Policy Research Working Paper 10868 Export and Labor Market Outcomes A Supply Chain Perspective—Evidence from Viet Nam Deeksha Kokas Gladys Lopez-Acevedo Ha Vu Poverty and Equity Global Practice August 2024 Policy Research Working Paper 10868 Abstract Are changes in the labor market in response to changes in exports in the worker’s own industry) and indirect exposure exports contained specifically within exporting industries, (from increased exports in other industries that use inputs or do they disperse throughout the economy through supply from the worker’s industry). Estimates of the repercussions chain linkages? This paper studies the case of Viet Nam, from increasing exports on labor market outcomes show an example of a successful export-led growth economy, to that both direct and indirect exposure significantly increase examine this question. Combining UN COMTRADE data, workers’ wages and employment, while reducing inactiv- input-output tables from the Global Trade Analysis Project, ity and inequality. Wage premiums for attending college and 2010 to 2019 annual labor force survey data for Viet decrease, and the gender wage gap narrows. Wages increase Nam, the study constructed a measure of each worker’s more for the lowest-income workers and employment gains total exposure to export shocks. The measure accounts for accrue more to unskilled workers, while employment changes due to both direct export exposure (increase in decreases for more skilled workers. This paper is a product of the Poverty and Equity Global Practice. 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 gacevedo@worldbank.org. 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 Export and Labor Market Outcomes: A Supply Chain Perspective - Evidence from Viet Nam Deeksha Kokas a Gladys Lopez-Acevedo b Ha Vu c a Nanyang Technological University b Poverty and Equity Global Practice, World Bank * c Economic Frontiers, International Institute for Applied System Analysis * STATMEMENTS AND DECLARATIONS: This paper is a product of the Poverty and Equity Practice at the World Bank Group. It was funded by the UFT Trade Facilitation Program and the Australia World Bank Group Strategic Partnership in Viet Nam Phase 2 (ABP2). We thank the following colleagues for their support during this process Raymond Robertson, Matthew Wai-Poi, Rinku Murgai, Judy Yang, Dorsati Madani, Sacha Dray and Deborah Elisabeth Winkler. The views expressed herein are those of the authors and do not necessarily reflect the views of the World Bank. 1 Introduction In the past two decades, the notable uptick in the integration of developing nations into global trade and value chains (GVCs) has sparked heightened interest among policy makers and researchers regarding the implications for labor markets. Consequently, a substantial body of literature has emerged seeking to unravel the intricate relationship between trade dynamics and localized labor market outcomes. Standard theory about the benefits of trade assumes perfect mobility of factors across geo- graphical regions and industries within a country. However, recent empirical evidence has shed light on the limitations of these conventional theoretical models. Topalova (2010) [9] uses data from India to indicate that regions more exposed to trade liberalization have experienced slower poverty reduction and muted consumption growth, diverging from the predictions of traditional trade frameworks. Assuming segmented labor markets, and exploiting cross-market variation in import exposure, a key study by Autor, Dorn, and Hanson (2013) [2] confirms that the ”China Shock” led to significant decline in employment and wages in more exposed U.S. regions. This sparked more studies on the repercussions of tariff changes or import competition on local labor markets (Pierre and Schott 2016 [7]; Acemoglu et al. 2016 [1]; Dix-Carneiro and Kovak 2013 [3]). It is crucial to note that trade affects not only tradable but also non-tradable sectors within the same local labor markets; while increased import competition or market access directly im- pacts specific tradable sectors, there are also indirect effects on non-tradable sectors such as retail, healthcare, or hospitality in the same region. With few exceptions, current literature largely ignores estimating these indirect effects propagated through domestic production linkages within a country [10]. While a few studies have examined the indirect effects of import shocks on local labor mar- kets, there has been limited exploration of indirect effects of exports. A central motivation for this empirical inquiry stems from the well-documented fact that as economies undergo structural trans- formation—that is, as they move from less agriculture to more services and manufacturing—the 1 proportion of domestic services or inputs in total outputs tends to rise (McCaig 2013 [6]; Ghosh 2021). At first glance, this might suggest that a larger proportion of employment remains unaf- fected by trade, given the larger non-tradable component of the services sector. However, more trade can even indirectly influence non-tradable industries that serve as inputs to tradeable sectors. This paper focuses on estimating the total effect on labor market outcomes in response to an export shock at the provincial level in Viet Nam from 2009 to 2019. In order to fully capture the overall impact of trade, we need to consider crucial supply chain linkages previous research has overlooked. Our analysis goes beyond solely focusing on directly exporting industries and accounts for industries indirectly affected by the rising demand for exports. To accomplish this, we adopt a multi-stage approach. Initially, we employ an instrumental variable (IV) methodology to isolate an exogenous component of trade driven solely by foreign demand. The chosen IV is the proportion of a given trading partner country’s share in a specific commodity relative to Viet Nam’s total export value in that commodity, adjusted by the observed GDP growth of the partner country. This IV demonstrates a strong predictive capacity for Viet Nam’s export values while maintaining plausible independence from any supply-side determinants of exports. Subsequently, leveraging the predicted export exposure obtained in the first stage, we construct a matrix delineating the direct and indirect exposure of each industry in each province. This is accomplished by utilizing input-output tables that document the flow of intermediate goods among different sectors in the economy. The predicted export exposure calculated earlier is then dis- tributed proportionally to the labor share of each industry within each province. Finally, we estimate the direct and indirect effects of exports by conducting a regression analy- sis of these computed exposures against various labor market outcomes of interest at the provincial level: wage levels, income disparities, the premium on college education wages, gender wage differentials, employment rates, rates of inactivity, informality in employment, and female em- ployment rates. Additionally, we conduct a detailed examination of how these labor effects differ 2 across gender, income levels, educational attainment, and employment sectors to elucidate which demographic groups benefit and which are adversely affected by this process. Our analysis goes beyond solely focusing on directly exporting industries and accounts for industries indirectly af- fected by the rising demand for exports. This study looks at the specific context of Viet Nam for a few important reasons. First, Viet Nam exemplifies the success of Asia’s export-driven growth model, making it a prime candidate for study due to the wealth of empirical evidence available. Over the past two decades, Viet Nam has witnessed substantial increases in real income, a reduction in poverty (excluding the Covid-19 pandemic in 2020), and an alignment of import-export activities with GDP, reflecting its integra- tion into global value chains. Improvements in the labor market have accompanied this economic progress, including lower unemployment rates and increased female workforce participation. Fig- ure 1 encapsulates Viet Nam’s Poverty Rate Reduction driven by its export focus. Figure 3 in the Appendix provides more trends in Trade, Labor, and other Socioeconomic Indicators of Viet Nam for a side-by-side visual comparison. 3 Figure 1: Trend in Poverty Rate in Viet Nam Source: World Bank staff calculations Figure 2: Domestic Service Sector Share in Total Output of Domestic Non-Service Sector in Viet Nam (2020) Source: World Bank staff calculation with Asian Development Bank data. 4 Second, the prominence of domestic services within Viet Nam’s non-tradable sectors is notable, representing an average of 10 percent of total output. The connection between non-service and ser- vice sectors is vital for amplifying the impact of exports on labor markets. In Viet Nam, domestic non-service industries heavily rely on domestic services as inputs. Figure 2 illustrates that around 50 percent of Vietnamese non-service sectors use local services, making up more than 15 percent of their end output. Neglecting these indirect export effects on domestic services and input supply sectors overlooks a crucial link tying local labor markets to foreign demand changes. Recognizing and understanding these connections is crucial for grasping the wider effects of export-oriented economic activities in Viet Nam. A significant proportion of non-service sector-country clusters exhibit a substantial reliance on local services, with approximately 50 percent showcasing a local services usage share exceeding 8 percent, and a minority displaying shares as high as 20 percent. Against this background, this study aims to examine the direct and indirect effects of exports, which are supported by supply chain connections, on the labor market. Third, Viet Nam heavily relies on the advantageous demographic structure of its labor force to propel economic advancement, characterized by a great proportion of the popula- tion being of working age as well as a significant spatial variation across sectoral specialization. Therefore, this investigation aims to leverage these variances to scrutinize how the potential effects vary across diverse demographic strata within the labor force: urban, rural, youthful, highly-skilled, and female cohorts, among others. We find that both direct and indirect exposure to exports has a significant impact on labor mar- ket outcomes, especially for those with no to little education and in the lowest income bracket. The wage premium for attending college decreases, and the gender wage gap narrows. With respect to employment variables, direct exposure to exports increases employment and reduces inactivity, with findings remaining consistent when accounting for supply chain linkages. The gains in em- ployment concentrate among workers with no schooling, while the employment rate falls for more skilled workers. 5 In this paper, we estimate the total impact of changes in exports driven by foreign demand shocks, rather than tariff changes or import competition, on labor market outcomes including in- come and employment variables. This direct and indirect export-induced demand has been studied in Goutam et al. (2017) [5]; however, as only employment variables are in focus, many other questions regarding wages and heterogeneity have been left unanswered. The paper proceeds as follows. Section 2 presents a conceptual framework that allows us to examine the local labor market repercussions of exports, incorporating a supply chain perspective. Section 3 discusses the data used and how we constructed the export exposure analysis using Input- Output linkages. Section 4 describes the empirical strategy and Section 5 presents the empirical results. Section 6 concludes. 2 Conceptual Framework We apply a standard shift-share approach that assesses the effects of trade shocks on labor markets. Relevant works in the literature include Autor, Dorn and Hanson (2013) and Dix-Carneiro and Kovak (2015) [3][2]. More directly related to exploring the effects of exports on local labor markets oes, Lopez-Acevedo, and Robertson (2023) [8] [4]. are studies by Robertson et. al. (2021) and G´ Unlike the papers above, however, our index in this paper is not one of export exposure, but one of total export receipts exposure. By exploiting the input-output structure of production, we account for both direct and indirect payments to factors of production, and use trade data to move closer to regional production, which is what we would ideally like to observe. Let s, d be industry index (s stands for source sector and d stands for destination sector), and let γs,d denote the intermediate use shares of a good of industry s in the production process of a good of industry d . Under the assumptions of perfect competitive or monopolistic competitive product markets, a constant fraction of total sales will be paid to the factors of production. If domestic factor markets are competitive, there are no mark-ups or mark-downs on factor prices. 6 Under those assumptions, then, up to a first-order approximation, the value of export sales can be distributed through the production network in the following fashion: Pd Qd ∝ ∑ γs,d Ps Qs + VAd Value of Export of sector d s Value added of sector d Value of intermediate use of sector d Therefore, we can account for total payments to each source sector s by summing over payments to sector s from every sector d in addition to the value added of sector s: ∆Xs,t +h ∝ (∑ γs,d Ps,t +h Qs,t +h − ∑ γs,d Ps,t Qs,t ) + ∆VAs,t +h d d or in words, Total Export Exposure ∝ Indirect Export Exposure + Direct Export Exposure So far, we have defined these relationships in terms of input-output linkages. To turn to the empir- ical effects on labor markets, we now define local labor markets exposure to total export receipts. Let r denote different regions in the country, exposure to total export receipts growth at regional level is defined as: Lr,s,t Lr,s,t ∆Xr,s,t +h ≡ ∑ ≡∑ ⋅ ∆Xs,t +h s Ls,t s ∑r Lr,s,t where Xr,s,t denotes total export exposure of industry s to region r at period t , as defined above; Lr,s,t Lr,s,t denotes total employment of industry s in region r at time t . The term Ls,t measures the share of region r in the national employment of industry s. 7 3 Data The goal of the paper is to assess the direct and indirect effects of export expansion on local labor market outcomes in Viet Nam while accounting for supply chain linkages. To do this, we exploit variation in export expansion across provinces and industries between 2010 and 2019 and combine export data from the United Nations Commodity Trade Statistics (UNCOMTRADE) data, input- output coefficient matrix from Global Trade Analysis Project (GTAP) data, and information on local labor market outcomes from Viet Nam’s Labor Force Survey (LFS) data. Details on each dataset and cleaning techniques are described below. 3.1 Labor Force Data Our main source of labor market data is the LFS provided by General Statistics office of Viet Nam (GSO) between 2010 and 2019, a period during which it was implemented every year. The LFS observations collect information in a host of areas including key labor market, household, and individual demographic characteristics. Our analysis looks at two main sets of outcome variables: wage outcomes and employment outcomes. The wage outcome data sets include real annual wages, real annual income, college degree wage premium and gender wage premium. The employment outcome sets include em- ployment rate, inactive rate, informality status, and female labor force participation. All of the outcomes are constructed from survey questionnaires, of which the wage outcomes are calculated at the province x sector level, while the employment outcomes are aggregated at the province level because we do not have sector information for those who are not employed. Over the period 2010 to 2019, several changes were introduced in the Viet Namese LFS, together with updates in concepts and definitions. These have been standardized to make key labor market outcomes, administrative geographies, as well as industry classifications, comparable over time. 8 3.2 Construction of Export Exposure Using Input-Output Linkages Any changes in the foreign export demand for products of a particular sector will have dual effects. First, it will lead to a direct increase in demand for output in that sector. Secondly, it indirectly affects the upstream sectors that supply inputs to the directly impacted sector. Not accounting for these linkages will underestimate the export exposure at the province level, as some provinces may not have a concentration of industries directly exporting but still be supplying to exporting sectors. To account for these value chain linkages, the literature represents uses Leontief inverse of an input-output production matrix for an economy. The method clearly tracks the use of intermediate inputs by each sector (Goutam et al. 2017; Acemoglu et al. 2016; Acemoglu et al. 2012). To explore potential effects of exports through domestic inputs, we employ the 2011 Viet Nam Input-Output table to calculate the input shares of each industry. These shares are determined by dividing the input usage by the gross output (which includes the value added in the own sector with own-sector inputs). We then multiply the resulting shares by the exports of the final sector aggre- gated over the input industry to obtain the total value of exports for each input sector (representing the cumulative effect of servicing multiple exporting sectors). In this sense, non-traded sectors that are assigned a value “zero” for exports will also have an implied value and will be used to estimate the total export exposure index at the province level using the following index. The total export exposure (accounting for supply chain linkages) is measured as the growth in exports in industry i between time periods, t and t + 1, captured by the term ∆Wi,t +h = Wi,t − Wi,t +h . This change is allocated to each province r in Viet Nam using the share of provinces in total national employment in each industry i. Li,r,t Li,r,t ∆Xr,t +h = ∑ ∆Wi,t +h = ∑ ∆Wi,t +h i Li,t i ∑r Li,r,t To construct the total exposure index at the province level in Viet Nam, we utilize several databases. Initially, we gather data on export value from the UNCOMTRADE database. To account for de- 9 mand generated in other sectors as a result of exports, and thus calculate the overall exposure index, we incorporate the 2011 input-output (I-O) GTAP tables. We chose 2011 I-O table instead of a more recent year available (2016) because it reflects the economy at the beginning of the stud- ied period, which guarantees that all the shocks computed are ex-ante. An important underlying assumption with this decision is that there isn’t any significant changes in the sectoral structure. Admittedly, this is a major limit of the paper that we have to accept due to data constraint.1 We begin by computing the input-output coefficients from the GTAP I-O tables, which capture the interdependencies between sectors in an economy. We match these coefficients with trade data the UNCOMTRADE data to compute the total export value for each sector, accounting for indirect changes in export demand through input-output linkages. The next step is to link these total export data with the LFSs. We utilize concordance tables available from UNSD that translate International Standard classification (ISIC) rev 3.1. codes into HS codes. By leveraging this concordance, we merge the micro-data on labor force variables at the industry and area level in Viet Nam with total export data. Once the integrated labor and trade data is prepared, we are able to calculate the total trade exposure index based on provinces, as previously explained. The starting point for the analysis is the idea that the impact of a trade shock differs across regions, depending on each province-industry composition. A fundamental principle for this approach is the existence of segmented labor markets. Existing labor mobility barriers or rigidities (such as commuting costs or lack of transport infrastructure) allow us to observe variations in local labor market outcomes and, as a result, to estimate the effects of differentiated exposure to trade. One heuristic method for assessing labor-market inte- gration involves examining the standard deviation of wages across regions and over time. This heuristic measure is used because various factors can prevent wage equalization across regions. To investigate the level of labor-market integration in Viet Nam, we calculate province and industry- province premiums, the existence of which can indicate segmented labor markets. Tables 1 and 2 1 In future exercises, we will attempt to test this assumption using a later I-O table 10 in the Appendix clearly show that wages are not equal across provinces and industry-provinces in Viet Nam, providing strong support for the existence of segmented labor markets during our study period. 4 Identification The goal of our empirical strategy is to understand how rising export expansion affects real wages, informality, and female labor force participation, exploiting data on cross-regional exposure to total exports in Viet Nam between 2010 and 2019. To this effect, we consider the following simple linear regression model: ∆Yr,t +h = β0 + β1 ∆Xr,t +h + β2 Kr,t + εr,t where ∆Yr,t +h is the change in outcomes of interest, may it be employment rate, informality rate, female participation rate, average annual income average annual wage, college premium or gender wage gap, among others, identified at province r over the period from time t to t + h. ∆Xr,t +h is our main independent variable, which stands for the change at regional level of total export exposure, as defined in the previous section. The key coefficient of interest is β1 , which measures the effects of total trade exposure on the outcome after accounting for the I-O structure, Kr,t is the vector of ex-ante control variables including individual demographic background taken from the LFS such as urban dummy, gender, marital status, age group, education level, social security ownership, among others. A relevant issue that needs to be addressed is potential endogeneity in the export exposure covariate. Since we observe changes in labor outcomes and exports simultaneously, we cannot identify which one is driving the other. To ensure truly exogeneity of our export exposure, we need a variable that predicts exports from Viet Nam based solely on its trading partners internal 11 demand growth, rather than supply-side determinants. Hence, we construct our instrument using time-series regressions of Viet Nam exports to its trading partners on the trading partner’s GDP by industry at the four-digit level as follows: Li,r,t Q j,i,t ∆Zr,t +h ≡ ∑ ⋅∑( ⋅ ∆Y j,t +h ) i Li,t j Qi,t Q j,i,t Q j,i,t where, Qi,t = ∑ j Q j,i,t denotes country j’s share of industry i’ export; ∆Y j,t +h is the change in real GDP in destination country j. Predicted values or exports from these regressions serve as a proxy for Viet Nam’s exports to its trading partners explained exclusively by the export market’s domestic aggregate demand. These predicted exports combine with I-O coefficients to generate total exports accounting for supply chain linkages. Subsequently, we use these total exports to generate provincial export exposure in Viet Nam. Then, estimation will take the form of two-stage least squares, with the first stage being: ∆Xr,t +h = α ˜ Kr,t + ε ˜ Zr,t +h + δ ˜ + β∆ ˜ r,t and the second stage: ˆr,t +h + β2 Kr,t + εr,t ∆Yr,t +h = β0 + β1 ∆X ˆr,t +h is the predicted value obtained from the first stage regression: where ∆X ˆr,t +h = α ∆X ˆ Kr,t ˆ Zr,t +h + δ ˆ + β∆ 12 5 Results 5.1 Impact of Exports on Wages Table 6 presents the outcomes of the two-stage least squares regression, detailing the relationship between changes in income-related variables—wages, income, gender wage differentials, and the college wage premium—and shifts in exposure to exports, instrumented by alterations in exposure to foreign demand. All models incorporate standard errors clustered at the province level and con- trol for various socio-demographic factors such as age, gender, education level, urban-rural status, economic sector, and hours of work. It is important to note that although Wages are measured annually, we have included Hours of Labor (Weekly average) in the regression to control for the amount of time the workers spend at work. Thus, the results can also reflect an effect that is sim- ilar to the effect on hourly wage. Another detail that is worth noting is that the Income variable often reflects a more positive improvement compared to the Wage variable for individuals, as it encompasses additional sources of non-wage income such as bonuses, dividends, and personal gifts. Overall, all measures of exposure—Total Exposure, Direct Exposure, and Indirect Exposure—exhibit statistically significant effects on income-related variables. Direct Exposure demonstrates the most substantial improvement, with a US$32.5 increase in annual wages and a US$36.31 increase in an- nual income for every US$1,000 rise in annual Direct Exposure per worker. While Indirect Expo- sure yields a similar effect on wages (an increase of US$31.14 per unit of exposure), its influence on income is relatively lower (an increase of US$23.22 per unit of exposure) compared to Direct Exposure. Additionally, variables reflecting labor market inequality—namely, the college wage premium and gender wage gap—decrease in response to export exposure, with the college premium exhibit- ing a higher degree of significance. Specifically, a US$1,000 increase in annual direct exposure per worker results in about a US$28 reduction in return on attending college. The effect of indirect 13 exposure on this premium is a reduction of about US$30.87. To give a benchmark, the national mean annual wage throughout the 2010-2019 period is US $1,654, meaning that an effect of US $30 is equivalent to approximately 1.81 percent. Compared to the average wage of US $1,867.5 in the urban area and US $1,433.8, this effect translates to 1.6 percent and 1.8 percent respectively. A noteworthy observation across various outcomes in the analysis is that Total Exposure, which represents the sum of Direct and Indirect Exposure, consistently yields smaller absolute effects compared to either Direct or Indirect Exposure. This could look puzzling at first glance, however, there are a few econometric reason why this is the case. First, one reason could be that the controls could absorb differently the effects on labor outcomes that are not caused by direct and indirect exposure. It’s likely that direct exposure is more correlated with some of the controls than indirect exposure or vice versa. For example, it could be that workers in tradable sectors tend to be younger and less educated. Furthermore, there could also be larger error terms in the indirect exposure regression, and indeed this may be why we see much less significance for indirect exposure. 5.2 Impact of Exports on Employment Regarding employment outcomes, Table 7 illustrates that both Total Exposure and Indirect Expo- sure exert statistically significant effects on the employment rate at the provincial level. Specif- ically, for every US$1000 increase in Total Exposure per worker, the likelihood of employment rises by 0.2 percent point, while Indirect Exposure increases the employment rate by 0.52 per- cent point. Interestingly, Direct Exposure exhibits either non- or minimally significant effects on the employment rate and other employment metrics such as the rate of informal employment and female employment rate. Conversely, both Total Exposure and Indirect Exposure significantly re- duce informality rates and augment female participation in the labor force. All of these trends imply that exports have caused healthy benefits to the labor market and created a more efficient and balanced working environment for workers. 14 5.3 Heterogeneity by Education Tables 10 and 11 provide insights into how wage and employment outcomes vary across different segments of the population based on their level of education. Our findings suggest that workers across all educational strata benefit from trade. Notably, individuals with a college education pre- mium derive the greatest absolute benefit, with a US$42.38 increase in wages for every US$1,000 rise in Total Exposure, compared to a US$14.84 gain for workers with no formal education. How- ever, as shown in Table 3, mean annual wages increase considerably as education levels increase. Thus, in a relative terms, these effect among workers with a college degree translates to 1.23 per- cent, lower than that of 1.55 percent among workers with no education. This disparity becomes more pronounced when examining the impact of Direct Exposure and Indirect Exposure on workers with a college degree, yielding increases of US$67.11 and US$72.13, respectively, for each type of exposure. Interestingly, the indirect effect is only significant for less educated labor while the direct effect is most significant for more educated labor. This suggests a higher concentration of skilled workers in exporting sectors and unskilled workers in non-exporting sectors. Yet generally, the absolute discrepancy between changes in wages resulting from Direct and Indirect exposure is minimal across all educational levels. On the contrary, changes in employment outcomes exhibit divergent patterns across different educational groups. While the employment rate among workers with little to no formal education (primary school level) tends to increase, that of college graduates decreases in response to higher export exposure, whether direct or indirect. For instance, a US$1000 increase in annual Total Expo- sure per worker can elevate the employment rate of workers with no formal education by 0.15 percent point while concurrently reducing the likelihood of employment for college graduates by 0.19 percent point. This outcome, coupled with the discussed variance of education on wages, suggests a nuanced interpretation. Specifically, it implies a more competitive environment for highly-skilled workers, juxtaposed with an expansion in economic opportunities for low-skilled 15 workers following an export-driven boost. 5.4 Heterogeneity by Economic Sector Tables 12 and 13 provide a detailed examination of wage and employment outcomes across various economic sectors, namely Household Farm, Household Business, Private Sector, State Agency, and Foreign Sector. While employment demonstrates significant improvement across all sectors, Household Farm experiences the most substantial gains, with approximately a 4.87 percent point increase for every US$1000 rise in total exposure, followed by Household Business, Private sector and State Agency with 4.12, 4.05 and 3.78 percentage points, respectively. In terms of wages, the findings also reveal notable wage increases in most sectors except Foreign sector, regardless of the type of exposure. Particularly, Household Business witnesses a sharper rise in wages, amounting to approximately US$53.26 per worker for every unit increase in Total Exposure, compared to a US$32.8 increase in wage observed in the Household Farm sector, US$29.6 in Private sector and US$37.6 in State Agency. The reason why no significant increase in wages was observed in Foreign Sector in spite of an increase in employment in the same sector probably roots in corporate culture where multinational employers tend to anchor their pay rates around homeland culture and regulations. 5.5 Heterogeneity by Income Level To gain a comprehensive understanding of how trade influences income inequality within the labor market, we extend our analysis to examine various income quantiles among workers. Tables 14 and 15 present the outcomes of these regressions. Our findings indicate that the most significant wage improvements occur among workers in the lowest income quantile, which aligns with the observation that the sectors most affected by trade tend to predominantly employ labor from this demographic segment. 16 Conversely, we observe enhanced employment opportunities across all income quantiles, albeit with diminishing effects as income levels rise. Specifically, the likelihood of employment increases the most among workers in the lowest income quantile, with an about 1.77 percent point rise for every US$1,000 increase in Total Exposure. This effect gradually diminishes across the second and third lowest quantiles, reaching as low as a 0.14 percent point increase in employment probability in the highest worker income bracket. 5.6 Heterogeneity by Tradability In our final analysis, we speculate about the spillover of direct trade exposure into indirect ex- posure, a phenomenon closely intertwined with the tradability nature of sectors. In this section, we explore the divergent effects of direct and indirect exposure on labor outcomes between trad- able and non-tradable sectors. We classify industries into “non-tradable” sectors if their export exposure is “zero”, as indicated by the summary statistics in Table 4. These sectors include Min- ing Extraction, Chemicals, Pharmaceuticals & Medical Products, Rubber and Plastics, Electrical Equipment, Electricity, Gas Manufacturing and Distribution, Water Supply, Construction, Whole- sale and Retail Trade, Accommodation and Food Service, Land Pipelines and Transportation, Wa- ter Transportation, Air Transportation, Warehousing, Other Financial Inter-mediation, Insurance, Real Estate Activities, Other Government Services, Education, and Human Health & Social Work. Broadly, these sectors are characterized by their focus on supplementary goods and domestic ser- vices. Conversely, we designate sectors with “non-zero” export exposure as “tradable”. Having established this differentiation based on tradability, we conduct regression analysis on samples restricted to these two categories. This approach enables us to derive insightful interpreta- tions regarding how direct and indirect exposure to exports affect labor outcomes. As anticipated, Table 16 indicates that direct exposure exerts a notably pronounced and robust impact on both wage and employment outcomes within tradable sectors, whereas indirect exposure demonstrates 17 a more modest effect on wages while still fostering employment growth and reducing informality. This observation aligns logically with the understanding that tradable sectors, in addition to pro- ducing goods for direct export, may also provide inputs for other traded sectors, thereby benefiting from indirect exposure. Table 17 initially presents a perplexing observation, revealing an unexpected trend wherein both direct and indirect exposure reduces employment and wages within non-tradable sectors, with indirect exposure exhibiting a stronger negative effect. However, upon closer examination, this finding unveils contrasting patterns regarding how trade influences these two distinct sectors. On one hand, trade has the potential to expand the overall economic “pie”, thereby reshaping wage and employment dynamics in both tradable and non-tradable sectors. Conversely, an intriguing offsetting effect, wherein trade acts as a driving force, can pull labor directly from non-tradable to tradable sectors within the confines of fixed labor supply. This second nuance might have a greater significance given the ample evidence of relatively full employment in Viet Nam and the limited benefits or unemployment insurance available, rendering staying in non-tradable sectors a far less attractive option compared to seeking opportunities in rapidly expanding tradable sectors. 6 Conclusion The intrinsic connection between direct exporting sectors and the indirect supplying sectors (that furnish production inputs to exporters), is crucial for transmitting the benefits of foreign demand- driven exports not only to tradable sectors, such as agriculture and manufacturing, but also to non-tradable sectors, such as services, where such effects are less anticipated. Leveraging this relationship within the supply chain, we have devised a theoretical framework and an empirical approach to re-examine the ramifications of trade on the labor market, incorporating considerations of both direct and indirect export exposure. Our analysis provides comprehensive insights into the impact of trade on various aspects of 18 the labor market, shedding light on nuanced patterns across different demographic and economic segments. The results indicate significant labor benefits from exposure to exports on income- related variables such as wages, income, and labor market inequality, with directly exposed sectors and provinces enjoying the greatest benefit. Moreover, both exposure types contribute to a de- crease in labor market inequality measures such as the college wage premium and the gender wage gap. Employment out- comes also vary, with both total and indirect exposure positively affecting employment, informality, and female employment rates, while direct and total exposure reduce informality rates and increase female labor force participation. Across education levels, workers benefit from an increase in trade, with college-educated in- dividuals deriving the greatest wage benefits. Changes in employment outcomes vary, with em- ployment increasing for workers with lower education but decreasing for college graduates in re- sponse to higher export exposure. Employment improves across all economic sectors, with the Household Farm sector experiencing the most substantial gains, while notable wage increases are observed only in the Household Farm and Household Business sectors, suggesting differential im- pacts across economic sectors. Wage improvements are most significant among workers in the lowest income quantile, while enhanced employment opportunities are observed across all income levels, albeit with diminishing effects as income rises. Direct exposure causes a pronounced in- crease on both wage and employment outcomes within tradable sectors, while indirect exposure exhibits a more modest increase in wages but still fosters employment growth. In contrast, both di- rect and indirect exposure reduce employment and wages within non-tradable sectors, highlighting the intricate interplay between trade dynamics and labor markets. Considering both direct and indirect export exposure leads to a more nuanced understanding of how trade shapes employment, wage dynamics, and labor market inequality across various sectors and demographic groups. These insights are crucial for policymakers and stakeholders seeking to navigate the complexities of globalization and ensure inclusive economic growth. 19 References [1] Daron Acemoglu et al. “Import Competition and the Great US Employment Sag of the 2000s”. In: The University of Chicago Press Journal (2016). URL: https://www.journals. uchicago.edu/doi/full/10.1086/682384. [2] David H. Autor, David Dorn, and Gordon H. Hanson. “The China Syndrome: Local Labor Market Effects of Import Competition in the United States”. In: American Economic Review (2013). URL: https://www.aeaweb.org/articles?id=10.1257/aer.103.6.2121. [3] Rafael Dix-Carneiro and Brian K. Kovak. “Trade Liberalization and the Skill Premium: A Local Labor Markets Approach”. In: American Economic Review (2015). URL: https: //www.aeaweb.org/articles?id=10.1257/aer.p20151052. oes, Gladys Lopez-Acevedo, and Raymond Robertson. “Gender-Segmented Labor [4] Carlos G´ Markets and Trade Shocks”. In: IZA Discussion Papers 15892, Institute of Labor Economics (IZA) (2023). URL: https://ideas.repec.org/p/iza/izadps/dp15892.html. [5] Prodyumna Goutam et al. “Does Informal Employment Respond to Growth Opportunities? Trade-Based Evidence from Bangladesh”. In: RAND Corporation. Working Paper (2017). URL : https://www.rand.org/content/dam/rand/pubs/working_papers/WR1100/ WR1198/RAND_WR1198.pdf. [6] Brian McCaig and Nina Pavcnik. “Export Markets and Labor Allocation in a LowIncome Country”. In: American Economic Reviewn. 108 (7): 1899–941 (2018). URL: https : / / www.aeaweb.org/articles?id=10.1257/aer.20141096. [7] Justin R. Pierce and Peter K. Schott. “The Surprisingly Swift Decline of US Manufacturing Employment”. In: American Economic Review (2016). URL: https://www.aeaweb.org/ articles?id=10.1257/aer.20131578. 20 [8] Raymond Robertson et al. “International Trade and Labor Markets: Evidence from the Arab Republic of Egypt”. In: IZA Discussion Paper No. 14413, Institute of Labor Economics (IZA) (2021). URL: https://ssrn.com/abstract=3860594. [9] Petia Topalova. “Factor Immobility and Regional Impacts of Trade Liberalization: Evidence on Poverty from India”. In: American Economic Journal: Applied Economics 2 (2010). URL: https://www.aeaweb.org/articles?id=10.1257/app.2.4.1. [10] Zhi Wang et al. “Reexamining the Effects of Trading with China on Local Labor Markets: A Supply Chain Perspective.” In: National Bureau of Economic Research. Working Paper 24886 (2018). URL: https://www.nber.org/papers/w24886. 21 7 Appendix 22 Figure 3: Trends in Trade, Labor and Socioeconomic Indicators in Viet Nam (2001-2020) Source: World Bank staff calculations and World Development Indicators. 23 Table 1: Summary Statistics - Wage by Industry Annual Wage (USD) Mean SD Max Veg & Fruit 2190.6 2473.4 208472.7 Cattle 994.6 1240.0 30110.4 Other Animal Products 703.8 1582.5 136806.2 Forestry 1237.4 1253.2 34336.4 Fishing 2156.0 5402.4 488124.5 Coal 2166.8 2276.8 178792.3 Oil 3639.6 4216.8 78177.3 Gas 3637.2 3574.2 22994.9 Vegetable Oils 2452.2 1451.1 8860.1 Milk 3476.3 3259.3 56608.8 Sugar 2320.5 1238.7 16982.7 Other Food 1838.8 2421.2 206651.2 Beverages and Tobacco products 2674.6 2524.1 65147.7 Textiles 2186.9 2215.9 104535.3 Wearing Apparel 2116.8 1779.8 329155.7 Leather 2416.3 1469.5 78181.3 Lumber 1712.1 1942.6 77494.2 Paper 2321.6 2065.0 43778.0 Petroleum & Coke 2833.0 2969.4 51662.8 Non-Metallic Minerals 2322.9 2118.4 110933.0 Iron & Steel 2894.9 4050.3 191528.7 Non-Ferrous Metals 2817.1 3114.8 52825.2 continued . . . 24 . . . continued Annual Wage (USD) Mean SD Max Fabricated Metal Products 2653.6 2078.0 36482.7 Motor vehicles and parts 2997.1 1733.0 26059.1 Other Transport Equipment 2789.5 1742.2 27361.2 Electronic Equipment 3008.2 1566.1 39618.9 Other Machinery & Equipment 2863.6 1858.9 27908.5 Other Manufacturing 2420.0 2197.0 154988.4 Electricity 3151.2 1985.9 54722.5 Gas Distribution 4065.2 6767.9 65667.0 Water 2313.1 1938.1 36482.7 Construction 2379.5 2098.7 438983.7 Trade 2303.6 3439.3 470131.5 Other Transport 2715.4 2267.6 71294.7 Water transport 3735.7 4735.5 82660.5 Air transport 5041.3 3968.1 46906.4 Communications 3022.8 2148.4 52825.2 Other Financial Intermediation 3698.3 2521.4 77494.2 Other Business Services 2936.9 3025.5 309976.8 Recreation & Other Services 1856.4 1804.6 57732.2 Other Services (Government) 2638.2 2089.5 438875.9 continued . . . 25 Table 2: Summary Statistics - Wage by Province Annual Wage (USD) Mean SD Max Hanoi 2265.1 2787.9 437779.8 Ha Giang 1157.5 1632.7 42260.2 Cao Bang 1030.9 1469.3 42788.4 Bac Kan 1114.3 2384.2 438875.9 Tuyen Quang 1200.3 2571.5 438173.8 Lao Cai 1467.1 2056.8 156354.5 Dien Bien 1222.4 1704.9 50948.0 Lai Chau 1176.4 1701.1 47542.7 San La 1266.9 2072.2 206651.2 Yen Bai 1238.5 2561.0 438874.2 Hoa Binh 1501.1 3232.6 456932.7 Thai Nguyen 1535.7 1818.0 191528.7 Lang Sn 1393.8 2741.8 451460.4 Quang Ninh 1983.2 2136.9 65667.0 Bac Giang 1665.3 1701.6 52825.2 Phu Tho 1360.3 1517.3 37194.3 Vinh Phuc 1734.8 2186.2 82660.5 Bac Ninh 2071.9 2299.8 156354.5 Hai Duong 1779.6 1788.6 54245.9 Hai Phong 1956.2 2129.7 61995.4 Hung Yen 1807.0 2983.4 278979.1 Thai Binh 1512.0 1661.5 57851.2 continued . . . 26 . . . continued Annual Wage (USD) Mean SD Max Ha Nam 1518.7 1574.2 43778.0 Nam Dinh 1559.1 1865.5 59412.2 Ninh Binh 1541.4 1711.5 54245.9 Thanh Hoa 1473.9 2194.4 244955.4 Nghe An 1518.2 2493.6 330523.8 Ha Tinh 1371.8 1689.7 52118.2 Quang Binh 1451.1 2328.7 330250.1 Quang Tri 1511.8 1983.9 188202.5 Thua Thien Hue 1545.3 1680.7 62541.8 Da Nang City 2195.1 2470.1 114660.0 Quang Nam 1465.6 1733.0 79237.8 Quang Ngai 1470.0 1737.8 47542.7 Binh Dinh 1606.0 2155.0 110933.0 Phu Yen 1431.8 2604.0 438983.7 Khanh Hoa 1662.0 2355.7 331618.2 Ninh Thuan 1458.7 1740.9 78177.3 Binh Thuan 1754.6 2142.4 78177.3 Kon Tum 1521.5 2317.4 329155.7 Gia Lai 1320.1 1493.9 35130.7 DakLak 1345.3 2065.3 154988.4 Dak Nong 1531.5 2150.2 114974.7 Lam Dong 1705.3 2852.3 329429.3 Binh Phuoc 1786.5 2234.7 75281.0 continued . . . 27 . . . continued Annual Wage (USD) Mean SD Max Tay Ninh 1719.2 2142.9 208472.7 Binh Duong 2144.4 3369.9 470131.5 Dong Nai 2093.4 2145.9 52825.2 Ba Ria - Vung Tau 1951.4 2523.7 180819.8 Ho Chi Minh City 2364.0 2699.0 178792.3 Long An 1939.9 2907.5 209039.9 Tien Giang 1805.6 2821.0 180819.8 Ben Tre 1624.5 5523.1 469063.6 Tra Vinh 1389.5 1882.7 115181.2 Vinh Long 1348.5 1714.4 77494.2 Dong Thap 1632.3 2555.7 154988.4 An Giang 1559.3 1971.8 67161.6 Kien Giang 1783.7 2756.5 309976.8 Can Tho 1676.6 3515.8 488124.5 Hau Giang 1407.5 1712.6 53160.5 Soc Trang 1412.7 2263.3 97565.2 Bac Lieu 1774.0 3151.3 329976.5 Ca Mau 1553.3 2613.8 158695.2 continued . . . 28 Table 3: Summary Statistics - Wage by Education Wage (USD) Mean SD Max No education 955.1 1591.2 390886.3 Primary school 1387.0 2197.9 438983.7 Secondary 2077.2 2783.9 469063.6 High school 2023.7 2795.9 488124.5 College 2136.1 2130.6 309976.8 University & over 3457.4 2989.7 451460.4 Table 4: Summary Statistics - Export Exposure by Industry Annual Change Export Exposure (million USD) Export Total Direct Indirect Exposure Exposure Exposure Exposure Veg & Fruit 43.98 225.58 40.99 184.58 Cattle 4.45 9.12 3.56 5.56 Other Animal Products 4.95 5.89 3.39 2.50 Forestry 2.53 35.16 2.31 32.85 Fishing -17.99 447.85 11.69 436.16 Coal -0.09 21.68 -0.08 21.76 Oil 0.01 0.18 0.01 0.17 Gas 0.00 0.00 0.00 0.00 Other Mining Extraction 0.00 0.00 0.00 0.00 Vegetable Oils -2.13 -1.78 -1.78 0.01 Milk 4.45 3.04 2.99 0.05 continued . . . 29 . . . continued Annual Change Export Exposure (million USD) Export Total Direct Indirect Exposure Exposure Exposure Exposure Sugar and molasses -2.56 -1.25 -1.42 0.17 Other Food 30.65 196.60 192.92 3.68 Beverages and Tobacco 0.03 0.56 0.29 0.27 Textiles 5.56 28.06 4.35 23.72 Wearing apparel 53.35 46.36 39.28 7.08 Leather 36.49 1584.48 1580.67 3.81 Lumber 11.35 7.38 5.49 1.89 Paper 73.12 190.41 170.32 20.09 Petroleum & Coke -0.02 0.27 -0.01 0.28 Chemicals 0.00 0.00 0.00 0.00 Pharmaceuticals & medicinal 0.00 0.00 0.00 0.00 Rubber and plastics 0.00 0.00 0.00 0.00 Other non-metallic mineral 2.09 10.26 1.51 8.75 Iron & Steel 2.44 0.54 0.49 0.05 Non-Ferrous Metals 0.92 1.12 0.57 0.55 Fabricated metal 12.81 11.13 10.81 0.32 Computer and electronic -147.13 -97.61 -122.86 25.24 Electrical equipment 0.00 0.00 0.00 0.00 Machinery and equipment 12.50 10.19 9.94 0.25 Motor vehicles 0.08 2.37 0.07 2.30 Other transport equipment 1.31 6.25 1.14 5.10 Other Manufacturing 69.10 1432.10 1318.29 113.81 continued . . . 30 . . . continued Annual Change Export Exposure (million USD) Export Total Direct Indirect Exposure Exposure Exposure Exposure Electricity 0.00 40.24 0.00 40.24 Gas manufacture, distribution 0.00 0.04 0.00 0.04 Water supply 0.00 0.59 0.00 0.59 Construction 0.00 0.05 0.00 0.05 Wholesale and retail trade 0.00 546.83 0.00 546.83 Accommodation and Food Service 0.00 0.00 0.00 0.00 Land pipelines and transport 0.00 15.44 0.00 15.44 Water transport 0.00 0.39 0.00 0.39 Air transport 0.00 0.64 0.00 0.64 Warehousing 0.00 0.00 0.00 0.00 Information and communication 0.06 25.26 12.51 12.76 Other Financial Intermediation 0.00 3.46 0.00 3.46 Insurance 0.00 0.00 0.00 0.00 Real estate activities 0.00 0.00 0.00 0.00 Other Business Services -0.00 29.06 0.12 28.93 Recreation, Other Services -19.66 178.78 178.47 0.32 Other Services (Government) 0.00 10.34 0.00 10.34 Education 0.00 0.00 0.00 0.00 Human health and social work 0.00 0.00 0.00 0.00 continued . . . 31 Table 5: Summary Statistics - Export Exposure by Province Annual Change Export Exposure (million USD) Direct Indirect Export Exposure Total Exposure Exposure Exposure An Giang 500.99 89.24 49.19 31.10 Bac Giang 405.61 55.67 37.59 13.42 Bac Kan 337.79 30.52 16.22 11.27 Bac Lieu 584.02 98.62 22.75 58.19 Bac Ninh 442.86 139.10 106.50 24.30 Ba Ria - Vung Tau 377.51 85.98 50.16 28.11 Ben Tre 437.36 72.32 39.86 23.79 Binh Dinh 407.73 99.50 67.57 23.60 Binh Duong 460.68 310.43 273.29 27.52 Binh Phuoc 464.60 101.85 67.20 25.95 Binh Thuan 442.99 80.47 43.35 27.99 Can Tho 417.10 80.51 45.17 25.71 Cao Bang 323.90 18.86 7.71 8.70 Ca Mau 629.74 110.99 24.15 68.26 Dak Nong 357.52 26.06 13.59 9.73 DakLak 318.88 35.54 18.67 13.05 Da Nang City 378.28 103.32 65.53 27.94 Dien Bien 371.92 20.92 9.84 8.66 Dong Nai 389.74 185.41 158.48 20.42 Dong Thap 435.84 85.31 51.65 24.13 Gia Lai 342.27 30.51 16.43 10.80 continued . . . 32 . . . continued Annual Change Export Exposure (million USD) Direct Indirect Export Exposure Total Exposure Exposure Exposure Ha Giang 368.58 18.39 9.91 6.79 Hai Duong 431.11 95.74 70.50 18.25 Hai Phong 427.47 131.39 100.84 22.97 Hanoi 354.86 90.02 59.74 23.24 Hau Giang 471.94 65.50 40.07 19.53 Ha Nam 503.85 111.70 82.51 21.32 Ha Tinh 390.94 59.86 32.64 20.42 Ho Chi Minh City 385.15 110.53 77.05 25.94 Hoa Binh 346.89 28.73 15.52 9.76 Hung Yen 450.38 84.07 58.92 18.82 Khanh Hoa 449.72 75.53 38.64 27.67 Kien Giang 508.43 81.99 31.10 39.33 Kon Tum 368.79 39.54 19.90 15.38 Lai Chau 430.70 23.26 12.57 8.35 Lam Dong 363.86 47.76 27.40 15.10 Lang Son 389.87 32.58 18.15 10.60 Lao Cai 433.41 28.21 12.54 11.74 Long An 415.50 127.11 101.93 18.95 Nam Dinh 444.56 104.69 73.39 22.37 Nghe An 344.61 40.90 22.76 14.10 Ninh Binh 437.69 69.68 46.19 17.53 Ninh Thuan 530.56 72.38 35.61 27.90 continued . . . 33 . . . continued Annual Change Export Exposure (million USD) Direct Indirect Export Exposure Total Exposure Exposure Exposure Phu Tho 435.20 55.56 34.30 15.87 Phu Yen 401.89 75.45 43.02 24.75 Quang Binh 435.60 52.76 24.18 21.66 Quang Nam 385.30 71.23 46.67 18.49 Quang Ngai 383.37 53.07 28.41 19.69 Quang Ninh 381.45 63.06 20.97 31.31 Quang Tri 457.58 66.60 34.64 24.22 San La 375.94 20.82 9.62 8.86 Soc Trang 503.88 72.29 30.68 32.48 Tay Ninh 452.32 140.53 108.46 23.71 Thai Binh 431.36 67.48 44.45 17.25 Thai Nguyen 380.89 48.62 27.82 16.42 Thanh Hoa 376.16 62.34 43.88 14.16 Thua Thien Hue 409.91 86.51 49.81 27.18 Tien Giang 429.61 110.81 83.40 20.15 Tra Vinh 440.07 103.72 71.46 23.06 Tuyen Quang 394.78 44.28 26.95 13.47 Vinh Long 488.18 98.44 68.94 21.59 Vinh Phuc 444.16 85.49 59.36 19.33 Yen Bai 443.67 40.89 22.68 13.71 continued . . . 34 Table 6: Impact of trade exposure on Wages, 2010 - 2019 Change in labor outcomes WAGE INCOME COLLEGE PREMIUM GENDER WAGE GAP (1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (11) (12) TOTAL 19.86∗∗∗ 19.93∗∗∗ -17.90∗∗∗ -5.79∗ EXPOSURE (5.29) (5.10) (4.76) (2.37) DIRECT 32.50∗∗ 36.31∗∗ -28.01∗∗ -10.65∗ EXPOSURE (11.65) (10.56) (10.21) (5.29) 35 INDIRECT 31.14∗∗ 23.22∗ -30.87∗∗∗ -6.55 EXPOSURE (9.32) (10.34) (8.32) (3.61) Observations 504 504 504 504 504 504 504 504 504 504 504 504 Adjusted R2 0.3 0.3 0.3 0.3 0.3 0.3 0.3 0.3 0.3 0.3 0.3 0.3 *** p<.001, ** p<.05, * p<.01. Standard errors clustered at Province level. Total Exposure = Direct Exposure + Indirect Exposure. Additional controls include Age, Gender, Education level, Urban-Rurual dummy, Economic Sector and Hours of work. Wage and wage premium measured in US dollars per year, Export Exposure measured in thousand US dollar per worker. Unemployed, Inactive, Informal and Female Employed measured in percentage point. Table 7: Impact of trade exposure on Employment, 2010 - 2019 Change in labor outcomes EMPLOYED INACTIVE INFORMAL FEMALE EMPLOYED (1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (11) (12) TOTAL 0.20∗∗ -0.72∗∗∗ -0.43∗∗ 0.14∗∗∗ EXPOSURE (0.06) (0.09) (0.16) (0.04) DIRECT 0.23 -0.91∗∗∗ -0.04 0.15∗ EXPOSURE (0.11) (0.18) (0.22) (0.07) 36 INDIRECT 0.52∗∗∗ -1.71∗∗∗ -2.08∗∗∗ 0.36∗∗∗ EXPOSURE (0.10) (0.17) (0.27) (0.06) Observations 504 504 504 504 504 504 504 504 504 504 504 504 Adjusted R2 0.2 0.1 0.2 0.4 0.3 0.4 0.2 0.2 0.2 0.2 0.2 0.2 *** p<.001, ** p<.05, * p<.01. Standard errors clustered at Province level. Total Exposure = Direct Exposure + Indirect Exposure. Additional controls include Age, Gender, Education level, Urban-Rurual dummy, Economic Sector and Hours of work. Wage and wage premium measured in US dollars per year, Export Exposure measured in thousand US dollar per worker. Unemployed, Inactive, Informal and Female Employed measured in percentage point. Table 8: Impact of trade exposure on Wages, 2010 - 2019 Change in labor outcomes WAGE INCOME COLLEGE PREMIUM GENDER WAGE GAP (1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (11) (12) TOTAL EXPOSURE 26.99 ∗∗∗ 25.74 ∗∗∗ -24.15 ∗∗∗ -8.09 ∗∗ (5.85) (6.65) (5.15) (2.52) HANOI x TOTAL -26.65∗∗ -22.91∗ 22.15∗ 10.39∗ (9.81) (10.44) (8.74) (3.99) HCM x TOTAL 5.37 -0.87 -2.99 -4.96 (16.82) (12.19) (15.36) (7.81) DIRECT EXPOSURE 49.45 ∗∗∗ 54.43∗∗ -42.22∗∗∗ -16.78∗∗ (13.70) (16.68) (11.47) (6.17) HANOI x DIRECT -43.99∗ -44.78∗ 35.86∗ 18.07∗ 37 (17.76) (19.61) (15.30) (7.56) HCM x DIRECT 6.85 -10.81 -4.75 -6.92 (27.40) (22.85) (24.59) (11.90) INDIRECT EXPOSURE 41.26 ∗∗∗ 31.42 ∗∗ -39.29∗∗∗ -10.04∗ (10.72) (10.97) (9.76) (4.12) HANOI x INDIRECT -73.28 ∗ -62.77 58.12∗ 27.75∗∗ (28.82) (31.69) (24.82) (10.18) HCM x INDIRECT -9.07 -7.23 11.88 0.14 (20.15) (17.54) (18.02) (13.42) Observations 504 504 504 504 504 504 504 504 504 504 504 504 Adjusted R2 0.3 0.3 0.3 0.3 0.3 0.3 0.3 0.3 0.3 0.3 0.3 0.3 *** p<.001, ** p<.05, * p<.01. Standard errors clustered at Province level. Total Exposure = Direct Exposure + Indirect Exposure. Additional controls include Age, Gender, Education level, Urban-Rurual dummy, Economic Sector, Hours of work, Dummies for Hanoi Economic Area and HCMC Economic Area. Wage and wage premium measured in US dollars per year, Export Exposure measured in thousand US dollar per worker. Unemployed, Inactive, Informal and Female Employed measured in percentage point. Table 9: Impact of trade exposure on Employment, 2010 - 2019 Change in labor outcomes EMPLOYED INACTIVE INFORMAL FEMALE EMPLOYED (1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (11) (12) TOTAL EXPOSURE 0.31 ∗∗∗ -0.87 ∗∗∗ -0.78 ∗∗∗ 0.20 ∗∗∗ (0.06) (0.08) (0.15) (0.03) HANOI x TOTAL -0.19 0.29 0.97∗∗∗ -0.10 (0.12) (0.18) (0.26) (0.07) HCM x TOTAL -0.44 ∗∗∗ 0.47 ∗∗ 0.14 -0.29∗∗∗ (0.11) (0.15) (0.37) (0.07) DIRECT EXPOSURE 0.51 ∗∗∗ -1.40 ∗∗∗ -0.72 ∗ 0.32∗∗∗ (0.11) (0.18) (0.29) (0.07) HANOI x DIRECT -0.41∗ 0.74∗∗ 1.28∗∗ -0.23∗ 38 (0.17) (0.27) (0.41) (0.11) HCM x DIRECT -0.74 ∗∗∗ 1.01 ∗∗∗ 0.39 -0.48∗∗∗ (0.17) (0.24) (0.65) (0.10) INDIRECT EXPOSURE 0.55 ∗∗∗ -1.61∗∗∗ -2.06 ∗∗∗ 0.37∗∗∗ (0.12) (0.19) (0.31) (0.07) HANOI x INDIRECT 0.15 -0.76∗ 0.57 0.13 (0.25) (0.38) (0.72) (0.18) HCM x INDIRECT -0.79 ∗∗ 0.14 -1.31 -0.48∗ (0.27) (0.27) (0.72) (0.22) Observations 504 504 504 504 504 504 504 504 504 504 504 504 Adjusted R2 0.2 0.2 0.2 0.4 0.3 0.4 0.2 0.2 0.2 0.2 0.2 0.2 *** p<.001, ** p<.05, * p<.01. Standard errors clustered at Province level. Total Exposure = Direct Exposure + Indirect Exposure. Additional controls include Age, Gender, Education level, Urban-Rurual dummy, Economic Sector, Hours of work, Dummies for Hanoi Economic Area and HCMC Economic Area. Wage and wage premium measured in US dollars per year, Export Exposure measured in thousand US dollar per worker. Unemployed, Inactive, Informal and Female Employed measured in percentage point. Table 10: Impact of trade exposure on Wages by Education, 2010 - 2019 Change in labor outcomes NO EDUCATION PRIMARY SCHOOL SECONDARY SCHOOL HIGH SCHOOL COLLEGE (1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (11) (12) (13) (14) (15) TOTAL 14.84∗∗ 17.00∗∗ 19.77∗∗ 16.42∗∗ 42.38∗∗∗ EXPOSURE (4.73) (5.70) (6.59) (4.94) (10.37) DIRECT 22.47∗ 28.03∗ 32.42∗ 31.91∗∗ 67.11∗∗∗ EXPOSURE (8.85) (11.05) (12.43) (9.93) (14.25) 39 INDIRECT 27.30∗∗ 26.38∗ 31.07∗ 15.13 72.13∗ EXPOSURE (9.76) (10.48) (14.51) (9.98) (27.29) Observations 504 504 504 504 504 504 504 504 504 504 504 504 502 502 502 Adjusted R2 0.2 0.2 0.2 0.2 0.2 0.2 0.2 0.2 0.2 0.2 0.2 0.2 0.2 0.2 0.2 *** p<.001, ** p<.05, * p<.01. Standard errors clustered at Province level. Total Exposure = Direct Exposure + Indirect Exposure. Additional controls include Age, Gender, Education level, Urban-Rurual dummy, Economic Sector and Hours of work. Wage and wage premium measured in US dollars per year, Export Exposure measured in thousand US dollar per worker. Unemployed, Inactive, Informal and Female Employed measured in percentage point. Table 11: Impact of trade exposure on Employment by Education, 2010 - 2019 Change in labor outcomes NO EDUCATION PRIMARY SCHOOL SECONDARY SCHOOL HIGH SCHOOL COLLEGE (1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (11) (12) (13) (14) (15) TOTAL 0.15∗∗ 0.12∗ 0.06 0.03 -0.19∗∗∗ EXPOSURE (0.05) (0.05) (0.06) (0.05) (0.02) DIRECT 0.19∗∗ 0.16∗ 0.11 0.01 -0.29∗∗∗ EXPOSURE (0.07) (0.08) (0.09) (0.09) (0.04) 40 INDIRECT 0.37∗∗ 0.24∗ 0.05 0.12 -0.36∗∗∗ EXPOSURE (0.12) (0.10) (0.13) (0.08) (0.06) Observations 504 504 504 504 504 504 504 504 504 504 504 504 504 504 504 Adjusted R2 0.1 0.1 0.1 0.1 0.1 0.1 0.2 0.2 0.2 0.0 0.0 0.0 0.2 0.2 0.2 *** p<.001, ** p<.05, * p<.01. Standard errors clustered at Province level. Total Exposure = Direct Exposure + Indirect Exposure. Additional controls include Age, Gender, Education level, Urban-Rurual dummy, Economic Sector and Hours of work. Wage and wage premium measured in US dollars per year, Export Exposure measured in thousand US dollar per worker. Unemployed, Inactive, Informal and Female Employed measured in percentage point. Table 12: Impact of trade exposure on Wages by Economic Sector, 2010 - 2019 Change in labor outcomes HH Farm HH Business Private State Foreign (1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (11) (12) (13) (14) (15) TOTAL 32.82∗∗∗ 53.26∗∗∗ 29.61∗∗∗ 37.57∗∗ 16.30 EXPOSURE (7.35) (7.47) (5.70) (12.21) (15.63) DIRECT 50.57∗∗∗ 84.82∗∗∗ 53.79∗∗∗ 59.47∗∗ 43.78∗ EXPOSURE (14.52) (12.62) (11.58) (21.03) (17.06) 41 INDIRECT 53.74∗∗∗ 81.99∗∗∗ 29.51∗ 58.76∗ -19.03 EXPOSURE (12.38) (16.26) (12.95) (24.90) (47.32) Observations 504 504 504 480 480 480 418 418 418 430 430 430 328 328 328 Adjusted R2 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.1 0.1 0.0 0.0 0.0 0.0 *** p<.001, ** p<.05, * p<.01. Standard errors clustered at Province level. Total Exposure = Direct Exposure + Indirect Exposure. Additional controls include Age, Gender, Education level, Urban-Rurual dummy and Hours of work. Wage and wage premium measured in US dollars per year, Export Exposure measured in thousand US dollar per worker. Unemployed, Inactive, Informal and Female Employed measured in percentage point. Table 13: Impact of trade exposure on Employment by economic sector, 2010 - 2019 Change in labor outcomes HH Farm HH Business Private State Foreign (1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (11) (12) (13) (14) (15) TOTAL 4.87∗∗∗ 4.12∗∗∗ 4.05∗∗∗ 3.78∗∗∗ 3.98∗∗∗ EXPOSURE (0.47) (0.48) (0.49) (0.46) (0.51) DIRECT 7.69∗∗∗ 6.68∗∗∗ 6.61∗∗∗ 6.31∗∗∗ 6.56∗∗∗ EXPOSURE (1.13) (1.15) (1.14) (1.09) (1.18) 42 INDIRECT 7.54∗∗∗ 5.98∗∗∗ 5.77∗∗∗ 5.10∗∗∗ 5.59∗∗∗ EXPOSURE (0.86) (0.76) (0.78) (0.77) (0.93) Observations 504 504 504 504 504 504 498 498 498 504 504 504 357 357 357 Adjusted R2 0.1 0.1 0.1 0.1 0.1 0.1 0.1 0.1 0.1 0.1 0.1 0.1 0.1 0.1 0.1 *** p<.001, ** p<.05, * p<.01. Standard errors clustered at Province level. Total Exposure = Direct Exposure + Indirect Exposure. Additional controls include Age, Gender, Education level, Urban-Rurual dummy and Hours of work. Wage and wage premium measured in US dollars per year, Export Exposure measured in thousand US dollar per worker. Unemployed, Inactive, Informal and Female Employed measured in percentage point. Table 14: Impact of trade exposure on Wages by Income level, 2010 - 2019 Change in labor outcomes Income quantile 1 Income quantile 2 Income quantile 3 Income quantile 4 (1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (11) (12) TOTAL 14.54∗∗ -12.00 -4.18 -2.78 EXPOSURE (5.31) (6.82) (3.85) (3.62) DIRECT 20.40 -12.88 -1.38 0.61 EXPOSURE (11.09) (7.32) (6.65) (5.54) 43 INDIRECT 30.12∗∗∗ -33.73 -18.49∗ -15.53 EXPOSURE (7.28) (21.41) (8.70) (8.86) Observations 504 504 504 452 452 452 478 478 478 504 504 504 Adjusted R2 0.3 0.3 0.3 0.0 0.0 0.0 0.2 0.2 0.2 0.1 0.1 0.1 *** p<.001, ** p<.05, * p<.01. Standard errors clustered at Province level. Total Exposure = Direct Exposure + Indirect Exposure. Additional controls include Age, Gender, Education level, Urban-Rurual dummy, Economic Sector and Hours of work. Wage and wage premium measured in US dollars per year, Export Exposure measured in thousand US dollar per worker. Unemployed, Inactive, Informal and Female Employed measured in percentage point. Table 15: Impact of trade exposure on Employment by Income level, 2010 - 2019 Change in labor outcomes Income quantile 1 Income quantile 2 Income quantile 3 Income quantile 4 (1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (11) (12) TOTAL 1.77∗∗∗ 0.32∗∗∗ 0.23∗∗∗ 0.14∗∗ EXPOSURE (0.25) (0.07) (0.05) (0.05) DIRECT 2.91∗∗∗ 0.44∗∗∗ 0.29∗∗ 0.15∗ EXPOSURE (0.58) (0.12) (0.09) (0.07) 44 INDIRECT 2.73∗∗∗ 0.69∗∗∗ 0.55∗∗∗ 0.39∗∗∗ EXPOSURE (0.43) (0.16) (0.11) (0.10) Observations 504 504 504 452 452 452 478 478 478 504 504 504 Adjusted R2 0.3 0.3 0.3 0.1 0.1 0.1 0.1 0.1 0.1 0.0 0.0 0.1 *** p<.001, ** p<.05, * p<.01. Standard errors clustered at Province level. Total Exposure = Direct Exposure + Indirect Exposure. Additional controls include Age, Gender, Education level, Urban-Rurual dummy, Economic Sector and Hours of work. Wage and wage premium measured in US dollars per year, Export Exposure measured in thousand US dollar per worker. Unemployed, Inactive, Informal and Female Employed measured in percentage point. Table 16: Impact of trade exposure on Labor market in Tradable sectors, 2010 - 2019 Change in labor outcomes WAGE INCOME EMPLOYED INFORMAL (1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (11) (12) TOTAL 11.40∗∗ 21.26∗∗∗ 0.55∗∗∗ -0.78∗ EXPOSURE (4.09) (5.11) (0.09) (0.32) DIRECT 19.00∗∗ 37.91∗∗∗ 0.66∗∗∗ -0.18 EXPOSURE (7.05) (9.32) (0.16) (0.47) 45 INDIRECT 17.13 26.54∗ 1.37∗∗∗ -3.61∗∗∗ EXPOSURE (9.72) (11.65) (0.19) (0.49) Observations 504 504 504 504 504 504 504 504 504 504 504 504 Adjusted R2 0.3 0.3 0.3 0.3 0.3 0.3 0.2 0.2 0.2 0.2 0.1 0.2 *** p<.001, ** p<.05, * p<.01. Standard errors clustered at Province level. Total Exposure = Direct Exposure + Indirect Exposure. Additional controls include Age, Gender, Education level, Urban-Rurual dummy, Economic Sector and Hours of work. Wage and wage premium measured in US dollars per year, Export Exposure measured in thousand US dollar per worker. Unemployed, Inactive, Informal and Female Employed measured in percentage point. Table 17: Impact of trade exposure on Labor market in Nontradable sectors, 2010 - 2019 Change in labor outcomes WAGE INCOME EMPLOYED INFORMAL (1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (11) (12) TOTAL -11.30∗ -6.83 -0.05∗∗∗ -1.17∗∗∗ EXPOSURE (4.27) (5.33) (0.01) (0.23) DIRECT -16.26∗∗ -7.58 -0.06∗∗ -1.16∗∗ EXPOSURE (5.69) (7.66) (0.02) (0.37) 46 INDIRECT -22.80∗ -18.61 -0.11∗∗∗ -3.46∗∗∗ EXPOSURE (11.33) (13.41) (0.03) (0.41) Observations 441 441 441 441 441 441 441 441 441 441 441 441 Adjusted R2 0.2 0.2 0.2 0.3 0.3 0.3 0.2 0.2 0.2 0.2 0.2 0.2 *** p<.001, ** p<.05, * p<.01. Standard errors clustered at Province level. Total Exposure = Direct Exposure + Indirect Exposure. Additional controls include Age, Gender, Education level, Urban-Rurual dummy, Economic Sector and Hours of work. Wage and wage premium measured in US dollars per year, Export Exposure measured in thousand US dollar per worker. Unemployed, Inactive, Informal and Female Employed measured in percentage point.