical and optical equipment, 2009 (basic prices) Compensation per hour ($)  Source: Meng Ye and Wei 2017. Improving the accounting frameworks for analyses of global value chains • 171 the United States, Japan, Germany, and France – were the main it is not an accurate representation of the position of distribu- beneficiaries in the postfabrication stage of this GVC. China’s tion services when decomposing basic prices. Indeed it stands ICT goods exported to the United States, Japan, and Germany to reason that for very fragmented chains, distribution services had to be delivered to their domestic consumers mainly through would be needed throughout the production process and, so, those countries’ domestic wholesale and transportation service would be further away from the consumer than retail distribution industries.” services (which are almost entirely related to the provision of ser- However, therein (the bolded text) lies the misunderstanding vices to final demand consumers). It is only because, in practice, between the basic price and market price concept. Decomposi- estimates of the position of distribution services (i.e. distance to tions of the value of a good purchased as final domestic demand consumer) are calculated for the sector as a whole that results in into source industries using input-output tables in basic prices distribution services appearing close to consumers. This reflects do not capture the: the fact that distribution services provided to final consumers • final contribution made by domestic wholesale and transpor- make up the majority of overall distribution services, and, so, tation service providers delivering an import to final domestic swamp results for the overall position of the sector. This some- consumers; what intuitive result appears to have led many to conclude that • international distribution costs involved in shipping the good the distribution service component in decompositions of basic into the country; nor indeed the prices reflects the final distribution service at the end of the • shipping costs from the factory to the customs frontier of the chain – but this is not the case. exporting country. This is why Degain et al. estimate the contribution of the dis- 3.3 Marketing, design and R&D services tribution activities at generally no greater than 20%, while this Thus, an aggregated view of the position of the distribution chapter finds significantly higher estimates (around 40% when sector in global value chains is unlikely to accurately reflect the the decomposition is for an export, as in Figure 8.13, and signifi- position of intermediate services in a given production process cantly higher when the decomposition relates to the price paid when input-output tables used decompositions in basic prices. by the final consumer, as in Figure 8.14). But, because the remuneration for marketing, design and R&D To re-emphasize, what decompositions in basic prices do cap- services is also often bundled within the final distribution margin, ture (at least in theory) is the contribution of distribution activ- our understanding of the contribution of other underlying activ- ities related to transactions in intermediates, before the very ities – recorded as distribution activities – may be similarly last transaction recorded in input-output tables at basic prices. affected, i.e. their position in global value chains, estimated using So, for example, they include any distribution activities related input-output tables, may not necessarily align with where they to the intermediate consumption of any firm (whether those appear in the physical production process.18 intermediates were imported, in which case decompositions This is particularly relevant for the position of high-value tasks would include any related international distribution margin, or such as research and development and design. These should of produced domestically). This is because the production func- course appear at the beginning of the production process, but tion (input-output coefficients) of any given industry will always where they appear in input-output based estimates depends show total intermediate consumption at market prices, even if greatly on a number of factors. Chiefly these relate to whether all the separate components are broken down into basic price these activities are conducted by separate production entities or components. However, these decompositions will not capture whether they are conducted within the firm. Further complicat- any distribution margins related to final demand transactions ing matters is the industrial classification of the firm itself, dis- (whether household final consumption, general government final cussed in more detail below. consumption – although in practice this is not generally an issue If the R&D and design activities are conducted by separate as in most countries general government final consumption only units classified to these specific activities in input-output tables, records transactions in services – capital formation or, indeed, then input-output based approaches will be able to capture their exports, including exports of intermediates). appropriate position and indeed value contribution within GVCs. This reveals another potential problem with analyses that However, often these activities are conducted in-house for which present the position of these distribution activities within global there is no observable transaction, and in these cases their con- value chains. In all of these studies distribution activities find tribution is included within the value added of the main activity themselves positioned very close to the final consumer. This is, of the firm. For example, a retailer may outsource production of course, an accurate reflection of their overall positions when of clothing, but the value generated through brand, design, seen as a whole (i.e. in market prices), as an overall view would and R&D may instead (and often) appear as distribution margin. include the distribution services provided to final domestic Input-output based measures will therefore record (but not sep- demand (household and government consumption, consumption arately) the positions of the underlying R&D and design activi- of non-profit institutions serving households, and capital forma- ties in the same position as the firm’s main activity (distribution), tion). However, this is not an accurate reflection of the position which will not typically be at the beginning of the value chain.19 of these activities when they refer to the provision of distribu- This of course is not an issue unique to these types of tasks; tion services used to service intermediate flows – in other words any in-house activity not separately identifiable in input-output 172 • Technological innovation, supply chain trade, and workers in a globalized world tables is treated in this way (as are secondary activities that words, decompositions of goods in basic prices (and in particu- are separately identifiable when input-output tables are con - lar hi-tech goods) may, in practice, typically significantly under- structed on an industry by industry as opposed to product by estimate the contribution of R&D, marketing, design etc. to the product basis). production process (as they will instead be recorded as a sepa- But whilst this is a more generic problem with input-out- rate transaction of “direct” purchases of distribution services). put tables, it is perhaps most pertinent when it comes to R&D, design and marketing activities, where in-house production 3.4 A new perspective on the role of imports remains significant (certainly when considering the very high Another area, among many, where a purchaser’s price perspec- distribution margins on exports seen in Figures 8.11A and B). tive can provide an important complementary view to a basic Further exacerbating this is the increasing importance within price concept concerns the role of imports. One highly sensi- global value chains of factory-less producers, who outsource tive indicator produced in TiVA-type analysis is the domestic physical production whether at home or abroad, but control the content of a country’s imports, typically used to highlight the overall production process (focusing control on activities such potentially counter-productive impact of tariffs as they may as specification, design, R&D, marketing), which to some extent affect upstream domestic exporters. In the United States, the is a reflection of the upgrading process underpinning GVCs. US content of its total goods imports amounts to, on average, Current international standards for the classification of firms 6% in recent years (Figure 8.17). But bringing the imports into (ISIC Rev 4) classify factory-less firms that own no material inter- the country, in turn, generates distribution services, whether mediate inputs in the production process to the distribution the imports are for intermediate consumption, final domestic sector. As such the value added by these factory-less firms will consumption, or indeed for direct re-exports. 20 materialize in input-output tables as distribution margins, and, Conventional input-output approaches, using the basic so, are allocated to a separate activity to the good being pro- price concept, de facto decouple and break the link between duced when input-output tables are recorded in basic prices. these costs and the imported good. But a purchasers’ price In other words the value of the goods (whose production and approach treats the distribution services as integral, revealing, sales are controlled by these firms) will reflect the (contractor’s) in turn, much higher US “dependencies” (or US “content”) of factory-gate price but these prices will not include the intellec- its imports. Indeed changing the price basis, and decompos- tual property, design, brand etc. owned by the factory-less firm. ing the purchasers price value of an imported good reveals that These will instead materialize in the wholesale prices the firm the US content of its total goods imports (or rather the US val- charges to other intermediaries or indeed the final retail price ue-added generated by consumption of imports) amounted to if the factory-less firm sells the products through its own chain 30%21 of the overall price of those imports (excluding any con - of retailers. sumption taxes). For imports of textiles, the US content was as Further complicating matters, factory-less firms that own high as 50% for consumption by US households and 20% for some material intermediate inputs (even if they have no actual exports, compared to the 3% shown in TiVA. role in the physical transformation of those inputs) are classi- Indeed, the total value of distribution margins provided by fied to the activity of the good being produced. In these cir- US domestic operators in taking imports from the customs fron- cumstances input-output tables should record transactions tier to their next destination (to industries, final consumers, or between the factory-less firm and its contractors following the as re-exports) amounted to close to 900 billion USD dollars in recommendations for the treatment of goods for processing 2016, equivalent to 5% of GDP. In value-added terms, as the transactions in the System of National Accounts. But in practice distribution sector also requires imports for production, dis- this may not be the case, especially if the value of the material tribution activities added 840 USD billion to US GDP in 2016 intermediate inputs purchased by the principal is marginal, in on account of transportation and sales of imports, supporting which case national accountants may instead choose to record 9 million jobs, including 6.3 million in the wholesale and retail the output of the principal as if it were a distribution service, sector, and 1.0 million in the transportation sector, with signifi- (i.e. excludes the factory gate price of the good) even if the cant contributions from upstream industries (0.2 million in man- industry of the principal remains classified to manufacturing. ufacturing, and 1.6 million in all other activities) (Figure 8.18). Following the example of the iPod above therefore, In many other countries the contribution of distribution input-output tables that decompose the basic price value of services (as recorded in official supply-use statistics) to the the iPod will not record the contribution from Apple’s R&D, domestic economy through sales of imports is significantly design, brand etc. to the good itself if Apple is classified as a higher (Figure 8.19). Unsurprisingly, the contribution is larger, distributor (e.g. as a factory-less firm, in which case the contri- the smaller the economy (and the higher the dependency on bution will be shown separately under consumption of distri- imports). In Lithuania for example, where gross imports were bution margins) and may not do so in practice even if Apple equivalent to 78% of GDP in 2014, and the value added of the is classified as a manufacturer. Whatever the classification, any distribution and transportation sectors accounted for 28% of retail margins incurred by final demand consumers, whether GDP, the domestic value added generated through sales of charged by independent retailers or Apple stores, will never imports in the economy accounted for 22% of GDP. Of partic- be included in the decomposition of the basic price. In other ular interest is the contribution to GDP made via distribution Improving the accounting frameworks for analyses of global value chains • 173 FIGURE 8.17 US value-added content of imports at the frontier (% of basic price) and as percent of consumer’s price (excluding taxes) (2016)  60% 50% 40% 30% 20% 10% 0% t s od es ds s re g g s c en t ed als s t c r tc qu tc al en en be al ile al et et rin in tu cl re oo in e Fo m et et et in m xt m b d i um tu ul em eh er p m pe M m Ru lg M oo Te p p ric i ac t u ui i v le pu n Pa Ch ta eq W Ag uf eq no or tro To m le at an ot d rt Pe er Co ic ca an m M po br th tri er ry ns Fa O ec th ne tra At the frontier (Basic Price) El O hi er ac th M O Total all goods all consumers (inc margins) Intermediate Consumption (inc margins) HHFC (inc margins) Exports (inc margins) Note: MHHFC, Household final consumption. Source: Calculations based on OECD-WTO TiVA and OECD Supply-Use Table database. FIGURE 8.18 Jobs supported and value added via sales and export of imports in the US, by source (2016)  7 600 6 500 5 400 4 300 3 200 2 100 1 0 0 Import content Manufacturing Transportation Other activities Wholesale, retail of distribution Jobs (millions) Value-added (USD billions) (RHS) Source: Calculations based on OECD-WTO TiVA, OECD Supply-Use Table database and OECD National Accounts. 174 • Technological innovation, supply chain trade, and workers in a globalized world FIGURE 8.19 Domestic value added generated and persons employed through sales of imports, by source of demand % of GDP and persons employed  0,25 Other Exports HHFC IC Persons employed, % of total 0,2 0,15 0,1 0,05 0 th p. ng . Sw ce Ca en G m ite ech a Po ile Sl tria he ia Au e D ico Be m Es ia a a ak a Tu s Ki Rep lia s k Fi aly M y Ire y H gal nd Au d e Un Cz ad nd ni ite ore i ec ar r Re e et ven tv Sl ton iu do ga an an at Ch ra ed rk ex ua It m s la La re rtu lg n rla St st Un f K Fr nl un en o d ov o Li p. d N Re Note: Data for Australia are 2014, Austria 2012, Belgium 2010, Canada 2013, Chile 2015, Czech Republic 2017, Denmark 2014, Estonia 2014, Finland 2014, France 2014, Greece 2010, Hungary 2014, Ireland 2011. Italy 2014, Republic of Korea 2010, Lithuania 2014, Latvia 2010, Mexico 2013, Netherlands 2014, Portugal 2013, Slovak Republic 2012, Slovenia 2015, Sweden 2015, Turkey 2012, United Kingdom 2010, United States 2016. Source: OECD Supply-Use table database and OECD-WTO TiVA and OECD National Accounts. HHFC: household final consumption. IC: Intermediate consumption. services related to re-exports, accounting, for example, for over 3.5 Developing market-price input-output frameworks 3% of GDP in the Netherlands. Total persons employed22 (pro- Despite all the commentary above, it’s important to reiterate viding distribution and upstream services) are generally higher that decompositions of basic price transactions into the origins than shares of GDP, reflecting the lower labor productivity23 of their value contribution are not wrong, nor are they without seen in the distribution sector compared to other activities in the meaning. However, care is needed in their interpretation. economy. There are a number of areas where care is needed, but key is None of the above is to say of course that higher import the fact that they do not provide a view from the purchaser’s per- prices, or lower imports, will necessarily reduce the domestic spective. In this respect therefore, they cannot provide a whole value (direct and upstream) generated by distribution activities view of the value chain (in particular the distribution, marketing, nor the jobs supported, as consumers will be able to substitute retail channel at the end of the chain), nor are they necessarily production with domestically produced equivalents (where these well-equipped to provide insights on the contribution of design, exist). But if the higher import prices occur through, for example, marketing and R&D (for example because they are bundled with tariff measures, this may reduce the overall purchasing power of distribution services or because they are performed in-house by consumers (in addition to the potential reduction in competitive- manufacturers) nor on the actual positions of various activities ness of producers, including exporters) which is likely to have a within value chains. volume effect. This would, in turn, reduce value added gener- In addition, basic price decompositions can also introduce ated and jobs sustained through distribution activities related to asymmetric results for chains that are to all extent and purposes, the sale of imports. identical. For example, if a Korean producer used a Japanese Improving the accounting frameworks for analyses of global value chains • 175 shipping company to ship parts to be assembled in China before exclusively on their ability to trade internationally but you would being shipped and sold to US households, the decomposition of not be able to identify this in a standard input-output table at the import price recorded in input-output tables at basic prices basic prices (which would show they had no imports). in the United States would include all costs incurred up to the Developing such a complementary view in practice is, how- point that the goods left the factory gate in China – in other ever, far from trivial (see Ahmad 2019, forthcoming). It would, in words they would exclude the costs incurred in shipping the effect, require a very different presentation of the role of distrib- goods across the Pacific, which would be treated as a separate utors in the accounting framework. They would be shown either import of distribution services by US households from Japan. as providers of intermediate services, resulting in changing the Typically, the distribution costs involved in shipping the good value of output of industries from basic prices to purchasers’ from the Chinese factory gate to the Chinese frontier (which prices (excluding taxes on products), or they would be shown would also be recorded as a direct import of distribution services as purchasers of the goods they sold. Thus, the accounts would by the US) would also be excluded. However, if the same goods need to record the value of their output inclusive of the value of were assembled in Mexico, the basic price for the imports into the goods that they sell, and not just their margins. Both cases the US would include shipping costs across the Pacific and the are complex, posing, in turn, difficulties for analyses and indeed distribution costs incurred in China (as, theoretically, these would in compilation. be included in the intermediate consumption costs of the Mex- ican assembler24). As such, even if the assembly costs in Mexico and China were identical and the shipping route (i.e. Republic 4. Conclusions of Korea-China-Mexico-US) and costs were also identical, the Japanese content of the US imports would be higher for goods Basic price approaches to the development of global input-out- assembled in Mexico compared to the same goods assembled put tables provide important insights on the nature of global in China. value chains and have helped transform our understanding on Perhaps the main shortcoming with the basic price concept, international trade today. However they can be prone to signifi- however, is that it breaks the link between the good being sold cant misinterpretation, as shown in many of the studies that use and the final distribution services that are reliant on it. That is, them to infer positions of activities in global value chains. But, as any upstream domestic distribution services involved in ship- shown above, this is not the only area where misinterpretation ping a good across borders before it is eventually consumed can occur; for example through their removal of the distribution back in that same country for final consumption will be (at least margin on goods transported from the factory gate to the cus- in theory) recorded in the home-country’s content of its imports. toms frontier, they provide a view of trade in goods that is sig- However, the same distribution services used to ship the prod- nificantly different to that seen by analysts of trade, which often uct to the country’s frontier before it is finally consumed will not hampers their take-up, and indeed can impact on analyses (for be recorded in the home-country’s content of imports (the dif- example in calculations of the impact of tariffs, whose price is ference between the C.I.F. and the basic price), nor (generally) typically C.I.F. or F.O.B.). will any domestic distribution services engaged in shipping the Perhaps chief in this respect is the application of basic price good from its frontier to its final domestic consumer (the differ- models to questions that require a consumption perspective ence between the purchasers price and the CIF price, ignoring (which is, to some extent, at the heart of many of the applications taxes and subsidies). As such, there is a clear case to be made of standard Leontief analyses, which often look at the impact of an (as in Figure 8.17) for complementary insights based on the pur- increase in final demand on production). But a significant part of chasers’ price. the actual consumption price (be that a market price or a CIF price) The perspective necessarily needs to be complementary to, on which taxes and tariffs are applied includes significant distribu- and not as a replacement for, the basic price concept, as a pur- tion margins, and pure basic price models that treat distributors as chaser’s price perspective cannot meet all needs. For example, providing direct services to customers, break these links. in looking at, say, the multiplicative impact of tariffs on imports, one still needs to have a view of the actual price of the imports and not the actual price paid by the ultimate consumer after dis- tribution margins are included. Even here, however, while the basic price concept is better it is also imperfect, as tariffs are typ- ically imposed on the CIF price and not the basic price, and when they are not CIF prices they are typically the FOB price, and the difference as shown in Figure 8.11 above can be significant. The idea for a complementary view in this respect is a means of supporting a broader narrative, whether that be on the full upstream impact of exports, the domestic spillover from imports or the positions (and interpretation of positions) of industries within GVCs. Import-export wholesalers, for example, depend 176 • Technological innovation, supply chain trade, and workers in a globalized world Notes the wholesale and retail sector provided 96% of all output in 2016 in the corresponding product. 1. This chiefly relates to the fact that no statistical information system 15. However at the same time because of the decoupling, in practice, at in the world actually has this information for all firms (by product least with current estimates of TiVA, there is an impact on the source produced and consumer) but even if this were the case, the need to of the distribution services, as, typically, the allocation (before balanc- preserve confidentiality of respondents to statistical business surveys, ing in a global input-output) to partner country sources of the imports would make it impossible to release such firm-level data for public is based on the partner shares observed for actual direct imports (and consumption. also, often, as part of the balancing process, exports) of these same 2. http://www.oecd.org/daf/inv/investment-policy/trade-invest- services. ment-gvc.htm 16. Dedrick, Kraemer, Linden (2008): “Who Profits from Innovation in 3. Albeit a relocation that satisfies the accounting rules regarding eco- Global Value Chains? A Study of the iPod and notebook PCs”. nomic, as opposed to legal, ownership. See the 2008 System of 17. Mudambi, R. (2008). “Location, Control and Innovation in Knowl- National Accounts. edge-Intensive Industries”. Journal of Economic Geography, 8(5), 4. It also includes taxes and subsidies on production. 699-725. 5. Not all labor compensation will necessarily stick in the economy, for 18. For example if an Apple store pays explicit cross-border royalties for example for cross-border workers. the use of intellectual property (such as design, software) to an Apple 6. Such as land and other intangible assets not recognised as Intellectual subsidiary abroad every time an iPhone is sold, the position of the Property Products in the SNA. intellectual property will appear close to the end of the value chain 7. http://www.cso.ie/en/csolatestnews/pressreleases/2017pressre- using standard input-output estimation methods, despite the fact that leases/pressstatementmacroeconomicreleasesyear2016andquar- the design and software are fundamentally at the beginning of the ter12017/ value chain. 8. See Ahmad and Araujo 2011 “Measuring Trade in Value-Added 19. See also Chen, Los and Timmer (2018), Factor Incomes in Global and Income using Firm-Level data”. Available at: http://sitere- Value Chains: The Role of Intangibles, NBER Working Paper, 25242, sources.worldbank.org/INTRANETTRADE/Resources/Internal-Train- which attempts to estimate the underlying contribution made by ing/287823-1256848879189/6526508-1283456658475/7370147 intangibles. -1308070299728/7997263-1308070314933/PAPER_8_Ahmad_ 20. Of interest with respect to the treatment of re-exports is the consid- Araujo.pdf erable margin associated with the distribution services (e.g. handling, 9. Where results have been generated using national tables only – in transportation etc.) for re-exports. In the United States, around 200 other words the domestic content of imports is recorded as zero. billion USD of its total 2.3 trillion of imports in 2016 in C.I.F. prices, 10. http://www3.inegi.org.mx/sistemas/tabuladosbasicos/LeerArchivo. reflected re-exports. The handling (transportation etc.) of these aspx?ct=44462&c=33654&s=est&f=4 imports for re-export generated 33 billion USD of distribution mar- 11. http://scholar.harvard.edu/files/jorgenson/files/4a.1_paper.pdf gins. In basic price input-output systems that exclude re-exports and 12. See the 2008 System of National Accounts allocate bilateral flows on the basis of their final destination, it is not 13. Note that some care is needed in interpreting the margin values pre- possible to separately differentiate this activity from other distribution sented here. The varying degree, across countries, of implementation services, masking the role of re-exports. Allocations of bilateral flows of the 2008 SNA guidance on merchanting transactions may affect on the basis of country of consignment, with a separate distinction for cross-country comparability and may also explain the very high esti- re-exports, even if only in basic prices may be a better approach for mates of margins in some countries. For example in countries with the construction of global input-output tables. significant merchanting activities (typically recorded as a distribu- 21. Indeed, this may be an underestimate as the calculations for percent- tion margin) there will be a positive entry for the margin (merchant- ages of “basic prices plus margins” shown here do not account for ing service) exported, including under goods transactions, but there international transport margins (which can also be provided by US will not be a corresponding value of the exports of the goods being transporters). TiVA estimates exclude these costs in the basic price of merchanted (unless the periods when the merchant acquires and sells the imported good, but the US Supply-Use tables used to generate the goods differ, in which case margin ratios in the period when the the “market” price equivalent estimates use imports at C.I.F. prices. goods are acquired will be biased upwards as the acquired goods will 22. Note that persons employed rather than jobs (as in Figure 8.18) are appear as a negative export). shown here as fewer countries provide estimates of jobs by activity 14. Note that in industry by industry output tables distribution margins 23. Labor productivity measures should preferably be calculated on an provided directly by the exporting industry are included in the output “hours worked” basis. But for the purposes of this paper, persons and the value added of the industry. Figure 8.13 assumes that the employed and jobs are used to better reveal the number of individu- additional margins are provided only by the domestic distribution als dependent on sales of imports. industry and so will present marginally upward biased estimates of the 24. This would be the case whether the Mexican firm actually purchased additional contribution made by the sector; typically the contribution the goods from the Korean producer or was merely a contractor, and made by the distribution sector represents nearly all of the domesti- so is unaffected by the changes in the 2008 SNA concerning goods cally-produced distribution activity. For example, in the United States sent abroad for processing. Improving the accounting frameworks for analyses of global value chains • 177 References Degain, C, B.Meng and Z.Wang (2017), “Recent trends in global trade and global value chains”, Global Value Chain Development Report (2017), Ahmad, N and S. Araujo (2011) “Measuring Trade in Value-Added and Chapter 2 Income using Firm-Level data”. Fetzer, J, T. Howells, L.Z Jones, E.H. Strassner, and Z.Wang (2016) “Estimat- Ahmad, N and J. Ribarksy (2014), “Trade in Value Added, Jobs and ing Extended Supply-Use Tables in Basic Prices with Firm Heterogeneity Investment” for the United States: A Proof of Concept”, paper prepared for the the Ahmad N, (2018), “Accounting for Globalisation: Frameworks for Inte- Fourth World KLEMS Conference Madrid, Spain grated International Economic Accounts” Miao, G and F. Fortanier (2017), “Estimating Transport and Insurance costs Ahmad, N. (2019), “A new look at Trade in Value-Added and Global Value of International Trade” OECD Statistics Working Paper, No. 80 Chains: A view from the Consumption Perspective” OECD Statistics Mudambi, R. (2008). Location, Control and Innovation in Knowledge-Inten- Working Paper, forthcoming. sive Industries. Journal of Economic Geography, 8(5), 699-725. Chen,W, B.Los and M.Timmer (2018), “Factor Incomes in Global Value Ming Y, B. Meng, and S.Wei (2015), “Measuring Smile Curves in Global Chains: The Role of Intangibles”, NBER Working Paper, 25242, Value Chains” IDE_Jetro Discussion paper, 530. Dedrick,J, K.Kraemer, and G.Linden (2008): Who Profits from Innovation in Saborío, G.M. (2015), “Costa Rica: An Extended Supply-Use Table”. Paper Global Value Chains? A Study of the iPod and notebook PCs. prepared for 23rd IIOA Conference México City. De La Cruz, J. R.Koopman, Z.Wang and S.Wei (2011) “Estimating Foreign Statistics Denmark: Nordic Countries in Global Value Chains (2017) Value-Added in Mexico’s Manufacturing Exports” United Nations (2009): System of National Accounts, 2008. APPENDIX 1 Chapter Authors’ conference Final Programme 8 October 2018, WTO 154 rue de Lausanne, Geneva, Switzerland 9:00 - 9:15 Registration 9:15 - 9:45 Welcome & opening remarks David Dollar (Brookings Institution) Emmanuelle Ganne (WTO) 9:45 - 10:30 Presentation of Chapter 1 – Update on trends in GVCs Presenter Zhi Wang (UIBE) Discussant Nadim Ahmad (OECD) 10:30 - 11:15 Presentation of Chapter 2 – Labor market effects of GVCs in developed countries Presenter Victor Stolzenburg (WTO) Discussant Michael Ferrantino (WBG) 11:15 - 11:30 Coffee break 11:30 - 12:15 Presentation of Chapter 3 – Labor market effects of GVCs in developing countries Presenter Claire Hollweg (WBG) Discussant Stela Rubinova (WTO) 12:15 - 13:00 Presentation of Chapter 4 – The future of manufacturing, automation, and developing countries Presenter Daria Taglioni (WBG) & Satoshi Inomata (IDE-JETRO) Discussant Yuqing Xing (National Graduate Institute for Policy Studies, Tokyo) 13:00 - 14:00 Lunch break 14:00 - 14:45 Presentation of Chapter 5 – GVCs and digital technologies Presenter Michael Ferrantino & Emine Elcin Koten (WBG) Discussant Satoshi Inomata (IDE-JETRO) 179 180 • Technological innovation, supply chain trade, and workers in a globalized world 14:45 - 15:30 Presentation of Chapter 6 – The digital economy, SMEs, and GVCs Presenter Emmanuelle Ganne & Kathryn Lundquist (WTO) Discussant Abdul Abiad (ADB) 15:30 - 16:15 Presentation of Chapter 7 – Should high domestic value added in exports be an objective of policy? Presenter David Dollar (Brookings Institution) Discussant Gaaitzen de Vries (Groningen Growth and Development Centre) 16:15 - 16:30 Coffee break 16:30 - 17:15 Presentation of Chapter 8 – Issues in GVC measurement Presenter Nadim Ahmad (OECD) Discussant Christophe Degain (WTO) 17:18 –18:15 Review of the Executive Summary Presenter David Dollar (Brookings Institution) 17:18 - 18:15 Wrap-Up Zhi Wang (UIBE) APPENDIX 2 Technological Innovation, Supply Chain Trade, and Workers in a Globalized World Global Value Chain Development Report 2019 Background Paper Conference GuoBin Hotel, Beijing, March 22-23, 2018 Pre-conference of China Development Forum Organized by RCGVC_UIBE and China Development Research Foundation. Co-sponsored by China National Science Foundation and Bill & Melinda Gates Foundation March 22, Guobin Hotel 8:30 - 9:00 Opening Remark 9:15 - 9:45 Welcome & opening remarks Speakers Dr. Chen Deming, Former Minster of MOFCOM Dr. Robert Koopman, Chief Economist, The WTO Professor Zhao Zhongxiu, Vice President, UIBE Mr. Lu Mai, General Secretory, CDRF 9:00 - 10:00 The Multilateral Nature of Bilateral Trade in the Age of Global Value Chains Speaker Wang Fei, Wang Zhi, UIBE, Wei Shang-jin, Columbia University, and Zhu Kunfu , UIBE Discussant Satoshi Inomata, IDE-JETRO 10:00 - 11:00 Did global value chains contribute to rising labor market polarization? Speaker Cosimo Beverelli, Victor Stolzenburg and Stela Rubinova , WTO Discussant Ma Hong, Tsinghua University 11:00 - 11:15 Tea break Participant All 11:15 - 12:15 Global Value Chain Participation and Labor Market Outcomes at the Macro and Micro Level Speaker Claire Hollweg, Jose Guilherme Reis and Deborah Winkler, World Bank Group Discussant Jiyoung KIM, IDE-JETRO 181 182 • Technological innovation, supply chain trade, and workers in a globalized world 12:15 - 13:30 Lunch Participant All 13:30 - 14:30 Accounting for Globalization: Frameworks for Integrated International Economic Accounts Speakers Nadim Ahmad, OECD Discussant David Dollars, Brookings institution 14:30 - 15:30 Re-examining the Impact of the China Trade Shock on Local US Labor Markets: A Supply Chain Perspective Speakers Zhi Wang, Shang-jin Wei, Xinding Yu and Kunfu Zhu, RCGVC Discussant Cosimo Beverelli, WTO 15:30 - 15:45 Tea break Participant All 15:45 - 16:45 From China with Love Speakers Mauro Boffa, Gianluca Santoni, and Daria Taglioni, World Bank Group Discussant Gaaitzen de Vries, University of Groningen, the Netherlands 16:45 - 17:45 The Changing Structure of the Global Value Chains and Domestic Firms’ Productivity: Evidence from Japanese and Chinese Firm-Level Data Speakers Yoshihiro Hashiguchi Keiko Ito, Chiara Criscuolo, Jonathan Timmis, IDE-JETRO Discussant Nick Hope, Stanford Center for International Development 17:45 - 18:45 The Future of Supply – How Might Industry 4.0 Impact Global Value Chains ? Speakers Michael Ferrantino, World Bank Group Discussant Emmanuelle Ganne, WTO March 23, Guobin Hotel 8:00 - 9:00 Are the Geese Still Flying? Evidence from Manufacturing FDI Mary Hallward-Drimeier and Gaurav Nayyar, World Bank Group Speakers Mary Hallward-Drimeier and Gaurav Nayyar, World Bank Group Discussant Jiandong Ju, Tsinghua University 9:00 - 10:00 Regional divergence in China: the Perspective of Value Chain Speakers Shan-tong Li, He jianwu (DRC) Discussant Deborah Winkler, World Bank Group 10:00 - 10:15 Tea break Participant All 10:15 - 11:15 Hollowing Out and Slowing Growth: A Perspective from Heterogeneous Technological Change Speakers Wenbo Zhu, RCGVC Discussants Heiwai Tang, John Hopkins University 11:15 - 12:15 Value added in China’s mobile phone handset industry Speakers Tim Sturgeon, MIT, Eric Thun, OXFORD, Daria Taglioni, World Bank Group Discussants Yuqing Xing, National Graduate Institute for Policy Studies, Japan 12:15 - 13:15 Lunch Participant All 13:15 - 14:15 Formerly Assembled, But Now Designed in China? An Exploration of Chinese Activities in Global Value Chains Speakers Quanrun Chen, Yuning Gao, Jiansuo Pei, Gaaitzen de Vries, Fei Wang Discussants Bo Meng, IDE-JETRO Authors’ conferences • 183 14:15 - 15:15 The effect of production fragmentation on Skill reallocation: Is it felt equally across levels of development? Speakers David Dollar, Matthew Kidder and Bilal M. Khan, RCGVC Discussants Liping Zhang, Research Center at The State Council 15:15 - 15:30 Tea break Participant All 15:30 - 16:30 How can the digital economy help small and medium-sized traders in developing countries to integrate into the global economy? Speakers Rainer Lanz, Kathryn Lundquist, Andreas Maurer and Robert Teh, WTO Discussants Xin Cheng, Alibaba Research Institute 16:30 - 17:30 Corruption, Import Liberalization, and Productivity in China: A Firm-Level Analysis Speakers Jiankun LU, Bo MENG, Hongsheng ZHANG, Shang-Jin WEI Discussants Miaojie Yu, CCER 17:30 - 18:30 Keynote Speech: The impact of new technology on future Jobs Speakers Michael Spence, Nobel Prize Laureate in Economics 18:30 - 18:45 Closing remarks Speakers Dr. Robert Koopman, Chief Economist, WTO Dr. David Dollar, Senior Fellow, Brookings Institution Wang Zhi, Professor and Director, RCGVC This report takes stock of the evolution of global value chains (GVCs) in light of technological developments, such as robotics, big data and the Internet of Things. It discusses how these technologies are reshaping GVCs and examines the effect of these changes on labor markets in developed and developing economies and on supply chain management. The report discusses how technological developments are creating new opportunities for the participation of small and medium-sized enterprises in global value chains and reviews issues related to GVC measurement. The report is a follow-up to the first Global Value Chain Development Report, which revealed the changing nature of international trade when analyzed in terms of value chains and value-added trade.