The World Bank Economic Review, 36(2), 2022, 329–360 https://doi.org10.1093/wber/lhab017 Article Determinants of Global Value Chain Participation: Downloaded from https://academic.oup.com/wber/article/36/2/329/6359883 by LEGVP Law Library user on 08 December 2023 Cross-Country Evidence Ana Margarida Fernandes, Hiau Looi Kee, and Deborah Winkler Abstract The past decades have witnessed big changes in international trade with the rise of global value chains (GVCs). Some countries, such as China, Poland, and Vietnam rode the tide, while other countries, many in the Africa region, faltered. This paper studies the determinants of countries’ GVC participation, based on a panel database of more than 100 countries from 1990 to 2015. Results from a three-pronged empirical approach show that factor endowments, geography, political stability, liberal trade policies, foreign direct investment and domestic industrial capacity are very important in determining GVC participation. These factors matter more for GVC trade than traditional trade. JEL classification: F13, F14, F23, O2 Keywords: global value chain, factor endowments, trade policy, foreign direct investment, institutions 1. Introduction In the early 1990s, Argentina tried to develop a homegrown auto industry, hiding behind an average tariff of more than 13 percent. Over the past two decades, Argentina’s auto exports have stagnated at a dismal 0.2 percent of global auto exports. Around the same period, General Motors (GM), one of the world’s largest automakers, set up GM Poland to import Opel cars for the large Polish domestic market. In 1994, production activities of GM Poland started, and today Poland is one of the world’s major auto exporting countries. Similar to the auto industry in Poland, Vietnam’s electronics export sector expanded sharply in less than a decade, fueled by foreign direct investment (FDI). Today, Vietnam is the world’s second largest Ana Margarida Fernandes is Lead Economist at the Development Research Group at the World Bank; her email address is afernandes@worldbank.org. Hiau Looi Kee (corresponding author) is Lead Economist at the Development Research Group at the World Bank; her email address is hlkee@worldbank.org. Deborah Winkler is Senior Consultant in Macroeconomics, Trade and Investment Global Practice at the World Bank; her email address is dwinkler2@worldbank.org. This paper is a substantial revision of the background paper “Factors Affecting Global Value Chain Participation across Countries,” prepared for the World Development Report 2020, Trading for Development in the Age of Global Value Chains. The authors thank Pol Antràs, Caroline Freund, Penny Goldberg, Aaditya Mattoo, and other colleagues for helpful comments. We are grateful to Alejandro Rojas, Bishakha Barman, and Natalia Camelo for research assistance and to Farid Toubal, Mary Hallward-Driemeier, and Gaurav Nayyar for sharing data on English language and foreign direct investment. We acknowledge financial support from the World Bank’s Multidonor Trust Fund for Trade and Development and the Strategic Research Program on Economic Development. Finally, we thank editor Nina Pavcnik and two anonymous referees for their helpful comments. The findings of this paper are those of the authors and do not necessarily represent the views of the World Bank, or its member countries. A supplementary online appendix for this article can be found at The World Bank Economic Review website. © The Author(s) 2021. Published by Oxford University Press on behalf of the International Bank for Reconstruction and Development / THE WORLD BANK. All rights reserved. For permissions, please e-mail: journals.permissions@oup.com 330 Fernandes, Kee, and Winkler Figure 1. Global Value Chain Participation across World Regions in the 2010s. Downloaded from https://academic.oup.com/wber/article/36/2/329/6359883 by LEGVP Law Library user on 08 December 2023 Source: Authors’ analysis based on data from Eora. See the supplementary online appendix for a description of the variables used. Note: Backward global value chain (GVC) participation measures the import content of exports relative to total exports. Averages shown cover the period 2010–2015. smartphone exporter, producing 40 percent of Samsung’s global mobile phone products and employing 35 percent of its global staff. Ten years ago, Vietnam barely exported electronics products. What sets Argentina, Poland, and Vietnam apart is their very different participation in global value chains (GVCs). In fact, the meteoric rises of Poland and Vietnam and the faltering of Argentina are not unique. China’s World Trade Organization (WTO) accession in 2001 ushered a new wave of GVCs which gave rise to “Factory Asia” (Baldwin 2016), while large parts of the Africa, South Asia, and Latin America regions are being left behind with little integration into GVCs (see fig. 1). At the heart of GVC participation are the international fragmentation of production across countries and durable firm-to-firm relationships that promote access to capital and inputs along chains (Antràs 2016; 2020). The efficiency gains and technology diffusion within GVCs explain the boost to incomes and the reductions in poverty in the participating countries, such as Bangladesh, China, and Vietnam (World Bank 2019). Empirical evidence across and within countries confirms that GVC participation fosters productivity, value-added, and growth (Constantinescu, Mattoo, and Ruta 2019; World Bank 2019; Stolzenburg, Taglioni, and Winkler 2019; Pahl and Timmer 2020).1 But what factors determine GVC participation across countries? Do factors that affect traditional trade have differential impacts on GVC trade? This paper studies the determinants of GVC participation based on a panel dataset covering more than 100 countries over the past three decades. The time period reflects the growing international fragmentation of production, while the sample includes countries in all geographical regions and stages of development. This paper exploits the large cross-country variation in the dataset which is uniquely suitable to estimate the marginal impact and relative importance of the different determinants on GVC participation. For GVC participation, we rely on the backward GVC participation measure of Borin and Mancini (2019), which captures the import content of exports of a country, originally referred to as the vertical special- ization index in Hummels, Ishii, and Yi (2001). For determinants, we rely on the trade literature which has emphasized the following: (a) factor endowments, (b) geography, (c) domestic industrial capacity, (d) trade policy and FDI, (e) institutional quality, (f) connectivity, and (g) macroeconomic factors. The unified framework of this paper further enables us to draw conclusions regarding which determinants 1 At the firm level, considering firms that import intermediate inputs and export as participating in GVCs, the evidence clearly shows that they benefit in terms of higher productivity, capital intensity, and employment relative to firms that engage only in exports (Muûls and Pisu 2009; Wagner 2012; Kasahara and Lapham 2013; World Bank 2019; Banh, Wingender, and Gueye 2020). The World Bank Economic Review 331 affect GVC trade more than traditional trade, due to their differential impacts on the intense firm-to-firm interactions necessary for GVC trade.2 To address the challenges in establishing causality in a cross-country setting we take a three-pronged approach. First, we rely on instrumental variable (IV) estimation for cross-country regressions to address the potential endogeneity of tariffs and FDI. The use of IV estimation also mitigates biases due to the construction of the GVC participation variables, which hinges on the use of intercountry input-output Downloaded from https://academic.oup.com/wber/article/36/2/329/6359883 by LEGVP Law Library user on 08 December 2023 (ICIO) tables that may have measurement problems. IV results are valid to the extent that the exclusion re- strictions are satisfied. Second, we use a difference-in-difference framework following Rajan and Zingales (1998) for cross-country cross-sector regressions of GVC participation to sidestep endogeneity concerns for other determinants, such as endowments or institutions, that may not be completely addressed in the country-level IV regressions. Third, we undertake event studies focused on isolating the effects of trade and FDI liberalization episodes on GVC participation. The results from the cross-country IV regressions show that the key determinants of backward GVC participation are, in order of importance, factor endowments, geographical location, political stability, tariffs and FDI inflows, and domestic industrial capacity.3 Moreover, broad sector analysis suggests that the findings are largely driven by manufacturing. The results from the difference-in-difference framework based on cross-country cross-sector regressions further confirms that institutional quality, factor endow- ments, trade policy, FDI, and connectivity matter for GVC participation. Finally, the results from event studies also support the finding that trade and FDI liberalization due to WTO or European Union (EU) accessions significantly increase backward GVC participation. In combination, the results from this three- pronged approach paint a consistent picture that actionable government policies, such as trade and FDI liberalization, are important in determining GVC participation. In addition, these results are robust to the use of alternative measures of tariffs, factor endowments and institutional quality, as well as to a change in the set of IVs considered, the exclusion of China and Mex- ico from the estimating sample due to their processing trade regimes which may cause mismeasurement in their GVC participation, and the estimation of country-level regressions separately for each decade. We also analyze forward GVC participation, which captures the domestic value-added that is used in a bilateral partner’s export production, and find those determinants to be also important in explaining it. Finally, we extend our analysis to show that most determinants have larger impacts on GVC trade than on traditional trade. This paper contributes to several strands of the trade literature. First, the paper relates to the empirical literature on the determinants of GVC participation (Baldwin and Taglioni 2012; Brooks and Ferrarini 2012; Noguera 2012; Blyde 2014; Cheng et al. 2015; Kowalski et al. 2015; United Nations 2015; Buelens and Tirpak 2017; Balié et al. 2019; Ignatenko, Raei, and Mircheva 2019). We expand the existing analyses in terms of country, time, and variable coverage, but also with regards to the methodology by addressing potential endogeneity concerns using IV and difference-in-difference estimations, and event studies.4 By embedding the determinants in a unified framework, we are able to estimate their marginal contributions and identify their relative importance for GVC participation. Second, the paper relates to the literature on the measurement of GVC participation initiated with the seminal work of Hummels, Ishii, and Yi (2001), Koopman, Wang, and Wei (2014), Wang, Wei, and Zhu (2013), Wang et al. (2016), and Johnson and 2 For theoretical models on GVCs and international production fragmentation, please refer to Feenstra and Hanson (1997), Feenstra (1998), Antràs and Helpman (2004; 2008), Romalis (2004), Nunn (2007), Chor (2010), Johnson and Noguera (2012), Antràs (2016), and Antràs and De Gortari (2020). 3 The order of importance of the different determinants is based on their contribution to the R-squared of our baseline regression. 4 The studies closest to ours that obtain results aligned with ours are Kowalski et al. (2015) and Ignatenko, Raei, and Mircheva (2019). These papers are very comprehensive in their selection of factors affecting GVC participation. How- ever, while these studies establish strong correlations, they do not address potential endogeneity concerns. 332 Fernandes, Kee, and Winkler Noguera (2017). Our study uses a measure of GVC participation which is closely related to the original Hummels, Ishii, and Yi (2001) measure of vertical specialization and avoids a double-counting problem as introduced and developed by Borin and Mancini (2015; 2019).5 Finally, the paper makes an empirical contribution to the literature on international production sharing and GVCs reviewed by Feenstra (1998) and Timmer et al. (2014).6 This paper is organized as follows. In GVC Measures and Determinants, we define the GVC partici- Downloaded from https://academic.oup.com/wber/article/36/2/329/6359883 by LEGVP Law Library user on 08 December 2023 pation measures and we provide the conceptual framework that motivates our selection of determinants for GVC participation. The three-pronged empirical strategy and results are presented in Empirical Spec- ifications: Three-Pronged Approach and Results. Robustness Checks and Extensions presents robustness checks and extensions. Concluding Remarks concludes. 2. GVC Measures and Determinants 2.1. Defining and Measuring GVC Participation From imports of pistons used as intermediates in car manufacturing in Morocco to Chilean exports of copper used in refrigerators produced by firms in China and Mexico, GVC participation is multifaceted and diverse across countries. This paper draws on the decomposition of gross exports by Borin and Mancini (2015; 2019).7 The decomposition breaks down gross exports into three components: (a) gross exports that are unassociated with GVCs (which cross just one country border and are entirely absorbed in the immediate destination country), (b) gross exports that are associated with forward GVC participation (which cross at least one more country border downstream), and (c) a residual term which the authors associate with backward GVC participation. This paper mainly focuses on backward GVC participation (the residual term) as the key outcome variable of interest. A country’s backward GVC participation measures the import content of its exports. It can be expressed in levels, or in shares relative to the country’s total gross exports, which allows us to study the “intensity” of GVC participation. Comparing the factors that affect GVC participation shares with their influence on GVC participation levels and on export levels allows us to study which determinants matter more for GVC trade beyond traditional exports. The backward GVC participation measure was first introduced by Hummels, Ishii, and Yi (2001), which they referred to as the “vertical specialization index.” This index measures the import content of exports and takes into account indirect effects (based on the Leontief inverse matrix) where imported in- puts are embodied in domestic output, sometimes at several stages of production in the GVC, before being used as inputs for exports. Hummels, Ishii, and Yi (2001) constructed it using national input- output (IO) tables that distinguish only between imported and domestic inputs. More recently, Borin and Mancini (2015; 2019) updated this measure which encompasses foreign value-added as well as do- mestic and foreign double counting, based on a bilateral source country-based decomposition of gross exports using ICIO tables that also include source and destination countries.8 Domestic double counting refers to inputs imported by a country that initially originated in that country but were exported, pro- cessed in a foreign country, and then reimported and incorporated into the export good. Foreign double counting refers to inputs that were produced in a foreign country and are imported a second time (or 5 Other recent work on measurement of GVC participation includes Miller and Temurshoev (2017) and Fortanier et al. (2020). 6 See also Baldwin (2012) for a discussion on GVC participation as a new industrialization strategy. 7 Details on the construction of the Borin and Mancini (2019) measures are provided in supplementary online appendix S1. 8 This decomposition uses a source-based approach focusing on the origin of value-added, as opposed to a sink-based approach which focuses on the country of final demand for domestic value-added. The World Bank Economic Review 333 more) in processed form before being embodied in the last shipment of the domestic country’s export good. In this paper, we mainly use the backward GVC participation measure of Borin and Mancini (2019) because of its tight connection to the Hummels, Ishii, and Yi (2001) measure of vertical specialization, drawing on ICIO tables from the Eora database.9 One of the advantages of the Borin and Mancini (2019) approach is that they derive the GVC measures at a bilateral level which allows for much more detail in Downloaded from https://academic.oup.com/wber/article/36/2/329/6359883 by LEGVP Law Library user on 08 December 2023 terms of where imported goods that are reexported originate and where goods are ultimately absorbed by final demand. Their construction of GVC participation measures ensures that trade flows cross at least two country borders, which is a key characteristic of a GVC. We aggregate Borin and Mancini (2019)’s bilateral backward GVC participation measures to the country level for our analysis. We also consider measures of backward GVC participation at the country level decomposed into four broad sectors: agriculture, mining, manufacturing, and services. In order to obtain these measures, we split the country-level numerator into these four sectors and divide each by the country’s total exports. These measures by broad sector shed light on which sectors are driving the aggregate cross-country results. Finally, we use Borin and Mancini (2019)’s backward GVC participation measures aggregated at the country-sector level to study cross- country cross-sector determinants of GVC participation. As an extension, we also consider forward GVC participation. A country’s forward GVC participation measures the domestic value-added in exports that is used by the country’s bilateral partner countries for export production and can be expressed in levels or relative to the country’s total gross exports. It captures the portion of domestic value-added in exports that is not directly consumed by the bilateral partner (which is the final stage in the value chain).10 For example, finished apparel exports from Bangladesh that are exported to and consumed in the United States would not be included in Bangladesh’s forward GVC participation measure. We aggregate Borin and Mancini (2019)’s bilateral forward GVC participation measures to the country level and country-sector level for our analysis. The main data source for the GVC participation measures is the most recent release of the Eora database.11 Our cross-country cross-sector analysis relies on GVC participation measures for eight man- ufacturing subsectors in the Eora database, listed in the supplementary online appendix.12 There are several limitations to using ICIO tables to measure GVC participation. A major limita- tion is related to the distorted statistics that ICIO tables are based on.13 For lower-income countries 9 Borin and Mancini (2019) show (in their appendix C) that this measure is theoretically equivalent to the vertical spe- cialization index of Hummels, Ishii, and Yi (2001). For a technical discussion of these and other GVC measures, see Koopman, Wang, and Wei (2014), Wang, Wei, and Zhu (2013), Wang et al. (2016), Aslam, Novta, and Rodrigues-Bastos (2017), and Borin and Mancini (2015; 2019). 10 This forward GVC participation measure is not mechanically inversely related to the backward GVC participation mea- sure. In terms of the aforementioned gross export decomposition, forward participation (component (b)) and backward participation (component (c)) do not sum to gross exports; there is some additional variation coming from gross exports that are unassociated with GVCs (component (a)). 11 As described in Lenzen et al. (2013), the Eora database covers 190 countries over the period 1990–2015 using a 26- sector harmonized classification. The number of countries in our estimating sample is at most 121 due to missing data of the determinants and due to the exclusion of countries with problematic (negative) values for the GVC participation measures (Liberia, Moldova, Mauritius, South Sudan, Sudan, and Zimbabwe). 12 In robustness checks we use as alternative sources the TIVA database 2016 and 2018 editions covering 64 countries over the periods 1995–2011 and 2005–2015, respectively, the WIOD 2013 release covering 40 countries over the period 1995–2011, and the WIOD 2016 release covering 43 countries over the period 2000–2014, described in Timmer et al. (2015; 2016). These results are available upon request. Further details, summary statistics, and correlations among the Eora-based GVC participation measures are provided in the supplementary online appendix. One interesting fact to note from the correlations is that the cross-country correlation between backward and forward GVC participation is negative. 13 United Nations (2019) describes the major limitations and proposes solutions to help substantially improve the IO coefficients used in current ICIO tables. 334 Fernandes, Kee, and Winkler that often do not produce national supply-use tables, the Eora data are based on interpolations and estimations, and are therefore subject to measurement errors. There is also a concern regarding the lack of firm-level heterogeneity in the conceptual measure of GVC participation which may lead to errors.14 These measurement errors in the Eora-based GVC measures may bias the t-statistics of the least squares regression coefficients toward zero. Following Hausman (2001), we will use IV estimation to help mitigate these concerns (in addition to addressing endogeneity concerns discussed in the section Cross-Country Downloaded from https://academic.oup.com/wber/article/36/2/329/6359883 by LEGVP Law Library user on 08 December 2023 Regressions).15 2.2. Conceptual Framework This section provides the conceptual framework for our analysis, motivating the choice of determinants of GVC participation, drawing heavily on Antràs (2020). We frame our discussion around seven broad types of determinants: (a) factor endowments, (b) geography, (c) domestic industrial capacity, (d) trade policy and FDI, (e) institutional quality, (f) connectivity, and (g) macroeconomic factors. Importantly, we discuss why we expect determinants to differentially impact GVC trade, which are trade flows by firms that use imported inputs for their exports, relative to traditional trade, which encompasses trade flows by firms that only export but do not use imported inputs. Some features of GVC participating firms, namely the fact that they import inputs, export, and engage in durable firm-to-firm linkages, set them apart from firms that only export or import. At the aggregate level, this may lead certain determinants to affect GVC participation and traditional trade differently. For example, trade costs matter more for GVC firms due to their importing of inputs and multiple border crossings. In addition, the relationship- specificity within GVCs and governance of GVCs by a lead firm also sets GVC firms apart from firms that only export. Country characteristics can shape these differences across firms. Trade costs are influenced by a country’s geography, trade and FDI policy, market size, connectivity, and macroeconomic factors. Relationship specificity in GVCs depends on contract enforcement, which is affected by a country’s polit- ical stability and the rule of law, but also by the availability of skills. Finally, factor endowments influence importantly firms’ specialization and thus positioning in the supply chain. The variable definitions and detailed data sources for the determinants, as well as summary statistics and correlations, are provided in the supplementary online appendix. 2.2.1. Factor Endowments Factor endowments are crucial in determining international specialization (Heckscher–Ohlin model) for traditional trade and GVC trade alike, but they also shape the positioning of countries in GVCs and thus their extent of backward and forward GVC participation (Antràs 2020).16 We focus on three types of endowments: (a) land and natural resources, (b) labor, which is separated into low skilled and middle to high skilled, and (c) capital. An abundance of land and natural resources (such as copper and iron ore) in a country is naturally linked to high forward GVC integration, since agricultural products and commodities are used in a variety of downstream production processes that typically cross several borders. Low-skilled labor in lower-income countries is often an entry point to downstream assembly-type stages of production associated with high content of imported inputs in a country’s exports (high backward GVC participation) 14 The GVC literature emphasizes governance of lead firms and their relationship with suppliers as a major feature of GVC trade but such information is unavailable in current ICIO tables. Recent work by Fortanier et al. (2020) addresses this issue by combining international IO tables with multinational activities for 19 OECD countries. Bems and Kikkawa (2020) directly study the issue of aggregation bias in the construction of GVC measures by using a rich Belgian value- added tax transaction database. 15 Hausman (2001) suggests that compared to least squares (LS) estimates, IV estimates are not biased towards zero in the presence of measurement error in the left-hand side variable. 16 Empirical evidence on the influence of factor endowments on country-sector trade patterns is widely available following the study by Romalis (2004). The World Bank Economic Review 335 and exports of final goods (low forward GVC participation). But advancing to more skill-intensive tasks in the value chain increases forward GVC participation. The supply of skilled labor matters more strongly for firms participating in relational GVCs than for firms that just export because exchanges of information and communication become more important when customized inputs and technology are involved (Antràs 2020). Finally, the production processes in GVCs require capital investments, hence a larger endowment with capital is expected to increase GVC participation. Downloaded from https://academic.oup.com/wber/article/36/2/329/6359883 by LEGVP Law Library user on 08 December 2023 2.2.2. Geography Trade costs due to geography and distance can determine which country to import products from and can shape a country’s positioning in GVCs.17 Trade costs can affect GVC trade more strongly than traditional trade for at least two reasons. First, due to the larger number of trade links in a GVC, overall performance is determined by the strength of the weakest link in the supply chain. That is, trade costs for a lead GVC firm can quickly increase if some suppliers are located in remote countries. Second, higher trade costs affect not only prices of export goods, as is the case for traditional trade, but also that of imported inputs in a GVC (Antràs 2020). In a world of GVCs where intermediate inputs cross borders multiple times, trade costs compound along the supply chain. In sequential (or snake-like) GVCs, trade costs have a higher incidence on down- stream stages than on upstream stages. Countries at a greater distance from GVC hubs face a higher level of trade frictions. If trade frictions are iceberg in nature, they compound as the gross value of trade flows increases further downstream along a GVC. This may encourage more remote countries to specialize in upstream stages and more central countries to specialize in downstream stages (Antràs 2020). Empiri- cal evidence shows that bilateral GVC links are strongly positively correlated with geographic proximity and that countries’ backward GVC participation is negatively associated with their longer distance to the closest manufacturing hub.18 We expect geography—measured by geographical distance to the GVC hubs China, Germany, and the United States—to play an important role in determining GVC participation. 2.2.3. Domestic Industrial Capacity It is well established by gravity models that countries with a larger market size have more traditional trade.19 One way to measure market size is using domestic industrial capacity given that the traditional way of using gross domestic product (GDP) is clearly endogenous. But domestic industrial capacity has an ambiguous relationship with GVC participation. On the one hand, countries with a larger domestic industrial capacity may have more capacity for exports, which should increase backward GVC partici- pation if imported inputs are used in their exports. Furthermore, such countries’ larger market size could result in higher demand for final goods for domestic consumption, and thus lead them to specialize in downstream stages of production embodying more imported inputs. This would increase backward GVC participation if some of the final goods are not fully absorbed domestically and are exported. On the other hand, to minimize cross-hauling of semiprocessed goods in different stages, countries often specialize in contiguous stages of production in GVCs. Countries with a larger domestic industrial capacity may have a larger set of contiguous stages which reduce the use of imported inputs relative to domestic inputs in their exports, thus lowering backward GVC participation. At the same time, a larger domestic industrial capacity implies a larger domestic supplier base which reduces search frictions and shields from possible production disruptions. Such greater potential to source inputs locally could decrease imports and lower backward GVC participation (while increasing domestic value-added and forward GVC participation).20 17 See, e.g., Eaton and Kortum (2002) and Antràs and De Gortari (2020). 18 See Kowalski et al. (2015) and Buelens and Tirpak (2017). 19 See, e.g., Arkolakis, Costinot, and Rodriguez-Clare (2012). Larger economies are expected to export more as they produce more and to import more given their higher income (Antràs 2020). 20 See the discussion in Antràs (2020) and Kee (2015) for evidence of this mechanism in Bangladesh. 336 Fernandes, Kee, and Winkler Thus, the overall effect of domestic industrial capacity on GVC participation is ambiguous and can only be determined empirically. In this paper, we proxy for domestic industrial capacity using value-added of the manufacturing sector. 2.2.4. Trade Policy and Foreign Direct Investment Trade policy affects traditional trade but may play a larger role for GVC trade. Regulatory barriers on Downloaded from https://academic.oup.com/wber/article/36/2/329/6359883 by LEGVP Law Library user on 08 December 2023 imports and exports, such as tariffs or quotas, increase trade costs, with consequences for countries’ participation and positioning in GVCs, as laid out in the section Geography. Reducing such barriers can have an amplified benefit for GVC trade—especially when production stages are organized sequentially across borders—by lowering not only the price of export goods but also the input costs faced.21 Lower tariffs imposed by partner countries can also reduce the costs of exporting.22 Deep preferential trade agree- ments (PTAs) that go beyond traditional market access issues and include policy areas such as movement of capital, investment, visas, and intellectual property rights can also foster GVC participation.23 Countries can attract FDI to overcome relative scarcity of capital, technology, and knowledge, and thus integrate into GVCs. When tight control over foreign production processes is necessary (perhaps because of weak contractual enforcement or weak protection of intellectual property), multinational lead firms might prefer vertical integration of suppliers over an arm’s length relationship, resulting in FDI flows and intra-firm trade. In fact, it is hard to imagine a GVC in which a multinational firm is not involved at some stage of the production chain. But FDI inflows are not homogenous; rather there are different types of FDI with different implica- tions for GVC participation. An import-substituting or market-seeking FDI that mainly serves the domes- tic market is unlikely to affect GVC participation or traditional exports. An export-oriented efficiency- seeking FDI that mainly serves export markets increases imports of inputs and exports and thus increases traditional trade and especially GVC trade. Emerging evidence shows that tariffs imposed at home and in foreign markets are negatively associated with backward GVC participation, while openness to FDI is positively associated with backward GVC participation.24 Liberal trade policies and strong FDI have also been identified as important factors for moving up GVCs, based on firm-level evidence for China and Bangladesh.25 Hence, we expect tariffs and FDI (as well as deep PTAs) to be significant in determining GVC participation. But these determinants may be subject to endogeneity concerns that we address in detail in Instrumental Variables. 2.2.5. Institutional Quality What distinguishes GVC trade from traditional trade is the intense firm-to-firm interactions characterized by contracting and specialized products and investment (Antràs 2016; 2020). Weak contract enforcement 21 See Yi (2003; 2010) and Antràs and De Gortari (2020) for large magnification impacts of tariffs when intermediates trade and multistage production are considered in general equilibrium trade models. Caliendo and Parro (2015) additionally examine linkages across sectors and derive larger welfare gains from trade liberalization relative to models with no trade in intermediates and sectoral linkages. 22 Trade preferences given to Bangladesh’s exports by the EU induced greater firm entry into exports to the EU and then growth in exports to all markets (Cherkashin et al. 2015). But evidence shows that in the long run, trade preferences per se are insufficient for export success. Complementary domestic policies are needed, including low import tariffs, reduced regulatory burden, and enhanced connectivity (Fernandes et al. 2019a). 23 See Orefice and Rocha (2014), Kowalski et al. (2015), Johnson and Noguera (2017), and Laget et al. (2018). 24 See evidence by Cheng et al. (2015) and Kowalski et al. (2015) for tariffs and by Kowalski et al. (2015) and Buelens and Tirpak (2017) for, respectively, FDI openness at the country level and FDI stocks at the bilateral level. Cheng et al. (2015) show that higher FDI restrictiveness is related to lower GVC participation in low-tech manufacturing. The importance of lower tariffs on intermediates to foster export performance is also confirmed at the firm level (Bas and Strauss-Kahn 2015; Pierola, Fernandes, and Farole 2018). 25 See Kee (2015) and Kee and Tang (2016). The World Bank Economic Review 337 is thus a significant deterrent not only for traditional trade, but even more so for GVC trade.26 Because the performance of a GVC depends on the strength of the weakest link in the supply chain, production delays driven by weak contract enforcement might be particularly harmful in GVCs. In addition, the presence of relationship-specific investments (e.g., the customization of products) and the exchange of large flows of intangibles (such as technology, intellectual property, and credit) reinforces the potential role of institutional quality as a significant determinant of relational GVC participation. Some emerging Downloaded from https://academic.oup.com/wber/article/36/2/329/6359883 by LEGVP Law Library user on 08 December 2023 evidence finds a correlation between a stronger rule of law and stronger GVC integration.27 We expect institutional quality to be more important in determining GVC participation than traditional trade. 2.2.6. Connectivity Transport costs remain, according to surveys of developing country suppliers, the main obstacle to enter- ing, establishing, or upgrading in GVCs.28 The geographic centrality of a country can attract downstream production stages in GVCs. But geographic centrality is more related to centrality in the transport net- work than to distance. Because centrality in the transport network influences trade costs and these are amplified in GVCs, logistics and communication infrastructure, port and customs efficiency, and infor- mation technology networks could be even more important for GVC trade than for traditional trade (see also the earlier discussion on geography and trade policy). The quality of the national road infrastruc- ture also matters for timely delivery to global markets. Moreover, the use of the internet and a common language could also facilitate GVC participation. Studies show a stronger role of logistics performance for trade in parts and components than for trade in final goods, and provide evidence that unpredictable land transport keeps most Sub-Saharan African countries out of GVCs.29 Given that our sample includes a wide range of countries with diverse logistics infrastructure and languages used, we are able to examine whether the connectivity matters more strongly for GVC participation than for traditional trade. 2.2.7. Macroeconomic Factors Macroeconomic factors, in particular related to real exchange rates and financial development, can play a different role for GVC trade relative to traditional trade. Real exchange rate appreciations tend to lower traditional exports but have a more muted impact on GVC participation due to its embedding of imported inputs.30 The degree of financial development of countries is a source of comparative advantage and could affect GVC trade more strongly than traditional trade.31 Firms that participate in GVCs face 26 A body of work establishes institutional quality as a comparative advantage factor in determining export patterns: Acemoglu, Antràs, and Helpman (2007), Levchenko (2007), Nunn (2007), Costinot (2009), and Chor (2010). 27 See Kowalski et al. (2015). 28 See OECD and WTO (2013). Hummels and Schaur (2013) find that a day of delay in transit due to a different transport mode choice has a tariff equivalent of 0.6 to 2.1 percent, and the most sensitive flows are for a type of GVC trade, that is, trade in parts and components. Similar magnitudes for the cost of a one-day delay in inland transit are found by Djankov, Freund, and Pham (2010). 29 See Ansón et al. (2017) for estimates of the sensitivity of bilateral trade in parts and components and in final goods to logistics performance. See Christ and Ferrantino (2011) for evidence on Sub-Saharan Africa. A positive correlation between broad infrastructure—including communication, electricity, roads, and power—and overall GVC participation in manufacturing is provided by Cheng et al. (2015). Linguistic proximity is shown to matter for bilateral GVC links in Buelens and Tirpak (2017) and Ignatenko, Raei, and Mircheva (2019). At the firm level, evidence on the role of unpredictability in imports’ border clearance times, regional road density, and internet access for export performance are shown, respectively, by Vijil, Wagner, and Woldemichael (2019), Rodríguez-Pose et al. (2013), and Fernandes et al. (2019b). 30 Amiti, Itskhoki, and Konings (2014) show theoretically and empirically low aggregate exchange rate pass-through to export prices as large exporters are simultaneously large importers, with high market share and high import intensity. Exchange rate volatility is shown to be negatively related to bilateral GVC trade by Ignatenko, Raei, and Mircheva (2019). 31 See Beck (2003) and Manova (2008; 2015). 338 Fernandes, Kee, and Winkler not only sunk entry costs to export but also sunk entry costs to import inputs. Thus, a country’s financial development may matter more strongly for GVC participants, due to larger flows of credit, especially in relational GVCs. Our sample coverage of developed and developing countries allows us to study how macroeconomic factors may impact traditional trade and GVC participation. 3. Empirical Specifications: Three-Pronged Approach and Results Downloaded from https://academic.oup.com/wber/article/36/2/329/6359883 by LEGVP Law Library user on 08 December 2023 This paper studies factors that influence GVC participation by exploiting variation across countries and over time both in GVC participation and in the determinants. Given that it is notoriously difficult to establish causality at a country level, our empirical strategy consists of a three-pronged approach. We first run reduced-form cross-country regressions, based on both decadal least-squares between estimation as well as IV estimation to address potential endogeneity and measurement issues. We then apply a difference- in-difference framework to country-sector regressions to further control for other endogeneity issues that cannot be satisfactorily addressed by the IV estimation. Finally, to highlight the importance of trade and FDI liberalization in affecting GVC participation, we rely on event studies based on countries’ accession to the WTO and EU during the sample period. 3.1. Cross-Country Regressions 3.1.1. Least-Squares between Estimation The main specification to estimate the impact of country decadal averages of the determinants on country decadal average GVC participation is given by Yct = β0 + β1 Xct + It + εct , (1) where c is a country, t a decade, Xct includes the determinants described in Conceptual Framework, It are decade fixed effects, and ε ct is an independent and identically distributed error (i.i.d.). The variable Yct is the backward GVC participation share described in Defining and Measuring GVC Participation. The construction of the backward GVC participation share allows us to interpret coefficients as follows: a significant positive coefficient on a determinant indicates a stronger impact of the determinant on GVC trade (the numerator) than on traditional exports (the denominator) while an insignificant coefficient on a determinant implies the determinant does not differentially impact GVC trade relative to traditional exports. As a baseline, we use least-squares between (LS-BE) estimation for the cross-country panel regression specified in equation (1), where the panel includes up to three observations per country, each covering a decadal average. The coefficients are identified via cross-country variation in GVC participation and the determinants within a decade. There are two justifications for this approach. First, GVC participation and some determinants change very slowly within countries from year to year. This is the key reason why we do not rely on a country-year panel and within effects estimation for equation (1).32 Decadal averages of GVC participation and determinants exhibit more meaningful variation than year-to-year observations and they may wash out measurement issues in GVC participation due to errors in ICIO tables (see Defining and Measuring GVC Participation). Moreover, the use of decadal averages allows to maximize country coverage as countries remain in the estimating sample even if GVC participation or some determinants are observed only in a few of a decade’s years. Second, in contrast to small within- country changes, GVC participation measures and determinants exhibit large cross-country variability. LS-BE estimation exploits this variability to identify the impacts of the different determinants on GVC participation. The decade fixed effects in equation (1) allow us to account for technological shocks or the global financial crisis affecting all countries. 32 Most coefficients in equation (1) cannot be precisely estimated in a within effects cross-country panel regression. The World Bank Economic Review 339 However, some determinants, namely tariffs and FDI inflows, could be endogenous or simultaneously determined with GVC participation. There are opposing forces at play regarding reverse causality from GVC participation to tariffs and FDI inflows. First, countries with stronger GVC participation may lower tariffs and attract FDI so GVC firms can have access to cheaper imported inputs or better domestic in- puts produced by FDI firms.33 This force confounds with the direct impact of tariffs and FDI on GVC participation and causes their LS-BE coefficients to be too large in magnitude. Second, countries with Downloaded from https://academic.oup.com/wber/article/36/2/329/6359883 by LEGVP Law Library user on 08 December 2023 stronger participation in GVCs rely more heavily on imports which may increase competition for domes- tic producers. Political economy pressures could lead to stronger tariff protection and FDI restrictions.34 This force causes LS-BE coefficients on tariffs and FDI to be too small in magnitude. To add to ambigu- ous reverse causality, potential measurement errors in FDI may bias LS-BE estimates towards zero. In particular, the fact that our FDI measure is unable to disentangle types of FDI affecting GVC partici- pation differently may bias LS-BE coefficients on FDI toward zero. Import-substituting FDI that mainly serves the domestic market may not affect GVC participation while export-oriented FDI increases GVC participation strongly. Ultimately, depending on which force dominates, LS-BE estimates of the impacts of tariffs and FDI on GVC participation could be biased upward or downward.35 3.1.2. Instrumental Variables To address the potential endogeneity of tariffs and FDI inflows in determining GVC participation we rely on IV estimation. IV estimation also addresses concerns of measurement error in the determinants and GVC participation measures (Hausman 2001). To obtain consistent estimates, the IVs need to be jointly significant in explaining tariffs and FDI inflows with meaningful first-stage coefficients and F-statistics, that is, they need to be relevant. To satisfy the exclusion restrictions these IVs should not be correlated with the second-stage regression errors. That is, the IVs should only affect GVC participation through tariffs and FDI but should not have significant direct impacts on GVC participation once tariffs and FDI are included in the second-stage regressions, conditional on all the other determinants.36 In this aggregate cross-country setting, it is highly challenging to find reasonable IVs that are both relevant and meet the exclusion restrictions. We therefore rely on existing trade theories and on solid empirical evidence to guide our choice of meaningful IVs and we conduct econometric tests to establish the strength of our identification. Below we describe the set of IVs that used to explain average tariffs imposed on manufacturing products and FDI inflows for more than 100 countries in our regressions. The detailed data sources for the IVs are provided in the supplementary online appendix. 3.1.3. Import Elasticity Classic trade theory since Bickerdike (1907) asserts that the optimal tariff set by a welfare-maximizing government for a country with market power is higher than zero, while the optimal tariff for a small open economy is zero. This positive relationship between tariffs and the market power of a country has long been referred to as the terms-of-trade theory of trade policy and is empirically well established.37 For our 33 See Blanchard (2010; 2015) for models where FDI and GVC linkages lead countries to lower their tariffs. 34 For tariffs, this is the classic endogeneity of trade policy hypothesis formalized in Grossman and Helpman (1994) and tested in Goldberg and Maggi (1999). 35 The current IV approach does not address the potential for reverse causality in determinants other than tariffs and FDI inflows. For example, backward GVC participation could have crowding-out effects on domestic firms and domestic industrial capacity. Likewise, GVC participation could be a key force that prompts a deepening in a country’s stock of capital equipment and machinery. These types of concerns are tackled in the difference-in-difference country-sector regressions. 36 According to Wooldridge (2012), instrument exogeneity means that the IV should have no partial effect on the dependent variable, after controlling for the right-hand side variables, and the IV should be uncorrelated with omitted variables. 37 See Broda, Limao, and Weinstein (2008) and Bagwell and Staiger (2011) for evidence of this relationship for a sample of pre-WTO accession countries. 340 Fernandes, Kee, and Winkler analysis we need a proxy for market power that does not directly affect GVC participation in order to satisfy the exclusion restriction. Our choice is to use import demand elasticity estimates from Kee, Nicita, and Olarreaga (2008) which are available for more than 100 importing countries, thus ensuring very good coverage of our sample.38 These elasticities represent the slope of the import demand function of the countries and are based on a GDP function that controls for countries’ factor endowments.39 While the elasticities are estimated at the country-HS 6-digit product level, our instrument for manufacturing Downloaded from https://academic.oup.com/wber/article/36/2/329/6359883 by LEGVP Law Library user on 08 December 2023 import tariffs measures the import demand elasticity for each country’s manufacturing sector obtained as the average of the import demand elasticities of all its manufacturing products. Countries with more elastic import demand may impose lower tariffs in order to minimize the deadweight loss which leads to a negative relationship between import elasticity and tariffs. The exclusion restriction requires that the import elasticity does not have a direct impact on GVC par- ticipation other than through tariffs, conditioning on all other determinants. In theory, the import demand elasticity may depend on preferences, income, and domestic substitutes. Given that the estimated elasticity is multilateral in nature (not bilateral partner country specific), and that in our second-stage regressions we control for domestic industrial capacity and for factor endowments which may affect domestic sub- stitutes as well as for skills and institutional quality which may reflect income and preferences, IV results are valid to the extent that the exclusion restriction is satisfied. 3.1.4. Population Terms-of-trade theory implies that countries with a larger total population may have more market power and thus impose higher tariffs. Further, political economy models also show that population size is rel- evant in determining a country’s optimal tariffs.40 We allow for the effects of import elasticity and total population to be related, so their interaction term is also included as an additional IV. Moreover, total population plays a dual role in our IV strategy as more populous countries are more labor abundant and may attract larger FDI inflows. Such FDI will likely be either export oriented or efficiency seeking and employ local workers to produce for export markets, thus leading to higher GVC participation. Hence, total population is relevant in addressing measurement problems in FDI by isolating export-oriented FDI inflows that matter for GVC participation. GVC participation should not affect a country’s total population, and populous countries (such as India and Brazil) do not necessarily have higher GVC participation. There may be concerns that popu- lation could affect GVC participation through other channels, such as business environment, domestic suppliers, and labor. However, since controls for institutional quality, domestic industrial capacity and factor endowments are included in our regressions, such concerns are mitigated. Thus, conditioning on these determinants, it is plausible that population affects GVC participation only through tariffs and FDI inflows and satisfies the exclusion restriction.41 3.1.5. Statutory Corporate Tax Rates Foreign affiliates of a multinational company (MNC) are subject to corporate income taxes in the FDI host country. Higher corporate tax rates in the host country reduce the after-tax return to investment for the MNC and hence may discourage stronger investment flows. But by also stimulating domestic activity, 38 An alternative choice would be to use estimates for the export supply elasticity an importing country faces. Broda, Limao, and Weinstein (2008) obtain such estimates but for a sample covering only 15 countries. 39 These elasticity estimates are used to study the effect of WTO tariff commitments by Bagwell and Staiger (2011) and the general equilibrium impact of trade liberalizations in GTAP models. 40 See the median voter model of Jiao and Wei (2020). 41 Our use of an elasticity measure and population size as IV for tariffs finds a theoretical justification in the model by Blanchard, Bown, and Johnson (2016), where insights for trade policy in a world with both GVCs and endogenous trade policy (protection-for-sale) are derived. The World Bank Economic Review 341 lower corporate tax rates could actually discourage multinational investments to set up GVC operations with strong backward linkages.42 Gravity regressions for MNC affiliates’ location choices and investments drawing on economic geog- raphy models provide extensive evidence that high corporate income tax rates are negatively associated with FDI inflows.43 Using statutory corporate tax rates as an instrument for FDI inflows could satisfy the exclusion restriction, as a priori there is no clear relationship between corporate tax rates and GVC par- Downloaded from https://academic.oup.com/wber/article/36/2/329/6359883 by LEGVP Law Library user on 08 December 2023 ticipation other than through the FDI channel, conditioning on institutional quality, domestic industrial capacity, and endowments. 3.1.6. Transitional Economy Status The late 1980s ushered the fall of communism and gave rise to a group of countries that embraced market capitalism and abandoned central planning, collectively referred to as transitional economies (Interna- tional Monetary Fund 2000). Most of these economies are the former Soviet Union and its satellite states, including Poland, Hungary, and Bulgaria, while other countries are in East Asia, such as China and Viet- nam. These transitional economies opened up engaging in tariff and FDI liberalization, and several were subsequently successful in participating in GVCs. But countries like Albania, Mongolia, Russia, Tajikistan, and Turkmenistan have maintained low backward GVC participation, proving that being a transitional economy per se—rather than through tariff and FDI liberalization—does not guarantee strong GVC par- ticipation. We use countries’ transitional economy status as an instrument for tariffs and FDI inflows. Since a priori GVC participation is not correlated with the transitional economy status of countries, this variable satisfies the exclusion restriction, conditioning on institutional quality, factor endowments, and domestic industrial capacity. 3.1.7. Instrumental Variables Summary and Tests In summary, we have two endogenous variables in our reduced-form specification equation (1): tariffs and FDI inflows. We use five excluded exogenous variables as instruments: import elasticity, total population, the interaction between import elasticity and total population, statutory corporate tax rate, and transi- tional economy status. We test for weak IVs in the first-stage regressions, based on the Kleibergen–Paap Wald F-statistic, which indicates that these IVs jointly have significant explanatory power for tariffs and for FDI inflows if the F-statistic is higher than the critical value given by Stock and Yogo (2005). We also test that (the instrumented) tariffs and FDI inflows can jointly explain GVC participation, based on the weak-instrument-robust Anderson–Rubin Wald test in the second-stage regressions. We expect the IV es- timates of the coefficients on tariffs and FDI inflows to be larger in magnitude than the LS-BE estimates of those coefficients if measurement errors and political economy forces dominate the confounding reverse causality forces.44 3.1.8. Results: Least-Squares between Effects Regressions Column (1) of table 1 shows the baseline LS-BE estimates of equation (1). Most coefficients have the expected signs and are significant. Factor endowments jointly matter for backward GVC participation, with a strong F-statistic of 6.84, which is significant at the 99 percent confidence level. Larger land and/or 42 The reasoning would be close to that described in Domestic Industrial Capacity, whereby the availability of a larger set of domestic suppliers could be negatively correlated with the backward GVC participation share. 43 See, e.g., Desai, Foley, and Hines Jr (2004), Head and Mayer (2004), and Mutti and Grubert (2004). 44 We also tried using lagged tariffs and FDI, as well as neighboring countries tariffs and FDI as instruments. In these cases, the second-stage results are very similar, with IV estimates of the coefficients on tariffs and FDI being larger in magnitude than the corresponding LS-BE ones. Reed (2015) shows that lagged endogenous variables are valid IVs if the lagged endogenous variables do not affect the current period dependent variable and if the lagged endogenous variables are correlated with the current period endogenous variables. These results are available upon request. 342 Fernandes, Kee, and Winkler Table 1. Determinants of Backward Global Value Chain Participation Shares (1) (2) (3) (4) (5) Regressions LS-BE LS-BE LS-BE LS-BE LS-BE Avg. tariff rate (%) −0.006*** −0.004* −0.006*** −0.005** −0.006*** (0.002) (0.002) (0.002) (0.002) (0.002) FDI inflows (log) 0.023** 0.019* −0.009 0.017* 0.022** Downloaded from https://academic.oup.com/wber/article/36/2/329/6359883 by LEGVP Law Library user on 08 December 2023 (0.010) (0.011) (0.013) (0.010) (0.010) Distance to GVC hubs (log) −0.104*** −0.111** −0.119*** −0.116*** −0.103*** (0.036) (0.048) (0.044) (0.038) (0.036) Political stability index 0.030* 0.024 0.040** 0.034* 0.025 (0.016) (0.019) (0.019) (0.017) (0.017) Domestic industrial capacity (log) −0.027*** −0.027*** 0.000 −0.015* −0.027*** (0.008) (0.009) (0.011) (0.009) (0.008) Resources rents/GDP (%) −0.003** −0.002** −0.002 −0.003*** −0.003** (0.001) (0.001) (0.001) (0.001) (0.001) Capital/GDP (log) 0.044* 0.025 0.033 0.034 0.050* (0.025) (0.029) (0.026) (0.027) (0.025) Land/GDP (log) −0.020*** −0.014** −0.014 −0.019*** −0.020*** (0.006) (0.007) (0.009) (0.007) (0.006) Medium/high-skilled labor/GDP (log) 0.012 0.005 −0.005 0.015 0.006 (0.015) (0.016) (0.016) (0.017) (0.016) Low-skilled labor/GDP (log) 0.009 0.011 0.018 0.007 0.011 (0.015) (0.016) (0.016) (0.016) (0.015) Exch. rate appreciation 0.000 0.000 −0.000 0.214 0.000 (0.000) (0.000) (0.000) (0.403) (0.000) NAFTA — 0.046 — — — (0.068) EU — 0.079* — — — (0.044) MERCOSUR — 0.020 — — — (0.071) ASEAN — 0.133** — — — (0.056) No. of trade agreement partners (log) — −0.011 — — — (0.038) Avg. depth of trade agreements (log) — 0.009 — — — (0.023) Time to import (log) — — — 0.000 — (0.001) Female labor participation (%) — — — — 0.001 (0.001) Observations 290 266 194 191 290 R-squared 0.528 0.546 0.464 0.505 0.532 Number of countries 121 119 88 109 121 Decade fixed effects Yes Yes Yes Yes Yes Source: Authors’ analysis based on data from Eora, CEPII, Doing Business Database, International Labor Organization (ILO), Penn World Tables 9.0, United Nations Conference on Trade and Development (UNCTAD), World Governance Indicators, World Development Indicators, and Hofmann, Osnago, and Ruta (2017). See the supplementary online appendix for a description of the variables used. Note: The table shows the estimated coefficients from least-squares between (LS-BE) regressions with backward global value chain participation shares as the dependent variable. LS standard errors in parentheses. Column (3) estimates are based on a sample excluding high-income countries. FDI is foreign direct investment, GDP is gross domestic product, NAFTA is the North American Free Trade Agreement, EU is the European Union, MERCOSUR is the Southern Common Market, and ASEAN is the Association of Southeast Asian Nations. ***, **, and * indicate significance levels of 1%, 5%, and 10%, respectively. The World Bank Economic Review 343 natural resources endowments are linked to lower backward GVC participation shares, while abundance in capital is associated with higher backward GVC participation shares. But geography and domestic industrial capacity also matter. A shorter distance to the GVC hubs is positively correlated with backward GVC participation shares. Countries with larger domestic industrial capacity (i.e., a potentially larger pool of domestic suppliers) exhibit lower backward GVC participation shares, as domestic inputs may be used to replace imported inputs. Downloaded from https://academic.oup.com/wber/article/36/2/329/6359883 by LEGVP Law Library user on 08 December 2023 Lower tariffs and larger FDI inflows are significantly linked to higher backward GVC participation shares across countries. Better institutional quality measured by a higher score in the political stability index is linked to higher backward GVC participation shares. Exchange rate appreciation is unrelated to backward GVC participation, which is not surprising, given that appreciation stimulates imports by reducing its prices, but can also reduce exports due to higher export prices.45 The R-squared in column (1) indicates that the different determinants considered explain more than half of the variation in backward GVC participation shares across countries. In columns (2)–(5) of table 1, we modify the baseline regression by including alternative or additional control variables or by changing the estimating sample. First, in column (2) we include alternative mea- sures of trade openness related to membership in PTAs (NAFTA, EU, MERCOSUR, ASEAN) and deep integration efforts. The estimated coefficients show that countries’ EU or ASEAN membership is linked to significantly higher backward GVC participation shares. Second, to understand whether the patterns identified are driven by a particular type of country, we drop high-income countries as defined by the World Bank income classification from the estimating sample in column (3).46 This sample exhibits sub- stantially less cross-country variation but several patterns identified in column (1) still hold, namely the strong negative impact of tariffs and distance to GVC hubs and the strong positive impact of political stability on backward GVC participation. Third, we add as a proxy for non-tariff trade costs and con- nectivity a measure of the time to import in column (4) and a control for the importance of females in the labor force in column (5). The baseline results in column (1) remain robust. We also included other control variables, such as the credit-to-GDP ratio, measures of the quality of logistics infrastructure, in- ternet infrastructure, and the prevalence of spoken English in the regressions, but these variables were all insignificant.47 To understand which sectors within the economy are driving the aggregate results presented in table 1, we estimate equation (1) separately for each of the backward GVC participation measures by broad sector—agriculture, mining, manufacturing, and services. The results presented in table 2 show clearly that the findings at the aggregate level are largely driven by GVC participation in man- ufacturing. All the determinants identified above influence backward GVC participation in manufac- turing in the same direction as in our baseline estimates. Low-skilled labor positively affects GVC participation in agriculture, while middle- or high-skilled labor negatively affects GVC participation in mining. GVC participation in services responds positively to stronger political stability, but nega- tively to domestic industrial capacity and to the abundance of land. The fact that FDI impacts back- ward GVC participation of the manufacturing sector but not of the services sector suggests that our 45 This result is consistent with the low exchange rate pass-through finding of Amiti, Itskhoki, and Konings (2014). 46 Since many countries change income status during our sample period, we use a time-varying World Bank income clas- sification to identify high-income countries. 47 These results are available upon request. A different type of robustness check uses GVC participation measures based on alternative TIVA and WIOD databases. These databases’ time and, importantly, country coverage is much more limited than Eora’s, as they focus on high- and middle-income countries. Thus, the cross-country variation in GVC participation and the determinants is much smaller. But the broad patterns of results are similar to those in tables 1 and supplementary online appendix table S3.21 though weaker in significance. The estimates that remain significant for backward GVC participation are the negative impact of natural resource endowments, distance to GVC hubs and domestic industrial capacity, and the positive impact of political stability. 344 Fernandes, Kee, and Winkler Table 2. Determinants of Sectoral Backward GVC Participation Shares (1) (2) (3) (4) Regressions LS-BE LS-BE LS-BE LS-BE Sector Agri. Comm. Manu. Serv. Avg. tariff rate (%) −0.000 −0.000 −0.004*** −0.001 (0.000) (0.000) (0.002) (0.000) Downloaded from https://academic.oup.com/wber/article/36/2/329/6359883 by LEGVP Law Library user on 08 December 2023 FDI inflows (log) 0.001 −0.001 0.023*** 0.001 (0.001) (0.001) (0.009) (0.003) Distance to GVC hubs (log) −0.002 0.004 −0.098*** 0.005 (0.004) (0.004) (0.031) (0.010) Political stability index 0.001 −0.001 0.016 0.008* (0.002) (0.002) (0.014) (0.004) Domestic industrial capacity (log) −0.001 0.001 −0.020*** −0.007*** (0.001) (0.001) (0.007) (0.002) Rents from resources/GDP (%) 0.000 0.001*** −0.004*** −0.000 (0.000) (0.000) (0.001) (0.000) Capital/GDP (log) −0.002 0.001 0.056*** −0.003 (0.003) (0.003) (0.021) (0.007) Land/GDP (log) 0.000 0.001 −0.016*** −0.003* (0.001) (0.001) (0.005) (0.002) Medium/high-skilled labor/GDP (log) 0.002 −0.003** 0.004 0.001 (0.002) (0.002) (0.013) (0.004) Low-skilled labor/GDP (log) 0.004** −0.001 0.004 0.006 (0.002) (0.002) (0.013) (0.004) Exch. rate appreciation 0.000 0.000 0.000 −0.000 (0.000) (0.000) (0.000) (0.000) Observations 290 290 290 290 R-squared 0.300 0.656 0.605 0.362 Number of countries 121 121 121 121 Decade fixed effects Yes Yes Yes Yes Source: Authors’ analysis based on data from Eora, CEPII, International Labor Organization (ILO), Penn World Tables 9.0, United Nations Conference on Trade and Development (UNCTAD), World Governance Indicators, and World Development Indicators. See the supplementary online appendix for a description of the variables used. Note: The table shows the estimated coefficients from least-squares between (LS-BE) regressions with backward global value chain (GVC) participation shares for different sectors as the dependent variables. LS standard errors in parentheses. Agri. is agriculture, Comm. is commodities, Manu. is manufacturing, Serv. is services, FDI is foreign direct investment, and GDP is gross domestic product. ***, **, and * indicate significance levels of 1%, 5%, and 10%, respectively. effects capture mostly export-oriented FDI. Overall, the results of the LS-BE regressions suggest that tariffs and FDI inflows are important in determining GVC participation of a country. Factor endow- ments are also important, and so are political stability, proximity to GVC hubs, and domestic industrial capacity. 3.1.9. Results: Two-Stage Least-Squares Regressions To address potential inconsistency in table 1’s estimates due to endogeneity of tariffs and FDI inflows and measurement error in determinants and GVC participation measures, we estimate the baseline specifica- tion in column (1) by two-stage least-squares relying on import elasticity, total population, their interac- tion, corporate tax rates, and transitional economy status as IVs. The corresponding first-stage between regressions are presented in columns (1) and (2) of table 3. All IVs exhibit the expected signs and are statistically significant. Countries with more elastic import demand have lower tariffs, but the negative impact of the import elasticity on tariffs diminishes with country size. The separate impact of total popu- lation on tariffs is positive and significant at the 90 percent level, while countries with larger populations The World Bank Economic Review 345 Table 3. First-Stage Regressions for Tariffs and Foreign Direct Investment (1) (2) (3) (4) LS-BE LS-BE LS LS Regressions Avg. tariff rate (%) FDI inflows (log) Avg. tariff rate (%) FDI inflows (log) Import elasticity −37.038*** 0.243 −38.720*** 0.782 (10.604) (1.864) (10.259) (1.880) Downloaded from https://academic.oup.com/wber/article/36/2/329/6359883 by LEGVP Law Library user on 08 December 2023 Import elasticity × population (log) 2.307*** −0.053 2.351*** −0.068 (0.606) (0.107) (0.594) (0.111) Population (log) −1.154 1.174*** −0.991 1.113*** (1.387) (0.244) (1.508) (0.332) Statutory corporate tax rate −0.075 −0.025** −0.022 −0.041*** (0.071) (0.013) (0.111) (0.013) Transitional economy status −4.318*** 0.343 −4.641*** 0.408* (1.500) (0.264) (1.391) (0.239) Distance to GVC hubs (log) −0.826 0.385 −1.733 0.641* (2.144) (0.377) (1.925) (0.336) Political stability index −0.890 0.439*** −0.854 0.421*** (0.818) (0.144) (0.785) (0.157) Domestic industrial capacity (log) −2.221** −0.127 −2.364* −0.072 (0.937) (0.165) (1.289) (0.217) Rents from resources/GDP (%) −0.016 −0.000 −0.042 0.007 (0.053) (0.009) (0.041) (0.011) Capital/GDP (log) 0.997 0.091 0.784 0.170 (1.169) (0.205) (1.256) (0.200) Land/GDP (log) −0.029 −0.050 −0.003 −0.055 (0.298) (0.052) (0.355) (0.074) Medium/high-skilled labor/GDP (log) −0.778 −1.093*** -1.233 −0.932*** (1.154) (0.203) (1.436) (0.257) Low-skilled labor/GDP (log) 0.112 −0.019 0.227 −0.065 (0.718) (0.126) (0.558) (0.105) Exch. rate appreciation 0.001 −0.001 0.001 −0.000 (0.004) (0.001) (0.002) (0.000) Observations 290 290 121 121 R-squared 0.480 0.863 0.465 0.851 Number of countries 121 121 121 121 Sanderson–Windmeijer F-test NA NA 12.69*** 7.02*** Kleibergen–Paap Wald test for weak IV NA 10.39** Decade fixed effects Yes Yes No No Source: Authors’ analysis based on data from CEPII, International Labor Organization (ILO), Penn World Tables 9.0, Tax Foundation (2019), United Nations Con- ference on Trade and Development (UNCTAD), World Governance Indicators, World Development Indicators, and Kee, Nicita, and Olarreaga (2008). See the supple- mentary online appendix for a description of the variables used. Note: The table shows the estimated coefficients from least-squares between (LS-BE) or LS regressions with either average tariff rate or FDI inflows as the dependent variable. LS standard errors in parentheses for columns (1)–(2); robust standard errors in parentheses for columns (3)–(4). GVC is global value chain, FDI is foreign direct investment, and GDP is gross domestic product. ***, **, and * indicate significance levels of 1%, 5%, and 10%, respectively. also attract more FDI inflows. But FDI inflows respond negatively to higher corporate tax rates. Finally, transitional economy status is associated with lower tariffs. Column (1) of table 4 presents the second-stage IV between estimates. Compared to the LS-BE co- efficients in column (1) of table 1, the coefficients on tariffs and FDI inflows are larger in magnitude, suggesting that political economy-driven reverse causality and measurement errors in FDI and GVC par- ticipation are important in biasing the LS-BE coefficients and their t-statistics towards zero. Jointly, the null hypothesis that tariffs and FDI inflows are not important in explaining GVC participation is strongly rejected. Likewise, the null hypothesis of factor endowments not being important in explaining GVC 346 Fernandes, Kee, and Winkler Table 4. Determinants of Backward Global Value Chain Participation Shares, IV Estimates (1) (2) (3) (4) Regressions IV-BE IV LS LS Avg. tariff rate (%) −0.011*** −0.009*** −0.005*** −0.004** (0.004) (0.003) (0.001) (0.002) FDI inflows (log) 0.038* 0.030* 0.018* 0.014 Downloaded from https://academic.oup.com/wber/article/36/2/329/6359883 by LEGVP Law Library user on 08 December 2023 (0.020) (0.018) (0.010) (0.010) Distance to GVC hubs (log) −0.090** −0.112*** −0.118*** −0.080* (0.038) (0.041) (0.044) (0.046) Political stability index 0.017 0.031** 0.040*** 0.034** (0.019) (0.014) (0.014) (0.015) Domestic industrial capacity (log) −0.040*** −0.031** −0.022** −0.023 (0.015) (0.012) (0.008) (0.027) Rents from resources/GDP (%) −0.003** −0.003*** −0.003*** −0.003** (0.001) (0.001) (0.001) (0.001) Capital/GDP (log) 0.045* 0.033 0.034 0.030 (0.026) (0.024) (0.023) (0.023) Land/GDP (log) −0.021*** −0.019*** −0.019** −0.021*** (0.006) (0.007) (0.007) (0.007) Medium/high-skilled labor/GDP (log) 0.018 0.009 0.005 −0.010 (0.016) (0.014) (0.014) (0.030) Low-skilled labor/GDP (log) 0.011 0.017 0.015 0.015 (0.016) (0.018) (0.019) (0.019) Exch. rate appreciation −0.000 −0.000 0.000 −0.000* (0.000) (0.000) (0.000) (0.000) Import elasticity — — — 0.084 (0.250) Import elasticity × population (log) — — — −0.009 (0.014) Population (log) — — — 0.024 (0.037) Statutory corporate tax rate — — — 0.000 (0.001) Transitional economy status — — — 0.039 (0.034) Observations 290 121 121 121 R-squared 0.483 0.472 0.503 0.530 Number of countries 121 121 121 121 F-stat for excluded IV — — — 1.41 AR Wald test for weak IV robust inference — 3.43*** — — KP rk LM statistic for underidentification — 22.07*** — — Decade fixed effects Yes No No No Source: Authors’ analysis based on data from Eora, CEPII, International Labor Organization (ILO), Penn World Tables 9.0, Tax Foundation (2019), United Nations Conference on Trade and Development (UNCTAD), World Governance Indicators, World Development Indicators, and Kee, Nicita, and Olarreaga (2008). See the supplementary online appendix for a description of the variables used. Note: The table shows the estimated coefficients from instrumental variables between estimates (IV-BE) or least squares estimates (LS) regressions with backward global value chain (GVC) participation shares as the dependent variable. LS standard errors in parentheses for column (1); robust standard errors for columns (2)–(4). FDI is foreign direct investment and GDP is gross domestic product. AR stands for Anderson-Rubin, KP stands for Kleibergen-Paap, LM stands for Lagrange Multiplier. ***, **, and * indicate significance levels of 1%, 5%, and 10%, respectively. participation is also strongly rejected. Similar to the results in column (1) of table 1, countries with larger endowments of land and/or natural resources exhibit significantly lower backward GVC participation shares, while those with stronger abundance in capital exhibit significantly higher shares. Moreover, coun- tries located farther away from GVC hubs and those with larger domestic industrial capacity exhibit significantly lower backward GVC participation shares. The World Bank Economic Review 347 3.1.10. Issues with Weak IVs and Statistical Inference A concern with the second-stage IV estimates is that they could be biased and the statistical inferences made above become invalid if the IVs are weak. That is, if the IVs included cannot explain the two en- dogenous variables, tariffs and FDI inflows, then the second-stage results are questionable. Unfortunately, the econometric theory of IV between estimators with two endogenous variables is not yet developed. The existing econometric theory focuses on cross-sectional IV estimation, with i.i.d. error terms as dis- Downloaded from https://academic.oup.com/wber/article/36/2/329/6359883 by LEGVP Law Library user on 08 December 2023 cussed in Andrews, Stock, and Sun (2019). To properly test for weak IVs, we collapse our cross-country decadal panel database into a single-period cross-country database, by averaging each variable over time within each country. In addition to enabling us to properly test for weak IVs, this cross-sectional speci- fication allows for heteroskedasticity to be addressed. Columns (3)–(4) of table 3 present the first-stage cross-sectional regressions, with standard errors robust to heteroskedasticity. Comparing column (3) to column (1) and column (4) to column (2) in table 3, it is clear that the first-stage cross-sectional regres- sions are very similar to the first-stage between regressions. The first-stage F-statistics for tariffs and FDI individually are strongly significant. Jointly, the Kleibergen–Paap F-statistic for the first-stage is 10.39, which is larger than the critical value of 8.78, as given in table 1 of Stock and Yogo (2005), and also larger than the rule of thumb of 10 widely used in empirical research relying on IV estimation. In table 4, column (2) presents the second-stage estimates of the cross-sectional regression, which are very similar to those of the IV between regression in column (1). The Anderson–Rubin Wald test for weak-instrument-robust inference for the second stage is 3.43, which is significant at the 99 percent level, rejecting the null hypothesis that tariffs and FDI jointly are not important in determining GVC partici- pation. The high Kleibergen–Paap rk lagrange multiplier statistic of 22.07 also rejects the null hypothesis that the regression is underidentified. Our preferred estimates in column (2) imply that an increase in a country’s manufacturing tariffs by 10 percentage points will cause its backward GVC participation share to decrease by 38 percent, while an increase in the country’s FDI inflows by 10 percent will cause its backward GVC participation share to increase by 1.3 percent, both at the sample mean. The estimates in column (2) also show that an increase in a country’s distance to GVC hubs by 10 percent is linked to a decrease in its backward GVC participation share by 4.8 percent at the sample mean. For completeness and ease of comparison, column (3) presents the second-stage cross-sectional re- gression LS coefficients which are smaller in magnitude for tariffs and FDI than the corresponding IV coefficients. Finally, in column (4) the five IVs are directly included in the second-stage cross-sectional regression and we cannot reject the null hypothesis that jointly these IVs are not significant and do not have a direct impact on backward GVC participation shares.48 3.1.11. Relative Importance of the Determinants Overall, the between and the cross-sectional IV estimates paint the same picture that tariffs and FDI inflows are important in determining backward GVC participation of a country. In addition, factor en- dowments, political stability, geographical proximity to GVC hubs, and domestic industrial capacity are crucial in affecting backward GVC participation. To assess the marginal contributions of the different de- terminants, we conduct the thought exercise described below, based on the IV cross-sectional estimation results in column (2) of table 4, and we report the results in table 5. First, we partial out all the included ex- ogenous variables such as distance, political stability, and factor endowments for a regression that includes only the endogenous variables: tariffs and FDI. The R-squared of 0.056 gives the marginal contribution of tariffs and FDI in explaining backward GVC participation. Next, we add back the other determinants one 48 There may be a concern that we are overidentifying the models with five IVs for two endogenous variables. We perform an overidentification test, under the assumption that the model is correctly identified and the error terms satisfy ho- moskedasticity. We fail to reject the null hypothesis that the five IVs are exogenous at a 95 percent level, with a Hansen J-statistic of 6.452. As a robustness check, we also dropped the transitional economy status as an IV. All the results are very similar and are available in the supplementary online appendix. 348 Fernandes, Kee, and Winkler Table 5. Marginal Contributions of Determinants (1) (2) (3) (4) Included variables R-squared Marginal R-squared Share of contribution (%) Tariff + FDI 0.056 0.056 13 Tariff + FDI + distance to GVC hubs 0.145 0.089 21 Tariff + FDI + political stability 0.132 0.076 18 Downloaded from https://academic.oup.com/wber/article/36/2/329/6359883 by LEGVP Law Library user on 08 December 2023 Tariff + FDI + dom. ind. cap. 0.073 0.017 4 Tariff + FDI + endowments 0.238 0.182 43 Sum 0.42 100 Source: Authors’ analysis based on data from Eora, CEPII, International Labor Organization (ILO), Penn World Tables 9.0, United Nations Conference on Trade and Development (UNCTAD), World Governance Indicators, and World Development Indicators. See the supplementary online appendix for a description of the variables used. Note: The table shows the R-squared from several least-squares between (LS-BE) regressions with backward global value chain (GVC) participation shares as the dependent variable. The R-squared in column (2) is constructed based on regressions that include only the variables listed in column (1). The marginal R-squared in column (3) measures the change in R-squared in column (2) when compared to the R-squared of including only average tariff rate and FDI inflows which is equal to 0.056. Column (4) calculates the share of contribution of the additional variables to the sum of all marginal R-squared. FDI is foreign direct investment and GVC is global value chain. at a time and we assess the changes in the R-squared. After adding distance to GVC hubs, the R-squared increases to 0.145, indicating the marginal R-squared contribution of distance is 0.089 (=0.145−0.056). Adding all factor endowments, the new R-squared is 0.238, suggesting that their marginal contribu- tion is 0.182 (=0.238−0.056). Overall, factor endowments can explain 43 percent of the backward GVC participation shares in the IV cross-sectional regression, which is the most important determinant, followed by distance (21 percent), political stability (18 percent), tariffs and FDI (13 percent), and do- mestic industrial capacity (4 percent), while exchange rate appreciation has no explanatory power. 3.2. Difference-in-Difference Country-Sector Regressions The cross-country regression models described in the previous sections use IVs to establish causality be- tween the determinants and GVC participation. However, there may be a concern that this strategy is not sufficient as other determinants, such as factor endowments, institutions, and connectivity may also be endogenous. As an alternative approach, we apply a country-sector difference-in-difference regression model inspired by the Rajan and Zingales (1998) framework using country-sector GVC participation measures with year-to-year variability. This framework relies on interactions between a country’s factor endowments and the intensity with which a sector uses a particular factor and has been used to link countries’ sectoral trade performance to various sources of comparative advantage: factor endowments (Romalis 2004), institutions (Levchenko 2007; Nunn 2007; Costinot 2009), and financial development (Beck 2003; Manova 2008).49 The difference-in-difference model to estimate the role of several determi- nants in explaining GVC participation and exports at the country-sector level is given by Ycst = β0 + β1 Xct -3 × ms + β2 Zcst -3 + Ict + Ist + εcst , (2) where c, s, and t are country, sector, and year, respectively, Ycst is GVC participation or gross exports in log levels, Xct -3 includes country factor endowments, ms is a set of sectoral intensities of use of these endowments from a benchmark country (the United States), and ε cst is an i.i.d. error. The term Zcst -3 captures country-sector trade policies, which include tariffs imposed at home on imports of final goods and on imports of intermediate inputs, and tariffs imposed by destination markets, and in some 49 Theoretical underpinnings for this framework in a trade context are provided by Chor (2010) who extends the seminal Eaton and Kortum (2002) trade model and estimates the importance of various sources of comparative advantage for trade patterns in an integrated framework. Ito, Rotunno, and Vézina (2017) examine whether GVC trade is more responsive than gross exports to such interaction terms intended to capture comparative advantage forces. The World Bank Economic Review 349 specifications a country-sector FDI measure.50 The specification controls for country-year fixed effects Ict that account for differences across countries both in time-varying macroeconomic conditions, such as aggregate FDI inflows, political stability, and domestic industrial capacity, as well as in time-invariant factors like geography. The regression also controls for sector-year fixed effects Ist that account for global shocks to technology and productivity across sectors. The interaction terms between X and m reflect the idea that a country’s GVC participation or exports Downloaded from https://academic.oup.com/wber/article/36/2/329/6359883 by LEGVP Law Library user on 08 December 2023 in a given sector differs depending on how intensively that sector uses a particular country endowment (all else being constant). For example, the interpretation of a positive effect of the interaction term with skilled labor is that increases in a country’s skilled labor endowments generate a stronger increase in GVC participation or exports of sectors that rely more intensively on skills compared to sectors that rely less intensively on skills (all else being constant).51 This difference-in-difference model helps to mitigate the endogeneity problems that arise in cross- country regressions in three ways. First, it is unlikely that strong sectoral GVC participation causes changes in the country-level determinants. Second, the model assumes that sectors differ in their intensity of use of endowments largely for technological reasons, which allows results to be given a causal interpretation.52 Third, using a three-year lag structure for determinants and including the country-year and sector-year fixed effects mitigates the risk of reverse causality. Moreover, the use of the lag structure allows GVC participation and exports to adjust slowly to changes in the determinants.53 However, a remaining concern may be that a country that is keen to promote GVC participation in, for example, skill-intensive sectors might make corresponding investments in education or training to expand its skill endowment. This would be consistent with a positive correlation between GVC participation and an interaction term between sectoral skill intensity and a country’s skill endowments. Using a lagged measure of skill endowments alleviates but does not fully resolve this concern. Still, we believe that this comparative advantage-type approach enables us to build a case that several determinants are likely to have a causal impact on GVC participation and exports at the country-sector level, thus complementing our previous cross-country analysis. Table 6 provides the estimates for equation (2). Columns (1)–(3) show that factor endowments and trade policy determine a country-sector’s backward and forward GVC participation levels as well as tra- ditional exports. Specifically, countries relatively more endowed with skilled labor have a comparative advantage in skill-intensive sectors, and they exhibit stronger backward and forward GVC participa- tion. Countries with larger abundance in capital exhibit larger exports and stronger GVC participation in capital-intensive sectors. Countries with larger natural resources endowments export more and have stronger GVC participation of natural-resource-intensive sectors. Countries with stronger institutional quality have a comparative advantage in exports and in GVC participation of contract-intensive sectors. Importantly, conditioning on countries’ factor endowments, trade policy plays an important role for GVC participation and exports. Sectors that face lower tariffs in their destination markets or lower tariffs on intermediate inputs exhibit stronger GVC participation and higher exports. To compare the strength of different determinants, fig. 2 shows the standardized beta coefficients cor- responding to the estimates in table 6. Hypothetical examples are useful to provide an economic in- terpretation of the coefficients. If Ghana increased its skilled labor share (7.5 percent in 2012) to the 50 The sectoral intensities and the tariffs and FDI measures used are described in the supplementary online appendix. 51 Our cross-country cross-sector specification follows closely the Nunn (2007) and Chor (2010) comparative advantage models that explain trade flows in levels. Hence our main results for this specification are based on GVC participation in levels. 52 This assumption is inspired by Rajan and Zingales (1998) and implies that there are no restrictions (other than cost) preventing access by certain sectors to skilled labor. 53 As a robustness check we use five-year and one-year lag structures and the results (available upon request) are qualita- tively similar. 350 Fernandes, Kee, and Winkler Table 6. Determinants of Country-Sector Global Value Chain Participation and Exports Backward global value chain Forward global value chain Exports participation (log) participation (log) (log) Dependent variables: (1) (2) (3) 3-year lag skilled labor endowment × skill intensity 0.067*** 0.059*** 0.069*** (0.006) (0.006) (0.006) Downloaded from https://academic.oup.com/wber/article/36/2/329/6359883 by LEGVP Law Library user on 08 December 2023 3-year lag capital endowment × capital intensity 0.006*** 0.006*** 0.005*** (0.000) (0.000) (0.000) 3-year lag nat. resource endowment × nat. resource intensity 0.103*** 0.111*** 0.113*** (0.009) (0.009) (0.009) 3-year lag rule of law index × contract intensity 1.495*** 1.677*** 1.560*** (0.056) (0.055) (0.053) 3-year lag average output tariff −0.573* 1.201*** 0.387 (0.304) (0.308) (0.287) 3-year lag average market access tariff −6.373*** −6.995*** −6.121*** (0.363) (0.384) (0.370) 3-year lag average input tariff −3.265*** −6.474*** −4.514*** (0.773) (0.849) (0.777) Observations 14,387 14,387 14,387 R-squared 0.922 0.911 0.916 Country × year fixed effects Yes Yes Yes Sector × year fixed effects Yes Yes Yes Source: Authors’ analysis based on data from Eora, International Labor Organization (ILO), National Bureau of Economic Research (NBER), US Census Bureau’s Center for Economic Studies (CES) Manufacturing Industry Database, Penn World Tables 9.0, United Nations Conference on Trade and Development (UNCTAD), World Governance Indicators (WGI), Braun (2003), Felbermayr, Teti, and Yalcin (2019), and Nunn (2007). See the supplementary online appendix for a description of the variables used. Note: The table shows estimated coefficients for each variable in the first column from least squares regressions using backward global value chain (GVC) participation in levels, forward GVC participation in levels, and gross exports as dependent variables. Robust standard errors in parentheses. ***, **, and * indicate significance levels of 1%, 5%, and 10%, respectively. cross-country median (20 percent), its backward GVC participation level would grow by 32 percent and its exports would grow by 33 percent, at the sample mean of sectoral skill intensity. If Mozambique in- creased its rule of law index to the cross-country median, its backward GVC participation would grow by 40 percent and its exports would grow by 42 percent, at the sample mean of sectoral contractual intensity.54 We estimate three extensions of equation (2). First, we add to equation (2) country-sector time-varying FDI. Given that, to the best of our knowledge, country-sector-year level FDI inflows data are not avail- able for a wide range of countries, we use FDI announcement data from the fDi Markets Database for the period 2003–2015 as a proxy for FDI inflows, following the approach of Hallward-Driemeier and Nayyar (2019). Specifically, we cumulate the number of cross-border greenfield FDI projects an- nounced from 2003 onwards for each country-sector. The estimates provided in supplementary online appendix table S3.3 show a positive and significant effect of FDI announcements on country-sector GVC 54 While the comparative advantage framework used in our specifications does not apply to GVC participation as a share of gross exports, we nonetheless obtained estimates using such shares as dependent variables. The results (available upon request) show that countries more abundant in capital and with stronger institutional quality exhibit higher backward and forward GVC participation shares in sectors more intensive in those factors. Lower tariffs in destination markets also increase those shares. Forward GVC participation shares are higher in natural-resource-intensive sectors in countries more abundant in natural resources. Skilled labor and natural resource endowments have a negative impact on backward GVC participation shares for sectors more intensive in those factors and so do tariffs on intermediate inputs. This finding is expected given the stronger (in magnitude) impacts those variables have on exports relative to backward GVC participation levels in table 6. The World Bank Economic Review 351 Figure 2. Standardized Coefficients for Determinants of Country-Sector Global Value Chain Participation and Exports. Downloaded from https://academic.oup.com/wber/article/36/2/329/6359883 by LEGVP Law Library user on 08 December 2023 Source: Authors’ analysis based on data from Eora, International Labor Organization (ILO), National Bureau of Economic Research (NBER), US Census Bureau’s Center for Economic Studies (CES) Manufacturing Industry Database, Penn World Tables 9.0, United Nations Conference on Trade and Development (UNCTAD), World Governance Indicators (WGI), Braun (2003), Felbermayr, Teti, and Yalcin (2019), and Nunn (2007). See the supplementary online appendix for a description of the variables used. Note: The graph shows standardized coefficients for each variable on the y-axis from three separate regressions using forward global value chain (GVC) participation in levels, backward GVC participation in levels, and gross exports as dependent variables. The regressions use a three-year lag of each of the determinants and control for country-year fixed effects and sector-year fixed effects. Standardized coefficients refer to how many standard deviations the dependent variable will change per standard deviation increase in the determinant. ***, **, and * indicate significance levels of 1%, 5%, and 10%, respectively. participation levels and on gross exports. Second, we add to equation (2) an interaction term between a country’s internet penetration and sectoral intensity of technology use. The results in supplementary on- line appendix table S3.4 show that countries with a higher internet penetration exhibit significantly higher GVC participation and gross exports in sectors that use technology more intensively. Third, we add to equation (2) interaction terms between the country’s GDP per capita and each of the sectoral factor inten- sities. This specification helps to ensure that the country characteristics are not merely picking up the role of the country’s level of economic development. The estimates shown in supplementary online appendix table S3.5 are qualitatively unchanged relative to those in table 6. 352 Fernandes, Kee, and Winkler Overall the results from these cross-country cross-sector regressions are consistent with the previous results based on IV cross-country regressions, highlighting the importance of endowments, institutional quality, tariffs, and FDI in determining GVC participation. 3.3. Event Studies As an alternative approach to the cross-country IV regression models, we conduct event studies to isolate Downloaded from https://academic.oup.com/wber/article/36/2/329/6359883 by LEGVP Law Library user on 08 December 2023 the impact of trade and FDI liberalization episodes on backward GVC participation.55 We define a trade and FDI liberalization event as the moment when a country either accedes to the WTO or joins the EU during the 1990–2015 sample period. The event year 0 is the year when the country experiences the liberalization event. For countries that join both the WTO and the EU during the sample period, such as Poland, we focus on the latter event, which is Poland joining the EU in 2004. For each country, we construct a GVC backward participation index which equals 1 in year 0 as backward GVC participation sharecτ GVC indexcτ = , backward GVC participation sharec0 where τ measures the year relative to the event. By dividing the backward GVC participation sharecτ by the backward GVC participation sharec0 , we can compare the change in backward GVC participation before and after the liberalization event within a country, and remove the influence of country-specific macroeconomic factors, such as endowments, geography, and institutions. Figure 3 plots the evolution of the GVC index for a selection of countries over the period ranging from five years before to five years after the trade and FDI liberalization event. The graphs show clearly that the GVC indexes trend upward after the event for all countries, while there is no clear evidence of any pre-trend before the event. As a first simple way to formally test the impact of the trade and FDI liberalization event, we perform a two-sample t-test that compares the average GVC index before and after the event year for the full sample of countries.56 The average GVC index before or on event year 0 is 1.01 with a standard error of 0.01. The fact that the average GVC index is not statistically different from 1 before the event indicates that there is no pre-trend in the GVC index. The average GVC index after year 0 is 1.06, with a standard error of 0.01. The fact that the index is statistically higher than 1 after the event indicates that backward GVC participation increases with trade and FDI liberalization. The difference between the GVC index before and after the event is 0.04 with a standard error of 0.01, which results in a t-statistic of 4.34 that rejects the null hypothesis of no difference in the GVC index before and after the liberalization event. As a second way to formally test for the impact of trade and FDI liberalization events on backward GVC participation, we estimate the following regression: GVCcτ = γ0 + γ1 Tτ + Ic + υcτ , (3) where Tτ is a treatment dummy variable that equals 1 after the event year 0, Ic are country fixed effects, and GVCcτ is either the backward GVC participation index or the backward GVC participation share. Table 7 presents the LS estimates of equation (3). Columns (1) and (3) show that on average the backward GVC participation index increases by 7.5 percent, or 1.5 percentage points, within a country after the trade and FDI liberalization event. Columns (2) and (4) add calendar year fixed effects to equation (3) 55 We thank a referee for suggesting this approach. 56 For the event studies, the only variable used is the backward GVC participation share from Eora which is available for 177 countries in 1990–2015. Hence our country sample is larger than in the rest of the analysis (where only 121 countries had information on all determinants). In the sample, 27 countries have no event while 150 countries have at least one liberalization event. For countries with no event, the year 0 used to construct the backward GVC participation index is 2015. The World Bank Economic Review 353 Figure 3. Backward Global Value Chain Participation Index for Selected Countries before and after Trade and FDI Liberalization Event. Downloaded from https://academic.oup.com/wber/article/36/2/329/6359883 by LEGVP Law Library user on 08 December 2023 Source: Authors’ analysis based on data from Eora. See the supplementary online appendix for a description of the variables used. Note: The event year 0 is the year of the trade and FDI liberalization. The backward global value chain (GVC) participation index is equal to 1 in event year 0. The actual calendar years for the event are 2001 for China, 2004 for Hungary, Nepal, Poland, Slovakia, Slovenia, and 2007 for Vietnam. Table 7. Trade and Foreign Direct Investment Liberalization Event Studies Backward GVC index Backward GVC share Dependent variables: (1) (2) (3) (4) After liberalization 0.075*** 0.056*** 0.015*** 0.006** (0.010) (0.016) (0.002) (0.003) Observations 4,602 4,602 4,602 4,602 R-squared 0.012 0.053 0.016 0.078 Number of countries 177 177 177 177 Country fixed effects Yes Yes Yes Yes Year fixed effects No Yes No Yes Source: Authors’ analysis based on data from Eora. See the supplementary online appendix for a description of the variables used. Note: The table shows estimated coefficients on an indicator for the period after a liberalization event from regressions using a backward global value chain (GVC) participation index (defined as the backward GVC participation share divided by its value in the year of the liberalization event) or backward GVC participation shares as dependent variables. LS standard errors in parentheses. ***, **, and * indicate significance levels of 1%, 5%, and 10%, respectively. 354 Fernandes, Kee, and Winkler to control for common time trends and international shocks that affect all countries in a given year. The finding that GVC participation increases post-liberalization remains robust. Overall, these event studies’ results are consistent with our previous findings based on cross-country IV regressions and country-sector difference-in-difference regressions, indicating the importance of lower tariffs and larger FDI inflows for GVC participation. Downloaded from https://academic.oup.com/wber/article/36/2/329/6359883 by LEGVP Law Library user on 08 December 2023 4. Robustness Checks and Extensions 4.1. Robustness Checks We subject our cross-country IV regressions to a variety of robustness checks. First, we address the pos- sibility that GVC participation may be measured with error for countries with a strong presence of pro- cessing exports, that is, production that relies on imported inputs but is entirely exported, as is the case in export processing zones. For such countries, the proportionality assumption made in the ICIO-based construction of GVC participation measures by Borin and Mancini (2019) is likely to be violated. We ex- clude from the estimating sample the two countries with the largest presence of processing exports, China and Mexico, and provide the results in supplementary online appendix tables S3.6 and S3.7. Second, we use alternative proxies for several of our determinants: (a) we replace the average tariff rate on manufac- turing products by an average tariff rate on intermediate inputs only (where these are defined according to the United Nations Broad Economic Category classification) providing results in supplementary online appendix tables S3.8 and S3.9, (b) we rescale the factor endowment variables with total population in- stead of GDP providing results in supplementary online appendix tables S3.10 and S3.11, (c) we replace the political stability index with the rule of law index providing results in supplementary online appendix tables S3.12 and S3.13. Third, we address the concern of an overidentified IV model given the five IVs for two endogenous variables by excluding the transitional economy status from the set of IVs providing results in supplementary online appendix table S3.14. Fourth, we reestimate all regressions decade by decade (instead of using the decadal averages) providing results in supplementary online appendix tables S3.15–S3.17 and S3.18–S3.20. The main findings on the importance of tariffs and FDI inflows but also factor endowments, institutional quality, and proximity to GVC hubs and domestic industrial capacity are robust to all these checks. 4.2. Determinants of Forward GVC Participation To complement our previous analysis on backward GVC participation, we estimate equation (1) using forward GVC participation shares as the dependent variable to assess their determinants. Column (1) of table S3.21 presents the baseline LS-BE estimation for forward GVC participation (as in column (1) of table 1 for backward GVC participation).57 Columns (2)–(3) of supplementary online appendix table S3.21 provide IV estimates for forward GVC participation shares (as in table 4 for back- ward GVC participation). The corresponding first-stage results are the same as those in table 3 for which we were able to reject the null hypothesis of weak IVs. The IV estimates in columns (2)–(3) of table S3.21 show that factor endowments play an important role for forward GVC participation, but in the opposite direction to what was found for backward GVC participation. Countries with stronger endowments of land and/or natural resources exhibit significantly higher forward GVC participation. These endowments lead to high forward participation because com- modities are used in a variety of downstream production processes that typically cross several borders. In contrast to backward GVC participation, capital stock has no effect on forward GVC participation, while labor matters. Countries with a stronger abundance in low-skilled labor exhibit lower forward GVC 57 Unreported results show that findings are maintained, adding alternative or additional controls or changing the sample as in columns (2)–(5) of table S3.21. The World Bank Economic Review 355 participation, whereas countries with a larger supply of medium- and high-skilled labor exhibit higher forward participation.58 As expected, shorter distance to GVC hubs, as well as larger domestic industrial capacity, is associated with higher forward GVC participation. Forward GVC participation is higher in countries with higher tariffs, while it is not significantly linked to FDI inflows. Perhaps surprisingly, political stability has a negative and significant effect on forward GVC participation. This finding likely reflects a sample composition effect as countries with the highest Downloaded from https://academic.oup.com/wber/article/36/2/329/6359883 by LEGVP Law Library user on 08 December 2023 forward GVC participation are low income or fragile and conflict countries richly endowed with natural resources but receiving weak FDI inflows and having deficient institutions.59 Exchange rate appreciation is unrelated to forward GVC participation, possibly as higher export prices hurt both domestic value- added embodied in bilateral partners’ exports (numerator) as well as gross exports (denominator). Note that the IV estimates of the coefficients and the corresponding t-statistics presented in columns (2) and (3) are larger in magnitude for tariffs, consistent with the previous findings, indicating that measurement errors and political economy-driven reverse causality are important. 4.3. GVC vs Traditional Trade In a final extension, we examine explicitly whether the various determinants affect GVC trade differently from traditional exports. Recall that this question was examined indirectly in the GVC participation shares regressions where a statistically significant positive coefficient indicates a stronger impact of the determinant on GVC trade (the numerator) than on traditional exports (the denominator) while an in- significant coefficient implies a similar impact of the determinant on the two types of trade. Table S3.22 shows the IV between estimates of equation (1) using as the dependent variables backward and forward GVC participation in log levels in columns (1)–(2) and gross exports in column (3). The IV cross-sectional estimates are presented in columns (4)–(6) while LS-BE estimates are omitted due to space constraints. The noticeable patterns emerging from table S3.22 are as follows. Several determinants are more im- portant in explaining backward GVC participation in levels than traditional exports, as seen from the comparison of coefficients in column (1) to those in column (3) or of coefficients in column (4) to those in column (6). Specifically, tariffs, distance to GVC hubs, and land endowments affect backward GVC par- ticipation more negatively than traditional trade, while domestic industrial capacity has a smaller positive impact on GVC trade than on traditional trade. This explains the negative effects of land endowments, domestic industrial capacity, distance, and tariffs on backward GVC participation shares in table 1. By contrast, the positive effects of capital endowments, political stability, and FDI inflows on backward GVC participation shares in table 1 imply that these factors have stronger positive effects on GVC trade than on traditional exports, as is confirmed by comparing the effects in columns (1) and (3) or in columns (4) and (6) of table S3.22. The stronger impact of FDI on backward GVC trade than on exports suggests that most of the effect that we are capturing is of the export-oriented type of FDI. Some determinants are more important in explaining forward GVC participation in levels than tra- ditional exports, as evidenced by contrasting coefficients in columns (2)–(3) or in columns (5)–(6) of table S3.22. Domestic industrial capacity has a larger positive effect on forward GVC participation than on traditional trade, while FDI inflows have a smaller positive impact on GVC trade than on traditional trade. This explains the positive effect of domestic industrial capacity and the negative effect of FDI in- flows on forward GVC participation shares in table S3.21. The significant negative effect of low-skilled 58 While this finding might appear counterintuitive, it is important to note that it is conditional on other endowments and institutional quality. Empirical evidence suggests that forward participation is highest for countries abundant in natural resources, falls strongly for countries engaged in limited manufacturing GVCs, and increases for more sophisticated GVC participation (World Bank 2019). 59 The five countries with the highest average forward GVC participation shares in the 2010s are Libya, Democratic Republic of Congo, Guinea, Algeria, and Iraq. 356 Fernandes, Kee, and Winkler labor, political stability, and distance on forward GVC participation shares in table S3.21 imply that such factors have more negative or smaller positive effects on forward GVC trade than on traditional exports. 5. Concluding Remarks The topic of GVC participation is not new, having been the object of much theoretical and empirical Downloaded from https://academic.oup.com/wber/article/36/2/329/6359883 by LEGVP Law Library user on 08 December 2023 interest in recent years. The objective of this paper is to provide a unified empirical framework to test jointly the role of different determinants highlighted in the literature as being important for trade in gen- eral, and more so for GVC trade in particular, based on data for more than 100 countries over the last three decades. Results based on cross-country IV estimation and difference-in-difference country-sector regression estimation suggest that factor endowments, geographical distance, political stability, trade pol- icy and FDI, and domestic industrial capacity are all very important in explaining GVC participation. Event studies confirm the importance of trade and FDI liberalization for GVC participation. Some of these determinants, such as trade policy, FDI, geographical distance, and factor endowments affect GVC trade more strongly than traditional trade. These findings are robust to alternative controls and country samples, and to different measurements of GVC participation and determinants. By embedding the de- terminants in a unified framework, we are able to estimate their marginal contributions and assess their relative importance for GVC participation. Given that GVCs are characterized as trade with intense firm-to-firm interactions via contracting and specialized products and investment, it is perhaps not surprising that factors such as tariffs and FDI, which affect traditional trade, have an amplified effect on GVC trade. However, other factors, such as domestic industrial capacity, are found to play a different role for GVC trade, due to a unique ability to cultivate domestic suppliers and allow countries, such as China, to climb up value chains (by increasing domestic value-added). We hope that the patterns we identified will be useful in guiding future theoretical and empirical studies of GVC formation and growth. 6. Data Availability The derived data generated in this research are available in the article and in its online supplementary material. References Acemoglu, D., P. Antràs, and E. Helpman. 2007. “Contracts and Technology Adoption.” American Economic Review 97 (3): 916–43. Amiti, M., O. Itskhoki, and J. Konings. 2014. “Importers, Exporters, and Exchange Rate Disconnect.” American Economic Review 104 (7): 1942–78. Andrews, I., J. Stock, and L. Sun. 2019. “Weak Instruments in Instrumental Variables Regression: Theory and Practice.” Annual Review of Economics 11 (1): 727–53. Ansón, J., J. Arvis, M. Boffa, M. Helble, and B. Shepherd. 2017. “Time, Uncertainty, and Trade Flows.” Working Paper 673, Asian Development Bank, Manila, Philippines. Antràs, P. 2016. Global Production: Firms, Contracts and Trade Structure, Princeton University Press. —–. 2020. “Conceptual Aspects of Global Value Chains.” World Bank Economic Review 34 (3): 551–74. Antràs, P., and A. De Gortari. 2020. “On the Geography of Global Value Chains.” Econometrica 88 (4): 1553–98. Antràs, P., and E. Helpman. 2004. “Global Sourcing.” Journal of Political Economy 112 (3): 552–80. —–. 2008. “Contractual Frictions and Global Sourcing.” In The Organization of Firms in a Global Economy, edited by E. Helpman, D. Marin, and T. Verdier. Cambridge, MA: Harvard University Press. Arkolakis, C., A. Costinot, and A. Rodriguez-Clare. 2012. “New Trade Models, Same Old Gains?” American Eco- nomic Review 102 (1): 94–130. The World Bank Economic Review 357 Aslam, A., N. Novta, and F. Rodrigues-Bastos. 2017. “Calculating Trade in Value Added.” IMF Working Paper WP/17/178. Bagwell, K., and R. Staiger. 2011. “What Do Trade Negotiators Negotiate About? Empirical Evidence from the World Trade Organization.” American Economic Review 101 (4): 1238–73. Baldwin, R. 2012. “Trade and Industrialization after Globalization’s Second Unbundling: How Building and Joining a Supply Chain Are Different and Why It Matters.” In Globalization in an Age of Crisis: Multilateral Economic Cooperation in the Twenty-First Century. University of Chicago Press. Downloaded from https://academic.oup.com/wber/article/36/2/329/6359883 by LEGVP Law Library user on 08 December 2023 —–. 2016. Globalization’s Three Unbundlings. Harvard University Press. Baldwin, R., and D. Taglioni. 2012. “Gravity Chains: Estimating Bilateral Trade Flows When Parts and Components Trade Is Important.” NBER Working Paper No. 16672. Balié, J., D. Del Prete, E. Magrini, P. Montalbano, and S. Nenci. 2019. “Does Trade Policy Impact Food and Agriculture Global Value Chain Participation of Sub-Saharan African Countries?” American Journal of Agricultural Economics 101 (3): 773–89. Banh, H., P. Wingender, and C. A. Gueye. 2020. “Global Value Chains and Productivity: Micro Evidence from Esto- nia.” IMF Working Paper No. 20/117. Bas, M., and V. Strauss-Kahn. 2015. “Input-Trade Liberalization, Export Prices, and Quality Upgrading.” Journal of International Economics 95 (2): 250–62. Beck, T. 2003. “Financial Dependence and International Trade.” Review of International Economics 11: 296–316. Bems, R., and K. Kikkawa. 2020. “Measuring Trade in Value Added with Firm-Level Data,” mimeo. Bickerdike, C. 1907. “Review of A. C. Pigou’s Protective and Preferential Import Duties.” Economic Journal 17 (65): 98–108. Blanchard, E. 2010. “Reevaluating the Role of Trade Agreements: Does Investment Globalization Make the WTO Obsolete?” Journal of International Economics 82 (1): 63–72. —–. 2015. “A Shifting Mandate: International Ownership, Global Fragmentation, and a Case for Deeper Integration under the WTO.” World Trade Review 14 (1): 87–99. Blanchard, E., C. Bown, and R. Johnson. 2016. “Global Supply Chains and Trade Policy,” NBER Working Paper No. 21883. Blyde, J. (ed.) 2014. Synchronized Factories: Latin America and the Caribbean in the Era of Global Value Chains. Inter-American Development Bank, Washington, DC. Borin, A., and M. Mancini. 2015. “Follow the Value Added: Bilateral Gross Export Accounting.” Economic Working Papers, Bank of Italy. —–. 2019. “Measuring What Matters in Global Value Chains and Value-Added Trade.” Policy Research Paper 8804, World Bank, Washington, DC. Braun, M. 2003. “Financial Contractibility and Assets’ Hardness.” mimeo UCLA. http://www.anderson.ucla.edu/ faculty/matias.braun/. Broda, C., N. Limao, and D. Weinstein. 2008. “Optimal Tariffs and Market Power: The Evidence.” American Eco- nomic Review 98 (5): 2032–65. Brooks, D., and B. Ferrarini. 2012. “Vertical Gravity.” ADB Working Paper 303, Manila, Philippines. Buelens, C., and M. Tirpak. 2017. “Reading the Footprints: How Foreign Investors Shape Countries’ Participation in Global Value Chains.” Comparative Economic Studies 59 (4): 561–84. Caliendo, L., and F. Parro. 2015. “Estimates of the Trade and Welfare Effects of NAFTA.” Review of Economic Studies, 82 (1): 1–44. Cheng, K., S. Rehman, D. Seneviratne, and S. Zhang. 2015. “Reaping the Benefits from Global Value Chains.” IMF Working Paper WP/15/204. Washington, DC: International Monetary Fund. Cherkashin, I., S. Demidova, H. L. Kee, and K. Krishna. 2015. “Firm Heterogeneity and Costly Trade: A New Esti- mation Strategy and Policy Experiments.” Journal of International Economics 96 (1): 18–36. Chor, D. 2010. “Unpacking Sources of Comparative Advantage: A Quantitative Approach.” Journal of International Economics 82 (2): 152–67. Christ, N., and M. Ferrantino. 2011. “Land Transport for Exports: The Effects of Cost, Time, and Uncertainty in Sub-Saharan Africa.” World Development 39 (10): 1749–59. Constantinescu, C., A. Mattoo, and M. Ruta. 2019. “Does Vertical Specialisation Increase Productivity?” World Econ- omy 42 (8): 2385–402. 358 Fernandes, Kee, and Winkler Costinot, A. 2009. “On the Origins of Comparative Advantage.” Journal of International Economics 77: 255–64. Desai, M., C. Foley, and J. Hines Jr. 2004. “Foreign Direct Investment in a World of Multiple Taxes.” Journal of Public Economics 88 (12): 2727–44. Djankov, S., C. Freund, and C. Pham. 2010. “Trading on Time.” Review of Economics and Statistics 92 (1): 166–73. Eaton, J., and S. Kortum. 2002. “Technology, Geography, and Trade.” Econometrica 70 (5): 1741–79. Feenstra, R. 1998. “Integration of Trade and Disintegration of Production in the Global Economy.” Journal of Eco- nomic Perspectives 12 (4): 31–50. Downloaded from https://academic.oup.com/wber/article/36/2/329/6359883 by LEGVP Law Library user on 08 December 2023 Feenstra, R., and G. Hanson. 1997. “Foreign Direct Investment and Relative Wages: Evidence from Mexico’s Maquiladoras.” Journal of International Economics 42 (3/4): 371–93. Felbermayr, G. F. T., and E. Yalcin. 2019. “Rules of Origin and the Profitability of Trade Deflection.” Journal of International Economics 121: 103248. Fernandes, A., A. Forero, H. Maemir, and A. Mattoo. 2019a. “Are Trade Preferences a Panacea? The African Growth and Opportunity Act and African Exports.” Policy Research Working Paper 8753, World Bank, Washington, DC. Fernandes, A., A. Mattoo, H. Nguyen, and M. Schiffbauer. 2019b. “The Internet and Chinese Exports in the Pre-Ali Baba Era.” Journal of Development Economics 138: 57–76. Fortanier, F., G. Miao, A. Kolk, and N. Pisani. 2020. “Accounting for Firm Heterogeneity in Global Value Chains.” Journal of International Business Studies 51, 432–53. Grossman, G. M., and E. Helpman. 1994. “Protection for Sale.” The American Economic Review 84 (4): 833–50. Hallward-Driemeier, M., and G. Nayyar. 2019. “Have Robots Grounded the Flying Geese? Evidence from Greenfield FDI in Manufacturing.” Policy Research Paper 9097, World Bank, Washington, DC. Hausman, J. 2001. “Mismeasured Variables in Econometric Analysis: Problems from the Right and Problems from the Left.” Journal of Economic Perspectives 15 (4): 57–67. Head, K., and T. Mayer. 2004. “Market Potential and the Location of Japanese Investment in the European Union.” Review of Economics and Statistics 86 (4): 959–72. Hofmann, C., A. Osnago, and M. Ruta. 2017. “Horizontal Depth: A New Database on the Content of Preferential Trade Agreements.” Policy Research Working Paper Series 7981, The World Bank. Hummels, D., J. Ishii, and K.-M. Yi. 2001. “The Nature and Growth of Vertical Specialization in World Trade.” Journal of International Economics 54 (1): 75–96. Hummels, D., and G. Schaur. 2013. “Time as a Trade Barrier.” American Economic Review 103 (7): 2935–59. Ignatenko, A., F. Raei, and B. Mircheva. 2019. “Global Value Chains: What Are the Benefits and Why Do Countries Participate?” IMF Working Paper No. 19/18. International Monetary Fund. 2000. Transition Economies: An IMF Perspective on Progress and Prospects. Ito, T., L. Rotunno, and P. Vézina. 2017. “Heckscher-Ohlin: Evidence from Virtual Trade in Value Added,” Review of International Economics 25 (3): 427–46. Jiao, Y., and S.-J. Wei. 2020. “Natural Barriers and Policy Barriers.” NBER Working Paper 28171. Johnson, R., and G. Noguera. 2012. “Accounting for Intermediates: Production Sharing and Trade in Value Added.” Journal of International Economics 86 (2): 224–36. —–. 2017. “A Portrait of Trade in Value-Added over Four Decades.” Review of Economics and Statistics 99 (5): 896–911. Kasahara, H., and B. Lapham. 2013. “Productivity and the Decision to Import and Export: Theory and Evidence.” Journal of International Economics 89 (2): 297–316. Kee, H. L. 2015. “Local Intermediate Inputs, and the Shared Supplier Spillovers of Foreign Direct Investment.” Journal of Development Economics 112: 56–71. Kee, H. L., A. Nicita, and M. Olarreaga. 2008. “Import Demand Elasticities and Trade Distortions.” Review of Eco- nomics and Statistics 90 (4): 666–82. Kee, H. L., and H. Tang. 2016. “Domestic Value Added in Exports: Theory and Firm Evidence from China.” American Economic Review 106 (6): 1402–36. Koopman, R., Z. Wang, and S. Wei. 2014. “Tracing Value-Added and Double Counting in Gross Exports.” American Economic Review 104 (2): 459–94. Kowalski, P., J. Gonzalez, A. Ragoussis, and C. Ugarte. 2015. “Participation of Developing Countries in Global Value Chains: Implications for Trade and Trade-Related Policies.” Trade Policy Papers 179, Organisation for Economic Co-operation and Development, Paris. The World Bank Economic Review 359 Laget, E., A. Osnago, N. Rocha, and M. Ruta. 2018. “Deep Trade Agreements and Global Value Chains.” Policy Research Working Paper 8491, World Bank, Washington, DC. Lenzen, M., D. Moran, K. Kanemoto, and A. Geschke. 2013. “Building Eora: A Global Multi-Regional Input-Output Database at High Country and Sector Resolution.” Economic Systems Research 25 (1): 20–49. Levchenko, A. 2007. “Institutional Quality and International Trade.” Review of Economic Studies 74 (3): 791–819. Manova, K. 2008. “Credit Constraints, Equity Market Liberalizations and International Trade.” Journal of Interna- tional Economics 76 (1): 33–47. Downloaded from https://academic.oup.com/wber/article/36/2/329/6359883 by LEGVP Law Library user on 08 December 2023 —–. 2015. “Global Value Chains and Multinational Activity with Financial Frictions.” In The Age of Global Value Chains: Maps and Policy Issues, edited by J. Amador and F. di Mauro. CEPR E-book, http://www.voxeu.org/content/age-global-value-chains-maps-and-policy-issues. Miller, R., and U. Temurshoev. 2017. “Output Upstreamness and Input Downstreamness of Industries/Countries in World Production.” International Regional Science Review 40 (5): 443–75. Muûls, M., and M. Pisu. 2009. “Imports and Exports at the Level of the Firm: Evidence from Belgium.” The World Economy 32 (5): 692–734. Mutti, J., and H. Grubert. 2004. “Empirical Asymmetries in Foreign Direct Investment and Taxation.” Journal of International Economics 62 (2): 337–58. Noguera, G. 2012. “Trade Costs and Gravity for Gross Trade and Value Added Trade.” UC Berkeley and Columbia University, Job Market Paper. Nunn, N. 2007. “Relationship-Specificity, Incomplete Contracts, and the Pattern of Trade.” Quarterly Journal of Economics 122 (2): 569–600. OECD and WTO. 2013. Aid for Trade at a Glance 2013: Linking to Value Chains. Paris: Organisation for Economic Co-operation and Development. Orefice, G., and N. Rocha. 2014. “Deep Integration and Production Networks: An Empirical Analysis.” World Econ- omy 37 (1): 106–36. Pahl, S., and M. Timmer. 2020. “Do Global Value Chains Enhance Economic Upgrading? A Long View.” Journal of Development Studies 56 (9): 1683–1705. Pierola, M., A. Fernandes, and T. Farole. 2018. “The Role of Imports for Exporter Performance in Peru.” World Economy 41 (2): 550–72. Rajan, R., and L. Zingales. 1998. “Financial Dependence and Growth.” American Economic Review 88 (3): 559–86. Reed, W. 2015. “On the Practice of Lagging Variables to Avoid Simultaneity.” Oxford Bulletin of Economics and Statistics 77 (6): 897–905. Rodríguez-Pose, A., V. Tselios, D. Winkler, and T. Farole. 2013. “Geography and the Drivers of Firm Exports in Indonesia.” World Development 44 (C): 225–40. Romalis, J. 2004. “Factor Proportions and the Structure of Commodity Trade.” American Economic Review 94 (1): 67–97. Stock, J., and M. Yogo. 2005. “Testing for Weak Instruments in Linear IV Regression.” In Identification and Inference for Econometric Models: Essays in Honor of Thomas J. Rothenberg, edited by J. Stock and D. Andrews, chapter 5. Cambridge University Press. Stolzenburg, V., D. Taglioni, and D. Winkler. 2019. “Economic Upgrading through Global Value Chain Participation: Which Policies Increase the Value-added Gains?” In Handbook on Global Value Chains, edited by S. Ponte, G. Gereffi, and G. Raj-Reichert, chapter 30, pp. 483–505. Northhampton, MA: Edward Elgar Publishing. Timmer, M., E. Dietzenbacher, B. Los, R. Stehrer, and G. de Vries. 2015. “An Illustrated User Guide to the World Input–Output Database: The Case of Global Automotive Production.” Review of International Economics 23 (3): 575–605. Timmer, M., A. Erumban, B. Los, R. Stehrer, and G. de Vries. 2014. “Slicing Up Global Value Chains.” Journal of Economic Perspectives 28 (2): 99–118. Timmer, M., B. Los, R. Stehrer, and G. de Vries. 2016. “An Anatomy of the Global Trade Slowdown Based on the WIOD 2016 Release.” GGDC research memorandum number 162, University of Groningen. United Nations. 2015. “Global Values and Interconnectedness of Asia-Pacific Economies.” In Supporting Participation in Value Chains, Asia-Pacific Trade and Investment Report 2015, 103–38. —–. 2019. Handbook on Accounting for Global Value Chains: GVC Satellite Accounts and Integrated Business Statis- tics, UN Statistics Division. 360 Fernandes, Kee, and Winkler Vijil, M., L. Wagner, and M. Woldemichael. 2019. “Import Uncertainty and Export Dynamics.” Policy Research Working Paper 8793, World Bank, Washington, DC. Wagner, J. 2012. “International Trade and Firm Performance: A Survey of Empirical Studies Since 2006.” Review of World Economics 148 (2): 235–67. Wang, Z., S. Wei, X. Yu, and K. Zhu. 2016. “Characterizing Global Value Chains.” mimeo www.cepr.org/ sites/default/files/Wang%20Zhi.pdf. Wang, Z., S. Wei, and K. Zhu. 2013. “Quantifying International Production Sharing at the Bilateral and Sector Levels.” Downloaded from https://academic.oup.com/wber/article/36/2/329/6359883 by LEGVP Law Library user on 08 December 2023 NBER Working Paper, No. 19677. Wooldridge, J. 2012. Introductory Econometrics: A Modern Approach, 5e. Mason, Ohio: South-Western, Cengage Learning. World Bank. 2019. Trading for Development in the Age of Global Value Chains, World Development Report 2020. Washington, DC: World Bank. Yi, K.-M. 2003. “Can Vertical Specialization Explain the Growth of World Trade?” Journal of Political Economy 111 (1): 52–102. Yi, K. M. 2010. “Can Multistage Production Explain the Home Bias in Trade?” American Economic Review 100 (1): 364–93.