Policy Research Working Paper 10667 Are Global Value Chains Women Friendly in Developing Countries? Evidence from Firm-Level Data Marize Kalliny Chahir Zaki Middle East and North Africa Region Office of the Chief Economist January 2024 Policy Research Working Paper 10667 Abstract Despite the efforts made to increase women’s inclusion female employees, especially production ones. A less robust in the economy, they are still underrepresented in trade negative effect is found for the impact on being a female top in general and in global value chains in particular. Thus, manager. These effects are moderated by the inclusion of this paper aims at examining the impact of global value gender provisions in trade agreements and by the character- chains on women’s trade participation as entrepreneurs and istics of the investment climate (especially tax policy, access employees. It also analyzes how this effect is moderated to finance, and corruption). These results remain robust through external (gender provisions in trade agreements) after controlling for the endogeneity of global value chains and internal (investment climate variables) factors. The using an instrumental variable approach and a propensity analysis uses firm-level data for 154 developing economies score estimation method where the treatment is being part and emerging markets with a special focus on the Middle of a global value chains. Thus, global value chains can be East and North Africa region, being one of the regions perceived as a tool that boosts women’s empowerment in with the lowest female labor force participation. The main emerging economies, especially in the Middle East and findings show that global value chains integration increases North Africa region. the likelihood of being a female owner and the share of This paper is a product of the Office of the Chief Economist, Middle East and North Africa Region. It is part of a larger effort by the World Bank to provide open access to its research and make a contribution to development policy discussions around the world. Policy Research Working Papers are also posted on the Web at http://www.worldbank.org/prwp. The authors may be contacted at chahir.zaki@feps.edu.eg. The Policy Research Working Paper Series disseminates the findings of work in progress to encourage the exchange of ideas about development issues. An objective of the series is to get the findings out quickly, even if the presentations are less than fully polished. The papers carry the names of the authors and should be cited accordingly. The findings, interpretations, and conclusions expressed in this paper are entirely those of the authors. They do not necessarily represent the views of the International Bank for Reconstruction and Development/World Bank and its affiliated organizations, or those of the Executive Directors of the World Bank or the governments they represent. Produced by the Research Support Team Are Global Value Chains Women Friendly in Developing Countries? Evidence from Firm-Level Data 1 Marize Kalliny 2 Chahir Zaki 3 JEL Classification : F12, F16, F23, J16 Keywords: Global Value Chains, Gender, Empowerment, Firm-level, MENA 1 The authors gratefully acknowledge the financial and analytical support from the Office of the Chief Economist for the Middle East and North Africa (MNACE) under the regional Labor and Gender Research Programs (TTLs: Nelly Elmallakh and Nazmul Chaudhury). The findings, interpretations, and conclusions expressed in this paper are entirely those of the authors. They do not necessarily represent the views of the International Bank for Reconstruction and Development/World Bank and its affiliated organizations, or those of the Executive Directors of the World Bank or the governments they represent. The authors are grateful to Mhamed Ben Saleh, Enrico Marvasi and the anonymous referees for constructive feedback. 2 PhD Candidate at Université Clermont Auvergne, CNRS, IRD, CERDI, F-63000 Clermont-Ferrand, France. Assistant Lecturer of Economics, Faculty of Economics and Political Science – Cairo University (on leave). Email: marize.kalliny@feps.edu.eg 3 Chaired Professor of Economics, University of Orléans and Labor d’Economie d’Orléans, Economic Research Forum and EMANES. Email: chahir.zaki@univ-orleans.fr 1. Introduction Women have always played a significant and pivotal role in the economy through their remarkable presence in business, agriculture, industry and even through their unpaid care work. Despite the efforts made to increase women’s inclusion in the economy, around 2.4 billion women do not have access to equal economic opportunities as men according to the World Bank’s Women, Business and the Law report (2022). Furthermore, based on a report published on women by the United Nations Development Programme (UNDP), women perform 66% of the world’s work, produce 50% of the food, but earn only 10% of the income. In addition, different stages of the production process are increasingly being fragmented across different countries through global value chains (GVCs), especially in emerging economies. For this reason, this study attempts to assess the impact of GVCs on women’s empowerment in international trade. Indeed, in recent years, the world has witnessed a growing interest in the impact of globalization on economic patterns in general and gender equality in particular. A significant focus has been on the gender impact of GVCs as they have proven their power to generate employment, drive development, and increase income. These value chains can help women by providing more income that can support their economic empowerment but can also downgrade them to poorly paid and undervalued jobs. Hence, considering gender issues and addressing them is critical in order to take advantage of the potential of GVCs, which, in turn, guarantees a better achievement of sustainable economic and social goals by 2030. In the previous literature, the relation between women’s empowerment and international trade was ambiguous. The first strand of studies was in line with the neoclassical theory based on Becker (1959) and according to which discrimination is costly. Hence, the increase of industry competitiveness due to trade participation reduces the incentives to discriminate against women especially in concentrated industries than in competitive ones. For instance, Boler et al. (2015) use matched employer-employee data from the Norwegian manufacturing sector and prove that trade participation has narrowed the gender wage gap in exporting firms relative to non-exporters. Similarly, Juhn et al. (2014) and Black and Brainerd (2004) reach the same conclusion using various datasets. Nevertheless, contradictory results have been proven in other studies. For instance, in a study based in India, Taiwan, and China, Berik et al. (2004) show that competition resulting from international trade increases wage discrimination against female workers, which does not go in line with the neoclassical theory. Moreover, using difference-in-difference estimation and data from the Demographic Census for 1991 and 2000 fielded by the Brazilian Census Bureau, Gaddis and Pieters (2017) show that trade liberalization decreased the male and female labor force participation rate. The effect is significantly larger on men, which means that liberalization reduced the gender gap in employment and participation rates. Nevertheless, the results show no evidence that women got any benefit from the competitive effects of liberalization as their employment and participation rate did not increase relative to those of men. The literature also shows that the limited or negative employment effect was due to other concomitant factors such as the effect of anti-sweatshop activism (Harrison and Scorse 2010) and the constitution of minimum wage for export tariff privileges (del Carpio et al 2015). By studying the opposite dimension of the subject, Karam and Zaki (2021) argue that female labor participation in the MENA region has a positive significant impact on both trade margins. Furthermore, their results show that female ownership positively affects the probability of exports of large firms. 2 Regarding the relation between GVCs and women’s empowerment, some strands of the literature studied the impact of GVCs on gender wage disparity. For instance, Deb (2021) uses the Trade in Value Added Database (TiVA) by the Organization for Economic Co-operation and Development (OECD) and conclude that neither backward nor forward linkages were able to improve the relative wages of female workers in India. Nevertheless, Jenkins (2005) confirmed that women wages and working conditions are better in GVCs. Furthermore, few papers have discussed the impact of GVCs on women’s employment in general. GVC participation has been proven to increase female employment especially for developing countries (Shepherd and Stone, 2013; Bamber and Staritz, 2016). Yet, it is important to note that the impact of GVCs on female empowerment can be moderated through several external and internal factors. At the external level, the scope of drafting of regional and bilateral trade agreements has been largely expanded to accommodate several Sustainable Development Goals. A large set of agreements prioritized environmental matters through the inclusion of climate change and environmental protection provisions (Martínez-Zarzoso, 2018). Labor rights related provisions also gained a large interest as starting from 2016, more than 136 countries negotiated at least one free trade agreement (FTA) that encloses labor rights related provisions (ILO, 2017; Harrison, 2019). However, only few strides have been made in order to include more gender related provisions in FTAs since among all FTAs in force, only 20% of them include explicit chapters or clauses that endeavor to achieve gender equality and to empower women (Monteiro, 2021). Similarly, according to the gender and trade report (UNCTAD, 2020), only 74 out of 500 RTAs (around 15%) include provisions that prioritize gender issues. Yet, it is worth mentioning that it is not the number of provisions or length of chapters including gender considerations that matter but their enforcement. Bahri (2021) shows how some RTAs such as Canada-Chile and Canada-Israel Agreements include whole chapters that address gender issues; however, there is a lack of legal obligations that ensure their proper implementation. Conversely, other RTAs such as the Stabilization and Association Agreement between the European Communities and the Republic of Montenegro, where gender provisions are included in the chapter on development and not standalone chapters for gender issues, are more efficient due to the existence of strong legal obligations that enforce the parties to respect the clauses and to stop any discrimination based on gender. At the internal level, and given the discrimination against women on the labor market, they might face more barriers when it comes access to finance, permits to start a business, tax policy, etc. This is why it is important to see how the effect of GVCs on women’s empowerment is moderated by the obstacles they might face. Against this background, there are no previous empirical studies, to our knowledge, that tackle the impact of GVCs on women’s entrepreneurship in developing countries and especially the Middle East and North Africa (MENA) region, which is characterized by both low female labor force participation and low integration into GVCs. In addition, no empirical studies have been conducted to assess the impact of gender provisions on women’s engagement in international trade in the MENA region. Indeed, compared to the other regions, the MENA region is ranked the lowest in the world for attaining gender equality based on the World Economic Forum’s Global Gender Gap Report (2021). Moreover, based on the Women Peace and Security (WPS) Index, the MENA region’s performance was very poor as it comprises 12 of the 25 worst performing countries globally (Danon and Collin, 2021). Therefore, the contribution of this paper is twofold. First, it focuses on the impact of GVCs on women’s trade participation as entrepreneurs and employees. 3 Second, we analyze how this effect is moderated through external (gender provisions in trade agreements) and internal (investment climate variables) factors. To do so, we use firm-level data for 154 developing economies and emerging markets with a special focus on the Middle East and North Africa region. Our main findings show that GVC integration increases the likelihood of being a female owner and the share of female employees, especially production ones. A less robust negative effect is found regarding the impact on being a female top manager. These effects are moderated by the inclusion of gender provisions in trade agreements and by the characteristics of the investment climate (especially tax policy, courts, access to finance and corruption). These results remain robust after we control for the endogeneity of GVC using an instrumental variable approach and a propensity score estimation method where the treatment is being part of a GVC. Thus, GVCs can be perceived as a tool that boosts women’s empowerment in emerging economies, especially in the MENA region. The remainder of the paper is organized as follows. Section 2 presents the data we use and some stylized facts on GVCs and women’s participation. Section 3 is dedicated to the methodology and the econometric specification. Section 4 presents the empirical results. Section 5 provides robustness checks. Section 6 concludes and provides some policy implications to increase women’s employment through trade and GVC channels. 2. Data and Stylized Facts To explore the nexus between firms’ integration into GVCs and women’s trade participation as entrepreneurs and employees, firm-level pooled data from the WBES is used. These surveys cover a broad range of business environment topics such as access to finance, trade, corruption, competition, and infrastructure for 143,598 firms in 154 developing economies and emerging markets. The manufacturing and services sectors are the primary business of interest in these surveys conducted in a range of time that varies from 2006 to 2021 (see Appendix 1). The objective of this section is to provide some descriptive statistics related to the nexus between GVCs and women’s participation. To define GVCs, we follow the definition of Dovis and Zaki (2020) where the least strict definition includes firms that export and import simultaneously (GVC1). Second, two stricter definitions are related to firms who are simultaneously exporters and importers and have either an international certification (GVC2) or a share of its capital owned by a foreign firm (GVC3). The strictest definition combines the four criteria altogether (GVC4). This variable is a dummy variable that takes a value of one if the firm is part of a GVC and zero otherwise. Based on these definitions, around one-third of firms are part of GVCs with most of them being two-way traders and only 2.3% that are two-way trader, have a foreign capital and an international certification (see Table A1 in Appendix 3). It is important to note that these two definitions help us measure GVC participation at the extensive margin level, not the intensive margin one. At the regional level (see Figure A1 in Appendix 3), Europe and Central Asia (ECA) is the most integrated region in GVCs (for all definitions) and South Asia the least integrated one. The MENA region, while being slightly better than South Asia, is still far from other top performers such as ECA and Latin American and the Caribbean. This confirms previous findings of the literature on GVCs in the MENA region that is characterized by an unfriendly business climate (Dovis and 4 Zaki, 2020), political connections (Kruse et al., 2021 and Aboushady and Zaki, 2022) and the presence of different trade barriers (whether tariffs or non-tariff measures, see Karam and Zaki (2021)). As it was mentioned before, this paper attempts to examine the nexus between women’s labor participation and GVCs integration. LAC has the highest share of full-time female employees followed by EAP and ECA and on the other extreme South Asia and the MENA region. In addition, for all the regions production workers are much higher than non-production ones (see Table A2 in Appendix 3). While the former are mainly working in the manufacturing sector, the latter are in the services one. This is confirmed by Table 1 that shows that being integrated in a GVC is positively associated to a larger number of females, whether production or non-production workers but more production ones. This result applies to the different GVCs definitions bearing in mind that for the most restrictive definition, the results are also driven by the firm size (as larger firms are more likely to be part of GVCs). Such a positive association is rather good news given that GVCs should mainly take place in the manufacturing sector, which is likely to create more jobs for female production workers (blue collars) that are abundant in emerging economies. Table 1. The Average Number of Female Employees and GVCs GVC1 GVC2 GVC3 GVC4 No Yes No Yes No Yes No Yes Female Employees 19.81 43.19 20.60 79.27 21.12 87.86 21.30 108.81 Female Production Workers 14.35 72.83 21.45 85.13 23.95 129.54 26.50 126.43 Female Non-Production Workers 5.30 18.48 6.20 25.86 7.66 28.18 7.90 35.07 Source: Constructed by the authors using the WBES. Note: GVC1 refers to firms that export and import simultaneously, GVC2 = GVC1 + international certification, GVC3= GVC1+ share of its capital owned by a foreign firm, GVC4 combines the four criteria altogether. Yet, it is important to look at women’s participation from a broader lens by taking into consideration, not only female workers, but whether the owner or the manager of the firm are females. These two measures can give a clearer picture of women’s empowerment as they are associated to more female power within the firm. Thus, Figure 1 compares firms that are part of GVCs whose owner or manager are women. Three remarks are worth mentioning. First, generally, the share of firms that are owned or managed by females is limited, compared to those owned or managed by males. Second, females that own a firm that is part of a GVC are higher than those who just manage it. This is closely related to the concept of empowerment as it captures the effect on women’s empowerment more than management given that the manager is, at the end of the day, an employee in the firm that takes orders from the owner, while the owner is an entrepreneur that takes financial and business risks on his/her own (Karam and Zaki, 2021). Third, these shares decrease with more restrictive definition of GVCs (for ownership it decreased from 36% to 26% of firms and for management from 13% to 12%). 5 Figure 1. Share of Female Ownership and Management in Firms integrating into GVCs 100% 90% 64% 64% 73% 74% 87% 89% 87% 88% 80% 70% 60% 50% 40% 30% 36% 36% 20% 27% 26% 10% 0% 13% 11% 13% 12% GVC1 GVC2 GVC3 GVC4 GVC1 GVC2 GVC3 GVC4 (a) Female Ownership (b) Female Top Manager Yes No Source: Constructed by the authors using the WBES. Note: GVC1 refers to firms that export and import simultaneously, GVC2 = GVC1 + international certification, GVC3= GVC1+ share of its capital owned by a foreign firm, GVC4 combines the four criteria altogether. When we look at the relationship between female participation, firm size, and GVC integration, two remarks are worthwhile. First, for large firms and the strict definition of GVC, females (whether they are owners or managers) tend to be under-represented. Second, generally, for less restrictive definitions of GVC, we do not observe significant differences between small, medium, and large firms when it comes to female ownership and management. As for regional differences, Figure 3 shows that females are doing better in terms of ownership compared to management when it comes to GVC integration. Indeed, for all GVC definitions in different regions, the share of firms that are part of GVCs and that are owned by females is greater than the one of firms managed by females with a slightly lower figures for the most restrictive definition (GVC4) as it is shown in Figure 2. In addition, for the most restrictive definition, EAP followed by SAR are the best performers whereas MENA and LAC are the worst in terms of female ownership and management (see Figure 3). This is due to the low labor cost that helps develop value chains, especially in the textile and garments sector (Kumar, 2017). Indeed, South Asia has the second-highest level of GVC exports out of total exports among developing regions, chiefly thanks to final and intermediate apparel products (Lopez-Acevedo et al., 2017). This is why, as it was mentioned before, the MENA region’s problem of women’s participation is also reflected in their integration into GVCs as owners or managers. 6 Figure 2. The Percentage of Firms Owned/ Managed by Female integrating into GVCs, by size 100% 90% 38.0% 52.5% 51.2% 63.3% 33.0% 49.2% 54.5% 63.7% 80% 70% 60% 36.0% 36.8% 50% 32.1% 33.0% 40% 34.4% 29.8% 30% 27.3% 25.2% 31.0% 20% 25.2% 10% 16.7% 17.8% 15.7% 13.1% 11.1% 0% 9.4% GVC1 GVC2 GVC3 GVC4 GVC1 GVC2 GVC3 GVC4 (a) Owned (b) Managed Small Medium Large Source: Constructed by the authors using the WBES. Note: GVC1 refers to firms that export and import simultaneously, GVC2 = GVC1 + international certification, GVC3= GVC1+ share of its capital owned by a foreign firm, GVC4 combines the four criteria altogether. Figure 3. The Percentage of Firms Owned/ Managed by Female integrating into GVCs, by region 100.0% 90.0% 80.0% 51.2% 70.0% 64.0%64.3% 65.7% 71.1% 62.5% 69.6% 72.3% 74.7% 70.7% 74.1% 60.0% 81.8% 79.4% 86.8% 89.0% 84.9% 78.4% 87.0% 88.5% 92.7% 88.1% 93.6% 50.0% 95.1% 95.0% 40.0% 30.0% 48.8% 20.0% 36.0%35.7% 34.3% 28.9% 37.5% 30.4% 27.7% 25.3% 29.3% 25.9% 10.0% 18.2% 20.6% 13.2% 11.0% 15.1% 21.6% 13.0% 11.5% 7.3% 11.9% 6.4% 0.0% 4.9% 5.0% ECA ECA ECA ECA SSA SAR SSA SAR SAR SAR EAP LAC EAP LAC SSA EAP LAC SSA EAP LAC MENA MENA MENA MENA GVC1 GVC4 GVC1 GVC4 (a) Female Ownership (b) Female Top Manager Yes No Source: Constructed by the authors using the WBES. Note: SAR stands for South Asia, MENA Middle East and North Africa, EAP East Asia and Pacific, SSA Sub-Saharan Africa, LAC Latin America, and the Caribbean and ECA Europe and Central Asia. We argue that the effect of GVC on women’s participation is moderated through internal and external factors. Generally, both factors might be gender blind as rules and regulations (to start a business or in trade agreements) do not have a gender lens (see Figure 4). However, their impact differs across individuals, and must be either gender mainstreamed or gender focused to improve women economic empowerment in a more explicit way. This is the case of gender provisions in trade agreements or rules that might be more gender-friendly. 7 Figure 4. Levels of Gender Inclusion in Business Environment Reform → Gender Focused Policy priorities targeted at enhancing WEE → Gender Mainstreaming Analysis gender sensitive Reform design and implementation are inclusive → Gender Aware Data disaggregated and efforts to minimize harm. No explicit priority on WEE and greater inclusion in policymaking → Do No Harm Focused on minimizing risk/unintended consequences. No positive equity goals Gender Blind No inclusion of gender lens in analysis, design or implementation Source: ILO (2021). Note: WEE stands for Women Economic Empowerment. At the internal level, Table 2 presents the share of firms reporting each variable as the biggest obstacle. A large heterogeneity is observed across different regions. For instance, while firms report that access to finance is a major obstacle in Sub-Saharan Africa, those in the MENA region complain mainly about business licensing, permits, corruption, and political instability. In Latin America, business and licensing, crimes and disorder, customs and trade regulations, and practices of the informal sector seem to be major obstacles. Finally, in Europe and Central Asia, problems related to inadequately educated workforce, labor regulations, tax rate, and administration are onerous. Obviously, such problems that affect the business environment might also hinder women’s participation. This is why we take different obstacles into consideration to see how they can amplify the impact of GVC on women’s participation. 8 Table 2. Percent of Firms Choosing Their Biggest Obstacle (%) All Countries EAP ECA LAC MENA SAR SSA Access to finance 14.2 13.6 9.4 9 12.1 12.1 23.9 Land 2.9 5.6 1.7 1.4 1.5 4.6 4.3 Business licensing and permits 2.5 2.8 2.7 4.4 3.5 1 1.4 Corruption 6.5 6.6 3.8 9.7 9.7 10.5 7.3 Courts 0.9 0.8 1.2 1 1 0.4 0.6 Crime, theft, and disorder 2.9 3 1.7 6.7 1.6 2.7 3.3 Customs and trade regulations 3.3 2.9 2.6 4.2 3.4 1.8 4.5 Electricity 8.4 6.5 3.5 5.4 10.9 21.8 13.9 Inadequately educated workforce 10.3 7.6 19.7 7.8 8.8 2.9 1.8 Labor regulations 3.5 2.9 5.3 4.5 2.4 4.8 1 Political instability 11.9 15.3 9 11.1 19.5 20.7 10.8 Practices of the informal sector 11.5 14.4 11.6 17.4 6.8 4.6 10.6 Tax administration 3.9 2.4 4.5 4.4 2.7 2.2 4.3 Tax rates 13.2 10.6 18.6 10.9 11.7 6 9.7 Transportation 3.8 5 4.8 2.2 4.3 3.7 2.6 Source: Constructed by the authors using the WBES. Note: EAP stands for East Asia and Pacific, ECA Europe and Central Asia, LAC Latin America and the Caribbean, MENA Middle East and North Africa, SAR South Asia and SSA Sub-Saharan Africa. When the gender dimension is considered, some obstacles turn to be more impeding than others. As such barriers are primarily faced by owners and managers, we do not include in this part female employees. Thus, Figure 5 shows that access to finance, tax rates, practices from informal competitors, political instability, inadequately educated labor force and corruption are the highest obstacles for firms that either owned or managed by women. In addition, electricity turns to be impeding for firms that are owned by women. This is why, we are going to focus only on these barriers in the empirical part, as it will be shown later. 4 Figure 5. Percent of Firms Choosing Their Biggest Obstacle and Gender (%) (a) Female owners (b) Female managers Courts Courts Access to land Business licensing Business licensing Access to land Trade regulations Trade regulations Tax administration Electricity Crime, theft and disorder Crime, theft and disorder Transportation Tax administration Labor regulations Labor regulations Corruption Transportation Electricity Corruption Political instability Inadeq. educ. labor force Inadeq. educ. labor force Political instability Informal comp. Tax rates Informal comp. Access to finance Access to finance Tax rates 0% 5% 10% 15% 20% 0% 5% 10% 15% 20% Source: Constructed by the authors using the WBES. 4 The results of other barriers are available upon request. 9 In addition, at the external level, having a gender provision in an RTA that takes into account gender issues and include an explicit mention of gender, sex, women, girls, the international instruments promoting women’s rights (such as the Convention on the Elimination of All Forms of Discrimination Against Women (CEDAW), the Beijing Declaration and Platform for Action for the Rights of Women and Girls, the Buenos Aires Declaration on Trade and Women’s Economic Empowerment or the UN Sustainable Development Goal 5 on gender equality (SDG 5)) can increase women’s participation, and thus might amplify the positive effect of GVC. Figure 6 shows that the number of trade agreements including such provisions has been increasing significantly since 2000. Yet, it is important to note that, while the inclusion of such provisions is necessary, it might not be sufficient with a weak enforcement. After presenting these different stylized facts, the next section provides empirical evidence on the association between women’s participation and GVCs integration. Figure 6. Evolution of RTAs with Gender Provisions over Time Source: WTO (2022). 3. Methodology Using the WBES, we examine the impact of firms’ integration into GVCs on women’s empowerment as follows: = + + + + + (1) Where Female is measured by three variables: first, whether the top manager of the firm is female (dummy variable equals to one and zero otherwise); second, whether the owner of the firm is female (dummy variable equals to one and zero otherwise); third, the number of full-time female employees (we also distinguish between the proportion of permanent full-time female production and non-production workers). GVC is measured using several dimensions: export status, import status, international certification, and type of ownership, as it has been explained before. The first definition (GVC1) is the most lenient as it encloses firms that are exporters and importers at the 10 same time. The second definition (GVC4) is stricter as it combines the four dimensions together: firms that are simultaneously exporters and importers, that also have an international certification and a foreign ownership of its capital (Dovis and Zaki, 2020). The subscripts i, j, c, g and t denote firms, sector, country, region and year respectively. is a vector that includes firms’ characteristics that are expected to affect female entrepreneurs and women’s participation in the workforce of the firm such as firms’ age, size, share of government ownership and city of operation. is the difference between the survey year and the year in which the establishment began operation. is the share of government ownership that is likely to attract women given that public employment is more women friendly than private one. is a dummy variable that takes the value 1 if the firm is operating in the main business city and is expected to positively affect women’s participation due to agglomeration economies. is a categorical variable that takes the value 1 for small firms, 2 for medium firms and 3 for large ones. Large firms get advantage of their size and engage in economies of scale which, in turn, allows them to enjoy lower costs of production and might hire more women. Appendix 2 summarizes the definition of different variables. Given that we pool data for different countries and years, we include year, country, and sector fixed effects ( , and ) to control for unobservables. is the disturbance term. Our estimations are run using a pooled Ordinary Least Squares (OLS) estimation method (when the dependent variable is a continuous variable, namely share of females’ workers) and a Linear Probability Model (when the dependent variable is binary, which is the case of a female manager or a female owner). We extend the analysis in three ways. First, we investigate whether the impact of firms’ integration into GVCs on female participation is conditional on some external factors such as gender provisions in RTAs. The literature on the efficiency of gender provisions in regional trade agreements (RTAs) is relatively scarce. Hence, there is little empirical evidence that supports the idea that gender-related provisions or labor provisions with clauses related to gender equality significantly promote gender equality and women’s empowerment in the workplace. In that vein, López Mourelo and Samaan (2018) run a difference-in-differences model using the Cambodian Socioeconomic Survey (CSES) conducted by the National Institute of Statistics of the Ministry of Planning over the period 1993-2012 in order to assess the average effect of the 1999 Cambodia- US Bilateral Textile Agreement (CUSBTA) on the gender wage gap. The CUSBTA encloses labor provisions, which mainly aim to improve working conditions through two main pillars. First, the elimination of discrimination between employees and especially those based on gender basis. Second, the reduction of gender-wage gap. The results of the study show that labor provisions decrease gender wage gap and gender discrimination in workplace only in the textile sector. However, the gap is still increasing in the other manufacturing sectors that are not concerned by these provisions. Furthermore, the impact of these provisions is proven to be significant and important only during the period of the agreement and while the International Labor Organization (ILO) is monitoring the proper implementation of these provisions. However, the impact of the provisions starts to decrease during the post-agreement period. In our paper, the impact of gender related provisions in RTAs is studied using the World Trade Organization (WTO) database that compiles provisions related to women’s empowerment and gender equality. In this dataset, 11 provisions are filtered by parties, date of signature, date of entry into force and the type of gender issues they address. The second extension pertains to internal factors measured by the investment climate variables. Indeed, females might face more barriers because of discrimination in access to finance, access to land, etc. This is why we make use of firms’ own perception regarding the main obstacles affecting their operation by identifying each problem as the main obstacle. As it was mentioned before, we focus on the most impeding barriers, namely, access to finance, tax rates, practices from informal competitors, political instability, inadequately educated labor force, corruption, and electricity. These obstacles are expected to have a negative impact on women’s engagement in international trade as well as their integration into GVCs (Christian et al., 2013; Staritz and Reis, 2013; Barrientos, 2014; Doss, 2014; Bamber and Staritz, 2016). Since the time dimension in the WBES is very weak, the inclusion of firm-level fixed effects will not be possible. Therefore, following Dovis and Zaki (2020), firms’ perception about obstacles to doing business is calculated using industry, country, and year averages minus firms’ own responses to control for endogeneity. The third extension checks the robustness of our results in two ways. First, a Propensity Score Matching (PSM) model that solves for endogeneity is estimated. This method assumes the conditional exogeneity of the treatment (GVCs participation in our case) or the selection on the observables only as it consists of finding a proper counterfactual group by matching a firm participating into GVCs with a non-participant firm with similar pre-intervention characteristics. The average treatment effect (ATE) will be estimated as follows: = ( / = 1, ) − ( / = 0, ) (2) where is the outcome measured using four dimensions: if the firm’s owner is female, if the top manager is female, the number of full-time female employees and the proportion of permanent full-time female production and non-production workers. T=1 if the firm is receiving the treatment (participating in GVCs in our case) and 0 if not. The vector represents the observables that are similar between the treatment and control group such as firm’s age, size, share of government ownership and city of operation in addition to the dummies mentioned before. Yet, as PSM is based on observables, the endogeneity that is due to unobservables is not controlled for. This is we rely on an Instrumental Variable approach (IV) to control for the endogenous characteristics of firms’ integration into GVC. The IV must satisfy two main criteria. First, it must be highly correlated with the endogenous variable (GVC in our case). Second, it should not be correlated with the error term and does not affect women’s participation directly. Following Dovis and Zaki (2020), a shift-share of GVC aggregated by country-year-sector-geographical zone (where the firm is located) minus the firm’s own performance is used as an instrument. GVC corrected from individual firm idiosyncrasies is expected to affect firms’ trade performance without having any direct impact on women’s participation. The endogeneity problem is tackled following a Two-Stage Least Squares (2SLS) technique. The first stage predicts GVCs as follows: = Ln() + Ln() + + + ℎ_ℎ_ + + + +∈ (3) 12 Where ℎ _ℎ_ is a shift share of firms’ integration into GVCs calculated using industry, country, year, and geographical zone averages minus firms’ own performance. ∈ is the error term. Different tests are performed to assess the validity and the strength of the instruments. 5 4. Empirical Results Our empirical analysis focuses on women’s participation measured by three variables: whether the owner is a female, whether the manager is a female and the share of females in the total number of workers. For each variable, we will run two different sets of regressions for two definitions of GVCs (GVC1 and GVC4): first, we run the regressions for all the regions (including the MENA region) as it is shown in Table 3a). Second, we run the same regressions but for the MENA region only (see Table 3b). The extensions mentioned above are presented as follows: first, we show how GVC impact on women’s participation is moderated by gender provisions in trade agreements (in Table 4). Second, we examine how GVC impact is moderated by some internal factors that affect the investment climate (for all regions in Table 5); and for the MENA region in Table 6. Finally, we control for the endogeneity of GVCs using a PSM approach (Table 7) and an IV estimation method (Table 8). 4.1. Gender and GVCs 6 Table 3 presents our baseline regression. Regarding our control variables, larger, older firms, located in the main city and having a higher share of government ownership are more likely to be owned by females for all the regions (Table 3a) and the MENA region (Table 3b). Larger firms are generally better performing, listed and might have a diverse board. Li and Chen (2018), using a panel data from listed non-financial firms in China, find that gender diversity in the board has a positive impact on firm performance. Similar results are confirmed by Said et al. (2021) for Egyptian firms. As per the sector of operation, a higher share of government ownership increases the likelihood of female ownership. Generally, the public sector remains a larger employer of women than the private one. In addition, females also enjoy a higher wage premium in the public sector compared to those employed in the private sector. As per our variable of interest, Tables 3a (overall) and 3b (MENA region) shows that the least restrictive definition of GVC is positively associated with a higher probability of having a women as the owner of the firm or with the share of female employees (especially production ones) whereas a deeper integration into GVC (GVC4) reduces the probability of female ownership. In addition, GVC integration exerts a negative impact on the likelihood of a women being the top manager of the firm (for GVC1) and an insignificant impact with the stricter definition (GVC4). Three remarks are worth mentioning. First, there are fundamental differences between female owner and manager given that an owner physically owns the business, while the manager is an employee of the business and works for the owner. In addition, the owner is more affected by 5 The minimum Eigenvalue is higher than all the critical values at 10% and the p-value is significant at 1%. Therefore, we reject the null-hypothesis according to which the instruments are weak. 6 Regressions for Non-Production Workers are presented in Table A3 (Appendix 4). 13 profits and losses, while the manager earns a salary and is not affected by external conditions or fluctuating sales. Management is also operational in the sense that it is concerned with the ongoing activities of the business (Woods and Joyce, 2003). Thus, from an empowerment perspective, being an owner gives more power to a woman, compared to the management position. This is why our results show that GVC might be associated to more empowerment given the positive effect on female ownership, whereas the one for top management is negative. Second, GVC can also improve women’s participation as it increases the share of female employees. While this is rather good news, this result must be cautiously analyzed as we do not measure the quality of jobs associated to these GVC. For instance, on 24 April 2013, the collapse of the Rana Plaza building in Dhaka, Bangladesh, raised several issues regarding the supply chain of garments and the working conditions associated to them (Koenig and Poncet, 2022). Indeed, the collapse of this building that housed five garment factories killed more than a thousand people and injured more than 2,500. This is why, while GVC can improve job creation (Kumar, 2017), job quality is still questionable. Third, most of the emerging economies are abundant in blue collar or production employees. In addition, most of the sectors where they have a comparative advantage (processed food, textile, ready-made garments) are intensive in blue collar workers. Thus, the positive effect of GVC on female production workers can help address the Sustainable Development Goal number five (promoting gender equality) by increasing female labor force participation, especially for production ones. This result corroborates the results of Guha-Khasnobis et al. (2022) who find that, in India, stronger forward linkage has created employment opportunities for the unskilled workers. In the same vein, Kumar (2017) show that lower skilled, young, female workers account for the largest share of jobs that are created in labor-intensive value chains (especially, apparel, footwear, and electronics). When the regressions are run separately for the MENA region (Table 3b), the positive effects on female owners and female employees are confirmed but with higher magnitudes, especially for production workers. This is in line with the findings of Aboushady and Zaki (2021a) who argue that exports and innovation in core production techniques increase the demand for skilled production (blue-collar) workers in the manufacturing sector rather than non-production workers (white collars). However, a major problem in the MENA region is limited employability and skill shortages in blue-collar workers. Additionally, female workers are concentrated in low value- added sectors and in the informal sector. Therefore, inclusive trade policy that promotes GVCs needs to be coupled by public private cooperation to enhance vocational training and improve the skills of blue-collars, especially women (Aboushady and Zaki, 2021b). This will help address the low female labor force participation that has been well-documented in the literature on the MENA region (Assaad and Artz, 2005 and Assaad and Krafft, 2015, Assaad and Boughazala, 2018). 14 Table 3. GVCs and Female Labor Force Participation – Baseline Results (a) All regions included Female Ownership Female Top Manager Female Employees Female Production Workers Variables GVC1 GVC4 GVC1 GVC4 GVC1 GVC4 GVC1 GVC4 Ln(Age) 0.047*** 0.047*** -0.005** -0.005** 0.075*** 0.077*** 0.014 0.019* (0.003) (0.003) (0.002) (0.002) (0.008) (0.008) (0.010) (0.010) Ln(Gov own.) 0.015*** 0.015*** 0.0005 0.0003 0.015 0.017 0.081*** 0.083*** (0.004) (0.004) (0.003) (0.003) (0.014) (0.014) (0.019) (0.019) Main city 0.015*** 0.015*** 0.011*** 0.011*** 0.114*** 0.115*** -0.033** -0.024 (0.004) (0.004) (0.003) (0.003) (0.010) (0.010) (0.015) (0.015) Medium 0.002 0.004 -0.033*** -0.034*** 0.997*** 0.998*** 0.717*** 0.755*** (0.003) (0.003) (0.003) (0.003) (0.009) (0.009) (0.011) (0.011) Large 0.003 0.013*** -0.053*** -0.055*** 2.462*** 2.460*** 2.189*** 2.271*** (0.005) (0.005) (0.004) (0.003) (0.020) (0.020) (0.021) (0.020) GVCs 0.013*** -0.098*** -0.011*** -0.006 0.093*** 0.424*** 0.323*** 0.402*** (0.004) (0.011) (0.003) (0.008) (0.014) (0.059) (0.017) (0.044) Constant 0.163*** 0.165*** 0.194*** 0.193*** 0.839*** 0.841*** 0.552*** 0.579*** (0.009) (0.009) (0.007) (0.007) (0.024) (0.024) (0.032) (0.032) Country dum. Yes Yes Yes Yes Yes Yes Yes Yes Year dum. Yes Yes Yes Yes Yes Yes Yes Yes Sector dum. Yes Yes Yes Yes Yes Yes Yes Yes Observations 83,949 83,949 84,341 84,341 41,224 41,224 41,234 41,234 R-squared 0.115 0.116 0.099 0.099 0.548 0.548 0.498 0.495 (b) MENA region Female Ownership Female Top Manager Female Employees Female Production Workers Variables GVC1 GVC4 GVC1 GVC4 GVC1 GVC4 GVC1 GVC4 Ln(Age) 0.037*** 0.036*** 3.85e-05 -3.26e-07 0.067*** 0.072*** -0.069*** -0.064*** (0.006) (0.006) (0.004) (0.004) (0.024) (0.024) (0.023) (0.023) Ln(Gov own.) -0.017 -0.016 -0.005 -0.005 0.074 0.082 -0.010 -0.013 (0.011) (0.011) (0.005) (0.005) (0.056) (0.055) (0.065) (0.065) Main city 0.023*** 0.024*** 0.019*** 0.018*** 0.200*** 0.209*** -0.035 -0.030 (0.008) (0.008) (0.005) (0.005) (0.033) (0.033) (0.033) (0.034) Medium 0.023*** 0.027*** 0.0005 -0.0001 0.848*** 0.857*** 0.448*** 0.497*** (0.080) (0.008) (0.005) (0.005) (0.030) (0.030) (0.025) (0.025) Large 0.047*** 0.061*** -0.009 -0.011* 2.311*** 2.320*** 1.449*** 1.578*** (0.011) (0.011) (0.006) (0.006) (0.067) (0.066) (0.051) (0.051) GVCs 0.031*** -0.070** -0.007 -0.003 0.228*** 0.753*** 0.417*** 0.645*** (0.011) (0.029) (0.006) (0.017) (0.043) (0.236) (0.045) (0.153) Constant 0.047** 0.051*** 0.050*** 0.049*** 0.445*** 0.447*** 0.602*** 0.619*** (0.019) (0.019) (0.012) (0.012) (0.073) (0.073) (0.073) (0.073) Country dum. Yes Yes Yes Yes Yes Yes Yes Yes Year dum. Yes Yes Yes Yes Yes Yes Yes Yes Sector dum. Yes Yes Yes Yes Yes Yes Yes Yes Observations 11,714 11,714 11,730 11,730 4,857 4,857 6,863 6,863 R-squared 0.109 0.109 0.019 0.019 0.503 0.503 0.471 0.465 Robust standard errors in parentheses, *** p<0.01, ** p<0.05, * p<0.1 Considering the sectoral activity of firms, the regressions for all the regions (Table A4a in the appendix) and for the MENA region (Table A4b) show that firms operating in the textile and garments sector are more likely to have a female owner and manager as well as a higher number of female production workers. Moreover, female employees are hired intensively in firms operating in leather sector. In the MENA, textile and leather seem to be the most female intensive, while the other sectors have a negative bias against women. These results are in line with the findings of Frederick et al. (2022) who argue that female employment is highest in apparel manufacturing sector. They find that textiles and leather sectors are the most important employer of women across developing countries as the percent of all female employment working in textiles 15 and leather ranges from 2% in the Arab Republic of Egypt to 16% in Cambodia. As for the interaction with GVCs, textile and leather sectors integrating into GVCs hire more women and are likely to have a female owner or manager. This can be due to the fact that the increase of industry competitiveness (especially that several emerging countries have a comparative advantage in these sectors) due to trade participation reduces the incentives of these firms to discriminate against women (Becker, 1959). 4.2. Moderating Factors 7 The previous analysis is extended by examining how the effect of GVC on women’s participation can be moderated through internal and external factors. 4.2.1. Gender Provisions in Trade Agreements As it was mentioned before, while the global economy witnessed a proliferation of gender provisions in regional trade agreements, only 20% of them include explicit chapters or clauses that endeavor to achieve gender equality and to empower women (Monteiro, 2021). This might make GVC integration more women friendly and thus might increase women’s participation in international trade through several channels. First, gender inequalities and discrimination against women are the most addressed issues in RTAs as they guarantee equitable access for men and women to opportunities generated by the RTA. Second, some provisions address the participation of women in economic activities, while fewer provisions promote women’s access to productive resources, such as credit and financial services, land, and technology as they might affect their participation in international trade. Finally, a handful of provisions address issues related to women’s leadership and decision-making roles (WTO, 2022). Table 4a shows the results of GVC, gender provisions, and their interaction on women’s participation for all countries. The positive effect of GVC on female owners and female employees is still confirmed in most of the regressions, with an insignificant impact on GVC4 (mainly due to a limited number of firms who are deeply integrated into GVCs). Moreover, gender provision per se exert a positive impact on female ownership and employees, and a negative impact on managers and production workers. As per the MENA region, Table 4b shows that gender provisions’ impact remains positive for female employees. In contrast, it becomes non-significant for the other variables measuring female empowerment. When GVCs are interacted with gender provisions, we also find an insignificant effect in the MENA region. This might be due to the de jure inclusion of gender provisions in trade agreements without real enforcement. Indeed, the number of provisions or length of chapters including gender considerations matter much less than their enforcement. In addition, the vast majority of gender provisions are non-binding in nature. This is why it is important to distinguish whether such provisions are enforced or not or are subject to a dispute settlement mechanism or not. In a nutshell, the positive effect of gender provision on female owners and employees is thus promising given that trade policy-related factors could be mobilized to make GVC more women friendly. Yet, more efforts are needed to make such provisions better enforceable and monitored. 7 Regressions for Non-Productions Workers are presented in Table A5 (Appendix 4) and Table A17 (Appendix 5). 16 Table 4. GVC, Female Labor Force Participation, and Gender Provisions in RTAs (a) All Regions Female Ownership Female Top Manager Female Employees Female Production Workers Variables GVC1 GVC4 GVC1 GVC4 GVC1 GVC4 GVC1 GVC4 (1)GVC 0.038*** -0.046 0.001 -0.001 0.131*** 0.463*** 0.378*** 0.525*** (0.009) (0.032) (0.006) (0.015) (0.024) (0.101) (0.056) (0.068) (2)Gender Provisions 0.050*** 0.050*** -0.048*** -0.048*** 0.066*** 0.065*** -0.246*** -0.232*** (0.007) (0.007) (0.006) (0.006) (0.014) (0.013) (0.028) (0.029) GVC*Gender Prov. -0.001*** -0.001*** -0.0004*** -0.0001 -0.001** -0.001 -0.002** -0.003** (0.0002) (0.001) (0.0001) (0.0003) (0.001) (0.002) (0.001) (0.001) Country dummies Yes Yes Yes Yes Yes Yes Yes Yes Year dummies Yes Yes Yes Yes Yes Yes Yes Yes Sector dummies Yes Yes Yes Yes Yes Yes Yes Yes Observations 83,949 83,949 84,341 84,341 41,227 41,227 41,237 41,237 R-squared 0.116 0.116 0.100 0.099 0.548 0.548 0.499 0.495 (b) MENA region Female Ownership Female Top Manager Female Employees Female Production Workers Variables GVC1 GVC4 GVC1 GVC4 GVC1 GVC4 GVC1 GVC4 (1)GVC 0.0770** -0.0453 0.00582 -0.00854 0.299*** 0.427 0.556*** 1.061** (0.0312) (0.0684) (0.0137) (0.0437) (0.0736) (0.361) (0.101) (0.445) (2)Gender Provisions 0.00585 0.00377 0.00417 0.00349 0.114*** 0.111*** 0.0116 -0.00751 (0.00891) (0.00923) (0.00311) (0.00246) (0.0402) (0.0393) (0.0374) (0.0458) GVC*Gender Prov. -0.00393 -0.000245 -0.00295 0.00352 0.00269 0.155 -0.0217 -0.0926 (0.00881) (0.0158) (0.00394) (0.0113) (0.0244) (0.128) (0.0371) (0.0938) Year dummies Yes Yes Yes Yes Yes Yes Yes Yes Sector dummies Yes Yes Yes Yes Yes Yes Yes Yes Observations 11,714 11,714 11,730 11,730 4,861 4,861 6,864 6,864 R-squared 0.064 0.060 0.011 0.010 0.456 0.454 0.472 0.465 Robust standard errors clustered by country and year in parentheses, *** p<0.01, ** p<0.05, * p<0.1 Notes: - We control for firms’ age, size, city of operation and the share of government ownership. - Country dummies have been removed from panel (b) given the high collinearity between countries and provisions. Moreover, our sample drops when we focus only on the MENA region. - The intercept is included. 4.2.2. Internal Factors In addition to external factors, there is growing evidence that an adverse business environment impedes firms’ performance and hence negatively affect women’s participation. Thus, the impact of GVC on women can also be moderated through the characteristics of the investment climate. This can be explained by two main reasons. Investment climate affects trade performance and GVC integration. Dovis and Zaki (2020) show that the number of days that are required to pay taxes, the number of procedures that are necessary to register property, and the time to export and to import have a significantly negative relationship with the likelihood of a firm’s integration into a GVC. In the same vein, Aboushady and Zaki (2019), using the WBES for Egypt, show that access to finance, tax payments and competition from the informal sector affect the firms’ decision to become exporters, which is a part of GVCs. Second, several constraints hinder women’s participation. Indeed, ILO (2021) argues that women’s access to finance and markets, their land and property rights, and business registration and informality are key issues to be addressed to increase women’s participation in the labor market. In addition, the World Bank’s Women, Business and the Law index providing the link between legal gender equality and women’s economic inclusion shows that the MENA and South Asia regions have the lowest index score (World Bank 2020b). The workplace indicator shows that, in many countries, the law does not prohibit gender-based employment discrimination that 17 covers mainly four areas: the existence of limited laws that stipulate equal remuneration, laws that hinder women from working similar night hours as men, laws that limit female participation in at least one industry or at certain jobs deemed as dangerous specially. Moreover, it shows that many countries such as Egypt, Turkey, Pakistan, and Bangladesh do not allow woman to get a job in the same place as a man. Thus, it is worth investigating how these barriers might reduce the positive impact of GVC on women’s participation. Table 5 shows the impact of GVC, different obstacles and their interaction at the world level. Globally, institutional barriers limit female participation in the workforce. First, female owners are negatively affected by access to finance and by the inadequate educate labor force. Access to finance is of particular importance as World Bank (2021) shows that, although 115 out of 190 mapped economies do not prohibit discrimination in access to credit based on gender (needing husband’s approval or signature for financial transactions), female entrepreneurs can still face several discriminatory practices from banks and credit facilities (because of lack of collaterals due to the lack of resources for instance). As per the inadequate labor force, it also exerts a negative impact on female owners, pointing out the skills mismatch that characterizes several developing countries. Finally, it is worthy to note that in most of the cases the interaction of GVC with the obstacles is positive and statistically significant, which shows that GVC firms might face less obstacles (as, on average, they are more productive). Thus, the net effect of GVC on female ownership remains positive and statistically significant. As per female managers, and in addition to inadequate labor force, political instability and corruption turn to be the most impending barriers, while the rest of the barriers are statistically insignificant. Such a finding has been documented in the literature, as women are perceived as more vulnerable and less likely to know and claim their rights. Thus, this will make them less confident in seeking legal redress and thus subject to abusive corruption. This is why, to empower women at the leadership level, more transparent and enforced rules are needed to have good governance. Access to finance and inadequate labor force are also exerting a negative impact on female employees. Finally, it is important to note that the GVC positive impact on female employees (and for production employees) is attenuated by most of the barriers as the interaction coefficient is negative with access to finance, electricity, and inadequate labor force (for GVC4 at the world level). Such a finding is of particular importance given that, as it was mentioned before, female production workers are abundant and hired intensively in several sectors where the emerging economies and have a comparative advantage. This is why, to increase the impact of GVC on this vulnerable category, it is crucial to improve the business environment. Thus, globally, improving skills to let them better match the labor market demand is a key issue to improve female labor force participation, management, and ownership. Moving to the MENA region (Table 6), the picture looks slightly different as tax rates have a negative impact on female owners and female employees. There is strong evidence that tax policy has a gender bias (Stotsky, 1996 and AWID, 2013). Indeed, even if tax systems do not include explicit gender biases, there are several implicit biases in dealing with tax collectors, tax procedures, and fair implementation of rules. One of the important barriers that affects female managers and production employees in the competition coming from the informal sector. The latter is still a major problem in the region. Indeed, cheaper products offered by the informal sector may harm the performance of formally registered firms, and thus affect women’s employment, namely 18 managers and production employees. Most of the interaction between GVC and such barriers are not statistically significant. Table 5. Female Labor Force Participation, GVCs and Barriers – All regions Finance Tax.Rates Pol. Inst. Corrup. Comp. Elec. Inad.Edu (a) Female Ownership (1) GVC -0.0154 0.0174 -0.0126 -0.0144 0.000437 0.00378 0.0281** (0.0121) (0.0143) (0.0123) (0.0110) (0.0128) (0.0145) (0.0124) (2) Obstacle -0.0788* 0.0114 -0.0160 -0.0617 -0.0276 0.0274 -0.0724* (0.0457) (0.0423) (0.0433) (0.0523) (0.0406) (0.0385) (0.0395) GVC1 (1)*(2) 0.160*** -0.0178 0.104*** 0.125*** 0.0681 0.0368 -0.0817 (0.0543) (0.0466) (0.0394) (0.0408) (0.0543) (0.0494) (0.0514) Observations 83,896 83,896 83,896 83,896 83,896 83,896 83,896 R-squared 0.115 0.115 0.115 0.115 0.115 0.115 0.115 (1) GVC -0.145*** -0.0469 -0.111*** -0.117*** -0.136*** -0.0665 -0.0303 (0.0311) (0.0410) (0.0330) (0.0330) (0.0302) (0.0416) (0.0426) (2) Obstacle -0.0529 0.00979 0.00591 -0.0399 -0.0202 0.0342 -0.081** (0.0449) (0.0408) (0.0420) (0.0519) (0.0393) (0.0377) (0.0382) GVC4 (1)*(2) 0.317** -0.228* 0.0666 0.112 0.240 -0.130 -0.364** (0.139) (0.124) (0.0963) (0.108) (0.147) (0.132) (0.157) Observations 83,896 83,896 83,896 83,896 83,896 83,896 83,896 R-squared 0.116 0.116 0.116 0.116 0.116 0.116 0.116 (b) Female Manager (1) GVC -0.0192** -0.00447 -0.0175* -0.023*** -0.0123 -0.0170 -0.0122 (0.00861) (0.00944) (0.00899) (0.00843) (0.00833) (0.0111) (0.0102) (2) Obstacle -0.0219 -0.0235 -0.0773** -0.0896** 0.00312 -0.0228 -0.0704* (0.0393) (0.0348) (0.0373) (0.0386) (0.0427) (0.0311) (0.0394) GVC1 (1)*(2) 0.0490 -0.0249 0.0282 0.0559** 0.00969 0.0253 0.00989 (0.0339) (0.0285) (0.0223) (0.0275) (0.0346) (0.0337) (0.0431) Observations 84,287 84,287 84,287 84,287 84,287 84,287 84,287 R-squared 0.099 0.099 0.099 0.100 0.099 0.099 0.099 (1) GVC 0.00141 0.0161 0.00753 0.0106 0.00932 0.0186 -0.00947 (0.0152) (0.0182) (0.0154) (0.0158) (0.0154) (0.0223) (0.0199) (2) Obstacle -0.00961 -0.0248 -0.0705* -0.0767** 0.00867 -0.0138 -0.0699* (0.0395) (0.0346) (0.0374) (0.0389) (0.0411) (0.0306) (0.0401) GVC4 (1)*(2) -0.0454 -0.0971* -0.0629 -0.0924* -0.0918 -0.102 0.0219 (0.0679) (0.0554) (0.0403) (0.0509) (0.0702) (0.0677) (0.0776) Observations 84,287 84,287 84,287 84,287 84,287 84,287 84,287 R-squared 0.099 0.099 0.099 0.099 0.099 0.099 0.099 (c) Female Employees (1) GVC 0.0913** 0.131*** 0.0467 0.0448 0.0831** 0.0694* 0.0570 (0.0376) (0.0370) (0.0352) (0.0348) (0.0378) (0.0373) (0.0386) GVC1 (2) Obstacle -0.568** -0.240 -0.0555 -0.241 -0.405 0.558*** -0.826*** (0.243) (0.228) (0.205) (0.233) (0.245) (0.213) (0.271) 19 (1)*(2) 0.0149 -0.158 0.196* 0.225* 0.0579 0.108 0.201 (0.160) (0.136) (0.114) (0.116) (0.156) (0.130) (0.172) Observations 41,200 41,200 41,200 41,200 41,200 41,200 41,200 R-squared 0.548 0.548 0.548 0.548 0.548 0.548 0.548 (1) GVC 0.589*** 0.421*** 0.459*** 0.433*** 0.424*** 0.618*** 0.677*** (0.102) (0.116) (0.111) (0.0964) (0.109) (0.115) (0.144) (2) Obstacle -0.550** -0.267 -0.0307 -0.212 -0.392 0.585*** -0.784*** (0.244) (0.224) (0.207) (0.234) (0.248) (0.210) (0.277) GVC4 (1)*(2) -0.972** 0.0159 -0.168 -0.0473 0.00781 -0.895** -1.373** (0.437) (0.393) (0.497) (0.453) (0.553) (0.397) (0.582) Observations 41,200 41,200 41,200 41,200 41,200 41,200 41,200 R-squared 0.548 0.548 0.548 0.548 0.548 0.549 0.549 (d) Female Production Workers (1) GVC 0.339*** 0.529*** 0.298*** 0.366*** 0.379*** 0.310*** 0.377*** (0.0530) (0.113) (0.0730) (0.0523) (0.0923) (0.0748) (0.0673) (2) Obstacle 0.257 0.537*** -0.295 0.412* -0.117 0.151 0.602*** (0.206) (0.192) (0.220) (0.236) (0.220) (0.197) (0.230) GVC1 (1)*(2) -0.0835 -0.784** 0.0974 -0.186 -0.304 0.0499 -0.287 (0.220) (0.336) (0.339) (0.261) (0.352) (0.318) (0.234) Observations 41,205 41,205 41,205 41,205 41,205 41,205 41,205 R-squared 0.498 0.499 0.498 0.499 0.498 0.498 0.499 (1) GVC 0.411*** 0.390*** 0.453*** 0.437*** 0.371*** 0.453*** 0.592*** (0.0921) (0.102) (0.0850) (0.0858) (0.0914) (0.104) (0.0969) (2) Obstacle 0.211 0.274 -0.243 0.381 -0.239 0.160 0.585** (0.206) (0.220) (0.232) (0.258) (0.249) (0.174) (0.227) GVC4 (1)*(2) -0.0681 0.0491 -0.245 -0.213 0.189 -0.210 -1.005** (0.451) (0.413) (0.299) (0.329) (0.395) (0.391) (0.451) Observations 41,205 41,205 41,205 41,205 41,205 41,205 41,205 R-squared 0.495 0.495 0.495 0.495 0.495 0.495 0.495 Robust standard errors clustered by country and year in parentheses, *** p<0.01, ** p<0.05, * p<0.1 Notes: i) Each regression controls for firms’ age, city of operation and share of government ownership ii) All the regressions include country, year, sector, and size fixed effects. iii) Country-year-sector averages are used to reduce the risk of endogeneity between the business environment and firm-level. iv) Each column for each GVC definition represents a separate regression. v) All the variables are in log. vi) The intercept is included. 20 Table 6. Female Labor Force Participation, GVCs and Barriers – MENA region Finance Tax.Rates Pol. Inst. Corrup. Comp. Elec. Inad.Edu. (a) Female Ownership (1) GVC -0.0312 -0.073*** -0.00214 -0.00956 0.0297 -0.0194 0.0285 (0.0264) (0.0230) (0.0388) (0.0373) (0.0357) (0.0261) (0.0185) (2) Obstacle -0.126 -0.234** 0.0158 -0.0172 -0.0365 -0.151 -0.00455 (0.0819) (0.0903) (0.103) (0.0788) (0.0718) (0.0892) (0.0990) GVC1 (1)*(2) 0.251*** 0.416*** 0.0777 0.110 0.00616 0.185** 0.0164 (0.0845) (0.0842) (0.0706) (0.0836) (0.140) (0.0693) (0.109) Observations 11,712 11,712 11,712 11,712 11,712 11,712 11,712 R-squared 0.110 0.111 0.109 0.109 0.109 0.110 0.109 (1) GVC -0.172*** -0.142** -0.0885 -0.0992 -0.120 -0.118** -0.0993* (0.0541) (0.0648) (0.0963) (0.0815) (0.0764) (0.0573) (0.0516) (2) Obstacle -0.0763 -0.160* 0.0337 0.00850 -0.0363 -0.118 -0.00169 (0.0867) (0.0912) (0.102) (0.0829) (0.0741) (0.0895) (0.0874) GVC4 (1)*(2) 0.451** 0.281 0.0449 0.0793 0.214 0.184 0.148 (0.176) (0.253) (0.193) (0.196) (0.321) (0.195) (0.218) Observations 11,712 11,712 11,712 11,712 11,712 11,712 11,712 R-squared 0.109 0.109 0.109 0.109 0.109 0.109 0.109 (b) Female Top Manager (1) GVC -0.0328 -0.0324* -0.0119 -0.0218 -0.0324* -0.0223 -0.034*** (0.0199) (0.0186) (0.0210) (0.0195) (0.0175) (0.0144) (0.0105) (2) Obstacle -0.0487 -0.0802 -0.0421 0.0470 -0.142*** -0.0407 -0.0202 (0.0491) (0.0625) (0.0704) (0.0497) (0.0324) (0.0422) (0.0781) GVC1 (1)*(2) 0.105 0.103* 0.0125 0.0409 0.112 0.0576 0.164*** (0.0660) (0.0582) (0.0403) (0.0426) (0.0660) (0.0437) (0.0576) Observations 11,728 11,728 11,728 11,728 11,728 11,728 11,728 R-squared 0.019 0.019 0.019 0.019 0.020 0.019 0.020 (1) GVC -0.0681 -0.0835* -0.0223 -0.0689 -0.0931 -0.0142 0.00623 (0.0400) (0.0481) (0.0566) (0.0537) (0.0631) (0.0231) (0.0350) (2) Obstacle -0.0309 -0.0639 -0.0402 0.0522 -0.127*** -0.0298 0.0222 (0.0493) (0.0594) (0.0695) (0.0482) (0.0319) (0.0396) (0.0706) GVC4 (1)*(2) 0.285* 0.309 0.0451 0.176 0.379 0.0411 -0.0483 (0.165) (0.199) (0.112) (0.139) (0.271) (0.106) (0.179) Observations 11,728 11,728 11,728 11,728 11,728 11,728 11,728 R-squared 0.019 0.019 0.019 0.019 0.020 0.019 0.019 (c) Female Employees (1) GVC 0.217* 0.0746 0.151 0.182** 0.174** 0.197** 0.177* (0.110) (0.111) (0.112) (0.0866) (0.0816) (0.0789) (0.0876) (2) Obstacle 1.055*** -0.766** 0.701* 0.187 0.385 0.0464 -0.612 GVC1 (0.313) (0.362) (0.356) (0.430) (0.357) (0.463) (0.514) (1)*(2) 0.0457 0.654 0.184 0.127 0.250 0.129 0.335 (0.360) (0.387) (0.219) (0.193) (0.281) (0.230) (0.383) Observations 4,857 4,857 4,857 4,857 4,857 4,857 4,857 21 R-squared 0.504 0.504 0.504 0.503 0.504 0.503 0.504 (1) GVC 0.584 0.194 0.419 0.598 0.348 1.014** 1.655** (0.453) (0.714) (0.683) (0.487) (0.479) (0.374) (0.777) (2) Obstacle 1.066*** -0.666* 0.765** 0.288 0.438 0.0985 -0.484 (0.321) (0.359) (0.355) (0.442) (0.327) (0.448) (0.498) GVC4 (1)*(2) 0.903 2.308 0.994 0.486 1.839 -1.199 -4.934 (2.439) (2.558) (1.885) (1.695) (2.251) (1.545) (2.999) Observations 4,857 4,857 4,857 4,857 4,857 4,857 4,857 R-squared 0.503 0.503 0.503 0.503 0.503 0.503 0.503 (d) Female Production Workers (1) GVC 0.241 0.172 0.369 0.266 0.307* 0.160 0.294*** (0.186) (0.170) (0.231) (0.207) (0.152) (0.150) (0.0929) (2) Obstacle -0.218 -0.784 -0.698 -0.176 -1.047** -0.663** -0.114 (0.509) (0.467) (0.496) (0.535) (0.506) (0.298) (0.337) GVC1 (1)*(2) 0.694 0.965 0.114 0.405 0.441 0.888** 0.725 (0.592) (0.638) (0.434) (0.468) (0.521) (0.417) (0.495) Observations 6,861 6,861 6,861 6,861 6,861 6,861 6,861 R-squared 0.472 0.472 0.472 0.472 0.472 0.473 0.472 (1) GVC 0.698** 0.238 0.652 0.688* 0.840** 0.386 0.881*** (0.295) (0.368) (0.480) (0.402) (0.384) (0.317) (0.315) (2) Obstacle 0.00541 -0.719 -0.672 -0.0633 -1.026* -0.537* 0.144 (0.511) (0.460) (0.541) (0.543) (0.516) (0.290) (0.394) GVC4 (1)*(2) -0.222 1.560 -0.0109 -0.112 -0.806 0.956 -1.156 (1.247) (1.344) (1.088) (0.986) (1.174) (1.155) (1.146) Observations 6,861 6,861 6,861 6,861 6,861 6,861 6,861 R-squared 0.465 0.466 0.465 0.465 0.466 0.466 0.465 Robust standard errors clustered by country and year in parentheses, *** p<0.01, ** p<0.05, * p<0.1 Notes: i) Each regression controls for firms’ age, city of operation and share of government ownership ii) All the regressions include country, year, sector and size fixed effects. iii) Country-year-sector averages are used to reduce the risk of endogeneity between the business environment and firm-level. iv) Each column for each GVC definition represents a separate regression. v) All the variables are in log. vi) The intercept is included. 22 5. Robustness Checks 8 As it was mentioned before, we use a PSM 9 (see Table 7) where the treatment is being part of a GVC. Clearly, this method assumes that the selection to be treated (being part of GVC) is based on observables only. The average treatment effect shows that the positive effect on female ownership (for GVC1), female employees and female production workers and the negative one on female top manager (for GVC1) are similar to those of the baseline regression (in Table 3). Matching statistics and the results are presented in Appendix 4. Tables A9-A15 shows that there is a high level of common support for the two definitions of GVC and for the total sample and the MENA region. This is also confirmed by Figure A2 and A3. Thus, our PSM results converge to those of the baseline. Table 7. GVCs and Female Labor Force Participation – PSM (a) All regions Fem. Ownership Fem. Top Manager Fem. Employees Fem. Prod. Workers GVC1 GVC4 GVC1 GVC4 GVC1 GVC4 GVC1 GVC4 ATE 0.028*** -0.026 -0.007 0.021 0.088*** 0.539*** 0.274*** 0.241** (0.0066) (0.0435) (0.005) (0.030) (0.027) (0.108) (0.027) (0.126) ATT 0.011** -0.096*** -0.012*** -0.007 0.107*** 0.435*** 0.337*** 0.333*** (0.006) (0.012) (0.004) (0.009) (0.023) (0.094) (0.031) (0.062) Observations 83,949 83,949 84,341 84,341 41,227 41,227 41,237 41,237 (b) MENA region Fem. Ownership Fem. Top Manager Fem. Employees Fem. Prod. Workers GVC1 GVC4 GVC1 GVC4 GVC1 GVC4 GVC1 GVC4 ATE 0.012 -0.091** -0.021*** -0.029** 0.235*** 0.236 0.250*** 0.110 (0.012) (0.041) (0.005) (0.013) (0.079) (0.351) (0.062) (0.302) ATT 0.036*** -0.056* -0.005 -0.003 0.260*** 0.650* 0.490*** 0.435** (0.014) (0.033) (0.006) (0.019) (0.069) (0.350) (0.106) (0.222) Observations 11,714 11,714 11,730 11,730 4,861 4,861 6,864 6,864 Standard errors in parentheses *** p<0.01, ** p<0.05, * p<0.1 Note: All the regressions include country, year and sector dummies. Moving to the second robustness check, an Instrumental Variable approach (IV) is used in order to control for the endogenous characteristics of firms’ integration into GVC. A shift-share of GVC aggregated by country-year-sector-geographical zone minus the firm’s own integration into GVC is used as an instrument. GVC corrected from individual firm idiosyncrasies is expected to affect firms’ trade performance without having any direct impact on women’s participation. Table 8 shows that our results regarding the positive effect on female ownership and female production workers are robust and become stronger. Hence, our previous estimates of GVC must be interpreted as lower bounds due to the downward bias resulting from the endogeneity problem. Moreover, the effect of GVC on top management is still insignificant. When the same IV approach is applied to the MENA region, we find a positive effect of GVC on female owners and female workers (especially production ones), with an insignificant impact on female managers. This 8 Results of IV-First Stage are presented in Table A6 in Appendix 4. Results of IV regressions for Non-Production Workers are presented in Table A7. 9 To check the robustness of the results, propensity scores are estimated using alternative matching methods with different choice of bandwidths (Kernel pair matching with replacement and cross validation with respect to the means of x). Matching statistics and the results are presented in Appendix 5. 23 provides evidence for the causal link between GVC and female empowerment (measured by ownership and employment). 10 Table 8. GVCs and Female Labor Force Participation – IV Approach (a) All regions Fem. Ownership Fem. Top Manager Fem. Employees Fem. Production Workers Variables GVC1 GVC4 GVC1 GVC4 GVC1 GVC4 GVC1 GVC4 Ln(Age) 0.047*** 0.047*** -0.005** -0.005** 0.076*** 0.080*** 0.007 0.023** (0.003) (0.003) (0.002) (0.002) (0.008) (0.008) (0.011) (0.012) Ln(Gov own.) 0.015*** 0.016*** 0.001 0.001 0.014 0.010 0.082*** 0.085*** (0.004) (0.004) (0.003) (0.003) (0.012) (0.012) (0.016) (0.017) Main city 0.015*** 0.015*** 0.011*** 0.011*** 0.112*** 0.108*** -0.044*** -0.028* (0.004) (0.004) (0.003) (0.003) (0.010) (0.011) (0.015) (0.016) Medium 0.00136 0.00482 -0.0347*** -0.0336*** 0.995*** 0.978*** 0.654*** 0.703*** (0.004) (0.004) (0.003) (0.003) (0.010) (0.012) (0.020) (0.022) Large 0.009 0.017 -0.056*** -0.050*** 2.459*** 2.381*** 2.024*** 2.013*** (0.007) (0.011) (0.006) (0.009) (0.017) (0.029) (0.042) (0.085) GVC 0.019 -0.143 0.001 -0.072 0.130 3.297*** 0.767*** 2.965*** (0.020) (0.136) (0.016) (0.109) (0.0816) (0.895) (0.102) (0.820) Observations 82,937 82,937 83,307 83,307 40,812 40,812 40,628 40,628 (b) MENA region Female Ownership Female Top Manager Female Employees Fem. Production Workers Variables GVC1 GVC4 GVC1 GVC4 GVC1 GVC4 GVC1 GVC4 Ln(Age) 0.036*** 0.039*** -0.002 0.003 0.059** 0.081*** -0.066*** -0.018 (0.006) (0.007) (0.004) (0.004) (0.025) (0.027) (0.025) (0.034) Ln(Gov own.) -0.018 -0.017 -0.006 -0.008 0.053 0.092** -0.015 -0.080 (0.011) (0.011) (0.007) (0.007) (0.042) (0.043) (0.044) (0.059) Main city 0.020** 0.023*** 0.016*** 0.016*** 0.163*** 0.202*** -0.045 -0.041 (0.008) (0.008) (0.005) (0.005) (0.036) (0.033) (0.034) (0.040) Medium 0.011 0.0238*** -0.00554 -0.00407 0.796*** 0.835*** 0.330*** 0.452*** (0.010) (0.009) (0.006) (0.005) (0.038) (0.037) (0.045) (0.042) Large 0.00530 0.0358 -0.0283** -0.0413** 2.209*** 2.245*** 1.062*** 1.187*** (0.022) (0.029) (0.013) (0.018) (0.066) (0.094) (0.105) (0.150) GVC 0.156*** 0.349 0.0524 0.523* 1.051*** 3.559 1.320*** 6.101*** (0.059) (0.476) (0.036) (0.296) (0.354) (2.831) (0.232) (2.029) Observations 11,610 11,610 11,626 11,626 4,820 4,820 6,797 6,797 Standard errors in parentheses, *** p<0.01, ** p<0.05, * p<0.1 Note: - All the regressions include country, year, and sector fixed effects. - The intercept is included. 6. Conclusion and Policy Implications This paper aims at examining the impact of GVCs on women’s trade participation as entrepreneurs and employees. We also analyze how this effect is moderated through external (gender provisions in trade agreements) and internal (investment climate variables) factors. To do so, we use firm- level data for 154 developing economies and emerging markets with a special focus on the Middle East and North Africa region. Our main findings show that GVC integration increases the likelihood of being a female owner and the share of female employees, especially production ones. A less robust negative effect is found regarding the impact on being a female top manager. These effects are moderated by the inclusion of gender provisions in trade agreements and by some characteristics of the investment climate, namely corruption, access to finance and tax policy. These results remain robust after we control for the endogeneity of GVC using an instrumental 10 Results of the first stage are in Table A16 in Appendix 4. 24 variable approach and a propensity score estimation method where the treatment is being part of a GVC. From a policy perspective, this topic is of particular importance as it addresses two important, and correlated, challenges in the MENA region that are low female labor force participation and a weak integration into GVCs. Indeed, if MENA countries are to improve firms’ insertion into GVCs, female labor participation can increase given that there are several sectors that are female intensive and that have a comparative advantage in the MENA region such as the textile, ready-made garments, processed food, and electronics sectors. Hence, from the Sustainable Development Goals (SDG) perspective, our paper is related to two goals, namely promoting inclusive and sustainable industrialization, and fostering innovation (SDG9) and promoting gender equality (SDG5). Yet, to move forward, three recommendations are worth taking into consideration. First, at the conception and the implementation of trade and industrial policies, it is important to mainstream gender as both trade agreements and GVCs can be used as tools to improve women’s participation. Yet, given that the majority of gender provisions are non-binding, it is important to have enforcement mechanisms that guarantee the implementation of gender provisions in trade agreements. Second, as the positive impact of GVCs on female owners or female production workers is attenuated by some obstacles (namely corruption, access to finance, and tax policies), it is important to address such barriers to maximize the impact of GVCs. This will require deeper and more structural reforms to improve the business environment. Finally, one of the important obstacles that hinder women’s empowerment is inadequate labor force participation. This is why investing in education (especially technical and vocational education for production workers) would provide women with the required skills that are relevant to firms that operate in GVCs. References [1]. Aboushady, N., & Zaki, C. (2022). Assessing EU–Middle East trade relations: Patterns, policies and imbalances. In Routledge Handbook of EU–Middle East Relations (pp. 432-446). Routledge. [2]. Aboushady, N. and Zaki, C. (2021a) “Do Exports and Innovation Matter for the Demand of Skilled Labor?”, International Review of Applied Economics, vol. 35 no. 1, pages 25-44. [3]. Aboushady, N. and Zaki, C. (2021b) “Trade and SDGs in the MENA Region: Towards a More Inclusive Trade Policy”, ERF Policy Brief No. 56. [4]. Aboushady, N. and Zaki, C. (2019) “Investment climate and Trade Margins in Egypt: Which Factors Do Matter?”, Economics Bulletin, vol. 39(4), pages 2275-2301. [5]. Assaad, R., & Arntz, M. (2005). Constrained geographical mobility and gendered labor market outcomes under structural adjustment: Evidence from Egypt. World Development, 33(3), 431- 454. [6]. Assaad, R., & Boughzala, M. (Eds.). (2018). The Tunisian labor market in an era of transition. Oxford University Press, USA. [7]. Assaad, R., & Krafft, C. (Eds.). (2015). The Egyptian labor market in an era of revolution. OUP Oxford. [8]. AWID (2013): Tax Justice and Women: Examining gender inequalities in the tax system. March 2013 25 [9]. Bahri, A. (2021). Making Trade Agreements Work for Women Empowerment. Latin American Journal of Trade Policy, 4(11), 6-24. [10]. Barrientos, S. (2014). Gender and global value chains: challenges of economic and social upgrading in agri-food. Robert Schuman Centre for Advanced Studies Research Paper No. RSCAS, 96. [11]. Bamber, P., & Staritz, C. (2016). The gender dimensions of global value chains. Geneva: International Center for Trade and Sustainable Development. [12]. Becker, G. S. (1959). The economics of discrimination. University of Chicago press. [13]. Berik, G., Rodgers, Y. V. D. M., & Zveglich, J. E. (2004). International trade and gender wage discrimination: Evidence from East Asia. Review of Development Economics, 8(2), 237- 254. [14]. Black, S. E., & Brainerd, E. (2004). Importing equality? The impact of globalization on gender discrimination. ILR Review, 57(4), 540-559. [15]. Boler, E. A., Javorcik, B. S., & Ulltveit-Moe, K. H. (2015). Globalization: a woman's best friend? Exporters and the gender wage gap. CEPR Discussion Paper No. 10475. London: Centre for Economic Policy Research. [16]. Christian, M. M., Evers, B. J., & Barrientos, S. (2013). Women in value chains: Making a difference. Capturing the Gains, Revised Summit Briefing, (6.3). [17]. Deb, K. (2021). Global Value Chains in India and Their Impact on Gender Wage Disparity. Foreign Trade Review, 00157325211024003. [18]. Del Carpio, X., Nguyen, H., Pabon, L., & Wang, L. C. (2015). Do minimum wages affect employment? Evidence from the manufacturing sector in Indonesia. IZA Journal of Labor & Development, 4(1), 1-30. [19]. Danon, Z., & Collin, S. R. (2021). Women in the Middle East and North Africa: Issues for Congress. In Congress Research Service. [20]. Doss, C. (2014). If women hold up half the sky, how much of the world’s food do they produce?. Gender in agriculture (pp. 69-88). Springer, Dordrecht. [21]. Dovis, M., & Zaki, C. (2020). Global Value Chains and Local Business Environments: Which Factors Really Matter in Developing Countries?. Review of Industrial Organization, 57(2), 481-513. [22]. Frederick, S., Lopez-Acevedo, G., Robertson, R., & Bahena, M. V. (2022). From Jobs to Careers: Apparel Exports and Career Paths for Women in Developing Countries. World Bank Publications. [23]. Gaddis, I., & Pieters, J. (2017). The gendered labor market impacts of trade liberalization evidence from Brazil. Journal of Human Resources, 52(2), 457-490. [24]. Guha-Khasnobis, B., Aditya, A. & Chandna, S. (2022) Employment and global value chain participation: the Indian experience. International Journal of Economic Policy Studies. https://doi.org/10.1007/s42495-022-00092-7 [25]. Harrison, A., & Scorse, J. (2010). Multinationals and anti-sweatshop activism. American Economic Review, 100(1), 247-273. [26]. Harrison, J. (2019). The labour rights agenda in free trade agreements. The Journal of World Investment & Trade, 20(5), 705-725. [27]. ILO. (2017). Handbook on Assessment of Labour Provisions in Trade and Investment Arrangements: International Labor Office, Geneva. [28]. ILO. (2021). Putting Gender at the Heart of Business Environment Reform. International Labour Office, Geneva. 26 [29]. Jenkins, M. (2005). Economic and social effects of export processing zones in Costa Rica. ILO Working Paper, 97. [30]. Juhn, C., Ujhelyi, G., & Villegas-Sanchez, C. (2014). Men, women, and machines: How trade impacts gender inequality. Journal of Development Economics, 106, 179-193. [31]. Karam, F., & Zaki, C. (2021). On women participation and empowerment in international trade: Impact on trade margins in the MENA region. The Journal of International Trade & Economic Development, 30(3), 384-406. [32]. Koenig, P., & Poncet, S. (2022). The effects of the Rana Plaza collapse on the sourcing choices of French importers. Journal of International Economics, 137, 103576. [33]. Kowalski, P., et al. (2015). Participation of Developing Countries in Global Value Chains: Implications for Trade and Trade-Related Policies. OECD Trade Policy Papers, No. 179, OECD Publishing, Paris, https://doi.org/10.1787/5js33lfw0xxn-en. [34]. Kruse, H. W., Martínez-Zarzoso, I., & Baghdadi, L. (2021). Standards and political connections: Evidence from Tunisia. Journal of Development Economics, 153, 102731. [35]. Kumar, R. (2017) “Global Value Chains: a way to create more, better and inclusive jobs”, the World Bank, https://blogs.worldbank.org/jobs/global-value-chains-way-create-more- better-and-inclusive-jobs [36]. Li, H., & Chen, P. (2018). Board gender diversity and firm performance: The moderating role of firm size. Business Ethics: A European Review, 27(4), 294-308. [37]. Lopez-Acevedo, G., Medvedev, D., & Palmade, V. (2017). South Asia's Turn: Policies to Boost Competitiveness and Create the Next Export Powerhouse. World Bank Publications. [38]. López Mourelo, E., & Samaan, D. (2018). Can labor provisions in trade agreements promote gender equality? Empirical evidence from Cambodia. Review of Development Economics, 22(1), 404-433. [39]. Martínez-Zarzoso, I. (2018). Assessing the Effectiveness of Environmental Provisions in Regional Trade Agreements: An Empirical Analysis. OECD Trade and Environment Working Papers, No. 2018/02, OECD Publishing, Paris, https://doi.org/10.1787/5ffc615c-en. [40]. Monteiro, J. A. (2021). The evolution of gender-related provisions in regional trade agreements (No. ERSD-2021-8). WTO Staff Working Paper. [41]. Said, M., Galal, R., & Sami, M. (2022). Gender diversity, productivity, and wages in private Egyptian firms. Applied Economics, 54(38), 4433-4448. [42]. Shepherd, B., & Stone, S. (2013). Global production networks and employment: a developing country perspective (OECD Trade Policy Papers, No. 154). OECD Publishing. [43]. Staritz, C., & Reis, J. G. (2013). Global value chains, economic upgrading, and gender: Case studies of the horticulture, tourism and call center industries. The World Bank International Trade Department, Gender Development Unit. [44]. Stotsky, M. J. G. (1996). Gender bias in tax systems. International Monetary Fund. [45]. UNCTAD (2020). Gender and trade: assessing the impact of trade agreements on gender equality Canada-EU comprehensive economic and trade agreement. [46]. World Bank. (2020b). Women, Business and the Law 2020. 6th ed. Washington, DC: World Bank. [47]. World Bank (2021), Universal Financial Access 2020. June. [48]. World Bank (2022), Women, Business, and the Law 2022. [49]. World Economic Forum (March 2021), Global Gender Gap Report. [50]. WTO (2022). Informal Working Group on Trade and Gender. INF/TGE/COM/4. 27 [51]. Woods, A., & Joyce, P. (2003). Owner-managers and the practice of strategic management. International Small Business Journal, 21(2), 181-195. 28 Appendix 1: List of Countries and years of the WBES Country Survey years Country Survey years Afghanistan 2008-2014 Congo, Dem. Rep. 2006-2010-2013 Albania 2007-2013-2019 Denmark 2020 Angola 2006-2010 Djibouti 2013 Antigua and Barbuda 2010 Dominica 2010 Argentina 2006-2010-2017 Dominican Republic 2010-2016 Armenia 2009-2013-2020 Ecuador 2006-2010-2017 Austria 2021 Egypt, Arab Rep. 2013-2016-2020 Azerbaijan 2009-2013-2019 El Salvador 2006-2010-2016 Bahamas 2010 Eritrea 2009 Bangladesh 2007-2013 Estonia 2009-2013-2019 Barbados 2010 Eswatini 2006-2016 Belarus 2008-2013-2018 Ethiopia 2011-2015 Belgium 2020 Fiji 2009 Belize 2010 Finland 2020 Benin 2009-2016 France 2021 Bhutan 2009-2015 Gabon 2009 Bolivia 2006-2010-2017 Gambia, The 2006-2018 Bosnia and Herzegovina 2009-2013-2019 Georgia 2008-2013-2019 Botswana 2006-2010 Germany 2021 Brazil 2009 Ghana 2007-2013 Bulgaria 2007-2009-2013-2019 Greece 2018 Burkina Faso 2009 Grenada 2010 Burundi 2006-2014 Guatemala 2006-2010-2017 Cambodia 2013-2016 Guinea 2006-2016 Cameroon 2009-2016 Guinea-Bissau 2006 Cabo Verde 2009 Guyana 2010 Central African Republic 2011 Honduras 2006-2010-2016 Chad 2009-2018 Hungary 2009-2013-2019 Chile 2006-2010 India 2014 China 2012 Indonesia 2009-2015 Colombia 2006-2010-2017 Iraq 2011 Congo, Rep. 2009 Ireland 2020 Costa Rica 2010 Israel 2013 Croatia 2007-2013-2019 Italy 2019 Cyprus 2019 Jamaica 2010 Czechia 2009-2013-2019 Jordan 2013-2019 Côte d'Ivoire 2009-2016 Kazakhstan 2009-2013-2019 Kenya 2007-2013-2018 Romania 2009-2013-2019 Kosovo 2009-2013-2019 Russian Federation 2009-2012-2019 Kyrgyz Republic 2009-2013-2019 Rwanda 2006-2011-2019 LaoPDR 2009-2012-2016-2018 Samoa 2009 Latvia 2009-2013-2019 Senegal 2007-2014 Lebanon 2013-2019 Serbia 2009-2013-2019 29 Lesotho 2009-2016 Sierra Leone 2009-2017 Liberia 2009-2017 Slovak Republic 2009-2013-2019 Lithuania 2009-2013-2019 Slovenia 2009-2013-2019 Luxembourg 2020 Solomon Island 2015 Madagascar 2009-2013 South Africa 2007-2020 Malawi 2009-2014 South Sudan 2014 Malaysia 2015-2019 Spain 2021 Mali 2007-2010-2016 Sri Lanka 2011 Malta 2019 St Kitts and Nevis 2010 Mauritania 2006-2014 St Lucia 2010 Mauritius 2009 St Vincent and Grenadines 2010 Mexico 2006-2010 Sudan 2014 Micronesia 2009 Suriname 2010-2018 Moldova 2009-2013-2019 Sweden 2014-2020 Mongolia 2009-2013-2019 Tajikistan 2008-2013-2019 Montenegro 2009-2013-2019 Tanzania 2006-2013 Morocco 2013-2019 Thailand 2016 Mozambique 2007-2018 Timor-Leste 2009-2015-2021 Myanmar 2014-2016 Togo 2009-2016 Namibia 2006-2014 Tonga 2009 Nepal 2009-2013 Trinidad and Tobago 2010 Netherlands 2020 Tunisia 2013-2020 Nicaragua 2006-2010-2016 Türkiye 2008-2013-2019 Niger 2009-2017 Uganda 2006-2013 Nigeria 2007-2014 Ukraine 2008-2013-2019 North Macedonia 2009-2013-2019 Uruguay 2006-2010-2017 Pakistan 2007-2013 Uzbekistan 2008-2013-2019 Panama 2006-2010 Vanuatu 2009 Papua New Guinea 2015 Venezuela, RB 2006-2010 Paraguay 2006-2010-2017 Vietnam 2009-2015 Peru 2006-2010-2017 West Bank and Gaza 2013-2019 Philippines 2009-2015 Yemen, Rep. 2010-2013 Poland 2009-2013-2019 Zambia 2007-2013-2019 Portugal 2019 Zimbabwe 2011-2016 Source: Constructed by the authors using the WBES. 30 Appendix 2: Variables Definition Variable Definition Ln (Age) Ln of the difference between the year in which the most recent survey is released and the year in which the establishment began operation Ln (Gov own) Ln of the share of government ownership Main City Dummy variable that takes the value 1 if the firm is operating in the main business city Small Dummy variable that takes the value 1 if the number of employees is below 20 Medium Dummy variable that takes the value 1 if the number of employees is between 20 and 99 Large Dummy variable that takes the value 1 if the number of employees is greater than or equal 100 Female Ownership Dummy variable that takes the value 1 if the firm has a female owner Female Top Manager Dummy variable that takes the value 1 if the top manager of the firm is female Ln (Female) Ln of the number of full-time female employees Ln (Femaleproduction) Ln of the number of full-time female production workers Ln (Femalenonprod) Ln of the number of full-time female non-production workers GVC1 Dummy variable that takes the value 1 if the firm is exporting and importing at the same time GVC2 Dummy variable that takes the value 1 if the firm is exporting and importing at the same time and if it has an international quality certification GVC3 Dummy variable that takes the value 1 if the firm is exporting and importing at the same time and if the share of private foreign ownership of the firm is greater than 10% GVC4 Dummy variable that takes the value 1 if the firm is exporting and importing at the same time, if it has an international quality certification and if the share of private foreign ownership of the firm is greater than 10% Gender Provisions The number of gender-related provisions in regional trade agreements. Source: Constructed by the authors using the WBES. 31 Appendix 3: Descriptive Statistics Table A1. Firms integrating into GVCs Number of firms Percentage of total firms GVC1 28,681 19.97% GVC2 12,485 8.69% GVC3 6,083 4.24% GVC4 3,288 2.29% Total 50,537 35.18% Source: Constructed by the authors using the WBES. Note: GVC1 refers to firms that export and import simultaneously, GVC2 = GVC1 + international certification, GVC3= GVC1+ share of its capital owned by a foreign firm, GVC4 combines the four criteria altogether. Table A2. The Average Number of Female Employees, by region 11 Region Female Employees Female Production Workers Female Non-Production Workers SSA 14.17 14.99 5.10 EAP 28.49 72.15 13.85 ECA 24.09 27.84 8.28 LAC 39.77 25.59 12.98 MENA 11.03 17.55 5.10 SAR 9.33 42 4.25 Source: Constructed by the authors using the WBES. Note: SAR stands for South Asia, MENA Middle East and North Africa, EAP East Asia and Pacific, SSA Sub-Saharan Africa, LAC Latin America and the Caribbean and ECA Europe and Central Asia. Figure A1. The Distribution of Firms’ integration into GVC, by region 100% 90% 80% 34.4% 38.6% 43.4% 49.5% 70% 60% 22.7% 50% 23.2% 22.1% 40% 19.4% 18.8% 13.1% 30% 12.3% 8.2% 20% 8.7% 8.9% 15.8% 16.5% 10% 8.1% 8.1% 5.9% 6.3% 6.1% 0% 5.3% 2.0% 2.5% GVC1 GVC2 GVC3 GVC4 SAR MENA EAP SSA LAC ECA Source: Constructed by the authors using the WBES. Notes: (i) SAR stands for South Asia, MENA Middle East and North Africa, EAP East Asia and Pacific, SSA Sub- Saharan Africa, LAC Latin America and the Caribbean and ECA Europe and Central Asia. (ii) GVC1 refers to firms that export and import simultaneously, GVC2 = GVC1 + international certification, GVC3= GVC1+ share of its capital owned by a foreign firm, GVC4 combines the four criteria altogether. 11 Given that some firms might have one of the variables missing, the sum of production and non-production does not necessarily add to the total number of employees. 32 Appendix 4: Empirical Results Table A3. GVCs and Female Non-Production Workers – Baseline Results (a) All regions (b) MENA region Variables GVC1 GVC4 GVC1 GVC4 Ln(Age) 0.102*** 0.105*** 0.0647*** 0.0675*** (0.00721) (0.00719) (0.0167) (0.0168) Ln(Gov own.) 0.0697*** 0.0706*** 0.151*** 0.150*** (0.0153) (0.0154) (0.0541) (0.0537) Main city 0.104*** 0.110*** 0.122*** 0.125*** (0.0100) (0.0101) (0.0243) (0.0244) Medium 0.619*** 0.643*** 0.467*** 0.494*** (0.00763) (0.00747) (0.0181) (0.0177) Large 1.775*** 1.817*** 1.601*** 1.673*** (0.0142) (0.0140) (0.0392) (0.0379) GVC 0.229*** 0.446*** 0.227*** 0.336*** (0.0117) (0.0284) (0.0337) (0.114) Constant 0.125*** 0.142*** 0.00424 0.0139 (0.0222) (0.0221) (0.0524) (0.0524) Country dummies Yes Yes Yes Yes Year dummies Yes Yes Yes Yes Sector dummies Yes Yes Yes Yes Observations 41,353 41,353 6,863 6,863 R-squared 0.538 0.537 0.469 0.466 Robust standard errors in parentheses, *** p<0.01, ** p<0.05, * p<0.1 33 Table A4. GVC & Female Labor Force Participation – by Sector (a) All regions included Fem. Ownership Fem Top Manager Fem. Employees Fem. Production Workers Variables GVC1 GVC4 GVC1 GVC4 GVC1 GVC4 GVC1 GVC4 (1) GVC -0.020 -0.098*** -0.021** 0.0141 0.228 1.679** 0.668*** 0.426 (0.015) (0.035) (0.010) (0.0262) (0.252) (0.793) (0.113) (0.376) (2) Textile 0.074*** 0.077*** 0.098*** 0.093*** -0.133 -0.0126 0.764*** 1.094*** (0.010) (0.009) (0.008) (0.007) (0.142) (0.132) (0.0623) (0.0606) (3) Leather -0.046** 0.0017 -0.058*** -0.038*** 0.754*** 0.980*** -0.156** -0.178** (0.021) (0.021) (0.010) (0.011) (0.245) (0.0848) (0.0706) (0.0693) (4) Wood -0.052*** -0.042** -0.053*** -0.049*** -0.809*** -0.807*** -0.467*** -0.587*** (0.018) (0.017) (0.013) (0.012) (0.185) (0.185) (0.0523) (0.0552) (5) Publishing -0.055 -0.048 -0.075*** -0.075*** -0.292* -0.338** -0.148 -0.213 (0.037) (0.035) (0.022) (0.020) (0.149) (0.133) (0.150) (0.155) (6) Chemicals -0.016 -0.006 -0.025** -0.018* -0.462* -0.380 -0.156** -0.287*** (0.015) (0.013) (0.011) (0.010) (0.278) (0.262) (0.0762) (0.0751) (7) Rubb&Pla -0.053*** -0.050*** -0.050*** -0.052*** -0.421 -0.421 -0.224* -0.247** (0.017) (0.015) (0.013) (0.012) (0.438) (0.436) (0.121) (0.121) (8) Machinery -0.117*** -0.103*** -0.088*** -0.086*** -0.65*** -0.779*** -0.620*** -0.701*** (0.012) (0.010) (0.009) (0.007) (0.153) (0.131) (0.0730) (0.0720) (9) Fab. Metals -0.089*** -0.078*** -0.083*** -0.076*** -0.972*** -0.820*** -0.539*** -0.618*** (0.011) (0.010) (0.008) (0.007) (0.144) (0.130) (0.0532) (0.0567) (10) Furniture -0.144*** -0.119*** -0.090*** -0.086*** -0.801*** -0.789*** -0.320*** -0.358*** (0.018) (0.017) (0.013) (0.012) (0.146) (0.111) (0.0913) (0.0947) (11) Electro. -0.096*** -0.088*** -0.083*** -0.073*** -0.789*** -0.787*** (0.022) (0.019) (0.018) (0.016) (0.279) (0.278) (12) Oth. Man. -0.060*** -0.047*** -0.053*** -0.045*** -0.350*** -0.325*** -0.520*** -0.551*** (0.007) (0.006) (0.005) (0.005) (0.0841) (0.0790) (0.0402) (0.0411) (13) Services -0.0162** -0.0146** -0.00258 -0.003 -0.192** -0.189*** -0.556*** -0.680*** (0.007) (0.006) (0.005) (0.005) (0.0768) (0.0725) (0.0797) (0.0777) (14) Wholesale -0.049*** -0.042*** -0.047*** -0.047*** -0.254*** -0.246*** -0.242 -0.357** (0.0127) (0.0120) (0.010) (0.009) (0.0801) (0.0758) (0.164) (0.161) (1)*(2) 0.0254 0.113* -0.011 -0.0733* 0.402 0 0.661*** 0.798* (0.0210) (0.0597) (0.0160) (0.0433) (0.358) (0) (0.150) (0.460) (1)*(3) 0.227*** -0.0735 0.104*** -0.0116 0 0 -0.355 2.411*** (0.0597) (0.0452) (0.0343) (0.0287) (0) (0) (0.269) (0.633) (1)*(4) 0.0485 -0.220*** 0.0191 -0.137*** 0 0 -1.339*** -1.315** (0.0490) (0.051) (0.0323) (0.0484) (0) (0) (0.201) (0.556) (1)*(5) 0.0290 -0.0762 -0.0199 0.0121 -0.481 0 -0.535 1.782*** (0.116) (0.0497) (0.0272) (0.0331) (0.312) (0) (0.885) (0.404) (1)*(6) 0.0495* 0.117* 0.0396* 0.0177 0.974** 0 -0.830*** 0.214 (0.0290) (0.0638) (0.0209) (0.0475) (0.452) (0) (0.211) (0.669) (1)*(7) -0.000906 -0.0454 -0.00354 -0.0465 0 -0.185 -0.396 1.145* (0.0371) (0.0742) (0.0247) (0.0531) (0) (0.832) (0.386) (0.660) (1)*(8) 0.0394* -0.0367 0.0188 -0.0131 -0.626** 0 -0.685*** 0.0955 (0.0225) (0.0498) (0.0148) (0.0350) (0.318) (0) (0.216) (0.723) (1)*(9) 0.0446* -0.0378 0.0320** -0.0231 0.344 0 -0.962*** -0.392 (0.0233) (0.0546) (0.0151) (0.0361) (0.329) (0) (0.228) (0.707) (1)*(10) 0.137*** 0.230 0.0150 -0.0513 -0.123 0 -0.879* 0 (0.0456) (0.178) (0.0282) (0.0403) (0.311) (0) (0.513) (0) (1)*(11) 0.00799 -0.0318 0.0390 -0.0304 0 -1.394 (0.0400) (0.0675) (0.0331) (0.0584) (0) (1.018) (1)*(12) 0.0460*** -0.0308 0.0342*** -0.0119 -0.130 0 -0.336*** -0.153 (0.0162) (0.0383) (0.0114) (0.0287) (0.263) (0) (0.127) (0.452) (1)*(13) 0.0142 0.0453 -0.0132 -0.0412 -0.173 -1.828** -0.443** 0.00548 (0.0163) (0.0440) (0.0116) (0.0323) (0.252) (0.806) (0.216) (0.389) (1)*(14) 0.0319 -0.212*** -0.0102 -0.0148 -0.127 -1.286 -0.177 3.109*** (0.0328) (0.0383) (0.0197) (0.107) (0.259) (0.795) (0.686) (0.440) Observations 83,949 83,949 84,341 84,341 41,227 -0.727 6,864 6,864 R-squared 0.111 0.111 0.090 0.090 0.528 (0.872) 0.478 0.456 34 (b) MENA region Female Ownership Female Top Manager Female Employees Female Production Workers Variables GVC1 GVC4 GVC1 GVC4 GVC1 GVC4 GVC1 GVC4 (1) GVC 0.0388 -0.0352 -0.00345 0.0286 0.120 1.013*** 0.668*** 0.426 (0.0301) (0.0787) (0.0147) (0.0477) (0.448) (0.297) (0.113) (0.376) (2) Textile 0.00911 0.0302* 0.0180 0.0234** 0.435 0.726* 0.764*** 1.094*** (0.0187) (0.0172) (0.0123) (0.0106) (0.438) (0.402) (0.0623) (0.0606) (3) Leather -0.00764 0.0153 -0.026** -0.0243** 0.936** 1.112*** -0.156** -0.178** (0.0238) (0.0241) (0.0126) (0.0118) (0.395) (0.210) (0.0706) (0.0693) (4) Wood 0.00446 -0.0007 -0.0150 -0.0138 -0.467*** -0.587*** (0.0205) (0.0197) (0.0116) (0.0107) (0.0523) (0.0552) (5) Publishing 0.0405 0.0423 0.0167 0.0143 -0.400* -0.325 -0.148 -0.213 (0.0635) (0.0622) (0.0403) (0.0377) (0.230) (0.220) (0.150) (0.155) (6) Chemicals 0.0404 0.0245 0.00497 0.0103 -0.385 -0.316 -0.156** -0.287*** (0.0247) (0.0216) (0.0148) (0.0132) (0.234) (0.222) (0.0762) (0.0751) (7) Rubb&Pla 0.0401 0.0721 0.0405 0.0275 0.349 0.416** -0.224* -0.247** (0.0465) (0.0456) (0.0325) (0.0278) (0.221) (0.207) (0.121) (0.121) (8) Machinery -0.0215 -0.0200 -0.0175 -0.0209* -0.620*** -0.701*** (0.0198) (0.0194) (0.0127) (0.0109) (0.0730) (0.0720) (9) Furniture 0.000521 0.00501 -0.00651 -0.00624 -0.539*** -0.618*** (0.0208) (0.0203) (0.0127) (0.0117) (0.0532) (0.0567) (10) Elect. -0.0173 -0.0159 0.0132 0.0107 -0.480** -0.412* -0.320*** -0.358*** (0.0359) (0.0351) (0.0244) (0.0227) (0.243) (0.233) (0.0913) (0.0947) (11)OtherMan. 0.0163 0.0165 -0.00255 0.00197 -0.260 -0.115 -0.520*** -0.551*** (0.0138) (0.0129) (0.00809) (0.00723) (0.245) (0.224) (0.0402) (0.0411) (12) Services 0.0142 0.00990 0.0160** 0.0147** -0.0605 0.0216 -0.556*** -0.680*** (0.0118) (0.0114) (0.00767) (0.00710) (0.220) (0.204) (0.0797) (0.0777) (13)Wholesale 0.0337 0.0298 -0.00478 -0.00635 -0.0620 0.0110 -0.242 -0.357** (0.0251) (0.0229) (0.0138) (0.0119) (0.224) (0.209) (0.164) (0.161) (1)*(2) 0.0571 0.142 0.0224 0.0410 0.957 0 0.661*** 0.798* (0.0422) (0.109) (0.0234) (0.0736) (0.801) (0) (0.150) (0.460) (1)*(3) 0.134 -0.154* 0.00498 -0.0292 0 0 -0.355 2.411*** (0.0845) (0.0835) (0.0307) (0.0494) (0) (0) (0.269) (0.633) (1)*(4) -0.0525 -0.23*** 0.00326 -0.0686 -1.339*** -1.315** (0.0616) (0.0819) (0.0267) (0.0501) (0.201) (0.556) (1)*(5) -0.0332 -0.201** -0.0570 -0.0865 0 0 -0.535 1.782*** (0.231) (0.0998) (0.0423) (0.0605) (0) (0) (0.885) (0.404) (1)*(6) -0.0950* -0.133 0.0106 -0.0772 1.648*** 0.745** -0.830*** 0.214 (0.0491) (0.0985) (0.0287) (0.0494) (0.455) (0.319) (0.211) (0.669) (1)*(7) 0.253* 0.719*** -0.0286 0.405 0 0 -0.396 1.145* (0.133) (0.0911) (0.0670) (0.352) (0) (0) (0.386) (0.660) (1)*(8) -0.00438 -0.0275 -0.0186 -0.0392 -0.685*** 0.0955 (0.0570) (0.140) (0.0182) (0.0487) (0.216) (0.723) (1)*(9) 0.0406 0.118 -0.00412 -0.0577 -0.962*** -0.392 (0.0693) (0.298) (0.0271) (0.0499) (0.228) (0.707) (1)*(10) -0.00377 -0.183** -0.0526* -0.0876* 1.469*** 0.560* -0.879* 0 (0.137) (0.0851) (0.0274) (0.0525) (0.458) (0.325) (0.513) (0) (1)*(11) -0.0191 -0.115 0.0102 -0.0878* 0.231 -1.36*** -0.336*** -0.153 (0.0352) (0.0953) (0.0173) (0.0481) (0.496) (0.380) (0.127) (0.452) (1)*(12) -0.0254 -0.0976 -0.0196 -0.0306 0.127 -0.209 -0.443** 0.00548 (0.0357) (0.0984) (0.0179) (0.0688) (0.450) (0.415) (0.216) (0.389) (1)*(13) -0.0272 -0.30*** -0.0114 -0.0840 -0.00772 -1.056** -0.177 3.109*** (0.0551) (0.0943) (0.0234) (0.0525) (0.457) (0.412) (0.686) (0.440) Observations 11,714 11,714 11,730 11,730 4,861 4,861 6,864 6,864 R-squared 0.108 0.107 0.016 0.017 0.496 0.495 0.478 0.456 Robust standard errors in parentheses, *** p<0.01, ** p<0.05, * p<0.1 Note: - All the regressions control for firm’s size, age, share of government ownership and city of operation. - Country and year fixed effects are included. - The intercept is included. 35 Table A5. Gender Provisions, GVCs and Female Non-Production Workers (a) All regions (b) MENA GVC1 GVC4 GVC1 GVC4 (1)GVC 0.229*** 0.448*** 0.358*** 0.096 (0.0296) (0.0541) (0.088) (0.229) (2)Gender Provisions -0.121*** -0.113*** 0.026 0.010 (0.0250) (0.0253) (0.045) (0.051) GVC*Gender Prov. 2.24e-05 -4.01e-05 -0.020 0.074* (0.000429) (0.000829) (0.022) (0.056) Country dummies Yes Yes No No Year dummies Yes Yes Yes Yes Sector dummies Yes Yes Yes Yes Observations 41,356 41,356 6,864 6,864 R-squared 0.538 0.537 0.469 0.466 Robust standard errors in parentheses, clustered by country and year, *** p<0.01, ** p<0.05, * p<0.1 Note: - We control for firm’s age, size, main city and the share of government ownership. - The intercept is included. Table A6. IV Approach – First Stage (a) All regions (b) MENA region Variables GVC1 GVC4 GVC1 GVC4 IV:Shift_Share_GVC 0.481*** 0.227*** 0.404*** 0.168*** (0.008) (0.010) (0.022) (0.026) Ln(Age) 0.015*** -0.001 0.001 -0.007*** (0.002) (0.001) (0.006) (0.002) Medium 0.084*** 0.012*** 0.092*** 0.007** (0.003) (0.001) (0.007) (0.003) Large 0.251*** 0.073*** 0.306*** 0.056*** (0.004) (0.001) (0.010) (0.003) Main city 0.010*** -0.0001 0.015** 0.002 (0.003) (0.001) (0.008) (0.003) Ln(Gov own) 0.013*** 0.002 0.012 0.005 (0.003) (0.001) (0.010) (0.004) Country dummies Yes Yes Yes Yes Year dummies Yes Yes Yes Yes Sector dummies Yes Yes Yes Yes Observations 82,937 82,937 11,610 11,610 Underidentification test P-Val 0.000 0.000 0.000 0.000 Cragg-Donald Wald F-statistic 3849.44 518.03 334.91 41.36 Standard errors in parentheses, *** p<0.01, ** p<0.05, * p<0.1 Notes: - The intercept is included. - The minimum Eigenvalue is higher than all the critical values at 10%. 36 Appendix 5: Propensity Score Matching Table A7. First Stage – Probit Estimation (a) All regions (b) MENA region Variables GVC1 GVC4 GVC1 GVC4 Medium 0.436*** 0.613*** 0.526*** 0.612*** (0.013) (0.036) (0.036) (0.121) Large 1.048*** 1.374*** 1.258*** 1.396*** (0.016) (0.037) (0.043) (0.122) Main city 0.130*** 0.077*** 0.138*** 0.125 (0.014) (0.030) (0.037) (0.083) Ln(Age) 0.063*** -0.064*** 0.018 -0.222*** (0.010) (0.021) (0.028) (0.065) Ln(Gov own) 0.063*** 0.049*** 0.038 0.074 (0.013) (0.022) (0.040) (0.066) Country dummies Yes Yes Yes Yes Year dummies Yes Yes Yes Yes Sector dummies Yes Yes Yes Yes Observations 84,292 77,973 11,446 10,220 Standard errors in parentheses, *** p<0.01, ** p<0.05, * p<0.1 Note: The intercept is included. 37 [1] Nearest Neighborhood Matching Method Figure A2. Overlap (common support) of propensity scores between the treated and untreated group (a) Treatment: GVC1, All regions (b) Treatment: GVC4, All regions (c) Treatment: GVC1, MENA region (d) Treatment: GVC4, MENA region Table A8. Common Support, Outcome: Female Ownership (a) GVC1, All regions (b) GVC4, All regions Treatment Common Total Treatment Common Total assignment support assignment support On support On support Untreated 66,790 66,790 Untreated 75,726 75,726 Treated 17,097 17,097 Treated 1,842 1,842 Total 83,887 83,887 Total 77,568 77,568 (c) GVC1, MENA region (d) GVC4, MENA region Treatment Common Total Treatment Common Total assignment support assignment support On support On support Untreated 9,235 9,235 Untreated 10,017 10,017 Treated 2,195 2,195 Treated 187 187 Total 11,430 11,430 Total 10,204 10,204 38 Table A9. Common Support, Outcome: Female Top Manager (a) GVC1, All regions (b) GVC4, All regions Treatment Common Total Treatment Common Total assignment support assignment support On support On support Untreated 67,059 67,059 Untreated 76,103 76,103 Treated 17,220 17,220 Treated 1,857 1,857 Total 84,279 84,279 Total 77,960 77,960 (c) GVC1, MENA region (d) GVC4, MENA region Treatment Common Total Treatment Common Total assignment support assignment support On support On support Untreated 9,246 9,246 Untreated 10,033 10,033 Treated 2,200 2,200 Treated 187 187 Total 11,446 11,446 Total 10,220 10,220 Table A10. Common Support, Outcome: Number of Full-Time Female Employees (a) GVC1, All regions (b) GVC4, All regions Treatment Common Total Treatment Common Total assignment support assignment support On support On support Untreated 35,900 35,900 Untreated 28,901 28,901 Treated 5,207 5,207 Treated 334 334 Total 41,107 41,107 Total 29,235 29,235 (c) GVC1, MENA region (d) GVC4, MENA region Treatment Common Total Treatment Common Total assignment support assignment support On support On support Untreated 3,954 3,954 Untreated 3,101 3,101 Treated 622 622 Treated 35 35 Total 4,576 4,576 Total 3,136 3,136 39 Table A11. Common Support, Outcome: Number of Full-Time Female Production Workers (a) GVC1, All regions (b) GVC4, All regions Treatment Common Total Treatment Common Total assignment support assignment support On support On support Untreated 29,503 29,503 Untreated 37,563 37,563 Treated 11,600 11,600 Treated 1,417 1,417 Total 41,103 41,103 Total 38,980 38,980 (c) GVC1, MENA region (d) GVC4, MENA region Treatment Common Total Treatment Common Total assignment support assignment support On support On support Untreated 5,270 5,270 Untreated 6,147 6,147 Treated 1,575 1,575 Treated 152 152 Total 6,845 6,845 Total 6,299 6,299 Table A12. GVCs and Female Labor Force Participation – PSM (a) All regions Fem. Fem. Fem. Fem. Production Fem. Production Ownership Top Manager Employees Workers Workers GVC1 GVC4 GVC1 GVC4 GVC1 GVC4 GVC1 GVC4 GVC1 GVC4 Treatment 0.040*** -0.060*** -0.04*** -0.034*** 0.338*** 1.157*** 0.952*** 1.466*** 0.870*** 1.415*** (0.004) (0.011) (0.003) (0.009) (0.018) (0.067) (0.017) (0.043) (0.012) (0.031) Constant 0.309*** 0.329*** 0.169*** 0.160*** 1.645*** 1.698*** 1.151*** 1.381*** 0.919*** 1.136*** (0.002) (0.002) (0.001) (0.001) (0.006) (0.007) (0.009) (0.008) (0.006) (0.006) Obs. 83,887 77,568 84,279 77,960 41,107 29,235 41,103 38,980 41,215 39,175 (b) MENA region Fem. Fem. Fem. Fem. Production Fem. Production Ownership Top Manager Employees Workers Workers GVC1 GVC4 GVC1 GVC4 GVC1 GVC4 GVC1 GVC4 GVC1 GVC4 Treatment 0.111*** 0.008 -0.005 -0.0009 0.634*** 1.664*** 1.250*** 1.801*** 0.953*** 1.290*** (0.009) (0.029) (0.005) (0.017) (0.054) (0.213) (0.041) (0.126) (0.030) (0.093) Constant 0.170*** 0.195*** 0.055*** 0.054*** 1.206*** 1.399*** 0.661*** 0.962*** 0.601*** 0.831*** (0.004) (0.004) (0.002) (0.002) (0.020) (0.023) (0.020) (0.020) (0.015) (0.014) Obs. 11,430 10,204 11,446 10,220 4,576 3,136 6,845 6,299 6,845 6,299 Standard errors in parentheses *** p<0.01, ** p<0.05, * p<0.1 Note: All the regressions include country, year and sector dummies. 40 Table A13. Common Support, Outcome: Number of Full-Time Female Non-Production Workers (a) GVC1, All regions (b) GVC4, All regions Treatment Common Total Treatment Common Total assignment support assignment support On support On support Untreated 29,691 29,691 Untreated 37,758 37,758 Treated 11,524 11,524 Treated 1,417 1,417 Total 41,215 41,215 Total 39,175 39,175 (c) GVC1, MENA region (d) GVC4, MENA region Treatment Common Total Treatment Common Total assignment support assignment support On support On support Untreated 5,270 5,270 Untreated 6,147 6,147 Treated 1,575 1,575 Treated 152 152 Total 6,845 6,845 Total 6,299 6,299 [2] Kernel: Pair Matching with Replacement Figure A3. Quality of Kernel Matching: Cumulative Distribution Functions of Propensity Scores before and after the PSM 41 Table A14. PSM – Matching Statistics (Kernel pair matching with replacement) (a) All regions Matched Treatment: GVC1 Yes No Total Bandwidth Treated 16232 865 17097 0.0002875 Untreated 64300 2552 66852 0.0001005 Combined 80532 3417 83949 Treatment: GVC4 Treated 1747 95 1842 0.0002037 Untreated 72133 9974 82107 0.0002 Combined 73880 10069 83949 (b) MENA region Matched Treatment: GVC1 Yes No Total Bandwidth Treated 2087 108 2195 0.0012 Untreated 8741 778 9519 0.0008998 Combined 10828 886 11714 Treatment: GVC4 Treated 178 9 187 0.0002743 Untreated 9869 1658 11527 0.001471 Combined 10047 1667 11714 Table A15. GVCs and Female Non-Production Workers – PSM (Kernel Pair Matching with Replacement) (a) All regions GVC1 GVC4 ATE 0.243*** 0.491*** (0.019) (0.118) ATT 0.206*** 0.406*** (0.022) (0.039) Observations 41,356 41,356 (b) MENA region GVC1 GVC4 ATE 0.188*** 0.341 (0.047) (0.295) ATT 0.252*** 0.317*** (0.052) (0.146) Observations 6,864 6,864 Standard errors in parentheses *** p<0.01, ** p<0.05, * p<0.1 Note: All the regressions include country, year and sector dummies. 42 Table A16. GVCs and Full-Time Female Non-Production Workers – IV Approach (a) All regions (b) MENA region Variables GVC1 GVC4 GVC1 GVC4 Ln(Age) 0.0986*** 0.111*** 0.0659*** 0.107*** (0.00726) (0.00851) (0.0179) (0.0266) Ln(Gov own.) 0.0708*** 0.0713*** 0.155*** 0.0923** (0.0106) (0.0123) (0.0315) (0.0461) Main city 0.0986*** 0.105*** 0.117*** 0.110*** (0.0102) (0.0117) (0.0243) (0.0317) Medium 0.589*** 0.590*** 0.431*** 0.457*** (0.0134) (0.0160) (0.0319) (0.0330) Large 1.689*** 1.542*** 1.475*** 1.334*** (0.0284) (0.0627) (0.0754) (0.118) GVCs 0.456*** 3.151*** 0.524*** 5.147*** (0.0694) (0.607) (0.166) (1.599) Country dummies Yes Yes Yes Yes Year dummies Yes Yes Yes Yes Sector dummies Yes Yes Yes Yes Observations 40,750 40,750 6,797 6,797 Standard errors in parentheses, *** p<0.01, ** p<0.05, * p<0.1 Note: The intercept is included. 43 Table A17. Female Non-Production Workers, GVCs and Barriers Finance Tax rates Pol. Inst. Corrup. Comp. Elec. Inad. Edu. (a) All Regions (1) GVC 0.277*** 0.311*** 0.229*** 0.264*** 0.299*** 0.259*** 0.285*** (0.0325) (0.0470) (0.0336) (0.0296) (0.0415) (0.0378) (0.0358) (2) Obstacle -0.0342 0.0453 0.00585 0.0534 -0.182 -0.198* 0.281** (0.114) (0.131) (0.120) (0.151) (0.136) (0.112) (0.141) GVC1 (1)*(2) -0.257** -0.313** 0.00194 -0.148 -0.377** -0.106 -0.292** (0.125) (0.134) (0.140) (0.130) (0.166) (0.137) (0.133) Observations 41,324 41,324 41,324 41,324 41,324 41,324 41,324 R-squared 0.538 0.538 0.538 0.538 0.538 0.538 0.538 (1) GVC 0.507*** 0.523*** 0.524*** 0.499*** 0.491*** 0.511*** 0.603*** (0.0449) (0.0647) (0.0500) (0.0456) (0.0468) (0.0610) (0.0680) (2) Obstacle -0.110 -0.0615 0.0250 0.0251 -0.297* -0.221** 0.245* (0.112) (0.150) (0.134) (0.163) (0.162) (0.104) (0.145) GVC4 (1)*(2) -0.411 -0.349 -0.372 -0.315 -0.296 -0.271 -0.834*** (0.302) (0.283) (0.238) (0.246) (0.266) (0.203) (0.264) Observations 41,324 41,324 41,324 41,324 41,324 41,324 41,324 R-squared 0.537 0.537 0.537 0.537 0.537 0.537 0.537 (b) MENA region (1) GVC 0.339** 0.206 0.255** 0.245* 0.464*** 0.307** 0.392*** (0.140) (0.134) (0.116) (0.138) (0.130) (0.139) (0.0908) (2) Obstacle -0.467** -0.133 -0.816** -0.684* -0.227 -0.283* -0.0198 (0.190) (0.298) (0.319) (0.336) (0.183) (0.163) (0.210) GVC1 (1)*(2) -0.435 0.0825 -0.0635 -0.0464 -0.985** -0.281 -0.972** (0.397) (0.456) (0.207) (0.320) (0.456) (0.332) (0.363) Observations 6,861 6,861 6,861 6,861 6,861 6,861 6,861 R-squared 0.470 0.469 0.469 0.469 0.470 0.469 0.470 (1) GVC 0.643** 0.762* 0.509 0.632* 0.436 0.632* 1.046*** (0.292) (0.428) (0.332) (0.358) (0.387) (0.331) (0.164) (2) Obstacle -0.519*** -0.148 -0.833** -0.683** -0.471*** -0.347** -0.190 (0.186) (0.258) (0.306) (0.326) (0.158) (0.160) (0.209) GVC4 (1)*(2) -1.299 -1.610 -0.390 -0.768 -0.414 -1.109 -3.474*** (1.006) (1.389) (0.894) (0.971) (1.319) (0.884) (0.658) Observations 6,861 6,861 6,861 6,861 6,861 6,861 6,861 R-squared 0.466 0.466 0.466 0.466 0.466 0.466 0.468 Robust standard errors clustered by country and year in parentheses, *** p<0.01, ** p<0.05, * p<0.1 Notes: i) Each regression controls for firms’ age, city of operation and share of government ownership. ii) All the regressions include country, year, sector and size fixed effects. iii) Country-year-sector averages are used to reduce the risk of endogeneity between the business environment and firm-level. iv) Each column for each GVC definition represents a separate regression. v) All the variables are in log. vi) The intercept is included. 44